Carotid Intima Media Thickness Predicts Cardiovascular Diseases in Chinese Predialysis Patients with Chronic Kidney Disease
Cheuk-Chun Szeto,
Kai-Ming Chow,
Kam-Sang Woo,
Ping Chook,
Bonnie Ching-Ha Kwan,
Chi-Bon Leung and
Philip Kam-Tao Li
Department of Medicine & Therapeutics, Prince of Wales Hospital, Chinese University of Hong Kong, Shatin, Hong Kong, China
Address correspondence to: Dr. Cheuk-Chun Szeto, Department of Medicine & Therapeutics, Prince of Wales Hospital, Chinese University of Hong Kong, Shatin, N.T., Hong Kong, China. Phone: +852-2632-3126; Fax: +852-2637-3852; E-mail: ccszeto{at}cuhk.edu.hk
Received for publication October 31, 2006.
Accepted for publication April 3, 2007.
Patients with chronic kidney disease (CKD) have a high riskfor cardiovascular disease. Ultrasound measurements of the intimamedia thickness (IMT) in the carotid arteries is a strong predictorfor cardiovascular events in the general population and dialysispatients. However, it is unclear whether carotid IMT is usefulfor the prediction of cardiovascular events in predialysis patientswith CKD. The prediction power of carotid ultrasonography forcardiovascular event, rate of renal function decline, and all-causemortality was tested in a cohort of 203 Chinese patients withstages 3 to 4 CKD. The average IMT was 0.808 ± 0.196mm; 121 (59.6%) patients had atherosclerotic plaques visualized.IMT correlated with patient age (r = 0.373, P < 0.001), serumLDL level (r = 0.164, P = 0.021), Charlson’s comorbidityscore (r = 0.260, P < 0.001), and serum C-reactive protein(r = 0.279, P < 0.001). Carotid IMT was significantly higherin patients with diabetes than in those without diabetes (0.930± 0.254 versus 0.794 ± 0.184; P = 0.002). At 48mo, the cardiovascular event-free survival was 94.4, 89.8, 77.7,and 65.9% for IMT quartiles I, II, III, and IV, respectively(log rank test, P = 0.006). By multivariate analysis with theCox proportional hazard model, each higher quartile of IMT conferred41.6% (95% confidence interval 6.4 to 88.4%; P = 0.017) excesshazard for developing cardiovascular event. The actuarial survivalat 48 mo was 96.3, 98.0, 95.7, and 85.7% for IMT quartiles I,II, III and IV, respectively (log rank test, P = 0.127), andthe difference was not statistically significant after Cox proportionalhazard model to adjust for confounders. Carotid IMT did notcorrelate with the rate of renal function decline in these patients.Carotid IMT is a strong predictor of cardiovascular diseasein Chinese predialysis patients and may be usefully appliedfor risk stratification in this group of patients.
Patients with chronic kidney disease (CKD) are at high riskfor developing cardiovascular disease (CVD) (1,2). CVD sharesmany similar risk factors with CKD, such as diabetes and hypertension(3). Nonetheless, after accounting for traditional risk factorsbased on the Framingham Heart Study, CKD remains an independentrisk factor for CVD (4). Longitudinal studies have establishedthat cardiovascular events occur more frequently than renalevents in CKD, and mortality rates are in fact higher than therates of reaching ESRD (5). However, there are few publisheddata on the incidence of asymptomatic atherosclerosis and possiblerisk factors for atherosclerosis in predialysis patients withCKD.
It is generally believed that the atherosclerotic changes inthe carotid artery mirror general atherosclerosis (6). High-resolutionB-mode ultrasound scan, a valid noninvasive method for assessmentof asymptomatic atherosclerosis, has been widely used to studycarotid atherosclerosis in the general population (7). Ultrasoundmeasurements of the intima media thickness (IMT) in the carotidarteries were used as an indicator of coronary atherosclerosis(6). There is, in fact, a close relationship between carotidartery wall morphology and CVD (6,8,9). IMT and plaque occurrencein the carotid arteries are strong predictors for cardiovascularevents in the general population (10). In several recent studies,ultrasound measurements of IMT in carotid arteries were alsoused as the indicator of coronary atherosclerosis in dialysispatients. For example, Benedetto et al. (11) found that carotidartery IMT represented an independent predictor of cardiovasculardeath in dialysis patients. Nishizawa et al. (12), Kato et al.(13), and Ekart et al. (14) all reported that carotid arteryIMT was an independent predictor of cardiovascular mortalityin hemodialysis patients. However, the prognostic value of carotidIMT has not been evaluated in predialysis patients. The aimof this study was to determine whether IMT predicts the cardiovascularmortality in predialysis patients with CKD and whether carotidDoppler ultrasound could be applied for the stratification ofcardiovascular risk in these patients.
