Fracture Risk and Bone Mineral Density Reduction Associated with Proton Pump Inhibitors

Yuen Ting Lau, Pharm.D.; Nasiya N. Ahmed, M.D.


Pharmacotherapy. 2012;32(1):67-79. 

In This Article

Risk of Bone Mineral Density Reduction

Two previously discussed cohort studies also collected data on BMD changes associated with PPI use. In the WHI study, BMD at the total hip, spine, and total body was measured in 10,833 patients from three study centers.[11] Data were available in 61% of the participants for this analysis. The baseline BMD at any site was not different between the PPI users and nonusers, regardless of the duration of PPI use. Although the BMD in year 3 increased from baseline in both PPI users and nonusers because of concurrent calcium and vitamin D supplementation, the nonusers had a slightly larger increase in 3-year hip BMD than did the PPI users (0.74%, 95% CI 0.01–1.51%). However, when the duration of PPI therapy was included in the analysis, the difference in BMD change at the hip was no longer statistically significant (p=0.43). The 3-year change in BMD at the spine and total body was not affected by PPI exposure (p=0.94 and p=0.08, respectively).

In the SOF and MrOS studies, the total hip BMD was measured at baseline and at a second visit, which was completed in 54% of the patients in SOF and 74% of the patients in MrOS at an average of 4 years after the baseline visit.[12] The baseline BMD was significantly lower in PPI users in MrOS (p=0.05) and was not different between PPI users and nonusers in SOF (p=0.37). The annualized percent change of BMD was not different in SOF (p=0.08) or MrOS (p=0.39) between PPI users and nonusers.

A third study used the Canadian Manitoba Bone Mineral Density Database to evaluate the risk of osteoporosis and BMD decline associated with PPI exposure.[18] To assess the risk of osteoporosis, patients with T scores of −2.5 or lower at the hip or lumbar region during 2000–2007 were identified. Cases of hip (2193 patients) and lumbar (3956 patients) osteoporosis were analyzed separately, and each case was matched with up to three controls who had T scores of −1.0 or greater at the respective site. Compared with no PPI use, any cumulative dose of a PPI over 5 years was not associated with an increased risk of hip or lumbar spine osteoporosis. The AORs for hip osteoporosis were 1.00 (95% CI 0.82–1.22), 1.02 (95% CI 0.69–1.50), and 0.86 (95% CI 0.55–1.34) associated with PPI exposure of less than 750 doses, 750–1500 doses, and 1500 doses or more over 5 years, respectively. Similar results were found with the risk of lumbar spine osteoporosis. In addition, those who underwent two BMD assessments between 2001 and 2006 (2549 subjects) were evaluated in a longitudinal arm of the study. Subjects who had an average standard PPI daily dose of more than 0.5 dose/day or more than 1.0 dose/day between the two BMD assessments were compared with those who had an average standard daily dose of less than 0.5 dose/day. Standard daily doses were defined as omeprazole 20 mg, rabeprazole 20 mg, esomeprazole 20 mg, lansoprazole 30 mg, and pantoprazole 40 mg. Use of a PPI at either dosage did not affect the annualized rate of decline in BMD at the lumbar spine or the total hip (p>0.2 for all comparisons).