QUALITY CONTROL LAB DATA MUST BE RETAINED BY PHARMACEUTICAL FIRMS

QUALITY CONTROL LAB DATA MUST BE RETAINED BY PHARMACEUTICAL FIRMS even if the lab results are determined by the company to be invalid, FDA Division of Manufacturing and Product Quality Director Paul Vogel told a Parenteral Drug Association forum on retesting and laboratory investigations on Sept. 10. Vogel cautioned that a QC lab "should not discount or ignore suspect data without scientific justification that the results are not representative of the material." However, Vogel added that even when lab results are shown to be invalid, "the initial results . . . must be kept" along with the company's justification for invalidating those results. Vogel emphasized that good manufacturing practices regulations require "laboratory records to include complete data derived from all tests. The evaluation of out-of-specification laboratory data has been a focal point for industry and FDA since the U.S. v. Barr Labs ruling last February in Newark federal court. Judge Alfred Wolin's 80-page ruling, which rested heavily on the court's interpretation of cGMP requirements for lab testing and failure investigations, has provided support for a strengthened FDA compliance program in the QC lab practice area. FDA field office management prepared a summary of the court ruling and issued it to the agency's district offices this spring. Relevant aspects of the decision were also incorporated into a revision of FDA's "Guide to Inspections of Pharmaceutical Quality Control Laboratories," which was released in August. Reflecting the Barr ruling, the new draft outlines FDA's expectations for conducting failure investigations and the use of retesting, averaging and blend analysis. Vogel pointed to lab data evaluation issues as critical to FDA's assessment of a firm's GMP compliance status. He emphasized that "the inadequate response to out-of-specification results is a common theme of many of the major enforcement actions taken by FDA over the past two-to-three years." Warner-Lambert, for instance, is among firms that have faced serious enforcement action following FDA findings in this area; the firm's handling of stability data was a key problem cited during the inspections that led to a court-adjudicated consent decree in August ("Tbe Tan Sheet" Aug. 23, p. 16). Industry participants at the meeting expressed concern that FDA may be placing overly tight strictures on the use of retesting and resampling, "outlier" tests, and averaging. Wyeth-Ayerst QC Associate Director Kenneth Dilloway commented that retesting and resampling may be used by firms to conduct a more thorough and accurate follow-up investigation as opposed to overriding or invalidating original test results. For example, he noted, resampling may help identify problems with the original sample, container, or sampling procedure that otherwise would not be detected. Responding to the industry comments, Vogel said that FDA's main concern is that a company is doing a "quality investigation" and assured the group that the agency "will look at what you do in context." He acknowledged that "you are never going to have absolute certainty" in investigating lab data. However, he cautioned, FDA will expect firms to have a "reasonable basis" for their conclusions. That rationale "is what investigators are going to look at," he said. Vogel identified resampling and the use of averaging when assaying for potency as issues that will got further attention at his office. The agency is also interested in "learning more" from proponents on the use of outlier tests, Vogel said, "especially when applied to chemical tests designed to measure uniformity." FDA panelists at the PDA workshop generally emphasized that firms should have written procedures that detail how evaluation techniques are going to be applied, and documentation showing that the procedures were followed in a consistent fashion. Dilloway suggested that the pharmaceutical industry put together a proposal for the U.S. Pharmacopeia on the best methods for handling out-of-spec results. Although USP has resisted addressing batch release criteria in the compendia, Dilloway maintained that inclusion of additional information on evaluating chemical test results would help resolve regulatory and legal disputes. The information could be included as a modification of the chapters on dissolution and content uniformity, he suggested. FDA Division of Manufacturing and Product Quality staffer David Barr reported that his office has begun working on a guideline addressing the handling of out-of-spec laboratory results. He predicted the guideline would not be available before spring 1994. PMA also has a technical committee focusing on statistical issues raised by the Barr decision. A report from that committee is expected to be released in six to nine months.