FDA Sending More Drug GMP Warning Letters to Foreign Sites

FDA warning letters are up at foreign sites and down at domestic sites as agency ramps up presence abroad. Inadequate OOS investigations and faulty testing of drug components were top violations; contamination was a common issue; firms cited included Novartis, Mylan, Smithkline Beecham, Lonza and Noven.

FDA is showing its enforcement muscle overseas with an increasing number of drug GMP warning letters going to facilities abroad. Of the 18 drug GMP letters issued in the second part of calendar year 2011, 10 went to foreign drug and API manufacturers.

These included three to China, two to Germany, one to India, one to the UK, one to Mexico, one to Poland, and one to Switzerland. Yet the number of drug GMP warning letters has decreased slightly from calendar year 2010, from 50 to 40.

Of the 40 drug GMP letters issued in calendar year 2011, 20 or half went to foreign drug and API manufacturers (see chart).

Thanks to budget increases, FDA has increased its inspectional presence overseas to meet the challenges of globalization, has opened 11 international offices and has hired international inspectors to staff these offices. The total number of foreign and domestic inspections for all FDA-regulated products increased from 16,236 in fiscal year 2009 to 18,109 in FY 2010 (Also see "FDA Continues Aggressive Enforcement as Drug GMP Warning Letters Mount" - Pink Sheet, 1 Apr, 2011.).

This translates into more inspectors conducting international inspections, which is likely to result in more warning letters.

Top citations not much changed

The most frequently cited GMP provision in warning letters that FDA issued in the past six months is deficient out-of-specification investigations, but moving up a bit in terms of top GMP failings are inadequate testing of drug components, including identity testing for contaminants.

Also, data integrity still remains an issue. One firm was chastised for backdating records, API manufacturer Yag Mag Labs in Hyderabad, India, and another for forging signatures, Deibel Laboratory in Lincolnwood, Ill.

The other prominent failures noted were inadequate control of the microbiological laboratory, and the lack of a strong management role in quality.

Other prominent findings were inadequate written procedures, inadequate control of automatic, mechanical and electronic equipment and inadequate laboratory controls.

In four warning letters, manufactures were requested to notify the agency if drug shortages would result from the warning letters while manufacturers fix problems, reflecting the agency’s new concern on the availability of crucial and life-saving drugs to patients. These requests went to International Isotopes, Novartis, Luitpold and Lonza Biologics.

About one-third of the drug GMP letters were issued to API manufacturers, signifying the continued importance that FDA places on the pharmaceutical supply chain.

Here is a breakdown of the recipients of the warning letters:

• six went to active pharmaceutical ingredient manufacturers, five of them foreign manufacturers;
• five went to injectable manufacturers;
• three went to contract manufacturers;
• one to a radiopharmaceutical manufacturer;
• one to a transdermal patch manufacturer; and
• two to solid oral dose and liquid manufacturers.

Yag-Mag backdated records

A Sept. 12 warning letter to API manufacturer Yag-Mag Labs in Hyderabad, India, chastised the firm for backdating records, and identified major problems with the manufacturing facility itself as well as inadequate cleaning. The firm’s products remain on an import alert until these problems are resolved.

This backdating of records occurred from April 2009 through March 2010. FDA noted that “batch records were admittedly written months after the batches were manufactured and released for shipment.”

FDA reminded the company that “batch records must be completed at the time of batch manufacture, and backdating of records is not an acceptable practice. This finding that manufacturing batch records are untrustworthy represents a basic systemic failure by your company.”

The agency also observed residues and corrosion on processing equipment, materials and processing equipment exposed to outside elements; poorly identified and leaking pipes; recessed ground-level floors susceptible to flooding and observed with substantial standing water; and inadequate, unsanitary restroom facilities.

FDA further noted that the company did not clean its facility on a routine basis. “There was no evidence that any cleaning between batches or between product changeovers occurred for non-dedicated equipment. It is a critical responsibility of your firm to clean non-dedicated equipment between production of different materials to prevent cross-contamination, using validated procedures.”

The agency recommended that the firm hire a third party auditor with experience in detecting data integrity problems. “Your firm’s broad-based cGMP compliance issues will require significant remediation and substantial investment of time and resources.”

