A COMMUNITY STUDY OF SNORING AND SLEEP-DISORDERD BREATHING - HEALTH OUTCOMES

Four hundred forty-one subjects 34 to 69 yr of age were recruited from a random sample of the community. They answered a questionnaire and w ere monitored in their homes for steep-disordered breathing (SDB). Thi s report concerns the association between observed SDB and arterial hy pertension and vascular disease. Hypertension was defined as self-repo rt of a diagnosis of hypertension made by a physician, current treatme nt for hypertension, or a systolic pressure greater than 150 mm Hg or a diastolic pressure greater than 90 mm Hg. Coronary artery disease wa s defined by self-report of angina or myocardial infarction or ''heart attack.'' There were few cases of stroke or claudication, and a categ ory of ''occlusive vascular disease'' was defined by self-report of co ronary artery disease or of ''blocked arteries'' or stroke. Subjects w ere classified as snorers (n = 289) or nonsnorers (n = 73) by self-rep ort of regular snoring, and as having SDB (n = 79) if more than 15 abn ormal respiratory events were recorded per hour of recording. There we re significant increases in the prevalence of hypertension, coronary a rtery disease, and occlusive vascular disease from nonsnorers (26, 7, and 10%, respectively) through snorers (39, 12, and 17%) to subjects w ith SDB (57, 20, and 28%). The crude odds ratio for SDB versus nonsnor ers was 3.8 (95% CI, 1.9 to 7.5) for hypertension, 3.5 (1.2 to 10.0) f or coronary artery disease, and 3.7 (1.5 to 9.1) for occlusive vascula r disease. Adjustment for age, sex, body mass index, current alcohol c onsumption, and smoking lowered the odds ratio to 1.5 (0.7 to 3.3) for hypertension, 1.4 (0.4 to 4.5) for coronary artery disease, and 1.5 ( 0.5 to 4.2) for occlusive vascular disease. We concluded that the cons istent reduction in odds ratios with adjustment suggests substantial c onfounding. Although the residual associations are clinically importan t, more complete and accurate adjustment for confounding might reduce the odds ratios further.