Pediatric Kidney Transplantation: Analysis of Donor Age, HLA Match, and Posttransplant Non-Hodgkin Lymphoma: A Collaborative Transplant Study Report

Abstract

Background. 

The impact and relationship of donor age, human leukocyte antigen (HLA) matching, and posttransplant non-Hodgkin lymphoma in pediatric kidney recipients are not completely understood.

Methods. 

We analyzed Collaborative Transplant Study data from 9209 pediatric kidney transplant recipients to examine the effects of donor age and HLA match on graft survival and the relationship between HLA match and occurrence of non-Hodgkin lymphoma.

Results. 

Survival rates using donors aged 11 to 17, 18 to 34, or 35 to 49 years were similar. Cox regression analysis showed that two HLA-DR mismatches were associated with lower graft survival in transplants performed during 1988 to 1997 (P<0.001) but not during the 1998 to 2007 period (P=0.95). A hierarchical relationship was observed for the effect of increasing numbers of combined HLA-A+B+DR mismatches on graft survival during the 1988 to 1997 (P<0.001) and the 1998 to 2007 period (P<0.001). An association between two HLA-DR mismatches and non-Hodgkin lymphoma was demonstrated by multivariate analysis (hazard ratio for 2 vs. 0–1 DR mismatches 2.04, P=0.021), and the result was consistent during both 10-year periods.

Conclusion. 

We recommend that (1) kidneys from deceased donors up to 49 years be allocated to children, (2) an acceptable HLA-A+B+DR match be attempted in patients with relatively common HLA phenotypes, and (3) transplants with two HLA-DR mismatches be avoided to reduce the risk of posttransplant non-Hodgkin lymphoma.

Kidney transplantation offers particular advantages for children with end-stage renal disease, because if children are transplanted early after the diagnosis of end-stage kidney failure, body growth improves after transplantation compared with chronic dialysis (^{1, 2}). To minimize the duration of pretransplant dialysis, donor kidneys from deceased organ donors are preferentially allocated to children. The aim is, of course, to transplant children with kidneys that will function for as long as possible. Thus, in the United States, for example, kidneys from donors younger than 35 years of age are prioritized for pediatric recipients in the expectation that organs from these relatively young donors will survive well. Although no strict age specifications have been established by Eurotransplant, a loose policy of age matching is applied routinely. In addition, the best possible human leukocyte antigen (HLA) match is selected by Eurotransplant because long-term graft survival is believed to benefit from improved histocompatibility (³). In contrast, in the United States, with the aim to eliminate inequalities in organ distribution to racial minorities, matching for HLA-A and -B has been abandoned, and only HLA-DR is considered during the kidney allocation process (^{4, 5}). Indeed, a recent analysis showed no advantage for HLA-DR matching in children, and a recommendation was made to disregard the HLA match entirely when allocating donor kidneys to pediatric recipients (⁴).

We used the database of the international Collaborative Transplant Study (CTS) to examine the effects of donor age and HLA match on graft survival after kidney transplantation in children. Furthermore, we addressed the question of whether recent findings showing an association between HLA incompatibility and posttransplant non-Hodgkin lymphoma is also observed in children, for whom non-Hodgkin lymphoma is a more frequent occurrence.

