LIVER TRANSPLANTATION 20:425–436, 2014
ORIGINAL ARTICLE
Operative Outcomes of Adult Living Donor Liver Transplantation and Deceased Donor Liver Transplantation: A Systematic Review and Meta-Analysis Ping Wan,1* Xin Yu,2* and Qiang Xia1 Department of Liver Surgery, Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China; and 2Department of Critical Care Medicine, Xiangya Hospital, Central South University, Changsha, China
1
Living donor liver transplantation (LDLT) has emerged as an alternative to deceased donor liver transplantation (DDLT) because of the increasing number of patients waiting for liver transplantation (LT). However, whether it can achieve operative outcomes similar to those achieved with DDLT for adult patients remains controversial. We conducted this metaanalysis to compare the operative outcomes of LDLT and DDLT recipients. A literature search was performed to identify clinical controlled studies comparing LDLT and DDLT that were published before October 2013. Four perioperative outcomes [duration of the recipient operation (DRO), red blood cell (RBC) transfusion requirement, length of the hospital stay, and cold ischemia time (CIT)] and 5 postoperative complication outcomes (biliary complications, vascular complications, intraabdominal bleeding, perioperative death, and retransplantation) were the main outcomes assessed. Nineteen studies with a total of 5450 patients were included in the meta-analysis. In comparison with DDLT, LDLT was associated with a significantly longer DRO and a shorter CIT. We found that biliary complications [odds ratio (OR) 5 3.08, 95% confidence interval (CI) 5 1.97-4.81, P < 0.001], vascular complications (OR 5 2.16, 95% CI 5 1.32-3.54, P 5 0.002), and retransplantation (OR 5 1.76, 95% CI 5 1.09-2.83, P 5 0.02) occurred more frequently for LDLT recipients, and the subgroup analysis indicated that the biliary complication rate decreased dramatically with greater LDLT experience. No significant difference was observed in RBC transfusion requirements, the lengths of hospital stays, intra-abdominal bleeding rates, or perioperative mortality between LDLT and DDLT recipients. In conclusion, LDLT is associated with a higher rate of surgical complications after transplantation. A reduction of postoperative complication rates can be achieved as centers gain greater experience with LDLT. However, LDLT is still an excellent alternative to DDLT because it facilitates access to LT. Liver Transpl C 2014 AASLD. 20:425-436, 2014. V Received November 18, 2013; accepted January 5, 2014.
Abbreviations: A2ALL, Adult-to-Adult Living Donor Liver Transplantation Cohort Study; CI, confidence interval; CIT, cold ischemia time; DDLT, deceased donor liver transplantation; DRO, duration of the recipient operation; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; LDLT, living donor liver transplantation; LT, liver transplantation; MELD, Model for End-Stage Liver Disease; NA, not available; OR, odds ratio; PRISMA, Preferred Reporting Items for Systematic Reviews and MetaAnalyses; RBC, red blood cell; SD, standard deviation; SRTR, Scientific Registry of Transplant Recipients; UNOS, United Network for Organ Sharing; WMD, weighted mean difference. There are no conflicts of interest to report. This study was supported by the Training Program for Superb Academic Leaders in the Shanghai Health System (XBR2011029) and the Special Fund for the Building of Leading Talent Teams in Shanghai. *These authors contributed equally to this work. Address reprint requests to Qiang Xia, M.D., Ph.D., Department of Liver Surgery, Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University, 1630 Dongfang Road, Shanghai 200127, China. Telephone: 186-21-68383775; FAX: 186-21-58737232; E-mail: [email protected] DOI 10.1002/lt.23836 View this article online at wileyonlinelibrary.com. LIVER TRANSPLANTATION.DOI 10.1002/lt. Published on behalf of the American Association for the Study of Liver Diseases
C 2014 American Association for the Study of Liver Diseases. V
426 WAN, YU, AND XIA
Liver transplantation (LT) is performed for irremediable end-stage liver disease. A shortage of donor organs continues to be the main obstacle to the treatment of this group of patients, even though the surgical techniques have progressed immensely. Living donor liver transplantation (LDLT) has emerged as an alternative to deceased donor liver transplantation (DDLT) because of the increasing number of patients waiting for LT. To those with end-stage liver disease, it offers the potential advantages of a shortened waiting time, an optimal donor graft, and the ability to optimize the recipient’s health.1,2 In the United States, the adoption of the Model for End-Stage Liver Disease (MELD) by the United Network for Organ Sharing (UNOS), on February 27, 2002, dramatically altered the allocation system for cadaveric organs.3,4 Results from the Adult-to-Adult Living Donor Liver Transplantation Cohort Study (A2ALL) further demonstrated that compelling survival benefits were associated with the receipt of LDLT in comparison with waiting for a deceased donor in the MELD liver allocation era.5 It is generally accepted by most researchers that LDLT can lead to long-term survival rates for adult patients comparable to those associated with DDLT, and this has been shown in many comparative cohorts.6,7 However, LDLT is criticized for its underlying higher rate of biliary complications in comparison with DDLT,8 and it remains characterized by its technical complexity and ethical controversies. Although the outcomes of perioperative results and postoperative complications have been investigated in several controlled trials, whether a distinct disparity exists between the 2 techniques in terms of these issues remains controversial. We, therefore, sought to compare the perioperative outcomes and postoperative surgical complication rates associated with LDLT and DDLT in adult patients through a meta-analysis of the published data.
MATERIALS AND METHODS Literature Search and Study Selection We identified publications by searching major medical databases such as MEDLINE, Embase, and the Cochrane Library for relevant articles published before October 2013. To supplement the primary search, we performed an additional search with Google Scholar. The search strategy consisted of a combination of the following terms: living donor liver transplantation, deceased donor liver transplantation, LDLT, DDLT, living donor, cadaveric donor, and liver transplantation. Relevant articles were also identified from the reference lists of previous articles. The study protocol was approved by the Science and Research Office of Ren Ji Hospital (Shanghai). Two authors (P.W. and X.Y.) carried out searches independently. There were no search restrictions with respect to language or publication type.
