Objective:Several clinical obstructive sleep apnea syndrome(OSAS)phenotypes associated with heterogeneous cardiovascular risk profiles have been recently identified.The purpose of this study was to identify clusters amongst these profiles that allow for the differentiation of patients.Methods:This retrospective study included all moderate-to-severe OSAS patients referred to the sleep unit over a 5-year period.Demographic,symptom,comorbidity,polysomnographic,and continuous positive airway pressure(CPAP)adherence data were collected.Statistical analyses were performed to identify clusters of patients.Results:A total of 567 patients were included(67％ men,54±13 years,body mass index:32±7 kg/m2,65％ Caucasian,32％ European African).Five clusters were identified:less severe OSAS(n=172);healthier severe OSAS(n=160);poorly sleeping OSAS patients with cardiometabolic comorbidities(n=87);younger obese men with sleepiness at the wheel(n=94);sleepy obese men with very severe desaturating OSAS and cardiometabolic comorbidities(n=54).Patients in clusters 3 and 5 were older than those in clusters 2 and 4(P=0.034).Patients in clusters 4 and 5 were significantly more obese than those in the other clusters(P=0.04).No significant differences were detected in terms of symptoms and comorbidities.Polysomnographic profiles were very discriminating between clusters.CPAP adherence was similar in all clusters but,among adherent patients,daily usage was more important in cluster 1(less severe patients)than in cluster 5.Conclusion:This study highlights that the typical sleepy obese middle-aged men with desaturating events represent only a minority of patients in our multi-ethnic moderate-to-severe OSAS cohort of 33％ females.