Patient Selection
The study was approved by the Clinical Research Ethical Committeeof the Chinese University of Hong Kong, and informed consentwas obtained from each participant. Recruitment criteria were(1) plasma creatinine >150 µmol/L or estimated GFR(eGFR) <60 ml/min per 1.73 m2 by a standard equation (15)and (2) age 18 to 70 yr. Patients who were on long-term dialysis,had a history of kidney transplantation, or had had recent myocardialinfarction or coronary bypass surgery (within 6 mo) or stroke(within 1 yr) were excluded. In total, we recruited into thestudy 203 predialysis patients who had CKD and attended therenal clinic of our hospital, which is a tertiary referral centerof the region.
For each recruited patient, background clinical information;underlying renal diagnosis; and the presence of diabetes, hypertension,or a history of cardiovascular disease were recorded. Hypertensionwas defined as systolic BP >130 mmHg or a diastolic BP >80mmHg or requiring antihypertensive therapy. The definition ofcardiovascular disease included angina, class III to IV congestiveheart failure, a history of myocardial infarction, cerebrovascularaccident, or amputation for vascular disease. The modified Charlson’sComorbidity Index was used to calculate a comorbidity score.
Laboratory Variables and Information on Smoking
Serum creatinine concentration was determined by clinical analyzersusing the Jaffe kinetic method (Hitachi 911 or 747; BoehringerMannheim, Mannheim, Germany). Our previous study on peritonealdialysis patients showed that the interassay coefficient ofvariation of these methods was <5% (16). Although patientswere recruited according to the eGFR as determined by the originalModification of Diet in Renal Disease (MDRD) equation (15),eGFR is calculated and presented in this report by the modifiedMDRD equation validated in Chinese patients (17). Proteinuriawas measured by a turbidimetric technique with the Modular Analytics(Roche Diagnostics, Rotkreuz, Switzerland). Serum cholesterol(total, LDL, and HDL cholesterol), triglycerides, C-reactiveprotein (CRP), calcium, and phosphate were measured by routinelaboratory methods. Blood was withdrawn after at least 12 hof fasting. Serum CRP was measured by our in-house sensitiveELISA. Information on smoking habit was obtained by direct questioning.
Ultrasound Imaging and Analysis
The method has been described previously (18). Briefly, B-modeultrasound examinations were performed with an Acuson 128XP/10mainframe with a 7-MHz scanning frequency linear array transducer;an ATL 3000 mainframe with a high-resolution, linear array scanner(medium frequency 7.5 MHz); or an Interspec Apogee CX200 mainframewith a 7.5-MHz transducer (all from Advanced Technology Laboratories,Bothell, WA. All ultrasound systems therefore used similar scanningfrequency and had similar resolution (approximately 0.12-mmtheoretical resolution in each case). All scans were performedby two operators after a predetermined, standardized scanningprotocol for the right and left carotid arteries, as describedby Blankenhorn et al. (19), using images of the far wall ofthe distal 10 mm of the common carotid arteries. Three scanningangles were used in each case: Anterior oblique, lateral, andposterior oblique. The image was focused on the posterior wall,and images were recorded from the angle that showed the greatestdistance between the lumen-intima interface and the media-adventitiainterface as described previously (18). All scans were recordedon super-VHS videotape for subsequent off-line analysis.
