In 601 Sleep Heart Health Study participants with OSA who were free of hypertension at baseline, the highest quartiles of SBII and hypoxic burden were associated with higher odds of incident HTN whereas AHI and ODI were not, but the continuous and overall spline analyses were nonsignificant and adding SBII or HB did not significantly improve discrimination, so cautious interpretation and further validation are warranted.
About one third of participants with OSA who were free of hypertension at baseline developed incident hypertension over roughly 5 years.
In the primary fully adjusted model, the highest quartile of SBII was associated with higher odds of incident hypertension.
In the primary fully adjusted model, the highest quartile of hypoxic burden (HB) was associated with higher odds of incident hypertension.
In the T90-adjusted sensitivity model, SBII and HB associations remained significant, whereas AHI and ODI quartiles remained nonsignificant.
Adding SBII or HB to a clinical base model did not significantly improve discrimination for incident hypertension.
Continuous and overall restricted cubic spline analyses of the indices were nonsignificant, so the findings are limited to the highest quartiles.
The authors interpret marked elevation of SBII and HB as providing a more comprehensive characterization of OSA-related nocturnal hypoxemia relevant to incident HTN than AHI or ODI.
This study followed 601 people with moderate-to-severe obstructive sleep apnea (OSA) who did not have high blood pressure at the start, using data from the Sleep Heart Health Study. Over about 5 years, roughly one in three (32.3%) developed hypertension. The researchers compared standard measures of OSA severity, the apnea-hypopnea index (AHI) and oxygen desaturation index (ODI), with two newer measures that capture how much and how deeply blood oxygen drops during sleep: the sleep breathing impairment index (SBII) and hypoxic burden (HB). People in the highest quarter for SBII or HB had about 70% higher odds of developing high blood pressure than those in the lowest quarter, after accounting for other factors. The standard measures, AHI and ODI, did not show a significant link. However, the association was seen only when comparing the top group with the bottom group, not when the measures were analyzed as continuous scores. Adding SBII or HB to a model with ordinary clinical risk factors also did not significantly improve its ability to predict who would develop hypertension. This research suggests that the amount of oxygen lowering during sleep may matter for blood pressure risk beyond what the standard sleep apnea counts show. The results are exploratory and modest, so they need confirmation in other studies before the newer measures could be considered useful for predicting hypertension.
Wang L, Hui X, Huang R, Xiao Y. (2026). Conventional and novel hypoxemia-related indices for obstructive sleep apnea severity and incident hypertension: a prospective cohort study.. Frontiers in neurology. https://doi.org/10.3389/fneur.2026.1973505