FDA Recalls - Insulin Makers Listed

Why aren't insulins on the list?

DRUG RECALL DATA FOR 2000

DRUG RECALL DATA FOR 2000 show the influence of adverse FDA inspection and sample findings. A majority of recalls conducted recently have been prompted by FDA rather than manufacturer quality systems -- a disturbing trend from the agency's point of view. FDA compliance concerns are reflected, in particular, in a significant number of recalls during 2000 involving the potential for product contamination from faulty aseptic processing and the mishandling of beta-lactam products.

Six contamination-related recalls drew a Class I rating as did nine involving products with unapproved claims or ingredients. Error and accident reports shed light on manufacturing problems biologic producers are having, but only a limited percentage result in recalls. [An eight-page tabulation of drug recalls listed in FDA "Enforcement Reports" during 2000 begins on p. 12.]

Recalls Prompted By FDA On The Increase A majority of drug product recalls conducted over the last two years have been initiated due to FDA inspection and sample findings. According to data collected by the compliance office at the Center for Drug Evaluation and Research (CDER), over 60% of the recalls conducted during the agency's past two fiscal years (FY 1999 and 2000) were prompted by FDA findings. The recent data contrasts sharply with that from the previous three years (FY 1996-98), during which two-thirds of the recalls were generated by the findings of the company involved and only one-third by FDA's activity.

FDA's compliance office views the upsurge in FDA-prompted recall actions as indicating that companies are either not reporting problems that they find to FDA through the voluntary system, or are not finding them, reflecting growing weakness in company quality assurance operations.

CDER Compliance Officer Richard Friedman expressed concern about this trend at an FDA tutorial on aseptic processing that followed the annual University of Georgia GMP conference in mid-March. "This is a remarkable turnaround in problem detection and may reveal an emerging problem with the effectiveness of QC unit oversight in the industry," Friedman commented. The issue of the voluntary nature of the recall system has been a recurrent focal point in Congress's review of agency operations. FDA's response to the Congressional inquiry on this issue in the past has been that keeping the system voluntary encourages companies to find and correct their own problems. However, the recent data may push the agency to rethink the issue.

Of course, "voluntary" may in reality mean that the firm chooses to conduct the recall instead of facing a seizure action by the agency. Many of the agency-prompted recalls sprang from cGMP deviations found during FDA inspections. Where the GMP problems are widespread and of sufficient significance, the result may be the recall of many products from the implicated manufacturing or repacking facility.

A prime example during the past year was a multi-product recall conducted in March by the repacker Pharmaceutical Corporation of American (PCA). PCA pulled back more than 200 repackaged prescription (Rx) and over the counter (OTC) drug products after FDA investigators cited the firm for a number of GMP deviations, including the lack of adequate controls to prevent beta-lactam antibiotic cross-contamination. The firm was repacking the drug products into small containers for distribution to hospitals and other healthcare facilities. PCA is no longer in the repackaging business as a result of its compliance problems. [EDITOR'S NOTE: The exceptional nature of the PCA recall action led FDA to count it only once in its recall data tracking system, whereas "The Gold Sheet" counted each separate chemical entity involved.] PCA's compliance problems began with the recall in late 1999 of a lot of carbamazepine, a portion of which was found to be labeled as tetracycline.

FDA inspected PCA following the recall and then sent a warning letter to the firm in February 2000. Investigators went to the facility to determine how the carbamezepine/tetracycline mix-up occurred, the letter explained, "as well as where the firm's quality systems failed in allowing the release of the misbranded product." The warning letter alleged that PCA was operating "with significant deviations" from GMPs. The letter focused entirely on the firm's quality control procedures for product labeling. Following the warning letter, FDA reinspected the repacker and found additional GMP problems, including the handling of beta-lactam products alongside other dosage units without adequate precautions against cross-contamination. These findings prompted the recall of multiple strengths and brands of over 240 separate chemical entities. Among the products involved in the recall were a number manufactured by big pharma companies, including Rhone-Poulenc Rorer, Abbott, Glaxo, and Lilly. Schering Recalls Aerosols Missing Active Ingredient Other recalls attributed specifically to GMP deficiencies in FDA's listings for 2000 were initiated by Schering-Plough, Wyeth-Ayerst, OMJ, Anabolic, and Leiner. The latter three launched recalls after findings that they had not properly validated or revalidated procedures and/or specifications. Schering-Plough recalled several hundred lots of albuterol inhalation aerosol and beclomethasone inhalation aerosol because they may have included canisters that had no active ingredient. The action was initiated by the firm in December 1999 and extended in the spring of 2000. The Schering recalls have generated attention from the Washington, D.C.-based consumer activist organization, Public Citizen. The group was supplied anonymously with an FD-483 from an inspection completed in January 2001 that covered Schering's Kenilworth, New Jersey facility and referenced the recalled products. The group also was supplied with an external audit of the facility, conducted by Rockville, Maryland-based AAC Consulting Group between February and April 2000.

