Energy expenditure in obstructive sleep apnea: validation of a multiple physiological sensor for determination of sleep and wake.

PURPOSE

Obstructive sleep apnea (OSA) may be associated with increased energy expenditure (EE) during sleep. As actigraphy is inaccurate at estimating EE from body movement counts alone, we aimed to compare a multiple physiological sensor with polysomnography for determination of sleep and wake, and to test the hypothesis that OSA is associated with increased EE during sleep.

METHODS

We studied 50 adults referred for routine overnight polysomnography. In addition to polysomnography, the SenseWear Pro3 Armband(TM) (Bodymedia Inc.) was placed on the upper right arm. Epoch-by-epoch agreement rate between the measures of sleep versus wake was calculated. Linear regression analyses were performed for EE against apnea-hypopnea index (AHI), 3% oxygen desaturation index (ODI), body mass index (BMI), waist-hip ratio (WHR), gender, age, and average heart rate during sleep.

RESULTS

The epoch-by-epoch agreement rate was high (79.9 ± 1.6%) and the ability of the SenseWear to estimate sleep was very good (sensitivity, 88.7 ± 1.5%). However, it was less accurate in determining wake (specificity 49.9 ± 3.6%). Sleep EE was associated with AHI, 3% ODI, BMI, WHR, and male gender (p < 0.001 for all). Stepwise multiple linear regression however revealed that BMI, male gender, age, and average heart rate during sleep were independent predictors of EE (Model R (2) = 0.78).

CONCLUSIONS

The SenseWear armband provides a reasonable estimation of sleep but a poor estimation of wake. Furthermore, in a selected population of OSA patients, increasing OSA severity is associated with increased EE during sleep, although primarily through an association with increased BMI. However, as our data are not adjusted for fat-free mass and the SenseWear has yet to be validated for EE in OSA patients, these data should be interpreted with caution.