Inclusion and Exclusion Criteria The inclusion criteria were as follows: (1) published clinical studies comparing LDLT to DDLT in adult
LIVER TRANSPLANTATION, April 2014
patients and (2) studies with at least 1 of the aforementioned outcomes of interest. The exclusion criteria were the following: (1) reviews, case reports, letters, and editorials; (2) studies without a control group of DDLT recipients; (3) studies without available data; (4) studies investigating only patients with hepatocellular carcinoma (HCC); (5) studies investigating emergency LT or patients with acute liver failure; (6) studies focusing on deceased donor split LT; and (7) studies based on data from UNOS or Scientific Registry of Transplant Recipients (SRTR) databases. Moreover, several studies based on overlapping cohorts from the same institutions reported different outcome data because of their respective research aims. Those studies might have been included in the metaanalysis simultaneously, but only 1 study with available data and of better quality was used in each synthetic analysis for a single outcome.
Outcomes of Interest Perioperative outcomes and postoperative surgical complications were evaluated in the meta-analysis. The duration of the recipient operation (DRO), the allogeneic red blood cell (RBC) transfusion requirement, the length of the hospital stay, and the cold ischemia time (CIT) were the main perioperative outcomes to be assessed. Specific postoperative complications, including biliary complications, vascular complications, intra-abdominal bleeding, perioperative death, and retransplantation, were recorded and compared between LDLT and DDLT recipients.
Data Extraction and Quality Assessment Three observers (P.W., X.Y., and Q.X.) independently extracted data from each study with standardized forms and entered the data into a database. The extracted information included study characteristics (first author, region, year of publication, source journal, study design, and study period), population characteristics (sample size and ages, sexes, and diagnoses of patients), and 9 outcome parameters (DRO, RBC transfusion requirement, length of the hospital stay, CIT, biliary complications, vascular complications, intra-abdominal bleeding, perioperative death, and retransplantation). The NewcastleOttawa quality assessment scale9 was used for the quality assessment of each study, and this was also carried out independently by the 3 observers.
Statistical Analysis The meta-analysis was performed with RevMan 5.0.0 according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.10 Weighted mean differences (WMDs) with 95% confidence intervals (CIs) were calculated for continuous variables, whereas odds ratios (ORs) with 95% CIs were calculated for dichotomous variables. Data for continuous variables in the form of means and
LIVER TRANSPLANTATION, Vol. 20, No. 4, 2014
Figure 1.
WAN, YU, AND XIA 427
PRISMA 2009 flowchart providing information about the different phases of the systematic review.
standard deviations (SDs) were required for the statistical analysis. However, some of the published clinical trials did not report the means and SDs of parameters but rather reported the size of the trial and the medians and ranges. From these available statistics, estimates of the means and SDs were obtained with formulas proposed by Hozo et al.11 P values < 0.05 were considered to indicate statistical significance in the meta-analysis. Heterogeneity among studies was assessed with the Q statistic (P < 0.10 was considered to represent statistically significant heterogeneity) and the I2 statistic (I2 > 50% was considered to represent significant heterogeneity).12 If there was no significant heterogeneity, a fixed-effect model was used. Otherwise, a random-effect model was used, and a subgroup analysis was performed to explore and explain the differences in the results of different studies. Potential publication bias was assessed with funnel plots. We also conducted a sensitivity analysis in which each trial was excluded in turn to evaluate the influence of a single trial on the pooled estimate.
RESULTS Study Characteristics As shown in Fig. 1, the systematic literature search identified 196 relevant references. After screening titles and abstracts, we excluded 147 ineligible or irrelevant articles according to the exclusion criteria. The full text of the 49 remaining articles was retrieved for a formal review. After independent reviews, 30 full-text articles were excluded for the following reasons: (1) 11 studies were based on overlapping cohorts from the same institution, (2) 4 studies were based on UNOS or SRTR databases, and (3) 15 studies lacked available data. Finally, 19 published clinical cohort studies (4 prospective cohorts and 15 retrospective cohorts) were included in the meta-analysis.6,7,13-29 The characteristics of the 19 included studies are presented in Table 1. All the studies were well designed to compare the 2 arms: LDLT and DDLT. Ten of them investigated patient populations with
2004 2004 2005 2006 2007 2007 2007 2008 2008 2009 2009 2009 2009 2010 2010 2011 2011 2013 2013
Garcia-Retortillo et al.13 Van Vlierberghe et al.14 Schiano et al.15 Liu et al.16 Shah et al.17† Al-Sebayel et al.18 Schmeding et al.19 Selzner et al.20† Freise et al.21‡
G omez et al.22 Gallegos-Orozco et al.23 Fisher et al.7‡ Lai et al.24‡ Shin et al.25 Merion et al.26‡
Li et al.27§ Jain et al.28 Reichman et al.6† Jiang et al.29§
Spain Belgium United States Hong Kong Canada Saudi Arabia Germany Canada United States (A2ALL) Argentina United States United States United States Korea United States (A2ALL) Mainland China United States Canada Mainland China
Region
Retrospective Retrospective Retrospective Retrospective
Retrospective Retrospective Prospective Retrospective Retrospective Retrospective
Prospective Retrospective Prospective Prospective Retrospective Retrospective Retrospective Retrospective Retrospective
cohort cohort cohort cohort
cohort cohort cohort cohort cohort cohort
cohort cohort cohort cohort cohort cohort cohort cohort cohort
Study Design
NOTE: References with the same marker (†, ‡, or §) were based on overlapping cohorts. *The data are presented as means and SDs or as medians and ranges.