In early March, Public Citizen sent a letter to HHS Secretary Thompson calling for a criminal investigation of the GMP problems at Schering relating specifically to the albuterol recalls. The group alleged that Schering became aware of many of the deficiencies in the plant when the AAC consultants audited the facility -- several months before the FDA inspection. Citing the AAC findings, Public Citizen noted that Schering managers and supervisors indicated that there was an imbalance between quality and production, leaning considerably toward the side of production. Included in the January FDA-483 report were observations alleging that the product quality review methods for albuterol delivery and particle size "were inadequate in that the methods exhibit various unidentified extraneous peaks." FDA also maintained that the "methods for total content of albuterol per can assay and estimation of degradation of products were also inadequate in that there was a lack of resolution between typical unknown peaks from neighboring active or placebo peaks." The methods were used to test and release product batches, as well as stability samples, from October, 1999, to December, 2000 -- the period of time during which the recalled products were produced.

The consumer activist group alleged that Schering "knowingly shipped millions of the 59 mil. units of albuterol...between the time the company became aware of the seriously flawed manufacturing processes and the time the recall was finally accomplished." Public Citizen further urged HHS to investigate the company for continuing to ship other prescription drug products while fully aware of the serious violations of GMP regulations during their production. OMJ was cited for failing to validate new manufacturing specifications for its ophthalmic solution, Vasosulf (sulfacetamide/phenylepherine).

Phenylpropanolamine/chlorpheniramine/methscopola mine tablets were recalled by Anabolic after FDA findings of incomplete process validation. In Leiner's case, hundreds of lots of pseudoephedrine tablets were recalled because the firm allegedly failed to validate the sequence of addition of components to the process. The cough/cold medicine was sold under various pharmacy labels. Leiner also recalled pseudoephedrine/chlorpheniramine tablets in 2000 because it did not properly adhere to a validated procedure. Specifically, the firm reduced the approved blend time and added an amount of an ingredient in excess of the approved formulation. Gabriel Skin Care Distributors recalled a number of topical products after findings that it did not perform active ingredient testing. The products involved in the action included blemish removal gels and sunscreens. Wyeth Recalls Accompany Compliance Problems Wyeth-Ayerst's GMP compliance difficulties with FDA are reflected in a series of recalls this past year. Overall, Wyeth-Ayerst recalled 10 products in 2000, six of which were manufactured at its aseptic processing facilities in West Chester and Marietta, Pennsylvania. In the past six years, several Wyeth plants have failed FDA inspections. At the Marietta facility, the extent and seriousness of deficiencies found by agency investigators during multiple inspections culminated in a seizure of products by FDA and the signing of a consent decree last fall ("The Gold Sheet" February 2001). The facility produces sterile drugs and vaccines, including the influenza vaccine. At the West Chester facility, a 1999 inspection resulted in a warning letter ("The Gold Sheet" March 2000). Plants in Canada, New Jersey, New York, and Virginia also have received warning letters since 1996 ("The Gold Sheet" March 1999). Following up on the 1999 inspection at West Chester, FDA investigators audited the plant for six weeks last summer and found additional GMP problems. As a result of the findings, the firm recalled two injectable products, meperidine and heparin, for "lack of sterility assurance." Prior to the inspection, Wyeth had recalled a lot of the meperidine produced at the West Chester facility and several lots of the product made at its Marietta facility because of sterility assurance issues. Sterility assurance concerns also prompted Wyeth to recall three other injectable products manufactured at the Marietta facility.

