Designation: F3127 − 16

Standard Guide for

Validating Cleaning Processes Used During the Manufacture
of Medical Devices1
This standard is issued under the fixed designation F3127; the number immediately following the designation indicates the year of
original adoption or, in the case of revision, the year of last revision. A number in parentheses indicates the year of last reapproval. A
superscript epsilon (´) indicates an editorial change since the last revision or reapproval.

1. Scope D543 Practices for Evaluating the Resistance of Plastics to

1.1 This guide provides considerations for validating clean- Chemical Reagents
ing processes for medical devices during initial fabrication and E2857 Guide for Validating Analytical Methods
assembly prior to initial use. Validated cleaning processes are F619 Practice for Extraction of Medical Plastics
important for achieving consistency in function and consis- F2459 Test Method for Extracting Residue from Metallic
tency in biocompatibility. The considerations include but are Medical Components and Quantifying via Gravimetric
not limited to, validation approach, equipment design, proce- Analysis
dures and documentation, analytical methods, sampling, devel- F2847 Practice for Reporting and Assessment of Residues
opment of limits, and other issues. on Single Use Implants
G121 Practice for Preparation of Contaminated Test Cou-
1.2 Inclusions: pons for the Evaluation of Cleaning Agents
1.2.1 This guide describes the validation of critical cleaning G122 Test Method for Evaluating the Effectiveness of
processes for medical devices to reduce contaminants to Cleaning Agents
acceptable levels prior to packaging. G131 Practice for Cleaning of Materials and Components by
1.3 Exclusions: Ultrasonic Techniques
1.3.1 Reusable medical devices. 2.2 ANSI/AAMI/ISO Standards:3
1.3.1.1 Validation of cleaning operations for reusable medi- ISO 10993-5 Biological Evaluation of Medical Devices—
cal devices is not within the scope of this standard guide. Part 5: Tests for Cytotoxicity, In Vitro Methods
Although cleaning of reusable medical devices is beyond the ISO 10993-11 Biological Evaluation of Medical Devices—
scope of this guide, many of the principles outlined in this Art 11: Tests for Systemic Toxicity
guide may be applicable to the validation of cleaning opera- ISO 10993-17 Biological Evaluation of Medical Devices—
tions for reusable devices. Part 17: Establishment of Allowable Limits for Leachable
1.3.2 Cleaning of medical devices in health care facilities. Substances
1.3.2.1 Validation of cleaning processes in patient/health ISO 11737-1 Sterilization of Medical Devices—
care facilities is not within the scope of this standard guide. Microbiological Methods—Part 1: Determination of a
1.4 This standard does not purport to address all of the Population of Microorganisms on Products
safety concerns, if any, associated with its use. It is the ISO 14971 Medical Devices—Application of Risk Manage-
responsibility of the user of this standard to establish appro- ment to Medical Devices
priate safety and health practices and determine the applica- AAMI ST72 Bacterial Endotoxins—Test Methodologies,
bility of regulatory limitations prior to use. Routine Monitoring, and Alternatives to Batch Testing
AAMI TIR30 A Compendium of Processes, Materials, Test
2. Referenced Documents Methods, and Acceptance Criteria for Cleaning Reusable
2.1 ASTM Standards:2 Medical Devices
2.3 United States Pharmacopoeia (USP) – General Chap-
ters:
1
This guide is under the jurisdiction of ASTM Committee F04 on Medical and
USP <85> Bacterial Endotoxins Test
Surgical Materials and Devices and is the direct responsibility of Subcommittee USP <87> Biological Reactivity Tests, In Vitro
F04.15 on Material Test Methods. USP <88> Biological Reactivity Tests, In Vivo
Current edition approved April 1, 2016. Published May 2016. DOI: 10.1520/ USP <1225> Validation of Compendial Procedures
F3127-16
2
For referenced ASTM standards, visit the ASTM website, www.astm.org, or
contact ASTM Customer Service at [email protected]. For Annual Book of ASTM
3
Standards volume information, refer to the standard’s Document Summary page on Available from American National Standards Institute (ANSI), 25 W. 43rd St.,
the ASTM website. 4th Floor, New York, NY 10036, http://www.ansi.org.

Copyright © ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. United States

1
 F3127 − 16
2.4 International Conference on Harmonization of Techni- 3.1.12 monitoring, v—verification testing at predefined in-
cal Requirements for Registration of Pharmaceuticals for tervals.
Human Use (ICH): 3.1.13 no observed adverse effect level (NOAEL),
ICH Q2 Validation of Analytical Procedures: Text and Meth- n—greatest concentration or amount of a substance found by
odology experiment or observation which causes no detectable adverse
ICH Q9 Quality Risk Management alteration of morphology, functional capacity, growth,
development, or life span of the target organism under defined
3. Terminology conditions of exposure.
3.1 Definitions: 3.1.14 operational qualification (OQ), n—establishing by
3.1.1 analyte, n—a substance (usually a residue) for which objective evidence process control limits and action levels
an analysis is being performed. The residue determination may which result in product that meets all predetermined require-
be qualitative, quantitative, specific, non-specific, and/or it may ments.
involve compositional identification. The analyte may be
3.1.15 process qualification (PQ), n—establishing by objec-
determined as an extract or directly on the surface of the device
tive evidence that the process, under anticipated conditions,
or portion (subassembly) of the device.
consistently produces a product which meets all predetermined
3.1.2 blank, n—an analytical sample taken to establish the requirements.
background value for an analytical measurement which may be
3.1.16 recovery study, n—a laboratory study combining the
subtracted from an experimental value to determine the “true”
sampling method and analytical method to determine the
value.
quantitative recovery of a specific residue for a defined surface.
3.1.3 clean, n—having an level of residues and environmen-
3.1.17 residue, n—a substance present at the surface of an
tal contaminants which do not exceed a maximum permissible
implant or embedded therein that is not explicitly recognized
level for the intended application.
and defined as part of the implant specification. It includes
3.1.4 cleaning, v—removal of potential contaminants from processing-based residues as well as contamination by envi-
an item to the extent necessary for further processing or for ronmental factors (adsorbates).
intended use.
3.1.18 tolerable intake (TI), n—estimate of the average
3.1.5 cleaning process, n—a process that is used to remove daily intake of a substance over a specified time period, on the
any product, process-related material and environmental con- basis of body mass, that is considered to be without appreciable
taminant introduced as part of the manufacturing process. harm to health.
3.1.6 cleaning validation, n—the documented evidence pro-
viding a high degree of assurance that a cleaning process will 4. Summary of Practice
result in products consistently meeting their predetermined 4.1 This guide provides an approach for validating the
cleanliness requirements. removal of contaminants and residues introduced during the
3.1.7 cleaning verification, n—a one-time sampling and intermediate process steps so that the terminal cleaning process
testing to ensure that a medical device has been properly can result in a consistently clean medical device.
cleaned following a specific cleaning event. 5. Significance and Use
3.1.8 contaminant, n—any material that potentially ad- 5.1 This guide describes an approach to validate a cleaning
versely impacts the assembly, the functioning of the device, system for a medical device. It is based on the manufacturer’s
and/or shows undesirable interaction with the host. A contami- accurate and comprehensive understanding of their internal
nant may be a single component or any combination of manufacturing and cleaning processes.
components. Examples of possible types of contaminants
include: (1) biological or non-biological in nature; (2) living or 5.2 This guide is not intended to provide a detailed plan or
dead; (3) particles or thin films; (4) solid, liquid, or vapor; (5) road map, but will provide considerations that can be used by
organic or inorganic. the device manufacturer to develop a detailed plan for perform-
ing cleaning validation.
3.1.9 first use, n—the initial contact with biological materi-
als or fluids. 5.3 In cleaning validation, as with other types of validations,
3.1.10 installation qualification (IQ), n—establishing by there are multiple ways to achieve a compliant, scientifically
objective evidence that all key aspects of the process equip- sound and practical cleaning validation program.
ment and ancillary system installation adhere to the manufacut- 5.4 There are several reference documents identified in
er’s approved specification and the recommendations of the Appendix X3 that describe cleaning validation approaches for
supplier of the equipment are suitably considered. non-medical devices (including cleaning for oxygen-enriched
3.1.11 lowest observed adverse effect level (LOAEL), environments, pharmaceuticals, semiconductors). Any of these
n—lowest concentration or amount of a substance found by reference documents could provide guidance for a well defined
experiment or observation which causes detectable adverse process for establishing a manufacturer’s minimum expecta-
alteration of morphology, functional capacity, growth, tion of a specific cleaning validation program.
development, or life span of the target organism under defined 5.5 This guidance specifically targets cleaning validation for
conditions of exposure. medical devices, in-process and at terminal cleaning so that the

