Abstract
Background and objectives
Alemtuzumab is a humanized anti-CD52 monoclonal antibody used as induction in kidney transplantation (KTX) since 2003. Few studies have evaluated long-term outcomes of this agent or changes in outcomes over time.
Design, setting, participants, & measurements
A retrospective cohort study was performed examining United States registry data from 2003 to 2014 of primary KTX recipients receiving induction with alemtuzumab (AZ; n=5521) or antithymocyte globulin (ATG; n=8504) and maintenance immunosuppression with tacrolimus and mycophenolate mofetil and early withdrawal of steroids. The primary outcome was overall death-censored graft survival (DCGS), and secondary outcomes were overall patient survival and 1-year acute rejection. Multivariate models were fit with donor, recipient, and transplant covariates. Because poorer outcomes with AZ may occur from a learning curve impact with the use of a new medication, transplant year was categorized into three time periods to evaluate outcomes over time (2003–2005, 2006–2008, ≥2009), and an interaction term of induction type with transplant year category was included in all models to test for era impacts.
Results
On multivariate analysis of DCGS there was a significant interaction between AZ and era (P<0.001). AZ was significantly associated with inferior DCGS in the earliest 2003–2005 era (adjusted hazard ratio [aHR], 2.21; 95% confidence interval [95% CI], 1.72 to 2.84) but not in the middle 2006–2008 era (aHR, 1.14; 95% CI, 0.96 to 1.36) or the most recent 2009–2014 era (aHR, 1.08; 95% CI, 0.90 to 1.29) compared with ATG. Risk-adjusted patient survival (aHR, 1.32; 95% CI, 1.08 to 1.61; aHR, 1.26; 95% CI, 1.09 to 1.46; and aHR, 1.10; 95% CI, 0.93 to 1.29 by era, respectively) and acute rejection (adjusted odds ratio [aOR], 1.17; 95% CI, 0.96 to 1.42; aOR, 0.94; 95% CI, 0.82 to 1.07; aOR, 0.89; 95% CI, 0.81 to 0.98 by era, respectively) with AZ was comparable with ATG in the most recent era; however, there was no significant interaction with time (P=0.13 and P=0.06, respectively).
Conclusions
Current alemtuzumab utilization is associated with comparable graft and patient survival and acute rejection compared with ATG. Graft survival with alemtuzumab has improved over time.
Keywords: renal transplantation, acute allograft rejection, chronic graft deterioration
Introduction
For nearly three decades there has been interest in developing kidney transplant immunosuppression protocols which avoid or quickly eliminate exposure to steroids to avoid the harmful sequelae of prolonged steroid therapy. Whereas successes with early steroid withdrawal have been limited by the risk of acute rejection and the perceived or actual reduction of long-term outcomes, some recent trials have been published with encouraging results with early graft survival and acute rejection incidence similar to steroid-based protocols, especially when concomitant antibody induction is administered (1), dosages of cyclosporine are high (2,3), or maintenance therapy consists of tacrolimus and/or mycophenolate mofetil (4,5).
Over the last decade, alemtuzumab (AZ) (Campath-1H) has been used in kidney transplant immunosuppressive protocols at a number of centers (5–14). Various single and multicenter studies, both randomized and nonrandomized, are reasonably consistent in demonstrating the effectiveness of AZ as an induction or preconditioning agent in renal transplant recipients usually in the context of steroid avoidance or early withdrawal. However, few studies have reported long-term outcomes of AZ, most are limited to single-center experiences and small sample sizes, even the largest studies lack power to demonstrate meaningful impacts on graft survival, and few have compared AZ with other agents using the same maintenance treatment.
The long-term efficacy of AZ in the context of early steroid elimination remains unclear. We aimed to more comprehensively investigate the impact of AZ with early steroid elimination using registry data of kidney transplant recipients from 2003 to 2014, the time period in which AZ was used and an era characterized by the predominant use of tacrolimus and mycophenolate as integral parts of maintenance immunosuppression. We also sought to determine if a learning curve impact with this new agent occurred over time.
Materials and Methods
Sources of Data
We used data from the Scientific Registry of Transplant Recipients (SRTR). The SRTR data system includes data on all donors, wait-listed candidates, and transplant recipients in the United States, submitted by the members of the Organ Procurement and Transplantation Network. The Health Resources and Services Administration, US Department of Health and Human Services provides oversight to the activities of the Organ Procurement and Transplantation Network and SRTR contractors. This study was approved by the Institutional Review Board of the Albert Einstein College of Medicine. The clinical and research activities being reported adhere to the Declaration of Helsinki and are consistent with the principles as outlined in the Declaration of Istanbul on Organ Trafficking and Transplant Tourism.
