Rehabilitation of pioglitazone

Abstract

Background In 2012 an editorial in the British Medical Journal stated that “it can confidently be assumed that pioglitazone increases the risk of bladder cancer”.1 Yet now, the recently announced results of a 10-year study mandated by the FDA have failed to demonstrate any association between pioglitazone and bladder cancer2 and, because of its many beneficial effects on glucose homeostasis and potential cardiovascular protective effects, the place of pioglitazone in the treatment of diabetes warrants reconsideration. During pre-clinical studies, an excess of bladder cancers were found in male but not female rats treated with pioglitazone.3 Of note, these bladder cancers could be prevented by acidification of the urine which prevents pioglitazone crystal formation.4 As a result of the findings in rats, the FDA requested a 10-year study of pioglitazone in humans to assess safety with regard to bladder cancer.5 The 8-year data have been published online6 and the 10-year results recently were made public.2 The main results of this study fail to show any association between pioglitazone and risk of bladder cancer.7 Another large, recently reported study involving six populations, including 1.01 million diabetic individuals from six countries across the world, has come to the same conclusion.8 Previous studies suggesting a link between pioglitazone and bladder cancer have been re-examined.7 The link between pioglitazone and bladder cancer in many of these retrospective observational studies is likely to be explained by the fact that patients treated with pioglitazone in the various databases were different from those not treated with pioglitazone, with whom they were compared, i.e. the pioglitazone-treated patients already were at higher risk of bladder cancer from other causes.7 Importantly, major risk factors for bladder cancer, i.e. smoking and proteinuria, were not available for most of these retrospective analyses. Initial concern about a potential link between pioglitazone and bladder cancer was derived from the PROactive study, where an apparent excess of bladder cancers was observed for pioglitazone (14) versus placebo (6, p=NS).7,9-11 The PROactive study investigators concluded that, because most of the bladder cancers occurred during the first year following initiation of pioglitazone therapy, the drug could not plausibly be related to the development of bladder cancer.9 Further, the total numbers of bladder cancers (n=20) was small, making it difficult to draw any meaningful conclusion about the statistically insignificant difference between the treatment groups. It has been suggested that pioglitazone might be a tumour promoter and in this way caused the excess of bladder cancer during the first year of PROactive,12 although there is no experimental evidence to support such an effect of pioglitazone. The actual data regarding the number of months into the trial when these bladder cancer cases were diagnosed has been published: most appeared so early into the trial (two cases were diagnosed 13 and 14 days into the trial respectively, one at one month, another at three months and a fifth at four months) that they could not possibly have been related to pioglitazone treatment.10,11 Links between pioglitazone and bladder cancer in meta-analyses of randomised controlled trials depend entirely on inclusion of these bladder cancer cases in the first year of PROactive, which we now know could not have been related to pioglitazone.10,11 Lastly, and most importantly, the 6-year follow up data of PROactive have been published.13 After 6 years there were 23 cases of bladder cancer in the pioglitazone-treated group versus 22 cases in the placebo-treated group. Thus, in 2015, it is highly unlikely that there is any link between pioglitazone and bladder cancer at all in humans.7,10,11