Journal Logo

Original Clinical Science—General

Induction Therapy in Elderly Kidney Transplant Recipients With Low Immunological Risk

Masset, Christophe MD1,2; Boucquemont, Julie PhD3; Garandeau, Claire MD1,2; Buron, Fanny MD4; Morelon, Emmanuel MD, PhD4; Girerd, Sophie MD5; Ladrière, Marc MD5; Mourad, Georges MD, PhD6; Garrigue, Valérie MD6; Cassuto, Elisabeth MD7; Albano, Laetitia MD7; Foucher, Yohann PhD1,3; Dantal, Jacques MD, PhD1,2; for the DIVAT Consortium*

Author Information
doi: 10.1097/TP.0000000000002804



The choice of induction therapy that targets graft recipient immune T cells remains controversial in kidney transplantation. Thymoglobulin or antithymocyte globulins (ATG) is a polyclonal antibody directed primarily against T cells, which also depletes other immunological cellular subsets.1 While it is known that ATG provides better outcomes for high immunological risk patients,2,3 there is no consensus on whether it is beneficial in low immunological risk patients. Basiliximab (BSX) is a chimeric mouse-human monoclonal antibody (Simulect) that targets the CD25 α chain (CD25) of the IL-2 receptor of T cells.

Furthermore, there is no difference between ATG and BSX regarding biopsy-proven acute rejection in patients following a corticosteroid withdrawal or corticosteroid-free therapy regimen.4 Nevertheless, ATG is associated with side effects, especially a higher risk of infections due to prolonged lymphopenia induced by its depleting effect.5-7 In addition, ATG use has been associated with cytokine release syndrome,8 fever, chills, pain, serum sickness, headache, diarrhea, hypertension, thrombocytopenia, and dyspnea and has been shown to increase malignancy, especially posttransplant lymphoproliferative disorder (PTLD).9-11 Based on a meta-analysis and regarding the side effects of ATG,12 the International KDIGO consensus recommends using BSX as induction therapy in low immunological risk patients.

Despite these guidelines, ATG is still the primary induction therapy used in North American countries13 and many others. This may be to avoid steroid use and delay the introduction of calcineurin inhibitors (CNI) without increasing rejection rates.14,15 The protective effect of ATG could be coupled to low-dose CNI therapy, especially in recipients receiving an expanded criteria donor (ECD) graft who are more susceptible to CNI-induced nephrotoxicity. Finally, data showed that ATG may prevent delayed graft function (DGF)16 and DGF of >6 days is strongly associated with reduced long-term survival of kidney allografts.17

Elderly patients are more likely than others to receive ECD grafts,18 which may therefore support the use of ATG in this population. However, they also have a higher risk for malignancy and infections due to reduction in immune system function that occurs with age, supporting the use of BSX, especially if they are at low immunological risk. The best induction therapy in these patients therefore remains unclear.19-21 While kidney transplants for elderly recipients (>65 y) represent around 20% of the total kidney transplants performed today in France,22 few studies have compared patient and graft outcomes based on ATG or BSX in this population.23-25 Clinical trials investigating the choice of induction therapy have neither considered elderly recipients26,27 nor reflected routine settings.28,29 We thus need observational studies in this population cohort to evaluate the most relevant induction therapy. Using a French multicentric cohort, we compared a range of posttransplantation outcomes between ATG and BSX in elderly kidney recipients at low immunological risk.


Studied Population

The included patients were adults ≥65 years receiving a first kidney transplantation from brain-dead donors, treated with either ATG or BSX as induction therapy between 2010 and 2017. Multiple organ transplants were not considered. We only included patients with nondetectable anti-human leucocyte antigen (HLA) class I and II antibodies using Luminex assay (donor specific and nonspecific antibodies) at the time of transplantation.

Data were extracted from the French DIVAT cohort (, approved by the CNIL, no. 914184) consisting of recipients monitored in Nantes, Lyon, Montpellier, Nancy, and Nice. The quality of the DIVAT data bank is validated by an annual cross-center audit. All participants gave informed consent.

The patients in the BSX group received 20 mg of Simulect at day 0 and day 4. The patients in the ATG group received Thymoglobulin, administrated at 1.5 mg/kg/day (maximum 75 to 100 mg/d, depending on centers). All patients received initial corticotherapy during the first days of transplantation, followed by maintenance immunosuppressive therapy.

Available Data

Donor features included age, gender, creatinine level, allograft status (extended criteria donor or standard criteria donor), cause of death, and cytomegalovirus (CMV) and Epstein-Barr virus status. Recipient characteristics included age, gender, body mass index, comorbidities (diabetes, hypertension, dyslipidaemia, neoplasia, vascular, cardiovascular history), duration on waiting list, preemptive transplantation, CMV serology status, and initial renal disease. Transplantation parameters were the cold ischemia time and number of HLA-A-B-DR incompatibilities. Patients lost to follow-up were right-censored for mid- or long-term time-to-event. We assumed that the corresponding information was noninformative. For missing data, we excluded patients for which value spread too much from the initial date (>3 mo for 1-y analysis).