Study Population
SRTR data were accessed to identify all adult, first-time, deceased-donor kidney-only recipients from 2003 to 2014. These years coincided with the utilization of AZ as induction therapy in the United States. Exclusions were donor age <6 years and recipient previous organ transplant, hepatitis C or HIV seropositive, and research study participation. Patients were selected if administered induction therapy exclusively with AZ or antithymocyte globulin (ATG) and received maintenance immunosuppression with tacrolimus and mycophenolate mofetil without steroids at the time of hospital discharge. Patients receiving other types of maintenance immunosuppression, multiple induction agents, and those with incomplete information on maintenance immunosuppressive drugs at discharge were excluded. Patients who experienced allograft failure before discharge were excluded because the United Network for Organ Sharing ascertains the primary maintenance immunosuppressive regimen at the time of discharge from the index transplant hospitalization. This resulted in a final analytic sample size of 14,025 patients. To manage potential bias because of changes in therapy in response to clinical course, we used an intention-to-treat analysis that maintained individuals in their initially assigned treatment group throughout follow-up.
Outcomes
The primary outcome was time to death-censored graft survival (DCGS; defined as return to dialysis or retransplantation). Secondary outcomes were patient survival after transplantation and 1-year acute rejection (indicated by reported acute rejection or treatment for acute rejection in the follow-up forms at 3, 6, or 12 months).
Covariates
We included the following covariates in the multivariable models: recipient age (continuous); sex; race (black, other); diabetes mellitus (defined as diabetes mellitus as primary diagnosis or by history); polycystic kidney disease; duration of maintenance dialysis before transplantation (<3 years, ≥3 years, missing); number of human leukocyte antigen-A, human leukocyte antigen-B, and DR mismatches (HLAMM) (≤3, >3); panel-reactive antibody level (>30%, ≤30%, missing); body mass index (<18.5, 18.5–30, 31–40, >40 kg/m2, missing); cold ischemia time (≤24 hours, >24 hours, missing); insurance status (nonprivate, other); comorbidity (drug-treated chronic obstructive pulmonary disease, peptic ulcer disease, angina, cerebrovascular disease, peripheral vascular disease, or malignancy); transplantation year category (2003–2005, 2006–2008, ≥2009); and donor/recipient weight ratio. Also included were donor age (6–17, 18–39, 40–59, ≥60 years), sex, race (black, other), cause of death (cerebrovascular accident, other), hypertension diabetes mellitus, terminal serum creatinine >1.5 mg/dl, and donation after circulatory death. Body mass index was calculated as weight (kg)/height (m)2; outliers (<10 and >60) were coded as missing. The appropriate functional form of model covariates was determined by exploratory data analysis in unadjusted models and perceived impact on clinical meaningfulness. To evaluate outcomes over time, transplant year was categorized a priori into three time periods, with the first time period chosen as 2003–2005 because AZ use within the study cohort was <30% during this time, the second time period being the next 3 years in sequence (2006–2008), and the last time period including the remaining years of the study (2009–2014).
Statistical Analyses
Univariate associations between exposure groups were examined using the chi-squared tests for categorical variables and t tests for continuous variables whose distributions approximated normality. Survival distributions for mortality and graft failure were examined with Kaplan–Meier curves and compared using the log-rank test. Cox proportional hazards models for DCGS and patient survival were fit to estimate hazard ratios (HR) and 95% confidence intervals (95% CI) for exposure groups after accounting for potential confounders. For variables that had missing data >1%, a missing category was created to conduct main multivariable analyses; complete case analysis was also conducted in sensitivity analyses. Other sensitivity analyses for the primary end point included (1) restriction to a maximum of 3 years follow-up, (2) adjustment of the SEM of the HR (sandwich estimator) to account for dependence of observations derived from the kidneys from the same center, and (3) inclusion of only centers with a high volume of steroid withdrawal patients. We additionally examined the utility of propensity or a mated analysis to address the potential of confounding by indication. The odds of known acute rejection at 1 year were assessed using logistic regression, including all covariates, and only included patients with at least 1 year of graft survival on the basis of their status as being at risk for these events. Acute rejection data were missing or unknown in 19 patients; these patients were not included in the analysis, resulting in a final analytic size of 10,793.
All multivariate models were additionally fit with an interaction term of induction type with transplant year category. Exposure groups and all other covariates were examined for adherence to the proportional hazard assumption. No important departures from proportionality were observed. Ties in the failure time were handled using the Breslow method. Time to outcome was defined as time from the date of transplant until date of outcome (death or graft failure), censored for loss to follow-up and end of study period (October 2, 2014).
All statistical analyses were conducted using the SAS system version 9.2 (SAS Institute, Cary, NC). Statistical significance was identified by a P value of <0.05, and all confidence intervals also used a 95% threshold. All P values are two-sided.
