A systematic review and meta-analysis of sleep following mild traumatic brain injury: A synthesis of the literature according to age and time-since-injury

Abstract

This systematic review and meta-analysis (PROSPERO: CRD42021287081) compared sleep in individuals with prior mTBI against that of controls. PubMed, Scopus, and EMBASE were searched through to November 2, 2024. Mean differences were calculated for objective (sleep duration, sleep efficiency, sleep stage [N1-3, REM] proportions) and subjective (Pittsburgh Sleep Quality Index [PSQI]) sleep measures, with analyses conducted based on age (adult ≥18y; child <18y) and time-since-injury/recovery phase (<1w, <1m, 1–12m, ≥1y). Qualitative syntheses were also conducted according to age and recovery phase. Sixty-nine studies met criteria. In children and adults, mean differences in objective sleep duration and sleep efficiency were not statistically significant for any recovery phase. Relative to controls, adults with mTBI had proportionately more N1 (Mean diff. [95%CI], 1.28 % [0.63 %,1.93 %], p = 0.014) and less N3 (−1.16 % [-2.08 %,-0.23 %], p = 0.033) sleep at 1–12-month follow-up, and less REM (−3.37 % [-6.28 %,-0.46 %], p = 0.023) at ≥1 year follow-up; however, these statistical differences did not remain when studies confounded by selection bias were removed. Subjectively, relative to controls, adults with mTBI had higher PSQI scores at <1-week (2.53 [1.43,3.62] points, p = 0.005), <1-month (2.26 [1.47, 3.04], p < 0.001), 1-12-month (2.86 [1.26,4.45], p < 0.001) and ≥1-year (2.62 [1.96, 3.28], p < 0.001) follow-up. The literature suggests individuals with prior mTBI have poorer subjective sleep than controls, but supportive objective data are limited. A limitation to consider when interpreting these findings is that a minority (45 %) of studies had low risk of bias. More research examining objective sleep post-mTBI is needed, especially for children with mTBI.