Background and aims Sleep is an essential physiological function in newborn development. Polysomnography is the gold standard for sleep analysis. Extended recordings are difficult with this method. To evaluate actigraphy’s reliability to determine sleep-wake patterns in newborns in comparison with polysomnography.

Methods Prospective, monocentric study. 48 infants sleep patterns were recorded and assigned into two groups: group 1: 24 preterm neonates at 34–36 weeks gestational age (GA); group 2: 24 term neonates. Polysomnography (PSG) and 2 actigraphs (ACT) Actiwatch Mini^® [on arm (arm-ACT), on leg (leg-ACT)] were used during a 3-hour period. Primary endpoint: agreement rate (AR) PSG and leg-ACT with Medium activity threshold setting. Secondary endpoint: AR arm-ACT and leg-ACT. AR’s threshold was set at 85% for validation purposes. Effect of ACT activity threshold setting on a sample of 11 newborns was evaluated.

Results GA, birth weight and age at the recording: 34.5 weeks ± 0.5 and 39.2 ± 1.1, 2368 g ± 336 and 3393 ± 439, 6.4 days ± 2.8 and 2.54 ± 0.72 respectively for group 1 and 2. Group 1: AR PSG and leg-ACT was 67% ± 17 [95% CI, 60–74] and group 2: 58% ± 17 [95% CI, 51–65]. Group 1: AR arm-ACT and leg-ACT was 78% ± 12 [95% CI, 73–83] and group 2: 85% ± 10 [95% CI, 81–89]. ACT activity threshold setting did not have an impact on the results.

Conclusions ACT recording, a few days after birth, is not a reliable method for sleep pattern studies in preterm and term neonates.