Diabetes and sleep disturbances: findings from the Sleep Heart Health Study

To test the hypothesis that diabetes is independently associated with sleep-disordered breathing (SDB), and in particular that diabetes is associated with sleep abnormalities of a central, rather than obstructive, nature.

Using baseline data from the Sleep Heart Health Study (SHHS), we related diabetes to 1). the respiratory disturbance index (RDI; number of apneas plus hypopneas per h of sleep); 2). obstructive apnea index (OAI; >or=3 apneas/h of sleep associated with obstruction of the upper airway); 3). percent of sleep time < 90% O(2) saturation; 4). central apnea index (CAI; >or=3 apneas [without respiratory effort]/h sleep); 5). occurrence of a periodic breathing (Cheyne Stokes) pattern; and 6) sleep stages. Initial analyses excluding persons with prevalent cardiovascular disease (CVD) were repeated including these participants.

Of the 5874 participants included in this report, 692 (11.8%) reported diabetes or were taking oral hypoglycemic medications or insulin and 1002 had prevalent CVD. Among the 4872 persons without CVD, 470 (9.6%) had diabetes. Diabetic participants had worse CVD risk factor profiles than their nondiabetic counterparts, including higher BMI, waist and neck circumferences, triglycerides, higher prevalence of hypertension, and lower HDL cholesterol (P < 0.001, all). Descriptive analyses indicated differences between diabetic and nondiabetic participants in RDI, sleep stages, sleep time <90% O(2) saturation, CAI, and periodic breathing (P < 0.05, all). However, multivariable regression analyses that adjusted for age, sex, BMI, race, and neck circumference eliminated these differences for all sleep measures except percent time in rapid eye movement (REM) sleep (19.0% among diabetic vs. 20.1% among nondiabetic subjects, P < 0.001) and prevalence of periodic breathing (odds ratio [OR] for diabetic subjects versus nondiabetic subjects 1.80, 95% CI 1.02-3.15). Additionally, adjusted analyses showed diabetes was associated with nonstatistically significant elevations in the odds of an increased central breathing index (OR 1.42, 95% CI 0.80-2.55). Addition to the analysis of the 1002 persons with prevalent CVD (including 222 people with diabetes) did not materially change these results.

These data suggest that diabetes is associated with periodic breathing, a respiratory abnormality associated with abnormalities in the central control of ventilation. Some sleep disturbances may result from diabetes through the deleterious effects of diabetes on central control of respiration. The high prevalence of SDB in diabetes, although largely explained by obesity and other confounders, suggests the presence of a potentially treatable risk factor for CVD in the diabetic population.