Cleaning Validation Optimization: A Case Study

Fred Ohsiek and David W. Vincent , B.Sc., MPH, Ph.D

How to Perform Cleaning Validation Optimization

Introduction

There are three stages in the process validation lifecycle (process design, process qualification, and continued process verification) [1]. The first stage, process design, is challenging for pharmaceutical and biopharmaceutical firms to plan and execute on time.
In particular, CIP cycle optimization is a struggle to perform due to perceived lengthy timelines and scheduling priorities, so removing it from the commissioning scope is common practice. This will cost firms enormous money and valuable production time, not to mention sustainability (i.e., water conservation, chemical waste, and energy usage).
Cleaning failures are not a cleaning validation problem; they result from a poor understanding of the cleaning process.
Cyle development (CD) and automated clean-in-place (CIP) recipe optimization are usually synonymous. CD typically entails using a CIP recipe from a legacy product, a sister facility, or a vendor or extending the coverage test duration. If the CIP recipe cleans the equipment (i.e., verified via rinse and/or swab samples) without alarming, it is considered developed and ready for cleaning validation (CV).

In contrast, Optimization determines the optimal temperature, detergent concentration, cleaning action (i.e., cascading or impingement), and time needed to clean the equipment.
CIP optimization is usually penciled in when planning commissioning, qualification, and validation (CQV) activities but is later removed from the scope for various reasons. The main reason is due to timeline restraints. As the schedule shrinks, so does the number of development batches. Some other reasons are lack of talent (i.e., individual with the skill set), resources, understanding the importance; or the CIP system is not qualified or even operational.
The reasons can be overcome if the firm understands how quickly it can be completed. In this article, the “how” of Optimization will be discussed. If the “how” is done properly, the additional time and number of batches needed to optimize should be negligible.

Regulatory Expectations

The development of cleaning processes is a regulatory expectation [2].
CV is a process to prove that your cleaning process is reproducible. This is normally accomplished by performing three (3) validation runs. Statistically, three (3) runs provide some assurance that the cleaning process is reproducible, but without fully understanding it, it isn’t easy to justify three (3) runs. Just performing three [3] runs may not be enough. Cleaning failures after validation may indicate that. The more a facility understands the cleaning process, the lower the risk of a cleaning failure after validation.

How to Expedite CIP Optimization

To expedite the CIP optimization, several activities are required. Most activities will be performed eventually, but they can support Optimization if performed early enough.
Completing the following activities in a similar sequence will accelerate the optimization process. Some activities will need to be performed sequentially.

Before Commencing with Optimization

Usually, if a CIP system is not designed properly or balanced correctly, it will consistently abort or alarm out. This is not always the case. A CIP system can run inconsistently for years. Regular aborts, warnings, and irregular cleaning events seem to be acceptable. However, it can indicate that the CIP process was not fully or properly developed and commissioned.
When grouping equipment for cleaning validation, the emphasis is typically on similar equipment dimensions, shapes, and interior components. What is forgotten is cleaning critical process parameters (CPP). The regulatory expectation for grouping is they have the same cleaning recipe. Two (2) like-for-like vessels can have the same CIP recipe but have different CPPs due to supply piping distance and insulation coverage.
Even if equipment is not grouped, operating CIP processes as intended is important. Aborts and having to acknowledge warnings during CIP frequently consume valuable manufacturing time and promote early equipment failure. Irregular cleaning events are future deviations that deplete valuable quality assurance, quality control, validation, operation, and engineering department’s bandwidth.
To facilitate CIP optimization, consider the recommendations in Table 1. Though the CPPs will most likely change during Optimization, understanding the current state, condition, capabilities, and underlying issues will reduce optimization surprises and uncertainty.

Table 1: CIP Recommendations Before Optimization

Experienced Individuals (SME)

For startup projects or new equipment/process additions, firms will search for individuals with similar project experience. When selecting CIP engineers, verify they have the experience and expertise to manage and execute the CIP optimization. Here is a list of reasonable questions:

1. How are development, qualification, and validation connected?
2. Tell me about your CIP design and development experience.
3. Describe your previous CIP system optimization projects.
4. What are the key elements to optimizing?
5. What are some prerequisites to optimizing the CIP system?
6. What is the sampling plan strategy you implemented during execution?

Product Characterization

Product characterization is needed to support cleanability coupon studies and calculate the cleaning limits, which will identify the worst-case component or soil.

