Conventional and novel hypoxemia-related indices for obstructive sleep apnea severity and incident hypertension: a prospective cohort study.

Abstract

Background and objective: Obstructive sleep apnea (OSA) is associated with hypertension (HTN), but the apnea-hypopnea index (AHI) and oxygen desaturation index (ODI) may not fully characterize nocturnal hypoxemia relevant to HTN development. We evaluated the associations of the sleep breathing impairment index (SBII) and hypoxic burden (HB) with incident HTN and compared their discriminative performance with AHI and ODI.

Methods: This prospective study included 601 participants with OSA (AHI ≥ 15 events/h) from Sleep Heart Health Study who were free of HTN at baseline. SBII, HB, AHI and ODI were analyzed both in quartiles and as continuous variables. Multivariable logistic regression and restricted cubic spline analyses were used to evaluate their associations with incident HTN. Discrimination performance was assessed by comparing a clinical base model with models that separately incorporating each quartile-categorized index.

Results: During a median follow-up of approximately 5 years, 194 (32.3%) participants developed HTN. In the primary fully adjusted model, the highest quartiles of SBII (OR, 1.71; 95% CI, 1.02-2.90; p = 0.044) and HB (OR, 1.69; 95% CI, 1.02-2.79; p = 0.041) were associated with higher odds of incident HTN. In the T90-adjusted sensitivity model, both associations were significant, whereas AHI and ODI quartiles remained nonsignificant. The clinical base model had a C-statistic of 0.614; adding SBII or HB, did not significantly improve discrimination.

Conclusion: Exploratory quartile-based analyses suggested that the highest quartiles of SBII and HB were associated with higher odds of incident HTN among participants with OSA, whereas AHI and ODI were not. These findings indicate that marked elevation in SBII and HB may provide a more comprehensive characterization of OSA-related nocturnal hypoxemia relevant to incident HTN. However, the nonsignificant continuous and overall spline analyses and the limited incremental discrimination warrant cautious interpretation and further validation.