Abstract
ABSTRACT. The impact of graft loss on acute coronary syndromes (ACS) after renal transplantation has not been studied in a national population. It was hypothesized that ACS might be more frequent after graft loss, as many of the benefits of a functioning allograft on metabolism and volume regulation would be lost. Data from the 2000 United States Renal Data System (USRDS) was used to conduct an historical cohort study of ACS in 14,237 patients who received renal transplants between April 1, 1995, and June 30, 1998, (followed until April 28, 2000) with valid information from CMS Form 2728, excluding patients with hospitalized ACS before renal transplant. Cox nonproportional regression models were used to calculate the time-dependent adjusted hazard ratio (AHR) of graft loss (censored for death) for time-to-first hospitalization for ACS (International Classification of Diseases 9th Modification Diagnosis Codes [ICD9] code 410.x or 411.x) occurring after transplant. The incidence of ACS was 12.1 per 1000 patient-years (PY) in patients after graft loss versus 6.5 per 1000 PY after transplantation (excluding patients with graft loss). As a time-dependent variable, graft loss had an AHR of 2.54 (95% confidence interval, 1.09 to 5.96; P = 0.031 by Cox regression). Other risk factors associated with ACS included diabetes, older recipient, and male recipient. Allograft rejection was NS. Renal transplant recipients share some of the risk factors for ACS with the general population. In addition, graft loss was identified as a unique risk factor for ACS in this population. E-mail: kevin.abbott@na.amedd.army.mil
Cardiovascular disease is a substantial health problem in renal transplant recipients (1–3). Renal transplant recipients have many conventional risk factors for acute cardiovascular disease, including hypertension, hyperlipidemia, and posttransplant diabetes mellitus. The relationship between these risk factors and ischemic heart disease after renal transplantation has been assessed in single-center studies (4–6). However, many of these conventional risk factors may be influenced or mediated by transplant immunosuppression (7). Renal transplant recipients are also at much greater risk of deteriorating renal function than the general population (8). Renal insufficiency has recently been identified as an independent risk factor for recurrent coronary artery disease in the general population (9). Improvement in renal function in patients with dialysis-dependent renal failure, such as occurs after renal transplantation in patients on the renal transplant waiting list, has been independently associated with a lower risk of hospitalized acute coronary syndromes (10).
Graft loss represents a possible intersection between the risk of immunosuppression and renal insufficiency and has been associated with both death(11) and hospitalized congestive heart failure (12) after renal transplantation. We hypothesized that ACS might be more frequent after graft loss, as many of the benefits of a functioning allograft on metabolism and volume regulation would be lost, and might interact with the detrimental effects of transplant immunosuppression. Therefore, we conducted an historical cohort study, using a national registry (the 2000 United States Renal Data System [USRDS]), of the independent association of graft loss after renal transplantation with time-to-hospitalization for acute coronary syndromes.
Materials and Methods
We conducted an historical cohort study of the independent associations of time-to-ACS as a primary diagnosis at hospital discharge occurring before versus after incident graft loss (censored for death) in renal transplant recipients. Hospitalizations were chosen because they were more accessible in the database and less subject to interpretation than outpatient cases of ACS, especially because the USRDS database has no information on radiographic studies and ACS are not managed on outpatient basis in the United States. The variables included in the USRDS data files, as well as data collection methods and validation studies, are listed at the USRDS website, under “Researcher’s Guide to the USRDS Database,” Section E, “Contents of all the SAF’s (Standard Analysis Files)” (13), and published in the USRDS. The demographics of the renal transplant population have been previously described (2001 USRDS report). SAF.TXUNOS and SAF.TXFUUNOS were selected for base and follow-up data, respectively. SAF.HOSP was used for hospitalization information, and SAF.PATIENTS was used for dates of death. Cause-of-death data are incomplete for renal transplant recipients. Although other authors have used sensitivity analyses to address this limitation (14), this requires several assumptions. We instead elected to use hospitalization data for our primary statistical analysis, because valid hospital coding is required for physician reimbursement, and hospitalization data may provide additional information on morbidity. SAF.MEDEVID was used for information on comorbidity within 10 yr of dialysis initiation and laboratory data at the time of dialysis initiation starting on or after April 1, 1995. Hospitalization data for transplant recipients may be unreliable ≥3 yr posttransplant, when hospitalization reporting to Medicare is no longer required. In contrast to dialysis patients, however, Medicare reporting starts immediately after transplant. The present study limited analysis to the first kidney transplant occurring in an individual recipient between April 1, 1995, and June 30, 1998, with valid data from CMS Form 2728, counting hospitalizations for ACS up to 3 yr posttransplant. Patients were followed until April 28, 2000. Recipients of organs other than kidneys were excluded. Hospitalizations for ACS occurring at any time after renal transplant, including after graft failure, were counted in analysis in an intention to treat fashion.
