Telephone: 214-820-2050; FAX: 214-818-6491
Implications of a positive crossmatch in liver transplantation: A 20-year review
Article first published online: 29 MAR 2012
Copyright © 2012 American Association for the Study of Liver Diseases
Volume 18, Issue 4, pages 455–460, April 2012
How to Cite
Ruiz, R., Tomiyama, K., Campsen, J., Goldstein, R. M., Levy, M. F., McKenna, G. J., Onaca, N., Susskind, B., Tillery, G. W. and Klintmalm, G. B. (2012), Implications of a positive crossmatch in liver transplantation: A 20-year review. Liver Transpl, 18: 455–460. doi: 10.1002/lt.22474
- Issue published online: 29 MAR 2012
- Article first published online: 29 MAR 2012
- Accepted manuscript online: 5 DEC 2011 02:51AM EST
- Manuscript Accepted: 10 NOV 2011
- Manuscript Received: 3 JUN 2011
Whether a positive crossmatch result has any relevance to liver transplantation (LT) outcomes remains controversial. We assessed the impact of a positive crossmatch result on patient and graft survival and posttransplant complications. During a 20-year period, 2723 LT procedures with crossmatch results were identified: 2479 primary transplants and 244 retransplants. The rates of positive B cell and T cell crossmatches were 10.1% and 7.4%, respectively, for primary transplants and 14.6% and 6.4%, respectively, for retransplants (P = 0.049 for a B cell crossmatch). Across all primary transplants, females (P < 0.001) and patients with autoimmune hepatitis (P < 0.001) had greater frequencies of positive crossmatches. There was no effect from race or age. For both primary transplants and retransplants, patient survival and graft survival were not affected by the presence of a positive crossmatch. With respect to posttransplant complications, there were no differences in rejection episodes (hyperacute, acute, or chronic) or technical complications (biliary and vascular) between negative and positive crossmatch groups. However, there were significant differences in the pathological findings of preservation injury (PI) on liver biopsy samples taken at the time of transplantation and within the first week of transplantation (P = 0.003 for B cells and P = 0.03 for T cells). In summary, a positive crossmatch had no significant impact on patient survival or graft outcomes. However, there was a significantly higher incidence of PI in primary LT recipients with a positive crossmatch. This finding is important for a broader understanding of PI, which may include a significant immunological component. Liver Transpl 18:455–460, 2012. © 2011 AASLD.
Initial experiences with kidney transplantation showed that transplants with ABO-incompatible blood types resulted in immediate graft failure.1-3 Thus, the crossmatch test was developed to determine which donor-recipient combinations were incompatible and thus minimize the risk of hyperacute rejection. In liver transplantation (LT), early observations revealed that liver allografts were resistant to hyperacute antibody-mediated rejection, and this led to the question whether a crossmatch result was relevant to or required for hepatic transplantation.4, 5 Since then, the authors of numerous retrospective studies have attempted to determine the significance of crossmatching, with most suggesting inferior graft survival with a donor-specific positive T cell crossmatch.6-10 Conversely, other analyses have reported no differences in outcomes.11, 12
With respect to graft complications, the Mayo group in 1988 observed a greater frequency of vanishing bile duct syndrome in positive crossmatch recipients.13 Approximately a decade ago, the Pittsburgh group reported an increased incidence of biliary complications and chronic rejection in recipients with a positive T cell crossmatch.14 Finally, in a recent report by Opelz et al.,15 a positive T cell crossmatch in primary LT recipients was found to be significantly associated with impaired graft survival. At our high-volume transplant center, we sought to determine whether a positive crossmatch had any consequences.
PATIENTS AND METHODS
We reviewed all adult patients with available crossmatch results who underwent LT at our institution between October 1987 and December 2007. Patient data were obtained from our prospectively maintained research database with appropriate institutional review board approval. Crossmatch testing of donors' isolated T and B lymphocytes and recipients' sera was performed with different methods during the 20-year review period. For the purposes of this study, crossmatch results obtained with the anti-human globulin technique (for T cells) and the standard National Institutes of Health technique (for B cells) are reported. Crossmatch data were not used prospectively and were generally made available after transplantation.
Liver biopsy was performed routinely after graft reperfusion, and the presence of preservation injury (PI) was coded (if it was applicable). After transplantation, biopsy was performed as clinically indicated. PI was defined by the following 4 criteria: (1) cytoplasmic eosinophilic degeneration or apoptosis of hepatocytes, (2) nuclear pyknosis, (3) perihepatocyte collections of lymphocytes and neutrophils, and (4) feathery ballooning of cytoplasm.16
Categorical data were analyzed with Fisher's exact test (2-tailed) for 2 × 2 tables and with the likelihood ratio chi-square test for larger tables. Continuous variables were compared with the 2-sample Wilcoxon test. Survival curves were estimated with the Kaplan-Meier method and were compared with the log-rank test. A P value < 0.05 was considered statistically significant. All analyses were performed with SAS 9.1.3.