2
 F3127 − 16
result is a consistently clean medical device that meets the cleaning agents to medical devices. Chemical compatibility of
performance expectations for that device. the cleaning process should be determined prior to cleaning
process validation.
6. General Requirements (4) In some instances, the structure of the device or the
surface of the device may cause liquid or vapor-phase residue
6.1 This guidance for the validation of cleaning processes is
to be entrapped. Such occurrences are generally not considered
divided into 3 sets of activities: understanding the upstream
to constitute a materials compatibility problem, if the residue is
manufacturing process, documenting the cleaning process, and
readily removed with extensive rinsing and/or drying (bake-
establishing the measurement tools used to evaluate cleanliness
out). However, given the potential negative impact on perfor-
and to establish the cleaning performance criteria.
mance and/or interaction with the host, the design and mate-
6.2 Preliminary process characterization, whether in the rials of construction may qualitatively and quantitatively
laboratory or on the manufacturing floor, provides the data impact the rinsing and/or drying portions of the cleaning
necessary to establish cleaning parameter control ranges. process.
7.3.2.2 While the discussion of device design (design for
7. Cleaning Validation Approach cleanability) is critical to a cleaning validation, a full discus-
7.1 A typical approach to a cleaning validation includes: sion is not within the scope of this guide.
7.1.1 An assessment of the risks and benefits of the cleaning 7.3.3 Risk Analysis:
process and the impact of the cleaning processes on the 7.3.3.1 The risks and benefits associated with a specific
medical device and on downstream processes. cleaning process should be addressed. There are a number
7.1.2 Identification of contaminants from raw materials and approaches to evaluating the risks associated with a cleaning
manufacturing and processing operations (e.g. machine oils) process, including those described in ISO 14971 and ICH Q9.
that could be residuals on the medical device. 7.3.3.2 The process risks evaluated should include the risk
7.1.3 Establishment of allowable limits for contaminants to the patient.
(determining “How clean is clean?”) based on the product and 7.3.3.3 All cleaning operations should be considered, in-
process needs. Acceptance criteria for “clean” should be stated cluding processes conducted by contract manufacturers.
with scientific justification for the criteria. (1) Some cleaning operations may not be termed cleaning;
7.1.4 A validation of the analytical methods used to measure and the terminology may be specific to a given technical field.
the residues or contaminants. Passivation, surface preparation, and surface modification may
7.1.5 A qualification or determination of the sampling or may not have a cleaning function. The manufacturer should
techniques used for evaluating the cleanliness of a medical determine the function and efficacy of each process.
device. (2) If an in-process cleaning operation is considered to be
7.1.6 A determination that statistical requirements and critical and therefore should be validated, acceptance limits for
documentation are adequate to conclude that the result of this in-process operation may be established by considering the
testing meets the output specification of the process. effect of residue levels after this operation on the final residue
levels of the device following the final cleaning step. For
7.2 A general process flow for a cleaning validation pro- example, a manufacturer may perform an OQ on this in-
gram is represented by the Fig. 1: process step to see what in-process residue levels start to
7.3 Definition of the Cleaning Process: impact the final residue levels beyond their acceptable levels.
7.3.1 The definition of the process should include an evalu- By reducing the in-process residue levels below this limit, the
ation of the device, the equipment to be used for the cleaning manufacturer can establish the process conditions for validat-
process, the process parameters, the process chemicals, and the ing this in-process operation.
manufacturing materials that should be removed by the pro- 7.3.3.4 Risks that should be considered include the impact
cess. on the subsequent process yields or the potential for carryover
7.3.2 Device Design: of residue to the next process or the final product.
7.3.2.1 The design, material composition, and intended end 7.3.4 In-process cleaning operations that are not critical to
use of the device have a significant impact on the suitability of subsequent processes or the final product could be included in
a cleaning process. A non-exhaustive list of examples are other process validation activities or, if appropriately justified,
provided: may not need to be validated.
(1) A cleaning process that will not reach a blind hole in a 7.3.5 Cleaning Process Development:
medical device will not get the blind hole clean. 7.3.5.1 The process development should include the devel-
(2) Densely populated electronics assemblies may not be opment of a process flow chart.
readily accessed by cleaning chemistries. As a result, conduc- 7.3.5.2 The process flow chart should begin with the process
tive and non-conductive residue may remain. steps immediately after the previous validated cleaning step
(3) The cleaning process should not have an adverse effect (all steps subsequent to the previous validated cleaning step are
on the materials of construction of the medical device, the residue inputs to the current cleaning step). The process flow
cleaning equipment, or the functionality of the medical device. chart should end after the cleaning operation and should
For example, for plastic devices, ASTM D543 may be used for include an evaluation of the impact of the cleaned device on the
guidance on how to determine the suitability of specific subsequent operations.

3
 F3127 − 16

FIG. 1 Process Flow for a Cleaning Validation

7.3.5.3 The process flow chart and an appropriate list of device. Without knowing the contact materials, the definition
materials should be detailed enough to identify all of the of an adequate cleaning process is incomplete.
materials (including metalworking fluids, polishing (1) The device manufacturer should work with the suppli-
compounds, glove contaminants, cleaning agents, etc.) that ers of process materials to assure that a consistent composition
come in contact with the in-process component or medical is obtained. Identifying the composition of process materials

