Polysomnography Following Traumatic Brain Injury: A Systematic Review and Meta-Analysis

Abstract

Introduction

Sleep disturbances are common following traumatic brain injury (TBI) worsening morbidity and other neuropsychiatric symptoms. Post-TBI alterations in sleep architecture require further study.

Objectives

(1) To evaluate polysomnographic measures of sleep architecture in participants with history of TBI compared to controls and as meta-analyses of pooled means. (2) To evaluate effects of timing and severity of TBI on polysomnographic outcomes.

Methods

PRISMA compliant systematic review was conducted of MEDLINE, PsycINFO, EMBASE and Scopus. Inclusion criteria: 1) reporting polysomnography in the context of TBI and 2) operationalizing TBI using clear/formalized criteria. Data were pooled in random-effects meta-analyses with outcomes expressed as mean differences (MD).

Results

In participants with TBI, sleep was comprised of 19.39% REM sleep, 8.13% N1, 51.18% N2, and 17.53% N3, as determined by meta-analyses of single means. Total sleep time was reduced in chronic (>6 months) TBI compared to acute-intermediate TBI (<6 months) (p=0.01). Compared to controls, participants with TBI differed with increased N1 sleep (MD=0.64%; 95%CI=0.02,1.25; p=0.04), reduced sleep efficiency (MD=-1.65%; 95%CI=-3.18,-0.12; p=0.03), and reduced sleep latency on the multiple sleep latency test (MD=-5.90mins; 95%CI=-10.09,-1.72; p<0.01). On sub-group analyses, participants with mild TBI differed from controls with reduced total sleep time (MD=-29.22mins, 95%CI=-54.16,-4.27; p=0.02). Similarly, participants with acute-intermediate TBI exhibited increased sleep latency compared to controls (MD=8.96mins; 95%CI=4.07,13.85; p<0.01) and differed significantly from participants with chronic TBI (X²(1,N=608)=6.54; p=0.01).

Conclusions

Sleep architecture is altered following TBI with potential implications regarding functional outcomes and recovery. These alterations appear to differ based on severity of injury and time since injury.

Disclosure

No significant relationships.