Trajectories of Glasgow Outcome Scale-Extended after traumatic brain injury: an analysis of the TRACK-TBI cohort.

Abstract

Recovery following traumatic brain injury (TBI) is highly heterogeneous. Characterising longitudinal functional trajectories may enable more individualised care and inform the trial design. This study aimed to identify distinct patterns of functional recovery in the first 12 months post-injury using the Glasgow Outcome Scale - Extended (GOSE). Patients with available GOSE data from Transforming Research and Clinical Knowledge in Traumatic Brain Injury, a prospective, multicentre study, were included. A two-stage multiple imputation procedure addressed missingness, and trajectories were modelled using polytomous variable latent class analysis (poLCA) in both complete-case and imputed datasets. Among the 2,100 participants with TBI, poLCA identified seven distinct recovery trajectories based on GOSE scores. These included (1) death (6.2%), (2) persistent GOSE 2-5 (5.7%), (3) improvement from GOSE 3-5 to GOSE 5 by 3 months with no further gains (9.0%), (4) rapid improvement from GOSE 3-5 to GOSE 7-8 sustained through follow-up (7.0%), (5) gradual improvement from GOSE 3-6 to GOSE 5-7 over 6 months (15.3%), (6) rapid early improvement from GOSE 4-7 to GOSE 6-8 by 3 months and then plateauing (21.8%) and (7) progression from GOSE 5-8 to GOSE 7-8 within 3 months, with stable outcomes thereafter (25.1%). Recovery class membership was significantly associated with the initial Glasgow Coma Scale scores, CT severity (Marshall and Rotterdam scores) and presence of psychiatric comorbidities. Our findings support a shift towards trajectory-based stratification in clinical care and research. Incorporating these patterns into prognostic modelling may improve outcome prediction, personalise rehabilitation and refine eligibility criteria for interventional trials.