Pablo Software Solutions
DEPARTMENT OF HEALTH & HUMAN SERVICES
Certified Mail
Return Receipt Requested
June 27, 2000
Bernard J. Kelley
President
Merck Manufacturing Division
1 Merck Drive
P.O. Box 100
Whitehouse Station, NJ 08889

Dear Mr. Kelley:
During our February 1 through March 16, 2000 inspection of your manufacturing facility located at 126 East Lincoln Avenue, Rahway, New Jersey, our investigator documented deviations from current good manufacturing practices applied to the manufacture of Active Pharmaceutical Ingredients (API's). These deviations cause your drugs for human use to be adulterated within the meaning of Section 501(a)(2)(B) of the Federal Food, Drug and Cosmetic Act. The deviations included:
Out-of-specification data were averaged with passing data to facilitate the release of Dorzolamide HCI Pure lot HRDT090, Dorzolamide HCI Crude lot HRDR090, and Impenem non-sterile lots HRI460 and HRI610. In addition, for each lot an average of several test results was reported without mention of the result being an average of multiple test results.
Out-of-specification assay results for lot HRVF030 of Rofecoxib Pure were invalidated on the premise that a sampling error had occurred. No documentation or data was presented or discovered during the inspection to support that conclusion. This lot was repeatedly tested (6 times) and the third through sixth re-tests also gave out-of-specification results. The third and fourth re-tests were invalidated due to sample preparation. again there was no evidence of sample preparation error and the procedure is not defined in the analytical method.
www.cancer.net.nz
Re-sampling and re-testing was performed on intermediate API lots without justification for: xxxxxxxx pre-reaction solution for Amitriptyline lots HRAR210A,210B, 220A, 260A, and 280A for water content failures using Karl Fischer determinations; Rofecoxib Pure lots HRVD690, HRVF650, HRVD460, HRVF190 and HRVF200 for extraneous matter failures; Rofecoxib Pure lot HRVD240 for a degradation peak; and ten Bromosulfone (starting material for Rofecoxib) lots for elevated impurity levels.
Two lots of Rofecoxib Pure (HRVF770 and HRVF850) were released for distribution despite analytical findings that they were contaminated with polypropylene fibers in excess of the action limit. Six lots of Rofecoxib Pure also demonstrated polypropylene fiber contamination but were deemed acceptable and released. Twenty-three lots of Rofecoxib pure were subjected to the same sifting process, later identified as the cause of the polypropylene contamination. The sampling procedure was inadequate, as it did not include an acceptable number of drums to be sampled. Also, the extraneous material action limit used in this investigation is inadequate, in that no instructions are present if extraneous material in excess of the specification limits is found.
Manufacturing investigations for nine Imipenem batches were not completed timely, ranging from ten months' duration to as long as two years.
The above deviations.................................reply should be directed to Kirk D. Sooter, Compliance Officer, U.S.Food an Drug Administration, 10 Waterview Boulevard, Third Floor, Parsippany, New Jersey 07054. If you have any questions about this letter, please contact Mr. Sooter at 973-526-6008.
WARNING LETTER
Public Heath Service

Central Region m389zn

Food and Drug Administration
Waterview Corporate Centre
10 Waterview Blvd., 3rd Floor
Parsippany, NJ 07054
Telephone (973) 26-6008
File # 00-NWJ-35
Use of a batch of intermediate material rejected due to contamination with a foreign substance in manufacturing two Amitriptyline finished API lots (ERF5390 and ERF 5411). As of the date of the inspection, portions of both lots remain under your direct control and approved for distribution.
Sincerely yours,

Douglas I. Ellsworth
District Director