Prevalence, Course and Impact of HLA Donor-Specific Antibodies in Liver Transplantation in the First Year


Julie K. Heimbach,


The presence of preformed donor-specific HLA antibodies (DSA) in liver transplant recipients is increasingly recognized; however, the prevalence of DSA and their impact on early allograft function remains unknown. We prospectively followed serum DSA levels of 90 consecutive liver transplant recipients from baseline to 4 months. Twenty recipients (22.2%) had preformed DSA. No antibody-targeting treatments were undertaken. Seven days after transplantation, DSA levels decreased markedly in all but three patients. Day 7 protocol biopsies showed diffuse C4d deposition along the portal stroma, central vein, subendothelial and stromal space in the patients with persistent high DSA levels. The rate of acute cellular rejection was not significantly different in patients with DSA. The transaminase and bilirubin levels remained comparable during the first year despite the presence of DSA. The three patients with persistently high DSA levels continue to have normal allograft function. We conclude that in most cases, DSA disappear after liver transplant, however in rare instances where they persist, there is evidence of complement activation in the liver allograft, without significant clinical impact in the first year.


anti-human globulin augmented complement-dependent cytotoxicity


alcoholic liver disease


alanine aminotransferase


aspartate aminotransferase




cold ischemia time


donor risk index


donor-specific HLA antibodies


hematoxylin and eosin


hepatocellular carcinoma


hepatitis C virus


model for end-stage liver disease


mean fluorescence intensity


nonalcoholic steatohepatitis


primary biliary cirrhosis


primary sclerosing cholangitis


warm ischemia time




The presence of preformed donor-specific HLA antibodies (DSA) in patients undergoing kidney (1,2) and heart (3) transplantation is closely associated with higher graft loss rates. Thus, strategies to prevent poor outcomes in these patients, such as better matching and pretransplant desensitization have been developed. In contrast, the role of preformed HLA DSA in liver transplantation remains controversial. Despite the early recognition of their potential role in liver allograft rejection (4), positive pretransplant cross-match has not been associated with adverse outcomes in large, retrospective analyses (5–8). Likewise, no histopathological criteria exist for antibody-mediated rejection of the liver allograft. In the absence of consistent data correlating DSA to poor graft survival, it has become common practice to perform liver transplant without cross-match results, to prevent prolonged ischemia time.

Recently, several groups have reported that the preformed DSA could be associated with both early (9–11) and chronic (12) liver rejection. These retrospective studies clearly demonstrate a link between DSA and poor liver allograft survival, however they are skewed toward recipients with poor outcomes. Thus, in order to ascertain whether presensitization needs to be included in the decision-making process at the time of organ acceptance, and whether the evidence for a clinical entity of antibody-mediated rejection in sensitized liver transplant recipients, similar to those in kidney transplant, exist, one needs to identify in a systematic, prospective cohort of consecutive liver allograft recipients: (1) the prevalence of preformed HLA DSA, (2) kinetics of the DSA levels after transplantation and (3) the impact of these antibodies on the liver allograft function and histology in correlation with their circulating levels at predetermined time points. Here, our aim was to answer these questions in a prospective analysis of consecutive liver transplants in our center.

Materials and Methods


This study was conducted under approval of the Institutional Review Board of the Mayo Foundation and Clinic. Ninety-four consecutive liver transplant recipients between July 1, 2009 and May 31, 2010 were enrolled in the study. Four perioperative deaths (within 30 days of transplantation; all with no DSA) were excluded from the analyses. All patients received methylprednisolone (500 mg i.v.) intraoperatively and standard triple maintenance immunosuppression, including tacrolimus (target daily trough level of 6–10 ng/mL), prednisone taper (first 120 days) and mycophenolate mofetil (first 60 days). On postoperative day 7, all patients underwent a protocol biopsy of the liver. In case of cellular rejection, patients were administered bolus methylprednisolone (three doses of 1000 mg i.v., every 48 h), in addition to maintenance immunosuppression.

Cross-match and DSA detection

Routine baseline flow cytometric cross-match and T-cell complement-dependent cytotoxicity (AHG-CDC) were done, as described previously (13), in all recipients. Mean channel shifts of ≥52 for the T cell, and ≥106 for the B cell as compared to the negative control on 1024 scale were considered a positive cross-match. LABScreen® Single Antigen class I and class II (One Lambda Inc., CA, USA) beads were used for detection of DSA. These were then read on a Luminex (One Lambda Inc.) platform. The results of the DSA levels normalized for background are reported as the mean fluorescence index (MFI). MFI ≥ 2000 was considered positive for a single DSA. In cases where more than one DSA was positive, the cumulative MFI value (i.e. each class I DSA + each class II DSA MFI) was used for the total DSA.

