Volume 29, Issue 9 , Pages 966-972, September 2010
Balancing rejection and infection with respect to age, race, and gender: Clues acquired from 17 years of cardiac transplantation data
Article Outline
Background
Donor and recipient risk factors for rejection and infection have been well characterized. The contribution of demographic factors, especially age at the time of transplantation to morbidity and mortality due to rejection and infection, is much less well understood.
Methods
Using parametric hazard analysis and multivariate risk-factor equations for infection and rejection events, we quantitatively determined the relationship of fundamental demographic variables (age, race and gender) to infection and rejection. These analyses were conducted with respect to date of transplant and age at the time of transplantation. The patient group consisted of all primary heart transplants performed at the University of Alabama at Birmingham during the years 1990 to 2007 (n = 526).
Results
Risk factors for rejection within 12 months post-transplantation were date of transplant (p < 0.0001) and age at the time of transplantation (young adults 10 to 30 years of age, p < 0.0001). Risk factors for infection were date of transplant (p < 0.0001) and age at the time of transplantation (young children and older adults, p < 0.0001). There were three immunosuppressive eras in 1990 to 2007. Notably, although the proportion of patients experiencing rejection and infection events decreased during each successive immunosuppressive era, the relative relationship of infection to rejection, as well as age at the time of transplantation, remained similar into the most recent era. The maximal frequency of rejection events and rejection death occurred among patients transplanted at ages 10 to 30 years. Conversely, the frequency of infection events was minimal within the same group. In the oldest and youngest patients receiving transplants, infection was the predominant cause of death and rates of rejection events decreased.
Conclusions
These data show that evolving immunosuppressive strategies have successfully reduced rejection and infection frequencies, and those patients transplanted at 30 to 60 years of age have the lowest frequency of rejection/infection events. However, individuals transplanted at younger or older ages, especially non-white recipients in the 10- to 30-year age group, experience significantly more infection or rejection. Therefore, programs should increase the level of surveillance in these patients and consider modification of immunosuppressive regimens in order to lower the frequency of infection and rejection events.
Keywords: rejection, infection, outcomes, multivariable, hazard, immuno-suppersion, age, mulitinstitutional
During the last two decades, numerous multi-institutional studies1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 have identified risk factors for infection and rejection after cardiac transplantation. These factors include donor-related variables, such as donor age and ischemic time, and recipient variables largely encompassing the clinical status of the recipient at the time of transplant. However, in addition to separate risk factors identified for each of these complications, intrinsic biologic differences among recipients (including age at the time of transplant, gender and race) may affect the immunologic response and therefore the balance of risk toward infection or rejection. This can be seen in multivariate and univariate analyses in which it has been reported that younger recipient age at transplant among adults is a risk factor for rejection,7 but among young children the tendency for rejection appears lower, particularly in infants and neonates.13, 14 These basic relationships were described in the 1990s and still hold true in the most recent era in which rejection rates have decreased because of evolving clinical protocols and the implementation of new immunosuppressive regimens.14, 15, 16, 17 The most recent era is also characterized by a wider range of demographic variables, as transplantation is more often offered to younger and older patients.
Over the years, there has been much discussion regarding “tailored immunosuppression,” which we will define here as the adjustment of maintenance immunosuppression with respect to individual patient variables with the intention of reducing the risk for morbid events. By this definition, most transplant physicians provide some version of tailored immunosuppression based on clinical indications. Among the majority of patients, most of whom are between the ages of 40 to 60 years, we will show that this goal has been largely achieved. However, in patients who fall outside this range, there exists considerable opportunity for improvement. This is a complex endeavor, involving interactions of age, race and gender superimposed on rejection and infection, which, in clinical transplantation, are not independent of one another. The observed basic relationships of rejection and infection may persist in the presence of “optimal survival” as we further tailor immunosuppressive modalities in an attempt to balance these risks while achieving the best possible survival.
