Establishing Scientifically Justified Acceptance Acceptan ce Criteria for the
Cleaning Validation of APIs Destin A. LeBlanc
Establishing Scientifically Justified Acceptance Criteria for the
Cleaning Validation of APIs Destin A. LeBlanc
M S I R E T S
Setting limits based on sound scientific principles is critical for cleaning validation protocols. Residues in the manufacture of active pharmaceutical ingredients (APIs) must be considered directly because of possible effects on any subsequently manufactured API and, ultimately, regarding how such APIs are used in finished drug products. The author presents equations for calculating various aspects of residue limits for APIs.
ost published cleaning validation protocols have focused on the residue acceptance criteria for finished drug products. The FDA guidance document for cleaning validation simply states that the residue limits must be logical, practical, achievable, and verifiable (1). Most presentations of residue limits focus on finished drug products and use a variation of the methods proposed by Fourman and Mullen (2–4). These presentations are based on the possible contamination of a product that is subsequently manufactured in the same equipment. Unfortunately, no clear guidance exists for residue limits in cleaning validation for the manufacture of an active pharmaceutical ingredient (API), except for the directive that the limit should be scientifically sound (5). Many pharmaceutical companies have established formulas for limits of API residues based on the dosing of subsequently manufactured APIs. The presentation of residue limits in this article reflects that strategy, but it goes further into the development of such calculations. Such an exploration of that development is not necessary to the final calculation but can serve to elucidate the contribution of factors that appear, either directly or indirectly, in any final calculation. The immediate concern about setting residue limits for cleaning APIs is that the residue can be transferred to an API subsequently manufactured in the same equipment. The ultimate issue is that the level of any residue present in the subsequently manufactured API must be evaluated in light of the effects of the residue on a finished drug product in which it is incorporated. Without taking into account how the subsequently manufactured API is used in a finished drug product (as well as factors such as dosage, routes of administration, and patient population), establishing scientifically justified residue acceptance criteria for the cleaning validation of APIs will be difficult.
Effects of API cleaning process residues on finished drugs is vice-president of technical support at Steris Corporation, 7405 Page Avenue, St. Louis, MO 63133, tel. 314.290.4790, fax 314.290.4650, e-mail destin_leblanc@ steris.com. Destin A. LeBlanc
Although one may approach residue data calculations in several ways, the principle is the same — the effects of any residue in API manufacturing must be evaluated in light of the effect(s) of that residue once it is present in a finished drug product and used by the patient. Presenting an overall calculation of residue limits is possible. However, this article separates the various parts of data
Calculations of residue limits in API manufacturing
handling so that the contributions of each component BL1 can be understood the limit of the target residue in any finished drug more clearly. Four product in which the residue ultimately may be calculations can be found made to express BL2 limits at various the limit of the residue in the API used in the points in the overall manufacture of those finished drug products analysis. This approach may be imBL3 portant because the limit of that residue per surface area of the using worst-case cleaned equipment used to manufacture the API limits can distinBL4 guish between, for the limit of the residue in the sample actually example, limits in analyzed by an appropriate analytical technique. the final API and limits per surface area of equipment. Key to terms The first task is to determine the API limit of the target active pharmaceutical ingredient residue in any finAPIA ished drug product the active pharmaceutical ingredient that initially in which the is manufactured in a specific piece of equipment residue ultimately may be found APIB (BL1). The second a second API that is manufactured in the same step is to calculate equipment (following cleaning) and then the limit of the converted into one or more finished drug products residue in the API ProdB,ProdB1, ProdB2 used in the manufinished drug products made with API B. facture of those finished drug products (BL2). The third is to calculate the limit of that residue per surface area of the cleaned equipment used to manufacture the API (BL3). The fourth step is to determine the limit of the residue in the sample actually analyzed by an appropriate analytical technique (BL4) (see sidebar “Calculations of residue limits in API manufacturing”). This paper will consider each calculation separately and then combined. The author assumes that cleaning is done following the manufacture of one API (APIA) (see sidebar “Key to terms”). At least one target residue, most likely the API itself, will be identified. In the same equipment (following cleaning), a second API (APIB) will be made. This second API will then be converted into one or more finished drug products (ProdB).
