The Effect of Antilymphocyte Induction Therapy on Renal Allograft Survival

Methods

Study Identification and Selection

The selection of studies has been described else-where [17]. Of 247 potentially eligible studies identified from MEDLINE, 7 met the inclusion criterion of reporting allograft survival data from a randomized trial that compared polyclonal or monoclonal antilymphocyte antibodies used in the period immediately after transplantation with conventional immunosuppression in adult recipients of cadaveric renal transplants. No additional published or unpublished trials were identified. Because the data from two trials [10, 12] were not available, our study is based on data from patients in five previously published randomized, controlled trials [11, 13-16].

No industry funds were used directly in support of our study. Two randomized trials included in this meta-analysis were initially supported with industry funds: The study by Norman and colleagues [14] was paid for by the R.W. Johnson Pharmaceutical Research Institute (Raritan, New Jersey), and the study by Abramowicz and coworkers [13] was supported in part by Cilag Benelux, Belgium. No company was involved, directly or indirectly, in the design or conduction of this meta-analysis or in the decision to submit this manuscript for publication.

Data Used for Analysis

The baseline individual patient-level information included age, sex, ethnicity, height, weight, presence of diabetes mellitus, history of previous transplantation, panel-reactive antibody levels, number of previous blood transfusions, number of donor-recipient HLA mismatches, cold ischemic time, delayed allograft function, and survival time. Survival time was defined as the interval between placement of the allograft and return to dialysis, retransplantation, death, or end of follow-up, whichever occurred first. For our primary analysis, death was considered an allograft failure. To confirm that our analyses were stable, we also considered death a censoring event rather than an allograft failure. These two approaches produced almost identical results; thus, only our primary analysis is presented here. Because we used individual-level data, follow-up was longer than in our previous meta-analysis. Mean follow-up (±SD) ranged from 36.7 ± 16 months to 46.1 ± 11 months (maximum, 60 months).

Data from each study were extracted by one of the authors and converted into a common format. Categorical variables were created for age (18 to 34, 35 to 49, 50 to 64, and ≥ 65 years of age), previous transplants (zero and nonzero), panel-reactive antibody level (0% to 19%, 20% to 79%, and 80% to 100%), and cold ischemic time (<24 hours or ≥ 24 hours) by using the strata used by UNOS (United Network of Organ Sharing) [32]. Two of the categories for panel-reactive antibody level (20% to 79% and 80% to 100%) were combined because of small numbers in those categories. Patients with panel-reactive antibody levels of 20% or more were considered to be presensitized; patients with levels less than 20% were considered to be unsensitized.

Because it is known that fewer than 2% of the patients enrolled by Abramowicz and coworkers [13] were diabetic or nonwhite (Vereerstraeten P. Personal communication), all of the patients in that study were considered white and nondiabetic. Delayed allograft function was defined as the need for dialysis during the first week after transplantation in all studies except one [11], in which it was defined as a decrease in creatinine level occurring more than 1 day after transplantation.

Each database was examined for missing data, and each observation was checked for internal consistency. The distributions of demographic and prognostic variables at baseline (for example, ethnicity, previous transplantation, or panel-reactive antibody level) were described overall and for each treatment group. Treatment groups were compared by using a chi-square test stratified by study to confirm the success of randomization [33]. Inconsistencies between these results and the results in the published reports were resolved with the authors of the studies.

Statistical Analysis

All analyses were done by using the intention-to-treat approach; that is, patients were analyzed according to the treatment group to which they had been randomly assigned. Possible selection bias resulting from the inclusion of five of the seven eligible randomized trials was evaluated. Using data from the published reports, we calculated the rate ratios of allograft failure by use of a discrete-time version of the proportional hazards model [34] for both the subset of studies for which individual patient-level data were available [11, 13-16] and the subset for which data were unavailable [10, 12]. The rate ratios compared induction therapy with no induction therapy at 2 years of follow-up. Each rate ratio was compared with the other and with the pooled rate ratio from all seven studies [17].

The unadjusted distribution of allograft survival was estimated for treatment groups within each study individually and for the seven studies overall by using the Kaplan-Meier product-limit estimator [35]. Treatments were compared by using the log-rank test [36], stratified by study, at 2 and 5 years. In addition, Cox proportional-hazards regression was used to estimate the overall effect of induction therapy on allograft survival in both unadjusted and adjusted analyses. Proportional hazards regression was also used to determine other factors associated with survival and to explore for interactions between treatment group and other factors [34].

