Renal Dysfunction Is a Strong and Independent Risk Factor for Mortality and Cardiovascular Complications in Renal Transplantation

Authors


*Corresponding author: Bengt Fellström, bengt.fellstrom@medsci.uu.se

Abstract

Renal transplant recipients (RTR) have shortened life expectancy, primarily due to premature cardiovascular disease (CVD). Traditional CVD risk factors are highly prevalent. In addition, several non-traditional risk factors may contribute to the high risk. The aim of the study was to evaluate the effects of renal dysfunction on mortality and cardiovascular complications in 1052 placebo-treated patients of the Assessment of LEscol in Renal Transplantation (ALERT) trial. Follow-up was 5–6 years and endpoints included cardiac death, non-cardiovascular death, all-cause mortality, major adverse cardiac event (MACE), non-fatal myocardial infarction (MI) and stroke. The effects of serum creatinine at baseline on these endpoints were evaluated. Elevated serum creatinine in RTR was a strong and independent risk factor for MACE, cardiac, non-cardiovascular, and all-cause mortality, but not for stroke or non-fatal MI alone. Serum creatinine was associated with increased mortality and MACE, independent of established CVD risk factors. Graft loss resulted in increased incidences of non-cardiovascular death, all-cause mortality, MACE and non-fatal MI. In conclusion, elevated serum creatinine is a strong risk factor for all-cause, non-cardiovascular and cardiac mortality, and MACE, independent of traditional risk factors, but not for stroke or non-fatal MI alone.

Introduction

Despite novel immunosuppressive agents and improvements in overall renal transplantation results, renal transplant recipients (RTR) have a shortened life expectancy (1,2). In addition to increased susceptibility to infections and malignant diseases, these patients mainly die of premature cardiovascular disease (CVD) (1,3–6). The prevalence of CVD risk factors is high in RTR and many patients have pre-existing CVD, diabetes, hypertension or dyslipidaemia at the time of transplantation. Following transplantation, immunosuppressive therapy may aggravate existing factors or promote the development of new risk factors (7–11). The duration of dialysis therapy prior to transplantation has also been shown to have a negative influence on patient survival (12).

It has been shown in patients with chronic renal failure, that classical Framingham risk factors are highly prevalent but qualitatively or quantitatively different in their relationship with CVD outcomes. This is believed to be due to the influence of non-traditional risk factors (13–15). The incidence and characteristics of CVD in renal patients differs from the general population (8,16). Coronary arteries in patients with end-stage renal disease are characterized by increased media thickness and plaques are markedly calcified compared to mainly fibroatheromatous lesions in non-uremic patients (17). In RTR serum creatinine has a strong association with graft failure, which is associated with increased cardiac and all-cause mortality (18–20). In the present study we investigated the impact of renal transplant dysfunction at baseline and graft loss during follow-up on endpoints such as cardiovascular events and mortality in the the Assessment of LEscol in Renal Transplantation (ALERT) trial. All endpoints were adjudicated by a Critical Events Committee (CEC).

Methods

The ALERT trial was an investigator-initiated and investigator-led, randomized, double-blind, parallel group study designed to investigate the effects of fluvastatin on cardiac and renal endpoints in RTR. The ALERT study design, baseline data and outcomes have been previously published (21,22).

Participants

Briefly, 2102 adult RTR were recruited from nephrology and transplant clinics in Belgium, Denmark, Finland, Norway, Sweden, Switzerland, the UK and Canada. The patients had received renal or combined renal and pancreas transplants >6 months prior to randomization. All patients were on cyclosporin-based immunosuppression, but no one received statins prior to inclusion. Total fasting cholesterol ranged from 4–9 mmol/L (4–7 mmol/L for those with previous cardiac event). Patients, who had had an acute rejection episode in the previous 3 months, or who had a predicted life expectancy of less than 1 year were excluded. The patients recruited to the study were of low risk, reflected in a low event rate and the use of statins in high-risk patients even in 1996, when recruitment began (22). The recorded endpoints included cardiac death, non-cardiovascular mortality, all-cause mortality, non-fatal myocardial infarction (MI), stroke, and MACE (Major Adverse Cardiac Event defined as cardiac death, non-fatal MI or coronary revascularization procedure). CEC consisted of two nephrologists and two cardiologists who were unaware of treatment assignment. All endpoints were adjudicated by the CEC and classified after agreement by consensus majority vote (22). The present analysis was performed in 1052 patients in the placebo arm only, because this was considered to be the ‘cleanest’ approach, preferable to including and adjusting for treatment arm.