Mathematical models are useful in assessing potential changes in outcome based on a set of known initial clinical variables. Previous models of outcomes due to infection or rejection have not been constructed that consider the interaction of both events and take into consideration demographic variables as well as immunosuppressive era. Therefore, the objective of these studies was to develop mathematical models of the interaction between rejection and infection as a function of age at the time of transplant to produce plots to assist in the decision-making process in which immunosuppressive regimes are adjusted to balance morbid events. We used 17 years of outcomes data to construct these models, which indicate the extent to which “tailoring” immunosuppression has succeeded, and to determine the circumstances under which such efforts have failed. We show that, in contrast to previous models that have examined rejection or infection in isolation, there are significant factors, particularly age, which strongly affect outcomes relative to the balance of rejection and infection.
Methods
Patient population
Between January 1, 1990 and December 31, 2007, 526 consecutive primary heart transplants were performed at a single institution, and are the subject of this analysis. Multi-organ transplants and retransplants are excluded. This study received approval from the institutional review board.
Definitions
Acute rejection was defined as any clinical event leading to temporary augmentation of immunosuppression, which usually consisted of a short course of oral or intravenous high-dose steroids and/or cytolytic therapy. A rejection episode was reported to end at the time of the last biopsy that did not trigger additional immunosuppression. By this definition, the frequency of biopsies after the beginning of a rejection episode will influence the rejection frequency and the calculated rejection duration. Since the frequency of biopsies was variable over the 18-year experience observed in the study, no follow-up biopsy was considered until day 10 after the start of rejection. If the biopsy resulted in increased immunosuppression, then it was considered part of the same rejection episode. Throughout the experience, treatment for rejection was typically initiated based on a group of clinical parameters, including endomyocardial biopsy of Grade 3A or higher (or Grade 1B or 2 in the early post-transplant period), an echocardiogram indicating acute right or left ventricular dysfunction, abnormal hemodynamics, or a group of symptoms considered indicative of acute rejection and resulting in the augmentation of immunosuppression. Hemodynamic compromise was defined as a decrease in rejection fraction to <0.45 as judged by echocardiography, with or without inotropic support, when the prior rejection fraction was documented to be normal (>0.55).
Infection was defined as an infectious process requiring intravenous therapy or as a potentially life-threatening infection requiring oral therapy.
Age groups were defined as follows: children (<10 years of age); young adults (10 to <30 years); mature adults (30 to <60 years); and senior adults (≥60 years).
Immunosuppression protocols
Immunosuppression protocols evolved over the 17-year period of this study. Throughout the experience, maintenance “triple therapy” immunosuppression consisted of cyclosporine or tacrolimus-based therapy combined with prednisone and a third agent. Prednisone was tapered after the first 3 weeks and eventually discontinued or maintained at low levels, depending on rejection frequency and severity. Early in the experience, the third agent was azathioprine, with an evolution toward mycophenolate mofetil in later years.
Routine treatment of acute rejection episodes consisted of prednisone or methylprednisolone for 3 days. In cases of persistent, recurrent or severe rejection, cytolytic therapy in the form of OKT3 or anti-thymocyte globulin (with or without plasmapheresis) was added. Later in the experience, this frequently included conversion from cyclosporine to tacrolimus and azathioprine to mycophenolate mofetil.
Immunosuppressive eras were determined on the basis of the frequency of use of six immunosuppressive drugs with respect to transplant date: cyclosporine, tacrolimus, azathioprine, mycophenolate mofetil, rapamycin and anti–interleukin-2 receptor humanized monoclonal antibodies. The percentages of patients receiving each of these drugs as initial immunosuppression were plotted as a function of transplant date. Three eras were identified based on the proportion of patients receiving each of each of these drugs. Era 1, spanning from 1990 through 1996, was ended by a switch to predominant usage of mycophenolate mofetil. Era 2, spanning from 1997 through 2000, ended with a switch to predominant usage of tacrolimus and the humanized IL-2 receptor monoclonal antibodies.
