Red Cell Distribution Width and Risk of Coronary Heart Disease Events

Transcription

1 Red Cell Distribution Width and Risk of Coronary Heart Disease Events Sandip K. Zalawadiya, MD a, Vikas Veeranna, MD a,b, Ashutosh Niraj, MD, MS a,b, Jyotiranjan Pradhan, MD a,b, and Luis Afonso, MD a,b, * Red cell distribution width (RDW) has emerged as a powerful predictor of all-cause mortality in variety of cardiovascular settings. However, no data are available associating RDW with coronary heart disease (CHD) risk in a healthy and nationally representative multiethnic population. A total of 7,556 participants of the National Health and Nutrition Examination Surveys 1999 to 2006 (age years, 60% women) were divided into 3 categories according to their 10-year Framingham risk of hard CHD events: <10% (n 6,173, reference category), 10% to 20% (n 1,093, intermediate-risk category), and >20% (n 290, high-risk category). Unadjusted and adjusted multivariate logistic regression analyses were performed evaluating RDW as a predictor of CHD risk. Each unit increase (0.1) in RDW posed a statistically significant greater odds of being in the intermediate-risk category (odds ratio 1.35, 95% confidence interval 1.27 to 1.45, p <0.001) and high-risk category (odds ratio 1.38, 95% confidence interval 1.25 to 1.53, p <0.001) compared to the reference category, after adjusting for race, body mass index, estimated glomerular filtration rate, hemoglobin A1c, C-reactive protein, hemoglobin, and mean corpuscular volume. Additional adjustments with serum iron, vitamin B 12, and folic acid levels did not affect the association. Subsequently, we divided participants into 2 categories according to their anemia status (as defined by the World Health Organization) to evaluate its effect. An RDW level greater than the seventy-fifth percentile in both anemic and nonanemic participants was a significant predictor of greater CHD risk while RDW of the seventy-fifth percentile or less in anemic participants failed to predict CHD (compared to nonanemic participants with similar RDW as the reference category). In conclusion, a higher RDW appears to be a powerful independent predictor of future CHD risk Elsevier Inc. All rights reserved. (Am J Cardiol 2010;106: ) The red cell distribution width (RDW), a recently described novel risk marker, has been shown to be predictive of morbidity and mortality in variety of cardiovascular settings, including heart failure, stable coronary artery disease, and acute myocardial infarction. 1 4 It is an objective measure of the heterogeneity in red blood cell size (coefficient of variability of red blood cell volume) obtained from the red blood cell size distribution curves and is routinely reported as part of a standard complete blood count. To our knowledge, no study has examined the relation between an elevated RDW, also referred to as anisocytosis, and incident coronary heart disease (CHD) risk in the general population. We, therefore, sought to determine whether an elevated RDW is associated with a greater risk of hard CHD events in a large, multiethnic cohort of nationally representative subjects. Methods The National Health and Nutrition Examination Survey (NHANES) is intended to provide information regarding the Department of a Internal Medicine and b Division of Cardiology, Wayne State University Detroit Medical Center, Detroit, Michigan. Manuscript received April 20, 2010; manuscript received and accepted June 2, *Corresponding author: Tel: (313) ; fax: address: lafonso@med.wayne.edu (L. Afonso). health status of a nationally representative sample of the civilian noninstitutionalized United States population. The study participants were interviewed at home and subsequently underwent a physical examination, including phlebotomy, at a mobile examination center after providing informed consent. We evaluated the study participants of the NHANES 1999 to 2000, 2001 to 2002, 2003 to 2004, and 2005 to The details of the NHANES studies from 1999 to 2006 have been previously published. 5 The National Center for Health Statistics ethics/institutional review board approved the study protocols for NHANES 1999 to A total of 41,474 participants were screened and 13,569 subjects identified for further analysis, after applying the exclusion criteria. The exclusion criteria were as follows: age 20 years (n 21,162) or 79 years (n 1,876); history of self-reported coronary artery disease (n 664), myocardial infarction (n 358), congestive heart failure (n 197), or stroke (n 367); self-reported diabetes mellitus (n 1,303); hemoglobin A1c 6.5 (n 1,867); and angina pectoris (n 111). The participants were also excluded from the analysis if they had missing values for the study variables, including systolic blood pressure (n 791), history of hypertension (n 17), history of dyslipidemia (n 47), chronic kidney disease (n 20), C-reactive protein (CRP) (n 95), blood pressure medication use (n 4,452), current cigarette use (n 7), family history of /10/$ see front matter 2010 Elsevier Inc. All rights reserved. doi: /j.amjcard