The principal outcome was the patient and graft survival, defined by the time between the transplant and first event requiring return to dialysis, preemptive retransplantation, or death with a functioning graft. Second, we studied cumulative probabilities of all infectious causes (bacterial, BkV viremia or BkV nephropathy, CMV, or mycotic), specific infections (bacterial, BkV, and CMV), first acute rejection episode (all biopsy-proven acute rejections including borderline lesions), malignancy, posttransplant diabetes (PTD) (patients with a diabetes diagnosed before transplantation were excluded), and cardiovascular complications (defined as occurrence of cardiac insufficiency, coronaropathy disease, rhythmic complication, or valvopathy). For patients with Bk virus infection, we considered only viremia and BkV nephropathy (all patients with viremia underwent a diagnosis biopsy or were considered as BkV nephropathy on clinical elements if biopsy was not possible).

Achievement of protocol biopsies were performed according to center practice policy. According to KDIGO recommendations, we considered indicated biopsies for rejection by the occurrence of 1 criterion among increasing of creatininemia (>25%) without any explanation, DGF >10 days, occurrence of de novo Donor Specific Anti HLA-Antibody, new onset of proteinuria, or unexplained proteinuria >3 g per day. Occurrence of de novo donor specific antibody (DSA) detected by Luminex assay and the estimated glomerular filtration rate (eGFR; estimated by MDRD) at 1 year posttransplantation were evaluated (patients who died or lost to follow-up before the first anniversary were excluded). We also studied the DGF occurrence, defined as the need for at least 1 dialysis session within the 7 days following the transplant. For this last outcome, the preemptive transplant patients and patients treated by peritoneal dialysis were excluded because the DGF definition was not evaluable.

Statistical Analysis

The characteristics at the time of transplantation between ATG and BSX groups were compared using Chi-square or Fisher exact tests for categorical variables and Student t-tests for continuous variables. Except for eGFR, to consider possible confounding variables, we weighted the models on the propensity scores,30 which were obtained by a multivariable logistic regression, including the center among the explicative variables. The variables significantly associated with the outcome were retained (P < 0.2). Splines on continuous covariates were used to ensure the log-linearity assumption. Stabilized weights were computed for estimating the average treatment effect on the entire population.31 Diagnostics of the models were assessed graphically for the positivity assumption and from the standardized differences for the correct model specification. Influential values were detected by a Cook distance >1 in absolute value.

For the time-to-event outcomes, cause-specific Cox models were estimated by maximizing the partial weighted likelihood with robust estimator for the variance.32 Proportionalities of hazards were checked graphically with log minus log survival curves. The adjusted survival curves were obtained using the weighted Kaplan-Meier estimator and compared using the adjusted log-rank test.33,34 We estimated the restricted mean survival time (RMST) at 3 years, that is, the mean time without the event for a cohort followed up to 3 years, by the area under the survival curve up to 3 years posttransplantation. The corresponding restricted mean time lost (RMTL) at 3 years was obtained by 3 minus the RMST.

For the DGF and de novo DSA outcomes, logistic models were also performed by maximizing the partial weighted likelihood with robust estimator for the variance. For the 1-year eGFR, a multiple linear model was performed.

Statistical analyses were performed using Plug-Stat software ( based on the R software.35 Tables S1-S14 (SDC, describe the variables significantly associated with the outcomes in the pseudo-samples after weighting on propensity scores. Note that respecting the methodology in causal inference, we did not consider the corticotherapy and tacrolimus residual dosages in the propensity scores. Obviously, physicians adapt the maintenance therapies according to the initial induction therapy. Figures S1-S4 (SDC, represent the positivity assumption validations.

Ethics Statement

All patients were included and extracted from the DIVAT database after informed consent. To respect confidential medical information, all data were anonymized before analysis.


Description of the Cohort

The study flow-chart is presented in Figure 1. The characteristics of the 383 studied patients at transplantation are described in Table 1. Two hundred and four patients were in the BSX group (53.3%) versus 179 in the ATG group (46.7%). Nineteen patients were lost to follow-up (8 in ATG group and 11 in BSX group). Use of ATG between the centers was heterogeneous, ranging from 16.2% to 73.2%. The average administration of ATG was 6.8 days (2 to 18 d), and average total dose of thymoglobulin was 5.22 mg/kg (SD = ±1.81), ranging from 1.1 to 12 mg/kg (n = 146 patients). The mean recipient age was 71.0 years in the BSX group versus 70.5 years in the ATG group (P = 0.373). About 25% of patients had a history of neoplasia before transplantation in both groups (25.5% versus 23.5%, P = 0.646). Among the characteristics significantly different between the 2 groups was the higher percentage of preemptive transplants in the BSX group (21.1% versus 10.1%, P = 0.004). Patients in the BSX group also had more frequent histories of dyslipidemia (62.7% versus 51.4%, P = 0.025) but less frequent CMV positive serologies (54.7% versus 68.0%, P = 0.008). Almost all grafts resulted from extended criteria donor with a similar cold ischemia time in both groups (15.3 versus 15.9 h, P = 0.282). Finally, we noticed a trend to a larger use of machine perfusion in the BSX group (59.8% versus 48%, P = 0.068).

