Atheroma Progression in Chronic Kidney Disease - Semantic Scholar

Atheroma Progression in Chronic Kidney Disease Claudio Rigatto,* Adeera Levin,† Andrew A. House,‡ Brendan Barrett,§ Euan Carlisle,储 and Adrian Fine* *Department of Internal Medicine and Section of Nephrology, University of Manitoba, Winnipeg, Canada; †Department of Medicine and Division of Nephrology, University of British Columbia, Vancouver Canada; ‡Department of Medicine and Division of Nephrology, University of Western Ontario, London, Canada; §Department of Medicine and Division of Nephrology, Memorial University of Newfoundland, St. John’s, Canada; and 储Department of Medicine, McMaster University, Hamilton, Canada Background and objectives: Cardiovascular events are 10 to 100 times more frequent in chronic kidney disease (CKD). We tested the hypothesis that the rate of atherosclerotic plaque growth is faster in severe versus moderate CKD. Design, setting, participants, & measurements: We performed a prospective cohort study in 318 prevalent CKD patients with initial creatinine clearance (CCr) between 20 and 50 ml/min/1.73 m2. Baseline clinical and laboratory data were obtained on all patients. Plaque area was determined every 6 mo using bilateral carotid ultrasonography. Plaque area distribution was normalized using a cube root transformation. Unadjusted and adjusted associations between CCr quintiles and rate of change in the transformed plaque area were assessed using multiple linear regression. Results: The rate of plaque progression appeared lower in patients with the lowest CCr. Median rate of plaque growth was 0.4 mm2/yr in the lowest quintile of CCr (< 23 ml/min/1.73 m2) versus 5.0 mm2/yr in the highest quintile (> 43 ml/min/1.73 m2). This association remained significant after adjustment for potential confounders. A secondary analysis using quintiles of Modification of Diet in Renal Disease (MDRD) GFR confirmed the absence of increased plaque growth at low GFR, although a reduced rate of growth in the lowest quintile of MDRD GFR was not observed. Conclusion: We did not observe accelerated plaque growth at low levels of renal function. We suggest that mechanisms other than plaque growth are responsible for the observed excess of cardiovascular disease in CKD patients. Clin J Am Soc Nephrol 4: 291–298, 2009. doi: 10.2215/CJN.01840408

C

ardiovascular disease rates are very high in patients with renal disease (1), but the underlying mechanisms are incompletely understood. Traditional cardiovascular risk factors do not explain the increased risk (2), and observational studies have observed paradoxical or absent associations between classical risk factors (such as cholesterol and BP) and mortality in dialysis patients. A large randomized controlled trial, the 4D Study (Die Deutsche Diabetes Dialyze Studie), did not demonstrate benefit from cholesterol reduction with statins in diabetic dialysis patients (3). The ALERT (Assessment of Lescol in Renal Transplantation) study found that statin therapy in kidney transplant patients reduced the rates of several secondary cardiovascular outcomes, although differences between groups in the primary outcome did not achieve statistical significance (4). These results are not perfectly consistent with the paradigm of “accelerated atherosclerosis” proposed by Lindner and colleagues over 30 years ago (5). The present study tested the hypothesis that the rate of plaque growth is progressively increased at progressively lower creatinine clearance (CCr) in patients with impaired kidReceived April 19, 2008. Accepted October 30, 2008. Published online ahead of print. Publication date available at www.cjasn.org. Correspondence: Dr. Claudio Rigatto, University of Manitoba, 409 Tache Avenue, Winnipeg, Manitoba R2H 2A6, Canada. Phone: 204-237-2121; Fax: 204-2332770; E-mail: [email protected] Copyright © 2009 by the American Society of Nephrology

ney function. We asked three fundamental questions: (1) how much intimal plaque is present in patients with reduced kidney function, (2) what is the rate of progression of atherosclerosis in such patients, and (3) is progression faster in patients with more severe depression in renal function? We studied a cohort of patients with chronic kidney disease not on dialysis, reasoning that risk factor associations might be more easily discerned before end stage renal disease.

Materials and Methods The study was conducted in five university-affiliated Canadian academic nephrology centers (Winnipeg, Manitoba; London, Ontario; Vancouver, British Columbia; Hamilton, Ontario and St John’s, Newfoundland). The study was approved by the research ethics boards of all institutions involved. Informed consent was obtained for all participants.