Initially, a model was fit to the individual patient data for 2 years of follow-up to confirm the results of our previous meta-analysis. The model was then fit by using the entire length of follow-up available in the individual-level data (5 years) to explore the long-term effects of induction therapy on allograft survival. Because we were uncertain whether the effect of induction therapy was constant over time, we looked for a time-by-treatment interaction by using a time-varying indicator variable, distinguishing the periods 0 to 2 years and 2 to 5 years [37]. The cut-point of 2 years was chosen to correspond to the availability of 2 years of follow-up in the published data. Finally, a model was fit for the subgroup of patients whose allografts survived for at least 2 years.

Several approaches were used to select variables for inclusion in the multivariable regression models. In general, the models were fit to include the maximal number of candidate variables while minimizing missing data. Variables representing treatment and study were forced into all models. Other predictors of survival were first selected by using both forward and backward stepwise elimination methods. Entry and elimination criteria were set at a value of P = 0.15. In addition, variables were allowed to enter models if their inclusion altered the rate ratio for the treatment variable by 15% or more. Standard methods were used to calculate multivariable rate ratios on the basis of model variable co-efficients [34]. Terms representing interactions between the use of induction therapy and selected variables (recipient ethnicity, delayed allograft function, diabetes mellitus, HLA mismatch, presensitization, previous transplantation, cold ischemic time, and study) were included in separate models. These interaction terms tested whether the benefit of induction therapy differed across categories of these variables.

Because data on cold ischemic time, recipient age, recipient sex, delayed allograft function, and number of previous blood transfusions were lacking for a substantial number of patients, these variables were individually examined in models adjusted for use of induction therapy, recipient ethnicity, diabetes mellitus in the recipient, number of HLA-DR mismatches, presensitization, previous transplantation, and study. The same models were also used to evaluate the potential confounding effect of cold ischemic time, recipient age, recipient sex, delayed allograft function, and number of previous blood transfusions on the effect of induction therapy.

All P values are two-sided. All analyses, with the exception of those that involved time-varying covariates, were done by using Stata, version 5.0 (Stata Corp., College Station, Texas). The analysis incorporating a treatment-by-time interaction was done by using SAS software, version 6.08 (SAS Institute, Inc., Cary, North Carolina).

Results

When the data from the five published reports were combined, the rate ratio for allograft failure at 2 years from the discrete-time survival models was 0.69 (95% CI, 0.47 to 1.01; P = 0.053). This did not differ from the rate ratio derived from the seven studies included in our previous meta-analysis [17], which was 0.69 (CI, 0.49 to 0.97; P = 0.03), or the rate ratio derived from the two studies that could not be included in the current analysis, which was 0.70 (CI, 0.31 to 1.58; P > 0.2). These results indicate that excluding two of the seven potentially eligible randomized, controlled trials did not create substantial bias.

Descriptions of the individual study designs and characteristics of the treatment and control groups are given in Table 1 and (Table 6). Individual-level data on 628 patients from five studies [11, 13-16] were included in this analysis. Five-year follow-up data that were not included in the original published reports were available for the studies by Belitsky and colleagues [11] and Abramowicz and coworkers [13]. During this extended follow-up period, the immunosuppressive protocols and the intensity of patient follow-up were the same as they were in the original observation period (Belitsky P. Personal communication; Abramowicz D. Personal communication).

Unadjusted Analysis

The unadjusted Kaplan-Meier curves comparing induction therapy and conventional immunosuppression over 5 years are shown in Figure 1. The rate of allograft survival at 5 years was 69.0% in the group receiving induction therapy and 64.4% in the group not receiving this therapy. When we used the log-rank test stratified by study site, this unadjusted analysis showed a benefit from induction therapy that did not reach conventional levels of statistical significance (P = 0.13). The proportion of patients with functioning allografts, the unadjusted rate ratios for allograft failure, and the results of log-rank tests for individual studies at 2 and 5 years are shown in Table 2. These results are similar to those in the published reports.

