Share this article on:

What Is the Optimal Timing of Liver Transplantation for Children With Biliary Atresia? A Markov Model Simulation Analysis

Arnon, Ronen*; Leshno, Moshe; Annunziato, Rachel; Florman, Sander*; Iyer, Kishore*

Journal of Pediatric Gastroenterology & Nutrition: September 2014 - Volume 59 - Issue 3 - p 398–402
doi: 10.1097/MPG.0000000000000428
Original Articles: Hepatology and Nutrition

Objectives: Biliary atresia (BA) is the most common liver disease leading to liver transplantation (LT) during childhood. The optimal timing of listing and LT for children with failed portoenterostomy (PE) is not clear. The aim of our study was to determine the optimal timing of LT for children with BA and failed PE by using a Markov model simulation analysis.

Methods: A Markov model was constructed presenting the progression of the severity of liver status for patients with BA who underwent PE before 60 days old. Three treatment strategies were compared: LT for moderate liver disease (MLD), LT for severe liver disease (SLD), and no LT, and 10,000 patients were simulated in each strategy. Health states were defined as LT, early repeat LT (≤30 days after LT), late repeat LT (>30 days after LT), status post-LT (period after first LT), and death.

Results: For patients with an available liver for transplantation (living donors), LT at MLD was associated with an increase of 17.4% additional expected life-years (LY) as compared with LT at SLD. Patient survival rates after 10 years were 84.7% and 75.5% in the MLD and SLD strategies, respectively. For the patients with no LT, the survival rate after 10 years was 48.1%. When the probability of deceased donor LT was lower than 50% from time of listing at 3 months, there was no increase in expected LY of MLD strategy. “No LT” resulted in approximately 60% reduction of expected LY compared with LT for patients with SLD.

Conclusions: Our model suggests that early listing and transplantation is beneficial in patients with an available liver for transplantation. For patients in whom the probability for LT is low, there appears to be no advantage to early listing. A validation of the present model in a “real” cohort of patients with BA is needed.

*RMTI, Mount Sinai Medical Center, New York, NY

Faculty of Management and Tel-Aviv Souraski Medical Center, Tel-Aviv University, Tel-Aviv, Israel

Psychology Department, Fordham University, Bronx, NY.

Address correspondence and reprint requests to Ronen Arnon, MD, Mount Sinai Medical Center, New York, NY (e-mail:

Received 10 December, 2013

Accepted 30 April, 2014

The authors report no conflicts of interest.

© 2014 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,