Predictors of Infarction in Tuberculous Meningitis in Indian Patients

Abstract

BackgroundStroke is a devastating complication of tuberculous meningitis (TBM) and is an important determinant of its outcome. We propose a model which would help to predict development of infarction or cerebrovascular events in patients of TBM. MethodsA prospective study with n=129 patients of TBM were evaluated for predictors and outcomes of stroke. A diagnostic grid was formulated with clinical, laboratory and radiology as parameters to predict the vascular outcomes. All patients were followed up for mortality and disability on the basis of modified rankin score (mRS). MRI & CSF cytokines TNF–alpha, IFN- gamma & IL-6,8, 10 were measured at baseline and 3 months. The diagnosis of TBM included definite, probable & possible types and stage I & II with early and late onset of symptoms respectively. ResultsThe mortality was 16.2% and 19.4% of all patients developed stroke. The mean GCS, barthel index and mRS at admission was 57.03± 9.5,10.2±2.3 & 3.3±0.03 respectively mild to moderate infection and functional limitation. Barthel index (BI) happened to be a strong predictor [F=32.6, p=0.001, t=15.5, βeta coefficient =0.002] followed by biomarker TNF-α [F=18.9, p=0.02, t= -2.07, βeta coefficient=-0.04]. N=25 patients developed stroke with TNF-α, IL-6, IFN -γ showing statistically significant increase in all the stroke affected TBM (95% CI; 4.5 to 1.2; p=0.003). At 3 months, it was observed that mRS was statistically significant between stage I & II (95% CI; 5.4 to 2.1; p=0.04). ConclusionsOur data revealed that 19.4% patients developed vascular events during the hospital stay or follow up. We recruited late onset TBM as compared to early onset. BI, TNF-α, IL6 are most potent predictors of stroke post TBM.