SLR - November 2009 - Sara Karamloo
Reference:
Douglas, I.J., Evans, S.J., Pocock, S., Smeeth, L. (2009). The Risk of Fractures Associated with Thiazolidinediones: A Self-controlled Case-Series Study. Public Library of Science Medicine, 6, e1000154.
Scientific Literature Review
Reviewed by: Sara Karamloo, DPM
Residency Program: Yale New Haven / VA CT
Podiatric Relevance:
This study investigates the risk of fracture, including those of the foot, in patients prescribed thiazolidinediones, an anti-diabetic class of drugs.
Methods:
The General Practice Research Database from the UK contains information including consultations, diagnoses, prescribed medications, and demographic data on 6 million patients. It undergoes thorough evaluations to ensure reliability of the information used to conduct research. In this study, patients with both a diagnosis of fracture and a first recorded thiazolidinedione prescriptionwere studied as part of a self-controlled case-series, utilizing a Poisson distribution. For each patient, the length of exposure to the mediciation and follow-up time were examined, and fractures were classified with regard to anatomical site. Gender differences, the differential effects of rosiglitazone and pioglitazone, exposure duration, and the effect on specific fracture site were further examined.
Results:
The GPRD identified 1,356 patients prescribed rosiglitazone, 389 prescribed pioglitazone and 74 prescribed both at different times; all with at least one fracture. The mean age was 65.4 at first thiazolidinedione exposure. There were 905 patients with an arm, foot, wrist, or hand fracture, 150 with hip fractures, 66 with spine fractures and 733 patients with fractures at other sites. Those with hip and spine fractures were older at first thiazolidinedione prescription and rate of fracture increased approximately 5% annually. When comparing periods of exposure, the adjusted rate ratio for the association between thiazolidinediones and fracture was 95%, with no relation to gender. The duration of thiazolidinedione exposure, however, accounted for an increase in the rate ratio for all fractures, increasing from 1.26 during the first year to 2.00 during years four to seven. With respect to the site of the fracture, thiazolidinedione users had a rate ratio of 1.28 for arm, wrist, hand, or foot fractures, 2.09 for hip fractures, and 2.72 for spine fractures. When comparing the two drugs, a test for interaction revealed no evidence that the effect varied by type (p= 0.47) and rate ratios for fractures of specific sites were also similar.
Conclusions:
This study reveals an increased risk of fractures in male and female patients taking thiazolidinediones, both rosiglitazone and pioglitazone. This is an important conclusion considering the vast number of diabetics with foot and ankle pathology on these medications.