In July, one lot each of digoxin, dimenhydrinate, and diphenhydramine were recalled when cracked cartridges were uncovered. The cracking was caused by contact with a rail in the filling line which held the bottles into position prior to the capping operations. The cracks prevented a proper seal, potentially compromising the sterility of the drug product. Wyeth's recall problems were not isolated to the two Pennsylvania facilities in 2000. Across the Delaware River, the firm's Lederle plant in Cherry Hill, N.J., recalled a lot of injectable morphine sulfate for which a label declaring an incorrect strength was used. A dissolution failure prompted the recall of several lots of Premarin and Premphase conjugated estrogen tablets, which are manufactured at Wyeth's Rouses Point, N.Y., facility. Recalls also were initiated by Wyeth's Carolina and Guayama, Puerto Rico plants. Vial breakage was the reason given for the call back of piperacillin/tazobactam injectable products manufactured at the Carolina facility. The Guayama action involved Lodine (etodolac) capsules, which were recalled after findings of possible cross-contamination with the beta-blocker, acebutolol HCl. This action received FDA's Class I risk classification, indicating that the mistake could result in serious health consequences or death.

Contamination Problems On The Upswing

Lack of certainty about sterility, potential cross-contamination, and confirmed microbial contamination of both sterile and non-sterile products have been a growing problem area over the past few years. Overall, contamination-related issues accounted for 53 recalls last year. A large recall action involving a number of companies was initiated in 2000 after some lots of Clinipad Corporation's alcohol and povidone iodine sterilizing products were found to be contaminated with the microbes Pseudomonas aeruginosa, Stenotrophomas maltophilia, and Coagulase Negative Staphyloccus. The discovery of the contamination problem in December 1999 prompted Clinipad to launch a Class I recall of all lots of the products manufactured over the prior three years. FDA noted in a March 1999 warning letter to Clinipad that, during four separate inspections between 1992 and February 1999, investigators had observed that cleaning procedures were not properly validated. The warning letter also cited the firm for faulty process validation for a number of products, including those involved in the recall. The letter noted that the firm had committed to complete process validation by the fall of 1999. The recalled topical sterilizing solutions were distributed in different fluid ounce containers and ready-to-apply swabsticks, prep pads, towelettes, and pouches. Two injection kits containing Clinipad products were recalled as well -- Pharmacia & Upjohn's Caverject (alprostadil) and Aventis' Lovenox (glatiramer). The two recalls received a Class II rating from FDA. A large number of medical device manufacturers were also embroiled in the Clinipad situation. Hundreds of thousands of individual units produced by over 140 device companies contained suspect Clinipad products. Medical device manufacturers caught up in the Clinipad action included Bard, Abbott Laboratories, Boston Scientific, Cook and Becton Dickinson. FDA noted that the Clinipad products were "widely distributed" to blood banks, hospitals, and clinical and retail pharmacies, either separately or as part of pre-packaged kits and trays from a number of manufacturers under various names. Connecticut-based Clinipad went out of business soon after the recalls.

Microbial contamination was implicated in three other Class 1 recalls in 2000: Rich Distributing called back methylsufonylmethane eye and ear drops (Pseudomonal fluorescen); H.L. Bouton recalled saline solution (Burkholderia cepacia); and Care-Tech Labs pulled back its Techni-Care cream (Pseudomonas aeruginosa). A Class II recall by Symbiotics followed the discovery of Pseudomonas putida and Tsukamurell paurometabloum contamination in Colostrum cream, manufactured by Vege-Kurl. In addition to the Wyeth actions, recalls for lack of sterility assurance were conducted by Eli Lilly Italia, American Pharmaceutical Partners, Bayer, and Miza. The recalls all were rated Class II with the exception of Miza's, which received the least serious, Class III rating.

FDA compliance officials have been closely tracking non-sterility problems in drug products. CDER's Friedman noted at the recent Georgia conference that the recall data for sterile drug products over the past five years shows that there were "a significant number...due to non-sterility as well as lack of assurance of sterility" -- the two categories his office uses "as a barometer of sterile drug quality in the recall context." FDA's review of this data, Friedman reported, shows that "all drugs recalled due to non-sterility over the last five years have been produced by aseptic processing." In addition, he noted that only one product produced by terminal sterilization was included among those recalled for "lack of sterility assurance." In that case, the drug firm "apparently forgot to irradiate the product and had to pull it off the market immediately." By FDA's count for fiscal year 1999, 83 drug recalls fell into the two sterility problem categories, with 28 of these involving products that were actually found to be non-sterile.

