0 This Document is licensed to 2 This Document is licensed to 4 This Document is licensed to 5 Sampling for 6 Pharmaceutical Water, 7 Steam, and Process Gases 8 Disclaimer: 9 The ISPE Good Practice Guide on Sampling for Pharmaceutical Water, Steam, and Process Gases aims to minimize 10 sampling errors, which could prevent out of specification results tainted by extrinsic factors such as environmental, 11 atmospheric, and human contact. This Guide is solely created and owned by ISPE. It is not a regulation, standard 12 or regulatory guideline document. ISPE cannot ensure and does not warrant that a system managed in accordance 13 with this Guide will be acceptable to regulatory authorities. Further, this Guide does not replace the need for hiring 14 professional engineers or technicians. 15 Limitation of Liability 16 In no event shall ISPE or any of its affiliates, or the officers, directors, employees, members, or agents of each 17 of them, or the authors, be liable for any damages of any kind, including without limitation any special, incidental, 18 indirect, or consequential damages, whether or not advised of the possibility of such damages, and on any theory of 19 liability whatsoever, arising out of or in connection with the use of this information. 20 © Copyright ISPE 2016. All rights reserved. 21 All rights reserved. No part of this document may be reproduced or copied in any form or by any means – graphic, 22 electronic, or mechanical, including photocopying, taping, or information storage and retrieval systems – without 23 written permission of ISPE. 24 All trademarks used are acknowledged. 25 ISBN 978-1-936379-89-7 26 GOOD PRACTICE GUIDE: 28 This Document is licensed to 29 Page 2 30 ISPE Good Practice Guide: 32 Sampling for Pharmaceutical Water, Steam, and Process Gases 34 Preface 35 Product Quality is of paramount importance in all industries, but it is of particular importance in the Life Sciences 36 industry and sampling is a critical part of this process. When any step of the sampling process isn’t done properly, 37 errors may be introduced and misrepresentative data can cost companies thousands if not millions of dollars by 38 creating costly and unnecessary investigations and corrective actions. 39 Sampling can help us understand if a system is delivering the critical utility to use points in the facility or if the 40 characteristics are changing between the generation and use of the utility. 41 Sampling is an involved and complicated process and very little regulatory guidance exists on this topic. The ISPE 42 Good Practice Guide: Sampling for Pharmaceutical Water, Steam, and Process Gases provides expert guidance 43 on all aspects of sampling from valve design, the number, location, and placement of sample valves, sampling 44 technique, frequency, and sample storage including delivery to the testing laboratory. 45 This Guide is an indispensable tool to all users of water, steam, or process gases and impacts facilities, production, 46 and quality control personnel within a facility. As such, this Guide applies to manufacturers of pharmaceuticals, 47 medical devices, biologics, cosmetics, and related products as well as equipment manufacturers, vendors, and other 48 industries outside of the pharmaceutical world. 50 This Document is licensed to 51 ISPE Good Practice Guide: 52 Page 3 53 Sampling for Pharmaceutical Water, Steam, and Process Gases 55 Acknowledgements 56 This Good Practice Guide was produced by a dedicated Task Team of Subject Matter Experts (SMEs) led by Brian 57 Hagopian, CPIP (Clear Water Consulting, Inc.). The work was supported by the ISPE Critical Utilities Community of 58 Practice (COP). 59 The authors and contributors to this Good Practice Guide are listed below, but the following people deserve special 60 recognition for their extensive involvement in developing and vetting the content for this guide: Joseph Manfredi and 61 Teri C. Soli, PhD for the Water Chapter, Brian Pochini, CPIP for the Steam Chapter, and Ruby Ochoa for the Process 62 Gas Chapter. 64 Chair 65 Brian Hagopian, CPIP 66 Clear Water Consulting, Inc. 67 USA 69 Chapter 1: Introduction 70 Brian Hagopian, CPIP (Chapter Lead) 71 Clear Water Consulting, Inc. 72 USA 74 Chapter 2: Pharmaceutical Water 75 Michael Baumstein 76 Pfizer, Inc. 77 USA 78 Rod Freeman 79 Beckman Coulter, Inc. 80 USA 81 Brian Hagopian, CPIP 82 Clear Water Consulting, Inc. 83 USA 84 Jeppe Kjems 85 CU Engineering 86 Denmark 87 Joseph Manfredi (Chapter Lead) 88 GMP Systems, Inc. 89 USA 90 Aravind Palinvelu 91 Roche 92 Singapore 93 Teri C. Soli, PhD 94 Soli Pharma Solutions, Inc. 95 USA 96 Michael Tomaselli 97 Filters, Water, and Instrumentation, Inc. 98 USA 100 Chapter 3: Pharmaceutical Steam 101 Michael Baumstein 102 Pfizer, Inc. 103 USA 104 Andre Gill, PE 105 Andre Gill Engineering 106 USA 107 Brian Hagopian, CPIP 108 Clear Water Consulting, Inc. 109 USA 110 Joseph Manfredi 111 GMP Systems, Inc. 112 USA 113 Brian Pochini, CPIP (Chapter Lead) 114 Sanofi 115 USA 116 Teri C. Soli, PhD 117 Soli Pharma Solutions, Inc. 118 USA 119 Philip Sumner, PE 120 Pfizer, Inc. 121 USA 122 Nancy Tomoney 123 West-Ward Pharmaceuticals 124 USA 125 Anders Widov 126 Wiphe AB 127 Sweden 129 Chapter 4: Process Gases 130 Michael Baumstein 131 Pfizer, Inc. 132 USA 133 Brian Hagopian, CPIP 134 Clear Water Consulting, Inc. 135 USA 136 Ruby Ochoa 137 Trace Analytics, LLC 138 USA 139 Aravind Palinvelu 140 Roche 141 Singapore 142 Michael Vestermark 143 Novo Nordisk Biopharm 144 Denmark 145 Peter Vishton (Chapter Lead) 146 Independent Consultant 147 USA 149 Glossary 150 Michelle Gonzalez, PE 151 Amgen, Inc. (retired) 152 USA 154 This Document is licensed to 155 Page 4 156 ISPE Good Practice Guide: 158 Sampling for Pharmaceutical Water, Steam, and Process Gases 159 600 N. Westshore Blvd., Suite 900, Tampa, Florida 33609 USA 160 Tel: +1-813-960-2105, Fax: +1-813-264-2816 161 www.ISPE.org 163 Special Thanks 164 The Sampling Guide Task Team would like to express thanks to ISPE for technical writing and editing support from 165 Gail Evans (ISPE Guidance Documents Technical Writer and Editor) for her thorough review and countless hours of 166 editing assistance in the final preparation of this Guide. 167 The Task Team would also like to thank the dozens of technical reviewers who commented on the initial draft of this 168 Guide and provided the valuable feedback necessary to ensure that this Guide became a clear and concise reference 169 document. 170 Company affiliations are as of the final draft of the Guide. 171 Cover photo: Shutterstock. 173 This Document is licensed to 174 ISPE Good Practice Guide: 175 Page 5 176 Sampling for Pharmaceutical Water, Steam, and Process Gases 178 Table of Contents 179 1 180 Introduction.......................................................................................................................7 182 1.1 183 Background..................................................................................................................................................7 185 1.2 186 Overview.....................................................................................................................................................10 188 1.3 189 Scope and Purpose....................................................................................................................................10 191 1.4 192 Benefits....................................................................................................................................................... 11 194 1.5 195 Objectives................................................................................................................................................... 11 197 1.6 198 Key Concepts/Terms.................................................................................................................................. 11 199 2 Pharmaceutical Water.................................................................................................... 13 201 2.1 202 Introduction.................................................................................................................................................13 204 2.2 205 Determining Sampling Locations................................................................................................................19 207 2.3 208 Developing Sampling Plans........................................................................................................................24 210 2.4 211 Sample Valve Design.................................................................................................................................38 213 2.5 214 Sampling Techniques.................................................................................................................................40 216 2.6 217 Handling of Samples..................................................................................................................................44 219 2.7 220 Parametric (Real Time) Release................................................................................................................48 221 3 Pharmaceutical Steam................................................................................................... 51 223 3.1 224 Introduction to Pharmaceutical Steam........................................................................................................51 226 3.2 227 Generation and Distribution of Pharmaceutical Steam...............................................................................51 229 3.3 230 Sampling Locations....................................................................................................................................56 232 3.4 233 Sampling Plans (Frequency and Duration).................................................................................................63 235 3.5 236 Sample Valve Design.................................................................................................................................68 238 3.6 239 Pure Steam Sampling Techniques.............................................................................................................68 241 3.7 242 Sample Handling........................................................................................................................................76 244 3.8 245 Other Factors Influencing Sampling Strategies..........................................................................................76 246 4 Process Gases.................................................................................................................. 81 248 4.1 249 Introduction.................................................................................................................................................81 251 4.2 252 Sampling Locations....................................................................................................................................83 254 4.3 255 Sampling Plan (Tests Performed, Frequency and Duration)......................................................................88 257 4.4 258 Sample Valve Design.................................................................................................................................91 260 4.5 261 Gases Sampling Techniques for Compressed Air and Process.................................................................91 263 4.6 264 Sample Handling........................................................................................................................................94 266 4.7 267 System Monitoring......................................................................................................................................94 268 5 Appendix 1 – Specification Summary for Various Non-Pharmacopeial 270 Water Grades.................................................................................................................. 95 271 6 Appendix 2 – Examples of Water System Sampling Point Locations.................... 97 272 7 Appendix 3 – Factors Influencing Pure Steam Generator Performance............... 101 274 7.1 275 Source Water............................................................................................................................................101 277 7.2 278 Steam Generator Mist Elimination Capability...........................................................................................102 280 7.3 281 Non-condensable Gas Removal Capability..............................................................................................103 283 7.4 284 Blow Down Adjustment.............................................................................................................................103 286 7.5 287 Potable Water Chloramine Use................................................................................................................103 289 7.6 290 Anti-scaling Steam Additives....................................................................................................................103 292 7.7 293 Monitoring Locations and Frequency.......................................................................................................104 295 This Document is licensed to 296 Page 6 297 ISPE Good Practice Guide: 299 Sampling for Pharmaceutical Water, Steam, and Process Gases 300 8 Appendix 4 – References.............................................................................................105 301 9 Appendix 5 – Glossary.................................................................................................107 303 9.1 304 Acronyms and Abbreviations....................................................................................................................107 306 9.2 307 Definitions.................................................................................................................................................109 309 This Document is licensed to 310 ISPE Good Practice Guide: 311 Page 7 312 Sampling for Pharmaceutical Water, Steam, and Process Gases 313 1 314 Introduction 315 1.1 316 Background 317 Effective sampling is of paramount importance to the success of any pharmaceutical critical utility system. 318 Unfortunately, there is very little regulatory guidance with regards to sampling. Extracting a representative sample 319 from a utility system is an involved and complicated process and error may be introduced due to: 320 • 321 Environmental conditions 322 • 323 Sample valve design 324 • 325 Sample point location 326 • 327 Technique 328 • 329 Container type 330 • 331 Sample handling 332 • 333 Sample storage and transport 334 • 335 Other factors 336 The impact of these factors is widespread and occurs throughout industry. Errors due to sampling are so common 337 that the 1993 US FDA “Guide to the Inspection of High Purity Water Systems” [1] specifically addresses sampling 338 errors in recommendations to their field inspectors by stating that, “it is expected that [Water For Injection systems] be 339 essentially sterile. Since sampling frequently is performed in non-sterile areas and is not truly aseptic, occasional low 340 level [bacteria] counts due to sampling errors may occur.” 341 Improper sampling may be responsible for generating Out Of Specification (OOS) data and prompting investigations 342 when a system is actually producing and delivering acceptable quality. Conversely, improper sampling may also 343 create a more dangerous scenario where the quality is actually OOS but the data incorrectly indicates that the 344 quality is acceptable. In either case, there could be negative impact on company image, cost, productivity, ethics, 345 and regulatory liability, resulting from improper sampling. Identifying the elements of proper sampling and eliminating 346 the underlying reasons for these negative consequences formed the primary driving force behind this ISPE Good 347 Practice Guide. 348 1.1.1 349 What is a sample? 350 Simply stated, the intent of sampling is to take a small but representative portion of a much larger stream, where the 351 sample collected accurately represents the content of the larger stream. The sample collected should not be altered 352 or changed in any way because the sampling process, but this is an almost impossibly difficult proposition as all 353 sampled utilities come into contact with air, containers, etc. during the sampling process. 354 1.1.2 355 Why do we sample? 356 It is most desirable to monitor quality parameters using in-line, on-line, or at-line monitors, which minimize and, in 357 some cases, eliminate, sample handling issues completely. However, when the quality attribute of concern cannot be 358 analyzed with an instrument or where an instrument is not present, a sample has to be removed from a stream and 359 evaluated off-line. In most cases, these attributes are microbiological or endotoxin in nature, but sampling for Total 360 Organic Carbon (TOC) and conductivity are commonly done as well. 362 This Document is licensed to 363 Page 8 364 ISPE Good Practice Guide: 366 Sampling for Pharmaceutical Water, Steam, and Process Gases 367 The goal of sampling really depends on why the sample is being collected. Sampling generally occurs for one or 368 more of the following reasons: 369 • 370 Regulatory requirements 371 • 372 Quality control purposes 373 • 374 Process control purposes 375 • 376 Technical, investigational or diagnostic purposes 377 • 378 Test method validation purposes 379 Each of these purposes is described below. 380 Regulatory monographs establish compendial requirements for Critical Quality Attributes (CQAs) in a critical utility 381 system. Industry must demonstrate that the utility’s CQAs are being met through instrumentation or sampling. There 382 are usually established sampling frequencies, and some of these frequencies may be established by authorities 383 and regulating agencies. In this sampling guide, we will refer to sampling for these purposes as sampling to meet 384 regulatory requirements. 385 In some cases, the utility being sampled is an ingredient or comes into direct contact with a pharmaceutical product. 386 In this case, sampling is done to ensure that the quality of the utility meets established quality parameters. In these 387 instances, samples must be taken in the exact way that the utility is removed from the delivery system during 388 manufacturing. Stated a bit differently in the case of taking a water sample, samples shall be collected from the end 389 of a hose or pipe in a manner that replicates exactly how the utility is used during manufacturing. Samples need to be 390 tested for all quality attributes that are specified for the grade of the utility. There are usually established frequencies 391 for sampling, and frequencies may be established by regulatory authorities as well as by internal Standard Operating 392 Procedures (SOPs). In this Good Practice Guide, we will refer to sampling for these purposes as sampling for 393 quality control purposes. 394 In other cases, sampling may be done regularly, but at a different frequency than samples taken for quality control 395 purposes. Samples of this type would be taken to confirm that a particular process is operating under a state of 396 control. Here, the goal of sampling is completely different in that the purpose is to determine if a specific contaminant 397 is being reduced to the levels established for that particular treatment process. The acceptance level for the 398 contaminant being removed may be different from the final product quality attribute level because the attribute is 399 being tested midway through a multi-step purification process. In this case, it would be a waste of resources to test 400 the sample for anything other than the contaminant that the process has been designed to remove. There is no 401 requirement to test for anything other than the established quality attributes for the treatment step. Sampling for all 402 quality attributes of the finished material and this location would provide data with little meaning. When sampling for 403 these purposes, it is critically important that the sample removed from a system be as representative as possible of 404 the material flowing through the system. Great care should be taken to ensure that samples are not contaminated. 405 Careful consideration should be taken in flushing or preparing these lines, for example by spraying with alcohol prior 406 to or immediately following sample collection. In this guide, we will refer to sampling for these purposes as sampling 407 for process control purposes. 408 If other instances, sampling may be done only as needed, e.g., if an Alert Level is reached for a critical quality 409 attribute, investigation into the root cause of the excursion would suggest sampling from locations that are not part of 410 the regular testing program. The frequency of sampling for these purposes will vary from the frequency of sampling 411 for all other purposes and would typically occur on an as needed basis as established in SOPs. In this sampling 412 guide, we will refer to sampling for these purposes as sampling for technical, investigational, or diagnostic 413 purposes. 415 This Document is licensed to 416 ISPE Good Practice Guide: 417 Page 9 418 Sampling for Pharmaceutical Water, Steam, and Process Gases 419 While there is no established regulatory requirement, a robust sampling program should periodically go through a 420 method validation process where duplicate samples are taken from a small number of sample points to document the 421 consistency of the data and therefore the accuracy of the method. Sampling for these reasons represents sampling 422 for method validation purposes. 423 Note: there is no purpose classified as sampling for informational purposes only. While industry has used this 424 description in the past, and many in industry continue to use this designation, every sample point should have a 425 specific purpose and the authors of this guide are recommending abandoning the “for informational purposes only” 426 designation as being too vague and non-specific. 427 When developing a sampling program and considering sampling at a particular location, the purpose the sample point 428 will serve should be considered. Sample valves should always be installed for a specific purpose and this purpose 429 should be documented. The purpose of a particular sampling point is used to establish the frequency as well as 430 the attributes monitored. For each sampling point, a risk-based approach should be utilized to establish or adjust 431 sampling frequencies based on data collected over time. As an example, an effective sampling program might sample 432 for bacteria and endotoxin after a carbon filter both before and immediately following a sanitization. 433 The results derived from an effective sampling program provide us with greater knowledge and understanding of 434 these critical utilities. Sampling brings a greater understanding of when a piece of equipment is operating properly 435 and when it may require adjustment or routine service in order to keep a system operating under a state of control. 436 Results of an effective sampling program can be used to establish normal operating levels for a purification process 437 or distribution system, or for the end product quality. In most cases, data collection and review allows for the 438 establishment and future adjustment of Alert (2σ) and Action (3σ) Levels based on actual operating conditions. 439 Sampling can help us understand if a system is delivering the critical utility to use points in the facility or if the 440 characteristics are changing between the generation and use of the utility. 441 1.1.3 442 Who is responsible for sampling? 443 Ultimately, the quality control function is responsible for sampling, but other stakeholders may become involved with 444 sampling including quality assurance, utilities, facilities, manufacturing, production, or regulatory. 445 1.1.4 446 How often do we need to sample? 447 Sampling frequency depends on the system and the grade of the critical utility. In the case of Water For Injection 448 (WFI) grade water, it is an expectation that water samples will be taken from each Point Of Use (POU) and sample 449 location in the distribution system at least once per week during validation. Industry has continued to utilize this 450 same sampling frequency beyond the completion of the Performance Qualification, although it is not mandated. By 451 contrast, purified water has no such preconceived sampling requirements, so the end user is not mandated to follow 452 any specific schedule during or following validation, although many companies also having WFI systems default 453 to the same frequency used on WFI systems. Pharmaceutical steam, compressed air, and process gases have 454 different frequencies during validation and ongoing operation. When no sampling frequency is specified, a risk-based 455 approach may be utilized in conjunction with the purpose of sampling to establish initial sampling frequencies. As 456 data is generated and an operating baseline is established, the collected data should be reviewed to determine if the 457 operating limits, Alert and Action Levels, and/or the sampling frequency may be adjusted (increased or decreased) 458 based on data generated on system operation and as established within operational SOPs. 459 1.1.5 460 Where do we need to sample? 461 Sampling locations are determined by the purpose of sampling and the information expected from the sampling 462 location. Careful consideration and sound engineering judgment should be utilized to distinguish between sampling 463 locations for Regulatory Compliance, Quality Control, Process Control, or Investigational and Diagnostic purposes. 465 This Document is licensed to 466 Page 10 467 ISPE Good Practice Guide: 469 Sampling for Pharmaceutical Water, Steam, and Process Gases 470 1.1.6 471 Are all quality attributes uniformly distributed in a system? 472 The answer is a resounding no, and this is one of the most important paradigm shifts that the reader should 473 understand when defining and evaluating a sampling program. In pharmaceutical water systems, if a quality attribute 474 is a chemical constituent (e.g., nitrates, conductivity, dissolved metals, TOC) then the constituent is considered 475 uniformly or homogeneously distributed (Second Law of Thermodynamics). However, if the quality attribute is 476 biological in nature (e.g., bacteria, endotoxin), then the quality attribute will not be uniformly distributed. Non-uniform 477 or heterogeneously distributed attributes are much more challenging to sample accurately as these attributes may 478 vary from point to point in a system. A significant portion of this Guide is devoted to providing guidance on the most 479 effective methods for accurate sample collection involving these heterogeneously distributed quality attributes. 480 In pharmaceutical steam systems, compressed air, and process gas systems, attributes may or may not be uniformly 481 distributed and sampling programs must take this into account. 482 1.2 483 Overview 484 This Guide affects users of water, steam, compressed air, or process gases and impacts facilities, production, and 485 quality control personnel within a facility. This Guide applies to manufacturers of pharmaceuticals, medical devices, 486 biologics, cosmetics and related products, as well as equipment manufacturers, vendors, and other industries outside 487 of the pharmaceutical arena. 488 1.3 489 Scope and Purpose 490 This Guide has been assembled to provide assistance to designer engineers, process engineers, facilities engineers, 491 quality assurance staff, quality control staff, production staff, and any other group, team, or individual that may be 492 impacted by the results of a sampling program. 493 This Guide covers the critical utilities of pharmaceutical water, pharmaceutical steam, and pharmaceutical process 494 gases in the chapters that follow. 495 While the focus of this Guide is geared to the pharmaceutical, biotechnology, and other life science industries, 496 the ideas, concepts, and methods detailed in this Guide may prove valuable and applicable to other industries 497 incorporating the utilities covered in this Guide. In the life science industries, it is important to remember that the 498 product manufacturer is ultimately responsible for all quality attributes that impact raw material and finished product 499 quality. 500 This Guide covers all aspects of sampling from valve design, the number, location, and placement of sample valves, 501 sampling technique, frequency, and sample storage including delivery to the testing laboratory. 502 The scope of this Guide stops when the sample is delivered to the laboratory for analysis. This Guide does not 503 include the actual testing performed after the sample is turned over to the laboratory. Further, this Guide does not 504 cover the interpretation or analysis of data generated by the sampling programs referenced in this Guide. While this 505 data is usually collected and compiled to establish alert (e.g., standard deviations (2σ) from the mean) and action 506 (e.g., three standard deviations (3σ) from the mean) levels for the parameter being measured, a thorough discussion 507 of data interpretation is outside the scope of this Guide. 508 This Guide does not cover Heating, Ventilation, and Air Conditioning (HVAC) systems for cleanrooms, environmental 509 monitoring, Active Pharmaceutical Ingredients (APIs), or finished products. 511 This Document is licensed to 512 ISPE Good Practice Guide: 513 Page 11 514 Sampling for Pharmaceutical Water, Steam, and Process Gases 515 1.4 516 Benefits 517 This Guide aims to: 518 • 519 Provide industry with standards and best practices for sampling in a comprehensive single reference. 520 • 521 Help to minimize sampling errors and their contribution to Out of Specification (OOS) results, which can cause 522 costly work and production stoppages, investigations, and disruptions in manufacturing, Quality Control (QC)/ 523 Quality Assurance (QA), and engineering organizations. 524 This Guide is expected to be of benefit to laboratory, QC/QA, engineering, facilities, and operations personnel. 525 1.5 526 Objectives 527 This Guide aims to establish good practices for sampling to minimize sample contamination from human contact, 528 error, atmospheric, or environmental conditions which could alter laboratory results and provide inaccurate data. 529 1.6 530 Key Concepts/Terms 531 Alert Level 532 A level or range that, when exceeded, indicates that a process may have drifted from its normal operating range. 533 May be established at two standard deviations (2σ) from the mean. Constitutes a warning and does not necessarily 534 require a corrective action. 535 Action Level 536 A level established that, when exceeded, indicates that a process from its normal operating range. Typically 537 established at three standard deviations (3σ) from the mean. Corrective actions should be taken to bring the process 538 back into its normal operating range. 539 Contaminant 540 Any foreign component present in another substance; used synonymously with impurity in this Guide. 541 Impurity 542 A foreign agent that is introduced as part of a process; used synonymously with contaminant in this Guide. 544 This Document is licensed to 546 This Document is licensed to 547 ISPE Good Practice Guide: 548 Page 13 549 Sampling for Pharmaceutical Water, Steam, and Process Gases 550 2 Pharmaceutical Water 551 2.1 552 Introduction 553 Water is one of the most widely used materials in the pharmaceutical industry and is commonly used as a: 554 • 555 Product or Ingredient 556 • 557 Cleaning agent 558 • 559 Sterilizing media 560 • 561 Analytical agent 562 • 563 Formulation aid 564 When water is used for pharmaceutical purposes it needs to be tested to ensure that it meets monograph 565 requirements or other specifications for applicable CQAs. When water is extracted for sampling purposes at any point 566 in a system for any reason, proper sampling is the key to ensuring that the collected sample accurately represents the 567 quality of the water at that particular point in the system. When water is sampled in the same way as it is used during 568 manufacturing, sampling is considered the best indicator available of the quality of water going into a product. 569 2.2.1 570 Background 571 Extracting water from a system into a container for subsequent laboratory analysis is the most common sampling 572 utilized by pharmaceutical companies to determine the levels of CQAs. This testing methodology is also the most 573 frequently used method for determining the microbiological and endotoxin content of water. As the accuracy and 574 reliability of instrumentation has increased, establishing the equivalency between the conventional “sample and 575 laboratory test” method and the “continuous monitoring with an instrument” method may allow an increasing number 576 of CQAs to be verified continuously. This can substantially reduce the frequency of water sampling from a system. 577 In order to determine if an attribute is a candidate for monitoring with process instrumentation, the uniformity of 578 distribution of the attribute within a water system needs to be understood. 579 In pharmaceutical water systems, CQAs or other impurities may or may not be uniformly distributed in a distribution 580 system. The homogeneous or heterogeneous nature of distribution of CQAs in a system should be understood prior 581 to formulating a sampling plan. 582 2.1.1.1 583 Homogeneously Distributed Impurities 584 Chemical impurities, small particulates, and other soluble impurities are typically expected to be homogeneously 585 distributed throughout a distribution system because of the solubility and nature of these impurities, and because 586 of chemical equilibrium processes. Since the impurity levels at any one location in a distribution system should be 587 identical to the impurity levels at every other location within the distribution system, sampling for homogeneously 588 distributed impurities may not be required at every POU and may be performed at a limited number of locations, 589 provided that there is no likelihood of downstream sources of contamination. 590 Homogeneously distributed impurities are candidates for continuous monitoring using on-line, in-line, or at-line 591 instruments if the impurity levels at the instrumentation locations have been validated through sampling and 592 laboratory analysis to be predictive of the levels at all POUs. When validated in this manner, the resulting data may 593 be used for quality control or water release purposes. Sampling and testing of these impurities is typically done at the 594 return of a distribution loop, but may be done in other locations. Impurities falling into the homogeneously distributed 595 impurity category include: 597 This Document is licensed to 598 Page 14 599 ISPE Good Practice Guide: 601 Sampling for Pharmaceutical Water, Steam, and Process Gases 602 • 603 Ionic impurities (detected non-specifically as electrical conductivity or resistivity) 604 • 605 Organic impurities (detected non-specifically as TOC) 606 - 607 Nitrates (EP [2] requirement for Purified Water (PW) and Water For Injection (WFI)) 608 • 609 Heavy metals (EP [2] requirement for PW1) 610 Localized sources of chemical impurities may be introduced to a system downstream of a particular point within a 611 distribution system. For example, chemical contamination sources after a POU valve, such as from hoses, gaskets, 612 or a POU heat exchanger may contaminate the water as it exits the system at one particular location. Chemical 613 contamination sources within a sub loop, parallel loop, or other isolated portion of the system could contaminate 614 the water in only that portion of the system. The purity of the water in these sub loops and POU locations may not, 615 therefore, be accurately represented by on-line, in-line, or at-line instrumentation installed in the main distribution 616 system. In system sub loops, additional on-line, in-line, or at-line instrumentation may be required in order to 617 accurately measure these impurities, or manual testing by sampling may be utilized at POUs within these sub loops. 618 2.1.1.2 619 Heterogeneously Distributed Impurities 620 Bacteria and endotoxin are attributes that are not uniformly distributed throughout a water system. The reason is that 621 most bacteria present in the system are not free floating, but are part of a biofilm that is attached to system surfaces 622 within a water system. It is generally understood that the majority of bacteria present in a system (99.9 to 99.99%) 623 reside in biofilm. It is the small fraction (0.01 to 0.1%) of free floating or planktonic bacteria that contribute to the 624 bacteria and endotoxin levels detected by sampling. 625 Biofilm development occurs more rapidly in areas of low or no flow where bacteria are able to attach to surfaces 626 more easily. Bacteria and endotoxin populations may, therefore, vary widely because of variable local flow conditions. 627 Examples of low flow areas that allow bacteria to attach to a distribution piping system include: 628 • 629 Within any non-flowing side connections of any length, including dead legs 630 • 631 On the upstream or wet side of valve sealing surfaces 632 • 633 On the water stopping sealing surfaces within the valve 634 • 635 On the downstream side of the valve surfaces (which may remain wet or moist) 636 • 637 On the fitting and gasket used to connect a hose or other connector 638 • 639 Within the hose or connector to the actual location of water use 640 • 641 Within any additional external device connected to the system through which the water flows, such as: 642 - 643 Flow meters 644 - 645 Pressure regulators 646 - 647 Heat exchangers 648 - 649 Check valves 650 - 651 Hoses, fittings, adapters, etc. 652 1 According to the EP [2] if Purified Water meets the conductivity limits for WFI, then heavy metals testing is not required. 654 This Document is licensed to 655 ISPE Good Practice Guide: 656 Page 15 657 Sampling for Pharmaceutical Water, Steam, and Process Gases 658 Localized low flow conditions throughout a system and at POUs creates differing conditions and variable bacteria and 659 endotoxin populations at these locations. For this reason, sampling is required of each POU location to confirm the 660 presence or absence of bacteria. Sampling the supply and return may indicate the general condition of the distribution 661 system, but is simply not sufficient to establish the bacterial characteristics and population of the entire distribution 662 system. 663 2.1.2 664 Purpose of Sampling 665 The purpose of sampling should be identified and documented for each potential sample location within a system as 666 early as possible during the design process. This process should justify and document the requirement for sampling 667 and identify the reason for sampling at a particular location. 668 The purpose of any potential sampling location should fall into one or more of the following categories: 669 1. 670 Regulatory requirements 671 2. 672 Consensus standards 673 3. 674 Process control purposes 675 4. 676 Quality control purposes 677 5. 678 Technical, investigational, or diagnostic purposes 679 The justification for a potential sampling location should also consider: 680 • 681 The impurities of interest at the potential location 682 • 683 The type of valve and other accessories that are best suited for or are contra-indicated for the potential location 684 • 685 Should sampling should be performed at this location at an established frequency 686 • 687 Should sampling be performed on an event based frequency (i.e., reaching an alert or action level) 688 • 689 Is the potential location readily accessible to staff 690 • 691 Can the sample be taken safely by staff at the potential location 692 • 693 Any special considerations for sample handling after a sample is taken 694 The purpose of a sampling location may impact the impurities of concern as well as the sampling frequency. 695 2.1.2.1 696 Regulatory Requirements 697 All pharmacopoeias define two bulk grades of water for pharmaceutical purposes:2 698 1. 699 Purified Water (PW) 700 2. 701 Water For Injection (WFI) 702 2 Highly Purified Water (HPW) is an additional pharmaceutical grade of water as defined by the European Pharmacopoeia. 704 This Document is licensed to 705 Page 16 706 ISPE Good Practice Guide: 708 Sampling for Pharmaceutical Water, Steam, and Process Gases 709 While there are some differences between the various pharmacopoeias on the exact specification for some 710 impurities, the general consensus is that the following impurities should monitored: 711 • 712 Conductivity 713 • 714 TOC 715 • 716 Endotoxin (for WFI only) 717 Some pharmacopoeias also require testing for heavy metals and nitrates. Consult applicable monographs for exact 718 specification. 719 While bacteria testing is not a monograph requirement, all pharmaceutical companies sample and test for bacteria. 720 The pharmaceutical industry references the 1993 US FDA “Guide to the Inspection of High Purity Water Systems” [1] 721 which stipulates the following maximum bacteria action limits for pharmaceutical water systems: 722 • 723 Bacteria action limits at or below 100 cfu per ml for PW 724 • 725 Bacteria action limits at or below 10 cfu per 100 ml for WFI 726 The pharmaceutical industry has universally adopted these action limits as specifications for PW and WFI and 727 pharmaceutical water systems are sampled and tested for: 728 • 729 Conductivity 730 • 731 TOC 732 • 733 Endotoxin (WFI only) 734 • 735 Bacteria 736 2.1.2.2 Consensus Standards 737 Consensus standards may be utilized as a source for water quality specifications if the use of water within the facility or 738 a portion of a facility is consistent with those standards. Consensus standards such as the International Organization 739 for Standardization (ISO) [3], the American Society for Testing and Materials (ASTM) [4], the Association for the 740 Advancement of Medical Instrumentation (AAMI) [5], or the Clinical and Laboratory Standards Institute (CLSI) [6], have 741 application specific standards for water purity. These water quality specifications may be found in Appendix 1. 742 2.1.2.3 Process Control Purposes 743 Monitoring a water purification system for process control purposes involves monitoring the performance and success 744 of the various treatment processes for specific attributes of concern and includes following the quantitative trends 745 of those or related attributes of interest. When data trends from this monitoring indicate that some aspects of the 746 water purification system or other physical or procedural process controls are not performing as intended, then 747 investigations and targeted remedial interventions can be instituted to bring the process back into normal operational 748 control before the finished quality of the water is sufficiently affected to adversely impact its quality specifications. 749 It is considered appropriate to monitor only those attributes related to the monitoring or control of a given function 750 in the purification process. For example, it may be appropriate to monitor a carbon bed for the reduction of chlorine, 751 the reduction of TOC, and the effluent (outflow) microbial level since carbon beds are installed to impact these 752 attributes (microbial levels being an adverse effect of installing a carbon bed) and high levels of these contaminants 753 could negatively impact the operation and maintenance of downstream unit operations. However, it would not 754 be appropriate to monitor hardness or conductivity at this location, as the carbon bed is not impacted by, nor is 756 This Document is licensed to 757 ISPE Good Practice Guide: 758 Page 17 759 Sampling for Pharmaceutical Water, Steam, and Process Gases 760 it intended to impact, these attributes. The selection of the appropriate attributes to monitor should be tailored to 761 the unit operation’s function as well as potential downstream impact. The attributes selected for process control 762 monitoring may be different than attributes selected to meet compendial requirements. 