Mediation effect of hypoxia in the relationship between obstructive sleep apnea and cardiovascular death risk

Purpose This study aimed to explore the role of hypoxia in the relationship between obstructive sleep apnea (OSA) and cardiovascular disease (CVD) death risk based on data from the Sleep Heart Health Study (SHHS). Methods Multivariate logistic regression analysis was used to analyze the association between OSA, hypoxia, and CVD death risk. Causal mediation analysis was performed to assess the role of hypoxia. The severity of OSA was evaluated by the apnea-hypopnea index (AHI), and the hypoxia was quantified by the percentage of sleep time with less than 90% oxygen saturation (PCTST90). Results Of these 5,145 participants, 989 had no OSA, 2,110 had mild OSA, and 2,046 had moderate-to-severe OSA. After adjusting all confounders, mild OSA [odds ratio (OR): 1.800; 95% confidence interval (CI), 1.192-2.802], moderate-to-severe OSA (OR: 1.745; 95%CI, 1.148-2.758), 0 < PCTST90 < 1 (OR: 1.668; 95%CI, 1.184-2.385), and PCTST90 >= 1 (OR: 1.649; 95%CI, 1.148-2.400) were associated with an increased death risk of CVD. Furthermore, participants with mild OSA (OR: 3.742; 95%CI, 3.183-4.398) and moderate-to-severe OSA (OR: 19.671; 95%CI, 16.303-23.734) had a higher risk of hypoxia than those without OSA. Causal mediation analysis indicated that the average direct effect (ADM) of moderate-to-severe OSA and average causal mediation effect (ACME) of hypoxia on CVD death risk were 0.024 (95%CI, 0.004-0.040) and 0.013 (95%CI, 0.005-0.020), respectively, and the average mediating effect ratio was 33.94%. Conclusion Hypoxia played a mediating role in the increased death risk of CVD caused by OSA, and the mediating effect of hypoxia did not account for a large proportion.