Impact of Cumulative 6 mg/kg Antithymocyte Globulin on Early Posttransplant Outcomes in Kidney Transplant Recipients with Delayed Graft Function

Abstract

Introduction: Delayed graft function in kidney transplant is associated with an increased risk of rejection and graft loss. Use of rabbit antithymocyte globulin induction in delayed graft function has been correlated with less rejection compared to basiliximab, but optimal dosing remains unknown. Program Evaluation Aims: The purpose of this evaluation was to retrospectively assess the short-term effectiveness and tolerability of a clinical protocol that increased the net state of immunosuppression in delayed graft function kidney transplant recipients using cumulative 6 mg/kg rabbit antithymocyte globulin induction. Design: This retrospective cohort included 88 kidney transplant recipients with delayed graft function, transplanted between January 2017 and March 2021, who either received cumulative 4.5 mg/kg pre-protocol or 6 mg/kg post-protocol rabbit antithymocyte globulin. Outcomes evaluated were biopsy-proven acute rejection and incidence of graft loss, infection, and cytopenia at 6 months. Results: A significant reduction of biopsy-proven acute rejection incidence occurred post-protocol implementation (10/33, 30.3% vs 6/55, 10.9%; P = .04). Of those with rejection, significantly less post-protocol patients were classified as acute cellular rejection (9/10, 90.0% vs 2/6, 33.3%; P = .04). No death-censored graft loss was observed in either group. Rates of cytopenia and infection were similar pre- versus post-protocol implementation. Conclusion: Increasing the exposure to rabbit antithymocyte globulin and maintenance immunosuppression in delayed graft function kidney transplant recipients was tolerable and significantly reduced rejection occurrence at 6 months.

Introduction

Delayed graft function (DGF) is a complication of kidney transplant recipients, where lack of immediate graft function posttransplant leads to the requirement of dialysis within 1 week. ¹ It has been associated with poor clinical outcomes, including up to a 40% increased risk of graft loss and higher first-year acute cellular rejection (ACR) incidence – 16% compared to 10% in non-DGF recipients.^1,2 Common risk factors include donation after cardiac death (DCD), older donor age, cold ischemia time (CIT) > 24 h, Black race, retransplant, and higher kidney donor profile index.^3,4 The incidence in the United States (US) is increasing with the expansion of the donor pool criteria, which includes donors of higher DGF risk. ⁵ This in part is attributable to the growing demands of kidney transplantation and its associated mortality benefit over patients remaining on long-term dialysis. ⁶

Rabbit antithymocyte globulin (rATG), a kidney transplant induction agent comprising 56% of all US transplant inductions, has been correlated with lower rates of biopsy-proven acute rejection (BPAR) compared to basiliximab induction in DGF recipients.^2,7 This mainly is due to its multifaceted mechanism of action in ameliorating ischemia-reperfusion injury, a key process underlying the development of acute tubular necrosis. Specifically, rATG causes T-cell depletion, resulting in a potent immunosuppressive effect and reduced renal epithelial cell damage. Dose-ranging studies have found cumulative rATG dose ≤6 mg/kg to be effective in preventing kidney transplant rejection while minimizing cytopenia and infection. ⁸ However, no available literature specifically evaluates rATG optimal dosing in DGF. A recent systematic review and meta-analysis of kidney recipients demonstrated cumulative rATG dose of ≤4.5 mg/kg to be correlated with significantly less BPAR at 6% per patient-years, compared to 12% in >4.5 mg/kg to 6 mg/kg and 16% in >6 mg/kg. ⁸ This analysis failed to distinguish induction therapy from rejection treatment as part of cumulative rATG dose and did not adjust for patient-specific rejection risk factors. Rates of cytomegalovirus (CMV) and BK viremia were lower with cumulative doses of ≤4.5 mg/kg, while the incidence of graft loss and DGF showed no dose-response correlation. The aim of this evaluation was to assess the effectiveness and tolerability of an increased net state of immunosuppression in adult recipients with DGF, using cumulative 6 mg/kg rATG induction and fixed-dose mycophenolate mofetil (MMF).

Design/Methods

Design and Setting

This was a single-center retrospective cohort study approved by the institutional review board. Electronic health records (EHR) were screened for adult kidney recipients with DGF transplanted between January 2017 and March 2021 at a 700-bed academic medical center affiliated with a transplant center.