Year
Reference
128 35 145 70
30 32 107 86 77 384
22 17 11 124 153 45 20 46 384
LDLT
221 65 145 191
357 168 465 403 45 215
95 26 15 56 350 77 269 155 216
DDLT
Sample Size (n)
TABLE 1. Characteristics of the Included Studies
43 6 9 51 6 7 54 6 8 40 6 8
NA 54 6 9 49 6 12 51 6 12 52 (47-57) 49 6 11
59 (24-68) 56 6 4 54 (50-66) 48 (18-68) 51 6 10 47 (2-63) 56 6 9 52 (32-68) 50 6 11
LDLT
45 6 10 50 6 7 54 6 8 44 6 9
NA 53 6 6 52 6 9 54 6 11 52 (43-60) 51 6 10
59 (38-66) 60 6 7 53 (44-65) 48 (27-66) NA 44 (11-63) 51 6 10 53 (36-71) 51 6 10
DDLT
Recipient Age (Years)*
Patient
Benign HCV-related Mixed Benign (HBV-related)
Mixed HCV-related Mixed Mixed Mixed Mixed
HCV-related HCV-related HCV-related Mixed Mixed Mixed HCV-related HCV-related Mixed
Diagnosis
428 WAN, YU, AND XIA LIVER TRANSPLANTATION, April 2014
LIVER TRANSPLANTATION, Vol. 20, No. 4, 2014
WAN, YU, AND XIA 429
TABLE 2. Newcastle-Ottawa Scoring System for Cohort Studies Reference 13
Garcia-Retortillo et al. Van Vlierberghe et al.14 Schiano et al.15 Liu et al.16 Shah et al.17 Al-Sebayel et al.18 Schmeding et al.19 Selzner et al.20 Freise et al.21 G omez et al.22 Gallegos-Orozco et al.23 Fisher et al.7 Lai et al.24 Shin et al.25 Merion et al.26 Li et al.27 Jain et al.28 Reichman et al.6 Jiang et al.29
Selection Stars
Comparability Stars
Outcome Stars
Total Stars
4 3 4 4 3 3 3 3 3 3 3 4 3 3 3 3 3 3 3
2 2 1 2 2 2 2 2 2 0 2 1 1 2 2 2 2 2 2
3 3 3 3 3 3 3 3 3 3 3 3 3 1 3 3 3 3 3
9 8 8 9 8 8 8 8 8 6 8 8 7 6 8 8 8 8 8
Figure 2. Meta-analysis of controlled trials comparing DROs for LDLT and DDLT recipients on the basis of a random-effect model. There was significant heterogeneity in the results of the 5 studies. The pooled results showed a significant difference between LDLT and DDLT in terms of DRO, and DRO for LDLT was longer than that for DDLT.
various primary diagnoses, whereas 7 studies and 2 studies focused on patients with hepatitis C virus (HCV)–related diseases and patients with benign liver diseases, respectively. Three of the studies from Canada (conducted by Shah et al.,17 Selzner et al.,20 and Reichman et al.6) and 2 from mainland China (conducted by Li et al.27 and Jiang et al.29) were based on overlapping cohorts. Moreover, the data of 2 multicenter studies from the United States (conducted by Freise et al.21 and Merion et al.26) were derived from the A2ALL database, which included 9 US transplant centers with LDLT experience. Two other US singlecenter studies conducted by Fisher et al.7 and Lai et al.24 were based on 2 different cohorts, but their study populations may have overlapped with patients in the A2ALL study. We extracted different outcome data from the overlapping cohorts, and thus no overlapping data were involved in the synthesis for a single outcome. The quality assessment of the included studies is shown in Table 2, and all studies got 6 or more stars.
Perioperative Outcomes DRO Five studies reported DRO, and all of them showed that it was significantly longer for LDLT versus DDLT. The heterogeneity test indicated that there was significant heterogeneity in the results of the 5 studies (P 5 0.003, I2 5 76%), a random-effect model showed a significant difference between LDLT and DDLT in terms of DRO, and DRO was longer for LDLT versus DDLT (WMD 5 2.80, 95% CI 5 2.18-3.42, P < 0.001; Fig. 2).
Allogeneic RBC Transfusion Five studies provided data on the requirement for allogeneic RBC transfusions, with all of them showing no significant difference between LDLT and DDLT. Significant heterogeneity among the 5 studies was found for RBC transfusion requirements (P 5 0.006, I2 5 72%); therefore, a random-effect model was used to combine the data. Pooled results revealed that LDLT and DDLT
430 WAN, YU, AND XIA
LIVER TRANSPLANTATION, April 2014
Figure 3. Meta-analysis of controlled trials comparing allogeneic RBC transfusion requirements for LDLT and DDLT recipients on the basis of a random-effect model. There was significant heterogeneity in the results of the 5 studies. The pooled results showed no significant difference between LDLT and DDLT in terms of allogeneic RBC transfusion requirements.
Figure 4. Meta-analysis of controlled trials comparing lengths of hospital stay for LDLT and DDLT recipients on the basis of a fixedeffect model. There was no significant heterogeneity in the results of the 4 studies. The pooled results showed no significant difference between LDLT and DDLT in terms of lengths of hospital stay.
groups were comparable in terms of RBC transfusion requirements (WMD 5 20.85, 95% CI 5 23.37 to 1.66, P 5 0.51; Fig. 3).
Length of the Hospital Stay The length of the hospital stay was reported in 4 studies, and all of them suggested that no significant difference existed between LDLT and DDLT recipients. There was no significant heterogeneity observed in the length of the hospital stay among the 4 studies (P 5 0.12, I2 5 48%). Pooled results of a fixed-effect model were found to be equivalent between LDLT and DDLT for the length of the hospital stay (WMD 5 21.37, 95% CI 5 24.37 to 1.64, P 5 0.37; Fig. 4).
