Monami M et al. (2008) Rosiglitazone and risk of cancer: a meta-analysis of randomized clinical trials. Diabetes Care 31: 1455–1460

Although preliminary data indicate that some thiazolidinediones have anticancer effects, studies in patients receiving the antidiabetic drug rosiglitazone indicate that this particular thiazolidinedione might actually confer an increased risk of malignancy. Monami et al. have conducted a meta-analysis of randomized, controlled trials of rosiglitazone and have found that this treatment is not associated with an increased risk of cancer.

The authors searched the GlaxoSmithKline website and the MEDLINE database. They identified 80 randomized, controlled trials of >24 weeks duration that compared patients receiving rosiglitazone (n = 16,332) with individuals receiving placebo or other active drugs (n = 12,522). All trials reported at least one adverse event, 302 of which were cases of cancer.

The risk of cancer in the rosiglitazone-treated group was not significantly different to that in the comparator group (Mantel-Haenszel weighted odds ratio 0.91, 95% CI 0.71–1.16; P = 0.44). Furthermore, cancer risk was not significantly different when a single trial that had recorded more than half of all malignancies was excluded from the analysis or when only trials of ≥52 weeks duration were studied. The incidence of cancer was significantly lower in the rosiglitazone group than in the comparator group: the cumulative cancer incidence densities were 0.23 (95% CI 0.19–0.26) vs 0.44 (95% CI 0.34–0.58) cases per 100 patient-years, respectively (P <0.05).

The authors conclude that rosiglitazone is safe with respect to risk of incident cancers, but caution that these data do not confirm a protective effect of this agent.