Meta-Analysis and Meta-Regression of Outcomes for Adult Living Donor Liver Transplantation Versus Deceased Donor Liver Transplantation

Arianna Barbetta; Mayada Aljehani; Michelle Kim; Christine Tien; Aaron Ahearn; Hannah Schilperoort; Linda Sher; Juliet Emamaullee


American Journal of Transplantation. 2021;21(7):2399-2412. 

In This Article


Systematic Review

The literature review is summarized in a PRISMA diagram (Figure 1). After removal of duplications, 5364 abstracts were screened and 374 were selected for full-text review. A total of 19 studies from countries including Canada, China, France, Germany, South Korea, Italy, and Saudi Arabia were included in this meta-analysis (summarized in Table 1). Seventeen studies were from single centers and two included multi-center data. All studies but one were retrospective, while three had a matched-pair design and one was prospective. No randomized controlled studies were identified. The quality risk assessment for these studies determined that all met criteria for fair or good quality, and none showed poor design (Table S4).

Figure 1.

PRISMA diagram of systematic review


A total of 1821 LDLT and 8706 DDLT recipients were pooled from the published studies for inclusion in the meta-analysis; study and patient population characteristics are summarized in Table 1. When U.S. SRTR data were added, 4571 LDLT and 66,826 DDLT recipients were analyzed. For the entire study population, the mean age was 54.0 ± 9.9 years (51.2 ± 11.4 for LDLT vs 54.2 ± 9.7 for DDLT, p < .001) and 29.6% were female (33.8% of LDLT vs 28.9% of DDLT, p < .001). The most common etiology of liver disease was hepatocellular liver disease (autoimmune hepatitis, NASH or alcoholic liver disease; collectively 34.6%), followed by HCC (29.2%), viral hepatitis (26.3%), and cholestatic liver disease (7.7%).

Examination of our first primary outcome, patient survival, revealed superior overall patient survival for LDLT recipients when compared to the DDLT recipients (p < .0001, Figure 2). Specifically, LDLT recipients had a 17% reduction (95% CI 10–24) in the risk of mortality at 1-year posttransplant when compared to the DDLT group (HR 0.83 [95% CI 0.76–0.90]; p < .0001, Figure 2A). The survival benefit for LDLT recipients was also observed at both 3- and 5-years posttransplant (3 year: HR 0.85 [95% CI 0.79–0.92] and 5 year: HR 0.87 [95% CI 0.81–0.93], p < .0001 at both intervals, Figure 2B,C). Graft survival was studied as a secondary outcome. At all time points, graft survival was comparable between LDLT and DDLT recipients (1 year: HR 0.94 [95% CI 0.84–1.02], p = .14, 3 year: HR 0.96 [95% CI 0.89–1.03] p = .25, and 5 year: HR 0.95 [95% CI 0.88–1.01], p = .12) (Figure 3).

Figure 2.

Forest plot of hazard ratios for overall patient survival at 1 year (A), 3 years (B), and 5 years (C) posttransplant. LDLT favored patient survival when compared to DDLT at all time points

Figure 3.

Forest plot of hazard ratios for overall graft survival at 1 year (A), 3 years (B), and 5 years (C) posttransplant. LDLT and DDLT had equivalent graft survival at 1, 3, and 5 years posttransplant

Next, secondary outcomes were analyzed among sub-cohorts of studies that included the specified variables. Two preoperative outcomes were studied: MELD score at transplant and waiting time (days). As shown in Figure 4A, MELD score at transplant was lower for LDLT recipients when compared to DDLT recipients (MD −2.54 [95% CI −5.02, −0.06] p = .04). LDLT recipients had a shorter waiting time when compared to DDLT recipients (MD −71.43 [95% CI −101.42, −41.44], p < .0001, Figure 4B). Post–operative technical complications including HAT and biliary complications were analyzed. While there was no difference between the two groups in the risk of HAT (OR 2.07 [95% CI 0.84–5.09], p = 0.11, Figure 5A), LDLT recipients experienced an approximately two-fold increase in the risk of biliary complications (OR 2.14 [95% CI 1.76–2.59], p < .001, Figure 5B). Pooled analysis for the risk of postoperative infection and length of hospital stay showed no difference between LDLT and DDLT recipients (OR 0.67 [95% CI 0.42–1.09], p = .11 [Figure 5C] and MD −3.80 [95% CI −8.36, 0.76], p = .10 [Figure 5D], respectively). Finally, LDLT recipients showed a lower risk of rejection when compared to DDLT recipients (OR 0.72 [95% CI 0.55–0.95], p = .02, Figure 5E).

Figure 4.

Forest plot of preoperative variables. (A) MELD at transplant and (B) time on waiting list. LDLT favored lower MELD at transplant and less time on the waiting list

Figure 5.

Forest plot of postoperative variables. (A) Hepatic artery thrombosis, (B) biliary complications, (C) risk of infection, (D) length of stay, (E) rejection rate. LDLT was equivalent to DDLT for rates of postoperative HAT (A), infections, and length of stay (D). LDLT were more likely to have biliary complications (B) and had a lower risk of rejection when compared to DDLT (E) [Color figure can be viewed at]

Meta-regression Analysis

A meta-regression analysis was completed to explore potential relationships between MELD at transplant, time on waitlist and biliary complications and 1-year patient survival (Table 2). MELD score and time on waitlist were expressed as weighted mean differences between LDLT and DDLT means, whereas biliary complications were expressed as difference of rate of occurrence in the LDLT versus DDLT. MELD score at LT was the sole variable that demonstrated a relationship with 1-year patient survival (Figure 6). These data indicate that as MELD score difference increased, survival at 1-year post-LT decreased. Time on waitlist and biliary complications had no impact on 1-year patient survival. The inclusion of MELD score as a moderator in the meta-regression of 1-year patient survival explained most of the observed heterogeneity in the relative risk of death (R 2 0.56, p = .02, Figure 6).

Figure 6.

Random effects meta-regression showing how results of meta-analysis examining 1-year patient survival are influenced by the difference in MELD score between LDLT and DDLT. Each dot represents an individual study, the solid line represents the regression prediction, and the dotted lines the 95% confidence intervals