Results
We identified 14,025 eligible patients transplanted from 201 centers between 2003 and 2014 who received maintenance with tacrolimus and mycophenolate mofetil at transplant discharge and induction with either AZ (n=5521) or ATG (n=8504). Patients receiving AZ were more likely to be younger, black, nonprivately insured, obese, sensitized (panel-reactive antibody >30%), and with a pretransplant dialysis duration >3 years. They were less likely to be male, have diabetes mellitus, have a comorbidity, or receive a poorly matched kidney (HLAMM>3). Donor characteristics were similar except that AZ recipients were less likely to receive imported kidneys and more likely to receive kidneys from donors that were black and with hypertension (Table 1). The between-group differences in the case mix did not change substantively across eras (Table 2). At 1-year follow-up, 77.9% remained on two maintenance medications, and 17.4% were receiving steroids. Steroid use at 1 year was significantly lower in the AZ group (9.6%) compared with the ATG group (20.0%) within the earliest era only. Between-group differences in maintenance medication number were significantly weighted toward a lower medication number in the AZ groups within all eras.
Table 1.
Donor and recipient characteristics by induction with alemtuzumab or antithymocyte globulin
| Characteristic | AZ (n=5521) | ATG (n=8504) | P |
|---|---|---|---|
| Donor age, y | |||
| 6–17 | 7.3 | 7.5 | 0.45 |
| 18–39 | 36.8 | 37.4 | |
| 40–59 | 45.2 | 45.2 | |
| ≥60 | 10.7 | 9.9 | |
| Donor black | 16.7 | 12.1 | <0.001 |
| Donor male | 58.7 | 59.1 | 0.61 |
| Donor hypertension | 32.4 | 30.4 | 0.02 |
| Donor diabetes | 8.5 | 8.1 | 0.43 |
| Donor death because of CVA | 38.0 | 37.0 | 0.23 |
| Donor DCD | 14.2 | 14.2 | 0.98 |
| Donor serum Cr>1.5 mg/dl | 17.8 | 18.2 | 0.47 |
| Donor/recipient weight ratio | 1.02±0.37 | 1.03±0.37 | 0.15 |
| Donor, nonlocal | 21.9 | 27.5 | <0.001 |
| Recipient age, y | 52.8±12.5 | 54.3±13.0 | <0.001 |
| Recipient diabetes mellitus | 30.5 | 32.2 | 0.03 |
| Recipient ESRD because of PCKD | 10.4 | 10.7 | 0.55 |
| Recipient black | 35.8 | 25.1 | <0.001 |
| Recipient male | 61.0 | 63.0 | 0.02 |
| Recipient comorbidity present | 15.9 | 19.6 | <0.001 |
| Recipient dialysis >3 y | 58.4 | 51.4 | <0.001 |
| Recipient BMI, kg/m2 | |||
| <18.5 | 1.4 | 1.6 | <0.001 |
| 18.5–30 | 57.9 | 61.9 | |
| 31–40 | 37.4 | 34.2 | |
| >40 | 3.4 | 2.3 | |
| Recipient PRA >30% | 19.5 | 15.6 | <0.001 |
| Recipient CIT >24 h | 19.7 | 22.6 | <0.001 |
| Recipient HLAMM >3 | 71.1 | 73.3 | <0.01 |
| Recipient insurance, nonprivate | 74.0 | 71.2 | <0.001 |
| 1-y follow-up maintenance steroidsa | 16.3 | 18.1 | 0.02 |
| 1-y follow-up number of maintenance medicationsa | |||
| 1 maintenance medication | 3.4 | 1.8 | <0.001 |
| 2 maintenance medications | 78.9 | 77.3 | |
| 3 maintenance medications | 16.5 | 18.6 | |
| 4–6 maintenance medications | 1.2 | 2.3 |
Values are mean±SD, percentage, or as otherwise indicated. CVA, cerebrovascular accident; DCD, donation after circulatory death; Cr, creatinine; PCKD, polycystic kidney disease; BMI, body mass index; PRA, panel reactive antibody; CIT, cold ischemia time; HLAMM, human leukocyte antigen mismatch; AZ, alemtuzumab; ATG, antithymocyte globulin.
Only includes recipients with 1-year graft survival.
Table 2.
