The study cohort consisted in a monocentric cohort of kidney transplant patients from Lyon University Hospital, France. It was used to retrospectively compare the impact of induction treatment in patients receiving kidneys from controlled circulatory death donors. All patients receiving a kidney transplantation from a DCD donor between January 1, 2015 and December 31, 2023 were enrolled. Data were prospectively collected. The validation cohort (DIVAT, Données Informatisées et VAlidées en Transplantation) included 506 adult recipients of DCD kidney transplants from 9 French centers [Necker (Paris), Saint Louis (Paris), Foch (Suresnes), le Kremlin-Bicêtre, Saint-Etienne, Nantes, Nancy, Montpellier, Nice] between January 1, 2015, and December 31, 2023.

In France, the Maastricht III program, which defines the rules for transplants using controlled DCD donors, imposes strict conditions. Donors must be strictly under the age of 70 and must not have acute organ failure. The agonic phase must last less than 3 h and the functional warm ischemia phase (time interval during which the organs are hypoperfused and then no longer perfused at all due to circulatory failure) less than 120 min for the kidneys. Ex vivo hypothermic perfusion is mandatory and must last at least 2 hours, while CIT must be < 18h. Recipients can only receive a first, HLA-compatible transplant.

Patients who received both available induction treatments were excluded from the analysis.

Delayed graft function was defined as the need for at least one hemodialysis session after the transplantation.

Light microscopy was performed on systematic or “for cause” kidney biopsy samples fixed in formalin-acetic alcohol after staining with haematoxylin and eosin. Histological lesions were scored according to the Banff 2022 classification [21].

Qualitative variables were expressed as percentages and compared with the chi-square test or Fisher’s exact test when the conditions of application of chi-square were not met. Quantitative variables were expressed as mean ± SD and compared using Student’s t-test. All tests were two-sided. When repeated measurements of the same parameter were performed over time, two-way ANOVA or mixed effects model were used with Sidak’s multiple comparisons test. No imputation was performed. Analyses were performed using variable-specific complete-case analyses, excluding only observations with missing data for the variable under consideration. Incidence and survival data were analysed by Kaplan-Meier curve and compared using a log-rank test. A multivariable Cox proportional hazards model was used to assess patient and graft survival in the validation cohort. All data were analysed using Prism (GraphPad) or the R software. All analyses were performed using R Statistical Software (v4.5.1); R Core Team [22].

The study was carried out in accordance with French legislation on biomedical research and the Declaration of Helsinki. All patients gave informed consent for the utilization of clinical data [Données Informatiques Validées en Transplantation (DIVAT cohort)] for research purpose (www.divat.fr, CNIL (Commission Nationale de l’Informatique et des Libertés) no. 914184, ClinicalTrials. gov: NCT02900040).

Between January 2015 and December 2023, 163 patients received a transplant from a DCD donor. All received induction treatment with ATG (Thymoglobulin, Sanofi ®, target dose 5 mg/kg) or an IL-2RA, basiliximab (Simulect, Novartis ®, 20 mg day 0 and day 4). As they did not tolerate ATG, 5 patients were switched to basiliximab and received both molecules; they were therefore excluded from the study. Of the remaining 158 patients, the vast majority of those transplanted before the onset of COVID-19 pandemic in France (March 2020) have received ATG as an induction, as specified in the National protocol [23]. To limit the risk of COVID-19 induced mortality in the post-transplantation period [24, 25], almost all patients transplanted after March 2020 were induced with basiliximab (Figure 1A). However, the change in induction treatment was not accompanied by changes in the maintenance immunosuppressive regimen. In particular, the target residual doses of calcineurin inhibitors were the same for all patients (Figure 1B; of note, the few patients (n = 5; 3%) on ciclosporin were excluded from this analysis), regardless of the induction treatment.

Taking advantage of this situation, we underwent the present study aiming at comparing the effect of induction on the prognosis of transplants with DCD donors. Therefore, transplant outcomes (in terms of renal parameters and complications) were compared between the 64 patients induced with ATG and the 94 patients induced with basiliximab, whose characteristics are presented in Table 1. Briefly, the mean recipient age was 56.2 ± 11.7 and 52.0 ± 12.6 years and the mean donor age was 50.6 ± 12.6 and 50.8 ± 13.6 years, in ATG and basiliximab groups, respectively. The type of initial kidney disease leading to end-stage renal disease was similar in both groups. The mean follow-up was 3.0 ± 1.8 years; nine patients died during follow-up.