Description of the entire cohort according to induction therapy
Flow-chart of the study. DGF, delayed graft function; eGFR, estimated glomerular filtration rate; PTD, posttransplant diabetes.

In the BSX group, 57.3% of patients received corticotherapy by month 3, whereas this was 82.9% for the ATG patient group (P < 0.001) with average dosage of 9.4 and 8.4 mg/day, respectively (P = 0.376). By 1 year, 62.5% and 78.2% of patients (P = 0.005) received a corticotherapy with average dosages of 6.5 and 5.8 mg/day (P = 0.078) in the BSX and ATG groups, respectively. In both groups, there was a large use of CNI during the first year posttransplantation (n = 319 patients, 96.4% versus 99.4%, P = 0.091 on mo 3; n = 289 patients, 89.2% versus 90.6%, P = 0.841 on mo 12 on ATG and BSX groups, respectively). Trough levels of Tacrolimus were similar at 1 year posttransplantation, whereas we observed a higher mean value on month 3 in the BSX group compared with the ATG group (9.48 versus 7.30 ng/mL, P = 0.023). Finally, patients underwent a maintenance therapy by mycophénolate mofétil (MMF) or MPA were similar between groups during the first year, whereas their average dosage was higher in the ATG group on month 3 (1378 versus 1166 mg, P = 0.001 in the ATG and BSX groups, respectively, expressed on equivalent MMF). All these results are summarized in Figure 2.

Representation of maintenance therapies used among groups. A, CNI use (ciclosporin in gray and white and tacrolimus in black and striped, respectively) and levels of residuals dosages of tacrolimus (with corresponding SEM, SEM) among ATG and BSX groups. B, MPA and MMF use (MPA in gray and white and MMF in black and striped, respectively) and average of equivalent MMF dosages among groups (with corresponding SEM). C, Percentage of patients underwent a steroid therapy among groups and their average dosage (with corresponding SEM). * represents a P < 0.05; ** represents a P < 0.01; *** represents a P < 0.001; n = 319 patients on mo 3, 302 patients on mo 6, and 289 patients on mo 12. ATG, antithymocyte globulins; BSX, basiliximab; CNI, calcineurin inhibitors; SEM, standard error of the mean.

The median follow-up time in the cohort was 2.0 years (range 0.0–8.2). During follow-up, 42 deaths with a functioning graft (28 in the BSX group) and 43 returns to dialysis (24 in the BSX group) were observed. We observed 197 patients with at least 1 infection, 63 patients with at least 1 acute rejection, and 64 patients with a de novo malignancy.

Patient and Graft Survival

The confounder-adjusted patient and graft survival curves are presented in Figure 3A. The patient and graft survival at 1 and 3 years posttransplantation were 87% (95% CI, 82%-93%) and 74% (95% CI, 65%-84%) in the ATG group versus 89% (95% CI, 84%-94%) and 68% (95% CI, 60%-78%) in the BSX group.

Confounder-adjusted probabilities of events according to the time posttransplantation and the induction therapy. A, Patient and graft survival. B, Cumulative probability of acute rejection episode. C, Cumulative probability of infection. D, Cumulative probability of malignancy. E, Cumulative probability of PTD. F, Cumulative probability of cardiac complications. ATG, antithymocyte globulins; BSX, basiliximab; PTD, posttransplant diabetes.

The confounder-adjusted RMST at 3 years posttransplantation was 2.47 years (95% CI, 2.29-2.66) in the ATG group; that is, the mean time of patient and graft survival was 2.47 years for the ATG cohort followed up 3 years posttransplantation. Comparatively, the RMST was 2.51 years (95% CI, 2.37-2.65) for the patients treated by BSX. This means that patients receiving ATG had a life expectancy with a functioning graft decreased by 0.48 months (0.04 y) compared with similar patients receiving BSX for an observation period of 3 years. Overall, the corresponding hazard ratio (HR) equaled 0.96 between the BSX group compared with the ATG group (95% CI, 0.58-1.60).

