OBJECTIVES: Some studies have illustrated that the relative efficacy and safety of widely used drugs can vary over time substantially. Possible influencers may include changes in population characteristics, inclusion or exclusion criteria such as patients with prior exposure to the active reference, or variation in the placebo response. Sumatriptan was considered as an illustrative case to examine this phenomenon, as it has steadily been used as a reference treatment over time for acute migraine episodes.

METHODS: A systematic literature review of randomized controlled trials (RCTs) of adults suffering from acute migraine episodes evaluating oral sumatriptan versus placebo was performed in Medline for the period from before and shortly after approval (1991) to the following 15 years. Meta-analyses were performed using the inverse-variance weighted average method to estimate the odds-ratios of the occurrence of pain-free at 2 hours and of any adverse event, by time period.

RESULTS: Out of the 67 RCTs identified, pain-free at 2 hours and any adverse event outcomes were reported in 25 and 28 studies, respectively. The odds ratio (95% CI) for the occurrence of pain-free for sumatriptan versus placebo was first evaluated at 3.13 (1.67-5.86) with meta-analysis based on the first RCTs assessing sumatriptan before or shortly after approval (1991-1994) and increased up to 4.14 (3.67-4.67) using also RCT data post-approval (1995-2006). No specific variation was observed in the relative tolerability effect of sumatriptan over placebo over time.

CONCLUSIONS: This research provides new evidence that the relative effect of widely used drugs such as sumatriptan can evolve over time substantially. This issue is particularly problematic when network meta-analysis or indirect comparisons are subsequently performed to evaluate the potential of a new drug, compared to widely prescribed older drugs. Assessment of this effect should be evaluated when making comparative analyses.