Reassessing the Impact of Donor HLA-C Genotype on Long-Term Liver Transplant Survival


* Corresponding author: Gerhard Opelz,


HLA-C is the major inhibitory ligand for killer immunoglobulin-like receptors (KIRs) that are expressed on natural killer (NK) cells. Based on their KIR specificity, HLA-C alleles can be divided into two groups, termed HLA-C1 and HLA-C2. Donor HLA-C group has recently been identified by Hanvesakul et al. (Am J Transplant 2008) as a critical determinant of clinical outcome following liver transplantation: Possession of at least one HLA-C group 2 allele by the donor was associated with significantly improved long-term graft and patient survival, presumably due to an inhibition of host NK cell function. To verify this study, we performed genotyping of 913 deceased liver donors for the relevant KIR epitopes of HLA-C and correlated the presence or absence of donor HLA-C2 genotype with graft and patient survival. In our study, donor HLA-C2 genotype had no impact on 10-year graft or patient survival. We cannot confirm a major role of donor HLA-C2 genotype on long-term allograft survival after liver transplantation.


Natural killer (NK) cells are large granular lymphocytes that play an important role in innate immunity (1). They express a variety of inhibitory and activating receptors including killer immunoglobuline-like receptors (KIRs) (2). Most KIRs bind to classical human leukocyte antigen (HLA) class I ligands (HLA-A3, -A11, HLA proteins carrying a Bw4 motif and HLA-C). Based on the amino acid residue at position 80 (either asparagine or lysine) that determines their NK specificity, HLA-C alleles can be divided into two groups, termed HLA-C1 and HLA-C2. HLA-C1 alleles bind to KIR2DL2/L3 while HLA-C2 alleles are recognized by KIR2DL1. Upon engagement of HLA-C, these KIRs can mediate inhibition of NK cell responses (2,3).

In contrast to kidney transplantation (4), HLA-matching for the loci HLA-A, -B and -DR does not appear to provide a benefit to graft survival following liver transplantation (5). However, Hanvesakul et al. (6) recently reported that donor HLA-C genotype had a profound impact on clinical outcome in a study of 416 liver transplants. A single HLA-C2 allele present in the donor conferred a 16.2% improvement in graft survival, and HLA-C2 homozygous allografts had a striking 26.5% graft survival benefit at 10 years compared with HLA-C1 homozygous allografts. The authors postulated that expression of an HLA-C2 allele by the allograft is associated with potent inhibition of host NK cell function through HLA-C2-KIR2DL1 interaction, leading to a reduction in graft injury. These impressive results, which if confirmed would have substantial clinical implications, prompted us to address the association between donor HLA-C2 genotype and clinical outcomes in the liver transplant cohort of the collaborative transplant study (CTS).

Materials and Methods

Liver donors and recipients

This study was based on DNA samples and clinical data provided by transplant centers that participate in the CTS (see Acknowledgment for listing of participating centers). We included all deceased donor liver transplants from Caucasian donors into Caucasian recipients on which donor DNA, as well as clinical transplant follow-up data were available. A total of 913 transplants were tested; 855 (93.6%) were first grafts. Patient demographics are shown in Table 1. Genotyping was performed in Belfast, Northern Ireland, without knowledge of clinical outcomes. Data analysis was carried out in Heidelberg, Germany. The centers included in this analysis provided written assurance of compliance with local ethical and consent guidelines for the use of data for scientific analysis. The study was approved by the ethics committee of the University of Heidelberg, Germany.

Table 1.  Demographic characteristics of study population
CharacteristicMissing (%)Donor HLA C2P
Negative n = 331Positive n = 582
  1. Statistics performed using Fisher's exact test, *Mann–Whitney U-test, Mantel–Haenszel test.

Geographic origin0.0   
 Europe 264 (80%)486 (84%)0.18
 North America  67 (20%) 96 (16%) 
Transplant year0.0   
 1988–1997 147 (44%)207 (36%) 0.009
 1998–2007 184 (56%)375 (64%) 
Transplant number0.0   
 First transplant 312 (94%)543 (93%)0.67
 Retransplant 19 (6%)39 (7%) 
Recipient gender0.0   
 Female 128 (39%)192 (33%)0.10
 Male 203 (61%)390 (67%) 
Recipient age0.8   
 yrs, mean ± SD 48.5 ± 14.547.7 ± 14.4 0.26*
Donor age0.4   
 yrs, mean ± SD 38.8 ± 16.639.4 ± 16.5 0.50*
Cold ischemia time4.5   
 hrs, mean ± SD 8.2 ± 3.18.1 ± 3.0 0.55*
Original disease0.9   
 Acute 13 (4%)20 (3%)0.95
 Autoimmune  51 (16%) 81 (14%) 
 Cirrhosis 213 (65%)387 (67%) 
 Malign 22 (7%)34 (6%) 
 Metabolic 11 (3%)23 (4%) 
 Other 17 (5%)33 (6%) 
General evaluation8.4   
 Good 206 (67%)368 (70%)0.17
 Moderate  57 (19%)108 (20%) 
 Poor  44 (14%)53 (10%) 
 Normal 249 (80%)448 (83%)0.41
 Urgent  62 (20%) 95 (17%) 
Graft size0.0   
 Full 316 (95%)556 (96%)1.00
 Reduced or split 15 (5%)26 (4%) 