Cumulative Probability of Acute Rejection Episode

Two centers practiced protocol biopsy and 3 did not. Totally, 234 protocol biopsies were performed on 136 patients (54 in the ATG group and 82 in the BSX group). The confounder-adjusted cumulative probabilities of acute rejection episodes are presented in Figure 3B. The probability at 1 year posttransplantation was 17% (95% CI, 11%-23%) in the BSX group versus 16% (95% CI, 9%-22%) in the ATG group. Among them, 22.5% were diagnosed on protocol biopsies (25% in ATG group and 20% in BSX group), and about 65% of these protocol biopsies revealed Borderline Rejection. The confounder-adjusted RMTL at 3 years posttransplantation was 0.55 years (95% CI, 0.36-0.75) in the ATG group, that is, patients treated by ATG lost 0.55 years on average during the first 3 years of follow-up. Comparatively, the RMTL was 0.56 years (95% CI, 0.40-0.72) for the patients treated by BSX. Overall, the corresponding HR equaled 1.03 between the BSX group compared with the ATG group (95% CI, 0.61-1.76).

In addition, most of patients who presented a rejection episode during first year had Borderline lesions (15 in the ATG group and 10 in the BSX group), some had T cell–mediated rejection (4 in the ATG group and 12 in the BSX group), some had antibody-mediated rejection (3 in the ATG group and 1 the in BSX group), and a few ones mixed rejections (4 in the ATG group and 1 in the BSX group).

Cumulative Probability of Infection

The confounder-adjusted cumulative probabilities of infection are presented in Figure 3C. The probability at 1 year posttransplantation was 52% (95% CI, 44%-59%) in the BSX group versus 51% (95% CI, 42%-59%) in the ATG group. The HR for time to first infection equaled 1.08 between the BSX group compared with the ATG group (95% CI, 0.76-1.54). Regarding different causes of infections, the related cause-specific HRs were 0.64 (95% CI, 0.33-1.26) for BkV infection, 0.69 (95% CI, 0.43-1.12) for CMV infection, and 1.22 (95% CI, 0.81-1.85) for bacterial infection. Among all patients who presented BkV viremia, only 2 patients (1 in the ATG group and 1 in the BSX group) had histological lesions of BkV nephropathy on biopsy. All others patients presented a nonsymptomatic BkV viremia.

For 1 center, we had access to the evolution of immune cells during the first year following induction therapy. As illustrated in Figure 4, we observed a trend to higher white cells count in the BSX group during the first 3 months posttransplantation compared with ATG (6800 versus 5800/mm3 on mo 3). Moreover, regarding total lymphocytes, as expected, there was a lower count in patients receiving ATG than BSX, remaining during the first 6 months (1070 versus 490/mm3 on mo 1; 1230 versus 580/mm3 on mo 6). Unfortunately, the small number of patients having a follow-up of immune cells after transplantation did not permit a statistical analysis.

Evolution of white cell count (A) and Lymphocytes count (B) depending of the induction therapy (based on the analysis of available data). ATG, antithymocyte globulins; BSX, basiliximab.

Cumulative Probability of Malignancy

During our cohort follow-up, we observed a total of 38 malignancies among BSX group (3 PTLD, 9 cutaneous squamous cell carcinoma, 16 basal cell carcinoma, and 15 other malignancies). In the ATG group, there were 26 malignancy events (4 PTLD, 8 cutaneous squamous cell carcinoma, 11 basal cell carcinoma, 6 others malignancies). Among the 94 patients with cancer history at the time of their transplantation, 23 malignancies were diagnosed during the first 3 years (10 were treated by ATG and 13 by BSX) and none of these were recurrence.

The confounder-adjusted cumulative probabilities of malignancy are presented in Figure 3D. The probability at 1 year posttransplantation was 7% (95% CI, 3%-11%) in the BSX group versus 6% (95% CI, 2%-10%) in the ATG group. The probability at 3-year posttransplantation was 23% (95% CI, 14%-31%) in the BSX group versus 16% (95%CI, 6%-26%) in the ATG group. The confounder-adjusted RMTL at 3 years posttransplantation was 0.28 years (95% CI, 0.13-0.44) in the ATG group versus 0.33 years (95% CI, 0.21-0.45) for the patients treated by BSX. Overall, the corresponding HR equaled 1.13 between the BSX group compared with the ATG group (95% CI, 0.58-2.19). The small number of neoplasia subtypes did not allow a robust confounder-adjusted statistical analyses.

De Novo DSA at 1 Year Posttransplantation

The observed number of patients with de novo DSA at 1 year were 4 in the ATG group (5.8%) and 7 in the BSX group (4.8%). The small number of events did not permit an evaluation of the adjusted odds ratio.

