by Barbara Unger, GMP Quality Expert and GMP Regulatory Intelligence Editor-in-Chief


Part of a comprehensive GMP Intelligence program is the monitoring of enforcement actions, including FDA 483s, warning letters, recalls, import alerts, consent decree agreements, and EU reports of GMDP noncompliance. This article presents the most recent GMP inspection data from CDER and MHRA.  The CDER data are from inspections conducted in FY2016 and the MHRA data come from inspections conducted in 2015.

The CDER drug inspection observations supplement the information we previously published regarding CDER Drug GMP warning letters from the same time.  The analysis herein includes data from the FY2016 483 observations and compares results with those from the three previous fiscal years. Raw data come from the FDA website but are presented in a different manner. For example, I have combined the frequencies of all observations that cite 21 CFR211.192, into a single number. In the FDA data, there are multiple line items for 211.192, each with a different frequency.  For example, in the FDA listing the most frequently cited item is 211.22(d), ‘procedures not in writing, fully followed.’  When the full collection of times that 211.192 and 211.42(c) are cited, however, they become number one and number two respectively with 211.22(d) becoming the third most frequent citation.  FDA uses the term ‘frequency’ which seems to be the number of times a given citation was identified in the 483 collection supporting these data.

Only 483s that were issued through the Turbo EIR system are considered in these data which provides a distinct limitation.  No 483s issued to API manufacturers or issued outside of the Turbo EIR system are included. This becomes important to consider with the FDA’s increased focus on API manufacturers, particularly outside the US.

We also looked at the GMP inspection data from the MHRA.  The recently published 78-slide deck presents GMP Inspection Deficiency Data Trends for 2015.  The MHRA provides graphic representation of the observations linked to the Chapters or Annexes that were cited. In this respect, it differs from the FDA published data that do not provide graphics. 


  • The number of 483s included in this analysis remains reasonably constant over the past four fiscal years even though it does not represent all drug inspections conducted by the FDA, particularly inspections of sites that manufacture APIs.
  • Deficiencies in investigations remains at the top on this list over the past four years. We as an industry cannot seem to get this quite right.
  • In general, the regulations cited and their relative order has remained reasonably constant over the past four fiscal years. Even though a few items have changed place, none of the numbers are striking.
  • MHRA issued 51 ‘critical’ deficiencies, in a total of 304 inspections of which 79 (26%) were overseas inspections and 224 (74%) were in the UK.
  • The EU GMP Guide Chapters and Annexes that had critical observations include: Chapters 1, 4, 6, and 8; Annex 1 and 11. Scaling of the graphs made it impossible to accurately determine the specific number of each.
  • The MHRA cited a total of 3176 deficiencies in the 303 inspections. Criticals constituted 2% of the deficiencies, Majors were 29%, and Others were 69%.


The following data are based on inspections generated using the FDA Turbo-EIR system.  The number of 483s remained quite similar over the four years in question, with FY2014 having the fewest.  483s issued to API manufacturers or issued outside of the Turbo EIR system are not included.

Table 1 shows only the most frequent group of inspection observations, the FDA tabulation shows all.  Table 1 is organized in the order of those observations with the highest to lowest frequency for this year.

TABLE 1: Inspection Observations Issued Through Turbo-EIR System per Fiscal Year.  These are shown in the order of highest to lowest for FY2016.

CITATION  SHORT DESCRIPTION  2013 2014  2015 2016
Total 483s issued thru Turbo EIR system for ‘Drugs’ 690 645 678 691
211.192 Production record review, investigations of discrepancies 239 209 250 227
211.42(c) Requirement for adequate facilities to prevent contamination or mix-ups 94 125 235 227
211.22(d) Quality unit responsibilities should be in writing and should be followed 168 148 165 153
211.160(b) Development of scientifically sound specifications 199 165 246 133
211.166(a) Expiration dating should be supported by appropriate studies 104 82 126 124
211.113(b) Validation of aseptic processes including sterilization 119 109 157 118
211.100(a) Written procedures shall describe production and process controls 135 107 123 110
211.67(b) Procedures shall describe cleaning and maintenance of equipment 83 80 91 102
211.188 Batch production records 114 74 110 100
211.25(a) Staff shall have training, education and experience to perform their jobs 132 115 119 99
211.67(a) Equipment shall be cleaned and maintained 71 94 113 94
211.68(a) Automatic equipment including computers shall be routinely calibrated and inspected 69 64 72 80
211.165(a) Specifications shall be established for drug product 66 64 80 73
211.100(b) Activities shall be documented at the time of performance 84 62 72 70
211.110(a) Processes shall be validated 79 74 85 65

Figure 1 below shows the data from Table 1 graphed over four fiscal years, 2013 – 2016.  While there is some variation from year to year, the frequency with which specific regulations are identified remains generally constant.  Figure 2 shows additional detail of several of the areas where the frequency of the observation did show some variation between FY2015 and FY2016.

