Introduction Recent data suggest that serum ferritin (SF) concentration predicts mortality in patients awaiting liver transplantation (LT).

Aim To test SF concentration as a predictor of waiting list (WL) mortality.

Method Retrospective analysis of all patients assessed at our centre for LT over a 4-year period.

Results Patients with acute liver failure (138), amyloid (15), multiple organ transplants (14) and re transplantation (18) were excluded. Of the remaining 422 listed patients for LT, 45 died on the wait list (11%), 26 (6%) were de-listed and 350 (83%) transplanted. Men comprised 64%. SF was analyzed as a continuous and categorical variable. Patients were classified into 3-groups according to assessment SF (<200, 200–400 and >400 ug/l), (Walker et al Hepatology 2010;15:1683). There was a significant difference in gender distribution (p=0001), listing model for end-stage liver disease (MELD) (p=0.0001), etiology of cirrhosis (p=0001) between groups, with alcoholic cirrhosis and HCV representing 45% and 35% respectively of SF >400ug/l group. Univariate analysis demonstrated age at listing (HR 0.97, p=0.013), listing MELD (HR 0.86, p=0.0001), listing Na (HR 1.10, p=0.0001), SF (HR 1.0,p=0.002), SF 200-400 (HR2.86, p=0.001) as predictors of WL mortality; whilst SF>400 was insignificant (HR 2.14,p=0.078). On multivariate analysis, only MELD score (p=00001), age at listing (p=0.002) and serum Na (0.011) were significant predictors of listing outcome. Serum Ferritin analyzed as continuous or categorical variable failed to predict WL mortality. ROC curve analysis showed AUROC for MELD of 0.73 whilst that for SF of 0.65. This analysis determined a cut-off for SF of 266ug/l as significant predictor of WL mortality. Interestingly, SF 200–400 (HR 3.9, p=0.002) and SF >266 ug/l (as determined by ROC curve analysis, HR 2.9, p=0.006) were significant indicators of 1-year post-LT survival. However, age, gender, presence of HCC, MELD score and Na level at transplant failed to predict 1-year post-LT survival.

Conclusion Although SF may appear to predict listing outcome, it failed to do so independent of established predictors of outcome such as age, MELD and serum Na in this cohort. However, SF predicted 1-year post-LT mortality. SF merits further study in larger cohorts of patients.