CDER compliance officials are also concerned about a recent uptrend in recalls of non-sterile products resulting from microbial contamination. The majority of these cases occur when companies do not properly control water and air systems or do not examine their raw materials for microbial contamination. The agency has been cautioning industry to be concerned about the presence of microbes in non-sterile products, especially when the drug is used for wound care. FDA is also noticing an increase in cases like that involving Pharmaceutical Corporation of America where adequate controls are not established to prevent cross-contamination with penicillin and derivative products.

In addition to Wyeth and PCA, repackagers Oribi and Spectrum conducted recalls during 2000 stemming from cross-contamination concerns. In Oribi's case, 18 bulk chemicals were packaged for pharmacy compounding on equipment also used to package penicillin or cephalosporin products. The Oribi action elevated the number of active ingredients recalled in 2000 to 23 compared with seven in 1999 and only one in 1998. Spectrum likewise recalled a bulk drug, dihydroergotamine powder, after packaging it on equipment used also for cefadroxil. The PCA, Oribi, and Spectrum actions received FDA's Class II rating. The requirement that firms separate the manufacture of penicillin and related products from the manufacture of other drug products has been in effect since the 1970s, but is one that may be getting overlooked by a new generation of production and quality officials, FDA officials point out. Another problem area that caused a number of recalls in 2000 was the failure of products to meet impurity or degradation specifications (included in the "other product specification problems" category in the box on pp. 12-20). Last year, seven different manufacturers pulled products out of distribution because of degradation or impurity failures. Recalls due to drug degradation were performed by Morton Grove (clindamycin), E. Fougera (clobetasol), and Altana (desoximetasone). Impurity failures prompted recalls by Glaxo Wellcome (chlorambucil), Abbott (cyclosporine bulk), Novopharm (glyburide), and BASF (ibuprofen). PCA Action Helps Boost Yearly Recall Total 40% The multi-product PCA recall significantly skewed the calendar year 2000 recall count, which totaled 464 -- a 40% increase over the previous year. The recall total for 1999 was elevated by recall actions involving 145 sterile drug products manufactured by Shein subsidiaries Steris and Marsam ("The Gold Sheet" February 2000).

Excluding the multi-product actions by PCA and Shein, the number of recalls for the two years would have approached the baseline of about 200 recalls over the previous few years. While the number of recalls falling into the "manufacturing/testing method deficiencies" category spiked due to the PCA action in 2000, the numbers in the other problem area categories generally remained consistent with 1999. The other exception in the two-year comparison was the category of "non-compliance with NDA or monograph requirements." The number of recalls in this category last year more than doubled the 1999 tally primarily because of the sale of unapproved or undeclared drugs by a number of dietary supplement companies. These actions resulted in 13 of 31 NDA/monograph-related recalls in 2000. Nine of the 13 recalls attributed to the use of unapproved or undeclared drug products received the most serious, Class I rating. The use of undeclared glyburide was specifically cited in four of these cases. Overall, FDA assigned a Class I rating to 15 drug product recalls in 2000 -- the second consecutive year in which Class I's reached double digits. Prior to 1999, the last time the number of Class I-rated recalls reached double digits was 1990, when industry was beleaguered by various labeling and packaging mishaps. FDA compliance officials have noticed a positive trend in the recall data regarding labeling and packaging mistakes. A source of many serious recall actions in the early 1990's, labeling-related recalls have been less frequent in recent years. In 1996, labeling mix-ups accounted for only 23 recalls -- a 10-year low for the category. Since then, the number of such recalls has hovered in the thirties, with 34 reported by FDA in 2000. Prior to 1996, labeling-related recalls averaged over 50 per year. Labeling Mix-Ups Plague Biologic Firms In 2000 Labeling and packaging mishaps were a problem for non-blood biological producers during 2000. Of the 11 non-blood biologic recall actions listed in FDA enforcement reports during the year, six were attributed to labeling mix-ups involving allergenic products. Four of these were conducted by Ohio-based Allergy Laboratories, and one each by Bayer and Hollister-Stier.

Potency problems prompted Bayer to carry out additional recalls last year. The firm recalled two antihemophilic products, recombinant Kogenate and recombinant Helixate. Aventis Pasteur also called back two lots of its rabies vaccine for failing potency testing. In March, Genentech recalled two dozen lots of its Pulmozyme (dornase alfa inhalant) due to findings of "miniscule" punctures at the tip of the single-use ampules in which the cystic fibrosis therapy was packaged. The punctures compromised the integrity of the container closure and allowed product leakage. Later in the year, an FDA inspection prompted Genentech to recall an additional lot of Pulmozyme based on findings that an unapproved re-filtration step was utilized. (This action did not appear in FDA's enforcement reports during 2000.) The re-filtration step was performed to remove metal particulates that had been found in the concentrated bulk drug substance. The firm received a warning letter following the inspection and the recall that highlighted the re-filtration issue ("The Gold Sheet" February 2001).