During the 20-year study period, we identified 2723 LT recipients with reported crossmatch results: 2479 underwent primary LT, and 244 underwent retransplantation. For primary transplants, the rates of positive B cell and T cell crossmatches were 10.1% and 7.4%, respectively. For retransplants, the rates were 14.6% (P = 0.049) and 6.4% (P = 0.87), respectively.
In primary graft recipients, a positive crossmatch result was more likely in females and in recipients with a diagnosis of autoimmune hepatitis (AIH; Table 1). The effect of female sex was lost with retransplantation. Neither age nor African American race was a factor.
|Positive T Cell Crossmatch (n = 135)|
|Primary Grafts (n = 126)||P Value||Retransplants (n = 9)||P Value|
|Sex: female [n (%)]||107 (85)||<0.001||5 (56)||0.29|
|Race: African American versus non–African American [n (%)]||7 (5.6)||0.39||2 (22.2)||0.62|
|Diagnosis: AIH versus non-AIH [n (%)]||13 (10.3)||<0.001||0||>0.99|
|Positive B Cell Crossmatch (n = 273)|
|Primary Grafts (n = 241)||P Value||Retransplants (n = 32)||P Value|
|Sex: female [n (%)]||176 (73)||<0.001||15 (47)||0.33|
|Race: African American versus non–African American [n (%)]||17 (7.1)||0.80||4 (12.5)||0.75|
|Diagnosis: AIH versus non-AIH [n (%)]||21 (8.7)||<0.001||1 (3.2)||>0.99|
As shown in Fig. 1, neither patient survival nor graft survival was adversely affected by the presence of a positive crossmatch in primary LT recipients. The same result was found for patients who underwent retransplantation (data not shown). To determine whether any differences existed by the period of transplantation, we divided our study into 4 time periods: 1987-1992 (era 1), 1993-1997 (era 2), 1998-2002 (era 3), and 2003-2007 (era 4). As expected, there were no differences in patient survival between the time periods for primary graft or retransplant recipients. However, decreased graft survival was noted for primary graft recipients with a positive T cell crossmatch in era 3 (P = 0.009).
As for posttransplant complications, rejection episodes (hyperacute, acute, and chronic) and technical complications were not more frequent in the positive crossmatch cohort (Table 2). However, PI was more commonly identified by postreperfusion liver biopsy and by liver biopsy performed within the first week of transplantation in primary graft recipients. This finding was not duplicated for retransplant patients.
|Positive B Cell Crossmatch||Positive T Cell Crossmatch||Positive B Cell Crossmatch||Positive T Cell Crossmatch|
|Hyperacute rejection||>0.99||>0.99||Not applicable#||Not applicable#|
|Acute rejection (<1 year)||0.31||0.41||>0.99||>0.99|
|Chronic rejection (ever)||0.60||0.51||0.75||>0.99|
|PI (≤1 week)||0.003||0.03||0.54||0.25|
|Vascular complications (ever)*||0.41||0.91||0.84||0.74|
|Biliary complications (3-12 months)†||0.50||0.60||0.75||0.60|
|Graft-versus-host disease (ever)||>0.99||0.62||1.0||Not applicable#|
When we compared the results of liver function tests shortly after transplantation, we found significant elevations in the alkaline phosphatase and total bilirubin levels of the positive crossmatch cohort on day 7 (Table 3). There were no differences in the transaminase levels on posttransplant days 1, 3, and 7.
|B Cell||T Cell|
|Positive Crossmatch||Negative Crossmatch||P Value||Positive Crossmatch||Negative Crossmatch||P Value|
|Posttransplant day 1|
|Aspartate aminotransferase (U/L)||546||595||0.63||652||611||0.28|
|Alanine aminotransferase (U/L)||463||495||0.33||494||500||0.71|
|Alkaline phosphatase (U/L)||77||74||0.33||77||74||0.62|
|Gamma-glutamyl transpeptidase (U/L)||64||63||0.61||63||63||0.96|
|Total bilirubin (mg/dL)||3.4||3.2||0.17||3.5||3.2||0.81|
|Posttransplant day 3|
|Aspartate aminotransferase (U/L)||149||174||0.26||173||163||0.66|
|Alanine aminotransferase (U/L)||279||313||0.31||314||305||0.36|
|Alkaline phosphatase (U/L)||100||94||0.26||107||96||0.06|
|Gamma-glutamyl transpeptidase (U/L)||106||118||0.39||124||121||0.50|
|Total bilirubin (mg/dL)||3.1||2.6||0.31||2.8||2.6||0.68|
|Posttransplant day 7|
|Aspartate aminotransferase (U/L)||41||44||0.27||38||41||0.26|
|Alanine aminotransferase (U/L)||119||131||0.12||128||131||0.91|
|Alkaline phosphatase (U/L)||153||138||0.05||170||142||0.002|
|Gamma-glutamyl transpeptidase (U/L)||263||243||0.15||273||229||0.001|
|Total bilirubin (mg/dL)||3.1||2.5||0.002||2.6||2.1||0.04|
Contradicting the findings in the majority of reports on the subject of crossmatching and LT,6-10, 17, 18 we found no difference in overall patient or graft survival between positive and negative crossmatch recipients. We were also unable to duplicate the findings of Takaya et al.,14 who noted increased biliary complications and chronic graft rejection in positive crossmatch recipients. In fact, we did not find any statistical associations between positive crossmatch cases and graft complications.
