We systematically reviewed published epidemiological studies evaluating the risk of bladder cancer with pioglitazone use. Some studies found an increased risk of bladder cancer with pioglitazone use and others did not. The following studies are presented to illustrate the inconsistent findings among all reviewed studies.
The manufacturer of pioglitazone, Takeda Pharmaceuticals, conducted a 10-year prospective cohort study in diabetic patients included in the Kaiser Permanente of Northern California (KPNC) database. The study included members of the KPNC diabetes registry between January 1, 1997, and December 31, 2002. The patients were followed until December 31, 2012. In June 2011, we issued a Drug Safety Communication to inform the public about the five-year interim study results, which found no significant increase in the risk for bladder cancer in patients ever exposed to pioglitazone compared to patients never exposed to pioglitazone (hazard ratio [HR]=1.2; 95% confidence interval [CI]=0.9-1.5). However, the risk of bladder cancer increased with increasing dose and duration of pioglitazone use. We have now reviewed the final 10-year results of this study.1 The final study included a total of 158,918 patients who never took pioglitazone (never users) and 34,181 patients who had taken pioglitazone at some point (ever users). Never users were followed for an average of 8.9 years, whereas ever users were followed for an average of 6.1 years. The investigators identified 1,075 newly diagnosed cases of bladder cancer in never users and 186 cases in ever users. Ever use of pioglitazone compared to never use of pioglitazone was not associated with an increased risk for bladder cancer, with a fully adjusted HR of 1.06 (95% CI=0.89-1.26). The study also suggested a modest trend towards higher risk with increasing duration of use, but this trend was not statistically significant, which means it could have been due to chance. Compared to the interim 5-year results, these final 10-year results found weaker associations that were not statistically significant. However, the directions of the associations remained unchanged.
Between May 2001 and April 2002, the investigators for the PROspective pioglitAzone Clinical Trial In macroVascular Events (PROactive) trial enrolled type 2 diabetic patients with extensive macrovascular disease from 18 European countries.3 Patients were randomized to double-blind treatment with pioglitazone (n=2,605) or placebo (n=2,633), in addition to their existing antidiabetic drugs. The patients completed their final clinical trial visits between November 2004 and January 2005. The investigators conducted an observational follow-up study to investigate the occurrence of cardiovascular outcomes and malignancies after the PROactive clinical trial period ended.4 Among 4,873 patients who completed the final PROactive trial visit, 73.9% enrolled in the observational study. Among patients who entered the follow-up study, the median total follow-up (clinical trial and observational periods) was 12.8 years. The imbalance in the number of bladder cancer cases observed in patients using pioglitazone during the clinical trial period (relative risk [RR]=2.83; 95% CI=1.02-7.85) did not persist during the combined 12.8-year period (HR=1.00; 95% CI=0.59-1.72).
Tuccori et al.2 conducted a retrospective cohort study that assessed the association between pioglitazone use and bladder cancer. The investigators used the United Kingdom Clinical Practice Research Datalink (CPRD) to identify a cohort of patients newly treated with antidiabetic drugs between January 1, 2000, and July 31, 2013, with follow-up until July 31, 2014. There were 145,806 patients (n=10,951 initiators of pioglitazone) with a mean follow-up of 4.7 years, during which 622 patients received a diagnosis of bladder cancer. Of these, 54 developed bladder cancer after pioglitazone exposure. The fully adjusted HR for bladder cancer with pioglitazone use compared with no thiazolidinedione use was 1.63 (95% CI=1.22-2.19). Statistically significant trends in the risk of bladder cancer were observed with increasing cumulative duration of use and cumulative dose of pioglitazone.
Overall, we conclude that pioglitazone may be associated with an increased risk in urinary bladder cancer, and we have updated the drug labels to include information about these additional studies.