All scans were then analyzed with a computerized edge-detectionsystem that was previously described and validated (20). Observerswere blinded to the patient’s identity and demographicfeatures. Two end-diastolic frames were selected, digitized,and analyzed for mean IMT, and the average reading from thesetwo frames was calculated for both right and left carotid arteries.Images were digitized with the use of a frame grabber (VideoAssociates Labs, Austin, TX) and an IBM-compatible computerinterfaced with a Panasonic AG7350 super-VHS videocassette recorder(Berkshire, UK). Edge-detection software automatically identifiesintimal and medial points from the region of interest of thefar wall of the common carotid artery as defined by the observer.As reported in our previous study (18), the mean differenceof repeated measurements between the two operators of our study(K.S.W. and P.C.) was 0.02 ± 0.04 mm, and the coefficientof variation for mean IMT measurements was 3.0%.
Clinical Follow-Up
All patients were followed for at least 4 yr. The clinical managementwas decided by individual clinician and not affected by thestudy. Primary end point was a composite one that consistedof cardiovascular death, nonfatal myocardial infarction or stroke,hospital admission for unstable angina, coronary intervention,congestive heart failure, transient ischemic attack, cerebrovascularaccident, or peripheral vascular disease that required surgicalreconstruction or amputation. Secondary end points includedrate of GFR decline, ESRD, and all-cause mortality. ESRD wasdefined as the need for long-term dialysis, preemptive kidneytransplantation, or death as a result of uremia.
Statistical Analyses
Statistical analysis was performed with SPSS for Windows software(version 10.0; SPSS, Chicago, IL). Data are expressed as means± SD unless otherwise specified. Data were compared byt test, 2 test, or Pearson correlation coefficient as appropriate.The relationship between IMT and clinical outcomes was furthertested by stratification of patients into quartiles accordingto the IMT values: Quartile I, <0.70 mm; quartile II, 0.70to <0.80 mm; quartile III, 0.80 to <0.90 mm; and quartileIV, 0.90 mm. Survival rates were analyzed using Kaplan-Meiersurvival curves. The Cox proportional hazards model was usedto identify independent predictors of primary cardiovascularend point and actuarial survival (21). IMT quartile; age; gender;serum LDL; Charlson’s comorbidity score; diabetic status;systolic and diastolic BP; cigarette smoking; hemoglobin; baselineGFR; proteinuria; and serum phosphate, albumin, and CRP levelswere taken as independent variables for modeling. Backward stepwiseelimination was applied to remove insignificant variables. P< 0.05 was considered statistically significant. All probabilitieswere two-tailed.
We studied 203 predialysis patients with CKD. The demographicand baseline clinical information is summarized in Table 1;baseline serum biochemistry is summarized in Table 2. Therewere 28 (13.8%) active smokers. Of the 203 patients, 123 (60.6%)received angiotensin-converting enzyme inhibitor or angiotensinreceptor blocker therapy, and 18 (8.9%) received aspirin.
IMT in Chinese Patients with CKD
The average IMT of the study population was 0.808 ± 0.196mm; 121 (59.6%) patients had atherosclerotic plaques visualizedby ultrasound in their carotid arteries. Univariate analysisshowed that IMT correlated with patient age (r = 0.373, P <0.001), serum LDL level (r = 0.164, P = 0.021), Charlson’scomorbid score (r = 0.260, P < 0.001), and serum CRP (r =0.279, P < 0.001; Figure 1). Carotid IMT was significantlyhigher in patients with diabetes than without diabetes (0.930± 0.254 versus 0.794 ± 0.184; P = 0.002) but lowerin patients who were receiving angiotensin-converting enzymeinhibitor therapy than the others (0.757 ± 0.157 versus0.842 ± 0.211 mm; P = 0.003). However, IMT was not relatedto gender, body mass index, BP, cigarette smoking, renal function,proteinuria, or calcium-phosphate product (data not shown).Eleven of the 12 patients who had a history of cerebrovasculardisease had atherosclerotic plaques visualized by ultrasoundin their carotid arteries.
Figure 1. Relation between carotid intima media thickness (IMT) and patient age (A), serum LDL level (B), Charlson’s comorbidity score (C), and serum C-reactive protein (CRP) level (D). Data are compared by Pearson correlation coefficient. Dashed lines represent 95% confidence interval of the trend line.
