Clinical assessment of sleep among pediatric burn patients does not correlate with polysomnography

A vast amount of sleep research relies on subjective, observational assessments of wakefulness and sleep. The authors had the unique opportunity to compare observational and polysomnographic (PSG) determinations of sleep in a randomized crossover study of sleep-inducing medication in a group of pediatric burn patients. Pediatric burn patients were randomized to one of two regimens with sleep-inducing agents over a 2-week period. PSG was conducted for three consecutive nights each week, between 7-13 and 14-20 days postburn. The first night of monitoring each week was conducted without medication to serve as a baseline. Observational sleep assessments (awake, drowsy, or asleep) were simultaneously recorded every 15 minutes. PSG concordance with observation was based on the PSG sleep stages identified during the 2 minutes before the observations. If all 30-second epochs in the two minutes were designated as sleep stage 1 or above, then the PSG record was categorized as asleep. If all epochs demonstrated wakefulness, an awake status was recorded. Otherwise, the corresponding PSG finding was classified as mixed. Forty patients were enrolled into the study, with a mean age of 9.4 ± 0.6 years, TBSA burn of 50.1 ± 2.9%, and third-degree burn surface area of 43.2 ± 3.6%. Patients were judged according to observational criteria to be awake 9% of the nocturnal study period compared with PSG recordings indicating that the patients were awake 52.3% of the time. The correlation between observation and PSG was poor regardless of sleep agent administration. In conclusion, observational determination of wakefulness in pediatric burn patients correlates poorly with PSG; therefore, PSG is vital in the accurate evaluation of sleep-inducing medications among burn patients.