Maria Eng is a behavioral epidemiologist and adjunct professor at Johns Hopkins University. She has a Phar.D. and M.P.H. degrees.

Maria Eng

Disclaimer: The opinions expressed in this article do not necessarily reflect FDA’s position on the matter being discussed.

Since 2007, foreign drug inspections have increased three-fold, from 200 to 600 in 2013 (1, 2). As more drugs are produced overseas, particularly in areas where regulatory oversight may differ compared to that in the United States or Europe, it is crucial to ensure that products entering the supply chain are safe, effective, and of high quality.

	FDA has been ramping up inspections of international drug manufacturers during the past decade to improve drug safety. The heparin recall of 2007–2008, which resulted when oversulfated chrondroitin sulfate was used to substitute for the active ingredient in heparin, illustrates the serious harm that substandard pharmaceutical drug products can cause to consumers. The contamination originated from pig slaughterhouses in China, affecting mostly the US market, but also citizens from 10 other countries worldwide (3).

	To prevent situations like this, FDA regulates manufacturers of API and finished dosage form (FDF) drug products through standards, regulations, and guidelines promulgated from the Federal, Food, Drug, and Cosmetic Act. Through statutory law, authorities given to FDA allow for regulations known as current good manufacturing practices (cGMPs) to assure control of manufacturing processes. Throughout the drug process, manufacturers must meet minimum requirements for identity, strength, quality, and purity of drug products by monitoring system practices and operations (4). By ensuring quality control at each stage of the drug production, contamination, deviations, and failures are more likely to be prevented.  Areas such as buildings and facilities, equipment, personnel qualification and training, starting materials, laboratory, packaging and labeling, and production units are all evaluated during a drug inspection (4). If manufacturers are not compliant with cGMPs, they can be issued a FDA Form 483 listing areas of noncompliance, in order of significance.

	This project analyzed all FDA inspections performed outside of the US from January through December 2013, to pinpoint problem areas for compliance.  The majority of inspections were conducted in Europe, at 33.76% and Asia, at 63.39%. The International Conference on Harmonization’s (ICH’s) Q7, 

Good Manufacturing Practice Guide For Active Pharmaceutical Ingredients

 (5, 6) was utilized to categorize results.

	Between January 1 through December 31, 157 API plant inspections found 839 observations of noncompliance within ICH Q7.  The five most problematic areas, in order of descending frequency, were found to be in ICH Q7 sections 11, Laboratory Controls; 12, Validation; 6, Documentation and Records; 5, Process Equipment; and 8, Production and In-Process Controls. Based on a Pharmaceutical Inspection Co-operation Scheme (PIC/S) questionnaire, these data were analyzed to determine which operations were most prone to noncompliance in API facilities outside the US. The project’s overall goal was to determine which types of deficiencies API manufacturers were most commonly cited for during inspections. By understanding the types and trends of violations found from inspectional outcomes, future resources and knowledge could be directed in specific areas, to help manufacturers improve quality control and assurance (QC and QA).


	Data for the project were compiled from an internal database known as the mission accomplishment and regulatory compliance services (MARCS) compliance management system (CMS). It is an internal, web-based enterprise architecture application that links to other FDA resources and is used to manage compliance-related work activities (7). Records and documents associated with inspections that were conducted by FDA are placed under work activities and case numbers in the MARCS-CMS. By utilizing the advanced work search function and the advanced output options, work ID, work type (CDER-Evaluate Foreign GMP Inspection), firms, inspection profiles, work country, inspection start date, inspection end date, and inspection initial classification were fields used to generate the data.  The calendar year of January 1 to December 31, 2013 was used to assemble relevant data into an Excel spreadsheet.  Formatting and sorting of cells was then completed to distinguish between API and FDF sites.

