Selecting Cleaning Agents, Parameters for CGMP Processes, Part 2

Note: This is Part 2 of a two-part article on cleaning agents for cGMP processes.
Read Part 1

Other Factors Affecting Cleaning Performance

Surface
The chemistry and nature of the surface can determine the extent of adhesion of the residue to the surface as well as the extent of wetting and soil removal by the cleaning solution. The surface energy of clean metals and oxides is very high. This allows fluids to easily wet and spread on the surface and into the micro-irregularities. For example the contact angle for deionized water (surface energy 73 mJ/m² on a clean metal oxide surface (many hundreds of mJ/m²) is less than 5º while that on polypropylene (surface energy 33 mJ/m²) is greater than 100º. Thus water tends to ball up on a material like polypropylene that has no polar component. Surfaces with a high surface energy will, however, try to lower their surface energy by adsorbing low energy material. such as hydrocarbons. The reactivity of surfaces to soils will depend on the tendency of the surface atoms to act as electron acceptors or electron donors. The surface finish can also be an important factor. A rough surface increases the surface area of contact of the soil and the substrate and, like cracks and crevices, can be viewed as micro-deadlegs.

Soil Levels
High soil levels could saturate a solvent or deplete the surfactant or other components in a cleaning formulation, rendering one or several of the cleaning mechanisms ineffective. This is particularly possible when using small volumes of cleaning solution for large surface areas, as with the use of spray devices in a CIP system or when cleaning a series of vessels in a train with the same cleaning solution. One way to estimate the typical soil levels in a vessel is to measure the amount of actives or soil in the used cleaning solution in a pre-validation trial. This estimate could then be used to determine the worst case ratio of soil to cleaning solution to be used for a laboratory cleaning evaluation. An unacceptably high ratio could lead to problems such as inadequate cleaning, redeposition of soil, and in the case of solids, plugging of spray devices.

Soil Condition
The condition of the soil on the surface is an important factor in determining cleaning performance. In most cases, if the soil is cleaned while still "wet" or fresh, cleaning is easier than if the soil were allowed to dry. In this regard, the time before cleaning begins becomes important and should be documented. In some cases, the production time or campaign time may also have to be considered as there could be certain areas of the process system, such as pipes, that may have been exposed to raw material, intermediate or product only at the beginning of the manufacturing process. In certain cases the soil may be baked onto the surface during or after manufacture, making it more difficult to clean. In other cases, such as in tablet presses, the pressure used can be an important factor in determining the condition of the soil.

Mixing
It is desirable for the cleaning solution to be well-mixed to assure optimum use of the cleaning solution and prevents localized saturation spots. In a static soak, or in areas of minimal mixing and agitation where diffusion is the primary mechanism of removal of the soil away from the surface, the cleaning solution very close to the surface of the vessel could become saturated with the soil. There would be a decreasing concentration gradient of the soil in the cleaning solution as one moved away from the surface. The surface is then exposed constantly to a depleted or spent cleaning solution, which is clearly undesirable. Note that mixing is not necessarily the same as action or impingement discussed earlier. One could have high levels of action and yet inadequate mixing. Consider the case of a long pipe with a heavy cake of soil that is cleaned with a turbulent once--through flow of a dilute detergent. The flow rate and velocity, and consequently action or impingement, would be the same at the beginning of the pipe as it is at the end of the pipe as long as the pipe diameter is the same. The system is, however, not a well-mixed system because at the beginning of the pipe the soil is constantly being cleaned by fresh cleaning solution, while at the other end, it is exposed to a spent solution that is loaded with the soil that is transported through the pipe. In this case one would expect the end of the pipe to be cleaned last, perhaps even less, than the other end. Where solution homogeneity cannot be easily achieved (as in the case of the pipe example), surface swabbing locations and parameter selection (such as cleaning time or cleaning agent concentration) should account for such worst case locations.

Rinsing
The water rinsing cycle that follows aqueous wash cycles is an important factor that determines performance for critical cleaning applications. Rinsing is similar to washing, except that the residue in the case of rinsing is more easily removable and is solubilized in water readily. This allows for a compromise in one or more parameters such as a reduction in temperature or rinsing time. Rinsing should, however, be performed for a long enough period to obtain complete coverage of the entire surface and removal of all holdup cleaning solutions. If the cleaning agent is easily rinsable, any difficulty encountered in rinsing the cleaning agent out of the system is often indicative of either solution holdup or poor coverage. Therefore, such a situation is often an indication of a poor soil washing step as well. Rinsing should be done as quickly as possible after the washing process to prevent the soil from drying back onto the surface. The quality of the final rinse water should be at least as good as that used in the manufacturing process.

Selecting the Right Chemistry and Parameters
Cleaning agent selection for cGMP applications involves evaluation of several important factors. Besides effectiveness, these may include consistent quality, consistent and easy long-term availability, safety, environmental issues, analytical methods, and overall cost. Determining the most effective cleaning agent for different types of pharmaceutical product soils is best achieved through laboratory evaluation.

