Objective: Transjugular intrahepatic portosystemic stent-shunt (TIPSS) is increasingly used to treat complications of portal hypertension, but proven tools for risk assessment of early mortality are lacking.
Design: The prospective evaluation of a new 60-day mortality score.
Patients and methods: In a tertiary medical centre, 30 consecutive TIPSS patients were analysed for early mortality predictors, such as Child-Pugh score, TIPSS urgency (elective: > or = 36 h or emergency: < 36 h after variceal bleeding), comorbidity (Acute Physiology and Chronic Health Evaluation [APACHE]-II) and clinical data. Main predictors (P< 0.01) in this group (group-1: Child-Pugh score 10A, 10B, 10C) were graded (1, 2 or 3 points representing low, medium and high risk, respectively) and summarized as a Bonn TIPSS early mortality (BOTEM) score. This score was then tested prospectively in the next 73 TIPSS patients (group-2: Child-Pugh score 14A, 42B, 17C).
Results: Group 1 early mortality (30%) depended primarily on bilirubin (P< 0.005), APACHE-II (P < 0.001) and TIPSS urgency (P< 0.001). Added risk points (1, 2, 3) for bilirubin (< 3 mg/dl, 3-6 mg/dl, > 6 mg/dl, respectively), APACHE-II (< 10, 10-20, > 20 points, respectively) and urgency (elective, emergency, active bleeding, respectively) represented individual BOTEM score points. BOTEM was the best mortality predictor (P< 0.001); < or = / > 6 score points was the optimal cut-off, with 56% sensitivity, 100% specificity, 100% positive predictive value, 84% negative predictive value and 87% accuracy. In group 2, early mortality (8.2%) was again best predicted by BOTEM (P < 0.01) with the same cut-off and 67% sensitivity, 99% specificity, 80% positive predictive value, 97% negative predictive value and 96% accuracy.
Conclusion: BOTEM score based on bilirubin, comorbidity and TIPSS-urgency predicts rather reliably post-TIPSS 60-day mortality and might optimize TIPSS treatment.