- Email: [email protected]
Cardiovascular disease in patients with chronic kidney disease: Getting to the heart of the matter
Cardiovascular Disease in Patients With Chronic Kidney Disease: Getting to the Heart of the Matter Adeera Levin, MD, Ognjenka Djurdjev, MSc, Brendan Barrett, MD, Ellen Burgess, MD, Euan Carlisle, MD, Jean Ethier, MD, Kailash Jindal, MD, David Mendelssohn, MD, Sheldon Tobe, MD, Joel Singer, MD, and Christopher Thompson, MD ● The high prevalence of cardiovascular disease (CVD) in patients with kidney disease is well described. This Canadian, multicenter, observational cohort study reports the prevalence and risk factors of CVD associated with kidney disease, in a cohort of patients with established chronic kidney disease (CKD), who are followed-up by nephrologists. This analysis sought to answer 2 questions: (1) in patients with established CKD, are the prevalence and progression of CVD accounted for by conventional or uremia-related risk factors, and (2) to what extent can progression to renal replacement therapy (RRT) be explained by CVD versus traditional risk factors for kidney disease? This study population consists of 313 patients (predominantly men) who had a mean age of 56 years and a mean creatinine clearance of 36 mL/min. Thirty percent were diabetic. The overall prevalence of CVD was 46%, and was independent of severity of kidney dysfunction (P ⴝ 0.700). The median follow-up time was 23 months, for a total of 462 patient years. We note the overall incidence of CVD events (new CVD or worsening of CVD) was 47/244 (20%). The best predictors of new CVD events among those without preexisting CVD were diabetes (odds ratio [OR] ⴝ 5.35, P ⴝ 0.018) and age (OR ⴝ 1.26, P ⴝ 0.08). In those with preexisting CVD, low diastolic pressure (DP) (OR ⴝ .72, P ⴝ 0.004) and high triglycerides (OR ⴝ 1.48, P ⴝ 0.019) at baseline were independent predictors of progression of CVD. We could not determine an independent impact of kidney function on CVD in the overall cohort. Furthermore, we determined that the presence of CVD itself confers an increased risk for progression to RRT (relative risk [RR] ⴝ 1.58, P ⴝ 0.047), adjusted for kidney function. This is the first in-depth analysis of CVD in a cohort of patients with established chronic kidney disease who are not on dialysis. The question regarding the impact of the altered biology of uremia in contributing to CVD progression remains unanswered, and clearly needs further study. However, the findings do raise the issue of whether aggressive treatment of CVD and risk factors might, in fact, reduce progression to RRT. Further large-scale, observational studies as well as interventional studies are needed to more clearly understand the complex biology of cardiovascular and kidney disease progression. © 2001 by the National Kidney Foundation, Inc. INDEX WORDS: Cardiovascular disease (CVD); chronic kidney disease (CKD); progression; cardiovascular events; risk factors.
From the Division of Nephrology, University of British Columbia, Vancouver, British Columbia, Canada; Centre for Health Evaluation and Outcomes Sciences, St. Paul’s Hospital and University of British Columbia, Vancouver, British Columbia, Canada; Division of Nephrology, Memorial University, St. John’s, Newfoundland; Division of Nephrology, University of Calgary, Calgary, Alberta, Canada; Division of Nephrology, McMaster University, Hamilton, Ontario, Canada; Division of Nephrology, Universite de Montreal, Hopital St. Luc, Montreal, Quebec, Canada; Division of Nephrology, Queen Elizabeth II Health Sciences Centre, Dalhousie University, Halifax, Nova Scotia, Canada; Division of Nephrology, St. Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada; Division of Nephrology, Sunnybrook Hospital, University of Toronto, Toronto, Ontario, Canada; Division of Cardiology, St. Paul’s Hospital, University of British Columbia, Vancouver, British Columbia, Canada. Received and accepted as submitted on July 17, 2001. Address reprint requests to Adeera Levin, MD, FRCPC, Associate Professor of Medicine, St. Paul’s Hospital, Providence Wing, 1081 Burrard Street, Room 6010A, Vancouver, BC, Canada V6Z1 Y8. E-mail: [email protected] © 2001 by the National Kidney Foundation, Inc. 0272-6386/01/3806-0037$35.00/0 doi:10.1053/ajkd.2001.29275 1398
T
HE HIGH PREVALENCE and incidence of cardiovascular disease (CVD) in both dialysis patients and transplant patients is well described,1-5 as is the adverse impact of CVD on long-term outcomes.6-8 Risk factors recognized to contribute to CVD in the general population, hypertension and diabetes, are present in patients with chronic kidney disease (CKD). Numerous studies to date have shown an independent effect of kidney disease itself on CVD outcomes.7-11 Contribution of kidney dysfunction itself, and of other factors associated with kidney disease, to the poor outcomes is unknown. In chronic kidney disease populations, both on and not on dialysis, there are specific uremia-related risk factors that have been associated with CVD outcomes: most notable is the consistent association of anemia and left ventricular hypertrophy (LVH) and congestive heart failure (CHF).6,12-14 Abnormal calcium phosphate metabolism and hyperparathyroidism have also been associated with myocardial fibrosis, hypertrophy, and vascular calcification.15-18 However, there is limited
American Journal of Kidney Diseases, Vol 38, No 6 (December), 2001: pp 1398-1407
SYMPOSIUM PROCEEDINGS
understanding of the differential impact of each risk factor on kidney function and cardiovascular function, and the interaction between established CVD and kidney disease. The issue of reverse causality and survival bias confounds most analyses performed to date.19 We report here the results of data collected for the Canadian Multi-centre Study in Patients with Renal Insufficiency, commenced in 1994. This study is an observational cohort study funded by the Kidney Foundation of Canada. We have previously reported on the risk factors associated with left ventricular (LV) growth and LVH.12 We identified the increased prevalence of LVH at each level of creatinine clearance. Further, we found that anemia and systolic blood pressure (BP) were independent risk factors for cardiac growth and hypertrophy in patients with varying degrees of kidney dysfunction. In this same group of patients with established chronic kidney disease, this study attempts to answer additional questions: (1) Is there an independent impact of uremia-associated risk factors, such as hemoglobin (Hgb), intact parathyroid syndrome (iPTH), and calcium phosphate product, after correcting for the level of kidney function, on the prevalence and incidence of CVD in patients with chronic kidney disease; and (2) is there an independent impact of the presence of CVD on the progression of kidney disease, in addition to established risk factors such as hypertension, proteinuria, age, and sex? MATERIALS AND METHODS
Study Population and Design This Canadian, multicenter, cohort study enrolled patients who had been identified by nephrologists as having CKD. The details of the study have been reported elsewhere.12 Consecutive patients in 7 different geographic locations across Canada, who had calculated creatinine clearances between 75 and 10 mL/min (calculated by using the Cockcroft-Gault formula), who were expected by their physicians to live and to be off dialysis for at least 12 months of follow-up, were enrolled in the study. Informed consent was obtained from all patients, and the ethics board of each institution approved the study. There were no attempts to modify therapy during the study. All patients had clinical, laboratory, and cardiovascular (CV) symptoms reviewed at 1-year intervals. Hospitalization events and medication use were also evaluated. The data was captured by using standardized data collection sheets, and was then entered into a central database for analysis.
1399
Definitions of Cardiovascular Disease CV disease was defined by both history of CV event or condition and/or current symptoms according to New York Heart Association (NYHA) and Canadian Cardiovascular Society (CCS) classifications for heart failure and angina, respectively. CV event history included history of myocardial infarction (MI), angina, coronary artery bypass graft (CABG) or angioplasty, transient ischemic attack (TIA) or cardiovascular accident (CVA), peripheral vascular disease (PVD), and congestive heart failure (CHF). Research assistants administered standardized questionnaires that describe symptoms associated with each class of heart failure or angina to each patient, and symptoms were graded by using conventional NYHA and CCS classifications. Change in cardiovascular status was defined as either change in NYHA or CCS classification, or a cardiac-related hospitalization.
Chronic Kidney Disease: Calculations and Definitions Serum creatinine level was measured in each patient’s own laboratory, which was maintained through the study period. Calculations using the formula of Cockcroft Gault12 were used to determine the level of creatinine clearance. Four categories of kidney function were identified: mild (⬎50⬍75 mL/min), moderate (35–49 mL/min), moderate severe (25–34 mL/min), or severe (10–24 mL/min). Renal replacement therapy (RRT) was defined as the need for either dialysis or transplantation.
Outcomes of Interest The major outcomes of interest for this analysis included kidney survival (time to RRT) and change in CVD status.
Statistical Analysis The baseline clinical characteristics of patients were compared by using the analysis of variance (ANOVA) or KruskalWallis test for continuous variables, depending on distribution, and the 2 test for categoric variables. We used the 2 test for discrete variables and the t test or Wilcoxon rank-sum test, where appropriate, for continuous variables, to compare patients with and without CVD at baseline, as well as patients with change and without change of CVD status. Univariate odds ratios (ORs) were adjusted for degree of kidney function by using logistic regression. The logistic regression was used to determine the association between patient baseline characteristics and presence of CVD at baseline, and then to predict change in CVD status. Time to RRT therapy was calculated based on the time from study enrollment to the dialysis start or transplantation date. Renal survival was estimated by using the Kaplan-Meier method. The survival curves by presence of CVD were compared by using the log-rank test. We used proportional hazards regression models to assess the effect of baseline patient characteristics on time to RRT after controlling for baseline kidney function. The Cox regression model was used to identify independent and significant predictors of time to RRT. All multivariate models were developed by using a backward elimination
1400
LEVIN ET AL
technique, including the following variables: patient age, sex, diabetes status, creatinine clearance, systolic and diastolic BP, total cholesterol, triglycerides, hemoglobin, phosphate-calcium product, iPTH, and presence of CVD at baseline. A P value of less than 0.05 for 2-sided univariate tests was considered statistically significant. We set the P value for multivariate analyses to 0.10 to account for the hypothesisgenerating purpose of this analysis. Statistical analyses were performed by using SAS software, version 6.12 (SAS Institute, Cary, NC).
RESULTS
Baseline Characteristics The original study enrolled 447 patients, of whom 417 had an initial evaluation of cardiac status. We excluded 104 patients from this analysis because of missing baseline values for key variables of interest. Thus, this analysis Table 1.
