Increased Immunosuppressive Vulnerability in Elderly Renal Transplant Recipients1,2

The results of renal transplantation have improved steadily over the last 10 years. This improvement is in large part attributable to improvements in immunosuppressive pharmacotherapy. Several phase III clinical trials have demonstrated decreased acute rejection rates with relatively little, if any, increase in infectious complications in a broad cross-section of patients (1–12). It is possible, however, that in certain subpopulations of patients, such as elderly renal transplant recipients, the risk/benefit ratio of acute rejection versus infection may be different from that of the general renal transplant population. If this were the case, the appropriate immunosuppressive regimen for that population should be reconsidered. Renal transplantation is a relatively safe option for renal replacement therapy in elderly patients with end-stage renal disease (ESRD). In 1976, Tersigni et al. (13) reported a series of nine ESRD patients above the age of 60 years successfully treated with renal transplantation. Shortly thereafter, Wedel et al. (14) published a series of 41 geriatric renal transplant recipients, and pointed out that the risk to these patients was not graft loss from rejection but, rather, death with a functioning graft. This author also reported that there was an increased risk for serious infectious complications in the older age group. With the advent of cyclosporine and more selective immunosuppression, the results of renal transplantation improved in high-risk groups, such as elderly patients. In 1989, Pirsch et al (15) concluded that cadaveric renal transplantation with cyclosporine immunosuppression was safe and an effective therapeutic modality in elderly ESRD patients. Subsequent studies have reinforced that concept (16) but have also reinforced the idea that elderly patients have a degree of immune incompetence (17) and require less aggressive immunosuppressive therapy (17). That theory was supported by the continuing observation of lower rejection rates (17) and lower incidence of chronic rejection (18, 19), but higher risk of infections (19, 20) noted in elderly transplant recipients. Reinforcement of the practice of transplantation in elderly patients came from a study that showed significantly greater survival probability in ESRD patients over the age of 60 who received transplants as opposed to matched patients who remained on dialysis (9). Recent data demonstrate a very similar 5-year graft (54–74%) and patient survival (52–74%), confirming the improvement made but also the concept that most patients in this high-risk group die with functioning grafts (19). In this study, posttransplant morbidity was attributed primarily to infectious complications and an increased prevalence of malignancy (19). We demonstrated previously, in a single-center study, that intensification of immunosuppressive therapy in elderly transplant recipients increased infectious complications without decreasing the incidence of acute rejection or improvement in graft survival (21). Given the above data, it is possible that the vulnerability to immunosuppression of older renal transplant patients may be very different from that of younger patients. To test this hypothesis, we analyzed a large group of renal transplant recipients with regard to their balance of acute rejection versus death due to infection. PATIENTS AND METHODS Patients and study design. We retrospectively analyzed adult patients (age >18 years) registered in the USRDS and UNOS Renal Transplant Scientific Registry, who had received a renal transplant between 1988 and 1997. Multiorgan recipients were excluded from the study. The patients were divided into age groups to evaluate the effect of age on the primary and secondary study end points. Primary study end points chosen to evaluate the risk/benefit ratio of the renal transplant were death secondary to infectious cause, death secondary to opportunistic infection, and acute rejection. Secondary study end points were patient survival, graft survival, and death censored graft survival. Gender, ethnicity, HLA mismatch, percent panel-reactive antibody (PRA), delayed graft function, cold ischemia time, induction therapy, dialysis time before transplantation, disease leading to ESRD, immunosuppressive regimen, cytomegalovirus (CMV) risk group, living versus cadaveric donor, and era effect were used to correct for possibly significant confounders in the analysis. Ethnicity was divided into Caucasians, African Americans, and others. The HLA mismatch was considered for DR and AB separately. Delayed graft function was defined as need for dialysis in the first week after transplantation. The original disease leading to the ESRD of the transplant recipient was divided into the following seven groups: glomerulonephritis (GN), hypertension (Htn), diabetes (DM), polycystic kidney disease (PKD), obstructive uropathy, uncertain diagnosis, and other causes of renal disease. Immunosuppressive regimens were evaluated as cyclosporine (CsA) versus tacrolimus (FK506) versus other, and azathioprine (Aza) versus mycophenolate mofetil (MMF) versus other. Based on the serological status of the donor and the recipient, the patients were divided into four CMV risk groups as follows: positive donor into positive recipient, negative