Because of the limited number of biologic products recalled each year, a more meaningful assessment of quality problems in the industry can be made by reviewing data from "error and accident" reports.

The Center for Biologics Evaluation and Research (CBER) tracks error and accident (E&A) data by the type of companies filing the reports and the problem areas cited in them. The center also notes when E&A reports are forwarded to district offices for follow-up and evaluation as potential recall situations. The data is on FDA's website at www.fda.gov/cber/efoi/error.htm. The last report posted summarizes data for the first three quarters of FY 2000 (ending June 30). Overall, 110 non-blood E&A reports were submitted by industry in this period. Of these, CBER concluded that eight referred to "non-reportable" problems. CBER groups the non-blood E&A reports into six categories: reagent performance, sterility compromised, labeling, storage/distribution, processing, and miscellaneous. In FY 2000, the majority (26%) the E&A reports involved labeling errors, coinciding with the relative frequency of this problem as a recall cause. Another quarter of the reports involve processing errors. Sterility problems were the next most frequent reason for E&A filings (18%). Over the nine-month period, therapeutic biologic manufacturers accounted for 23 of the 110 reports filed with CBER. Processing problems were the cause of nine of the therapeutic product reports. Failing to meet a product specification, mixing of purified water into a WFI distillate stream, and failing to sanitize a gowning room were among the specific reasons included in this processing category.

Five of the E&A reports from therapeutic biologic manufacturers fell into the miscellaneous category. The specific problem areas included stability failures, lack of vacuum in vials, and vial breakage. The categories of sterility compromised, storage/distribution, and labeling each had three reports. Of the 110 reports submitted altogether, only nine were from vaccine manufacturers. Stability and specification-related problems prompted more than half of these. One of the reports referenced a "non-reportable" error because the product was not in distribution. Almost a third of the E&A reports (32) were filed by in-vitro diagnostic manufacturers for which labeling was the primary problem area (14 reports). Allergenic manufacturers filed 27 reports; ten of these referenced labeling problems and nine referenced sterility issues. Blood derivative manufacturers filed 19 reports.

According to the CBER data, only a limited number of E&A reports were passed on to the district offices as recall candidates. Of the 110 FY 2000 E&A reports, 14 generated recall inquiries. For allergenics and therapeutics, the ratio of E&A reports to recall follow-up was about the same, 27/6 and 23/4, respectively. Only four of 32 reports filed by IVD firms were sent along to the districts, and none of those involving vaccines or derivatives. Biologic Companies Cited For Not Reporting E&As Team biologics is finding that the failure to comply with accident and error reporting requirements ranks with process validation as a frequent inspection issue. At the fall 2000 PDA/FDA conference, then Team Biologics expert investigator Marsha Major asserted that failure to file accident and error reports is cited on almost "every 483" issued by team investigators. In May, error and accident reports will officially be renamed "biological product deviation" reports, reflecting FDA's recent revision to the relevant statute (21 CFR 606.14). The revision was completed last year and published in the Federal Register in November. At the same time, FDA published a new rule (21 CFR 606.17) that applies identical deviation reporting requirements to blood and blood component operations. Besides changing the report names, the new regulations clarify that deviation reporting is limited to only those products which have entered into the distribution chain. FDA considers products which are no longer in the control of the licensed manufacturer to be in distribution.

"Control" is defined by the agency as "having responsibility for maintaining a product's continued safety, purity, and potency, and compliance with applicable product and establishment standards and cGMP requirements." In the joint-preamble to the rules, FDA stresses that, in cases where product does not leave the control of the manufacturer, investigations into product deviations would continue to be required per the GMPs even though deviation reports are not required. The final rules establish two other criteria that must be applicable to an event in order to trigger the filing of a biological deviation report. The first is that the event involves a deviation from cGMPs, applicable regulations, applicable standards or established specifications, or represents an unexpected or unforeseen occurrence "that may affect the safety, purity, or potency" of the product involved.