All process materials (i.e., active ingredients, impurities, excipients, buffers, solvents) within each process step/manufacturing phase should be assessed. The assessment should include solubilities, chemical properties (i.e., hydrophobic or hydrophilic, pH, stability, reactiveness to detergent solutions), Health-Based Exposure Limit (HBEL) toxicity evaluation (i.e., Allowable Daily Exposure [ADE], Permitted Daily Exposure [PDE], or Threshold of Toxicological Concern [TTC]), handling hazards, and route of administration.

Equipment Characterization

Understanding the equipment will identify hard-to-clean locations and potentially reduce future cleaning issues.

A general description of each process step (e.g., process duration, any description of any physical changes during the process [e.g., lyophilization, filter drying]) should be included. Within each process step, equipment needs to be reviewed carefully. Each piece of equipment or part should be evaluated for difficulty in cleaning. This includes materials of construction (MOC) of direct product contact.

From equipment drawings, cut sheets, and P&IDs, determine:

• Hard-to-clean (HTC) locations
• Sampling location options (i.e., sample ports, equipment disassembly, accessibility)
• Worst-case scenarios (i.e., DHT, CHT, process steps/timing, multi-circuits)
• Visual inspection access

Note: HTC locations should include common industry locations such as gasket seals, under agitators (especially magnetic drive agitators), under domes, and piping elbow.

Keep future potential CV equipment grouping in mind during the review and assessment. If there is a general idea of what equipment will be grouped, the equipment groups, CIP recipe creation, and control can be coordinated. For example, equipment grouped should all have the same recipe and same control type (i.e., flow-controlled or pressure-controlled), not mixed. This will facilitate and expedite the grouping process.

Cleaning Limits

Knowing the cleaning limits early in stage 1 will help determine the level of risk, type of analytical instruments and methods needed, and required sampling methods (i.e., rinse and/or swab).
The cleaning limit is calculated from the HBEL. If products are bracketed using a product matrix approach, toxicity, and hardest-to-clean components must be considered when determining worst-case products and justifying the cleaning limit. For example, the CV is performed on the HTC product using the lowest cleaning limit.

Analytical Method Selection

Once the cleaning limit is determined, it can be decided if a specific or non-specific analytical method is needed. A specific method may be required if the cleaning limit is too low. The type of analytical instruments available onsite or operational will also play a part.
Analytical methods’ role in CIP optimization should also be considered. It will depend on the type of testing performed between cleaning cycles (refer to Strategic Sampling Plan section) and the type of inline CIP testing capabilities (i.e., conductivity, pH, or TOC). Because specific methods (e.g., HPLC) tend to take longer to develop, facilities may consider using non-specific methods (i.e., TOC) for Optimization and then specific methods (i.e., HPLC) for CV. It will depend on the cleaning limit vs. the analytical method’s sensitivity (limit of detection [LOD] limit of quantitation [LOQ]).
A validated analytical method is not required for development and optimization testing, but there needs to be confidence in analytical results. Therefore, the analytical instruments should be validated. If they are not, the testing results may not be accurate or reproducible.

Choosing a Detergent Provider

Choosing the best detergent provider might sound gimmicky, but it is critical. It is important to consider their capabilities, such as soiling methods (coupons only or also small-scale equipment), type of soil removal verification (i.e., visually only, water break, gravimetric, or analytical), analyst experience and knowledge base, and test result turnaround.
Handling and managing active ingredients (especially potent compounds) should be confirmed unless it is known that the excipients are the hardest to clean.
It is also important to know if they will supply analytical methods for testing detergents and toxicity evaluation (i.e., HBEL).
Most detergent providers have some form of capabilities mentioned, but their process and support capabilities must be vetted.

Cleanability Coupon Studies

Cleaning agent selection is important, but it is more important to understand how easy or difficult it is to remove the soil from equipment surfaces. [4] cited a case study where temperature and product cleanability (nature of the soil) were the only statistically significant cleaning CPPs. Agitation, detergent concentration, time, water characteristics, the individual performing the study, and surface (MOC) were not statistically significant.
Additionally, it was noted that numerous pharmaceutical and biopharmaceutical products could be cleaned with just water at ≥ 70°C.
If the firm is going to rely on a detergent provider to perform the cleanability coupon studies, the firm should make three (3) requests:
1. That water is included as a side-by-side for every temperature tested.
2. Include water-only testing at ≥ 70°C if the firm’s cleaning process can withstand that temperature.
3. Ensure the provider is willing to perform enough testing to meet any justifications being made (i.e., DHT).