Characteristic
Age yr Gender, % men Race, % Caucasian Diabetes CrCl mL/min 24-hr urine protein g/L Systolic BP mm Hg Diastolic BP mm Hg Pulse pressure mm Hg Total cholesterol g/L HDL mmol/L Triglycerides mmol/L Hemoglobin g/L Ca*PO4 mmol/L *mmol/L Log PTH pmol/L LVMI g/m2 ACEI alone or in combination -blockers alone or in combination Any lipid-lowering agent Any Hx or Sx of CVD Hx of any CVD Hx of angina Hx of CHF Hx of MI Hx of PVD NYHA I–II III–IV CCS I–II III–IV
includes 313 predominantly male (67%) and Caucasian (87%) patients, who had a mean age of 56 years (Table 1). The mean calculated creatinine clearance (Ccr) was 36.2 ⫾16.8 mL/min. Twenty nine percent of the group were diabetic, and the prevalence of the causes of kidney disease paralleled those in national registries: hypertension (27%), glomerulonephritis (18%), diabetes (25%), and polycystic kidney disease (PCKD) (8%). Table 1 shows the baseline demographics of the whole cohort, and compares a number of pertinent variables relative with baseline kidney function. Sex, hemoglobin, pulse pressure, calcium-phosphate product, iPTH, high-density lipoprotein (HDL), LVMI were different between groups with different kidney function.
Baseline Characteristics by Level of Renal Function
All n ⫽ 313
⬎50 mL/min n ⫽ 62
35–49 mL/min n ⫽ 83
56.3 ⫾ 14.5 211 (67.4%) 267 (87.0%) 91 (29.1%) 36.2 ⫾ 16.8 1.04 ⫾ 1.36 144.4 ⫾ 23.4 85.4 ⫾ 11.6 58.9 ⫾ 20.8 5.70 ⫾ 1.34 1.15 ⫾ 0.51 2.36 ⫾ 1.69 128.1 ⫾ 19.5 2.78 ⫾ 0.62 2.22 ⫾ 0.85 111.3 ⫾ 35.3 155 (49.7%)
53.5 ⫾ 13.4 54 (87%) 54 (88%) 21 (34%) 62.6 ⫾ 10.5 0.96 ⫾ 1.24 143.9 ⫾ 20.9 87.3 ⫾ 11.8 56.5 ⫾ 19.1 5.76 ⫾ 1.19 1.04 ⫾ 0.54 2.57 ⫾ 1.44 141.5 ⫾ 16.6 2.50 ⫾ 0.49 1.64 ⫾ 0.58 110.0 ⫾ 35.2 37 (59.7%)
54.9 ⫾ 15.3 64 (77%) 74 (91%) 24 (29%) 41.9 ⫾ 4.3 0.97 ⫾ 1.22 142.7 ⫾ 21.6 85.1 ⫾ 11.3 57.0 ⫾ 21.8 5.58 ⫾ 1.16 1.06 ⫾ 0.31 2.46 ⫾ 2.27 134.3 ⫾ 19.3 2.66 ⫾ 0.56 1.88 ⫾ 0.67 106.8 ⫾ 28.8 47 (56.6%)
74 (23.7%) 42 (13.5%) 143 (45.7%) 120 (38.3%) 52 (16.7%) 23 (7.6%) 39 (12.7%) 38 (12.4%)
15 (24.2%) 10 (16.1%) 27 (43.5%) 22 (35.5%) 11 (18.0%) 5 (8.3%) 10 (16.4%) 8 (13.1%)
20 (24.1%) 10 (12.1%) 41 (49.4%) 32 (38.6%) 16 (19.5%) 6 (7.5%) 10 (12.2%) 9 (11.0%)
16 (21.6%) 12 (16.2%) 29 (39.2%) 26 (35.14%) 8 (10.8%) 3 (4.2%) 10 (13.7%) 10 (13.9%)
23 (24.7%) 10 (10.9%) 46 (48.9%) 40 (42.6%) 17 (18.1%) 9 (9.8%) 9 (9.9%) 11 (12.1%)
0.970 0.676 0.788 0.743 0.472 0.596 0.686 0.953
76 (24.3%) 7 (2.2%)
15 (24.2%) 1 (1.6%)
25 (30.1%) 0 (0%)
13 (17.6%) 2 (2.7%)
23 (24.5%) 4 (4.3%)
0.331
70 (22.4%) 6 (1.9%)
16 (25.8%) 1 (1.6%)
20 (24.1%) 1 (1.2%)
13 (17.6%) 3 (4.1%)
21 (22.3%) 1 (1.1%)
0.712
Abbreviations: Hx, history; Sx, symptoms.
25–34 mL/min n ⫽ 74
56.5 ⫾ 14.7 47 (64%) 59 (81%) 16 (22%) 29.9 ⫾ 3.0 1.03 ⫾ 1.75 141.0 ⫾ 24.7 84.8 ⫾ 12.7 56.1 ⫾ 21.8 5.98 ⫾ 1.64 1.26 ⫾ 0.65 2.45 ⫾ 1.53 125.7 ⫾ 16.6 2.73 ⫾ 0.54 2.34 ⫾ 0.79 106.1 ⫾ 34.7 37 (50.0%)
10–24 mL/min n ⫽ 94
P Value
59.3 ⫾ 14.1 46 (49%) 80 (87%) 30 (32%) 18.8 ⫾ 3.8 1.22 ⫾ 1.25 149.3 ⫾ 24.4 84.9 ⫾ 10.9 64.4 ⫾ 21.2 5.56 ⫾ 1.30 1.22 ⫾ 0.48 2.07 ⫾ 1.29 115.6 ⫾ 15.3 3.12 ⫾ 0.65 2.80 ⫾ 0.81 120.2 ⫾ 39.8 34 (36.6%)
0.0604 0.001 0.268 0.383 0.0001 0.7165 0.0847 0.5519 0.0242 0.1592 0.0166 0.2424 0.0001 0.0001 0.0001 0.0474 0.002
SYMPOSIUM PROCEEDINGS Table 2.
1401
Conventional and Uremia-Associated Risk Factors in Patients With and Without CVD at Baseline
Variables
Age (yr) Gender, % men Diabetes CrCl (calculated) mL/min 24-hr urine protein g/L Systolic BP mm Hg Diastolic BP mm Hg Pulse pressure mm Hg Total cholesterol mmol/L HDL mmol/L Triglycerides mmol/L Hemoglobin g/L PO4*Ca mmol/L *mmol/L log PTH pmol/L LVMI g/m2 ACEI alone or in combination Beta blockers alone or in combination Any lipid-lowering agent
CVD (⫺) N ⫽ 170
CVD (⫹) N ⫽ 143
P Value
53.5 ⫾ 14.9 108 (63.5%) 37 (21.8%) 35.9 ⫾ 16.2 1.08 ⫾ 1.48 143.5 ⫾ 23.2 87.2 ⫾ 11.7 56.3 ⫾ 20.4 5.77 ⫾ 1.54 1.19 ⫾ 0.55 2.36 ⫾ 1.39 128.0 ⫾ 19.3 2.78 ⫾ 0.61 2.27 ⫾ 0.88 105.59 ⫾ 31.83 88 (51.8%) 34 (20.0%) 17 (10.0)
59.6 ⫾ 13.4 103 (72.0%) 54 (37.8%) 36.6 ⫾ 17.5 1.01 ⫾ 1.24 145.3 ⫾ 23.7 83.3 ⫾ 11.2 62.0 ⫾ 21.0 5.62 ⫾ 1.06 1.09 ⫾ 0.46 2.36 ⫾ 1.99 128.2 ⫾ 19.9 2.80 ⫾ 0.63 2.16 ⫾ 0.82 117.94 ⫾ 37.95 67 (47.2%) 40 (28.2%) 25 (17.7%)
0.0002 0.110 0.002 0.7000 0.6945 0.4950 0.0029 0.0154 0.3209 0.1168 0.9808 0.9196 0.7663 0.2246 0.0034 0.420 0.091 0.047
The use of angiotensin-converting enzyme inhibitor (ACEI) therapy was the lowest in those with the worst kidney function. The prevalence of CVD was similar at each level of kidney function.
the multivariate analysis and adjusted for severity of kidney disease to identify older age, diabetes, and low diastolic BP as factors associated with increasing the probability of the CVD presence at baseline (Table 3).
Prevalence of Cardiovascular Disease at Baseline: A Comparison of Conventional and Uremia-Related Risk Factors Comparisons between patients with and without CVD at baseline are presented in Table 2. Those with CVD at baseline were likely to be older, diabetic, have lower diastolic BPs, higher pulse pressures, and larger values for LVMI. However, in this cohort, the presence or absence of CVD at baseline is not different with respect to creatinine clearance (35.9 mL/min versus 36.6 mL/min, P ⫽ 0.700). Again, the use of ACEI was not different as a function of presence or absence of CVD, but lipid-lowering agents were used more often in those with CVD than those without (18% versus 10%, P ⫽ 0.047). After adjustment for the level of renal function, the same risk factors (ie, age [OR ⫽ 1.17, P ⫽ 0.0002], diabetes [OR ⫽ 2.18, P ⫽ 0.0022], diastolic BP [OR ⫽ .81, P ⫽ 0.0002], pulse pressure [OR ⫽ 1.07, P ⫽ 0.013], and LVMI [OR ⫽ 1.11, P ⫽ 0.0033]), were associated with presence or absence of CVD at baseline. We used
Cardiovascular Disease—Change in CVD Status During Follow-Up: Impact of Preexisting CVD Of the original 313 patients included in this analysis, 36 patients were lost or refused further follow-up, 8 patients died, 24 patients started dialysis, and 1 patient was transplanted before at least 1 cardiovascular follow-up evaluation. Hence, 244 patients had evaluable cardiovascular follow-up: median duration of follow-up was 23 months, with a total of 5,547 months or 462 patient years of follow-up. During the follow-up Table 3. Multivariate Analysis of Risk Factors Associated With the Presence of CVD at Baseline
Variables
Odds Ratio
95% Confidence Interval
P Value
Age 5 years Male gender Diabetes CrCl 5 mL/min Diastolic BP 5 mm Hg
1.143 1.355 2.024 1.025 0.898
1.047–1.248 0.793–2.313 1.206–3.397 0.952–1.103 0.808–0.998
0.0028 0.2663 0.0076 0.5169 0.0464
1402
LEVIN ET AL
period, 48 of 244 (20%) patients developed new or worsening cardiac symptoms or were hospitalized for cardiac disease. Of the 135 patients who had no preexisting CVD, 10 (7.4%) developed new CVD; this is in contrast to the 109 patients with preexisting cardiac disease, of which 38 (35%) developed a new event or worsening of CVD (P ⫽ 0.001). Thus, 21% of the 48 patients with new or worsening cardiac disease had no previous CVD. Table 4 describes the results of a univariate analysis comparing the association of conventional and uremia-related variables with change in CVD status, in patients with and without preexisting CVD. Note that age and diabetes at baseline predict change in CVD status in both groups of patients, whereas higher serum triglyceride levels and lower diastolic BP predict worsening of CVD among patients with preexisting CVD over the follow-up period. There were no apparent associations of uremia-specific factors such as hemoglobin, calcium phosphate product, iPTH level, or of level of kidney dysfunction itself, with change of CVD status. The same baseline characteristics were associated with change in CVD status, depending on the presence or absence of CVD at baseline, after adjustment for level of kidney function. A multivariate analysis revealed that diabetes increases the probability of change in CVD status in those without preexisting CVD (Fig 1). Low diastolic BP and Table 4.