4
 F3127 − 16
includes, at a minimum, obtaining a Material Safety Data Sheet 7.3.6.5 Cleaning processes are generally comprised of mul-
(MSDS). However, the focus of an MSDS is worker safety tiple steps. Each step of the process should have a function and
issues and therefore may not reveal ingredients that may have a set of parameters that are controlled within defined ranges to
an undesirable interaction with the process, with the device, or ensure effective residue or contaminant removal. The process
with the patient. Further, many process materials, notably parameters for each step of the process should be identified and
metalworking fluids and cleaning agents, may be complex specified in detail and should be based on empirical evidence.
blends where individual components are present at levels that 7.3.6.6 Factors to identify and specify in detail may include
do not have to be listed on the MSDS. the use and type of detergents, solvent grade and lot
7.3.5.4 The device manufacturer should work with the information, the presence of an acid cleaning step, the concen-
suppliers of process materials to develop a plan for managing tration of cleaning agents, the contact time of cleaning agents,
product changes. This is in recognition that products may be feed pressure or flow rate, cleaning temperature, sonication
reformulated in response to environmental mandates or worker energy, ultrasonic frequency, spray pressures, required length
safety issues. These new formulations have the potential to or volume of rinse steps, required conditions for drying and/or
have an adverse impact on the product. bakeout, length of time or number of parts between tank clean
7.3.5.5 Based on the process flow and the risk analysis, a out cycles and the wait time between cleaning steps in addition
validation plan that identifies all validation activities required to other process specific parameters.
to demonstrate the suitability and effectiveness of the cleaning 7.3.6.7 Each cleaning process line should be considered
process should be developed. The validation plan should independently. The burden of validation for multiple cleaning
provide rationale for product type groupings, process lines might be reduced based on identical cleaning equipment
definition, sample size selection, numbers of runs, types of and processes (i.e., process equivalency). Each firm is respon-
analyses, and acceptance criteria. sible for determining and justifying the specific criteria for
7.3.6 Process Qualification: cleaning equivalency between cleaning processes.
7.3.6.1 The plan should consider the requirements of use 7.3.6.8 The process qualification (PQ) combines the actual
and can incorporate risk management to prioritize certain facility, utilities, equipment (each now qualified), and the
activities and to identify a level of effort in both the perfor- trained personnel (including required training programs) with
mance and documentation of qualification activities. The plan the commercial manufacturing process, control procedures,
should identify the following items: and components to produce commercial batches. A successful
(1) The studies or tests to use, PQ should confirm the process design and demonstrate that the
(2) The criteria appropriate to assess outcomes, cleaning process performs as expected. The decision to begin
(3) The timing of qualification activities, manufacturing should be supported by data from commercial-
(4) The responsibilities of relevant departments and the
scale batches.
quality unit, and
(5) The procedures for documenting and approving the 7.3.6.9 Data from laboratory and pilot studies can provide
qualification. additional assurance that the commercial cleaning process
performs as expected.
7.3.6.2 The project plan should also include the require-
ments for the evaluation of changes. Qualification activities 7.3.6.10 The approach to PQ should be based on sound
should be documented and summarized in a report with science, the overall level of product and process understanding,
conclusions that address criteria in the plan. and demonstrable control. The cumulative data from all rel-
7.3.6.3 Installation Qualifications (IQ) should be performed evant studies (e.g., designed experiments; laboratory, pilot, and
on all equipment used in the cleaning process prior to any commercial batches) should be used to establish the process
validation activities. At a minimum the IQ should include conditions for the PQ. To understand the production cleaning
verifications that utility systems and equipment are built and process sufficiently, the manufacturer will need to consider the
installed in compliance with the design specifications (e.g., effects of scale. However, it is not typically necessary to
built as designed with proper materials, capacity, and explore the entire operating range at production scale if
functions, and properly connected and calibrated). assurance can be provided by process design data. Previous
credible experience with sufficiently similar products and
7.3.6.4 The operational qualification (OQ) establishes the
processes can also be helpful. In addition, objective measures
ability of the processing equipment to execute the cleaning
(e.g., statistical metrics) are strongly recommended wherever
operation within the allowable process parameters. At a mini-
mum the OQ should include verification that utility systems feasible and meaningful to achieve adequate assurance.
and equipment operate in accordance with the process require- 7.3.6.11 In most cases, PQ will have a higher level of
ments in all anticipated operating ranges. This should include sampling, additional testing, and greater scrutiny of process
challenging the equipment or system functions while under performance than would be typical of routine production. The
load comparable to that expected during routine production. It level of monitoring and testing should be sufficient to confirm
should also include the performance of interventions, stoppage, uniform product quality throughout the batch. The sample size
and start-up as is expected during routine production. Operat- should be statistically justified for each objective acceptance
ing ranges should be shown capable of being held as long as criterion. A minimum of three production lots should be
would be necessary during routine production. Worst-case evaluated to capture production variation prior to cleaning.
product should be tested at the process challenge conditions. 7.3.7 Routine Monitoring:

5
 F3127 − 16
7.3.7.1 An output of the cleaning validation should include (12) Bubbling parameters,
establishment of ongoing routine process monitoring at prede- (13) Spray parameters (when used),
termined intervals. (14) Current density in electrolytic descaling systems,
7.3.7.2 The collection and evaluation of information and (15) Required length or volume of rinse steps, and change-
data about the performance of the cleaning process, should out cycle (max number of parts cleaned or cleaning cycles
allow detection of undesired process variability. Evaluating the performed prior to a change),
performance of the cleaning process can identify problems and (16) Required drying conditions,
determines whether action should be taken to correct, (17) Rack configurations,
anticipate, and prevent problems so that the cleaning process (18) Rack quantities (min and max quantities in the racks,
remains in control. and min and max quantities of racks should be considered for
7.3.7.3 An ongoing program to collect and analyze product validations. Standard loading conditions will be defined, along
and process data that relate to product quality should be with worst-case loading conditions. Note that there should be
established. The data collected should include relevant clean- evidence to justify worst case conditions), and
ing process parameter monitoring, trends and quality of incom- (19) Wait times between process steps.
ing materials or components, in-process material, and cleanli- 7.3.9.3 The documentation of the cleaning validation should
ness of finished products. include:
7.3.7.4 The data should be statistically trended and re- (1) Process flow diagrams,
viewed. The information collected should verify that the device (2) Process risk assessments,
cleanliness is being appropriately controlled throughout the (3) Validation plans (including, but not limited to, catego-
process. rization of products, sample size selection and rationale,
7.3.7.5 The methods used for monitoring the cleaning pro- numbers of runs, types of analyses, acceptance criteria),
cess should be included in the cleaning validation process. (4) IQ, OQ and PQ protocols and reports,
7.3.8 Re-Validation: (5) A written statement providing a conclusion about the
7.3.8.1 Another output of the validation activities should be suitability of the process to clean effectively,
a schedule for periodic consideration of re-validation of the (6) Criteria for routine monitoring, and
cleaning processes. (7) Criteria for re-validation.
7.3.8.2 Any changes in the process flow (addition of new 7.4 Acceptance Limits:
equipment, changes to the process parameters, changes to 7.4.1 The process cleanliness requirement should be defined
upstream processes or processing materials, changes to the and documented. The process output requirement as well as
cleaning agents, etc.) should be assessed to determine whether expected end use and risk analysis factor into the definition of
re-validation should be performed and the extent of the cleanliness.
re-validation. 7.4.2 The output requirements (measurements of residue
7.3.8.3 A periodic review of deviations from the original levels) of the cleaning process should be determined,
validated cleaning process should be conducted to evaluate if a established, and justified by the manufacturer. These criteria
re-validation is required. The review should be thorough for “clean,” or acceptance limits, should be stated with
enough to determine if the deviations are enough to warrant scientific justification (see Appendix X1).
re-validation. 7.4.3 There are many ways to establish acceptance limits for
7.3.8.4 Routine monitoring data used with periodic reviews a cleaning process.
could provide data to justify continued processing without 7.4.3.1 For existing processes, analysis of current compo-
revalidation. nents or product, analysis of product taken from the field,
7.3.9 Documentation: and/or analysis of product returned due to expiration can be
7.3.9.1 The process inputs for the cleaning process should helpful in establishing a baseline result that reflects the current
be defined and documented. state. The current state may provide an acceptable rationale of
7.3.9.2 The documentation of the cleaning process should suitability, assuming no associated complaints or adverse
include, but not be limited to, the following, as defined and events that can be tied to manufacturing material residues or
pertinent to the user’s process: contaminants.
(1) Water quality (and conditioning/treatment), 7.4.3.2 For new processes, or processes with limited prod-
(2) Solvent quality, uct clinical history, several techniques can be used to determine
(3) Makes, models and serial numbers of the equipment, the suitability of cleaning including quantifiable specific and
(4) Verification of preventative maintenance of tanks to non-specific methods and qualitative methods.
prevent contamination build up, 7.4.3.3 ISO 10993-17 provides a method for calculating the
(5) The concentration of cleaning agents, tolerable intake (TI) limits of leachable substances based on a
(6) Cleaning agent type (Brand and manufacturer), substance’s “No Observed Adverse Effect Level” (NOAEL)
(7) The contact time of cleaning agents, and “Lowest Observed Adverse Effect Level” (LOAEL). These
(8) Feed pressure or flow rate of cleaning agents, calculated TI’s can be converted into a cleaning requirement.
(9) Cleaning temperature, The method for establishing limits of leachables requires a
(10) Cleaning agitation requirements, detailed knowledge of all leachable contaminants that come
(11) Verified delivered ultrasonic power (when used), into contact with the component or device. It is based on a