MATERIALS AND METHODS

The analysis was based on the data of the international CTS (⁶). Nine thousand two hundred nine kidney transplants from deceased donors performed during the 20-year period from 1988 to 2007 were analyzed. Subanalysis was carried out for transplants undertaken during the two 10-year periods from 1988 to 1997 and 1998 to 2007. All recipients were aged 0–17 years at the time of transplantation. Recipients of multiorgan grafts, including combined kidney and pancreas, were excluded. HLA typing for the HLA-A, -B, and -DR loci was performed at laboratories of the participating centers and reported to the CTS study center shortly after transplantation when the transplant was registered for the study. Clinical data were recorded at 6, 9, and 12 months and annually thereafter. Centers were included in the analysis of non-Hodgkin lymphoma only if the completeness and accuracy of cancer data were confirmed in writing; 7802 patients were available for lymphoma analysis. Information on pretransplant Epstein-Barr virus (EBV) serostatus was available on 705 (9%) of these recipients. Graft survival and cumulative posttransplant lymphoma incidence rates were computed using the Kaplan-Meier method. Mantel Cox log-rank test was used for trend analysis of HLA mismatches and graft survival and for the incidence of non-Hodgkin lymphoma. Cox multivariate regression analysis was performed to account for the possible influence of the following confounding factors: geographical origin (continent), year of transplant, transplant number (first or retransplant), recipient age, donor age stratified for age groups, gender and race, preformed panel reactive antibodies, cold ischemia time, time on dialysis before transplantation, general evaluation as a candidate for transplantation at the time of transplant, and immunosuppressive medication (type of calcineurin inhibitor, type of antimetabolite, type of antibody induction agent, and use of steroids). A back-step elimination algorithm was used to exclude confounding factors with a threshold of P more than 0.1. Non-Hodgkin lymphomas occurring up to the time of graft loss, but not after removal of a failed graft, were included in the analysis of lymphomas. We restricted the analysis to non-Hodgkin lymphomas and excluded other conditions usually summarized under the loose definition posttransplant lymphoproliferative disease, such as infectious mononucleosis, monoclonal gammopathies, or benign lymphatic hyperplasias (⁷). Classical Hodgkin lymphomas and multiple myelomas were also excluded. Hazard ratios (HRs) were calculated using the software package SPSS version 16.0 (SPSS Inc, Chicago, IL). P values less than 0.05 were considered significant.

RESULTS

A total of 9209 transplants were included in the analysis. Demographic and baseline data are summarized in Table 1. All characteristics that were not evenly distributed between the transplants with 0, 1, or 2 HLA-DR mismatches (i.e., those showing a statistically significant P value between groups) were included as potential confounders in Cox multivariate analyses.

TABLE 1: Demographic and baseline characteristics of pediatric kidney transplants according to HLA-DR mismatch

The analysis of graft survival according to donor age for transplants performed during 1988 to 1997 or 1998 to 2007 showed (a) the expected improvement in graft survival for transplants performed in the more recent time period regardless of donor age and (b) surprisingly good graft survival among transplants from donors aged 35 to 49 years during the 1998 to 2007 period (Fig. 1A and B). Statistically, the survival rate of grafts from 35- to 49-year-old donors did not differ from that of donors aged 11 to 17 or 18 to 34 years. Transplants from donors less than or equal to 10 years or more than 49 years, however, showed inferior outcomes. It was notable that the proportion of donors aged 10 years or younger decreased from 34.3% of donors during 1988 to 1997 to 18.7% during 1998 to 2007 (P<0.001). This was predominantly because of a decline in the number of very young donors: from the first to the second decade, the proportion of donors aged 0–5 years decreased from 16.4% to 7.5% and 6- to 10-year-old donors decreased from 17.9% to 11.3%, whereas 11- to 17-year-old donors increased from 19.9% to 24.0%.

FIGURE 1.: Graft survival of deceased-donor pediatric kidney transplants according to donor age. Donor age categories and numbers of transplants studied are indicated. Five-year graft survival is shown for transplants performed during 1988 to 1997 (A) and 1998 to 2007 (B). Graft survival for 5 to 10 years after transplantation is shown for transplants performed during 1988 to 1997 (C), with curves starting at 100% at year 5.

The second donor category of special interest, the 35- to 49-year-old donors increased from 15.4% in 1988 to 1997 to 22.5% in 1998 to 2007 (P<0.001), indicating that centers participating in the CTS generally did not follow the recommendation of United States to avoid donors older than 35 years (⁴). In an attempt to identify possible disadvantages associated with use of donors aged 35 to 49 years, we examined 1-year serum creatinine levels as a qualitative indicator of graft function. Among patients transplanted during 1998 to 2007, the incidence of 1-year serum creatinine less than 130 μmol/L was lower among recipients who received a graft from a donor aged 35 to 49 years (67.6%) compared with those who received a graft from a donor aged 18 to 34 years (77.7%). Conversely, creatinine values more than 260 μmol/L were more frequently observed among patients who received a kidney from 35- to 49-year-old donors (1.9%) compared with 18- to 34-year-old donors (1.1%). Because this was of concern, we examined whether the long-term survival of grafts from donors aged 35 to 49 years was likely to deteriorate more rapidly than that of grafts from donors aged 18 to 34 years by assessing 5- to 10-year graft survival rates for transplants performed from 1988 to 1997. The survival curves did not indicate that the rate of graft loss accelerated over time in the older donor group (Fig. 1C). Accordingly, all transplants involving grafts from donors up to the age of 49 years were included in subsequent analysis steps.