high triglyceride levels were independent and significant predictors of change in CVD status in those with preexisting CVD (Fig 2). We performed a separate multivariate analysis for all patients who had a change in NYHA class heart failure versus patients who had change in CCS class angina. The important predictors of NYHA class change were older age (OR ⫽ 1.21 per 5 years, P ⫽ 0.0588), lower diastolic BP (OR ⫽ .79 per 5 mm Hg, P ⫽ 0.0175), higher triglyceride levels (1.44 per mmol/L, P ⫽ 0.0118), and presence of CVD at baseline (OR ⫽ 4.91, P ⫽ 0.0011). The change in CCS class was associated with lower baseline kidney function (OR ⫽ .84 per 5 mL/min, P ⫽ 0.0624), diabetes (OR ⫽ 3.96, P ⫽ 0.0035), lower diastolic BP (OR ⫽ .79 per 5 mm Hg, P ⫽ 0.0284), and baseline CVD (OR ⫽ 3.88, P ⫽ 0.0111). Loss of Kidney Function: Conventional and Other Predictors of Time to Renal Replacement Therapy Table 5 describes risk factors associated with time to RRT, adjusted for baseline kidney function. In addition to known risk factors for progressive kidney function decline (age, sex, diabetes, proteinuria, BP, and kidney function), there are uremia-specific factors—hemoglobin, iPTH, calcium phosphate product— that may also be associated with decline. Furthermore, the presence of CVD, as well as low HDL and cholesterol
Patient Characteristics Associated With CVD Status in Patients With and Those Without CVD at Baseline No Preexisting CVD N ⫽ 135
Variables
Age (yr) Gender, % men Diabetes CrCl mL/min 24-hr urine protein g/L Systolic BP mm Hg Diastolic BP mm Hg Pulse pressure mm Hg Total cholesterol mmol/L HDL mmol/L Triglycerides mmol/L Hemoglobin g/L Ca*PO4 mmol/L *mmol/L Log PTH pmol/L LVMI g/m2
Preexisting CVD N ⫽ 109
No CVD n ⫽ 125
CVD n ⫽ 10
P Value
Stable n ⫽ 71
Worsening n ⫽ 38
P Value
52.3 ⫾ 15.4 80 (64.0%) 21 (16.8%) 37.8 ⫾ 16.3 0.84 ⫾ 1.01 141.2 ⫾ 23.8 86.2 ⫾ 11.8 55.0 ⫾ 20.7 5.66 ⫾ 1.27 1.21 ⫾ 0.60 2.36 ⫾ 1.32 130.6 ⫾ 19.3 2.70 ⫾ 0.61 2.24 ⫾ 0.84 103.28 ⫾ 32.17
61.8 ⫾ 13.1 5 (50.0%) 5 (50.0%) 33.7 ⫾ 16.1 1.08 ⫾ 1.29 149.8 ⫾ 27.9 87.0 ⫾ 14.1 62.8 ⫾ 22.8 6.34 ⫾ 1.44 1.11 ⫾ 0.30 2.54 ⫾ 1.22 121.8 ⫾ 17.1 2.87 ⫾ 0.37 1.97 ⫾ 0.76 115.61 ⫾ 34.95
0.0592 0.499 0.023 0.4440 0.5246 0.2808 0.8492 0.2562 0.1102 0.3604 0.6816 0.1659 0.3926 0.3339 0.2738
58.4 ⫾ 13.3 51 (71.8%) 19 (26.8%) 37.8 ⫾ 17.3 0.75 ⫾ 1.05 147.0 ⫾ 24.8 85.8 ⫾ 10.7 61.2 ⫾ 22.9 5.57 ⫾ 0.90 1.16 ⫾ 0.43 2.03 ⫾ 1.30 127.5 ⫾ 19.0 2.77 ⫾ 0.59 2.06 ⫾ 0.79 115.7 ⫾ 37.35
63.2 ⫾ 10.7 26 (68.42%) 17 (44.7%) 32.8 ⫾ 13.1 1.48 ⫾ 1.53 141.8 ⫾ 24.3 78.3 ⫾ 11.1 63.5 ⫾ 19.79 5.89 ⫾ 1.19 1.04 ⫾ 0.60 3.04 ⫾ 3.09 129.0 ⫾ 18.2 2.83 ⫾ 0.48 2.26 ⫾ 0.78 125.38 ⫾ 42.64
0.0564 0.709 0.057 0.1224 0.0278 0.2980 0.0008 0.5950 0.1228 0.5958 0.0183 0.6867 0.5827 0.1957 0.2414
SYMPOSIUM PROCEEDINGS
1403
Fig 1. Factors associated with change of CVD status in patients without history or symptoms of CVD at baseline.
levels at baseline, are associated with increased risk for kidney failure leading to RRT. There were several factors that independently and significantly predicted decline of kidney function in the multivariate analysis, shown in Fig 3: older age, male gender, diabetes, low baseline kidney function and hemoglobin, and high diastolic blood pressure and iPTH level. In terms of the impact of CVD on time to RRT, there was only a trivial difference after adjustment for all of the earliermentioned factors (univariate relative risk [RR] ⫽
1.58, multivariate RR ⫽ 1.52), indicating the importance of this factor. Figure 4 shows the impact of CVD presence or absence at baseline on time to RRT. DISCUSSION
This analysis of a cohort of patients with known chronic kidney disease describes the high prevalence of cardiac disease (46%) and of conventional cardiovascular risk factors. The analysis does not show that uremia-specific risk fac-
Fig 2. Factors associated with change of CVD status in patients with history or symptoms of CVD at baseline.
1404
LEVIN ET AL
Table 5. Associations of Baseline Characteristics, Adjusted for Kidney Function, With Time to RRT
Variables
Risk Ratio
95% Confidence Interval
P Value
Age 5 yr Gender, male Diabetes CrCl 5 mL/min 24-hr urine protein g/L Systolic BP 5 mm Hg Diastolic BP 5 mm Hg Pulse pressure 5 mm Hg Total cholesterol mmol/L HDL mmol/L Triglycerides mmol/L Hemoglobin 5 g/L Ca*PO4 mmol/L *mmol/L log PTH pmol/L Hx. or Sx. of CVD
0.879 1.957 2.181 0.708 1.550 0.990 1.120 0.997 0.819 0.614 0.935 0.863 1.651 1.613 1.580
0.816-0.947 1.204-3.181 1.376-3.457 0.639-0.785 1.377-1.746 0.942-1.041 1.004-1.250 0.950-1.046 0.680-0.988 0.351-1.074 0.790-1.106 0.804-0.926 1.229-2.217 1.199-2.168 1.006-2.482
0.0007 0.0068 0.0009 0.0001 0.0001 0.6991 0.0417 0.9042 0.0371 0.0876 0.4315 0.0001 0.0009 0.0016 0.0470
tors predict either the prevalence of CVD or the change in CVD status. Previous publications have described a similarly high prevalence of CVD risk factors and have determined that chronic kidney disease itself was a risk factor for CV outcome and mortality.10,20,21 All of these studies compared patients with kidney disease with those without kidney disease, did not adjust for specific levels of kidney function, and did not measure specific variables associated with kidney disease—hemoglobin, iPTH, calcium, phos-
phate—that may mediate the impact of CKD on outcomes. Notably, Culleton et al22 reported findings similar to ours in the Framingham cohort, despite a higher prevalence of CVD in patients with CKD at baseline, there was no impact of CKD on long-term outcomes after controlling for conventional risk factors. That uremia-specific factors did not appear to exert an effect on CVD prevalence/incidence independent of conventional cardiac risk factors could be owing to study limitations, rather than to the absence of such an effect. Furthermore, the prevalence of ischemia-related CVD (angina, CVA, MI, PVD) at baseline is similar in this population to that in other renal populations at later stages of disease (ie, on dialysis and posttransplant).1,3,23,24 This finding supports the notion recently discussed in depth by Parfrey25 in an editorial: cardiac function can be affected by processes different from simple atheroscleroticinduced ischemia, and CVD with progressive kidney disease may be secondary to the influence of both atherogenic processes and factors that modulate cardiac muscle structure function. We did not measure LV ejection fractions in this study, rather, we only have a measure of cardiac hypertrophy—LVMI and pulse pressure. Pulse pressure, as an indicator of vascular stiffness, has recently been described as predicting morbidity and mortality, and contributing to cardiac dysfunction26,27 in cardiac populations. The current
Fig 3. Independent predictors of time to RRT (results of Cox proportional hazards modeling).
