A well-conducted meta-analysis has found that rosiglitazone and pioglitazone▼ approximately double the risk of fractures in women, but not men.
Health professionals should heed MHRA warnings and follow NICE guidance, and not commence or continue a glitazone in people at higher risk of fractures, in addition to considering the cardiovascular risks of glitazones
What is the background to this?
A risk of fracture with rosiglitazone seems to have been first noted in the ADOPT trial in 2006. Letters to health professionals were sent by the manufacturer of rosiglitazone and pioglitazone▼ in early 2007, and in October 2007 the MHRA drew attention to this in Drug Safety Update.
What does this study claim?
This meta-analysis of 10 randomised controlled trials (RCTs), each lasting at least one year, involving 13,715 patients found that the odds ratio (OR) for fractures with glitazones compared with control (placebo or other oral hypoglycaemic drugs) was 1.45 (95% confidence interval [CI] 1.18 to 1.79; P<0.001). Five RCTs (11,401 patients) provided data on fractures by sex and found a statistically significant increased risk of fractures among women (OR 2.23, 95%CI 1.65 to 3.01; P<0.001) but not among men (OR 1.00, 95%CI 0.73 to1.39; P=0.98).
This increased fracture risk seems clear and is another consideration relating to the use of glitazones, in addition to their cardiovascular risks. NICE recommends that glitazones should not be started or continued in people at higher risk of fractures; diabetes itself increases the risk of fracture. A limitation of this study is that the site of the fractures (wrist, hip, spine, etc) was not examined.
The authors of the study applied the increased risk to the baseline risk of fractures in three different age groups of women with diabetes, taken from clinical studies. Among women with an average age of 56 and recent diagnosis of diabetes, they estimated a number needed to harm (NNH) of 55 over one year (95%CI 34 to 103); that is, for every 55 women treated with a glitazone for one year, one extra woman develops a fracture who would not have done had they not received a glitazone. This is equivalent to an extra 18 (95%CI 10 to 29) fractures per 1000 patient years. Among older post-menopausal women (mean age 65 years) the NNH was 31 (95%CI 19 to 57), equivalent to an extra 32 (95%CI 18 to 53) fractures per 1000 patient years. Among an older cohort of women (mean age 72 years) the NNH was 21 (95%CI 14 to 39), equivalent to an extra 48 (95%CI 26 to 71) fractures per 1000 patient years.
Citation: Loke YK, Singh S, Furberg CD, et al. Long-term use of thiazolidinediones and fractures in type 2 diabetes: a meta-analysis. CMAJ 2009;180(1). DOI:10.1503/cmaj.080486 (Published early online, 10th December 2008)
Patients: 13,715 patients with type 2 diabetes or impaired glucose tolerance, from 10 RCTs of glitazones which lasted 52 weeks or more and which reported fracture outcomes.
Intervention and comparison: Rosiglitazone or pioglitazone▼ versus placebo or active comparator (other oral hypoglycaemic drugs e.g. metformin or sulphonylurea).
Outcome: the OR for fractures with glitazones compared with control (placebo or other oral hypoglycaemic drugs) was 1.45, (95%CI 1.18 to 1.79; P<0.001). Five RCTs (11,401 patients) provided data by sex and found a statistically significant increased risk of fractures among women (OR 2.23, 95%CI 1.65 to 3.01; P<0.001) but not among men (OR 1.00, 95%CI 0.73 to1.39; P=0.98).
Sponsorship: no external funding