Luitpold had black particle problem

An Aug. 31 warning letter to parenteral drug maker Luitpold Pharmaceuticals Inc. of Shirley, N.Y., found that the firm failed to conduct a thorough investigation for black particles found in potassium phosphate product. The warning letter was sent following a Feb. 9 to March 15 inspection.

FDA said that the failure to investigate cases of particle contamination and product failures are repeat violations from October 2010, October 2009 and November 2008 inspections.

“Without adequate justification, your firm’s Deviation Report 2457 stated the root cause was glass delamination resulting from a manufacturing variation in the 5 ml [redacted] glass. … Your firm failed to extend the investigation to other unexpired lots of potassium phosphate made with the same [redacted] glass.”

Luitpold recalled three products last year because of manufacturing problems, including multitrace-5 concentrate for high aluminum levels detected in products; ketorolac tromethamine injection for particles found in product; and magnesium sulfate for injection for discolorations (Also see "Drug Product Recalls In 2010: Categorized By Problem Area" - Pink Sheet, 1 May, 2011.).

The firm’s failure to resolve particulate contamination has been a “persistent and serious issue at your firm for many years,” said FDA.

The firm was also criticized for failing to routinely calibrate, inspect, or check automatic equipment in accordance with an adequately written program,

Jelfa cited for inadequate micro controls

A July 14 warning letter to injectable manufacturer Jelfa Pharmaceuticals in Jelenia Gora, Poland, focused on the firm’s inadequate investigation of batch failures and inadequate controls for preventing microbiological contamination. The company is a subsidiary of Valeant Pharmaceuticals International Inc.

FDA investigators said that a batch failed a sterility test and instead of initiating an investigation, the firm repeated the test on a new sample to confirm the original result prior to initiating an investigation.

“Please note,” said FDA, “that when microbial growth is observed, a lot should be considered non-sterile and an investigation conducted. An initial positive test would be invalid only in an instance in which microbial growth can be unequivocally ascribed to laboratory error.”

FDA advised the manufacturer to read the agency’s October 2006 guidance document, “Investigating Out-of-Specification Test Results for Pharmaceutical Production.”

The company was also criticized for having inadequate controls for preventing microbiological contamination.

For example during the aseptic filling of two injection batches on the filling line, employees were observed following poor aseptic techniques.

“Specifically movement inside the Class A area were not slow and deliberate; operators and an engineer were observed with exposed facial skin during the filling operation; and a forceps was observed in a Class B (ISO 6) area and was then used to remove fallen ampoules from the aseptic processing line in the Class A (ISO 5) areas.”

Also the tubing ends used to connect the solution tanks to the filling line were not protected prior to sterilization and disinfectant efficacy studies were not completed for three of the disinfectants used to sanitize surfaces in the sterility testing suite and production aseptic core filling line.

These violations were similar to others FDA noted in February 2007 and July 2008 inspections.

“We are particularly concerned with your firm’s failure to implement a robust quality system,” FDA said. “Repeated citations from prior inspections indicate that your quality control unit is not exercising its responsibilities, and may not have the appropriate authority to carry out its responsibilities.”

FDA recommended that the firm hire a third party consultant with GMP expertise to assess the firm’s facility, procedures, processes, and systems to help resolve these issues.

Sichuan risked cross-contamination

API maker Sichuan Pharmaceutical Co. of China drew the attention of FDA in a Sept. 9 warning letter for inadequately assessing cross-contamination risks between adjacent workshops.

FDA inspected the firm’s manufacturing facility located at Pengzhou, a city near Chengdu, June 23 to 29, 2010, and FDA said the company’s Aug. 5, 2010, and Dec. 13, 2010, responses lacked sufficient corrective actions.

From September 2008 to July 2009 the firm manufactured an API which was adjacent to a workshop producing a separate API. “However, you failed to have adequate controls and monitoring program to prevent cross-contamination between these adjacent workshops.”

In addition, the firm manufactures an API in a facility that was previously used to manufacture another API without conducting an assessment of the cross-contamination risks.

The agency further noted that analytical testing of a product for possible contamination “is not sufficient to ensure adequate conditions.”

FDA first approved Sichuan Pharma to export APIs in 1992 for the antibiotic tetracycline.

