ABSTRACT: The prognostic role of comorbidity in patients with severe obstructive sleep apnoea syndrome (OSAS) and their role for risk stratification remain incompletely studied and poorly defined. We studied 1,592 patients with severe OSAS diagnosed by polysomnography. The primary outcome was all-cause mortality. We used Poisson regression to estimate the standardized mortality ratio (SMR), compared with the general population. The association of comorbidities with all-cause mortality was assessed using multivariable Cox regression analysis. Then, we calculated pseudo-R² to measure the proportion of variation explained by the model. Finally, we applied recursive-partitioning analysis to cluster the patients into risk groups according to individual comorbidities. A total of 11,721 person-years of follow-up were examined during which 390 deaths (3.33 deaths/100 person-years) occurred. The median follow-up was 7 (4-10) years. The SMR was 1.47 (1.33-1.63). Age, sex, obesity, cardiovascular diseases (CVD), moderate-to-severe chronic obstructive pulmonary disease (COPD), chronic kidney disease (CKD) and malignancy were independently associated with mortality risk. The model had a C index of 0.811 and explained 63.2% of the variation in mortality risk. Recursive-partitioning analysis allowed distinguishing three clinical phenotypes differentially associated with mortality risk. The combination of CKD with CVDs or with moderate-to-severe COPD conferred the highest risk. The SMR was 2.56 (2.13-3.08) in the high-, 1.59 (1.37-1.85) in the intermediate-, and 0.91 (0.75-1.11) in the low-risk category. Severe OSAS is associated with increased risk for all-cause death. Age and comorbidity are crucial contributors to increased mortality in severe OSAS. Clustering patients according to comorbidities allows identifying clinically meaningful phenotypes.