On the association between statin and fracture: A Bayesian consideration
Received 1 June 2006; received in revised form 22 August 2006; accepted 9 November 2006. published online 22 December 2006.
Abstract
Background
The association between statin use and fracture risk is controversial, due to conflicting findings from previous studies. This study utilized the Bayesian approach to combine existing evidence and update the association with consideration of potential bias.
Methods
Data on the association between statin use and fracture incidence from 11 observational studies and 4 RCTs were synthesized by both empirical Bayesian analysis and fully Bayesian random-effects meta-analysis models.
Results
Empirical Bayesian analysis showed that statin use was associated with a reduction in hip fracture risk (OR=0.57, 95% credible interval (CrI): 0.46–0.71) and for non-vertebral (OR=0.69, 95% CrI, 0.63–0.74). These results were comparable with results from the fully Bayesian random-effects meta-analysis only for hip fracture (OR 0.56, 95% CrI, 0.42–0.73), but not for non-vertebral fracture (OR 0.77, 95% CrI, 0.58–1.03). The probability that statin use reduces fracture risk by at least 20% was 0.995 for hip fracture and 0.61 for non-vertebral fracture. Under the assumption that bias over-estimates the true OR by 20%, there is still a probability of 0.97 that statin use reduces hip fracture risk by at least 20%; however, the effect on non-vertebral fracture was much less robust with a probability of 0.27.
Conclusions
Results of this Bayesian consideration are highly consistent with the hypothesis that statin use reduces hip fracture, but the association between statin use and non-vertebral fracture remains uncertain. The Bayesian approach presented here has the ability to help updating existing evidence as new data becomes available.
Bone and Mineral Research Program, Garvan Institute of Medical Research, St Vincent's Hospital, 384 Victoria Street, Darlinghurst, Sydney, NSW 2010, Australia
Faculty of Medicine, University of New South Wales, Sydney, NSW, Australia
ABSTRACT