Statistical analysis
Patient data were collected using coded event forms that included recipient and donor demographic data and pre- and post-transplant clinical data. The forms also incorporated post-transplant events including death, rejection and infection. The information was entered into a computerized database and was subsequently verified using software-based tools and by inspection of the raw data by qualified medical personnel. Causes of death were individually verified by the responsible transplant physicians. Data were analyzed using an array of parametric methods as well as hazard function analysis of the instantaneous risk for a given event across time.18 Risk factors were identified using multivariate analyses. All statistics were performed using SAS software (SAS Institute, Inc., Cary, NC) and custom scripts executed within the SAS software package.
Results
Demographics
The demographics of this population are depicted in Table 1. Of these recipients, 407 (77%) were male. Approximately 85% of patients were white, with all but 5 (1.2%) of the remainder of apparent African descent (black). Seventy-nine percent of all patients were between 40 and 69 years of age (Table 2). Of the remaining patients, all but 2 (1% , >70 years) were <40 years old and were distributed among age groups as shown in Table 2.
Table 1. Patients by Race and Gender
| Race | Gender | ||||
|---|---|---|---|---|---|
| Male | Female | ||||
| n | Percent (of 407) | n | Percent (of 119) | Total | |
| White | 345 | 84.8% | 89 | 74.8% | 434 |
| Black | 57 | 14.0% | 30 | 25.2% | 87 |
| Other | 5 | 1.2% | 0 | — | 5 |
| Total | 407 | 100% | 119 | 100.0% | 526 |
Table 2. Patients by Age and Gender
| Age (years) | Gender | |||
|---|---|---|---|---|
| Male | Female | |||
| n | Percent (of 407) | n | Percent (of 119) | |
| 10 | 30 | 7% | 14 | 12% |
| 10–19 | 16 | 4% | 10 | 8% |
| 20–29 | 9 | 2% | 14 | 12% |
| 30–39 | 30 | 7% | 16 | 13% |
| 40–49 | 79 | 19% | 18 | 15% |
| 50–59 | 145 | 36% | 25 | 21% |
| 60–69 | 96 | 24% | 22 | 19% |
| 70+ | 2 | 1% | 0 | — |
Rejection and infection with respect to immunosuppressive era
Three immunosuppressive areas (see Methods) were identified during the study period. Era 1 spanned the years 1990 to 1996, and was characterized by the use of cyclosporine, azathioprine and methylprednisolone. In Era 2 (1997 to 2000), there was a shift from azathioprine to mycophenolate mofetil, occasional use of tacrolimus and the introduction of rapamycin and humanized anti–IL-2 receptor monoclonal antibodies. In Era 3 (2001 to 2007), tacrolimus became predominant and anti–IL-2 receptor antibody use became routine. As depicted in Figure 1, Figure 2, both cumulative rejection and infection decreased significantly during the experience (p < 0.0001).
Figure 1.
Cumulative incidence of rejection in recipients stratified by immunosuppression era. The circles represent non-parametric estimates of cumulative rejection episodes. Vertical bars represent the 70% confidence limits of the non-parametric estimate. The solid line is a parametric estimate resulting from the statistical model of cumulative rejection episodes. The dotted lines represent the 70% confidence limits of the parametric estimate.
Figure 2.
Cumulative incidence of infection in recipients stratified by immunosuppression era. The circles represent non-parametric estimates of cumulative rejection episodes. Vertical bars represent the 70% confidence limits of the non-parametric estimate.
Relation of demographic variables to cumulative rejection and infection
We performed a risk-factor analysis to model the relationships of demographic variables to cumulative rejection and infection. The shapes of the relationship curves were estimated by a parametric hazard analysis of cumulative episodes across time. Variables in the multivariate risk-factor analysis included age at time of transplant, race, gender and date of transplant. Interactions of these same variables were also examined. The results of the multivariate analysis (Table 3) show that date of transplant and black race as risk factors for rejection. It is of interest to note that young adults (10 to <30 years of age) were at higher risk for rejection, whereas children (<10 years) or senior adults (≥60 years) at the time of transplant had a higher risk for infection. This analysis included the entire experience.