Zhejiang Casing did not review records

An Aug. 5 warning letter to API manufacturer Zhejiang Casing Animal By-Products in Tonglu, China, focused on the failure of the quality unit to review lab records before releasing API for distribution. The firm contracts with a firm to ensure that the product is free of contamination through proton nuclear magnetic resonance (HNMR) testing.

The quality unit, for example, failed to review HNMR laboratory control records prior to releasing the API for distribution. Another problem was that the raw data to support the release of the lots to the U.S. distributor were not available for review from the contract laboratory during the inspection.

Also, FDA said that “your contract testing laboratory for HNMR only provides you with partial spectra zoomed in on the range for [redacted] and [redacted]. We are concerned that your quality unit is not reviewing the entire spectrum for other potential impurities or contaminants.”

The firm was also cited for failing to have a system in place to evaluate suppliers and to adequately test each batch of incoming materials intended to be used in drug production. Further, the firm does not perform identity tests on each batch of incoming raw material prior to release of the materials for production.

Nanjing Maohai lacked oversight

As in the Sichuan warning letter, FDA criticized API maker Nanjing Maohai of Nanjing, China, for failing to exercise oversight over its contract laboratory in reviewing batch production and lab records prior to releasing product. The warning letter was issued on Aug. 5 following an Oct. 25-28, 2010, inspection.

As a result, the firm lacked the raw data to demonstrate that its API shipped to the U.S. met acceptance criteria.

“Your firm sends samples of [redacted] produced by your firm, to your contract laboratory [redacted] for [redacted] testing. The material is released without adequate review of laboratory control records. … In addition, your contract testing laboratory [redacted] for proton nuclear magnetic resonance (HNMR) only provides you with a partial spectra zoomed in on the range for [redacted] and [redacted]. It is important for your firm to review the entire spectrum for other potential impurities or contaminants.”

The company was also criticized for failing to investigate complaints. For example, complaints for four lots of product were not properly documented or investigated.

The inspection also documented another complaint regarding an unknown impurity found in a lot when tested using HNMR. No complaint investigation was conducted regarding this unknown peak.

The company was also criticized for failing to have a system in place to evaluate suppliers and to adequately test each batch of incoming materials. “Your firm currently accepts certificates of analysis (COA) from unevaluated suppliers in lieu of testing each batch of an incoming component to ensure conformity with the established specifications.”

Noven faulted for process validation

Noven Pharmaceuticals Inc., a manufacturer of the fentanyl patch, based in Carlsbad, Calif., drew FDA’s attention in an Aug. 25 warning letter for faulty validation practices. For example, the final process validation report failed to include and evaluate the impact of brown particles found in a sample of fentanyl solution. The firm was inspected from Jan. 31 to Feb. 14.

The warning letter states that “your final process validation report failed to include and evaluate the impact of all the combined deviations that occurred during process validation.”

The report, for example, failed to include such deviations as the discovery of brown particles in a laboratory sample of fentanyl adhesive mass solution (FAMS), and aborting the cutting/packaging operations during the first process validation lot due to a broken cutting and packing machine.

The firm also failed to adequately validate the FAMS mixing process to assure blend uniformity after a modification was made to raw material specification. The firm, for example, conducted a revalidation of the FAMS mixing process but failed to include the fentanyl API in the validation.

“In your response, your firm states that the API is substituted with a placebo to conduct feasibility or investigative studies for troubleshooting purposes, and that you believed the process (without fentanyl) would be appropriate in this instance. Your response, however, is not adequate because you have yet to provide any evidence to support your claims that the studies, including the use of a placebo, demonstrate robustness of the actual mixing process.”

The company was also criticized for not having a strong enough sampling plan. “For each product size of 25/50/75/100 mcg/hr dosage strengths, only forty [redacted] samples were tested through the checkweigher to qualify the equipment as a subsystem. There is no statistical basis for the selected sampling size.”

Uriel had glass particles in ampoules

FDA on Aug. 18 sent a warning letter to Uriel Pharmacy of East Troy, Wis., for not having adequate controls in the microbiology laboratory to prevent contamination and for inadequate investigation of batch failures. An inspection was conducted from Oct. 18 to Nov. 1, 2010. The firm is a maker of over-the-counter liquids and solid oral dosage forms.