CIT Seven studies reported CITs, and all of them showed that CIT was significantly longer for DDLT versus LDLT. A random-effect model was used to combine the data because of evident heterogeneity among the studies (P < 0.001, I2 5 99%). According to the pooled data, CIT for the DDLT group was significantly longer than that for the LDLT group (WMD 5 2346.28, 95% CI 5 2417.10 to 2275.46, P < 0.001; Fig. 5).
Postoperative Complications Biliary Complications Biliary complication rates were reported in 8 of the included studies. In 7 studies, LDLT was associated
with significantly higher rates of biliary complications in comparison with DDLT; and only 1 study showed comparable results between the 2 groups. Significant heterogeneity was shown in the 8 studies in terms of biliary complication rates (P 5 0.008, I2 5 63%). Therefore, pooled results of a random-effect model revealed that a larger percentage of patients suffered from biliary complications in the LDLT group versus the DDLT group (OR 5 3.08, 95% CI 5 1.97-4.81, P < 0.001; Fig. 6).
Vascular Complications Four studies compared the vascular complication rates of LDLT and DDLT groups. Two reported that LDLT was associated with a significantly higher vascular complication rate than DDLT, whereas no significant difference was observed between the 2 groups in the remaining studies. The results of the 4 studies showed no heterogeneity (P 5 0.73, I2 5 0%), and pooled results of a fixed-effect model revealed that vascular complications tended to be more common in the LDLT group versus the DDLT group (OR 5 2.16, 95% CI 5 1.32-3.54, P 5 0.002; Fig. 7).
Intra-Abdominal Bleeding Three studies reported intra-abdominal bleeding rates after LT. The results of 2 studies showed no significant difference between LDLT and DDLT recipients, whereas the other study suggested a significantly higher intra-abdominal bleeding rate associated with
LIVER TRANSPLANTATION, Vol. 20, No. 4, 2014
WAN, YU, AND XIA 431
Figure 5. Meta-analysis of controlled trials comparing CITs for LDLT and DDLT recipients on the basis of a random-effect model. There was significant heterogeneity in the results of the 7 studies. The pooled results showed a significant difference between LDLT and DDLT in terms of CIT, and CIT for LDLT was shorter than that for DDLT.
Figure 6. Meta-analysis of controlled trials comparing the biliary complication rates after LDLT and DDLT on the basis of a randomeffect model. There was significant heterogeneity in the results of the 8 studies. The pooled results showed a significant difference between LDLT and DDLT in terms of biliary complication rates, and biliary complications occurred more frequently in LDLT recipients versus DDLT recipients.
Figure 7. Meta-analysis of controlled trials comparing the vascular complication rates after LDLT and DDLT on the basis of a fixedeffect model. There was no significant heterogeneity in the results of the 4 studies. The pooled results showed a significant difference between LDLT and DDLT in terms of vascular complication rates, and vascular complications occurred more frequently in LDLT recipients versus DDLT recipients.
DDLT recipients. There was no significant heterogeneity in the intra-abdominal bleeding rates of the 3 studies (P 5 0.23, I2 5 32%), and a fixed-effect model showed that the intra-abdominal bleeding rate for the LDLT group was similar to the rate for the DDLT group (OR 5 0.72, 95% CI 5 0.43-1.20, P 5 0.21; Fig. 8).
Perioperative Death Four studies provided data for perioperative mortality after LT, and all of them showed that there was no significant difference between LDLT and DDLT recipients. The heterogeneity test showed no significant heterogeneity in the perioperative mortality rates of the 4
432 WAN, YU, AND XIA
LIVER TRANSPLANTATION, April 2014
Figure 8. Meta-analysis of controlled trials comparing intra-abdominal bleeding rates after LDLT and DDLT on the basis of a fixedeffect model. There was no significant heterogeneity in the results of the 3 studies. The pooled results showed no significant difference between LDLT and DDLT in terms of intra-abdominal bleeding rates.
Figure 9. Meta-analysis of controlled trials comparing perioperative mortality rates after LDLT and DDLT on the basis of a fixed-effect model. There was no significant heterogeneity in the results of the 4 studies. The pooled results showed no significant difference between LDLT and DDLT in terms of perioperative mortality.
Figure 10. Meta-analysis of controlled trials comparing retransplantation rates after LDLT and DDLT on the basis of a fixed-effect model. There was no significant heterogeneity in the results of the 7 studies. The pooled results showed a significant difference between LDLT and DDLT in terms of retransplantation rates, and retransplantation occurred more frequently for LDLT recipients versus DDLT recipients.
studies (P 5 0.12, I2 5 49%). The pooled results of a fixed-effect model were found to be comparable for LDLT and DDLT recipients (OR 5 1.23, 95% CI 5 0.801.89, P 5 0.34; Fig. 9).
Retransplantation Data for retransplantation rates were provided by 7 studies. Five of these studies showed no significant
difference between LDLT and DDLT groups with respect to retransplantation rates, whereas 2 studies suggested that retransplantation occurred more frequently in LDLT recipients. No significant heterogeneity was found in the results of the 7 studies (P 5 0.24, I2 5 25%), and a fixed-effect model was used to combine the data. According to the pooled data, LDLT was associated with a significantly higher retransplantation rate than DDLT (OR 5 1.76, 95% CI 5 1.09-2.83, P 5 0.02; Fig. 10).
