Fiscal Year (FY) 2019 was a fascinating year for drug GMP warning letters in the diversity of topics addressed, depth of focus, and trends in enforcement actions. This article presents a comprehensive summary of the drug GMP warning letters issued in FY2019, including an evaluation of trends since FY2013. The data presented for FY2019, ending September 31, 2019, is based on drug GMP warning letters posted by the FDA no later than January 20, 2020.
This article is data dense and includes many figures, so we are publishing it in a two-part series.
- Part 1 begins by discussing warning letter trends between FY2013 and FY2019, including an in-depth look at manufacturers by country and data integrity citation trends.
- In Part 2 we take an in-depth look at the types of drug product manufacturers that received warning letters in FY2019, as well as the time between inspection and warning letter issuance. We also review other notable topics regarding warning letters issued in FY2019 and offer conclusions to the data.
The term “compounding pharmacy” as used here includes outsourcing facilities and, for the purpose of this analysis, is considered a separate category from other drug manufacturers based on their legal foundation. These sites are located in the United States, but they are not considered with data from U.S. drug manufacturing sites in most analyses in this article.
Outsourcing facilities were established by the FDA under an amendment to the Food Drug & Cosmetic Act in November of 2013. The data from these sites are included in Figures 1 and 2. All other tables and figures omit this market segment data. All data is reflected in fiscal years.
The narrative and figures in this report address four broad areas, spread among Part 1 and Part 2, including trends between FY2013 and FY2019 in:
- Type of manufacturer (API, dosage form, API and dosage form, compounding pharmacy/outsourcing facility) and country associated with the site that is the subject of the warning letters.
- The time interval between the inspection and issuance of a warning letter including data presented by compounding pharmacy sites, US sites, and OUS sites.
- Particular targets of warning letters, including over-the-counter (OTC) and homeopathic drug products, drug product manufacturers, API manufacturers, and Human Cell Therapy Product (HCT/P) manufacturers.
- Notable topics, including but not limited to data integrity associated warning letters, nitrosamine contamination of Angiotensin II Receptor Blockers (ARBs), API re-packagers, and the first warning letter citing failure to comply with requirements in the Drug Supply Chain Security Act amendments to the FD&C Act.
Five high-level conclusions can be drawn from the FY2019 warning letter data; additional conclusions are provided at the end of Part 2.
- Warning letters issued to firms in the US far exceed the number issued to firms OUS for the first time in the 8 years for which I present data. (Figures 1 and 2)
- Warning letters to compounding pharmacies continue to decrease following a trend that began in FY2017 after a high was reached in FY2016. (Figures 1 and 2)
- Warning letters that include a data integrity component continue to decrease, a trend that began in FY2017 after they reached a high in FY2016. (Figures 6 and 7)
- The intervals between inspection and warning letters continue to decrease with the exception of warning letters issued to compounding pharmacies where the interval increased in FY2019. (Part 2)
- A notably diverse group of market segments and topics are included in the FY2019 warning letter collection including but not limited to failure to comply with the DSCSA, failure of re-packagers to include complete information on CoAs, GMP deficiencies that led to nitrosamine contamination of Angiotensin II Receptor Blocker (ARB) APIs and drug products, warning letters issued in the OTC market segment, and finally, warning letters issued to human cell therapy producers. (Part 2)
Overall Warning Letter Data
Figure 1 and Figure 2 show that drug GMP warning letters more than doubled from FY2015 to FY2016 and continued to increase through FY2019. The increase continues but has been more gradual in the past three fiscal years. Figure 1 also shows that while the FDA continues a focus on compounding pharmacies, the number of warning letters issued to these entities has decreased significantly for the past three years. The number of warning letters issued to these firms in FY2019 is approximately 34 percent of the number issued at a high point in FY2016. I doubt that this is because they have become more GMP compliant. Two factors may have contributed to this decrease:
- Many firms have decided to cease the compounding of sterile injectable drugs after inspection by FDA
- FDA has taken a new and intense focus on OTC firms in the past two years. Further, regarding outsourcing facility compliance with GMPs, the FDA published draft guidance describing how outsourcing facilities might comply with GMP regulations in December 2018 and published a 51-page second draft on January 22, 2020.