donor into negative recipient, positive donor into negative recipient, and negative donor into positive recipient. The study endpoints of death due to infection and the subgroup of patients with death due to opportunistic infection (opportunistic infection was defined as infection due to CMV, fungus, parasite, or PCP) were recorded for the first 24 months of follow-up. Acute rejections were recorded over the initial 6 months after transplantation. The secondary study endpoints were analyzed for a 24-month follow-up period. The follow-up was limited to the initial 24 months after transplantation to increase the precision of the recorded data. Statistical analysis. A Cox proportional hazard model was used to investigate the relationship between age group and the primary study end points of death due to infection and opportunistic death due to infection. The above-listed variables were used as covariates in these models. Conditional backward elimination was used to identify significant covariates and reduce the model to contain the least amount of covariates necessary. Categorical covariates with an intrinsic order, such as age group and era, were analyzed by simple contrasts. Categorical covariates without an intrinsic order, such as CMV risk groups, were analyzed by a deviation contrast. Simple contrast compared each predictor variable except the first to the reference category represented by the first category. Deviation contrasts compared each predictor variable category except for the first to the overall effect. The cumulative hazard by age group, as calculated by the Cox proportional hazard model at mean of covariates was displayed in graphical format. Logistic regression was used to investigate the relation of acute rejection to the above-mentioned covariates. Demographic data, divided by age group, were presented as mean±SD or frequencies, as appropriate. Follow-up data such as patient death and graft loss were reported as frequencies and in percentages divided by age group. Test results were considered statistically significant at P <0.05. For all statistical analysis, SPSS software (Version 9.0 for Windows 95, SPSS, Inc., Chicago, IL) was used. RESULTS Demographics. A total of 73,707 primary renal transplant recipients who had received transplants between 1988 and 1997 were evaluated. The demographic characteristics of this study population divided by age group are presented in Table 1. Table 1: DemographicsUnivariate description of study end points. The primary and secondary study end points are reported in Table 2. Graft loss censored for patient death, patient death, patient death secondary to infectious causes, and patient death secondary to opportunistic infectious cause progressively increased with increasing age group. The percentage of patients who died during follow-up increased 5 times from the lowest to the highest age group. Death due to infection increased 6 times from 0.8% in the lowest age group to 4.8% in the highest age group. Opportunistic death due to infection increased 8.7 times from 0.06% in the 18–29 year age group to 0.52% in the over 65-year group. Conversely, the incidence of acute rejection during the first 6 months of follow-up decreased progressively from 28% in the youngest age group to 19.7% in the oldest patient group. Figure 1 shows the decreasing incidence of acute rejection by age during the first 6 months of follow-up and the increasing incidence of death due to infection during the first 24 months of follow-up. Table 2: Graft loss, patient death within the first 2 yr of follow-upFigure 1: Incidence of infectious death versus acute rejection.Multivariate analysis. The results of the Cox proportional hazard model for death due to infection during the first 24 months of follow-up are shown in Table 3. Age was the greatest relative risk for death due to infection during the first 24 months of follow-up (Fig. 2). The relative risk increases progressively and significantly for each increasing age group, compared with the reference group (18–29 years), with the oldest patients (above age 65) having the greatest (6.2-fold) relative risk of death due to infection within the first 24 months of follow-up (P <0.001). Table 3 shows all other significant covariates for which the model corrected. Gender, cold ischemia time, and delayed graft function were eliminated from the model as not significant for death due to infection. CMV serology with positive donor transplant into a negative recipient was an independent risk for death due to infection (relative risk=1.37, P <0.001). Compared with azathioprine, MMF was more protective against death due to infection (relative risk=0.77, P =0.02), whereas cyclosporine versus FK506 did not show a significant difference. Higher panel reactive antibody (PRA), AB, and DR mismatch, increased time on dialysis, and diabetes mellitus also were associated with a higher relative risk of death due to infection. Living donor status, and African-American race, were associated with a decrease in relative risk of death due to infection. There was a strong era effect, with more recent years demonstrating a decreased risk of death due to infection. Table 3: Cox proportional hazard model describing the risk factors for infectious death after transplantationaFigure 2: Cumulative hazard of infectious death during the first 24 mo after