Toward equitable transplant protocols: Thymoglobulin vs. Basiliximab in African-American kidney transplantation shows higher infection risk without survival benefit

Abstract

In kidney transplantation, induction therapy is often selected based on perceived immunologic risk for acute rejection, commonly considering African American (AA) ethnicity, panel reactive antibody (PRA) levels > 20%, and retransplantation.

Aims

Compare post-transplant outcomes in African American kidney transplant recipients between two induction agents, rabbit anti-thymocyte globulin (rATG) and basiliximab and followed long-term.

Methods

This single-center retrospective study included 173 AA KTRs followed for a mean of 39.7 ± 26.3 months. Patients received induction with either rATG (n = 71, 41%) or basiliximab (n = 102, 59%). All recipients were maintained on calcineurin inhibitor–based triple immunosuppression. Backward stepwise regression was used to obtain final adjusted models, and time to graft loss was analyzed using Cox proportional hazards regression (STATA 15).

Results

There were no significant differences in overall graft survival (p = 0.75), or death censored graft survival (p=0.28), between the two induction groups. The incidence of first biopsy-proven acute rejection (BPAR) was similar (p = 0.14). Compared with basiliximab, rATG induction significantly increased the risk of viral infections (p = 0.019), bacterial infections (p = 0.025), and overall infections (p = 0.035). Infections were a risk factor for late graft loss (p = 0.039).

Conclusion

Among AA kidney transplant recipients, the use of depleting induction therapy with rATG did not confer benefit for graft or patient survival in the intermediate and long-term but was associated with a significantly higher risk of post-transplant infections. These findings suggest that the use of depleting antibody induction should be carefully balanced against infection risk and that risk stratification based solely on ethnicity should be reconsidered.

Introduction

In 2005, a seminal study suggested that the choice of induction agent for initial immunosuppressive preconditioning in kidney transplantation may not substantially influence long-term graft outcomes among African American (AA) kidney transplant recipients (KTRs). ¹ However, limited long-term data exists to support the preferential use of interleukin-2 receptor antagonist (IL2RA) therapy as the initial induction strategy in this population. In clinical practice, induction therapy is typically selected based on perceived immunologic risk for acute rejection. This risk assessment commonly includes AA ethnicity, panel reactive antibody (PRA) levels, and re-transplantation history. The Kidney Disease: Improving Global Outcomes (KDIGO) guidelines recommend IL2RA, such as basiliximab, as first-line induction therapy, and suggest lymphocyte-depleting agents—such as rabbit anti-thymocyte globulin (rATG)—for recipients considered at high immunologic risk. ² These recommendations, however, were largely derived from studies conducted in earlier eras of maintenance immunosuppression. The use of induction therapy in kidney transplantation has increased from approximately 30% to over 70% in the past decade, contributing to significant improvements in short-term transplant outcomes. ³ lymphocyte-depleting agents—such as rabbit anti-thymocyte globulin (rATG)is a polyclonal antibody directed against human T lymphocytes and remains the most used induction agent in the United States. ⁴ Basiliximab, (IL2RA) is a chimeric mouse-human monoclonal antibody that binds to the α-chain (CD25) of the interleukin-2 receptor on activated T lymphocytes, preventing further immune activation.^5,6

Despite their widespread use, the optimal choice of induction therapy in AA KTRs remains uncertain. This population is often perceived as having a heightened immunological response and an elevated risk of rejection and graft loss, yet direct evidence supporting the routine use of depleting induction based solely on ethnicity is limited.

The present study aims to evaluate whether the type of induction therapy—rATG versus basiliximab—impacts rejection rates, infection risk, and long-term graft survival among AA kidney transplant recipients in the modern era of calcineurin inhibitor (CNI)-based immunosuppression. By comparing these two induction strategies within a racially homogenous cohort, this study seeks to clarify whether ethnicity alone warrants the use of depleting induction agents in AA transplant recipients.

Patients and methods

This was a retrospective cohort study evaluating outcomes among African American (AA) first-time kidney-only transplant recipients who received induction therapy with either rabbit antithymocyte globulin (rATG) or interleukin-2 receptor antagonist (IL2RA) at our center between January 2012 and December 2018. The study protocol was approved by both the Detroit Medical Center and Wayne State University Institutional Review Boards with ethical oversight. (IRB Approval Number: IRB-20-03-1934). A waiver of written informed consent was formally granted in accordance with 45CFR46.116(d). The IRB determined that the research met the criteria for a waiver because it involved no more than minimal risk to participants, did not adversely affect their rights or welfare, and could not practicably be carried out without the waiver. Additionally, a waiver of HIPAA Authorization was granted as the research could not be conducted without access to and use of protected health information.

