Novartis Vaccines and Diagnostics.

Cambridge, MA, Jan. 24 (CBER)

FDA conducted an inspection of Novartis Vaccines and Diagnostics in Marburg, Germany, Sept. 20-27. FDA investigators documented a number of significant deviations from GMP in the manufacture of the firm's Rabies Vaccine (RabAvert) and Diphtheria and Tetanus Toxoids Adsorbed concentrate (without preservative). These include deviations from the applicable requirements of the FD&C Act as well as requirements of the BLA.

The sterility of one media lot used in the RabAvert vaccine upstream process could not be assured, yet final product lots that used that media lot were further processed and submitted to FDA for lot release. Two contamination events were linked to this media lot:

* A RabAvert lot was observed contaminated on Oct. 16, 2006, with Candida guilliermondii.

* A RabAvert lot passed sterility testing on Oct. 19, 2006. However, on June 20, as part of an investigation into inactivation failures, a lot was retested for safety and sterility and found non-sterile. Fifty percent of the volume was retested and failed sterility on July 7. The contaminant was identified as Candida guilliermondii.

The root cause for sterility failures of both lots was determined to be "low levels of contamination of individual bottles" of a media lot with Candida guilliermondii during aseptic filling of the media lot, yet 15 additional lots that came into contact with this media lot did not undergo a second sterility test as was performed during the investigation of previously manufactured lots.

In the "Master Plan Stability of Media, MPS-001" the media/buffers assigned to each group were not defined, and the rationale for determining which media or buffer represented the worst case for each group was not documented in the study. Also, the actual storage containers were not represented for media and buffer solutions used in the production of Rabies Vaccine.

Stability study Q312-001, which was conducted to establish an expiry period for Diphtheria and Tetanus Toxoids Adsorbed concentrate (without preservative), was inadequate because containers and closure systems used for stability sample storage were not representative of the final container.

Validation study for the use of certain connectors for the media formulation area of a building was inadequate because the final report did not address the reason for the deviations or the potential effect of the deviations on the study. The final report inaccurately stated that media filled were incubated according to the protocol. The samples were actually held for a number of days not in compliance with the protocol.

The final report did not completely address the deviation in the batch record for the media fill study in which media was observed leaking. The impact of this event to the process validation was not addressed.

The firm failed to follow written procedures to assure proper inactivation of the rabies virus suspension.

The agency also noted that Sterility Failure investigations were incomplete. Twelve out of 96 bottles from a media lot remained after the sterility failure finding in October 2006. All 12 remaining bottles of the media lot were discarded from production without subjecting the media to sterility testing to determine the extent of the contamination. The retained sample for filling of a media lot was not tested for sterility.

In 2005, six Rabies Vaccine lots intended for use in Rabipur were found non-sterile; four were linked to contamination of media. The failure investigation was closed in January 2006. Corrective actions were recommended and several were implemented, but final correction was not implemented until January 2007. From January 2006 to January 2007, two media lots were found to be contaminated and two RabAvert vaccine lots were linked to contamination of a media lot.

As part of a failure investigation into inactivation failures, 21 Rabies Vaccine lots that came in contact with a media lot were re-tested for sterility; one lot was found non-sterile upon retest.

Three rabies vaccine batches that failed viable rabies virus testing after the virus inactivation process were rejected and the firm suspended production of rabies vaccine in February 2007. The investigations were incomplete because they did not identify a root cause for the virus inactivation failures and did not include an evaluation of the cleaning processes and procedures for product contacting equipment to determine if equipment cleaning is effective in preventing cross contamination.

Appropriate validation studies had not been conducted for critical processes. The cleaning procedure for the centrifuges was revised May 14, 2004, to remove the requirement for disinfection for a specified number of hours after each use/prior to cleaning. Additionally, the cleaning procedure for inactivation, SOP 102448, was revised May 17, 2004. These new cleaning procedures were not validated to establish the impact of the changes on the cleaning process.

The deficiencies described in this letter "are indicative of your quality control unit not fulfilling its responsibility to assure the identity, strength, quality, and purity of your components/in-process materials." FDA requested that the company describe how it will attain GMP compliance with regard to bulk lot production and process controls and investigations, including how it will use all of the relevant information to conduct thorough investigations to ensure that adequate steps are taken to evaluate whether deviations impact product, and to implement effective corrective and preventive actions."