6
 F3127 − 16
review of toxicological data that establishes a “no adverse residues in questions. Using this method the LOD can calcu-
effect level” for a material or agent. The calculations determine lated from the regression curve:
a tolerable intake value for specific materials or agents. LOD 5 y-intercept13*SE ~ standard error of the regression line!
7.4.3.4 For manufacturing materials that do not have well
(1)
studied toxic responses, appropriate data may need to be
developed to justify the suitability of residue limits. ANSI/ 8.2.1.3 Samples that are at a level at or below the limit of
AAMI/ISO 10993-5, ANSI/AAMI/ISO 10993-11, USP <87> detection are referred to as “Non-Detectable”.
and USP <88> provide guidance on methods to establish 8.2.2 The limit of quantitation (LOQ) is generally defined as
suitable limits for manufacturing materials that are not well 10 times the standard deviation of the blank.
studied. 8.2.3 For instrumental methods, this is often considered to
7.4.4 Visual inspection techniques, which should be the first be 10 times the average value of the noise.
cleanliness inspection step, are often used to evaluate the 8.2.4 An alternative method for determining LOQ is based
aesthetics like “visually clean” (at some defined level of on detectability through analysis of serial dilutions of the
magnification and under defined lighting conditions), visible residues in questions. Using this method the LOQ can calcu-
debris or residue, consistent color, discoloration, or presence of lated from the regression curve:
surface imperfections.
LOQ 5 y-intercept110*SE ~ standard error of the regression line!
7.4.5 There is often a requirement to be microbiologically
(2)
clean. Most of the time the biologically clean requirement is
associated with the finished product. It can also apply to 8.2.5 Samples that are at a level at or above the limit of
in-process cleaning operations to minimize the carryover to detection, but below the limit of quantitation, are referred to as
subsequent operations. See ANSI/AAMI ST72, USP <87> and “Non-Quantifiable”.
USP <88> for guidance on methods to evaluate biological
8.3 The specificity and limit of detection (sensitivity) of the
contamination.
analytical method used to detect residuals or contaminated
7.4.6 Note that there are conditions and cleaning parameters
should be determined.
in which the cleaning agent, itself can leave or create unac-
ceptable residues/contaminants or alter the surface of the 8.4 If levels of contamination or residual are not detected, it
component. The cleaning agent should be treated exactly like does not mean that there is no residual contaminant present
any other process residue or contaminant. Acceptance criteria after cleaning. It only means that the levels of contaminant
for residual cleaning agents should be established just as they greater than the sensitivity or detection limit of the analytical
are for any process material, and analytical techniques shall be method are not present in the sample.
established for measuring the residual cleaning compounds. 8.5 All methods of evaluation of process output (whether
Manufacturers of cleaning agents can sometimes contribute quantitative or qualitative, or specific or non-specific) should
appropriate certification and testing or testing methods. The be evaluated to establish method suitability (adequate limits of
composition of some complex cleaning agent blends may have detection and quantification), accuracy, precision, linearity,
to be changed in response to safety and/or environmental range, reliability, and robustness. For example, visual exami-
regulatory considerations, and such changes may result in nation may not be adequate to identify the presence of
undesirable cleaning and/or unacceptable surface residue. microgram quantities of aqueous cleaning agent residue. Test
Therefore, part of the quality program should include provi- suitability should be demonstrated and justified based on data.
sions for notification of such changes by suppliers. ASTM E2857, USP <1225>, and ICH Q2 are standards that
describe analytical method validations.
8. Analytical Methods
8.6 The analytical method should be challenged in combi-
8.1 Use of appropriate analytical methods is essential to any nation with the sampling method used to show that contami-
cleaning validation program. Analytical methods should be nants can be recovered from the device and at what level, (e.g.,
demonstrated to adequately detect the residues of concern at or 50% recovery; 90% recovery) they can be recovered.
preferably below the acceptable limits. Additionally, adequate NOTE 1—ASTM F2459 requires 75% recovery on the gravimetric
recovery should be defined and demonstrated to justify the analysis.
appropriateness of the method (see Practice F2847). Selection 8.7 Inspection processes that only yield a pass/fail result
of an analytical method depends on the nature and level of the cannot be qualified using standard Repeatability and Repro-
expected residue after the cleaning process. ducibility Testing (R&R) techniques, so in these cases fault
8.2 If a method results in a “Non-Detectable” or “Non- seed testing (or other options for qualifying pass/fail testing)
Quantifiable” response at a level that is higher than the can be used. Fault seed testing can be conducted by randomly
acceptable limits, then it is not an appropriate method. testing both acceptable and unacceptable product, and verify-
8.2.1 The limit of detection (LOD) is generally defined as 3 ing that the inspection process yields the desired disposition.
times the standard deviation of the blank. The inspector should not know which product is acceptable,
8.2.1.1 For instrumental methods, this limit is often consid- and ideally should be unaware that the process is being tested.
ered to be 3 times the average value of the noise. Acceptance criteria are then based on the criticality of the
8.2.1.2 An alternative method for determining the LOD is attribute being inspected. For automated processes, generally
based on detectability through analysis of serial dilutions of the all fault seeded product should be rejected.

7
 F3127 − 16
8.8 It is important to establish analytical method suitability 8.13.2 The analytical method validation should include the
before any conclusions can be made about a cleaning valida- sampling technique as a confounding factor for interference
tion based on the sample results. determination and for recovery studies.
8.9 Specific Analytical Methods: 8.13.3 Direct Surface Sampling:
8.9.1 Specific analytical methods are those which measure a 8.13.3.1 Direct surface sampling using surface analytical
specific residue in the presence of expected interferences. techniques like Photoelectron Spectroscopy (PES), Time of
8.9.2 The advantage of a specific analytical method is that it Flight Secondary Ion Mass Spectroscopy (TOF-SIMS), Energy
provides specific measurements of the major residue of con- Dispersive Spectroscopy (EDS) and micro-FT-IR can provide
cern. direct sampling of surfaces. These techniques can have the
8.9.3 Examples of methods that can be specific are: advantage that they provide immediate results about specific
8.9.3.1 Gas Chromatography with a Mass Spectrometer sites on surfaces. They also can provide direct evidence at the
Detector (GC/MS), worst-case locations or at the best-case locations. While there
are no specific recovery issues, direct surface sampling can
8.9.3.2 Infrared Spectroscopy, including micro-Fourier
have the disadvantage that the techniques may not provide an
Transform Infra-Red (FTIR) Spectroscopy,
overall picture of the device. It may lead to erroneous conclu-
8.9.3.3 High Pressure Liquid Chromatography (HPLC),
sions because of sampling bias. Direct surface sampling is
8.9.3.4 Enzyme-linked Immunosorbent Assay (ELISA)
dependent on how and where sampling sites are chosen.
assays, and
Imaging techniques, based on direct surface sampling, can also
8.9.3.5 Gel Electrophoresis (sodium dodecyl sulfate (SDS) provide an overall view of the relative distributions of residues
polyacrylamide gel electrophoresis (PAGE)). and contaminants. Each of these techniques has technique-
8.10 Non-Specific Analytical Methods: specific requirements for the sample surface such as a require-
8.10.1 There are several non-specific analytical methods ment for surface flatness, the ability to withstand high vacuum,
that can be useful for detecting the presence of residues or depth of penetration, and access to the desired sample location
contaminants. due to equipment limitations. The costs associated with these
8.10.2 Non-specific methods measure a general property techniques can also be a limitation.
which could be a combination of residues or contaminants. 8.13.4 Swab sampling is also a direct surface sampling
8.10.3 The advantage of non-specific methods is that it technique that is reasonably cost-effective. It provides infor-
provides a measurements of total levels of residues or contami- mation about the specific sites selected and swabbed. Swab-
nants of a given type, organic, inorganic, biologic, particulate, bing protocols that reduce swab sampling bias should be
etc. developed. A limitation of swab sampling is that the swab
8.10.4 Examples of methods that are not specific to a should “release” the residue or contaminant. The ability of the
particular contaminant include: swab to release the residue should be considered in the
8.10.4.1 Gravimetric Analysis, recovery study. There are also potential interferences from the
8.10.4.2 Total Organic Carbon (TOC), swab (based on swab material composition) that should be
8.10.4.3 Total Protein, considered and minimized. Swabbing is a manual operation so
8.10.4.4 Conductivity, procedures should be established to develop sampling consis-
8.10.4.5 Visual Inspection, and tency.
8.10.4.6 Water contact angle. 8.13.5 Rinse sampling and extraction by immersion in-
volves the use of a solvent to contact all surfaces of a sampled
8.11 Microbiological Test Methods: item to quantitatively remove the residue or contaminant. The
8.11.1 Control of the bioburden and endotoxin in a cleaning solvent can be water, water-based or organic, depending on the
process is important to ensure that subsequent sterilization or relative solubility of the residue or contaminant and the
sanitization procedures achieve the necessary sterility assur- composition of the medical device. Different solvents can be
ance. Methods to evaluate residual bioburden include ANSI/ used to evaluate residues of different solubility on the same
AAMI ST72, USP <85>, and ANSI/AAMI/ISO 11737-1. sample groups.
8.11.2 Depending on the medical device, both bioburden
and endotoxin are monitored and controlled during the manu- NOTE 2—If a sample is used for one rinse or extraction, the exact same
part should not be used again with a different solvent.
facturing and cleaning processes.
Sampling and extraction can be assisted by ultrasonic
8.12 Biocompatibility Testing: agitation, reflux, bubbling, or with heat. The residue in the
8.12.1 Biocompatibility testing (e.g., cytotoxicity as de- collected rinse solution is measured using either a specific or
scribed in ISO 10993-05) can be appropriate for determining if non-specific method. Particle collection and quantification can
the output of a cleaning process meets its specified require- occur per ASTM F2459.
ments. 8.13.5.1 Rinse sampling has the advantage that it reaches
8.13 Sampling: often inaccessible locations on the device or product. It
8.13.1 Preparation of samples of residues and contaminants provides an average or overall picture of the cleanliness of the
for cleaning validation testing for analytical testing is as critical device or product.
as the test itself. If a sample is prepared inappropriately, the 8.13.5.2 The disadvantage of rinse sampling is that it is
result will also not be appropriate. dependent on the solubility of the residue. For reasons of cost