Population

The transplant center performed approximately 450 kidney transplants during the evaluation period with majority being deceased donors, receiving lymphocyte-depleting induction, and triple maintenance immunosuppression, including corticosteroid minimization (prednisone 5 mg daily). Population was predominantly Hispanic (73%) and male (60%) with primary kidney disease etiologies of diabetes and hypertensive nephrosclerosis. Additionally, most had a calculated panel reactive antibody <20% with a higher-than-average DGF incidence (20%) during the evaluation period.

Sampling

Recipients with DGF of any race/ethnicity receiving rATG induction between January 2017 and March 2021 were the target population of the program evaluation. Eligible patients were identified through an internal database. Delayed graft function was defined as need for dialysis within 7 days following transplantation. This included patients who required dialysis for hyperkalemia within 24 h posttransplant. Indications for rATG induction included DCD donor, CIT >24 h, or high immunologic risk. High immunologic risk was defined as pre-formed donor-specific antibody (DSA), retransplant, Black race, or positive cross-match. Recipients with DGF were excluded if not receiving rATG or undergoing multiorgan or dual en bloc kidney transplantation.

Data Collection

Patient demographics, baseline information, and outcomes were collected via EHR. Outcomes were evaluated at 6 months posttransplant and included BPAR, all-cause and death-censored graft loss, cytopenia, and infection. Biopsy-proven acute rejection did not include borderline ACR as defined by the Banff criteria. ⁹ Cytopenia criteria were met with any of the following on 2 consecutive instances: white blood cell count <3000/mm³, automated absolute neutrophil count <1000/mm³, or platelet count <50 000/mm³. Types of infection assessed included viral, fungal, and bacterial. CMV viremia was defined as presence of viral replication via plasma polymerase chain reaction (PCR) ≥900 IU/mL, BK viremia as PCR ≥137 IU/mL, and fungal and bacterial as any positive cultures. Other outcomes assessed included all-cause readmission (defined as first readmission incidence posttransplant), time to serum creatinine nadir, and presence of DSA (mean fluorescence intensity defined as weak:1000-2999; moderate: 3000-9999; and strong: ≥10 000) at 6 months.

Protocol Management

Pre-protocol immunosuppression consisted of cumulative 4.5 mg/kg rATG induction and actual body, weight-based MMF dosing (1000 mg twice daily (bid) if ≥80 kg, 750 mg bid if 79–51 kg, and 500 mg bid if ≤50 kg). A DGF protocol, aimed to increase immunosuppression exposure and surveillance, was implemented in July 2019. The protocol included cumulative 6 mg/kg rATG induction, fixed-dose MMF (1000 mg bid), and a renal biopsy 8 to 14 days posttransplant for increased rejection surveillance. Further biopsy was performed on a for-cause basis. Adjusted body weight was used for rATG (actual body weight was used for underweight patients). Doses were rounded to nearest 25 mg with a maximum 150 mg dose cap and held or modified per package insert recommendations. Additional post-protocol rATG dose was administered outpatient between post-operative Day 7–14. Protocol modifications could be made at the discretion of transplant providers based on compelling patient-specific factors. Other maintenance immunosuppression irrespective of DGF protocol included tacrolimus (goal of 8-12 ng/mL) and an early corticosteroid taper (prednisone 5 mg daily by post-operative Day 5).

Institutional monitoring protocols were as follows: weekly complete metabolic panel and complete blood cell count for the first 2 months posttransplant, then every 2 weeks until 4 months posttransplant prior to proceeding to monthly; CMV PCR at 3 and 6 months posttransplant; BK serum PCR at 1 and 6 months posttransplant; DSA at 4 and 6 months posttransplant. Additional laboratory monitoring was performed if clinically indicated.

Data Analysis

Descriptive statistics were used to analyze baseline characteristics. Categorical data were analyzed using chi-squared or two-sided Fisher's exact test, as appropriate. Continuous data were analyzed using logistic regression. Normality of data was assessed using goodness-of-fit test. Tests were statistically significant for P-values <.05. A subset analysis was performed between patients receiving cumulative ≤4.5 mg/kg and >4.5 mg/kg rATG induction for pertinent outcomes.