The study adhered to the International Conference On Harmonisation Guidelines and for good clinical practice and was done in accordance with the Declaration of Helsinki. All participants provided written informed consent, and the ethics committee at each participating center approved the trial.

Statistical analysis

The statistical analysis plan of the main study has been described previously (22). Univariate and multivariate Cox proportional hazards models were used to analyze the impact of serum creatinine on predefined endpoints, calculated as risk ratio (RR) per 100 μmol/L (≈1.12 mg/dL) serum creatinine increment. Logistic regression model was used to calculate the probability of experiencing events at creatinine levels. Multivariate risk factor analysis was made for endpoints with potentially important variables including age, diabetes, previous transplant rejection, smoking, previous coronary heart disease, peripheral vascular disease, cerebrovascular disease, left ventricular hypertrophy, number of transplants, LDL-cholesterol, HDL-cholesterol, polycystic kidney disease, serum creatinine, systolic blood pressure, pulse pressure, HLA-DR mismatch, body mass index, proteinuria, time on dialysis prior to transplantation and time since transplantation. Between-group differences were assessed with χ2. All analyses were performed using the SAS statistics package. p < 0.05 was regarded significant.

Results

Between June 1996 and October 1997, 2102 patients were recruited to the ALERT trial, of whom, 1052 were randomly assigned to the placebo arm and followed for 5–6 years. During that time, there was a ‘drop-in’ rate of 14% for statin use in this arm, compared with 7% in the active treatment (fluvastatin 40–80 mg) arm, mainly occurring late in the study. The patients in the placebo arm of the study experienced 54 cardiac deaths. In addition, there were 66 patients with definite MI and 65 patients died of non-cardiovascular causes. The baseline demographic and clinical characteristics of the placebo group are presented in Table 1.

Table 1.  Baseline demographic and clinical characteristics of the placebo group (n (%) or mean ± SD)
Demographic and
clinical characteristics
Placebo
(n = 1052)
  1. ACE = angiotensin-converting enzyme, AIIRA = angiotensin-II-receptor blocker.

  2. *Taken at least once during study.

Age, years50.0 ± 11.0
Male686 (65.2%)
Diastolic blood pressure, mmHg85.6 ± 10.0
Systolic blood pressure, mmHg144.0 ± 19.1
Body mass index, kg/m225.8 ± 4.6
LDL-cholesterol, mmol/L (mg/dL)4.1 ± 1.0 (159 ± 39)
HDL-cholesterol, mmol/L (mg/dL)1.4 ± 0.4 (54 ± 15)
Triglycerides, mmol/L (mg/dL)2.2 ± 1.5 (196 ± 134)
Serum creatinine, μmol/L (mg/dL)142± 51 (1.58 ± 0.58)
Diabetes mellitus199 (19%)
Time taking renal replacement therapy, months89 ± 58
Time on dialysis prior to transplantation, months28 ± 42
First transplantation900 (85.6%)
Type of last transplant: live donor229 (21.8%)
Type of last transplant: cadaveric donor822 (78.1%)
Hypertension777 (73.9%)
Current smoker185 (17.6%)
History of angina pectoris77 (7.3%)
Previous myocardial infarction34 (3.2%)
History of cerebrovascular disease60 (5.7%)
History of peripheral vascular disease78 (7.4%)
Known family history of coronary heart disease124 (11.8%)
Concomitant immunosuppressive therapy*:
Azathioprine680 (64.6%)
Prednisolone848 (80.6%)
Cyclophosphamide10 (1.0%)
Mycophenolate mofetil159 (15.1%)
Other224 (21.3%)
Concomitant cardiovascular medication*:
Any cardiovascular drug999 (95%)
Acetylsalicylic acid353 (33.6%)
Dipyridamole26 (2.5%)
Coumarin or warfarin94 (8.9%)
β-blockers627 (59.6%)
Calcium antagonists738 (70.2%)
ACE inhibitors or AIIRA529 (50.3%)
Diuretics573 (54.5%)
α-blockers170 (16.2%)
Other373 (35.5%)

The probabilities of reaching endpoints in relation to baseline creatinine levels are depicted in Figure 1. The increases in cardiac death, non-cardiovascular death, all-cause mortality and MACE were most obvious at serum creatinine levels above 200 μmol/L (Figure 1). We used this cut-off to divide the placebo group into two. In the high creatinine group, the incidence of cardiac death, non-cardiovascular death, all-cause mortality and MACE were clearly increased compared to RTR with a serum creatinine below 200 μmol/L (Figure 2). For example, the cardiac death rate was 4% in those with a serum creatinine below 200 μmol/L and 13.2% (p < 0.0001) in those with a serum creatinine above 200 μmol/L. The corresponding values for non-cardiovascular death were 4.7% and 16% (p < 0.0001); for all-cause mortality, 10.1% and 34.9% (p < 0.0001) and for MACE, 11.5% and 18.9% (p < 0.0001), respectively (Figure 2).