Study Population Between 1999 and 2003, we enrolled 318 patients with 24-h urine CCr between 20 and 50 ml/min/1.73 m2 at the participating centers. Study entry criteria based on CCr were developed and implemented before the widespread acceptance of the current chronic kidney disease (CKD) staging criteria and Modification of Diet in Renal Disease (MDRD) estimation formulas. Our study patients correspond approximately to CKD stages 3 and 4 as defined in the current Kidney Disease Outcomes Quality Initiative (KDOQI) guidelines. Patients were excluded if they had active autoimmune disease or malignancy. ISSN: 1555-9041/402–0291

0.06 0.9 0.1 0.7

Pa

43 40 44 29 29 10 28 17 28 11 3 (3 to 4) 12 (11 to 22) 16 (0 to 75) 1.0 (0 to 11)

5.1 (1.2) 2.9 (1.0) 1.2 (0.4) 2.5 (1.5)

43 43 38 30 30 4 24 8 19 6 3 (3 to 4) 12 (11 to 19) 39 (8 to 141) 6.0 (0 to 20)

5.0 (1.1) 3.1 (1.0) 1.2 (0.4) 2.3 (1.8)

55 40 36 16 22 4 22 9 21 0 3 (3 to 5) 18 (12 to 26) 20 (0 to 88) 3.9 (0 to 15)

5.2 (1.3) 3.1 (1.1) 1.2 (0.4) 2.4 (1.4)

46 27 36 19 32 3 21 3 17 0 3 (3 to 5) 13 (11 to 25) 19 (2 to 51) 5 (0 to 18)

5.3 (1.1) 3.1 (1.0) 1.2 (0.4) 2.3 (1.6)

0.6 0.2 0.5 0.2 0.8 0.4 0.9 0.007 0.5 0.03 0.06 0.07 0.08 0.02

0.6 0.6 0.2 0.9

a Comparison between quintiles using ANOVA for normally distributed variables, Kruskal-Wallis H test for non-normally distributed measures, and ␹2 test for categorical values. bDocumented prior myocardial infarction, ischemic stroke or peripheral arterial disease. cAngiotensin converting enzyme inhibitor. dAngiotensin II receptor blocker. eIQR, interquartile range.

44 46 41 21 30 8 27 22 16 13 3 (3 to 5) 13 (11 to 23) 32 (0 to 84) 0.4 (to 2 to 7)

5.0 (1.2) 3.2 (1.2) 1.1 (0.4) 2.3 (2.0)

5.2 (1.2) 3.1 (1.1) 1.2 (0.4) 2.4 (1.6)

(ⱖ43) 61 (11) 33 21 54

5

Clinical Journal of the American Society of Nephrology

46 61 38 23 27 6 25 12 20 9 3 (3 to 5) 13 (11 to 24) 23 (1 to 84) 3.0 (0 to 14)

25 13 6 11 14 140 (23) 78 (9) 38 (4) 1.2 (0.4 to 2.3) 119 (17) 2.29 (0.16) 1.37 (0.33) 3.9 (1.5)

21 16 8 7 29 140 (22) 79 (11) 38 (4) 1.0 (0.3 to 2.5) 126 (17) 2.3 (0.1) 1.2 (0.3) 3.6 (1.4)

(35 to ⬍ 43) 56 (14) 31 42 52

4

22 19 22 17 0.8 17 23 12 21 0.4 7 6 13 6 0.4 8 4 7 3 0.4 21 36 28 17 0.04 138 (21) 144 (24) 141 (22) 139 (21) 0.6 78 (11) 77 (12) 81 (10) 79 (10) 0.4 38 (4) 38 (5) 39 (4) 39 (4) 0.5 1.1 (0.3 to 2.7) 1.0 (0.4 to 2.6) 0.8 (0.2 to 2.6) 0.6 (0.2 to 2.3) 0.4 122 (16) 124 (14) 129 (17) 132 (16) ⬍0.001 2.26 (0.14) 2.33 (0.14) 2.32 (0.12) 2.37 (0.15) 0.01 1.25 (0.27) 1.16 (0.24) 1.18 (0.29) 1.11 (0.20) ⬍0.001 3.6 (1.5) 3.6 (1.4) 3.6 (1.4) 3.5 (1.9) 0.6

(29 to ⬍5) 61 (12) 40 28 62

(23 to ⬍29) 57 (12) 35 35 53

(⬍ 23) 57 (11) 37 33 49

58 (12) 35 32 54

Age (y) (SD) Female gender (%) Diabetes (%) Smokers (%) Renal diagnosis (%) diabetic nephropathy hypertension glomerulonephropathy polycystic kidney disease History of cardiovascular disease (%)b Systolic BP (mmHg) (SD) Diastolic BP (mmHg) (SD) Serum albumin (g/L) (SD) Proteinuria (g/day), median (IQR) Hemoglobin (g/L) (SD) Serum calcium (mmol/L) (SD) Serum phosphate (mmol/L) (SD) PTH (LN pmol/L) (SD) Lipids (mmol/L) (SD) total cholesterol LDL HDL Triglycerides Medications (%) ACEIc or ARBd diuretic calcium blocker beta blocker HMG-CoA reductase inhibitors fibric acid derivatives phosphate binders vitamin D3 ASA erythropoietin Median number of plaque measurements (IQR)e Median follow-up, months (IQR) Median plaque area, mm2 (IQR) Median rate of plaque growth, mm2/yr (IQR)