We closely examined the primary graft failures in era 3, in which a positive T cell crossmatch result positively correlated with decreased graft survival. There were no differences in acute rejection episodes, primary nonfunction, chronic rejection, or technical complications between the positive and negative T cell crossmatch groups. Furthermore, neither the use of induction agents nor the type of immunosuppression positively correlated with graft failure. Three of the 12 graft failures (25%) in era 3 occurred within a month of the start or end of the designated time period. The randomization of these eras may very well have created the statistical significance seen.
Our rate of positive T cell crossmatches (7.4%) is similar to the rates in other reports, which have ranged from 5% to 12% with various lymphocytotoxic techniques. Interestingly, there was a higher incidence of positive B cell crossmatches with retransplants. This may be a result of more class I antibodies being absorbed by the primary graft during the initial transplant, which allowed greater class II reactivity during retransplantation. No other differences were found between positive T cell and B cell crossmatches.
Who is more likely to have a positive crossmatch? We contemplated whether female sex, African American race, or an autoimmune diagnosis could be a risk factor. Patients with autoimmune disease have a greater number of circulating antibodies, rendering them more immunogenic. We found this to be true with respect to crossmatching. In kidney transplantation, African Americans are presumed to have a greater propensity for graft rejection, thus requiring more intense immunosuppression.19-21 However, and a positive crossmatch did not find an association between African American race in LT. Finally, we found that female patients were more likely to have a positive crossmatch. We can speculate that this may be due to previous pregnancies, which made the female patients more sensitized. However, this disadvantage was lost with retransplantation perhaps because male patients became more sensitized with exposure to the initial graft.
Our results suggest an immunological cause—a positive crossmatch—as a pertinent risk factor for PI. What is called PI is a multifactorial process that can cause various degrees of graft dysfunction (from mild, self-limiting transaminitis to primary nonfunction). Poor outcomes for grafts with PI have been previously described; these include increased biliary complications,22 increased acute rejection,23 and decreased graft survival24 and patient survival.25 Although PI is principally dependent on graft warm and cold ischemia, it is also greatly affected by the quality of the donor graft, recipient factors, and intraoperative (donor and recipient) events. According to Table 2, PI was not found to be a factor in retransplant cases with a positive crossmatch. It can be surmised that the use of younger, more ideal grafts in the retransplant setting decreased the probability of PI, so a positive crossmatch was no longer a factor.
During the study period, University of Wisconsin solution was principally used as our preservation medium (Euro-Collins solution was used in only 39 recipients). Thus, a reliable comparative analysis could not be performed to determine whether the type of preservation solution affected the occurrence of PI.
Our cases of PI (identified pathologically) did not present immediately after transplantation with elevated aminotransferase levels, which have been previously used to describe its various degrees of severity.24, 25 Our results suggest that elevated canalicular enzymes on posttransplant day 7 may be a better marker of PI.
The limitations of this analysis are few but are worth mentioning. This was a retrospective study looking at the results of our positive crossmatch cases, although our data were prospectively collected. Additionally, our liver biopsy samples were not coded for PI severity (mild, moderate, or severe), and this may have strengthened (or weakened) our results. This study crossed both the cyclosporine and tacrolimus eras of immunosuppression; however, we did not find any differences in acute rejection in the positive crossmatch cohort. Finally, it is plausible that relevant correlations resulting from a positive crossmatch (eg, chronic rejection and primary nonfunction) were hidden in this large cohort study, and a more detailed filtered analysis may be required.
At present, LT is performed despite crossmatch positivity. Because of the urgency of these transplants and the limits of graft cold ischemia, it is unrealistic to postpone an operation until a negative crossmatch is found. In this review, we have been unable to find a significant reason to defer transplantation even if a positive crossmatch is present before transplantation. Nonetheless, the identification of cases with a positive crossmatch may help teams to better prepare for the finding of PI after transplantation. Alternatively, if a positive crossmatch donor is identified before transplantation, modifications in procurement (eg, limited cold ischemia time and novel agents for ischemia/reperfusion prevention) and in posttransplant approaches (eg, increased immunosuppression, induction with cytotoxic antibodies, and prostaglandin use) can be prearranged. Further investigations in which donor-specific antibodies are identified in these cases may substantiate our results and favor prospective crossmatching for LT.
The authors are grateful to Linda Jennings, Ph.D., for her expert statistical analysis and valued feedback throughout the study period.