Kidney transplantation is a lifesaving therapy for patients with end-stage kidney disease, but kidney transplant recipients (KTRs) can suffer from substantial morbidity due to prolonged use of immunosuppressive medications. The goal of immunosuppression therapy is to reduce rejection and thereby prolong graft survival.¹ The therapeutic regimen attempts to balance keeping rejection rates low while minimizing the impact of over-immunosuppression and drug toxicities.¹ However, complications of long-term immunosuppressive therapy including infections, cardiovascular disease, and cancer, are major issues.

Cancer is a common complication and major cause of death in transplant recipients.^2,3 KTRs have a 28% higher risk of any type of cancer compared to the general population.⁴ Specifically, KTRs have an elevated risk of kidney cancer, non-Hodgkin lymphoma, and melanoma,⁴ resulting in large part from decreased immune surveillance due to maintenance immunosuppression, effects of oncogenic viruses, and, for kidney cancer, the presence of native end-stage kidneys.

The most common maintenance immunosuppression for KTRs is based on the use of a calcineurin inhibitor (CNI), antiproliferative agents, and glucocorticoids. Although CNI-based therapy improves early graft survival, the nephrotoxic effect of CNIs result in chronic graft damage.⁵ Belatacept, a T-lymphocyte co-stimulation blocker, inhibits T-cell activation by preventing CD28 binding to B7 proteins on antigen presenting cells. In this way, belatacept is intended to replace CNIs and prevent rejection while avoiding CNI-associated toxicities.⁶

Belatacept was approved in 2011 for use in KTRs to prevent organ rejection. In clinical practice, belatacept is preferred over CNIs or mammalian target of rapamycin (mTOR) inhibitors for KTRs who have experienced or are at risk of toxicity from those agents, such as nephrotoxicity, recurrent infections, or neurological side effects. This preference extends to recipients who are at relatively low risk of rejection, such as older individuals, those well-matched by human leukocyte antigen, and recipients of living donor transplants. Although the efficacy of belatacept, including lower nephrotoxicity compared with cyclosporine (a CNI), is established,⁷ the overall benefits and risks compared to tacrolimus, the CNI in predominant use today and the cornerstone of KTR maintenance therapy, are less studied.⁸

In particular, uncertainty exists regarding the long-term effect of belatacept use on cancer incidence among KTRs.⁸ All medications that induce immunosuppression can lead to loss of immune control of oncogenic viral infections (e.g., Epstein–Barr virus [EBV]). Several randomized controlled trials reported higher incidence of serious viral infections, including cytomegalovirus and EBV, among belatacept users compared with CNI users.^9,10 EBV-naïve solid organ recipients face a higher risk of developing posttransplant lymphoproliferative disorder (PTLD) compared to those who are latently infected and EBV-seropositive at the time of transplantation, owing to an inadequate immune response to primary EBV infection after transplantation.¹¹ Belatacept is associated with increased risk of PTLD in KTRs who are EBV-seronegative at the time of transplantation.⁹ This association led to a black box warning that belatacept is contraindicated in KTRs who are EBV-seronegative or whose EBV serostatus is unknown before transplantation.⁶

Another important safety consideration is the mechanism by which belatacept downregulates T-cell activation. T-cells in tumors can express CTLA-4, which binds to B7 proteins on antigen presenting cells, suppressing T-cell activation and facilitating evasion of the immune system by the cancer. Blocking CTLA-4 promotes antitumor immunity. For example, ipilimumab, a monoclonal antibody that blocks CTLA-4, improves survival for several cancers, including metastatic melanoma.¹² Thus, the use of belatacept in KTRs to downregulate T-cell activation resembles the pathway tumors exploit to evade immune surveillance. Prior studies of belatacept provided limited data for an assessment of cancer outcomes beyond PTLD.^7,9,13–17

To address these gaps in knowledge, we compared cancer risk between KTRs treated with belatacept or tacrolimus in a large population sample in the US. We also examined the association of belatacept use with other key posttransplant outcomes, namely, patient and graft survival.