3

2

Creatinine clearance quintiles (ml/min/1.73 m2) 1

Entire cohort

Variable

Table 1. Baseline characteristics of cohort grouped by creatinine clearance quintile

292 Clin J Am Soc Nephrol 4: 291–298, 2009

Clin J Am Soc Nephrol 4: 291–298, 2009

Plaque Growth in CKD

293

Data Collection

Other Variables

Clinical data were obtained by oral history combined with review of all available medical records, including electrocardiograms and echocardiograms. All clinical data were collected on standardized forms and faxed to a central data management center (Dr. Levin, Centre for Health Outcomes and Evaluation, University of British Columbia, Vancouver), where the data were entered into an electronic database. Timed Urine Collections. Twenty-four hour urine collection for CCr and protein excretion was obtained at baseline. Patients did their own collection. Criteria for adequate collection were as follows: urine volume greater than 1 L and urine creatinine excretion greater than 0.13 ⫾ 0.02 mmol/kg/d. Patients with suspected inadequate collections were reminded of proper collection protocols, and the collection was repeated.

The following clinical and laboratory variables were recorded at baseline: age, gender, diabetic status, race, and original renal disease; smoking history and alcohol use; previous history of hypertension, MI, coronary revascularization, congestive heart failure, ischemic stroke, and amputation for peripheral vascular disease; height, weight, BP, hemoglobin and serum calcium, phosphate, albumin, urea, creatinine, fasting lipid profile (total cholesterol, LDL, HDL cholesterol, and triglycerides), and parathyroid hormone levels. Physical exams and chart reviews were performed by the participating centers.

Measurement of Carotid Intimal Plaque Area Carotid intimal plaque area measurements were scheduled every 6 mo. They were performed by a single trained technician at each center, using the same ultrasound equipment to minimize technique-related variation. The technique used has been published and validated by others, and is known to be predictive of cardiovascular events in non-CKD populations (6 – 8). Briefly, high-resolution duplex ultrasound was used to image the right and left common, internal, and external carotid arteries. Plaque was defined as a local intimal thickening of at least 1 mm. The plane of measurement chosen was that in which the largest extent of plaque was visualized in a given arterial segment. Total plaque area was defined as the sum of cross-sectional areas of all plaques seen between the clavicle and the angle of the jaw in both carotids.

Statistical Analysis Continuous variables were expressed as mean ⫾ SD, or, where indicated, median and interquartile range (IQR). Categorical variables were expressed as number and percentages. Bivariate comparisons were made using t tests, Kruskal-Wallace, and ␹2 tests as appropriate. Multivariate Modeling of Baseline Plaque. We first explored multivariate associations between biologically related subsets of variables and baseline plaque. A final overall model was created by including all variables significant at P ⱕ 0.2 on bivariate testing and applying backward selection to remove nonsignificant variables. As a final step, variables known to be biologically associated with atherosclerosis were reintroduced into the model to make sure that no significant associations were missed. Modeling of Plaque Progression in Patients with Plaque at Baseline. Plaque area was heavily skewed. A cube root transformation was used to curb departures from normality, as done by Bennett et al. (8). The rate of change of plaque area was determined for each patient as the slope of the least squares regression line fitted to all transformed

Table 2. Bivariate comparisons of patients with (n ⫽ 242) and without (n ⫽ 76) detectable baseline plaque Variable

Age (years)

With plaque

62 (10)a

Without plaque

Variable

With plaque

Without plaque

P

⬍0.001

5.1 (1.2)

5.1 (1.2)

0.8

3.1 (1.1)

3.0 (1.0)

0.8

38 (4.4)

39 (4.2)

0.01

48

43

0.6

43 41 24 31

26 28 18 22

0.01 0.04 0.35 0.2

Female gender (%)

34

37

0.7

Diabetes (%) Smoker (%)b

38 57

14 42

⬍0.001 0.03

History of cardiovascular disease (%)c Systolic BP (mmHg) Diastolic BP (mmHg) Pulse pressure (mmHg) Creatinine Clearance (ml/min/1.73 m2) Proteinuria (g/day)

29

4

⬍0.001

Total cholesterol (mmol/L) LDL cholesterol (mmol/L) Albumin (g/L) Treatment variables (%) ACEI or ARBc

143 (22) 79 (11) 64 (20) 33 (10)

130 (19) 78 (11) 51 (76) 32 (11)

⬍0.001 1.0 ⬍0.001 0.3

Diuretic Calcium blocker Beta blocker Statind

1.5 (2.6)

1.1 (1.6)

0.2

Hemoglobin (g/L) Total Calcium (mmol/L) Phosphate (mmol/L)

125 (17) 126 (17) 2.3 (0.14) 2.3 (0.14) 1.2 (0.3) 1.2 (0.2)

0.8 0.5 0.8

Ca ⫻PO4 (mmol2/L2) LN PTH 4

2.6 (0.5) 3.8 (1.3)

0.2 0.002

Fibric acid derivatives ASA Erythropoietin Calcium-based PO4 binderse Vitamin D3

a

46 (9)

P

2.9 (0.7) 3.2 (1.5)

5.8

6.6

24 8.7 21

8 9.2 34

0.002 0.8 0.03

23

0.001

8

0.8

Numbers in brackets are standard deviations. bCurrent or past smoker cDocumented prior myocardial infarction, ischemic stroke or peripheral arterial disease. dNatural logarithm transform. eCalcium carbonate or calcium acetate.