Renal Function at 1 Year Posttransplantation

For this study, among the 383 included patients, 20 were excluded because of a follow-up shorter than 1 year posttransplantation, 8 were excluded due to death before 1 year posttransplantation, and 76 patients were excluded because of missing data (Table S13, SDC, Before 1 year posttransplantation, 27 returns to dialysis were observed and the corresponding eGFR levels were fixed at 5 mL/min/1.73m2. As illustrated in Figure S5 (SDC,, the observed means of the 1-year eGFR were 36.8 mL/min/1.73m2 (95% CI, 33.5-40.0) in the ATG group versus 38.1 mL/min/1.73m2 (95% CI, 35.4-40.8) in the BSX group. After adjustment, the corresponding difference in the eGFR level at 1 year posttransplantation was 0.67 mL/min/1.73m2 (95% CI, −4.61 to 5.95; P = 0.804).

Delayed Graft Function

For this study, among the 383 included patients, 61 patients were excluded because of preemptive transplantations, 33 because they were treated by peritoneal dialysis, and 9 because of missing data (Table S14, SDC, The observed DGF percentages were 32% (95% CI, 25%-40%) in the ATG group versus 29% (95% CI, 21%-36%) in the BSX group. The corresponding confounder-adjusted percentages were 28% (95% CI, 20%-36%) versus 29% (95% CI, 21%-39%), respectively. The adjusted odds ratio for the BSX group compared with the ATG group equaled 1.06 (95% CI, 0.58-1.95).

Cumulative Probability of PTD

We excluded 123 patients with a history of diabetes at transplantation from this analysis (Table S12, SDC, The confounder-adjusted cumulative probabilities of PTD are presented in Figure 3E. The probability at 1 year posttransplantation was 23% (95% CI, 16%-30%) in the BSX group versus 15% (95% CI, 7%-21%) in the ATG group. The confounder-adjusted RMTL at 3 years posttransplantation was 0.41 years (95% CI, 0.21-0.60) in the ATG group. Comparatively, the RMTL was 0.71 years (95% CI, 0.51-0.92) for the patients treated by BSX. Overall, the corresponding HR equaled 1.88 between the BSX group compared with the ATG group (95% CI, 1.00-3.56).

Cumulative Probability of Cardiac Complication

The confounder-adjusted cumulative probabilities of cardiac complications are presented in Figure 3F. The probability at 1 year posttransplantation was 22% (95% CI, 16%-28%) in the BSX group versus 34% (95% CI, 25%-42%) in the ATG group. The confounder-adjusted RMTL at 3 years posttransplantation was 1.03 years (95% CI, 0.79-1.27) in the ATG group. Comparatively, the RMTL was 0.81 years (95% CI, 0.63-0.98) for the patients treated by BSX. Overall, the corresponding HR equaled 0.72 between the BSX group compared with the ATG group (95% CI, 0.47-1.11). Figure 5 summarizes all the relative risks for studied outcomes comparing ATG and BSX in elderly recipients with low immunological risk.

Summary of the RRs of using basiliximab instead of thymoglobulin for the studied outcomes. ATG, antithymocyte globulins; BSX, basiliximab; CMV, cytomegalovirus; HR, hazard ratio; OR, odds ratio; RR, relative risk.


With populations aging, more and more kidney transplants are performed for elderly recipients. The choice of induction therapy for nonimmunized elderly kidney transplant recipients remains an important consideration. In our study, the use of ATG was very heterogeneous between the different centers, ranging from 16.2% to 73.2% in this population, suggesting a lack of clinical consensus on best practices for nonimmunized elderly recipients.

We did not observe any statistically significant differences between patients treated by ATG and BSX in terms of patient and graft survival, infections, acute rejection, malignancy, 1-year de novo DSA, 1-year eGFR, and DGF, in agreement with other studies.36 In the Harmony Trial,4 the investigators included patients with low immunological risk with a rapid corticosteroid withdrawal and demonstrated the absence of superiority of ATG in comparison with BSX for patient/graft survival, rejection, or infection rates. Other studies have reported similar findings for induction therapy among patients with low immunological risk.6,15

However, all of these studies were conducted on the entire population of transplanted patients, and the average age was ≤50. In a retrospective study conducted in elderly recipients,23 no statistically significant difference between ATG and BSX was reported. The authors hypothesized that this result may be explained by a lower ATG dose compared with the planned one. However, our data showed that even with a complete ATG dose (average of 5.2 mg/kg total dose), there was no difference between ATG and BSX in terms of patient and graft survival.

Several studies reported a higher rate of viral infections, especially CMV3,37 with full depletant induction. In our study, ATG did not induce significatively more CMV infections, but this can be related to an insufficient number of patients. Likewise, there was a minor trend to higher cardiovascular complications in ATG group, which could be linked to CMV status.38

We did not observe any difference between groups regarding the occurrence of malignancy, while a previous study reported that ATG can be associated with an increased risk of PTLD.9 This is very valuable data, especially in the context of an elderly population: about 25% of our patients had a history of malignancy and ATG did not only lead to similar neoplasia outcomes, but also did not seem to increase the recurrence of a pretransplant neoplasia.

