Vaccine Manufacturing Déjà vu: A 483 Review

June 24, 2008
Noncompliance appears to focus on the same issues: inadequate SOPs, root cause analysis and CAPA

In 2003, MHRA’s and FDA’s investigation of Chiron’s Fluvirin flu vaccine manufacturing facility in the U.K. brought to light some important cGMP deficiencies. However, since then, 483’s for a number of vaccine manufacturing facilities (including two subsequent reinspections of Chiron’s Liverpool plant) have continued to point out similar GMP problems focusing on faulty standard operating procedures (SOPs), inadequate testing, validation and root cause analysis, and lack of specificity in batch records and alarming.

This article takes a very quick look at some patterns in noncompliance found in vaccine manufacturing 483’s of fairly recent vintage (you’ll find some of them on www. and a more extensive listing in ComplianceGuru’s (West Newbury, Mass.) online library.

Among the FDA inspections examined:

  • Merck’s April 2008, in West Point, Pa.
  • MedImmune’s in March 2007
  • Chiron’s May, 2006 and October, 2004 in the U.K.
  • Sanofi ’s April 2006.

For many of these facilities, the most fundamental problem appears to be a lack of robust SOPs for key manufacturing practices, particularly cleaning validation and bioburden testing, notes Michael Gregor, principal of Compliance Gurus, who focuses on validation and auditing, and who tracks 483s closely, offering many recent documents in an online library.

“In the past two years, we’ve seen a notable problem with SOP quality. Companies think they’re covering all bases, but their SOPs are often mere placeholder documents,” Gregor says. Documents are sometimes written by those without manufacturing perspective, and without adequate scientific knowledge or review, he notes. Several recent vaccine manufacturing 483’s note insufficient level of investigation into adverse reactions.

“Companies need to have better CAPA programs and processes in place, since CAPAs at some facilities tend to go unresolved for over a year,” Gregor says. “In general,” he adds, “there is a lack of SOPS defining CAPA processes, and also a failure to distinguish between deviations and CAPA.” Commercial compliance software solutions may be a contributing factor, he says. Some facilities may be using these packages right out of the box, Gregor says, when they require significant customization in order to meet FDA requirements.

In addition, he sees a pervasive lack of validation for such supporting systems as LIMS and improperly qualified systems. Gregor has also noted inadequate facility design and qualification for features such as cleanroom airlock and improper qualification for such basic equipment as the programmable logic controllers (PLCs) used to control valves and other equipment. 483’s noted lack of sufficient employee training in cGMPs and associated problems with gowning, use of eye protection and gloves; some found inadequate testing and alarm setting for purified water systems.

In a March 2007 483 of MedImmune’s facility, for instance, it was found that “specified [microbial] limits were not achievable, so no limits were applied.” Other 483 issues for all manufacturers included:

  • Inadequate auditing of raw material or equipment suppliers
  • Poorly specified alarm levels for bioburden levels, and failure to study reasons for exceeding alarm levels
  • Insufficient environmental monitoring, cleaning and sterility testing, and inadequate steps taken to identify and control mold

A 2004 reinspection of Chiron’s Fluvirin plant in Liverpool found that SOPs were not followed during sterility testing and that data from some failed batches were left out of sterility test and bioburden root cause analyses. In addition, FDA inspectors noted, environmental monitoring of cleanrooms was insufficient, while water quality was not investigated as a potential contributor to bioburden.

“There was no overall investigation to ensure adequate CAPA,” according to the document. Two years later, inspectors still noted that CAPA was inadequate for environmental monitoring trends in key sections of the plant. FDA’s 2007 inspection of MedImmune’s FluMist facility found inadequate documentation that disinfectants or manufacturing equipment cleaning validation had been reviewed.

Inspectors also noted lack of specificity in batch records, for example, the validated hold time for ultracentrifugation was not in the record, and the fact that there had been no documentation of the effect of fill volumes on container integrity. There was also no procedure governing the control of keys in the warehouse area, and uncontrolled data sheets were used without tracking or serial numbers —just a date stamp, which would make root cause analysis difficult.

In addition, the inspectors found, Quality Control testing SOPs lacked specificity---for example, there were no documented time, temperature or humidity limits for incubators, refrigerators or freezers that would trigger a material being out of specification after a specific period of time.

Redacted 483’s are freely available, and more newspapers are publishing them online. Not every 483 cited led to quality problems or recalls, and no doubt, manufacturers (and their competitors) have attended to these issues diligently. Or have they? Will there be more 483’s in the future, and what will they point out?

About the Author

Agnes Shanley | Editor-in-Chief