Each week we take one recent FDA warning letter and work through it the way a quality unit should: what the observation actually said, what the firm proposed in its Form 483 response, why FDA found that response inadequate, and what a defensible answer would have required.
Every issue ends with a short diagnostic you can apply to your own open responses.
A sterile injectables maker answered four media fill failures with better simulations and intervention tracking. FDA’s reply: a better test does not fix a line that generates the risk, and the design flaws had no dates.
A contract manufacturer answered three findings with a new programme, two procedures and a CAPA number, all inside three weeks. FDA rejected all three, for the same reason each time: not one of them contained a specification.
A sterile contract manufacturer closed its product impact conclusions on what had not happened since. FDA answered that three silences produced by three deficient systems are not corroboration — they are the finding, restated.
A contract lab scoped its response to the samples FDA had no jurisdiction over. One quality system and one set of procedures meant there was no line to draw — and the firm’s own defence sized the review for it.
Three findings, three commitments, three rejections in almost the same words. What a corrective action with a completion date leaves unanswered — and what is already on the market.
A firm found fourteen more affected batches than FDA did and invalidated all sixteen sterility tests. FDA still rejected the response, because a review scoped by sampling cannot find the edge of the problem.
Four inspections in seven years cited the same cleaning deficiencies. Why a repeat observation changes who has to answer, and what that answer must contain.
Three years of validation cycles passed with biological indicator growth in nearly every one. FDA’s finding was not the growth. It was the limit that kept calling it a pass.
A batch released with no microbiological testing, because no customer had asked for it. FDA’s answer: a purchase request is not a specification.
Six corrective actions built on retraining, six rejections, and the same four reasons in rotation. What training alone can never supply.
Insect larvae on an ISO 5 monitoring plate, a risk rating of low, and a released lot. FDA read the disposition itself as proof the quality unit was not exercising its authority.
A failing result, no organism identified, a passing resample, and a released lot. FDA rejected the invalidation because a resample that passes is not a root cause.
A firm listed its prior corrective actions to demonstrate diligence. FDA read the same list as proof its CAPA system could not detect its own ineffectiveness.
Three findings, three attributions to individual error, three rejections. The firm’s own written procedure had directed the behaviour it blamed on the analyst.
Veritas Quality Consultants works with pharmaceutical, medical device, and biologics manufacturers on Form 483 responses, root cause analysis, and CAPA development — built to FDA expectations inside the 15-day window.
Request a consultation