FDA Focuses Latest Warnings on Accountability and Root Cause

Quest for accountability and root cause seen in latest FDA warning letters. FDA, Congress pressure McNeil, J&J. 'Phantom recall' debated. Is propofol too hard for Teva, Hospira to make? Perrigo raked over tablet mix-ups. Braintree learns to fear spiders. Merck unit goes OOS. Splitting hairs over Shamrock's tablet splitting? Pierre Fabre: Three batches and you're out?

A common theme running through many of the warning letters and inspection reports that FDA unveiled last month was an emphasis on corporate accountability.

Another was a growing sense of frustration over a continuing inability of many drug product manufacturers to realize why certain types of process failures were persisting and how to put a stop to them.

Many of the letters focus on the need to identify and correct the root causes of manufacturing quality problems.

McNeil, J&J get lesson in corporate accountability

FDA used last month's highest profile case - that of McNeil Consumer Healthcare - to underscore corporate accountability for drug manufacturing quality.

The agency has grown increasingly concerned about the company's overall state of compliance due to a series of inspection findings and recalls at two McNeil facilities over the past year.

In early 2010, alarmed by a report that a six-year-old child died after taking several McNeil products, FDA conducted focused inspections at the Fort Washington, Pa., and Las Piedras, Puerto Rico, facilities. FDA tested the products the child used and found nothing connecting them to the child's death.

Nevertheless, the agency's rising anxiety led it to call a meeting with corporate management in February.

At the Feb. 19 meeting, senior compliance staff with FDA's center for drugs and its field organization told the president of McNeil, a high-level corporate official from the parent company, Johnson & Johnson, and quality assurance executives of both companies, to step it up on compliance, which they assured the agency they were doing.

FDA followed that up in April by inspecting the Fort Washington facility, even though it wasn't due for another biennial surveillance inspection until June 2011.

That inspection triggered a plant shutdown and a major recall on April 30 of 6 million bottles of 40 types of oral liquid medicines for children.

The House Committee on Oversight and Government Reform launched an investigation of the recall on May 5. And on May 27, the committee held a hearing featuring FDA's principal deputy commissioner, Joshua Sharfstein, and the head of J&J's consumer group, Colleen Goggins.

In his testimony, Sharfstein said that based on the pattern of problems, FDA is working with McNeil "to address its systemic quality issues." It is monitoring implementation of a corrective action plan that includes quality system enhancements, organizational changes and senior management oversight.

Also FDA told the news media in a later update that it was conducting a companywide investigation of McNeil's drug manufacturing practices to see if similar problems exist throughout the company.

At the hearing, Sharfstein had a message for other manufacturers. He said the agency "will focus on entire companies and their systems in addition to focusing on specific violations, individuals and sites, much as we are doing in the McNeil situation."

The committee's chairman, Edolphus Towns, D-N.Y., told the hearing he wants not only to give FDA authority to mandate recalls, but also to empower the agency to order a halt in drug production.

'Phantom recall' drew fire

A highlight of the hearing was what Towns described as a "phantom recall."

The committee shared a document obtained from FDA that contained instructions for employees of San Diego-based WIS International to carry out a Motrin purchase project last June for a reverse logistics unit of Inmar Inc., Winston-Salem, N.C., on behalf of J&J.

Each worker was to buy all Motrin IB caplet eight-count vials in specific stores. "You should simply 'act' like a regular customer while making these purchases," the memo said. "THERE MUST BE NO MENTION OF THIS BEING A RECALL OF THE PRODUCT! If asked, simply state that your employer is checking the distribution chain of this product and needs to have some of it purchased for the project."

Since the hearing, there has been some back and forth between the committee and J&J concerning what the company knew about this activity and what its intentions were.

In any case, McNeil recalled the coated Motrin IB ibuprofen tablets last November, apparently at FDA's urging. FDA enforcement reports show that McNeil made the tablets at its Las Piedras facility, and recalled them due to dissolution failures seen during three-month stability testing.

An FDA inspection of the Las Piedras plant that was ongoing at that time attributed the dissolution failures to a pair of electrical failures and a broken bag incident during granulation of those lots, which FDA said also affected a third lot that McNeil did not recall.

Particles, organisms and invalidated processes

FDA's April 19-30 inspection of McNeil's Fort Washington facility raised several issues, according to a redacted Form 483 report and Sharfstein's congressional testimony.

The report faulted McNeil for failing to take corrective action after observing particulates in 26 batches since May 2009 and receiving some 46 consumer complaints about them since June 2009.