Although we had very few data on lymphocytes following induction therapy, our data suggest a persistent lymphopenia during the first year after ATG therapy comparing with BSX. Moreover, these data were from a center that practiced low doses of ATG (75 mg/d during 2 d), so we suggest that this is valid for most of patients who underwent an ATG induction in our cohort.

We observed similar infectious and neoplasia outcomes following ATG and BSX, while a prolonged lymphopenia in the ATG group. This may be linked to functionality of lymphocytes, that is known to decrease in elderly population, added to the maintenance therapy.39 This is concordant with recent data, which tend to show that depleting induction therapy has no effect on PTLD risk with current maintenance immunosuppressive regimens.11

Rejection rates at 1 year were also similar for both groups (16% and 11% for ATG and BSX, respectively), but were slightly higher than previously described,12 possibly due to the fact that we considered all rejection episodes including borderline episodes (representing about 50% in each group). It is noteworthy that if ATG is commonly used to decrease rejection rates, its impact on T cells subsets is not fully understood and some have demonstrated that the homeostatic repopulation observed after ATG induction influence negatively CD4+FoxP3+ Treg cells and also could favor the upregulation of markers associated with rejection.40-42 As expected, we observed a low occurrence of de novo DSA at 1 year (5.8% and 4.8% in ATG and BSX groups, respectively, without significant difference). This rate is lower than described by others,43 confirming a low immunological risk population with possibly a less effective immune system linked to the age.

In our cohort, we detected more PTD in the BSX group than the ATG group, whereas the number of patients who underwent a steroid therapy was smaller in BSX group. Nevertheless, steroid dosages were rather low in both groups (<10 mg/d), and the increased number of PTD in the BSX group could be more related to the use of higher dosages of CNI, traduced by significant higher trough levels.

Finally, the occurrence of DGF was similar in our study, whereas others have shown a benefit by using ATG.16 Kyllönen et al44 concluded that intraoperative administration of ATG lead to lower DGF than BSX, probably because of the avoidance of leukocyte sequestration in the allograft by reducing adherence to the antigen-presenting endothelial cells.45 Thus, it has been found that intraoperative ATG results in better reduction in DGF than postoperative administration.46 In our study, the lack of data on the exact time of ATG administration precludes any conclusion on the effects of ATG on DGF. Moreover, we observed a trend to a higher use of machine of perfusion in the BSX group, which could have minimized the potential difference of DGF between groups. The low observed eGFR at 1 year in both groups is probably related to the fact that all patients received a graft from an ECD (mean donor age was 72.7 y) because of their age and that they underwent a CNI therapy.47

Our study suffers from some limitations. First, one cannot exclude possible unobserved confounders. The randomized clinical trial remains the design of reference for evaluating the theoretical effect of ATG versus BSX. But because elderly recipients are often poorly represented in such trials, cohort-based studies with representative patients as in real-life practices are the most relevant studies to evaluate the effect of ATG versus BSX. Second, the sample size of our study is relatively small, which might explain the statistically nonsignificant trends for a higher incidence of CMV and BkV infections related to ATG induction therapy and the lack of difference between rejection rates between groups (added to nonsystematic surveillance biopsies). In addition, the potential effect of the center as a confounder has been considered. Indeed, the center was an explicative variable of the propensity scores, ensuring the balanced repartition of centers between both the ATG and BSX. Finally, regarding our mid-term results, we still need further long-term studies to validate the robustness of our conclusion, especially for late related outcomes.

In conclusion, whereas elderly kidney transplant recipients have weaker immune reconstitution following ATG induction therapy, there is no difference for patient and graft outcomes between ATG and BSX as induction therapy, neither for infections nor neoplasia outcomes. This highlights the fact that with most recent maintenance therapies permitting lower doses of ATG, added to prophylaxis and preventive strategies of infections (especially the CMV), induction therapy would not be the leading factor linked to infectious nor neoplasia complications. However, using a nondepletant therapy could require higher CNI trough levels and thus increase occurrence of PTD.


The authors thank the members of the DIVAT consortium for their involvement in the study, physicians who helped recruit patients, and all patients who participated in this study. We also thank the clinical research associates who participated in the data collection. Data were collected from the French DIVAT multicentric prospective cohort of kidney and/or pancreatic transplant recipients (, No. CNIL 914184). The analysis and interpretation of these data are the responsibility of the authors.