Eligible participants were self-identified AA kidney transplant recipients aged ≥18 years who received induction with either rATG or IL2RA, based on immunologic risk assessment only at the time of transplantation. Variables collected for analysis included recipient characteristics such as age, sex, race, panel reactive antibody (PRA) level, dialysis duration before transplantation, and comorbidities including diabetes, hypertension, and cause of end-stage renal disease (ESRD). The expected post-transplant survival (EPTS) score was also recorded. ⁷

Donor variables included donor age, sex, human leukocyte antigen (HLA) mismatch, DR match, kidney source (living or deceased), cause of death, donation after cardiac death (DCD) status, CMV and hepatitis serologies, cold ischemia time, and kidney donor profile index (KDPI). Recipients with graft loss from surgical complications, primary nonfunction of unclear etiology, multi-organ transplants, or a history of multiple renal transplants were excluded. Induction therapy groups consisted of either IL2RA, administered as two 20 mg doses on postoperative days 1 and 4 (Group 1), or rATG, administered at a total dose of up to 6 mg/kg (1.5 mg/kg per dose for a total of four doses), adjusted for leukopenia or thrombocytopenia (Group 2). In our center, patients who self-identify as African American are not considered at higher immunologic risk solely based on ethnicity. Instead, induction therapy selection is guided by established clinical risk factors for rejection, including degree of sensitization (e.g., PRA >20%), risk of delayed graft function, and history of transplantation. In accordance with our institutional practice, rabbit anti-thymocyte globulin (rATG) was preferentially used in recipients considered to be at higher immunologic risk, particularly those with higher levels of sensitization. All recipients received standard triple maintenance immunosuppression consisting of a calcineurin inhibitor (CNI), mycophenolic acid, and prednisone. Tacrolimus was initiated after graft recovery—typically on postoperative day 2 or 3—with target trough levels of 10–12 ng/mL for the first six months and 8–10 ng/mL thereafter. Mycophenolic acid was prescribed up to a maximum of 2 g/day, adjusted for gastrointestinal intolerance or leukopenia. Corticosteroid therapy included intravenous methylprednisolone 250 mg on the day of surgery, followed by 125 mg on postoperative days 1 and 2, 80 mg on day 3, and an oral prednisone taper from 20 mg daily, reduced by 5 mg weekly to a maintenance dose of 5 mg daily for the duration of graft function.

For patient’s intolerant to tacrolimus, cyclosporine was substituted.

Antimicrobial prophylaxis included valganciclovir for cytomegalovirus (CMV) prevention—900 mg daily for 6 months (dose-adjusted for renal function) in CMV D+/R− recipients and CMV R+ recipients valcyte dose was 450 mg daily in for 3 months. CMV surveillance was performed using a protocolized approach CMV PCR was obtained monthly for the first 12 months post-transplant, for early detection of breakthrough and post prophylaxis viremia. CMV infection was defined as detection of viral DNA by PCR regardless of symptoms during follow-up. Additional CMV PCR testing was performed when clinically indicated, including in patients presenting with viral-like illness, gastrointestinal symptoms, or leukopenia (particularly neutropenia). Tissue biopsy for CMV disease was performed in selected symptomatic cases at the discretion of the treating physicians. BK virus (BKV) was monitored by surveillance as follows; BKV Urine PCR two weeks post-transplant, then monthly for first twelve months, thereafter BKV Urine PCR was done every six monthly for two years, followed by annually, If BK Viuria is >5 million copies/ml or there is a BKViuria with a rise in serum creatinine, BKV Plasma PCR was obtained. When the BKV plasma PCR became positive, Plasma BKV PCR was obtained every two weeks while actively modifying immunosuppression and then monthly until two consecutive negative results were obtained for two months.

All recipients also received Pneumocystis jirovecii (PJP) prophylaxis with trimethoprim-sulfamethoxazole single-strength (400/80 mg) daily for six months. For those with sulfa allergies, second-line agents were substituted. Prophylaxis was reinitiated if recipients received depleting antibody therapy for treatment of rejection. Diagnosis of PJP pneumonia was based on clinical syndrome and confirmation with PJP PCR in BAL. Antifungal prophylaxis consisted of nystatin suspension 5 ml (100,000 units/ml) PO swish and swallow 4 times daily for one month. The diagnosis of fungal and bacterial infections was based on clinical, microbiologic, and radiographic findings consistent with infection, and confirmed by appropriate culture or tissue biopsy.