LIVER TRANSPLANTATION, Vol. 20, No. 4, 2014
WAN, YU, AND XIA 433
TABLE 3. Subgroup Analysis Studies Outcome
Subgroup
(n)
Effect Estimate (95% CI)
P Value
Heterogeneity
DRO
Prospective cohort Retrospective cohort Overall
2 3 5
3.60 (2.98-4.22) 2.32 (2.06-2.59) 2.80 (2.18-3.42)
ORIGINAL ARTICLE
Operative Outcomes of Adult Living Donor Liver Transplantation and Deceased Donor Liver Transplantation: A Systematic Review and Meta-Analysis Ping Wan,1* Xin Yu,2* and Qiang Xia1 Department of Liver Surgery, Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China; and 2Department of Critical Care Medicine, Xiangya Hospital, Central South University, Changsha, China
1
Living donor liver transplantation (LDLT) has emerged as an alternative to deceased donor liver transplantation (DDLT) because of the increasing number of patients waiting for liver transplantation (LT). However, whether it can achieve operative outcomes similar to those achieved with DDLT for adult patients remains controversial. We conducted this metaanalysis to compare the operative outcomes of LDLT and DDLT recipients. A literature search was performed to identify clinical controlled studies comparing LDLT and DDLT that were published before October 2013. Four perioperative outcomes [duration of the recipient operation (DRO), red blood cell (RBC) transfusion requirement, length of the hospital stay, and cold ischemia time (CIT)] and 5 postoperative complication outcomes (biliary complications, vascular complications, intraabdominal bleeding, perioperative death, and retransplantation) were the main outcomes assessed. Nineteen studies with a total of 5450 patients were included in the meta-analysis. In comparison with DDLT, LDLT was associated with a significantly longer DRO and a shorter CIT. We found that biliary complications [odds ratio (OR) 5 3.08, 95% confidence interval (CI) 5 1.97-4.81, P < 0.001], vascular complications (OR 5 2.16, 95% CI 5 1.32-3.54, P 5 0.002), and retransplantation (OR 5 1.76, 95% CI 5 1.09-2.83, P 5 0.02) occurred more frequently for LDLT recipients, and the subgroup analysis indicated that the biliary complication rate decreased dramatically with greater LDLT experience. No significant difference was observed in RBC transfusion requirements, the lengths of hospital stays, intra-abdominal bleeding rates, or perioperative mortality between LDLT and DDLT recipients. In conclusion, LDLT is associated with a higher rate of surgical complications after transplantation. A reduction of postoperative complication rates can be achieved as centers gain greater experience with LDLT. However, LDLT is still an excellent alternative to DDLT because it facilitates access to LT. Liver Transpl C 2014 AASLD. 20:425-436, 2014. V Received November 18, 2013; accepted January 5, 2014.
Abbreviations: A2ALL, Adult-to-Adult Living Donor Liver Transplantation Cohort Study; CI, confidence interval; CIT, cold ischemia time; DDLT, deceased donor liver transplantation; DRO, duration of the recipient operation; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; LDLT, living donor liver transplantation; LT, liver transplantation; MELD, Model for End-Stage Liver Disease; NA, not available; OR, odds ratio; PRISMA, Preferred Reporting Items for Systematic Reviews and MetaAnalyses; RBC, red blood cell; SD, standard deviation; SRTR, Scientific Registry of Transplant Recipients; UNOS, United Network for Organ Sharing; WMD, weighted mean difference. There are no conflicts of interest to report. This study was supported by the Training Program for Superb Academic Leaders in the Shanghai Health System (XBR2011029) and the Special Fund for the Building of Leading Talent Teams in Shanghai. *These authors contributed equally to this work. Address reprint requests to Qiang Xia, M.D., Ph.D., Department of Liver Surgery, Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University, 1630 Dongfang Road, Shanghai 200127, China. Telephone: 186-21-68383775; FAX: 186-21-58737232; E-mail: [email protected] DOI 10.1002/lt.23836 View this article online at wileyonlinelibrary.com. LIVER TRANSPLANTATION.DOI 10.1002/lt. Published on behalf of the American Association for the Study of Liver Diseases
C 2014 American Association for the Study of Liver Diseases. V
426 WAN, YU, AND XIA
Liver transplantation (LT) is performed for irremediable end-stage liver disease. A shortage of donor organs continues to be the main obstacle to the treatment of this group of patients, even though the surgical techniques have progressed immensely. Living donor liver transplantation (LDLT) has emerged as an alternative to deceased donor liver transplantation (DDLT) because of the increasing number of patients waiting for LT. To those with end-stage liver disease, it offers the potential advantages of a shortened waiting time, an optimal donor graft, and the ability to optimize the recipient’s health.1,2 In the United States, the adoption of the Model for End-Stage Liver Disease (MELD) by the United Network for Organ Sharing (UNOS), on February 27, 2002, dramatically altered the allocation system for cadaveric organs.3,4 Results from the Adult-to-Adult Living Donor Liver Transplantation Cohort Study (A2ALL) further demonstrated that compelling survival benefits were associated with the receipt of LDLT in comparison with waiting for a deceased donor in the MELD liver allocation era.5 It is generally accepted by most researchers that LDLT can lead to long-term survival rates for adult patients comparable to those associated with DDLT, and this has been shown in many comparative cohorts.6,7 However, LDLT is criticized for its underlying higher rate of biliary complications in comparison with DDLT,8 and it remains characterized by its technical complexity and ethical controversies. Although the outcomes of perioperative results and postoperative complications have been investigated in several controlled trials, whether a distinct disparity exists between the 2 techniques in terms of these issues remains controversial. We, therefore, sought to compare the perioperative outcomes and postoperative surgical complication rates associated with LDLT and DDLT in adult patients through a meta-analysis of the published data.
MATERIALS AND METHODS Literature Search and Study Selection We identified publications by searching major medical databases such as MEDLINE, Embase, and the Cochrane Library for relevant articles published before October 2013. To supplement the primary search, we performed an additional search with Google Scholar. The search strategy consisted of a combination of the following terms: living donor liver transplantation, deceased donor liver transplantation, LDLT, DDLT, living donor, cadaveric donor, and liver transplantation. Relevant articles were also identified from the reference lists of previous articles. The study protocol was approved by the Science and Research Office of Ren Ji Hospital (Shanghai). Two authors (P.W. and X.Y.) carried out searches independently. There were no search restrictions with respect to language or publication type.