IMT and Primary Cardiovascular End Point
The average follow up was 52.4 ± 11.5 mo. Forty-six (22.7%)patients developed the primary composite end point. The eventsthat contributed to the primary composite end point were myocardialinfarction (three cases), stroke (one case), nonfatal myocardialinfarction (three cases), nonfatal stroke (10 cases), hospitalizationfor congestive heart failure (12 cases), hospitalization foracute coronary syndrome (16 cases), and amputation for peripheralvascular disease (one case). At 48 mo, the event-free survivalwas 94.4, 89.8, 77.7, and 65.9% for IMT quartiles I, II, III,and IV, respectively (log rank test, P = 0.006; Figure 2).
Figure 2. Kaplan-Meier plot of cardiovascular event-free survival (A) and actuarial survival (B). Patients were divided to quartiles of carotid IMT: Quartile I, <0.70 mm; quartile II, 0.70 to <0.80 mm; quartile III, 0.80 to <0.90 mm; and quartile IV, 0.90 mm.
By multivariate analysis with the Cox proportional hazard modelto adjust for confounders, the independent factors for event-freesurvival were IMT quartile, baseline GFR, and Charlson’scomorbidity score. The result of the Cox model analysis is summarizedin Table 3. In this model, each higher quartile of IMT conferred40.2% (95% confidence interval 2.9 to 91.0%; P = 0.032) excesshazard for developing the primary cardiovascular end point.
IMT and Renal Outcome
The average rate of GFR decline in our patients was –0.373± 0.703 ml/min per 1.73 m2 per month. During the studyperiod, 58 (28.6%) patients progressed to ESRD. They were treatedby peritoneal dialysis (42 cases), by hemodialysis (five cases),by preemptive kidney transplantation (nine cases), or conservatively(two cases). There was no difference in the rate of GFR declineor ESRD-free survival among IMT quartiles (Table 4).
IMT and Overall Survival
During the follow-up period, 17 (8.4%) patients died. The causesof death were myocardial infarction (four cases), cerebrovascularaccident (three cases), infection (five cases), malignancy (threecases), and ESRD (two cases). At 48 mo, the actuarial survivalwas 96.3, 98.0, 95.7, and 85.7% for IMT quartiles I, II, III,and IV, respectively (log rank test, P = 0.127; Figure 2). Bymultivariate analysis with the Cox proportional hazard modelto adjust for confounders, the only independent factors foractuarial survival were patient age, gender, baseline GFR, serumCRP, and Charlson’s comorbidity score. The result of theCox model analysis is summarized in Table 3. In this model,carotid IMT was not an independent predictor of actuarial survival.
In this study, we found that carotid IMT is a strong predictorof cardiovascular disease in this group of patients. Althoughit is well reported that premature atherosclerosis in commonin patients with renal failure (1–5) and carotid IMT isa strong predictor of cardiovascular event in patients who areon long-term dialysis (11–14), our results indicate thatIMT by carotid Doppler ultrasound is a valuable tool for thestratification of cardiovascular risk in predialysis patientswith CKD.
It is important to note that the patients in our study wererecruited according to the eGFR as determined by the originalMDRD equation (15); eGFR is calculated and presented in thisreport by the modified MDRD equation as validated in Chinesepatients (17). We originally intended to study stages 3 and4 CKD, but 24 (11.8%) of the recruited patients actually hadstage 2 CKD according to the modified MDRD equation. However,the result of multivariate analysis and conclusion would notbe much affected if we used the original MDRD equation for calculation(data not shown).
The absolute values of carotid IMT in our patients were smallerthan those reported by Benedetto et al. (11) and Nishizawa etal. (12) but similar to Kato et al. (13) and Ekart et al. (14).It is important to note that these studies examined patientswho were on long-term dialysis (11–14), whereas our patientshad only stages 3 to 4 CKD. Taken together, our results indicatethat increase in carotid IMT occurs early in the course of CKD.Although it is commonly believed that atherosclerosis is lesssevere in Asian patients, neither previous studies in Japanesepatients (12,13) nor our study showed a lower IMT as comparedwith a white population (11,14).