DSA was monitored post-transplant weekly for the first 2 weeks, then monthly for the first 12 months. Detection of de novo DSA accompanied by graft dysfunction or rising titers prompted biopsy. In cases of de novo DSA without graft dysfunction, immunosuppression was optimized, and DSA was repeated in one month. IVIG (2 g/kg per month ×3 doses) was considered as clinically indicated. Optimization of immunosuppression included full-dose MPA (total 3 g/day), tacrolimus trough 8–10 ng/mL, and prednisone 5 mg daily.

Recipients with unexplained rises in serum creatinine underwent ultrasound evaluation and allograft biopsy when indicated. All acute rejection episodes were diagnosed by indication biopsy and classified according to the Banff 97 criteria. ⁸ Treatment of rejection followed institutional protocols and included steroid boluses or depleting antibody therapy for steroid-resistant episodes. Maintenance immunosuppression was optimized post-rejection. Concomitant antibody-mediated rejection was managed with plasmapheresis and intravenous immunoglobulin (IVIG) according to center policy. The study outcomes evaluated were overall graft survival, death-censored graft survival, patient survival, time to first biopsy-proven acute rejection (BPAR), and incidence of post-transplant infections (viral, bacterial, or fungal) between induction groups. Graft loss was defined as return to dialysis or death with a functioning graft.

Statistical analysis

Quantitative variables were summarized as means with standard deviations, and categorical variables were presented as frequencies and proportions. Comparisons between induction groups (r ATG vs. IL2RA) were performed using Fisher’s exact test for categorical variables and either the t-test or Wilcoxon rank-sum test for continuous variables, as appropriate.

Cox proportional hazards regression was used to model time to graft loss and time to first rejection. Time to graft loss was dichotomized as ≤12 months versus >12 months. Logistic regression was applied to evaluate predictors of dichotomized graft loss and to model the occurrence of infections. All statistical analyses were performed using STATA version 15 (Stata Corp, College Station, TX, USA).

Results

The demographic characteristics are summarized in Table 1. The median follow-up duration was 64.8 months. Baseline demographic variables were largely similar between the IL2RA (group 1) and rATG (group 2) cohorts. A total of 102 recipients received IL2RA, and 71 recipients received rATG for induction therapy. The causes of ESRD were not significantly different between the two groups: hypertension (n = 39 [38%] vs. n = 32 [45%]), diabetes (n = 30 [29%] vs. n = 20 [28%]), glomerulonephritis (n = 9 [9%] vs. n = 9 [13%]), FSGS (n = 12 [12%] vs. n = 9 [13%]), and other causes (n = 12 [12%] vs. n = 1 [1%]) for group 1 versus group 2, respectively.

Table 1.

Summary of demographic variables by induction agents Thymoglobulin (r ATG) vs Simulect (IL2RA).