FDA said the firm failed to validate cycles to demonstrate their effectiveness at rendering products sterile. In addition, the firm did not document the parameters such as time, temperature and pressure used to sterilize the products.

The firm was also criticized for using distilled water that does not meet the specifications of water for injection to manufacture sterile injectables and sterile eye drops.

The firm also failed to validate the use of a sterilizing filter with its products and processes and to sterilize drug product containers and closures.

FDA said, “Your firm has failed to validate the depyrogenation process (i.e. heating to 350 degrees Fahrenheit in a convection oven) to demonstrate that any glassware used in the manufacture of your sterile injectables are rendered non-pyrogenic.”

In response, company officials said that they were using pyrogen-free containers and closures, Yet FDA was not satisfied. “At this time, we cannot determine the adequacy of your response because your response is unclear. Please clarify whether you are purchasing pyrogen-free materials and/or depyrogenating materials in-house. Also, please provide details regarding the depyrogenation cycle and other controls utilized to prevent contribution of endotoxin load.”

The firm also failed to launch an investigation for visible particulates found in ampoules until Oct. 27, 2010, after the start of the inspection, even though the firm identified that there was a problem with particulates on Oct. 15, 2010.

“In your response, your firm states that the inspection delayed the documentation of the event and implementation of the action plan. In addition, your firm was in the process of hiring a QA director with the responsibility of responding to discovered manufacturing errors. Your response, however, is inadequate because your firm has failed to ensure, as described under 21 CFR 211.25(c), that your firm has an adequate number of qualified personnel to perform daily operations.”

On Oct. 21, 2010, the company announced a recall of specific ampoules due to the presence of foreign substances. Glass particles were found in some ampoules manufactured from Sept. 1 through Oct. 15, 2010. ( (Also see "FDA Recalls -- March 16, 2011" - Pink Sheet, 21 Mar, 2011.)

Deibel forged signatures

FDA on Sept. 1 issued a warning letter to contract testing laboratory Deibel Laboratory in Lincolnwood, Ill., for inadequate testing of batches to ensure that they are free of objectionable microorganisms. The firm also had inadequate laboratory practices, including having forged signatures on laboratory worksheets.

In terms of the first finding, FDA found inadequate controls in the microbiology lab. For example, there were drug products tested without showing evidence of microorganism detection. These included the following products: 0.12 percent chlorhexidine gluconate prescription oral rinse, levothyroxine sodium raw material, 1.5 percent chlorhexidine gluconate scrub, 1 percent furosemide syrup, benzalkonium chloride solution, piperazine adiptate paste and gentamicin sulfate USP raw material.

“In your response, your firm states that Standard Operating Procedures DLI-QM-SOP-034 will be updated to include a section on validations in which clients will be asked if they would like validation of the methods performed for the formulation. Your response is not adequate because it does not indicate whether you intend to validate all of the microbiological methods for each of the marketed and tested drug products. For commercially marketed drug products, you should determine if the laboratory method used is actually validated, reliable and adequate to support microbial growth. It should not be left to the client’s discretion.”

FDA also noted that the firm had inadequate laboratory practices, including forging signatures in laboratory notebooks.

The agency said that “it is common practice in your firm to have another employee record the test results or the initials and signatures of the analyst who actually performed the testing on the raw data worksheets. Your firm’s management was aware of this practice.”

Although the firm now has now revised its procedures by stating that writing another analyst’s initials or signature is considered “forgery,” FDA said that this response is inadequate.

“There is no evidence to show that the revisions will be effectively implemented. In addition, your revised procedure still allows for employees to enter data on each others’ paperwork, which is a practice that continues to undermine your data integrity and increases the possibilities of errors in the laboratory record.”

Lonza did not investigate failures

API oncology maker Lonza Biologics of Portsmouth, N.H., received a warning letter on Sept. 1, 2011, for inadequately investigating batch failures and for inadequate analytical methods.

FDA inspectors visited Lonza’s Hopkinton, Mass., plant from April 4 to May 6, 2011.