Inclusion and Exclusion Criteria The inclusion criteria were as follows: (1) published clinical studies comparing LDLT to DDLT in adult
LIVER TRANSPLANTATION, April 2014
patients and (2) studies with at least 1 of the aforementioned outcomes of interest. The exclusion criteria were the following: (1) reviews, case reports, letters, and editorials; (2) studies without a control group of DDLT recipients; (3) studies without available data; (4) studies investigating only patients with hepatocellular carcinoma (HCC); (5) studies investigating emergency LT or patients with acute liver failure; (6) studies focusing on deceased donor split LT; and (7) studies based on data from UNOS or Scientific Registry of Transplant Recipients (SRTR) databases. Moreover, several studies based on overlapping cohorts from the same institutions reported different outcome data because of their respective research aims. Those studies might have been included in the metaanalysis simultaneously, but only 1 study with available data and of better quality was used in each synthetic analysis for a single outcome.
Outcomes of Interest Perioperative outcomes and postoperative surgical complications were evaluated in the meta-analysis. The duration of the recipient operation (DRO), the allogeneic red blood cell (RBC) transfusion requirement, the length of the hospital stay, and the cold ischemia time (CIT) were the main perioperative outcomes to be assessed. Specific postoperative complications, including biliary complications, vascular complications, intra-abdominal bleeding, perioperative death, and retransplantation, were recorded and compared between LDLT and DDLT recipients.
Data Extraction and Quality Assessment Three observers (P.W., X.Y., and Q.X.) independently extracted data from each study with standardized forms and entered the data into a database. The extracted information included study characteristics (first author, region, year of publication, source journal, study design, and study period), population characteristics (sample size and ages, sexes, and diagnoses of patients), and 9 outcome parameters (DRO, RBC transfusion requirement, length of the hospital stay, CIT, biliary complications, vascular complications, intra-abdominal bleeding, perioperative death, and retransplantation). The NewcastleOttawa quality assessment scale9 was used for the quality assessment of each study, and this was also carried out independently by the 3 observers.
Statistical Analysis The meta-analysis was performed with RevMan 5.0.0 according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.10 Weighted mean differences (WMDs) with 95% confidence intervals (CIs) were calculated for continuous variables, whereas odds ratios (ORs) with 95% CIs were calculated for dichotomous variables. Data for continuous variables in the form of means and
LIVER TRANSPLANTATION, Vol. 20, No. 4, 2014
Figure 1.
WAN, YU, AND XIA 427
PRISMA 2009 flowchart providing information about the different phases of the systematic review.
standard deviations (SDs) were required for the statistical analysis. However, some of the published clinical trials did not report the means and SDs of parameters but rather reported the size of the trial and the medians and ranges. From these available statistics, estimates of the means and SDs were obtained with formulas proposed by Hozo et al.11 P values < 0.05 were considered to indicate statistical significance in the meta-analysis. Heterogeneity among studies was assessed with the Q statistic (P < 0.10 was considered to represent statistically significant heterogeneity) and the I2 statistic (I2 > 50% was considered to represent significant heterogeneity).12 If there was no significant heterogeneity, a fixed-effect model was used. Otherwise, a random-effect model was used, and a subgroup analysis was performed to explore and explain the differences in the results of different studies. Potential publication bias was assessed with funnel plots. We also conducted a sensitivity analysis in which each trial was excluded in turn to evaluate the influence of a single trial on the pooled estimate.
RESULTS Study Characteristics As shown in Fig. 1, the systematic literature search identified 196 relevant references. After screening titles and abstracts, we excluded 147 ineligible or irrelevant articles according to the exclusion criteria. The full text of the 49 remaining articles was retrieved for a formal review. After independent reviews, 30 full-text articles were excluded for the following reasons: (1) 11 studies were based on overlapping cohorts from the same institution, (2) 4 studies were based on UNOS or SRTR databases, and (3) 15 studies lacked available data. Finally, 19 published clinical cohort studies (4 prospective cohorts and 15 retrospective cohorts) were included in the meta-analysis.6,7,13-29 The characteristics of the 19 included studies are presented in Table 1. All the studies were well designed to compare the 2 arms: LDLT and DDLT. Ten of them investigated patient populations with
2004 2004 2005 2006 2007 2007 2007 2008 2008 2009 2009 2009 2009 2010 2010 2011 2011 2013 2013
Garcia-Retortillo et al.13 Van Vlierberghe et al.14 Schiano et al.15 Liu et al.16 Shah et al.17† Al-Sebayel et al.18 Schmeding et al.19 Selzner et al.20† Freise et al.21‡
G omez et al.22 Gallegos-Orozco et al.23 Fisher et al.7‡ Lai et al.24‡ Shin et al.25 Merion et al.26‡
Li et al.27§ Jain et al.28 Reichman et al.6† Jiang et al.29§
Spain Belgium United States Hong Kong Canada Saudi Arabia Germany Canada United States (A2ALL) Argentina United States United States United States Korea United States (A2ALL) Mainland China United States Canada Mainland China
Region
Retrospective Retrospective Retrospective Retrospective
Retrospective Retrospective Prospective Retrospective Retrospective Retrospective
Prospective Retrospective Prospective Prospective Retrospective Retrospective Retrospective Retrospective Retrospective
cohort cohort cohort cohort
cohort cohort cohort cohort cohort cohort
cohort cohort cohort cohort cohort cohort cohort cohort cohort
Study Design
NOTE: References with the same marker (†, ‡, or §) were based on overlapping cohorts. *The data are presented as means and SDs or as medians and ranges.