Analytic Variables and Outcome Measures
The outcome variables were based on International Classification of Diseases 9th Modification Diagnosis Codes (ICD9) for procedures at hospital discharge for ACS: 410.x or 411.x, because these conditions have similar pathogenesis and treatment (15). To ensure these were incident and not previous hospitalizations, only hospitalizations with a primary discharge procedure code for ACS within 3 yr after the date of renal transplantation were selected. Patients with myocardial infarction after dialysis are generally not considered suitable candidates for surgery until at least 6 mo after the event (16) and were therefore excluded. Patient characteristics and treatment factors were those at the date of transplant. Information on use or results of cardiac catheterization, treadmill testing, echocardiography, electrocardiogram, serum isoenzymes, serum albumin levels, or chest radiographs were not available. BP levels and blood lipid levels were also not available. The USRDS information on medications did not include total dose, and because almost all recipients were on corticosteroid therapy (17), analysis of corticosteroids was not included in the present analysis. Patient characteristics and treatment factors were those at the date of transplant.
All analyses were performed using SPSS 9.0 TM (SPSS, Inc., Chicago, IL). Files were merged and converted to SPSS files using DBMS/Copy (Conceptual Software, Houston, TX). Statistical significance was defined as P < 0.05. For continuous variable, values above or below 3 SD from the mean were excluded from analysis. For categorical variables, missing or ambiguous values were excluded from analysis. Univariate analysis of factors associated with primary hospitalizations for ACS was performed with χ2 testing for categorical variables and t test for continuous variables, respectively. Variables with P < 0.10 in univariate analysis for a relationship with development of a first hospitalization for ACS after renal transplantation in the study period were entered into multivariate analysis as covariates. An exception was made for factors thought to have a clinical reason to be associated with ACS, in accordance with established epidemiologic principles (18).
Covariates used in analysis included donor and recipient age, race, gender, weight, body mass index (calculated from height and weight), induction, maintenance, and antirejection immunosuppressive medications, previous transplant, delayed graft function, network, state of transplant, hospitalization with a primary procedure code for coronary revascularization before transplant (including percutaneous transluminal angioplasty, stenting, or coronary artery bypass surgery), and duration of dialysis before transplantation. Because multiple episodes of rejection in the first year after transplant have been associated with cardiovascular events in previous studies (4,17), we analyzed rejection separately as all cases occurring during the study period and as multiple rejection episodes occurring in the first year after transplant. Because exact dates of graft loss were known, graft loss was analyzed as a time-dependent variable in a nonproportional Cox regression model, coded as 1 for all events occurring after graft loss versus 0 for all events occurring before graft loss or in transplant recipients who did not experience graft loss. Time-dependent analysis of rejection episodes was not performed because the exact dates of rejection were not specified in the USRDS database, other than those occurring within defined blocks of follow-up. Causes of end-stage renal disease were diabetes, hypertension, and glomerulonephritis. Comorbidity and laboratory data were obtained from CMS Form 2728. Continuous variables were also analyzed as quartiles to assess for a nonlinear relationship between these values and ACS.
Graft survival time was calculated as the time from the date of transplant until the date of graft loss after transplant. Dates of graft loss before the study period (from prior transplants) were not included in analysis. Patients were censored at death, most recent follow-up, or the end of the study. Time-to-hospitalization for ACS was calculated as the time of transplant until the first hospitalization for ACS, with recipients censored (removed from analysis) at time of death, loss to follow-up, or the end of the study period. Survival time was calculated as the time form the date of transplant until the date of death, with recipients censored at most recent follow-up or the end of the study period. Graft loss was not censored. Multivariate analysis excluded all patients with missing values, resulting in substantially smaller models than the entire study population.