Donor and recipient characteristics by induction with alemtuzumab or antithymocyte globulin for each era
| Characteristica | 2003–2005 | 2006–2008 | 2009–2014 | |||
|---|---|---|---|---|---|---|
| AZ | ATG | AZ | ATG | AZ | ATG | |
| Donor age, y | b | |||||
| 6–17 | 7.4 | 11.4 | 8.8 | 7.4 | 6.7 | 6.8 |
| 18–39 | 30.7 | 36.3 | 37.6 | 37.2 | 37.2 | 37.3 |
| 40–59 | 47.2 | 40.4 | 42.1 | 44.8 | 46.1 | 46.4 |
| ≥60 | 14.7 | 11.9 | 11.5 | 10.7 | 10.0 | 9.1 |
| Donor black | 14.3 | 12.7 | 14.7 | 12.3 | 17.7b | 12.0 |
| Donor male | 56.8 | 57.8 | 58.1 | 60.8 | 59.0 | 58.5 |
| Donor hypertension | 29.4 | 28.8 | 30.2 | 29.3 | 33.6 | 31.2 |
| Donor diabetes | 7.6 | 5.6 | 7.4 | 8.5 | 9.0 | 8.4 |
| Donor death because of CVA | 49.6b | 43.6 | 40.2 | 39.0 | 35.9 | 34.7 |
| Donor DCD | 7.9 | 7.1 | 12.7 | 10.9 | 15.5b | 17.2 |
| Donor serum Cr>1.5 mg/dl | 17.0 | 14.2 | 18.9 | 18.6 | 17.4 | 18.8 |
| Donor/recipient weight ratio | 1.04±0.40b | 1.03±0.36 | 1.01±0.36b | 1.04±0.37 | 1.02±0.37 | 1.02±0.38 |
| Donor, nonlocal | 31.2 | 28.4 | 32.3 | 34.6 | 17.0b | 24.0 |
| Recipient age, y | 53.6±12.1 | 52.5±13.1 | 53.6±12.6b | 53.9±12.8 | 52.8±12.4b | 54.9±13.0 |
| Recipient diabetes mellitus | 38.1 | 33.8 | 31.9 | 31.7 | 29.1b | 32.1 |
| Recipient ESRD because of PCKD | 9.8 | 9.8 | 9.2b | 11.5 | 10.9 | 10.5 |
| Recipient black | 26.8 | 24.6 | 35.8b | 24.3 | 36.8b | 25.6 |
| Recipient male | 56.8 | 60.0 | 60.5 | 63.4 | 61.7b | 63.3 |
| Recipient comorbidity present | 27.8 | 24.6 | 19.5b | 23.3 | 13.3b | 16.9 |
| Recipient dialysis >3 y | 58.4 | 50.0 | 55.9b | 43.3 | 59.2b | 55.6 |
| Recipient BMI, kg/m2 | b | b | ||||
| <18.5 | 3.7 | 2.9 | 1.4 | 1.3 | 1.1 | 1.5 |
| 18.5–30 | 62.9 | 65.2 | 60.6 | 64.6 | 56.4 | 60.0 |
| 31–40 | 29.5 | 29.8 | 33.8 | 32.2 | 39.5 | 35.9 |
| >40 | 3.9 | 2.0 | 4.1 | 2.0 | 3.0 | 2.6 |
| Recipient PRA >30% | 22.0b | 15.7 | 18.5b | 12.3 | 19.5b | 17.3 |
| Recipient CIT >24 h | 29.2b | 20.1 | 28.3b | 22.0 | 15.5b | 23.4 |
| Recipient HLA MM >3 | 61.9b | 67.6 | 67.4 | 69.3 | 73.4 | 76.2 |
| Recipient insurance, nonprivate | 69.3 | 70.7 | 69.5 | 67.2 | 76.3b | 73.1 |
| 1-y follow-up maintenance steroids | 9.6b | 20.0 | 16.1 | 17.1 | 17.4 | 18.2 |
| 1-y follow-up number of maintenance medications | ||||||
| 1 maintenance medication | 3.6b | 1.2 | 5.8b | 1.8 | 2.2b | 2.1 |
| 2 maintenance medications | 85.0b | 77.4 | 77.8b | 80.4 | 78.5b | 75.2 |
| 3 maintenance medications | 10.4b | 20.2 | 15.8b | 16.8 | 17.8b | 19.4 |
| 4–6 maintenance medications | 1.1b | 1.2 | 0.6b | 1.1 | 1.5b | 3.3 |
Values are mean±SD or percentages. CVA, cerebrovascular accident; DCD, donation after circulatory death; Cr, creatinine; PCKD, polycystic kidney disease; BMI, body mass index; PRA, panel reactive antibody; CIT, cold ischemia time; HLAMM, human leukocyte antigen mismatch; AZ, alemtuzumab; ATG, antithymocyte globulin.
Data not shown for recipients missing information on BMI (n=690), PRA (n=290), CIT (n=278), HLAMM (n=2), dialysis duration (n=94), donor creatinine (n=16), donor DCD (n=1), donor/recipient weight ratio (n=130), and 1-year follow-up form immunosuppression medication (n=174).
Statistically significant difference between the AZ and ATG groups (P<0.05).
DCGS and Patient Survival
DCGS time to event curves of AZ versus ATG for the entire cohort and within each of the three eras are provided in Figure 1. On multivariate analysis the interaction term between AZ and era was significant for DCGS (P<0.001). AZ was significantly associated with inferior DCGS in the earliest 2003–2005 era (adjusted hazard ratio [aHR], 2.21; 95% CI, 1.72 to 2.84; P<0.001) but not in the middle 2006–2008 era (aHR, 1.14; 95% CI, 0.96 to 1.36; P=0.12) or the most recent 2009–2014 era (aHR, 1.08; 95% CI, 0.90 to 1.29; P=0.43) compared with ATG (Table 3). Risk-adjusted patient survival was inferior with AZ relative to ATG in the earlier two eras (aHR, 1.32; 95% CI, 1.08 to 1.61 and aHR, 1.26; 95% CI, 1.09 to 1.46, respectively) and was comparable in the most recent era (aHR, 1.10; 95% CI, 0.93 to 1.29); however, there was no significant interaction with time (P=0.13) (Table 4).