SYMPOSIUM PROCEEDINGS
1405
Fig 4. Kaplan-Meier curves of time to RRT by CVD status at baseline.
analysis shows that both LVMI and pulse pressure are significantly different between those patients with and without CVD at baseline. Progressive cardiovascular disease and progressive decline of kidney function do have a number of risk factors in common, including hypertension and diabetes. There are some risk factors specific to each condition and some that occur as a consequence of 1 of these impacting another. It may be that the purported uremic factors exert their influence through affecting cardiac function, not the atherosclerotic process as previously thought. Interestingly, it is possible that these same factors that affect cardiac function may also accelerate decline of kidney function, both through activation of cytokines and growth factors, and that ongoing endothelial cell dysfunction may contribute simultaneously to both processes.16,30 There is substantial animal and cellular data to support the impact of these factors on progressive kidney disease.28,31 However, the current analysis is unable to determine the relative impact of each of these. This study further shows the association of conventional risk factors with kidney function decline and also shows an additional association of CVD presence with kidney survival. The data also suggest that anemia and hyperparathyroidism may contribute to progression of kidney disease, independent of kidney function level, sex, diabetes, or age.32 The presence of CVD at baseline confers a 50% increase in probability of
RRT. The explanation for this may be complex and multifactorial. Worsening cardiac function may lead to reduced blood flow to the kidney, thus, adding a prerenal insult to kidneys with little reserve, and to activation of cellular mechanisms that lead to sclerosis and fibrosis. Alternatively, cardiac dysfunction may lead to the addition of medications/interventions that may further cause kidney injury or reduce function. These hypotheses are not mutually exclusive. Interestingly, a recent report by Silverberg et al33 describes the treatment of heart failure and associated anemia with erythropoietin. Importantly, the treatment group had significantly impaired kidney function (serum creatinine 2.3 mg/dL) before therapy. After correction of anemia and concomitant with improvement in cardiac status there was improvement of kidney function.33 The current analysis, showing that CV disease and conventional risk factors for cardiac disease adversely affect the progression of kidney disease, provides hope that attention to CV health and risk factor modification might improve longterm outcomes for the progression of kidney disease. In this cohort, as in others,8,34-36 there is suboptimal use of cardiovascular medications such as ACEIs, -blockers, and lipid-lowering medications. It may be that both cardiologists and nephrologists caring for patients with established kidney disease actually practice therapeutic nihilism.35 This inadequate therapy of cardiac disease may accelerate the rate of kidney func-
1406
tion decline in susceptible individuals. Thus, the high prevalence of those with established cardiac disease commencing dialysis may be the consequence of failure to institute treatment before dialysis. This study is unique in that it describes the prevalence of CVD and its attendant risk factors in patients with kidney disease before dialysis known to nephrologists in Canada. Others have described unselected groups of general populations, or used population databases or post hoc analyses of large studies to review the impact of kidney disease on CV outcomes.7,9,10,22 These approaches suffer from a number of problems. There is a lack of uniformity in defining the presence and severity of kidney disease. There is little information about the stability of kidney dysfunction before the study starts, and there is no documentation of the etiology of kidney dysfunction or of uremia-associated biochemical and hematologic abnormalities. There is ample evidence to support the problems of use of serum creatinine alone as a marker for kidney disease.37 The current analysis, in contrast, focuses on individuals with established, nonreversible kidney disease, by using calculated estimates of kidney function. By examining those variables specifically associated with kidney disease, we are able to explore the hypothesis that there is an impact of these variables versus conventional risk factors versus kidney disease per se on CVD. This study also describes the complex interplay between CVD and progressive CKD. This is important if therapeutic strategies to treat one disease may actually be justified in terms of their impact on another aspect of this disease. Our study, like all prospective observational studies, precludes the establishment of causality. Furthermore, owing to missing values and loss of patients to follow-up, the true burden of illness in this population may be underestimated. This study did not capture events after RRT, or specific cardiac events. We believe, however, that the similarity of the prevalence of CVD in this cohort to that reported in other similar cohorts34,38 makes it unlikely that there is significant underestimation of CVD. The issue of reverse causality, surviving off RRT to actually have a CVD event, and survival bias (ie, those with worse CVD were put onto RRT as a means of controlling symptoms) cannot be addressed
LEVIN ET AL
with this analysis, or in this cohort. Importantly, this data can be used to generate hypotheses and develop interventional studies based on biologically plausible associations of specific risk factors. Ideally, we should establish follow-up in 2 distinct cohorts of patients: those with established kidney disease without CVD, and those with established cardiac disease without kidney dysfunction, and determine through careful and systematic measurement of known risk factors which of these are predisposed to developing into cardiac disease and to kidney disease, respectively. In summary, the prevalence of CVD in the population with CKD is high, and is attributable to both ischemic disease and myocardial dysfunction. This study shows an underuse of medications known to improve cardiac outcome in the general population. This may contribute to the poor cardiac outcomes seen in this group. Furthermore, the presence of CVD appears to predict an increased probability of RRT; thus, the appropriate treatment of cardiac disease may in fact lead to delay of RRT, and thus serve to lessen the burden of illness not only for the individual patients, but also for health care systems. ACKNOWLEDGMENT The authors would like to acknowledge the support of the Kidney Foundation of Canada, the extensive assistance of Dr. Vadim Minster and Ms. Cheyenne Reese for data organization and manuscript preparation, and all of the research assistants in each of the participating centers for their extensive data collection.
REFERENCES 1. Foley RN, Parfrey PS, Harnett JD, Kent GM, Martin CJ, Murray DC, Barre PE: Clinical and echocardiographic disease in patients starting end-stage renal disease therapy. Kidney Int 47:186-192, 1995 2. Canadian Organ Replacement Registry, Canadian Institute for Health Information (CIHI), Don Mills, Ontario, Canada, 1999 3. Parfrey PS: Cardiac disease in dialysis patients: Diagnosis, burden of disease, prognosis, risk factors and management. Nephrol Dial Transplant 15:58-68, 2000 4. Sarnak MJ, Levey AS: Epidemiology, diagnosis, and management of cardiac disease in chronic renal disease. J Thromb Thrombolysis 10:169-180, 2000 5. Kasiske B, Guijarro C, Massy Z, Wiederkehr M, Ma J: Cardiovascular disease after renal transplantation. J Am Soc Nephrol 7:158-165, 1996 6. Foley RN, Parfrey PS, Harnett JD, Kent GM, Murray DC, Barre PE: The impact of anemia on cardiomyopathy, morbidity and mortality in end-stage renal disease. Am J Kidney Dis 28:53-61, 1996
SYMPOSIUM PROCEEDINGS
7. Chertow GM, Normand S-LT, Silva LR, McNeil BJ: Survival after acute myocardial infarction in patients with end-stage renal disease: Results from the Cooperative Cardiovascular Project. Am J Kidney Dis 35:1044-1051, 2000 8. Herzog CA: Poor long-term survival of dialysis patients after acute myocardial infarction: Bad treatment or bad disease? Am J Kidney Dis 35:1217-1220, 2000 9. Mann J, Gerstein HC, Pogue J, Baosch, Yusef S: Renal insufficiency as a predictor of cardiovascular outcomes and the impact of ramipril: The HOPE randomized trial. Ann Intern Med 134:629-636, 2001 10. Hemmelgarn BR, Ghali WA, Quan H, Brant R, Norris CM, Taub KJ, Knudtson ML: Poor long-term survival after coronary angiography in patients with renal insufficiency. Am J Kidney Dis 37:64-72, 2001 11. Beattie JN, Soman SS, Sandberg KR, Yee J, Borzak S, McCullough PA: Determinants of mortality after myocardial infarction in patients with advanced renal dysfunction. Am J Kidney Dis 37:1191-1200, 2001 12. Levin A, Thompson CR, Ethier J, Carlisle EJ, Tobe S, Mendelssohn D, Burgess E, Jindal K, Barrett B, Singer J, Djurdjev O: Left ventricular mass index increase in early renal disease: Impact of decline in hemoglobin. Am J Kidney Dis 34:125-134, 1999 13. O’Riordan E, Foley RN: Effects of anaemia on cardiovascular status. Nephrol Dial Transplant 15:19-22, 2000 14. Levin A, Foley RN: Cardiovascular disease in chronic renal insufficiency. Am J Kidney Dis 36:24-30, 2000 15. Block GA, Port FK: Re-evaluation of risks associated with hyperphosphatemia and hyperparathyroidism in dialysis patients: Recommendations for a change in management. Am J Kidney Dis 35:1226-1237, 2000 16. Pannier B, Guerin AP, Marchais SJ, Metivier F, Safar ME, London GM: Postischemic vasodilation, endothelial activation, and cardiovascular remodeling in end-stage renal disease. Kidney Int 57:1091-1099, 2000 17. Goodman WG, Goldin J, Kuizon BD, Yoon C, Gales B, Sider D, Wang Y, Chung J, Emerick A, Greaser L, Elashoff RM, Salusky IB: Coronary-artery calcification in young adults with end-stage renal disease who are undergoing dialysis. N Engl J Med 342:1478-1483, 2000 18. Amann K, Ritz E, Wiest G, Klaus G, Mall G: A role of parathyroid hormone for the activation of cardiac fibroblasts in uremia. J Am Soc Nephrol 4:1814-1819, 1994 19. Baigent C, Burbury K, Wheeler D: Premature cardiovascular disease in chronic renal failure. Lancet 356:147152, 2000 20. Anderson RJ, O’Brien M, MaWhinney S, VillaNueva CB, Moritz TE, Sethi GK, Henderson WG, Hammermeister KE, Grover FL, Shroyer AL: Mild renal failure is associated with adverse outcome after cardiac valve surgery. Am J Kidney Dis 35:1127-1134, 2000 21. McCullough PA, Soman SS, Shah SS, Smith ST, Marks KR, Yee J, Steven Borzak S: Risks associated with renal dysfunction in coronary care unit patients. J Am Coll Cardiol 36:679-684, 2000 22. Culleton BF, Larson MG, Wilson PWF, Evans JC, Parfrey PS, Levy D: Cardiovascular disease and mortality in
1407