Liver allograft biopsies

Protocol needle core liver biopsies were obtained on postoperative day 7 in all recipients, regardless of the DSA status. The formalin-fixed paraffin embedded (FFPE) specimens were cut as standard 4-micron-thick sections, placed on charged slides and stained with hematoxylin and eosin (H&E) for routine diagnostic evaluation. Standard immunohistochemistry for C4d (American Research Products Inc., Belmont, MA, USA) was also performed on FFPE specimens. All slides were evaluated by a liver pathologist, who was blinded to the clinical data and DSA status. The C4d staining was assessed for location (subendothelial, venular, sinusoidal and portal) and intensity (mild, moderate or severe). Hepatic arteriolar C4d staining was not assessed, as its deposition on internal elastic lamina has been reported in normal, nontransplant liver tissue before (10). Accompanying cellular rejection was reported using standard criteria from Banff schema for grading liver allograft rejection. Twenty normal liver specimens (nontransplant liver resections) were stained for C4d as control.

Statistical analysis

Categorical variables were compared using Fisher's exact test. Paired Student's t-tests were used to compare mean antibody levels; p-values <0.05 were considered significant.


Patient characteristics and DSA levels

Ninety consecutive patients who had liver transplant were included in the study. Twenty of these had DSA (DSA+), 16 of who had positive cross-match at the time of transplant (10 with T- and B-cell, 5 with isolated B-cell and 1 with isolated T-cell cross-match) (Figure 1). Four patients with the highest T- and B-cell flow cross-match levels had positive T-cell AHG-CDC. There were 12 recipients with DSA against class I HLA, 5 against class II and 3 against both (Table 1). Not surprisingly, a greater fraction of the DSA+ recipients were female, undergoing second transplants and had history of blood transfusions (Table 2). There was no difference in the age, model for end-stage liver disease (MELD) score, indication for transplant or the donor characteristics between the two groups. Standard immunosuppression was administered to all recipients and no alterations were made for the DSA+ group. On day 7, the DSA levels decreased in 90% of patients (18/20) (Figure 2). Overall, only 3 patients had circulating cumulative DSA of above 20 000 MFI on day 7. At 4 months, only one patient (5%) had persistently elevated levels of DSA, whereas no antibody was detectable in 85% (17/20) of the group. Two patients (10%) had decreased but detectable levels of DSA.

Figure 1.

Preformed DSA and B-cell cross-match levels of individual liver transplant recipients. (A) The cumulative preformed DSA levels (MFI) of all, and (B) B-cell (left panel) and T-cell (right panel) flow cross-match levels (MFI) of DSA+ patients. *Patients with positive T-cell AHG-CDC.

Table 1.  DSA levels of presensitized recipients
Patients Day 0Day 74 months
  1. 1MFI < 2000.

  2. 2These DSA were not detectable at 1 year posttransplant.

 1A11 70251
 B810 498
 B44 8969
 2B716 229
 B6218 593 
 3DR7 2958
 4B55 3336
 5DQ7 4534 2164
 DQ910 315 7855
 6A2 2890
 7A24 5415
 8B60 9304
 9B27 7955
10A213 839
 A24 4666
11DQ618 52512 619
12DQ715 443
13A1 5195 5105
 B44 2162
 DR8 304110 5542104
 DQ4 266910 5145758
14DQ410 417 3994
 DQ9 8970 2436
15B61 2417
16A215 718 4019
 B819 544 5703
 B6010 232
 DQ2 6699
17A1 7666
 B35 3381
18A11 927882362
 A2 8001
 B7 3051
 B27 4978
 DR17 336656622
 DQ8 55012
19A3 4515
 B7 6483 265824592
 C713 259 3781
 B35 3435
 DR1114 111 7686
 DRW52 4027 6537
 DQ7 2094 5191
20B44 3026
Table 2.  Patient and donor characteristics of the two groups
 DSA negativeDSA positivep-Value
N70 (78%)20 (22%) 
Age (years)52.2 ±15.2 (1–71)47.9 ± 13.6 (2–65)0.26
Gender (F:M)28:4216:40.002
Race (Caucasian:other)66:419:11
Retransplant (n)550.04
Preoperative transfusion (n)29140.04
MELD (range)20.4 (6–44)18.7 (6–37)0.45
 HCC, underlying cirrhosis1020.73
 DRI (range)1.48 ± 0.39 (0.86–2.44)1.61 ± 0.44 (1.08–2.44)0.21
 CIT, mins (range)328 ± 108 (40–535)290 ± 124 (64–467)0.19
 WIT, mins (range)45±11 (18–75)44 ± 9 (26–65)0.86
Follow-up, days (range)456 ± 86 (294–584)440 ±93 (307–569)0.47
Figure 2.