Factor	Simulect	Thymoglobulin	p-value
N	102	71	​
AGE at Txp, mean (SD)	49.5 (11.2)	47.6 (10.8)	0.29
Donor AGE, mean (SD)	31.8 (12.5)	33.6 (13.5)	0.36
PEAK PRA, median (IQR)	0.0 (0.0, 0.0)	25.0 (0.0, 79.0)	<0.001
EPTS score >20			0.80
0	27 (26.5%)	20 (28.2%)
1	75 (73.5%)	51 (71.8%)
Cause of ESRD			0.13
HTN	39 (38.2%)	32 (45.1%)
DM	30 (29.4%)	20 (28.2%)
GN	9 (8.8%)	9 (12.7%)
FSGS	12 (11.8%)	9 (12.7%)
Other	12 (11.8%)	1 (1.4%)
Deceased donor			0.67
0	9 (8.8%)	5 (7.0%)	​
1	93 (91.2%)	66 (93.0%)	​
PEAKPRA>50			<0.001
0	97 (95.1%)	48 (67.6%)	​
1	5 (4.9%)	23 (32.4%)
History of first rejection (Y/N)			0.14
0	75 (73.5%)	59 (83.1%)	​
1	27 (26.5%)	12 (16.9%)
Rx Thymoglobulin used for rejection (Y or N)	​	​	0.051
0	91 (89.2%)	69 (97.2%)	​
1	11 (10.8%)	2 (2.8%)
KDPI categorized			0.076
1	32 (35%)	15 (24%)	​
2	41 (45%)	25 (40%)
3	19 (21%)	23 (37%)
DCD transplant			0.36
0	86 (84.3%)	56 (78.9%)	​
1	16 (15.7%)	15 (21.1%)
Viral infection			0.012
0	81 (79.4%)	44 (62.0%)	​
1	21 (20.6%)	27 (38.0%)
Bacterial infection			0.045
0	56 (54.9%)	28 (39.4%)	​
1	46 (45.1%)	43 (60.6%)
Fungal infection			0.17
0	92 (90.2%)	68 (95.8%)	​
1	10 (9.8%)	3 (4.2%)
Total admissions for infections			0.36
0	60 (58.8%)	34 (47.9%)	​
1	22 (21.6%)	19 (26.8%)
2	20 (19.6%)	18 (25.4%)
Cold Time >24			0.62
0	83 (81.4%)	59 (84.3%)	​
1	19 (18.6%)	11 (15.7%)	​
h/o of any infection			0.045
0	44 (43.1%)	20 (28.2%)	​
1	58 (56.9%)	51 (71.8%)
Graft event			0.701
0	83 (81.4%)	56 (78.9%)	​
1	19 (18.6%)	15 (21.1%)
Years on dialysis, median (IQR)	5.0 (2.9, 7.1)	5.4 (3.1, 7.0)	0.65

The mean age at transplant was 49 ± 11 years in the IL2RA group and 47 ± 11 years in the rATG group. Median (Q1, Q3) dialysis vintage was 5.0 (2.9–7.1) years and 5.4 (3.1–7.0) years, respectively. Cold ischemia time >24 hours occurred in 18.6% versus 15.7% of cases, with no significant difference between groups. Other donor-related characteristics were also comparable: donation after cardiac death (15.7% vs. 21.1%), deceased donor transplantation (91.2% vs. 93%), and mean donor age (31.8 ± 12 years vs. 33 ± 13 years) for group 1 and group 2, respectively. As expected, given selection criteria for depleting antibody use, group 2 had a significantly higher proportion of recipients with PRA > 50% (n = 23 [32.4%]) compared with group 1 (n = 5 [4.9%], p < 0.001). Overall, the incidence of first biopsy-proven acute rejection (BPAR) was similar between groups (IL2RA = 27 [27%] vs. rATG = 12 [17%], p = 0.14).

In the multivariable Cox regression model for time to graft loss (Table 2), there was no significant difference in overall graft survival between IL2RA and rATG after adjustment for covariates (HR = 0.88, 95% CI 0.407–1.922, p = 0.75). Compared with hypertension, ESRD due to glomerulonephritis (GN) was associated with a lower risk of overall graft loss (p = 0.03). In the death-censored Cox regression model (Table 3), graft survival remained similar between induction groups (HR = 0.53, 95% CI 0.18–1.56, p = 0.25). Higher donor age at transplant was an independent predictor of inferior death-censored graft survival (p = 0.037).

Table 2.

Adjusted cox regression models for graft loss (N = 169, Failure = 34).

Variable	Hazard ratio	95% CI		P-value
Time on dialysis > 3yrs	0.698	0.267	1.827	0.464
Any infection	0.928	0.368	2.341	0.875
AGE at Transplant	0.990	0.950	1.031	0.623
Deceased Donor	1.273	0.281	5.766	0.754
Induction - Thymoglobulin	0.885	0.409	1.917	0.757
History of first rejection Y-N	1.235	0.507	3.008	0.642
Donor AGE	1.034	1.000	1.068	0.049
DCD	1.399	0.566	3.460	0.467
Cold Time 24	1.543	0.636	3.748	0.338
Cause of ESRD
HTN	Reference
DM	0.458	0.187	1.122	0.088
GN	0.202	0.045	0.904	0.036
FSGS	0.337	0.069	1.642	0.178
Other	0.352	0.044	2.817	0.325
High risk	1.217	0.373	3.966	0.745

Note. Multivariate analysis for overall graft survival.

Table 3.

Adjusted cox regression model for death censored graft loss.

