Aventis receives 13-item 483 for inadequate controls, GMP issues.

Aventis Behring, Bardley, IL, Chicago District/Team Biologics

FDA's Chicago District sent "Team Biologics" investigators David Perkins, Stephen Brown and Jason Chancey for a compliance follow-up inspection of Aventis Behring's Bradley, IL, bioanalytical laboratory. Following the 2002 audit, the firm was slapped with a 13-item 483 for inadequate controls for inspecting incoming glass vials, deficiencies in investigating failures, process deviations and unexplained discrepancies.

The firm was previously inspected in April 2001, which resulted in a 26-item 483 for systemic problems regarding failure investigations and validation studies, an issue with glass vial breakage, metal particles and deficiencies with lyophilization and cleaning validations, according to FDA records.

FDA has historically had problems with glass vial breakage at the facility when it was known as Centeon. The breakage of vials was one issue that landed Centeon into a consent decree with FDA over its plasma fractionation business.

And, the 2002 inspection revealed that Aventis' controls for the inspection of incoming glass vials were inadequate to detect defective vials used to fill albuminar and sterile water for injection.

According to the EIR, "the inspection showed that lots of albuminar filled using defective vials causing loose glass in solution were released based on AQL sampling even though 100% re-inspection on one lot of product which was rejected detected a significant increase in the number of defects (loose glass in solution)."

In addition, the inspectors reviewed the firm's technical complaints regarding receipt of broken 100mL albumin 25% vials, and revealed that there were numerous similar reports over the last 12 months. Investigation into the breakage led the firm to conclude that the removal of the IV set and redesign of the carton and shipper would correct the issue. Investigators noted that "drop-ship tests" were performed before the packaging changes were suggested or implemented. Additionally, a "testing report for the 100mL albumin vials does not note the inclusion of the IV set," the EIR stated.

Technical complaints also were received regarding the "milky appearance of the reconstituted solution" of monoclate (anti-hemophilic factor), and an adverse event of "pain and burning." Samples of the solution were tested by the firm, which initially determined that the flocculate was either not protein or was a highly denatured protein. "The firm did not attempt to further identify the flocculate until I inquired as to what the flocculate was actually composed of," Chancey wrote in the EIR.

In addition, the EIR noted that "lot W98801 had numerous deviations during its manufacture, including: utilization of glass vials that had passed their sterilization expiration date, lost nonviable particulate data, black particulate secondary to the sucrose raw material, vacuum failures, and the 'bubbling of the lyophilized cake."

The following are highlights from the 483 and corresponding responses from a July 2002 response letter from Gene Estes, senior vice president, World-wide Quality for Aventis Behring:

* There are no established criteria for re-inspection of glass vials used in the manufacture of certain albumin 25% lots. "There is no rationale for not performing 100% inspections." For example, the 483 stated that certain pallets "rejected upon re-inspection were all observed to contain defects identical to those that resulted in the destruction of albumin 25% lots X904707 and X904807, sterile water lot X65507, and the online destruction of glass vials dispositioned for use in albumin 25% lot X904907."

Response: "Aventis Behring understands the importance of the quality of glass to our manufacturing process and has undertaken several initiatives to proactively address the need to improve the quality of incoming glassware," Estes wrote. "Prior to the April 2001 inspection, Aventis Behring had identified deficiencies in the quality of the glass vials being supplied by [name purged]. We believe that this issue is best addressed by improving the quality of these components at the source rather than relying solely on the inspection process to cull out defects." To that end, Estes laid out the firm's plan, which includes: creating a glass quality task force, creating a supplier quality department, establishing a quality engineering department, assigning global glass quality initiatives and establishing a primary packaging quality steering committee.

* There is no documentation of sampling, sample inspection results, or corrective actions taken for previously rejected 50mL glass vials that were eventually used in the manufacture of albumin 25% lots. Additionally, there "are no assurances that the [unnamed] pallets that were used in manufacturing [certain lots] contain anything but glass vials previously rejected for containing critical defects."

Response: "The glass manufacturer evaluated the lot by re-sampling the glass on a pallet by pallet basis, taking 200 samples per pallet. These samples were inspected by the glass manufacturer with the intention that any pallets found to contain critical defects would be 100% re-inspected. In this instance, no further defects were found in the samples ... the glass manufacturer, therefore, provided a new pre-delivery sample (PDS) to Aventis," Estes stated in this letter.

He continued: "Aventis Behring inspected the new PDS samples and after finding no critical defects, accepted the lot. Based on the results of the initial PDS inspection, the vendor's additional sampling and the second PDS inspection, the estimated defect rate was [number purged]. It should be noted that the glass in question was processed via multiple inspections, including the Incoming Quality Assurance AQL inspection, the additional sample inspection as part of the deviation, an empty vial inspection, an online pre-capping inspection, and the final container inspection followed by a Quality Assurance post-inspection AQL."

* "Investigations of failures, process deviations and unexplained discrepancies are not always initiated, are not always complete, and/or final conclusions are not always supported by data." Five specific examples were given, including the loose glass in solution mentioned earlier.

Response: "In response to these observations, specific deviation handling procedures have been modified," Estes wrote. He laid out the firm's plan for better addressing deviation management processes, which include: deviation categorization, product impact assessment, corrective action/preventive action, data-handling enhancements and performance metrics. Estes also out-lined new SOPs that would better define requirements for lots that exceed inspection reject limits, as well as require the initiation of a deviation investigation whenever capping rejects exceeded the established limits. In addition, follow-up assessments would be required to support conclusions made in investigations, as well as better documentation mechanisms for addressing these issues.

* Investigations into technical complaints are not completed in a timely manner and may remain open beyond the expiration date of the product. "Documentation of 30-day and 120-day interim reports for open technical complaints as required in SOP-QA-424, Revision 2.0 is not consistent. As of 5/22/02, technical complaints originating in 2001 were open an average of 213 days with 32/122 complaints still open."

Response: "Aventis Behring commits that it will revise SOP-QA-24 entitled "Processing Technical Complaints on Licensed Biological and Pharmaceutical Products" to ensure that technical complaints are completed in a timely manner," Estes wrote in his letter to FDA. He outlined a corrective action plan that includes: implementation of metrics, completion of the current backlog and data analysis leading to system improvement. He added that Aventis Behring has closed all complaints originating in 2001, and planned to close all 2002 complaints by July 15, 2002.

* The number and type of glassware rejected during empty vial inspection prior to washing is not routinely documented.

Response: "Aventis Behring will revise its manufacturing procedures to require documentation of the total glassware rejects at the pre-wash glassware inspection step," Estes wrote. "In addition, a Quality Engineering Study has been designed to collect data on the total glassware reject levels and the type of rejects at the pre-wash glassware inspection step for each empty glass vial configuration."

FDA also cited Aventis' water for injection (WFI) system exceeding endotoxin alert or action limits because one sample failed.

In addition, FDA wrote: There's limited assurance that drug containers are always clean." The 483 lists six examples of unclean glass vials and equipment.

Also, Aventis was cited for stability samples not stored in a "manner that provides a 'worst-case' scenario in that they are stored upright. There are no assurances that the existing stability demonstrates that the current [purged] sample orientation provides complete information on product sterility, chemical absorption or otherwise supports expiration dating," FDA added.

The firm also was hit for not having "shipping validation studies for finished product."

Aventis did not return calls requesting comment by deadline.

Aventis Behring, Bradley, IL, 5/20-6/6/02, Doc. 109473M, $62.50 plus retrieval.