Figure 1.

Endpoint probabilities at baseline creatinine levels. All p-values have been calculated by univariate Cox regression.

Figure 2.

Incidences for recorded endpoints in RTR with baseline serum creatinine <200 μmol/L (N = 946) and >200 μmol/L (N = 106).***p < 0.0001.

It is well established in other populations that renal dysfunction is a risk factor for CVD outcomes. We therefore performed a risk-factor analysis of baseline serum creatinine among placebo-treated patients, which revealed that each 100 μmol/L increase was associated with increased risks for cardiac death, non-cardiovascular death, all-cause mortality and MACE. Multivariate risk factor analyses with the predefined endpoints as dependent variables were also performed and revealed that increased serum creatinine at baseline was an independent risk factor for cardiac death, non-cardiovascular death, all-cause mortality and MACE (Table 2). We also made the same analysis of endpoints in relation to serum creatinine at baseline in the treament arm and found an association to the risk for cardiac death (p = 0.0025), all-cause mortality (p = 0.0095) and MACE (p < 0.0001) in univariate analyses, and cardiac death (p = 0.0342), all-cause mortality (p = 0.0160), MACE (p = 0.0003) and non-fatal MI (p = 0.0410) in multivariate analyses.

Table 2.  Univariate and multivariate RR (95% CI) per 100 μmol/L increase in baseline serum creatinine
 Creatinine increase by 100 μmol/L
Univariate RR (95% CI)Multivariate RR (95% CI)
MACE1.63 (1.23–2.17), p = 0.00071.89 (1.42–2.55), p < 0.0001
Non-fatal MI1.12 (0.69–1.82), p = 0.6465-
Stroke1.30 (0.75–2.25), p = 0.355-
Cardiac death2.29 (1.58–3.32), p < 0.00012.94 (2.01–4.31), p < 0.0001
Non-cardiovascular death1.95 (1.34–2.83), p = 0.00052.30 (1.54–3.43), p < 0.0001
All-cause mortality2.12 (1.66–2.70), p = 0.00012.50 (1.90–3.29), p < 0.0001

In the placebo group, 137 patients experienced renal graft loss during the follow-up. Graft loss increased the risk for non-cardiovascular death, all-cause mortality, non-fatal MI and MACE. The incidence of non-cardiovascular death was 5.5% in patients with a functioning graft and 10.9% in graft failure (p = 0.0128). Correspondingly, the all-cause mortality rate was 11.8% and 21.9% (p = 0.0010), the non-fatal MI rate 5.4% and 12.4% (p = 0.0014), and the MACE rate 11.5% and 20.4% (p = 0.0032) in patients with a functioning graft and patients with graft loss, respectively (Figure 3).

Figure 3.

Incidences for recorded endpoints in RTR with functioning graft (N = 915) and with graft loss (N = 137) during follow-up.*p < 0.02; **p < 0.005; ***p < 0.001.

Discussion

This study investigated renal function measured by serum creatinine at baseline, as a risk factor for mortality and cardiovascular complications in RTR in the placebo arm of the ALERT trial. It is clearly evident that elevated serum creatinine at baseline was a strong and independent risk factor for cardiac, non-cardiovascular and all-cause mortality, and MACE in RTR.

CVD accounts for more than 50% of all deaths in kidney transplant recipients. The CVD mortality rate is 3–5 times higher than in an age-matched general population, but significantly lower than in an age-matched dialysis population (3,8). Renal transplantation adds on to the unfavorable metabolic and hemodynamic conditions the patients have been exposed to prior the transplantation (9). The onset or aggravation of diabetes, dyslipidemia or hypertension often occurs and can be partly associated to the immunosuppressive regimen (10,11,23). Up to 10% of RTR develop post-transplant diabetes (24). In the present study, we demonstrated that elevated baseline serum creatinine, indicating renal dysfunction, was a strong and independent risk factor for all-cause mortality and CVD endpoints in RTR. Renal transplant dysfunction is clearly a strong indicator of future graft loss. In the ALERT study the hazard ratio for graft loss was 5.4 per 100 μmol/L increase in baseline serum creatinine leading, in turn, to an even larger risk of premature patient death (19,25). The present study does not allow us to draw any conclusions about the specific mechanisms underlying the associations between serum creatinine and mortality or cardiovascular complications in RTR, although it should be emphasized that the risk increment by renal dysfunction was independent of other established or traditional risk factors. Renal dysfunction may contribute to the increased risk by its associations with non-traditional CVD risk factors such as inflammation, oxidative stress, malnutrition, endothelial dysfunction and calcium-phosphate imbalance. The study does not reveal the reasons for lack of associations to stroke and non-fatal MI.