Similar to the report of Kato et al. (13), we found that carotidIMT correlates with patient age and serum CRP level. In thisstudy, carotid IMT also correlates with diabetic status, serumLDL level, and Charlson’s comorbidity score, which arenot unexpected findings. It is important to note that thesefactors in total account for only 36% of the overall variabilityof carotid IMT. We did not observe any relation between carotidIMT and other traditional risk factor of atherosclerosis, suchas BP, cigarette smoking, or serum calcium-phosphate product.It is possible that there was unidentified selection bias inour study population, although previous studies did not finda relation between carotid IMT and BP or serum calcium-phosphateproduct in long-term hemodialysis patients (13,14). Becauseof the limitations of budget and study design, we did not examinethe role of other novel risk factors of premature atherosclerosis,such as serum homocysteine and lipoprotein(a).
More than 20% of our patients developed a primary end pointduring the study period, whereas 3% of the patients had fatalor nonfatal myocardial infarction. The incidence of cardiovascularevent in our study is similar to that reported by other groups(2,22). In a study by Levin (2) on predialysis CKD, a changein cardiac status (worsening of heart failure or anginal symptoms)occurred in 20% of patients (2). Wattanakit et al. (22) reportedthat the incidence of myocardial infarction was 0.8% per yearin patients with stage 3 CKD. It is important to note that increasein carotid IMT probably represents generalized, including coronary,atherosclerosis. In fact, the absolute risk for coronary heartdisease is actually higher than that of cerebrovascular diseasein patients with greatly elevated carotid IMT. As expected,the absolute mortality of our study population was substantiallylower than that in previous studies on dialysis patients (11–14).The relatively small number of primary event restricted thepossible scope of multivariate analysis as well as the strengthof our conclusion. Our study is, in fact, underpowered to detecta difference in actuarial survival, although we did find a similartrend of better actuarial survival in patients with lower IMT.Post hoc estimation of sample size (23) shows that 600 patientsare required to provide 80% statistical power for the survivalanalysis.
The high incidence of cardiovascular disease in our study populationmay partly be explained by the less than optimal control oftraditional risk factors, such as BP and serum cholesterol level(see Tables 1 and 2). The suboptimal BP control probably alsoexplains the higher rate of renal function decline in our patients(approximately 3.6 ml/min per yr) as compared with patientswith ideal BP and diabetes control (approximately 2 ml/min peryr) (24). Although we did not observe any effect of BP controlon the risk for cardiovascular event or rate of renal functiondecline, our study does not have adequate statistical powerto, neither did we aim to, detect such an effect. The latestguidelines state that BP and serum total cholesterol shouldbe below 130/80 mmHg and 5.2 mmol/L, respectively (25–27).Neither of the targets was completely achieved in our patients.Nonetheless, it is well reported that optimal BP control inlarge cohorts of patients with renal failure is difficult inreal-life practice (28,29).
Because 46 patients developed a primary composite end point,theoretically at most five independent variables could be includedin the Cox model without overfitting. In other words, our studydoes not have enough statistical power to analyze all 15 variablesthat we put in the Cox model, and it is possible that an importantpredictor of cardiovascular event or patient survival couldhave been missed by our study (i.e., type II statistical error).For the same reason, we did not include in the Cox model othervariables that were used to derive the Charlson’s comorbidityindex but were also potential primary outcomes. It has beensuggested that a propensity score to combine all of the potentialconfounders could be used for the Cox model analysis (30), butthis method is not without bias (31).
Carotid IMT is a strong predictor of cardiovascular diseasein Chinese predialysis patients and may be usefully appliedfor risk stratification in this group of patients.
Weiner DE, Tabatabai S, Tighiouart H, Elsayed E, Bansal N, Griffith J, Salem DN, Levey AS, Sarnak MJ: Cardiovascular outcomes and all-cause mortality: Exploring the interaction between CKD and cardiovascular disease.
Am J Kidney Dis 48
: 392
–401, 2006[CrossRef][Medline]
Levin A: Clinical epidemiology of cardiovascular disease in chronic kidney disease prior to dialysis.
Semin Dial 16
: 101
–105, 2003[CrossRef][Medline]
Parikh NI, Hwang SJ, Larson MG, Meigs JB, Levy D, Fox CS: Cardiovascular disease risk factors in chronic kidney disease: Overall burden and rates of treatment and control.