Lyon E. Hériot: Lionel Badet, Maria Brunet, Fanny Buron, Rémi Cahen, Sameh Daoud, Coralie Fournie, Arnaud Grégoire, Alice Koenig, Charlène Lévi, Emmanuel Morelon, Claire Pouteil-Noble, Thomas Rimmelé, Olivier Thaunat; Montpellier: Sylvie Delmas, Valérie Garrigue, Moglie Le Quintrec, Vincent Pernin, Jean-Emmanuel Serre; Nancy: Pascal Eschwege, Luc Frimat, Sophie Girerd, Jacques Hubert, Marc Ladriere, Emmanuelle Laurain, Louis Leblanc, Pierre Lecoanet, Jean-Louis Lemelle; Nantes: Gilles Blancho, Julien Branchereau, Diego Cantarovich, Agnès Chapelet, Jacques Dantal, Clément Deltombe, Lucile Figueres, Claire Garandeau, Magali Giral, Caroline Gourraud-Vercel, Maryvonne Hourmant, Georges Karam, Clarisse Kerleau, Aurélie Meurette, Simon Ville, Christine Kandell, Anne Moreau, Karine Renaudin, Anne Cesbron, Florent Delbos, Alexandre Walencik, Anne Devis; Nice: Laeticia Albano, Elisabeth Cassuto.