Figure 1.
Kaplan-Meier plots of overall death-censored graft survival by induction with alemtuzumab (AZ) or anti-thymocyte globulin (ATG). (A) All years, (B) years 2003–2005, (C) years 2006–2008, and (D) years 2009–2014.
Table 3.
Multivariable Cox model of overall death-censored graft survival
| Parameters (Reference Group) | Adjusted Hazard Ratio (95% CI)a |
|---|---|
| Donor age, y (18–39) | Reference |
| 6–17 | 0.99 (0.78 to 1.27) |
| 40–59 | 1.35 (1.17 to 1.56) |
| ≥60 | 1.93 (1.58 to 2.36) |
| Donor, black (other) | 1.32 (1.14 to 1.53) |
| Donor, sex male (female) | 0.85 (0.76 to 0.95) |
| Donor, history of hypertension | 1.23 (1.08 to 1.41) |
| Donor, history of diabetes mellitus | 1.44 (1.20 to 1.72) |
| Donor, death because of cerebrovascular accident | 1.05 (0.92 to 1.19) |
| Donor, donation after circulatory death | 1.12 (0.94 to 1.33) |
| Donor, terminal serum creatinine >1.5 mg/dl | 1.10 (0.95 to 1.28) |
| Donor: recipient weight ratio | 0.81 (0.67 to 0.97) |
| Donor, nonlocal | 1.16 (1.01 to 1.34) |
| Recipient, age, continuous per year | 0.98 (0.98 to 0.99) |
| Recipient, diabetes mellitus | 1.04 (0.92 to 1.18) |
| Recipient, ESRD because of PCKD (other) | 0.67 (0.53 to 0.84) |
| Recipient, race black (other) | 1.32 (1.17 to 1.48) |
| Recipient, sex male (female) | 1.19 (1.05 to 1.35) |
| Recipient, comorbidity present (not present) | 1.01 (0.87 to 1.16) |
| Recipient, dialysis time pretransplant >3 y (≤3) | 1.03 (0.91 to 1.16) |
| Recipient BMI, kg/m2 (18.5–30) | |
| <18.5 | 1.02 (0.63 to 1.66) |
| 31–40 | 1.19 (1.05 to 1.36) |
| >40 | 1.31 (0.96 to 1.78) |
| Recipient PRA >30% (PRA≤30) | 1.22 (1.05 to 1.42) |
| Recipient, cold ischemia time ≥24 h (<24 h) | 1.09 (0.94 to 1.26) |
| Recipient HLA, mismatches >3 (≤3) | 1.37 (1.19 to 1.58) |
| Recipient insurance, nonprivate | 1.31 (1.14 to 1.50) |
| Transplant year 2003–2005×AZ (ATG) | 2.21 (1.72 to 2.84) |
| Transplant year 2006–2008×AZ (ATG) | 1.14 (0.95 to 1.36) |
| Transplant year 2009–2014×AZ (ATG) | 1.08 (0.90 to 1.29) |
PCKD, polycystic kidney disease; BMI, body mass index; PRA, panel reactive antibody; HLA, human leukocyte antigen; 95% CI, 95% confidence interval.
For death-censored graft survival, 13,876 observations out of 14,025 were used in the model.
Table 4.
Adjusted outcomes by induction group
| Clinical Outcome | Alemtuzumab | ATG |
|---|---|---|
| Overall patient survival, aHR (95% CI) | 1.32 (1.08 to 1.61) | Reference group |
| Years 2003–2005 | ||
| Overall patient survival, aHR (95% CI) | 1.26 (1.09 to 1.46) | Reference group |
| Years 2006–2008 | ||
| Overall patient survival, aHR (95% CI) | 1.10 (0.93 to 1.29) | Reference group |
| Years 2009–2014 | ||
| 1-y acute rejection, aOR (95% CI) | 1.17 (0.96 to 1.42) | Reference group |
| Years 2003–2005 | ||
| 1-y acute rejection, aOR (95% CI) | 0.94 (0.82 to 1.07) | Reference group |
| Years 2006–2008 | ||
| 1-y acute rejection, aOR (95% CI) | 0.89 (0.81 to 0.98) | Reference group |
| Years 2009–2014 |
For overall patient survival 13,752 observations out of 14,025 were used in the model. aHR, adjusted hazard ratio; 95% CI, 95% confidence interval; aOR, adjusted odds ratio; ATG, antithymocyte globulin.
