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


This meta-analysis identified and analyzed a global population of 4571 LDLT and 66,826 DDLT recipients across a broad range of liver disease diagnoses, programs, and countries. The results confirm that LDLT recipients experience superior patient survival at 1-, 3-, and 5-years posttransplant when compared to DDLT recipients. LDLT resulted in equivalent graft survival when compared to DDLT at all time points. Preoperative MELD and waiting time favored LDLT recipients, and lower MELD at transplant was strongly associated with posttransplant survival on meta-regression. Moreover, despite a higher rate of biliary complications, LDLT recipients had a similar rate of HAT, risk of postoperative infection, and overall length of hospital stay and less rejection when compared to DDLT. Collectively, these data suggest that LDLT can offer several advantages when compared to DDLT.

The primary outcome of this meta-analysis, overall patient survival, identified a reduced risk of mortality of 17%, 15%, and 13% at 1, 3, and 5 years posttransplant respectively for LDLT recipients (Figure 2). Prior single center or consortium studies have also suggested that LDLT confers an overall survival advantage.[34–38] This finding is likely multifactorial, as shown by analysis of secondary outcomes, specifically preoperative variables indicating that LDLT recipients experience a shorter waiting time and are transplanted at a lower MELD (Figure 4). Indeed, meta-regression examining the correlation between MELD at transplant and patient survival confirmed a strong relationship exists (Figure 6). Other factors that likely contribute to superior outcomes for LDLT were not studied in this analysis. Generally, LDLT is an elective surgery and thus programs have the opportunity to screen and choose an ideal donor, schedule the procedure for the daytime with a highly specialized team, plan for anatomic variants, and optimize a recipient for surgery. Furthermore, a living donor allograft is not exposed to brain death, which may negatively affect both graft and patient survival.[39,40]

Analysis of the first secondary outcome, overall graft survival, demonstrated that graft survival is comparable between LDLT and DDLT for all time points (Figure 3). This is an important finding, as it suggests that the risk for early graft loss for DDLT and LDLT are equivalent. That being said, the risk profile for each type of donor is different. LDLT is a highly technical procedure, and as a consequence, poses a greater risk for procedure-related complications including vascular complications, biliary stricture or leak, early allograft dysfunction, or ultimately early graft loss requiring re-transplant. In countries with a predominant LDLT experience and thus lower rate overall rate of technical complications, such as Japan or Korea, national registry data have shown that 1-year graft survival modestly favors DDLT over LDLT.[12,16] Prior studies have reported variable outcomes for graft survival, ranging from equivalence between LDLT and DDLT, to improved graft survival for LDLT when compared to DDLT.[13,29,38,41–43]

Our analysis established that LDLT recipients had a lower MELD at transplant when compared to DDLT recipients, and this was associated with improved survival rates on meta-regression. This is consistent with the North American A2ALL cohort, which reported a lower MELD at transplant for LDLT recipients, with only 16% of LDLT recipients with MELD >20 at the time of transplant compared to 43% of DDLT recipients.[2] While LDLT candidates benefit from being transplanted at a lower MELD, studies have reported acceptable outcomes following LDLT even for higher MELD patients. A prior study comparing LDLT and DDLT with MELD >30 showed an improved overall patient survival for LDLT, even for patients with hepatorenal syndrome.[41] Similarly, single center studies from Taiwan and India have demonstrated that 5-year overall survival for LDLT with MELD >30 is comparable to the outcome in patients with MELD <30.[44,45]

A second preoperative variable that may influence patient survival is time on the waiting list. Even when including U.S. data, which showed a modestly longer waiting time for LDLT recipients, our comprehensive meta-analysis confirmed an overall shorter waiting time for LDLT recipients, which was not associated with overall survival on meta-regression (Figure 4; Table 2). Specific factors contributing to longer waiting time for LDLT recipients in the U.S. were beyond the scope of our study, but it is likely that variable local access to LDLT in different states and additional time for LDLT referral and donor evaluation are involved. Shorter waiting time for LDLT recipients may specifically benefit patient populations that may be disadvantaged in current allocation schemes: children, women, and patients with HCC.[46–48]