294


Table 3. Multivariate logistic models of baseline plaque Model

Base variable model (R2 ⫽ 0.34)a age (years) diabetes history of cardiovascular disease Blood pressure model (R2 ⫽ 0.16)b pulse pressure (mmHg) Renal model (R2 ⫽ 0.10)c proteinuria (g/day) Lipids and nutrition variables (R2 ⫽ 0.10)d Alb (g/L) Divalent ion metabolism model (R2 ⫽ 0.09)e PTH (natural log) Overall model (R2 ⫽ 0.37)f age history of cardiovascular disease pulse pressure Alb (g/L) PTH (natural log)

Odds ratio (95 % CI)g

1.13 (1.09 to 1.17) 3.37 (1.42 to 8.00) 3.05 (0.81 to 11.5)

1.04 (1.02 to 1.06) 1.18 (1.00 to 1.39)

0.92 (0.86 to 0.98)

1.09 (0.88 to 1.36 1.13 (1.09 to 1.17) 5.77 (1.45 to 23.0) 1.03 (1.005 to 1.05) 0.87 (0.79 to 0.95) 1.22 (0.92 to 1.63)

Entered variables: aBase model: age, gender, diabetes, history of cardiovascular disease, smoking history; bBlood pressure model: systolic and diastolic BP, pulse and mean arterial pressures, history of hypertension; cRenal model: Creatinine clearance, proteinuria, hemoglobin; dLipids and nutrition model: Total cholesterol, LDL, HDL, triglycerides, serum albumin; eDivalent ion metabolism model: PTH, Calcium, Phosphate, calcium ⫻ phosphate product; fOverall model: All variables achieving significance at P ⬍ 0.2 level on bivariate analyses inTable 2; gAll coefficients are adjusted for treatment center. plaque area estimations for that patient. In the 10% of patients with only two ultrasounds, the simple slope was used instead. This slope was used as the dependent variable in linear regression analyses. CCr was analyzed in quintiles because the relationship with rate of progression was not linear. The following models were created: model 1: plaque area progression as a function of CCr quintile; model 2: Model 1 plus adjustment for baseline plaque and study center; Model 3: model 2 plus adjustment for any variable noted to be significant at P ⱕ 0.2 on the bivariate comparisons, any variables known to be associated with atherosclerosis in non-CKD populations, as well as adjustment for number of plaque measurements and follow-up duration. Finally, a sensitivity analysis replacing CCr with MDRD GFR quintiles was performed.


presented in Table 1. Median follow-up was 13 (IQR ⫽ 13) months. None of the participants died during the observation period, which in some individuals extended to 48 mo, yielding a mortality rate lower than reported for other stages 3 and 4 CKD patients (see Discussion). Median number of plaque estimates was three (IQR ⫽ 2) per patient. The distribution of carotid ultrasounds was: 34 patients (10%) with two carotid ultrasounds, 163 patients (52%) with three ultrasounds, 79 (25%) with four to five ultrasounds, and 42 (13%) with six to nine ultrasounds. Study patients were younger (mean age, 58 to 60 yr) than in other published CKD cohorts (mean age, 65 to 67 yr) (9 –10), but were similar in terms of gender distribution, cause of renal disease, and comorbidities. Statistically significant differences between quintiles were noted for hemoglobin, serum calcium, and serum phosphate, as expected. A lower proportion of patients with coronary artery disease in the first and fifth quintiles was apparent. Statistical trends toward a higher number of plaque estimates in patients with higher quintile of CrCl, toward longer follow-up times in quintile 4, and toward higher median plaque area in quintile 3 were noted.

Determinants of baseline plaque: Seventy-six (76) of 318 patients had no detectable plaque at baseline, whereas 242 had measurable plaque. Median plaque area overall was 23 (IQR ⫽ 1 to 84) mm2. In patients with plaque at baseline, it was 41 (IQR ⫽ 13 to 112) mm2. Older age, diabetes, history of cardiovascular disease (documented prior myocardial infarction, ischemic stroke, or peripheral arterial disease), history of smoking, high BP (systolic, pulse and mean arterial), and elevated PTH were all associated with plaque in bivariate comparisons (Table 2). Of the demographic and clinical variables, age, diabetes, and history of cardiovascular disease were significantly associated with presence of plaque, and remained so in the final model (Table 3). Of the BP variables, elevated pulse pressure was most closely associated with plaque and was significant in the final model. Lower serum albumin, higher serum creatinine, and elevated 24-h urine protein excretion were all associated with plaque within the renal function group, but only low serum albumin remained statistically significant in the final model. None of the measures of lipid status was associated with plaque. Within the divalent ion metabolism group, PTH level (log transformed) demonstrated the strongest associations.