763 Process control sampling points may be strategically located throughout the water purification system. They may 764 be located at the beginning and end of distribution loops or near critical POUs which are typically not used by 765 manufacturing. They may also be located at POU valves in the distribution system that is used by manufacturing. In 766 these cases, sampling may be limited only to final product CQAs such as conductivity, TOC, bacteria, and endotoxin 767 (WFI only). POU valves with built in upstream sampling ports may be utilized for process control monitoring. 768 Biological attributes such as bacteria and endotoxin that are not homogeneously distributed throughout a water 769 system have the potential to contaminate the attribute being monitored at a process control sampling location as 770 water is sampled. In these cases, efforts should be made to prevent or minimize sample contamination from these 771 attributes. 772 Process control indicating trigger values, such as Alert and Action Levels, are typically used to assist in identifying 773 when the process is deviating from its normal operating capabilities. 774 Alert Levels 775 Alert Levels should be set at levels that are near the top edge of normal data trends. Alert levels may be based on 776 analysis and trending of historic performance data and are usually set at two standard deviations (2σ) from normal 777 operating levels. 778 An occasional Alert Level excursion that is barely above the Alert Level could be considered a normal expectation, 779 as it is within the upper reaches of the normal operating trends for the system. Though such a level could be an early 780 indication of a growing control problem, it could also be simply the variability of normal operating data. Responses to 781 Alert Level excursions should include notification of appropriate quality, manufacturing, and maintenance personnel, 782 along with closer attention being paid to the routine data. Frequent excursions beyond the Alert Level should not be 783 expected and might also be considered out of trend. 784 When data begins to drift outside of the normal trends and expected performance, notification of those who may 785 potentially be affected, as well as resampling, may be appropriate to verify if the data trends are consistently higher (a 786 possible developing control problem) or not (a possible sampling inconsistency or normal data variability). 787 Where Alert Level excursions occur concurrently at multiple locations or sequentially at the same location, this may 788 be an indication of a need for investigation and remediation and may be escalated to an Action Level excursion. 789 Action Levels 790 Historic data trending should be used to establish Action Levels, which are typically set at three standard deviations 791 (3σ) from normal operating levels. Action Levels should be established at values that are: 792 • 793 Above the normal data trends and indicative of routine process controls beginning to fail 794 • 795 Far enough from the specification so that remedial action(s) can bring the process back under control before a 796 specification excursion occurs 797 Resampling and retesting should be one of the first corrective actions (along with the usual notification of appropriate 798 parties such as for Alert Level excursions) to see if the excursion conditions are an indication of a true process control 799 problem or the result of a sampling problem. 800 Where high values are confirmed by retesting, or if the original value is following Alert Level excursions, appropriate 801 remediation activities should be instituted to restore control and determine the cause(s) of the excursions, so that 802 corrective measures can preclude recurrence and remediate any potential impact to quality specifications. 804 This Document is licensed to 805 Page 18 806 ISPE Good Practice Guide: 808 Sampling for Pharmaceutical Water, Steam, and Process Gases 809 2.1.2.4 Quality Control Purposes 810 When monitoring for quality control of the water, the intent is to: 811 1. 812 Exactly duplicate the quality of water that is being used for manufacturing by collecting samples at routine 813 intervals 814 2. 815 Collect water samples at the beginning of a period when the water is being drawn from the system for 816 manufacturing use 817 Exact duplication of the preparatory procedures utilized by manufacturing should be used during sampling, and 818 should include the use of the same: 819 • 820 Outlet sanitization practice, if applicable 821 • 822 Hose and gasket 823 • 824 Outlet flush procedure (if any) 825 • 826 Attached devices such as heat exchangers, reducers, adapters, check valves, and flow meters 827 The purpose of including all of the potential sources of contamination in the flow path for sampling is to attempt 828 to duplicate the water quality delivered to the manufacturing process, including any heterogeneously distributed 829 contaminants coming from the associated biofilms on or within these components. 830 Hard piped or otherwise permanent connections between the water system and manufacturing equipment, as well as 831 automatic valves, create challenges for obtaining samples that duplicate the way that water is used by manufacturing. 832 One engineering solution may be to utilize a three way divert valve that may be installed on the manufacturing 833 equipment. This could facilitate sampling by diverting water to obtain a sample. Alternatively, there could be a sample 834 port on, or near, the equipment isolation valve that could allow a sample to be collected when the water is first being 835 discharged into the processing tank or other POU. Another approach (if feasible) may be to simply disconnect the 836 interconnecting piping and direct the water flow toward a sample container, instead of allowing it to flow into the 837 manufacturing equipment. 838 While POU valves with a built in upstream sample port may be used when sampling for process control purposes, 839 they are typically not used when sampling for quality control purposes. The only exceptions that may allow use of a 840 built-in sampling port on the POU valve would be: 841 • 842 The conduit is routinely sanitized with a validated procedure prior to water use for manufacturing with sanitization 843 occurring by hot water, steam, or chemicals so that the conduit provides no added contamination to the loop 844 water 845 • 846 There is no feasible engineering solution that allows sampling exactly as used during manufacturing 847 • 848 The relevant attributes at that particular POU were not microbial, i.e., only the homogeneously distributed 849 chemical impurities, e.g., for a washer whose cycle ends with a hot (> 60°C air) drying step that would kill any 850 microbial contaminants in the transferred water 851 The test data from quality control sampling should be used for comparison with the quality control requirements or 852 specifications for the water. 854 This Document is licensed to 855 ISPE Good Practice Guide: 856 Page 19 857 Sampling for Pharmaceutical Water, Steam, and Process Gases 858 Alert, Action, and Out Of Specification (OOS) Levels 859 Sampling could be performed from the same sample location for process control and quality control purposes. 860 Different Alert and Action Levels could be established; however, process control levels may typically be established at 861 tighter values than those for quality control. 862 For an excursion above a specification limit, an investigation should be performed to determine, correct, and prevent 863 the cause of the excursion from occurring in the future. In this case, the particular investigation would be an OOS 864 investigation and the impact of this water on the product should be assessed. 865 2.1.2.5 Sampling for Technical, Investigative, or Diagnostic Reasons 866 Sampling performed for technical, investigative, or diagnostic reasons may include: 867 • 868 Gathering data to better understand (or pinpoint) the root cause of a process deviation or specification excursion 869 • 870 To determine the extent of contamination within a system 871 • 872 To spot check the long-term efficacy of a particular remedial action (e.g., from a CAPA response) by checking 873 specific attribute levels at a specified location 874 Additional sampling points, installed for technical, investigative, or troubleshooting purposes may not be routinely 875 sampled. For instance, if a Reverse Osmosis (RO) membrane was prematurely scaling with hardness, it would 876 be appropriate to sample upstream and downstream of a water softener to determine if hardness reduction levels 877 were being achieved throughout the softener use cycle or if the water hardness increased since the equipment was 878 originally set up. 879 Data from technical, investigative, or diagnostic sampling can also be used to establish a baseline for non-routinely 880 measured water attributes (such as the chemical attributes of the potable feed water during initial commissioning 881 or validation) which may be valuable in future investigations. Samplings of this nature are typically done only 882 occasionally and only when needed for long term profiling, or as part of an investigation. 883 2.2 884 Determining Sampling Locations 885 Locating sample points to ensure samples are taken from suitable process locations before and after unit operations, 886 at points of use and elsewhere, should assure a successful water system monitoring program. Since many 887 technologies and unit operations may be employed to produce pharmaceutical grades of water, sampling locations 888 are best identified during the conceptual and detailed design phases. A Risk Analysis (RA) may be performed to help 889 determine sample locations. The RA can provide a rationale for the sampling points in a water system and can help to 890 define the intended purpose of each sampling location, based on established CQAs and Critical Process Parameters 891 (CPPs). 892 These sample locations should provide samples that are representative of water quality at the sample location. This 893 cannot be accomplished without serious consideration to the physical location of these valves and the methods that 894 should be utilized to obtain the sample. Sample valve locations should be addressed not only on a P&ID but should 895 also be addressed during the equipment and piping installation. Components, hardware, or other items that interfere 896 with correct sampling techniques should be avoided through careful planning and appropriate field execution. 897 Considerations should include the disposition of waste from flushing, location of sample ports relative to other 898 equipment, local environmental issues, and any other appropriate concerns. Design in consultation with QA, QC, and 899 Microbiology Subject Matter Experts (SMEs) is desirable. 901 This Document is licensed to 902 Page 20 903 ISPE Good Practice Guide: 905 Sampling for Pharmaceutical Water, Steam, and Process Gases 906 For example, sample valves located in return loop piping that are inaccessible to samplers, sample valves blocked 907 by other equipment or with otherwise limited access, and/or sample valves that cannot be utilized in a reasonable 908 fashion, do not accomplish the goal of ensuring that the samples are representative. Improper location may put a 909 sampler’s safety at risk or may yield data with a high degree of variability, and may result in an increase in alert limit, 910 action limit, and OOS findings. 911 Sampling locations should include all appropriate sections of the system including: 912 • 913 Source water 914 • 915 Pretreatment unit operations 916 • 917 Primary and final treatment 918 • 919 Storage 920 • 921 Distribution systems (main loops and sub loops) 922 • 923 POU and any special POU treatments 924 Sampling locations should be tabulated. Table 2.1 provides an example of such a table. 925 Table 2.1: Examples of some Sample Points in a Purified Water System 926 Drawing No. 927 Tag 928 Function/ 929 Classification 930 Location 931 (Room) 932 Frequency 933 Remarks 934 PW-1 935 PW-01 936 Carbon Bed 937 Inlet 938 Mechanical 939 Monthly 940 Chlorine, TOC, bacteria for 941 process control 942 PW-1 943 PW-02 944 Carbon Bed 945 Outlet 946 Mechanical 947 Monthly 948 Chlorine, TOC, bacteria for 949 process control 950 PW-2 951 PW-03 952 Deionizer 953 Outlet 954 Mechanical 955 Quarterly 956 Sodium, silica for diagnostic 957 purposes 958 PW-2 959 PW-04 960 Distribution 961 Loop Supply 962 Mechanical 963 Daily/Weekly 964 TOC, conductivity, bacteria, for 965 quality control purposes 966 PW-3 967 PW-05 968 POU Valve #1 969 Production 970 Room #1 971 Weekly 972 TOC, conductivity, bacteria, for 973 regulatory requirements and 974 quality control 975 2.2.1 976 Source Water 977 Source water feeding any compendial pharmaceutical water system should meet local drinking water standards such 978 as those specified by the WHO, the US Environmental Protection Agency (EPA), drinking water regulations of the EU 979 or Japan, or the drinking water regulations from other competent authorities. Source water that does not meet potable 980 or drinking water standards should be brought into compliance with recognized drinking water regulations prior to 981 being used to feed a pharmaceutical water system. 982 Chlorination (or chloramination) of municipal water has become common in many areas of the world, for 983 microbiological control. If residual chlorine levels are inadequate for microbial control, rechlorination may be required 984 and should be defined in a SOP and should be monitored on a regular basis through an appropriately located sample 985 port. If rechlorination is required, sampling immediately afterward is also indicated. 987 This Document is licensed to 988 ISPE Good Practice Guide: 989 Page 21 990 Sampling for Pharmaceutical Water, Steam, and Process Gases 991 Complete chemical analysis of source water that confirms it meets the appropriate potable or drinking water 992 requirements may be performed on a periodic basis to trend incoming contaminant levels and to make system 993 adjustments if required. If a source water storage tank is employed, additional testing, after storage, is advisable 994 before water enters the water purification system. 995 2.2.2 996 Pretreatment Unit Operations 997 Sampling locations should be included before and after each unit operation for diagnostic purposes so that monitoring 998 can be performed as required. Sampling before and after individual unit operations helps in establishing a unit 999 operation’s performance. 1000 The following unit operations are examples of monitoring locations that may be considered in a purification system for 1001 assessing the performance of the purification step: 1002 • 1003 Before and after media filtration (e.g.: multi-media, cartridge, etc.) 1004 • 1005 Before and after softeners 1006 • 1007 Before and after Activated Carbon Filters3 1008 • 1009 Before and after Ultraviolet units 1010 • 1011 Before and after chlorination or dechlorination units 1012 • 1013 Before and after Reverse Osmosis (RO), Continuous Electro Deionization (CEDI), or ion exchange units 1014 • 1015 Before and after any other unit operations within the pretreatment system boundary 1016 2.2.3 1017 Primary and Final Treatment 1018 Primary and final treatment for compendial or non-compendial water may include several different treatment 1019 technologies. See the ISPE Baseline® Guide: Water and Steam Systems (Second Edition) [7]. 1020 Examples of primary and final treatment unit operations equipment include: 1021 • 1022 Reverse osmosis 1023 • 1024 Continuous electro deionization 1025 • 1026 Ion exchange deionization 1027 • 1028 Microfiltration 1029 • 1030 Ultrafiltration 1031 • 1032 Nanofiltration 1033 • 1034 Distillation 1035 • 1036 Ultraviolet treatment 1037 • 1038 Heat exchangers 1039 3 Microbial monitoring may additionally be critical in this location because activated carbon beds tend to promote microbial growth. 1040 1041 This Document is licensed to 1042 Page 22 1043 ISPE Good Practice Guide: 1044 1045 Sampling for Pharmaceutical Water, Steam, and Process Gases 1046 Sampling locations should be provided before and after each treatment unit operation as appropriate. 1047 Attributes to be evaluated are based on equipment performance expectations and may include: 1048 • 1049 Conductivity 1050 • 1051 TOC 1052 • 1053 Chlorine 1054 • 1055 Ammonia 1056 • 1057 Hardness 1058 • 1059 Silica 1060 • 1061 Silt Density Index (SDI) 1062 • 1063 Microbial 1064 • 1065 Endotoxin 1066 Attributes may be selected for determining the performance of a unit operation and should have Alert and Action 1067 Levels that mirror the performance expectations of the respective unit operation. 1068 Final treatment unit output water should be monitored using in-line, on-line or at-line analyzers where possible. 1069 Conductivity and TOC analyzers are usually used for these purposes. If on-line instruments are available, then off-line 1070 sampling may not be required for process control purposes. Attributes such as pH, chlorine, and ammonia may also 1071 be monitored if they are defined as CPPs. 1072 Where multiple final treatment units are present in a facility, there should be a sample point before and after each. 1073 Each point should be monitored routinely for process control purposes. One sample point located between unit 1074 operations can serve to sample the outlet of one process and the inlet of another. 1075 2.2.4 1076 Storage Tanks 1077 Sampling locations directly on a storage tank may provide valuable information if any microbial buildup (biofilm) is 1078 occurring in the tank and also allows testing of the stored water quality since water velocity through storage tanks 1079 provides minimal turbulence when compared to distribution lines. A sample valve installed for direct sampling from 1080 the storage tank may be required for validation or for diagnostic purposes and may or may not be used for routine 1081 sampling. 1082 Sampling valves located at storage tanks should be easily accessible for sampling and should minimize the risk of 1083 contamination of the sample and also the storage tank itself during sampling. 1084 Tank water quality can also be assessed indirectly through monitoring of the primary generation supply (incoming 1085 water to storage), the loop return (water returning to storage), and the tank outlet which feeds the pumps and 1086 subsequently the use points. These samples indicate the water quality prior to entry to the storage tank, after 1087 returning from the loop, and at the storage tank outlet, which provides an indirect indication of the impact the storage 1088 tank is having on water quality. 1089 1090 This Document is licensed to 1091 ISPE Good Practice Guide: 1092 Page 23 1093 Sampling for Pharmaceutical Water, Steam, and Process Gases 1094 2.2.5 1095 Distribution Systems (main loops and sub loops) 1096 Quality control sampling locations in the main distribution system should include all POUs, with sampling occurring 1097 in a manner identical to how the water is used by manufacturing as discussed above. Process control sampling 1098 locations should be located before the first and after the last POU and at other specified locations. Sampling the loop 1099 return after the last POU ensures process control as this location may be the worst case location for water quality in 1100 the entire distribution system. Water systems with multiple return loops should have sampling locations in each of the 1101 return loops. For one way (non-recirculating) branch systems, all POUs should be sampled. 1102 When temperature changes are required, the distribution system should be designed such that the pharmaceutical 1103 water is maintained at a higher pressure than heating/cooling media such as in a heat exchanger to prevent leakage 1104 from the non-sanitary utility to the sanitary side of the heat exchanger. If the pressure of the pharmaceutical water is 1105 lower than the heating/cooling media, then a double tube sheet heat exchanger design should be employed or the 1106 conductivity/TOC of the pharmaceutical water should be monitored through in-line instrumentation or regular testing 1107 of the water at a use point downstream of the heat exchanger. 1108 Sampling location selection should also consider those locations in the distribution loop where possible contamination 1109 could occur, e.g., locations downstream of a heat exchanger, flow monitor, or other equipment. 1110 Each sub loop returning to the main loop return should have a sampling point before the sub loop connects to the 1111 main loop return. 1112 In the distribution loop of a pharmaceutical water system, all use point valves and sample valves should be of a 1113 sanitary design. Sample points in a laboratory system may not be sanitary and may contain dead legs such as may 1114 be present in a gooseneck faucet. In addition, sample points located between components in a water purification 1115 system and used for process control purposes may not be sanitary and may consist of ball valves or other valves that 1116 should not be used in a pharmaceutical water distribution system. 1117 2.2.6 1118 Points of Use 1119 A Points Of Use (POU) is defined as the location where the water delivered from the water distribution system is 1120 actually used. POUs may be located on the main loop or in sub loops. 1121 When sampling for quality control sampling purposes, the POU is typically not the valve on the distribution system 1122 itself from which the water is taken. Instead, it is the end of the delivery system that transfers the water from the 1123 distribution system to the process (e.g., sinks, autoclaves, washers, process tanks). Sampling at POUs should 1124 include all components and surfaces along this water delivery pathway (e.g., heat exchangers, hoses, fittings) that 1125 the water needs to travel through to reach the process. 1126 When the outlet of a POU cannot be sampled, the next most representative sampling location should be identified. 1127 Rationale for the selection should be provided in the risk assessment. Validation and/or routine monitoring data 1128 should conclude that the sample location is representative of the POU. 1129 Direct sampling of water from a drop, e.g., to a sink in a wash room where the water is normally used does not 1130 require any additional sampling locations as the POU serves both purposes. Water collected for sampling should be 1131 collected in a way that represents how water is normally taken from the system during routine use. 1132 Every POU in a water distribution system that is not active should be locked out or tagged out of service and should 1133 be appropriately maintained to prevent contamination of the POU or the distribution system. 1134 1135 This Document is licensed to 1136 Page 24 1137 ISPE Good Practice Guide: 1138 1139 Sampling for Pharmaceutical Water, Steam, and Process Gases 1140 2.3 1141 Developing Sampling Plans 1142 The frequencies and durations of sampling plans have historically coincided with the phases of the validation 1143 program for the water system. Using this rationale, the final phase of validation forms the basis for ongoing sampling 1144 frequencies with the goal of ensuring that the system is maintained in a validated state. However, the use of risk 1145 analysis tools coupled with periodic data review may be used to alter the frequency of sampling when the data 1146 provides overwhelming support for a change (increase or decrease) in sampling frequency. Regular data review 1147 should be documented as part of the planning process. 1148 A sampling plan should define and document: 1149 • 1150 Potential sampling points and the purpose for each 1151 • 1152 Sample locations intended for continuous, regular, routine, periodic, or as needed monitoring 1153 • 1154 Attributes to be tested at each sampling location 1155 • 1156 The frequency at which samples are required for any given attribute 1157 • 1158 The purpose and use of the resulting data for those attributes 1159 • 1160 When sample valves and use point valves may be used as the same valve 1161 Sampling plans are required for both new and renovated systems and may involve sampling locations throughout the 1162 water purification and distribution systems, including source water. During renovations or modifications, evaluate the 1163 changes that will be made and use risk analysis tools to help design a sampling plan appropriate for the significance 1164 of the modification and to assess the usability of the water during this time. 1165 In specific circumstances, adequate scientific and statistical rationale may be required to not sample at an installed 1166 sampling point. 1167 Sample locations installed for technical, investigative, or diagnostic purposes may not be included in a routine 1168 sampling plan and may not be sampled on a regular basis. The purpose and use of these sample locations should be 1169 defined, and they should: 1170 • 1171 Be used on an as need basis 1172 • 1173 Be suitably maintained 1174 • 1175 Not pose a contamination risk to the system 1176 Designating sampling locations For Information Only (FIO) or for future use should be avoided. 1177 2.3.1 1178 Background Philosophy 1179 Sample plan development should be based on the expected sampling required during the life cycle of the system. 1180 A typical life cycle scenario for water sampling activities may include activities during the following steps: 1181 • 1182 Conceptual and detailed design 1183 • 1184 Construction and installation 1185 • 1186 Commissioning 1187 1188 This Document is licensed to 1189 ISPE Good Practice Guide: 1190 Page 25 1191 Sampling for Pharmaceutical Water, Steam, and Process Gases 1192 • 1193 Performance qualification (including worst case test and special scenario testing) 1194 • 1195 Ongoing testing during production 1196 • 1197 System changes and modifications (major and minor) 1198 • 1199 Periodic System Evaluation (PSE) 1200 Sampling guidance is provided in Tables 2.2, 2.3, and 2.4 regarding specific sampling locations, testing frequencies, 1201 and the type of monitoring or testing for various activities including commissioning and qualification/verification. This 1202 information may apply to new and renovated systems as well as some unique scenarios. 1203 Automated electronic monitoring of conductivity and TOC can provide a comprehensive database from which 1204 chemical water quality trends may be observed over time. When analyzed, this information may allow for minimization 1205 or elimination of manual chemical testing of the water during routine operations. However, during the Performance 1206 Qualification, confirmation that the on-line or in-line readings are representative of water quality used should be 1207 verified if this data is going to be used for product release purposes. Where localized, outlet specific contamination 1208 could occur (e.g., POU heat exchangers), a downstream instrument, or grab samples would be required for release 1209 purposes. 1210 For investigative purposes, “TOC can be used as a process control attribute to monitor the performance of unit 1211 operations comprising the purification and distribution system”. [8] 1212 Water may be delivered at several different temperatures for use at a single POU. The water sample collected from this 1213 point should be collected at the worst case temperature for microbiological control. For example, if water is delivered at 1214 both 65°C and 20°C (149°F and 68°F) from the same POU, the sample should be collected at 20°C (68°F). 1215 2.3.2 1216 Regulatory Requirements Related to Monitoring Frequency and Duration 1217 No well-defined regulatory requirements exist regarding routine sampling frequency and duration. Various 1218 pharmacopoeias provide specifications for microbiological and chemical contaminants for the various grades of water, 1219 but do not define the frequency or duration of sampling. General expectations that can be found in various compendia 1220 and consensus standards are provided as follows: 1221 USP General Information Chapter <1231> Water for Pharmaceutical Purposes, [9] states: 1222 1223 Water systems should be monitored at a frequency that is sufficient to ensure that the system is in control, and 1224 continues to produce water of acceptable quality”; and “The sampling plan should take into consideration the 1225 desired attributes of the water being sampled”. 1226 EP Guidelines to GMP Vol. 4 Annex 1 “Manufacture of Sterile Medicinal Products” [10] states: 1227 1228 “Water sources, water treatment equipment and treated water should be monitored regularly for chemical and 1229 biological contamination and, as appropriate for endotoxins”. 1230 The JP16 General Information, G8 Water, Quality Control of Water for Pharmaceutical Use, Section 4.2 Sampling [11] 1231 states: 1232 1233 “Sampling frequency should be established based on the data from validation studies on the system.” 1234 ICH Q7 Section 4.20 [12] states: 1235 1236 “All utilities that could impact on quality (e.g., steam, gases compressed air ........etc.) should be qualified and 1237 appropriately monitored and action should be taken when limits are exceeded”. 1238 1239 This Document is licensed to 1240 Page 26 1241 ISPE Good Practice Guide: 1242 1243 Sampling for Pharmaceutical Water, Steam, and Process Gases 1244 The 1993 US FDA “Guide to the Inspection of High Purity Water Systems” states [1]: 1245 1246 “The results of the Commissioning & Qualification are designed to demonstrate that all of the process steps and 1247 components within a Purified Water or Water For Injection system are capable of consistently providing water 1248 meeting the necessary water quality requirements”. 1249 The FDA guide also provides the most specific guidelines for sampling frequency during system validation, but does 1250 not stipulate frequency beyond validation and recognizes that more than one approach may be acceptable. 1251 To summarize, routine sampling should always be adequate for the intended purpose. It is necessary to look into the 1252 design and operational conditions for each specific water system with frequency and duration decisions based on RA 1253 including rationales for the decisions made. 1254 2.3.3 1255 Commissioning and Pre-validation 1256 Water testing of the pretreatment, generation and distribution systems for PW and WFI systems during 1257 commissioning is an exercise in Good Engineering Practice. This activity may contribute to developing operational 1258 and maintenance practices and procedures that will be confirmed at a later time during qualification [13] and may 1259 provide baseline information for the system that may be useful in the future. 1260 The testing activities may include chemical testing, or monitoring as appropriate, for specific unit operations and 1261 microbial testing to identify the indigenous microbial bioburden of the system. The duration of commissioning 1262 activities may vary depending upon the complexity and size of the system (e.g., 1 to 5 days). 1263 Once the entire system is operating, the effluent of each unit operation should be monitored or tested at least once 1264 during the commissioning phase in order to provide baseline values for the system. 1265 It may be necessary to perform extended pre-validation microbiological sampling activities in order to achieve process 1266 knowledge and assure system robustness. 1267 2.3.4 1268 Source Water 1269 USP General Information Chapter <1231> Water for Pharmaceutical Purposes [9] states: 1270 1271 “To ensure adherence to certain minimal chemical and microbiological quality standards, water used in the 1272 production of drug substances or as source or feed water for the preparation of the various types of purified 1273 waters must meet the requirements of the National Primary Drinking Water Regulations (NPDWR) (40 CFR 141) 1274 issued by the U.S. Environmental Protection Agency (EPA) or the drinking water regulations of the European 1275 Union or Japan, or the WHO drinking water guidelines”. 1276 Various nomenclature may be in use across the globe for drinking water including: potable water, tap water, city 1277 water, and feed water, however, the quality intent remains consistently the same. 1278 Note: there may be other additional regional/local area requirements or expectations. 1279 WHO Technical Report Series No. 929, Annex 3 [14] recommends daily sampling of incoming potable water during 1280 the early validation phase of the pharmaceutical water system to verify its quality, but this may not be required in all 1281 areas of the world. 1282 If there is little or no knowledge of the feed water quality at the system inlet, it is recommended to take daily samples 1283 for analysis including organism(s) of concern if such local requirement exists. If the potable water quality is already 1284 well known and the historical data shows good, stable water quality, a rationale can be made to perform less source 1285 water testing during the Performance Qualification (PQ) period. Typically, the potable/drinking water supplier is 1286 responsible for the potable water quality up to the property boundary, thereafter the responsibility is transferred to the 1287 1288 This Document is licensed to 1289 ISPE Good Practice Guide: 1290 Page 27 1291 Sampling for Pharmaceutical Water, Steam, and Process Gases 1292 property owner/site operator. Keeping in mind that pharmaceutical product manufacturers are ultimately responsible 1293 for end product quality, the burden of proof falls on the pharmaceutical manufacturer to assure, through suitable 1294 testing, and that potable water quality is maintained, as required, up to the inlet of the water system. 1295 An impact assessment should be performed on potable water depending on the attributes, the treatment technology 1296 employed, and the end use of the water. 1297 Depending on the end product or the process purifications steps etc. it may be possible to classify the potable water 1298 with different criticality levels. A RA with rationales should be made in order to define the criticality of the potable 1299 water. The outcome of the assessment will tell if Good Engineering Practices or validation/verification of the potable 1300 water supply is necessary. 1301 2.3.4.1 Pretreatment 1302 Pretreatment is a term that represents designated treatment steps, and typically prior to primary/final treatment. This 1303 may also be designated by the unit operations involved such as filtering, softening, or RO. It may be convenient to 1304 refer to pretreated water as feed water for the production of PW, WFI and/or pure stream (PS). When used as feed 1305 water for WFI or PS, pretreated water does not necessarily need to comply with PW requirements to ensure WFI/PS 1306 quality water is obtained. It should be noted that the Chinese Pharmacopoeia (ChP) [15] requires that supply water to 1307 a WFI still must comply with PW specification. The USP, EP, or JP does not require PW quality as feed for production 1308 of WFI. 1309 In order to ensure that the pretreatment fulfills its purpose, sampling is carried out specifically for each purification 1310 step as mentioned in the ISPE Good Practice Guide: Approaches to Commissioning and Qualification of 1311 Pharmaceutical Water and Steam Systems (Second Edition) [13] and includes: 1312 • 1313 For a sand, mixed media, or similar filter, the SDI may be measured before and after the equipment. The worst 1314 case sampling time would be just before the automatic back flush. 1315 • 1316 For a water softener, hardness may be measured before and after the softener. The worst case sampling time 1317 may be immediately before softener regeneration. 1318 • 1319 For a carbon filter, the chlorine concentration may be measured before and after the carbon filter. A worst case 1320 sampling time may be immediately before the back flush. 1321 • 1322 For a RO system, conductivity may be measured before and after the RO unit 1323 • 1324 For CEDI, the conductivity may be measured before and after the CEDI 1325 For all purification steps microbiological testing may be required, including before and after the RO if applicable. 1326 Acceptable microbiological levels should be defined for each purification step when applicable. There should at least 1327 be an Action Level with consideration for adding an Alert level if appropriate. 1328 If, during RA, the pretreatment is determined to be a no impact system, samples should be taken for record purposes 1329 during Performance Qualification (PQ) of PW/WFI. There is no regulatory requirement, but if the pretreatment is 1330 determined to be indirect impact, a rationale should be developed for the sampling frequency during the PQ of the 1331 pretreatment. 1332 2.3.4.2 Purified Water 1333 Depending on the number of POUs and sample points in the distribution system consideration could be given 1334 to rotational sampling for chemical tests since chemical contaminants are uniformly distributed in the water. It is 1335 generally recommended that several samples be taken from each use point during the PQ period. It should be noted 1336 that the worst case testing or special testing may influence the results and should be taken into account. The number 1337 of worst case tests chosen may influence the length of the PQ. 1338 1339 This Document is licensed to 1340 Page 28 1341 ISPE Good Practice Guide: 1342 1343 Sampling for Pharmaceutical Water, Steam, and Process Gases 1344 Define the microbiological levels as required by intended use of the water. There should be an Alert Level and an 1345 Action Level. Testing frequency should be determined based on a suitable RA. 1346 2.3.4.3 Water For Injection 1347 POU sampling plans should rotate through all use points on the system, with the expectation that samples are 1348 collected on a daily basis from various use points, and that all use points are sampled on a routine basis based on 1349 risk analysis. System representative locations (e.g., supply and/or return) may be sampled on a daily basis for the 1350 process control purposes and to aid in the investigation of OOS results from individual use points. 1351 During the PSE of a WFI (or PW) system, sampling frequency may be evaluated. Increases or decreases in sampling 1352 frequency should be based on historical data and should only occur when data indicates that more or less frequent 1353 sampling is justified based on performance of the use point or system. Sample frequency should not be arbitrarily 1354 adjusted without adequate scientific and statistical justification. A decrease in sampling frequency may result in 1355 increased regulatory scrutiny. Decreased sampling frequency also puts more product at risk during investigations 1356 because of longer periods between acceptable test results surrounding a deviant test result. In short, careful 1357 consideration should be given before decreasing sampling frequency relative to the increased number of product 1358 batches that will be put at risk. 1359 2.3.5 1360 Sampling for Performance Qualification/Verification 1361 2.3.5.1 Purpose 1362 The purpose of PQ/verification is to document that the stated water quality is delivered from the water system during: 1363 1. 1364 Typical (nominal) use conditions 1365 2. 1366 Worst case use conditions (system limits) 1367 3. 1368 Special scenarios 1369 and that the water complies with the requirements for the grade of pharmaceutical water type specified. 1370 The period for the PQ depends on whether it applies to a completely new system or to modifications made to an 1371 existing system. 1372 2.3.5.2 Typical Use Conditions (Nominal) 1373 All PQs include a period of normal operation or simulated normal operation to demonstrate that the water quality 1374 during normal operation complies with the requirements for the actual specified pharmaceutical water quality. Normal 1375 operational usage ranges are defined with rationales. 1376 2.3.5.3 Worst Case Use Conditions 1377 A worst case scenario tests the limits of the system and is defined as the most extreme situations during the 1378 operation of the water system. 1379 Below are a few examples of worst case conditions: 1380 • 1381 Conditions of no water consumption (e.g., during a weekend or shut down) can result in a rise in the level of 1382 bioburden and conductivity 1383 • 1384 Maximum use of the water confirms the system or unit operation is capable of delivering suitable water quality 1385 during a period of maximum delivery 1386 1387 This Document is licensed to 1388 ISPE Good Practice Guide: 1389 Page 29 1390 Sampling for Pharmaceutical Water, Steam, and Process Gases 1391 • 1392 If routine sanitization occurs, just before start of and after the sanitization 1393 • 1394 After a maintenance procedure or a periodic event such as derouging and/or passivation 1395 2.3.5.4 Special Scenarios 1396 Special scenarios are situations which may occur during normal use either as unplanned or as planned scenarios. 1397 Often these special scenarios occur at inconvenient times, and it would be advantageous to validate these scenarios 1398 to avoid delays when they occur. 1399 Validation of special scenarios can be part of the ordinary PQ of the pharmaceutical water system, or it can be 1400 performed as a specific PQ for only this purpose. 1401 The validation of a special scenario may need to be performed in a worst case scenario and be documented by water 1402 quality sampling data. The outcome and conclusion of the validation, based on the data, can decrease time spent on 1403 the procedure during daily operation, when going from the special scenario to normal condition, due to less sampling, 1404 sanitization, etc. 1405 Examples of special scenarios are provided below: 1406 • 1407 Power failure: to document that the pharmaceutical water quality is unchanged after restarting after a defined 1408 period of time without any sanitization. A prerequisite for this test is no breach of integrity of the water system. 1409 • 1410 Emergency or unplanned maintenance procedures: including unusual or unexpected component failure, system 1411 damage from internal or external action, or other situations that may require action but cannot be anticipated. 1412 Validation of all emergency or unplanned maintenance procedures is not practical. 1413 Acceptance criteria should be set up for the special scenario properly because if it is not set properly, the PQ may 1414 fail. For example, if after a power failure the water retains its quality characteristics for three hours, then a power 1415 failure of up to three hours can be followed by normal operation without any sanitization, provided the system integrity 1416 is not otherwise compromised and will only require documentation of the actual duration of the power failure. A 1417 special scenario test may be performed for a longer period than the period used as the limit, if there is a need for 1418 incorporation of a safety factor. The three-hour power failure scenario mentioned above is only an example and that 1419 shorter or longer times may actually be validated. 1420 2.3.5.5 Risk-Based Approach 1421 A risk-based approach to C&Q/verification may be utilized to develop the monitoring and testing strategy. For 1422 example, risk analyses or impact assessments may be utilized to determine what equipment or functions of the 1423 water system are considered critical [13]. There may be unit operations or process steps within the generation or 1424 distribution system that require testing during initial commissioning as well as during ongoing routine operation (e.g., 1425 UV treatment, pumps, heat exchangers, etc.) that may not be considered critical. 1426 If determined to be non-critical, such equipment operation (e.g., multimedia filters, softeners) may be tested only 1427 during commissioning as part of the C&Q/Verification process [13]. Testing relating to the continued operation of 1428 a unit process may become part of a routine monitoring program based upon the impact to the system and the 1429 finished quality of the water as detailed in SOPs and subject to Good Engineering Practices (GEP). Frequency of 1430 routine testing could be adjusted based on the consistency of equipment performance. These tests assure proper 1431 maintenance and operation of the water system (e.g., regeneration of resin beds, sanitization of equipment, etc.) [13]. 1432 A CQA is a test defined as critical, based on the need to meet a specific compendia specification or other defined non- 1433 compendial requirements [13]. For example, conductivity may be a CQA of water when measured at the outlet of a PW 1434 generation system. Sampling the conductivity at this point would initially occur during C&Q/verification activities [13]. 