The firm failed to conduct an investigation for ONTAK after an unidentified peak was observed co-eluting with the primary product peak, FDA said. There were also inadequate investigations of critical deviations after a number of ONTAK drug substance lots failed to meet release specifications.

The firm also failed to validate analytical test methods used for potency testing. For example the firm failed to validate and quantify chromatographic peaks for potency and robustness.

“In your response, your firm states that the same RPLC column type has been used for 15 years for the [redacted] method. In the last two years, there has been a change in the resolution (i.e. column-to-column variability outside the previously observed range.”

The warning letter also raised concerns about the firm’s quality control unit. The letter noted that the firm issued an outdated Certificate of Analysis for ONTAK that was manufactured on Sept. 4, 2009, after receiving the DNA test results on March 3, 2011. The outdated Certificate of Analysis did not include a revised purity specification for the [redacted] as required by the BLA.

“In your response, you firm states that you agree that there was a failure to check the amended Certificate of Analysis against the current specification as required by your disposition procedure. The failure was due to human error and your firm will train appropriate staff on this aspect of the procedure. Your response, however is inadequate because there is no assurance that that other staff did not use the outdated Certificate of Analysis.”

Dental Technologies didn’t test for DEG

Dental Technologies, a contract manufacturer of prescription drugs based in Lincolnwood, Ill., came under fire in a Sept. 15 warning letter for inadequate testing of drug components and for faulty batch investigations.

For example, the company was not testing drugs to ensure that glycerin raw materials are free of diethylene gycol and ethylene glycol.

FDA warned that “diethylene glycol and ethylene glycol are both dangerous contaminants that have been found in glycerin raw materials.”

The firm has also not investigated the failure of a batch or any of its components to meet its specifications.

The firm’s investigation into microbial testing for the presence of pseudomonas aeruginosa identified in water samples collected on Dec. 20, 2010, and Dec. 27, 2010, was inadequate.

“Your investigation concluded that the microbial contamination occurred at the water delivery spigots. Your investigation however, failed to include information regarding the swab tests results of the spigots and the test results of the spigots following sanitation.”

FDA also noted that the firm failed to follow written procedures for handling written and oral complaints. The company evaluated patient complaints for nausea and vomiting and concluded that the patients were hypersensitive to fluoride.

Subsequently, your firm’s corrective action included discontinuance of the Cherry flavor 2 percent NaF Rinse product and the destruction of the remaining lots.

“Your firm’s investigations, however, failed to evaluate the manufacturing process, raw materials or packaging components that could have contributed to the patient reactions.”

SmithKline had faulty micro controls

An Oct. 7 warning letter to SmithKline Beecham targeted inadequate controls in the microbiology laboratory as well as faulty laboratory control procedures. Inspectors paid a visit to the Worthing, UK, site in March 2011.

FDA inspectors found that the firm had not established appropriate written procedures to prevent microbiological contamination.

“The qualification of your disinfectant [redacted] failed to demonstrate that it is suitable and effective to remove microorganisms from different surfaces. Specifically, this disinfectant failed to meet qualification criteria when challenged with multiple organisms.”

Further, a media fill conducted in January 2011 resulted in two contaminated units.

The firm, said FDA, attributed the contamination to stopper bags left inside the Class 100 area for a long period of time, through a shutdown that took place prior to the media fill in January 2011.

FDA said that the company had intended to use the media fill data to extend the sterility holding times for product contact components, without the approval of its quality unit.

The agency also found fault with the firm’s gown monitoring for aseptic processes.

The inspection documented that the firm conducts personnel monitoring for the classified manufacturing rooms by only sampling the hood, goggles and sleeves, and that more sampling is needed. “We are concerned about your current gowning monitoring approach as operators may perform substantial interventions into the Restricted Access Barriers (RABs), where sterile product is exposed, several times per week.”

FDA inspectors also said that the quality control unit does not adequately exercise its responsibilities to approve procedures or specifications that may impact the identity, strength, quality and purity of the drug product.

For example, the firm’s visual inspection certification program for the finished product does not “adequately challenge” the technician performing the inspection. The program only requires that one of the five critical defects be included in the challenge set.

Although the firm identifies five areas as critical defects: a vial with a cracked neck, a missing cap, missing stopper, high/low weight, and a foreign body, only a missing cap defect is included in the program.