Analysis of HLA-DR matching in relation to graft survival showed that transplants with 0 or 1 HLA-DR mismatch had virtually identical survival rates during 1988 to 1997 (5-year graft survival 64.2% vs. 64.4%) and 1998 to 2007 (79.6% vs. 78.3%). Grafts with two HLA-DR mismatches fared less well (Fig. 2). However, because matching for HLA-DR is not independent of matching for HLA-A and HLA-B (organ exchange organizations attempt to match as best as possible for all three HLA loci), Cox regression analysis including all three loci is required for estimating the true influence of HLA-DR. The Cox result, considering the potential confounders listed under Methods section, showed that two HLA-DR mismatches exerted a significant impact on survival rates for transplants performed during 1988 to 1997 (HR=1.27, P<0.001) but not anymore during the 1998 to 2007 period (HR=1.01, P=0.95). In previous analyses, we found that the long-term effect of HLA matching is most appropriately assessed by considering mismatches for HLA-A, -B, and -DR together (^{8, 9}). Therefore, we examined the 5-year graft survival according to the number of HLA-A+B+DR mismatches. When “relatively good” HLA matches (defined as 0–3 HLA-A+B+DR mismatches) were compared with “relatively poor” matches (4–6 mismatches), an HR of 1.19 (95% confidence interval [CI] 1.08–1.32; P<0.001) was obtained for the 1988 to 1997 period and HR 1.26 (95% CI 1.06–1.50; P=0.009) for the period 1998 to 2007 (Fig. 3A and B). A hierarchical relationship was observed for the effect of increasing numbers of mismatches on graft survival (Fig. 3C), and this effect was statistically significant both for the 1988 to 1997 and 1998 to 2007 periods (Mantel Cox log rank trend analysis: P less than 0.001 and P less than 0.001, respectively).

FIGURE 2.: Influence of human leukocyte antigen-DR mismatch on graft survival of pediatric kidney transplants. Transplants from deceased donors younger than 50 years were analyzed. Statistical significance: log-rank P less than 0.001 for grafts performed (A) 1988 to 1997 and P=0.12 for grafts performed (B) 1998 to 2007.

FIGURE 3.: Impact of human leukocyte antigen (HLA)-A+B+DR mismatch on survival of deceased-donor pediatric kidney transplants performed during 1998 to 2007. Five-year graft survival is shown for transplants performed during 1988 to 1997 (A), 1998 to 2007 (B), and 1988 to 2007 (C). Graft survival was significantly associated with the number of HLA mismatches (Mantel Cox log-rank P<0.001 in all three analyses).

We considered whether recent findings regarding the occurrence of non-Hodgkin lymphoma in relation to HLA-DR mismatching may also apply to pediatric kidney transplantation (¹⁰). Analysis of the total study cohort (1988–2007) showed a strikingly higher rate of non-Hodgkin lymphoma in patients who received a graft with two HLA-DR mismatches (log-rank P=0.003, Fig. 4). Subanalysis of the consecutive decades 1988 to 1997 and 1998 to 2007 demonstrated that this effect was consistent over time (Fig. 5). Cox regression analysis confirmed the association between two HLA-DR mismatches and non-Hodgkin lymphoma in children (HR for 2 DR mismatches vs. 0–1 mismatches 2.04, 95% CI 1.11–3.74; P=0.021). Two mismatches at the HLA-A or -B loci were not associated with an increased rate of non-Hodgkin lymphoma (data not shown). In addition to two HLA-DR mismatches, very young recipient age 0 to 5 years (HR=2.09, 95% CI 1.13–3.88, P=0.019) and use of lymphocyte depleting antibodies (HR=3.22, 95% CI 1.87–5.55, P<0.001) were shown to be associated with increased incidence of lymphoma.

FIGURE 4.: Cumulative incidence of posttransplant non-Hodgkin lymphoma according to the number of HLA-DR mismatches. Pediatric transplants from deceased donors younger than 50 years were analyzed. Statistical significance: Mantel Cox log-rank P=0.003.

FIGURE 5.: Association of posttransplant non-Hodgkin lymphoma with human leukocyte antigen (HLA)-DR mismatches in two consecutive decades. The incidence of non-Hodgkin lymphoma was significantly higher with two HLA-DR mismatches during both the (A) 1988 to 1997 (log-rank P=0.043) and the (B) 1998 to 2007 periods (P=0.017).