Procedure

Data were collected using Microsoft Excel with password protection, followed by extraction into JMP 14® (Cary, NC) for data analysis. Demographic data were presented as nominal variables, except age, CIT, kidney donor profile index, calculated panel reactive antibodies, and hospital length of stay reported as continuous variables. Outcomes were presented as nominal variables, except cumulative rATG dose and time to and serum creatinine nadir as continuous variables.

Results

A total of 88 DGF kidney recipients were included, 33 and 55 patients within the pre- and post-protocol groups, respectively. Baseline characteristics were similar between groups with median age 57 [45–65] years and majority being Hispanic (43.2%) males (61.4%), all with a negative crossmatch (Table 1). Total of 27 (49.1%) patients deviated from protocolized management at the discretion of the provider: MMF doses in 11 (20.0%) patients were lowered due to cytopenia; 16 (29.1%) patients received ≤4.5 mg/kg rATG induction post-protocol. Exceptions to cumulative 6 mg/kg rATG included patients with active infection (n = 3), on immunosuppressants at the time of transplant (n = 2), and rATG infusion-related adverse effects (n = 3). The remaining eight patients did not receive cumulative 6 mg/kg induction due to death prior to completing therapy or provider-specific decisions, including dialysis for hyperkalemia and increased risk of COVID-19 during pandemic surges. Two pre-protocol patients received <4.5 mg/kg rATG due to untreated hepatitis C and advanced age. The mean cumulative induction rATG dose was 4.4 mg/kg (0.4) and 5.5 mg/kg (1.0) pre- versus post-protocol, respectively. When including rejection treatment, the mean cumulative rATG dose increased to 5.0 mg/kg (1.6) pre-protocol and remained the same post-protocol, 5.5 mg/kg (1.2).

Table 1.

Baseline Characteristics of Study Sample.

Baseline Characteristics	Pre-Protocol (N = 33)	Post-Protocol (N = 55)	P-Value
	Median [IQR]	Median [IQR]
Age, years median [IQR]	54 [41–66]	50 [51–65]	.11
	N (%)	N (%)
Male	24 (72.7)	31 (54.6)	.12
Race
Hispanic/Latino	18 (54.5)	20 (36.4)	.12
White, Non-Hispanic	9 (27.3)	27 (49.1)	.05
Black, Non-Hispanic	6 (18.2)	7 (12.7)	.54
Asian, Non-Hispanic	0 (0)	1 (1.8)	1.00
Etiology of Kidney Disease
Diabetes	20 (60.6)	24 (43.7)	.07
Other	11 (33.3)	13 (23.6)	.33
Hypertension	2 (6.1)	12 (21.8)	.07
Diabetes & hypertension	0 (0)	6 (10.9)	.08
Donor Type
Donation after circulatory death	17 (51.5)	27 (49.1)	1.00
Deceased donor, non-DCD	16 (48.5)	28 (50.9)	1.00
Cytomegalovirus Risk
Moderate (R+)	21 (63.6)	40 (72.7)	.48
High (D+/R-)	7 (21.2)	10 (18.2)	.78
Low (D-/R-)	5 (15.2)	5 (9.1)	.49
History of prior transplant	3 (9.1)	8 (14.3)	.52
cPRA >80%	3 (9.1)	11 (20.0)	–
	Median [IQR]	Median [IQR]
Cold ischemic time, hours	23.4 [17–36.9]	32 [23.4–38.6]	.05
KDPI, %	52 [30–66]	49 [34–66]	.96
cPRA, %	0 [0–19]	0 [0–30]	.36
Initial length of stay, days	6 [5–8]	6 [5–8]	.39

cPRA = calculated panel reactive antibodies; D = donor; DCD = donation after circulatory death; KDPI = kidney donor profile index; R = recipient