A total of 135,175 individuals received a kidney-only transplant in the US between January 16, 2011, and December 31, 2019. Of these, 126,678 were included in the study, including 1,514 who used belatacept and 125,164 who used tacrolimus as part of their baseline maintenance immunosuppressive regimen (SRTR cohort, Figure 1). Among these, 65,280 resided in an area covered by a TCM cancer registry and did not have a pretransplant cancer (TCM cohort: 815 on belatacept and 64,465 on tacrolimus). Finally, 7,570 tacrolimus users were PS-matched with belatacept users in the SRTR cohort, and 4,057 tacrolimus users were PS-matched to belatacept users in the TCM cohort. The PS-matched group of tacrolimus users was more similar to the belatacept users than the overall group of tacrolimus (Supplementary Table 1). Our PS models were well calibrated as the sum of cross-predicted probabilities was very close to the observed frequency (data not shown) and Hosmer and Lemeshow goodness-of-fit tests were nonsignificant (TCM cohort p=0.32; SRTR cohort p=0.49). The AUC was 0.70 for the TCM cohort and 0.67 for the SRTR cohort, showing good discrimination of the PS model.

The characteristics of KTRs in the TCM and SRTR cohorts included in our primary analyses are presented in Table 1. The median age at transplantation was 57 years in belatacept users and 56 years in tacrolimus users. Most participants were male, and the most frequent racial/ethnic group was non-Hispanic White. Forty-one percent of the transplants among belatacept users occurred between 2014 and 2016. Among belatacept users, 8.5% in the TCM cohort and 6.0% in the SRTR cohort were EBV-seronegative at transplantation. Fifty-four percent of belatacept users received induction therapy with polyclonal antibodies, and 46% received IL2 antagonists. More than 98% of KTRs in both groups received mycophenolate as part of their baseline maintenance immunosuppression. The proportion of KTRs treated for acute rejection in the first 18 months posttransplant was higher in belatacept compared to tacrolimus users (SRTR cohort: 14.5% [N=220] vs. 6.4% [N=499], p<0.001).

Among KTRs in the TCM cohort, the median duration of follow-up was 2.3 years for belatacept users and 2.5 years for tacrolimus users. The overall incidence of cancer was 10.1 and 12.6 per 1,000 person-years for belatacept users and tacrolimus users, respectively. An estimated 3.6% of KTRs in both groups were diagnosed with cancer within 3 years after transplantation; these proportions increased to 4.7% in belatacept users and 6.4% in tacrolimus users 5 years after transplantation (Figure 2A, Supplementary Table 3). All cancers in the belatacept group occurred within 6 years posttransplant (Figure 2A, Supplementary Table 3). Among specific cancer outcomes, colorectal cancer was the most common in belatacept users followed by stomach cancer and non-Hodgkin lymphoma (Table 2). Prostate cancer and non-Hodgkin lymphoma were the most common in tacrolimus users (Table 2). We did not find a significant difference in overall cancer risk associated with belatacept use (adjusted HR 0.83, 95% confidence interval [CI] 0.53–1.30) (Table 2, Figure 2A). No association was observed between belatacept and risk for specific cancer types, although the number of events in these analyses was small (Table 2).

In the SRTR cohort, the median duration of follow-up was 4.0 years for both groups. An estimated 8.3% of belatacept users and 6.2% of tacrolimus users died within 3 years after transplantation; these proportions increased to 13.9% and 11.3%, respectively, at 5 years, and 19.1% and 17.9%, respectively, at 7 years (Figure 2B, Supplementary Table 3). Mortality rates were higher in belatacept users (31.0 vs. 25.9 per 1,000 person-years; adjusted HR 1.22, 95%CI 1.04–1.43; Table 2). The risk of death associated with belatacept was 20% increased among EBV-seropositive KTRs (adjusted HR 1.20, 95%CI 1.01–1.41; Table 3) and nearly two-fold increased among EBV-seronegative KTRs (1.98, 1.12–3.50; Table 3). Of the 1,024 deaths in the SRTR cohort, 743 (73.6%) occurred through December 31, 2019 (Table 4) and had a COD available. Belatacept was not associated with any specific COD (Table 4).