1435 1436 This Document is licensed to 1437 Page 30 1438 ISPE Good Practice Guide: 1439 1440 Sampling for Pharmaceutical Water, Steam, and Process Gases 1441 A risk-based approach may also be utilized for determining the sampling approach to qualify changes made to an 1442 existing validated water system. 1443 2.3.5.6 Performance Qualification Sampling 1444 The PQ sampling protocol should describe the purpose, scope, and strategy (including justification). The sampling 1445 plan and the expected data from in-line/on-line measuring equipment (e.g., for temperature, pressure, conductivity, 1446 TOC) should be described in the protocol. The prerequisites and the acceptance criteria for the performance testing 1447 should also be stated in the protocol. The protocol should also consider adjusting sampling as appropriate based on 1448 production activities (e.g., if there is no production during weekends or if production varies significantly). 1449 All stakeholders involved in the PQ, (production, engineering, quality control and quality assurance) should have an 1450 opportunity to review and comment on the protocol. 1451 Communication and information sharing between the different participants before, during, and after the PQ are 1452 essential for efficient and successful qualification. Feedback on changes (e.g., missing samples) during the 1453 qualification should immediately be communicated to the relevant participants and documented in a deviation. 1454 PQ sampling should be performed for a newly installed water system or for significant changes made to an existing 1455 qualified system. When minor changes are made to a qualified system, reduced sampling requirements may be 1456 considered. See the ISPE Good Practice Guide: Approaches to Commissioning and Qualification of Pharmaceutical 1457 Water and Steam Systems (Second Edition) [13] and the ISPE Baseline® Guide: Water and Steam Systems (Second 1458 Edition) [7] for additional information. 1459 The following provides basic guidance regarding sampling plans during formal validation. Sampling during 1460 commissioning provides information that may be used to develop SOPs (e.g., for routine system operation, sampling, 1461 cleaning, sanitization, and maintenance). 1462 Initial Sampling 1463 Initial sampling is the most intensive and typically lasts from two to four weeks. Commonly, this period can be 1464 satisfied by monitoring and testing all sample and use points in the distribution system and selected other points 1465 daily for between 10 to 20 consecutive working days depending upon the design of the system (e.g., overall size, 1466 complexity, or number of use points). If the weekends involve no water consumption, the first day of the week may 1467 represent a worst case scenario for sampling. 1468 During initial sampling, if consistency in operation is apparent, representative microorganisms that are visually distinct 1469 (i.e., physical colony characteristics) and detected during microbiological monitoring may be identified. Since all 1470 parts of a water systems are not always completely sterile by design, inherent populations of microorganisms may 1471 be found and these populations can vary from system to system. Typically, there are only a few distinct types of 1472 microorganisms that make up the resident population. 1473 Initial sampling activities are performed to demonstrate that the production and delivery of water consistently meets 1474 the specified compendial requirements. The initial sampling phase of PQ typically starts after Installation Qualification 1475 (IQ) and Operational Qualification (OQ) report approval. Initial sampling should demonstrate consistent quality before 1476 start of the next phase. 1477 Intermediary Sampling 1478 Intermediary sampling is less intensive and may involve reduced sampling frequency. This phase typically lasts for 1479 10 to 20 working days. The purpose of intermediary sampling is to further demonstrate consistent production and 1480 delivery of water (e.g., repeatability) of the required quality within the established ranges when using SOPs. 1481 Intermediary sampling may provide additional worst case data since use points are not as frequently active. 1482 1483 This Document is licensed to 1484 ISPE Good Practice Guide: 1485 Page 31 1486 Sampling for Pharmaceutical Water, Steam, and Process Gases 1487 Extended Sampling 1488 The extended sampling portion of the water system qualification activities continue for not less than the remaining 1489 balance of one year’s time (one year minus the time used for initial and intermediary sampling) [13]. Collectively, the 1490 validation process includes an extended sampling study of the system and allows for the system to be challenged by 1491 evaluating its effectiveness of delivering water of acceptable quality despite seasonal variations in the potable water 1492 feed to the system, the slow development of the system’s natural flora, and any other variations that may occur [1]. 1493 Typically, the extended sampling plan is reduced in frequency [13], from initial and intermediary sampling activities. 1494 Extended sampling monitoring typically forms the basis for the routine monitoring and testing program established 1495 once the validation has been completed. During this phase, there may be subtle shifts in the types of microorganisms 1496 inherently present within the system due to seasonal changes. Only those isolates from the microbiological 1497 testing samples that are visually distinct (i.e., physical colony characteristics) should be identified if not previously 1498 characterized. 1499 Table 2.2: Examples of Suggested Source Water and Equipment Testing and Frequencies (Note A) 1500 Equipment Sample 1501 Location 1502 Suggested 1503 Commissioning and 1504 Pre-validation Tests and 1505 Frequencies 1506 Suggested Initial Phase 1507 Tests and Frequencies 1508 Suggested Intermediary 1509 Phase Tests and 1510 Frequencies 1511 Suggested Extended 1512 Tests and Frequencies 1513 (Note B) 1514 Potable Source Water 1515 Supply at Inlet to Pre- 1516 Treatment 1517 Qualification of a new water source in the facility may involve initial and periodic testing of the source water to 1518 verify and ensure continuing compliance with potable water requirements. If the water is from a reliable source, at a 1519 minimum, obtain and review testing certificates from the supplier to establish full compliance with appropriate potable 1520 water regulations. Based on a risk assessment, it may be necessary to verify, via testing, the supplied water entering 1521 the facility complies with all potable water requirements. Local/country requirements should also be followed for 1522 water used in the manufacture of products sold in those areas. 1523 Media Filter 1524 Confirm SDI re-duction 1525 Quarterly or on-line for process control purposes, or at a frequency commensurate 1526 with the criticality of the process to the production of the desired grade of water. 1527 Organic Scavenger Bed 1528 Confirm TOC re-duction 1529 Quarterly or on-line for process control purposes, or at a frequency commensurate 1530 with the criticality of the process to the production of the desired grade of water. 1531 Water Softener 1532 Confirm hardness 1533 reduction 1534 Quarterly or on-line for process control purposes, or at a frequency commensurate 1535 with the criticality of the process to the production of the desired grade of water. 1536 Carbon Filter for 1537 Dechlorination 1538 • 1539 Confirm free and total 1540 chlorine reduction 1541 • 1542 Monitor 1543 microbiological levels 1544 • 1545 Quarterly or on-line for process control purposes, or at a frequency commensurate 1546 with the criticality of the process to the production of the desired grade of water. 1547 • 1548 Monitoring microbial levels may be important as needed for process control 1549 purposes. 1550 Chemical Feed for 1551 Declorination 1552 • 1553 Confirm free and total 1554 chlorine reduction 1555 • 1556 Confirm sufficient 1557 dosage without 1558 overdosing (Note H) 1559 Quarterly or on-line for process control purposes, or at a frequency commensu-rate 1560 with the criticality of the process to the production of the desired grade of water. 1561 Carbon Filter for Organics 1562 Reduction 1563 • 1564 Confirm TOC 1565 reduction 1566 • 1567 Monitor 1568 microbiological levels 1569 • 1570 Quarterly or on-line for process control purposes, or at a frequency commensurate 1571 with the criticality of the process to the production of the desired grade of water. 1572 • 1573 Monitoring microbial levels may also be important. 1574 Ultrafiltration for SDI or 1575 TOC Reduction (Note I) 1576 Confirm SDI and/or TOC 1577 reduction 1578 Quarterly or on-line for process control purposes, or at a frequency commensu-rate 1579 with the criticality of the process to the production of the desired grade of water. 1580 Deionization as a 1581 Pretreatment Process 1582 • 1583 Confirm conductivity 1584 reduction 1585 • 1586 Monitor 1587 microbiological levels 1588 • 1589 Quarterly or on-line for process control purposes, or at a frequency commensurate 1590 with the criticality of the process to the production of the desired grade of water 1591 (Note F). 1592 • 1593 Monitoring microbial levels may also be important. 1594 1595 This Document is licensed to 1596 Page 32 1597 ISPE Good Practice Guide: 1598 1599 Sampling for Pharmaceutical Water, Steam, and Process Gases 1600 Table 2.2: Examples of Suggested Source Water and Equipment Testing and Frequencies (Note A) (continued) 1601 Equipment Sample 1602 Location 1603 Suggested 1604 Commissioning and 1605 Pre-validation Tests and 1606 Frequencies 1607 Suggested Initial Phase 1608 Tests and Frequencies 1609 Suggested Intermediary 1610 Phase Tests and 1611 Frequencies 1612 Suggested Extended 1613 Tests and Frequencies 1614 (Note B) 1615 Reverse Osmosis 1616 • 1617 Confirm conductivity 1618 and TOC reduction 1619 • 1620 Confirm pH levels if 1621 pH is adjusted 1622 • 1623 Monitor 1624 microbiological levels 1625 • 1626 Quarterly or on-line for process control purposes, or at a frequency commensurate 1627 with the criticality of the process to the production of the desired grade of water 1628 (Note F). 1629 • 1630 Monitoring microbial levels may also be important. 1631 CEDI or other 1632 Deionization Process as a 1633 Final Treatment 1634 • 1635 Confirm conductivity 1636 reduction 1637 • 1638 Confirm chemical 1639 reduction (Note D) 1640 • 1641 Monitor TOC and 1642 microbiological levels 1643 • 1644 Conductivity and TOC 1645 daily or on-line (Note 1646 C) 1647 • 1648 Chemical testing daily 1649 (Note D) 1650 • 1651 Microbiological testing 1652 daily (Note E, F) 1653 • 1654 Conductivity and 1655 TOC at a frequency 1656 between daily and 1657 weekly or on-line 1658 (Note C, J) 1659 • 1660 Chemical testing 1661 between daily and 1662 weekly (Note D, J) 1663 • 1664 Microbiological testing 1665 between daily and 1666 weekly (Note E, F, J) 1667 • 1668 Conductivity and TOC 1669 at regular intervals 1670 or on-line (Note C), 1671 typically at least once 1672 per week 1673 • 1674 Chemical (Note D) 1675 and microbiological 1676 (Note E, F) testing 1677 at regular intervals, 1678 typically at least once 1679 per week 1680 Ozone as part of the 1681 Purification System (Note 1682 F, G) 1683 Monitor ozone and 1684 microbiological levels 1685 • 1686 Ozone levels on-line 1687 daily 1688 • 1689 Microbiological testing 1690 daily (Note E, F) 1691 • 1692 Ozone levels on-line 1693 • 1694 Microbiological testing 1695 between daily and 1696 weekly (Note E, F, J) 1697 • 1698 Ozone levels on-line 1699 • 1700 Microbiological testing 1701 at regular intervals, 1702 typically at least once 1703 per week (Note E, F, 1704 J) 1705 Ultraviolet Treatment for 1706 Microbiological Control 1707 or TOC Reduction (Note 1708 F, K) 1709 Confirm microbio-logical 1710 reduction 1711 Microbiological testing 1712 daily (Note E) 1713 Microbiological testing 1714 between daily and weekly 1715 (Note E, F, J) 1716 Microbiological testing at 1717 regular intervals (Note E, 1718 F), typically at least once 1719 per week 1720 Ultraviolet Treatment 1721 for Ozone Destruction 1722 (Note F) 1723 Monitor ozone levels 1724 Quarterly or on-line for process control purposes, or at a frequency commensurate 1725 with the criticality of the process to the production of the desired grade of water. 1726 Ultrafiltration for Endotoxin 1727 Control (Note F) 1728 Confirm absence of 1729 detectable endotoxin 1730 (Note L) 1731 Endotoxin testing daily 1732 Endotoxin testing 1733 between daily and weekly 1734 (Note J) 1735 Endotoxin testing at 1736 regular intervals, typically 1737 at least once per week 1738 Microbially Retentive 1739 Filtration (Note F) 1740 Confirm acceptable 1741 microbiological levels 1742 Microbiological testing 1743 daily (Note E) 1744 Microbiological testing 1745 between daily and weekly 1746 (Note E, F, J) 1747 Microbiological testing at 1748 regular intervals (Note E, 1749 F), typically at least once 1750 per week 1751 Distillation 1752 Confirm conductivity, 1753 TOC, endotoxin, 1754 microbiological, and 1755 chemical (Note D) content 1756 • 1757 Conductivity and TOC 1758 daily or on-line (Note 1759 C) 1760 • 1761 Chemical testing daily 1762 (Note D) 1763 • 1764 Microbiological testing 1765 daily (Note E, F) 1766 • 1767 Conductivity and 1768 TOC at a frequency 1769 between daily and 1770 weekly or on-line 1771 (Note C, J) 1772 • 1773 Chemical testing 1774 between daily and 1775 weekly (Note D, J) 1776 • 1777 Microbiological testing 1778 between daily and 1779 weekly (Note E, F, J) 1780 • 1781 Conductivity and TOC 1782 at regular intervals 1783 or on-line (Note C), 1784 typically at least once 1785 per week 1786 • 1787 Chemical (Note D) 1788 and microbiological 1789 (Note E, F) testing 1790 at regular intervals, 1791 typically at least once 1792 per week 1793 1794 This Document is licensed to 1795 ISPE Good Practice Guide: 1796 Page 33 1797 Sampling for Pharmaceutical Water, Steam, and Process Gases 1798 Table 2.2: Examples of Suggested Source Water and Equipment Testing and Frequencies (Note A) (continued) 1799 Equipment Sample 1800 Location 1801 Suggested 1802 Commissioning and 1803 Pre-validation Tests and 1804 Frequencies 1805 Suggested Initial Phase 1806 Tests and Frequencies 1807 Suggested Intermediary 1808 Phase Tests and 1809 Frequencies 1810 Suggested Extended 1811 Tests and Frequencies 1812 (Note B) 1813 Finished Water leaving 1814 Purification System 1815 Confirm conductivity, 1816 chemical (Note D), 1817 TOC, microbiological, 1818 and endotoxin levels, if 1819 applicable 1820 • 1821 Conductivity and TOC 1822 daily or on-line (Note 1823 C) 1824 • 1825 Chemical testing daily 1826 (Note D) 1827 • 1828 Microbiological testing 1829 daily (Note E) 1830 • 1831 Endotoxin testing, if 1832 applicable, daily 1833 • 1834 Conductivity and 1835 TOC at a frequency 1836 between daily and 1837 weekly or on-line 1838 (Note C) 1839 • 1840 Chemical testing at 1841 a frequency between 1842 daily and weekly 1843 (Note D) 1844 • 1845 Microbiological 1846 testing at a frequency 1847 between daily and 1848 weekly (Note E) 1849 • 1850 Endotoxin testing, 1851 if applicable at a 1852 frequency between 1853 daily and weekly 1854 • 1855 Conductivity and TOC 1856 at regular intervals 1857 or on-line (Note C), 1858 typically at least once 1859 per week 1860 • 1861 Chemical (Note D), 1862 microbiological (Note 1863 E) and endotoxin (if 1864 applicable) testing 1865 at regular intervals, 1866 typically at least once 1867 per week 1868 Note A: This table lists several of the most commonly used treatment processes and may contain treatment steps that are not present in all 1869 systems. There may also be treatment processes in use that are not listed in this table. In all cases, the extent and frequency of sampling should 1870 be commensurate with the criticality of the process to the production of the desired grade of water. 1871 Note B: The frequency and extent of sampling during the extended phase of the performance qualification typically serves as the initial frequency 1872 and extent for future validation maintenance testing, but is subject to adjustment based on a review of accumulated data and the use of risk 1873 analysis tools. 1874 Note C: Verification of the equivalency of on-line readings to those obtained from sampling should be established in order to utilize on-line readings 1875 as suitable for quality control release. 1876 Note D: Chemical testing may or may not be required as determined by applicable pharmacopoeias. 1877 Note E: Microbiological identification testing may be performed to provide a profile of the resident micro flora within the water system. 1878 Note F: Depending on the criticality of the process to the desired impurity/contaminant content of the water, less frequent sampling and testing may 1879 be performed for processes determined to be less critical. 1880 Note G: Refer to the ISPE Good Practice Guide: Ozone Sanitization of Pharmaceutical Water Systems [16] for a more complete review of ozone 1881 and its use in the pharmaceutical industry. 1882 Note H: A slight excess is desirable, but overdosing may create undesirable and unintended side effects. 1883 Note I: In the event that ultrafiltration is used for microbial reduction, testing should reflect the impurity/contaminant reflects process operation. 1884 Note J: Testing frequency dependent on criticality. 1885 Note K: When an ultraviolet treatment unit is used for TOC reduction, organic molecules are oxidized and converted into molecules that 1886 actually increase the conductivity of the water. Typically, the increase in conductivity is not monitored because there is an ion exchange process 1887 immediately downstream of the ultraviolet treatment unit that removes the oxidized organic material, reducing both the conductivity and the TOC. 1888 Note L: The limit of detection shall be commensurate with the level of endotoxin appropriate for the process. 1889 1890 This Document is licensed to 1891 Page 34 1892 ISPE Good Practice Guide: 1893 1894 Sampling for Pharmaceutical Water, Steam, and Process Gases 1895 Table 2.3: Examples of Purified Water Distribution System Water Testing and Frequencies (Note A) 1896 Purified Water 1897 Distribution System 1898 (Sample Location) 1899 Commissioning Tests 1900 and Frequencies 1901 Initial Phase Tests and 1902 Frequencies 1903 Intermediary Phase 1904 Tests and Frequen-cies 1905 Extended Phase Tests 1906 and Frequencies 1907 (Note B) 1908 Downstream of 1909 Distribution Pump 1910 discharge prior to 1911 beginning in loop unit 1912 operations 1913 Confirm conductivity, 1914 chemical (Note E), TOC 1915 and micro-biological 1916 levels 1917 • 1918 Conductivity and TOC 1919 daily or on-line (Note 1920 D) 1921 • 1922 Chemical testing daily 1923 (Note E) 1924 • 1925 Microbiological testing 1926 daily (Note F) 1927 • 1928 Conductivity and 1929 TOC at a frequency 1930 between daily and 1931 weekly or on-line 1932 (Note D) 1933 • 1934 Chemical testing at 1935 a frequency between 1936 daily and weekly 1937 (Note E) 1938 • 1939 Microbiological 1940 testing at a frequency 1941 between daily and 1942 weekly (Note F) 1943 • 1944 Conductivity and TOC 1945 at regular intervals 1946 or on-line (Note D), 1947 typically at least once 1948 per week 1949 • 1950 Chemical (Note E) 1951 and microbiological 1952 (Note F) testing at 1953 regular intervals, 1954 typically at least once 1955 per week 1956 Sample Port following 1957 the beginning in loop 1958 operations or the first use 1959 point 1960 Confirm conductivity, 1961 chemical (Note E), TOC 1962 and microbiological levels 1963 • 1964 Conductivity and TOC 1965 daily or on-line (Note 1966 D) 1967 • 1968 Chemical testing daily 1969 (Note E) 1970 • 1971 Microbiological testing 1972 daily (Note F) 1973 • 1974 Conductivity and 1975 TOC at a frequency 1976 between daily and 1977 weekly or on-line 1978 (Note D) 1979 • 1980 Chemical testing at 1981 a frequency between 1982 daily and weekly 1983 (Note E) 1984 • 1985 Microbiological 1986 testing at a frequency 1987 between daily and 1988 weekly (Notes F) 1989 • 1990 Conductivity and TOC 1991 at regular intervals 1992 or on-line (Note D), 1993 typically at least once 1994 per week 1995 • 1996 Chemical (Note E) 1997 and microbiological 1998 (Note F) testing at 1999 regular intervals, 2000 typically at least once 2001 per week 2002 Critical Use Point (Note G) 2003 Confirm conductivity, 2004 chemical (Note E), TOC 2005 and microbiological levels 2006 • 2007 Conductivity and TOC 2008 daily or on-line (Note 2009 D) 2010 • 2011 Chemical testing daily 2012 (Note E) 2013 • 2014 Microbiological testing 2015 daily (Note F) 2016 • 2017 Conductivity and 2018 TOC at a frequency 2019 between daily and 2020 weekly or on-line 2021 (Note D) 2022 • 2023 Chemical testing at 2024 a frequency between 2025 daily and weekly 2026 (Note E) 2027 • 2028 Microbiological 2029 testing at a frequency 2030 between daily and 2031 weekly (Note F) 2032 • 2033 Conductivity and TOC 2034 at regular intervals 2035 or on-line (Note D), 2036 typically at least once 2037 per week 2038 • 2039 Chemical (Note E) 2040 and microbiological 2041 (Note F) testing at 2042 regular intervals, 2043 typically at least once 2044 per week 2045 Non-critical Use Point 2046 (Note H) 2047 Confirm conductivi-ty, 2048 chemical (Note E), TOC 2049 and micro-biological 2050 levels 2051 • 2052 Conductivity and TOC 2053 daily or on-line (Note 2054 D) 2055 • 2056 Chemical testing daily 2057 (Note E) 2058 • 2059 Microbiological testing 2060 daily (Note F) 2061 • 2062 Conductivity and 2063 TOC at a frequency 2064 between daily and 2065 monthly or on-line 2066 (Note D) 2067 • 2068 Chemical testing at 2069 a frequency between 2070 daily and monthly 2071 (Note E) 2072 • 2073 Microbiological 2074 testing at a frequency 2075 between daily and 2076 monthly (Note F) 2077 • 2078 Conductivity and TOC 2079 at regular intervals 2080 or on-line (Note D), 2081 typically at least once 2082 per month 2083 • 2084 Chemical (Note E) 2085 and microbiological 2086 (Note F) testing at 2087 regular intervals, 2088 typically at least once 2089 per month 2090 2091 This Document is licensed to 2092 ISPE Good Practice Guide: 2093 Page 35 2094 Sampling for Pharmaceutical Water, Steam, and Process Gases 2095 Table 2.3: Examples of Purified Water Distribution System Water Testing and Frequencies (Note A) (continued) 2096 Purified Water 2097 Distribution System 2098 (Sample Location) 2099 Commissioning Tests 2100 and Frequencies 2101 Initial Phase Tests and 2102 Frequencies 2103 Intermediary Phase 2104 Tests and Frequen-cies 2105 Extended Phase Tests 2106 and Frequencies 2107 (Note B) 2108 Last POU or end of loop 2109 sample port (Note I) 2110 Confirm conductivity, 2111 chemical (Note E), TOC, 2112 and microbiological levels 2113 Confirm any other 2114 physical attributes such 2115 as flow rate, temperature, 2116 pressure, etc. at minimum 2117 and maximum water use 2118 scenarios as defined 2119 in a User Requirement 2120 Specification 2121 • 2122 Conductivity and TOC 2123 daily or on-line (Note 2124 D) 2125 • 2126 Chemical testing daily 2127 (Note E) 2128 • 2129 Microbiological testing 2130 daily (Note F) 2131 • 2132 Conductivity and 2133 TOC at a frequency 2134 between daily and 2135 weekly or on-line 2136 (Note D) 2137 • 2138 Chemical testing at 2139 a frequency between 2140 daily and weekly 2141 (Note E) 2142 • 2143 Microbiological 2144 testing at a frequency 2145 between daily and 2146 weekly (Note F) 2147 • 2148 Conductivity and TOC 2149 at regular intervals 2150 or online (Note D), 2151 typically at least once 2152 per week 2153 • 2154 Chemical (Note E) 2155 and microbiological 2156 (Note F) and other 2157 physical attribute 2158 testing at regular 2159 intervals, typically at 2160 least once per week 2161 Note A: For any equipment that may be present as part of the PW distribution system, refer to Table 2.2. 2162 Note B: The frequency and extent of sampling during the extended phase of the performance qualification typically serves as the initial frequency 2163 and extent for future validation maintenance testing, but is subject to adjustment based on a review of accumulated data and the use of risk 2164 analysis tools. 2165 Note C: There are many variations on distribution system designs. This portion of the table is intended to capture considerations in “chase the tail” 2166 system configurations having no storage tank. In these instances, the incoming water to this system should be verified as suitable. 2167 Note D: Verification of the equivalency of on-line readings to those obtained from sampling should be established in order to utilize on-line readings 2168 as suitable for quality control release. 2169 Note E: Chemical testing may or may not be required as determined by applicable pharmacopoeias. 2170 Note F: Microbiological identification testing is performed to provide a profile of the resident micro flora within the water system. Only those isolated 2171 from the microbiological samples taken at POU that are visually distinct (i.e., physical colony characteristics) should be identified to minimize 2172 redundancy. Microbial identification is not required where microorganisms with identical colony characteristics have been previously identified 2173 Note G: A critical use point is defined as a point providing water that is either an ingredient or has direct product contact or one that is used for final 2174 rinse 2175 Note H: A non-critical use point could be a point providing water that does not directly contact the product or process, such as pre-rinse or non- 2176 product contact use, or a use point that has been taken out of service and is no longer used for manufacturing but has not been removed from the 2177 system. 2178 Note I: There may be additional equipment such as heat exchangers, ultraviolet units, etc., between the last POU and the storage tank. In these 2179 cases, it may also be appropriate to monitor each of these unit operations before the water is returned to the storage tank. 2180 2181 This Document is licensed to 2182 Page 36 2183 ISPE Good Practice Guide: 2184 2185 Sampling for Pharmaceutical Water, Steam, and Process Gases 2186 Table 2.4: Examples of WFI Distribution System Water Testing and Frequencies 2187 WFI Distribution System 2188 (Sample Location) 2189 Commissioning Tests 2190 and Frequencies 2191 Initial Phase Tests and 2192 Frequencies 2193 Intermediary Phase 2194 Tests and Frequencies 2195 (Note B) 2196 Extended Phase Tests 2197 and Frequencies (Note 2198 A, B) 2199 Downstream of 2200 Distribution Pump 2201 discharge prior to 2202 beginning in loop unit 2203 operations 2204 Confirm conductivity, 2205 chemical (Note D), 2206 TOC, endo-toxin and 2207 microbiological levels 2208 • 2209 Conductivity and TOC 2210 daily or on-line (Note 2211 C) 2212 • 2213 Chemical testing daily 2214 (Note D) 2215 • 2216 Endotoxin testing 2217 daily 2218 • 2219 Microbiological testing 2220 daily (Note E) 2221 • 2222 Conductivity and 2223 TOC at a frequency 2224 between daily and 2225 weekly or on-line 2226 (Note C) 2227 • 2228 Chemical testing at 2229 a frequency between 2230 daily and weekly 2231 (Note D) 2232 • 2233 Endotoxin testing at 2234 a frequency between 2235 daily and weekly 2236 (Note B) 2237 • 2238 Microbiological 2239 testing at a frequency 2240 between daily and 2241 weekly (Notes B, E) 2242 • 2243 Conductivity and TOC 2244 at regular intervals 2245 or on-line (Note C), 2246 typically at least once 2247 per week 2248 • 2249 Chemical (Note D), 2250 endotoxin (Note B), 2251 and microbiological 2252 (Notes B, E) testing 2253 at regular intervals, 2254 typically at least once 2255 per week 2256 Sample Port following 2257 the beginning in loop 2258 operations or the first use 2259 point 2260 Confirm conductivity, 2261 chemical (Note D), 2262 TOC, endo-toxin and 2263 microbiological levels 2264 • 2265 Conductivity and TOC 2266 daily or on-line (Note 2267 C) 2268 • 2269 Chemical testing daily 2270 (Note D) 2271 • 2272 Endotoxin testing 2273 daily 2274 • 2275 Microbiological testing 2276 daily (Note E) 2277 • 2278 Conductivity and 2279 TOC at a frequency 2280 between daily and 2281 weekly or on-line 2282 (Note C) 2283 • 2284 Chemical testing at 2285 a frequency between 2286 daily and weekly 2287 (Note D) 2288 • 2289 Endotoxin testing at 2290 a frequency between 2291 daily and weekly 2292 (Note B) 2293 • 2294 Microbiological 2295 testing at a frequency 2296 between daily and 2297 weekly (Note B, E) 2298 • 2299 Conductivity and TOC 2300 at regular intervals 2301 or on-line (Note C), 2302 typically at least once 2303 per week 2304 • 2305 Chemical (Note D), 2306 endotoxin (Note B), 2307 and microbiological 2308 (Notes B, E) testing 2309 at regular intervals, 2310 typically at least once 2311 per week 2312 Critical Use Point (Note F) 2313 Confirm conductivity, 2314 chemical (Note D), 2315 TOC, endotoxin, and 2316 microbiological levels 2317 • 2318 Conductivity and TOC 2319 daily or on-line (Note 2320 C) 2321 • 2322 Chemical testing daily 2323 (Note D) 2324 • 2325 Endotoxin testing 2326 daily 2327 • 2328 Microbiological testing 2329 daily (Note E) 2330 • 2331 Conductivity and TOC 2332 daily or on-line (Note 2333 C) 2334 • 2335 Chemical testing daily 2336 (Note D) 2337 • 2338 Endotoxin testing 2339 daily 2340 • 2341 Microbiological testing 2342 daily (Note E) 2343 • 2344 Conductivity and TOC 2345 at regular intervals 2346 or on-line (Note C), 2347 typically at least once 2348 per week 2349 • 2350 Chemical (Note D), 2351 endo-toxin (Note B), 2352 and microbiological 2353 (Note B, E) testing 2354 at regular intervals, 2355 typically at least once 2356 per week 2357 2358 This Document is licensed to 2359 ISPE Good Practice Guide: 2360 Page 37 2361 Sampling for Pharmaceutical Water, Steam, and Process Gases 2362 Table 2.4: Examples of WFI Distribution System Water Testing and Frequencies (continued) 2363 WFI Distribution System 2364 (Sample Location) 2365 Commissioning Tests 2366 and Frequencies 2367 Initial Phase Tests and 2368 Frequencies 2369 Intermediary Phase 2370 Tests and Frequencies 2371 (Note B) 2372 Extended Phase Tests 2373 and Frequencies (Note 2374 A, B) 2375 Non-Critical Use 2376 Point (Note G) 2377 Confirm conductivity, 2378 chemical (Note D), 2379 TOC, endotoxin, and 2380 microbiological levels 2381 • 2382 Conductivity and TOC 2383 daily or on-line (Note 2384 C) 2385 • 2386 Chemical testing daily 2387 (Note D) 2388 • 2389 Endotoxin testing 2390 daily 2391 • 2392 Microbiological testing 2393 daily (Note E) 2394 • 2395 Conductivity and 2396 TOC at a frequency 2397 between daily and 2398 weekly or on-line 2399 (Note C) 2400 • 2401 Chemical testing at 2402 a frequency between 2403 daily and weekly 2404 (Note D) 2405 • 2406 Endotoxin testing at 2407 a frequency between 2408 daily and weekly 2409 (Note B) 2410 • 2411 Microbiological 2412 testing at a frequency 2413 between daily and 2414 weekly (Note B, E) 2415 • 2416 Conductivity and TOC 2417 at regular intervals 2418 or on-line (Note C), 2419 typically at least once 2420 per month 2421 • 2422 Chemical (Note D), 2423 endo-toxin (Note B), 2424 and micro-biological 2425 (Note B, E) test-ing 2426 at regular intervals, 2427 typically at least once 2428 per month 2429 End of loop Sample 2430 Port or last POU 2431 (Note H) 2432 Confirm conductivity, 2433 chemical (Note D), 2434 endotoxin, TOC, and 2435 microbiological levels 2436 Confirm any other 2437 physical attributes such 2438 as flow rate, temperature, 2439 pressure, etc. as defined 2440 in a User Requirement 2441 Specification 2442 • 2443 Conductivity and TOC 2444 daily or on-line (Note 2445 C) 2446 • 2447 Chemical testing daily 2448 (Note D) 2449 • 2450 Microbiological testing 2451 daily (Note E) 2452 • 2453 Conductivity and 2454 TOC at a frequency 2455 between daily and 2456 weekly or on-line 2457 (Note C) 2458 • 2459 Chemical testing at 2460 a frequency between 2461 daily and weekly 2462 (Note D) 2463 • 2464 Endotoxin testing at 2465 a frequency between 2466 daily and weekly 2467 (Note B) 2468 • 2469 Microbiological 2470 testing at a frequency 2471 between daily and 2472 weekly (Note B, E) 2473 • 2474 Conductivity and TOC 2475 at regular intervals 2476 or on-line (Note C), 2477 typically at least once 2478 per week 2479 • 2480 Chemical (Note D), 2481 endotoxin (Note B), 2482 and microbiological 2483 (Note B, E) testing 2484 at regular intervals, 2485 typically at least once 2486 per week 2487 Note A: The frequency and extent of sampling during the extended phase of the performance qualification typically serves as the initial frequency 2488 and extent for future validation maintenance testing, but is subject to adjustment based on a review of accumulated data and the use of risk 2489 analysis tools. 2490 Note B: For water for injection systems, samples should be taken daily from a minimum of one POU, with all points of use tested weekly.4 2491 Note C: Verification of the equivalency of on-line readings to those obtained from sampling should be established in order to utilize on-line readings 2492 as suitable for quality control release. 2493 Note D: Chemical testing may or may not be required as determined by applicable pharmacopoeias. 2494 Note E: Microbiological identification testing is performed to provide a profile of the resident micro flora within the water system. Only those 2495 isolated from the microbiological samples taken at points of use that are visually distinct (i.e., physical colony characteristics) should be identified 2496 to minimize redundancy. Microbial identification is not required where microorganisms with identical colony characteristics have been previously 2497 identified. 2498 Note F: A critical use point is defined as a point providing water that is either an ingredient or has direct product contact or one that is used for final 2499 rinse. 2500 Note G: A non-critical use point may be a use point that is no longer used for manufacturing and not planned for removal, one that is not currently 2501 used for manufacturing but may be in the future, or a point where the water from the point has no direct product contact such as sanitant 2502 preparation in a clean area. 2503 Note H: There may be additional equipment such as heat exchangers, venturi, etc. between the last POU and the storage tank. In these cases, it 2504 may also be appropriate to monitor each of these unit operations before the water is returned to the storage tank. 2505 4 1993 US FDA “Guide to the Inspection of High Purity Water Systems” [1] 2506 2507 This Document is licensed to 2508 Page 38 2509 ISPE Good Practice Guide: 2510 2511 Sampling for Pharmaceutical Water, Steam, and Process Gases 2512 It may be possible to spend more time/sampling effort during commissioning and pre-validation in order to gain 2513 improved process knowledge and thereby execute initial and intermediary sampling each at the lower end of the 2514 recommended time frame. 2515 The 1993 US FDA “Guide to the Inspection of High Purity Water Systems” [1] describes the suggested sampling 2516 frequencies and durations; however, the total period required for the actual PQ should be based on a documented 2517 risk evaluation with rationales for the decisions, but for not less than one year. 2518 The FDA guide describes the need for validation of all steps involved in purification, particularly for new purification 2519 systems, due to the lack of historical data. Data from the validation will provide baseline information which can be 2520 used to evaluate operational trending changes, which in turn can provide operational details including that the system 2521 may be in need of maintenance. 2522 POU sampling is expected; however, some POUs cannot be sampled for practical reasons or physical limitations, 2523 based on the POU design and/or installation. In the cases where POU sampling is not possible, a representative 2524 sample point may be substituted instead of the actual POU, however rationales for any substitute sampling sites 2525 should be established before start of the PQ. 2526 When POU valves, which may be inaccessible for sampling once permanently connected, are initially installed these 2527 POU valves may be included in the initial sampling and compared with data from the substitute sampling points for 2528 confirmation of suitability. 2529 2.3.5.7 Performance Qualification Reporting 2530 At the conclusion of the PQ intermediary sampling, a summary report should be prepared that includes 2531 recommendations for the PQ extended sampling plan (sampling sites, parameters and frequencies). 2532 At the conclusion of the PQ sampling (all portions) the final report, including any recommendations, should be made 2533 available to the participants of the PQ as well as to current and future users of the pharmaceutical water system, as this 2534 information may provide useful input regarding their daily use of the water and the associated SOPs, instructions, etc. 2535 At least annually, as part of the PSE, data that has been generated from the sampling of a water system should be 2536 reviewed. This review affords the opportunity to see seasonal trends, make changes to the sampling frequency (i.e., 2537 daily to weekly or weekly to monthly) based on the accumulated data, and adjust Alert Levels and/or Action Levels 2538 previously established [13]. 2539 2.4 2540 Sample Valve Design 2541 Good sample valve design should not allow contamination of the sample itself and should take into account both the 2542 internal function and external operation of the valve. Design considerations should include the method of connection, 2543 the length of the flow path, water retention within the valve body upstream of the seat/seal, and the amount of water 2544 held within the exit path after the valve has been closed. This water downstream of the valve is either trapped or held 2545 by capillary properties of water. Valve outlet configuration should allow for drainage while not obstructing the sampling 2546 process. Good sample valve outlet fitting design should not be so constructed as to facilitate water holdup in the 2547 valve resulting in microbial growth. 2548 Any hardware added after the sample valve often poses a significant risk of contaminating samples by introducing 2549 extrinsic contaminants that are not representative of the actual water produced and delivered by the system. 2550 Sample valves should also be suitable for use in the selected system such that they should be appropriate for 2551 the pressure, temperature, and flow of the system. Sample valves should also be selected based on the location 2552 within the system. For example, sample valves in the feed water, pretreatment, and even in sections of the primary 2553 treatment may not need to meet sanitary requirements. However, if used for microbial sampling, these valves should 2554 not allow contamination of the sample. 2555 2556 This Document is licensed to 2557 ISPE Good Practice Guide: 2558 Page 39 2559 Sampling for Pharmaceutical Water, Steam, and Process Gases 2560 Sample valves should be constructed of acceptable materials, have a suitable finish, and have a system compatible 2561 method of mounting/connection. Valves in the non-sanitary portion of the system may be glued, threaded, or 2562 otherwise installed provided they do not contaminate or otherwise compromise the sample. If valves with industrial 2563 connections (i.e., threads, flanges, etc.) are used for chemical sampling, they are not likely to cause concern. 2564 However, these same valves will likely contaminate microbial samples and should be avoided when sampling for 2565 bacteria. 2566 Sample valves designed to sanitary requirements should be employed when feasible and appropriate. Sanitary 2567 diaphragm valves are a well-accepted standard within the pharmaceutical industry although other designs have been 2568 successfully employed. Ball valves and other non-sanitary valves should be avoided and only implemented when 2569 design constraints (high pressure, the requirement to meter flow rate, or other constraints) necessitate their use. In 2570 these cases, additional microbial control consideration may be required. Sanitary diaphragm valves may be the most 2571 robust and functional but offer less in the way of options often motivating users to consider alternate designs. 2572 The most common connection for sample valves in the sanitary portion of the system is via sanitary clamp, regardless 2573 of whether used for chemical or microbial sampling, although some may be welded or occasionally mounted via other 2574 specialty fitting. A variety of different configurations are available for the outlet of the sample valve, but valves should 2575 be selected to be compatible with the sampling procedures and methods planned. Some sample valves have been 2576 designed to allow for the introduction and entrapment of alcohol or other sanitant to create an environment that is 2577 hostile to microbial growth. 2578 The use of sanitary components becomes more critical after the final purification process of the pharmaceutical water 2579 generation system. Within the distribution system where the compendial water claim is established, sanitary design is 2580 required. 2581 When sampling for quality control purposes, it is usually accepted that regulators prefer samples are taken from 2582 the actual POU where the water is delivered to the process. In instances where a use point is suitably hard piped 2583 as close as feasible to a process or vessel, a sample point should be incorporated as close to the actual POU as 2584 possible. If sampling between the outlet of the use point valve and equipment inlet is not possible, one option would 2585 be to incorporate a sample valve into the use point valve body upstream of the diaphragm, although readers were 2586 dissuaded from sampling for quality control purposes at this location earlier in the chapter. Care should be taken 2587 in the placement of all sample valves to ensure the location supports sampling while minimizing the opportunity for 2588 contamination, as well as addressing ergonomic and safety issues. 2589 A significant difficulty associated with the FDA preference for sampling at the use point is the likelihood that process 2590 piping sizes and/or pressures, and their connections, may be considerably larger than the ideal size for sampling. 2591 This creates the dilemma of whether to sample from the actual larger process connection or to make an adaption for 2592 sampling purposes. Since neither is ideal, the addition of an adapter can aid the sampling process provided it does 2593 not compromise the sample taken. 2594 Sample valves are available in straight through and angle body configurations. Selection of the most suitable design 2595 should be by individual location. Correct valve design at each location is critical so as to avoid the need to make 2596 modifications at a later time. 2597 There have been many observed cases of sample valves located in ways that make sampling impossible because 2598 sample containers cannot access the port, or because the outlet points upward. Numerous cases of sample valves 2599 creating safety hazards to samplers have been documented. Some valves may be more likely to retain water in the 2600 outlet so it may be prudent to purchase and test a particular valve design prior to installing it system wide. 2601 Sample valve design should also take into account other issues such as water temperature and sample location 2602 within the pipe (near the sidewall or midstream). In heated systems, valve handles made of stainless steel will often 2603 conduct heat making the valve handle too hot to operate. In this case alternate valve handle materials, such as 2604 plastics that can tolerate elevated temperatures, should be considered. Valve seals, such as “O” rings should also be 2605 suitable for the intended application. 2606 2607 This Document is licensed to 2608 Page 40 2609 ISPE Good Practice Guide: 2610 2611 Sampling for Pharmaceutical Water, Steam, and Process Gases 2612 Valve design should take into account both the internal and external issues related to sampling. Valves mounted 2613 onto tee fittings should always utilize short outlet fittings or fittings that incorporate a valve mount flush to the interior 2614 wall or into the flow mainstream. When it is desirable to take a sample further into the water stream, as in larger tube 2615 sizes, valves with an integral extension that protrudes into the process stream may be considered. These valves also 2616 pose a greater difficulty for cleaning so careful evaluation is critical to success. 