Mylan failed to complete CU test

An Oct. 13, 2011, warning letter to generic manufacturer Mylan NV at its Caguas, Puerto Rico, site found inadequate testing of drugs and inadequate investigation of batch failures.

The firm’s Caguas, Puerto Rico, facility was inspected from Jan. 24 to Feb. 24, 2011.

The company failed to complete the content uniformity test required by the U.S. Pharmacopeia after a lot of loperamide hydrochloride capsules 2 mg failed to conform to the acceptance criteria.

“Your firm’s response failed to include the scientific rationale (e.g. statistical justification) for not completing the CU test,” said FDA.

The firm also did not thoroughly investigate batch failures. “In several investigations your firm attributed out-of-specification (OOS) results to assignable causes [redacted] but failed to include sufficient supporting evidence to invalidate the initial data.”

Jenahexal did not measure endotoxins

An Oct. 19 letter to Jenahexal Pharma focused on inadequate sterility testing. The Jena, Germany, facility was inspected Oct. 25-29, 2010. The company is a subsidiary of EverPharma and makes homeopathic injectables.

“Your firm released homeopathic injectables for distribution in the United State without performing adequate release testing. Specifically you failed to test for bacterial endotoxin in the products you manufactured for your customer … such testing is essential to assure safety of your products.”

The company also failed to establish and document that the analytical methods used in the lab were adequate for their intended uses.

FDA said, “Your sterility test method validation … for the GUNA-HIP and GUNA ISCHIAL injectable drug products is inadequate because it lacks verification of bacteriostasis and fungistasis activity in the product. In addition, you did not validate the sterility test method for other products that you manufacture.”

The firm also failed to conduct stability testing for its products. “In your response you indicated that homeopathic drug products are exempt from the requirements under 211.137. While expiration dating is not required for homeopathic products, evidence of the product stability over the expected shelf life is required. Your response failed to provide an assessment of the stability of the homeopathic injectable products that you manufacture.”

International Isotopes: no identity tests

An Oct. 26 warning letter to radiopharmaceutical manufacturer International Isotopes Inc. in Idaho Falls, Idaho, said that the firm made a number of mistakes related to identity testing. The company was inspected July 18-22, 2011.

“For example, your firm does not perform identity tests for Sodium Iodide I-131. Each batch of incoming components must be tested for identity of each shipment.”

The firm has also failed to calibrate instruments, apparatus, gauges and recording devices at suitable intervals.

The firm was also criticized for having inadequate controls over computer systems to ensure that the input and the outputs are checked for accuracy.

FDA said the company’s custom software for its master batch production record, which it calls the “I-131 Database,” had not been validated.

Akzo kept using faulty equipment

API maker Akzo Nobel NV came under fire in a Nov. 16, 2011, warning letter for not adequately investigating the reason for black particles found in 13 batches of API and for not repairing equipment that led to the contamination.

The likely cause of the particles was leaking hydraulic oil in equipment at the company’s site in La Paz, Mexico. The facility was inspected in June 2011.

During the manufacture of these batches, production personnel reported hydraulic oil leaking from the equipment. Yet the firm did not perform an investigation to identify the contaminant or to determine the root cause of the leak.

The firm said that it had repaired the leaking equipment each time the leaking oil was detected and that it would revise its procedure for process deviations investigations to address documentation of the investigations, root cause analyses, and corrective and preventive actions.

Yet this response did not sit well with FDA. “Your response is inadequate because it does not describe the specific changes you will make to your procedures or the timeframes for their completion and implementation, nor did you describe any additional process controls to prevent contamination of the API. In addition, your firm has assumed that the black particles are hydraulic oil contamination, without an investigation or any identification of the contaminant. Furthermore, you did not describe what steps you will be taking with regard to the API manufactured using defective production equipment.”

The company was also criticized for having an inadequate maintenance procedure in place to prevent contamination.

“For example, the inspection revealed that between August 2010 and August 2011 at least 10 maintenance requests were submitted as a result of oil leaks detected during manufacturing. … However, your response is inadequate because it fails to explain why your firm continually authorized the use of manufacturing equipment known to be defective.”