Information on pretransplant EBV serostatus was available on a subgroup of 705 recipients, too few for conclusive analysis. In agreement with published data on lymphoma risk and EBV (¹¹), the 289 EBV-negative pediatric patients showed a 3-year cumulative incidence of non-Hodgkin lymphoma of 3.1%, when compared with an incidence of 1.2% in the 416 EBV-positive recipients, a result that did not reach statistical significance because of the relatively small patient population available for analysis (P=0.11). Importantly, the distribution of EBV-negative recipients was similar among patients receiving transplants with 0, 1, or 2 HLA-DR mismatches: 39%, 42%, and 42%, respectively, were EBV negative. Thus, there was no indication that the described association of non-Hodgkin lymphoma with two HLA-DR mismatches was related to a negative pretransplant EBV serostatus.

DISCUSSION

The current analysis was prompted by a recent report from the United States, which recommended that two HLA-DR mismatches should not be avoided when allocating donor grafts for pediatric kidney transplantation (⁴). We have consistently observed that HLA matching affects graft survival regardless of recipient age (^{8, 9}), and therefore, the report caused us concern in case we had overlooked some aspect of the association that was of particular relevance to children. Clarifying this issue is important because disregarding the HLA match could potentially shorten waiting times before transplantation, although avoiding two HLA-DR mismatches can be accomplished relatively simply and would only have a small effect on waiting time in a large exchange organization such as Eurotransplant. Extending the maximum donor age limit from 35 to 49 years would, in contrast, have a marked benefit for donor organ availability for children. The effect of HLA matching in pediatric kidney recipients has not been studied extensively. Other than two studies based on US registries that did not demonstrate a clear matching effect (^{4, 12}), we could identify only two single-center reports that showed conflicting results (^{13, 14}). The current analysis is based on the largest patient series studied so far.

Our results suggest that donors up to age of 49 years can be used for children without noticeable detriment to graft survival for up to 10 years posttransplant. Recipients of transplants from donors aged 35 to 49 years had somewhat less favorable 1-year serum creatinine values; however, graft survival was not affected up to 10 years of follow-up. We cannot exclude the possibility that graft survival beyond the tenth year may eventually become lower in the 35- to 49-year-old donor category; an answer to this question would require prolonged documentation of graft outcome data. Nevertheless, given the importance of providing children with functioning grafts during the critical period of growth and development, it would seem acceptable to allocate kidneys from 35- to 49-year-old donors for pediatric transplantation.

Our analysis of HLA matching did not show a statistically significant effect of HLA-DR mismatches on graft survival for transplants performed during 1998 to 2007. Thus, our data do not vary markedly from the reported US data, which were based on an analysis of transplants undertaken in 1996 to 2005 (⁴). The maximum HLA matching effect was seen in our population only when the HLA-A+B+DR loci were analyzed together (Fig. 3). A comparable analysis was not performed by Gritsch et al. (⁴). In practical terms, waiting for a good HLA-A+B+DR match may take far longer than simply avoiding two HLA-DR mismatches. If a patient possesses a rare HLA phenotype, waiting until a well-matched kidney becomes available may be unacceptable because even a poorly matched but well-functioning kidney can correct growth defects. The potential benefit of HLA matching must, therefore, be balanced against the potential disadvantage of prolonged waiting time. On the Eurotransplant waiting list, for example, children are given extra “points” for waiting time, so that excessively long waiting times are avoided (³). Patients with a relatively common HLA phenotype, who are reasonably likely to receive an appropriately matched kidney within an acceptable time period, will benefit from being placed on the waiting list of a large organ exchange organization, which increases the likelihood of finding a good HLA match (^15–17).