2617 Valve configurations vary widely and are available with alternate outlet connections ranging from hose barb, to 2618 sanitary clamp, to plain end, and are available in a variety of sizes. Valve inlet fittings, available in welded or clamp, 2619 can be provided to allow connection to 1/2, 3/4, 1, 1 1/2, and 2 inch OD supply line sizes (and may be ordered in 2620 larger sizes when required). Inlet fittings are also available eccentrically mounted to facilitate draining when used for 2621 dual purposes. Common configurations include: 2622 a. 2623 1/2" clamp inlet with 1/2" barb outlet 2624 b. 2625 1/2" clamp inlet with 1/4" barb outlet 2626 c. 2627 3/4" to 2" clamp inlet with 1/2" or 1/4" barb outlet 2628 d. 2629 1/2" clamp inlet with barbed outlet with separate or onboard port for sanitant injection 2630 When sampling hot water and clean/pure steam it is often necessary to reduce the temperature for safety reasons or 2631 based on compatibility with sample containers, media, etc. Cooling the water or condensing the steam reduces safety 2632 concerns and simplifies sampling but adds an inherent risk of microbial contamination. Hence, sampling through 2633 suitable heat exchange equipment using methods that minimize risk are appropriate and highly recommended. 2634 Equipment should be suitably sized to help to ensure successful operation and maintaining that equipment so as to 2635 not add contamination is paramount. 2636 2.5 2637 Sampling Techniques 2638 2.5.1 2639 Obtaining Representative Samples 2640 The goal of any sampling program is to obtain samples that accurately reflect the contents of the larger system and/ 2641 or to ensure water of suitable quality is used for manufacturing or other designated purposes. The vast majority of 2642 chemical constituents are uniformly distributed and relatively easy to sample with a minimal amount of rinsing and 2643 flushing. Sampling technique, including flushing, that is representative of use should be performed, as bacterial and 2644 endotoxin contaminants will not be uniformly distributed because bacteria primarily adhere to surfaces in a system. 2645 The 1993 US FDA “Guide to the Inspection of High Purity Water Systems” [1] states that WFI should essentially be 2646 sterile and that the allowance of bacteria counts of 10 cfu per 100 ml are to account for sampling error. 2647 2.5.2 2648 Use of Appropriate Sample Containers 2649 When taking samples, efforts should be taken to ensure that sampling containers do not introduce measurable 2650 amounts of the contaminant being monitored to the sample. This implies that several different sample containers 2651 may be required to sample any given location. For example, sampling for TOC should be done in certified low TOC 2652 containers since many containers will introduce extractables that contribute to the TOC value of a sample and 2653 increase the results. On the other hand, bacteria samples should be taken using pre-sterilized containers, endotoxin 2654 samples should be taken in depyrogenated or pyrogen free containers, and testing for conductivity should be 2655 accomplished using plastic containers. 2656 Sample container materials of construction should meet the same requirements as the system being sampled, 2657 specifically that they be non-additive, non-reactive, and non-absorptive. Sample containers should allow samples to be 2658 collected in an aseptic manner and minimize splashing and aeration upon collection. They should provide an adequate 2659 sealing mechanism and provide a means to record pertinent sample information. Sample information may include, but 2660 is not limited to, sample location/ID, date and time of acquisition, name of sampler, and testing to be performed. 2661 2662 This Document is licensed to 2663 ISPE Good Practice Guide: 2664 Page 41 2665 Sampling for Pharmaceutical Water, Steam, and Process Gases 2666 The collected sample should be able to quantitatively reflect the same attributes present in the water at that sampling 2667 location. Attributes and their levels in the sampled water should not substantially change during confinement and 2668 storage in the sample container. The selection of container and closure materials should be carefully considered, as 2669 the trace levels of the pertinent attributes in the water can be changed substantially by anything the water touches, 2670 including the air. 2671 The suitability of materials of container construction is determined by the attributes that will be tested. The walls of 2672 the container and the closure seal can positively contribute to the attribute of interest through leaching or desorption, 2673 increasing the attribute concentration to a higher value, or to a detectable level where it was formerly undetectable. 2674 The attributes of interest can also be reduced in concentration by when they are adsorbed to the internal surfaces of 2675 the container. Adsorption is typically a contact time phenomena based on the length of time the water is held in the 2676 container. 2677 When the attribute of interest is conductivity, the container should not substantially leach conductive ions. Therefore, 2678 clean plastic containers represent the best choice. Glass tends to leach sodium and silicate ions from its surface, so 2679 it is usually not a good choice except for samples having only a brief residence time within the container. Cleanliness 2680 should be considered, as residues from past uses, including soap residues, can have a significant and detrimental 2681 impact on the test result. Another potential impact on the sample’s conductivity is atmospheric carbon dioxide (CO2). 2682 It is almost impossible to prevent at least some CO2 intrusion during the sampling process, even if the container is 2683 filled smoothly and rapidly to the brim to minimize any trapped air headspace. A water sample will almost always give 2684 a higher conductivity than the on-line conductivity readings from the system at the time the sample was collected. The 2685 number of intruding ions from the dissolved CO2 gas and bicarbonate ions (HCO3 2686 -) to which it equilibrates may not be 2687 sufficient to cause an OOS value, but there can be sufficient conductivity generated to mask the conductivity trends 2688 of the system, making performance trending and process control much more difficult from sampled conductivity 2689 measurements. Conductivity based process control is best achieved using on-line conductivity or resistivity 2690 instrumentation. 2691 When the attribute of interest is TOC, the container surfaces should not leach substantial amounts of organic 2692 residues. The container of choice may be glass or plastic provided that they are certified low TOC. Many plastics 2693 leach sufficient organics to substantially affect the final test result. However, any container can leach organic residues 2694 associated with even the most careful handling and scrupulous cleaning. When selecting containers for TOC testing, 2695 it is best to use low TOC containers that are compatible with the TOC instrument’s sampling/staging mechanisms. 2696 Reusing these containers for TOC testing can pose challenges because of organic residues deposited on the internal 2697 container surfaces by the cleaning and drying practices. For example, these containers should never be exposed 2698 to cleaning soaps nor should they ever be dried by placing on a drying peg board or absorptive toweling. Even 2699 imperceptible organic vapors in the laboratory and glassware washing area are sufficient to compromise the trace 2700 quantitative analyses that TOC measurements are. It may be difficult to achieve the necessary container cleanliness 2701 to be suitable for TOC testing, so using specially cleaned low TOC containers from a commercial vendor is generally 2702 a recommended practice. 2703 Note: no sampling container can guarantee zero TOC leachables. As a general rule, the level of container 2704 cleanliness should be commensurate with the expected TOC levels in the sample. TOC readings taken using on-line 2705 instrumentation will always be lower than those obtained by sampling using containers. 2706 The materials of construction for microbial sampling containers have few functional restrictions. A container 2707 constructed of any material that can be sterilized is generally adequate. However, the bacteria in water samples tend 2708 to be surface seeking – the first step in biofilm formation. The rate of bacterial adsorption to the container walls could 2709 theoretically be affected by the chemical and topographic nature of the container surface. The microbial adsorptive 2710 properties of hydrophilic (glass) versus hydrophobic (plastic) and smooth versus rough surface differences are usually 2711 imperceptible over the typical elapsed period of time between sampling and testing, but there may be exceptions. For 2712 instance, the use of pre-sterilized plastic tissue culture bottles that have a coating to encourage tissue cell adhesion 2713 and proliferation is not recommended since it is hypothetically possible that this coating would encourage microbial 2714 cell adhesion as well as proliferation. These may not be labeled substantially differently from bottles without such a 2715 coating. 2716 2717 This Document is licensed to 2718 Page 42 2719 ISPE Good Practice Guide: 2720 2721 Sampling for Pharmaceutical Water, Steam, and Process Gases 2722 In microbial sampling, reused container cleanliness should be considered. Any growth promoting residues from the 2723 previous uses of the container could promote growth of bacteria collected in the water sample, thus affecting the 2724 actual microbial counts. Microbial sample containers should be clean and free of past use residues. 2725 The sample container material of choice for endotoxin testing is glass (or a suitably evaluated plastic alternative) 2726 because glass can be definitively depyrogenated by heat treating to 250°C or higher for at least 30 minutes. Plastic 2727 containers are not generally recommended for endotoxin sampling since the hydrophobic surface could encourage 2728 micelle adsorption to the surface resulting in removal of endotoxin from the bulk fluid phase. Such containers can 2729 only be depyrogenated by exhaustive rinsing and/or chemical processes which are usually not feasible for a user to 2730 implement, nor are they reliably validatable. Sterile disposable plastic labware available for microbial water sampling 2731 should be pyrogen-free; however, they should be used for water sampling for endotoxin testing with caution as it may 2732 not be possible to establish their pyrogen free nature. Endotoxin surface adsorption phenomenon is also possible 2733 with prolonged storage. 2734 Many other attributes could be tested that might be critical to the suitability of the water for use as well as the 2735 performance of a purification process. Examples include, but are not limited to the following: 2736 • 2737 Particulates 2738 • 2739 Specific heavy metal ions or other elemental impurities 2740 • 2741 Specific microbial contaminants 2742 • 2743 Trace sanitizer residues 2744 • 2745 Nitrates 2746 • 2747 Ammonia 2748 • 2749 Hardness 2750 • 2751 Free and/or total chlorine 2752 • 2753 Sodium 2754 • 2755 Silica 2756 • 2757 Turbidity 2758 • 2759 SDI 2760 • 2761 Other EPA Drinking Water Regulation contaminants 2762 2.5.3 2763 When Hoses Are Involved 2764 When the sample port is located in a tough to reach spot, hoses may be employed to make sampling easier. When 2765 sampling from a POU for quality control purposes, sampling should accurately reflect the way that water is removed 2766 from the system during actual production use using the actual manufacturing equipment employed (including hoses). 2767 Regardless of the purpose of sampling, it should be understood that while hoses may be utilized, they provide an 2768 opportunity to introduce contamination to the samples. For example, if a hose is not disconnected and allowed to fully 2769 drain and dry between uses, bacteria may grow inside the hose that would introduce contamination to the sample. 2770 Hoses used for sampling for process control or other purposes should be installed immediately before samples are 2771 taken and then removed immediately afterward or within a qualified period. 2772 2773 This Document is licensed to 2774 ISPE Good Practice Guide: 2775 Page 43 2776 Sampling for Pharmaceutical Water, Steam, and Process Gases 2777 2.5.4 2778 Training 2779 Personnel taking samples should be trained on proper sampling technique. The vast majority of variations in results 2780 can be traced back to variations in how samples are taken. Ideally, one person would be designated to take all samples 2781 and their technique would be regularly reviewed. Typically, this is not practical and several samplers are typically 2782 involved. Without rigorous training and continual review, variations in water quality data may reflect variations in 2783 sampling technique rather than actual variations in water quality. This introduces increased variability to collected data. 2784 2.5.5 2785 Use of Common Sampling Points for Process Control and Quality Control 2786 Sampling for process control purposes attempts to obtain samples of water that accurately reflect what is flowing 2787 through the pipe. Sampling for quality control purposes is focused on accurately reflecting how the water is used in 2788 production in order to duplicate the quality of water used during manufacturing. In some cases, a single sampling 2789 location may be used for both process and quality control purposes. 2790 2.5.6 2791 Sampling from Sinks 2792 Sink use point sampling should represent the actual use of the sink use point. Sampling should match the procedural 2793 use of the sink, including any flushing requirements and attachments or hoses used to direct the water. 2794 2.5.7 2795 Sample Point Preparation Before Taking Samples 2796 Samplers should understand that preparation of the sample point for one type of analysis may introduce extrinsic 2797 contamination for a subsequent sample taken minutes later from the same point. The order of sampling should be 2798 clearly defined and the reasons for this order should be clearly understood by the sampler. Sample point preparation 2799 may involve flushing, spraying, disinfecting, or other treatments. For example, spraying and disinfecting a sample 2800 point with isopropyl alcohol (IPA) prior to taking a sample for bacterial testing could increase the value of a sample 2801 taken subsequently for TOC measurement. 2802 Sample point preparation is usually performed in order to permit taking an accurate bacteria and endotoxin sample. 2803 Samples for bacteria and endotoxins should be taken last in the sequence and samples for chemical constituents, 2804 metals, and TOC should be taken first. 2805 2.5.8 2806 Sampling Technique 2807 Sampling technique is dictated by the purpose of sampling. If sampling for process control or diagnostic purposes, 2808 any sample point contamination should be thoroughly flushed away so that the sample accurately represents the 2809 water flowing inside the piping system. If sampling for quality control purposes or to meet regulatory requirements 2810 or consensus standards, the goal is to sample in a way that exactly replicates the way that water is used for 2811 manufacturing. 2812 When sampling for process control or diagnostic purposes, several different types of valves could be encountered and 2813 may necessitate an alteration of the sample process. An example sampling technique consists of the following steps 2814 1. 2815 Locate a proper size receptacle or drain to receive the water being flushed from the valve to be sampled. 2816 The sample valve flushing procedure should not be modified to fit an available receptacle (make sure that the 2817 receptacle is large enough to receive the appropriate flush) 2818 2. 2819 Slowly open the valve until it is fully open (e.g., or meets 8 FPS) 2820 3. 2821 Allow water to flow at full force from the valve for at least 30 seconds then slowly close the valve to reduce the 2822 flow to a rate that is comfortable for taking the type and number of samples required. 2823 4. 2824 Collect samples for any required non-microbial constituents (e.g., conductivity, TOC) in appropriate containers. 2825 2826 This Document is licensed to 2827 Page 44 2828 ISPE Good Practice Guide: 2829 2830 Sampling for Pharmaceutical Water, Steam, and Process Gases 2831 5. 2832 If the sample valve is to be sprayed with IPA for bacteria and/or endotoxin samples, slowly close the sample valve. 2833 6. 2834 Spray the outlet of the sample valve with disinfectant. 2835 7. 2836 Wait 30 seconds 2837 8. 2838 Repeat steps 2 to 3 and collect bacteria and endotoxin samples in appropriate containers. 2839 9. 2840 Close sample valve and post treat if appropriate to remove or displace any retained water in the outlet flow path. 2841 10. Ensure samples are labeled and handled/processed properly. 2842 Alternatively, when sampling for quality control purposes, the goal is to sample in a way that exactly replicates the 2843 way that water is used for manufacturing (e.g., with manufacturing’s hoses, pre-use flushing rate and duration, SIP, 2844 sample water temperature). Water should be sampled as close as physically possible to the process equipment. 2845 As a general rule, and to minimize contamination during production water use, it is recommended that water be 2846 flushed (e.g., for at least 30 seconds at a velocity of at least 8 feet per second or an alternative validated technique) 2847 through the outlet and connectors to ensure that bacteria are flushed from the walls of the hose, discharge piping 2848 and outlet valve prior to use. Repeatability of flushing (both time and rate) should be achievable for water quality and 2849 data consistency. Unless the manufacturing use point valve is sprayed with a disinfectant such as IPA prior to use/ 2850 manufacturing, that valve should not be sprayed with IPA prior to sampling, and similarly for flushing. 2851 2.5.9 2852 Sample Point Treatment After Taking Samples 2853 If the sample point is used to collect samples for bacteria or for endotoxin testing, treating the sample valve after 2854 sampling should be performed, as water can become trapped downstream of the sample valve, creating an exposed 2855 location where bacteria and biofilm can grow before the next sample is taken from the sample point. 2856 Once all microbial related samples have been taken, the downstream portion of the valve and the outlet path should 2857 be dried using alcohol or another suitable disinfectant (compatible with materials of contact and appropriate for use). 2858 The disinfectant would be injected or sprayed into and onto the sample valve and allowed to evaporate. Alternatively, 2859 special valves that allow for sanitant to be charged into the valve and then capped retaining the sanitant may be 2860 considered. In closed systems, typical of many POU installations, sterile compressed air may be utilized to remove 2861 residual moisture downstream of a sampling point. 2862 The capping of wet sample points after sampling should be avoided because it traps water and allows bacteria and 2863 biofilm growth between samples. Expelling water from the outlet of the sample valve by flooding the valve outlet with 2864 alcohol following sampling and/or before capping is recommended because it excludes water and prevents microbial 2865 growth in the sample valve between samples. Whenever this practice is utilized, thorough flushing of the sample 2866 valve during subsequent sampling is required to prevent alcohol from impacting TOC or microbial results. 2867 2.6 2868 Handling of Samples 2869 If sample analysis is being performed by an in-line or on-line instrument or analyzer, then the sample fluid never 2870 leaves the closed confines of the water system en route to analysis. The analyzed water essentially never 2871 experiences the chemical and microbial contamination hazards introduced by manual sampling from the water 2872 system and transport to the analytical area or off-line laboratory instrument.5 Where justifiable, based on the nature 2873 of the attribute to be tested and the location in the water system, having in-line or on-line instrumentation that can 2874 measure a homogeneously distributed attribute is generally preferable to collecting a sample and transporting it to an 2875 off-line analyzer in a remote laboratory. 2876 5 The only exception is if the fluid to be analyzed has an inordinately long or stagnated contact time within side stream tubing, a multichannel valve or a 2877 manifold on the way to an on-line analyzer. 2878 2879 This Document is licensed to 2880 ISPE Good Practice Guide: 2881 Page 45 2882 Sampling for Pharmaceutical Water, Steam, and Process Gases 2883 Sampling for off-line analysis in the laboratory can affect the measured quality of a given attribute, exposing it to 2884 external contamination by any of several routes: 2885 • 2886 Through manual sampling procedures that could be poorly performed due to negligence, lack of training or 2887 inadequate procedural detail 2888 • 2889 From transfer between the closed confines of the water system and the sample container 2890 • 2891 Through contact of the inside of the sample container with the outside surface of a sample port 2892 • 2893 During its transport between the sample location and the laboratory 2894 • 2895 During storage prior to analysis 2896 • 2897 Through sample container opening and re-closure in the laboratory 2898 • 2899 During transport and/or restorage should additional attribute tests need to be performed from the same container 2900 or for a later repeat test or investigative retest 2901 Significant handling involving human interventions into the sampling container may require extended contact time 2902 between the sample and the container, strongly suggests that issues associated with sample handling are addressed 2903 and resolved during the planning process and before samples are collected. The human issues may be more subtle 2904 than extractables or leachables from the container and can ultimately influence the results obtained from the sample. 2905 Regardless of the attributes of concern for a given sample, the sample collection, transport and storage processes 2906 share the same general concerns as the specific attributes. These sampling issues should not significantly affect the 2907 quantitative levels of the attributes over the samples elapsed time within the container or through handling. 2908 The transport of the samples to the laboratory for testing should be executed so that sampling container contents 2909 spillage or coolant seepage into samples does not occur. This may involve protective bagging of sample containers 2910 and/or the application of additional barriers (such as protective tape) to protect the sealing surfaces from leakage 2911 or seepage during the rigors of transport. In addition, depending on the attributes of concern for the sample and 2912 its intended testing destination, temperature control may also be a consideration if it is necessary to preserve 2913 the attributes of the sample during the transport period. This may involve the use of insulated containers, as well 2914 as the use of coolant materials such as waterless cold packs. The use of water and ice should be avoided since 2915 sloshing and seepage of this melting coolant into inadequately sealed sampling containers could occur. Similarly, 2916 the use of ultra-cold coolants should be avoided since the freezing of the sampled water in the container is usually 2917 contraindicated for most types of samples. The intent of the coolant is to bring the temperature of the samples to 2918 a reasonably controlled range, usually lower than the ambient conditions of the original sample. However, during 2919 winter months or air cargo shipments, the ambient temperature of the shipping process could freeze the samples, 2920 so protection from external low temperatures should also be factored into the temperature control process during 2921 transport. Once the sample is sealed into the container and its transport begins, its storage in that container also 2922 begins; prompt transport to the testing facility should be considered. 2923 The duration and temperature of sample storage is a hotly debated issue, especially for attributes that could 2924 quantitatively change during that storage period. For this reason, understanding the mechanisms behind those 2925 changes can help the user craft storage conditions that will least affect the attribute of concern. The choice of 2926 sampling container materials may have an impact on the rate of those changes. In general, the less time the sampled 2927 water spends in the container, the more closely it will approximate its original quality at the time of sampling. Relevant 2928 pharmacopoeias should be consulted, as applicable, for guidance. 2929 2930 This Document is licensed to 2931 Page 46 2932 ISPE Good Practice Guide: 2933 2934 Sampling for Pharmaceutical Water, Steam, and Process Gases 2935 Suggested sample handling practices are discussed in the following sections. If deviating from these suggested 2936 practices with longer storage times or different temperatures, an assessment should be performed to determine 2937 whether more extreme storage practices will have an impact on the samples. This may be considered a validation 2938 of the user’s practice. However, getting less than detectable values for both analyses does not validate the user’s 2939 storage conditions. Non-zero values are usually needed from both analyzed samples in order to confidently validate 2940 the user’s storage practice. If sample storage practices are within the guidelines established, then it may not be 2941 necessary to validate the practices. 2942 2.6.1 2943 Conductivity 2944 When a water sample for off-line conductivity testing is collected, container cleanliness and atmospheric CO2 are the 2945 two greatest risks to preserving the original conductivity of the sample. If single use, disposable containers are not 2946 used, then reused container cleanliness should be controlled and assured by routine procedures. Trace conductive 2947 residues can leach from glass container surfaces. The solubility of the leaching materials is relatively poor, so the 2948 leaching process could be relatively slow, but it is nevertheless a factor, particularly if glass containers are used and 2949 if the conductivity of ultrapure water is being assessed against tight specifications. Sample containers should be filled 2950 as much as practical to minimize air entrapment in the headspace which may allow additional atmospheric CO2 to 2951 impact the sample. 2952 Plastic containers tend to avoid the ionic leachables associated with glass, but these containers are moderately 2953 permeable to atmospheric CO2. Permeation is a slow process that occurs over time and has a detectable impact 2954 on the conductivity of the samples. However, atmospheric CO2 exposure during sampling collection will be the 2955 overwhelming influence on conductivity. 2956 As a best practice, samples for conductivity testing should not be refrigerated during storage while waiting testing, 2957 because it increases the time required to run the Stage 2 conductivity test. Sample testing should occur as soon as 2958 possible. 2959 2.6.2 2960 Total Organic Carbon (TOC) 2961 A significant impact on the TOC of a solution during sample storage can be from organic residues leaching from the 2962 container or closure surface. If reusable containers are used for TOC sampling, then the risk of leachable residue is 2963 much greater than if single use, low TOC sample containers are used. Refrigeration may slow this rate of leaching. If 2964 certified low TOC containers are used, only that certified amount of TOC will leach into the sample fluid no matter how 2965 long it is stored. However, prolonged storage can lead to alteration of the soluble TOC interaction with the container, 2966 hence extended storage should be supported by a risk analysis and supporting data. 2967 For TOC samples only, if certified low TOC containers are used, the samples should be suitable for up to 2 to 3 days 2968 at room temperature or when refrigerated without suffering any substantial increase or decrease in TOC level. If 2969 reused sample containers are employed, the amount of TOC leaching into the water is unknown as is the leaching 2970 rate, so testing these sample containers within a few hours after collection is advised. 2971 2.6.3 2972 Microbial Count 2973 Various maximum hold times and storage temperatures are recommended by different references. The concern is 2974 that either microbial proliferation or microbial loss due to adsorption or starvation could occur if the elapsed time after 2975 sample collection is more than a few hours. Generally, refrigeration is recommended if testing cannot commence 2976 within about 2 hours of sample collection. Refer to the USP <1231> [9] or other relevant compendia for the markets 2977 being served for guidance regarding refrigerated hold times. 2978 2.6.4 2979 Bacterial Endotoxin 2980 Because at least a portion of the endotoxin in a sample could be of a particulate nature, such as whole cells or cell 2981 wall fragments in a water sample, the same guidelines as for microbial testing should be followed for endotoxin test 2982 samples. 2983 2984 This Document is licensed to 2985 ISPE Good Practice Guide: 2986 Page 47 2987 Sampling for Pharmaceutical Water, Steam, and Process Gases 2988 2.6.5 2989 Other Attributes 2990 Storage conditions for a sample requiring other attribute tests are dependent on the specific needs of the attribute to 2991 maintain the integrity of its original quantitative values. Analyte susceptibilities may include: 2992 • 2993 Loss or increase due to container surface adsorption or leaching 2994 • 2995 Natural chemical or physical degradation 2996 • 2997 Microbially related processes 2998 These susceptibilities should be understood in order to establish appropriate sample storage conditions between 2999 sampling and testing. 3000 2.6.6 3001 Avoiding Extrinsic Contamination Through Closure Design 3002 When a sample is collected, the ease and cleanliness of opening and closing the sample container should be 3003 considered. This activity occurs multiple times between sampling and eventual analysis in the laboratory. The process 3004 of opening and closing the container during sampling has specific precautions that should be exercised to avoid 3005 extrinsic contamination, particularly with the handling of the removed closure and the potential splashing of the water 3006 while a sample is being collected. When a sample arrives at a laboratory for testing, it is opened and then closed for 3007 the removal of an aliquot of water. This presents different concerns than the sample collection process, especially if 3008 the container will be entered and reclosed multiple times for multiple tests. 3009 The simplicity of container opening and closing (including sanitary concerns, if for microbial testing) should be 3010 considered to help to minimize extrinsic contamination that could alter test results. Generally, sample containers have 3011 a lid or cap that can be easily removed; however, the placement of fingers near, or on, the sealing surfaces of the 3012 closure should be avoided to prevent a source of extrinsic sample contamination, even when the hands are gloved 3013 and alcohol sanitized. Ideally, fingers should be able to grasp and open the closure at a reasonable distance from any 3014 sealing surface to avoid inadvertent contact with the sealing surfaces or the inside of the container when opening and 3015 reclosing the container. 3016 Sealing surfaces of the container should not contact the potentially contaminated external surfaces of the container 3017 or closure during opening and closing, as may occur with folded over flexible containers. In the case of microbial 3018 contamination, a sanitized surface, such as gloved hands or the external container surface, should not be assumed 3019 to be sterile. A minimal level of extrinsic contamination in a test may be sufficient to cause concern, particularly where 3020 the typical microbial counts are zero. 3021 2.6.7 3022 Chain of Custody 3023 Maintaining a chain of custody provides documented proof of where the samples are and have been and under what 3024 conditions, at what times, and handled by whom. This may be a significant consideration when there is an excursion 3025 in the value of one or more attributes. The sample’s history could have a bearing on how to interpret the test results. 3026 Chain of custody information needs to be proactively collected on every sample in order to have a retrospective 3027 interpretation. 3028 Chain of custody may also be a consideration when expected samples are missing, perhaps allowing the samples 3029 to be found and to understand what conditions they have been exposed to, in order to help assess their suitability 3030 before potential testing. Chain of custody should also be considered for samples being sent to contract laboratories or 3031 sister sites within the same company for testing. 3032 If the test results for samples become included in any legal disputes, chain of custody should be able to verify that the 3033 samples are legitimate and have maintained their integrity. The user may not know this in advance. 3034 3035 This Document is licensed to 3036 Page 48 3037 ISPE Good Practice Guide: 3038 3039 Sampling for Pharmaceutical Water, Steam, and Process Gases 3040 2.7 3041 Parametric (Real Time) Release 3042 Releasing products in real time based on monitoring and control of manufacturing in process parameters critical 3043 to finished product quality has been a long-term goal for pharmaceutical water systems. This approach employs 3044 Process Analytical Technology (PAT) and Statistical Process Control (SPC) and should be properly validated. It has 3045 been successfully used for product manufacturing processes to eliminate or greatly reduce the need for finished 3046 product testing and usually involves on-line and in-line analytical technologies to provide timely in process data on 3047 critical in process parameters that could affect the finished product quality. Generally, these parameters should be 3048 continuously monitored, so that deviations from acceptable ranges can be detected and process adjustments made 3049 to maintain the acceptable finished product quality. Speed and timeliness of these analyses is crucial, which is why 3050 real time measurements are needed for real time product release. 3051 The traditional and untimely alternative for release of pharmaceutical waters is to collect water samples and send 3052 them to a lab for testing prior to release, all by human mediated procedures. Eliminating the delays and potential for 3053 data aberrations caused by human interventions and procedures is the goal of Real Time Release (RTR) for water. 3054 Although desirable, RTR may not be possible in all scenarios. 3055 2.7.1 3056 Chemical Attributes 3057 Since 1996, with the changes made with USP 23rd Edition, 5th Supplement [17], USP has facilitated the use of on-line 3058 monitoring of TOC (USP <643> [8]) and in-line monitoring of Conductivity (USP <645> [18]) for satisfying the water’s 3059 compliance with PW and water for injection monograph requirements. The only provision for on-line and in-line release 3060 in both test chapters is that the data from these on-line, in-line, and at-line instruments should be representative of 3061 those attributes as measured at the POU. If used solely for process control, there is no need to establish this POU 3062 equivalence. Chemical attributes are homogeneously distributed hence the values measured at remote locations in a 3063 simple water distribution system usually are identical to the values measured at the POUs, except for the unavoidable 3064 issues associated with sample contamination during collection and off-line testing. With this equivalence validated, the 3065 on-line and in-line readings will truly reflect the quality of those attributes in the water that is used. 3066 However, if opportunities for non-equivalence exist, due to complex system designs (with sub or parallel loops, heat 3067 exchangers, etc.), and potentially TOC contributing hardware, then additional sampling points may be required. 3068 2.7.2 3069 Microbiological Attributes 3070 Microbiological attributes of pharmaceutical water are not homogeneously distributed. Although microbiological 3071 quality is not a USP monograph requirement (but may be in other compendia), suggested action levels are furnished 3072 in USP <1231> [9] which reflect the original guidance established by the FDA for their inspectors in 1993 [1]. 3073 These microbiological quality levels are a regulatory expectation when these waters are used in pharmaceutical 3074 manufacturing with direct or indirect product contact. 3075 Microbial counts in the water being used from POUs could vary significantly in the distribution system, as planktonic 3076 or free floating microorganisms originate from biofilms and biofilms are localized in the piping or hardware between 3077 the water distribution system and the POU. This localized phenomenon can cause rapid on-line microbial testing, as 3078 well as conventional spot microbial testing from any system sampling port, to be an inaccurate representation of the 3079 microbial quality of the water that is being used for manufacturing. 3080 Similar concerns exist with bacterial endotoxin, as it is produced by the Gram-negative microorganisms that 3081 can inhabit the water distribution system, the water purification system, and may be present in the feed water. 3082 Consequently, the results of any rapid on-line and at-line endotoxin testing or conventional endotoxin testing from any 3083 system sample port does not necessarily reflect the quality of the water that is being used for manufacturing.6 3084 6 Note: because of the extremely high level of microorganisms required to produce a significant change in the level of endotoxin and the special 3085 conditions needed to release that endotoxin into the water, it is uncommon for a well-designed and maintained system to exhibit varying endotoxin 3086 levels at different POUs. 3087 3088 This Document is licensed to 3089 ISPE Good Practice Guide: 3090 Page 49 3091 Sampling for Pharmaceutical Water, Steam, and Process Gases 3092 2.7.3 3093 Grab Sampling versus On-Line and At-Line Monitoring 3094 Because of the potential for localized, outlet specific contamination, grab samples for microbial and endotoxin testing 3095 have been unequivocally accepted from POUs as reflective of the quality of the water being used at the POU. 3096 However, manual sampling and conventional testing approaches can lead to days of delay in data availability and 3097 potential for artificially influenced good or bad data. 3098 If rapid on-line or at-line technologies were installed for microbial and endotoxin testing at a single point in the 3099 distribution system (similar to conductivity and TOC) the resulting data may be valuable for process control purposes, 3100 but could not be used for product release purposes. POU testing is required for release. Alternatively, if microbial 3101 and endotoxin detection instruments were installed at all POUs, they may be compared with grab sample data and 3102 eventually validated for product release purposes. Portable instruments may represent another approach since they 3103 could be moved to different POUs, reducing the equipment costs. 3104 2.7.4 3105 Rapid Microbial Detection Techniques 3106 It is not the purpose of this Guide to review the growing list of available rapid detection techniques for quantifying 3107 microbial levels in water. However, one way of discriminating the various approaches is to categorize them as being 3108 either destructive or non-destructive. In the destructive analyses, the microbial cell is killed by the analysis in the 3109 process of producing the signal of its presence. When a bacterial cell is killed, it can no longer be speciated, which is 3110 usually of great importance for quality control or release testing. 3111 This analytical lethality is not important for process control purposes as the change in the microbial counts over time 3112 is the important variable. There are also times for quality control testing when this analytical lethality is unimportant, 3113 or when the finding of any microbial count triggers the execution of a conventional cultural plate count procedure 3114 using the same sample within a short time (for example, < 4 hours) of the original rapid test. In the non-destructive 3115 analyses, the microbial cells are not killed by the analysis, making their identification possible when the detected 3116 bacteria are allowed to develop into visible colonies by culturing. 3117 Rapid microbial detection techniques can also be discriminated by whether or not there is a need to amplify the 3118 number of each of the original colony forming units (CFUs) so that a detectable signal can be elicited. Usually, 3119 this amplification process merely involves a relatively short incubation period, typically ¼ to ½ the full incubation 3120 time of a conventional cultivation approach. This allows several multiplications of the original CFU to occur. Other 3121 rapid microbial detection techniques not needing amplification to detect the original microbial cells are generally 3122 much quicker. Some of these techniques kill the cells and some do not. A few of these approaches have on-line 3123 instrumentation versions, but at the time of publication, most of these instruments are laboratory based. 3124 In general, the various rapid microbiological methods still have some delay in data availability. The on-line 3125 technologies, though claiming to have instantaneous data availability, are actually a rolling summation of many hours 3126 of prior sampling caused by the limited side stream flow rate. The quickest techniques which can generate data, in 3127 perhaps 1 to 2 hours, usually involve destructive techniques. Many of the slower techniques are non-destructive, 3128 but take longer (1 to 3 days) for data availability. Current technologies do not give essentially instantaneous data 3129 akin to conductivity or TOC testing. Emerging technological advances have the potential to provide instantaneous 3130 microbiological data in the relatively near future. 3131 2.7.5 3132 Rapid Endotoxin Detection Techniques 3133 The detection of the miniscule levels of endotoxin, which are considered toxic, in WFI can currently only be 3134 accomplished by the molecular amplification process involved with the bioassay employing Limulus Amebocyte Lysate 3135 (LAL) reagents. Analogs and chemical modifications of these reagents have allowed endotoxin to be quantitated 3136 photometrically in as little as 15 minutes. The LAL technology has traditionally been laboratory based, but has now 3137 been commercialized as a rapid portable version that can be easily employed at-line if needed. This technique is grab 3138 sample based with the associated limitations, but it is sufficiently rapid and performable with minimal training to make it 3139 a near real time at-line test, rivaling the test result turnaround time of some on-line TOC instruments. 3140 3141 This Document is licensed to 3142 Page 50 3143 ISPE Good Practice Guide: 3144 3145 Sampling for Pharmaceutical Water, Steam, and Process Gases 3146 2.7.6 3147 Future of Real Time Release for Pharmaceutical Water 3148 The localized formation of biofilm between the distribution loop outlets and the POUs has consigned the definitive 3149 microbial testing of the water used by manufacturing to be from that POU where the water is delivered to a process 3150 or piece of equipment. Though the chemical water attributes are generally not different between a POU and an on- 3151 line instrument location, the microbial water attributes may be. From an on-line instrument perspective, and with the 3152 rapid advances occurring with instrumentation, real time release of the water for microbial attributes may become a 3153 more accepted practice in the relatively near future. For WFI systems where endotoxin levels are also an important 3154 attribute, the use of the portable at-line endotoxin test kit essentially adds endotoxin to the list of attributes that can be 3155 assured as passing in the finished water with an essentially real time test. 3156 Other rapid microbial methods are available that could be used on grab samples from points of use. These tests are 3157 fairly rapid (currently about an hour) before test results are available. Some methods are destructive, some are non- 3158 destructive, but when the test results show zero counts, the test’s destructive properties are considered irrelevant. 