Hospitalizations for ACS were examined using multivariate analysis with stepwise nonproportional hazards Cox regression (19) for time until the first hospitalization for ACS during the study period, from which adjusted hazard ratios (AHR) for each covariate were calculated. Three additional analyses were performed because of potential bias. Because graft loss occurred after renal transplantation and risk of ACS might increase with time, analysis was also performed limited to recipients who experienced graft loss compared with a cohort of patients who did not experience graft loss but had survived at least as long as the median time to graft loss. This analysis was also performed excluding patients with a history of myocardial infarction on CMS Form 2728. Because of potential selection bias resulting from limiting analysis to recipients with valid data from CMS Form 2728, we also performed an analysis on the entire cohort of solitary renal transplant recipients from April 1, 1995, to June 30, 1998, adjusted for all factors in the models above except for variables in CMS Form 2728. To assess for differences in the study population versus the entire cohort of transplant recipients from April 1, 1995, to June 30, 1998, stepwise logistic regression was performed for factors associated with valid information from CMS Form 2728 (using a valid history of myocardial infarction) versus cases where this information was missing, using the same covariates as for Cox regression above.
Results
Of 34,977 recipients of solitary renal transplants from April 1, 1995, to June 30, 1998, 32,141 had valid follow-up times. Of these, 696 were hospitalized for ACS after initiation of dialysis and before transplant and were excluded. Of the remaining patients, 14,237 had data available from HCFA Form 2728 including comorbidity and laboratory data, specifically the presence or absence of a history of myocardial infarction before initiation of dialysis. Of 133 patients hospitalized with ACS after transplantation from this cohort, 83% were hospitalized once, 14% twice, 2.1% three times, and 1.3% more than three times after transplant. Table 1 shows the event rates for ACS after transplantation for both the entire cohort and for the study population. All-cause survival for the cohort was 96.2% at 1 yr and 89.9% at 3 yr. Death-censored graft survival was 97.2% at 1 yr and 93.8% at 3 yr.
Table 1. Event rates for acute coronary syndromes after renal transplantation, April 1, 1995, to June 30, 1998a
Rates of ACS are shown in Table 1. Unadjusted rates of ACS were higher after graft loss than before or in patients who did not experience graft loss. As shown, unadjusted rates of ACS were greater in the cohort of recipients transplanted between April 1, 1995, and June 30, 1998, than in the study population, which may be due to differences in the study population (Table 2).
Table 2. Factors associated with Acute Coronary Syndromes (ACS) after renal transplantation, renal transplant recipients presenting to ESRD April 1, 1995, to June 30, 1998
Table 2.—Continued
Characteristics of the study population and univariate associations with ACS are shown in Table 2. Compared with all patients transplanted between April 1, 1995, amd June 30, 1998, the study population included fewer recipients of cadaver kidneys, fewer African Americans, and fewer patients with delayed graft function, allograft rejection, and graft loss, whereas more patients had diabetes, younger age, more recent year of first ESRD treatment and transplant, and more repeat transplant recipients. Of patients with graft loss, 48% had previous rejection, and 18% of patients with rejection later developed graft loss. Factors significantly associated with ACS in univariate analysis included increased risk of ACS in recipients who were male, diabetic, older, who weighed more (but not those with higher BMI), history of coronary revascularization, longer cold ischemic time, graft loss, cadaveric donor, delayed graft function, history of cardiovascular disease, lower albumin, chronic lung disease, and Medicare eligibility.
Table 3 shows factors associated with ACS in Cox regression. As shown, graft loss was independently associated with ACS, along with older year of transplant and increasing duration of follow-up after transplant. Neither the year of first ESRD service nor the duration of waiting time before transplant was significant. Neither rejection occurring at any time during the study period nor multiple episodes of rejection in the first year were associated with ACS in either univariate or multivariate analysis. Among demographic factors, increased age and diabetes were significant. No prior comorbidities were significant after all other factors were assessed. The interaction between diabetes and female gender was statistically significant, but there were no other significant interactions, specifically between diabetes and graft loss or gender and graft loss.
Table 3. Cox regression analysis of factors associated with acute coronary syndromes after renal transplantationa
Results of stratified models, performed to assess for possible bias in our model, are shown in Table 4. Graft loss was still significant when all the above analyses were performed, excluding patients with a history of myocardial infarction on CMS Form 2728 (n = 13,936).
