Last month we discussed the strict applicability (or rather lack of strict applicability) of Q3A and Q3B to residues found in cleaning validation (CV) samples, specifically to “unknown” peaks in chromatograms. That said, it may be possible to utilize the principles found in those documents (using Q3B) to unknown peaks in CV chromatograms focused on residues of a prior API in drug product manufacture.
Before we go through possible calculations to demonstrate the applicability, it is important to point out some difference between the impurity limit in Q3B and CV limit calculations. Here are things to consider.
The purpose of CV calculations is to establish limits, and then the measured amounts in the protocol execution are compared to those limits. In Q3B the limits are not calculated (except to the extent that they depend on the maximum daily dose of the API); rather they are given in Q3B.
In CV calculations for drug products, the limit is most typically calculatedbased on the possible concentration (such as mcg/g or ppm) in the next drug product expressed as the concentration. The eventual result is most typically expressed as a concentration (mcg/g or ppm) in a rinse sample or an extracted swab sample. In Q3B the limits are given for impurities based on the concentration present in the drug substance (API) of the drug product. The measured concentration of the impurity in the drug substance is compared to the Q3B limit.
For purposes of this Memo, I am not focusing on acceptability of the prior API (drug substance), but on the acceptability of unknown peaks when that prior API is being measured.
Q3B gives thresholds for limits (reporting, identification and qualification) based on the maximum daily dose of the drug substance that the impurity might be in. For this Memo, focused on unknown peaks, I would focus on the identification threshold. The reason for that is that in CV studies, those peaks are not identified, so one purpose of this Memo is to help establish why those peaks do not require identification.
A key issue for both unknown peaks in CV studies and impurities in Q3B is the analytical method to measure the chemical amount present in either situation. How is it possible to measure the amount unless the chemical substance is known so that the relationship between the amount present and the chromatogram AUC (Area Under Curve) can be established? In past approaches to unknown peaks in CV protocols (see Cleaning Memo of February 2005 – 21 years ago!), it is assumed that the molar absorption coefficient of the unknown is the essentially the same as that of the target residue. In Q3B, there is a section on analytical methods which states that the analytical method should be validated for the “specified and unspecified” degradation products. That section also allows for use of an “appropriate reference standard or to the response of the new drug substance itself”. [Note 1: This issue could possibly be addressed by UV/PDA detection for HPLC, but I will leave that to the analytical chemists to determine if the analytical method for the drug substance can be used for quantitating the impurity based on the assumption that the unknown peak is a degradation product of the target protocol residue. Note 2: If the unknown peak is not such a degradation product, then other approaches to determine its source (to help identify it so it is no longer unknown) will be covered next month in the April Memo.]
So with these differences and issues, is there a way to establish whether unknown peaks in CV protocols would fall below the thresholds listed in Q3B?
Here is a summary of one approach for swab sampling:
First of all, I will use “expressions” similar to the “M” (for “Measured”) which were introduced in the Cleaning Memo of March 2025. The expression I will use for unknown peaks is “MI” (for “Measured Impurity”). The relevant terms for this discussion are MI-4b (typically in units like mcg/g or ppm) for the measured concentration in the extracted swab sample, and MI-1 for the measured concentration (typically in unit such as mcg/g or ppm) in the next drug product.
In the CV protocol, measure the concentration of the unknown species by assuming the molar absorption coefficient is the same as the target residue. This gives a MI-4b concentration for the unknown peak species equivalent to the M4b concentration as is done for the target residue.
Next convert this MI-4b (a concentration in the analytical sample) to a MI-1 (a concentration in the next drug product). For swab sampling, this is done by multiplying the MI-4b by the total surface area (TSA) of the equipment and then dividing by the minimum batch size (MBS) of the next drug product, and then multiplying that result by the swab extraction amount (SEA) and then dividing by the swabbed surface area (SSA). The net result is the MI-1 value expressed as a concentration (such as mcg/g) of the unknown peak species in the next drug product.
Next is to convert this MI-1 to a concentration in the drug substance (that is, the API) in that next drug product. That is done by dividing the MI-1 result from Step 3 by the fraction (F) of the drug substance expressed as a decimal in the next drug product. The result is a value of the concentration of the unknown peak species in the drug substance, which I will call UPC (for “unknown peak concentration”). This UPC is the concentration that should be compared to the percentage thresholds provided in Attachment 1 of Q3B. Note that if the Q3B thresholds are given in “percent” values, those percent values should be converted from percentages to mcg/g (remembering that 1% is 1000 mcg/g and 0.1% is 100 mcg/g).
If a TDI (Total Daily Intake) of the unknown species to be used for the comparison, this TDI is calculated by multiplying the UPC calculated in the previous step (and not the MI-1 calculated in Step 3) by the maximum daily dose (in grams) of the drug substance of the next product (and not the minimum dose as may be done for calculating limits for the CV protocol).
The threshold used for comparison could be either the reporting threshold or the identification threshold, either of which will depend on the maximum daily dose of the drug substance. For purposes of this evaluation, that drug substance considered is not the drug substance of the cleaned product, but rather the drug substance of the next product in the CV limit calculation procedure.
Here is an example of such a calculation where swab sampling is used. It follows the “steps” 1-5 above (read them carefully!), with the following input data:
MI-4b = 0.2 mcg/g
TSA = 1,000,000 square centimeters
MBS = 500,000 grams
SEA = 10 g
SSA = 100 square centimeters
F = 0.10
Note that the values for these inputs are not something new. They are the same values (except for F) used in a typical limit calculation for the CV protocol.
Based on those inputs, here is the calculation for UPC (in units of mcg/g):
Inputting the data above into Equation A gives the following result:
This value of the concentration of the unknown peak species in the drug substance of the next product is to be compared to thresholds in Attachment 1 of Q3B. In this case, the calculated UPC would be below any reporting threshold, any identification threshold, and any qualification threshold. If a comparison were to be made for the TDI, assuming the maximum daily dose of the drug substance in the next product were no more than 2 grams, the resultant TDI would be 0.8 mcg (calculated as 0.4 mcg/g X 2 grams), which would be below all the TDI thresholds. These conclusions in this paragraph apply only to this example with the specified inputs. It is possible that with extreme examples, either the percentage or TDI thresholds could be exceeded.
Caveat
I fully realize there are assumptions in this type of evaluation (like the molar absorbance coefficient of the unknown peak as compared to the target residue in the CV protocol), but it should be clear that any result below the qualification threshold should be acceptable. However, as stated in Attachment 1 of Q3B, “lower thresholds can be appropriate if the degradation product is unusually toxic.” A good example of this is nitrosamines in ARB medications (see Memo of March 2019 for more on this issue).
This type of calculation is offered just to present an alternative to approaches that have been used in the past to address unknown peaks. Those past approaches will be discussed again in the April Cleaning Memo.
