Effects of Long-term Proton Pump Inhibitor Use on Bone Mineral Density

Introduction: Proton pump inhibitors (PPIs) have become one of the most widely prescribed medications and many patients remain on them chronically. Over the last decade, several studies revealed adverse associations with prolonged PPI use, in particular, increased risk of osteoporotic fractures. These studies, however, failed to strictly study patients who had dual-energy X-ray absorptiometry (DEXA) scans performed after at least 5 years of PPI use. A clear association between PPIs and a decrease in bone mineral density (BMD) has not fully established. Thus, our study will aim to determine if prolonged PPI use is associated with a decrease in BMD on DEXA. Methods: For each patient seen at our University Tertiary-care outpatient clinic from 1999 to 2017 on a PPI with a DEXA, we determined: (1) basic demographic information and medical history; (2) PPI name, dose, start and end date; (3) DEXA with t-scores of femoral neck, lumbar spine, right hip. DEXA scans must have been performed at least 60 months after onset of PPI use. Patients were excluded if they had medical history or medication use with known BMD effects such as hypothyroidism, steroids, end stage renal disease, SSRIs. Age and gender adjusted control cohort was also obtained. Results: Of 2516 patients seen, only 30 (27 female (90%), 3 male (10%), age 69±10) met the inclusion criteria for our study. Average BMI was 31.4±10 (range 24-45); 53% African American. Fifteen (48.4%) were on omeprazole, 8 (25.8%) esomeprazole, 4 (12.9%) pantoprazole, and 3(9.7%) on other PPIs. Duration of PPI use was 104±10 months (range 63-181). Interval to DEXA was 88±25 months (range 60-168). T-score lumbar spine was -0.51±1.56 versus age-gender adjusted control -1.18±1.24 (P=0.071). Right hip was -0.69± 1.24 versus control -0.86±0.92 (P=0.56). Femoral neck was -1.43±0.84 versus control -1.26±0.89 (P=0.46). Conclusion: We were unable to show a correlation between PPI use and BMD. This result may have been due to our unexpectedly small sample size with over 1500 patients being excluded due to concurrent medications or medical history with known effects on BMD. Surprisingly, DEXA scans were underutilized by clinicians in our predominantly elderly female population which significantly limited our sample size. We suggest a prospective study examine the possible cumulative effect of PPIs with underlying medical conditions or medications on BMDs to ascertain if this is the possible missing connection.