Post-Marketing Surveillance

Post-marketing surveillance is everything a regulator and a sponsor do to keep characterising a medicine after it is licensed. It runs from passive spontaneous reporting through to obligations with teeth: post-marketing requirements the FDA can compel, post-marketing commitments a sponsor agrees to, and in Europe post-authorisation safety studies written into a risk management plan. Phase 4 trials are the formal, protocol-driven end of the same spectrum.

The reason it exists is a counting problem. A registration programme of a few thousand patients can only rule out events much commoner than roughly one in a thousand, because seeing at least one case with reasonable confidence takes about three times the reciprocal of the true incidence. Rodent thyroid C-cell tumour findings with long-acting GLP-1 receptor agonists produced exactly this situation, and were handled with a boxed warning plus a long-running cancer registry rather than a refusal.

Surveillance changes decisions because it can reverse them. Rofecoxib was withdrawn in 2004 over cardiovascular events invisible in the original programme, and sibutramine was withdrawn in 2010 after a dedicated outcome trial in a higher-risk population. Equally often the outcome is quieter: a new labelled warning, a contraindication, a restricted population, or an obligation discharged with nothing found.

The common misreading is treating years on the market as evidence of safety. Time only accumulates evidence where something was actively collecting it, and an approved drug with no active surveillance may simply have no data. The inverse error is loudest around unapproved compounds, where the absence of adverse event reports is presented as a clean record when no mechanism exists to generate one.