Aventis audit triggered by complaints of false lab data.

FDA may have found the entrance to the Kansas City, MO, plant of Hoechst Marion Roussel and, eventually, Aventis Pharmaceuticals, rather familiar by the time of the March 2000 audit performed by chemist Michele Obert and investigator Alan Mehl, both with the Kansas City District Office. According to the EIR from this inspection, the audit Analysis of 483s/EIRs for GMP validation issues

By Mark McCarty, contributing editorwas triggered by a complaint the District received alleging falsification of lab data.

Obert and Mehl uncovered nothing to support the claim and wrote that they found the "anonymous allegations to be highly improbable." The duo reviewed Aventis' complaint-handling system as well as files in the firm's DQRS (drug quality reporting system), but saw no reason to crank out a 483.

Hoechst merged with Rhone-Poulenc Rorer to form Aventis, but the firms had not completed the task when investigator Steven Barber, presumably of the Kansas City district, combed through the plant in February and March 1999 for a pre-approval and GMP audit.

Barber inked in some observations regarding contamination on a two-page 483 with detailed text covering six deviations, but the plant's personnel evidently managed to bring matters under control before the agency felt compelled to issue a warning letter.

Obert and Mehl stated in their EIR that operations at Kansas City were unaffected by the merger and noted that Aventis had commenced with an expansion and upgrade of manufacturing facilities in Kansas City during their visit.

The Checklist Aventis, Kansas City District

Computers Contamination issues The EIR from the 2000 review stated that a former employee had filed the complaint regarding falsified data in July 1999. The complainant alleged that Hoechst had released five lots of a product FDA declined to name, despite the fact that the proportion of an ingredient was off by 20%. The complainant also accused Aventis of covering its tracks by changing computer files.

Barber's 1999 inspection, which ran between Feb. 25 and April 2, resulted in a citation for "serious non-conformities," including inadequate investigation into the appearance of glass in finished products, distribution of products prior to completion of process validation and failure to report adverse drug events within required time frames.

Barber also wrote up Hoechst for an unvalid-ated cleaning procedure for a dozen TranStore containers and added that the cleaning procedure in use did not "remove all the known contaminants in the containers."

According to the EIR, Hoechst personnel noticed pipe sealant in the containers in April 1998. Their attempts to remove the sealant were apparently unsuccessful. Barber indicated that the sealant was present in denatured alcohol used in the manufacture of 45 lots of unnamed product(s).

Barber encountered a technical summary of the incident, which stated that "these contaminants were generated during the fabrication of the totes and remained in the containers due to inadequate cleaning procedures and inspection prior to use." The EIR provided no further detail.

The 483 cited Hoechst for releasing 12 lots of Pentasa (mesalamine) 250mg capsules and two lots of Cardizem (diltiazem) after establishing that specially-denatured alcohol (SDA-3A) used in their production had been contaminated with various substances, including pipe sealant. Other contaminants included Teflon, unidentified metal and tetraethylene glycol. Barber dated the release as Nov. 10, 1998.

Barber wrote that Hoechst performed a risk analysis employing a worst-case scenario for the contamination and concluded that the PermaLok pipe sealant presented "little or no risk." Barber did not state whether the firm performed similar analyses for the other contaminants.

However, Barber wrote that he informed the firm that "even though they cannot detect the contamination in the finished product, the raw material used in the manufacture was contaminated and therefore the finished product is contaminated by definition." Barber did not specify which raw material served as the vehicle for the contaminants.

Aventis Pharmaceuticals/Hoechst Marion Roussel, Kansas City, MO, 3/3, 7-8/00 (Aventis), 2/25...4/1/99 and 8/4...9/3/98 (Hoechst), Doc. 108987M, $ 18.50 plus retrieval.

ChemSource's validation of cleaning for APIs cited

FDA's inaugural review of the Guayama, PR, plant of ChemSource disclosed problems with cleaning validation. The maker of intermediates and APIs hosted the San Juan District Office's investigator Daryl DeWoskin and took in a three-page 483.

According to the firm's response to the 483, ChemSource did not get a warning letter. The firm was working to get its products listed as sources in drug applications at the time of the inspection.

Some of the firm's products listed in the EIR were ranitidine HCL and fluvoxamine maleate.

According to the EIR, ChemSource, was owned by Quimica Sintetica SA of Spain, Chemical Capital Corporation of an undisclosed location, and MOVA Pharmaceuticals of Puerto Rico, and was incorporated "around 1996." DeWoskin wrote that the firm closed its purchase of the Guayama plant from SmithKline Beecham in 1997.

DeWoskin filled the first page on the 483 with a citation covering the firm's cleaning procedures. DeWoskin zeroed in on what he described as a lack of sufficient data demonstrating that cleaning procedures for non-dedicated equipment were adequate to prevent contamination of APIs. The first section noted that the firm lacked documentation regarding approved specifications for acceptable residue limits on multi-use equipment. The only detail Dewoskin went into was that the lapse included "scientific rationale, such as safety."

Chemsource gave the acceptable residual limit in its response, but the figure was purged. The firm insisted that the limit "was taken from the available literature," but that it would revise the threshold to a more appropriate limit.

DeWoskin complained the firm had no residue recovery study data that showed the effectiveness of target residue removal, including for APIs, from production equipment contact surfaces.

The firm replied that "experiments performed up to this point to establish the appropriate analytical methods will be revised and completed" and that the results of the revision would be approved and distributed.

The Checklist ChemSource, San Juan District

Cleaning validation Analytical methods According to the EIR, DeWoskin asserted that "all final rinses...must be large enough to touch all the product contact surfaces and not too large that they dilute the residues to quantities that cannot be quantified." He was concerned that ChemSource's tracking of rinse volume may be lacking when he noted it needed to know how much rinse it had used.

As a specific citation, DeWoskin observed that ChemSource personnel were not including total rinse quantity in their calculations of residuals. He insisted that the values for residuals were not meaningful in the absence of rinse volume information.

The company responded that "the amount of rinse solvent used on each piece of equipment will be taken into consideration in each analytical method calculation when applicable."

ChemSource, Guayama, Puerto Rico, 4/19-5/4/99, Doc. 108988M, $8.00 plus retrieval.