Increase in Vertebral Fracture Risk in Postmenopausal Women Using Omeprazole
Proton pump inhibitors are taken by millions of patients for prevention and treatment of gastroesophageal diseases. Case-control studies have suggested that use of omeprazole is associated with an increased risk of hip fractures. The aim of this prospective study was to assess the risk of vertebral fractures in postmenopausal women using omeprazole. We studied 1,211 postmenopausal women enrolled in the Osteoporosis and Ultrasound Study from the general population. Information on omeprazole and other risk factors for fractures including prevalent fractures and bone mineral density was obtained at baseline. Vertebral fractures were assessed on X-rays obtained at baseline and at the end of the 6-year follow-up and analyzed centrally. At baseline, 5% of this population was using omeprazole. Age-adjusted rates for vertebral fractures were 1.89 and 0.60 for 100 person-years for omeprazole users and nonusers, respectively (P = 0.009). In the multivariate analysis, omeprazole use was a significant and independent predictor of vertebral fractures (RR = 3.50, 95% CI 1.14–8.44). The other predictors were age higher than 65 years (RR = 2.34, 95% CI 1.02–5.34), prevalent vertebral fractures (RR = 3.62, 95% CI 1.63–8.08), and lumbar spine T score ≤ −2.5 (RR = 2.38, 95% CI 1.03–5.49). Omeprazole use is associated with an increased risk of vertebral fractures in postmenopausal women. Further studies are required to determine the mechanism of the association between the underlying gastric disease, omeprazole use, and risk of osteoporotic fractures.
KeywordsOsteoporosis Proton pump inhibitor Vertebral fracture Postmenopausal women
Proton pump inhibitors (PPIs) are taken by millions of patients for gastroesophageal reflux disease  and prevention of gastroduodenal complications associated with the use of nonsteroidal anti-inflammatory drugs (NSAIDs) [2, 3, 4, 5, 6]. The efficacy of PPIs for this indication has been established in prospective comparative studies [3, 4], and the potential severity of gastrointestinal complications of NSAIDs has led to recommendations for concomitant administration of PPIs in at-risk patients [7, 8]. Moreover, there is an increased use of PPIs in patients taking oral bisphosphonates, which are effective antiosteoporotic drugs but can be responsible for dyspepsia .
Laboratory studies have shown that a PPI, omeprazole, can decrease bone resorption through a direct effect on osteoclast activity [10, 11, 12]. The vacuolar type H+-ATPases are present in high concentration on the ruffled border of the resorbing osteoclasts, and this proton pump extrudes protons during the bone-resorption process. This allows maintenance of the acidic environment required for mineral solubilization and hydrolysis of bone matrix . Moreover, administration of a selective inhibitor of the osteoclastic V–H+-ATPase is able to prevent bone loss in ovariectomized rats, which is an animal model representative of postmenopausal osteoporosis . These findings raise the possibility that PPIs may prevent osteoporosis and fractures.
In contrast with these experimental data, two case-control studies suggested that PPI therapy is associated with an increased risk of hip fractures [15, 16]. The number of hip fractures and other fractures is increasing dramatically because of the increasing number of osteoporotic patients . Excess mortality and morbidity related to osteoporotic fractures are well recognized and add to the economic cost burden . Thus, screening for risk factors for osteoporosis and fractures is recommended. Because of their widespread use, the potentially increased risk of fractures related to PPIs must be confirmed prospectively.
The aim of this study was to assess the association between exposure to PPI therapy and risk of fractures in a prospective study of postmenopausal women.
Patients and Methods
The Osteoporosis and Ultrasound Study (OPUS) is a multicenter prospective study of risk factors for fractures in postmenopausal women; it has been described in detail elsewhere , and the current investigation is a substudy of OPUS. The study population consisted of ambulatory European women recruited in five European centers from random population samples between 1999 and 2001 and followed for a mean of 6.1 years, as of March 2007. The initial OPUS study population consisted of 2,389 women 55–79 years of age, without any disorders precluding bone mineral density (BMD) measurements and general and cognitive inability that precluded completing questionnaire. For the current study, data of one center (n = 437 patients) were excluded as baseline information on omeprazole was not available. Human subject review committees at each participating intitution reviewed and approved the study.
Each participant filled out a questionnaire, which was a modified version of the European Vertebral Osteoporosis Study risk factor questionnaire  and validated in English, German, and French. It includes demographic information and a complete health history. Participants were asked about family history of osteoporosis, previous falls and fractures, medications, and nutritional and lifestyle aspects. Weight and height were measured at baseline, and body mass index (BMI, kg/m2) was calculated. BMD of the lumbar spine and the proximal femur was measured using dual-energy X-ray absorptiometry (DXA) in posterior–anterior (Hologic QDR-4500; Hologic, Bedford, MA) or in anterior–posterior (Lunar Expert devices; GE Lunar, Madison, WI) projection using standardized procedures and centralized quality control.