Adjusted Analysis

A multivariable model that adjusted for known prognostic variables for allograft survival was initially fit at 2 years of follow-up by using individual patient-level data. The rate ratio for allograft failure at 2 years was 0.62 (CI, 0.43 to 0.90; P = 0.012), confirming the effect seen in our previous meta-analysis (Table 3). In addition to induction therapy, greater donor-recipient HLA-DR mismatch was significantly associated with allograft failure at 2 years. Previous transplantation had an unexpected beneficial effect on allograft survival, but this did not achieve conventional levels of statistical significance. Although presensitization was not a significant predictor of allograft survival at 2 years, the effect of induction therapy differed in presensitized and unsensitized patients (P = 0.03). This finding was confirmed in a stratified analysis. The adjusted rate ratios of allograft failure for patients receiving induction therapy compared with conventional therapy were 0.12 (CI, 0.03 to 0.44; P = 0.002) in presensitized patients (85 patients with 15 failures) and 0.74 (CI, 0.50 to 1.09; P = 0.13) in unsensitized patients (511 patients with 100 failures). Among presensitized patients, these results translate to allograft survival rates at 2 years of 68.4% in the group receiving conventional therapy and 92.2% in the group receiving induction therapy. The Kaplan-Meier curves comparing induction therapy and conventional immunosuppression at 2 years for presensitized and unsensitized patients are shown in the top panel of Figure 2. No other interactions between induction therapy and prognostic variables were significant in the analysis at 2 years of follow-up.

In the multivariable model that examined 5 years of follow-up (Table 4), patients receiving induction therapy had somewhat greater allograft survival, but this difference did not achieve conventional levels of statistical significance (rate ratio, 0.82; P = 0.17). Diabetes mellitus in the recipient and greater donor-recipient HLA-DR mismatch were significant predictors of poor allograft survival. In addition, African-American and presensitized recipients had worse outcomes. A recipient's history of previous transplantation had an unexpected beneficial effect on allograft survival.

As at 2 years, the interaction between presensitization and induction therapy at 5 years was highly significant (P = 0.009). In the fully adjusted model stratified by presensitization, the rate ratios for allograft failure at 5 years were 0.20 (CI, 0.09 to 0.47; P < 0.001) in presensitized patients (85 patients with 33 failures) and 0.97 (CI, 0.71 to 1.32; P > 0.2) in unsensitized patients (510 patients with 163 failures). Among presensitized patients, allograft survival was 47.4% in the group receiving conventional therapy and 72.5% in the group receiving induction therapy. The Kaplan-Meier curves comparing induction therapy and conventional immunosuppression at 5 years for presensitized and unsensitized patients are shown in the bottom panel of Figure 2. No interactions between induction therapy and any other variable (including study site) were found.

The statistical and clinical significance of the association between induction therapy and allograft survival in the analysis at 2 years but not at 5 years suggests that the effect of induction therapy waned with time. Therefore, we examined the time-by-treatment interaction to determine whether the effect of induction therapy differed between the first 2 years of follow-up and the period beginning 2 years after transplantation. Because this interaction was statistically significant (P = 0.04), we examined these two time periods separately. The adjusted rate ratio for induction therapy was 0.62 (CI, 0.43 to 0.90; P = 0.012) during the first 2 years after transplantation and 1.13 (CI, 0.72 to 1.78; P > 0.2) among patients followed for 5 years whose allograft survived for 2 or more years (Table 5). Thus, the beneficial effect of induction therapy seemed to be limited to the first 2 years after transplantation.

Neither recipient sex nor number of previous blood transfusions was a statistically significant predictor of allograft survival. Younger recipient age, delayed allograft function, and prolonged cold ischemic time were significant predictors of worse allograft survival at 2 and 5 years. The potentially confounding influence of these variables on the effect of induction therapy was examined among the subset of patients for whom no data were missing. The effect of induction therapy was stable with and without multivariable adjustment for all of these variables, indicating a lack of confounding. No interactions between these variables and the effect of induction therapy on allograft survival were detected.

Discussion

This meta-analysis of individual patient-level data has confirmed and extended our previous work [17], delineating the effect of induction therapy on allograft survival at 5 years of follow-up and examining the effects of this therapy in subgroups of patients at high risk for allograft failure. In contrast to an improvement in allograft survival that was both statistically and clinically significant at 2 years, induction therapy was associated with only a small improvement in allograft survival in the entire study population at 5 years; this improvement did not reach conventional levels of statistical significance. The attenuation of the effect of induction therapy at 5 years resulted from an absence of benefit after 2 years of follow-up. Among the many high-risk subgroups studied, only presensitized patients derived substantial benefit from induction therapy. These subgroup findings were stable even at 5 years of follow-up, at which time induction therapy was associated with an increase in allograft survival of 25.1% among presensitized patients.

Cyclosporine treatment, like antibody induction therapy, decreases the incidence [38] and diminishes the severity [39] of acute transplant rejection. The period after transplantation during which induction therapy provides a benefit with respect to allograft survival is similar to that seen with cyclosporine. Randomized trials of immunosuppressive regimens, including cyclosporine, were associated with greater allograft survival, but only during the first year after transplantation [38, 40, 41]. The mechanism of the benefit of both cyclosporine and antibody induction therapy may therefore be related to early immunomodulation in the peritransplantation period.

