Bones and PPIs: Is the 'Evidence' for Harm Fracturing?

David A. Johnson, MD


November 13, 2012

The Relationship Between Proton Pump Inhibitor Use and Longitudinal Change in Bone Mineral Density: A Population-Based From the Canadian Multicentre Osteoporosis Study (CaMos)

Targownik LE, Leslie WD, Davison KS, et al
Am J Gastroenterol. 2012;107:1361-1369

Proton Pump Inhibitors and Bone Fractures

Literature analysis of the association between proton pump inhibitor (PPI) use and bone fractures reveals conflicting results. The earlier published reports linking PPI use to the development of hip fractures have been observational, case-control studies; therefore, they have greater potential for bias and less accurate incidence estimates.

Study Summary

The study by Targownik and colleagues involved a population-based sample of Canadians who underwent bone mineral density (BMD) testing of the femoral neck, total hip, and lumbar spine at baseline, and again 5 and 10 years later. In all, 8340 patients were included in the baseline analysis, with 4512 (55%) undergoing year-10 BMD testing. After adjusting for potential confounders, PPI use was associated with significantly lower baseline BMD at the femoral neck and total hip. However, the regression models did not demonstrate a significant association between continuous PPI use and the rate of change in BMD at any of the measurement sites from baseline to 5 years or 10 years. Moreover, there was no evidence of acceleration in the covariate-adjusted BMD loss at any measurement site after 5 and 10 years. Therefore, although PPI users had lower BMD at baseline than PPI nonusers, PPI use over 10 years was not associated with accelerated BMD loss.


Overall, the strength of association between PPI use and low BMD has been of low magnitude. Given that the estimates and even the upper bounds of most of the 95% confidence intervals (CI) of the odds ratios (OR) were well below 2, there is a strong possibility that these differences could have been a consequence of a channeling bias inherent in many observational studies. For example, a recent prospective study,[1] including 79,899 postmenopausal women from the Nurses' Health Study, showed that despite an OR of 1.36 (95% CI, 1.13-1.63) for PPI use, when accounting for a history of smoking (an independent risk for bone fracture) there was no significant association between PPI use and facture risk (OR, 1.06; 95% CI, 0.77-1.46). Furthermore, although the study showed an association between PPI use and hip fracture risk, there was no evidence to suggest a duration effect from long-term PPI use (OR, 1.30; 95% CI, 0.98-1.70) or short-term use (OR, 1.24; 95% CI, 1.19-1.28), even when looking at BMD. This suggests that the observed association was likely a consequence of confounding factors. 

The study by Targownik and colleagues provides even more convincing evidence to support the lack of harm. The finding of an element of risk at study entry, along with the finding that longer duration of exposure does not amplify this risk, is paradoxical. Accordingly, the data on bone density loss and osteoporotic fractures do not support discontinuing PPI therapy in patients taking PPIs for appropriate indications at appropriate doses. I recommend, however, that all patients be made aware of the data and that healthcare providers always review and reevaluate the decision for continued PPI therapy.