Table 4. Cox regression analysis of factors associated with acute coronary syndromes after renal transplantation, stratified modelsa
Figure 1 shows a Kaplan-Meier plot of the unadjusted time-to-hospitalization for ACS for the study population, comparing time-to-ACS after graft loss versus time-to-ACS after transplant excluding patients who eventually developed graft loss. Figure 2 shows a similar plot adjusted for age, race, gender, diabetes, history of myocardial infarction, year of first dialysis, and time since transplantation. Figures 3 and 4 show similar plots for the entire cohort. Figure 5 shows unadjusted mortality after ACS stratified by graft loss for the study population, which did not differ significantly in Cox regression. In analysis performed on the entire cohort of solitary renal transplant recipients from April 1, 1995, to June 30, 1998 (n = 32,141), graft loss was similarly associated with ACS as a time-dependent variable (AHR, 2.99; 95% CI, 2.18 to 4.10; P < 0.0001). In the larger cohort, other factors independently associated with ACS were diabetes, older recipient age, male gender, elevated body mass index, earlier year of dialysis, and maintenance rapamycin use.
Figure 1. Time to hospitalizations for acute coronary syndromes (ACS) in years, occurring after graft loss censored for death (GL) versus occurring after transplant excluding patients who developed graft loss (Tx) respectively. United States renal transplant recipients between April 1, 1995, and June 30, 1998, with valid data from CMS Form 2728 (n = 14,237). ACS developed much more rapidly after graft loss than after transplant (P = 0.004 by log rank test). Graft loss was also independently associated with a higher rate of ACS as a time-dependent variable in Cox regression analysis (Table 5).
Table 5. Patients at risk, Figure 1
Figure 2. Time-to-hospitalizations for ACS in years, occurring after graft loss censored for death (GL) versus occurring after transplant censoring events after graft loss (Tx), adjusted for age, race, gender, diabetes, history of myocardial infarction or ischemic heart disease, year of first dialysis, and follow-up time after transplant. United States renal transplant recipients between April 1, 1995, and June 30, 1998, with valid data from CMS Form 2728 (n = 14,237). ACS developed much more rapidly after graft loss than after transplant (P = 0.004 by log rank test). Graft loss was also independently associated with a higher rate of ACS as a time-dependent variable in Cox regression analysis.
Figure 3. Time-to-hospitalizations for ACS in years, occurring after graft loss censored for death (GL) versus occurring after transplant censoring events after graft loss (Tx). All United States renal transplant recipients between April 1, 1995, and June 30, 1998 (n = 32,141). ACS developed much more rapidly after graft loss than after transplant (P < 0.001 by log rank test) (Table 6).
Table 6. Patients at risk, Figure 3
Figure 4. Time to hospitalizations for acute coronary syndromes (ACS) in years, occurring after graft loss censored for death (GL) versus occurring after transplant excluding patients who developed graft loss (Tx), adjusted for age, race, gender, diabetes, year of first dialysis, and follow-up time after transplant. United States renal transplant recipients between April 1, 1995, and June 30, 1998 (n = 32,141). ACS developed much more rapidly after graft loss than after transplant (P < 0.001 by Cox regression). Graft loss was also independently associated with a higher rate of ACS as a time-dependent variable in Cox regression analysis.
Figure 5. Mortality after hospitalizations for ACS in years, occurring after graft loss censored for death (GL) and after transplant excluding patients with graft loss (Tx). United States renal transplant recipients between April 1, 1995, and June 30, 1998, with valid data from CMS Form 2728 (n = 14,237). Patients with GL died after ACS at a higher rate than Tx, but was not statistically significant (P = 0.42 by log rank test) (Table 7).
Table 7. Patients at risk, Figure 5
Because allograft rejection was not associated with ACS, the validity of its measurement in this study was tested by its potential associations with graft loss and all-cause mortality. As expected, allograft rejection (both as measured as all events occurring in the study and as multiple events in the first posttransplant year) was independently associated with both graft loss and all-cause mortality.
Mortality after ACS was 10.7% at 30 d and 38% at 2 yr in the study population. Mortality after ACS was higher in recipients who eventually experienced graft loss but was not statistically significant (Figure 2). Graft loss was also independently associated with all-cause mortality as a time-dependent variable (AHR, 6.39; 95% CI, 5.47 to 7.46). Graft loss was not significantly associated with specific causes of ACS, either acute myocardial infarction (ICD9 410.x) or unstable coronary syndromes (ICD9 411.x).
