Cerebral venous thrombosis in traumatic brain injury: a cause of secondary insults and added mortality

Abstract

OBJECTIVE Cerebral venous thrombosis (CVT) is increasingly recognized in traumatic brain injury (TBI), but its complications and effect on outcome remain undetermined. In this study, the authors characterize the complications and outcome effect of CVT in TBI patients. METHODS In a retrospective, case-control study of patients included in the Oslo University Hospital trauma registry and radiology registry from 2008 to 2014, the authors identified TBI patients with CVT (cases) and without CVT (controls). The groups were matched regarding Abbreviated Injury Scale 1990, update 1998 (AIS’98) head region severity score 3–6. Cases were identified by AIS’98 or ICD-10 code for CVT and CT or MR venography findings confirmed to be positive for CVT, whereas controls had no AIS’98 or ICD-10 code for CVT and CT venography or MR venography findings confirmed to be negative for CVT. All images were reviewed by a neuroradiologist. Rates of complications due to CVT were recorded, and mortality was assessed both unadjusted and in a multivariable logistic regression analysis adjusting for initial Glasgow Coma Scale score, Rotterdam CT score, and Injury Severity Score. Complications and mortality were also assessed in prespecified subgroup analysis according to CVT location and degree of occlusion from CVT. Lastly, mortality was assessed in an exploratory subgroup analysis according to the presence of complications from CVT. RESULTS The CVT group (73 patients) and control group (120 patients) were well matched regarding baseline characteristics. In the CVT group, 18% developed venous infarction, 11% developed intracerebral hemorrhage, and 19% developed edema, all representing complications secondary to CVT. Unadjusted 30-day mortality was 16% in the CVT group and 4% in the no-CVT group (p = 0.004); however, the difference was no longer significant in the adjusted analysis (OR 2.24, 95% CI 0.63–8.03; p = 0.215). Subgroup analysis by CVT location showed an association between CVT location and rate of complications and an unadjusted 30-day mortality of 50% for midline or bilateral CVT and 8% for unilateral CVT compared with 4% for no CVT (p < 0.001). The adjusted analysis showed a significantly higher mortality in the midline/bilateral CVT group than in the no-CVT group (OR 8.41, 95% CI 1.56–45.25; p = 0.032). CONCLUSIONS There is a significant rate of complications from CVT in TBI patients, leading to secondary brain insults. The rate of complications is dependent on the anatomical location of the CVT, and midline and bilateral CVT is associated with an increased 30-day mortality in TBI patients.