Selective benefit was seen in this study among presensitized patients. These findings are consistent with data from a retrospective cohort study by Opelz [42] that showed greater allograft survival among patients who received OKT3 (Orthoclone OKT3, Ortho Pharmaceutical Corp., Raritan, New Jersey) induction therapy.

We could not show an increased benefit from induction therapy for many subgroups postulated to benefit from it, including African-Americans [19, 43], patients with poor HLA matches [44-46], patients with previous transplantation [14, 47], and patients with delayed allograft function [48-50]. The numbers of patients in these subgroups were limited, however, reducing the power to detect these effects.

This meta-analysis also confirms the clinical importance of several factors that have been shown to have a negative effect on allograft survival, including African-American ethnicity of the recipient [18, 20, 21], diabetes mellitus in the recipient [18], increasing HLA-DR mismatch [18, 28-31], delayed allograft function [23-27], and prolonged cold ischemic time [18]. In contrast to findings in previous reports [18, 22], previous transplantation was associated with longer allograft survival. However, only two of the five studies included in this meta-analysis enrolled patients with previous transplantation.

Meta-analysis of the published literature is increasingly being used, allowing similar clinical trials to be combined quantitatively and thereby increasing the power and precision of the estimation of effect. However, these meta-analyses are limited to summary measures of the subset of variables and duration of follow-up presented in each published report and often cannot use survival analyses that can consider the incomplete follow-up of individual persons. Meta-analysis of individual patient-level data, as done in this study, can overcome these limitations. Even when follow-up is incomplete, individual patient-level data can be analyzed with survival analysis techniques that yield estimates of true rates of failure over time rather than less informative estimates of risk (such as the proportion of patients having failure at specific time points). In addition, by using individual patient-level data extended beyond the follow-up in the published literature, meta-analyses such as this one can evaluate long-term survival. Individual patient-level data also provide the opportunity to obtain information on variables that are potentially important predictors of outcome and are unavailable in meta-analyses of published literature. For example, in this study, we were able to evaluate potential predictors of allograft survival, such as recipient ethnicity and panel-reactive antibody levels; this was not possible in our previous meta-analysis, which used published data [17]. Each trial can also be assessed for the balance of these factors across treatment arms, and analyses can be adjusted for potential confounding. Finally, use of individual patient-level data allowed us to perform subgroup analyses that were not possible by using original published analyses.

Few meta-analyses combining data at the level of the individual patient have been published [51-55]. Several of these studies [51, 52, 55] compared their results with those of a meta-analysis done by using published data, principally as a way to validate these techniques. Fortin and colleagues [52] and Steinberg and associates [55] used both techniques to analyze the same group of patients and found similar associations between treatment groups. Stewart and Parmar [51], however, added trials, used different analytic methods, and extended follow-up for the meta-analysis of individual patient-level data. The meta-analysis of the published literature showed a greater effect of therapy (cisplatin-based chemotherapy for ovarian cancer) with regard to both effect size and statistical significance compared with meta-analysis of individual patient-level data. Although limited in number, these previous studies, along with this meta-analysis, suggest that the two types of meta-analysis can be complementary.

The limitations of our analysis include the limitations of the individual studies, including the absence of measures of quality of life, potential publication bias, and heterogeneity among studies. The issue of individual study quality was addressed in our previous analysis, in which we showed that excluding studies with lower quality scores increased the effect size of induction therapy [17]. In our current analysis, inclusion of these studies has the potential to bias the results toward the null; this suggests that we may have underestimated the beneficial effect of induction therapy. Potential publication bias has been examined through an extensive search of many sources for additional studies and unpublished data and is unlikely in light of the nonsignificant findings of all individual published randomized, clinical trials on allograft survival [10-16].

Heterogeneity of treatment effects among studies included in a meta-analysis suggests that pooling the results of those studies may be inappropriate. Extensive efforts were made to explore for heterogeneity among the studies included in our analyses. No significant interaction was seen between treatment with induction therapy and the indicator variables representing study site, indicating that no variation was detectable in the effect of induction therapy on survival across studies.

In summary, our analysis of individual patient-level data showed improved allograft survival at 2 years with the use of antilymphocyte antibody therapy in the peritransplantation period. The benefit waned after the first 2 years after transplantation, except among presensitized patients, whose allograft survival was improved substantially by induction therapy even at 5 years of follow-up. These results indicate that induction therapy may have a beneficial role in cadaveric kidney transplantation, especially in presensitized patients. Data on toxicity were not available in sufficient detail to allow for their assessment here. The decision to use induction therapy must balance the treatment's potential benefit against its costs ($6720 is the average wholesale pharmacy price for a 50-mg course of OKT3 [56]) and potential toxicity [57, 58], which need further examination before the routine use of antibody induction therapy can be recommended.