Specific causes of death were missing or unknown for 55% of recipients. However, factors associated with cardiac death were similar to those associated with ACS; specifically, graft loss was significantly associated with cardiac death (AHR, 3.94; 95% CI, 2.02 to 7.75; P = 0.0001).
Discussion
This historical cohort study of an essentially complete national population of renal transplant recipients demonstrated that hospitalized acute coronary syndromes were significantly more common after graft loss (AHR, 2.54) than either before graft loss or in transplant recipients who did not experience graft loss. Mortality after myocardial infarction (3) and demographic factors in the study population were similar to previous USRDS reports except for a lower proportion of cadaveric donors, as documented in Table 2. This may represent living donor recipients being overrepresented in our study due to their shorter waiting times for transplant and greater likelihood of having CMS Form 2728 completed before transplant within the time frame of the current study. It is possible this introduced bias in the study, because recipients of living donor have lower rates of graft loss and better survival than recipients of cadaver kidneys (2). Although previous studies have associated graft loss with increased all-cause mortality, none had found that graft loss had an independent association with acute coronary heart disease after renal transplantation. Graft loss was the only transplant-specific risk factor other than year of transplant and duration of follow-up after transplantation that persisted in multivariate analysis (Table 3). The significant negative association between duration of follow-up after transplantation and ACS can be explained by the increased mortality associated with ACS (Figure 5). Thus, patients who survived longer were much less likely to have ACS. This should not be interpreted to mean that ACS was less likely to occur with longer durations after transplant, as is clear from Figures 1 to 4.
Graft loss uniquely represents a possible convergence of the cardiovascular risks associated with both transplant immunosuppression and dialysis-dependent renal failure. Previous studies have found that over 50% of causes of death after graft loss were attributed to cardiovascular disease, just as in the entire renal transplant population, arguing against a predominant role for infectious death (11). Only one other study has shown an association of graft loss with cardiovascular outcomes (congestive heart failure) (12). Because prospective studies may not be available in renal transplant recipients who experience graft loss, registry data may be needed to suggest the mechanism of increased mortality in this population. The reasons why renal failure may be associated with increased rates of cardiovascular events have been the subject of a previous review (20). Authors of another review were “convinced that currently identified risk factors do not fully explain the abysmal cardiovascular risk of the uremic patient.” (21) The relative risk of graft loss for ACS of 2.54 in the present study is similar to the relative risk of renal transplantation in comparison with diabetic patients on maintenance dialysis on the renal transplant waiting list (0.38, the reciprocal of which is 2.63) (10). Because the present study was weighted toward recipients of living donor transplants, the impact of graft loss on ACS may be underestimated in the present study despite its use of comorbidity data.
The increased risks of older age, male gender, diabetes, and its interaction with female gender were also consistent with previous studies (4). Because BP and lipid levels were not measured, the Framingham prediction model could not be used, although the limitations of the Framingham model should be considered (22). Outcomes in this study were less sensitive and more specific in comparison with the Framingham model to maximize the specificity of risk factor analysis, mainly to avoid finding an association between graft loss and ACS when none existed.
In contrast, allograft rejection was not significantly associated with ACS, despite its potential association with all-cause mortality and graft loss. It has been postulated that allograft rejection (through renal insufficiency and possible volume overload) and/or its treatment with high-dose corticosteroids may raise BP and alter lipid levels (23), both risk factors for cardiovascular events. Kasiske et al. (4) found that two or more rejection episodes in the first post transplant year were associated with increased risk of ischemic heart disease outcomes, Rigatto et al. (5) also found that rejection was also associated with ischemic heart disease after renal transplantation. However, the national incidence of allograft rejection has declined considerably (8) since the time frame of both those studies, which included considerable numbers of patients transplanted before 1995. The lack of significance of rejection in the present analysis may also reflect heterogeneity in the way allograft rejection is diagnosed and treated in a national registry versus single-center studies (24), differences in endpoints between studies, or differences in the populations studied and the ways in which they react to rejection and its treatment. Neither study assessed the association between graft loss and coronary heart disease events after transplantation. Other factors independently associated with graft rejection, such as donor age and race (25), and use of antirejection antibodies were also NS in analysis of ACS, suggesting that either renal insufficiency itself or unidentified risk factors co-segregating with graft rejection are responsible for the increased risk of ACS.