3159 When the test results show some counts and microbial identifications become important, only the longer duration 3160 rapid microbiological methods (1 to 3 days depending on medium) that allow microbial identification or conventional 3161 cultivation tests will suffice. A 1 to 3 day delay in result availability is not suitable for a parametric or real time release 3162 approach, but currently emerging technologies may make real time release a reality. 3163 3164 This Document is licensed to 3165 ISPE Good Practice Guide: 3166 Page 51 3167 Sampling for Pharmaceutical Water, Steam, and Process Gases 3168 3 Pharmaceutical Steam 3169 3.1 3170 Introduction to Pharmaceutical Steam 3171 Types and uses of pharmaceutical steam:7 3172 1. 3173 Plant steam 3174 2. 3175 Chemical free steam 3176 3. 3177 Process steam 3178 4. 3179 Pure steam 3180 The majority of discussion contained in this chapter will focus on pure steam, as pure steam is of the greatest 3181 concern in the pharmaceutical industry. 3182 3.2 3183 Generation and Distribution of Pharmaceutical Steam 3184 Table 3.1 (Table 7.1 from ISPE Baseline® Guide: Water and Steam Systems (Second Edition) [7]) represents the 3185 typical industry and baseline practices for the production of pharmaceutical steam. They include the commonly 3186 accepted generation methods used to meet regulatory requirements in the pharmaceutical industry. The table is not 3187 intended to be definitive or all inclusive. 3188 Table 3.1: Typical Industry and Baseline Practices for the Production of Pharmaceutical Steam 3189 7 Definitions are contained in the Glossary. 3190 Intended Use of Steam 3191 Method of Steam Generation or Steam Type 3192 Parenteral and non-parenteral dosage form applications where 3193 steam is in direct contact with the drug. 3194 Use of a Pure Steam Generator (PSG). 3195 Critical step in the manufacture of API where steam is in direct 3196 contact with the API. 3197 Use of a PSG. 3198 Non-critical step in the manufacture of an API where added 3199 impurities will be removed in a subsequent step. 3200 A PSG is commonly used; however, chemical-free steam may be 3201 acceptable. 3202 Sanitization or sterilization of a high purity water system. 3203 A PSG is commonly used; however; chemical-free steam may be 3204 acceptable, followed by adequate flushing. 3205 Humidification for dosage form production where steam is in direct 3206 contact with the drug, where open processing occurs, or where 3207 chemical additives may be detrimental to the drug product. 3208 Use of a PSG. 3209 Humidification of non-critical HVAC systems, such as rooms and 3210 areas where the drug product is not directly exposed to the room 3211 atmosphere. 3212 A PSG is commonly used; however; chemical free steam or plant 3213 steam may be acceptable. 3214 Humidification of critical process cleanrooms. 3215 Use of a PSG. 3216 Heat source for non-critical and cGMP heat exchangers. 3217 Chemical free steam or plant steam. 3218 Deactivation of solid or liquid biologic process waste. 3219 Use of PSG or chemical free steam in a dedicated deactivation 3220 vessel. 3221 Sterilization of direct product contact production equipment, 3222 process vessels, containers, or packaged product. 3223 Use of a PSG. 3224 3225 This Document is licensed to 3226 Page 52 3227 ISPE Good Practice Guide: 3228 3229 Sampling for Pharmaceutical Water, Steam, and Process Gases 3230 Feed water requires proper pretreatment, as all generators are susceptible to scaling and corrosion. Pretreatment 3231 includes: 3232 • 3233 Minimizing or preventing scale formation to minimize or prevent corrosion. 3234 • 3235 Removal of objectionable volatiles, such as ammonia that are not removed by distillation/steam generating 3236 processes, and would carry over into the clean steam. 3237 • 3238 Utilization of the facility’s high purity water system (e.g., PW, WFI). (Note: steam generators may not require feed 3239 water to meet PW or WFI requirements.) This practice, however, ignores the ability of the pure steam generator 3240 to remove impurities. 3241 Pure steam may also be referred to as clean steam, pyrogen free steam, water for injection (WFI) steam, or USP 3242 purified water (PW) steam. Regardless of the term used, when condensed, the steam condensate must meet 3243 requirements for USP/EP WFI. Pure steam is predominantly used for sterilization. Steam used for sterilization used in 3244 autoclaves for international manufacturing should also meet the requirements of British Standard (BS) EN 285:2015 3245 [19] for non-condensable gases, degrees of superheat, and dryness: 3246 • 3247 This steam is characterized as having no additives and limited generated superheat, except when the generated 3248 pressure is significantly higher than the use pressure of the steam. 3249 • 3250 Condensate of steam must meet requirements for USP/EP WFI, have no buffer, and have a relatively low pH 3251 compared to that of plant steam. 3252 For further information, see the ISPE Baseline® Guide: Water and Steam Systems (Second Edition) [7]. 3253 3.2.1 3254 Regulatory Requirements 3255 3.2.1.1 3256 USP Monographs 3257 Table 3.3 summarizes requirements from the USP monographs for the production of clean steam, which states that 3258 condensate from a clean steam system must meet USP WFI water quality requirements. 3259 The USP monographs also state that dryness and non-condensable gas requirements should be determined by 3260 application. Table 3.4 includes some of the standards associated with pure steam and sterilization applications. 3261 Table 3.2: USP Pure Steam Condensate Purity Requirements 3262 USP Monograph Limits for Pure Steam 3263 Attribute 3264 USP Chapter 3265 Notes 3266 Conductivity (µS/cm at 25°C (77°F) 3267 <645> 3268 1 3269 TOC (mg/l) 3270 <643> 3271 2 3272 Bacterial Endotoxins (EU/ml) 3273 <85> 3274 3 3275 Notes: 3276 1. Stage 1 limit of 1.3 µS/cm at 25°C (77°F), Stage 2 limit of 2.1 µS/cm at 25°C (77°F), Stage 3 measure pH and 3277 refer to monograph table 3278 2. Instrument response (Rs-Rw) to 0.50 mg/l standard 3279 3. < 0.25 EU/ml 3280 3281 This Document is licensed to 3282 ISPE Good Practice Guide: 3283 Page 53 3284 Sampling for Pharmaceutical Water, Steam, and Process Gases 3285 3.2.1.2 British Standard EN 285:2015 [19] 3286 Additional pretreatment requirements may be necessary if the pure steam is to meet the requirement of British 3287 Standard (BS) EN 285:2015, specifically for non-condensable gases. Generators may need to be fitted with a feed 3288 water degassifier (also called a deaerator/decarbonator) or heated break tank (accomplished through the use of hot 3289 WFI). Hot WFI or PW, meeting Stage 1 conductivity requirements that has not been nitrogen blanketed, should have 3290 low enough levels of non-condensable gases to pass the BS EN 285:2015 test. The BS EN 285:2015 requirements 3291 for feed water and clean steam condensate are listed in the Table 3.3. 3292 Table 3.3: Requirements for Feed Water and Clean Steam Condensate [19] 3293 Determinant 3294 Feed Water 3295 Clean Steam Condensate 3296 Residue on Evaporation 3297 ≤ 10 mg/l 3298 N/A 3299 Silicate (SiO2) 3300 ≤ 1 mg/l 3301 ≤ 0.1 mg/l 3302 Iron 3303 ≤ 0.2 mg/l 3304 ≤ 0.1 mg/l 3305 Cadmium 3306 ≤ 0.005 mg/l 3307 ≤ 0.005 mg/l 3308 Lead 3309 ≤ 0.05 mg/l 3310 ≤ 0.05 mg/l 3311 Balance of heavy metals, except 3312 iron, cadmium, lead 3313 ≤ 0.1 mg/l 3314 ≤ 0.1 mg/l 3315 Chloride (Cl) 3316 ≤ 2 mg/l 3317 ≤ 0.1 mg/l 3318 Phosphate (P2O5) 3319 ≤ 0.5 mg/l 3320 ≤ 0.5 mg/l 3321 Conductivity (at 25°C (77°F)) 3322 ≤ 5 µS/cm 3323 Passes Stage 3 3324 pH value (degree of acidity) 3325 5 to 7.5 3326 5 to 7 3327 Appearance 3328 Colorless, clean without sediment 3329 Colorless, clean without sediment 3330 Hardness (∑ Ions of alkaline earth) 3331 ≤ 0.02 mmole/l 3332 ≤ 0.02 mmole/l 3333 Note: compliance should be tested in accordance with acknowledged analytical methods. 3334 Table 3.3 suggested maximum contaminants in feed water supplied to a dedicated steam generator and suggested 3335 maximum contaminants in clean steam supplied to the sterilizer from BS EN 285:2015 Informative Annex [19]. 3336 Steam is produced in specially designed non-fired generators or from the first effect of multi effect WFI stills, which 3337 should not use scale or corrosion inhibitor additives. The feed water should be pretreated to remove elements that 3338 contribute to scaling or corrosion, and the materials of construction of the generators are resistant to corrosion by 3339 steam containing no corrosion inhibitors. 3340 3341 This Document is licensed to 3342 Page 54 3343 ISPE Good Practice Guide: 3344 3345 Sampling for Pharmaceutical Water, Steam, and Process Gases 3346 3.2.2 3347 Consensus Standards 3348 Table 3.4: Consensus Standards and Recommended Practice for Pure Steam 3349 Consensus Standard or 3350 Guideline 3351 Title 3352 Specifications or Recommended 3353 Practice 3354 ANSI/AAMI ST79 [20] 3355 Comprehensive guide to steam 3356 sterilization and sterility assurance in 3357 health care facilities 3358 This recommended practice covers 3359 steam sterilization in health care 3360 facilities. 3361 ASME BPE-2014 [21] 3362 Bioprocessing Equipment 3363 Design recommendations for clean 3364 steam distribution systems. 3365 BS EN 285:2015 [19] 3366 Sterilization, Steam sterilizers, Large 3367 sterilizers 3368 Specifies design requirements and 3369 tests for large steam sterilizers primarily 3370 used in health care or commercially. 3371 HTM 2031 [22] 3372 Clean Steam for Sterilization 3373 Guideline for the quality of steam for 3374 sterilization. 3375 ISO 13408-5:2006 [23] 3376 Aseptic processing of health care 3377 products – Part 5: Sterilization in place 3378 Specifies the general requirements for 3379 sterilization in place applied to product 3380 contact surfaces of the equipment 3381 used in manufacture of sterile health 3382 care products by aseptic processing 3383 and offers guidance on qualification, 3384 validation, operation and control. 3385 PDA Technical Report 1 [24] 3386 Validation of Moist Heat Sterilization 3387 Processes Cycle Design, Development, 3388 Qualification and Ongoing Control 3389 Guidance of the development and 3390 validation of steam sterilization cycles 3391 in industrial applications. 3392 ISO 17665-1:2006 [25] 3393 Sterilization of Health Care Products -- 3394 Moist Heat -- Part 1: Requirements for 3395 the development, validation and routine 3396 control of a sterilization process for 3397 medical devices 3398 Requirements for the use of moist heat 3399 in the sterilization process, including 3400 development, validation and routine 3401 control of the sterilization process. 3402 ISO 11140-4:2007 [26] 3403 Sterilization of Health Care Products 3404 -- Chemical Indicators -- Part 4: Class 3405 2 indicators as an alternative to the 3406 Bowie and Dick type test for detection 3407 of steam penetration 3408 Requirements for use of Class 2 3409 indicators as an alternative to the Bowie 3410 and Dick type test for steam sterilizers. 3411 3.2.3 3412 Distribution Systems 3413 The distribution piping systems used for all grades of steam should follow good engineering practices. Pure steam 3414 distribution systems have additional requirements for contact materials due to the aggressive nature of pure steam. 3415 Corrosion resistant 316 or 316L grade stainless steel tubing or solid drawn pipe is typically utilized in new systems. 3416 However, legacy systems may be constructed of 304 and 304L grade stainless steel. 3417 The surface finishes recommended for pure steam distribution systems are not as strict as for pure water distribution 3418 systems because of the high operating temperatures and self-sanitizing nature of pure steam. Pure steam distribution 3419 systems are generally subjected to risk analysis and may installed after consideration of the following guidelines: 3420 3421 This Document is licensed to 3422 ISPE Good Practice Guide: 3423 Page 55 3424 Sampling for Pharmaceutical Water, Steam, and Process Gases 3425 • 3426 Mill finish or mechanically polished pipe may be used, but typical mechanical polishes of 35 to 20Ra are utilized 3427 for clean steam generators and distribution systems. 3428 • 3429 Internal surfaces, including orbital and mechanical welds may be passivated to improve corrosion resistance. 3430 • 3431 Electropolishing of internal surfaces may be utilized but may not be necessary because of the high operating 3432 temperatures and an increase in the formation of rouge. 3433 For further information, see ISPE Baseline® Guide: Water and Steam Systems (Second Edition), Chapter 10 [7]. 3434 General design practices for pure steam distribution systems include: 3435 • 3436 Allowance for thermal expansion and the ability for condensate to drain effectively. 3437 • 3438 Sloping of piping between 1/8 inch/foot (10 mm/m) to 1/4 inch/foot (20 mm/m) as outlined in ASME BPE-2014 3439 [21], but good engineering practices and risk analysis tools may be used to determine the appropriate slope. 3440 • 3441 Sloping of piping in the direction of steam flow and to a low point condensate trap for adequate condensate 3442 removal.8 3443 Figure 3.1: Isometric of Typical Clean Steam System 3444 Reprinted from ASME BPE-2014 (Figure SD-4.2.2-1), by permission of The American Society of Mechanical 3445 Engineers. All rights reserved. www.asme.org. 3446 8 Sloping of distribution piping in the direction opposite of flow is not recommended because condensate may be easily reintroduced into the pure steam. 3447 3448 This Document is licensed to 3449 Page 56 3450 ISPE Good Practice Guide: 3451 3452 Sampling for Pharmaceutical Water, Steam, and Process Gases 3453 Figure 3.2: Clean Steam POU Design Guidance 3454 Reprinted from ASME BPE-2014 (Figure SD-4.2.2-2), by permission of The American Society of Mechanical 3455 Engineers. All rights reserved. www.asme.org. 3456 3.3 3457 Sampling Locations 3458 The product or facility requirements should be used to define the CQAs (e.g., dryness, chemical properties) for pure 3459 steam. Since different CQAs may require specific sampling equipment installed in multiple locations, several sampling 3460 locations may be required to effectively sample a pure steam system. 3461 3.3.1 3462 Critical Quality Attributes Define Sampling Locations 3463 The level of non-condensable gases and dryness are two commonly defined CQAs and are defined in British 3464 Standards EN 285:2015 [19]. 3465 3.3.1.1 3466 Non-condensable Gases 3467 If non-condensable gas removal is required, these gases may be removed in the pretreatment to a pure steam 3468 generation system either by applying heat or using a degassifier because these gases pass directly through a pure 3469 steam generator and into the pure steam. A pretreatment degassifier is shown in Figure 3.3. Including a degassifier at 3470 the beginning of a pure steam distribution system may be considered. The degassifier can be as simple as a tee with 3471 the vent leg pointing upward and terminating with a balanced pressure air vent that opens to expel NCGs from the 3472 pure steam as shown in Figure 3.4. 3473 3474 This Document is licensed to 3475 ISPE Good Practice Guide: 3476 Page 57 3477 Sampling for Pharmaceutical Water, Steam, and Process Gases 3478 Figure 3.3: Using Heat to Remove Non-condensable Gases from the Pretreatment to a Pure Steam Generator 3479 Courtesy of MECO, http://www.meco.com. 3480 Figure 3.4: Installation of a Balanced Pressure Air Vent (V) at a High Point to Remove Non-condensable 3481 Gases from a Clean Steam Distribution System 3482 Non-condensable gases should be sampled with a sample valve suitable for its intended purposes. Sampling for 3483 chemical and endotoxin purposes would use a different sample valve. Proper sampling for non-condensable gases is 3484 described in detail in BS EN 285:2015, Chapter 22.1 [19], and is shown in Figure 3.5. 3485 3486 This Document is licensed to 3487 Page 58 3488 ISPE Good Practice Guide: 3489 3490 Sampling for Pharmaceutical Water, Steam, and Process Gases 3491 Figure 3.5: Schematic Diagram of Sampling Apparatus Used to Sample Pure Steam for 3492 3493 3494 Non-condensable Gases 3495 Permission to reproduce extracts from British Standards (EN 285:2015, Figure 7) is granted by BSI. 3496 British Standards can be obtained in PDF or hard copy formats from the BSI online shop: www.bsigroup.com/Shop or by contacting BSI Customer 3497 Services for hardcopies only: Tel: +44 (0)20 8996 9001, Email: cservices@bsigroup.com. 3498 The calculation of the concentration of non-condensable gas as a percentage should be made using the equation: 3499 Cn = (Vb / Vc) (100%) 3500 Where: 3501 Cn = concentration of non-condensable gases, in % 3502 Vb = volume of water displaced from the burette, in ml 3503 Vc = volume of water collected in the graduated cylinder, in ml 3504 3505 This Document is licensed to 3506 ISPE Good Practice Guide: 3507 Page 59 3508 Sampling for Pharmaceutical Water, Steam, and Process Gases 3509 3.3.1.2 Dryness 3510 Dryness refers to the level of steam saturation and is the ratio of vapor mass to the mass of the steam mixture. 3511 Dryness is a dimensionless value, with a value of 1.0 being representative the ideal of dry saturated steam. Dry 3512 saturated steam may have excess heat introduced, which creates superheated steam. The amount of superheating 3513 should be minimized for efficient steam sterilization, since superheated steam is not saturated and will create uneven 3514 temperature distribution at the surface being sterilized. The dryness of the pure steam may be adversely impacted by 3515 the system design, with a few examples including: 3516 • 3517 Incorrectly sloped lines 3518 • 3519 Inadequately sloped lines 3520 • 3521 Inadequate condensate removal (e.g., insufficient traps, incorrectly located traps) 3522 • 3523 Insufficient pipe insulation resulting in excess condensation 3524 • 3525 Incorrect POU design 3526 The same location that is used for measuring non-condensable gases may be used for measuring the dryness of the 3527 steam. Dryness should be greater than 0.95 for metal loads and 0.90 for other loads according to BS EN 285:2015, 3528 Chapter 13.3.3 [19]. The method for measuring steam dryness is described in BS EN 285:2015, Chapter 22.2 [19], 3529 and is shown in Figures 3.6 and 3.7. 3530 For dryness sampling, the sample should be taken from the center of the steam distribution pipe. 3531 Note: this sample does not take into account the moisture present as a film on the pipe wall or the condensate on the 3532 bottom of the pipe. It is, therefore, not an accurate value, but represents only an approximation. 3533 However, this sampling technique has become the most commonly used method, so changing to more exact 3534 sampling methods would lose the linkage between the established accepted values and the obtained values. As the 3535 industry continues to change and advance, parallels will surely be generated between the current industry accepted 3536 practices and more advanced sampling techniques as they are developed. This sampling technique can be said to 3537 represent a well-designed take off as this will also not include any of the condensate on the walls of the pipe. 3538 Figure 3.6: Pitot Tube Design Used to Sample Steam for Dryness 3539 Permission to reproduce extracts from British Standards (EN 285:2015, Figure 8) is granted by BSI. 3540 British Standards can be obtained in PDF or hard copy formats from the BSI online shop: www.bsigroup.com/Shop or by contacting BSI Customer 3541 Services for hardcopies only: Tel: +44 (0)20 8996 9001, Email: cservices@bsigroup.com. 3542 3543 This Document is licensed to 3544 Page 60 3545 ISPE Good Practice Guide: 3546 3547 Sampling for Pharmaceutical Water, Steam, and Process Gases 3548 Figure 3.7: Sampling Apparatus for Dryness Measurements 3549 Permission to reproduce extracts from British Standards (EN 285:2015, Figure 9) is granted by BSI. 3550 British Standards can be obtained in PDF or hard copy formats from the BSI online shop: www.bsigroup.com/Shop or by contacting BSI Customer 3551 Services for hardcopies only: Tel: +44 (0)20 8996 9001, Email: cservices@bsigroup.com. 3552 The calculation of the dryness concentration is then made using the following equation [19]: 3553 3554 (T2 – T1)(Cpw(ms – me) + A) 3555 3556 (T3 – T2)Cpw 3557 D = ____ – ____ 3558 3559 L (mf – ms) 3560 3561 L 3562 Where: 3563 L = latent heat of dry saturated steam at temperature T3, in kilojoules per kilogram 3564 me = mass of the Dewar flask and stopper, pipes and tube, in kilograms 3565 ms = mass of the Dewar flask, water charge stopper, pipes and tube, in kilograms 3566 mf = mass of the Dewar flask, water charge, condensate, stopper, pipes and tube, in kilograms 3567 T1 = initial temperature of the water in the Dewar flask, in degrees Celsius 3568 T2 = final temperature of the water and condensate in the Dewar flask, in degrees Celsius 3569 T3 = temperature of dry saturated steam delivered to the sterilizer, in degrees Celsius 3570 Cpw = specific heat capacity of water (4.18 kJ/kg K) 3571 D = dryness value of the steam 3572 A = effective heat capacity of the apparatus (0.24 kJ/K) 3573 3574 This Document is licensed to 3575 ISPE Good Practice Guide: 3576 Page 61 3577 Sampling for Pharmaceutical Water, Steam, and Process Gases 3578 3.3.1.3 Superheating 3579 Superheating must be below 25°C (77°F) when the pure steam is expanded to atmospheric pressure according to BS 3580 EN 285:2015, Chapter 13.3.4 [19]. Samples should be taken from the center of the steam pipe in accordance with the 3581 test method of BS EN 285:2015, Chapter 22.3 [19], as shown in Figures 3.8 and 3.9. 3582 Figure 3.8: Locating a Port for Sampling for Superheating 3583 Permission to reproduce extracts from British Standards (EN 285:2015, Figure 10) is granted by BSI. 3584 British Standards can be obtained in PDF or hard copy formats from the BSI online shop: www.bsigroup.com/Shop or by contacting BSI Customer 3585 Services for hardcopies only: Tel: +44 (0)20 8996 9001, Email: cservices@bsigroup.com. 3586 Figure 3.9: Sampling Apparatus for the Measurement of Superheating 3587 Permission to reproduce extracts from British Standards (EN 285:2015, Figure 11) is granted by BSI. 3588 British Standards can be obtained in PDF or hard copy formats from the BSI online shop: www.bsigroup.com/Shop or by contacting BSI Customer 3589 Services for hardcopies only: Tel: +44 (0)20 8996 9001, Email: cservices@bsigroup.com. 3590 The calculation is: 3591 enthalpy at atmospheric pressure: 2673.8 kJ/kg 3592 heat capacity of saturated steam: 2.029 kJ/kg/°K 3593 maximum allowed superheating according to BS EN 285:2015 [19]: 25°K 3594 superheat enthalpy: 2673.8 kJ/kg + (2.029 kJ/kg/°K) (25°K) = 2724.53 kJ/kg 3595 While we do not go into all the calculations here, saturated steam at 2 barG (~30 psi) has an enthalpy of 2723.2 kJ/ 3596 kg, which is less than the superheat enthalpy calculated above. Therefore, if the pressure is below 2 barG and the 3597 dryness value is below 1.0, the steam cannot get superheated and will always pass the superheat requirement for 3598 pure steam. 3599 3600 This Document is licensed to 3601 Page 62 3602 ISPE Good Practice Guide: 3603 3604 Sampling for Pharmaceutical Water, Steam, and Process Gases 3605 The formula for higher pressures (> 2 bar) is: 3606 D H + (1 – D) * h < 2724.53 kJ / kg 3607 Where: 3608 D = measured dryness 3609 H = enthalpy of steam at measured steam pressure at POU 3610 h = enthalpy of water at measured steam pressure at POU 3611 Figure 3.10: Positioning of Sample Collection for Wetness and Superheating Test 3612 3.3.1.4 Conductivity, TOC, and Endotoxins 3613 Condensate collected from pure steam sample points should comply with the specifications for WFI grade water. 3614 Such sample points should be placed at the outlet of the pure steam generator, at POUs, or as close to POUs as 3615 possible. There may be one sampling cooler at each sampling point, or coolers can be transported from sample 3616 point to sample point, depending on the design of the system. A simple method for sample collection for these 3617 purposes is to have a valve with a sanitary connection at each drainage point, allowing connection of a steam trap 3618 and condensate cooler without closing the system down. Design considerations for sample valves for these quality 3619 attributes is identical to those for water systems, which have been covered in Chapter 2 of this Guide. 3620 3621 This Document is licensed to 3622 ISPE Good Practice Guide: 3623 Page 63 3624 Sampling for Pharmaceutical Water, Steam, and Process Gases 3625 3.4 3626 Sampling Plans (Frequency and Duration) 3627 Pure steam is defined to be steam that does not contain any added substances and the condensate sampled from 3628 pure steam meets the requirements for USP [9]/EP [2] WFI water. In addition, pure steam that is used for sterilization 3629 applications of porous loads for international manufacturing also should meet the requirements of BS EN 285:2015 3630 [19] for dryness, non-condensable gases, and superheat the USP states that: 3631 3632 “Pure Steam is prepared from suitably pretreated source water analogously to either the pretreatment used for 3633 Purified Water or Water for Injection”. 3634 Pure steam that is used in aseptic applications, as well as applications where the steam or condensate directly 3635 contacts products, excipients, container closures, or equipment/components, should be appropriately sampled. 3636 The recommended frequencies for sampling pure steam should be determined based on the state of the system such 3637 as: 3638 • 3639 Restoration from scheduled/unscheduled shutdown 3640 • 3641 Maintenance 3642 • 3643 Validation 3644 • 3645 Normal routine operation 3646 The frequency of sampling can also be based upon the generation and distribution portions of the system. Physical, 3647 chemical, and biological elements (i.e., endotoxin) should be monitored for pure steam. There may be different 3648 sampling frequencies employed within these categories. 3649 Risk analysis tools may be used to develop a sampling rationale that takes into account: 3650 • 3651 The criticality of the POUs on the system 3652 • 3653 The frequency of use of each POU 3654 • 3655 The accessibility for sampling 3656 • 3657 Historical performance 3658 Figure 3.11 is a schematic representation of the typical components of a clean/pure steam system. 3659 3660 This Document is licensed to 3661 Page 64 3662 ISPE Good Practice Guide: 3663 3664 Sampling for Pharmaceutical Water, Steam, and Process Gases 3665 Figure 3.11: Simple Clean Steam System Diagram 3666 Note: in Figure 3.11 steam traps have not been shown for the purposes of clarity. 3667 3.4.1 3668 Commissioning and Qualification 3669 Commissioning is defined in the ISPE Good Practice Guide: Approaches to Commissioning and Qualification of 3670 Pharmaceutical Water and Steam Systems (Second Edition) [13] as: 3671 3672 “a well-planned, documented, and managed engineering approach to the start-up and turnover of facilities, 3673 systems, and equipment to the end-user that results in a safe functional environment that meets established 3674 design requirements and stakeholder expectations”. 3675 Qualification is intended to provide documented evidence with oversight by quality assurance personnel that a pure 3676 steam system will consistently produce the appropriate quality of steam to end users in accordance with regulatory 3677 expectations [13]. 3678 3.4.2 3679 Commissioning 3680 Sampling and monitoring of generation and distribution systems for pure steam systems during commissioning is 3681 considered GEP. Commissioning is where the intended operating parameters of the equipment are identified and 3682 tested. For further information on commissioning and qualification recommended practices see the ISPE Good 3683 Practice Guide: Approaches to Commissioning and Qualification of Pharmaceutical Water and Steam Systems 3684 (Second Edition) [13]. 3685 3686 This Document is licensed to 3687 ISPE Good Practice Guide: 3688 Page 65 3689 Sampling for Pharmaceutical Water, Steam, and Process Gases 3690 Sampling and testing activities include chemical testing as appropriate for the unit operation with the maximum 3691 acceptable chemical levels of the condensate meeting the applicable WFI monograph requirements, e.g., EP [2], 3692 USP <1231> [9], and JP16 [11]. Typically, the duration of commissioning activities may vary depending upon the 3693 complexity and size of the system. Once the entire system is operating, each unit operation should be tested at least 3694 once during the commissioning phase. The total time needed for sampling during commissioning is determined by the 3695 time needed to complete all unit operation testing at least once during the commissioning phase. Depending upon the 3696 depth and breadth of the commissioning plan, the subsequent qualification may serve only as additional proof that the 3697 system will consistently deliver the required attributes of pure steam. The data collected as part of the commissioning 3698 sampling phase may be used to justify the sampling plan developed for qualification, if the acceptance criteria have 3699 been preapproved by an appropriate SME and/or quality assurance representative [13]. 3700 The commissioning plan should sufficiently be rigorous, to test and challenge the CPPs of the system. During this 3701 phase, if any changes are made to defined CPPs, the commissioning phase should be repeated. It is strongly 3702 recommended that any changes be identified during this phase, as changes identified and initiated during the initial 3703 and/or intermediary phases of qualification are more costly to implement. 3704 3.4.2.1 Qualification 3705 Qualification typically consists of three parts: IQ, OQ, and PQ. Since most sampling is done during the PQ, the 3706 following discussion will focus on PQ testing. 3707 PQ is typically organized into three phases: 3708 1. 3709 Initial sampling 3710 2. 3711 Intermediary sampling 3712 3. 3713 Extended sampling 3714 Activities performed during commissioning may be leveraged if the work was suitably conducted and documentation 3715 is prepared according to prescribed site quality documentation practices.3 3716 Initial Sampling 3717 Initial sampling is intended to demonstrate that the system consistently operates within predetermined operating 3718 ranges and delivers pure steam of the required quality. Sampling points are typically located immediately after the 3719 pure steam generator, at POUs and at worst case locations, which are either the most distant from the pure steam 3720 generator or at the lowest points in the system. In addition, sampling and testing for the physical attributes (i.e., non- 3721 condensable gases, superheat, or dryness) should be performed while equipment (e.g., autoclaves) is operating. The 3722 initial sampling study should be performed for a minimum of three consecutive working days but may last for up to ten 3723 working days with samples being taken from each POU and at the outlet of the pure steam generation system.3 3724 Pure steam should not be used for production purposes during the initial sampling phase. The formal provisional 3725 release of the pure steam system should be performed by quality management after the review and assessment of all 3726 available results obtained during the initial sampling phase and the preparation of a summary report from this phase. 3727 In the case of larger systems with multiple branches, sampling at the ends of each individual pipe branch should be 3728 taken into consideration as shown in Figure 3.12. 3729 3730 This Document is licensed to 3731 Page 66 3732 ISPE Good Practice Guide: 3733 3734 Sampling for Pharmaceutical Water, Steam, and Process Gases 3735 Figure 3.12: Complex Clean Steam System Diagram 3736 Note: in Figure 3.12 steam traps have not been shown for the purpose of clarity. 3737 Intermediary Sampling 3738 Intermediary sampling is intended to demonstrate that a system consistently operates within predetermined operating 3739 ranges and delivers pure steam of the required quality. Intermediary sampling or system consistency/stability 3740 activities typically last for the same duration as the initial sampling phase and occur immediately following completion 3741 of the initial sampling phase. 3742 The sampling and testing frequency are nearly identical to the activities in initial sampling phase except that the outlet 3743 of the PSG need not be sampled daily, but the frequency may be reduced. One suitable approach would be to sample 3744 once at the start of the intermediary phase and again at the end of the intermediary phase. 3745 An assessment approved and documented by quality management, should also be available at the end of the 3746 intermediary sampling phase, which allows an uninterrupted progression into the extended phase of sampling. 3747 Pure steam produced during the intermediary phase may be provisionally released for manufacturing use, provided 3748 that the initial phase was assessed and documented by quality management to be successful and the system has 3749 been authorized for use. 3750 Summary Interim Report 3751 A summary interim report should be prepared following both the initial and intermediary sampling phases. The report 3752 should be based on all data gathered during these phases and provides the conclusion on the acceptability (or 3753 otherwise) of the system for routine production use during the extended phase of sampling. The summary interim 3754 report following the initial phase would be used to justify the provisional release of the system for manufacturing use. 3755 The summary interim report following the intermediate phase would be used to justify the full release of the system 3756 for manufacturing purposes. 3757 3758 This Document is licensed to 3759 ISPE Good Practice Guide: 3760 Page 67 3761 Sampling for Pharmaceutical Water, Steam, and Process Gases 3762 Extended Sampling 3763 Extended sampling is designed to demonstrate that the pure steam system, when operated over a longer period of 3764 time (typically for the balance of the year), will consistently produce pure steam of the required quality. The PQ of the 3765 pure steam system is considered completed when performance data for a full year (inclusive of all sampling phases) 3766 are obtained. The extended sampling plan further reduces sampling frequencies for the balance of the year, with the 3767 outlet of the pure steam generator being sampled on a weekly basis and other locations sampled on a rotating basis 3768 such that each location is sampled every two weeks. 3769 A report should be created documenting the overall qualification process (all phases) including a clear conclusion 3770 statement on whether the system was successfully qualified. The report should be reviewed and authorized by the 3771 relevant functions involved, including quality management. 3772 3.4.3 3773 Ongoing Routine Monitoring Program 3774 Following successful qualification of the pure steam system, routine ongoing monitoring should involve sampling and 3775 testing at a frequency and for the critical quality attributes that are supported by a documented risk analysis process. 3776 Some considerations may involve local or other regulatory compliance requirements as well as: 3777 • 3778 The use of the system, including processes where the steam is used (e.g., autoclaves) 3779 • 3780 The design and implementation of POUs (e.g., frequency of port use, cool points, dead legs, line slopes, types of 3781 valves, location of the port in the loop or on the line) 3782 • 3783 The material of construction of piping 3784 • 3785 The sampling technique and sample container utilized 3786 • 3787 The design of feed water and steam distribution systems, including, but not limited to: age of system, system 3788 use, feed water quality, proper steam trap placement, etc. 3789 • 3790 Routine maintenance (e.g., rouge inspections, passivation) 3791 • 3792 Qualification results 3793 The physical clean steam quality attributes (i.e., dryness, non-condensable gas limits, maximum superheat) should 3794 be monitored at established, regular intervals, but the frequencies may be different from the sampling and testing of 3795 the chemical and biological aspects of the steam condensate. 3796 The quality attributes (i.e., TOC, conductivity and bacterial endotoxins) should be measured in the pure steam 3797 condensate and tested according to WFI specifications. The items “Appearance” and “Nitrates” should be included in 3798 the testing schedule to assure compliance with the European Pharmacopoeia. Testing for total aerobic count for pure 3799 steam condensate is not necessary due to the lethal properties of steam. 3800 Routine monitoring frequencies should be set up after a thorough evaluation of the results obtained during the three 3801 phases of the qualification process. Depending upon the results, routine sampling schedules may vary significantly 3802 from those applied during the extended sampling phase (e.g., if data obtained during qualification shows that the 3803 system is in a state of control). However, if occasional adverse results occur (e.g., exceeding Alert and Action Levels, 3804 adverse trends), measures should be taken (e.g., system redesign and modification) to improve the situation, such as 3805 the detection and elimination of cold spots and/or areas where condensate can accumulate. 3806 An example of an Ongoing Sampling Plan appears in Table 3.5. 3807 3808 This Document is licensed to 3809 Page 68 3810 ISPE Good Practice Guide: 3811 3812 Sampling for Pharmaceutical Water, Steam, and Process Gases 3813 Table 3.5: Ongoing Monitoring Program Example 3814 Pure Steam 3815 Weekly 3816 Endotoxin, nitrates, TOC, conductivity 3817 Generator 3818 Monthly 3819 Endotoxin, nitrates, TOC, conductivity 3820 One-way distribution most distant 3821 points 3822 Biannually 3823 Endotoxin, nitrates, TOC, conductivity 3824 All use points 3825 Annually 3826 Steam quality (non-condensable gas, 3827 superheat, dryness) 3828 Autoclave use points 3829 Trending of data should be performed at regular intervals to detect any adverse trends, with intervals set depending 3830 upon the pure steam system’s quality and performance (e.g., ongoing, quarterly, semi-annually or annually). Testing 3831 frequencies may change based on the trending results obtained and authorized by appropriate change control 3832 functions. 3833 3.4.4 3834 System Shutdown 3835 Following a shutdown for an extended period as defined by SOPs, samples for chemical and biological (i.e., 3836 endotoxin) properties should be collected from the generator and all use points and tested before system use and 3837 prior to resumption of use for routine activities. 3838 3.5 3839 Sample Valve Design 3840 This is covered in Section 2.4 of this Guide. 3841 3.6 3842 Pure Steam Sampling Techniques 3843 3.6.1 3844 Obtaining Representative Samples 3845 The goal of the sampling program is to obtain samples that accurately reflect the contents of the steam system. When 3846 sampling a liquid, contaminant distribution should be accounted for as part of the sample evaluation. 3847 Pure steam cannot be evaluated in its vapor state; therefore, pure steam needs to be condensed in order to be 3848 evaluated. Based on compendial requirements (USP, EP, JP, etc.), condensed pure steam should be evaluated using 3849 the same criteria used to evaluate WFI. The techniques used are similar to those for water; however, there are two 3850 issues that can affect pure steam condensate sampling: 3851 1. 3852 The equipment used to condense the pure steam and collect samples 3853 2. 3854 The final temperature of the condensate 3855 3.6.1.1 3856 Equipment Considerations 3857 When condensing pure steam for sampling, the equipment should not contribute contaminants to the sample. The 3858 equipment should be designed to allow for complete cleaning and drainage. It should also be possible to sterilize 3859 the equipment and for it to be depyrogenated to ensure that no endotoxins are added to the sample. In addition, 3860 the equipment typically requires a cooling source (such as chilled water) to facilitate condensation. The amount of 3861 chilled water required may be significant, necessitating larger than expected equipment sizes and/or cooling media 3862 flow rates, because of the significant amount of latent heat that needs to be removed to condense the pure steam. 3863 Calculations should be performed to determine the exact requirement. 3864 3865 This Document is licensed to 3866 ISPE Good Practice Guide: 3867 Page 69 3868 Sampling for Pharmaceutical Water, Steam, and Process Gases 3869 Sampling heat exchangers can be sterilized in place, using the pure steam from the system. This can help to simplify 3870 the sampling process, but may result in a larger cooling load, as the mass of the sampling heat exchanger needs to 3871 be raised to the necessary temperature for sterilization, as a minimum. This methodology can result in an extended 3872 sampling duration requirement, possibly for each sample obtained. 3873 Heat exchangers used for pure steam condensation and subsequent sampling may need to meet the requirements 3874 of heat exchangers used for WFI sampling in order to mitigate the risk of undetectable cooling water contamination of 3875 the condensate sample. Requirements may include double tube sheet design and routine maintenance to ensure that 3876 cooling water contamination does not occur. The size of this equipment will depend on the: 3877 1. 3878 Pure steam pressure (and resulting temperature) 3879 2. 3880 Sample temperature desired (“delta T”) 3881 3. 3882 Temperature and flow of cooling media available 3883 If hoses or other components are employed during the sampling process, they should be suitably cleaned and dried 3884 when not in use, sterilized and possibly depyrogenated to prevent sample contamination. 3885 It is important to predetermine the type of equipment that will be required based on the application. A source of 3886 cooling media is required near each sample point unless a portable, packaged, closed loop chiller is used. For fixed 3887 applications, permanently piped coolant supply and return lines are required, unless once through cooling (requiring 3888 suitable drains) is selected. In addition, a suitable drain or receptacle should be available for flushing. 3889 Once there is sufficient condensate flowing, samples can be collected. Again, any flush time interval required should 3890 be qualified. The sampler should take the sample immediately; this need not be established based on the equipment 3891 used to condense the steam to get the best sample. 3892 Note: bioburden is not required of steam sampling per the USP/NF. However, establishing that the pure steam line has 3893 no thermophilic bioburden should be an evaluation done prior to or during the PQ. Vegetative bacteria will not be present 3894 in condensed pure steam unless they are outgrowths of spores, from inadequately designed or failed steam traps which 3895 prevent condensate removal, from poorly prepared condensers, or from poor sample collection and handling. 3896 Packaged Sampling Systems for Steam Collection 3897 Packaged sampling systems are commercially available for both fixed and portable applications. Additionally, individual 3898 components may be purchased for assembly in the field as needed. Figure 3.13 shows an example of such a kit. 3899 Figure 3.13: Steam Quality Sampling Kit Example 3900 Used with permission from Carltex Inc., http://www.carltex.com/. 3901 3902 This Document is licensed to 3903 Page 70 3904 ISPE Good Practice Guide: 3905 3906 Sampling for Pharmaceutical Water, Steam, and Process Gases 3907 Sample Containers 3908 When taking samples, efforts should be made to ensure that sampling containers do not introduce measurable 3909 contaminants. This may necessitate using several types of sample containers to sample a given POU. The 3910 temperature of the condensate sample should be considered if a material other than glass is utilized. 