Background: Brain natriuretic peptide (BNP) serum concentration has been shown to be a preoperative predictor of postoperative outcome in high risk surgery. Whether it is able to predict early post-liver transplantation (LT) mortality in cirrhotic patients is unanswered.
Methods: Prospective monocenter observational study including all consecutive patients who received LT for cirrhosis and for whom a preoperative BNP serum dosage was available between January 2011 and December 2014.
Results: During the period, 207 cirrhotic patients among 525 LT were studied. The ICU and 180-day mortality rates were, respectively, 6% and 8%. Pre-LT BNP concentration, adjusted on model of end-stage liver disease (MELD) score, was an independent factor of ICU and 180-day mortality rates (for each 50 pg/mL increase; hazard ratio, 1035 [1.022-1.049]; P < 0.001 and 1.035 [1.014-1057]; P = 0.001). According to the receiver operator characteristic curve with an accuracy of 0.79 (0.66-0.93), the optimal cutoff value of pre-LT BNP serum level to predict ICU mortality was 155 pg/mL with a negative predictive value of 99%. All patients with MELD score exceeding 25 and pre-LT serum BNP level less than 155 pg/mL survived, whereas patients combining MELD score exceeding 25 and pre-LT BNP concentration exceeding 155 pg/mL had a 27% ICU mortality rate (P = 0.03).
Conclusions: In cirrhotic patients, pre-LT BNP serum level was an independent predictor of post-LT ICU mortality. With its excellent negative predictive value, the use of this biomarker in combination with MELD score could be useful to better predict post-LT early outcome.