The cold-ischemia time was comparable between the groups (787 ± 214 min in ATG group vs. 769 ± 264 min in basiliximab group; unpaired t-test, p = 0.66; Table 1), and all kidneys were preserved on a perfusion machine. The rate of PNF was low (5; 3.16%) and did not differ according to the induction received (1 [1.56%] patient in ATG group, vs. 4 [4.25%] patients in basiliximab group; Fisher’s exact test, p = 0.65; Figure 2A). Among the remaining 153 patients, 15% experienced DGF, 11 in ATG group and 12 in basiliximab group (chi-square test, p = 0.48; Figure 2B). Thus, induction does not affect the early outcomes after transplantation.

Since some studies have indicated a lower risk of biopsy-proven rejection with ATG induction compared to IL-2RA [26], we sought to determine whether this also holds true for DCD transplants, considered more immunogenic [27]. In our cohort, patients had very low levels of immunization (19.6% with preformed anti-HLA antibodies, none with preformed donor-specific antibodies; chi-square test for comparison between groups, p = 0.86; Table 1), and the number of HLA-A/B/DR mismatches was also similar between the groups (3.7 ± 1.1 and 3.7 ± 1.2 in ATG and basiliximab groups, respectively; chi-square test, p = 0.86; Table 1). Overall, 5 patients experienced full blown T-cell mediated rejection (2 and 1 grade IA, 0 and 1 grade IIA, 0 and 1 grade III, in ATG and basiliximab groups, respectively) during the first year of follow-up, whereas one patient with concomitant BK nephropathy was excluded from this analysis. Over the same period of 1 year of follow-up, 6 other patients had borderline lesions (1 in ATG group and 5 in basiliximab group). The intensity of inflammatory lesions (borderline lesions or TCMR) did not differ between the groups (chi-square test, p = 0.45; Figure 2C). Of these 11 patients, all but two (with borderline lesions) were treated with corticosteroids and the control biopsy performed at 192 ± 83 days showed resolution of inflammation in all cases. Two of these 9 treated patients developed post-transplant diabetes mellitus. In addition, three patients in the basiliximab group developed microvascular inflammation (MVI) without DSA.

The evolution of renal function and proteinuria during the first year was comparable regardless of induction (Figures 2D,E). Patients had a mean one-year estimated glomerular filtration rate of 61.0 ± 19.6 mL/min/1.73 m² and 61.3 ± 20.3 mL/min/1.73 m² in ATG and basiliximab groups, respectively (unpaired t-test, p = 0.93; Figure 2D). At 1 year, among patients with an available proteinuria measurement, 10 of 59 and 12 of 80 had significant proteinuria ≥50 mg/mmol creatininuria in ATG and basiliximab groups, respectively (chi-square test, p = 0.76; Figure 2E).

At the systematic one-year evaluation, 102 patients had a biopsy (41 in the ATG group and 61in the basiliximab group, mean delay 383 ± 23 days post transplantation [range 275–455 days]). Among all biopsies for which the following Banff scores were available, there was no difference in terms of tubulointerstitial (i + t; evaluated on 91 biopsies) or microvascular (g + ptc; 89 biopsies) acute inflammatory lesions, nor in chronic glomerular lesions (cg; 90 biopsies) or interstitial fibrosis/tubular atrophy (IFTA; 88 biopsies; Figure 2F). Very few patients developed de novo donor-specific antibodies (2 in ATG group, 1 in basiliximab group, log-rank test, p = 0.36).

Finally, to integrate all the risks (PNF, DGF, rejection…) that could potentially affect graft function on the long term, we compared the death-censored graft survival of the two groups and found no difference (log-rank test, p = 0.90; Figure 2G).

Having demonstrated that graft survival was similar in the two groups of patients, we wanted to compare the safety of the two induction therapies.