a community-based cohort with mild renal insufficiency. Kidney Int 56:2214-2219, 1999 23. Parfrey PS, Harnett JD, Foley RN: Heart failure and ischemic heart disease in chronic uremia. Curr Opin Nephrol Hypertens 4:105-110, 1995 24. Rigato C, Jeffery J, Foley R, Brown S, Tribula C, Negrijin C, Parfrey P: Risk factors for de novo ischaemic heart disease in renal transplant recipients. J Am Soc Nephrol 11:705A, 2000 (abstr) 25. Parfrey PS: Is renal insufficiency an atherogenic state? Reflection on prevalence, incidence and risk. Am J Kidney Dis 37:154-156, 2001 26. Mitchell GF: Pulse pressure, arterial compliance and cardiovascular morbidity and mortality. Curr Opin Nephrol Hypertens 8:335-342, 1999 27. Safar ME: Systolic blood pressure, pulse pressure and arterial stiffness as cardiovascular risk factors. Curr Opin Nephrol Hypertens 10:257-261, 2001 28. Hunter JJ, Chien KR: Signaling pathways for cardiac hypertrophy and failure. N Engl J Med 341:1276-1283, 1999 29. Amann K, Kronenberg G, Gehlen F, Wessels S, Orth S, Munter K, Ehmke H, Mall G, Ritz E: Cardiac remodeling in experimental renal failure—an immunohistochemical study. Nephrol Dial Transplant 13:1958-1966, 1998 30. London GM, Fabiani F, Marchais SJ, et al: Uremic cardiomyopathy: An inadequate left ventricular hypertrophy. Kidney Int 31:973-980, 1987 31. Tonelli M, Bohm C, Pandeya S, Gill J, Levin A, Kiberd BA: Cardiac risk factors and the use of cardioprotective medications in patients with chronic renal insufficiency. Am J Kidney Dis 37:484-489, 2001 32. Hebert LA, Wilmer WA, Falkenhain ME, LadsonWofford SE, Nahman NS, Rovin BH: Renoprotection: One or many therapies? Kidney Int 59:1211-1226, 2001 33. Silverberg D, Wexler D, Blum M, Keren G, Sheps D, Laibovitch E, Brosh D, Laniado S, Schwartz D, Yachnin T, Shapira I, Gavish D, Baruch R, Koifman B, Kaplan C, Steinbruch S, Iaina A: The use of subcutaneous erythropoietin and intravenous iron for the treatment of the anemia of severe, resistant congestive heart failure improves cardiac and renal function and functional cardiac class, and markedly reduces hospitalizations. J Am Coll Cardiol 35:17371744, 2000 34. Taal M, Omer SA, Nadim MK, McKenzie HS: Cellular and molecular mediators in common pathway mechanisms of chronic disease progression. Curr Opin Nephrol Hypertens 9:323-331, 2000 35. Levin A, Stevens L, McCullough P: Cardiovascular disease: The killer in chronic kidney disease. Postgrad Med (in press) 36. Maschio G: How good are nephrologists at controlling blood pressure in renal patients? Nephrol Dial Transplant 14:2075-2077, 1999 37. Duncan L, Djurdjev O, Heathcote J, Levin A: Screening for renal disease using serum creatinine: Who are we missing? Nephrol Dial Transplant 16:1402-1404, 2001 38. Holland DC, Lam M: Predictors of hospitalization and death among pre-dialysis patients: A retrospective cohort study. Nephrol Dial Transplant 15:650-658, 2000
From the Division of Nephrology, University of British Columbia, Vancouver, British Columbia, Canada; Centre for Health Evaluation and Outcomes Sciences, St. Paul’s Hospital and University of British Columbia, Vancouver, British Columbia, Canada; Division of Nephrology, Memorial University, St. John’s, Newfoundland; Division of Nephrology, University of Calgary, Calgary, Alberta, Canada; Division of Nephrology, McMaster University, Hamilton, Ontario, Canada; Division of Nephrology, Universite de Montreal, Hopital St. Luc, Montreal, Quebec, Canada; Division of Nephrology, Queen Elizabeth II Health Sciences Centre, Dalhousie University, Halifax, Nova Scotia, Canada; Division of Nephrology, St. Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada; Division of Nephrology, Sunnybrook Hospital, University of Toronto, Toronto, Ontario, Canada; Division of Cardiology, St. Paul’s Hospital, University of British Columbia, Vancouver, British Columbia, Canada. Received and accepted as submitted on July 17, 2001. Address reprint requests to Adeera Levin, MD, FRCPC, Associate Professor of Medicine, St. Paul’s Hospital, Providence Wing, 1081 Burrard Street, Room 6010A, Vancouver, BC, Canada V6Z1 Y8. E-mail: [email protected] © 2001 by the National Kidney Foundation, Inc. 0272-6386/01/3806-0037$35.00/0 doi:10.1053/ajkd.2001.29275 1398
T
HE HIGH PREVALENCE and incidence of cardiovascular disease (CVD) in both dialysis patients and transplant patients is well described,1-5 as is the adverse impact of CVD on long-term outcomes.6-8 Risk factors recognized to contribute to CVD in the general population, hypertension and diabetes, are present in patients with chronic kidney disease (CKD). Numerous studies to date have shown an independent effect of kidney disease itself on CVD outcomes.7-11 Contribution of kidney dysfunction itself, and of other factors associated with kidney disease, to the poor outcomes is unknown. In chronic kidney disease populations, both on and not on dialysis, there are specific uremia-related risk factors that have been associated with CVD outcomes: most notable is the consistent association of anemia and left ventricular hypertrophy (LVH) and congestive heart failure (CHF).6,12-14 Abnormal calcium phosphate metabolism and hyperparathyroidism have also been associated with myocardial fibrosis, hypertrophy, and vascular calcification.15-18 However, there is limited
American Journal of Kidney Diseases, Vol 38, No 6 (December), 2001: pp 1398-1407
SYMPOSIUM PROCEEDINGS
understanding of the differential impact of each risk factor on kidney function and cardiovascular function, and the interaction between established CVD and kidney disease. The issue of reverse causality and survival bias confounds most analyses performed to date.19 We report here the results of data collected for the Canadian Multi-centre Study in Patients with Renal Insufficiency, commenced in 1994. This study is an observational cohort study funded by the Kidney Foundation of Canada. We have previously reported on the risk factors associated with left ventricular (LV) growth and LVH.12 We identified the increased prevalence of LVH at each level of creatinine clearance. Further, we found that anemia and systolic blood pressure (BP) were independent risk factors for cardiac growth and hypertrophy in patients with varying degrees of kidney dysfunction. In this same group of patients with established chronic kidney disease, this study attempts to answer additional questions: (1) Is there an independent impact of uremia-associated risk factors, such as hemoglobin (Hgb), intact parathyroid syndrome (iPTH), and calcium phosphate product, after correcting for the level of kidney function, on the prevalence and incidence of CVD in patients with chronic kidney disease; and (2) is there an independent impact of the presence of CVD on the progression of kidney disease, in addition to established risk factors such as hypertension, proteinuria, age, and sex? MATERIALS AND METHODS
Study Population and Design This Canadian, multicenter, cohort study enrolled patients who had been identified by nephrologists as having CKD. The details of the study have been reported elsewhere.12 Consecutive patients in 7 different geographic locations across Canada, who had calculated creatinine clearances between 75 and 10 mL/min (calculated by using the Cockcroft-Gault formula), who were expected by their physicians to live and to be off dialysis for at least 12 months of follow-up, were enrolled in the study. Informed consent was obtained from all patients, and the ethics board of each institution approved the study. There were no attempts to modify therapy during the study. All patients had clinical, laboratory, and cardiovascular (CV) symptoms reviewed at 1-year intervals. Hospitalization events and medication use were also evaluated. The data was captured by using standardized data collection sheets, and was then entered into a central database for analysis.
1399
Definitions of Cardiovascular Disease CV disease was defined by both history of CV event or condition and/or current symptoms according to New York Heart Association (NYHA) and Canadian Cardiovascular Society (CCS) classifications for heart failure and angina, respectively. CV event history included history of myocardial infarction (MI), angina, coronary artery bypass graft (CABG) or angioplasty, transient ischemic attack (TIA) or cardiovascular accident (CVA), peripheral vascular disease (PVD), and congestive heart failure (CHF). Research assistants administered standardized questionnaires that describe symptoms associated with each class of heart failure or angina to each patient, and symptoms were graded by using conventional NYHA and CCS classifications. Change in cardiovascular status was defined as either change in NYHA or CCS classification, or a cardiac-related hospitalization.
Chronic Kidney Disease: Calculations and Definitions Serum creatinine level was measured in each patient’s own laboratory, which was maintained through the study period. Calculations using the formula of Cockcroft Gault12 were used to determine the level of creatinine clearance. Four categories of kidney function were identified: mild (⬎50⬍75 mL/min), moderate (35–49 mL/min), moderate severe (25–34 mL/min), or severe (10–24 mL/min). Renal replacement therapy (RRT) was defined as the need for either dialysis or transplantation.
Outcomes of Interest The major outcomes of interest for this analysis included kidney survival (time to RRT) and change in CVD status.
Statistical Analysis The baseline clinical characteristics of patients were compared by using the analysis of variance (ANOVA) or KruskalWallis test for continuous variables, depending on distribution, and the 2 test for categoric variables. We used the 2 test for discrete variables and the t test or Wilcoxon rank-sum test, where appropriate, for continuous variables, to compare patients with and without CVD at baseline, as well as patients with change and without change of CVD status. Univariate odds ratios (ORs) were adjusted for degree of kidney function by using logistic regression. The logistic regression was used to determine the association between patient baseline characteristics and presence of CVD at baseline, and then to predict change in CVD status. Time to RRT therapy was calculated based on the time from study enrollment to the dialysis start or transplantation date. Renal survival was estimated by using the Kaplan-Meier method. The survival curves by presence of CVD were compared by using the log-rank test. We used proportional hazards regression models to assess the effect of baseline patient characteristics on time to RRT after controlling for baseline kidney function. The Cox regression model was used to identify independent and significant predictors of time to RRT. All multivariate models were developed by using a backward elimination
1400
LEVIN ET AL
technique, including the following variables: patient age, sex, diabetes status, creatinine clearance, systolic and diastolic BP, total cholesterol, triglycerides, hemoglobin, phosphate-calcium product, iPTH, and presence of CVD at baseline. A P value of less than 0.05 for 2-sided univariate tests was considered statistically significant. We set the P value for multivariate analyses to 0.10 to account for the hypothesisgenerating purpose of this analysis. Statistical analyses were performed by using SAS software, version 6.12 (SAS Institute, Cary, NC).
RESULTS
Baseline Characteristics The original study enrolled 447 patients, of whom 417 had an initial evaluation of cardiac status. We excluded 104 patients from this analysis because of missing baseline values for key variables of interest. Thus, this analysis Table 1.
Characteristic
Age yr Gender, % men Race, % Caucasian Diabetes CrCl mL/min 24-hr urine protein g/L Systolic BP mm Hg Diastolic BP mm Hg Pulse pressure mm Hg Total cholesterol g/L HDL mmol/L Triglycerides mmol/L Hemoglobin g/L Ca*PO4 mmol/L *mmol/L Log PTH pmol/L LVMI g/m2 ACEI alone or in combination -blockers alone or in combination Any lipid-lowering agent Any Hx or Sx of CVD Hx of any CVD Hx of angina Hx of CHF Hx of MI Hx of PVD NYHA I–II III–IV CCS I–II III–IV
includes 313 predominantly male (67%) and Caucasian (87%) patients, who had a mean age of 56 years (Table 1). The mean calculated creatinine clearance (Ccr) was 36.2 ⫾16.8 mL/min. Twenty nine percent of the group were diabetic, and the prevalence of the causes of kidney disease paralleled those in national registries: hypertension (27%), glomerulonephritis (18%), diabetes (25%), and polycystic kidney disease (PCKD) (8%). Table 1 shows the baseline demographics of the whole cohort, and compares a number of pertinent variables relative with baseline kidney function. Sex, hemoglobin, pulse pressure, calcium-phosphate product, iPTH, high-density lipoprotein (HDL), LVMI were different between groups with different kidney function.
