Proton Pump Inhibitors Adversely Induce Selective Changes in the Bone Mineral Density Detected by Dual-Energy X-ray Absorptiometry in Postmenopausal Women

Dual-energy X-ray absorptiometry (DEXA) is a universal tool that can detect bone loss and diagnose osteoporosis. Long term of using certain drugs contributed to the etiopathology of bone loss. Proton pump inhibitors (PPIs) users are at risk of developing osteoporosis. This study aimed to prove the selectivity of PPIs in inducing bone loss in postmenopausal women by determining the T-score of the axial spine and femur bone. A total number of 215 menopausal womenwere recruited from a Teaching hospital and private clinics from August 2018 to November 2019. The participants were grouped into Group I, n=150 (had no PPIs treatment) and Group II, n=65 (had treatment with PPIs). All the participants were subjected to DEXA investigation. Group II patients showed signi icantly lower T-score of the femur bone, while Group I patients showed a signi icantly lower T-score of lumbar vertebrae. The percentage of Group II patients had a T-score – 2.5 in femur ward bone is 35.4%, while the percentage of Group I patients had a T score -2.5 in the lumbar-3 vertebrae is 35.3%. Moreover, PPIs users showed an acceleration of bone loss despite the age, duration of menopause, bodymass index, andwaist-to-height ratio. We conclude that PPIs users are at risk of developing bone mass loss in the femur more than in the lumbar vertebrae

Moreover, the effect of PPIs on the BMD is usually reversible (Corley et al., 2010). Current studies claimed that hypochlorite due to PPI is the cause of bone fractures as hypohydrochloria and high gastrin levels reduced the mineral absorption and thereby producing a defect in the bone remodelling (Thong et al., 2019). Some authors claimed that PPIs directly altered bone regeneration and osseointegration (Mester et al., 2019). Brozek et al. studied the link between subsequent fractures with the PPIs doses and found low de ined daily dose of PPIs (<90 mg cumulative doses) is not linked with increased risk of later fracture (Brozek et al., 2019).
The effect of PPIs on the BMD in postmenopausal women without earlier or current evidence of bone fracture was not investigated in the previous studies. Moreover, there is no study investigated the direct effect of PPIs on the speci ic bones. Therefore, this cross-sectional study was carried on postmenopausal women using PPIs, and they had no clinical history of bone fracture looking for changes in the BMD which detected by DEXA investigation.

MATERIALS AND METHODS
The present study is an observational crosssectional study carried on patients treated regularly with a proton pump inhibitors for different medical conditions. This study carried on at the University of Anbar, College of Medicine/ Department of Pharmacology in collaboration with the private clinics and the Teaching Hospital in Anbar city-Iraq from August 2018 till February 2020.
The Ethical and Scienti ic Committees of the College of Medicine reviewed the interventions, investigations, and approved the study, according to the guidelines of investigations of osteoporosis patients. The patients are free to refuse participation or withdraw from the study at any time of the study.
The criteria of inclusion were postmenopausal women aged ≥ 50 years subjected to DEXA investigation. Their pro ile was assessed to get information about age, duration of menopause, smoking, concomitant diseases, and anthropometric measurements, and using PPIs. Patients with a history of earlier or current fractures, use of medicines (such as glucocorticosteroids, insulin sensitizers, anticoagulants, selective serotonin reuptake inhibitors), and terminal illnesses were excluded.
The patients were grouped into patients had no treatment with PPIs (Group I), and patients had treated with PPIs (Group II). Group II patients had a history of regular treatment with PPIs, including  The results are expressed as number (percentage) of T-score at a threshold of -2.5 and as median (interquartile range) of T-score. Pvalue was calculated by using a Chi-square test for categorized data. GroupI: patients had not treated with proton-pump inhibitors, Group II Patients had treated with proton pump inhibitors.
omeprazole, lansoprazole, and esomeprazole with different therapeutic regimen. A cutoff value of Tscore -2.5 indicated evidence of osteoporosis.

