Clinical prediction of periodic leg movements during sleep in children

Objective: To assess the utility of several symptoms and a questionnaire-based scale in the identification of children with periodic leg movements during sleep (PLMS). Background: PLMS may have important consequences in some children, but the extent to which a diagnosis can be established by clinical history is unknown. Methods: Subjects were patients aged 2-18 years who underwent polysomnography to assess for sleep-disordered breathing (SDB). Parents completed a Pediatric Sleep Questionnaire which contained items under consideration for inclusion in the desired scale. Results: Subjects (n = 113) had a mean age of 9.8 +/- 4.0 (SD) and 73 (65%) were male; 59 (52%) had SDB and 29 (26%) had five or more PLMS per hour of sleep (PLMI >= 5). Severity of SDB was not different among those with and without PLMI >= 5. Yes/no responses to several question-items - about restless legs, growing pains, leaving the bed at night, waking more than twice per night, waking feeling unrefreshed, and morning headaches - showed some association with PLMI >= 5 and were combined into a composite PLMS score artificially weighted toward the first two items. The PLMS score averaged 0.40 +/- 0.31 and ranged from 0.0 to 1.0; a 1 SD increase was associated with PLMI >= 5 (odds ratio 1.87, 95% confidence interval (1.15, 3.13), P = 0.014) after adjustment for age, sex, and SDB severity. Sensitivity of a PLMS score > 0.33 for PLMI >= 5 was 0.79, specificity was 0.56, positive predictive value was 0.38, and negative predictive value was 0.89. Internal consistency was reasonable (Cronbach's alpha = 0.71), as was test-retest reliability (rho = 0.62, P = 0.0026, n = 21 separate subjects). Conclusions: Restless legs, growing pains, sleep-maintenance insomnia, unrefreshing sleep, and morning headaches show moderate associations with polysomnographically-defined PLMS, but several other symptoms do not. These results require confirmation but suggest that clinical assessment and the PLMS score may be helpful but far from definitive. (C) 2001 Elsevier Science B.V. All rights reserved.