GF/RT rates were similar in both groups (19.6 vs. 20.6 per 1,000 person-years; adjusted HR 0.95, 95%CI 0.79–1.16). This association exhibited non-proportionality, i.e., belatacept was not associated with GF/RT in the first four years posttransplant (adjusted HR 1.16, 95%CI 0.93–1.44) but was associated with reduced risk after four years posttransplant (adjusted HR 0.54, 95%CI 0.35–0.83; p for interaction 0.002). This time-dependency of the association between belatacept and GF/RT is shown graphically in Figure 2C, where the curves diverge after 6 years posttransplant. The association between belatacept and GF/RT was similar among EBV-seropositive KTRs (Table 3, Supplementary Figure 2).

PTLD risk was 62% higher with belatacept use (adjusted HR 1.62, 95%CI 0.89–2.95; Table 2). In our analysis stratified by recipients’ EBV serostatus (Table 3), belatacept was associated with increased PTLD risk among KTRs who were EBV-seropositive at transplantation (adjusted HR 1.96, 95%CI 1.03–3.73) (Table 3). There were no PTLD events among EBV-seronegative KTRs, so the association could not be assessed in that group.

In our sensitivity analysis using the unmatched group of all tacrolimus users, we found similar associations with belatacept use for cancer overall, death, GF/RT, PTLD and cSCC (Supplementary Tables 4 and 5). However, in the unmatched analysis stratified by EBV serostatus, the associations between belatacept and mortality were no longer significant (EBV-seropositive KTRs: adjusted HR 1.14, 95%CI 0.98–1.33; EBV-seronegative KTRs: 1.50, 0.91–2.47; Supplementary Table 3). The additional sensitivity analyses that incorporated transplant center effects or individual matching produced similar results to the main analysis (Supplementary Table 6).

In this large study based on real-world evidence, KTRs who used belatacept as part of their initial maintenance immunosuppressive regimen did not experience an overall increased risk of cancer, but belatacept was associated with an increased risk of PTLD in EBV-seropositive KTRs compared to those who used tacrolimus. Of concern, belatacept use was also associated with an increased risk of death. In addition, we demonstrated a long-term nephroprotective effect of belatacept compared with tacrolimus, specifically, we showed a decreased risk of GF/RT among KTRs who received belatacept.

Our study provides reassurance that belatacept does not increase the risk of invasive cancer. A few randomized trials have reported null associations between belatacept use and cancer outcomes,^7,9,13 but the trials were too small to precisely estimate an effect. Our study had a much larger number of KTRs, leading to a more precise estimate for cancer overall, despite the low number of site-specific cancer events and the relatively short median follow-up duration. We also found that belatacept was not associated with risk of cSCC. In contrast, a recent small retrospective study found that, among KTRs with an initial diagnosis of cSCC, switching to belatacept from a CNI reduced the risk of developing a new cSCC.²¹

A new and notable finding in our study is the observed association between belatacept use and an increased risk of death. The five-year estimated mortality among belatacept users in our study is higher than that reported in the BENEFIT trial⁷ but comparable to that reported in BENEFIT-EXT among KTRs with extended criteria donors.⁹ These two randomized trials did not report a difference in mortality between KTRs treated with belatacept or cyclosporine. The difference in results between the trials and our study may be attributed to the fact that the trials recruited only carefully screened participants, whereas our study utilized real-world data on a much larger population of KTRs. Additionally, BENEFIT and BENEFIT-EXT compared belatacept with cyclosporine, while our study compared it with tacrolimus, the current standard-of-care immunosuppressant agent.

Various factors could contribute to the higher mortality among belatacept users, including the type, doses, and timing of other immunosuppressive agents.²² Another critical factor may be the rate of acute rejection. In the BENEFIT trial, acute rejection was more common in the belatacept group compared to the cyclosporine group.¹⁰ Similarly, in our study, treatment for acute rejection was more than twice as common in the first 18 months posttransplant among belatacept users than tacrolimus users. Such treatment, which involves, intensification of the immunosuppressive regimen, could have increased the risk of severe infection. However, our cause-specific mortality analysis did not identify any specific COD attributable to belatacept use. Given the different levels of statistical significance between our PS-matched analysis and the sensitivity analysis based on all tacrolimus users, we cannot exclude chance as an explanation for the apparent increase in mortality.

A novel finding in our study is the observed association between belatacept use and increased PTLD risk among EBV-seropositive KTRs. EBV plays a key role in the development of a large fraction of PTLD tumors.²³ Immunosuppressive agents decrease T-cell immune surveillance, resulting in EBV-induced proliferation of B-cells.^23,24 Our results suggest that belatacept has a greater effect than tacrolimus in causing loss of immune control of EBV, even in the setting of infection that is already established at the time of transplantation.

