Background: During hospitalization, many patients experience acute sleep deprivation and hyperglycemia. Given laboratory and epidemiologic studies demonstrate associations between sleep loss, hyperglycemia, and diabetes, we aimed to assess these relationships among hospitalized adults.
Methods: We conducted a prospective cohort study of patients 50 years or older on the General Medicine and Hem/Onc wards. Patients in isolation, with preexisting sleep disorders, with an ICU stay, were non-ambulatory, or residing in a nursing home were excluded. Participants reported sleep quality the past month prior to hospitalization via the Pittsburgh Sleep Quality Index (PSQI), excessive daytime sleepiness via the Epworth Sleepiness Scale, and risk of obstructive sleep apnea (OSA) via the Berlin Questionnaire. Wrist actigraphy measured hospital sleep duration and efficiency. Diabetic patients were defined by patient report, diabetes diagnosis in the record, or billing codes for diabetes. Morning fasting glucose was obtained from lab records, with our lab defining elevation as >109mg/dL. Multivariable logistic regression was used to test the association between sleep (duration and efficiency) and odds of hyperglycemia.
Results: From April 2010 to June 2014, 388 patients (58% female, 73% African-American) were enrolled. Roughly one-third (126, 33%) had diabetes. Compared to non-diabetics, diabetic patients reported worse sleep quality the prior month [PSQI>5: 75% diabetic vs 55% non-diabetic, p<.001], greater excessive daytime sleepiness [Epworth >9: 42% diabetic vs 29% nondiabetic, p=.003], and higher risk for OSA [42% diabetic vs 25% nondiabetic, p=.001]. Objective inpatient sleep measures did not differ between diabetic and non-diabetic patients. Glucose levels were available on 378 mornings for 198 patients [mean glucose 110.3 mg/dL (95% CI 106, 114)]. Using multivariate logistic regression, clustered by subject, and controlling for diabetes diagnosis and gender, for every hour of inpatient sleep lost, the odds of elevated subsequent morning blood glucose increased by 17% [OR 1.17 (95% CI 1.04, 1.31), p=.007]. Likewise, every 10% decrease in sleep efficiency was associated with a 26% increase in odds of elevated morning blood glucose [OR 1.28 (95% CI 1.15, 1.42), p<.001]. These associations remained significant when controlling for risk of OSA, obesity, and age.
Conclusions: Hospitalized diabetic patients reported worse sleep quality, greater excessive daytime sleepiness, and higher risk of OSA. While diabetic patients did not have worse objective measures of inpatient sleep, poor sleep among all hospitalized patients was associated with greater odds of hyperglycemia. These findings have important implications for hospitalists. Specifically, screening diabetic inpatients for sleep disorders may be high yield and improving sleep loss for all hospitalized patients could be a novel way to reduce the likelihood of hyperglycemia of hospitalization.