Vertebral fracture status was determined on lumbar and thoracic spine X-rays performed using a standardized procedure identical in all centers and a standardized assessment in a central facility.
Radiographs were performed at baseline and final visits using the same procedures and evaluated centrally by two radiologists. The procedure to assess fracture status was based on a semiquantitative reading. Vertebrae with deformities of nonosteoporotic origin (e.g., degenerative changes) or exhibiting potentially misleading appearances were not considered as fracture. Nonvertebral fracture information was collected from questionnaires.
Omeprazole Exposure and Potential Confounders
Participants were asked about treatments, including omeprazole, taken at baseline and during the previous year. Data on the use of gastrointestinal medications other than omeprazole for the same indication were not collected.
The following baseline covariates were ascertained: current use of thiazide diuretics, corticosteroids, thyroid hormone supplementation, calcium, vitamin D, and hormone replacement therapy (HRT). For this study we excluded women without baseline data on omeprazole (n = 24), women with baseline use of bisphosphonate or raloxifene, and those without X-ray follow-up.
Thus, a population of 1,211 postmenopausal women was used as the basis of the analysis of vertebral fracture risk and 1,364 for the analysis of nonvertebral fracture risk.
The characteristics of women taking omeprazole at baseline were compared with those of women not taking omeprazole using chi-squared tests or Fisher’s exact tests as appropriate. Numbers of incident fractures were compared in these two groups; incidence was calculated as (number of subjects with an incident fracture/total number of subjects)/time of follow-up, to obtain an indicative incidence in patient-years. A multivariate stepwise logistic regression model was then used to explain incident fractures (yes/no), controlling for potential confounding factors; all characteristics with a difference at baseline, using P < 0.20 as entry criterion, were entered into the model. Statistics were analyzed using the Statistical Analysis System (SAS Institute, Cary, NC) version 9.0, including the PROC LOGISTIC option.
Baseline characteristics of patients
65.9 ± 6.3
65.8 ± 6.6
Age groups (years), n (%)
BMI (mean ± SD)
28.5 ± 5.5
27 ± 4.4
History of clinical low-trauma fracture, n (%)
Prevalent vertebral fractures, n (%)
Breast cancer, n (%)
Gastric surgery, n (%)
Current use of
Steroids, n (%)
Thiazide, n (%)
Thyroid hormone, n (%)
HRT, n (%)
Calcium, vitamin D, n (%)
Cigarette smoking, n (%)
Alcohol (≥15 units/week), n (%)
Walk for exercise, n (%)
Self-reported falls in the past 12 months, n (%)
Self-reported health poor to fair, n (%)
Baseline lumbar spine T score (mean ± SD)
−1.19 ± 1.6
−0.81 ± 1.5
Baseline hip T score (mean ± SD)
−0.60 ± 1.3
−0.46 ± 1.1
Spine and/or hip T score < −1, n (%)
Spine and/or hip T score ≤ −2.5, n (%)
Serum CTX (mean ± SD)
0.24 ± 0.19
0.22 ± 0.16
Patients with falls in the previous year, n (%)
During follow-up, 49 (4.05%) patients had at least one incident vertebral fracture. Numbers of fractures were 1, 2, 3, and 4 in 38, 4, 5, and 2 patients, respectively. Age-adjusted rates for vertebral fractures were 1.89 and 0.60 for 100 person-years for omeprazole users and nonusers, respectively (P = 0.009). The age-adjusted relative risk (RR) of incident vertebral fractures for omeprazole users was 3.41 (95% confidence interval [CI] 1.45–8.01), P = 0.005.
In the multivariate analysis, omeprazole use was a significant and independent predictor of vertebral fracture (RR = 3.10, 95% CI 1.14–8.44, P = 0.027), and the other predictors were age higher than 65 years (RR = 2.34, 95% CI 1.02–5.34, P = 0.044), prevalent vertebral fractures (RR = 3.62, 95% CI 1.63–8.08, P = 0.001), and lumbar spine BMD T score <−2.5 (RR = 2.38, 95% CI 1.03–5.49, P = 0.04).
During follow-up, among 1,364 patients with available nonvertebral fracture data, 156 had a nonvertebral fracture, including 11 with hip fracture. The incidence was 12.68% and 11.37% (P = 0.73) in the omeprazole and control groups, respectively.
This is the first prospective study showing that use of omeprazole is associated with an increased risk for vertebral fracture in postmenopausal women. Omeprazole use was independent of other well-known risk factors, such as prevalent fractures, low BMD, and age. Moreover, the link between omeprazole use and incidence of vertebral fractures was strong, with a magnitude similar to that estimated for prevalent vertebral fractures.