Table 3. Risk Factors-Multivariable Analysis
| Risk factors | Cumulative rejections | Cumulative infections |
|---|---|---|
| Age | Young adult | Younger, older |
| (p < 0.0001) | (p < 0.0001, p < 0.0001) | |
| Gender | NS | NS |
| Race | Black | NS |
| (p < 0.0001) | ||
| Year of transplant | Earlier year | Earlier year |
| (p < 0.0001) | (p < 0.0001) |
The magnitude of the effect of race is illustrated in the cumulative rejection episodes, particularly during the first year (Figure 3), in which cumulative rejection was higher in blacks than whites for the same gender. As reflected in the analysis, black females were at highest risk for rejection, followed by black males, white females and white males (Figure 3). Race did not appear as a risk factor for infection by multivariate analysis (see Table 3), but actual cumulative infections were highest in white males by 8 years post-transplant (Figure 4).
Figure 3.
Cumulative incidence of rejection in recipients stratified by race and gender. The circles represent nonparametric estimates of cumulative rejection episodes. Vertical bars represent the 70% confidence limits of the non-parametric estimate.
Figure 4.
Cumulative incidence of infection in recipients stratified by race and gender. The circles represent non-parametric estimates of cumulative rejection episodes. Vertical bars represent the 70% confidence limits of the non-parametric estimate.
Rejection and infection with respect to age at time of transplant
As indicated by the multivariate analysis, striking differences were observed with respect to age at the time of transplant. Figure 5 shows a clear increase in cumulative rejection episodes in young adults (ages 10 to 30 years) with respect to other groups. In contrast, senior adults and children exhibited the lowest number of cumulative rejection episodes over time. Among these same groups of patients, this relationship was essentially inverted for infection, with the lowest rates of cumulative infection events experienced by young adults, and the highest infection rates displayed by senior adults transplanted at ≥60 years of age (Figure 6).
Figure 5.
Cumulative incidence of rejection in recipients stratified by age at the time of transplantation. The circles represent non-parametric estimates of cumulative rejection episodes. Vertical bars represent the 70% confidence limits of the non-parametric estimate.
Figure 6.
Cumulative incidence of infection in recipients stratified by age at the time of transplantation. The circles represent non-parametric estimates of cumulative rejection episodes. Vertical bars represent the 70% confidence limits of the non-parametric estimate.
To better understand these relative relationships, we performed a parametric hazard analysis of infection and rejection events across time and constructed mathematical models that enabled us to estimate the relative magnitude of cumulative rejection and infection within specific immunosuppressive eras as a continuous function of age at time of transplant. The general shape of these relationships curves can be seen in Figure 7a where cumulative rejection and infection is projected for a given patient at 1 year post-transplant in the year 1994, which is in the first immunosuppression era as discussed previously. Here we see that cumulative infection and rejection with respect to age are, as expected, largely inversely related. It is notable, however, that cumulative episodes are not constant with respect to age at time of transplant. For rejection, the lowest cumulative episodes are noted among the oldest and the youngest patients, whereas the inverse is seen for infection. In contrast, peak cumulative rejection and minimal cumulative infections are seen among young adults. For transplants performed in the most recent immunosuppressive era (Era 3), these relationships still exist even though the overall magnitude of cumulative events substantially decrease in the most recent era (Figure 7b). These differences with respect to age at the time of transplantation are exaggerated when data at 5 years post-transplantation are modeled (Figure 7c), suggesting that the highest number of rejections per patient could be found among young adults at 5 years post-transplantation, who also exhibited the lowest numbers of infection. In contrast, in children and senior adults, infection episodes predominate over rejection.
Figure 7.
The solution of multivariate equations depicting cumulative episodes of rejection or infection with respect to age at the time of transplantation in: (a) white recipients at 1 year post-transplantation, transplanted in 1994; (b) white recipients at 1 year post-transplantation, transplanted in 2004; and (c) white recipients at 5 years post-transplantation, transplanted in 2004. The dashed lines represent the 70% confidence intervals.
To determine whether the relative differences in rejection or infection within specific groups translate into differences in mortality, we performed competing outcomes analyses in which the causes of death were compared with respect to time post-transplantation, stratified by age at the time of transplantation. As seen in Figure 8a and b, the general shape of these relationships is similar to those shown in Figure 7, indicating that death due to rejection and infection was a reflection of the increased infection and rejection events per patient. Therefore, it is likely that a parametric hazard analysis of death due to rejection and infection would show similar results if enough data were available.