McNeil has described the particulates as fine, black metal particles.

Also, given concerns raised by various departments, FDA said McNeil's quality unit should not have let the company use a certain contaminated raw material to make various oral liquid Tylenol drug products for children and infants.

The company had recalled 8 million bottles of the products in August 2009 after learning that the raw material, microcrystalline cellulose, was contaminated with B. cepacia, a gram negative microbe that can pose a risk to immune-compromised people such as those who have cystic fibrosis (¹ (Also see "How FDA and Industry Aim to Find and Eliminate Drug Recalls' Root Causes" - Pink Sheet, 1 Oct, 2009.)).

However, there were some lots of the raw material that tested positive for the gram negative organism, but that were used anyway to make 27 lots of liquid Tylenol products that remained on the market, FDA learned during the inspection.

Although those products tested negative for B. cepacia, there was still a risk of contamination that the quality unit should have been able to control, FDA said.

There was also a problem with process validation after a manufacturing change for cherry-flavored Tylenol suspension drops for infants.

McNeil did not adjust the duration and speed of agitation for a hold tank after increasing the batch size relative to tank size, FDA said. After the agitator shut off, acetaminophen active ingredient became concentrated, resulting in end-of-run superpotent batch assay failures, the rather heavily redacted Form 483 report indicated.

Also, workers were only pulling stability samples from the beginning of packaging runs, when they should also have been sampling the middle and the end, so they did not have samples from the superpotent portion of the three affected batches.

There was also a batch of the suspension drops that was subpotent because the operator used a drum that was not full. A second operator should have noticed the drum was underweight but instead reported in the batch record that it was the correct weight.

Earlier complaints and root cause investigations

An FDA inspection of the Las Piedras plant that ran from Oct. 22, 2009, to Jan. 8, 2010, focused mainly on a problem that McNeil eventually traced to contamination of bottles by a degradant of a wood preservative used to treat storage pallets (² (Also see "Data Integrity Problems Continued to Surface in Recent Warning Letters" - Pink Sheet, 1 Feb, 2010.)).

The bottle contamination problem occurred when heat and humidity broke down the 2,4,6-tribromophenol preservative and fungicide, which is used in some countries including Brazil, where the pallets were made.

McNeil was slow to react due in part to its complaint-trend analysis process. It wasn't until August 2008 that the company realized complaints related to two lots of Tylenol arthritis caplets had increased three months earlier.

Also, the company's first root cause investigation ended in a blind alley. The company had assumed the problem was contamination from mold or fungi. But microbiological testing came back negative and anyway, the complaint rate dropped that fall.

The problem resurfaced the next summer, although once again, months went by before McNeil identified the complaint trend.

Once it noticed the complaints, McNeil began a new investigation, this time more extensive and broad ranging. The company saw that the five lots that generated the most complaints over the two-year period shared common shipments of bottles. After retaining an outside expert and Microanalytics, a contract laboratory with unique capabilities, McNeil was able to identify the contaminant, which was present at mere part-per-trillion levels.

McNeil recalled not only the Tylenol arthritis caplets but also a number of other products that might have been affected by the bottle contamination issue.

The inspectors found other problems as well. Unknown peaks in Motrin IB assays were attributed to acetaminophen contamination of disposable pipettes used in the lab for vial filling.

Also, the facility had reported 12 mix-ups of drug products in the packaging line over the year leading up to the inspection, as well as 123 mix-up-related consumer complaints - all of which it classified as isolated occurrences, because each one involved a different lot.

FDA also faulted McNeil for not signing off on the work of Microanalytics, the lab that identified the pallet degradant, and for not validating the test method Microanlytics used.

Problems with propofol

FDA issued two warning letters related to propofol manufacturing, one last month to Hospira and one in April to Teva.

Both manufacturers have expressed frustration over the inherent difficulty of manufacturing the emulsion product. It's so hard to make the anesthetic and the profit to be made from it is so small that Teva has decided to drop it from its portfolio.

Hospira plans to continue making propofol, although it has suspended shipments while it obtains FDA approval for recent manufacturing improvements, including the use of a smaller micron filter to remove microscopic particulates.

Speaking during an April 27 earnings call, Hospira CEO Chris Begley assured investors that he takes FDA's warning letter seriously and is working with FDA and third-party consultants to resolve the issues the agency raised.

Hospira battles stainless steel particles

FDA got after Hospira in an April 12 warning letter about following up on particulate contamination of propofol, Liposyn and Cleviprex and related corrective actions.