HLA-C group genotyping

DNA was extracted by the salting-out method (7) from peripheral blood or spleen tissue that had been sent by participating transplant centers to the CTS center in Heidelberg. DNA samples were shipped in a blinded fashion to the laboratory in Belfast where molecular typing was carried out by polymerase chain reaction with sequence-specific oligonucleotide probes. HLA-C1 and -C2 groups were defined using two probes that differentiated nucleotides at positions 77 and 80 as previously described (8).

Clinical data

Graft loss was defined as graft failure requiring retransplantation or resulting in patient death. No patients were excluded from this study for any reason.

Statistical analysis

Graft survival rates were computed according to the Kaplan–Meier method and expressed as mean percentage ± standard error. Graft or patient survival was compared for patients receiving HLA-C2 positive with those receiving HLA-C2 negative donor livers. Cox regression analysis was performed to determine hazard ratios (HR) considering the following confounders: geographic origin (continent), year of transplant, first or regraft, recipient and donor gender and age, cold ischemia time, original disease leading to transplantation, general evaluation of the patient by the transplant center at time of transplant as a candidate for transplantation, urgency and graft size (whole liver or reduced size). The p-values below 0.05 were considered significant. The software package SPSS (version 15.0, SPSS Inc., Chicago, IL) was used for all statistical analyses.


Effect of donor HLA-C group genotype on overall survival

Patients were divided into two groups: those who received liver grafts from HLA-C2 positive (C2 homozygous or C1/C2 heterozygous) and those who received liver grafts from HLA-C2 negative (C1 homozygous) donors. The genotype frequencies were checked for Hardy–Weinberg equilibrium and showed no significant deviation. As shown in Figure 1A, 10-year graft survival was 56.5 ± 2.5% in the presence and 53.4 ± 3.3% in the absence of donor HLA-C2 genotype (p = 0.37). Similarly, no effect of donor HLA-C2 genotype was seen on 10-year patient survival (59.7 ± 2.5% vs. 58.2 ± 3.3%, respectively, p = 0.62) (Figure 1B). Multivariate Cox regression analysis showed no significant effect for graft failure (HR 0.95 for the presence vs. the absence of HLA-C2 in the donor, 95% confidence interval (CI) 0.76–1.19, p = 0.69) or patient death (HR 1.05, 95% CI 0.82–1.33, p = 0.71). Because Hanvesakul et al. (6) reported a gene dose effect of HLA-C group 2 alleles on graft survival (highest survival rate for grafts from HLA-C2 homozygous donors), we compared the graft survival rates between three possible groups of donor HLA-C genotypes in our transplant cohort. No significant difference in 10-year graft survival was found between transplants with HLA-C2 homozygous donors (HR 1.12, CI 0.81–1.54, p = 0.50), HLA-C1/C2 heterozygous donors (HR 0.90, CI 0.71–1.15, p = 0.41) and transplants with HLA-C1 homozygous donors (HR 1.00, reference).

Figure 1.

Comparison of (A) graft survival and (B) patient survival in recipients of HLA-C2 positive or HLA-C2 negative liver allografts. HLA-C2 positive or negative donors and transplant numbers in follow-up at posttransplant intervals are indicated.

Subset analysis

Separate analysis was performed for recipients transplanted in the early period from 1992 to 1997, for which complete 10-year follow-up was available (n = 321). The presence of an HLA-C2 allele in the liver donor was not associated with better graft survival (HLA-C2 positive vs. HLA-C2 negative donor: HR 0.98, 95% CI 0.71–1.37, p = 0.93) (Figure 2A).

Figure 2.

Comparison of graft survival in (A) patients transplanted during the period 1992–1997 who received HLA-C2 positive or HLA-C2 negative liver allografts, and (B) patients with primary diagnosis cirrhosis who received HLA-C2 positive or negative liver allografts.

