Cleaning, sterility among items holding up ANDA.

Bioindustria Farmaceutici, Novi Ligure, Italy.

A May 1994 inspection of bulk pharmaceutical manufacturer Bioindustria Farmaceutici in connection with an ANDA resulted in a "no go" answer from FDA, along with an eight-item 483.

Investigator Daryl DeWoskin of the Baltimore district office topped the list of objections with an observation that the manufacturer had no specification for percent of theoretical yield, making it impossible to find out whether yields were within proper limits, according to the EIR.

DeWoskin moved on to the water supply, questioning microbial counts and faulting sampling procedures. He also reported that the firm gave him conflicting information about its water supply, making it impossible to determine whether the system was indeed validated.

The validation issue appeared again in connection with equipment cleaning. The EIR and 483 noted a lack of validated cleaning procedures where a variety of bulk pharmaceutical chemicals (BPCs) were made on non-dedicated equipment.

Batch records came under fire, with DeWoskin describing them in the documents as incomplete. The investigator noted that temperatures were not recorded for all manufacturing processes and found failure to record the review and approval of batch records or product release.

Temperature concerns showed up one more time in relation to stability samples, which were not stored in a controlled environment, according to the EIR and 483.

An observation that contamination of a BPC was not minimized stirred up controversy during discussion reported in the EIR, indicating that the agency and Bioindustria Farmaceutici had very different ways of looking at the issue.

For instance, when DeWoskin said that an employee should have worn protective clothing when handling intermediates, company personnel retorted that she did -- a lab coat and a hard hat. DeWoskin responded that the lab coat was worn around other products, while the hard hat protected the employee, not the product, the report stated.

What the investigator "did not want to hear"

Management argued further that small amounts of contamination would do no harm. According to the EIR, DeWoskin answered, "I did not want to hear that it is not a health hazard if a small amount of one drug gets in another.... I just wanted personnel to take precautions to minimize contamination, such as wearing protective clothing and sealing bags."

Another practice that turned up on the 483 concerned filling out batch records before processing occurred. Company employees routinely recorded projected completion times before pro-cedures actually were finished, the report said.

A training issue concluded the 483, with DeWoskin finding no evidence that employees involved in the manufacture of BPCs and intermediates had taken part in "on-going training." On April 14, 1995, FDA called the company response inadequate. The agency dismissed as unclear the firm's approach to correcting the problem involving theoretical yields. FDA also found the proposed protocol for validating cleaning procedures lacking in specificity and described requirements for storing stability samples.

Bioindustria Farmaceutici, Novi Ligure, Italy, 5/24-25/94, 107515M, $8 plus handling.