If a detergent has already been chosen, it is still important to perform the coupon studies. It will help understand how difficult or easily processed soil will be removed.
In cases with limited detergent provider support, the coupon studies can be performed relatively easily in the laboratory with very limited laboratory materials. [3] is a source for performing cleanability studies.

CIP Capabilities and Characterization

The cleanability coupon studies will provide information for choosing CIP skid capabilities and spray devices. The CIP skid must produce enough pressure and flow to clean pipes and pressure/flow to meet vendor requirements for spray device operating ranges.
Generally, rotary jet spray devices utilize impingement cleaning, and spray balls utilize cascading action. Rotary jet spray devices usually have a longer cleaning cycle duration than spray balls, but spray balls tend to use more water and do not have as strong impingement capabilities. Therefore, it is important to correlate the product’s cleanability with the type of spray device.
For pipe cleaning, hard-to-remove soil will require turbulent flow, not laminar flow. Additionally, dead legs and dead ends cleaning time depends on the level of turbulence. This means hard-to-clean (HTC) locations (i.e., elbows, high-rise piping, instruments, and probes) will take a long time or fail to get cleaned without adequate flow.
Therefore, it makes sense to perform cleanability studies very early. A startup project can be held up for long periods if re-engineering is required.

Design of Experiment (DOE) Studies

Coupon cleanability studies are a design of experiment (DoE) studies prerequisite.
Creating an optimal cleaning model via DoE studies is essential for expeditious, reliable, and successful optimization timelines.

Most detergent providers may not be willing or capable of conducting DoEs. DoEs are very time-consuming. Additionally, special equipment can produce rapid and reliable results [4]. Using a third party has a cost, but the reward (i.e., reduced optimization timeline) should outweigh the cost.

If a third party is used, it should be a laboratory specializing in CIP cleaning DoEs. Creating/executing DoE studies and statistically evaluating the data will require experienced individuals. Some third parties can drastically reduce the number of test runs using different modeling methods (i.e., definitive screening designs) instead of the traditional full factorial [4]. This can save time and money. In addition, a third party will be unbiased in cleaning agent selection (water vs. detergent).
The study can also support worst-case location, product bracketing, and equipment grouping. It is also a robust reference for cleaning rationale documents since it is performed by a third party with no expected biases.
The result of the DoE will provide the optimal cleaning agent, concentration, and temperature. It should also provide insight or expectations on cleaning performance and duration when creating the CIP recipe and setpoints for cleaning commercial equipment.

Analytical and Sampling Methods

It should be obvious to have analytical and sampling (rinse and/or swab) methods ready before starting the product process development studies. Still, some facilities will start the product manufacturing development studies without having the methods ready. Analytical instruments should be validated, and analytical methods should be established.
Rinse and/or swab recovery studies must not be completed and approved before initiating development/optimization. Still, there should be confidence that the sampling method will recover the target component.

Operational Qualified CIP System

The CIP skid and associated automation should be qualified and fit for the intended use. At minimum, the CIP process must be operational and functional (no aborts or alarms); and coverage tests (i.e., riboflavin tests) have been completed.
Initial product process development batches are often produced on the equipment, and the equipment is not sampled after cleaning. Samples are not taken because the CIP is not functional or the analytical and sampling methods are not ready. Clearly, this is not the best practice.

Coverage Tests

Coverage tests are performed to verify all interior surfaces of a vessel are contacted with cleaning solutions. Therefore, do not accept unsuccessful coverage tests. Though this seems to be a common industry practice. If an area of the vessel is not in contact with the CIP solution, then any inline rinse samples taken may not contain residue from those locations, causing a false negative test result.

The coverage test can also be used to confirm HTC locations. This can be done by inspecting for riboflavin between bursts. Even though the riboflavin was removed on the third burst, it can be argued that the location where it was not removed until the third burst is a HTC location. Another approach is conducting coverage tests at reduced pressure/flows.

Strategic Sampling Plan

The initial CIP recipe will use the cleaning CPP values determined from the DoE studies. The equipment should be tested after every batch cleaning. The cleaning CPPs are adjusted depending on the results of the sampling test data. If the DoE was performed correctly and the correct spray devices were chosen, only the cleaning duration will need to be adjusted.
During the Optimization of the CIP cleaning process, all cycles (i.e., pre-rinse, cleaning, post-cleaning rinse, and final rinse steps) should be evaluated for several reasons:

1. Optimizing every step makes sense if Optimization is going to be performed.
2. Each CIP step affects the next step. For example, if the cleaning step does not remove all soil from the equipment surface, the next rinse step will have to remove soil from equipment surfaces which is not its design.
3. There is a sustainability opportunity in each step.
4. Reduce cleaning issues during CV.
5. Cleaning understanding will greatly increase, demonstrating to the regulatory agencies that the firm understands its cleaning process.
6. The data will support future change requests.