Figure 8.
Depictions of the proportion of patients dying from rejection or infection with respect to age at the time of transplantation at (a) 1 year or (b) 5 years post-transplantation.
Discussion
Rejection and infection are particularly useful for estimating the effects of demographic variables because they are the best known indicators of functional immune status in transplant recipients, and are also the basis upon which clinical decisions concerning immunosuppression are often made. Therefore, analysis of the data with respect to transplant date and age at the time of transplant can yield information about changes in immunity in transplant recipients with respect to age and whether newer immunosuppressive protocols have resulted in optimal outcomes for all age groups. Our data show that current immunosuppressive protocols are largely optimal for the majority of cardiac transplant patients, who are mature adults. Once outside this age range, however, we observed an apparent departure from the optimum. Mathematical modeling of the relationship of infection/rejection events to age at the time of transplant shows that relative differences in these events with respect to age remained relatively constant over the 17-year experience, despite the fact that the absolute number of events was reduced in the most recent era. Therefore, the observed differences, in which rejection increases and infection decreases in young adults while the extremes in age experience the opposite effect, are likely due to intrinsic biologic differences in the immune response of the recipient. Of particular interest are young adults, ages 10 to 29 years, in whom rejection/death occurred in nearly 20% of recipients by 5 years post-transplantation, whereas infection events were minimal, which suggests an active immune response in these individuals despite immunosuppression. The root cause of these responses is beyond the scope of this investigation, although we can speculate that there could be a number of contributing factors, including compliance issues, increased immune reactivity or age-dependent differences in the metabolism of immunosuppressive drugs. Further investigation of larger numbers of patients in a multi-institutional setting could possibly result in further insights, such as the role of center experience, which might be accessible by comparing relative rates of rejection/infection in similarly treated patients at widely different transplant dates.
A common feature of diseases in which the immune response, or lack thereof, plays a role in the pathology of diseases or susceptibility to them is that increased rates of morbidity and/or mortality were observed among the youngest and oldest segments of the population.19, 20 Similar phenomena have been observed in cardiac transplant patients as decreased rejection and increased rejection have been reported among the oldest and youngest recipients.13, 14, 15, 21, 22, 23, 24 We observed similar characteristics in this dataset in which rates of infection increased in the youngest and oldest patients, with a concomitant decrease in rejection. This observation is consistent with most previous reports of outcomes in patient groups large enough for sufficient statistical power. Most did not simultaneously analyze both events with respect to age across all three immunosuppressive eras; however, within specific subsets of patients, stratified by age, some similar observations were made. Among these studies, a decrease in rejection13, 14 or an increase in infection in infants was reported.16 Similarly, older patients (typically classified as >55 years of age at the time of transplant) also were reported to experience lower rates of rejection17, 25, 26 or increased rates of infection.22, 25, 26 Two studies are notable because of their larger numbers or the number of years in which the data were collected. The largest study, reported by Weiss and colleagues, included information from 14,401 patients in the UNOS database to show that patients >60 years of age experienced reduced rejection and increased infection rates.26 Information on immunosuppression and follow-up is limited in the UNOS database, but similar results were obtained in a large study by Tjang and colleagues using data obtained from 1,262 patients transplanted from 1998 to 2004.25 The collection of follow-up data, including morbidity, mortality, immunosuppression and patient demographics across 17 years of experience, allowed us to create mathematical models of the relationships between infection and rejection as a function of age at time of transplant as well as year of transplant.
In conclusion, our data show that evolving immunosuppressive strategies have successfully reduced rejection and infection frequencies in all age groups. Those patients transplanted between 30 and 60 years of age experience the lowest frequency of rejection or infection events. Individuals transplanted at younger ages, especially non-white recipients in the 10- to 30-year age group, experience significantly more rejection. In contrast, individuals transplanted at older ages (>60 years) experience significantly more infection. Therefore, programs should increase the level of surveillance in these patients and consider tailoring of immunosuppressive regimens to lower the frequency of infection and rejection events among the different age groups.