DSA levels from baseline to 4 months after liver transplantation in DSA+ patients. Each patient is represented by a line, DSA levels are obtained prior to transplant (d0), on day 7 (d7) and at 4 months (4 m).


All recipients had day 7 protocol biopsies. These were evaluated by routine H&E, as well as immunohistochemistry for C4d. Twenty-eight percent of all patients (31/90) had acute cellular rejection, and this was not significantly different in DSA+ (9/20) versus DSA– (22/70) (p = 0.19) patients. No C4d staining was found in control liver specimens (Figure 3A). Portal subendothelial and stromal C4d deposition was seen rarely in DSA– patients (5/70), accompanied by acute cellular rejection in each case. In DSA+ patients, C4d staining in these locations was more common (5/20, p = 0.04). Similarly, central venular and sinusoidal C4d deposition was significantly more common in DSA+ patients (1/70 vs. 4/20, p = 0.008). Strikingly, diffuse Cd4 staining (defined as present in all four locations) was observed only in three patients with acute cellular rejection (Figures 3C–F), whose total MFI was over 20 000 at the time of the biopsy. Diffuse staining was not seen in any DSA– patients.

Figure 3.

C4d immunohistochemistry on biopsies obtained on day 7. (A) No C4d deposition is seen in normal liver (control). (B) Low-power magnification of liver allograft in a DSA+ patient, with diffuse C4d staining. Only patients with DSA levels above 20 000 (MFI) had diffuse C4d deposition along the (C) portal subendothelial space, (D) portal stromal, (E) central vein and (F) sinusoidal space.

Clinical outcomes

All recipients were followed for at least 1 year after transplant. Two patients in each group died during this period, with well-functioning allografts (Table 3). The rate of graft loss was not different in the DSA+ group. There was no difference in the serum bilirubin and transaminase levels between DSA– and DSA+ groups on day 7, at 4 months and 1 year, with the exception of serum alanine aminotransferase (ALT) levels on day 7. The three patients with acute cellular rejection and high DSA levels received high-dose steroid with subsequent normalization of serum bilirubin and transaminase levels. Since their liver function tests remained normal until their 1 year follow-up, no further liver biopsies were obtained.

Table 3.  Clinical outcomes, liver functional and structural tests up to 1 year after liver transplantation
 DSA negativeDSA positivep-Value
Death, n220.21
Graft loss, n410.69
Bilirubin, mg/dL (range)
 Day 09.4 (0.3–53.6)8.7 (0.3–4.04)0.80
 Day 73.3 (0.5–19.5)4.6 (0.4–19.4)0.25
 4 Months2.2 (0.1–44.8)0.6 (0.2–1.1)0.33
 1 Year0.8 (0.1–4.3)0.6 (0.1–1.2)0.13
AST, u/L (range)
 Day 098.9 (28–614)82.1 (34–294)0.51
 Day 750.3 (16–226)57.2 (21–129)0.46
 4 Months60.2 (9–655)35.9 (16–82)0.32
 1 Year47.5 (13–458)34 (13–88)0.4
ALT, u/L (range)
 Day 083.9 (8–1596)62.3 (22–246)0.62
 Day 7142.4  (17–471)200.8  (43–581)0.02
 4 Months65.9 (8–587)45.4 (13–161)0.43
 1 Year44.6 (6–293)45.4 (14–143)0.94


Over the past two decades, the outcomes in liver transplant have improved significantly, in part due to better understanding of cellular alloimmunity and development of new immune surveillance techniques to detect and intervene timely with cellular rejection. In contrast, the potential impact of preformed HLA DSA on the liver allograft remains controversial and the practice of transplanting patients with positive cross-match has not changed. In the past 2 years, several groups have provided convincing evidence that DSA could be linked to poor liver allograft survival (10,12), and may be associated with chronic, ductopenic rejection (14). Due to the retrospective nature of these studies, however, there is the inevitable selection bias toward recipients with the worst outcomes. While there is statistical significance, not all presensitized patients have evidence of rejection, and similarly not all patients with chronic, ductopenic rejection have high circulating DSA levels.