Acute Rejection within 1 Year
The overall prevalence of 1-year acute rejection was 8.1% in the AZ group (11.4%, 6.9%, and 8.2% by era) and 8.5% in the ATG group (9.0%, 7.3%, and 9.2% by era). On multivariable analysis, the interaction term between AZ and era for the outcome of acute rejection was not significant (P=0.06). AZ was associated with comparable acute rejection in the first two eras (aOR, 1.17; 95% CI, 0.96 to 1.42 and aOR, 0.94; 95% CI, 0.82 to 1.07) and a lower adjusted hazard of acute rejection in the most recent era (aOR, 0.89; 95% CI, 0.81 to 0.98) relative to ATG. Of recipients who developed acute rejection, unadjusted DCGS was significantly worse in the AZ group for the entire cohort (P<0.001) and in the earliest (P<0.001) and middle (P<0.01) eras but not the most recent era (P=0.46) (Figure 2).
Figure 2.
Kaplan-Meier plots of overall death-censored graft survival amongst recipients who developed Acute Rejection by induction with alemtuzumab (AZ) or anti-thymocyte globulin (ATG). (A) All years, (B) years 2003–2005, (C) years 2006–2008 and (D) years 2009–2014.
Sensitivity Analyses
Additional analyses were performed to evaluate the robustness of the primary results. The results did not change in magnitude or direction, with exclusion of dependent variable patients with missing data, restriction of the multivariate model to a maximum of 3 years follow-up, or adjustment of the SEM of the HR (sandwich estimator) to account for dependence of observations derived from the kidneys from the same center.
To test whether the poorer outcomes with AZ may be caused by a center volume impact, we performed a subanalysis restricting the original cohort to include only centers with the most early steroid withdrawal experience, that is, the top 90th percentile of transplant centers by volume of steroid withdrawal patients (range, 1–828; 90th percentile, 213 patients). The subset analysis of 21 centers, all of which performed at least 213 kidney transplantations under AZ or ATG induction without steroids (8850 patients), recapitulated the DCGS findings (early era: aHR, 2.38; 95% CI, 1.71 to 3.31; middle era: aHR, 1.10; 95% CI, 0.89 to 1.37; recent era: aHR, 0.89; 95% CI, 0.71 to 1.12).
To address potential confounding by indication, we examined the utility of a propensity score to summarize the predicted probability of being prescribed AZ using a logistic regression model that included all study covariates and center identification with the aim of balancing the distributions of patient characteristics across nonusers versus users of AZ. However, the top probability was 55%, and most of the probability of receiving AZ was explained by transplant year and center identification. Therefore, because most of the probability of AZ use was on the basis of transplant year and center but not patient characteristics, a propensity analysis on the basis of patient characteristics was not helpful. For the same reason (center clustering of AZ), a mated analysis was not possible because of the paucity of mate kidneys that did not receive AZ.
Discussion
This analysis of adult primary kidney transplant recipients in the United States under tacrolimus- and mycophenolate mofetil–based immunosuppression without steroids by time of transplant hospitalization discharge shows that current induction treatment with AZ is similar to that of thymoglobulin in terms of graft and patient survival and is somewhat superior to thymoglobulin for acute rejection, as reflected by the most recent era in our analysis. Graft survival with AZ was associated with poor early outcomes, which improved over time.
Our early era results are similar to previous registry reports of kidney transplants performed during a similar time period (before 2007), which demonstrated 1-year acute rejection to be similar with AZ compared with IL2-RAb despite fewer maintenance immunosuppression medications administered (6). At approximately 4-years follow-up, DCGS either trended worse (7) or was significantly worse (6). The poorer early results with AZ have been attributed to a learning curve with the use of this medication because in early trials this agent was commonly used with irrational exuberance in terms of immunosuppressive drug withdrawal (15). Our finding that the AZ recipients were receiving fewer immunosuppression medications at follow-up and had poorer graft survival after development of acute rejection (except in the most recent era) support this assertion. Although we adjusted for donor and recipient confounders and restricted the analysis to patients receiving the same maintenance therapy, intrinsic biases inherent in the comparison of medication regimens cannot be completely accounted for. In particular, when there are known or perceived differences in the efficacy of medication regimens, the assignment of patients to regimens and the adjustment of medication dosage are often not random and cannot be completely accounted for with statistical adjustment (16). Hence, in this analysis, which reflects the common practice of AZ usage in the United States, the poorer outcomes in the early era of AZ usage is possibly caused by underimmunosuppression rather than a diminished effectiveness of AZ.