3911 Steam collection kits that support collection of samples to be tested under BS EN 285:2015 [19] are commercially 3912 available for purchase. Any collection containers procured from a supply company not selling certified kits should 3913 conform to the requirements in BS EN 285:2015 [19]. Metal parts on apparatus must be constructed of sanitary grade 3914 stainless steel and passivated. Non-metal apparatus should be glass or HDPE or heat resistant plastic material with 3915 smooth surfaces. Measuring apparatus should be certified as clean and suitable for use accordance with ASTM 3916 apparatus standards. 3917 There should be a plentiful supply of prepared sterilized bioburden sampling containers and appropriate clean 3918 chemical collection containers. 3919 If containers are reused, they should be carefully washed and rinsed, with a final distilled or deionized water rinse 3920 prior to sterilization. Plastic containers will require steam sterilization in a validated cycle. Glass containers can be 3921 sterilized by steam or dry heat, depending on the container closure. That cycle must also be properly validated. 3922 Types of Samples and Containers 3923 For most high purity and pure steam samples, three to four groups of samples are taken: 3924 1. 3925 TOC 3926 TOC samples should be taken in TOC collection vials, which are made of glass, as plastic containers can introduce 3927 TOC to the sample results. A TOC vial sample rack should be used that securely holds the unused and filled TOC 3928 vials. This can help to prevent accidental breakage during collection and handling. 3929 The TOC vials typically hold a 40 ml sample and should be filled completely before the closure is applied. 3930 2. 3931 Conductivity and pH 3932 Typically, glass (borosilicate), sterile polypropylene plastic, or inert fluoropolymer containers are the most common 3933 container material for collecting condensed steam samples. Glass containers may introduce trace residues to 3934 the sample that will change both the conductivity and pH of the collected sample. Plastics tend to avoid the ionic 3935 leachables, but most are somewhat permeable to atmospheric CO2, which will alter the conductivity of the sample. 3936 Container leaching should be evaluated prior to implementation, but the best practice would be to test the samples as 3937 quickly as possible after sampling. Typically, containers should be able to hold at least 300 ml of sample, to allow for 3938 Conductivity Stage 1, Stage 2, and Stage 3 testing. 3939 3. 3940 Bioburden 3941 Bioburden samples should be taken in sterile containers, typically an inert plastic such as polypropylene or a suitable 3942 fluoropolymer. Typically, high purity water testing includes a total plate count. Some companies also include a coliform 3943 count, although sampling for coliform is not needed in most systems. 3944 4. 3945 Endotoxin 3946 Endotoxin sampling should be done using pyrogen free containers of either glass or inert plastic. If considering 3947 plastic, the type of plastic container is important to minimize the potential for endotoxin adsorption issues that could 3948 result in false negative results. Sampling for endotoxin testing requires collecting approximately 10 to 20 ml in sterile 3949 pyrogen free containers. 3950 3951 This Document is licensed to 3952 ISPE Good Practice Guide: 3953 Page 71 3954 Sampling for Pharmaceutical Water, Steam, and Process Gases 3955 3.6.1.2 Final Temperature of the Condensate 3956 The condensate temperature for steam sampling should be < 50°C (122°F) and preferably near ambient temperature. 3957 It is preferable that a temperature gauge be installed near the condensate line, if not one cannot be located at the 3958 condensate inlet on the condensing equipment. 3959 It is recommended that sample conditioning condensers, coolers, and associated fittings and valves be constructed of 3960 316 stainless steel and passivated. 3961 Any special sampling equipment required that can make steam contact during collection (e.g., POU gaskets, sanitary 3962 fittings, hose barbs, steam condensing equipment) should be pyrogen free or must have been depyrogenated. 3963 3.6.2 3964 Training Sample Collectors 3965 Operators should be properly trained and should have demonstrated their ability to collect samples following the 3966 approved sample collection SOP. There should be a particular emphasis on safety training when sampling pure 3967 steam, its condensate, or other hot liquid. Pressurized pure steam, often at pressures of 30 to 45 psig (135°C to 3968 145°C (~275°F to 293°F)) and its condensed water (at temperatures near 82°C (180ºF)) can cause severe burns and 3969 potentially death. 3970 During their training samplers should have observed a trained sampler who demonstrates the collection procedure for 3971 each type of collection, as well as the proper handling of each type of collected sample prior to hand off to the testing 3972 function. When collecting samples during training, there should be a duplicate set of samples taken by the trainer. 3973 The test results from the new trainee should be within a predefined acceptance level to those of the trainer as defined 3974 in the appropriate standard operating procedures. Any major discrepancies in trainee’s test samples (physical or 3975 visual) may require additional resampling and should be noted and discussed with the trainee, including the possibility 3976 of a requirement for retraining. 3977 Based on industry accepted best practice, trainees may be required to demonstrate that they can collect a minimum 3978 of three distinct sets of samples, one set per day for each type of collection and each unique location from which they 3979 will be required to collect samples. 3980 Trainees should not be allowed to conduct unobserved sampling and handling until the trainer is satisfied that they 3981 can sample and collect to an appropriate standard, without impacting the water sample test results. 3982 3.6.3 3983 Preparation for Sample Collection 3984 Materials should be prepared before beginning the actual sampling process. An inventory of containers/equipment 3985 should be maintained, to avoid delays for lack of materials. In some cases, preparation should be made the day 3986 before sampling. 3987 Examples of typical items that may be staged for use include, but are not limited to: 3988 • 3989 Rolling railed cart, stainless or cleanable plastic 3990 • 3991 Portable handle container 3992 • 3993 Plastic/wire baskets and TOC racks 3994 • 3995 Sample containers 3996 • 3997 Labels, writing instruments, e.g., waterproof markers and indelible ink pens 3998 • 3999 Laminated sampling plan 4000 4001 This Document is licensed to 4002 Page 72 4003 ISPE Good Practice Guide: 4004 4005 Sampling for Pharmaceutical Water, Steam, and Process Gases 4006 It is important that the pure steam collection and testing SOP is reviewed carefully prior to sampling. This can help 4007 to ensure that final test results are valid within the limits of the analysis. The sampler should be aware of the types 4008 of sample collection containers. All collection containers should be capable of being sealed to prevent leakage or 4009 contamination of the sample. 4010 3.6.3.1 Sample Point Preparation 4011 Sample point preparation for sampling pure steam may or may not include flushing, spraying, disinfecting, or other 4012 treatments, based on the temperature, and assuming proper design. However, if particulate contamination from the 4013 local environment may be a problem, physical cleaning (i.e., wiping) of the sample outlet may be indicated, taking into 4014 account the temperature of the sampling point. 4015 3.6.3.2 Use of Common Sampling Points for Process Control and Quality Control 4016 Sampling for quality control purposes should reflect how the pure steam is used in production. In some cases, a 4017 single sample point in a piping system may be used for both purposes. 4018 3.6.3.3 Multiple Sample Points 4019 If a heat exchanger is dedicated to a specific POU for pure steam sampling and is permanently installed minimum 4020 setup will be required for each sampling event allowing utility connections and sanitary piping to be permanently 4021 installed and making sample taking fairly routine. 4022 However, some heat exchanger units are mounted on portable carts either alone or with on board chillers, piping 4023 and controls. Portable units minimize capital outlay but often require special provisions at each sample location 4024 for connection of the portable unit to utilities and sanitary piping. In addition, there is an associated challenge of 4025 sanitizing the portable system between samples, although in some cases steaming at each subsequent sample 4026 location may be suitable for ensuring cleanliness and sterility. 4027 3.6.3.4 Duplicate Samples 4028 An important but often times overlooked part of any quality control program is to perform a predetermined percentage 4029 of all analytical tests in duplicate for a period of time to confirm the validity of established procedures, reagents and 4030 techniques. 4031 When duplicate samples are taken, at least double the normal test volume should be collected. 4032 Duplicate samples should be taken from the same sample location, but in two separate sample containers. Both 4033 containers should be labeled with the same sample identification, but can be distinguished by adding “-1” and “-2” 4034 after the sample number on the two containers. 4035 3.6.3.5 Blanks 4036 A blank containing no sample should be submitted to the laboratory for each type of container used during sampling. 4037 The blank collection container should be from the same lot as the collection container used for that date. All 4038 precautions observed in the preparation of containers for sampling shall also be observed for containers used in the 4039 preparation of analytical blanks. 4040 Variability in the blank can call the accuracy of any analysis into question. In the event of unusual test results, the 4041 laboratory can use a qualified high purity water sample to analyze any issues in the blank as part of an investigation. 4042 4043 This Document is licensed to 4044 ISPE Good Practice Guide: 4045 Page 73 4046 Sampling for Pharmaceutical Water, Steam, and Process Gases 4047 3.6.4 4048 Sampling Procedures 4049 3.6.4.1 General Considerations 4050 For routine operation of critical systems that have completed PQs, there should be detailed sampling procedures 4051 approved for use. The departments responsible for sample collection and handling should have procedures detailing 4052 how to collect, label and handle the collected samples. It may not be necessary to have individual procedures for 4053 each sample point and one sampling and collection procedure may be used cover all samples, collections, and 4054 assays. Since additional apparatus may be required for sampling of pure steam quality attributes, it is a good practice 4055 to keep pure steam sampling procedures separated from those used to collect pure water samples. 4056 When systems are new and undergoing PQ or have been modified and as part of a requalification PQ with a new 4057 sampling plan, it is acceptable to use the draft procedures. 4058 Care should be taken to protect the sampler from scalding. Proper heat resistant gloves should be present in the 4059 collection kit and worn to protect the sampler from accidental burning. 4060 If the POU is located in a dirty area, clean the sample port with 3% hydrogen peroxide or 70% sterile IPA prior to 4061 flushing the sample point. The cleaning procedure must be qualified prior to routine use and must account for the fact 4062 that the use of sterile IPA will impact TOC results. 4063 Samplers should practice the good hand hygiene and health practices expected to be followed in any GMP facility. 4064 Appropriate GMP gowning for the area should be worn. Bouffant caps and beard covers should be used to prevent 4065 the sampler’s hair from contaminating the sample during collection. 4066 Samplers should wear powder free gloves before taking samples or handling sample containers. 4067 Samplers should open the sample valve and flush the POU by running steam/condensed water to waste for the 4068 qualified flush time interval. This will allow for adequate flushing of the pipe between the sampling point and the 4069 condensing equipment. Sampling procedures may vary based on sample valve location, valve design, and the 4070 attribute being sampled. Sample valves should be of a sanitary design although laboratory points may not need to be 4071 as hygienically designed as those supplying pure steam to manufacturing. 4072 Heat exchangers and sampling components should be suitably cleaned and/or sterilized. If reusable sample 4073 containers are utilized, cleaning procedures should be qualified prior to sampling. 4074 When sampling for process control purposes, sampling should be performed in accordance with established 4075 procedures which should define flush times, hold times, and other variables. 4076 When sampling for quality control purposes, sampling should replicate the way that the steam is used for 4077 manufacturing. Converting steam to liquid at higher flow rates may be required and the sizing of cooling equipment 4078 and media should be adjusted accordingly. 4079 3.6.4.2 Sample Labeling 4080 Samplers should label all sample containers with: 4081 1. 4082 Sample identification: 4083 a. 4084 POU (include system ID and building, if required) 4085 2. 4086 Type of sample (e.g., condensed steam) and indicate if a duplicate sample 4087 3. 4088 Date and time of sampling 4089 4090 This Document is licensed to 4091 Page 74 4092 ISPE Good Practice Guide: 4093 4094 Sampling for Pharmaceutical Water, Steam, and Process Gases 4095 4. 4096 Analysis required 4097 5. 4098 Sampler’s name 4099 3.6.4.3 Sampling Procedure for TOC, Conductivity/pH, Bacteria, and Endotoxin 4100 1. 4101 Samplers should flush the POU for the qualified flush time interval and at the qualified flush flow rate. 4102 2. 4103 Samplers should wear clean powder free gloves. 4104 3. 4105 Samplers should remove and set aside the sample container closure, and then manually open the valve on the 4106 condenser. 4107 4. 4108 To prevent the sample from becoming contaminated, samplers should use caution to not let anything, except the 4109 sample, touch the inside of the container or the inside of the closure. 4110 5. 4111 Samplers should fill each container completely, except for conductivity/pH samples where approximately 1 inch 4112 of air space should be left at the top to allow for adequate agitation prior to analysis. 4113 6. 4114 Samplers should securely close the sample containers to prevent leakage or contamination. 4115 If there is any question whether a sample has become contaminated, it should be discarded and a new sample taken. 4116 3.6.4.4 Non-condensable Gases 4117 Figure 3.5 shows sampling while running a steam sterilization cycle on a large steam sterilizer. It is recommended 4118 that a sequence of tests should be performed to determine the variability in the level of non-condensable gases. 4119 1. 4120 Fill the 2 l container and burette with cold, deaerated water. Invert the burette with great care to ensure no air is 4121 trapped in the burette. 4122 2. 4123 Remove the steam sampling pipe from the 2 l container and open the needle valve to purge all air from the pipe. 4124 Return sampling pipe to container, allowing it to bubble into the funnel. Add deaerated water to the 2 l container 4125 until water flows out through the overflow pipe. 4126 3. 4127 Place the graduated cylinder to collect overflow. Adjust the needle valve to allow a continuous flow of steam into 4128 the funnel, ensuring that the steam enters the funnel in a way that any non-condensable gases will be collected 4129 in the burette. 4130 4. 4131 Close the sampling valve. 4132 5. 4133 Start the sterilization cycle and when steam is flowing to the sterilizer, open the sampling valve, allowing a 4134 continuous sample of steam into the funnel at a rate sufficient to cause a small amount of steam hammer. 4135 6. 4136 Allow the steam sample to condense in the funnel. Any non-condensable gases will collect in the top of the 4137 burette. Collect the overflow in the graduated cylinder and observe the temperature of the water in the 2 l 4138 container. When the temperature in the container is between 70°C to 75°C (158°F to 167°F), close the needle 4139 valve. 4140 7. 4141 Record the volume of water displaced in the burette and the volume of water collected in the graduated cylinder. 4142 4143 This Document is licensed to 4144 ISPE Good Practice Guide: 4145 Page 75 4146 Sampling for Pharmaceutical Water, Steam, and Process Gases 4147 8. 4148 Calculate the concentration of non-condensable gases using the following equation: 4149 Cn (percent) = (Vb / Vc) × 100 4150 Where: 4151 Cn = the concentration of non-condensable gases, in percent 4152 Vb = the volume of non-water displaced from the burette, in ml 4153 Vc = the volume of water collected in the graduated cylinder, in ml 4154 The percentage of non-condensable gases should be equal to or less than 3.5% (V/V). 4155 3.6.4.5 Dryness 4156 Steam should be tested for, and should pass the non-condensable gases test, prior to being tested for dryness. Refer 4157 to Figure 3.6 and 3.7 and understand that this test assumes the use of a large steam sterilizer operating in the textile 4158 cycle. 4159 1. 4160 Assemble the apparatus as shown in Figure 3.7. 4161 2. 4162 Weigh the assembled Dewar flask and record the mass as Me. 4163 3. 4164 After several sterilization cycles, select the textile cycle with a temperature of 134°C (273°F). 4165 4. 4166 Remove the stopper assembly and place 650 ml of water into the Dewar flask, with the water at a temperature at 4167 or below 27°C (81°F). Weigh the new assembly and record the mass as Ms. 4168 5. 4169 Place a standard test pack (details in BS EN 285:2015 [19]) in the sterilizer. 4170 6. 4171 Introduce a temperature sensor into the Dewar flask and note the temperature of the fluid in the flask, record this 4172 value as T1. 4173 7. 4174 When the sterilizer steam valve first opens, attach the rubber tube (3) to the pitot tube (1) ensuring free drainage 4175 of condensate into the Dewar flask. Record the steam temperature as T3. 4176 8. 4177 When the water in the Dewar flask reaches approximately 80°C (176°F), disconnect the rubber tube, agitate the 4178 flask to thoroughly mix the contents and record the temperature as T2. 4179 9. 4180 Weigh the Dewar flask again, record this mass as Mf. 4181 Dryness is calculated using the recorded values according to the following equation: 4182 Dryness (D) = ((T2 – T1)(Cpw(Ms – Me) + A))/L(Mf – Ms) – ((T3 – T2)Cpw)/L 4183 Where: 4184 L = the latent heat of saturated steam at temperature T3, in kilojoules per kilogram 4185 Cpw = the specific heat capacity of water (4.18 kioljoules/kilogram/°K) 4186 A = the effective heat capacity of the apparatus (0.24 kilojoules/°K) 4187 The dryness should be 0.95 or above for metal loads and 0.90 for other types of loads. 4188 4189 This Document is licensed to 4190 Page 76 4191 ISPE Good Practice Guide: 4192 4193 Sampling for Pharmaceutical Water, Steam, and Process Gases 4194 3.6.4.6 Superheat 4195 Refer to Figures 3.8 and 3.9. 4196 1. 4197 Assemble the apparatus as indicated in Figures 3.8 and 3.9. 4198 2. 4199 Run a sterilization cycle with an empty chamber. 4200 3. 4201 At the completion of this cycle, immediately place a full textile load and run another sterilization cycle. At the 4202 completion of the sterilization cycle, check the temperature recordings and confirm that the temperature did not 4203 differ by more than 3°C (37°F) from that measured in the steam pipe during the dryness test. 4204 3.7 4205 Sample Handling 4206 Clear instructions should be provided to personnel collecting and handling pure steam condensate samples. A sample 4207 holding time limit should be established prior to the PQ to confirm the maximum time between collection and testing. 4208 Samples should be transferred to the laboratory and placed on test within an hour of collection, to help increase the 4209 accuracy of the test results. If multiple samples are being collected and the collection process takes longer than an 4210 hour, the collector should divide up the sampling process, deliver samples to the testing laboratory, and then return to 4211 sampling. 4212 Refrigerating samples will help to inhibit microbial growth, but gas absorption may create inaccurate test results. 4213 3.8 4214 Other Factors Influencing Sampling Strategies 4215 Sampling strategies may be influenced by other factors such as: 4216 • 4217 Geographic location of the plant 4218 • 4219 Incoming water quality 4220 • 4221 Pretreatment equipment 4222 • 4223 PSG design 4224 Each of these factors should be carefully evaluated for impact on steam quality as sampling strategies are developed. 4225 The geographic location of the plant may lead to variation in incoming water quality as well as variations in the type of 4226 chemical treatment in use (e.g., chlorine, chloramine, ozone, etc.). In some instances, ammonia may be present as a 4227 byproduct of the chloramine process. Ammonia and carbon dioxide are particularly troublesome since they will pass 4228 through a pure steam generator and will influence the non-condensable gas content as well as the conductivity. When 4229 such conditions exist case contaminant removal should be addressed through proper pretreatment and confirmed 4230 with appropriate sampling. 4231 With no specific regulatory requirements dictating input water quality to a PSG, pretreatment equipment may vary 4232 widely. In general, less pretreatment equipment requires more rigorous sampling strategies to ensure that issues are 4233 detected and addressed before they impact PS quality. Conversely, in a situation where WFI is used as feed water to 4234 a PSG, sampling of the PS may be reduced because of the sampling already performed on the WFI feed water. 4235 PSG designs may vary significantly based on incoming water quality and pretreatment equipment and may include 4236 mist eliminators, non-condensable gas removal, and adjustments to the blow down rate. 4237 4238 This Document is licensed to 4239 ISPE Good Practice Guide: 4240 Page 77 4241 Sampling for Pharmaceutical Water, Steam, and Process Gases 4242 A more detailed discussion of this topic is covered in Appendix 3. 4243 The purification system generating the water feeding a PSG typically needs the same monitoring and maintenance 4244 attention as pretreatment systems feeding distillation units that generate WFI. This monitoring can help to optimize 4245 control of the pretreatment unit operations and steam generation process Monitoring could also serve to alert the user 4246 to take remedial action to avoid specification failures in the steam. 4247 Though a water phase change process is reasonably efficient at separating pure water vapor from the chemical 4248 contaminants that stay behind in the evaporator, the separation is not absolute due to potential entrainment of mist 4249 from the incoming water (and its chemical contaminants) along with the generated steam. 4250 It is commonly accepted that the phase change and mist elimination processes of distillation will purify that feed water 4251 by at least three orders of magnitude (≥ 1000-fold or ≥ 99.9% impurity reduction), allowing no more than 0.1% of the 4252 feed water impurities to remain in the distillate or steam (see the ISPE Baseline® Guide: Water and Steam Systems 4253 (Second Edition) [7]). 4254 However, there may be important distinctions between the generation of WFI by distillation units and the generation 4255 of PS by steam generators. These distinctions relate to the nature of their source water, the unit operations designed 4256 to prepare that water for the generation of steam, and the associated mist elimination and blow down designs, all of 4257 which could impact the quality and performance of the PSG and pretreatment system. 4258 3.8.1 4259 Source Water 4260 3.8.1.1 4261 Pharmaceutical Grade Feed Water from WFI Systems 4262 Organizations may choose to feed their PSGs with WFI, usually from a POU valve that is part of the WFI loop, 4263 although this is not required by compendia or regulation. 4264 When the feed water to a PSG already has very low or undetectable levels of endotoxin, and is otherwise chemically 4265 pure, no additional PSG feed water pretreatment system monitoring is required. When properly operated, the 4266 chemical attributes of the PSG feed water are essentially the same as those of the steam it generates. 4267 The WFI system pretreatment process should be monitored and controlled in order for the distillation process to 4268 produce WFI. For further information on sampling WFI systems, see Chapter 2 of this Guide. 4269 3.8.1.2 Pharmaceutical Grade Feed from Purified Water Systems 4270 The source water should already be chemically pure and sufficiently free from chemical contaminants (e.g., hardness 4271 and silica) that could damage the steam generator equipment, as well as from high levels of inorganic impurities and 4272 TOC. Organizations may choose to feed their PSGs with water from a PW distribution loop that may also feed other 4273 manufacturing uses, although this is not required by compendia or regulation. 4274 However, with PW, there is no requirement that it is free from bacterial endotoxins; therefore, without specific 4275 microbial and endotoxin control precautions, excessive levels of endotoxin may be present. 4276 Since PSGs have a finite impurity removal capability (including endotoxin), the endotoxin levels in the PW feeding the 4277 PSG should be monitored to ensure that the removal capability of the steam generator is not overwhelmed. However, 4278 assuming a minimum 3-log endotoxin removal capability and proper equipment operation, the endotoxin level in the 4279 feed water would have to be 100 EU/mL or greater to be problematic. PW systems with these levels of endotoxin 4280 typically occur only where there is very poor microbial control in the purification and/or distribution systems. Thus, 4281 depending on the requirement for and success of microbial control in the PW system and its historical endotoxin 4282 performance, monitoring the endotoxin level in that feed water from the PW distribution system may not be necessary. 4283 Note: that this 3-log removal capability is a generalized capability that may vary. 4284 4285 This Document is licensed to 4286 Page 78 4287 ISPE Good Practice Guide: 4288 4289 Sampling for Pharmaceutical Water, Steam, and Process Gases 4290 Depending on the PSG design (as discussed below), the log reduction capability may be higher, or it could be lower, 4291 especially in the absence of mist elimination. Although endotoxin is typically not monitored in a PW system used for 4292 manufacturing, if this PW is used as feed water for some PSG designs, monitoring endotoxin in the PW at appropriate 4293 intervals would be indicated. 4294 3.8.1.3 Potable Water 4295 The USP monograph for PS requires that it is prepared from water that meets at least the drinking/potable water 4296 requirements established by the US, EU, Japan, or the WHO drinking water guidelines. This is the same requirement 4297 for PW and WFI. However, in lieu of providing the feed water to the steam generator from a PW or WFI distribution 4298 system outlet, the user may use a dedicated potable water pretreatment system to provide this feed water. Although 4299 there is no compendial requirement that this water first meets PW or WFI specifications, it should be suitable feed 4300 water for the steam generator in terms of its chemical content that could affect the steam generator’s maintenance 4301 (e.g., low levels of hardness, chlorine and silica content could create adverse effects), and final quality (i.e., low 4302 enough levels of conductivity, TOC, and endotoxin that it does not exceed the purification capabilities of the PSG, 4303 and absence of ammonia if chloramines are present in the potable water supply). This is discussed below. The 4304 source of the potable water supply could have a dramatic impact on its starting purity, and therefore, on the required 4305 pretreatment. In this case, a well-designed and well maintained pretreatment system would be required to assure 4306 proper PSG operation. 4307 Based on the initial cost and the cost of maintenance/operation, it may be less costly to include additional capacity 4308 in the feed water system for PW and WFI systems for the PS system, than to use a dedicated and separate 4309 pretreatment system for the PSG. 4310 3.8.2 4311 Steam Generator Mist Elimination Capability 4312 The USP monograph for pure steam specifically states that the pure steam generation process uses a process 4313 that “prevents source water entrainment”. There are many pure steam, or clean steam, systems using PSGs that 4314 have no specific mist elimination capability, particularly in company locations either outside the jurisdiction of USP 4315 or for systems constructed prior to the USP pure steam monograph becoming official on 1 April 2006 [27]. These 4316 generators likely also have poorer log reductions of feed water contaminants than their mist eliminating counterparts. 4317 Well designed and well operated systems using WFI as steam generator feed water essentially have no need for 4318 mist elimination since the feed water quality is already suitable for steam, so mist removal is of lesser concern. These 4319 users may have lower risk of unacceptable pure steam condensate relative to conductivity, TOC or endotoxin content 4320 due to the absence of mist elimination. Therefore, there may be no need to monitor the feed water for these critical 4321 attributes, even with a PSG without any mist elimination capability, as long as the WFI feed water complies with USP 4322 specifications and is not compromised by other aspects of the steam generation equipment or intermittent operation 4323 that may allow contaminants to develop in the PSG feed water from the pretreatment/purification system. 4324 When using PW as steam generator feed water it should be understood that excessive endotoxin levels could be 4325 present, unless controls are active. As discussed above, endotoxin in this feed water should be sufficiently monitored 4326 (in addition to all other PW specifications) to ensure that the endotoxin levels continue to be low enough to be 4327 effectively removed by the PSG. 4328 A standalone PSG feed water system that is either not well designed or not properly maintained carries a high risk of 4329 contaminant carry over into the PS. These systems can be significantly more susceptible to the variable quality of the 4330 incoming potable water and could also experience endotoxin and TOC generated in situ by the pretreatment system. 4331 Pretreatment systems feeding steam generators with no mist elimination are likely to need the rigorous monitoring of 4332 CQAs. 4333 4334 This Document is licensed to 4335 ISPE Good Practice Guide: 4336 Page 79 4337 Sampling for Pharmaceutical Water, Steam, and Process Gases 4338 3.8.3 4339 Dissolved Gas Removal Capability 4340 Without a dissolved non-condensable gas removal process prior to the PSG or as part of the steam generation 4341 process, these gases would come out of solution during the water vaporization process, mix with, and be present 4342 in the distributed steam. If little steam is actively being used and steam loss is primarily due to in situ condensation 4343 and removal at steam traps, these non-condensable gases can accumulate in the distributed steam piping system. 4344 If this steam is used for a porous load sterilization process, it could dramatically reduce the efficacy of the steam 4345 sterilization process. This is the main reason that the BS EN 285:2015 [19] and HTM 2010 [20] standards for non- 4346 condensable gases were created for steam used in sterilizers. 4347 There is no simple direct test for these dissolved gases in feed water, but there are correlations that may be used. 4348 Dissolved gases have a higher solubility in cold water than in hot water. When the feed water to the PSG is already 4349 hot (as with many hot WFI feed water systems), the hot water will contain very little dissolved gas; therefore, steam 4350 systems whose feed water is hot normally have few problems with non-condensable gases in the distributed steam. 4351 If the feed water to the generator is at ambient temperature where higher levels of dissolved gases will be present, 4352 it has the potential for containing dissolved gases (particularly CO2), and at relatively high concentrations. If this 4353 ambient feed water is extremely high quality (very low conductivity) due to nearly complete deionization, it is unlikely 4354 there will be sufficient dissolved CO2 to be problematic. There could, however, be other dissolved gases such as 4355 nitrogen and oxygen present, especially if the feed water uses any ozone process for microbial or TOC control. 4356 Control of dissolved gas in PSG feed water is usually accomplished by pretreatment unit operations performing gas 4357 stripping or carbon dioxide removal or by using high efficiency deionizers. The performance of these unit operations 4358 should be monitored when ambient water is fed to steam generators whose operation must comply with BS EN 4359 285:2015 [19] or when the application dictates. 4360 3.8.4 4361 Blow Down Adjustment 4362 Monitoring of the blow down discharge coupled with periodic testing of conductivity, TOC, and endotoxin in the 4363 feed water may be appropriate to verify a blow down rate aligned with the manufacturer’s recommendations and 4364 guidelines. 4365 3.8.5 4366 Potable Water Chloramine Use 4367 The use of chloramines in potable water is one of the approaches encouraged by EPA to avoid the in situ generation 4368 of carcinogenic disinfection byproducts such as trihalomethanes (THMs) and haloacetic acids (HAAs). All incoming 4369 water to a PSG (as well as to a distillation unit) should be dechlorinated to avoid corrosion of the metals in the PSG. 4370 When chloramines are removed in a pretreatment step, however, ammonia may be present. If the ammonia is not 4371 removed prior to the PSG (or the distillation process), it can vaporize along with the steam, remain in the steam, and 4372 condense along with the distillate or steam condensate when it is sampled from the distribution system sampling 4373 ports, causing it to fail its conductivity specifications. 4374 Ammonia that is not removed from, and is allowed to enter the distillation unit or PSG, may produce conductivity 4375 spikes in the resulting distillate or steam condensate that fails to meet its conductivity requirements. 4376 Additionally, a process that removes chlorine may not be as effective in removing chloramines. If this occurs in a 4377 standalone PSG pretreatment system, serious corrosion of the steam generator or distillation unit would result. 4378 Further, ammonia exists in different states that are pH dependent and easy to control with the proper equipment. 4379 Therefore, it is recommended that the user monitor both ammonia and chlorine levels (both free and total) after 4380 the dechlorination step in the pretreatment train. A change from chlorine to chloramines by the water utility without 4381 notification can cause irreparable equipment damage and noncompliant WFI or PS, which may not be discovered 4382 immediately based on sampling frequency, potentially affecting finished product quality and impacting patient safety. 4383 4384 This Document is licensed to 4385 Page 80 4386 ISPE Good Practice Guide: 4387 4388 Sampling for Pharmaceutical Water, Steam, and Process Gases 4389 Often, ammonia can be removed from pretreated water by a polishing softener, acidifying the feed water to an RO 4390 system, or by a full deionization process, but because of variations in ammonia removal efficiency, monitoring for 4391 ammonia could be an important process control effort to avoid specification failures. 4392 Both chlorine and chloramines are considered acceptable potable water disinfectants. But if the potable water 4393 supplier neglects to inform pharmaceutical users of an impending change from one disinfectant to the other, there 4394 could be far reaching implications in finished water or steam quality. Hence maintaining lines of communication with 4395 potable water suppliers is important. 4396 3.8.6 4397 Anti-scaling Steam Additives 4398 The use of steam additives to avoid metal corrosion in the steam generator and distribution piping is not allowed for 4399 PS systems. Many of these plant steam additives” tend to carry over into the steam and would cause the steam to fail 4400 its quality attributes. Depending on the nature of the additive, it might be detectable in the feed water as increased 4401 conductivity or TOC, but the best approach is by a specific test for each compound. 4402 The nature of these additives may make them difficult to remove, so monitoring for these additives after their removal 4403 process step is important for ensuring that the pure steam meets its specifications when recycled condensate is used. 4404 Better than that, if the plant steam condensate system is kept separate, the potential for a problem will be eliminated. 4405 Many plant steam systems capture the condensate from their steam traps and reuse this water to feed the plant 4406 steam generators. If the condensate collection system is shared between the plant steam and pure steam systems 4407 (this practice is not recommended). 4408 3.8.7 4409 Monitoring Locations and Frequency 4410 There are no pretreatment system monitoring requirements (other than for using potable water as the feed water) that 4411 are mandated by the compendia or FDA for pure steam systems. 4412 Pretreatment operation and monitoring should follow the guidelines established in this and other ISPE guidance 4413 documents for compendial water and steam systems (see the ISPE Baseline® Guide: Water and Steam Systems 4414 (Second Edition) [7] as well as the ISPE Good Practice Guide: Approaches to Commissioning and Qualification of 4415 Pharmaceutical Water and Steam Systems (Second Edition) [13]). 4416 Specific chemical attributes (including endotoxins) may need to be periodically monitored in the: 4417 • 4418 Incoming source water 4419 • 4420 Feed water to the steam generator 4421 • 4422 Between pretreatment unit operations 4423 The frequency should be appropriate to each steam system. 4424 4425 This Document is licensed to 4426 ISPE Good Practice Guide: 4427 Page 81 4428 Sampling for Pharmaceutical Water, Steam, and Process Gases 4429 4 Process Gases 4430 4.1 4431 Introduction 4432 This chapter contains information and best practices gathered from several major pharmaceutical organizations 4433 listing their current practices for assuring the quality of gases used in their operations. Testing may be performed by 4434 the pharmaceutical manufacturer or by a certified independent laboratory. 4435 The quality of product contact compressed gases should be assured through adequate design, appropriate 4436 controls (e.g., change control or change management procedures), and routine maintenance of the system. If 4437 adequate controls are in place, critical compressed air and other process gas systems typically require analytical 4438 testing on a periodic basis to confirm that the system is operating under a state of control. Routine release testing 4439 (e.g., identification testing) of compressed air or process gases furnished by a qualified supplier may be required 4440 for release of incoming raw materials. If it is determined that analytical testing of a distribution piping system is 4441 necessary, the specification limits for the testing should be suitable for the intended use of the gas. 4442 4.1.1 4443 Types of Process Gases Covered in this Guide 4444 This Guide considers the most commonly used gases.9 The process gases covered in this Guide include: 4445 • 4446 Compressed air 4447 • 4448 Argon 4449 • 4450 Carbon dioxide 4451 • 4452 Nitrogen 4453 • 4454 Oxygen 4455 Process gases may be generated on site or may be provided by a qualified supplier as portable tanks, compressed 4456 gas cylinders, cylinders containing liquefied gases, or other types of containers. There may be variations in the 4457 number of sample points as well as the frequency of sampling. Decisions should be made after a risk analysis is 4458 performed. 4459 4.1.2 4460 Regulatory Requirements and Consensus Standards 4461 No well-defined regulatory requirements exist regarding the frequency, extent, or method of routine sampling for 4462 compressed air or other process gases. None of the pharmacopeia monographs specifies limits for particles or 4463 total viable counts (bacteria), but some consensus standards and US FDA guidance documents provide additional 4464 information and should be consulted as part of the risk analysis process. 4465 Some monographs stipulate that the moisture content of compressed air and other process gases should be below 4466 -46°C (-50°F).10 USP monographs indicate no standard methods for air sampling. 4467 9 These gases reflect the gases covered in the ISPE Good Practice Guide: Process Gases [29]. 4468 10 Equivalent to 67 ppm for compressed air. 4469 4470 This Document is licensed to 4471 Page 82 4472 ISPE Good Practice Guide: 4473 4474 Sampling for Pharmaceutical Water, Steam, and Process Gases 4475 The 2004 US FDA “Guidance for Industry Sterile Drug Products Produced by Aseptic Processing – Current Good 4476 Manufacturing Practice” 2004 [30] provides the following nonbinding recommendation regarding the filtration of 4477 compressed air and other process gases: 4478 4479 “A compressed gas should be of appropriate purity (e.g., free from oil) and it’s microbiological and particle quality 4480 after filtration should be equal to or better than that of the air in the environment into which the gas is introduced. 4481 Compressed gases such as air, nitrogen, and carbon dioxide are often used in cleanrooms and are frequently 4482 employed in purging or overlaying.” 4483 ISO 8573 [31] is a universal standard that applies to all industrial uses and provides some information that is 4484 applicable to the pharmaceutical industry. 4485 4.1.3 4486 Critical Quality Attributes of Compressed Air and Process Gases 4487 A sensible and logical method to determine the requirements for compressed air or process gases is to review the 4488 role played by the gas in the actual process in which it is used. Compressed air and process gases are utilities that 4489 are widely used in the pharmaceutical industry for many different applications. Examples include: 4490 • 4491 Displacement or conveyance of product from one location to another 4492 • 4493 As a feed source to fermentation processes 4494 • 4495 Actuating pneumatic valves in classified and controlled areas and the resulting venting of exhaust gases into 4496 processing spaces where product contact may occur 4497 • 4498 Occupying head space in lyophilization vials, solution vials, ampoules, etc. 4499 • 4500 Occupying head space or acting as a blanket in a bioreactor 4501 • 4502 Drying of product in fluidized beds 4503 • 4504 Drying of product contact equipment surfaces 4505 • 4506 As a carrier in spray tablet coatings 4507 • 4508 As a carrier to suspend an API during micronizing or milling steps 4509 • 4510 As a carrier to pneumatically convey dry powders 4511 • 4512 As a blanket to provide an inert atmosphere within API vessels 4513 • 4514 As a blanket or to pressurize WFI storage tanks where the compressed air or process gas comes into contact 4515 with WFI water, which may have direct product contact 4516 Gaseous utilities may impact the product quality directly or indirectly as an: 4517 • 4518 Excipient 4519 • 4520 Material 4521 • 4522 Raw material 4523 • 4524 In process material 4525 4526 This Document is licensed to 4527 ISPE Good Practice Guide: 4528 Page 83 4529 Sampling for Pharmaceutical Water, Steam, and Process Gases 4530 • 4531 Inactive ingredient 4532 • 4533 Component or ingredient 4534 • 4535 Process aid 4536 The broadest of these definitions is component or ingredient. For components/ingredients, there are regulatory 4537 expectations of specifications, and whenever there are specifications, sampling programs should be devised to 4538 document that the utility is operating under a state of control and satisfying the component specifications, shown in 4539 Figure 4.1. 