This study is the first systematic, prospective analysis of the prevalence and impact of preformed HLA DSA on liver allograft function. Our results show that preformed DSA is common (found in 22.2% of our recipients). This incidence is consistent with the previous reports of cross-match studies in liver transplant recipients, ranging from 5.2% to 24.2% (9–11,15). In this study, the DSA levels dropped immediately postoperatively in majority of these recipients, and became undetectable by 4 months. This decrease has been attributed to absorption of circulating antibodies by the liver. Although a definitive mechanism of action is not known, several have been proposed: secretion of HLA antigens (16), phagocytosis by resident Kupffer cells (17) or simply dilution owing to large vascular bed (18). In fact, there is clinical evidence that the liver allograft, when transplanted simultaneously with kidney in presensitized recipients improves the outcomes of the latter, demonstrated by a reduction in the circulating DSA (19). Dar et al. recently reported that in six presensitized patients undergoing combined liver and kidney transplants, the HLA class I DSA were preferentially cleared from the circulation, while class II DSA persisted (20). This is in contrast to our findings, where persistence of both HLA class I and II DSA was rare. This difference may be due to the fact that all the patients reported by Dar et al. had combined liver–kidney transplants and their average follow-up was shorter than that of the present group. In the present cohort, one patient had anti-HLA class I, one had anti-HLA class II and one had both at 4 months. Two of these three patients (patient #18 and 19) had no detectable DSA at 1 year. The DSA in the third patient (patient #13) was not assessed at 1 year.

We have also found that diffuse deposition of C4d was seen only in recipients with circulating cumulative DSA levels higher than 20 000 MFI. C4d is the enzymatic cleavage product of the complement cascade, resulting from classical pathway of activation. In the kidney, its deposition in peritubular capillaries is accepted as an important marker for antibody-mediated rejection. It is also a sensitive surrogate for circulating DSA (21,22) in kidney transplant recipients. In liver transplant, the evidence for C4d deposition as a marker for antibody-mediated injury is not as strong and thus has not yet been established, however there is increasing interest in its utilization (10,14,23). As protocol biopsies are routinely paraffin-embedded in our institution, we elected to perform the C4d immunostaining on these, although the sensitivity of this method may be lower than C4d staining of fresh-frozen tissue. In order to optimize our staining, as control we used 20 paraffin-embedded normal, nontransplant liver biopsies, which did not demonstrate any C4d deposition. Among our transplant recipients, although C4d deposits were observed occasionally along the stroma and subendothelial space of those with no DSA, sinusoidal and venular deposition was unique to DSA+ patients. These findings are in concordance with the previously published reports (14,23–25), and suggest a more specific staining pattern in those with high DSA.

In our series, all three cases where diffuse C4d staining was observed had accompanying acute cellular rejection. In contrast, not all recipients with acute cellular rejection had C4d deposition. Without a definitive temporal relationship, it is not clear whether one causes the other, or whether they are completely different immunological phenomena. However, C4d expression has been reported to be significantly increased in acute rejection along the endothelial cells of portal veins and hepatic arteries (but not along the hepatic veins and sinusoids) (26), due presumably to upregulated HLA class I and II expression in the cellular infiltrate and the endothelial cells. Therefore, we speculate that in patients with high circulating DSA, the complement activation acts as chemotactic to cellular components of the immune system. This hypothesis also explains the clinical improvement after steroid bolus in these patients.

Despite the significant C4d staining on day 7 and accompanying circulating DSA, we found that the clinical outcomes were not inferior in DSA+ recipients, compared to DSA– ones, during the first year after transplantation. Specifically, we did not observe any increased incidence of cholestatic failure or ductopenia in this cohort. Because of the normal liver functional and structural tests at 4 months and 1 year, none of our DSA+ patients, including the three with diffuse C4d staining on day 7, underwent additional biopsies. Therefore, it is possible that ongoing C4d deposition in some cases were not identified. It is also possible that the allograft injury caused by the DSA may take longer than 1 year. This appears to be the case in kidney transplant, as chronic antibody-mediated rejection occurs slowly in kidney allografts, with average time of over 2 years from detection of DSA to graft failure (18,27), indicating a cumulative pathology. The difference, however, between the kidney and liver is that in the former, the DSA levels persist (28), most likely contributing to this cumulative injury. In contrast, our data show that the persistence of DSA is much less common after liver transplant. Nevertheless, given the possible long-term impact of the DSA, we feel that continued monitoring of the antibody levels is warranted in these patients.

In summary, we conclude that preformed HLA DSA is common in liver transplant recipients at baseline, and even in the rare instances in which high levels of DSA persist and lead to complement activation in the liver, they are well tolerated at least in the first year after transplantation. Given the difficulty of finding a liver allograft against whom the recipient has no DSA and the real possibility of increased wait-list mortality associated with turning down organ offers in patients with high MELD scores, we suggest that these data support our current practice of liver transplantation despite the presence of DSA. However, longer follow-up is needed in this cohort to determine the incidence of chronic antibody-mediated damage and whether or not specific therapy is needed to prevent late allograft loss.


The authors of this manuscript have no conflict of interest to disclose as described by the American Journal of Transplantation.