Sampling Recommendations

Pre-rinse Step:
The purpose of the pre-rinse (usually a once-through rinse) is to remove gross soil from equipment surfaces. The justification for sampling is to determine the optimal rinse volume or time needed to remove the gross amount of soil. This will reduce the number of cleaning agents required and support better cleaning.

The recommended sampling is to take several rinse samples (i.e., target soil components) during the rinse. The concentration of soil will decrease over time. Taking samples at different times will indicate when most of the soil has been removed. Therefore, equipment will not be over-rinsed (water waste) or under-rinsed (cleaning agent waste).
At least for the first run, perform a visual inspection on the equipment after the pre-rinse step in cases where soil was still left on equipment surfaces and not in the rinse sample.

Cleaning Step:

The purpose of the cleaning step is to remove all soil from equipment surfaces and suspend it in the cleaning solution.

The justification for sampling is that extended cleaning cycles increase operating costs (i.e., FTEs, energy, and production downtime. If the cleaning step is once-through, there is the additional cost of cleaning chemicals, water, and waste removal. In rare cases, for recirculation cycles soil can re-adhere to equipment surfaces.

The sampling recommendations are to take rinse samples (i.e., target soil component) throughout cleaning step(s) (e.g., beginning, middle, and end) to determine peak soil removal.

As with the pre-rinse, perform a visual inspection after the cleaning step. The equipment should be relatively clean, with no soil adhered to equipment surfaces. There most likely will be some loose soil. This is acceptable as the post-rinse should completely rinse it away.

The cleaning step testing may also provide insight into whether additional cleaning steps or other cleaning steps (e.g., acid detergent) are needed.

Post-Cleaning Agent Rinses:

The purpose of the post-cleaning step is to remove cleaning chemicals and loose soil, not soil adhered to equipment surfaces.

The justification for sampling is to optimize the rinse volume/time, reducing water, time, and waste.

The sampling recommendation is to review CIP conductivity trending, if there are inline instruments (i.e., TOC analyzer or conductivity probe) to detect cleaning agents. The final post-rinse before the final rinse should have a conductivity near the “Final Rinse” conductivity, as the final rinse should not be utilized to remove the residual cleaning agent.

Final Rinse:

The purpose of the Final Rinse is to verify cleaning agent(s) and (in some cases) process soil has been removed to acceptable levels.

The justification for sampling is to optimize the rinse volume/time, reducing water, time, and waste, especially if only the last step is process water. Process water has a considerably higher cost than city water.

The sampling recommendation is to review CIP conductivity trending.

Final Testing

Following the successful optimization run (all steps are balanced and optimized), perform testing similar to what will be performed for cleaning validation before commencing to CV.

Conclusion

Due to tight commissioning and validation timelines, there is usually a push to develop the cleaning cycles without optimizing. If all prerequisites are performed sequentially and timely, the additional time and number of batches needed to optimize are negligible.

Optimizing the cleaning process will save years of water, energy, and chemical waste. It will reduce cleaning time and increase manufacturing time. It will also reduce regulatory inspection risk because the cleaning will be understood and robust.

It is never too late to optimize your cleaning process.

References

[1] FDA Guidance for Industry: Process Validation: General Principles and Practices January 2011, FDA,www.fda.gov/downloads/Drugs/…/Guidances/UCM070336.pdf.
[2 Song, Ruijin, Alfredo Canhoto, Ph.D., and Andrew Walsh, Cleaning Process Development: Cleanability Testing and “Hardest-To-Clean” Pharmaceutical Products Pharmaceutical Online, January 2019
[3] ASTM G121 Standard Practice for Preparation of Contaminated Test Coupons for the Evaluation of Cleaning Agents. https://www.astm.org/g0121-18.html
[4] Song, Ruijin and Andrew Walsh, Cleaning Process Development: Using Design of Experiment to Determine Critical Process Parameters Outsourced Pharma, June 2023.
For more information or support on optimizing a CIP process, feel free to contact me personally at:
vt*************@********on.org or call office:  858.673.3606