Baseline Characteristics by Level of Renal Function
All n ⫽ 313
⬎50 mL/min n ⫽ 62
35–49 mL/min n ⫽ 83
56.3 ⫾ 14.5 211 (67.4%) 267 (87.0%) 91 (29.1%) 36.2 ⫾ 16.8 1.04 ⫾ 1.36 144.4 ⫾ 23.4 85.4 ⫾ 11.6 58.9 ⫾ 20.8 5.70 ⫾ 1.34 1.15 ⫾ 0.51 2.36 ⫾ 1.69 128.1 ⫾ 19.5 2.78 ⫾ 0.62 2.22 ⫾ 0.85 111.3 ⫾ 35.3 155 (49.7%)
53.5 ⫾ 13.4 54 (87%) 54 (88%) 21 (34%) 62.6 ⫾ 10.5 0.96 ⫾ 1.24 143.9 ⫾ 20.9 87.3 ⫾ 11.8 56.5 ⫾ 19.1 5.76 ⫾ 1.19 1.04 ⫾ 0.54 2.57 ⫾ 1.44 141.5 ⫾ 16.6 2.50 ⫾ 0.49 1.64 ⫾ 0.58 110.0 ⫾ 35.2 37 (59.7%)
54.9 ⫾ 15.3 64 (77%) 74 (91%) 24 (29%) 41.9 ⫾ 4.3 0.97 ⫾ 1.22 142.7 ⫾ 21.6 85.1 ⫾ 11.3 57.0 ⫾ 21.8 5.58 ⫾ 1.16 1.06 ⫾ 0.31 2.46 ⫾ 2.27 134.3 ⫾ 19.3 2.66 ⫾ 0.56 1.88 ⫾ 0.67 106.8 ⫾ 28.8 47 (56.6%)
74 (23.7%) 42 (13.5%) 143 (45.7%) 120 (38.3%) 52 (16.7%) 23 (7.6%) 39 (12.7%) 38 (12.4%)
15 (24.2%) 10 (16.1%) 27 (43.5%) 22 (35.5%) 11 (18.0%) 5 (8.3%) 10 (16.4%) 8 (13.1%)
20 (24.1%) 10 (12.1%) 41 (49.4%) 32 (38.6%) 16 (19.5%) 6 (7.5%) 10 (12.2%) 9 (11.0%)
16 (21.6%) 12 (16.2%) 29 (39.2%) 26 (35.14%) 8 (10.8%) 3 (4.2%) 10 (13.7%) 10 (13.9%)
23 (24.7%) 10 (10.9%) 46 (48.9%) 40 (42.6%) 17 (18.1%) 9 (9.8%) 9 (9.9%) 11 (12.1%)
0.970 0.676 0.788 0.743 0.472 0.596 0.686 0.953
76 (24.3%) 7 (2.2%)
15 (24.2%) 1 (1.6%)
25 (30.1%) 0 (0%)
13 (17.6%) 2 (2.7%)
23 (24.5%) 4 (4.3%)
0.331
70 (22.4%) 6 (1.9%)
16 (25.8%) 1 (1.6%)
20 (24.1%) 1 (1.2%)
13 (17.6%) 3 (4.1%)
21 (22.3%) 1 (1.1%)
0.712
Abbreviations: Hx, history; Sx, symptoms.
25–34 mL/min n ⫽ 74
56.5 ⫾ 14.7 47 (64%) 59 (81%) 16 (22%) 29.9 ⫾ 3.0 1.03 ⫾ 1.75 141.0 ⫾ 24.7 84.8 ⫾ 12.7 56.1 ⫾ 21.8 5.98 ⫾ 1.64 1.26 ⫾ 0.65 2.45 ⫾ 1.53 125.7 ⫾ 16.6 2.73 ⫾ 0.54 2.34 ⫾ 0.79 106.1 ⫾ 34.7 37 (50.0%)
10–24 mL/min n ⫽ 94
P Value
59.3 ⫾ 14.1 46 (49%) 80 (87%) 30 (32%) 18.8 ⫾ 3.8 1.22 ⫾ 1.25 149.3 ⫾ 24.4 84.9 ⫾ 10.9 64.4 ⫾ 21.2 5.56 ⫾ 1.30 1.22 ⫾ 0.48 2.07 ⫾ 1.29 115.6 ⫾ 15.3 3.12 ⫾ 0.65 2.80 ⫾ 0.81 120.2 ⫾ 39.8 34 (36.6%)
0.0604 0.001 0.268 0.383 0.0001 0.7165 0.0847 0.5519 0.0242 0.1592 0.0166 0.2424 0.0001 0.0001 0.0001 0.0474 0.002
SYMPOSIUM PROCEEDINGS Table 2.
1401
Conventional and Uremia-Associated Risk Factors in Patients With and Without CVD at Baseline
Variables
Age (yr) Gender, % men Diabetes CrCl (calculated) mL/min 24-hr urine protein g/L Systolic BP mm Hg Diastolic BP mm Hg Pulse pressure mm Hg Total cholesterol mmol/L HDL mmol/L Triglycerides mmol/L Hemoglobin g/L PO4*Ca mmol/L *mmol/L log PTH pmol/L LVMI g/m2 ACEI alone or in combination Beta blockers alone or in combination Any lipid-lowering agent
CVD (⫺) N ⫽ 170
CVD (⫹) N ⫽ 143
P Value
53.5 ⫾ 14.9 108 (63.5%) 37 (21.8%) 35.9 ⫾ 16.2 1.08 ⫾ 1.48 143.5 ⫾ 23.2 87.2 ⫾ 11.7 56.3 ⫾ 20.4 5.77 ⫾ 1.54 1.19 ⫾ 0.55 2.36 ⫾ 1.39 128.0 ⫾ 19.3 2.78 ⫾ 0.61 2.27 ⫾ 0.88 105.59 ⫾ 31.83 88 (51.8%) 34 (20.0%) 17 (10.0)
59.6 ⫾ 13.4 103 (72.0%) 54 (37.8%) 36.6 ⫾ 17.5 1.01 ⫾ 1.24 145.3 ⫾ 23.7 83.3 ⫾ 11.2 62.0 ⫾ 21.0 5.62 ⫾ 1.06 1.09 ⫾ 0.46 2.36 ⫾ 1.99 128.2 ⫾ 19.9 2.80 ⫾ 0.63 2.16 ⫾ 0.82 117.94 ⫾ 37.95 67 (47.2%) 40 (28.2%) 25 (17.7%)
0.0002 0.110 0.002 0.7000 0.6945 0.4950 0.0029 0.0154 0.3209 0.1168 0.9808 0.9196 0.7663 0.2246 0.0034 0.420 0.091 0.047
The use of angiotensin-converting enzyme inhibitor (ACEI) therapy was the lowest in those with the worst kidney function. The prevalence of CVD was similar at each level of kidney function.
the multivariate analysis and adjusted for severity of kidney disease to identify older age, diabetes, and low diastolic BP as factors associated with increasing the probability of the CVD presence at baseline (Table 3).
Prevalence of Cardiovascular Disease at Baseline: A Comparison of Conventional and Uremia-Related Risk Factors Comparisons between patients with and without CVD at baseline are presented in Table 2. Those with CVD at baseline were likely to be older, diabetic, have lower diastolic BPs, higher pulse pressures, and larger values for LVMI. However, in this cohort, the presence or absence of CVD at baseline is not different with respect to creatinine clearance (35.9 mL/min versus 36.6 mL/min, P ⫽ 0.700). Again, the use of ACEI was not different as a function of presence or absence of CVD, but lipid-lowering agents were used more often in those with CVD than those without (18% versus 10%, P ⫽ 0.047). After adjustment for the level of renal function, the same risk factors (ie, age [OR ⫽ 1.17, P ⫽ 0.0002], diabetes [OR ⫽ 2.18, P ⫽ 0.0022], diastolic BP [OR ⫽ .81, P ⫽ 0.0002], pulse pressure [OR ⫽ 1.07, P ⫽ 0.013], and LVMI [OR ⫽ 1.11, P ⫽ 0.0033]), were associated with presence or absence of CVD at baseline. We used
Cardiovascular Disease—Change in CVD Status During Follow-Up: Impact of Preexisting CVD Of the original 313 patients included in this analysis, 36 patients were lost or refused further follow-up, 8 patients died, 24 patients started dialysis, and 1 patient was transplanted before at least 1 cardiovascular follow-up evaluation. Hence, 244 patients had evaluable cardiovascular follow-up: median duration of follow-up was 23 months, with a total of 5,547 months or 462 patient years of follow-up. During the follow-up Table 3. Multivariate Analysis of Risk Factors Associated With the Presence of CVD at Baseline
Variables
Odds Ratio
95% Confidence Interval
P Value
Age 5 years Male gender Diabetes CrCl 5 mL/min Diastolic BP 5 mm Hg
1.143 1.355 2.024 1.025 0.898
1.047–1.248 0.793–2.313 1.206–3.397 0.952–1.103 0.808–0.998
0.0028 0.2663 0.0076 0.5169 0.0464
1402
LEVIN ET AL
period, 48 of 244 (20%) patients developed new or worsening cardiac symptoms or were hospitalized for cardiac disease. Of the 135 patients who had no preexisting CVD, 10 (7.4%) developed new CVD; this is in contrast to the 109 patients with preexisting cardiac disease, of which 38 (35%) developed a new event or worsening of CVD (P ⫽ 0.001). Thus, 21% of the 48 patients with new or worsening cardiac disease had no previous CVD. Table 4 describes the results of a univariate analysis comparing the association of conventional and uremia-related variables with change in CVD status, in patients with and without preexisting CVD. Note that age and diabetes at baseline predict change in CVD status in both groups of patients, whereas higher serum triglyceride levels and lower diastolic BP predict worsening of CVD among patients with preexisting CVD over the follow-up period. There were no apparent associations of uremia-specific factors such as hemoglobin, calcium phosphate product, iPTH level, or of level of kidney dysfunction itself, with change of CVD status. The same baseline characteristics were associated with change in CVD status, depending on the presence or absence of CVD at baseline, after adjustment for level of kidney function. A multivariate analysis revealed that diabetes increases the probability of change in CVD status in those without preexisting CVD (Fig 1). Low diastolic BP and Table 4.