The lack of significance of hematocrit, albumin and other laboratory values in the present analysis may be related to the inability of the present study to measure these values after the start of dialysis. This is significant considering the considerable waiting time from initiation of dialysis to receipt of a renal transplant, or changes from increased levels of erythropoietin.
This study has several limitations. Findings are associative, not causative. Laboratory, invasive, and radiologic studies could not be independently verified. Although ICD-9 codes are subject to errors in miscoding and duplication, their use provided the only means of assessing the frequency of acute coronary syndromes in the USRDS. Unfortunately, use of primary discharge codes for ACS might miss ACS occurring in the perioperative or immediate posttransplant phase, but use of non-primary diagnoses could have introduced historical, not active, diagnoses into the study. This is therefore one potential limitation that could have underestimated the early occurrence of ACS after transplant. ICD-9 codes have been used for determining rates of medical conditions in other studies (3) and are likely to be at least as reliable as outcomes used in recent studies (14,26,27). Regardless, use of hospitalizations and, as a result, ICD-9 codes to determine the frequency of acute coronary syndromes in the ESRD population is likely to produce an underestimate of disease prevalence, because both mild cases and pre-hospital fatalities are underrepresented. We specifically could not assess the frequency of sudden cardiac death, which may be a major presentation of unstable coronary syndromes. However, analysis based on causes of death in renal transplant recipients may be limited not only because of the large percentage of missing causes, but also because, in the absence of autopsy verification, sudden death cannot necessarily be distinguished between coronary artery plaque rupture and thrombosis versus other causes of sudden death common in renal transplant recipients (28–31).
Reliance on CMS Form 2728 for comorbidity data could have led to selection bias as shown in Table 2, although we also validated our findings in a larger cohort of renal transplant recipients. Limitations of Form 2728 have been addressed in other studies; however, the form’s accuracy is highest for cardiovascular diseases (32). We were unable to assess dialysis adequacy or dialysis modality after graft loss, although the utility of such information in comparison with patients with functioning renal allografts is unproven. We could not assess the impact of BP or lipid levels, which are important limitations. The short follow-up of the study disproportionately impacted patients who experienced graft loss. As a result, the beneficial effects of renal transplantation may be weakened in this study, and patients with early graft loss more frequently have acute complications and comorbidities not taken into account in this study that might have influenced the occurrence of ACS. Despite these limitations, our analysis is strengthened by the completeness and large size of the database, its population-based character, and its relatively complete follow-up. Because patients were at increased risk of all-cause mortality after graft loss, survival bias is an unlikely explanation for their higher rate of ACS.
In conclusion, analysis of the national renal transplant population confirms that while renal transplant recipients share some of the risk factors for ACS with the general population, the occurrence of dialysis-dependent renal failure after transplantation, as determined by graft loss, is a major risk factor for ACS in this population. This association may be accounted for by a higher rate of traditional risk factors, such as hypertension and hyperlipidemia after graft loss, as well as new risk factors, such as inflammation that may be increased by chronic immunosuppression, or as yet unidentified mechanisms. Because graft loss is so common after renal transplantation, it should also be considered an endpoint of interest in future studies of cardiovascular disease in renal transplant recipients.
Footnotes
-
The opinions are solely those of the authors and do not represent an endorsement by the Department of Defense or the National Institutes of Health. This is a U.S. Government work. There are no restrictions on its use.
- © 2002 American Society of Nephrology
References
Jump to section
More in this TOC Section
Cited By...
- Risks of Death and Graft Failure After Surgical Versus Percutaneous Coronary Revascularization in Renal Transplant Patients
- Outcomes Associated with Influenza Vaccination in the First Year after Kidney Transplantation
- Incidence, Predictors, and Associated Outcomes of Prostatism after Kidney Transplantation
- Variations in the Risk for Cerebrovascular Events after Kidney Transplant Compared with Experience on the Waiting List and after Graft Failure
- Incidence, Predictors, and Associated Outcomes of Atrial Fibrillation after Kidney Transplantation
- Incidence and Predictors of Myocardial Infarction after Kidney Transplantation
- Early Renal Insufficiency and Hospitalized Heart Disease after Renal Transplantation in the Era of Modern Immunosuppression