The threshold above which there was a clear influence of serum creatinine on outcomes in our study was around 200 μmol/L (2.3 mg/dL), similar to that reported in other studies (8). It has been shown that the prevalence of non-traditional CVD risk factors increased steeply when renal function was reduced below this level. Woo et al. have also reported the best survival in RTR with serum creatinine less than 200 μmol/L at three months after transplantation (20).

Even mild elevation of serum creatinine has been associated with increased mortality and cardiovascular risk in other populations (26,27). The CVD risk has been shown to increase progressively with deteriorating GFR and is increased significantly by the time serum creatinine is elevated (28–30). In low-risk populations or community studies, the relationship between the level of kidney dysfunction and CVD outcome is unclear (8). Our study firmly established the associations between serum creatinine and mortality and cardiac morbidity in RTR. Our results are in agreement with a report by Meier-Kriesche et al. analyzing United States Renal Data System (31). They show an independent association between renal dysfunction at 1-year post-transplantation and cardiovascular death. Since the present study was based on the placebo-group of the ALERT trial, the study population was homogenous with regard to immunosuppressive therapy and representative of the long-term RTR (22). It also has the advantage, compared with registry studies, that the follow-up period was longer, inclusion criteria strictly defined, reporting of events and data collection carefully monitored, endpoints independently validated and adjudicated by the CEC. Furthermore, it was also shown that not only cardiac death, but non-cardiovascular death, and thus all-cause mortality was strongly influenced by the degree of renal dysfunction, whereas stroke and non-fatal MI alone were not. A relative limitation of our study is the disparity in elapsed time (6–418 months) on renal replacement therapy, including transplantation, before baseline creatinine measurement. However, time on dialysis and time since transplantation were accounted for in the multivariate models.

Graft loss during follow-up was a risk factor for non-cardiovascular death, all-cause mortality, MACE and non-fatal MI, but not for stroke or cardiac death. After a graft loss the advantages of a functioning allograft over dialysis therapy disappear, which adds to the disadvantageous effects of immunosuppression, resulting in poor patient survival (32). The risk factors for graft loss in the placebo arm of the ALERT trial have been described previously, some of which are also CVD risk factors (25). Even though renal transplant dysfunction serves as a strong indicator for future graft loss, the effects of renal dysfunction and graft loss on the endpoints differed in our study. Unlike serum creatinine level, graft loss did not increase cardiac death risk but increased the risk for non-fatal MI. We believe this discrepancy is due to small number of patients with these complications.

In conclusion, elevated serum creatinine in RTR reflecting renal transplant dysfunction is a strong and independent risk factor for cardiac, non-cardiovascular and all-cause mortality, as well as MACE at 5-year follow-up. Creatinine is a risk factor independent of traditional CVD risk factors. Renal graft loss during follow-up increases the incidences of non-cardiovascular death, all-cause mortality, non-fatal MI and MACE. Thus, every effort should be made to preserve the function of a renal transplant.

Acknowledgments

We wish to thank all trial participants, physicians, and nurses in the participating centers for their important contribution to the study. Novartis provided the fluvastatin and matching placebo used in this study. The ALERT steering committee members received financial support from Novartis Pharma AG in the form of honoraria (excluding members who were investigators) and support for travel and accommodation expenses incurred by attending steering committee meetings. Claudio Gimpelewicz is a Novartis employee. Members of the steering committee have served as consultants for and received travel expenses, payment for lecturing, or funding for research from other pharmaceutical companies marketing lipid-lowering drugs, including Merck Sharpe and Dohme, Bristol-Myers Squibb, Astra-Zeneca, Schering, Bayer and Pfizer. Inga Soveri was supported by the Swedish Heart and Lung Foundation, Uppsala County Association Against Heart and Lung Diseases, and Swedish Society of Nephrology.