Variable	Hazard ratio	95% CI		P-value
Time on dialysis > 3yrs	1.539	0.343	6.896	0.573
Any infection	0.711	0.208	2.430	0.587
AGE at Txp	0.951	0.902	1.003	0.066
Deceased donor	1.525	0.128	18.175	0.739
Induction Thymoglobulin	0.554	0.190	1.616	0.280
History of first rejection Y-N	0.871	0.277	2.733	0.812
Donor AGE	1.047	1.003	1.094	0.038
DCD	1.421	0.441	4.585	0.556
Cold Time 24	2.348	0.724	7.618	0.155
Cause of ESRD
HTN	Reference
DM	0.557	0.174	1.779	0.323
GN	0.000	.	.	.
FSGS	0.513	0.087	3.023	0.460
Other	0.483	0.055	4.204	0.510
High risk	1.456	0.357	5.934	0.600

Note. Multivariate analysis for death censored graft survival.

The adjusted Cox model for risk of first biopsy proven acute rejection (BPAR) (Table 4) showed a protective effect of rATG induction for first biopsy proven acute rejection (HR = 0.455, 95% CI 0.214–0.968, p = 0.041). However, BPAR was not an independent risk factor for graft loss in multivariable analysis (HR = 0.88, 95% CI 0.28–2.78, p = 0.83). Overall, no significant difference was observed in graft outcomes between groups (p = 0.75). In contrast, infections were significantly associated with late graft loss. Recipients who experienced any infection had approximately a 2.5-fold higher risk of late graft loss (≥12 months post-transplant) after adjustment for covariates (OR = 2.45, 95% CI 1.10–5.44, p = 0.028). Donation after cardiac death was also independently associated with increased risk of graft loss (p = 0.027; Table 5).

Table 4.

Adjusted model for time to first rejection in AA.

Variable	Hazard ratio	95% CI		P-value
Time on dialysis > 3yrs	1.942	0.780	4.835	0.154
Any infection	0.626	0.278	1.405	0.256
AGEatTxp	0.921	0.886	0.958	0.000
Deceased donor	1.299	0.342	4.931	0.701
Induction Thymoglobulin	0.487	0.231	1.027	0.059
Donor AGE	1.031	1.000	1.063	0.047
DCD	0.648	0.230	1.826	0.412
Cold Time 24	2.238	0.883	5.671	0.089
Cause of ESRD
HTN	Reference
DM	2.834	1.128	7.116	0.027
GN	1.031	0.270	3.935	0.965
FSGS	3.821	1.274	11.465	0.017
Other	2.222	0.690	7.150	0.181
High risk	1.877	0.722	4.884	0.197

Table 5.

Adjusted logistic model for time to graft loss with infections (≥12 months vs <12 months) in AA.

Variable	Odds ratio	95% CI		P-value
any infection	2.706	1.051	6.971	0.039
Months on dialysis before transplant	1.008	0.996	1.021	0.182
AGE at Transplant	1.017	0.970	1.066	0.482
Cause ESRD
HTN	Reference
DM	1.141	0.397	3.277	0.807
GN	3.869	0.447	33.483	0.219
FSGS	0.469	0.126	1.741	0.258
Other	2.192	0.240	20.015	0.487
Deceased donor	0.680	0.118	3.914	0.666
Induction Thymoglobulin	0.828	0.335	2.043	0.682
PEAKPRA> 50	1.119	0.319	3.920	0.861
History of first rejection Y-N	2.170	0.659	7.149	0.203
Donor AGE	0.984	0.948	1.021	0.381
DCD	0.378	0.138	1.032	0.058
Cold Time 24	0.881	0.259	3.002	0.840

When compared with IL2RA, rATG induction was associated with a higher likelihood of viral infection (OR = 2.43, 95% CI 1.15–5.09, p = 0.019; Table 6). Similarly, rATG significantly increased the risk of bacterial infection (OR = 2.27, 95% CI 1.10–4.68, p = 0.026; Table 7) and any infection (OR = 2.23, 95% CI 1.07–4.91, p = 0.031; Table 8). Additional independent predictors of any post-transplant infection included history of end stage kidney disease from FSGS (p = 0.016), diabetes mellitus (p = 0.019), GN (p = 0.026), and higher donor age (p = 0.009).

Table 6.

Adjusted logistic model for viral infections.