Laboratory Evaluation
A good understanding of the chemistry of the soil being cleaned and of the cleaning agent mechanisms will help in the screening process. Evaluating agents and determining parameters such as cleaning time, temperature and concentration could be done by a laboratory cleaning study. The soil to be cleaned is coated onto small panels or coupons and allowed to air dry or is baked onto the surface in an oven to simulate worst case conditions that might exist in the process. This step requires a good understanding and review of the manufacturing process. The coupons can then be cleaned by various cleaning agents using the desired cleaning process such as agitated immersion in laboratory glassware or under flow conditions as dictated by the process being simulated. The cleaned coupon can then be analyzed for residue by an appropriate method.

Laboratory evaluation can give useful information on a variety of issues, including: the effectiveness of cleaning, the potential for the soil to precipitate and redeposit on the surface, the potential foaming issues of the solution in the presence of the residue, the ratio of the soil level to solution volume that the cleaning solution can handle, and the physical state of the soil that has been removed from the surface. Laboratory evaluations can also provide a documented rationale for selecting the cleaning agent and parameters. The final confirmation of performance can, however, only come from a field-process cleaning trial. Since these laboratory evaluations can be time consuming, preliminary screening is best left to suppliers of cleaning agents who have experience in doing these studies.

As discussed earlier, time, action, concentration, and temperature are interrelated. One could, for example, increase the cleaning agent concentration and reduce the cleaning time to get the same level of cleaning performance. However, of the various parameters and factors discussed above, some are easily controllable, while others are not. An evaluation of the process could be done and a suitable cleaning program could be designed based on the considerations discussed below.

Constraints on Parameters
For some of the parameters there could be constraints on the maximum level achievable in the process. This may include a time constraint for completing the cleaning process, a temperature constraint, or a constraint on worst-case impingement or action when dealing with existing processes. It is important to consider the worst case conditions in the process system. For example, if a mixing vessel is cleaned by flooding it with the cleaning solution and using a simple agitator for mixing (agitated immersion process), the level of mixing obtained at the dome could represent the worst case. In certain cases, a constraint may exist on a combination of parameters. An example is substrate compatibility. Isocorrosion curves are available for glass lined vessels where the combination of alkali concentration and temperature determines whether the cleaning agent is substrate-compatible.

There would also be constraints on the minimum level required for parameters related to cleaning effectiveness. As an example, for certain types of soils, a static soak, representing the lowest level of action, would not clean a residue no matter what the other parameters may be. The minimum level of any parameter required for cleaning could be determined from laboratory cleaning evaluations or prior experience. Despite these upper and lower constraints, there is usually a wide range available for selecting the level of these parameters, and consequently multiple combinations of parameter values that would give the same cleaning performance.

Consistency of Parameters
Maintaining consistency in the parameters that affect the cleaning process is essential for obtaining consistent cleaning. Certain parameters can be more easily controlled and monitored than others, depending on the cleaning process used. As an example, let us compare the cleaning of a tank with a manual scrubbing process versus a static soak. Only the wash cycle with respect to the four parameters time, action, concentration and temperature are considered for simplicity. Table 1 shows the constraints imposed by the process on the various parameters.

It may be possible to clean the tank using either a manual scrubbing process or a static soak process. By appropriate selection of the parameters, both of these processes could give the same average level of cleaning performance. It is clear from the above table, however, that a manual scrubbing process would have to rely almost entirely on action/impingement while operating at reduced levels of the other parameters. Static immersion cleaning is not a very efficient process due to its very low action or impingement levels. However, the very low levels of action can be compensated for by higher levels of other parameters, such as concentration or cleaning time.

The key issue is that the action may not be easily reproducible for manual cleaning. The process relies almost entirely on a parameter that cannot be consistently monitored or maintained in the system. On the other hand, in the case of the static soak, the action is low but consistent. To compensate for the low action, we could use higher levels of other parameters such as concentration or cleaning time, both of which can be easily monitored and controlled. This process could therefore be more easily reproducible as it relies on parameters that are reproducible.

Can the manual process then be validated? Yes, but not by getting consistency in the performance measured by residue levels, but rather by using a procedure that is an "overkill." The resultant residue levels after cleaning would then not be at a consistent level from one cleaning to another, but could be consistently below the acceptance criterion.

Whatever the application method that is used for cleaning may be, it will involve a certain level of compromise on parameters. It is important to understand the parameters that are critical to the cleaning process and mechanisms, and to select those parameters at levels that are consistently achievable.

Cost of Parameters
Each of the parameters selected, such as time, action, concentration, and temperature, have associated costs. Some of these parameters may have a greater component of fixed cost associated with them, while others may be operating or variable costs. For example, investing in the fixed cost of providing better impingement could reduce operating costs such as cleaning agent concentration. Cleaning time would have both a direct and indirect (or opportunity) cost, and is generally the most expensive parameter. In plants that operate at high capacity utilization levels, the cost of time would typically dwarf the cost of all other parameters combined.

Conclusion
Selection of appropriate cleaning agents and parameters can be achieved by a combination of laboratory cleaning evaluations and a good detailed evaluation of the manufacturing process with respect to the various parameters and factors that influence cleaning performance. An appropriate selection process could simplify validation efforts and optimize the cleaning process.

For more information: George Verghese, Senior Applications Engineer, STERIS Corp., P.O. Box 147, St. Louis, MO 63166. Tel: 314-535-3390.