Our recent era results are consistent with most data from prospective randomized trials wherein outcomes with AZ have been shown to be optimal when maintenance immunosuppression medications are the same as in control arms or at least both groups are exposed to similar calcineurin inhibitor levels. Most of the randomized trials comparing AZ with other induction agents (IL2R-Ab or ATG) or no induction have reported equal or lower biopsy-proven acute rejection (BPAR) in the AZ arms, with no unfavorable impact on early-term renal function or graft survival when both groups are exposed to similar calcineurin inhibitor levels. Margreiter et al. (8) found that AZ and tacrolimus monotherapy resulted in somewhat lower BPAR relative to AZ with dual/triple immunosuppression (20%/65 versus 32%/66, respectively; P=0.09). Another study by Welberry Smith et al. (9) using AZ and tacrolimus monotherapy also found lower BPAR relative to basiliximab, tacrolimus, and mycophenolate mofetil (10%/58 versus 24%/58; P=0.05). Cherukuri et al. (10) found that AZ resulted in significantly less BPAR compared with basiliximab with both groups having similar tacrolimus levels and mycophenolate mofetil dosage (10.3%/51 versus 24.1%/45, P<0.05). Farney et al. (11) demonstrated significantly lower acute rejection rates among recipients of AZ relative to ATG with triple maintenance therapy and early steroid withdrawal (16%/113 versus 26%/109; P=0.02). Although follow-up in most studies has been short, a prospective randomized study by Hanaway et al. (12) demonstrated optimal 3-year DCGS and BPAR in AZ-receiving cohorts on the same maintenance therapy with tacrolimus, MMF, and early steroid withdrawal as control groups. The study showed similar 1-year BPAR in high-risk recipients compared with ATG induction with the same maintenance therapy (10%/139 versus 13%/69, P=0.53) and lower 1-year BPAR in low-risk recipients compared with basiliximab (5%/335 versus 17%/335, P<0.001). At 3 years, DCGS was also similar in the high-risk (91.3% versus 91%) and low-risk groups (97% versus 94%) receiving AZ relative to control medications, respectively. A large retrospective analysis also found lower acute rejection rates and better early graft survival when full doses of tacrolimus and mycophenolate mofetil were administered in the context of steroid avoidance or early withdrawal (13); however, others have not (17). The prospective randomized trial that showed higher BPAR and lower graft survival in the AZ arm had minimized tacrolimus and mycophenolate mofetil only among those receiving AZ (14). In fact, through subset analysis, the authors found that the significantly poorer renal function observed in the AZ group was specifically a result of early underimmunosuppression of maintenance agents (14).
AZ usage in the United States is approximately 13% of all kidney transplants (18). The utilization of this agent may increase especially if it continues to be free or less expensive than alternatives. AZ is currently offered free of charge by Sanofi through its Campath Distribution Program. Given the unmet need to define optimal induction regimens in kidney recipients and the growing interest in AZ, our finding of acceptable long-term results with AZ is important.
Our results are subject to the limitations inherent in observational data. Because kidney transplant recipients are typically not randomly selected to receive specific types of immunosuppressive therapy, it is likely that certain groups in some unmeasured way may be systemically less (or more) healthy than those that received other types of therapy. Nevertheless, we found that transplants performed with AZ induction were somewhat more commonly associated with traditional risk factors for graft loss, and these differences in clinical profiles between treatment groups did not change by era. There is the possibility for residual confounding as a result of clinical or demographic factors not included in the analysis, such as some donor biopsy characteristics, transplant technique, and center-specific treatment protocols. Because induction type is reported on follow-up forms after discharge, ascertainment bias may exist because of exclusion of patients with graft failure or death before discharge. However, the proportion of patients excluded for this reason was small (3.5%). These data do not include information on dose administration, drug level, immunosuppression changes, or timing of steroid withdrawal, which may significantly modify the impacts of specific agents. Potential issues relating to the determination of acute rejection include errors of sampling or technique, subjective interpretation, and ascertainment bias.
Our findings suggest that in the recent era patients receiving AZ and maintenance immunosuppression with tacrolimus, mycophenolate mofetil, and steroid early withdrawal had comparable results with those receiving ATG and the same maintenance regimen in terms of graft and patient survival and improved results for acute rejection. Graft survival with the use of AZ has improved over time, suggesting a learning curve impact with the use of this agent.
Disclosures
Dr. Kayler and Patricia Friedman received support from Astellas to conduct this study.
Acknowledgments
The data reported here have been supplied by the Minneapolis Medical Research Foundation as the contractor for the Scientific Registry of Transplant Recipients (SRTR).
The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy of or interpretation by the SRTR or the US Government.
Footnotes
Published online ahead of print. Publication date available at www.cjasn.org.
