Analysis of 483s/EIRs for GMP validation issues.

Human Drugs

Forest tripped on stability, acceptance criteria data

Stability trending and acceptance criteria dominated FDA's audit of the analytical lab run by Forest Labs in Farmingdale, NY, which centerd on the firm's handling of bioequivalence studies for unnamed drugs.

Investigator Lawrence Farina and chemist Gary Lehr of the New York District Office inked in a six-page 483 (the EIR was not included in the FOIA release) after the October 1998 inspection that ended up on a March 12, 1999, warning letter (Doc. 7402W).

The firm followed up on the 483 with a letter from the firm's associate director of bioanalytical R&D, Dr. Timothy Getek, Ph.D., that rebutted or promised corrections of each item appearing in the EIR, evidently to little avail.

FDA and the firm found a sharp difference of opinion acceptance criteria in an analytical run the description of which was purged from the 483. Farina and Lehr wrote that the SOP in question allowed Forest to base its "calculation of accuracy of QCs using the validated QC value as the true value rather than using the theoretical value." In the 483, the FDAers insisted that the use of the theoretical value "would result in the majority of QCs not meeting the acceptance criteria in about 50% of the analytical runs."

In the Dec. 11, 1998, letter to the agency, Getek insisted that either value was sufficient to demonstrate reproducibility of the study runs. He commented further that "to use the theoretical value to conduct a stability trend analysis could lead to a false indication of the stability of the drug in the biomatrix." Getek emphasized that only known observed values were legitimate parameters for stability evaluation.

The Checklist Forest, New York District

Stability Acceptance criteria Getek added that the fact that use of the theoretical value would flunk 50% of the analytical runs "in itself lends evidence to the fact that the theoretical value is not the known observed value."

By the date of the warning letter, the issue had not been resolved. FDA pointed out that, "contrary to your response, the mean values of QCs estimated in the validation (i.e., validated values) only reflect the accuracy of the analytical method." The agency insisted that the theoretical concentration of QC samples "alone reflects the nominal value and should be used to set analytical run acceptance limits."

The 483 stated that "the acceptance criteria for validation on new QCs" was not specific in that the relevant SOP failed to specify "the number of new QCs that should be acceptable." The investigators wrote that in a December 1994 mini-validation run, two of six low QCs were greater than the acceptance criteria and that there was an 11% difference between the old and new QCs in that run.

Getek's response was that the SOP "should have read that [purged] of [purged] must pass at each level and that the criteria in the SOP had been modified to reflect the requirement. Getek also wrote that, "in regard to the [ISDN] QCs, the new QCs were approximately [purged] percent from the old QCs," which he stated was within the firm's acceptance criteria for measured versus theoretical values. This matter was not discussed in the warning letter.

FDA was not fond of the fact that Forest had been storing stock and mix solutions in identical conditions, a situation the investigators felt rendered inadequate the chromatographic comparison of the two for stability.

Getek made no promises regarding storage conditions in his reply, stating that "a comparison of the solutions at the beginning of the validation or study will be performed against the measurement of the solutions at the end of the study." He stated that if the firm established that should the solutions demonstrate instability, Forest would use reference material to produce fresh solutions "to determine the extent of degradation."

In the warning letter, the agency rebutted that the use of the same reference standard to evaluate stability was "not meaningful." FDA decried the fact that "the SOP has no provision for checking stability of stored stock solutions, although you use the same solutions for several studies." FDA pounded the point home in stating that "it is imperative that purity and stability of reference standards be adequately established and the actual purity be used to assure that the measured concentrations of the drug and its metabolites in human biological matrices are accurate."

At press time, the firm had not responded to requests for an update.

Forest Labs, Farmingdale, NY, 10/5-23/98, Doc. 108719M, $9.00 plus retrieval.

Impurities, validation net Lusochimica 483, warning FDA's audit of the Lomagna, Italy, plant of active pharmaceutical ingredient (API) maker Lusochimica was a pre-approval review of nine pending ANDAs, and although the firm managed to sidestep a warning letter, the investigation turned up a few validation problems.

Investigators Terri Dodds and Carol Gripp, apparently with the Center for Drugs, stated that the previous inspection, the date of which was not given, resulted in a rating of Official Action Indicated because Lusochimica failed to forward its proposed corrective actions within a time frame promised by the firm. The EIR noted that another Italian firm, the name of which FDA had purged, bought Lusochimica sometime around the time of the 1998 inspection. Also, a fire at the facility in February 1998 complicated matters further for the company.

Dodds and Gripp hit Lusochimica's prospective validation for an unnamed product because the firm had not completed validation of an unnamed part of the process. The EIR also noted that a new impurity cropped up during the first round of stability tests for one lot of the API, which Lusochimica evidently had released in July 1997 "to only one customer." The company managed to identify the impurity later on, but the agency purged the name of the aberrant substance from the EIR. According to Dodds and Gripp, Lusochimica had retrieved the entire lot by October that year.

The EIR noted that Lusochimica had not validated cleaning procedures for packing and sieving rooms. The FDAers noted a similar lapse for plastic containers used as utensils, hoses used in the transfer of solvents to reactor equipment and a flow meter.

The EIR indicated that Dodds and Gripp were less than thrilled with a retrospective validation for an unnamed substance. The FDAers wrote that one batch included in the review had been reprocessed after exceeding limits for a parameter that the agency redacted from the EIR and the firm had not included in the study. The inspection report stated that the batch in question was in quarantine and under investigation at the time of the inspection.

The Checklist Lusochimica, foreign audit

Impurities Retrospective validation Dodds and Gripp wrote that "several concerns were discussed with the firm regarding their retrospective validation to include the limited number of batches included, the exclusion of a batch that did not meet specifications, and no validation of the reprocessing step." However, according to the EIR, Lusochimica personnel had indicated that they were "committed to redoing the validation by conducting prospective validation, given the opportunity."

The EIR noted that the company had performed toxicology studies for the API containing the unidentified impurity. Lusochimica established that the LD50 (lethal dose for 50% of a population) for the drug substance was found not to be "significantly changed from the product not containing the impurity."

The agency also noted the firm's quality assurance manager, who was not named, had forwarded information regarding the new impurity to USP.

The EIR noted that officials with Lusochimica expressed their opinion that the problems with this API were related to the scale-up of production and not related to "continued development of the drug." However, the FDAers commented that it appeared to them that development was the source of the deviations.

The EIR noted further that "the firm understood that a full and complete revalidation of the process would be necessary with the proposed changes" and that Lusochimica "would make that commitment in their response."

Lusochimica, Lomagna, Italy, 10/19-21/98, Doc. 108861M, $10.00 plus retrieval.