Variable	Odds ratio	95% CI		P-value
Time on dialysis > 3yrs	0.690	0.296	1.608	0.390
AGE at Txp	1.013	0.972	1.055	0.550
Deceased donor	2.379	0.444	12.749	0.312
Induction Thymoglobulin	2.427	1.155	5.099	0.019
History of first rejection Y-N	1.445	0.580	3.601	0.429
Donor AGE	1.006	0.976	1.037	0.695
DCD	1.210	0.493	2.972	0.677
Cold Time_24	0.816	0.299	2.228	0.692
Cause ESRD
DM	0.992	0.393	2.502	0.986
GN	1.652	0.502	5.434	0.409
FSGS	2.560	0.833	7.866	0.101
Other	0.787	0.146	4.260	0.781
After stepwise selection
Time on dialysis > 3yrs	0.776	0.356	1.692	0.523
Induction Thymoglobulin	2.448	1.224	4.896	0.011

Note. Association between depleting induction therapy and post transplant Viral infection.

Table 7.

Adjusted logistic model for bacterial infections.

Variable	Odds ratio	95% CI		P-value
Time on dialysis > 3yrs	0.720	0.305	1.698	0.453
AGE at Txp	0.997	0.960	1.037	0.896
Deceased donor	0.413	0.099	1.720	0.224
Induction - Thymoglobulin	2.270	1.101	4.680	0.026
History of first rejection Y-N	1.376	0.572	3.309	0.476
Donor AGE	1.021	0.992	1.051	0.152
DCD	2.090	0.835	5.229	0.115
Cold Time_24	2.069	0.754	5.676	0.158
Cause of ESRD
DM	3.958	1.645	9.527	0.002
GN	6.039	1.651	22.086	0.007
FSGS	2.234	0.749	6.662	0.150
Other	2.299	0.598	8.833	0.225
After stepwise selection
Time on dialysis > 3yrs	0.587	0.273	1.261	0.172
Cause of ESRD
HTN	Reference
DM	3.960	1.781	8.808	0.001
GN	6.283	1.787	22.085	0.004
FSGS	2.009	0.731	5.522	0.176
Other	1.666	0.464	5.987	0.434
Induction -Thymoglobulin	2.200	1.104	4.384	0.025

Note. Association between depleting induction therapy and post transplant bacterial infection.

Table 8.

Adjusted logistic model for any infection (Viral, bacterial, or fungal).

Variable	Odds ratio	95% CI		P-value
Time on dialysis > 3yrs	0.690	0.279	1.708	0.423
AGE at Txp	1.005	0.965	1.046	0.814
Deceased donor	0.479	0.102	2.241	0.350
Induction - Thymoglobulin	2.299	1.078	4.902	0.031
History of first rejection Y-N	1.139	0.455	2.850	0.781
Donor AGE	1.031	1.002	1.062	0.038
DCD	1.830	0.687	4.869	0.226
Cold Time 24	2.109	0.700	6.355	0.185
Cause of ESRD
DM	2.920	1.190	7.165	0.019
GN	5.013	1.209	20.774	0.026
FSGS	4.521	1.319	15.491	0.016
Other	2.542	0.682	9.473	0.165
After stepwise selection
Time on dialysis > 3yrs	0.590	0.259	1.344	0.209
Cause of ESRD
HTN	Reference
DM	3.048	1.323	7.023	0.009
GN	5.059	1.246	20.536	0.023
FSGS	4.297	1.331	13.867	0.015
Other	1.991	0.555	7.137	0.291
Induction - Thymoglobulin	2.206	1.055	4.612	0.035
Donor AGE	1.038	1.009	1.066	0.009

During follow-up, there were 284 documented first bacterial infections, 22 fungal infections, and 174 viral infections. Among the 97 early post-transplant bacterial infections, the most common were urinary tract infections (UTI), pneumonia, skin and subcutaneous infections, and urosepsis. Opportunistic infections were rare, except for one case of Actinomyces turicensis thigh abscess. First viral infections were identified through surveillance or clinical suspicion, totaling 88 events—predominantly CMV viremia, herpes zoster, and BK virus infection. Fungal infections (n = 15) were mainly Candida UTIs and superficial wound infections. Infections were a significant determinant of late graft loss in univariate analysis. All abbreviations are listed as supplementary Material.

Discussion

Since 1990, it has been common practice to use induction agents to maximize immunosuppression early after kidney transplantation, until maintenance agents take effect. ⁹ The most used induction agents are rabbit anti-thymocyte globulin (rATG) and interleukin receptor (IL2RA) antagonists. Induction therapy reduces the short-term risk of acute rejection. Induction provides intense immunosuppression at the time of transplantation, bridging the period until maintenance therapy becomes effective.