Year
Reference
128 35 145 70
30 32 107 86 77 384
22 17 11 124 153 45 20 46 384
LDLT
221 65 145 191
357 168 465 403 45 215
95 26 15 56 350 77 269 155 216
DDLT
Sample Size (n)
TABLE 1. Characteristics of the Included Studies
43 6 9 51 6 7 54 6 8 40 6 8
NA 54 6 9 49 6 12 51 6 12 52 (47-57) 49 6 11
59 (24-68) 56 6 4 54 (50-66) 48 (18-68) 51 6 10 47 (2-63) 56 6 9 52 (32-68) 50 6 11
LDLT
45 6 10 50 6 7 54 6 8 44 6 9
NA 53 6 6 52 6 9 54 6 11 52 (43-60) 51 6 10
59 (38-66) 60 6 7 53 (44-65) 48 (27-66) NA 44 (11-63) 51 6 10 53 (36-71) 51 6 10
DDLT
Recipient Age (Years)*
Patient
Benign HCV-related Mixed Benign (HBV-related)
Mixed HCV-related Mixed Mixed Mixed Mixed
HCV-related HCV-related HCV-related Mixed Mixed Mixed HCV-related HCV-related Mixed
Diagnosis
428 WAN, YU, AND XIA LIVER TRANSPLANTATION, April 2014
LIVER TRANSPLANTATION, Vol. 20, No. 4, 2014
WAN, YU, AND XIA 429
TABLE 2. Newcastle-Ottawa Scoring System for Cohort Studies Reference 13
Garcia-Retortillo et al. Van Vlierberghe et al.14 Schiano et al.15 Liu et al.16 Shah et al.17 Al-Sebayel et al.18 Schmeding et al.19 Selzner et al.20 Freise et al.21 G omez et al.22 Gallegos-Orozco et al.23 Fisher et al.7 Lai et al.24 Shin et al.25 Merion et al.26 Li et al.27 Jain et al.28 Reichman et al.6 Jiang et al.29
Selection Stars
Comparability Stars
Outcome Stars
Total Stars
4 3 4 4 3 3 3 3 3 3 3 4 3 3 3 3 3 3 3
2 2 1 2 2 2 2 2 2 0 2 1 1 2 2 2 2 2 2
3 3 3 3 3 3 3 3 3 3 3 3 3 1 3 3 3 3 3
9 8 8 9 8 8 8 8 8 6 8 8 7 6 8 8 8 8 8
Figure 2. Meta-analysis of controlled trials comparing DROs for LDLT and DDLT recipients on the basis of a random-effect model. There was significant heterogeneity in the results of the 5 studies. The pooled results showed a significant difference between LDLT and DDLT in terms of DRO, and DRO for LDLT was longer than that for DDLT.
various primary diagnoses, whereas 7 studies and 2 studies focused on patients with hepatitis C virus (HCV)–related diseases and patients with benign liver diseases, respectively. Three of the studies from Canada (conducted by Shah et al.,17 Selzner et al.,20 and Reichman et al.6) and 2 from mainland China (conducted by Li et al.27 and Jiang et al.29) were based on overlapping cohorts. Moreover, the data of 2 multicenter studies from the United States (conducted by Freise et al.21 and Merion et al.26) were derived from the A2ALL database, which included 9 US transplant centers with LDLT experience. Two other US singlecenter studies conducted by Fisher et al.7 and Lai et al.24 were based on 2 different cohorts, but their study populations may have overlapped with patients in the A2ALL study. We extracted different outcome data from the overlapping cohorts, and thus no overlapping data were involved in the synthesis for a single outcome. The quality assessment of the included studies is shown in Table 2, and all studies got 6 or more stars.
Perioperative Outcomes DRO Five studies reported DRO, and all of them showed that it was significantly longer for LDLT versus DDLT. The heterogeneity test indicated that there was significant heterogeneity in the results of the 5 studies (P 5 0.003, I2 5 76%), a random-effect model showed a significant difference between LDLT and DDLT in terms of DRO, and DRO was longer for LDLT versus DDLT (WMD 5 2.80, 95% CI 5 2.18-3.42, P < 0.001; Fig. 2).
Allogeneic RBC Transfusion Five studies provided data on the requirement for allogeneic RBC transfusions, with all of them showing no significant difference between LDLT and DDLT. Significant heterogeneity among the 5 studies was found for RBC transfusion requirements (P 5 0.006, I2 5 72%); therefore, a random-effect model was used to combine the data. Pooled results revealed that LDLT and DDLT
430 WAN, YU, AND XIA
LIVER TRANSPLANTATION, April 2014
Figure 3. Meta-analysis of controlled trials comparing allogeneic RBC transfusion requirements for LDLT and DDLT recipients on the basis of a random-effect model. There was significant heterogeneity in the results of the 5 studies. The pooled results showed no significant difference between LDLT and DDLT in terms of allogeneic RBC transfusion requirements.
Figure 4. Meta-analysis of controlled trials comparing lengths of hospital stay for LDLT and DDLT recipients on the basis of a fixedeffect model. There was no significant heterogeneity in the results of the 4 studies. The pooled results showed no significant difference between LDLT and DDLT in terms of lengths of hospital stay.
groups were comparable in terms of RBC transfusion requirements (WMD 5 20.85, 95% CI 5 23.37 to 1.66, P 5 0.51; Fig. 3).
Length of the Hospital Stay The length of the hospital stay was reported in 4 studies, and all of them suggested that no significant difference existed between LDLT and DDLT recipients. There was no significant heterogeneity observed in the length of the hospital stay among the 4 studies (P 5 0.12, I2 5 48%). Pooled results of a fixed-effect model were found to be equivalent between LDLT and DDLT for the length of the hospital stay (WMD 5 21.37, 95% CI 5 24.37 to 1.64, P 5 0.37; Fig. 4).
