Identifying the Superior Reperfusion Technique in Liver Transplantation: A Network Meta-Analysis

Abstract EN

Objective.

To investigate the clinical effects of different reperfusion techniques in liver transplantation based on network meta-analysis.

Method.

Literature retrieval was conducted in globally recognized databases, namely, MEDLINE, EMBASE, and Cochrane Central, to address relative randomized controlled trials (RCTs) investigating the clinical effects of respective reperfusion techniques in liver transplantation.

Short- and long-term parametric data, including ICU stay, dysfunction rate (DFR), biliary complications (BC), 1-year graft survival (GS), and patient survival (PS), were quantitatively pooled and estimated based on the Bayesian theorem.

The P values of surface under the cumulative ranking (SUCRA) probabilities regarding each parameter were calculated and ranked by various techniques.

The Grades of Recommendations Assessment, Development and Evaluation (GRADE) criteria were utilized for the recommendations of evidence from pairwise direct comparisons.

Results.

Seven RCTs containing 6 different techniques were finally included for network meta-analysis.

The results indicated that retrograde vena cava (RVC) reperfusion possessed the highest possibility of revealing the best clinical effects on DFR (SUCRA, P=0.93), ICU stay (SUCRA, P=0.76), and GS (SUCRA, P=0.44), while portal-arterial reperfusion (simultaneous initialize) seemed to exhibit the most benefits in reducing BC (SUCRA, P=0.67) and enhancing PS rate (SUCRA, P=0.48).

Moreover, sensitivity analysis with the inconsistency approach clarified the reliability of the main results, and the evidence of the most direct comparisons was ranked low or very low.

Conclusions.

Current evidence demonstrated that RVC and portal-arterial reperfusion (simultaneously initialized) revealed superior clinical effects, compared to other interventions.

Investigation of these 2 techniques should be a future research direction, and more high-quality RCTs are expected.