	Through the class code filter, the following codes were defined as APIs: CBI (recombinant/non-recombinant protein drug substance of biologic origin), CEX (starting/intermediate derived from plant/animal extraction), CFN (non-sterile API by fermentation), CFS (sterile API by fermentation), CRU (non-sterile starting/intermediate [not plant/animal]), CSN (non-sterile API by chemical synthesis), CSS (sterile API by chemical synthesis), and CXA (purified API derived from plant/animal extraction).  Once API manufacturing sites were separated from FDF manufacturers, further grouping was used to distinguish between sites that received a FDA Form 483 of objectionable conditions: voluntary action indicated (VAI) or official action indicated (OAI) from those not receiving one; no action indicated (NAI).  The spreadsheet was further filtered to include VAI and OAI under the inspection initial classification column, and then each FDA Form 483 was pulled from the MARCS-CMS database.

	Observations listed in the FDA Form 483 were categorized against the ICH Q7 guideline (6). In addition to the introduction (Section 1), ICH Q7 features the following sections:

	2. Quality Management
	3. Personnel
	4. Buildings and Facilities
	5. Process Equipment
	6. Documentation and Records
	7. Materials Management
	8. Production and In-process Controls
	9. Packaging and Identification Labeling of APIs and Intermediates
	10. Storage and Distribution
	11. Laboratory Controls
	12. Validation
	13. Change Control
	14. Rejection and Re-use of Materials
	15. Complaints and Recalls
	16. Contract Manufacturers (including laboratories)
	17. Agents, Brokers, Traders, Distributors, Repackers, and Relabellers
	18. Specific Guidance for APIs Manufactured by Cell Culture/Fermentation
	19. APIs for Use in Clinical Trials.

	Each section is further broken down into multiple subsections that highlight more specific details.  For example, under Section 6, Documentation and Records, further subsections include:
	6.1 Documentation System and Specifications
	6.2 Equipment Cleaning and Use Record
	6.3 Records of Raw Materials, Intermediates, API Labeling and Packaging Materials
	6.4 Master Production Instructions
	6.5 Batch Production Records
	6.6 Laboratory Control Records
	6.7 Batch Production Record Review.

	Documentation of each citation was recorded under each corresponding subsection of ICH Q7.  Breakdown of data was evaluated in each section, subsection, geographic region, and country.  Foreign API drug manufacturers included countries outside the US, in Asia, Australasia, Europe, North America, and South America.


	Data were collected in mid-June 2014, using data from January 1 to December 31, 2013. At this time, 581 foreign drug inspections were conducted. Of the inspected sites, 270 (46.47%) were performed at API drug manufacturing facilities, while the remaining 311 were at FDF sites. In addition, 157 sites were issued a FDA Form 483 and classified initially as VAI or OAI at 58.15%, compared to manufacturers that did not receive one and were considered as NAI.

	A total of 839 observations were counted, ranging from 1 to 25 per site, with an average of 5.38 observations issued per site. Within regions, the average number of observations per site was 5.9 for Asia, 4.4 in Europe, 5 in North America, 3 in South America, and 1 in Australasia. Most of the violations, totaling 592, were found in Asian facilities; 233 violations were found in Europe; 10 in North America; three in South America and one in Australasia. 

 show the percentage of observations in each country in Asia and Europe.

Overall, ICH Q7 Section 11, Laboratory Controls, was the most frequently cited section, with 167 observations (19.90%). This was followed by 12, Validation, at 120 observations (14.30%); 6, Documentation and Records, with 109 (12.99%); 5, Process Equipment at 108 (12.87%); and 8, Production and In-Process Controls, with 62 observations (7.39%) (see 

	A further breakdown into subsections revealed that the most common overall observation (n=839) cited in ICH Q7 was for 11.1, General Controls, with 84 observations (10.01%). This subsection was followed by 5.2, Equipment Maintenance and Cleaning, with 61 observations (7.27%); 6.5, Batch Production Records, at 45 observations (5.36%); 2.2, Responsibilities of the Quality Unit(s), with 39 observations (4.65%); and 6.6, Laboratory Controls Records, with 36 observations (4.29%) (see 

In Asia, the top observations noted were for 11, Laboratory Controls, at 18.58%; 5, Process Equipment, at 14.19%; 12, Validation, at 14.02%; 6, Documentation and Records, at 12.84%; and 4, Building and Facilities, at 6.93%. Similar to Asia, in Europe, the most cited observation was for 11, Laboratory Controls, at 21.46%. This was followed by 12, Validation, at 15.02%; 6, Documentation and Records, at 14.16%; 8, Production and In-Process Controls, at 10.73%; and 5, Process Equipment, at 9.87%.

