Sleep apnea and sleep efficiency as predictors of survival after polysomnography

Abstract

Obstructive sleep apnea (OSA) has been associated with increased mortality. It is not known, however, how sleep apnea affects the relation between short and long sleep duration and mortality risk. Despite the existence of many studies, the association between sleep duration and mortality is still unclear due to the large number of confounders, such as OSA. In the context, the relationship between mortality and sleep efficiency in sleep apnea patients remained less studied. The present study aims to investigate the association between survival time, sleep apnea severity, and sleep efficiency in patients with sleep disorders who underwent full night polysomnography. A list with 17,778 names and demographics of people who underwent polysomnography between January 1985 and June 2012 was sent to the Health Information Group of the State Health Department. The search in the database of death certificates from 2000 to 2012 obtained 227 patients. Subjects had signed a form consenting in the anonymous use of their data. The subjects were divided at the median survival time in SHORT (<5 years) and LONG survival (>5 years). Among the deceased, 173 were men and 54 women, with ages of (mean ± SD) 58 ± 14 and 59 ± 12 years, respectively. Two cases had sleep apnea as cause of death. The mean survival time was similar for men and women (6.2 ± 4.3 and 6.3 ± 4.4 years). Comparing the SHORT and LONG survival, all variables examined were similar, including the apnea-hypopnea index (37 ± 28 and 39 ± 31 events/hour; P = 0.6), gender, age, body mass index, systolic and diastolic blood pressure, waist and neck circumference, average and minimum oxygen saturation, as well as time and cause of death. The number of years of survival only correlates significantly with sleep efficiency (rho = 0.17; P = 0.01). The sleep efficiency, however, was lower in the LONG survival group (76 ± 16% and 80 ± 16%; P = 0.049) and the total sleep time was 27 min shorter ( P = 0.01). In the linear regression to predict survival time, sleep efficiency remains the only variable with a significant beta value (−0.16; P = 0.03). The odds ratio for SHORT survival was 2.8 times higher in the group with sleep efficiency >84% (95% CI 1.5-5.0). In the present study, the only variable that explains years of survival is sleep efficiency. This may suggest that hyperexcitability evidenced in the polysomnography by reduced sleep efficiency is a feature that enhances survival. This study was supported in part by CNPq, CAPES, and FIPE-HCPA (Brazil).