1. Mourad G, Morelon E, Noël C, et al. The role of thymoglobulin induction in kidney transplantation: an update.Clin Transplant201226E450–E464
2. Noël C, Abramowicz D, Durand D, et al. Daclizumab versus antithymocyte globulin in high-immunological-risk renal transplant recipients.J Am Soc Nephrol2009201385–1392
3. Brennan DC, Daller JA, Lake KD, et al.; Thymoglobulin Induction Study GroupRabbit antithymocyte globulin versus basiliximab in renal transplantation.N Engl J Med20063551967–1977
4. Thomusch O, Wiesener M, Opgenoorth M, et al. Rabbit-ATG or basiliximab induction for rapid steroid withdrawal after renal transplantation (harmony): an open-label, multicentre, randomised controlled trial.Lancet20163883006–3016
5. Longuet H, Sautenet B, Gatault P, et al. Risk factors for impaired CD4+ T-cell reconstitution following rabbit antithymocyte globulin treatment in kidney transplantation.Transpl Int201427271–279
6. Mourad G, Rostaing L, Legendre C, et al. Sequential protocols using basiliximab versus antithymocyte globulins in renal-transplant patients receiving mycophenolate mofetil and steroids.Transplantation200478584–590
7. Kho MM, Bouvy AP, Cadogan M, et al. The effect of low and ultra-low dosages thymoglobulin on peripheral T, B and NK cells in kidney transplant recipients.Transpl Immunol201226186–190
8. Couvrat-Desvergnes G, Salama A, Le Berre L, et al. Rabbit antithymocyte globulin-induced serum sickness disease and human kidney graft survival.J Clin Invest20151254655–4665
9. Cherikh WS, Kauffman HM, McBride MA, et al. Association of the type of induction immunosuppression with posttransplant lymphoproliferative disorder, graft survival, and patient survival after primary kidney transplantation.Transplantation2003761289–1293
10. Lim WH, Turner RM, Chapman JR, et al. Acute rejection, T-cell-depleting antibodies, and cancer after transplantation.Transplantation201497817–825
11. Hertig A, Zuckermann A. Rabbit antithymocyte globulin induction and risk of post-transplant lymphoproliferative disease in adult and pediatric solid organ transplantation: an update.Transpl Immunol201532179–187
12. Webster AC, Ruster LP, McGee R, et al. Interleukin 2 receptor antagonists for kidney transplant recipients.Cochrane Database Syst Rev2010CD003897
13. Hart A, Smith JM, Skeans MA, et al. OPTN/SRTR 2016 Annual Data Report: Kidney.Am J Transplant201818Suppl 118–113
14. Cantarovich D, Rostaing L, Kamar N, et al.; FRANCIA Study Trial Investigators GroupEarly corticosteroid avoidance in kidney transplant recipients receiving ATG-F induction: 5-year actual results of a prospective and randomized study.Am J Transplant2014142556–2564
15. Sureshkumar KK, Thai NL, Hussain SM, et al. Influence of induction modality on the outcome of deceased donor kidney transplant recipients discharged on steroid-free maintenance immunosuppression.Transplantation201293799–805
16. Chapal M, Le Borgne F, Legendre C, et al. A useful scoring system for the prediction and management of delayed graft function following kidney transplantation from cadaveric donors.Kidney Int2014861130–1139
17. Giral-Classe M, Hourmant M, Cantarovich D, et al. Delayed graft function of more than six days strongly decreases long-term survival of transplanted kidneys.Kidney Int199854972–978
18. Huang E, Segev DL, Rabb H. Kidney transplantation in the elderly.Semin Nephrol200929621–635
19. Danovitch GM, Gill J, Bunnapradist S. Immunosuppression of the elderly kidney transplant recipient.Transplantation200784285–291
20. Peeters LEJ, Andrews LM, Hesselink DA, et al. Personalized immunosuppression in elderly renal transplant recipients.Pharmacol Res2018130303–307
21. Montero N, Pérez-Sáez MJ, Pascual J, et al.; DESCARTES Working Group; DESCARTES ERA-EDTA BoardImmunosuppression in the elderly renal allograft recipient: a systematic review.Transplant Rev (Orlando)201630144–153
22. Agence de la biomédecine[The medical and scientific report of sampling and transplantation in France.]Available at Accessed May 25, 2018
23. Hurst FP, Altieri M, Nee R, et al. Poor outcomes in elderly kidney transplant recipients receiving alemtuzumab induction.Am J Nephrol201134534–541
24. Jasiak N, Thielke J, Chen J, et al. Outcomes in elderly renal transplant recipients using thymoglobulin or interleukin-2 receptor antagonist induction [abstract].Am J Transplant201616Suppl 3Abstract C41
25. Gill J, Sampaio M, Gill JS, et al. Induction immunosuppressive therapy in the elderly kidney transplant recipient in the United States.Clin J Am Soc Nephrol201161168–1178
26. Lebranchu Y, Bridoux F, Büchler M, et al. Immunoprophylaxis with basiliximab compared with antithymocyte globulin in renal transplant patients receiving MMF-containing triple therapy.Am J Transplant2002248–56
27. Ayaz-Shah AA, Hussain S, Knight SR. Do clinical trials reflect reality? A systematic review of inclusion/exclusion criteria in trials of renal transplant immunosuppression.Transpl Int201831353–360
28. Zwarenstein M, Treweek S. What kind of randomized trials do we need?CMAJ2009180998–1000
29. Blosser CD, Huverserian A, Bloom RD, et al. Age, exclusion criteria, and generalizability of randomized trials enrolling kidney transplant recipients.Transplantation201191858–863
30. Austin PC. The performance of different propensity-score methods for estimating differences in proportions (risk differences or absolute risk reductions) in observational studies.Stat Med2010292137–2148
31. Robins JM, Hernán MA, Brumback B. Marginal structural models and causal inference in epidemiology.Epidemiology200011550–560
32. Lin DY, Wei LJ. The robust inference for the Cox proportional hazards model.J Am Stat Assoc1989841074–1078
33. Xie J, Liu C. Adjusted Kaplan-Meier estimator and log-rank test with inverse probability of treatment weighting for survival data.Stat Med2005243089–3110
34. Le Borgne F, Giraudeau B, Querard AH, et al. Comparisons of the performance of different statistical tests for time-to-event analysis with confounding factors: practical illustrations in kidney transplantation.Stat Med2016351103–1116
35. R: A Language and Environment for Statistical Computing [computer program]2010Vienna, AustriaR Development Core TeamAvailable at
36. Opelz G, Unterrainer C, Süsal C, et al. Efficacy and safety of antibody induction therapy in the current era of kidney transplantation.Nephrol Dial Transplant2016311730–1738
37. Clesca P, Dirlando M, Park SI, et al. Thymoglobulin and rate of infectious complications after transplantation.Transplant Proc200739463–464
38. Ducloux D, Courivaud C, Bamoulid J, et al. Polyclonal antithymocyte globulin and cardiovascular disease in kidney transplant recipients.J Am Soc Nephrol2014251349–1356
39. Halloran PF. Immunosuppressive drugs for kidney transplantation.N Engl J Med20043512715–2729
40. Bouvy AP, Klepper M, Kho MM, et al. The impact of induction therapy on the homeostasis and function of regulatory T cells in kidney transplant patients.Nephrol Dial Transplant2014291587–1597
41. Havenith SH, Remmerswaal EB, Bemelman FJ, et al. Rapid T cell repopulation after rabbit anti-thymocyte globulin (rATG) treatment is driven mainly by cytomegalovirus.Clin Exp Immunol2012169292–301
42. Krepsova E, Tycova I, Sekerkova A, et al. Effect of induction therapy on the expression of molecular markers associated with rejection and tolerance.BMC Nephrol201516146
43. Hourmant M, Cesbron-Gautier A, Terasaki PI, et al. Frequency and clinical implications of development of donor-specific and non-donor-specific HLA antibodies after kidney transplantation.J Am Soc Nephrol2005162804–2812
44. Kyllönen LE, Eklund BH, Pesonen EJ, et al. Single bolus antithymocyte globulin versus basiliximab induction in kidney transplantation with cyclosporine triple immunosuppression: efficacy and safety.Transplantation20078475–82
45. Chappell D, Beiras-Fernandez A, Hammer C, et al. In vivo visualization of the effect of polyclonal antithymocyte globulins on the microcirculation after ischemia/reperfusion in a primate model.Transplantation200681552–558
46. Goggins WC, Pascual MA, Powelson JA, et al. A prospective, randomized, clinical trial of intraoperative versus postoperative thymoglobulin in adult cadaveric renal transplant recipients.Transplantation200376798–802
47. Iordanous Y, Seymour N, Young A, et al. Recipient outcomes for expanded criteria living kidney donors: the disconnect between current evidence and practice.Am J Transplant200991558–1573

Supplemental Digital Content

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.