8
 F3127 − 16
and time, there can be a tendency to use only a single solvent. 8.13.5.6 There are available standards for some rinse sam-
Multiple solvent rinses, using solvents of differing solubility, pling techniques including Practice F619 and Test Method
can provide a more complete picture of the cleanliness of the F2459.
device or product because other residues, having different
polarities, can be identified. It is essential that solvent(s) which 9. Sample Size
are verified to be capable of dissolving all known residues
without affecting the medical device be chosen, as the results 9.1 The sample size required to reach a justifiable conclu-
could lead to an false positive. sion for a given cleaning validation is dependent on a valid
8.13.5.3 The method validations should consider the impact statistical approach, the analytical technique, the variability of
of volume reductions (evaporation of solvent to increase the the analytical technique, and the desired outcome of the test.
concentration of the analyte) of the extracting solvents to Factors that should be considered include the following:
ensure that the concentration of the extracting solvent does not 9.1.1 Will the results of the process evaluation be compared
contribute to the result. Documentation of extraction process to an existing process?
should include specifying the appropriate quantity and the 9.1.2 Will the results be used to develop a process?
appropriate quality of water and/or one or more of the correct
quality of extraction solvent(s). The extraction process (includ- 9.1.3 Will the results be used to qualify a new process?
ing temperature, force, and time) should also be developed and 9.2 In order to give some guidance on how sample sizes can
documented so that the residue is identifiable. In other words, be determined for a cleaning validation, some examples of
the residue cannot be swamped by artifactual interferences approaches to establish sample size are provided in Appendix
from the extraction media or the extraction process. Appropri- X2 and the validation example.
ate controls to establish the suitability of the solvent to extract
the residue should be evaluated and considered in a recovery 9.3 Note that the sample size justification and statistical
study. Potential interferences from the solvent should be procedures used to analyze the data should be based on sound
considered and minimized. Both positive (recovery) and nega- scientific principles and should be suitable for reaching an
tive controls should be specified. appropriate and justifiable conclusion.
8.13.5.4 In some instances, a volatile residue should be
9.4 If a statistically significant sample size is overly
considered. In such an instance, rather than an extraction in a
liquid, extraction in the vapor phase combined with head-space burdensome, a nonstatistically-based sample size may be
gas chromatography (GC) may be appropriate. justified.
8.13.5.5 Certain types of devices (e.g., porous devices and 9.5 Questions that should be considered and resolved prior
some coatings) may not be suitable for surface detections to writing the cleaning validation plan include:
methods.

Questions If Yes If No
Is the process controllable and adequately defined? The process is ready for Then define the process and
validation establish approaches to control
the process

If the process is not controllable and adequately defined, Develop the new process and Verify the output of the process
is there a need to develop a new process such that it is validate
controllable?

If the process will be fully verified, is the verification cost- Verify the output of the process Consider validating the process
effective?

Has the process been defined? Begin planning the validation Define the process

Has a process flow chart been developed? Define the inputs and outputs of Develop the flow chart
the process

Identified process inputs (raw material, manufacturing Consider the process parameter Identify all manufacturing
materials, subcomponents, fixtures, tooling, etc.)? specifications materials, cleaning agents,
fixtures, etc.

Manufacturing materials Address these items in the risk

• Blast grit analysis and in the validation plan
• Oils
• Polishing compound properties
• Coolant types
• Cleaning Solvents
• Tooling and fixture materials
• Masking materials
• Water/Air quality
• Other variables as deemed critical

Has the step at which inputs enter process been Address these items in the Complete the process flow chart
dentified? validation plan to include the inputs

9
 F3127 − 16

Questions If Yes If No

Have process parameter specifications (including limits) Use the process parameters in Develop designs of experiments
been defined? the OQ to determine appropriate process
parameters

Manufacturing processes variables (from Flow Chart) Address these items in the Consider the impact of each of
• Blasting validation plan the process variables with an
• Polishing appropriate design of experiments
• Machining to establish process parameters
• Mass Finishing
• Annealing
• Dwell times
• Drying
• Other variables as deemed critical

Have the necessary process controls been identified? Validate the process controls Develop appropriate process
control methods

Has the manufacturing process rework that would impact Verify that rework processes are Consider rework in process flow
the output of this process been defined? included in the validation and impact on validation

Has all equipment used in the process been identified? Consider impact through risk Identify all equipment and include
analysis and consider in validation in flow chart and validation
plan planning

Have all utilities required for the process been identified? Include in IQ Verify all utility requirements

Has all software used in the process been identified? Include in IQ Verify all utility requirements

Is all equipment suitable for use? Verify with IQ Consider repair or replacement
and re-execute the IQ

Has all software needed for the process been validated Include in validation plan or Include in validation plan
or will be validated ? validation plan justification

Is the detailed process flow chart complete? Plan validation activities Do not begin planning validation
until complete

Has the Process Failure Mode and Effects Analysis Include in Plan validation activities Do not begin planning validation
(PFMEA) been completed for the process? until complete

Has the risk to the patient or the risk to the downstream Then the result of an evaluation Then the impact on downstream
process been determined? could be used as part of the risk processes should be evaluated to
analysis establish acceptable cleaning
process limits

Does the cleanliness of the component as it exits the Then the impact on downstream Then the acceptable cleaning
cleaning process have an impact on a downstream processes should be evaluated to process limits can determined by
process? establish acceptable cleaning some other factor
process limits

Does the cleanliness of the component impact Then the impact on downstream Document justification in
downstream yields? processes should be evaluated to validation planning activities
establish acceptable cleaning
process limits

Does the cleanliness of the component or medical device Then the impact on performance Document justification in
at this stage in the manufacturing process have an should be evaluated to establish validation planning activities
impact on the performance of the medical device, or acceptable cleaning process limits
instrument, when used by the customer?