BL1 limit in final dosage form After identifying the target residue for the cleaned equipment and the final dosage form(s) of concern for the subsequently manufactured API, one can calculate the limit of that residue in the final dosage form. Assuming that the target residue is the bulk active APIA and that the subsequently manufactured bulk active (APIB) is formulated into a finished product (ProdB), the equation introduced in Reference 3 for finished drugs should be used, i.e.,
BL1
minimum daily dose of APIA 1,000,000 SF [1] maximum daily dose of ProdB
In this case, BL1 is in parts per million (ppm), and the two dosage numbers are the same units, such as mg or g. The minimum daily dose of API A is the minimum daily dose of APIA when administered in a finished drug product made with API A. If APIA is used in several dosage forms, the minimum among the various dosage forms should be used for this calculation. The maximum daily dose of Prod B is the maximum dosage of the drug product containing APIB and is independent of what constitutes that APIB. If APIB is used in more than one finished drug product, the maximum dosage of the finished product with the highest maximum dosage should be used. The 1,000,000 figure represents a conversion to ppm. SF is the safety factor, a value based on risk analysis, typically 0.001. The calculated BL1 using Equation 1 represents the maximum allowable concentration of the target residue in the finished drug product made from the subsequently manufactured API. For example, after the manufacture of APIA, the equipment is cleaned, and APIB is manufactured. APIB is formulated in finished drug ProdB1. If APIA is dosed at 20 mg of active from two to four times per day and if ProdB1 is dosed at 2000 mg of finished drug product one to three times daily, then Equation 1 becomes
BL1
20mg 2 1,000,000 0.001 6.67 ppm 2000mg 3
[1a]
If APIB also is used for manufacturing a second finished drug product (ProdB2) and this second product is dosed at 1500 mg two to three times daily, then the BL1 limit for this alternative situation would be
BL1
20mg 2 1,000,000 0.001 8.89 ppm 1500mg 3
[1b]
Although one may be tempted to select the lowest limit here, both results (the BL1 limits for both ProdB1 and ProdB2) should be carried forward to the next calculation. This BL1 calculation only gives the maximum allowable level of target residue (in this case APIA) in the finished product, ProdB.
BL2 limit in API The next step is to calculate the maximum allowable residue in the bulk active itself (that is, in APIB).This calculation is a simple one because it simply determines the target residue (APIA) as a portion of bulk active APIB — not of the finished drug (ProdB). This is accomplished by dividing the BL1 limit
by the fraction of active (APIB) in the finished drug (ProdB). This is expressed as BL2
BL1 100 %APIB in Prod
[2] BL2
Because the BL1 limit and the %APIB in ProdB may be different for ProdB1 and ProdB2, the BL2 limit should be calculated for each combination. Continuing with the same two examples used for calculations of BL1: If ProdB1 contains 2% of APIB, then the residue limit of APIA in APIB is calculated as
BL2
6.67 ppm 100 334ppm 2
[2a]
If ProdB2 contains 4% of APIB, then the residue limit of APIA in APIB is calculated as
BL2
8.89 ppm 100 222ppm 4
[2b]
In these examples, BL1 was lower when APIB was formulated in ProdB1, and BL2 was lower for the case involving Prod B2. If APIB can be used in either finished drug product, the lower BL2 value (in this case 222 ppm) should be used for further calculations.
Alternative BL2 calculation Although the equations presented for BL1 and BL2 help separate the factors that contribute to the BL2 calculation, a simplified equation also can be used. Because (daily dose ProdB) (%APIB in ProdB) =daily dose of APIB
dose for APIB is therefore 180 mg. Using this value for the maximum daily dose of APIB in Equation 3 and the minimum daily dose of 40 mg for APIA and a SF of 0.001, BL2 is calculated as 40mg 1,000,000 0.001 222ppm 180mg
Either Equation 2 or Equation 3 can be used to calculate BL2. Equivalent information is required for either calculation. For finished products with very low percentages of APIs, the calculated BL2 limits may be very high, even as high as 1000 ppm or greater. In such cases, establishing a default limit of 1000 ppm may be prudent so that residue limits are below the usual 0.1% threshold that requires analysis and characterization of impurities in APIs dosed at 2 g/day (6). Caution should be taken in adopting such a default limit because of the misuse of the 10-ppm default limit that is often used for finished drug products (2). This default value may be used only if the actual calculated value for BL2 is 1000 ppm. If the value is 1000 ppm, then that actual value should be used for subsequent calculations. This bears repeating so it is not misused — the important issue is that the default value cannot be adopted automatically; the default value is used only if the calculated value (the scientifically justified value) is 1000 ppm.
