Comparison of three actigraphic algorithms used to evaluate sleep in patients with obstructive sleep apnea

Purpose Actigraphy is a non-invasive and valid method to detect sleep/wake status. However, the technique lacks reliability in patients with sleep-disordered breathing and its results may depend on the algorithm employed. Methods We compared three currently used algorithms (the Cole-Kripke, Sadeh, and University of California San Diego [UCSD]) and determined which is the most reliable in patients with obstructive sleep apnea (OSA) assessing total sleep time. After identification of the most reliable algorithm, we compared total sleep time with the severity of obstructive sleep apnea. Results The mean total sleep time was not significantly different from that yielded by polysomnography when the UCSD algorithm was employed ( p = 0.798) and UCSD algorithm was associated with the smallest bias. The correlation levels (with polysomnographic data) were mild-to-modest when the results yielded by all algorithms were evaluated, but were highest when the UCSD algorithm was employed (UCSD, r = 0.498, p