Predictors of Developing De Novo Plaque Only 15 (19%) of the 76 patients without plaque eventually developed detectable de novo plaque, whereas the remaining 61 remained plaque free for the duration of follow-up. In bivariate comparisons, older age (51 versus 44 yr, P ⫽ 0.03) and lower serum calcium levels (2.2 versus 2.3 mmol/L, P ⫽ 0.03) were associated with de novo plaque at 1 yr. The small numbers did not permit multivariate modeling.

Results

Determinants of Plaque Progression in Patients with Plaque at Baseline

We enrolled 318 prevalent CKD patients in five Canadian centers over the period of the study. Baseline characteristics are

Plaque area did increase over time. Median rate of untransformed plaque area change was ⫹ 3.0 (IQR ⫽ 0, 15) mm2 per


year. This corresponds to a median plaque area increment in the cohort as a whole of 7%/yr, and a median plaque doubling time of 10 yr. In our primary analysis, plaque growth was statistically significantly slower in the lowest quintile of CCr (Table 4). Age, female gender, pulse pressure, calcium-based phosphate binder use, and PTH were also associated with rate of plaque area change in bivariate analyses. Center and baseline plaque were very strongly associated with change in plaque area and were included in all multivariate analyses. Follow-up duration and number of plaque area measurements were not significantly associated with rate of plaque progression. The relationship between CCr quintile and plaque progression remained statistically significant after adjustment for baseline plaque and center, inclusion of variables significant on bivariate testing, forced inclusion of factors known to influence atherosclerosis progression (age, sex, smoking status, diabetic status, BP, and lipid levels), duration of follow-up, and number of plaque estimates (Table 5). In a sensitivity analysis, we excluded center 5 (n ⫽ 107), which had significantly higher plaque progression rates. The ␤ coefficients of the model remained similar to the main analysis (table not shown). In a second sensitivity analysis, there was no association between rate of plaque growth and quintiles of MDRD GFR (11) (Table 6).

Discussion Our present study suggests that plaque growth is not accelerated at lower levels of renal function. Neither our primary


analysis, based on CCrs quintiles, nor a sensitivity analysis, based on MDRD GFR, demonstrated faster rates of plaque growth at lower levels of kidney function. Our primary analysis suggested that plaque growth progressed more slowly, not faster, in patients at the lowest quintile of CCr. In multivariate models, we were unable to attribute these results to confounding by other risk variables, to center effects, or to the modest differences observed in number of plaque measurements and follow-up duration between CCr quintiles. However, In a sensitivity analysis, a drop in rate of plaque growth in the lowest quintile of MDRD GFR was not observed. The reason for this discrepancy is not clear. The major pitfall of 24-h urine collections in routine use is inadequate collection leading to underestimated renal function. This risk was likely reduced but not eliminated by our study protocol. MDRD eGFR, although not dependent on urine collection, is sensitive to the method of creatinine measurement, which was not standardized against a reference lab and may have differed from center to center. We do not have data to permit post hoc adjustment of creatinine to a reference standard. It is possible that variations in creatinine assays may have increased variability and introduced misclassification bias of the MDRD GFR, diminishing observed differences between renal function quintiles. This variation would have less influence on the accuracy of 24-h urines, because urine and serum creatinine were measured with the same technique at each center, and would cancel out in the clearance calculation. Because of the differences

Table 4. Significant bivariate associations between exposure variables and rate of plaque growth Variable

Creatinine clearance quintile Q1 (⬍23 ml/min/1.73 m2) Q2 (23 to 29 ml/min/1.73 m2) Q3 (29 to 35 ml/min/1.73 m2) Q4 (35 to 43 ml/min/1.73 m2) Q5 (⬎43 ml/min/1.73 m2) (reference) Age (per decade) Female gender Pulse pressure (per 10 mmHg) PTH (natural log pmol/L) Medications HMG-CoA reductase inhibitors calcium-based phosphate binders Baseline plaque area ( per mm2/3) Center 1 (reference) 2 3 4 5

295

␤ coefficient (95% CI)

⫺0.22 (⫺0.34 to ⫺0.10) ⫺0.10 (⫺0.22 to 0.02) ⫺0.10 (⫺0.22 to 0.02) ⫺0.07 (⫺0.19 to 0.05) 0.0 (reference) ⫺0.07 (⫺0.11 to ⫺0.03) ⫺0.10 (⫺0.18 to ⫺0.02) ⫺0.02 (⫺0.04 to ⫺0.001) ⫺0.03 (⫺0.06 to ⫺0.02) 0.08 (⫺0.003 to 0.17) 0.12 (0.02 to 0.21) ⫺0.07 (⫺0.09 to ⫺0.05) 0.0 0.02 (⫺0.16 to 0.19) ⫺0.07 (⫺0.16 to 0.02) ⫺0.01 (⫺0.21 to 0.19) 0.20 (0.11 to 0.3)