Note that data from compounding pharmacies/outsourcing facilities are not included except in Figures 1 and 2. All other tables and figures omit this market segment data.
(Note: clicking on the figures will enlarge them.)
Figure 3 shows that FY2019 warning letters issued to API sites are approximately half the number issued in FY2018 which was down slightly from the peak in FY2016 and FY2017. The number of warning letters issued to drug product sites, however, increased substantially again this year to a total of 96 and represents 86 percent of all drug GMP warning letters issued to sites excluding compounding pharmacies. We will address later some reasons for the increase in the number of warning letters issued to this group.
Figure 4 presents the location of the firms outside the US that were the subject of warning letters in FY2013 through FY2019. The number of warning letters issued regarding sites in India increased slightly from FY2017 and FY2018. Firms in China, however, saw a dramatic decrease in warning letters from the previous year and the number is now more consistent with values from FY2016 and FY2017. Last year, sites that were the subject of warning letters were located in 10 countries outside the US, and this year that number fell to 9.
Figure 5 shows the distribution of all drug warning letters issued in FY2019 based on geographic region or countries where the firms are located. FY2019 is notable because this is the first year that warning letters issued to firms in the US exceed those issued regarding sites outside the US since I began monitoring this in FY2013. These data show that significantly more than half of the warning letters in FY2019 were issued to firms in the US.
China and India together continue to be locations for a significant number of warning letters issued OUS. This shift in the location of recipients of warning letters begs the question of whether the number of inspections outside the US decreased, and those inside the US correspondingly increased in FY2019. Dr. Woodcock presented data in her testimony before Congress on December 10, 2019, showing this is not true. The FDA conducted 698 inspections in the US, and 966 outside the US in FY2019.
Dr. Woodcock’s testimony states that “FDA’s Inspections of foreign drug manufacturing facilities increased sharply after 2006 and have exceeded inspections of domestic drug facilities since 2015.” Thus, even with fewer inspections conducted in the US the number of warning letters is markedly higher.
Data Integrity
Rather than writing a separate article on Data Integrity in warning letters, this year I’m consolidating that information here.
Figure 6 shows that the percentage of all warning letters with a data integrity component continues to decrease after reaching a high of almost 80% in FY2016. Warning letters to sites outside the US with a data integrity component also continue to decrease. Sites in the US, however, are up slightly over last year, perhaps due to the increased number of warning letters issued to firms here.
It is also interesting to note that 38 percent (20 total) of warning letters to OTC/homeopathic manufacturers included a data integrity component. In FY2019, the percentage of warning letters issued to sites in the US, OUS, and overall is remarkably similar at just under 50 percent.
Figure 7 shows the geographic distribution of warning letters with data integrity deficiencies. The US takes first place this year with 32, India was second with 9, and China was third with 6. The remaining 5 warning letters are divided among other countries.
Although the absolute number of warning letters with a data integrity component issued to India and China this year seems reasonably low, it is useful to evaluate these as a percentage of the total number of warning letters issued to firms in the country. The data below results from the calculation of data in Figure 4 and Figure 7.
- 56 percent of the warning letters issued to firms in India included a data integrity component.
- 40 percent of the warning letters issued to firms in China included a data integrity component.
- 47 percent of the warning letters issued to firms in the US included a data integrity component.
The deficiencies cited in this area have not changed much since this topic was first identified more than 15 years ago. Firms continue to fail to exercise adequate controls over computer systems, fail to review all analytical raw data, fail to review audit trails, have demonstrable disagreement between electronic data and paper data that address the same event including batch records and human-machine interfaces (HMI), discard both electronic and paper-based GMP data, and abort electronic analytical test runs without adequate reasons.
In Part 2, we will look at particular targets of warning letters, including over-the-counter (OTC) and homeopathic drug products, drug product manufacturers, API manufacturers, and Human Cell Therapy Product (HCT/P) manufacturers. Additionally, we will review the time interval trends between inspection and warning letters, notable topics that arose in FY2019, and conclusions to the data.