PTLD risk is greatest in individuals who are EBV-seronegative at the time of transplantation, among whom primary EBV infection after transplantation can lead to uncontrolled viral replication. BENEFIT-EXT reported an increased risk of PTLD in EBV-seronegative KTRs,²⁵ which resulted in a black box warning to avoid belatacept use in this subgroup of patients. As a result, very few EBV-seronegative KTRs were prescribed belatacept in the US after approval of the medication. Given the small number of EBV-seronegative KTRs and absence of PTLD events in this group, we were unable to assess the association between belatacept use and PTLD risk among EBV-seronegative KTRs.

Our study supports a long-term beneficial effect of belatacept on graft survival, which may be attributable to its less nephrotoxic nature.²⁶ In the BENEFIT trial of KTRs receiving kidneys from standard criteria donors, a less intensive belatacept regimen reduced the risk of graft loss by 43% compared with a cyclosporine-based regimen.⁷ Among KTRs who received kidneys from extended criteria donors in BENEFIT-EXT, belatacept was associated with improved long-term kidney function in comparison with cyclosporine, although the association with graft loss was not significant.^7,9 We observed a protective association between belatacept and GF/RT after 4–6 years posttransplant, consistent with the long-term nephrotoxic effects of CNIs leading eventually to graft failure.⁴

Our study has several important strengths. We utilized data from a large population-based sample of belatacept users, which allowed us to assess associations of belatacept with several important outcomes in a real-world setting. We compared belatacept with tacrolimus, which is now the most commonly used immunosuppressive medication in kidney transplantation. Finally, we were able to incorporate data from 34 population-based cancer registries (covering 55.4% of the US KTR population during this period) to assess the incidence of invasive cancers.

This study also has several limitations. First, exposure to belatacept and tacrolimus was determined from the baseline maintenance immunosuppressive regimen, i.e., at the time of initial hospital discharge following transplantation. Thus, our findings may be generalizable only to KTRs who receive belatacept as a baseline maintenance immunosuppressant medication. KTRs in our study might later have been converted to or from belatacept or another regimen, but we could not assess that. To the extent this conversion occurred, it most likely would have biased our results towards the null. Second, since this study is not a randomized trial, there might be differences between belatacept and tacrolimus users that contributed to their risk of the outcomes that we evaluated. We attempted to mitigate this issue through PS matching. Nonetheless, residual confounding by unmeasured factors could have affected the results.

A third limitation is that, due to the relatively recent introduction of belatacept, our study could only estimate the short-term association with cancer, while cancer incidence typically increases with longer follow-up after transplantation.²⁷ For other outcomes, our study included a much larger number of belatacept users at risk in each time period compared to prior studies.^7,9 PTLD can occur early (within 2 years following transplantation) or later.²⁸ In BENEFIT, all PTLD cases occurred within the first 2 years posttransplant. In our study, 10 of the 14 PTLD cases among belatacept users and 23 of the 47 cases among tacrolimus users occurred within 2 years, but we also had a substantial number of KTRs followed beyond this period. Fourth, due to the black box warning, only a small number of EBV-seronegative transplant recipients received belatacept, so we were limited in assessing this group. Finally, we excluded KTRs who concurrently received tacrolimus, cyclosporine, or mTOR inhibitors, to allow a straightforward comparison of belatacept with tacrolimus. Our results therefore do not apply directly to KTRs receiving belatacept in combination with certain other medications.

In conclusion, our study indicates that belatacept presents a similar overall cancer risk profile for KTRs, compared to a tacrolimus-based immunosuppressive regimen. We found that belatacept was associated with a long-term beneficial effect in reducing the incidence of GF/RT compared to tacrolimus. Of concern, however, we observed an increased risk of PTLD associated with belatacept use among EBV-seropositive KTRs, suggesting that this medication causes marked loss of EBV control even among individuals who have been infected with the virus before transplantation. Finally, the new finding in our study was the higher mortality in KTRs treated with belatacept, pointing to the need for further investigation to characterize adverse events in KTRs treated with belatacept.