Is there a point in the process where manufacturing Identify all manufacturing Identify all manufacturing
material removal from prior processes has been materials starting at this point materials from prior processes
validated?

Are there any Corrective Actions/Preventative Actions Consider failure mode in the Document justification in
(CAPA’s) associated with the process? process validation planning activities

Have acceptance criteria for contaminating Identify suitability of analytical Consider PFMEA
manufacturing materials been established? methods

Is there a clinical history for the medical device? Then the result of an evaluation Then base the risk analysis on
could be used as part of the risk preclinical data only
analysis

10
 F3127 − 16

Questions If Yes If No
Have there been complaints or adverse events Can be used in the risk analysis Sample size selection is based on
associated with inflammatory responses or infection and sample size selection risk to patient
associated with the medical device?

Have there been complaints or adverse events that Then the validation should include Determine if FMEA document
involved a failure or malfunction of the device that can be any corrective actions that result revision is necessary
associated with a cleaning operation? from the root cause analysis

Have the process outputs (product requirements) for Then the OQ can be executed They should be defined prior to
each step been defined? executing the OQ

Have all test methods used to verify product Validate their suitability Test methods should be identified
requirements are met been identified? and validated before OQ or PQ

All test method validations completed? Use for cleaning validation Validate the methods before OQ
or PQ

Has a process validation plan been developed? Execute plan Develop plan based on PFMEA
and process flow chart

Has the process scope been identified? Take scope into account when Determine the scope of the
preparing OQ and PQ validation plan

Has the product scope been identified? Take scope into account when Determine the scope of the
preparing OQ and PQ validation plan

Have appropriate sampling plans been established or No additional action required Develop appropriate sampling
justified? plan based on PFMEA and
process needs

Has it been determine if multiple pieces of equipment are Validate equipment or just not Determine if multiple pieces of
required? validating equipment are required

Have worst-case process conditions based on process Use for OQ Identify using design of
parameter specifications (including downstream rework experiments or similar
and/or reprocessing) been identified?

Is the worst case product identified? Use for OQ Identify using design of
experiments or similar

Have the sampling plans for all process requirements for Use for OQ Identify using design of
OQ been determined? experiments or similar

Have all anticipated sources of variation in the process Use for OQ Identify using design of
been identified? experiments or similar

Has it been verified that the resolution of the inspection Use for OQ Verify prior to starting OQ or PQ
method of all product requirements is adequate?

Has it been verified that the resolution of the equipment Use for OQ Verify prior to starting OQ or PQ
used to measure all process parameters is adequate?

Does the sampling plan include all anticipated sources of Use for OQ Consult the PFMEA
variation?

Has the sampling plan for all product requirements for Use for PQ Consider process risk or justify
PQ determined?

Does the sampling plan include all pieces of equipment? Use for PQ Justify or reconsider sampling
plan

Is an ongoing monitoring and control plan defined? No further action required Consider PFMEA

Is the revalidation plan defined? No further action required Consider PFMEA

11
 F3127 − 16
APPENDIXES

(Nonmandatory Information)

X1. A PERSPECTIVE ON “HOW CLEAN IS CLEAN ENOUGH?”

INTRODUCTION

Each manufacturer of medical devices has the responsibility to remove manufacturing materials to
ensure that they are removed or limited to an amount that does not adversely affect the device’s quality.
There is no specific standard or definition that can be attributed to “How Clean is Clean Enough?”
because it depends on the needs of the product and it should be determined by the manufacturer of the
device. The manufacturer’s rationale for the residue limits established (“How Clean is Clean
Enough?”) should be practical, achievable, and verifiable, and should be based on the manufacturer’s
knowledge of the materials and processes involved.

Visual inspection of incoming cleaning agents can often help determine if the material is
contaminated. Given the number of variables involved, there can be instances where the cleaning
chemistry has changed even though vendor certifications are acceptable. Therefore, the importance of
observation and visual inspection is essential. A formal visual inspection should be part of incoming
QC. Aspects may include color, clarity, or perceived changes such as odor. For example, in a complex
mixture (such as in many aqueous cleaning agents), a change in odor could indicate a change in
formulation or the unanticipated addition of a masking odor. In addition, in solvents or solvent blends,
a change in color or clarity could indicate a breakdown in the cleaning agent that may have resulted
from formulation changes or reaction with storage or transfer materials.

Changes in cleaning agents may be subtle and may not become apparent until the actual production
process. Therefore, a level of judicious wariness on the part of technicians is essential and they should
be considered as part of the quality management system. Instructions to report unusual changes in
cleaning performance, or unexpected reactivity should be part of the assembly instructions as well as
part of initial and ongoing training.

While visual determination of cleanliness is highly subjective, it can be somewhat systematized.4

Therefore, in addition to inspection of incoming cleaning agents, it is suggested that a program to
review the visual appearance of equipment and peripherals as received. Similarly, the appearance of
equipment and peripherals should be monitored periodically.

4
Forsyth et al, “Ruggedness of Visible Residue Limits for Cleaning-Part III: Visible Residue Limits for Different Materials of Construction,” Pharmaceutical Technology,
October, 2013.

X1.1 Example of an Approach to a Cleaning Validation: X1.2.2 Determine the method (assay) to be used to recover
X1.1.1 This description of a validation approach is an and test the residue or contaminant and validate the method.
example of a process that could be used as a basis for X1.2.3 Apply the test, or actual manufacturing, residue or
developing an appropriate and justifiable cleaning validation contaminant to 3 lots of worst-case devices using a technique
plan. It is not all encompassing. The details of the approach that mimics in-use processes.
should be established to meet the requirements for the product
X1.2.4 Allow the test residue or process contaminant resi-
or process and address risks identified during process devel-
due to dry under environmental conditions deemed appropriate
opment. Before executing this validation, verify that all equip-
(e.g., the drying cycle of an automated washing device). The
ment used in the process has been adequately qualified (IQ)
drying times should include a realistic minimum and extended
and that the process has been qualified (OQ). The process
times that reflect worst-case in-use process hold or dwell
qualification (PQ) to demonstrate consistency of the process,
conditions for the device.
should be completed before the cleaning process can be
considered validated. X1.2.5 Process the set of contaminated devices using the
cleaning process to be validated at nominal cleaning process
X1.2 Process Development: conditions. The sample size should be statistically valid for
X1.2.1 Prepare the test residue or contaminant that is each variable test or defined by a standard. For ASTM F2459
appropriate for the device or product that is being tested. gravimetric methods, this is normally six to ten samples per

12
 F3127 − 16
variable test. Other more precise instrumental methods may to capture all process variation. At least one of the lots should
yield lower sample sizes. See Appendix X2 for examples. be taken near the end of the tank clean-out cycle. Implants need
X1.2.6 Analyze the residual or contaminant levels on the 3 not be worst case but should represent the part families cleaned
lots of devices. in the cleaning process being validated. Each PQ lot shall have
a statistically valid number of samples and be analyzed for
X1.2.7 If samples have residual levels that are below the statistical confidence and capability independently of other PQ
LOD or LOQ, then the limits of detection or of quantitation for lots.
the method should be identified in the reporting of the results.
NOTE X1.1—If it is not possible to apply specific or known manufac- X1.3 Report:
turing materials, due to a complex matrix of contact materials, run the
worst-case product or products through the manufacturing process and X1.3.1 Reporting requirements are established by the orga-
represent process variability by gathering at least three lots. Multiple nization performing the cleaning validation.
methods that address the full range of contaminants would need to be used
to have adequate confidence in residue or contaminant removal. X1.3.2 The practice for reporting and assessment of resi-
X1.2.8 PQ processing should also analyze three lots of parts dues on single use implants provided by ASTM F2847 can also
taken at various times in the cleaning cycle between tank provide a template for reporting the result of the cleaning
cleanouts and processed on different shifts/days/weeks in order validation.

