Association Between Obstructive Sleep Apnea Severity, Nocturnal Hypoxemia, and DXA T-Scores in Postmenopausal Women: A Retrospective Cross-Sectional Study

Aim: Obstructive sleep apnea (OSA) causes intermittent hypoxemia and sleep fragmentation, both of which may affect bone metabolism, yet studies of OSA and bone mineral density (BMD) have produced inconsistent results. We examined whether OSA status and conventional polysomnographic indices of OSA severity and nocturnal oxygenation were associated with lumbar spine and total hip T-scores in postmenopausal women. Methods: We conducted a retrospective cross-sectional study of 101 postmenopausal women with polysomnography (PSG) and dual-energy X-ray absorptiometry (DXA) records; 40 had PSG-confirmed OSA and 61 were controls. Lumbar spine and total hip T-scores were the primary outcomes. The OSA and control groups were compared with linear regression adjusted for age and body mass index (BMI). Four conventional indices (apnea–hypopnea index [AHI], oxygen desaturation index [ODI], mean SpO2, and minimum SpO2) were each examined in a separate model adjusted for age and BMI. Extended models, which added calcium, phosphate, alkaline phosphatase [ALP], vitamin D, and parathyroid hormone [PTH] and entered all four indices together, were treated as sensitivity analyses and were calculated. Results: Women with OSA were older than controls (63.9 ± 7.8 vs. 60.1 ± 7.2 years) and had a higher BMI (36.0 ± 7.0 vs. 30.6 ± 4.9 kg/m2). The unadjusted T-score differences were 0.28 (95% CI −0.30 to 0.86) at the lumbar spine and 0.11 (95% CI −0.34 to 0.56) at the total hip; neither was statistically significant. None of the indices were statistically significant in the extended models. The lumbar extended model was significant overall (p = 0.023), whereas the total hip model was not (p = 0.117). Conclusions: We found no statistically significant associations between the conventional indices and lumbar spine or total hip T-scores. The findings should be interpreted cautiously because of the modest sample size, possible overfitting and collinearity in the extended models, and the lack of data on cumulative hypoxic burden. They do not show that OSA has no effect on bone health. Because fractures were not assessed, these data cannot inform screening or fracture-risk decisions.