There has been a persistent and serious particulate contamination problem at Hospira's Clayton, N.C., facility for several years, the agency said in the letter, which it posted online in May. The problem mainly involves stainless steel particles.

FDA said Hospira did not realize 16 lots of propofol and Liposyn that it made in 2007 were visibly contaminated until it belatedly performed a second annual retain inspection in November 2009, after the lots had expired.

Hospira has since conducted three major recalls totaling 78 lots of propofol, 121 lots of Liposyn and 24 lots of Cleviprex, all made at the Clayton facility between January 2008 and February 2010.

In response to FDA's Form 483 report of inspectional observations made during a Jan. 26 to Feb. 23 inspection of the Clayton facility, Hospira on March 26 said it would enhance particulate monitoring and complete revalidation activities. But this was not enough, FDA warned, "because it is unclear if your firm has determined the root cause of the problem and resolved it."

Hospira should have been quicker to verify the adequacy of corrective actions implemented in October 2009 when it resumed production after the three major recalls, FDA said.

As it was, Hospira did not discover that particulates were still a problem despite the corrective actions until three months later when inspecting retain samples Jan. 21. Hospira's response to that finding drew further criticism from the agency for its slow pace.

The warning letter also raised some validation-related issues from a Jan. 12-19 inspection of Hospira's Rocky Mount, N.C., facility.

FDA said Hospira had not validated the mixing processes involved in its mixing tank validation matrix for various products.

The agency had been raising concerns about process validations at the plant since 2005. It was still anxiously awaiting Hospira's validation prioritization plan and requested the company to at least provide an interim plan.

Teva's response to propofol 483 fell short

Meanwhile, FDA in April posted a warning letter it sent Dec. 11, 2009, to Teva Parenteral Medicines, Irvine, Calif., mainly regarding problems surrounding bacterial endotoxin contamination of propofol that came to the agency's attention during a July 13-24, 2009, inspection.

FDA raised its concerns in a Form 483 report that it redacted and posted online just a week after the inspection ended (³ 'The Gold Sheet,' September 2009).

Much of the warning letter was focused on Teva's initial response, which came Aug. 10, just 11 working days after the inspection ended.

Although FDA accepted some aspects of Teva's response, it found others to be inadequate.

In the response, Teva said it tightened its final product release specification for endotoxin levels and committed to increased in-process and finished product sampling and testing.

However, FDA was not satisfied with Teva's plan for tougher testing. By themselves, tighter specifications are not enough, the agency said. The company also must meet cGMP requirements to "implement adequate manufacturing practices and controls to prevent bacterial endotoxin contamination."

Teva had been sampling 13 vials of every lot, testing three for bacterial endotoxin and 10 for bioburden. FDA was concerned that this protocol never varied, regardless of lot size, even though some lots were much bigger than others.

Teva's Aug. 10 plan to increase sampling was not good enough for FDA. The plan did not reference "basic elements of statistical analysis," the agency said. It also lacked "a scientific rationale for sampling that would vary the amount of samples taken according to the lot size," and it did not establish a statistical confidence limit.

The 483 report had noted a similar issue with equipment cleaning - Teva was not making sure that cleaning actually reduced bioburden and endotoxin to acceptable levels.

In its Aug. 10 response, Teva said it would solve that problem by flushing the water for injection ports for a set period of time and by washing the reverse osmosis ports, adding that both corrective actions have historically reduced bacterial endotoxin levels. But FDA wanted to see validation studies supporting this contention.

The agency added that the continued recurrence of endotoxins "indicates that the WFO and RO water systems are not in a state of control."

FDA had similar concerns about Teva's plans for improving the cleaning of carts used to carry vial stoppers and for validating its procedures for fogging various areas to sanitize them.

Equipment maintenance was also an issue in the warning letter. Teva should have conducted a root cause investigation into three filter failures in one room, according to the FDA inspection report. Teva's Aug. 10 response was inadequate, FDA said in the warning letter. The company just added instructions about determining the location of the leak and inspecting for physical damage rather than determining the root cause of the failures, and preventing recurrences.

Tablet mix-ups spurred Perrigo warning

The focus of an April 29 warning letter to L. Perrigo Co., Allegan, Mich., differed somewhat from that of the Form 483 report of inspectional observations upon which it was based.

The warning letter, posted online May 25, highlighted line clearance issues reminiscent of those that came up last year with another Michigan drug maker, Detroit-based Caraco Pharmaceuticals Laboratories Ltd. (⁴ (Also see "Hamburg Promises Quicker Enforcement as FDA Spews Warning Letters" - Pink Sheet, 1 Aug, 2009.)).