4540 Figure 4.1: Establishing Sampling Programs for Components/Ingredients 4541 CQAs are best determined by risk analysis when considering the intended use of the compressed gas in the facility. 4542 Without restricting the type of gas under consideration, the CQAs are most generally defined as: 4543 • 4544 Moisture content 4545 • 4546 Total hydrocarbons 4547 • 4548 Particles 4549 • 4550 Viable particles (bacteria) 4551 During the design and construction phases, other potential contaminants may be introduced as distribution systems 4552 are built, such as cutting oils, cleaning agents, or other materials used during system construction or subsequent 4553 maintenance activities. A risk analysis should be conducted to assess the likelihood of these other contaminants 4554 being present, to assist with the mitigation of their sources, and to develop testing strategies that may be applicable. 4555 4.2 4556 Sampling Locations 4557 Sample points may be installed for varied reasons in a compressed air or process gas system, including: 4558 • 4559 Quality control purposes 4560 • 4561 Process control or diagnostic purposes 4562 4563 This Document is licensed to 4564 Page 84 4565 ISPE Good Practice Guide: 4566 4567 Sampling for Pharmaceutical Water, Steam, and Process Gases 4568 Sample points intended for quality control purposes should be determined based on the CQA being monitored and on 4569 a suitably developed sampling plan. 4570 Sample points intended for process control or diagnostic purposes should: 4571 • 4572 Be used as needed to ensure that processes are under control 4573 • 4574 Be regularly maintained 4575 • 4576 Not introduce a risk of contaminating the system 4577 • 4578 Not be included in the sampling plan 4579 The determination of sample point location should be initiated during system design. Sampling locations should also 4580 be reviewed again during system modification to ensure appropriate sample ports are installed to meet the quality 4581 control and process control monitoring requirements established for the system. 4582 Risk analysis tools may be utilized to define the sampling strategies for compressed air and process gas systems. 4583 Sample point locations may vary for quality control and process control purposes depending on which category the 4584 gas system falls under, i.e.: 4585 1. 4586 Compressed air systems 4587 2. 4588 Cylinder systems and liquefied gases provided in cylinders 4589 3. 4590 On site generation and bulk storage systems 4591 Compressed air systems are typically owned by the facility, whereas cylinder systems, liquefied gas systems, on site 4592 generation and bulk storage systems are typically rented or leased from qualified specialty gas suppliers. 4593 Expectations for sampling points for compressed air and process gas systems include: 4594 • 4595 Sampling points should be easily accessible 4596 • 4597 The location of a sampling point should avoid locations common to other clean utilities, e.g., high purity water or 4598 PS to prevent cross contamination 4599 • 4600 Sampling points should be clearly and uniquely identified on P&ID drawings and tagged in the field 4601 • 4602 A risk analysis should be performed to: 4603 - 4604 Determine the CQAs for the gas 4605 - 4606 Identify sampling points required for quality control 4607 - 4608 Determine CQAs from samples taken from each quality control sample point 4609 - 4610 Identify sampling points required for process control 4611 - 4612 Determine important process parameters for each process control sample point 4613 • 4614 Test results should be trended and reviewed annually, and sampling frequencies adjusted accordingly. 4615 4616 This Document is licensed to 4617 ISPE Good Practice Guide: 4618 Page 85 4619 Sampling for Pharmaceutical Water, Steam, and Process Gases 4620 Typically, a sampling location should reflect the quality of a gas at the POU. However, for a gas line fitted with a POU 4621 sterile filter, the sampling point should be located before the sterile filter. Sampling in this location determines the 4622 quality of gas delivered to the sterile filter, while simultaneously avoiding the risk of product contamination from a 4623 sample point downstream of the filter. 4624 4.2.1 4625 Sampling Locations for Pharmaceutical Compressed Air Systems 4626 Pharmaceutical compressed air is used in a pharmaceutical industry for many purposes. Pharmaceutical compressed 4627 air may also be referred to as compressed dry air or process air. Air that comes in direct contact with product and/or 4628 product contact surfaces can affect product quality. Compressed air quality should consistently meet specifications 4629 established for the process and should not adversely impact product quality. 4630 In a compressed air system, sampling points may be installed at specific locations for quality control and/or process 4631 control purposes. For example, a sampling port installed before a final in-line filter on the central system would 4632 be installed to document process control up to this point in the system, but may not be part of the quality control 4633 sampling system. The intended purpose of each sampling point should be specified in the quality plan for the system. 4634 Examples of process control sample point locations intended to assess the compressed air quality include: 4635 • 4636 After the compressor, dryer, and coalescing filter on the central system 4637 • 4638 Before and after any in-line filters on the central system and prior to the distribution system 4639 Examples of quality control sample locations include: 4640 • 4641 Following any central treatment process intended to control the levels of an identified CQA 4642 • 4643 The central supply header of the distribution system 4644 • 4645 The most distant POU from the central system on each floor 4646 • 4647 Prior to a POU final filter if the POU cannot be sampled 4648 Results from samples taken from distant ports would be compared to results from the supply header and used to 4649 determine if contamination issues exist in the distribution pipework. Excursions above Alert and Action Levels would 4650 likely trigger sampling at additional points. 4651 CQAs for the compressed air system should be determined by coupling risk analysis tools with the intended use of 4652 the compressed air in the facility, but typically involve sampling for: 4653 • 4654 Moisture content 4655 • 4656 Total hydrocarbons 4657 • 4658 Viable and nonviable particles 4659 With these CQAs, the following sampling locations represent sampling that would be performed for quality control 4660 purposes. Sampling for moisture is typically performed: 4661 • 4662 Following a drying process, before compressed air is introduced to the distribution piping system 4663 • 4664 At the farthest end of a distribution line serving a floor, prior to the farthest POU in the system 4665 4666 This Document is licensed to 4667 Page 86 4668 ISPE Good Practice Guide: 4669 4670 Sampling for Pharmaceutical Water, Steam, and Process Gases 4671 Sampling for total hydrocarbons (oil) is typically performed: 4672 • 4673 After the final coalescing or oil removal filter in the system (Note: hydrocarbon testing may be performed prior 4674 to these oil removal processes, or during intermediate steps if several are involved, but these samples would 4675 represent process control sample points) 4676 • 4677 At the end of a distribution line serving each floor, prior to the most distant POU in the system 4678 • 4679 At critical POUs 4680 Sampling for Non-Viable Particles and Viable Particles (NVP/VP) is typically performed: 4681 • 4682 After the final central system filter as compressed air is introduced to the distribution piping system 4683 • 4684 At the most distant end of a distribution line serving each floor, prior to the most distant POU in the system 4685 • 4686 At critical POUs 4687 During the construction phase other potential contaminants, such as cutting oils or cleaning agents used in system 4688 construction or subsequent maintenance activities, should be considered. 4689 A risk analysis should be performed to determine the likelihood of other contaminants which may be present and to 4690 assist with the mitigation of their sources, or the development of testing strategies where applicable. When a system 4691 has been verified through sampling and testing to be clean, testing frequencies within the distribution system may be 4692 reduced because distribution piping systems prevent the ingress of contaminants. For additional discussion on this 4693 matter, refer to the ISPE Good Practice Guide: Process Gases [29]. 4694 Note: Figure 4.2 is provided for illustrative purposes only and should not be considered as an industry standard or 4695 recommendation for system design. 4696 Figure 4.2: One Potential Compressed Air System Design with Sampling Points for Both Process Control (PC) 4697 and Quality Control (QC) Purposes 4698 Figure 4.2 shows a typical compressed air system design utilized in a pharmaceutical manufacturing facility. Sampling 4699 ports have been positioned throughout the system. Most sampling ports have been installed for process control 4700 purposes. Where possible, compressed air quality attributes may be monitored using on-line instrumentation. 4701 Sampling locations monitored for quality control purposes should be incorporated into an established sampling plan. 4702 The types of sampling and locations of sampling points include: 4703 • 4704 Moisture content monitoring by dew point indicator 4705 - 4706 following the central system drying process following the furthest final filter in the distribution piping system 4707 4708 This Document is licensed to 4709 ISPE Good Practice Guide: 4710 Page 87 4711 Sampling for Pharmaceutical Water, Steam, and Process Gases 4712 - 4713 at critical POUs 4714 • 4715 Hydrocarbon monitoring 4716 - 4717 following the central system microcoalescer 4718 - 4719 prior to the furthest POU in the distribution piping system 4720 - 4721 at critical POUs 4722 • 4723 Nonviable and viable particle monitoring 4724 - 4725 following the central system final filter 4726 - 4727 prior to the furthest final filter in the distribution piping system 4728 - 4729 at critical POUs 4730 4.2.2 4731 Sampling Locations for Process Gases Delivered in Cylinders and for Liquefied Gases 4732 Delivered in Cylinders 4733 A typical example of a cylinder based compressed gas system is shown in Figure 4.3 and that of a liquefied gas 4734 system are shown in Figure 4.4. Since there are fewer steps in each of these systems, fewer sample points are 4735 required as the finished product is supplied by a qualified supplier and piped to each POU by connecting the cylinder 4736 to the distribution piping system. Sampling ports or valves are usually installed at the points as indicated. Depending 4737 on the critical quality attributes established for the process gas, the following sample ports would be utilized for quality 4738 control purposes, with other indicated ports being used for process monitoring purposes: 4739 • 4740 Moisture (if required) by dew point indicator following final filter, prior to furthest final filter, or at critical POUs 4741 • 4742 Nonviable and viable particles (if required) would be sampled prior to the furthest final filter or at critical POUs 4743 In many cases, there is no POU filter and the actual POU would be used as the sampling location. 4744 Figure 4.3: Sampling Points for Gases Delivered in High Pressure Cylinders 4745 4746 This Document is licensed to 4747 Page 88 4748 ISPE Good Practice Guide: 4749 4750 Sampling for Pharmaceutical Water, Steam, and Process Gases 4751 Figure 4.4: Sampling Points for Gases Delivered from Liquefied Storage Tanks 4752 4.2.3 4753 On Site Generation and Bulk Storage Systems 4754 On site generation and bulk storage systems vary significantly depending on the process gas involved, but should 4755 follow the general rules outlined for compressed air systems discussed earlier in this chapter. Sample locations 4756 should be carefully reviewed and installed for both process control and quality control purposes in a similar fashion to 4757 those outlined for compressed air systems. 4758 4.3 4759 Sampling Plan (Tests Performed, Frequency and Duration) 4760 4.3.1 4761 General Considerations 4762 Compressed air or other process gases should be monitored at a frequency that would assure that gas does not 4763 adversely affect the environment or the product. Routine monitoring frequency should be similar to the frequencies 4764 established for the cleanroom, but these frequencies can be adjusted using the risk analysis tools once a base of 4765 historical data has been generated. 4766 4767 This Document is licensed to 4768 ISPE Good Practice Guide: 4769 Page 89 4770 Sampling for Pharmaceutical Water, Steam, and Process Gases 4771 4.3.2 4772 Risk Analysis for Sampling 4773 Risk analysis of process gas systems should be performed upon initial system design and construction, or following 4774 significant system modifications. This risk analysis should identify sampling points that may present a risk to: 4775 • 4776 Product 4777 • 4778 Environment 4779 • 4780 The integrity of the gas system 4781 These risks should be based on, but not limited to, system validation, routine monitoring, and system application. 4782 Risk analysis principles should include the inherent risk of both viable and nonviable particulate level requirements 4783 that may already be controlled through adequate system design and utilization of critical POU filtration and 4784 maintenance. 4785 Compressed air and process gas systems typically are designed, controlled, and maintained in a dry state (i.e., very 4786 low dew point), which inherently reduces the risk of microbial growth. Filter selection is dependent on the intended 4787 use of the compressed air or process gas at the POU, and the particulate requirements/tolerances are typically 4788 defined during the system design phase. 4789 4.3.3 4790 Background Philosophy 4791 Sampling and testing of compressed air and process gases can be broadly classified based on the phases of system 4792 life cycle, i.e.: 4793 • 4794 Commissioning (prequalification) 4795 • 4796 Initial performance qualification 4797 • 4798 Requalification during major or minor changes to the system 4799 • 4800 Routine monitoring 4801 4.3.3.1 Sampling for Commissioning (Prequalification) 4802 Sampling during commissioning or prequalification of a compressed air system is recommended for gas generation 4803 and distribution to prove the performance of individual unit operations like compressor or dryer. 4804 Process compressed gas distribution systems from cylinders do not normally require any sampling during 4805 commissioning. 4806 4.3.3.2 Sampling for Performance Qualification 4807 Qualification of compressed air or process gases intends to verify the distribution piping is clean and free of oil and 4808 contains acceptable levels of other impurities/contaminants. Compressed gas system qualification typically involves a 4809 minimum of three days of sampling and testing at specified sampling locations. 4810 CQAs that are monitored at each sampling location in a compressed air system include: 4811 1. 4812 Total particulates (viable and nonviable) 4813 2. 4814 Viable particulates 4815 4816 This Document is licensed to 4817 Page 90 4818 ISPE Good Practice Guide: 4819 4820 Sampling for Pharmaceutical Water, Steam, and Process Gases 4821 3. 4822 Total hydrocarbons (both aerosol droplets and vapors) 4823 4. 4824 Dew point (moisture content) 4825 Sampling for process gases provided in cylinders, or as liquefied gases, may require less sampling, as some of these 4826 CQAs may have been eliminated during the risk analysis process. Sampling may be performed from the sampling 4827 locations after the final filter and critical POU at the farthest end of the longest pipe run of each floor of the building 4828 served by the system 4829 The purity of the gas delivered by cylinders may be verified through Certificates of Analysis (CoA) provided by the 4830 manufacturer of the particular gas. If the process gas is considered a raw material, then raw material qualification is 4831 performed to provide assurance of the quality of the gas being used. Periodic testing, sampling, or third party testing 4832 may be required in cases where the typical CoA provided by the manufacturer has not indicated all the specification 4833 limits required for the raw material. It is important to remember that the pharmaceutical manufacturer is responsible 4834 for all quality attributes that impact raw material and finished product quality. 4835 4.3.3.3 Sampling after Major/Minor Changes in the System 4836 Qualification involving gas sampling and testing should be required for all new systems and also for any change to 4837 the design of the existing system. A major modification includes any change to a system that has the potential to 4838 affect the quality of the compressed air or process gas that is being delivered through the distribution system to the 4839 process. 4840 Examples of a major change to an existing compressed air system includes, but not limited to, changes in a unit 4841 treatment process, distribution piping changes, or additions to the piping system. 4842 Minor changes would involve the addition of a single valve or a minor system repair. Normal routine sampling and 4843 testing may continue but additional testing may be required based on the extent of the system alteration and the risk 4844 to product. Established procedures should define and identify the applicable test requirements. 4845 4.3.3.4 Routine Monitoring 4846 Qualified compressed air and process gas systems should be monitored at established frequencies. In cases where 4847 the gas is entering a classified area, it is required to at least meet the room classification limits established for the 4848 cleanroom environment. 4849 The frequency of routine quality control monitoring at representative points of use and other established points should 4850 be developed based on risk analysis tools. Suggested monitoring frequencies for CQAs identified by risk analysis 4851 processes are indicated in Table 4.1. It is important to note that these attributes may not apply to all gases being tested. 4852 Table 4.1: Recommended Routine Monitoring Frequencies for Compressed Air Systems 4853 Test 4854 Frequency/Location 4855 Nonviable Particles 4856 Tested every three months on a rotating basis for sampling locations following the central 4857 system final filter and at predetermined locations on the horizontal piping run for each floor 4858 Viable Particles 4859 Tested every three months on a rotating basis for sampling locations following the central 4860 system final filter and at predetermined locations on the horizontal piping run for each floor 4861 Dew Point 4862 (moisture) 4863 Monitored continuously using on-line dew point instrumentation following compressed air 4864 dryer 4865 Hydrocarbon (oils) 4866 Annual sampling and testing following the coalescing filter to verify the hydrocarbon 4867 content of the compressed air system 4868 4869 This Document is licensed to 4870 ISPE Good Practice Guide: 4871 Page 91 4872 Sampling for Pharmaceutical Water, Steam, and Process Gases 4873 The sampling location on a gas line fitted with a 0.2 µm/0.22 µm final hygienic filter should be located before the final 4874 filter. Sampling before the POU hygienic filter should represent the quality of air delivered to the POU or process 4875 and at the same time avoid any contamination risk posed by a sample valve located downstream of the final filter. A 4876 sample should represent the quality at a POU and the sampling location used should be demonstrated to represent 4877 the quality at POU. 4878 4.4 4879 Sample Valve Design 4880 Sample valves located within a central compressed air or process gas treatment system may be ball valves when 4881 installed prior to the central system final filter and prior to distribution. Sampling valves and POU valves located in 4882 the distribution piping system have stricter requirements and are usually diaphragm valves. For example, when the 4883 installation of a hygienic, pharmaceutical grade 0.22 μm filter is required in close proximity to the POU, it should be 4884 installed in a suitably clean housing attached to the POU valve to minimize contamination of the gas stream. The 4885 filter housing with the filter placed in it requires the ability to be sterilized and kept sterile until it is placed in the piping 4886 system for use. 4887 4.5 4888 Sampling Techniques for Compressed Air and Process Gases 4889 Sampling techniques will vary based on the contaminant being monitored. Each will be discussed below. 4890 4.5.1 4891 Total Particles, Non-viable and Viable 4892 The selection of a sampling method is dependent on the concentration and size range of the impurity/contaminant 4893 being monitored. Other evaluation criteria include sampling times, air volumes, and allowable outlet pressure 4894 restrictions. 4895 Sampling equipment should be within calibration, if applicable. Refer to equipment manufacturer’s guidelines for 4896 particle size and concentration measurement limits. 4897 Any gauges, adaptors, fittings, or tubing used to connect the sampling equipment to the sampling outlet should be 4898 clean and particle free, and should not introduce contamination to the collected sample. The preferred connection 4899 to sampling equipment should be as short as practical, constructed of stainless steel or other suitable material, and 4900 without sharp bends (approximately a 6 inch radius if bends are present), and non-restrictive. Softer metals such as 4901 brass or copper can produce particles when joined and should not be utilized. 4902 Methods include laser particle counting with a high-pressure diffuser, condensation nucleus counting coupled with 4903 scanning mobility particle sizing, differential mobility analysis, and membrane capture as covered in ISO 8573-4:2001 4904 [32]. Other acceptable methods may be available. 4905 It is important to note that flow meters, control valves, pressure gauges, quick connect fittings and polymeric tubing 4906 can shed particles into the sampling stream. If these items must be used, specify non-shedding materials to minimize 4907 the contamination of the collected sample. 4908 When closed, valves will capture particles; and when opened the captured particles will be released. Friction on 4909 the valve seals will also release particles. If a valve is required for flow regulation, specify a high purity diaphragm 4910 or bellows valve; seal material must not be composed of rubber or other polymeric materials, with the exception of 4911 Teflon. Valves and regulators should be fully opened to minimize particle shedding as partially opened components 4912 will shed more particles. 4913 Tubing materials for minimizing particle contamination in order of preference are as follows: stainless steel, 4914 conductive polymer, polyester, vinyl (if plasticizer does not interfere), polyethylene, copper, glass, Teflon, and 4915 aluminum. 4916 4917 This Document is licensed to 4918 Page 92 4919 ISPE Good Practice Guide: 4920 4921 Sampling for Pharmaceutical Water, Steam, and Process Gases 4922 4.5.2 4923 Solid Particles, Viable Bacteria 4924 Viable particles or microbial particles may be monitored using several different techniques. 4925 The slit to agar technique operates via a fixed rate air flow and fixed motor carriage to rotate an agar plate at 4926 constant speed (typically 30 minutes to complete one revolution). Plate position, specifically the media surface is 4927 positioned at a defined distance from the location of the air outlet slit. 4928 The impaction technique is where air is forced perpendicular to the agar surface through a diffuser of fixed holes to 4929 monitor bacterial levels. 4930 Hydrophilic membranes may be utilized. This involves passing a fixed volume of air (or process gas) through the 4931 membrane where the microorganisms are captured. Enumeration using conventional laboratory techniques can be 4932 used to determine bacterial counts. 4933 Real time monitoring is a rapidly evolving technique with great potential and applicability in the pharmaceutical 4934 industry. A sample of the gas stream is passed through an analyzer and the presence of microorganisms is directly 4935 monitored by one of several different techniques as developed by instrument manufacturers. Progress is being 4936 made at a rapid rate in this area by manufacturers as there is a huge demand for on-line, real time microbiological 4937 monitoring. Different analytical techniques may have varying capabilities and limits of detection, but rapid 4938 improvement in this field may change the landscape for real time microbiological monitoring in the very near future. 4939 4.5.3 4940 Water Vapor 4941 If moisture content testing is determined to be necessary, limits for moisture may include applicable EP 4942 pharmacopoeia limits. Moisture levels are typically controlled via adequate system design, operation and 4943 maintenance. Moisture content is typically monitored using on-line instrumentation providing real time information. 4944 For systems where the level of moisture is considered critical, the instrumentation and alarms associated with the 4945 monitoring of moisture levels should be commissioned and qualified. 4946 4947 “Moisture in process gases is usually measured as either dewpoint or pressure dewpoint. Methods include 4948 condensation devices, (Cooled Mirror Dewpoint, Electrolytic, and Impedance Hygrometers, and Polymer Film 4949 Relative Humidity Sensors), instrumentation devices (Chilled Mirror, Electrolytic, and Impedance Hygrometers, 4950 Polymer Film Relative Humidity (RH) Sensors).” [29] 4951 Other acceptable methods may be available. 4952 4953 “When using a hygrometer, the temperature and line pressure should be accounted for during dewpoint 4954 measurements. The instruments may be used in line for continuous monitoring or may be used to periodically 4955 monitor critical POUs. 4956 4957 The type used should be determined by the degree of accuracy required for downstream processes, 4958 maintenance and calibration schedules, costs, and equipment reliability. For on line testing of moisture, indicator 4959 tubes may be used. Ingress of external environmental moisture into the sample path should be avoided.” [29] 4960 4961 Sampling equipment should be maintained in a calibrated state, if applicable. Refer to equipment manufacturer’s 4962 guidelines for concentration measurement limits. 4963 As mentioned earlier, the sampling process and/or equipment should not introduce contaminants to the sample being 4964 collected. 4965 4966 This Document is licensed to 4967 ISPE Good Practice Guide: 4968 Page 93 4969 Sampling for Pharmaceutical Water, Steam, and Process Gases 4970 The following is an excerpt taken from ISO 8573-3:1999 Annex C, C.2.2, page 11 [33]: 4971 4972 “[Flush] sample lines and hygrometers with dry gas, or by evacuating to low pressure. Drive off stray residual 4973 water by baking assemblies if possible ... The lower the moisture content to be measured, the more dramatically 4974 the drying time multiplies... 4975 4976 Select impermeable materials, to avoid inward diffusion of moisture through sampling tubes and enclosures. 4977 Steel and other metals are practically impermeable. [Teflon] is only slightly permeable and will usually be 4978 satisfactory for dew points above -20°C…. Materials such as polyvinyl chloride (PVC), nylon and rubber are … 4979 not really satisfactory in any humidity range. 4980 4981 Surface finish is important [particularly] at very low humidities…. Polished or electropolished stainless steel is 4982 recommended for the best results…. Sample tubing should be as short in length as possible [and] surface area 4983 should be minimized …. [Minimize] the number of connections [where possible]. [Dead legs] in tubing [can trap 4984 moisture and} should be avoided.” 4985 When using chemical reaction tubes (detector tubes) to manually sample for moisture content, care should be taken 4986 to follow the manufacturer’s instructions. After the tips of the detector tube are broken the tube should be inserted 4987 immediately into the sampling device. With training and proper technique, the diffusion of water laden atmospheric 4988 air into the tube, as it is attached to the sampling device is negligible. After the detector tube is removed from 4989 sampling apparatus, immediately read the color change. Detector tube readings provide water concentration in mg/ 4990 m3. It is common practice to then convert the readings to ppm to determine either dew point at atmospheric, system, 4991 or line pressure based on the specifications established for the system. When comparing test results to air quality 4992 requirements make certain that units of measure are compatible [29]. 4993 4.5.4 4994 Total Hydrocarbon: Oil Aerosol and Oil Vapor 4995 Total hydrocarbon includes both oil aerosols11 and oil vapors12, which are detected differently. Both types are typically 4996 associated with oils introduced as part of a compressed air system as oils are typically not present in cylinder based 4997 systems. The selection of a sampling method is dependent on the anticipated level of oil expected from the point 4998 being sampled as well as whether the oil would be present as an aerosol or vapor. 4999 The total hydrocarbon content within compressed air systems is typically addressed through the use of design 5000 or engineering controls by installing oil free compressors, which eliminate the vast majority of hydrocarbon and 5001 related issues. Some compressors are oil free, but there may still be components within the compressor that utilize 5002 oil. Removal of oil aerosols from these systems is performed through the use of coalescing filters located in the 5003 distribution system. Where design or engineering controls are in place, oil aerosols are not typically expected to be 5004 present in the system and annual or less frequent testing or monitoring is required. 5005 Methods for sampling include the use of pre-weighed filter membrane(s) or detector tubes. When using filter 5006 membranes, install in a suitable holder with subsequent analysis by gravimetric or infrared techniques. Obtain a 5007 sufficient air volume to provide reliable results at the limit of detection. 5008 Should oil aerosols not be effectively removed, they may accumulate and coalesce into larger oil droplets and 5009 eventually into oil liquid that may become visible. If oil liquid is observed, the system must be immediately shut down 5010 and the cause of failure determined as oil liquid is intolerable in pharmaceutical manufacturing. 5011 For oil vapors, the selection of a sampling method is dependent on the established air quality limits. ISO 8573-1:2010 5012 [34] considers oil vapor measurement to be optional if the purity limit is > 1.0 mg/m3 due to the low impact on results 5013 higher than 1.0 mg/m3. Other evaluation criteria include sampling times, air volumes, and allowable outlet pressure 5014 restrictions. 5015 11 “Oil aerosol is the term for oils that can form an aerosol that will remain suspended in the air until it impinges against a surface.” [29] 5016 12 Also called total volatile hydrocarbons. 5017 5018 This Document is licensed to 5019 Page 94 5020 ISPE Good Practice Guide: 5021 5022 Sampling for Pharmaceutical Water, Steam, and Process Gases 5023 Typical oil vapor levels reported for breathing air include gaseous hydrocarbons with up to ten carbons (C1-10). ISO 5024 8573-1:2010 [34] defines oil to be a mixture of hydrocarbons composed of six or more carbon atoms (C6+). Method 5025 of collection varies depending on the expected molecular weight of the hydrocarbon vapors to be detected. Glass 5026 or stainless steel cylinders can be used for lower molecular weight hydrocarbons (C1-10). Commercially available 5027 charcoal tubes or cylinders packed with charcoal are suitable for collection of samples expected to contain higher 5028 molecular weight hydrocarbon vapors (C6+). 5029 4.6 5030 Sample Handling 5031 Samples should be analyzed as quickly as possible after collection. 5032 Samples should be clearly and uniquely labeled. Labels for internal analysis should include the following information: 5033 • 5034 Sample or location identification 5035 • 5036 Analysis to be performed 5037 • 5038 Type of preservation required, if any 5039 • 5040 Sampling media identification numbers 5041 • 5042 Sampling parameters (i.e., sampling time, flow rate, temperature, pressure, etc.) 5043 • 5044 Date and time sample collected 5045 • 5046 Sampling technician name and signature 5047 Labels for samples submitted to third parties for analysis should also include the company name and contact 5048 information. 5049 4.6.1 5050 Chain of Custody 5051 Chain of custody documentation establishes that the sample has been properly transmitted from sample collection 5052 through sample analysis and is a recommended practice. If a collected sample is submitted to an outside firm for 5053 analysis, sample transport will be required and the bill of lading or airbill will become part of the chain of custody. 5054 Chain of custody seals or evidence tape may be used but are not mandatory. 5055 4.7 5056 System Monitoring 5057 Monitoring approaches may vary between pharmaceutical companies for process gases and compressed air, but 5058 these variations are always attributed to the products being manufactured at a facility and the corresponding CQAs 5059 that have been identified for the gas during the risk analysis process. 5060 Performing a Hazard Analysis and Critical Control Points (HACCP) may be utilized for the creation of critical utility 5061 system monitoring programs. Typical routine monitoring programs and frequencies have been described in Table 4.1. 5062 5063 This Document is licensed to 5064 ISPE Good Practice Guide: 5065 Page 95 5066 Sampling for Pharmaceutical Water, Steam, and Process Gases 5067 Appendix 1 5068 Appendix 1 5069 5 Appendix 1 – Specification Summary for 5070 5071 Various Non-Pharmacopeial Water Grades 5072 Organization/ 5073 Reference 5074 ISO 3696 (1995) 5075 Water for Analytical Laboratory Use 5076 ASTM D1193-06 (2011)(1) 5077 Standard Specification for Reagent Water 5078 ASTM D5196 5079 (2006) 5080 CLSI 4th Ed 5081 (2006) 5082 Water Grade or Type 5083 Grade 1 5084 Grade 2 5085 Grade 3 5086 Type I 5087 Type II 5088 Type III 5089 Type IV 5090 Standard 5091 Guide for 5092 Bio-App 5093 lications 5094 Grade Water 5095 CLRW 5096 (Specified 5097 quantitative 5098 attributes 5099 only) 5100 Specified Source 5101 and Purification 5102 Approaches 5103 Grade 2 Source; 5104 RO +0.2 µm Filt, 5105 or DI +0.2 5106 µm Filt, 5107 or Re-Dist 5108 (in glass) 5109 Multiple-Dist 5110 or DI 5111 or RO+Dist 5112 Single-Dist 5113 or DI 5114 or RO 5115 < 20 µS/cm 5116 Source 5117 (Dist, equiv); 5118 MB-DI 5119 +0.2 µm Filt 5120 Distillation or 5121 Equiv. 5122 Distillation, 5123 DI, EDI, 5124 and/or RO 5125 +0.45 µm Filt 5126 Distillation, 5127 DI, EDI, 5128 and/or RO 5129 Drinking 5130 Water Source; 5131 Suitable 5132 Process(es) 5133 5134 pH value at 25°C 5135 (inclusive range) 5136 5137 5138 5.0 to 7.5 5139 5140 5141 5142 5.0 to 8.0 5143 5144 5145 Conductivity µS/cm 5146 @ 25°C, max 5147 0.1 5148 1.0 5149 5.0 5150 0.0555 5151 1.0 5152 0.25 5153 5.0 5154 5155 5156 Resistivity MΩ-cm 5157 @ 25°C, min 5158 5159 5160 5161 18 5162 1.0 5163 4.0 5164 0.2 5165 18.2 ± 1(2) 5166 10 5167 Temperature 5168 Compensated 5169 Conductivity 5170 Measurement? 5171 YES 5172 YES 5173 YES 5174 YES 5175 YES 5176 YES 5177 YES 5178 YES 5179 YES 5180 TOC (as C), max 5181 5182 5183 5184 50 µg/l 5185 (50 ppb) 5186 50 µg/l 5187 (50 ppb) 5188 200 µg/l 5189 (200 ppb) 5190 5191 20 µg/L 5192 (20 ppb) 5193 500 ppb 5194 Oxidizable 5195 Substances 5196 (Permanganate Red. 5197 Subst.) 5198 5199 5200 5201 5202 5203 5204 5205 5206 5207 Oxidizable matter 5208 O2 content mg/l, max 5209 5210 0.08 5211 0.4 5212 5213 5214 5215 5216 5217 5218 Absorbance at 5219 254 nm and 1 cm 5220 optical path length, 5221 absorbance units, 5222 max 5223 0.001 5224 0.01 5225 5226 5227 5228 5229 5230 5231 5232 Residue after 5233 evaporation on 5234 heating at 110°C, 5235 mg/Kg, max 5236 5237 1 5238 2 5239 5240 5241 5242 5243 5244 5245 Residue after 5246 evaporation on 5247 heating at 105°C, 5248 mg/100 ml, max 5249 5250 5251 5252 5253 5254 5255 5256 5257 5258 Silica (as SiO2) 5259 mg/l, max 5260 0.01 5261 0.02 5262 5263 5264 5265 5266 5267 5268 5269 Total Silica µg/l, max 5270 5271 5272 5273 3 5274 3 5275 500 5276 5277 5278 5279 Sodium µg/l, max 5280 5281 5282 5283 1 5284 5 5285 10 5286 50 5287 5288 5289 Chloride µg/l, max 5290 5291 5292 5293 1 5294 5 5295 10 5296 50 5297 5298 5299 Heterotrophic 5300 Bacteria Count 5301 cfu/ml, max 5302 5303 5304 5305 5306 5307 5308 5309 Grade A: 0.01 5310 (10 cfu/ 5311 1000 ml) 5312 Grade A: 0.01 5313 (10 cfu/ 5314 1000 ml) 5315 Grade A: 0.01 5316 (10 cfu/ 5317 1000 ml) 5318 Grade A: 0.01 5319 (10 cfu/ 5320 1000 ml) 5321 5322 5323 5324 Grade B: 0.1 5325 (10 cfu/ 5326 100 ml) 5327 Grade B: 0.1 5328 (10 cfu/ 5329 100 ml) 5330 Grade B: 0.1 5331 (10 cfu/ 5332 100 ml) 5333 Grade B: 0.1 5334 (10 cfu/ 5335 100 ml) 5336 1 5337 (100 cfu/ 5338 100 ml) 5339 10 5340 5341 5342 5343 Grade C: 10 5344 (1000 cfu/ 5345 100 ml) 5346 Grade C: 10 5347 (1000 cfu/ 5348 100 ml) 5349 Grade C: 10 5350 (1000 cfu/ 5351 100 ml) 5352 Grade C: 10 5353 (1000 cfu/ 5354 100 ml) 5355 5356 This Document is licensed to 5357 Page 96 5358 ISPE Good Practice Guide: 5359 Appendix 1 5360 Sampling for Pharmaceutical Water, Steam, and Process Gases 5361 Organization/ 5362 Reference 5363 ISO 3696 (1995) 5364 Water for Analytical Laboratory Use 5365 ASTM D1193-06 (2011)(1) 5366 Standard Specification for Reagent Water 5367 ASTM D5196 5368 (2006) 5369 CLSI 4th Ed 5370 (2006) 5371 Water Grade or Type 5372 Grade 1 5373 Grade 2 5374 Grade 3 5375 Type I 5376 Type II 5377 Type III 5378 Type IV 5379 Standard 5380 Guide for 5381 Bio-App 5382 lications 5383 Grade Water 5384 CLRW 5385 (Specified 5386 quantitative 5387 attributes 5388 only) 5389 Specified Source 5390 and Purification 5391 Approaches 5392 Grade 2 Source; 5393 RO+0.2 µm Filt, 5394 or DI+0.2 5395 µm Filt, 5396 or Re-Dist 5397 (in glass) 5398 Multiple-Dist 5399 or DI 5400 or RO+Dist 5401 Single-Dist 5402 or DI 5403 or RO 5404 < 20 µS/ 5405 cm Source 5406 (Dist, equiv); 5407 MB-DI 5408 + 0.2 µm Filt 5409 Distillation or 5410 equiv 5411 Distillation, 5412 DI, EDI, 5413 and/or RO 5414 + 0.45 µm Filt 5415 Distillation, 5416 DI, EDI, 5417 and/or RO 5418 Drinking 5419 Water Source; 5420 Suitable 5421 process(es) 5422 5423 Bacterial 5424 Endotoxins 5425 EU/ml or IU/ml 5426 5427 5428 5429 5430 5431 5432 5433 Grade A: 0.03 5434 Grade A: 0.03 5435 Grade A: 0.03 5436 Grade A: 0.03 5437 0.01 5438 5439 5440 5441 5442 Grade B: 0.25 5443 Grade B: 0.25 5444 Grade B: 0.25 5445 Grade B: 0.25 5446 5447 5448 5449 Grade C: 5450 Grade C: 5451 Grade C: 5452 Grade C: 5453 Nitrates ppm, max 5454 5455 5456 5457 5458 5459 5460 5461 5462 5463 Aluminium ppb, max 5464 5465 5466 5467 5468 5469 5470 5471 5472 5473 Heavy Metals ppm, 5474 max 5475 5476 5477 5478 5479 5480 5481 5482 5483 5484 Other Inorganic 5485 Attributes 5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 Particulate and 5496 Colloids 5497 Implied 5498 limitation by 5499 0.2 µm filter 5500 5501 5502 Implied 5503 limitation by 5504 0.2 µm filter 5505 5506 Implied 5507 limitation by 5508 0.45 µm filter 5509 5510 5511 Implied 5512 limitation by 5513 0.22 µm filter 5514 Nucleases, 5515 Proteases 5516 5517 5518 5519 5520 5521 5522 5523 Limited as 5524 needed 5525 for certain 5526 applications 5527 5528 Footnotes: 5529 Not Specified, Not Required, Not Applicable, or No Limit 5530 (1) Water may be produced with alternate technologies if specifications are met and water is appropriate for the application. 5531 (2) If in-line resistivity testing is not possible, then the total concentration of inorganic ions must not exceed 1µg/L for cations such as Aluminum, Ammonium, Arsenic, 5532 Cadmium, Chromium, Cobalt, Copper, Iron, Lead, Magnesium, Nickel, Potassium, Sodium, Titanium, Zinc, and anions such as Chloride, Nitrate, Phosphate, Sulfate, and 5533 Fluoride. 5534 5535 This Document is licensed to 5536 ISPE Good Practice Guide: 5537 Page 97 5538 Sampling for Pharmaceutical Water, Steam, and Process Gases 5539 Appendix 2 5540 Appendix 2 5541 6 Appendix 2 – Examples of Water System 5542 5543 Sampling Point Locations 5544 System Boundary designations. 5545 Note: these figures are found in Appendix 3 of the ISPE Good Practice Guide: Approaches to Commissioning and 5546 Qualification of Pharmaceutical Water and Steam Systems (Second Edition) [13]. 5547 Figure 6.1: POU Piping Direct Connect to Process Equipment System Classification Example 5548 (modified for this Guide) 5549 5550 This Document is licensed to 5551 Page 98 5552 ISPE Good Practice Guide: 5553 Appendix 2 5554 Sampling for Pharmaceutical Water, Steam, and Process Gases 5555 Figure 6.2: POU Hard-Piped to Process Equipment System Classification Example (modified for this Guide) 5556 5557 This Document is licensed to 5558 ISPE Good Practice Guide: 5559 Page 99 5560 Sampling for Pharmaceutical Water, Steam, and Process Gases 5561 Appendix 2 5562 Figure 6.3: POU Piping with Integral Heat Exchanger System Classification Example (modified for this Guide) 5563 5564 This Document is licensed to 5565 Page 100 5566 ISPE Good Practice Guide: 5567 Appendix 2 5568 Sampling for Pharmaceutical Water, Steam, and Process Gases 5569 Figure 6.4: POU Piping with Hose System Classification Example 5570 Figure 6.5: Water Distribution System Sub-Loop Example 5571 5572 This Document is licensed to 5573 ISPE Good Practice Guide: 5574 Page 101 5575 Sampling for Pharmaceutical Water, Steam, and Process Gases 5576 Appendix 3 5577 Appendix 3 5578 7 Appendix 3 – Factors Influencing Pure 5579 5580 5581 Steam Generator Performance 5582 The purification system generating the water feeding a PSG typically needs the same monitoring and maintenance 5583 attention as pretreatment systems feeding distillation units that generate WFI grade water. This monitoring can help to 5584 optimize control of the pretreatment unit operations and steam generation processes. Monitoring could also serve to 5585 alert the user to take remedial action to avoid Action Limits and specification failures in the steam. 5586 The process of generating steam from water is reasonably efficient at separating pure water vapor from the chemical 5587 contaminants that stay behind in the evaporator. However, the separation is not absolute due to potential entrainment 5588 of mist from the incoming water (and its chemical contaminants) and carryover of contaminants into the pure steam. 5589 It is commonly accepted that the phase change and mist elimination processes of distillation will purify feed water 5590 by at least three orders of magnitude (≥ 1000-fold or ≥ 99.9% impurity reduction), allowing no more than 0.1% of the 5591 feed water impurities to remain in the distillate or steam (see the ISPE Baseline® Guide: Water and Steam Systems 5592 (Second Edition) [7]). 5593 However, there may be distinctions between the generation of WFI by distillation units and the generation of PS 5594 by steam generators. These distinctions relate to the nature of their source waters, the unit operations designed to 5595 prepare that water for the generation of steam, and the associated mist elimination and blow down designs, all of 5596 which could impact the quality and performance of the PSG and pretreatment system. 5597 7.1 5598 Source Water 5599 7.1.1 5600 Pharmaceutical Grade Feed Water from WFI Systems 5601 Organizations may choose to feed their PSGs with WFI, usually from a POU valve that is part of the WFI loop, 5602 although this is not a regulatory requirement. 5603 When the feed water to a PSG already has very low or undetectable levels of endotoxin, and is otherwise chemically 5604 pure, no additional PSG feed water pretreatment system monitoring is required. When properly operated, the 5605 chemical attributes of the PSG’s feed water are very similar to as those of the steam it generates. 5606 The WFI system pretreatment process should be monitored and controlled in order for the distillation process to 5607 produce WFI. Information on sampling for water is contained in Chapter 2 of this Guide. 