Premium
Prognostication of inflammatory cells in liver transplantation: Is the waitlist neutrophil‐to‐lymphocyte ratio really predictive of tumor biology?
Author(s) -
McVey John C.,
Sasaki Kazunari,
Firl Daniel J.,
Fujiki Masato,
DiagoUso Teresa,
Quintini Cristiano,
Eghtesad Bijan,
Miller Charles C.,
Hashimoto Koji,
Aucejo Federico N.
Publication year - 2019
Publication title -
clinical transplantation
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.918
H-Index - 76
eISSN - 1399-0012
pISSN - 0902-0063
DOI - 10.1111/ctr.13743
Subject(s) - medicine , hepatocellular carcinoma , quartile , hazard ratio , gastroenterology , liver transplantation , proportional hazards model , neutrophil to lymphocyte ratio , lymphocyte , retrospective cohort study , oncology , transplantation , confidence interval
Objective The objective of this retrospective study was to characterize the neutrophil to lymphocyte ratio ( NLR ) on the waitlist and determine its prognostic utility in liver transplantation ( LT ) for hepatocellular carcinoma ( HCC ) with special focus on longitudinal data. Biomarkers such as pre‐operative NLR have been suggested to predict poor oncological outcomes for patients with HCC seeking LT . NLR 's utility is thought to be related to tumor biology. However, recent studies have demonstrated that a high NLR conveys worse outcomes in non‐ HCC cirrhotics. This study investigated the relationship between NLR , liver function, tumor factors and patient prognosis. Methods Patients with HCC undergoing LT were identified between 2002 and 2014 (n = 422). Variables of interest were collected longitudinally from time of listing until LT . The prognostic utility of NLR was assessed using Kaplan‐Meier and Cox Proportional Hazard regression. Associations between NLR and MELD ‐Na, AFP , and tumor morphology were also assessed. Results NLR demonstrated a positive correlation with MELD ‐Na at LT (R2 = 0.125, P < 0.001) and had parallel trends over time. The lowest NLR quartile had a median MELD ‐Na of 9 while the highest had a median MELD ‐Na of 19. There were minimal differences in AFP , tumor morphology, and rates of vascular invasion between quartiles. NLR was a statistically significant predictor of OS ( HR = 1.64, P = 0.017) and recurrence ( HR = 1.59, P = 0.016) even after controlling for important tumor factors. However, NLR lost its statistical significance when MELD ‐Na was added to the Cox regression model ( OS : HR = 1.46, P = 0.098) (recurrence: HR = 1.40, P = 0.115). Conclusions NLR is a highly volatile marker on the waitlist that demonstrates a significant correlation and collinearity with MELD ‐Na temporally and at the time of LT . These characteristics of NLR bring into question its utility as a predictive marker in HCC patients.