Clinical utility of oximetric parameters to identify a high‐risk phenotype of moderate‐severe Obstructive Sleep Apnea (OSA)

Abstract

Toanalyzed oximetric variables in patients with moderate-severe OSA to assess their predictive value regardingas hypertension, type 2 diabetes mellitus (T2DM),coronary heartdisease (CHD) and CV mortality. Using data from SantOSA cohort, we develop receiver operating characteristic curve and area under the curve (AUC) for each parameter,defining theproposed cutoff point in a training set. Then, in a validation set with a 5years follow-up, we evaluatetheclinicaldifferencesbetween groups usingtheproposed cutoff. We also calculated adjusted Hazard Ratios (HR) of mortality using a Cox regression model. About 965 patients with moderate-severe OSA (525 in training and 440 in validation group) were included.The bestAUC was achieved with T90 (AUC=0.66) and ODI (AUC=0.61). Proposed cutoffs of T90 were hypertension: 10%, T2DM: 20%, CHD: 15%, meanwhile, proposed cutoff ofODI was≥30 ev for hypertension and T2DM. Regarding CV mortality,T90≥20% was independentlyassociated with anadjusted HR 2.44 (CI, 1.21-4.94),P-value=0.01, meanwhile, ODI≥30 ev. reported and adjusted HR 1.59 (CI, 0.75-3.39),P-value=0.22. In patients with moderate-severe OSA, oximetric parameters, especially T90≥20% remained a predictor of mortality after adjusting for a range of demographic and disease predictors.