Incidence of Biopsy Proven Acute Rejection and Graft Loss

Significantly less BPAR was observed post-protocol (10/33, 30.3% vs 6/55, 10.9%; P = .04). Substantially more biopsies occurred post-protocol implementation (7/33, 21.2% vs 51/55, 92.7%; P < .01). When excluding patients who did not receive cumulative 6 mg/kg rATG post-protocol, BPAR incidence remained numerically lower but lost statistical difference (10/33, 30.3% vs 5/39, 12.8%; P = .09). The median time to BPAR was 32 days [14–130] and 143 days [49–308] for pre- versus post-protocol, respectively. Of those with BPAR, significantly less post-protocol patients were classified as ACR per Banff criteria (9/10, 90.0% vs 2/6, 33.3%; P = .03). Both post-protocol patients had Banff criteria Grade I ACR, while pre-protocol patients had either Grade I (n = 5) or Grade II (n = 4). Significantly more patients with BPAR were classified as antibody-mediated rejection (AMR) post-protocol (1/10, 10.0% vs 4/6, 66.7%; P = .03) (Figure 1). One post-protocol AMR patient had a history of retransplant and moderate DSA 3 months posttransplant, while two patients had a known history of medication noncompliance. Borderline rejection was seen in 4/33 (12.1%) and 8/55 (14.5%) pre- versus post-protocol, respectively (P = .46). One patient with ACR post-protocol had borderline rejection on surveillance biopsy.

Figure 1.

Banff classification of biopsy-proven acute rejection at 6 months posttransplant.

All-cause graft loss was similar pre- versus post-protocol (0 vs 5/55, 9.1%; P = .15). All post-protocol graft loss was due to patient deaths, four from COVID-19-associated mortality and one of unknown cause. Death-censored graft loss was therefore not observed in either group.

Incidence of Cytopenia and Infection

Rates of cytopenia were not statistically different between groups at 6 months. The most common type was thrombocytopenia (50.0%) pre-protocol and leukopenia (34.8%) post-protocol. When comparing the incidence of cytopenia between patients receiving ≤4.5 mg/kg versus >4.5 mg/kg, no statistical difference was observed; leukopenia 10.6% versus 12.8% (P = .26), neutropenia 0% versus 7.7% (P = .09), thrombocytopenia 12.8% versus 10.3% (P = .75), and multiple 6.4% versus 10.3% (P = .62).

Incidence of infection was also not statistically different pre- versus post-protocol with bacterial infections (20/33, 60.6% vs 38/55, 69.1%; P = .48) being most common (Table 2). More than 90% of bacterial infections were urinary tract infections (UTI). Several patients from both groups (n = 16) experienced multiple, concurrent infections in addition to UTI, including bacteremia, pneumonia, and skin and soft tissue. Five patients experienced non-urinary-related infections. Viral infections included CMV viremia (3/33, 9.1% vs 7/55, 12.7%; P = .74) and BK viremia (4/33, 12.1% vs 11/55, 20.0%; P = .39). Fungal infections (1/33, 3.0% vs 4/55, 7.3%; P = .65) included pulmonary aspergillosis (n = 1) pre-protocol, and COVID-19-associated invasive candidiasis (n = 3) or candidemia (n = 1) post-protocol.

Table 2.

Incidence of Infection at 6 Months Posttransplant.

Infection Type	Pre-Protocol (N = 33)	Post-Protocol (N = 55)	P-Value
	N (%)	N (%)
Bacterial infection	20 (60.6)	38 (70.4)	.48
Urinary tract	10 (50.0)	27 (71.1)	.64
Urinary tract plus other	8 (40.0)	8 (21.0)	.27
Other	2 (10.0)	3 (7.9)	1.00
Cytomegalovirus viremia	3 (9.1)	7 (12.7)	.74
BK viremia	4 (12.1)	11 (20.0)	.40
Fungal infection	1 (3.0)	4 (7.3)	.65

Other Outcomes of Interest

All-cause readmission at 6 months was 69.7% (23/33) pre-protocol and 61.8% (34/55) post-protocol (P = .64). Most frequent cause was renal-related for both groups, 65.2% (15/23) and 70.6% (24/34) pre- versus post-protocol, respectively. Acute rejection accounted for 26.7% (4/15) pre-protocol and 8.0% (2/24) post-protocol in recipients readmitted. Average hospital length of readmission was 4.6 days and 5 days, respectively. Median time to serum creatinine first nadir was 59 days [33–83] and 36 days [22–75] pre- versus post-protocol, respectively. Median serum creatinine nadir was 1.3 [1.2–1.5] and 1.5 [1.2–2.0], respectively. No statistical difference was found between groups for DSA presence at 6 months, however, not every patient had a histocompatibility report available for analysis (1/15, 6.7% vs 4/44, 9.1%; P = 1.00). The pre-protocol patient had no history of pre-formed DSA and presented with weak DSA 1 month posttransplant. Of the post-protocol patients who developed DSA within 6 months posttransplant, 2 had pretransplant DSA (strong and weak). Three developed weak de novo DSA and 1 moderate DSA.