References
- 1.Rostaing L, Cantarovich D, Mourad G, Budde K, Rigotti P, Mariat C, Margreiter R, Capdevilla L, Lang P, Vialtel P, Ortuño-Mirete J, Charpentier B, Legendre C, Sanchez-Plumed J, Oppenheimer F, Kessler M, CARMEN Study Group : Corticosteroid-free immunosuppression with tacrolimus, mycophenolate mofetil, and daclizumab induction in renal transplantation. Transplantation 79: 807–814, 2005 [DOI] [PubMed] [Google Scholar]
- 2.Hricik DE, Seliga RM, Fleming-Brooks S, Bartucci MR, Schulak JA: Determinants of long-term allograft function following steroid withdrawal in renal transplant recipients. Clin Transplant 9: 419–423, 1995 [PubMed] [Google Scholar]
- 3.Opelz G: Effect of the maintenance immunosuppressive drug regimen on kidney transplant outcome. Transplantation 58: 443–446, 1994 [DOI] [PubMed] [Google Scholar]
- 4.Nematalla AH, Bakr MA, Gheith OA, Elagroudy AE, Elshahawy M, Aghoneim M: Steroid-avoidance immunosuppression regimen in live-donor renal allotransplant recipients: A prospective, randomized, controlled study. Exp Clin Transplant 5: 673–679, 2007 [PubMed] [Google Scholar]
- 5.Chan K, Taube D, Roufosse C, Cook T, Brookes P, Goodall D, Galliford J, Cairns T, Dorling A, Duncan N, Hakim N, Palmer A, Papalois V, Warrens AN, Willicombe M, McLean AG: Kidney transplantation with minimized maintenance: Alemtuzumab induction with tacrolimus monotherapy--an open label, randomized trial. Transplantation 92: 774–780, 2011 [DOI] [PubMed] [Google Scholar]
- 6.Sampaio MS, Kadiyala A, Gill J, Bunnapradist S: Alemtuzumab versus interleukin-2 receptor antibodies induction in living donor kidney transplantation. Transplantation 88: 904–910, 2009 [DOI] [PubMed] [Google Scholar]
- 7.Huang E, Cho YW, Hayashi R, Bunnapradist S: Alemtuzumab induction in deceased donor kidney transplantation. Transplantation 84: 821–828, 2007 [DOI] [PubMed] [Google Scholar]
- 8.Margreiter R, Klempnauer J, Neuhaus P, Muehlbacher F, Boesmueller C, Calne RY: Alemtuzumab (Campath-1H) and tacrolimus monotherapy after renal transplantation: Results of a prospective randomized trial. Am J Transplant 8: 1480–1485, 2008 [DOI] [PubMed] [Google Scholar]
- 9.Welberry Smith MP, Cherukuri A, Newstead CG, Lewington AJ, Ahmad N, Menon K, Pollard SG, Prasad P, Tibble S, Giddings E, Baker RJ: Alemtuzumab induction in renal transplantation permits safe steroid avoidance with tacrolimus monotherapy: A randomized controlled trial. Transplantation 96: 1082–1088, 2013 [DOI] [PubMed] [Google Scholar]
- 10.Cherukuri A, Salama AD, Carter C, Smalle N, McCurtin R, Hewitt EW, Hernandez-Fuentes M, Clark B, Baker RJ: An analysis of lymphocyte phenotype after steroid avoidance with either alemtuzumab or basiliximab induction in renal transplantation. Am J Transplant 12: 919–931, 2012 [DOI] [PubMed] [Google Scholar]
- 11.Farney A, Sundberg A, Moore P, Hartmann E, Rogers J, Doares W, Jarrett A, Adams P, Stratta R: A randomized trial of alemtuzumab vs. anti-thymocyte globulin induction in renal and pancreas transplantation. Clin Transplant 22: 41–49, 2008 [DOI] [PubMed] [Google Scholar]
- 12.Hanaway MJ, Woodle ES, Mulgaonkar S, Peddi VR, Kaufman DB, First MR, Croy R, Holman J, INTAC Study Group : Alemtuzumab induction in renal transplantation. N Engl J Med 364: 1909–1919, 2011 [DOI] [PubMed] [Google Scholar]
- 13.Knechtle SJ, Fernandez LA, Pirsch JD, Becker BN, Chin LT, Becker YT, Odorico JS, D’alessandro AM, Sollinger HW: Campath-1H in renal transplantation: The University of Wisconsin experience. Surgery 136: 754–760, 2004 [DOI] [PubMed] [Google Scholar]
- 14.Ciancio G, Gaynor JJ, Guerra G, Sageshima J, Chen L, Mattiazzi A, Roth D, Kupin W, Tueros L, Flores S, Hanson L, Vianna R, Burke GW, 3rd: Randomized trial of three induction antibodies in kidney transplantation: Long-term results. Transplantation 97: 1128–1138, 2014 [DOI] [PubMed] [Google Scholar]
- 15.Kirk AD: The cam-path forward. Am J Transplant 13: 9–10, 2013 [DOI] [PubMed] [Google Scholar]
- 16.Opelz G, Döhler B, Laux G, Collaborative Transplant Study : Long-term prospective study of steroid withdrawal in kidney and heart transplant recipients. Am J Transplant 5: 720–728, 2005 [DOI] [PubMed] [Google Scholar]
- 17.Heilman RL, Khamash HA, Smith ML, Chakkera HA, Moss AA, Reddy KS: Delayed allograft inflammation following alemtuzumab induction for kidney transplantation. Clin Transplant 27: 772–780, 2013 [DOI] [PubMed] [Google Scholar]
- 18.Organ Procurement and Transplantation Network and Scientific Registry of Transplant Recipients 2010 data report. Am J Transplant 12[Suppl 1]: 1–156, 2012 [DOI] [PubMed] [Google Scholar]