In conclusion, there is little change in the overall frequency of inspection observations, as characterized by the regulation cited, between FY2013 and FY2016. This may have been different if all inspected sites, including API sites, had been included in the metrics.  The three most frequent observations in FY2016 cite 211.192 (investigations), 211.42(c) (validation of aseptic processes including sterilization), and 211.160(b) (scientifically sound specifications).  In several instances, though, the order of the observations did change in FY2016, and these are highlighted in gray in Table 1.  While 211.192 continued in first place for all four fiscal years, 211.42(c), Requirement for adequate facilities to prevent contamination or mix-ups moved from third place to second place, even though the actual number of those observations decreased. Citations against 211.160(b) Development of scientifically sound specifications went from second place to fourth place this year.  Observations citing 211.113(b) Validation of aseptic processes including sterilization dropped from fifth place to sixth place in 2016 and the actual number decreased significantly to FY2013 levels.  Finally, observations identifying 211.25(a) ‘Staff shall have training, education and experience to perform their jobs’ dropped from eighth place to tenth place this year.

MHRA Inspection Deficiencies

I won’t reproduce the graphics from the MHRA slide deck; do read those because they contain a wealth of information at a very granular level.  The areas with critical deficiencies include, in order of most frequent (% of critical deficiencies):  Quality Systems (53%), Complaints and Recalls (20%), Documentation (18%), Quality Control (8%), and Computerized Systems (2%).  Figure 3A shows the distribution of deficiencies among the three classifications by actual number.   Figure 3B shows this distribution of deficiencies among the three classifications by percentage.  Figure 4 shows the distribution in percentages, of critical deficiencies using the MHRA identified categories and the associated Chapter / Annex.

The MHRA inspections identified 51 total ‘critical’ observations for 2% of the total number of inspection identified deficiencies.  About a third of the deficiencies are in the ‘major’ category and almost 70% are in the ‘other’ category.  We cannot compare this with the US FDA inspection observations because the FDA does not classify their observations.  In the future, perhaps health authorities will adopt a common classification categories for inspection observations.

The Quality System was the subject of the majority of MHRA identified critical GMP deficiencies in 2015 followed by an almost equal percentage for Complaints and Recalls and Documentation.  Quality Control followed and a single critical deficiency was identified in Computerized Systems.

The fifty-one (51) critical observations cluster in four (4) of nine (9) EU GMP Chapters, and two (2) of nineteen (19) GMP Guide Annexes.  These include:

  • Chapter 1: Pharmaceutical Quality System
  • Chapter 4: Documentation
  • Chapter 6: Quality Control
  • Chapter 8: Complaints and Product Recalls
  • Annex 1: Manufacture of Sterile Medicinal Products
  • Annex 11: Computerized Systems


It is difficult to directly compare these areas with those identified by the FDA.  The FDA does not categorize the criticality of inspection observations as the MHRA and other health authorities do.  We can, however, say that with observations addressing ‘investigations’ at the top of the list, ‘quality unit responsibilities’ third on the list, and ‘staff training’ at number 10, it’s clear that the Quality System is a high priority for both the FDA and the MHRA.  Computer systems is identified separately by the MHRA in Annex 11, but in 483s this type of observation is generally identified as 211.68(b) that addresses controls over computer systems.  Part 11, Electronic Records; Electronic Signatures, is rarely, if ever, identified in either 483s or warning letters.

For those who use inspection observations as a means of monitoring and improving their own quality systems, both publications provide ample resources against which firms can measure their own potential vulnerabilities.

Learn more about how FDAzilla can help you achieve your quality and inspection preparation goals: get 483sInspector ProfilesEnforcement Analytics, and GMP Regulatory Intelligence. Contact me if you ever have questions at tony@fdazilla.com.

Leave a Reply

Your email address will not be published. Required fields are marked *