P

0.02 0.0001 0.1 0.1 0.2 0.001 0.02 0.03 0.03 0.06 0.02 ⬍0.0001 ⬍0.0001 0.9 0.1 0.9 ⬍0.0001

The following variables were not significant on bivariate testing and are not shown: Diabetes, smoking, history of cardiovascular disease, systolic and diastolic BP, albumin, proteinuria, hemoglobin, calcium, phosphate, lipids (Total, LDL, and HDL cholesterol, triglycerides), Medication use (ACEi or ARB, diuretic, calcium channel blockers, beta blockers, HMGCoA reductase inhibitors, fibrates, vitamin D, ASA, erythropoietin; number of plaque measurements, duration of follow-up

296



Table 5. Effect of multivariate adjustment on the association between baseline creatinine clearance and rate of plaque growth Model 1: Unadjusted (model P ⫽ 0.01, R2 ⫽ 0.065) Creatinine clearance quintile Unadjusted ␤ coefficient (95% CI) P ⫽ 0.01 ⫺0.22 (⫺0.34 to ⫺0.10) q1(⬍23 ml/min/1.73 m2) q2 (23 to 29 ml/min/1.73 m2) ⫺0.10 (⫺0.22 to 0.02) q3 (29 to 35 ml/min/1.73 m2) ⫺0.10 (⫺0.22 to 0.02) q4 (35 to 43 ml/min/1.73 m2) -0.07 (⫺0.19 to 0.05) q5 (⬎43 ml/min/1.73 m2) (ref.) 0.0 Model 2: Adjusted for baseline plaque and center (model P ⬍ 0.0001, R2 ⫽ 0.255) Creatinine Clearance Quintile Unadjusted ␤ Coefficient (95% CI) P ⫽ 0.01 ⫺0.19 (⫺0.29,⫺0.08) P ⫽ 0.03 q1(⬍23 ml/min/1.73 m2) q2 (23 to 29 ml/min/1.73 m2) ⫺0.09 (⫺0.19 to 0.02) q3 (29 to 35 ml/min/1.73 m2) ⫺0.07 (⫺0.18 to 0.03) q4 (35 to 43 ml/min/1.73 m2) ⫺0.06 (⫺0.16 to 0.04) q5 (⬎43 ml/min/1.73 m2) (ref.) 0.0 Model 3: Adjusted for center, baseline plaque, duration of follow-up, number of ultrasounds, age, gender, smoking status, diabetes, blood pressure, LDL cholesterol, HMG-CoA reductase inhibitors, PTH, calcium binder use (model P ⬍ 0.0001, R2 ⫽ 0.296) Creatinine Clearance Quintile Adjusted ␤ Coefficients P ⫽ 0.02 ⫺0.189 (⫺0.30, ⫺0.08) q1(⬍23 ml/min/1.73 m2) q2 (23 to 29 ml/min/1.73 m2) ⫺0.10 (⫺0.21, 0.01) q3 (29 to 35 ml/min/1.73 m2) ⫺0.07 (⫺0.18, 0.03) q4 (35 to 43 ml/min/1.73 m2) ⫺0.06 (⫺0.17, 0.04) q5 (⬎43 ml/min/1.73 m2) (ref.) 0.0

between the primary (CCr) and secondary (MDRD GFR) analyses, we cannot be certain that plaque growth truly slows down at low levels of renal function. Importantly, however, neither analysis supported our original hypothesis that plaque area grows faster at lower levels of renal function. Our findings are consistent with two other studies that measured plaque volume with intravascular ultrasound (IVUS). A meta-analysis by Nicholls et al. analyzed 989 patients participating in three cardiovascular studies monitoring plaque progression by IVUS (12). The investigators did not find any differences in plaque volume or plaque growth in patients with GFR ⬎ 60 ml/min versus those with GFR ⱕ 60 ml/min. A study by Gruberg et al. also failed to find a significant association between plaque volume on IVUS and decreased GFR, although they did find higher plaque burden in the dialysis subgroup (13). Our study reinforces these observations, extending them

to a prevalent CKD cohort. These data together suggest that the accelerated cardiovascular event rate in CKD is not due to plaque growth per se. Our findings (and those of Nicholls and Gruberg) are in apparent conflict with some (14 –17), but not all (18 –20), studies assessing the relationship between kidney function and carotid intimal medial thickness (IMT). Interstudy differences in populations, distribution of GFR, and medications could account for some of this variation. Cross-sectional analyses (15–20), moreover, do not measure progression. The finding that IMT is higher at reduced GFR does not necessarily imply that IMT increases faster at lower GFR. Such an association could result from confounding by shared causal mechanisms, such as hypertension or diabetes. Most importantly, the technique of IMT lacks specificity for plaque. IMT measures arterial wall thickness from the luminal