Arch Intern Med 166
: 1884
–1891, 2006[Abstract/Free Full Text]
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: 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[Free Full Text]
Foley RN, Murray AM, Li S, Herzog CA, McBean AM, Eggers PW, Collins AJ: Chronic kidney disease and the risk for cardiovascular disease, renal replacement, and death in the United States Medicare population, 1998 to 1999.
J Am Soc Nephrol 16
: 489
–495, 2005[Abstract/Free Full Text]
Craven TE, Ryu JE, Espeland MA, Kahl FR, McKinney WM, Toole JF, McMahan MR, Thompson CJ, Heiss G, Crouse JR 3rd: Evaluation of the associations between carotid artery atherosclerosis and coronary artery stenosis. A case-control study.
Circulation 82
: 1230
–1242, 1990[Abstract/Free Full Text]
Salonen R, Seppanen K, Rauramaa R, Salonen JT: Prevalence of carotid atherosclerosis and serum cholesterol levels in eastern Finland.
Arteriosclerosis 8
: 788
–792, 1988[Abstract/Free Full Text]
Salonen JT, Salonen R: Ultrasonographically assessed carotid morphology and the risk of coronary heart disease.
Arterioscler Thromb 11
: 1245
–1249, 1991[Abstract/Free Full Text]
Fabris F, Zanocchi M, Bo M, Fonte G, Poli L, Bergoglio I, Ferrario E, Pernigotti L: Carotid plaque, aging, and risk factors. A study of 457 subjects.
Stroke 25
: 1133
–1140, 1994[Abstract]
Benedetto FA, Mallamaci F, Tripepi G, Zoccali C: Prognostic value of ultrasonographic measurement of carotid intima media thickness in dialysis patients.
J Am Soc Nephrol 12
: 2458
–2464, 2001[Abstract/Free Full Text]
Nishizawa Y, Shoji T, Maekawa K, Nagasue K, Okuno S, Kim M, Emoto M, Ishimura E, Nakatani T, Miki T, Inaba M: Intima-media thickness of carotid artery predicts cardiovascular mortality in hemodialysis patients.
Am J Kidney Dis 41[Suppl 1]
: S76
–S79, 2003
Kato A, Takita T, Maruyama Y, Kumagai H, Hishida A: Impact of carotid atherosclerosis on long-term mortality in chronic hemodialysis patients.
Kidney Int 64
: 1472
–1479, 2003[CrossRef][Medline]
Ekart R, Hojs R, Hojs-Fabjan T, Balon BP: Predictive value of carotid intima media thickness in hemodialysis patients.
Artif Organs 29
: 615
–619, 2005[CrossRef][Medline]
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.
Ann Intern Med 130
: 461
–470, 1999[Abstract/Free Full Text]
Mak TW, Cheung CK, Cheung CM, Leung CB, Lam CW, Lai KN: Interference of creatinine measurement in CAPD fluid is dependent on glucose and creatinine concentrations.
Nephrol Dial Transplant 12
: 184
–186, 1997[Abstract/Free Full Text]
Ma YC, Zuo L, Chen JH, Luo Q, Yu XQ, Li Y, Xu JS, Huang SM, Wang LN, Huang W, Wang M, Xu GB, Wang HY: Modified glomerular filtration rate estimating equation for Chinese patients with chronic kidney disease.
J Am Soc Nephrol 17
: 2937
–2944, 2006[Abstract/Free Full Text]
Woo KS, Chook P, Raitakari OT, McQuillan B, Feng JZ, Celermajer DS: Westernization of Chinese adults and increased subclinical atherosclerosis.
Arterioscler Thromb Vasc Biol 19
: 2487
–2493, 1999[Abstract/Free Full Text]
Blankenhorn DH, Selzer RH, Crawford DW, Barth JD, Liu C-R, Liu C-H, Mack WJ, Alaupovic P: Beneficial effects of colestipol-niacin therapy on the common carotid artery.
Circulation 88
: 20
–28, 1993[Abstract/Free Full Text]
Adams MR, Nakagomi A, Keech A, Robinson J, McCredie R, Bailey BP, Freedman SB, Celermajer DS: Carotid intima-media thickness is only weakly correlated with the extent and severity of coronary artery disease.
Circulation 92
: 2127
–2134, 1995[Abstract/Free Full Text]
Cox D: Regression models and life-tables.