	The majority of inspections and observations occurred in India and China. In India, 39 sites were issued an FDA Form 483 and had 299 observations.  Across 41 inspections in China, 198 observations were issued to manufacturers.  These two countries accounted for 80% of inspections in the Asia region, and 83.95% of observations. There were similar situations with sections 11, Laboratory Control, and 12, Validation, which were the top two problem areas. Section 11 accounted for 17.39% of observations in India and 17.68% in China, and 12, Validation, for 14.38% of observations in India and 14.14% in China.

	In India, section 5, Process Equipment came in at 13.38%, 6, Documentation and Records, at 12.70%, and 4, Building and Facilities, and 7, Materials Management, tied at 7.69%.  In China, 6, Documentation and Records, was at 16.16%; 5, Process Equipment, at 15.15%; 2, Quality Management, and 4, Buildings and Facilities, tied at 6.57% (see 

The volume in North America, South America, and Australasia regions combined was an extremely small fraction of the total calculation, and accounted for only 2.55% with four inspections (n=157) and 1.67% with 14 observations (n=839).


	Across geographic regions of Asia and Europe and within China and India, deficiencies in section 11, Laboratory Control, were most frequently cited during inspection. In this section, API manufacturers failed to follow testing standard operating procedures, establish appropriate specifications with accepted reference standards and did not have knowledge of testing methods or investigate out-of-specifications or discrepancies.

	The next two cited sections for Asia and Europe had similar findings in 12, Validation, and 6, Documentation and Records. Validation plays another critical area of API drug manufacturing, because it is tied to production processes, cleaning procedures, analytical methods, in-process control procedures, and computerized systems.  Manufacturers must set up critical process parameters during the development stage so that ranges are established and operations can be reproducible based on the data. With validation of procedures and processes connected to so many parts of the manufacturing process, validation failures can negatively affect the quality of the API in every step to follow.

	Documentation and Records was another place that manufacturers were cited on when there was a failure to record data properly or retain records. Under subsections 6.5, Batch Production Records, and 6.6, Laboratory Control Records, firms did not have complete records or were not completing documentation contemporaneously. Documentation of every phase during a batch production must be completed accurately and legibly, to be checked against the master batch record.

	Differences within the top deviations were noted in Asia, where a higher percentage of observations were for 5, Process Equipment, whereas in Europe, Section 8, Production and In-Process Controls, had a higher number of deficiencies when inspected. Manufacturing sites in one region may have more challenges within a particular section of ICH Q7, and therefore, appear to have a higher percentage of deficiencies. This finding, however, could be attributed to selection bias, because the number of inspections in a region can vary from year to year.

	In Asia, subsection 5.2, Equipment Maintenance and Cleaning, was the area with the most deficiencies seen.  Equipment used in the manufacturing of APIs was not properly cleaned or given adequate maintenance. Improper cleaning of equipment between production cycles could affect the next product to be used, and result in contamination issues.

	With subsection 8.3, In-process Sampling and Controls, control procedures during manufacturing were deemed inadequate, as firms failed to monitor process steps that could cause variability in the quality characteristics of the API. India and China shared all of the top five observations, which included sections 11, Laboratory Controls; 12, Validation; 6, Documentation and Records; 5, Process Equipment; and 4, Buildings and Facilities.