However, the impact of induction therapy on long-term graft and patient survival in African American (AA) kidney transplant recipients(KTRs) remains unclear. Historically, AA KTRs have been considered at high immunologic risk for graft loss. ¹⁰ It is still uncertain whether induction therapy modifies the risk of inferior long-term outcomes in this group. Considerable debate persists regarding which antibody induction strategy provides the best outcomes for AA KTRs.

In one study comparing three induction agents—alemtuzumab, rabbit antithymocyte globulin, and basiliximab—over a median follow-up of 95 months, the incidence of biopsy-proven acute rejection was similar among all three groups (p = 0.34). ¹¹ More recent data suggest improved patient and graft survival with the use of depleting induction therapies in AAs. ¹² Nevertheless, due to the limited number of AAs enrolled in randomized controlled trials (RCTs), the relative effectiveness of depleting versus non-depleting induction therapy in this population is not well understood.

Ongoing disagreement persists within the transplant community about which induction agent yields the best outcomes in AA KTRs. There is also a paucity of long-term data assessing outcomes in this population within the current era of calcineurin inhibitor (CNI)– and mycophenolate mofetil (MMF)–based maintenance immunosuppression.

In our single-center retrospective analysis of predominantly AA KTRs followed for a median of 64.8 ± 4 months, there was no difference in long-term graft outcomes between those who received basiliximab (IL2RA, Group 1) and those who received rATG (Group 2) (p = 0.75). This finding is consistent with a prior study from the same institution in which AA KTRs were followed for a median of 19 ± 7 months; the type of induction agent (depleting vs. non-depleting) was not associated with risk of acute rejection or graft loss. ¹ Together, these findings suggest that the choice of induction agent may not significantly affect long-term graft outcomes in AA KTRs in the current CNI-based immunosuppression era.

In a large registry study with long-term follow-up, cytolytic induction therapy was associated with reduced rates of graft loss and acute rejection in AA KTRs. ¹⁰ However, baseline immunologic risk differed by induction type: those receiving cytolytic induction had more HLA mismatches, higher PRA, longer cold ischemia time (CIT), more retransplants, donation after cardiac death (DCD), and delayed graft function (DGF). ¹⁰ Another single-center study of 464 KTRs (2006–2015) found no significant difference in death-censored graft survival between AA and Caucasian recipients who received alemtuzumab induction and maintenance with CNI and MPA. ¹⁰ Moreover, AA ethnicity was not an independent predictor of rejection-free or graft survival. The authors concluded that alemtuzumab-based cytolytic induction therapy may help equalize outcomes between White and AA KTRs.

In low immunologic-risk AA KTRs, cytolytic induction therapy may not provide additional benefit. ¹³ A meta-analysis of low-risk KTRs comparing rATG and IL2RA found no difference in acute rejection or overall graft outcomes. ¹⁴

Infections remain a leading cause of graft loss after transplantation. A post hoc analysis from the FAVORIT study of 4,010 KTRs found that post-transplant infections were a leading cause of non-cardiovascular mortality. ¹⁵ While rATG is an effective induction agent, its optimal dose remains uncertain, and over-immunosuppression leading to infection is a key concern. In low-risk, non-AA KTRs, an rATG dose of 3 mg/kg is generally considered safe and comparable to basiliximab. A randomized trial of living donor KTRs found that low-dose thymoglobulin (3 mg/kg) had similar efficacy and safety to IL-2RA induction under standard maintenance therapy, with no difference in infection rates or graft outcomes at one year. ¹⁶

Similarly, Wong et al. compared rATG doses of 3 mg/kg and 4.5 mg/kg, finding both achieved sustained lymphocyte depletion without significant differences in graft outcomes. ¹⁷ Pankewycz et al. reported comparable patient and graft survival between low dose rATG (3.1 mg/kg) and IL2RA in kidney transplant recipients. ¹⁸

High-risk KTRs often receive higher induction doses, ranging from 4.5 mg/kg to 6 mg/kg¹⁹. A meta-analysis of 10 studies (n = 1,065) evaluated rATG doses from 4.5 mg/kg to 10 mg/kg and found that higher doses (> 9 mg/kg) were associated with increase in CMV and bacterial infections compared with 4.5 mg/kg dosing (OR = 1.2; 95% CI, 0.35–4.9) [19]. Overall, rATG doses between 4.5 mg/kg and 9 mg/kg were effective in preventing rejection without significant differences in infection rates. ¹⁹ Based on these findings, our center used thymoglobulin doses of 4.5–6 mg/kg for highly sensitized AA KTRs and basiliximab for low-risk recipients. ¹