As cirrhosis represents the main indication for liver transplantation, we carried out a separate analysis to assess whether donor HLA-C2 group genotype may have a beneficial effect in this subgroup of patients (n = 600). No improvement of 10-year graft survival was observed in cirrhotic patients who were transplanted from HLA-C2 positive donors compared with recipients of grafts from HLA-C2 negative donors (HR 1.12, 95% CI 0.85–1.49, p = 0.43, Figure 2B). Moreover, we found no significant impact on graft survival of donor HLA-C2 genotype in patients who had a positive or negative pretransplant serostatus for hepatitis C virus (HCV). Neither in the 31.2% of recipients who were HCV positive nor in the 68.8% who were HCV negative did the difference between grafts from HLA-C2 positive or negative donors reach statistical significance (HR 0.80, 95% CI 0.54–1.17, p = 0.25, and HR 1.06, 95% CI 0.79–1.43, p = 0.69, respectively).


Hanvesakul et al. (6) recently reported that possession of an HLA-C2 allele by the donor was associated with significantly higher 10-year graft and patient survival after liver transplantation. This would represent a single, clinically very useful prognostic parameter, easily determinable by routine genotyping. Recipient KIR genotype or recipient/donor compatibility for HLA-C group ligand was not found to influence transplant survival. The study was well designed, including the analysis of two cohorts that both showed a beneficial effect of donor HLA-C2 genotype on long-term survival. This potentially important finding prompted us to specifically validate the relevance of donor HLA-C group genotype in the liver transplant cohort of the CTS. All 913 transplants on whom donor DNA was available were included. We could not confirm the findings of Hanvesakul et al. In our study, donor HLA-C2 genotype did not have a significant impact on 10-year graft and patient survival. Analysis of patient subgroups, such as transplantations carried out during an earlier period (similar to the period 1992–1999 analyzed in the Birmingham study), cirrhotic patients, or consideration of pretransplant hepatitis C serostatus also did not show positive results. The reason for the discrepancy between our results and the report of the Birmingham group is unclear, but there are some notable differences between the two studies. Whereas all transplant recipients examined by Hanvesakul et al. were treated at one center, the CTS analysis contained transplants from 14 centers. The number of transplants studied by us was 913 as compared to the 416 transplants in the Birmingham study. We think that, with this study aimed specifically at validating the impact of a genetic polymorphism on graft outcomes, analysis of a large sample size outweighs the potential disadvantage of greater patient heterogeneity due to the participation of 14 transplant centers. The striking result reported by Hanvesakul et al. was strictly limited to donor HLA-C group genotype. In our attempt at validation we did not reinvestigate recipient KIR or HLA-C genotype, factors that had been found irrelevant by Hanvesakul et al.

The contradicting results of our present study and the study by the Birmingham group highlight the complexity of studies on genetic associations with transplant outcomes. Indeed, due to frequent difficulties in replicating initial findings in follow-up studies by independent groups, criticism and skepticism has been raised about gene association studies (9,10). Our entirely negative result was surprising to us because the Birmingham group reported a very strong effect of HLA-C2 alleles, showing a 26.5% reduction in graft loss in transplants with HLA-C2 homozygous donors compared with HLA-C1 homozygous donors (HR 7.2, p < 0.001). In our study, neither the entire cohort nor any subgroup showed a positive association of donor HLA-C2 alleles with graft survival. Not only at 10 years, but also at intervals up to 10 years, there was no significant difference in graft survival between HLA-C2 positive and HLA-C2 negative grafts (data not shown).

Liver transplantation is a complex medical procedure subject to the influence of multiple factors. Any proposal of a single genetic determinant being decisive for transplant outcome, albeit attractive, needs careful interpretation and validation. According to the results of this study, the role of donor HLA-C genotype in liver transplantation must be considered unresolved. To evaluate the implication of HLA-C genotype as a potential tool for liver allocation or risk-adapted modulation of immunosuppressive regimens, as suggested by the Birmingham group, requires further investigation.


We are very grateful to the following transplant centers participating in the liver transplant DNA typing project of the CTS for their invaluable support: Bern (Switzerland), Budapest (Hungary), Essen (Germany), Freiburg (Germany), Geneva (Switzerland), Heidelberg (Germany), Innsbruck (Austria), Leuven (Belgium), Portland (USA), Strasbourg (France), Torino (Italy), Toronto (Canada), Zagreb (Croatia) and Zurich (Switzerland).

Conflict of Interest Statement