Decision analysis of prophylactic treatment for patients with high-risk esophageal varices

Abstract

Background: Clinical decision analyses were conducted to quantify the uncertainty and to identify important factors in selection of prophylactic therapy for patients with esophageal varices. Methods: A Markov model compared variceal ligation, β-blockers, and “watchful waiting” strategies in terms of bleeding-free life years. Transition probabilities were obtained from meta-analyses of published data. A hypothetical 50-year-old white man with high-risk esophageal varices and cirrhosis served as the prototypical baseline case. Traditional n-way sensitivity analyses were applied to clarify the influence of each factor, and Monte Carlo probabilistic sensitivity analyses were used to investigate clinical uncertainty. Results: Probabilistic sensitivity analyses demonstrated that 77.0% of hypothetical cases had more bleeding-free life years after variceal ligation, whereas 23% had more when treated with β-blockers. On the basis of one-way sensitivity analyses, only 2 factors (variceal bleeding rates after ligation and treatment with β-blockers) influenced the strategy choice. Conclusions: Variceal ligation is an effective prophylactic therapy in many cases, but nearly one quarter of patients with high-risk esophageal varices and cirrhosis may benefit more from prophylactic treatment with β-blockers. Additional clinical studies identifying key variceal bleeding risk factors may lead to more effective clinical decision making for these patients. (Gastrointest Endosc 2000;52:707-14.)