	Almost equal in the number of inspections, Indian manufacturing sites were often cited more frequently. The average number of observations was slightly higher, with 7.7 compared to China with 4.8. The high number of inspections and observations found in India and China could be expected, since these two countries provide nearly 50% of APIs used in the US (8). Monitoring this number in the future can be useful to see if sites in India will continue to have more deviations with cGMPs, compared to China.

	The project also showed how inspection reporting could be standardized to facilitate trending and improve industry response.  Investigators wrote and formatted observations differently from one FDA Form 483 to the next. Depending on how the observation was written, the citation could be understood as a single or multiple count. For example, an observation may cite failure to follow a procedure in one unit, and then follow with another description stating failure in a different area. The two examples were listed under a single observation, yet, they related to two different units within ICH Q7.

	Another limitation in categorization of observations under ICH Q7 was when a single citation described the quality unit, laboratory, and documentation all in one observation.  Because the observation related to more than one section of ICH Q7, but was counted as a single observation, the classification of the observation could be placed in one section over another and subject to interpretation.

	While data were pulled half way into 2014, not all inspections that occurred in 2013 were included. If work activities were not entered into the MARCS–CMS database due to incomplete establishment inspection reports, missing FDA Form 483s would not be reflected in the total count.  Of foreign API manufacturing sites that were inspected from January 1 to December 31, 2013 and issued a FDA Form 483, the top areas of concern were for sections 11, Laboratory Control; 12, Validation; 6, Documentation and Records; 5, Process Equipment; and 8, Production and In-Process Controls.

	Today, most of the world’s APIs are manufactured in India and China, which were the countries with the most inspections as well as the highest number of observations. In 2013, Indian facilities received a higher number of observations than their counterparts in China. As more APIs are produced there, a rise in observations is expected.

	Steps could also be taken to standardize data collection. Currently, FDA has an automated reporting system, TurboEIR (Establishment Inspection Reports), which is used to standardize FDA Form 483.  However, with API products, FDA Form 483s are manually prepared (9).

	Having a program with the number and type of observations in place would establish a baseline for trending, and help focus resources and training. It could also provide insights to optimize control of API manufacturing, since any deviations from cGMPs could affect drug product integrity and potentially harm patients if they are not controlled. The manufacturing of APIs is a crucial point in the pharmaceutical drug process and all surrounding activities must be accurately performed in order to confirm the chemically active substance is safe and effective as it moves forward.


	1. A.C. von Eschenbach, MD., Statement, “FDA’s Foreign Drug Inspection Program,” House Committee on Energy and Commerce Hearing, Nov. 1, 2007.
	2. FDA, Data pull from Mission Accomplishment and Regulatory Compliance Services (MARCS) Compliance Management System (CMS) on June 16, 2014.
	3. J. Woodcock, MD, Statement, “FDA’s Ongoing Heparin Investigation,” House Committee on Energy and Commerce Hearing, Apr. 29, 2008.
	4. FDA, 

Guidance for Industry Quality Systems Approach to Pharmaceutical CGMP Regulations

 (Rockville, MD, September 2006).
	5. ICH, “FAQs About ICH

 (2014), www.ich.org/about/faqs.html, accessed July 29, 2013.
	6. ICH, Q7, 

Good Manufacturing Practice Guide For Active Pharmaceutical Ingredients 

CMS Software Users Guide For Operations & Maintenance 3

 (OM3) (2011).
	8. A. Edney, Drug Quality in China Still Poses Risks for U.S. Market, (Apr. 4, 2014).
	9. Inspection Observations, Fda.gov, (2013), www.fda.gov/ICECI/Inspections/ucm250720.htm, accessed July 30, 2014.

Pharmaceutical Technology's APIs, Excipients, and Manufacturing Supplement

:
	When referring to this article, please cite it as L. Tung and M. Eng, “Lessons Learned from FDA Inspections of Foreign API Facilities," 

Pharmaceutical Technology APIs, Excipients, and Manufacturing Supplement

Articles

A study of recent inspection observations allows for baseline trending and continuous improvement. (The opinions expressed in this article do not necessarily reflect FDA’s position on the matter being discussed.)