Disclosure statement
The authors have no conflicts of interest to disclose.
References
- Heart transplant coronary artery disease detected by coronary angiography: a multiinstitutional study of preoperative donor and recipient risk factors. J Heart Lung Transplant. 1998;17:744–753
- Outcome of listing for heart transplantation in infants younger than six months: predictors of death and interval to transplantation. J Heart Lung Transplant. 1997;16:1255–1266
- Infection after pediatric heart transplantation: results of a multiinstitutional study. J Heart Lung Transplant. 1997;16:1207–1216
- Heart transplant rejection with hemodynamic compromise: a multiinstitutional study of the role of endomyocardial cellular infiltrate. J Heart Lung Transplant. 1997;16:813–821
- Risk factors for early, cumulative, and fatal infections after heart transplantation: a multiinstitutional study. J Heart Lung Transplant. 1996;15:329–341
- Prognostic determinants of six-month morbidity and mortality in heart transplant recipients. J Heart Lung Transplant. 1996;15:124–135
- Risk factors for late recurrent rejection after heart transplantation: a multiinstitutional, multivariable analysis. J Heart Lung Transplant. 1995;14:409–418
- Influence of HLA mismatch on rejection after heart transplantation: a multiinstitutional study. J Heart Lung Transplant. 1994;13:583–595
- Cytomegalovirus after heart transplantation (Risk factors for infection and death: a multiinstitutional study). J Heart Lung Transplant. 1994;13:394–404
- Infection after heart transplantation: a multiinstitutional study. J Heart Lung Transplant. 1994;13:381–392
- Pretransplantation risk factors for death after heart transplantation: a multiinstitutional study. J Heart Lung Transplant. 1993;12:549–562
- Pretransplantation risk factors for acute rejection after heart transplantation: a multiinstitutional study. J Heart Lung Transplant. 1993;12:355–366
- Rejection is reduced in thoracic organ recipients when transplanted in the first year of life. J Heart Lung Transplant. 2002;21:311–318
- Late rejection episodes more than 1 year after pediatric heart transplantation: risk factors and outcomes. J Heart Lung Transplant. 2003;22:869–875
- Long-term results of heart transplantation in patients older than 60 years. J Thorac Cardiovasc Surg. 2003;126:224–231
- Infectious, malignant, and autoimmune complications in pediatric heart transplant recipients. J Pediatr. 2008;152:671–677
- Seventeen-year experience with 1,083 heart transplants at a single institution. Ann Thorac Surg. 2002;74:1558–1566
- . The decomposition of time-varying hazard into phases, each incorporating a separate stream of concomitant information. J Am Stat Assoc. 1986;81:615–624
- . The first pandemic of the 21st century: a review of the 2009 pandemic variant influenza A (H1N1) virus. Postgrad Med. 2009;121:43–47
- . Observations on mortality during the 1918 influenza pandemic. Clin Infect Dis. 2001;33:1375–1378
- Heart transplantation in patients seventy years of age and older: a comparative analysis of outcome. J Thorac Cardiovasc Surg. 2001;121:532–541
- An analysis of the effect of age on survival after heart transplant. J Heart Lung Transplant. 1999;18:668–674
- Long-term outcomes of heart transplantation in older recipients. J Heart Lung Transplant. 2008;27:830–834
- Heart transplantation in children: clinical outcomes in a single center. Ann Thorac Surg. 2007;84:1640–1644
- Impact of recipient's age on heart transplantation outcome. Ann Thorac Surg. 2008;85:2051–2055
- Outcomes in patients older than 60 years of age undergoing orthotopic heart transplantation: an analysis of the UNOS database. J Heart Lung Transplant. 2008;27:184–191
PII: S1053-2498(10)00284-6
doi:10.1016/j.healun.2010.05.003
© 2010 International Society for Heart and Lung Transplantation. Published by Elsevier Inc. All rights reserved.
Volume 29, Issue 9 , Pages 966-972, September 2010