BL3 limit per equipment surface area The next step is to calculate the actual residue limit per surface area of the cleaned equipment before the manufacture of API B. This step is basically no different from the surface-area calculations performed for finished drugs. This calculation takes into account not only the limit of the residue in the subsequently manufactured product but also the batch size and the equipment shared surface area. BL3 in g/cm2 is calculated as
BL3 [2c]
[3a]
BL2 (batch size of APIB) 1000 (shared surface area)
[4]
where the batch size is expressed in kg and the shared surface This simplified approach combines Equation 1 into Equation area is expressed cm2. The figure 1000 in this equation converts 2 to express BL2 as the batch size from kilograms to grams. The batch size is the total amount of API only and does not include functional aids such as solvents that may be used to facilitate manufacture. In other words, although 1000 L of solvent may be used to manu(minimum daily dose of APIA) 1,000,000 SF [3] facture a batch containing only 20 kg of APIs and is important BL2 maximum daily dose of APIB for determining product shared surface area, for calculating limits, only the amount of API itself is used for batch-size purposes. Continuing with the previous example, if the batch size is In this case, one must still select the maximum daily dose of 20 kg and if the shared surface area is 30,000 cm 2, then BL3 is APIB (from among its possible dosing regimens). The result is calculated as the same as Equation 2. For example, using the information from the examples pre222ppm 20kg 1000 2 [4a] viously given, APIB will be dosed at a maximum rate of 120 mg 148 g/cm BL3 2 30,000 cm daily (6000 mg 2%) of APIB in ProdB1, and it will be dosed at 180 mg daily (4500 mg 4%) in ProdB2. The maximum daily
In this case, the resulting BL3 limit of 148 g/cm2 is likely to be a level of surface contamination that could be observed readily in a visual examination (a typical visual limit is approximately 1–4 g/cm2) (2,7). Using this scientific approach, most cases involving residue limits for APIs could involve calculation of BL3 limits considerably above any visual observation of no detectable residues.
sample. These calculations may be combined into one equation. However, this should be done cautiously so that the principles underlying the calculation are fully understood. In both of the following presentations, BL2 is expressed in the simplified form given in Equation 3. An equation for BL3, the surface area limit in g/cm2, would be: 9
10 (batch size ofAPIB)
BL4 limit in the analyzed sample This step calculates the residue limit in the analyzed sample. Calculating the limit in the analyzed sample follows exactly the calculation for finished drug products. For illustration purposes, the following analysis assumes that surfaces are sampled by swabbing, desorbing the swab into a specified volume of an appropriate solvent, and then using an appropriate analytical technique to measure the target residue (APIA) in that solvent sample. The limit in the analyzed sample in g/mL is
(maximum daily dose ofAPIB)
(minimum daily dose of APIA)(SF ) (shared surface area)
[6]
An equation for BL4, the limit in the analytical sample in g/mL, would be 9
10 (batch size ofAPIB) (minimum daily dose of APIA) BL4
BL3 (swabbed surface area) volume of solvent
[5]
Some scientists also will include the swab recovery factor (percent recovery) in this equation. That recovery factor can be included in Equation 5. However, it is preferable to include it as part of the analytical data transformation so that residue limits do not change as analytical and sampling methods change. Regardless of how recovery factors are handled, one must be consistent to avoid a situation in which the recovery factor is included in both the BL4 limit calculation and the analytical data calculation. If the recovery factor was included in both calculations, product safety would not be compromised. However, meeting the resultant calculated BL4 residue limit would be more difficult. Continuing with the same example, if the surface area swabbed is 25 cm2 and the swab is desorbed into 5 mL of solvent, then the limit in the desorbed solvent sample (in g/mL) is
(maximum daily dose ofAPIB)(shared surface area) (swabbed surface area)(SF ) (volume solvent)
[7]
Care must be used in expressing units, with both doses in the same units (preferably mg), the batch size expressed in kg, surfaces areas expressed in cm2, and the desorbed solvent amount in mL for the 109 factor to be correct. For this presentation of data, SF is expressed as the decimal equivalent (that is, 0.001 rather than 1000).