To address the question of whether two HLA-DR mismatches should be avoided, our view must be widened to include not only the graft survival rate but also the incidence of posttransplant non-Hodgkin lymphoma. In light of recent evidence indicating that two HLA-DR mismatches are associated with an increased rate of non-Hodgkin lymphoma (¹⁰), we considered it important to examine this issue in our pediatric cohort. Pretransplant negative Epstein-Barr virus (EBV) serostatus, which because of the effect of primary infection with EBV is strongly associated with posttransplant non-Hodgkin lymphoma, is much more common among children than adults (¹¹). If HLA-DR mismatches were an additional lymphoma-promoting factor in pediatric transplants, the high risk of lymphoma in children could be modified by avoiding HLA-DR mismatches. The documentation of a significant relationship of HLA-DR mismatches with non-Hodgkin lymphoma in pediatric patients is statistically demanding because of the relatively small number of pediatric patients available for analysis. However, the current results virtually replicate those obtained in our previous analysis of the entire transplant population (¹⁰). In the current analysis, children who received a kidney graft from a donor with two HLA-DR mismatches had an incidence of non-Hodgkin lymphoma that was approximately twice as high as for recipients with 0 to 1 DR mismatch. At least based on the limited data available from 705 patients, negative pretransplant EBV serostatus was evenly distributed among patients with 0, 1, or 2 DR mismatches. Our previous analysis, based on a far larger patient population, allowed us to identify an association between two HLA-DR mismatches and an increased need for both rejection treatment and higher doses of maintenance immunosuppression drugs, factors believed to promote the occurrence of non-Hodgkin lymphoma (¹⁰). Extrapolating from the results shown here in Figure 5, one could expect that the simple measure of avoiding kidney transplants from donors with two HLA-DR mismatches would reduce the rate of posttransplant lymphoma from two to one for every 100 children.

In summary, on the basis of the results of this analysis, we recommend that (1) kidneys from deceased donors up to age of 49 years be allocated to pediatric recipients, (2) an acceptable HLA-A+B+DR match be attempted in patients with relatively common HLA phenotypes, and (3) transplants with two HLA-DR mismatches be avoided to reduce the risk of posttransplant non-Hodgkin lymphoma.

ACKNOWLEDGMENTS

The authors thank the following transplant centers and staff for their support: Aachen, Aberdeen, Adelaide, Amsterdam, Ankara, Antwerp, Auckland, Augsburg, Barcelona, Bari, Basel, Beilinson, Belfast, Belo Horizonte, Berlin, Bern, Birmingham (GB), Bochum, Bogota, Bologna, Bonn, Bremen, Brescia, Brighton, Brisbane, Bristol, Brussels, Budapest, Buenos-Aires, Cagliari, Cairo, Cali, Cambridge, Cape Town, Cardiff, Chicago, Christchurch, Cincinnati, Cleveland, Cologne, Coventry, Dallas, Debrecen, Dublin, Duesseldorf, Dundee, Edinburgh, Edmonton, Erlangen-Nuernberg, Essen, Exeter, Florence, Frankfurt, Freiburg, Gdansk, Geneva, Genoa, Gent, Glasgow, Goettingen, Goteborg, Grand Rapids, Groningen, Halle, Hamilton (CDN), Hamilton (NZ), Hann-Muenden, Hannover, Heidelberg, Helsinki, Homburg, Hong Kong, Innsbruck, Istanbul, Izmir, Jena, Kaiserslautern, Kansas City, Karachi, Katowice, Kaunas, Kent, Kiel, La Plata, Lausanne, Leeds, Leicester, Leipzig, Leuven, Liege, Lille, Lima, Limoges, Linz, Liverpool, Ljubljana, London, Louisville, Luxemburg, Lyon, Maastricht, Maceio, Madrid, Malmo-Lund, Manchester, Manila, Mannheim, Mar del Plata, Marburg, Martin, Medellin, Melbourne, Milan, Modena, Monterrey, Moscow, Muenster, Munich, Nancy, Nantes, New Delhi, New Orleans, New York, Newcastle (Australia), Newcastle u Tyne, Nice, Nijmegen, Nottingham, Orlando, Oviedo, Oxford, Padua, Palermo, Pamplona, Panama, Paris, Pato Branco, Pecs, Perth, Phoenix, Plymouth, Poitiers, Portland, Porto Alegre, Portsmouth, Prague, Quebec, Regensburg, Reims, Rennes, Ribeirao Preto, Rijeka, Rio de Janeiro, Rome, Rosario, Rostock, Rotterdam, Santa Fe, Santander, Santiago, Sao Paulo, Seoul, Sheffield, St. Etienne, St. Gallen, Stanford, Stoke on Trent, Strasbourg, Stuttgart, Sydney, Szeged, Toronto, Turin, Ulm, Uppsala, Utrecht, Valdivia, Valencia, Valhalla, Vancouver, Varese, Verona, Vicenza, Vilnius, Warsaw, Wellington, Winnipeg, Wuerzburg, Zagreb, and Zurich.