Cleaning Validation: What can a Quality Risk Management Approach Look Like?

With the implementation of the ICH Q9 document, risk management has become mandatory in almost all GMP areas. This also applies to cleaning validation. Now, what can a quality risk management approach look like in cleaning validation? The American Society of Testing and Material (ASTM) has recently issued a guideline on this.

The document entitled "Standard Guide for Science-Based and Risk-Based Cleaning Processes Development and Validation", labelled E3106-18, comprises 9 pages with 10 chapters. The first chapter Scope starts pointing out that the Guide applies a life cycle approach to cleaning validation, from development to validation up to cleaning process verification. This approach can be applied to all dosage forms, to active substances and also to cleaning during clinical supply production.

The first 2.5 pages deal both with the scope and referenced documents (Chapter 2) as well as definitions (Chapter 3). Interestingly, some of the definitions still contain discussion subpoints which further interpret the respective definition point. Chapter 4 (Significance and Use) refers to ICH Q8, 9, 10 and 11 and the FDA Process Validation Guidance. In Chapter 5, the use of cleaning development process and validation based on science, risk and statistics is recommended.

Chapter 6 Risk Assessment

The 3.5-page Chapter 6 addresses risk assessment. The following points should be taken into account in the risk assessment:

• Acceptable daily exposure (ADE) values, if available
• The threshold of toxicological concern (TTC) concept
• Microbiological contaminations
• Equipment design
• Handling errors

In the risk analysis, the aspects mentioned above are assigned to a risk. This assignment should also include:

• The development of the cleaning process
• A design review of the facility and equipment
• A review of the cleaning processes and
• Selection of the analysis method.

The risk analysis should also include risk reduction. An important point of the risk analysis is the characterisation of the process residues (solubilities, adhesion behaviour) and the influence of instrument design (material, dead ends, drainability) on the cleanability. If possible, historical data of cleaning results should be integrated. Great importance is attached to the development of the cleaning process. This development should include laboratory studies, the determination of the cleaning parameters and the selection of cleaning agents. Existing cleaning SOPs should also be subjected to a risk analysis. Equipment for cleaning should also have a suitable design. Interestingly, the Guide suggests Design of Experiment (DoE) to optimize the cleaning processes and even a "Cleaning Design Space". Manual cleaning should also be considered with a risk analysis (are there differences between different people?). Regarding automated cleaning systems, the risk of cross-contamination by the system itself should be considered in the risk analysis.

Of course, the grouping of processes and equipment should also be based on a risk analysis. These groupings can then also serve as a basis for factors in the DoE mentioned above. Also the times in which devices may stand dirty before they are cleaned ("Dirty Hold Time") and the time in which the plant may stand clean ("Clean Hold Time") should be considered on the basis of a risk analysis. Where there is no influence on the cleanability due to the holding times, no qualification activities - in this case called qualification activities - have to be carried out. If these hold times are exceeded, a new term is used: "expired equipment hold time" (EEHT). Particular emphasis is placed on monitoring the success of training with manual cleaning.

The result of all risk analyses should finally lead to a "Cleaning Control Strategy", which has to be evaluated regularly.

The next item Sampling covered in the chapter Risk Analysis is very extensive. Of course, a risk analysis should also determine the sampling strategy (representative sampling locations, number and methods). This risk analysis should also include statistical considerations. Somewhat surprisingly, in the case of process residues with low risk and a fully visible surface, even visual evaluations based on a risk analysis are considered sufficient for cleaning validation. Interestingly, this assessment is cited with reference to Annex 15. A direct sampling (swab) before an indirect sampling (rinse) is considered to be preferable. Of course, the choice should also be based on a risk analysis. Fourier Transform Infrared (FTIR), Near Infrared (NIR), Raman, fluorescence and UV spectroscopy are mentioned as possible methods for surface scanning sampling as part of a cleaning verification. Sampling techniques require recovery rates and extensive training of "samplers". Statistical techniques are recommended to determine the accuracy, precision and robustness of the sampling technique. The selection of the analytical method (specific or non-specific) should also be based on science and risk. A master plan for cleaning should form the basis for the cleaning control strategy.

In the risk assessment (evaluation), the cleaning data should be evaluated against the acceptance criteria of the risks. The risk assessment should already consider risk reduction if the risks are too high. The evaluation of cleaning data can be done via "maximum safe surface residue (MSSR) based on ADE data". According to the Guide, microbiological data can be collected using a comparable procedure.

The actual risk-reducing measures, if necessary, are described in chapter 7 on risk control. Possible reduction measures with regard to significance, probability of occurrence and detectability are listed in a separate table. Routine monitoring should also be developed as part of risk control. Routine monitoring could also take the form of statistical process control (SPC). Process analytical technologies (PAT) for the development, analysis and control of the cleaning process, up to continuous monitoring or even parametric releases of the equipment are also recommended. The results of risk-reducing measures should be documented in terms of risk acceptance.

Chapter 8 deals with the "Risk Review". Based on a risk analysis, cleaning and cleaning monitoring results should be evaluated regularly. On this basis, the review cycles can be adjusted if necessary. There is an indication when a new product is introduced. Using the toxicological data as a basis, the integration of this new product into the existing cleaning system should be considered.

Chapter 9 on risk communication points out that risk communication involves the exchange of information on cleaning risks and their control. This concerns the various parties involved. The company itself, various companies among themselves, contract manufacturers, regulatory authorities, etc. are mentioned. Contents of the information exchange should be

• Severity (ADE/PDE)
• Probability (experience from the past)
• Detectability (analytical methods)
• Controls (cleaning procedures and agents)

and always with a view to patient risk.

Chapter 10 includes a keyword index.

Conclusion: The guide shows the implementation of a life cycle approach based on quality risk management with regard to the development, validation and control of cleaning very well. New terms and abbreviations not yet listed in regulatory documents are remarkable. The integration of "modern" methods, such as DoE and PAT for cleaning validation is future-oriented. Interesting is the section regaring the visually clean-criterion as a solely acceptance criteron in cleaning validation with reference to Annex 15. Concerning this the EMA Q &A document could as the circumstances require help to discuss this with GMP inspectors. 

The Standard Guide for Science-Based and Risk-Based Cleaning Processes Development and Validation with the label E3106-18 can be purchased on the ASTM website.