These results are in line with data from case-control studies. Using the General Practice Research Database, comparison of 13,556 hip fractures cases and 135,386 controls shows that the use of PPI therapy for more than 1 year increases the risk of hip fracture, with an odds ratio (OR) of 1.44 (95% CI 1.30–1.59) . In a Danish study, PPI therapy was associated with an increased risk of both hip (OR = 1.45, 95% CI 1.28–1.65) and vertebral (OR = 1.60, 95% CI 1.25–2.04) fractures .
These results might appear unexpected with regard to the mechanism of action of PPIs. However, the proton pump inhibition action of the PPIs is site-specific and, while omeprazole inhibits the gastric proton pump, the inhibition of osteoclast H+-ATPase activity is much less pronounced . Part of the site specificity of the enzyme activity is related to its general structure, and different subunit isoforms play a role in the localization and activity of the proton pumps, as shown for bone and kidney . These observations have therapeutic implications as H+-ATPase inhibitors have a potential for being antiresorptive drugs, providing that they can act on bone-specific targets rather than on enzyme subunits with widespread expression .
It is expected that patients receiving acid-suppressive therapies have more comorbidities than others. Indeed, at baseline, omeprazole users were different from nonusers. They had a higher BMI, a lower alcohol intake, and a higher thiazide use; actually these three parameters may be considered as protective against osteoporotic risk. Omeprazole patients were not different for corticosteroid intake or previous gastric surgery, and these baseline characteristics cannot explain the observed effect of omeprazole on incident vertebral fractures. The main result is that bone status was different at baseline, with omeprazole users having more prevalent clinical and vertebral fractures and a trend for a lower spine BMD. There was no difference in hip BMD between users and nonusers, as previously shown in the Study of Osteoporotic Fractures (SOF) . There were no BMD data in the two large case-control studies recently published [15, 16]. In the SOF, hip BMD did not change during follow-up, although nonspine fracture risk was increased by 18% . Definite conclusions on any detrimental BMD effects of omeprazole cannot be drawn from these data.
The link between omeprazole therapy and osteoporosis may be the decrease in calcium absorption, related to acid suppression in the stomach. In rats, omeprazole has been associated with malabsorption of calcium phosphate and impaired BMD . Modifying the acidic environment in the gastrointestinal tract may change the calcium solubility. In women older than 65 years, omeprazole reduces the absorption of calcium carbonate taken under fasting conditions . In normal subjects, omeprazole significantly reduces calcium and phosphate urinary excretion after an oral load, and this gastric acid inhibition is more effective at reducing calcium rather than phosphate dietary salt absorption . In contrast with these data, it has been shown that changing the gastric pH by the use of omeprazole does not modify the net intestinal absorption of several minerals from food, and there is no evidence of deleterious effects of moderate hypochlorhydria . Moreover, the effect of calcium deficiency on the risk of fracture is controversial [28, 29, 30]. The link between low calcium intake and fractures needs the assessment of other intakes and nutrients. A long-lasting potential consequence of long-term use of omeprazole is the decrease in vitamin B12 store , which may be associated with low BMD and fractures [32, 33]. Use of omeprazole may have an indirect effect on bone metabolism through the decrease in absorption of folates, riboflavin, and vitamin B12, which in turn influences plasma total homocysteine, the level of which is associated with hip fracture risk and reduced BMD . Finally, the diet of omeprazole users may be different because of intolerances secondary to gastritis as suggested by our result on alcohol intake. The dietary acid load can act on bone metabolism [35, 36]. However, we do not have data about food tolerance or unhealthy lifestyle in our patients. Finally, the bone alteration may be related to the underlying gastric disease. The prevalence of Helicobacter pylori infection has been found to be high in male patients with osteoporosis , and this infection can be present in patients taking PPIs. Such an infection can induce a permanent inflammatory response, i.e. a gastric and systemic increase in indexes of inflammation. This mechanism has been suggested to explain some extragastric manifestations in H. pylori-infected patients . Experimental data suggest that gastric parietal cells have a potent endocrine role in secreting estrogens [38, 39]; atrophy of the gastric mucosa, which is observed in patients infected with cag A-positive H. pylori , leads to a reduced number of gastric parietal cells and may lead to decreased local production of estrogens. In both female and male rats, estrogens produced in the stomach directly induce expression and production of ghrelin , which is produced by epithelial cells lining the fundus of the stomach and is a ligand for the growth hormone secretagogue receptor. Experimental data indicate an effect of ghrelin on osteoblasts, leading to an increase in bone formation ; but in the general population of older adults, there is no current evidence for an influence of ghrelin on bone metabolism . Thus, the relevance of potential endocrine activities of gastric cells on bone metabolism is unknown.