Table 6. Association between four variable MDRD GFR quintiles and rate of plaque growth: Variable

␤ Coefficient

MDRD GFR quintile q1 (⬍18 ml/min/1.73 m2) q2 (18 to 23 ml/min/1.73 m2) q3 (⬎23 to 28 ml/min/1.73 m2) q4 (⬎28 to 34 ml/min/1.73 m2) q5 (⬎34 ml/min/1.73 m2) (ref.) Constant

0.03 (⫺0.17 to 0.22) ⫺0.11 (⫺0.31 to 0.08) ⫺0.02 (⫺0.22 to 0.17) 0.08 (⫺0.11 to 0.28) 0.0 0.24 (0.11 to 0.38)

P

0.4

0.001


margin to the adventitia (i.e., intima plus media). This measure cannot distinguish thickening caused by intimal plaque (atherosclerosis) from diffuse medial thickening related to arteriosclerosis, a point acknowledged by Preston (16). The increased IMT observed in some studies (14 –20) could reflect progressive arteriosclerosis rather than plaque growth in CKD, thus accounting for some of the differences between studies measuring IMT and those, including ours, measuring plaque area or volume (12–13). Moreover, because arteriosclerosis is known to be associated with left ventricular hypertrophy, decreased coronary perfusion, and downstream clinical events in ESRD patients and the general population (21–22), the association between IMT and clinical events in ESRD patients (23) could in part reflect the link between arteriosclerosis and cardiovascular outcome, rather than plaque growth and outcome. If plaque growth is not accelerated in CKD, then other mechanisms must explain the excess cardiovascular disease. Arteriosclerosis, as distinct from atherosclerosis, may be an important factor, as discussed above. It is possible too, as suggested by Nicholls et al. (12), that differences in plaque composition (e.g., calcification), endothelial dysfunction, and inflammation between CKD and non-CKD patients could alter plaque vulnerability and downstream clinical events independently of plaque growth. Our study has several important limitations. We did not study patients with normal or near-normal GFR, and thus cannot comment about plaque progression in these groups. We studied a referred cohort that experienced better clinical outcomes than reported in other studies, and it may not be possible to generalize our findings to population-based CKD cohorts or cohorts with higher event rates. We observed considerable between-center variation in plaque area, and despite adjustment for center, residual confounding may still be present. We measured only intimal disease with carotid intimal plaque area, and did not measure indices of other vascular changes. Further prospective studies of vascular remodeling in incident patient cohorts at all stages of CKD, using multiple metrics of vascular remodeling (e.g., plaque area or volume, IMT, compliance, calcification), are needed to better understand the evolution of vascular changes occurring in CKD.

Conclusion We did not find evidence that carotid plaque grew faster in CKD patients with severe versus moderate renal impairment. Our results are congruent with those of studies measuring rate of change of plaque volume. There was a suggestion that patients with the lowest quintile CCr had slower rates of progression than patients with higher CCrs. Further prospective studies of vascular remodeling in incident CKD cohorts are needed to understand the precise nature and clinical consequences of vascular changes in renal disease.

Acknowledgments The present study was supported by operating grants from the Kidney Foundation of Canada (Drs. Rigatto and Fine) and by an unrestricted educational grant from Baxter (Dr. Fine).


297

Dr. Rigatto received salary support as Biomedical Research Scholar of the Kidney Foundation of Canada from 2002 to 2004

Disclosures None.

References 1. Foley N, Parfrey P, Sarnack M: Clinical epidemiology of cardiovascular disease in chronic renal disease. Am J Kidney Dis 32: S112–S119, 1998 2. Sarnak MJ, Levey AS, Schoolwerth AC, Coresh J, Culleton B, Hamm LL, McCullough PA, Kasiske BL, Kelepouris E, Klag MJ, Parfrey P, Pfeffer M, Raij L, Spinosa DJ, Wilson PW; American Heart Association Councils on Kidney in Cardiovascular Disease, High Blood Pressure Research, Clinical Cardiology, and Epidemiology and Prevention: Kidney disease as a risk factor for development of cardiovascular disease: A statement from the American Heart Association Councils on Kidney in Cardiovascular Disease, High Blood Pressure Research, Clinical Cardiology, and Epidemiology and Prevention. Circulation 108: 2154 –2169, 2003 3. Wanner C, Krane V, Ma¨rz W, Olschewski M, Mann JFE, Ruf G, Ritz E, for the German Diabetes and Dialysis Study Investigators (4D): Atorvastatin in patients with type 2 diabetes mellitus undergoing hemodialysis. N Engl J Med 353: 238 –248, 2005 4. Holdaas H, Fellstrom B, Jardine AG, Holme I, Nyberg G, Fauchald P, Gronhagen-Riska C, Madsen S, Neumayer HH, Cole E, Maes B, Ambuhl P, Olsson AG, Hartmann A, Solbu DO, Pedersen TR; Assessment of Lescol in Renal Transplantation (ALERT) Study Investigators: Effect of fluvastatin on cardiac outcomes in renal transplant recipients: a multicentre, randomised, placebo-controlled trial. Lancet 361: 2024 –2231, 2003 5. Lindner A, Charra B, Sherrard DJ, Scribner BH: Accelerated atherosclerosis in prolonged maintenance hemodialysis. N Engl J Med 290: 697–701, 1974 6. Spence JD, Eliasziw M, DiCicco M, Hackam DG, Galir R, Lohmann T: Carotid plaque area: A tool for targeting and evaluating vascular preventive therapy. Stroke 33: 2916 – 2922, 2002 7. Spence JD, Hegele RA: Noninvasive phenotypes of atherosclerosis: Similar windows, but different views. Stroke 35: 649 – 653, 2004 8. Barnett PA, Spence JD, Manuck SB, Jennings JR: Psychological stress and the progression of carotid atherosclerosis. J Hypertens 15: 49 –55, 1997 9. Levin A, Djurdjev O, Beaulieu M, Er L: Variability and risk factors for kidney disease progression and death following attainment of stage 4 CKD in a referred cohort. Am J Kidney Dis 52: 661– 671, 2008 10. Parikh NI, Hwang SJ, Larson MG, Levy D, Fox CS: Chronic kidney disease as a predictor of cardiovascular disease (from the Framingham Heart Study). Am J Cardiol 102: 47–53, 2008 11. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D: A more accurate method to estimate glomerular filtration rate from serum creatinine: A new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 130: 461– 470, 1999