LDLT was associated with an increased incidence of arterial complications in the early era.[49,50] However, in this meta-analysis, no difference in risk of HAT was observed between LDLT and DDLT recipients. Studies from high-volume centers have confirmed this finding, as the rate of vascular complications has decreased over time, presumably as surgeons have gained experience and in some cases considered microvascular reconstruction.[13,15,51–53] A single center analysis of risk factors associated with HAT identified prolonged anastomosis time, perioperative blood transfusion, and graft to recipient weight ratio >1.15% as risk factors for early HAT.[54] One shortcoming of our analysis was the inability to effectively track HAT in the SRTR, and thus U.S. data were not included in examination of this variable.

Even with experience, early biliary complications are the recognized 'Achille's heel' in LDLT. Our meta-analysis confirmed that the risk of biliary complication was approximately two-fold higher in the LDLT group; however, there was no difference in graft survival between LDLT and DDLT and biliary complications did not negatively impact survival on meta-regression. A recent study from an experienced Japanese program reported a rate of biliary complications in LDLT of 17.3% and observed that multiple bile duct anastomoses and recurrent cholangitis prior to transplant were risk factors for biliary stricture or leak.[9] Our results are supported by a prior systematic review of biliary complications following LT, which identified MELD ≥35, multiple bile ducts, prolonged cold ischemic time, post–operative bile leak, and HAT as risk factors for biliary stricture for LDLT recipients on multivariable analysis.[19]

Postoperative infections and length of stay were similar among LDLT and DDLT in this meta-analysis. Prior single center studies have reported a higher incidence of bacterial infection in DDLT when compared to LDLT.[37,38,55] A Korean study identified receipt of a deceased donor allograft as an independent risk factor for postoperative infection (OR 5.5 [95% CI: 2.4–12.3]).[56] Length of stay is a difficult metric to study across different geographic regions, as practice patterns vary considerably. Even with regional variation, LDLT has been reported to be associated with a shorter length of stay in Canada (19 vs. 22 days), the U.S. (11 vs. 13 days), and China (42 vs. 45 days).[13,29,38]

This meta-analysis confirms that LDLT recipients have a lower risk of rejection when compared to DDLT (Figure 5E). Single center studies have shown that LDLT recipients experience a lower rate of biopsy-proven rejection at 24 months post-LT compared to DDLT recipients.[57,58] It has been postulated that prolonged cold ischemic time and exposure to the physiology of brain death can lead to inflammatory cell recruitment into the allograft, thereby disrupting liver immune homeostasis, a phenomenon that is reduced in LDLT.[59] A more recent study analyzing both A2ALL and OPTN data reported a lower risk of biopsy-proven acute rejection among biologically related LDLT when compared to non–biologically related LDLT and DDLT recipients, and more importantly, acute rejection was associated with increased risk of graft failure and death.[60] Thus, an additional factor that may relate to superior patient survival over time following LDLT is the lower rate of rejection episodes.

There are limitations to our study. By design, we required that eligible studies included a comparison cohort. As a consequence, studies from centers that exclusively performed either LDLT or DDLT were not included. While all available studies reporting outcomes of LDLT versus DDLT were included, data were screened by center to exclude studies that may have contained overlapping patient cohorts. The majority of the included studies were retrospective, and no randomized controlled studies were available. While 20 studies representing four continents were included, the U.S. data represent >50% of the LDLT and DDLT cohorts, which may have impacted some of the results. There were also inherent differences between LDLT and DDLT recipients in terms of age, sex, and etiology of underlying liver disease, that may have impacted our findings. Neither the SRTR analysis nor all studies examined reported on each of the secondary outcomes, potentially introducing bias and affecting the analysis. In particular, rejection, biliary and vascular complication are not consistently reported in the SRTR, limiting the possibility of including those data on analysis of secondary outcomes in this study. Additionally, there was heterogeneity among the studies, reflecting the differences in practice, protocols, and possibly in outcomes. Lastly, as per our study design, some outcomes were not considered, such as graft size or volume, technical details including anatomic variants, or the recurrence of disease and its impact on patient outcome.