X2. APPROACHES TO SAMPLE SIZE SELECTION

X2.1 Insufficient sample size may lead to erroneous conclu- TABLE X2.1 K Factor as a Function of Sample Size
sions. The sample size for any test should be established in the Sample Size, K Factor,
validation plan, and should be based on the criticality of the n 95%/95%

process. 3 7.656
4 5.145
5 4.202
X2.2 There are multiple ways of determining sample size 6 3.707
for cleaning validation, using a variety of statistical models. 7 3.399
Some example methods are shown below. Any approach to 8 3.188
9 3.031
determining the appropriate sample size should be accompa- 10 2.911
nied by a rationale. 11 2.815
12 2.736
X2.2.1 Variable Data based on tolerance interval 13 2.670
(Parametric, One-sided, Standard Deviation Unknown): 14 2.614
15 2.566
X2.2.1.1 Objective—To determine the sample size for medi- 16 2.523
cal device cleaning validation protocols based on the limit 17 2.486
value and estimates of the average and standard deviation. 18 2.453
19 2.423
X2.2.1.2 Definition—A one-sided interval, value, such that a 20 2.396
stated proportion, P, of the population will lie below the value 21 2.371
with a specified confidence 1-α. 22 2.350
23 2.329
X2.2.1.3 Assumed—The data are an independent random 24 2.309
sample from a single population. 25 2.292
(1) The data are normally distributed. 30 2.220
35 2.166
(2) The true mean of the population is not known. 40 2.126
(3) The standard deviation of the population is not known. 45 2.092
50 2.065
NOTE X2.1—Variables:
α is the risk of decision or significance level.
γ = 1-α, is the confidence level.
P is the proportion of the population.
X2.2.1.4 Procedure: K value from Table X2.1 is the minimum sample size to utilize
(1) Identify the data to be used for estimates of average assuming the values selected for Xe and Se are conservative
(Xe) and standard deviation (Se), and determine Xe and Se. values (that is, the actual average is lower and the actual
(2) Identify the selected limit value L (using whatever standard deviation is lower). Manufacturers may include a
criteria the firm is using). “safety” factor by estimating average and standard deviation
(3) The significance level α is 0.05 and the confidence level higher than expected or by selecting a value for n larger than
γ is 95% [or γ = 100(1-α)%]. the minimum value of n associated with Ke. This is to prevent
(4) Determine the factor Ke based on the formula: false failures. One technique is to use an upper confidence
K e 5 ~ L 2 X e ! ⁄S e bound of the process average and standard deviation, however,
(5) Refer to Table X2.1, and select a factor K that is at or this may lead to unpractical, large sample sizes when estimates
below the calculated Ke. The sample size n associated with that are made using small samples. Selection of Xe and Se is

13
 F3127 − 16
balanced between the cost of sampling more parts than needed X2.2.2.2 Table X2.2 provides select results from the equa-
versus cost of false process validation failures. tion for C=0, always rounded up to the next integer value to
(6) In cases where estimates of Xe and Se are unknown assure an adequate sample.
analysis of cleaning validations has shown a sample size of n=6 NOTE X2.4—Other tables can be used to identify appropriate sample
for a cleaning process analyzed with gravimetric methods is sizes for C greater than 0. These tables are also readily calculated using
typically sufficient. many statistics programs.
NOTE X2.2—Following collection of data in a cleaning validation X2.2.3 Variables Testing:
protocol, actual values may be evaluated statistically to determine that the
sample size is adequate. If the actual mean is Xa and the actual standard X2.2.3.1 For a one-sided hypothesis test (assuming that the
deviation is Sa, and the sample size is n, then using the value of K minimum requirement is a Not Detectable value) where there
associated with n in Table X2.1, calculate the value at the upper is a need to detect an increase in the population mean of one
confidence level (95%) as: Xa + (K)(Sa). If the calculated value is below standard deviation (any assigned difference can be detected and
the L, then the sample size is adequate to state “We are 0.95 confident that does not have to be limited to one standard deviation), the
95% of the population values will lie below L.” There are three possible
outcomes: following information is required: α, the significance level of
(a) If the calculated value is below L, then the validation the test, and β, the probability of failing to detect a shift of one
test passes the acceptance criteria. “We are 0.95 confident that standard deviation. To control the risk of accepting a false
95% of the population values will lie below L.” hypothesis, set not only α, the probability of rejecting the null
(b) If the calculated value at 50% confidence 95% of hypothesis when it is true, but also β, the probability of
population (K=1.645) is above L, then the validation test fails accepting the null hypothesis when in fact the population mean
the acceptance criteria. “We do have a high degree of assurance is µ+δ where δ is the maximum allowed error in one sample.
95% of the population values will lie below L.” X2.2.3.2 The minimum sample size, N, is shown below for
(c) If the calculated value at 95%/95% is above L, but the a one sided tests of hypotheses with σ assumed to be known:
calculated value for 50%/95% is below L, then the process is
good but more samples are needed to demonstrate the confi- N 5 ~ Z 12α 2 Z 12β ! 2 SD σ
δ
2
(X2.2)
dence required. “The process produces 95% of the population
below L, but more sampling is needed to demonstrate a high X2.2.3.3 The quantities Z1-α and Z1-β are the statistics from
degree of assurance” If this is the case and additional parts are the normal distribution.
available add more test parts to the sample and recalculate the
upper tolerance limit. If not, create a new sample. To calculate
N test 5 Z α 2 SD
σ
δ
2
(X2.3)

the total number of parts needed, use the mean and standard where:
deviation from the original study and the K factors in Table
X2.1 to determine an appropriate sample size. Ntest = number of samples needed for the test,
α = a predetermined, assumed confidence interval,
NOTE X2.3—A conservative estimate of Xe and Se should be used to
prevent false failures. One method is to calculate the upper 95%
Zα = statistic for a normal distribution,
confidence bound of the mean and standard deviation when calculating E = error between the specification acceptance value and
sample size. the true mean, of the population,
σ = historical standard deviation for the test, and
X2.2.1.5 References: δ = the maximum allowed error of one sample.
Natrella, Mary G., 1963, 1966, “Experimental Statistics;
Handbook 91,” NIST, Library of Congress Number: 63-60072. X2.2.4 t-Test Analysis:
X2.2.1.6 Additional Tables: X2.2.4.1 The sample standard deviation, s, and mean, x̄, will
Hahn, G. J. and Meeker, W. O., (1991). “Statistical approach the population (all cleaned samples ever produced)
Intervals.” John Wiley and Sons, New York, NY. standard deviation, σ, and mean, µ, as the number of tested
specimens n increases towards the total population number N.
X2.2.2 Attribute Testing (pass/fail):
X2.2.2.1 Sample sizes for attributes may be calculated using
the following equation, which is found in Reliability Statistics,
σ5 Œ Σ~xi 2 µ!2
N
;s 5 Œ Σ ~ x i 2 x̄ ! 2
n21
(X2.4)

by Robert A. Dovich.5 The resulting sample sizes assume no X2.2.4.2 As more samples are taken, the standard deviation
test failures within the sample: in the error between the true mean and the calculated mean is
ln~ 1 2 c !
n5 (X2.1) TABLE X2.2 Sample Size for Attributes
ln~ q !
c 0.90 0.95 0.975 0.99 0.999
where: q
n = the sample size (with no failures), 0.75 9 11 13 17 25
c = the confidence level (decimalized percentage), and 0.80 11 14 17 21 31
0.85 15 19 23 29 43
q = the quality level (yield, decimalized percentage 0.90 22 29 36 44 66
non-defective). 0.925 30 39 48 60 89
0.95 45 59 72 90 135
0.975 91 119 146 182 273
0.99 230 299 368 459 688
0.999 2302 2995 3688 4603 6905
5
ASQ Quality Press, Milwaukee, 1990.

