An environmental monitoring programme exists to answer one question honestly: is this environment actually under control right now. Two FDA warning letters issued about seven weeks apart in the summer of 2026 show what happens when that question gets answered with paperwork instead of evidence.
Neither firm was accused of skipping environmental monitoring altogether. Both had EM programmes, logbooks, and plates. The failures were in what happened after the plates were read, or in one case, whether the plates were ever actually there to read.
Multiple environmental monitoring and [redacted] samples were documented in logbooks as collected and incubated; however, the corresponding plates could not be located in incubators during the inspection. Colony forming unit counts had been recorded as if incubation was actively in progress.
What the two letters found
FDA inspected Reliance Life Sciences' Navi Mumbai facility in February 2026 and found environmental monitoring and other microbiology samples logged as collected and incubating, with the physical plates nowhere to be found in the incubators. When investigators pressed the point in management interviews, the firm's own analysts confirmed the samples had never been collected in the first place. The colony forming unit counts on file were not measurements of anything. FDA also found the lab's system for calculating microbial counts skipped a required dilution factor, which FDA said pointed to a risk of systematic underreporting of microbial findings. FDA recommended that the firm remove sterile drug products from the U.S. market, and the firm agreed to recall all of them.
Roughly seven weeks earlier, FDA cited International Medication Systems Limited after a very different kind of EM failure: the data was real, and alarming, and the firm released the batch anyway. A viable air monitoring plate positioned in the ISO 5 aseptic filling area came back too-numerous-to-count, with confluent bacterial growth and insect larvae. A personnel monitoring plate in the adjacent room, on the same day, showed the same thing. The firm's investigation attributed both results to insects somehow infiltrating sealed monitoring plates, a theory unsupported by any physical evidence of tampering or defect. The quality unit rated the contamination risk as low, based on other environmental results from that day looking acceptable, and released the epinephrine injection lot for distribution.
FDA's response to International Medication Systems made a point worth repeating: sterility testing on the finished batch cannot substitute for a credible EM investigation, because contamination is typically episodic, not uniform. A passing sterility test on a handful of units says nothing about the two rooms that returned confluent bacterial growth and insect larvae the same day.
What 21 CFR actually requires here
21 CFR 211.113(b) requires written procedures designed to prevent microbiological contamination of drug products that purport to be sterile, including validation of aseptic processes. 21 CFR 211.194(a) requires laboratory records to include complete data from every test performed. Neither regulation treats environmental monitoring as a box-checking exercise. An EM result is only meaningful if it reflects something that was actually measured, and an adverse result only protects patients if the investigation that follows takes it seriously instead of explaining it away.
What a defensible environmental monitoring programme actually contains
An EM programme that holds up under inspection has to connect every stage, from sample to result to decision, with nothing skipped or assumed:
- Sample collection that can be independently verified. A chain of custody for every plate, from placement through incubation through reading, so a logged result can always be traced back to a real, physical sample.
- Complete, accurate calculation of results. Every required factor, including dilution, applied consistently, with the raw data retained alongside the final figure.
- Established alert and action limits with a defined response for each. What happens automatically when a result crosses a limit, not left to case-by-case judgment after the fact.
- Investigations that default to skepticism, not convenience. An explanation for an adverse result has to be supported by physical evidence, not simply be the explanation that allows the batch to ship.
- A quality unit both able and willing to reject a lot. The authority to stop distribution has to be exercised when the evidence calls for it, not overridden by a passing test elsewhere in the process.
Both cases in these letters involved firms with an EM programme on paper. Reliance's programme could not withstand a simple check for whether the plates it referenced actually existed. International Medication Systems' programme produced a real, disturbing result and then reasoned its way past it.
Why "the sample may have been misplaced" doesn't hold up
Reliance's response to FDA offered that the unaccounted-for samples might have been misplaced. FDA's reply noted this explanation directly contradicted what the firm's own analysts had already admitted: the samples were never collected. When a firm's account of an EM discrepancy conflicts with its own staff interviews, the discrepancy stops being a data integrity question and becomes a credibility question that extends to every other result in the system.
Why an unsupported hypothesis is not a root cause
International Medication Systems' insect-infiltration theory illustrates a pattern FDA has flagged repeatedly: an investigation that lands on whichever explanation requires the least action. Written procedures already required plates to be inspected before use and handled under strict controls after sampling specifically to rule out contamination of the plate itself. An investigation that skips past those controls to reach a low-risk conclusion, without the evidence to support it, does not satisfy the requirement to identify an actual root cause.
Environmental monitoring only works as a control if an adverse result is allowed to mean something. A programme that produces numbers nobody can trace back to a real sample, or that explains away contamination severe enough to include insect larvae, is not functioning as a control at all. It is functioning as documentation that a control existed.
Key regulatory references
- FDA Warning Letter 320-26-117, Reliance Life Sciences Private Limited, August 18, 2026
- FDA Warning Letter 320-26-99, International Medication Systems Limited, July 2, 2026
- 21 CFR 211.113(b): Control of microbiological contamination
- 21 CFR 211.194(a): Laboratory records
- FDA Guidance: Sterile Drug Products Produced by Aseptic Processing: Current Good Manufacturing Practice
Explore our Environmental Monitoring GMP Map and EM Reviewer tool
Walk the EM programme end to end, from sampling through investigation, and try the free EM Reviewer tool for a quick check against common inspection gaps.