The other criterion is that the deviation occurs in the license-holder's facility or another facility under contract. Companies have 45 days to report a deviation, according to the new rule. FDA clarifies that the reporting clock begins on the date that "the manufacturer, its agent, or another person performing a manufacturing, holding, or distribution step under the manufacturer's control, first discovers information reasonably suggesting a reportable event has occurred." The draft version of the rule, released in 1997, generated a lot of discussion between representatives of CBER and industry at public workshops and meetings. FDA received 98 written comments from manufacturers, blood establishments, trade associations, professional associations, and other groups and individuals. In a lengthy preamble to the final rule, FDA addressed a number of issues raised by industry in the written comments on the draft revision. The agency reported that about a third of the comments generally supported the rule, less than a third "objected" to the entire rule, and several comments from individual transfusion and pharmaceutical companies "objected" to the application of the reporting system to their establishments. Regarding the change in nomenclature, FDA said that the terms "accident" and "error" are "immaterial to the purposes underlying the rule and appear to have caused confusion."

E&A reporting is meant to capture all cases that deviate "from cGMP, applicable standards or established specifications, or represent unexpected or unforeseeable events...regardless of whether or not they are considered 'errors' or 'accidents.'" Some industry commentators on the proposed draft felt that the reporting system was redundant to internal QA programs. In response to this concern, FDA explained that the reporting system provides information "that an individual establishment's QA program may not detect" and helps the agency track industry-wide quality trends.

The preamble points out that, "if an event occurs once a year in every establishment, it may not appear significant to any single establishment. The reporting system will allow FDA to recognize the significance of that event in a timely fashion and to take appropriate action to protect public health." The agency further noted that the information would "aid FDA, licensed manufacturers, and unlicensed blood establishments in appropriately targeting QA efforts to improve quality and reduce manufacturing problems." Besides the review of deviation reports when they are received by CBER, the preamble stresses that FDA investigators will review them during routine inspections and "examine all manufacturing deviations, not merely reportable deviations, to ensure that the establishment has followed all established [SOPs] related to investigation, follow-up, and reporting of deviations."

The reports will also help FDA target areas for future guidance. To the suggestion that the rule was imposing a new burden, FDA countered that CFR 600.14 required the reporting of errors and accidents prior to the revision, and that by limiting the reporting requirement to distributed products, the new rule "would decrease the existing burden." The revision was necessary, FDA said, "to ensure that all manufacturers understand their reporting requirements, to expedite biological product deviation reporting, and to enable FDA to monitor accurately the safety of biological products." There was some concern that FDA did not have the authority to extend the reporting requirement to deviations from "applicable standards" or "established specifications." In reply, FDA noted that the rule calls for the reporting of events that "may affect the safety, purity, or potency of a product." The agency maintained that such events "fall squarely within FDA's jurisdiction."

FDA did change some of the language of the rule in accord with industry concern. For example, the term "made available for distribution," which was used in the draft version, was changed to clearly indicate that the rule applied to "distributed products." FDA has received positive feedback from industry since the final rule was published. Although the new biological deviation report forms are not required until May, firms have already begun using them. CBER added a page to its website (www.fda.gov/CBER/biodev/biodev.htm) that contains information to assist firms in using the new rule. The center also plans to release two draft guidances on the new biologic rules. Eventually, FDA will accept biological product deviation reports electronically, and the electronic form will be made available on the website. FDA Is Updating Recall Processing Operations FDA's Office of Regulatory Affairs (ORA) expects to complete a conversion of the agency's recall processing operations to electronic media within six months -- the first phase in a comprehensive recall reengineering effort. Sandra Whetstone, director of the Division of Compliance Management and Operations within ORA's Office of Enforcement explains that the primary goal of the electronic initiative is to eliminate the inefficiencies and redundancies of the current recall program.

"All of our [field] recall coordinators will work on the web" using a "uniform [recall] submission form amenable to all centers," she said. Whetstone noted that currently field personnel must submit several different forms to the various FDA centers for every recall. For example, the field must prepare the initial recall alert, a classification recommendation, and updates on the action, as well as a notice about the termination of the action. "Oftentimes, lots of information has to be reentered," the ORA division director said. The electronic form will consolidate all the necessary recall data on a single entry, so the same information does not have to be entered multiple times. The ORA project is an extension of the Center for devices' recall reengineering initiative. Under the CDRH program, field personnel were given access to the device center's recall database in order to help them classify recalls without CDRH input. Use of a uniform electronic recall form also will facilitate creation of a single recall database for the agency which should aid FDA in the analysis of recall trends. Currently, the FDA centers have separate recall databases that may or not be compatible with each other. Whetstone estimates that it may take a contractor nearly a year to consolidate all the different databases into a single, searchable entity. FDA's electronic recall initiative also will strive to provide more and quicker recall information to the public.