Statistical Analysis
The results are expressed as a number, percentage, interquartile range and mean ± S.D. Difference between means of two groups of continuous data was analyzed using a two-tailed independent twosample t-test. The Chi-square test analyzed categorized data, and Pearson's correlation test did the correlations between age, duration of menopause, and anthropometric measurements with DEXA data. The P-value of ≤ 0.05 is the cutoff level of signi icance. Excel software (2010) program was used for data analyses. Table 1 showed that there are signi icant differences between Group I and II in the characteristics of the patients. The mean of Group II patients was lower than the corresponding mean of the Group I. There is no signi icant difference in the duration of the menopause, BMI, and WHeR between Groups I and II. Signi icantly higher percentages of smokers and patients presented with rheumatic illnesses and peptic ulcers were observed in Group II compared with the corresponding values of the Group I. The percentages of concomitant illnesses, including diabetes mellitus and hypertension, were signi icantly The results are expressed as a correlation factor (above value) and the p-value(below the value) in each cell. Group I: patients had not treated with proton-pump inhibitors, Group II Patients had treated with proton pump inhibitors. higher in Group I compared with the corresponding Group II. Table 2 shows signi icant differences between Groups I and II in the T-scores of the BMD. The percentages of Group I patients with T-score -2.5 were higher in the lumbar region (L2, L3, L1-L3, L2-L3) compared with the corresponding values of the Group II patients. The signi icantly higher percentages of patients who have T-score -2.5 in the femur were observed in Group II.

Relationship between T score and risk factors
In the Group I patients, T-scores of BMD of lumbar and femur regions were correlated signi icantly and inversely with the age and the duration of the menopause as signi icant positive correlations with the anthropometric measurements (BMI and WHeR) were observed (Table 3). In Group II, the correlations between the T-scores of the BMD at lumbar and femur regions were non-signi icant. As with Group I, the T-scores of the BMD at lumbar and femur regions were signi icantly and positively correlated with the anthropometric measurements in the Group II patients. Signi icant positive correlations between the T-scores and the BMD at L1, L2 and L1-L2 regions were observed in Group II. At the same time, there is a signi icant inverse correlation between T-score of BMD at a trochanter region of the femur with the duration of the menopause.

DISCUSSION
The present study shows that PPIs users have a signi icantly lower T score of femur bone compared with non-PPIs users who had a signi icant T score of lumbar spines. The percentage of patients who have T-scores that indicated osteoporosis is higher in femur bone (Group II) and lumbar vertebrae (Group I). Baseline data showed that the age Group II patients were less than the corresponding value of Group I, indicating that the osteoporotic process is accelerated in Group II. Ageing is a signi icant risk factor for the decline in the BMD of the total hip (Westbury et al., 2020). The percentage of current smokers among Group II is signi icantly higher than the corresponding percentage of the Group I patients, indicating that smoking is a confounding risk factor of osteoporosis among Group II patients. In the cohort study carried in the male population, smokers showed a signi icantly high percentage of hips and vertebrae bone fractures (Cho et al., 2020). Chronic concomitant diseases, e.g. diabetes, hypertension, and ischemic heart disease are not confounding risk factors of osteoporosis in the Group II patients as the percentages of these illnesses are less than the corresponding values of Group I. Hypertension Per se is not a risk factor of bone mass loss, but using thiazides or beta-blockers can significantly decrease the BMD (Hijazi and Alour i, 2020). Signi icantly higher percentages of peptic ulcer and rheumatic illnesses in the Group II patients link to the prescription of PPIs as therapeutic and prophylactic remedies. Anthropometric measurements of our patients indicated that the patients are overweight in both groups. This inding does not agree with earlier studies, which showed that there is an inverse relationship between BMI and BMD (Hijazi and Alour i, 2020). Moreover, our observation of higher WHeR (>0.5) in both groups is in agreement with other studies that showed a signi icant inverse correlation between WHeR and BMD (Jafari-Adli et al., 2019). Table 2 shows that Group II patients have signi icantly lower T scores of femur bones compared with corresponding values of Group I, and vice versa with the lumbar vertebrae. This observation indicates that PPIs induced selective adverse effects against bone minerals in the femur, and this inding agreed other earlier studies (Fattahi et al., 2019). Moreover, people had used PPIs more frequently are at risk of atypical femur fractures (Buitendijk et al., 2019). The correlations between the T-scores with age, duration of menopause, and body mass index in both groups do no show speci ic differences between Group I and II. Moreover, the positive correlation between WHeR with the T-score of the spine, indicating that WHeR is a signi icant risk factor of losing body mass in the spine. The T-score of the femur trochanter is signi icantly correlated with WHeR in Group II patients, indicating that PPIs accelerate bone loss in the femur bone. Previous studies showed that not all PPIs induced bone loss as (Bahtiri et al., 2016) reported that esomeprazole is independently reduced the bone mass of lumbar spine and femur, while omeprazole does not show a signi icant effect.

Study limitations
Risk factors of lower T score that reported in this study do no impact our indings, while the results of each PPIs user and the duration of using PPI may limit our indings which are far away from goals of this study. Different generic names of PPIs that used by patients may limit the results because not all PPIs induced bone loss to the same level.

CONCLUSIONS
We conclude that PPIs users are at risk of bone loss in the femur and osteoporosis rather than the axial bones in postmenopausal women.