14
 F3127 − 16
called the standard error of the mean, denoted σx̄, with the form
(which is based on the population standard deviation):
n5 S D
t cs
E
2
(X2.8)

σ s to determine the minimum sample size n required to obtain

σ x̄ 5 ;s x̄ 5 (X2.5)
=n =n a confidence interval having a margin of error no larger than E,
with specified the sample standard deviation s from pilot data.
X2.2.4.3 Using these computations, an estimate of the
sample size that will be required to accurately determine the X2.2.4.6 Values of E can be estimated based on weighing
true average residue remaining on the components within a resolution of the overall procedure. As tc depends on n, Eq
certain tolerance can be obtained. A confidence interval for the X2.8 needs to be solved iteratively using T-tables.
population mean µ can be constructed using the formula: X2.2.4.7 Following collection of residue levels, the sample
size n should be re-calculated based on the actual standard
x̄6z c σ x̄ (X2.6)
deviation computed from the test data. Additionally, the
X2.2.4.4 However, this discussion assumes that sufficient normality of the sample residue distribution should be verified
information is known about the population statistics to deter- to assure that Eq X2.8 is appropriate.
mine σ accurately. Simply replacing the population statistics X2.2.5 Guidance in selection of an appropriate confidence
(σ) with the sample statistics (s) for a “small” number of level and quality level is provided by Fig. X2.1.
specimens may not be sufficient. Normally, one determines the
sample size required by using accurate estimates of the NOTE X2.5—Fig. X2.1 is provided as a guide for selection of appro-
underlying population quantities. These values should be priate Quality Levels and Confidence Levels, based on the needs and
application of the test. A more detailed analysis can result from using
obtained either through historical data, or through a pilot study. appropriate OC (Operating Characteristic) curves found in several refer-
In this case we propose using a small pilot study to estimate the ences and statistics programs.
population values and use these for the sample size calculation.
X2.2.5.1 Sample sizes in the dark green zone can be
To account for the increased uncertainty by replacing σ with s,
considered highly conservative.
in Eq X2.7, we recognize that the form of the confidence
interval for µ is given by: X2.2.5.2 Sample sizes in the light green zone can be
considered conservative unless risk of patient harm is severe if
x̄6t c s x̄ (X2.7) failure were to occur.
where tc is the critical value from the t-distribution with df X2.2.5.3 Sample sizes in the yellow zone are dependent
degrees of freedom, where df=n-1. upon the criticality of the decision being made (e.g., if used for
X2.2.4.5 Critical values from a t-distribution with given df final product acceptance or final verification and validation
are routinely tabulated and computable in most statistics testing that confirms device function or performance). In such
computer packages. For the same level of confidence (e.g., cases the influence of Type II error and minimizing its impact
95%), the value of tc for any value of df is always larger than to user/patient risk should be considered.
the corresponding value of zc to reflect the extra uncertainty X2.2.5.4 Sample sizes in the orange zone should be evalu-
based on using a sample standard deviation instead of the ated to ensure they are commensurate with risk. These sample
population standard deviation. The formula is used: sizes may be used when there are technical limitations within

FIG. X2.1 Selection Matrix for Confidence Level and Quality Level

15
 F3127 − 16
the measurement process, sample size, or number of trials/ ments where variation is purposefully introduced and studied
replicates that cannot be overcome. (targeted testing) to better characterize and understand the
X2.2.5.5 Sample sizes in the red zone should only be used impact of boundary conditions as such designs generate even
in extenuating circumstances where Type II error is unlikely to more conservative estimates with regard to a given population.
contribute to user/patient risk. Sample size selections in this X2.2.5.8 Other sample size selection criteria may be used if
zone should not be used for final product acceptance, testing to they follow sound scientific principles, are applied
mitigate an intolerable risk or final verification/validation appropriately, and are appropriately justified. Regardless of the
without other supporting confirmatory studies.
sample size selection criteria chosen, careful consideration of
X2.2.5.6 Sample sizes in the blue zone should be used for
the following should be undertaken:
proof of concept, technical feasibility, characterization or
(1) Accuracy and precision (uncertainty) of the analytical
investigative/experimentation purposes and not for the pur-
poses of: final product acceptance, final verification/validation method,
testing, or for testing to mitigate intolerable risk to a tolerable (2) Statistical confidence of the test chosen with regard to
level. the decision needed, and
X2.2.5.7 Any of the above sample size selection criteria (3) How outliers will be identified, analyzed and resolved if
may be used in conjunction with a sound Design of Experi- encountered.

X3. OTHER CLEANING VALIDATION APPROACHES

X3.1 Cleaning validation approaches that have been devel- justification suitability should include the use of coupons in cleaning
oped and documented for other business needs include: validations.
ASTM G122 Test Method for Evaluating the Effectiveness
X3.1.1 Reusable Medical Devices: of Cleaning Agents
AAMI TIR30:2011 A Compendium of Processes, ASTM G131 Practice for Cleaning of Materials and Com-
Materials, Test Methods, and Acceptance Criteria for Cleaning ponents by Ultrasonic Techniques
Reusable Medical Devices
X3.1.5 Printed Circuit Board Manufacturing:
X3.1.2 Parenteral Drug Association: IPC-CH-65B: IPC Guidelines for Cleaning of Printed
Technical Report No. 49: Points to consider for Biotechnol- Boards and Assemblies
ogy Cleaning Validation X3.1.6 Other References and Handbooks:
Technical Report No. 29: Points to Consider for Cleaning Handbook for Critical Cleaning: Cleaning Agents and Sys-
Validation tems (Book 1) and Applications, Processes and Controls (Book
X3.1.3 Pharmaceutical Manufacturing: 2), Barbara Kanegsberg & Edward, Kanegsberg, ed., CRC/
The FDA Guide to Inspections: Validation of Cleaning Taylor & Francis, 2011.
Process “Cleaning and Contamination Control in Medical Devices”
by Barbara Kanegsberg and Edward Kanegsberg, in “Cleaning
X3.1.4 Oxygen-Enriched Systems and Components: and Cleaning Validation Volume 2” Paul Pluta ed., PDA DHI
ASTM G121 Practice for Preparation of Contaminated Test Technical Book, 2013.
Coupons for the Evaluation of Cleaning Agents Validated Cleaning Technologies for Pharmaceutical Manu-
NOTE X3.1—The use of coupons as surrogates may not always reflect facturing by Destin A. LeBlanc (Feb 28, 2000).
the complexity of the medical device design such as long narrow lumens, Cleaning Validation: Practical Compliance Solutions for
textured or rough surfaces, etc. Careful consideration of relevance and Pharmaceutical Manufacturing by Destin A. LeBlanc (2006).

ASTM International takes no position respecting the validity of any patent rights asserted in connection with any item mentioned
in this standard. Users of this standard are expressly advised that determination of the validity of any such patent rights, and the risk
of infringement of such rights, are entirely their own responsibility.

This standard is subject to revision at any time by the responsible technical committee and must be reviewed every five years and
if not revised, either reapproved or withdrawn. Your comments are invited either for revision of this standard or for additional standards
and should be addressed to ASTM International Headquarters. Your comments will receive careful consideration at a meeting of the
responsible technical committee, which you may attend. If you feel that your comments have not received a fair hearing you should
make your views known to the ASTM Committee on Standards, at the address shown below.

This standard is copyrighted by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959,
United States. Individual reprints (single or multiple copies) of this standard may be obtained by contacting ASTM at the above
address or at 610-832-9585 (phone), 610-832-9555 (fax), or [email protected] (e-mail); or through the ASTM website
(www.astm.org). Permission rights to photocopy the standard may also be secured from the Copyright Clearance Center, 222
Rosewood Drive, Danvers, MA 01923, Tel: (978) 646-2600; http://www.copyright.com/

16