J R Stat Soc 34
: 187
–201, 1972
Wattanakit K, Coresh J, Muntner P, Marsh J, Folsom AR: Cardiovascular risk among adults with chronic kidney disease, with or without prior myocardial infarction.
J Am Coll Cardiol 48
: 1183
–1189, 2006[Abstract/Free Full Text]
Harrell FE Jr, Lee KL, Matchar DB, Reichert TA: Regression models for prognostic prediction: Advantages, problems and suggested solutions.
Cancer Treat Rep 69
: 1071
–1077, 1985[Medline]
Bakris GL, Williams M, Dworkin L, Elliott WJ, Epstein M, Toto R, Tuttle K, Douglas J, Hsueh W, Sowers J: Preserving renal function in adults with hypertension and diabetes: A consensus approach.
Am J Kidney Dis 36
: 646
–661, 2000[Medline]
K/DOQI clinical practice guidelines on hypertension and antihypertensive agents in chronic kidney disease.
Am J Kidney Dis 43[Suppl 1]
: S1
–S290, 2004
Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL Jr, Jones DW, Materson BJ, Oparil S, Wright JT Jr, Roccella EJ: Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure.
Hypertension 42
: 1206
–1252, 2003[Abstract/Free Full Text]
Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults: Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III).
JAMA 285
: 2486
–2497, 2001[Free Full Text]
Schwenger V, Ritz E: Audit of antihypertensive treatment in patients with renal failure.
Nephrol Dial Transplant 13
: 3091
–3095, 1998[Abstract/Free Full Text]
Prosch LK, Saelen MG, Gudmundsdottir H, Dyrbekk D, Hunderi OH, Arnesen E, Paulsen D, Skjonsberg H, Os I: Blood pressure control is hard to achieve in patients with chronic renal failure: Results from a survey of renal units in Norway.
Scand J Urol Nephrol 39
: 242
–248, 2005[CrossRef][Medline]
Rubin DB: Estimating causal effects from large data sets using propensity scores.
Ann Intern Med 127
: 757
–763, 1997[Abstract/Free Full Text]
Austin PC: The performance of different propensity score methods for estimating marginal odds ratios.
Stat Med 2007
, in press
This article has been cited by other articles:
M. Chonchol, H. Gnahn, and D. Sander Impact of subclinical carotid atherosclerosis on incident chronic kidney disease in the elderly
Nephrol. Dial. Transplant.,
August 1, 2008;
23(8):
2593 - 2598.
[Abstract][Full Text][PDF]
F. A. Benedetto, G. Tripepi, F. Mallamaci, and C. Zoccali Rate of Atherosclerotic Plaque Formation Predicts Cardiovascular Events in ESRD
J. Am. Soc. Nephrol.,
April 1, 2008;
19(4):
757 - 763.
[Abstract][Full Text][PDF]
C.-C. Szeto, B. C.-H. Kwan, K.-M. Chow, K.-B. Lai, K.-Y. Chung, C.-B. Leung, and P. K.-T. Li Endotoxemia is Related to Systemic Inflammation and Atherosclerosis in Peritoneal Dialysis Patients
Clin. J. Am. Soc. Nephrol.,
March 1, 2008;
3(2):
431 - 436.
[Abstract][Full Text][PDF]
Top
Abstract
Introduction
2 test, or Pearson correlation coefficient as appropriate. The relationship between IMT and clinical outcomes was further tested by stratification of patients into quartiles according to the IMT values: Quartile I, <0.70 mm; quartile II, 0.70 to <0.80 mm; quartile III, 0.80 to <0.90 mm; and quartile IV, 
0.90 mm. Survival rates were analyzed using Kaplan-Meier survival curves. The Cox proportional hazards model was used to identify independent predictors of primary cardiovascular end point and actuarial survival (21). IMT quartile; age; gender; serum LDL; Charlson’s comorbidity score; diabetic status; systolic and diastolic BP; cigarette smoking; hemoglobin; baseline GFR; proteinuria; and serum phosphate, albumin, and CRP levels were taken as independent variables for modeling. Backward stepwise elimination was applied to remove insignificant variables. P < 0.05 was considered statistically significant. All probabilities were two-tailed. 