According to Whetstone, the agency's internal recall systems will be linked to FDA's website, allowing the agency "to publish in 'real time' public recall information." Currently, such information is published in the agency's weekly Enforcement Report weeks and sometimes months after an action has occurred. The agency envisions recall data being uploaded to its website in three phases as information becomes available. "Phase 1" data would include the same information generated by field coordinators when they send a recall alert to the FDA centers.

That information would include: the name of the product and recalling firm; a list of any identifying codes, such as lot numbers or catalog numbers; a description of the problem or reason for the recall; the area of product distribution; and the status of the recall. Additional links would be created to any press releases issued by FDA or the recalling firm on the action, and to any recall letters, if available, the firm sent to its customers. "Phase 2" data would be posted after the field made its initial recall recommendation, Whetstone explained. More information on the quantity of product involved would be available at this point and a clearer picture of what the "depth" of the recall should be -- that is, whether the action should stop at the distributor level or go further to individual customers. "Phase 3" data would be the information provided by the various FDA centers, such as the final recall classification and official recall number. FDA also intends to use its internet site to make the recall process more "transparent" for industry and consumers, Whetstone remarked.

For example, the agency plans to post recall definitions, industry guidance, press releases related to specific issues, and any "model" press releases developed by FDA for high-volume, reoccurring recall situations, for example with allergens. The recall internet site also will include a newly devised health hazard evaluation (HHE) form that will be used by all the centers to classify recall actions. Previously, each center has used separate, albeit similar, HHE criteria to classify recalls. The agency "relooked at our health hazard evaluation form and we developed one form for all of the various products -- foods, drugs, devices, and biologics," Whetstone said. "What we have done is to clarify the guidance that we give our [center recall] boards so that they will be more in tune with each other in different areas," the FDAer noted. "So the health hazards should all be done basically the same way." The HHE form and criteria as well as some model evaluations for various device, drug, biologics and food products at different risk levels will be displayed on FDA's website in order to provide consumers and industry an idea of the thought processes that go into a recall classification decision, Whetstone commented. The ORA official noted that, in addition to creating an electronic recall program, FDA is evaluating all its recall procedures "to see where we can improve." For example, the agency is reassessing how it counts recalls. "Right now we count recalls based on all of the products involved," Whetstone stated.

So, for example, a single event that affects 10 different size catheters is now counted as 10 recalls. FDA is "leaning toward" a "one event, one recall" counting method, she said, in line with how recalls are tabulated by the U.S. Department of Agriculture and Canadian authorities. The agency also would like to streamline its procedures for sending notification letters to firms involved in recalls. Letters are sent for various reasons, she noted, such as when FDA requires additional information or when the agency has concluded a recall is complete. Notifications can come from the districts, the centers, or ORA.

"We are looking at who is going to send the letters, how many letters do we send," and the feasibility of "combining recalls with warning letters," Whetstone said. Another area to be reviewed is FDA's procedures for terminating and auditing recalls. "When is a recall supposedly closed and what do we consider effective?" are two of the questions the agency plans to ask itself, Whetstone said. FDA also will look at who internally should be responsible for making a recall termination decision, which currently can be done by either the field or the centers.

Finally, the agency plans to assess its current approach to auditing the effectiveness of a company's recall actions, looking at such issues as how much auditing is necessary and what form those audits should take. The agency is expecting its recall reengineering efforts to result in increased efficiency, less duplication of effort, and a greater uniformity agency wide in classifying and processing recalls, Whetstone explained. In addition, she hopes that the added transparency in the process will "have a positive impact" on industry determinations of what and when to report. CDER's compliance office is intending to start outlining case histories of significant recalls for inclusion in the "cGMP Notes" found at the CDER website. Compliance officials believe that such case studies could help firms avoid manufacturing and compliance complications that often result in recalls and enforcement actions. While it is expected that firms would not share compliance-related information with each other, FDA has found that subsidiaries of large pharmaceutical companies are not communicating.

This lack of communication may have played a role in the multi-facility problems that companies like Wyeth and Schein have experienced over the past several years.