Decisions by regulatory agencies: are they evidence-based?

Decisions by regulatory agencies follow explicit regulations and should be evidence-based. An established practice has been that approval of a new drug requires two independent clinical trials documenting safety and efficacy for the drug's intended use. But are the regulatory agencies rigorously ensuring that decisions are evidence-based? Contradictory statements about the non-steroidal anti-inflammatory drugs (NSAIDs) from the European Medicines Agency (EMEA) and the U.S. Food and Drug Administration (FDA) have raised this question.An FDA Advisory Committee convened in February, 2005 to review primarily the three selective COX-2 inhibitors available in the U.S. It concluded overwhelmingly (32 votes to zero) that these agents increase the risk of thrombotic cardiovascular events[1]. The evidence from several placebo-controlled clinical trials was considered conclusive. The problem was interpreted as a class effect, although the degree of harm appeared to differ among the agents. The Advisory Committee recommended that celecoxib remain on the market with major restrictions applied to its use[1].EMEA was in agreement with the FDA and recommended suspension of valdecoxib[2]. It also added new contraindications and warnings to the other marketed coxibs. Contraindications were added for patients with established ischemic vascular disease and reinforced warnings were issued for patients with risk factors of heart disease. Based on the same available scientific evidence, the FDA did not follow the recommendations by its Advisory Committee[3]. Rather the FDA added only a Black Box warning vaguely stating that celecoxib "may (author's emphasis) cause an increased risk of serious cardiovascular events,..." and that "Patients with cardiovascular disease or risk factors for cardiovascular disease may (author's emphasis) be at greater risk"[4].The recommendations for the non-selective NSAIDs by FDA and EMEA issued in 2006 also went in a different direction. EMEA concluded that t