Prevalence, predictors, and treatment of chronic pain after complicated mild to severe traumatic brain injury: a systematic review and meta-analysis.

Chronic pain reduces health-related quality of life, but previous research on pain after traumatic brain injury (TBI) has been largely concussion-focused. This systematic review evaluated the prevalence, predictors, and treatment of chronic pain (bodily, headache, any chronic pain), as well as average pain intensity/interference (secondary objective) in adults ≥3 months after complicated mild to severe TBI. Studies were identified through electronic databases and screened against predefined selection criteria. Chronic pain prevalence and mean pain intensity/interference (observational studies), predictor associations (cohort studies), and treatment effects (clinical trials) were synthesized; data were pooled where able using generalized linear mixed-, random-, or fixed-effects models. We assessed heterogeneity using between-cluster variance and sensitivity analyses, risk of bias, publication bias, and evidence certainty (Grading of Recommendations Assessment, Development and Evaluation). Overall, 14,331 studies were screened, with 28 included across the 4 review topics (17,899 participants; 94.61% mod-sev TBI; 9.72% military). Pooled prevalences (95% CI) of chronic pain were as follows: "bodily" 51% (39-63; k = 5 studies; moderate certainty), "headache" 31% (20-45; k = 4; moderate certainty), "any" 40% (31-49; k = 9; low certainty). Pooled 36-Item Short Form Health Survey Bodily Pain score was 70.7 (66.1-75.4; k = 7). Chronic pain predictors were sex, ethnicity, military service, preinjury headache/migraine, penetrating injury, and poor psychosocial outcomes (sleep, depression, posttraumatic stress, anxiety). One trial (Botulinum toxin-A vs placebo) showed moderate short-term reductions in posttraumatic headache frequency (very low certainty). Chronic pain was double the general population rate post-TBI, with bodily pain more prevalent than headache. Rates were similar over time and between civilian and military populations; however, there was greater variance and lower certainty in military estimates. Future research involving high-quality treatment studies is needed.