This is the second of several Cleaning Memos dealing with selection of a worst-case product in a grouping approach for cleaning validation. Last month we considered use of the pharmacopeia solubility categories, and also presented what may be a more useful system based on how “solubility bands” are defined. Many companies may use evaluation systems involving calculation of a risk score that include a solubility parameter (based on the solubility of the active) in addition to other parameters such as toxicity (based on PDE/ADE values), potency (based on therapeutic dose of the active), and cleaning difficulty (based on operator experience). Such systems have been modified by different groups over the years, and have been widely used and accepted. That said, this Cleaning Memo will focus on the shortfalls (or problems) with those systems when used to select the worst-case drug product for pharmaceutical drug product manufacturing. It won’t be until subsequent months that we get around to exploring better options, so let’s get started with the critique of such systems as they are currently used.
Simplified Summary of Such Systems
The presentation below is just a generic one, and is not tied to any one system that I have evaluated. The systems include multiple parameters that may be related to the difficulty of cleaning. Here are examples of possible parameters, along with the rationale for why it contributes to the difficulty of cleaning.
A. Solubility of active: The lower the solubility, the more likely it will be left behind after the washing/rinse
process.
B. PDE/ADE: The lower toxicity number, the lower the residue limit will be for cleaning that particular drug
product.
C. Potency (Minimum Daily Therapeutic Dose): The lower the daily dose, the lower the residue limit will
be (if MTD is considered in a limit calculation)
D. Operator Difficulty of Cleaning: The more difficult in the judgment of operator for manual cleaning, the
more likely there is to be inadequate cleaning.
E. Concentration of the active in the formulation: The higher the concentration, the more likely higher
levels of residue will be left behind (assuming cleaning process is the same).
F. Solubility of excipients: The more insoluble the excipients, then the more difficult the product will be to
clean (assuming the same cleaning process).
The above list is not exhaustive. However, most commonly lists will include A, B and C, and to lesser extent D.
In such a system, each parameter will have a rating system (which could vary from two to six levels depending on the parameter) and with a weighting factor for that parameter. The weighting factor will depend on the impact of the parameter. For example, “A” may have a weight of 3 while E may have a weight of 1. Then for each product, the rating for each parameter is multiplied by the weighting factor, and the resulting weighted ratings are multiplied or added together to obtain an overall RPN. The product with the highest RPN is the worst case and is used for the validation protocol to represent the entire group of products. More typically in 2023 the residue limit for that worst-case product is set at the lowest limit of any product in the group. Note that there would also be a way to decide what is done if two or more products have an equally “highest” RPN, but it is not needed to cover those possibilities this month. Furthermore, care must be used in selecting both the rating values and the weighting factor for each parameter to make sure that the balance between the two is a good reflection of the overall risk. By this, I mean that if one parameter has ratings varying from 1 to 4, with a weighting factor of 2, and a second parameter has ratings varying from 1 to 2 with rating factor of 1, the impact of the two parameters differs not be a factor of two, but rather by a factor of four. That is, the highest weighted ratting possible with the first product is “8” (4 times 2), whereas for the second product the highest is only “2” (2 times 1).
Concerns with a RPN Approach as Typically Practiced
Here are possible concerns with such systems.
- While solubility of the active in water can be very important factor in liquid drug products (such as injectables), it is less important in oral solid dosage (OSD) drug products. The reason for this is that it is more likely that it is the solubility of the excipients in OSD products that make the drug product more difficult to clean. This should be obvious from the nature of immediate release (IR) products and extended release (ER) products. The ER products have excipients designed to slow down the release of actives when ingested by a patient. This difference in release of the active may be evident in the dissolution tests typically done as part of the QC for a manufactured product. It should be noted that the effect of excipients in slowing release may be different in fluids that simulate the stomach/gut as compared to water alone, but other things being equal, ER products are more difficult to clean than IR products (and both would be more difficult to clean as compared to orally dispersible products – at least for tablets).
- Toxicity (as defined by PDE/ADE) is relevant for calculating the residue limit for the cleaning of a product. However, is it really a factor in determining the inherent difficulty of cleaning for products in a cleaning validation group. Particularly in grouping approaches where the “most difficult to clean” product is evaluated in a protocol using the lowest limit of the active of all product in the group, that concern about the impact of a PDE value on a cleaning validation protocol is already addressed by this approach. In other words, is including toxicity as a factor for determining the highest RPN really appropriate if that toxicity factor will be addressed by selection of the lowest limit?
Here is another way to look at this situation for toxicity. Suppose we have two drug products. Product A has an active with a PDE/ADE of only 10 mcg/day, while Product B has an active with a PDE/ADE of 100. Other than this difference, all other parameters (such as solubility of active, concentration of active, and excipient formulation), are equal. Let’s say we determine the lowest residue limit of any product in the group (which may be the limit of a product other than “A” and “B”) and it is an L4b concentration of 2 ppm. So the question is, which product presents a more difficult cleaning situation? Is it Product A, which has to be cleaned to a limit of 2 ppm or is it Product B which has to be cleaned to the same limit of 2 ppm? I would conclude that they were equally difficult to clean. [Note that for the actual calculated limits specific for each of those two products, it is likely that the limit for Product A would be lower. But in a grouping approach, any product when cleaned by the specified cleaning procedure should also meet lowest limit of 2 ppm.
Here is a second example to illustrate this concern. Let’s assume we have two products again, Product C and Product D. Product C has a weighted toxicity rating of 4 but weighted solubility ratting of the active rating of 3, but Product D has a weighted toxicity rating of 3 and a weighted solubility rating of 4; all other parameters have the same weighted rating values. Let’s further assume that the lowest limit of any product in the group is 2 ppm. What is the conclusion as to which product I should choose as worst-case product to represent other products in a grouping approach? Are they equally difficult to clean? My answer is “No” provided I want to select the ”most difficult to clean” and evaluate that product at the lowest limit of any product in the group. So the question becomes “Is it more difficult to clean Product C to a limit of 2 ppm, or is it more difficult to clean product D to the same limit of 2 ppm?”. My answer is that the solubility of the active is the only difference, and with other things being equal, it would be more difficult to clean the active in Product D down to 2 ppm than it would be to clean Product C down to 2 ppm because of the relative difference in solubility of the active. Now that difference may be minor or may be major, but it does not affect the overall conclusion that the two products are not equally difficult to clean.
- Potency as used in many systems is defined as the daily minimum therapeutic dose (MTD) of the active. I will simplify things by pointing out that potency (as so defined) is similar to the toxicity as a parameter. It is related not so much to the inherent difficulty of cleaning, as it is to the calculation of limits (provided that limits are calculated using both PDE/ADE and 0.001 of the MTD values, and then selecting the lowest value to determine the residue limit of the active for that product. The effect of potency (as so defined) is really not the inherent difficulty of cleaning , but rather the lowest limit.
While the presentation this month discusses issues with this approach as currently practiced, I am not suggesting that this approach is an awful one that should be discarded. It is certainly an improvement over approaches used in the 1990s. However, over the next months we will cover options for further improvements.
Copyright © 2023 by Cleaning Validation Technologies