Discussion

This evaluation showed cumulative 6 mg/kg rATG induction and fixed-dose MMF in adult recipients with DGF lowered BPAR incidence without negatively affecting tolerability, even in the setting of increased biopsy rates. Of those with BPAR, post-protocol patients experienced significantly less ACR and lower grading per Banff criteria. Historically, the US Food & Drug Administration approved cumulative rATG 6 mg/kg to 10.5 mg/kg for acute rejection prophylaxis. Although effective, use of higher rATG doses resulted in higher incidences of cytopenia and infection. Smaller dose-ranging studies support cumulative ≤6 mg/kg rATG induction, however optimal dosing in DGF subset has not been analyzed. ¹⁰

Substantial BPAR reduction occurred by more than half post-protocol using cumulative 6 mg/kg rATG induction and fixed-dose MMF in DGF patients at 6 months posttransplant. Not all post-protocol patients received protocolized treatment, introducing a major study limitation affecting internal validity. The authors attempted to address limitations by stratifying patients receiving cumulative ≤4.5 mg/kg versus >4.5 mg/kg rATG. Numerically fewer BPAR were observed in protocolized patients but of no difference upon stratification. Nonetheless, this was limited by a small sample size and could have impacted statistical significance. Additionally, significantly more patients presented with delayed BPAR, representing AMR post-protocol, suggesting an offset in total incidence of rejection that could have been multifactorial. Acute cellular rejection discovered were Grade I and readmissions for treatment of rejection were numerically lower post-protocol. Lastly, this protocol allowed for earlier treatment of less severe rejection as demonstrated by less rATG used post-protocol for rejection. Therefore, a protocolized biopsy for DGF by 14 days posttransplant should be considered given the potential added benefits.

The difference in transplantation timeframe, pre- versus post-protocol, and the occurrence of COVID-19 pandemic post-protocol notably impacted outcomes. With respect to all-cause graft loss, results illustrated an unforeseen majority of COVID-19-associated mortality, although this was not statistically significant. When censored for death, no graft loss occurred pre- or post-protocol. The impact of an increased exposure to induction and maintenance immunosuppression on COVID-19 infection and mortality in DGF is beyond the scope of this evaluation. Recent literature has found lower doses of lymphocyte-depleting agents used for induction therapy amidst the COVID-19 pandemic, which may have increased the incidence of rejection and DGF but with no clear benefits on COVID-19 infection and mortality rates posttransplant. ¹¹

Incidences of cytopenia and infection occurred more frequently post-protocol but did not reach statistical significance. Previous studies have shown rATG >6 mg/kg to be associated with higher rates of infection, such as that of CMV and BK viremia, and cytopenia. ⁸ These patients were also concurrently on other immunosuppressants and cytopenia-inducing agents posttransplant. Cytopenia was further analyzed between patients receiving ≤4.5 mg/kg and >4.5 mg/kg rATG. No statistical difference was observed. In patients who received <6 mg/kg post-protocol, infections still occurred in 75% (12/16) of patients, suggesting rATG dose-response correlation may be more specific to the type of infection. Both bacterial and fungal infections were defined as any positive cultures and did not consider the presence of symptoms. This may have overestimated the overall infection rates in both groups. Further studies are needed to assess the impact of cumulative 4.5 mg/kg versus 6 mg/kg rATG induction on the frequency of infection and cytopenia.

Other limitations include retrospective design and small sample size, leading to potential for confounders. Larger sample sizes in prospective, randomized studies are therefore needed to address confounders. Additionally, outpatient medication adherence for maintenance immunosuppression was unable to be assessed. Lastly, the study duration was limited for post-protocol patients to assess for outcomes >6 months posttransplant.

Conclusion

An increased exposure to induction and maintenance immunosuppression in recipients with DGF significantly lowered the 6-month incidence of BPAR without negatively affecting tolerability. Further prospective, randomized studies of larger sample sizes are needed to assess less frequently occurring variables.