Labor used unqualified personnel

Following a July 18-21 inspection, contract lab Labor L+S AG in Bad Bocklet-Grossenbrach, Germany, received an FDA warning letter on Dec. 15 for not adequately investigating batch failures and for having untrained personnel conduct sterility tests.

In terms of the first failing, the firm’s investigations into missing environmental samples were deficient.

“Your investigation into these incidents concluded that these errors occurred ‘inadvertently.’ However, the investigation failed to establish a root cause and your quality unit failed to ensure the implementation of adequate corrective actions to prevent future recurrence.”

In addition, the investigator found several internal deviation reports in which the firm did not identify the product involved or the final test results associated with these investigations.

The company also came under fire for allowing inadequately trained and qualified personnel to perform sterility tests.

For example, the inspection revealed that one of the sterility testing technicians conducted sterility testing unsupervised in January 2010 prior to completing analytical testing training.

Novartis had series of quality problems

A Nov. 18 warning letter to Novartis AG, the parent company of generic subsidiary Sandoz, pointed to a series of quality problems at three Sandoz facilities in Canada, Colorado and North Carolina.

FDA officials said that many of the problems seen at these sites have still not been resolved despite earlier warning letters.

The warning letter was prompted by a series of inspections at these sites from April to August 2011.

At its Boucherville, Canada, site, the company was criticized for having inadequate controls to prevent microbiological contamination and for having an ineffective quality system in handling manufacturing problems.

The company, said FDA, failed to have adequate records on the total number of units incubated compared to the total number of units filled, which resulted in “inconsistent and inaccurate media fill results.”

This included six media fill lots manufactured from 2009 to 2011 where the number of units filled did not match the number being evaluated.

FDA officials said that the number of units evaluated in some media fill runs was smaller than what had been filled, and in other media fill runs, the number of units evaluated as greater than what had been filled. “Your firm lacked a justification for these discrepancies.”

Also at the Canadian facility the firm had also not thoroughly investigated batch failures, including crystals found in the finished product. Five out of eight lots reviewed during the inspection contained crystals in the vials.

“We are concerned that your firm lacks process understanding to consistently manufacture [redacted] Injectable.”

FDA also noted that a similar crystallization problem was observed in an earlier inspection conducted in July 2009 which should have been taken care of by management.

“Your failure to implement appropriate corrective actions and prevent future recurrence is indicative of an ineffective quality system.”

The warning letter also said that the Sandoz Canada facility did not file field alert reports within three days of discovering contamination in the distributed product.

At the firm’s Broomfield, Colo., facility, FDA inspectors found inadequate process controls. For example several pieces of equipment were used for producing multiple drug products “without validating the process using both pieces of equipment for each of the products.” FDA said this was also a repeat violation from the August 2008 warning letter issued to the Wilson, N.C., facility.

Another finding at the Colorado facility similar to the finding at the Canadian plant was that the firm has not thoroughly investigated the failure of a batch or any of its components to meet its specifications, whether or not the batch has already been distributed.

“Fourteen out of 79 (18 percent) laboratory investigations lacked documentation of either an investigation into other associated batches or products, or a corrective action. In-process specification for [redacted] uniformity was not met for three validated processes including [redacted] mg capsules…”

And at the Wilson, N.C., facility, inspectors also found inadequate process controls.

During the inspection, an investigator identified at least four products for which process validation was performed using different process validation methods.

FDA officials said that many of the problems seen in earlier inspections in 2008 had still not been resolved by upper management, particularly at its Canadian and North Carolina facilities.

In a tersely worded statement, FDA said that “it is apparent that Novartis International AG (Novartis) is not implementing global and sustainable corrective actions. … Corporate management has the responsibility to ensure that quality, safety and integrity of its products. Neither upper management at Novartis nor at Sandoz Inc. nor at Sandoz Canada Inc. ensured global, adequate, or timely resolution of the issues at these sites.”

Ironically, Novartis’ quality system was held up as an example of a system worth emulating at an Oct. 4, 2011, gathering of pharmaceutical quality professionals in Arlington, Va., a month before the warning letter was issued (Also see "How Pharma is Working to Change Cultures With Q10 Quality Systems" - Pink Sheet, 21 Dec, 2011.).