5608 7.1.2 5609 Pharmaceutical Grade Feed from PW Systems 5610 The source water which is fed to a PSG should already be chemically pure and sufficiently free from chemical 5611 contaminants, e.g., hardness and silica that could damage the steam generator equipment, as well as from high 5612 levels of inorganic impurities and TOC. Organizations may choose to feed their PSGs with water from a PW 5613 distribution loop that may also feed other manufacturing uses, although this is not a regulatory requirement. 5614 Because PW has no endotoxin requirement, without specific microbial and endotoxin control precautions, excessive 5615 levels of endotoxin may be present in the feed water. 5616 5617 This Document is licensed to 5618 Page 102 5619 ISPE Good Practice Guide: 5620 Appendix 3 5621 Sampling for Pharmaceutical Water, Steam, and Process Gases 5622 Since PSGs have a finite impurity removal capability (including endotoxin), the endotoxin levels in the PW feeding 5623 the PSG may be monitored to ensure that the removal capability of the steam generator is not exceeded. With 5624 an assumed 3-log endotoxin removal capability and proper equipment operation, the endotoxin level in the feed 5625 water would have to be in excess of 100 endotoxin units/ml to create conditions that would challenge the removal 5626 capabilities of a PSG. Endotoxin levels this high are typically not seen in a PW system where microbial control 5627 measures are in effect. For this reason, monitoring the endotoxin level in the PW distribution system may not be 5628 necessary. Depending on the PSG design and the presence of mist elimination, the actual log reduction capabilities 5629 of a PSG may be higher or lower and may impact the decision on endotoxin monitoring in the PW system. 5630 7.1.3 5631 Potable Water 5632 Like PW and WFI, the USP monograph for PS indicates that it should be prepared from a starting material that 5633 meets the drinking/potable water requirements established by the US, EU, Japan, or the WHO. The user may feed 5634 a PSG from a PW or WFI distribution system outlet or may use a dedicated potable water pretreatment system to 5635 provide feed water to the PSG. While there is no compendial requirement that the water feeding a PSG meets PW 5636 or WFI quality requirements, the water quality should meet the quality attributes stipulated by the PSG manufacturer 5637 to ensure successful operation and maintenance. Water quality attributes such as conductivity, TOC, endotoxin, 5638 ammonia, and other attributes should be monitored to ensure that they do not exceed the purification capabilities of 5639 the PSG. The source of the potable water supply could have a dramatic impact on its starting purity, and therefore, on 5640 the required pretreatment. 5641 Based on the initial cost and the cost of maintenance/operation, it may be more economical to include additional 5642 capacity in the feed water system for PW and WFI systems to support the capacity of the PSG system rather than to 5643 use a dedicated and separate pretreatment system for the PSG. 5644 7.2 5645 Steam Generator Mist Elimination Capability 5646 The USP monograph for pure steam specifically states that the pure steam generation process uses a process that 5647 “prevents source water entrainment.” [9] PSGs without mist elimination will have poorer log reductions of feed water 5648 contaminants and this should be considered when devising a sampling plan. 5649 Well designed and well operated systems using WFI as feed water have no need for mist elimination, as the feed 5650 water is already of suitable quality for steam. These uses may have lower risk for PS quality attributes of conductivity, 5651 TOC or endotoxin content because of the absence of mist elimination. There may be no need to monitor the feed 5652 water for these quality attributes, even with a steam generator without any mist elimination capability, as long as the 5653 WFI feed water complies with USP monograph requirements and is not compromised by other aspects of the steam 5654 generation equipment. 5655 When using PW as feed water to a PSG, endotoxin levels may be higher and mist elimination becomes a more 5656 important consideration. Ongoing sampling and monitoring ensures that the endotoxin levels continue to be low 5657 enough to be effectively removed by the PSG. 5658 A standalone PSG feed water pretreatment system that is either not well designed or not properly maintained carries 5659 a high risk of contaminant carry over into the pure steam. These systems could also experience endotoxin and TOC 5660 generated in situ by the pretreatment system. Regardless of the presence of mist elimination capabilities, these 5661 systems are likely to require the most rigorous monitoring of quality attributes. 5662 5663 This Document is licensed to 5664 ISPE Good Practice Guide: 5665 Page 103 5666 Sampling for Pharmaceutical Water, Steam, and Process Gases 5667 Appendix 3 5668 7.3 5669 Non-condensable Gas Removal Capability 5670 Without a dissolved gas removal process prior to or as part of the PSG, these gases will pass through a PSG and 5671 directly into the distributed steam. In low usage systems, non-condensable gases can accumulate in the distribution 5672 piping system if there is no means for their removal. The presence of non-condensable gases in PS will reduce the 5673 efficacy of steam sterilization process, with BS EN 285:2015 [19] and HTM 2010 [28] standards stipulating non- 5674 condensable gas requirements for steam used in sterilizers. 5675 There is no simple direct test for these dissolved gases in the feed water to a PSG. However, dissolved gases are 5676 more soluble in cold water and less soluble in hot water. Whenever the feed water to the PSG is hot WFI, the hot 5677 water will contain very little dissolved gas. Steam systems whose feed water is hot WFI have few problems with non- 5678 condensable gases in the distributed steam. 5679 However, if the feed water to the PSG is at ambient temperature, higher levels of dissolved gases will be present. 5680 Ambient temperature feed water has the potential for containing dissolved gases such as CO2, O2, and N2. If this 5681 ambient feed water is highly deionized, it will be extremely low in conductivity and there will be no CO2 present. Both 5682 nitrogen and oxygen may be present, with more oxygen presence if the feed water pretreatment system utilizes an 5683 ozone generating process for microbial or TOC control. 5684 The removal of dissolved gases from PSG feed water is typically accomplished using as vacuum degassification or 5685 high efficiency deionizers (CO2 removal only). 5686 7.4 5687 Blow Down Adjustment 5688 Monitoring of the blow down discharge coupled with periodic testing of conductivity, TOC, and endotoxin in the feed 5689 water may be appropriate to establish a blow down rate aligned with the manufacturer’s recommendations and 5690 guidelines. 5691 7.5 5692 Potable Water Chloramine Use 5693 Chloramine use in potable water disinfection is gaining popularity over chlorine because of chlorine’s in situ 5694 generation of carcinogenic disinfection byproducts such as THMs and HAAs. Chloramines are produced by adding 5695 ammonia to water containing free chlorine. When chloramines are removed by a pretreatment system, some residual 5696 ammonia may remain. Ammonia will pass through most pretreatment processes and will also pass through a PSG 5697 and contaminate the PS, causing it to fail its conductivity specifications. 5698 Processes designed initially to remove chlorine are usually not as effective at removing chloramine. Ammonia exists 5699 in different forms based on the pH of the water and ammonia removal is assured with the proper equipment. When 5700 chloramine is in use, both ammonia and chlorine should be monitored after a dechlorination step. Changes from 5701 chlorine to chloramine disinfection by a water municipality without notification may cause equipment damage and the 5702 production of non-compliant WFI or PS, potentially impacting finished product quality and impacting patient safety. 5703 7.6 5704 Anti-scaling Steam Additives 5705 The use of steam additives to minimize corrosion in a PSG and distribution piping system is strictly forbidden. Many 5706 plant steam additives are intended to carry over into the steam. If used in the feed water for a PSG, these additives 5707 would cause the PS to fail its quality attributes. This is typically not a problem since plant steam and clean steam 5708 condensate collection and return systems are kept separate and plant steam condensate is never returned to the 5709 feed of a PSG. 5710 5711 This Document is licensed to 5712 Page 104 5713 ISPE Good Practice Guide: 5714 Appendix 3 5715 Sampling for Pharmaceutical Water, Steam, and Process Gases 5716 7.7 5717 Monitoring Locations and Frequency 5718 There are no pretreatment system monitoring requirements stipulated by regulatory authorities. Pretreatment 5719 operation and monitoring should follow the guidelines established in this and other ISPE guidance documents for 5720 compendial water and steam systems (see the ISPE Baseline® Guide: Water and Steam Systems (Second Edition) 5721 [7] as well as the ISPE Good Practice Guide: Approaches to Commissioning and Qualification of Pharmaceutical 5722 Water and Steam Systems (Second Edition) [13]). 5723 Quality attributes may need to be periodically monitored in the: 5724 • 5725 Incoming source water 5726 • 5727 Between pretreatment unit operations 5728 • 5729 Feed water to the steam generator 5730 The frequency should be appropriate for each system. 5731 5732 This Document is licensed to 5733 ISPE Good Practice Guide: 5734 Page 105 5735 Sampling for Pharmaceutical Water, Steam, and Process Gases 5736 Appendix 4 5737 Appendix 4 5738 8 Appendix 4 – References 5739 1. 5740 FDA “Guide to Inspections of High Purity Water Systems,” July 1993, The Division of Field Investigations, Office 5741 of Regional Operations, Office of Regulatory Affairs, US Food and Drug Administration (FDA), www.fda.gov. 5742 2. 5743 European Pharmacopoeia (EP), EDQM Council of Europe, https://www.edqm.eu/en/ph-eur-9th-edition. 5744 3. 5745 International Organization for Standardization (ISO), www.ISO.org. 5746 4. 5747 American Society for Testing and Materials (ASTM) International, West Conshohocken, PA, www.astm.org. 5748 5. 5749 Association for the Advancement of Medical Instrumentation (AAMI), www.aami.org. 5750 6. 5751 Clinical and Laboratory Standards Institute (CLSI), http://clsi.org. 5752 7. 5753 ISPE Baseline® Pharmaceutical Engineering Guide, Volume 4 – Water and Steam Systems, International Society 5754 for Pharmaceutical Engineering (ISPE), Second Edition, December 2011, www.ispe.org. 5755 8. 5756 USP <643> Total Organic Carbon, United States Pharmacopeial Convention, www.usp.org. 5757 9. 5758 USP <1231> Water for Pharmaceutical Purposes, United States Pharmacopeial Convention – National 5759 Formulary (USP-NF), www.usp.org/usp-nf. 5760 10. EudraLex Volume 4 – Guidelines for Good Manufacturing Practices for Medicinal Products for Human and 5761 Veterinary Use, Annex 1: Manufacture of Sterile Medicinal Products, http://ec.europa.eu/health/documents/ 5762 eudralex/vol-4/index_en.htm. 5763 11. JP16 General Information, G8 Water, Quality Control of Water for Pharmaceutical Use, Section 4.2 Sampling, 5764 Japanese Pharmacopoeia (JP) – Sixteenth Edition, Pharmaceuticals and Medical Devices Agency (PMDA), 5765 http://www.pmda.go.jp/english/rs-sb-std/standards-development/jp/0005.html. 5766 12. International Council for Harmonisation (ICH), ICH Harmonised Tripartite Guideline, Good Manufacturing 5767 Practice Guide for Active Pharmaceutical Ingredients – Q7/Q7A, Step 4, 10 November 2000, www.ich.org. 5768 13. ISPE Good Practice Guide: Approaches to Commissioning and Qualification of Pharmaceutical Water and Steam 5769 Systems, International Society for Pharmaceutical Engineering (ISPE), Second Edition, July 2014, www.ispe.org. 5770 14. WHO Technical Report Series, No. 929, Annex 3: WHO Good Manufacturing Practices: Water for Pharmaceutical 5771 Use, World Health Organization (WHO), 2005, http://www.who.int/medicines/publications/pharmprep/en/. 5772 15. Chinese Pharmacopeia (ChP), www.usp.org. 5773 16. ISPE Good Practice Guide: Ozone Sanitization of Pharmaceutical Water Systems, International Society for 5774 Pharmaceutical Engineering (ISPE), First Edition, July 2012, www.ispe.org. 5775 17. United Sates Pharmacopeia (USP), 23rd Edition, 5th Supplement, United States Pharmacopeial Convention, 5776 www.usp.org. 5777 18. USP <645> Water Conductivity, United States Pharmacopeial Convention, www.usp.org. 5778 19. BS EN 285:2015 – Sterilization; Steam Sterilizers; Large Sterilizers, http://shop.bsigroup.com. 5779 5780 This Document is licensed to 5781 Page 106 5782 ISPE Good Practice Guide: 5783 Appendix 4 5784 Sampling for Pharmaceutical Water, Steam, and Process Gases 5785 20. ANSI/AAMI ST79:2010/A4:2013, Comprehensive guide to steam sterilization and sterility assurance in health 5786 care facilities, Association for the Advancement of Medical Instrumentation (AAMI), www.aami.org. 5787 21. ASME BPE-2014: Bioprocessing Equipment, American Society of Mechanical Engineers (ASME), www.asme. 5788 org. 5789 22. Health Technical Memorandum (HTM) 2031: Clean Steam for Sterilization, http://www.wales.nhs.uk/sites3/ 5790 Documents/254/HTM%202031%201997.pdf. 5791 23. ISO 13408-5:2006 Aseptic Processing of Health Care Products -- Part 5: Sterilization in Place, International 5792 Organization for Standardization (ISO), www.iso.org. 5793 24. Technical Report No. 1, Validation of Moist Heat Sterilization Processes Cycle Design, Development, 5794 Qualification and Ongoing Control, Revised 2007, Parenteral Drug Association (PDA), www.pda.org. 5795 25. ISO 17665-1:2006 Sterilization of Health Care Products -- Moist Heat -- Part 1: Requirements for the 5796 development, validation and routine control of a sterilization process for medical devices, International 5797 Organization for Standardization (ISO), www.iso.org. 5798 26. ISO 11140-4:2007 Sterilization of Health Care Products -- Chemical Indicators -- Part 4: Class 2 indicators as 5799 an alternative to the Bowie and Dick type test for detection of steam penetration, International Organization for 5800 Standardization (ISO), www.iso.org. 5801 27. United States Pharmacopeial Convention (USP), www.usp.org. 5802 28. Health Technical Memorandum (HTM) 2010: Sterilization, Part 3: Validation and Verification, http://www.wales. 5803 nhs.uk/sites3/Documents/254/HTM%202010%20Pt3%20Val%201994.pdf. 5804 29. ISPE Good Practice Guide: Process Gases, International Society for Pharmaceutical Engineering (ISPE), First 5805 Edition, July 2011, www.ispe.org. 5806 30. FDA Guidance for Industry: Sterile Drug Products Produced by Aseptic Processing – Current Good 5807 Manufacturing Practice, September 2004, US Food and Drug Administration (FDA), www.fda.gov. 5808 31. ISO 8573 – Compressed Air series, International Organization for Standardization (ISO), www.iso.org. 5809 32. ISO 8573-4:2001 Compressed Air -- Part 4: Test methods for solid particle content, International Organization for 5810 Standardization (ISO), www.iso.org. 5811 33. ISO 8573-3:1999 Compressed Air -- Part 3: Test methods for measurement of humidity, International 5812 Organization for Standardization (ISO), www.iso.org. 5813 34. ISO 8573-1:2010 Compressed Air -- Part 1: Contaminants and purity classes, International Organization for 5814 Standardization (ISO), www.iso.org. 5815 35. National Primary Drinking Water Regulations (NPDWR) (as cited in 40 CFR Part 141), United States 5816 Environmental Protection Agency (EPA), www.epa.gov. 5817 5818 This Document is licensed to 5819 ISPE Good Practice Guide: 5820 Page 107 5821 Sampling for Pharmaceutical Water, Steam, and Process Gases 5822 Appendix 5 5823 Appendix 5 5824 9 Appendix 5 – Glossary 5825 9.1 5826 Acronyms and Abbreviations 5827 AAMI 5828 Association for the Advancement of Medical Instrumentation 5829 ANSI 5830 American National Standards Institute 5831 API 5832 Active Pharmaceutical Ingredient 5833 ASME 5834 American Society of Mechanical Engineers (US) 5835 ASTM 5836 American Society for Testing and Materials (US) 5837 BPE 5838 Bioprocessing Equipment (ASME National Standard) 5839 CAPA 5840 Corrective Action and Preventive Action 5841 CFU 5842 Colony Forming Unit 5843 CEDI 5844 Continuous Electro De-Ionization 5845 CLSI 5846 Clinical and Laboratory Standards Institute (US) 5847 CoA 5848 Certificate of Analysis 5849 CPP 5850 Critical Process Parameter 5851 CQA 5852 Critical Quality Attribute 5853 CUP 5854 Critical Use Point 5855 EN 5856 European Norm 5857 EP 5858 European Pharmacopeia 5859 EPA 5860 Environmental Protection Agency (US) 5861 FDA 5862 Food and Drug Administration (US) 5863 GEP 5864 Good Engineering Practice 5865 HAA 5866 Haloacetic Acid 5867 HACCP 5868 Hazard Analysis and Critical Control Point 5869 HDPE 5870 High Density Polyethylene 5871 HPW 5872 Highly Purified Water 5873 HTM 5874 Health Technical Memorandum 5875 HVAC 5876 Heating, Ventilation, and Air Conditioning 5877 5878 This Document is licensed to 5879 Page 108 5880 ISPE Good Practice Guide: 5881 Appendix 5 5882 Sampling for Pharmaceutical Water, Steam, and Process Gases 5883 IPA 5884 Isopropyl Alcohol 5885 JP 5886 Japanese Pharmacopeia 5887 IQ 5888 Installation Qualification 5889 ISO 5890 International Organization for Standardization 5891 LAL 5892 Limulus Amoebocyte Lysate 5893 NCG 5894 Non-condensable Gas 5895 NPDWR 5896 National Primary Drinking Water Regulations (US FDA) 5897 OOS 5898 Out Of Specification 5899 OQ 5900 Operational Qualification 5901 PAT 5902 Process Analytical Technology 5903 POU 5904 Point Of Use 5905 PQ 5906 Performance Qualification 5907 PSE 5908 Periodic System Evaluation 5909 PSG 5910 Pure Steam Generator 5911 PTFE 5912 Polytetrafluoroethylene (Teflon®) 5913 PVC 5914 Polyvinyl Chloride 5915 PW 5916 Purified Water 5917 QA 5918 Quality Assurance 5919 QC 5920 Quality Control 5921 RA 5922 Risk Analysis 5923 RO 5924 Reverse Osmosis 5925 RMM 5926 Rapid Microbial Method 5927 RTR 5928 Real Time Release 5929 SDI 5930 Silt Density Index 5931 SIP 5932 Sterilize-In-Place 5933 SME 5934 Subject Matter Expert 5935 SOP 5936 Standard Operating Procedure 5937 5938 This Document is licensed to 5939 ISPE Good Practice Guide: 5940 Page 109 5941 Sampling for Pharmaceutical Water, Steam, and Process Gases 5942 Appendix 5 5943 SPC 5944 Statistical Process Control 5945 THMs 5946 Trihalomethanes 5947 TOC 5948 Total Organic Carbon 5949 USP 5950 United States Pharmacopeia 5951 UV 5952 Ultraviolet 5953 WFI 5954 Water for Injection 5955 WHO 5956 World Health Organization 5957 9.2 5958 Definitions 5959 Action Level 5960 Levels or ranges (actual attribute values) that, when exceeded, indicate that a process has drifted from its normal 5961 operating range. Exceeding an Action Level indicates that corrective action should be taken to bring the process back 5962 into its normal operating range. 5963 Action Limit 5964 Criteria established based on possible impact to product quality, outside the operating range (acceptance criteria). A 5965 documented response is usually required. (also see: Action Point) 5966 Action Point 5967 Used in determining when a parameter has drifted outside the operating range. 5968 Adsorption 5969 Nonspecific adherence of molecules in solution or suspension to cells, particles, or other molecules. (also see: 5970 Desorption) 5971 Alert Level 5972 Levels or ranges (actual attribute values) that, when exceeded, indicate that a process may have drifted from its 5973 normal operating condition. Alert Levels constitute a warning and do not necessarily require a corrective action. 5974 Alert Limit 5975 Criteria established with the intent of notification and possible corrective action prior to exceeding action limits; alert 5976 when a parameter is drifting toward extremes of the operating range. 5977 Alert Point 5978 Used in determining when a parameter is drifting toward extremes of the operating range. 5979 Aseptic 5980 Not sterile, but contaminants controlled within established acceptable limits. 5981 5982 This Document is licensed to 5983 Page 110 5984 ISPE Good Practice Guide: 5985 Appendix 5 5986 Sampling for Pharmaceutical Water, Steam, and Process Gases 5987 At-Line 5988 In water monitoring systems. (also see: Off-Line) 5989 Atmospheric Dew Point 5990 This term refers to what the dewpoint would be if fully depressurized to atmospheric conditions (1013 mbar at either 5991 15 or 20°C (59 or 68°F). (also see: Pressure Dew Point) 5992 Bacteria 5993 Single-celled microorganisms measured in high purity water by several means: culturing, high power microscope, or 5994 Scanning Electron Microscope (SEM). The value is reported as Colony Forming Units (CFU), or colonies per milliliter 5995 or per liter. The bacteria in the water act as particle contamination on the surface of the product, or as a source of 5996 detrimental by-products. (also see: Pyrogen) 5997 Bioburden 5998 The concentration of microbial matter per unit volume. Microbial matter includes viruses, bacteria, yeast, mold, and 5999 parts thereof. 6000 Biofilm 6001 A collection of microorganisms, extracellular polymeric products, and organic matter located at the interface in solid- 6002 liquid, gas-liquid, or liquid-liquid biphasic systems. 6003 Biologic 6004 A therapeutic agent derived from living organisms. 6005 Blowdown 6006 The withdrawal of water from an evaporating water system to maintain a solids balance within specified limits of 6007 concentration of those solids. 6008 Chemical Free Steam 6009 Non-direct impact steam produced from pretreated water with no volatile boiler additives. May be used for 6010 humidification but is not used for product contact. 6011 Chloramine 6012 A chlorine compound formed by reaction with organic amines or ammonia. 6013 Commissioning 6014 A well planned, documented, and managed engineering approach to the start-up and turnover of facilities, systems, 6015 and equipment to the end-user that results in a safe and functional environment that meets established design 6016 requirements and stakeholder expectations. 6017 Compendial 6018 Official; purported to comply with USP, EP, or JP. 6019 6020 This Document is licensed to 6021 ISPE Good Practice Guide: 6022 Page 111 6023 Sampling for Pharmaceutical Water, Steam, and Process Gases 6024 Appendix 5 6025 Conductivity 6026 A measure of flow of electrical current through water. This conductance is high with high Total Dissolved Solids 6027 (TDS) water and very low with ultrapure deionized water. Conductivity is the reciprocal of resistivity (C=1/R) and is 6028 measured in micromho/cm (µmho/cm) or microsiemens/cm (µS/cm). (also see: Resistivity) 6029 Contaminant 6030 Any foreign component present in another substance. For example, anything in water that is not H2O is a 6031 contaminant. 6032 Critical Quality Attribute 6033 Attribute of the water or steam product which usually relates to the identity, purity, or quality of the water or steam 6034 product. Some attributes for water and steam are, contained levels of TOC, conductivity, microbial activity, etc. 6035 Critical Utility 6036 Utility that has the identified potential to impact product quality or performance in a significant way. 6037 Critical Utility Sampling 6038 In the case of WFI, it is the expectation that water samples will be taken from each point of use (POU) and sample 6039 point in the distribution system, at least once per week, although the widespread implementation of risk analysis tools 6040 may cause this expectation to be challenged in the future. 6041 Dead Leg 6042 An area of entrapment in a vessel or piping run that could lead to contamination of the product. 6043 Desorption 6044 The opposite of adsorption; the release of adsorbed molecules, particles, or cells into the surrounding medium. (also 6045 see: Adsorption) 6046 Dew Point 6047 The dew point is the temperature to which a given parcel of humid air must be cooled, at constant barometric 6048 pressure, for water vapor to condense into water. The condensed water is called dew. The dew point is a saturation 6049 temperature. 6050 Distillation 6051 The process of separating water from impurities by heating until it changes into vapor and then cooling the vapor to 6052 condense it into purified water. 6053 Dry Air 6054 Air from which all water vapor and contaminants have been removed. Its composition by volume is: 6055 1. 6056 Nitrogen 6057 6058 78.08% 6059 2. 6060 Oxygen 6061 6062 20.95% 6063 3. 6064 Argon 6065 6066 0.93% 6067 4. 6068 Carbon Dioxide 6069 0.03 6070 5. 6071 Other gases 6072 0.00003 6073 (also see: Moist Air) 6074 6075 This Document is licensed to 6076 Page 112 6077 ISPE Good Practice Guide: 6078 Appendix 5 6079 Sampling for Pharmaceutical Water, Steam, and Process Gases 6080 Dryness 6081 Refers to the level of steam saturation, and it is the ratio of vapor mass to the mass of the steam mixture. Dryness 6082 has a value of 1.0 which is representative of dry saturated steam; otherwise is dimensionless. 6083 Endotoxins 6084 Pyrogens from certain Gram negative bacteria. Generally, highly toxic Lipopolysaccharide-protein complexes (fat, 6085 linked sugars, and protein) from cell walls. A marker for these bacteria with a reputation for persistent contamination 6086 because they tend to adhere to surfaces. (also see: Pyrogen) 6087 Extended Sampling 6088 In water monitoring systems – compared to Phase 3 of the ISPE Good Practice Guide: Approaches to Commissioning 6089 and Qualification of Pharmaceutical Water and Steam Systems (Second Edition) [13], in general terms this sampling 6090 phase allows for the system to be challenged by evaluating its effectiveness in delivering water of acceptable 6091 quality despite seasonal variations of the potable water feed to the system, the slow development of the mature 6092 system’s natural flora, and any other variations that may occur. Its duration may continue for not less than one year 6093 (subtracting Initial and Intermediary phases). (also see: Initial, Intermediary Sampling) 6094 Extractables 6095 Undesirable foreign substances that are leached or dissolved by water or process streams from the materials of 6096 construction used in filters, storage vessels, distribution piping, and other product contact surfaces. (also see: 6097 Leachables) 6098 Feedwater 6099 The water entering a treatment process. 6100 Flushing 6101 Cleansing for the removal of particulates or water soluble contaminants, brought about by the flowing of large 6102 quantities of water over the product and/or solution contact surfaces of system components. (also see: Rinsing) 6103 Gram-Negative Organism 6104 Any prokaryotic organism that does not retain the first stain (crystal violet) used in Gram’s staining technique. It does 6105 retain the second stain (safranin O) and therefore has a pink color when viewed under a light microscope. Retention 6106 of the stain is due to the structure of the cell wall. 6107 Hardness (water) 6108 The concentration of calcium and magnesium salts in water. Hardness is a term originally referred to the soap- 6109 consuming power of water; as such it is sometimes also taken to include iron and manganese. “Permanent hardness” 6110 is the excess of hardness over alkalinity. “Temporary hardness” is equal or less than the alkalinity. These also are 6111 referred to as “non-carbonated” or “carbonate” hardness, respectively. 6112 6113 This Document is licensed to 6114 ISPE Good Practice Guide: 6115 Page 113 6116 Sampling for Pharmaceutical Water, Steam, and Process Gases 6117 Appendix 5 6118 Highly Purified Water 6119 Water intended for use in the preparation of products where water of high biological quality is needed, except where 6120 Water for Injection is required. Highly Purified water is obtained from water that complies with the regulations on 6121 water intended for human consumption laid down by the competent authority. Current production methods include, for 6122 example, double-pass reverse osmosis coupled with other suitable techniques such as ultrafiltration and deionization. 6123 Highly Purified water meets the same quality standards as WFI but the production methods are considered less 6124 reliable than distillation and thus it is considered unacceptable for use as WFI. 6125 Hydrocarbons 6126 Organic compounds containing only carbon and hydrogen. Sometimes broadened to include compounds or mixtures 6127 of compounds with small amounts of oxygen also. 6128 Hydrophilic 6129 Having an affinity for water. Its opposite, non-water-wettable, hydrophobic. 6130 Hydrophobic 6131 The extent of insolubility; not readily absorbing water; resisting or repelling water, wetting, or hydration; or being 6132 adversely affected by water. Hydrophobic bonding is an attraction between the hydrophobic or non-polar portions of 6133 molecules, causing them to aggregate and exclude water from between them. 6134 Hygienic 6135 Of, or pertaining to, equipment and piping systems that by design, materials of construction, and operation provide for 6136 the maintenance of cleanliness so that products produced by these systems will not adversely affect human or animal 6137 health. 6138 Hygienic Design 6139 A system of design that meets standards, specification, codes, regulatory and industrial guidelines, and acceptable 6140 engineering design methods to reach a degree of sanitation required by food, pharmaceutical, and cosmetics 6141 processing. 6142 Hygroscopic 6143 The property of absorbing moisture from the air. 6144 Impurity 6145 A foreign agent or material either introduced as part of processing (such as buffers or salts added during 6146 chromatography) or intrinsic to the nature of bioprocessing (such as product variants and cellular debris). 6147 In-Line 6148 An integral part of the flow path. In a fluid stream, something is said to be in-line if the entire fluid stream flows directly 6149 through or past it. (also see: On-Line, Off-Line) 6150 In-Process Control 6151 Checks performed during production in order to monitor and, if appropriate, to adjust the process and/or to ensure 6152 that the intermediate or API conforms to its specifications. (also see: Process Control) 6153 6154 This Document is licensed to 6155 Page 114 6156 ISPE Good Practice Guide: 6157 Appendix 5 6158 Sampling for Pharmaceutical Water, Steam, and Process Gases 6159 Initial Sampling 6160 In water monitoring systems – compared to Phase 1 of the ISPE Good Practice Guide: Approaches to Commissioning 6161 and Qualification of Pharmaceutical Water and Steam Systems (Second Edition) [13] and based on risk assessment, 6162 this sampling phase can be satisfied by monitoring and testing all sample and use points in the distribution system 6163 and selected other points, daily for between 10 to 20 consecutive days depending upon the design of the system. 6164 (also see: Intermediary, Extended Sampling) 6165 Intermediary Sampling 6166 In water monitoring systems – compared to Phase 2 of the ISPE Good Practice Guide: Approaches to Commissioning 6167 and Qualification of Pharmaceutical Water and Steam Systems (Second Edition) [13], this sampling phase further 6168 demonstrates consistent production and delivery of water of the required quality within the established ranges, when 6169 using SOP’s. It also may provide additional worst case data since use points are not as frequently active. Its duration 6170 may be between 10 and 20 days. (also see: Initial, Extended Sampling) 6171 Leachables 6172 Compounds that are present in the drug formulation as a result of direct contact with the component under normal 6173 conditions. Leachables are typically a subset of extractables but may also include reaction products. (also see: 6174 Extractables) 6175 Leaching 6176 The release of plastic components or additives to the product. (contrast with: Sorption) 6177 Microfiltration 6178 A method of sterile filtration, clarification, or cell harvesting that removes particles in the 0.1 to 10.0 μm range. 6179 Microorganism 6180 Organisms (microbes) observable only through a microscope. Larger, visible types are called organisms. 6181 Moist Air 6182 A binary mixture of dry air and water vapor. Each component behaves as if the other is not present and each 6183 occupies the complete volume of the mixture. (also see: Dry Air) 6184 Non-condensable Gas 6185 Air and other gas which will not condense under the conditions of steam sterilization. 6186 Non-viable 6187 Opposite of viable, not alive. 6188 Off-Line 6189 In water monitoring systems – referring to measurement devices that are not directly coupled to the water stream. 6190 (also see: In-Line, On-Line) 6191 6192 This Document is licensed to 6193 ISPE Good Practice Guide: 6194 Page 115 6195 Sampling for Pharmaceutical Water, Steam, and Process Gases 6196 Appendix 5 6197 On-Line 6198 In water monitoring systems – Referring to measurement devices that are directly coupled to the water stream. (also 6199 see: In-Line, Off-Line) 6200 Process/Piping and Instrument Diagram (P&ID) 6201 A diagrammatic representation (drawing) of the piping, electrical and control systems required for a specific process 6202 and/or utility system. 6203 pH 6204 The negative log of the hydrogen ion concentration, is a measure of the concentration of hydrogen ions (H+) in a 6205 water-based solution. The more hydrogen ions that are present, the lower the pH and the more acidic the solution. 6206 Particulate 6207 Usually a solid particle large enough to be removed by filtration. Non-filterable solids are usually referred to as 6208 colloids. 6209 Particulates 6210 Discrete quantities of solid matter dispersed in water. 6211 Passivation 6212 The means of obtaining the loss of chemical reactivity exhibited by certain metals under special environmental 6213 conditions. More specifically, the state in which a stainless steel exhibits a very low corrosion rate. Passivation 6214 generates an oxide film that covers and protects the surface of the metal. 6215 Pharmaceutical 6216 A medicinal drug, or relating to or engaged in pharmacy or the manufacture and sale of pharmaceuticals. A 6217 pharmaceutical product is generally one that is made up using available chemical compounds. 6218 Planktonic 6219 Term used to describe aquatic microorganisms that float. 6220 Plant Steam 6221 Non-direct impact steam produced by an industrial type boiler. Corrosion control additives may be present. Typically 6222 used for non-product contact heating purposes. Also called utility steam. 6223 Point Of Use (POU) 6224 The location where the water delivered from the water distribution system, is actually used. 6225 6226 This Document is licensed to 6227 Page 116 6228 ISPE Good Practice Guide: 6229 Appendix 5 6230 Sampling for Pharmaceutical Water, Steam, and Process Gases 6231 Potable Water 6232 Water of suitable quality for drinking; It is not covered by a pharmacopeial monograph but must comply with either 6233 the NPDWR (U.S. Environmental Protection Agency’s National primary Drinking Water Regulations as cited in 40 6234 CFR Part 141 [35]), the drinking water regulations of the European union or Japan, or the WHO Drinking Water 6235 Guidelines. Unless otherwise specified, Drinking Water may be used in the early stages of cleaning pharmaceutical 6236 manufacturing equipment and product-contact components. Where compatible with the processes, the allowed 6237 contaminant levels in Drinking Water are generally considered safe for use in the manufacture of drug substances 6238 and other bulk pharmaceutical ingredients. Drinking Water is the minimum quality feed water that may be used for the 6239 production of bulk USP monographed pharmaceutical waters because its specifications establish a reasonable set of 6240 maximum allowable levels of chemical and microbiological contaminants with which a water purification system will 6241 be challenged. 6242 Pressure Dew Point 6243 This term refers to the dew point temperature (PDP) of a gas under full line pressure; it is usually found when 6244 measuring the dew point temperature of gases at pressure higher than atmospheric pressure. (also see: Atmospheric 6245 Dew Point) 6246 Pretreatment 6247 Initial water treatment steps performed prior to final processing to prolong the life of cartridges and filters and to 6248 protect downstream elements from premature failure. (also see: Polishing) 6249 Process Control 6250 (also see: In-Process Control) 6251 Process Steam 6252 Direct impact steam that, once condensed, meets the quality attributes of potable water. This steam may be used in 6253 manufacturing areas for direct injection heating and sterilization. 6254 Pure Steam (USP) 6255 Water that has been heated above 100°C (212°F) and vaporized in a manner that prevents source water entrainment. 6256 It is prepared from water complying with the U.S. EPA Primary Drinking Water Regulations, or with drinking water 6257 regulations of the European Union or Japan, or with WHO drinking water guidelines. It contains no added substance. 6258 The level of steam saturation or dryness, and the amount of non-condensable gases are to be determined by the 6259 Pure Steam application. Note: Pure Steam is intended for use where steam or its condensate comes in contact with 6260 the article of the preparation. 6261 Purified Water 6262 USP Purified Water prepared from water complying with the quality attributes of “Drinking Water” with conductivity in 6263 accordance with stage 1, 2 and 3 tests and Conductivity Tables. Total Organic Carbon is at 0.5 mg/l. Typically, less 6264 than 100 CFU/ml (10,000 CFU/100 ml) for microbiological acceptability. 6265 Pyrogen 6266 Trace organics which are used as markers of bacterial growth or contamination. Produced by various bacteria 6267 and fungi. Critical pharmaceutical and biotechnological processes have restrictions on contamination by these 6268 substances, usually at levels near the limit of detection. Primarily polysaccharide (made of linked sugars) in nature. 6269 Fever producing substances when administered parenterally to man and certain animals. (also see: Endotoxin) 6270 6271 This Document is licensed to 6272 ISPE Good Practice Guide: 6273 Page 117 6274 Sampling for Pharmaceutical Water, Steam, and Process Gases 6275 Appendix 5 6276 Qualification 6277 Action of proving and documenting that equipment or ancillary systems are properly installed, work correctly, and 6278 actually lead to the expected results. Qualification is part of validation, but the individual qualification steps alone do 6279 not constitute process validation. 6280 Quality Attribute 6281 A molecular or product characteristic that is selected for its ability to help indicate the quality of the product. 6282 Collectively, the quality attributes define identity, purity, potency and stability of the product, and safety with respect to 6283 adventitious agents. Specifications measure a selected subset of the quality attributes. 6284 Quality Control (QC) 6285 Process or group responsible for coordinating the activities associated with analytical test planning and execution. 6286 Resistivity 6287 The measure of the resistance to the flow of electrical current through high purity water. This is measured in millions 6288 of ohms-cm or Megohm-cm (Mohm-cm). Resistivity is the reciprocal of Conductivity (R = 1/C, 1 Mohm-cm = 1 µS/ 6289 cm). This provides an easy means of continuously measuring the purity of very low TDS water or ionic concentration. 6290 (also see: Absolute Purity Water). (also see: Conductivity) 6291 Reverse Osmosis (RO) 6292 A process that reverses (by the application of pressure) the flow of water in the natural process of osmosis so that 6293 it passes from the more concentrated to the more dilute solution. This is one of the processes used to reduce the 6294 ionic TDS, TOC, and suspended materials of feed water through a semipermeable membrane leaving dissolved and 6295 suspended materials behind. These are swept away in a waste stream to drain. 6296 Rinsing 6297 Action using a liquid, generally water, to remove and in some cases dissolve a soil. (also see: Flushing) 6298 Risk Assessment 6299 A systematic process of organizing information to support a risk decision to be made within a risk management 6300 process. It consists of the identification of hazards and the analysis and evaluation of risks associated with exposure 6301 to those hazards. 6302 Sampling 6303 To take a small but representative portion of a much larger stream where the sample collected, accurately represents 6304 the content of the larger stream. 6305 Sampling for informational purposes only 6306 While the biopharmaceutical industry has used this description in the past, and many in the industry continue to 6307 use it, every sample point should have a specific purpose; therefore, it is recommended that this designation be 6308 abandoned for being vague and non-specific. 6309 6310 This Document is licensed to 6311 Page 118 6312 ISPE Good Practice Guide: 6313 Appendix 5 6314 Sampling for Pharmaceutical Water, Steam, and Process Gases 6315 Sanitization 6316 That part of decontamination that reduces viable microorganisms to a defined acceptance level; normally achieved by 6317 using a chemical agent or heat to reduce microbial levels. 6318 Scale 6319 The precipitate that forms on surfaces in contact with water as the result of a physical of chemical change. 6320 Solute 6321 A substance, usually a solid or semisolid, that forms a chemically and physically homogeneous mixture with one or 6322 more other substances, usually a liquid. 6323 Sorption (1) 6324 The uptake of product components by the plastic materials. (contrast with: Leaching) 6325 Sorption (2) 6326 Bonding of a solute to a plastic packaging component as a physicochemical phenomenon related to the properties 6327 of the packaging material and the chemical properties of the active substance or other soluble substances in the 6328 preparation. 6329 Stainless Steel 6330 Steel to which a significant amount of chromium and nickel has been added to inhibit corrosion. 6331 Superheated Steam 6332 Steam whose temperature, at any given pressure, is higher than the indicated by the vaporization curve of water. 6333 Ultrafiltration 6334 Filter technology similar to reverse osmosis that is capable of filtering colloids and large molecular weight organics 6335 out of the water. The filter capability of ultrafiltration filters to 0.005 µm particle size. Ultrafiltration also will remove 6336 organic material down to about 1,000 to 10,000 molecular weight. 6337 Utility Systems 6338 Facility-wide systems not tailored to a specific process and that do not have contact with the drug substance or 6339 potential drug substance. 6340 Viable Particle 6341 Particle that consists of, or supports, one or more live microorganisms. 6342 Volumetric Humidity 6343 Thee water content of a compressed gas may be specified as mass per unit volume (mg-m-3) at 1013 mbar and at 6344 either 15 or 20°C (59 or 68°F). (also see: Moisture Content) 6345 6346 This Document is licensed to 6347 ISPE Good Practice Guide: 6348 Page 119 6349 Sampling for Pharmaceutical Water, Steam, and Process Gases 6350 Appendix 5 6351 Water For Injection (WFI) (USP) 6352 Prepared from water complying with the quality attributes of “Drinking Water.” Purified by distillation or a purification 6353 process that is equivalent or superior to distillation in the removal of chemicals and microorganisms. Conductivity in 6354 accordance with Stage 1, 2, and 3 tests and Conductivity Tables. Total Organic Carbon limit is at 0.5 mg/l. Typically 6355 less than 10 CFU/100ml for microbiological acceptability. Less than 0.25 USP EU/ml. 6356 Water Treatment 6357 Water treatment, also referred to as water conditioning, can consist of adding or removing chemicals to change 6358 the properties of water. In water softening, for example, sodium ions are substituted for metallic ions that cause 6359 “hardness” thus reducing the scale-forming tendencies of water. Water purification on the other hand, always consists 6360 of removing undesirable impurities. 6361 Worst Case Testing 6362 Testing which encompasses upper and lower limits, and circumstances which pose the greatest chance finding of 6363 errors. Synonymous: Most Appropriate Challenge Conditions. 6364 6365 This Document is licensed to 6366 600 N. Westshore Blvd., Suite 900, Tampa, Florida 33609 USA 6367 Tel: +1-813-960-2105, Fax: +1-813-264-2816 6368 www.ISPE.org 6369
Applying these regulatory guidelines in practice? Use our free validated calculators for risk management, sampling, and environmental monitoring.