Additional issues in API limits
Multiple finished dosage forms for API A. If APIA is used only in one dosage form, then these calculations may be relatively straightforward. They become more complex as that API is used in multiple dosage forms, which can then be dosed at various rates. As discussed in the section on BL2, the BL2 limits for each combination should be calculated and the lowest result should be used in the BL3 calculation. Different routes of administration for API A and APIB. Different 2 2 148 g/cm 25cm routes of administration would occur if, for example, APIA was [5a] 740 g/mL BL4 used only for oral dosage finished drugs and APIB was used only 5 mL for parenteral finished drug products. In such a case, one would need information about the parenteral application of APIA to In this case, the BL4 analytical sample limit is considerably calculate the appropriate limits. For example, the calculation higher than the limit in the bulk active itself (BL2) because of based on an application of SF to a no-effect intravenous dose the leveraging power of the sampling method (3). This effect is in animals may be suitable for calculating its limit in a parenteral similar to the situation for finished drug products. A similar finished drug product. Multiple subsequent API products. When calculating an actype of calculation can be used for collecting samples with a rinse sampling procedure (8). ceptable BL3 limit per surface area of cleaned equipment following the manufacture of one active (APIA), the case may arise Presentation as one equation where, depending on the manufacturing schedule, the cleaned These steps flow logically from determining the concentration equipment will be used to manufacture APIs other than APIB, limit of a residue that is allowable in a finished product, to cal- (e.g., APIC, APID, or APIE). This involves calculating BL1, BL2, culating the residue level allowed in the API itself, to calculat- and BL3 for each combination (i.e., APIA followed by APIB, APIA ing the allowable residue limit per surface area, and finally to followed by APIC, APIA followed by APID, and APIA followed by calculating the allowable residue limit in a specific analytical APIE). In such cases, the lowest BL3 limit is selected for valida-
tion purposes to ensure that any of the other APIs may follow APIA in the manufacturing order. If this is the situation, one must also address limits for API B in light of any of the other APIs following in the manufacturing order. An alternative is to consider restricting the manufacturing order, an option that may or may not be acceptable depending on manufacturing and economic factors. Cleaning before a final purification. Any limit established for an intermediate cleaning step of an API should be established according to how those residue levels contribute to the cleaning residue levels in the final API. If one can demonstrate that those residues (intermediates or cleaning agents) are effectively removed from the API during a final purification step (e.g., recrystallization) so that the residue level in the final API is consistent regardless of how much residue is present at any intermediate step and if that consistent level is below the BL2 limit as calculated above, then one can scientifically argue that those intermediate cleaning steps are not critical manufacturing steps. If they are not critical steps, then they should not require cleaning validation (9). By conducting lab studies in which the preparation immediately before final purification is deliberately spiked with a target residue, one can demonstrate that residue levels in the final API are independent of levels present at intermediate steps. However, one also should demonstrate this with pilot- or fullscale manufacture. If this approach is taken, then it should be addressed in the cleaning validation master plan and any specific decision about noncriticality should be documented clearly. In such case, the final process purification step of the API is clearly a critical process step and therefore requires process validation. Furthermore, despite the fact that cleaning may not be validated in such intermediate steps, standard operating procedures for cleaning should be written and followed. Such an overall strategy also does not relieve one of the obligation to validate the final cleaning at the time of product changeover.
Summary Residue limits for APIs for cleaning validation purposes can be calculated using principles similar to those used for residue limits for finished drug products. The effect of any residue left in cleaned equipment must be considered in regard to the transfer of that residue to the subsequently manufactured API and ultimately to the effects of that residue in any finished drug products made from that API. The fact that APIs are one step removed from the finished dosage form usually makes scientifically justified residue limits relatively high (as compared with limits for finished drug products). However, it is important to make the actual calculations to demonstrate compliance in the manufacture of APIs where cleaning validation is required.
References 1. FDA, “Guide to Inspections of Validation of Cleaning Processes,” Division of Investigations, Office of Regional Operations, Office of Regulatory Affairs, July 1993. 2. G.L. Fourman and M.V. Mullen, “Determining Cleaning Validation Acceptance Limits for Pharmaceutical Manufacturing Operations,” Pharm. Technol. 17 (4), 54–60 (1993). 3. D.A. LeBlanc, “Establishing Scientifically Justified Acceptance Criteria for Cleaning Validation of Finished Drug Products,” Pharm. Technol. 22 (10), 136–148 (1998). 4. D.A. LeBlanc, “Setting Acceptance Criteria,” in Validated Cleaning Technologies for Pharmaceutical Manufacturing (Interpharm Press, Englewood, CO, 2000), pp. 135–150 . 5. PhRMA Quality Committee, Bulk Pharmaceuticals Work Group, “PhRMA Guidelines for the Validation of Cleaning Procedures for Bulk Pharmaceutical Chemicals,” Pharm. Technol. 21(9), 56–73 (1997). 6. ICH Q3A(R), Draft Consensus Guideline, “Impurities in New Drug Substances,” 7 October 1999. 7. K.M. Jenkins and A.J. Vanderweilen, “Cleaning Validation: An Overall Perspective,” Pharm. Technol. 18 (4), 60–73 (1994). 8. D.A. LeBlanc, “Rinse Sampling for Cleaning Validation Studies,” Pharm. Technol. 22 (5), 66–74 (1998). 9. ICH Q7A, Draft Consensus Guideline, “Good Manufacturing Practice Guide for Active Pharmaceutical Ingredients,” 19 July 2000. PT
© Reprinted from PHARMACEUTICAL TECHNOLOGY, October 2000
Printed in U.S.A.
STERIS® Corporation 5960 Heisley Road Mentor, OH 44060-1834 USA 440-354-2600 ■
■
■
■
■
410-600-0012