H2 receptor antagonists (H2RAs) are prescribed for the same indications as PPIs. In rats, these drugs have antiresorptive properties, and cimetidine has been shown to prevent osteoclast differentiation induced by histamine [44, 45]. Patients with chronic use of H2RAs do not demonstrate low BMD , and fracture risk in such patients is unclear, either decreased  or increased [16, 47]. However, long-term therapy with H2RA seems to be associated with a lower risk of fracture than long-term therapy with PPIs . Information on H2RA use was not available in the OPUS.
Our study has several limitations. Data were obtained in postmenopausal women only and do not apply to men or premenopausal women. The prevalence of omeprazole use and the incidence of fractures were low. We could not account for previous use of the treatment, dose, and duration of exposure. Moreover, we did not seek information on changes in omeprazole intake during follow-up as the exposure was not longitudinally collected. Duration of omeprazole treatement is a strong determinant of the induced risk of fracture. Vestergaard et al.  did not observe a duration effect, but the maximum duration of potential PPI exposure was 5 years, in contrast with the study of Yang et al. , with a maximum duration of exposure of 15 years, who observed a strong association between increasing duration of use and increasing risk of hip fracture. Moreover, Targownik et al.  showed recently that increased risk of osteoporosis-related fractures is observed for duration of omeprazole exposure exceeding 5 years. The high prevalence of osteoporosis and fractures in patients receiving omeprazole in our study may be related to long-term use before baseline evaluation, but the methodology of our study does not allow confirmation of this hypothesis.
The strengths of our study include the sample size, the recruitment of subjects in the general population, and the assessment of vertebral fractures on standardized X-rays with central analysis. The prospective methodology of our study allows adequate assessment of potential confounders including the main risk factors for fractures. Further studies are required to determine the mechanism of the association between the underlying gastric disease, omeprazole use, and risk of osteoporotic fractures. Moreover, the potential effect of these data on antifracture efficacy of bisphosphonates  and other antiosteoporotic drugs warrants attention.
We acknowledge the contributions of the other members of the OPUS teams at the five participating centers: Rosie Reid, Lana Gibson, and Alison Stewart in Aberdeen; Antonia Gerwinn, Dr. Maren Glüer, Roswitha John, Roswitha Marunde-Ott, Monika Mohr, Regina Schlenger, Pia Zschoche, Dr. Reinhard Barkmann, Dr. Carsten Liess, Carsten Rose, and Wolfram Timm in Kiel; Therese Kolta and Nathalie Delfau in Paris; Margaret Paggiosi, Nicky Peel, Diane Shutt, Anne Stapleton, and Debbie Swindell in Sheffield. This project would not have been possible without the financial support of our sponsors: Aventis, Eli Lilly, Novartis, Procter & Gamble Pharmaceuticals, and Roche. We also thank the equipment manufacturers: DMS, IGEA, OSI/Osteometer Meditech, Quidel/metra.
- 5.Ekström P, Carling L, Wetterhus S, Wingren PE, Anker-Hansen O, Lundegardh G et al (1996) Prevention of peptic ulcer and dyspeptic symptoms with omeprazole in patients receiving continous non-steroidal anti-inflammatory drug therapy. A nordic multicentre study. Scand J Gastroenterol 31:753–758PubMedCrossRefGoogle Scholar
- 23.Yu EW, Shinoff C, Blackwell T, Ensrud K, Hillier T, Bauer DC (2006) Use of acid-suppressive medications and risk of bone loss and fracture in postmenopausal women. J Bone Miner Res 21(Suppl 1):S281Google Scholar
- 29.Grant AM, Avenell A, Campbell MK, McDonald AM, MacLennan GS, McPherson GC et al (2005) Oral vitamin D3 and calcium for secondary prevention of low-trauma fractures in elderly people (Randomised Evaluation of Calcium or vitamin D, RECORD): a randomised placebo-controlled trial. Lancet 365:1621–1628PubMedCrossRefGoogle Scholar
- 35.Hostmark AT (2003) Dietary acid load and self reported prevalence of hip fractures: a population based cross sectional study. Norsk Epidemiologi 13:193–198Google Scholar
- 36.Henriksen DB, Hartman B, Jeppesen PB, Miholic J, Christiansen C, Holst JJ (2004) The postprandial bone resorption process is regulated by gastrointestinal signal. J Bone Miner Res 19(Suppl 1):S92Google Scholar
- 49.de Vries F, Cooper A, Logan R, Cockle S, van Staa T, Cooper C (2007) Fracture risk in patients receiving concomitant bisphosphonate and acid-suppressive medication or bisphosphonate alone. Osteoporos Int 18(Suppl 3):S261Google Scholar