high triglyceride levels were independent and significant predictors of change in CVD status in those with preexisting CVD (Fig 2). We performed a separate multivariate analysis for all patients who had a change in NYHA class heart failure versus patients who had change in CCS class angina. The important predictors of NYHA class change were older age (OR ⫽ 1.21 per 5 years, P ⫽ 0.0588), lower diastolic BP (OR ⫽ .79 per 5 mm Hg, P ⫽ 0.0175), higher triglyceride levels (1.44 per mmol/L, P ⫽ 0.0118), and presence of CVD at baseline (OR ⫽ 4.91, P ⫽ 0.0011). The change in CCS class was associated with lower baseline kidney function (OR ⫽ .84 per 5 mL/min, P ⫽ 0.0624), diabetes (OR ⫽ 3.96, P ⫽ 0.0035), lower diastolic BP (OR ⫽ .79 per 5 mm Hg, P ⫽ 0.0284), and baseline CVD (OR ⫽ 3.88, P ⫽ 0.0111). Loss of Kidney Function: Conventional and Other Predictors of Time to Renal Replacement Therapy Table 5 describes risk factors associated with time to RRT, adjusted for baseline kidney function. In addition to known risk factors for progressive kidney function decline (age, sex, diabetes, proteinuria, BP, and kidney function), there are uremia-specific factors—hemoglobin, iPTH, calcium phosphate product— that may also be associated with decline. Furthermore, the presence of CVD, as well as low HDL and cholesterol
Patient Characteristics Associated With CVD Status in Patients With and Those Without CVD at Baseline No Preexisting CVD N ⫽ 135
Variables
Age (yr) Gender, % men Diabetes CrCl mL/min 24-hr urine protein g/L Systolic BP mm Hg Diastolic BP mm Hg Pulse pressure mm Hg Total cholesterol mmol/L HDL mmol/L Triglycerides mmol/L Hemoglobin g/L Ca*PO4 mmol/L *mmol/L Log PTH pmol/L LVMI g/m2
Preexisting CVD N ⫽ 109
No CVD n ⫽ 125
CVD n ⫽ 10
P Value
Stable n ⫽ 71
Worsening n ⫽ 38
P Value
52.3 ⫾ 15.4 80 (64.0%) 21 (16.8%) 37.8 ⫾ 16.3 0.84 ⫾ 1.01 141.2 ⫾ 23.8 86.2 ⫾ 11.8 55.0 ⫾ 20.7 5.66 ⫾ 1.27 1.21 ⫾ 0.60 2.36 ⫾ 1.32 130.6 ⫾ 19.3 2.70 ⫾ 0.61 2.24 ⫾ 0.84 103.28 ⫾ 32.17
61.8 ⫾ 13.1 5 (50.0%) 5 (50.0%) 33.7 ⫾ 16.1 1.08 ⫾ 1.29 149.8 ⫾ 27.9 87.0 ⫾ 14.1 62.8 ⫾ 22.8 6.34 ⫾ 1.44 1.11 ⫾ 0.30 2.54 ⫾ 1.22 121.8 ⫾ 17.1 2.87 ⫾ 0.37 1.97 ⫾ 0.76 115.61 ⫾ 34.95
0.0592 0.499 0.023 0.4440 0.5246 0.2808 0.8492 0.2562 0.1102 0.3604 0.6816 0.1659 0.3926 0.3339 0.2738
58.4 ⫾ 13.3 51 (71.8%) 19 (26.8%) 37.8 ⫾ 17.3 0.75 ⫾ 1.05 147.0 ⫾ 24.8 85.8 ⫾ 10.7 61.2 ⫾ 22.9 5.57 ⫾ 0.90 1.16 ⫾ 0.43 2.03 ⫾ 1.30 127.5 ⫾ 19.0 2.77 ⫾ 0.59 2.06 ⫾ 0.79 115.7 ⫾ 37.35
63.2 ⫾ 10.7 26 (68.42%) 17 (44.7%) 32.8 ⫾ 13.1 1.48 ⫾ 1.53 141.8 ⫾ 24.3 78.3 ⫾ 11.1 63.5 ⫾ 19.79 5.89 ⫾ 1.19 1.04 ⫾ 0.60 3.04 ⫾ 3.09 129.0 ⫾ 18.2 2.83 ⫾ 0.48 2.26 ⫾ 0.78 125.38 ⫾ 42.64
0.0564 0.709 0.057 0.1224 0.0278 0.2980 0.0008 0.5950 0.1228 0.5958 0.0183 0.6867 0.5827 0.1957 0.2414
SYMPOSIUM PROCEEDINGS
1403
Fig 1. Factors associated with change of CVD status in patients without history or symptoms of CVD at baseline.
levels at baseline, are associated with increased risk for kidney failure leading to RRT. There were several factors that independently and significantly predicted decline of kidney function in the multivariate analysis, shown in Fig 3: older age, male gender, diabetes, low baseline kidney function and hemoglobin, and high diastolic blood pressure and iPTH level. In terms of the impact of CVD on time to RRT, there was only a trivial difference after adjustment for all of the earliermentioned factors (univariate relative risk [RR] ⫽
1.58, multivariate RR ⫽ 1.52), indicating the importance of this factor. Figure 4 shows the impact of CVD presence or absence at baseline on time to RRT. DISCUSSION
This analysis of a cohort of patients with known chronic kidney disease describes the high prevalence of cardiac disease (46%) and of conventional cardiovascular risk factors. The analysis does not show that uremia-specific risk fac-
Fig 2. Factors associated with change of CVD status in patients with history or symptoms of CVD at baseline.
1404
LEVIN ET AL
Table 5. Associations of Baseline Characteristics, Adjusted for Kidney Function, With Time to RRT
Variables
Risk Ratio
95% Confidence Interval
P Value
Age 5 yr Gender, male Diabetes CrCl 5 mL/min 24-hr urine protein g/L Systolic BP 5 mm Hg Diastolic BP 5 mm Hg Pulse pressure 5 mm Hg Total cholesterol mmol/L HDL mmol/L Triglycerides mmol/L Hemoglobin 5 g/L Ca*PO4 mmol/L *mmol/L log PTH pmol/L Hx. or Sx. of CVD
0.879 1.957 2.181 0.708 1.550 0.990 1.120 0.997 0.819 0.614 0.935 0.863 1.651 1.613 1.580
0.816-0.947 1.204-3.181 1.376-3.457 0.639-0.785 1.377-1.746 0.942-1.041 1.004-1.250 0.950-1.046 0.680-0.988 0.351-1.074 0.790-1.106 0.804-0.926 1.229-2.217 1.199-2.168 1.006-2.482
0.0007 0.0068 0.0009 0.0001 0.0001 0.6991 0.0417 0.9042 0.0371 0.0876 0.4315 0.0001 0.0009 0.0016 0.0470
tors predict either the prevalence of CVD or the change in CVD status. Previous publications have described a similarly high prevalence of CVD risk factors and have determined that chronic kidney disease itself was a risk factor for CV outcome and mortality.10,20,21 All of these studies compared patients with kidney disease with those without kidney disease, did not adjust for specific levels of kidney function, and did not measure specific variables associated with kidney disease—hemoglobin, iPTH, calcium, phos-
phate—that may mediate the impact of CKD on outcomes. Notably, Culleton et al22 reported findings similar to ours in the Framingham cohort, despite a higher prevalence of CVD in patients with CKD at baseline, there was no impact of CKD on long-term outcomes after controlling for conventional risk factors. That uremia-specific factors did not appear to exert an effect on CVD prevalence/incidence independent of conventional cardiac risk factors could be owing to study limitations, rather than to the absence of such an effect. Furthermore, the prevalence of ischemia-related CVD (angina, CVA, MI, PVD) at baseline is similar in this population to that in other renal populations at later stages of disease (ie, on dialysis and posttransplant).1,3,23,24 This finding supports the notion recently discussed in depth by Parfrey25 in an editorial: cardiac function can be affected by processes different from simple atheroscleroticinduced ischemia, and CVD with progressive kidney disease may be secondary to the influence of both atherogenic processes and factors that modulate cardiac muscle structure function. We did not measure LV ejection fractions in this study, rather, we only have a measure of cardiac hypertrophy—LVMI and pulse pressure. Pulse pressure, as an indicator of vascular stiffness, has recently been described as predicting morbidity and mortality, and contributing to cardiac dysfunction26,27 in cardiac populations. The current
Fig 3. Independent predictors of time to RRT (results of Cox proportional hazards modeling).
SYMPOSIUM PROCEEDINGS
1405
Fig 4. Kaplan-Meier curves of time to RRT by CVD status at baseline.
analysis shows that both LVMI and pulse pressure are significantly different between those patients with and without CVD at baseline. Progressive cardiovascular disease and progressive decline of kidney function do have a number of risk factors in common, including hypertension and diabetes. There are some risk factors specific to each condition and some that occur as a consequence of 1 of these impacting another. It may be that the purported uremic factors exert their influence through affecting cardiac function, not the atherosclerotic process as previously thought. Interestingly, it is possible that these same factors that affect cardiac function may also accelerate decline of kidney function, both through activation of cytokines and growth factors, and that ongoing endothelial cell dysfunction may contribute simultaneously to both processes.16,30 There is substantial animal and cellular data to support the impact of these factors on progressive kidney disease.28,31 However, the current analysis is unable to determine the relative impact of each of these. This study further shows the association of conventional risk factors with kidney function decline and also shows an additional association of CVD presence with kidney survival. The data also suggest that anemia and hyperparathyroidism may contribute to progression of kidney disease, independent of kidney function level, sex, diabetes, or age.32 The presence of CVD at baseline confers a 50% increase in probability of
RRT. The explanation for this may be complex and multifactorial. Worsening cardiac function may lead to reduced blood flow to the kidney, thus, adding a prerenal insult to kidneys with little reserve, and to activation of cellular mechanisms that lead to sclerosis and fibrosis. Alternatively, cardiac dysfunction may lead to the addition of medications/interventions that may further cause kidney injury or reduce function. These hypotheses are not mutually exclusive. Interestingly, a recent report by Silverberg et al33 describes the treatment of heart failure and associated anemia with erythropoietin. Importantly, the treatment group had significantly impaired kidney function (serum creatinine 2.3 mg/dL) before therapy. After correction of anemia and concomitant with improvement in cardiac status there was improvement of kidney function.33 The current analysis, showing that CV disease and conventional risk factors for cardiac disease adversely affect the progression of kidney disease, provides hope that attention to CV health and risk factor modification might improve longterm outcomes for the progression of kidney disease. In this cohort, as in others,8,34-36 there is suboptimal use of cardiovascular medications such as ACEIs, -blockers, and lipid-lowering medications. It may be that both cardiologists and nephrologists caring for patients with established kidney disease actually practice therapeutic nihilism.35 This inadequate therapy of cardiac disease may accelerate the rate of kidney func-
1406
tion decline in susceptible individuals. Thus, the high prevalence of those with established cardiac disease commencing dialysis may be the consequence of failure to institute treatment before dialysis. This study is unique in that it describes the prevalence of CVD and its attendant risk factors in patients with kidney disease before dialysis known to nephrologists in Canada. Others have described unselected groups of general populations, or used population databases or post hoc analyses of large studies to review the impact of kidney disease on CV outcomes.7,9,10,22 These approaches suffer from a number of problems. There is a lack of uniformity in defining the presence and severity of kidney disease. There is little information about the stability of kidney dysfunction before the study starts, and there is no documentation of the etiology of kidney dysfunction or of uremia-associated biochemical and hematologic abnormalities. There is ample evidence to support the problems of use of serum creatinine alone as a marker for kidney disease.37 The current analysis, in contrast, focuses on individuals with established, nonreversible kidney disease, by using calculated estimates of kidney function. By examining those variables specifically associated with kidney disease, we are able to explore the hypothesis that there is an impact of these variables versus conventional risk factors versus kidney disease per se on CVD. This study also describes the complex interplay between CVD and progressive CKD. This is important if therapeutic strategies to treat one disease may actually be justified in terms of their impact on another aspect of this disease. Our study, like all prospective observational studies, precludes the establishment of causality. Furthermore, owing to missing values and loss of patients to follow-up, the true burden of illness in this population may be underestimated. This study did not capture events after RRT, or specific cardiac events. We believe, however, that the similarity of the prevalence of CVD in this cohort to that reported in other similar cohorts34,38 makes it unlikely that there is significant underestimation of CVD. The issue of reverse causality, surviving off RRT to actually have a CVD event, and survival bias (ie, those with worse CVD were put onto RRT as a means of controlling symptoms) cannot be addressed
LEVIN ET AL
with this analysis, or in this cohort. Importantly, this data can be used to generate hypotheses and develop interventional studies based on biologically plausible associations of specific risk factors. Ideally, we should establish follow-up in 2 distinct cohorts of patients: those with established kidney disease without CVD, and those with established cardiac disease without kidney dysfunction, and determine through careful and systematic measurement of known risk factors which of these are predisposed to developing into cardiac disease and to kidney disease, respectively. In summary, the prevalence of CVD in the population with CKD is high, and is attributable to both ischemic disease and myocardial dysfunction. This study shows an underuse of medications known to improve cardiac outcome in the general population. This may contribute to the poor cardiac outcomes seen in this group. Furthermore, the presence of CVD appears to predict an increased probability of RRT; thus, the appropriate treatment of cardiac disease may in fact lead to delay of RRT, and thus serve to lessen the burden of illness not only for the individual patients, but also for health care systems. ACKNOWLEDGMENT The authors would like to acknowledge the support of the Kidney Foundation of Canada, the extensive assistance of Dr. Vadim Minster and Ms. Cheyenne Reese for data organization and manuscript preparation, and all of the research assistants in each of the participating centers for their extensive data collection.