transplantation, by age group.The Cox proportional hazard model for opportunistic death due to infection during the first 24 months of follow-up (data not shown) displayed a similar pattern. The risk associated with age was stronger in this subgroup of infections, with an 8.7-fold relative risk in the age group above 65 years (P <0.001). Figure 3 shows the cumulative risk of opportunistic death due to infection over the first 24 months by age group corrected for all significant covariates. Figure 3: Cumulative hazard of opportunistic infectious death during the first 24 mo after transplantation, by age group.The binary logistic regression analysis for acute rejection during the first 6 months of follow-up confirmed the decrease of acute rejection with increasing age described in Table 2. Other significant risk factors for acute rejection were race (African-American), gender (female), cadaveric versus living donation, high PRA, no induction, azathioprine versus MMF, and tacrolimus versus cyclosporine. As shown in Figure 4, the relative risk of acute rejection decreases progressively and significantly with increasing age to almost half for the age group over 65 years, whereas the risk for death due to infection increased to more than 6 times for the highest age group. Figure 4: Relative risk of infectious death versus acute rejection.DISCUSSION Our study demonstrates a change in the vulnerability to immunosuppression as renal transplant recipients’ ages increase. The relative risk of death due to infection is markedly influenced by patient age, with older patients incurring a greater risk of this endpoint. Conversely, recipient age is negatively related to the incidence of acute rejection. This was also true for the absolute numbers with the incidence of death secondary to infection increasing linearly from 0.8% in the youngest age group to 4.8% in the oldest age group, and the acute rejection decreasing linearly from 28% to 19.7% (Fig. 1). In addition, this was confirmed by the multivariate analysis, which revealed a rising relative risk of death due to infection with increasing age and a progressively decreasing relative risk of acute rejection risk (Fig. 4). The analysis of the cumulative hazard for death due to infection over time divided by age group (Fig. 2) shows that the curves separate fairly soon after transplant. Whereas all curves show the steepest slope early after transplantation, between 0 and 12 months, the curves for the younger patients level out earlier than do those in the older transplant recipient groups. The multivariate analysis corrected for other variables that significantly affected the risk of death due to infection. As expected, the higher-risk CMV donor-to-recipient serologies affected the risk of death due to infection. Possibly, higher PRA predisposed to more death due to infection through increased immunosuppression in this patient group, and more need for anti rejection therapy. The same considerations are valid for the effect of AB and DR mismatch. Death from opportunistic infection followed the same trend as overall death due to infection, with older patients incurring a significantly greater risk of this adverse end point. Length of dialysis time before transplant was among the expected significant risk factors for fatal infectious outcome. Notable also was the significant and linear decreased relative risk of death due to infection over the years from 1988 to 1997, which decrease is probably ascribable to improved diagnostic and therapeutic tools. By multivariate analysis, the effect of age on both rejection and death due to infection was independent of baseline immunosuppression. In fact, because of the lower rejection rates in older patients and the resulting lower incidence of rejection treatment, it is likely that the older patients may have received less net immunosuppression, thus reinforcing the concept of their enhanced vulnerability. Our study confirms previous data indicating a decrease in acute rejection in older transplant recipients (17). The increase in death due to infection in older transplant recipients is not unexpected given the changes in immunologic function that occurs with aging (17, 22–27). It is possible that transplant per se is not changing this pattern of increased relative infectious vulnerability with age. However, given the nature of immunosuppressive therapy, we find this to be unlikely, and even if true, does not detract from the basic finding of altered risk/benefit ratio. We would like to make clear that this study does not indicate that it is unsafe to transplant older ESRD patients. In fact, several studies have demonstrated excellent results of renal transplantation in older patients (19). In addition, this study does not address the relative risk of transplantation versus dialysis in older patients, an issue that has been discussed elsewhere (9). In summary, the relative risk of death due to infection increases and the risk of acute rejection decreases as patients age. Our data suggest that older renal transplant recipients might need less pharmacologic immunosuppression than do younger patients to achieve similar outcomes. These data may be important in tailoring immunosuppressive protocols for people at the extremes of age, as well as helping form a realistic expectation of transplant outcomes in elderly transplant recipients.