CIT Seven studies reported CITs, and all of them showed that CIT was significantly longer for DDLT versus LDLT. A random-effect model was used to combine the data because of evident heterogeneity among the studies (P < 0.001, I2 5 99%). According to the pooled data, CIT for the DDLT group was significantly longer than that for the LDLT group (WMD 5 2346.28, 95% CI 5 2417.10 to 2275.46, P < 0.001; Fig. 5).
Postoperative Complications Biliary Complications Biliary complication rates were reported in 8 of the included studies. In 7 studies, LDLT was associated
with significantly higher rates of biliary complications in comparison with DDLT; and only 1 study showed comparable results between the 2 groups. Significant heterogeneity was shown in the 8 studies in terms of biliary complication rates (P 5 0.008, I2 5 63%). Therefore, pooled results of a random-effect model revealed that a larger percentage of patients suffered from biliary complications in the LDLT group versus the DDLT group (OR 5 3.08, 95% CI 5 1.97-4.81, P < 0.001; Fig. 6).
Vascular Complications Four studies compared the vascular complication rates of LDLT and DDLT groups. Two reported that LDLT was associated with a significantly higher vascular complication rate than DDLT, whereas no significant difference was observed between the 2 groups in the remaining studies. The results of the 4 studies showed no heterogeneity (P 5 0.73, I2 5 0%), and pooled results of a fixed-effect model revealed that vascular complications tended to be more common in the LDLT group versus the DDLT group (OR 5 2.16, 95% CI 5 1.32-3.54, P 5 0.002; Fig. 7).
Intra-Abdominal Bleeding Three studies reported intra-abdominal bleeding rates after LT. The results of 2 studies showed no significant difference between LDLT and DDLT recipients, whereas the other study suggested a significantly higher intra-abdominal bleeding rate associated with
LIVER TRANSPLANTATION, Vol. 20, No. 4, 2014
WAN, YU, AND XIA 431
Figure 5. Meta-analysis of controlled trials comparing CITs for LDLT and DDLT recipients on the basis of a random-effect model. There was significant heterogeneity in the results of the 7 studies. The pooled results showed a significant difference between LDLT and DDLT in terms of CIT, and CIT for LDLT was shorter than that for DDLT.
Figure 6. Meta-analysis of controlled trials comparing the biliary complication rates after LDLT and DDLT on the basis of a randomeffect model. There was significant heterogeneity in the results of the 8 studies. The pooled results showed a significant difference between LDLT and DDLT in terms of biliary complication rates, and biliary complications occurred more frequently in LDLT recipients versus DDLT recipients.
Figure 7. Meta-analysis of controlled trials comparing the vascular complication rates after LDLT and DDLT on the basis of a fixedeffect model. There was no significant heterogeneity in the results of the 4 studies. The pooled results showed a significant difference between LDLT and DDLT in terms of vascular complication rates, and vascular complications occurred more frequently in LDLT recipients versus DDLT recipients.
DDLT recipients. There was no significant heterogeneity in the intra-abdominal bleeding rates of the 3 studies (P 5 0.23, I2 5 32%), and a fixed-effect model showed that the intra-abdominal bleeding rate for the LDLT group was similar to the rate for the DDLT group (OR 5 0.72, 95% CI 5 0.43-1.20, P 5 0.21; Fig. 8).
Perioperative Death Four studies provided data for perioperative mortality after LT, and all of them showed that there was no significant difference between LDLT and DDLT recipients. The heterogeneity test showed no significant heterogeneity in the perioperative mortality rates of the 4
432 WAN, YU, AND XIA
LIVER TRANSPLANTATION, April 2014
Figure 8. Meta-analysis of controlled trials comparing intra-abdominal bleeding rates after LDLT and DDLT on the basis of a fixedeffect model. There was no significant heterogeneity in the results of the 3 studies. The pooled results showed no significant difference between LDLT and DDLT in terms of intra-abdominal bleeding rates.
Figure 9. Meta-analysis of controlled trials comparing perioperative mortality rates after LDLT and DDLT on the basis of a fixed-effect model. There was no significant heterogeneity in the results of the 4 studies. The pooled results showed no significant difference between LDLT and DDLT in terms of perioperative mortality.
Figure 10. Meta-analysis of controlled trials comparing retransplantation rates after LDLT and DDLT on the basis of a fixed-effect model. There was no significant heterogeneity in the results of the 7 studies. The pooled results showed a significant difference between LDLT and DDLT in terms of retransplantation rates, and retransplantation occurred more frequently for LDLT recipients versus DDLT recipients.
studies (P 5 0.12, I2 5 49%). The pooled results of a fixed-effect model were found to be comparable for LDLT and DDLT recipients (OR 5 1.23, 95% CI 5 0.801.89, P 5 0.34; Fig. 9).
Retransplantation Data for retransplantation rates were provided by 7 studies. Five of these studies showed no significant
difference between LDLT and DDLT groups with respect to retransplantation rates, whereas 2 studies suggested that retransplantation occurred more frequently in LDLT recipients. No significant heterogeneity was found in the results of the 7 studies (P 5 0.24, I2 5 25%), and a fixed-effect model was used to combine the data. According to the pooled data, LDLT was associated with a significantly higher retransplantation rate than DDLT (OR 5 1.76, 95% CI 5 1.09-2.83, P 5 0.02; Fig. 10).
LIVER TRANSPLANTATION, Vol. 20, No. 4, 2014
WAN, YU, AND XIA 433
TABLE 3. Subgroup Analysis Studies Outcome
Subgroup
(n)
Effect Estimate (95% CI)
P Value
Heterogeneity
DRO
Prospective cohort Retrospective cohort Overall
2 3 5
3.60 (2.98-4.22) 2.32 (2.06-2.59) 2.80 (2.18-3.42)