REFERENCES 1. Foley RN, Parfrey PS, Harnett JD, Kent GM, Martin CJ, Murray DC, Barre PE: Clinical and echocardiographic disease in patients starting end-stage renal disease therapy. Kidney Int 47:186-192, 1995 2. Canadian Organ Replacement Registry, Canadian Institute for Health Information (CIHI), Don Mills, Ontario, Canada, 1999 3. Parfrey PS: Cardiac disease in dialysis patients: Diagnosis, burden of disease, prognosis, risk factors and management. Nephrol Dial Transplant 15:58-68, 2000 4. Sarnak MJ, Levey AS: Epidemiology, diagnosis, and management of cardiac disease in chronic renal disease. J Thromb Thrombolysis 10:169-180, 2000 5. Kasiske B, Guijarro C, Massy Z, Wiederkehr M, Ma J: Cardiovascular disease after renal transplantation. J Am Soc Nephrol 7:158-165, 1996 6. Foley RN, Parfrey PS, Harnett JD, Kent GM, Murray DC, Barre PE: The impact of anemia on cardiomyopathy, morbidity and mortality in end-stage renal disease. Am J Kidney Dis 28:53-61, 1996
SYMPOSIUM PROCEEDINGS
7. Chertow GM, Normand S-LT, Silva LR, McNeil BJ: Survival after acute myocardial infarction in patients with end-stage renal disease: Results from the Cooperative Cardiovascular Project. Am J Kidney Dis 35:1044-1051, 2000 8. Herzog CA: Poor long-term survival of dialysis patients after acute myocardial infarction: Bad treatment or bad disease? Am J Kidney Dis 35:1217-1220, 2000 9. Mann J, Gerstein HC, Pogue J, Baosch, Yusef S: Renal insufficiency as a predictor of cardiovascular outcomes and the impact of ramipril: The HOPE randomized trial. Ann Intern Med 134:629-636, 2001 10. Hemmelgarn BR, Ghali WA, Quan H, Brant R, Norris CM, Taub KJ, Knudtson ML: Poor long-term survival after coronary angiography in patients with renal insufficiency. Am J Kidney Dis 37:64-72, 2001 11. Beattie JN, Soman SS, Sandberg KR, Yee J, Borzak S, McCullough PA: Determinants of mortality after myocardial infarction in patients with advanced renal dysfunction. Am J Kidney Dis 37:1191-1200, 2001 12. Levin A, Thompson CR, Ethier J, Carlisle EJ, Tobe S, Mendelssohn D, Burgess E, Jindal K, Barrett B, Singer J, Djurdjev O: Left ventricular mass index increase in early renal disease: Impact of decline in hemoglobin. Am J Kidney Dis 34:125-134, 1999 13. O’Riordan E, Foley RN: Effects of anaemia on cardiovascular status. Nephrol Dial Transplant 15:19-22, 2000 14. Levin A, Foley RN: Cardiovascular disease in chronic renal insufficiency. Am J Kidney Dis 36:24-30, 2000 15. Block GA, Port FK: Re-evaluation of risks associated with hyperphosphatemia and hyperparathyroidism in dialysis patients: Recommendations for a change in management. Am J Kidney Dis 35:1226-1237, 2000 16. Pannier B, Guerin AP, Marchais SJ, Metivier F, Safar ME, London GM: Postischemic vasodilation, endothelial activation, and cardiovascular remodeling in end-stage renal disease. Kidney Int 57:1091-1099, 2000 17. Goodman WG, Goldin J, Kuizon BD, Yoon C, Gales B, Sider D, Wang Y, Chung J, Emerick A, Greaser L, Elashoff RM, Salusky IB: Coronary-artery calcification in young adults with end-stage renal disease who are undergoing dialysis. N Engl J Med 342:1478-1483, 2000 18. Amann K, Ritz E, Wiest G, Klaus G, Mall G: A role of parathyroid hormone for the activation of cardiac fibroblasts in uremia. J Am Soc Nephrol 4:1814-1819, 1994 19. Baigent C, Burbury K, Wheeler D: Premature cardiovascular disease in chronic renal failure. Lancet 356:147152, 2000 20. Anderson RJ, O’Brien M, MaWhinney S, VillaNueva CB, Moritz TE, Sethi GK, Henderson WG, Hammermeister KE, Grover FL, Shroyer AL: Mild renal failure is associated with adverse outcome after cardiac valve surgery. Am J Kidney Dis 35:1127-1134, 2000 21. McCullough PA, Soman SS, Shah SS, Smith ST, Marks KR, Yee J, Steven Borzak S: Risks associated with renal dysfunction in coronary care unit patients. J Am Coll Cardiol 36:679-684, 2000 22. Culleton BF, Larson MG, Wilson PWF, Evans JC, Parfrey PS, Levy D: Cardiovascular disease and mortality in
1407
a community-based cohort with mild renal insufficiency. Kidney Int 56:2214-2219, 1999 23. Parfrey PS, Harnett JD, Foley RN: Heart failure and ischemic heart disease in chronic uremia. Curr Opin Nephrol Hypertens 4:105-110, 1995 24. Rigato C, Jeffery J, Foley R, Brown S, Tribula C, Negrijin C, Parfrey P: Risk factors for de novo ischaemic heart disease in renal transplant recipients. J Am Soc Nephrol 11:705A, 2000 (abstr) 25. Parfrey PS: Is renal insufficiency an atherogenic state? Reflection on prevalence, incidence and risk. Am J Kidney Dis 37:154-156, 2001 26. Mitchell GF: Pulse pressure, arterial compliance and cardiovascular morbidity and mortality. Curr Opin Nephrol Hypertens 8:335-342, 1999 27. Safar ME: Systolic blood pressure, pulse pressure and arterial stiffness as cardiovascular risk factors. Curr Opin Nephrol Hypertens 10:257-261, 2001 28. Hunter JJ, Chien KR: Signaling pathways for cardiac hypertrophy and failure. N Engl J Med 341:1276-1283, 1999 29. Amann K, Kronenberg G, Gehlen F, Wessels S, Orth S, Munter K, Ehmke H, Mall G, Ritz E: Cardiac remodeling in experimental renal failure—an immunohistochemical study. Nephrol Dial Transplant 13:1958-1966, 1998 30. London GM, Fabiani F, Marchais SJ, et al: Uremic cardiomyopathy: An inadequate left ventricular hypertrophy. Kidney Int 31:973-980, 1987 31. Tonelli M, Bohm C, Pandeya S, Gill J, Levin A, Kiberd BA: Cardiac risk factors and the use of cardioprotective medications in patients with chronic renal insufficiency. Am J Kidney Dis 37:484-489, 2001 32. Hebert LA, Wilmer WA, Falkenhain ME, LadsonWofford SE, Nahman NS, Rovin BH: Renoprotection: One or many therapies? Kidney Int 59:1211-1226, 2001 33. Silverberg D, Wexler D, Blum M, Keren G, Sheps D, Laibovitch E, Brosh D, Laniado S, Schwartz D, Yachnin T, Shapira I, Gavish D, Baruch R, Koifman B, Kaplan C, Steinbruch S, Iaina A: The use of subcutaneous erythropoietin and intravenous iron for the treatment of the anemia of severe, resistant congestive heart failure improves cardiac and renal function and functional cardiac class, and markedly reduces hospitalizations. J Am Coll Cardiol 35:17371744, 2000 34. Taal M, Omer SA, Nadim MK, McKenzie HS: Cellular and molecular mediators in common pathway mechanisms of chronic disease progression. Curr Opin Nephrol Hypertens 9:323-331, 2000 35. Levin A, Stevens L, McCullough P: Cardiovascular disease: The killer in chronic kidney disease. Postgrad Med (in press) 36. Maschio G: How good are nephrologists at controlling blood pressure in renal patients? Nephrol Dial Transplant 14:2075-2077, 1999 37. Duncan L, Djurdjev O, Heathcote J, Levin A: Screening for renal disease using serum creatinine: Who are we missing? Nephrol Dial Transplant 16:1402-1404, 2001 38. Holland DC, Lam M: Predictors of hospitalization and death among pre-dialysis patients: A retrospective cohort study. Nephrol Dial Transplant 15:650-658, 2000