Patient and graft survival were similar between the 2 groups (log-rank test, p = 0.58; Figure 3A). The number of bacterial complications (including bacteremia, pyelonephritis and bacterial pneumonia) during the first year was equivalent (chi-square test, p = 0.46; Figure 3B). Regarding viral complications, we observed no difference in the incidence of cytomegalovirus (CMV) viremia between the two groups (log-rank test, p = 0.94; Figure 3C). However, this result should be interpreted with caution because patients received protocolized prophylaxis against CMV with valganciclovir, the duration of which varied according to the serological status of the donor/recipient pair (6 months for the high-risk seropositive donor/seronegative recipient group [D+/R-] and 3 months for intermediate-risk group [R+]). Although the difference was not statistically significant, there were more D+/R- high-risk pairs in the basiliximab than in the ATG group (16% vs. 8% respectively; Table 1). Regarding EBV, there were only 4 high-risk pairs (D+/R-) and all received basiliximab as induction. In the rest of patients (D+/R+ or D-/R+), the incidence of EBV reactivation was comparable between groups (log-rank test, p = 0.75; Figure 3D) and the maximum viral load was similar (mean maximum viral load 3.1 ± 0.7 log vs. 3.1 ± 0.6 log in the ATG and basiliximab groups, respectively; unpaired t-test, p = 0.92; Figure 3E), but the duration of viremia was significantly longer in patients who had received ATG (log-rank test, p = 0.0028; Figure 3F). Despite this difference, no patient of the cohort developed EBV-related post-transplant lymphoproliferative disease during the follow-up period. For the analysis focusing on BK virus, we excluded from the analysis the 5 patients who experienced PNF and remained on dialysis. The results were reminiscent to those obtained for EBV. The incidence of viremia (log-rank test, p = 0.91; Figure 3G), and maximum viral loads (mean maximum viral load 3.8 ± 1.4 log vs. 4.0 ± 1.3 log in ATG and basiliximab groups, respectively; unpaired t-test, p = 0.71; Figure 3H), were similar in the two groups but patients from the ATG group exhibited longer duration of viremia (log-rank test, p = 0.0297; Figure 3I). However, this difference did not translate into a higher risk of BK virus-associated nephropathy (BKVAN), as only 4 patients developed BKVAN, 1 in the group ATG and 3 in the group basiliximab (log-rank test, p = 0.41).

Finally, the risk of cancer was equivalent in each group (log-rank test, p = 0.58; Figure 3J).

To assess the external validity of our findings, we analysed an independent validation cohort comprising 521 kidney transplant recipients from nine additional French centers who received a DCD graft between January 2015 and December 2023. Fifteen patients were excluded due to missing induction data. Among the 506 patients analysed, 408 received ATG and 98 received basiliximab. The characteristics of these patients are presented in Table 2. Of note, a higher proportion of patients in this cohort were immunised, with 61% having preformed anti-HLA antibodies and 15% having day-0 DSA. Furthermore, 7% of grafts in both groups were not preserved using a perfusion machine.

The distribution of induction regimens was more balanced over time than in the study cohort (Figure 4A). The rates of PNF, DGF (after exclusion of PNF), and rejections occurring during the first-year post-transplantation were similar between groups (Figures 4B,C). Kidney function and proteinuria at 1 year were also comparable between patients receiving ATG or basiliximab (Figure 4D).

Protocol biopsies were less systematically performed in this cohort, with fewer than one third of patients undergoing a one-year biopsy; therefore, no biopsy analysis at 1 year is reported.

Importantly, death-censored graft survival did not differ between the groups (log-rank test, p = 0.44; Figure 4E). Multivariable Cox regression adjusted for recipient age, HLA-A/B/DR mismatches and cold ischemia time confirmed that survival was comparable between basiliximab and ATG groups (HR 1.26, 95% CI 0.65–2.45; p = 0.49; Figure 4F).

From a safety perspective, the two groups were compared in terms of viral and malignancy-related complications. The incidences of CMV, EBV and BKV viremia were comparable (log-rank tests, p = 0.26, p = 0.42 and p = 0.2, respectively; Figures 5A–C). A non-significant trend towards an increased risk of cancer (all types combined) was observed in the ATG group (log-rank test, p = 0.12; Figure 5D).

Basiliximab-treated patients exhibited slightly more early inflammatory kidney lesions without long-term graft impact, while ATG-treated patients experienced prolonged EBV and BKV viremia without apparent severity. These non-significant differences balance the risks, offering no clear advantage for either agent. A pragmatic approach is to consider tolerance - measured here by the length of the initial hospital stay - and cost. Hospital stays were similar in both groups (mean 12 ± 6 days for ATG vs. 12 ± 7 days for basiliximab, p = 0.80; and 12 ± 8 days for ATG vs. 12 ± 7 days for basiliximab, p = 0.48; in the study and validation cohorts, respectively; Figure 6A). However, at our institution, ATG induction is, on average, 1.7 times (range 0.92–2.65) more expensive than basiliximab (€300 per ATG vial, median 15 vials per patient versus €1,315 per basiliximab vial, 2 vials per patient; Figure 6B). Therefore, we conclude that in the absence of a compelling clinical reason to prefer ATG, basiliximab appears to be a reasonable and cost-effective choice for induction in DCD kidney transplant recipients.