Distinct prognostic value of aspartate-to-alanine aminotransferase ratio (AAR) in traumatic brain injury versus hemorrhagic stroke: a cohort study of 1,069 patients

Background The liver-brain axis influences outcomes after acute brain injury, yet the prognostic significance of liver markers across varying injury types remains unclear. We examined the disease-specific prognostic value of the aspartate aminotransferase-to-alanine aminotransferase ratio (AAR) in traumatic brain injury (TBI), intracerebral hemorrhage (ICH), and aneurysmal subarachnoid hemorrhage (aSAH). Methods This retrospective cohort study included 1,069 consecutive patients (413 TBI, 490 ICH, 166 aSAH). The primary outcome was unfavorable neurological status at discharge (modified Rankin Scale 3–6). Multivariable logistic regression and formal interaction testing assessed the heterogeneity of AAR’s effect across disease types. Results Unfavorable outcomes occurred in 420 patients (39.3%). Higher AAR was associated with poor outcomes in the overall cohort (adjusted OR 1.34, 95% CI: 1.07–1.66). However, significant effect modification by disease type was observed (P for interaction = 0.001). In TBI, elevated AAR was a robust independent predictor (adjusted OR 2.15, 95% CI: 1.46–3.15, p < 0.001), with the magnitude of association escalating with injury severity (Severe TBI: adjusted OR 5.38). Conversely, AAR showed no significant prognostic value in ICH ( p = 0.746) or aSAH ( p = 0.810). A significant dose–response relationship between AAR and unfavorable outcome was observed across the overall cohort (P for trend < 0.001), and the prognostic effect of AAR in TBI escalated markedly with injury severity. Conclusion The prognostic value of AAR is highly heterogeneous. It serves as a powerful predictor in TBI—likely reflecting systemic metabolic stress—but lacks utility in hemorrhagic stroke. These findings suggest distinct pathophysiological mechanisms driving liver-brain interactions across neurosurgical conditions, cautioning against a “one-size-fits-all” biomarker approach.