However, it passed over tablet-thickness issues raised in the 483 report, posted online June 7, similar to those that bedeviled not only Caraco, but also KV Pharmaceuticals of St. Louis, and Actavis Totowa, of Totowa, N.J. (⁵ (Also see "Drug Recall Totals for 2008 Highlight Global Outsourcing and Other Risks" - Pink Sheet, 1 Mar, 2009.)).

The 483 report detailed three instances where Perrigo found oversized tablets multiple times during packaging of ibuprofen lots.

In each case, the first finding led the company to run a sample and check it for overlarge tablets. When none were found, it would resume packaging operations.

In one case, the second finding of an overlarge tablet led to a second sampling, which was also negative. In that case, oversized tablets were found a third time, and that led to a third sampling, which found two oversized tablets and led Perrigo to acknowledge a deviation.

According to the report, there had been 50 instances of oversized tablets over the previous year that the quality assurance unit had ranked as low risk, as well as some higher risk incidents.

The line clearance issues highlighted in the warning letter revolved around the filling operation. When Perrigo found a round brown ibuprofen tablet mixed with oval brown caplets, FDA said it should have checked the lot of round orange tablets it had packaged after the round brown tablets and before the oval brown caplets, to see if any round brown ones had gotten mixed with them.

The warning letter also highlighted packaging problems with Milk of Magnesia. There were two cases late last year in which Perrigo had to recall mislabeled Milk of Magnesia.

The warning letter also noted Perrigo's March 2009 recall of ibuprofen tablets due to contamination with metal shavings. Perrigo segregated the contaminated lot, but subsequently released and shipped a portion of it, forcing the recall.

FDA concluded the warning letter with a bit of a scolding: "Your firm has had an ongoing program since 2005 to address mix-ups. However, your firm continues to receive complaints regarding this issue, and despite past assurances that previous enhancements would control this problem, deviations continue. Your firm failed to implement sustainable corrective actions to prevent mix-ups as well as other continuing problems. As a result, we have concerns regarding the failure of your firm, including the QCU, to act proactively to ensure compliance with SOPs and the cGMP regulations."

FDA went on to advise Perrigo that it expects "that your corporate management will undertake a comprehensive evaluation of manufacturing operations to ensure compliance with cGMP."

Braintree ignored insects and spiders

Batch failure investigations were the focus of a May 10 warning letter to Braintree Laboratories, Braintree, Mass., that FDA posted on May 18.

FDA was not impressed with the way Braintree Laboratories handled 21 consumer complaints received in 2008 and 2009 concerning insects, insect parts and spiders in five finished drug products.

The agency learned during a Dec. 3, 2009, to Jan. 22, 2010, inspection that Braintree responded to the complaints simply by concluding that its manufacturing practices were not at fault. Instead there should have been a thorough investigation to find the root cause, FDA said.

Braintree asserted in a Feb. 10 response to the Form 483 report of FDA's inspectional observations that the insects and spiders "do not present inherent health hazard."

FDA countered in the warning letter that it viewed the contamination as "a significant product quality issue."

In addition to the corrective actions Braintree agreed to take, FDA suggested that it may need to audit suppliers of raw materials and packaging components.

There were also problems with dissolution failures of calcium acetate gelcaps that Braintree appeared to dismiss without appropriately blaming on testing problems.

In one case, Braintree got around dissolution failures of one lot's accelerated stability samples at the six-month time point by removing caplets from gelcaps and getting a passing result with the removed caplets. FDA said the entire marketed gelcap had to pass.

Braintree also continued to distribute another lot even though accelerated stability studies showed dissolution failures at the 12-month and 15-month time points. FDA wanted to know its rationale for keeping that lot on the market.

It was unclear to FDA whether Braintree had found and addressed the root cause of the dissolution failures. The agency was also concerned that the company did not properly understand its expectations and requirements for out-of-specification investigations, and asked Braintree to review its October 2006 OOS guidance document.

Merck's Intervet failed to report OOS finding

Merck's Intervet Inc. unit in Unterschleissheim, Germany, got into trouble with FDA's Center for Veterinary Medicine over a failure to promptly report out-of-specification test results, according to a March 12 warning letter that FDA posted May 4.

The reporting requirement at 21 CFR 514.80(b) was triggered when analytical testing showed that several batches of Vetsulin (porcine insulin zinc suspension) would not meet stability specifications for non-extractable insulin before the expiration date.

At that point, Intervet had three days to transmit a field alert report to FDA regarding the out-of-specification results.

