SANDOZ CAFERGOT 10% ASSAY FAILURE RATE REFLECTS INVALID MANUFACTURING PROCESS

SANDOZ CAFERGOT 10% ASSAY FAILURE RATE REFLECTS INVALID MANUFACTURING PROCESS, FDA told the company in a June 17 warning letter regarding the March 30-May 7 inspection of the Sandoz facility in East Hanover, N.J. and the company's May 19 response to the FD-483 inspection report. "Any process that yields over 10% failures is viewed by FDA to be an inadequate or invalid process," the warning letter states. FDA noted that four of the approximately 40 batches of the migraine drug manufactured last year "failed assay release limits." The agency also found problematic the observation that Sandoz based its core weight for the labeled amount of Cafergot tablets "on an assay value rather than by a predetermined specification." The warning letter maintains that "the analytical variability and the batch variability may give Sandoz a false sense of security when relying on analysis to determine the target core weight." The warning letter to Sandoz is one of three issued recently by FDA's Mid-Atlantic Region office for manufacturing facility quality control and validation problems. A warning letter was issued to Ciba on June 16. Merck received a similar letter around the same time. Sandoz said that, as of July 1, it had not yet responded to the FDA warning letter, but believes it has appropriate responses to each of the issues raised. Ciba said it submitted an action plan on June 18 in response to the FD-483 observations the firm had received earlier. Ciba maintains that the action plan which proposes modification of certain procedures, a better documentation system and an improved employee training program -- should address FDA's concerns as outlined in the warning letter. Ciba recently discontinued two products because of production process problems. In May, the firm informed physicians that it would discontinue distribution of Lithobid 300 mg tablets, the only slow-release version of the antidepressant on the market, because of its inability to ensure product consistency under current FDA manufacturing standards. Ciba's quality testing lab found that the lithium tablets were dissolving faster than specifications allowed. In a June 14 letter, the company informed doctors that it is discontinuing the manufacture of the antihypertensive Apresoline (hydralazine) ampuls effective immediately "due to production process issues." Ciba will continue manufacture of Apresoline tablets. The recent compliance actions against the major pharmaceutical manufacturers and several smaller firms (see related item, T&G-11) demonstrate that FDA is exercising a greater scrutiny in the quality control testing area. The agency received legal support for a strengthened compliance program in the quality control testing area of Good Manufacturing Practices from the February Newark, N.J. federal court decision on Barr Labs ("The Pink Sheet" Feb. 8, T&G-1). The court backed FDA positions on retesting of products that have failed assays and validation procedures. The court agreed with FDA that retesting should not be used to invalidate an out-of- specification result unless an investigation has determined that analyst error may have been the cause. In the Sandoz warning letter, FDA states that the firm's standard operating procedure for stability samples is inadequate because of retesting following product failure. "Subsequent testing allowed by the Sandoz [SOP] following a failure of the stability testing is unacceptable and gives the impression that Sandoz intends to test its products into compliance," the warning letter states. FDA asserted that Sandoz "has no standard operating procedures for investigation, documenting and invalidating out-of- specification test results." The agency cited one lot of DynaCirc (isradipine) capsules that was resampled, tested and released. The lot, FDA noted, was assayed five times initially by one analyst. Then, different isradipine capsules were assayed nine times by a different analyst. "Clearly the results are statistically different by the two analysts and cannot be averaged as Sandoz pointed out in its response [to the FD-483 inspection report]." In conclusion, the agency stated that "there was no adequate investigation into the failure to see which set of results were accurate and what if any corrective measures needed to be effected to prevent reoccurrence."