Glossary

Allograft: Tissue transplanted between two genetically different members of the same species.

Cold ischemic time: Period of cold storage between organ harvest and transplantation.

Delayed allograft function: Primary nonfunction of the transplanted kidney that results in anuria or oliguria and the need for dialysis in the period immediately after transplantation.

HLA mismatch: Human leukocyte molecules encoded by the HLA complex on the short arm of chromosome 6 play a central role in the recognition of foreign antigens by T cells. For renal transplantation, the alleles at three loci of the HLA complex (A, B, and DR) are evaluated to match recipients with donors. Because each locus is represented by two alleles, 0 to 6 HLA mismatches are possible between any allograft recipient and any donor.

Individual patient-level data: Demographic, risk factor, and outcome information obtained individually for each study patient.

Induction therapy: Antilymphocyte antibody therapy given in the period immediately after transplantation to prevent acute and chronic rejection.

Monoclonal antilymphocyte antibodies: A murine mouse monoclonal antibody directed against the CD3 complex on mature human T cells (Orthoclone OKT3, Ortho Pharmaceutical Corp., Raritan, New Jersey).

Panel-reactive antibody test: A test used to screen potential recipients for preformed anti-HLA cytotoxic antibodies. A patient's serum specimen is tested against a panel of common antigens; the percentage to which it reacts indicates the degree of sensitization.

Polyclonal antilymphocyte antibodies: γ-Globulin fractions from animals, such as horses or rabbits, that have been immunized with human lymphoid tissue, such as antilymphocyte globulin (Merieux, Lyon, France) and Minnesota antilymphoblast globulin (University of Minnesota, Minneapolis, Minnesota).

Presensitized patient: A patient with many preformed HLA cytotoxic antibodies and an increased likelihood for a positive cross-match when tested against a potential donor.

Published-level data: Demographic, risk factor, and outcome information obtained from a published report that is often limited to summary measures, such as the mean age of a cohort or the proportion of a cohort that is male.

Unsensitized patient: A patient with a low or undetectable number of preformed HLA cytotoxic antibodies and a decreased likelihood for a positive cross-match when tested against a potential donor.

Appendix

The members of the Anti-Lymphocyte Antibody Induction Therapy Study Group are J.M. Morales and the Spanish Monotherapy Study Group, Hospital 12 de Octobre, Madrid, Spain; Daniel Abramowicz, Michel Goldman, Luc De Pauw, Jean-Louis Vanherweghem, Paul Kinnaert, and Pierre Vereerstraeten, Hospital Erasme, Brussels, Belgium; Philip Belitsky, A.S. MacDonald, A.D. Cohen, J. Crocker, D. Hirsch, K. Jindal, and J. Lawen, Dalhousie University, Halifax, Nova Scotia, Canada; Douglas P. Slakey, Christopher P. Johnson, Robert D. Callaluce, Barry J. Browne, Yong-Ran Zhu, Allan M. Roza, and Mark B. Adams, Medical College of Wisconsin, Milwaukee, Wisconsin; Thomas P. Haverty and Jennifer Dehlinger, R.W. Johnson Pharmaceutical Research Institute, Raritan, New Jersey; Douglas J. Norman, Oregon Health Sciences University, Portland, Oregon; Lawrence Kahana, University of South Florida, Tampa, Florida; Frank P. Stuart Jr., Northwestern University Medical Center, Chicago, Illinois; James R. Thistlethwaite Jr., University of Chicago, Chicago, Illinois; Charles F. Shield III, St. Francis Regional Medical Center, Wichita, Kansas; Anthony Monaco, New England Deaconess Hospital, Boston, Massachusetts; Shu-Chen Wu, R.W. Johnson Pharmaceutical Research Institute, San Diego, California; and Allan Van Horn, deceased.

Dr. Berlin: University of Pennsylvania Medical Center, Center for Clinical Epidemiology and Biostatistics, 611 Blockley Hall, 423 Guardian Drive, Philadelphia, PA 19104-6021.

Dr. Feldman: University of Pennsylvania Medical Center, Center for Clinical Epidemiology and Biostatistics, 720 Blockley Hall, 423 Guardian Drive, Philadelphia, PA 19104-6021.