FDA went on to remind Intervet that it holds the company responsible for investigating and determining the causes of the OOS results, preventing their recurrence, and preventing the occurrence of other violations.

The agency warned Intervet that if it didn't correct the violation, FDA might take legal action, such as seizure and injunction.

Shamrock's tablet-splitting process criticized

Tablet-splitting was the main focus of FDA's April 8 warning letter to Shamrock Medical Solutions Group of Lewis Center, Ohio.

FDA inspected Shamrock's drug repackaging and relabeling facility in East Taunton, Mass., five times last year - twice in July and with one-day follow-ups in August, September and October.

The biggest worry was that the tablets Shamrock halved or quartered may have lost their identity, were likely of uncertain, and in cases unapproved, strengths, and had unknown stability profiles.

Of special concern to FDA was Shamrock's splitting of tablets for drug products that had narrow therapeutic ranges and that required monitoring of patient blood concentrations.

There should have been validation studies showing the split tablets met label specifications, FDA said. Shamrock also should have tested their stability to support expiration dating instead of simply assigning 12-month expiration dates to all repackaged products, including the split tablets. FDA noted that its Compliance Guide 480.200 allows repackagers a maximum of six months for expiration dating without supporting stability studies.

FDA cited Shamrock quality committee minutes

FDA took the unusual tack in the Shamrock warning letter of quoting from meeting minutes of the company's quality committee.

The agency supported an equipment cleaning and maintenance finding by quoting from the minutes of an Aug. 18, 2008, meeting during which the manufacturer of disposable silicone tubing sets advised against cleaning with isopropyl alcohol because it could "'breakdown [sic] the silicone parts of the tubing.'"

Shamrock nevertheless used isopropyl alcohol and distilled water to clean the tubing sets before reusing them to fill liquid oral drug products.

A search of FDA's drug GMP warning letters turned up only one other instance where an internal committee was even mentioned: A Feb. 15, 2005, letter to Medsep Corp. discussed a finding by Medsep's Sterilization Review Committee.

There was also a blood GMP warning letter dated July 17, 2003, that mentioned the minutes of the South Texas Blood and Tissue Center's quality review board, but only to substantiate how infrequently it met.

In the past, FDA has resisted the urge to review internal audits, and drug manufacturing quality managers have sought to shield management review meeting minutes from FDA inspectors ("The Gold Sheet," ⁶ (Also see "Quality Systems No Longer Theoretical as Industry Begins Implementing Concepts" - Pink Sheet, 1 Jan, 2008.) and ⁷ (Also see "Ranbaxy Internal Audit Disclosure Highlights FDA Policy Exceptions" - Pink Sheet, 1 Aug, 2008.)).

The concern has been that if employees thought FDA would be reviewing the audits or meeting minutes, they would be careful what they said, and this could thwart the internal communication needed to identify and address quality issues.

FDA exhorted Shamrock in the warning letter to adopt a global quality systems approach to GMP compliance. The agency noted that many of the issues it raised in the warning letter had already been cited during an inspection in 2007 of the company's Lewis Center, Ohio, facility. The company's corporate quality committee should have fixed those problems at both facilities, the agency said.

Pierre Fabre must start over on process validation

FDA wants Pierre Fabre Medicament Production of Idron, France, to find out what went wrong with a three-batch process validation study and do a second study before the agency will approve it as a new contract manufacturer for Eisai's Aloxi (palonostron).

The first validation batch failed to meet yield limits, FDA said in a March 26 warning letter that resulted from a Sept. 1-8, 2009, pre-approval inspection.

The company attributed the problem to a high volume of particles. However, FDA said, "The root cause for the increased number of particles in the product remains unclear."

With the root cause unidentified and unresolved, the second and third batches also had out-of-limit yields.

There were also out-of-limit fill weights in all three validation batches - before and after Pierre Fabre changed the fill weight specification. The company never investigated why.

In a Sept. 29 response to the Form 483 report of inspectional observations, Pierre Fabre told FDA that it planned to manufacture a "mock batch" to demonstrate validation so it could release the three validation batches to the market.

"We disagree that performing validation studies of an additional batch is sufficient to show that the process is validated," FDA said. "Provide us the root cause analysis of this deviation and any implemented corrective actions, and address our aforementioned concerns."

Because its validation studies showed a lack of reproducibility and failure to consistently manufacture Aloxi, FDA said Pierre Fabre should "conduct a thorough evaluation of your initial Aloxi validation studies and execute a second process validation study to ensure validation of the manufacturing process."

- Bowman Cox (⁸ [email protected])