298


12. Nicholls SJ, Tuzcu EM, Hsu A, Wolski K, Sipahi I, Schoenhagen P, Crowe T,. Kapadia SR,. Hazen SL, Nissen SE: Comparison of coronary atherosclerotic volume in patients with glomerular filtration rates ⬍60 versus ⬎60 ml/min/ 1.73 m2: A meta-analysis of intravascular ultrasound studies. Am J Cardiol 99: 813– 816, 2007 13. Gruberg L, Rai P, Mintz GS, Canos D, Pinnow E, Satler LF, Pichard AD, Kent KM, Waksman R, Lindsay J, Weissman NJ: Impact of renal function on coronary plaque morphology and morphometry in patients with chronic renal insufficiency as determined by intravascular ultrasound volumetric analysis. Am J Cardiol 96: 892– 896, 2005 14. Fathi R, Isbel N, Short L, Haluska B, Johnson D, Marwick TH: The effect of long-term aggressive lipid lowering on ischemic and atherosclerotic burden in patients with chronic kidney disease. Am J Kidney Dis 43: 45–52, 2004 15. Kawagishi T, Nishizawa Y, Konishi T: High resolution B-mode ultrasonography in evaluation of atherosclerosis in uremia. Kidney Int 48: 820 – 826, 1995 16. Preston E, Ellis MR, Kulinskaya E, Davies A, Brown E: Association between carotid artery intima-media thickness and cardiovascular risk factors in CKD. Am J Kidney Dis 46: 856 – 862, 2005 17. Shoji T, Emoto M, Tabata T: Advanced atherosclerosis in predialysis patients with chronic renal failure. Kidney Int 61: 2187–2192, 2002 18. Briet M, Bozec E, Laurent S, Fassot C, London GM, Jacquot


19.

20.

21.

22.

23.

C, Froissart M, Houillier P, Boutouyrie P: Arterial stiffness and enlargement in mild-to-moderate chronic kidney disease. Kidney Int 69(2): 350 –357, 2006 Leskinen Y, Lehtimaki T, Loimaala A, Lautamatti V, Kallio T, Huhtala H, Salenius JP, Saha H: Carotid atherosclerosis in chronic renal failure-the central role of increased plaque burden. Atherosclerosis 71: 295–302, 2003 Nanayakkara PW, Teerlink T, Stehouwer CD, Allajar D, Spijkerman A, Schalkwijk C, ter Wee PM, van Guldener C: Plasma asymmetric dimethylarginine (ADMA) concentration is independently associated with carotid intima-media thickness and plasma soluble vascular cell adhesion molecule-1 (sVCAM-1) concentration in patients with mild-tomoderate renal failure. Kidney Int 68: 2230 –2236, 2005 Pannier B, Guerin AP, Marchais SJ, Safar ME, London GM: Stiffness of capacitive and conduit arteries: Prognostic significance for end-stage renal disease patients. Hypertension 45(4): 592–596, 2005 Mattace-Raso FU, van der Cammen TJ, Hofman A, van Popele NM, Bos ML, Schalekamp MA, Asmar R, Reneman RS, Hoeks AP, Breteler MM, Witteman JC: Arterial stiffness and risk of coronary heart disease and stroke: The Rotterdam Study. Circulation 113: 657– 663, 2006 Kato A, Takita T, Maruyama Y, Kumagai H, Hishida A: Impact of atherosclerosis on long term mortality in chronic hemodialysis patients. Kidney Int 64: 1472–1479, 2003