We observed a higher risk of viral infections in patients receiving rATG compared to IL2RA (OR = 2.43; CI = 1.15–5.09; p = 0.019). Among these, 27% involved CMV viremia (with or without syndrome) and 48% involved BK viremia. CMV infection risk is influenced by donor–recipient serostatus and prophylaxis duration. It is well established that antilymphocyte antibody therapy increases CMV risk by 2–5-fold compared with basiliximab.^20–23 In this study induction with rATG was also independently associated with higher bacterial infection risk compared to IL2RA (OR = 2.27; 95% CI = 1.10–4.68; p = 0.02). Similarly a prospective, nonrandomized study of 213 KTRs, and followed over three years found higher bacterial infection rates with rATG (47.4%) versus IL2RA (28.4%) (p = 0.006). ²⁴ A meta-analysis also confirmed higher bacterial infection risk in rATG-treated patients compared with IL-2RA. ²⁵ In our adjusted model, overall infections—including viral, bacterial, and fungal—were significantly more common with rATG induction (OR = 2.53; 95% CI = 1.07–4.91; p = 0.03). Additionally we found that diabetes as the cause of ESRD independently increased bacterial infection risk post-transplant (p = 0.019), consistent with similar findings from a large NHS cohort of 19,103 patients showing > 70% higher infection-related mortality in diabetic KTRs. ²⁶

In our study, patients with infections had a higher risk of late graft loss (OR = 2.45; CI = 1.10–5.44; p = 0.039). This association was independent of rejection history, cause of ESRD, or dialysis duration. Similarly, a registry study of 210,327 first-time KTRs (1996–2014) found infection to be the second leading cause of death with a functioning graft within one year post-transplant. ²⁷

Other studies have shown similar results. A prospective RCT of 278 KTRs found that those receiving rATG had significantly more infections than those given IL2RA (p = 0.03) [27-28].

In our study, first biopsy-proven acute rejection (BPAR) rates differed between high immunological risk AA who received rATG and low risk AA who received IL-2RA groups (p = 0.04) but did not translate into long-term outcome differences (p = 0.75). Similar to our study, a RCT similarly found lower early BPAR rates with rATG at 6 months (p = 0.05), though these differences were not significant at 1 year (p = 0.10) ²² and without significant differences in patient or graft survival between groups.^22,28

Interpretation of Conclusions: Taken together, the above studies and our single-center long-term analysis of African American (AA) kidney transplant recipients (KTRs) suggest that in the current calcineurin inhibitor–based maintenance era, AA ethnicity alone may not represent an independent risk factor for graft loss. We also acknowledge the relatively small sample size, and the limited number of graft failure events (n = 34) may reduce the statistical power to detect differences in survival outcomes between the induction groups. Consequently, the absence of statistically significant differences may reflect a potential type II error, whereby modest but clinically meaningful differences could remain undetected.

In our study of primarily AA KTRs, high-dose rabbit anti-thymocyte globulin (rATG) induction was associated with significantly higher infection rates. In addition, infections in our study were associated with an increased risk of long-term graft loss. These findings highlight that the use of high-dose lymphocyte-depleting induction therapy should be carefully balanced against the potential for infection-related complications, which may adversely affect both graft and patient outcomes. While our findings should be interpreted cautiously given the single-center retrospective design, non-randomized treatment allocation, and modest sample size, they suggest that AA status alone may not be sufficient to guide induction therapy decisions. Rather, our results support the importance of individualized risk assessment that incorporates immunologic risk factors, comorbidities, and infection risk when selecting induction strategies. These observations should be considered hypothesis-generating and warrant confirmation in larger, prospective studies.

We acknowledge the limitations of our study due to its retrospective design, which precludes establishing a causal relationship between induction therapy and infection risk. Thus, larger studies with greater numbers of outcome events will be necessary to provide more definitive conclusions. However, the study has notable strengths: all patients were followed within a single transplant center, data were collected consistently by dedicated personnel, and the analysis uniquely focuses on AA KTRs while directly comparing two specific induction strategies.