2 Preventive Cardiology/RDW and CHD Risk 989 Table 1 Baseline characteristics of study population stratified according to coronary heart disease (CHD) risk Variable 10-year Incident CHD Risk p Value p Test for Trend 10% (n 6,173) 10 20% (n 1,093) 20% (n 290) Age (years) Men 30.9% 79.5% 81.0% Race Black 48.1% 56.7% 60.3% Hispanic 29.5% 21.0% 22.4% White 18.5% 18.4% 15.8% Other 3.8% 3.7% 1.3% Systolic blood pressure (mm Hg) Body mass index (kg/m 2 ) Waist circumference (cm) Blood hemoglobin (g/dl) Mean corpuscular volume ( m 3 ) Total cholesterol (mg/dl) Serum high-density lipoprotein (mg/dl) Hemoglobin A1c (%) Standardized serum creatinine (mg/dl) Estimated glomerular filtration rate (ml/min/1.73 m 2 ) Blood urea nitrogen (mg/dl) Serum C-reactive protein (mg/dl) Serum vitamin B 12 (pg/ml) % 8.5% 12.1% % 5.4% 5.4% Serum iron ( g/dl) Serum folic acid (ng/ml) Red cell distribution width (%) Continuous variables are presented as mean standard deviation; categorical variables as proportions (%). Figure 1. Spearman s correlation between RDW and FRS. coronary artery disease (n 230), hemoglobin A1c (n 42), or RDW (n 354). The remaining 7,556 participants were included in the final analysis. To further assess the effect of nutritional factors such as iron, vitamin B 12, and folic acid, we ran a subgroup analysis of 4,932 subjects, all of whom had complete information for these variables; data were missing for serum iron (n 1,809), serum vitamin B 12 (n 812), and serum folic acid (n 3). The Beckman Automated Coulter Counter method of counting and sizing, combined with an automatic diluting and mixing device for sample processing, and a single-beam photometer for hemoglobinometry, were used to measure the RDW, mean corpuscular volume, and hemoglobin respectively. Plasma hemoglobin A1c was measured on whole blood with high-performance liquid chromatography, and CRP levels were quantified with latex-enhanced nephelometry. To ensure the accuracy and reproducibility of each measured laboratory parameter, control specimens with a known assigned value were run with each sample, and testing was repeated on a 2% random specimen sample. Standardized creatinine was calculated according to the NHANES guidelines using the following formula: standard creatinine serum creatinine (mg/dl) The estimated glomerular filtration rate was calculated using the following formula: estimated glomerular filtration rate [175 (standardized creatinine) (age) (0.742, if the participant were female) (1.212 if the participant were black)]. 7 Detailed descriptions of the laboratory methods used in NHANES 1999 to 2006 have been previously published. 5 The Framingham risk score (FRS) was calculated for all subjects, assigning gender-specific points as recommended by the National Cholesterol Education Program Expert Panel on the Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) guidelines. 8 The participants were then divided into 3 categories of 10-year CHD risk: low risk ( 10%, reference

3 990 The American Journal of Cardiology ( Table 2 Red cell distribution width (RDW) and 10-year coronary heart disease (CHD) risk adjusted logistic regression analysis 10-year Framingham Risk Subjects (n 7,556) RDW and 10-yr Framingham Risk Adjusted OR Model 1 Model 2 Model 3 Model 4* 10% OR 6,173 1 (referent) 1 (referent) 1 (referent) 1 (referent) 95% CI p Value 10% vs 10 20% OR 1, % CI p Value % vs 20% OR % CI p Value Model 1, unadjusted; model 2, variables included in model 1 plus race, body mass index, hyperlipidemia, estimated glomerular filtration rate, hemoglobin A1c, C-reactive protein; model 3, variables included in model 2 plus hemoglobin and mean corpuscular volume; model 4, variables included in model 3 plus serum iron, serum folic acid, and vitamin B 12 levels. * Model 4 represents subgroup analysis after exclusion of subjects with missing data for serum iron and vitamin levels. category), intermediate risk (10% to 20%), and high risk ( 20%). The baseline characteristics were then compared across these categories. The categorical variables were compared using the chi-square-test (reported as proportions) and one-way analysis of variance test for continuous variables (reported as the mean SD). The test of trend was used to analyze the linearity of the association between the continuous variables and the Framingham risk categories. Logtransformations were applied to meet the assumption of normality for variables such as CRP, estimated glomerular filtration rate, and serum folic acid. The body mass index was categorized as normal ( 25 kg/m 2 ), overweight (25 to 29.9 kg/m 2 ), and obese ( 30 kg/m 2 ). The vitamin B 12 levels were divided into 3 categories: 200, 200 to 900, and 900 pg/ml. Spearman s correlation analysis was performed between the FRS and RDW values. To evaluate the role of RDW as a predictor of greater 10-year risk of hard CHD events, unadjusted and adjusted logistic regression analyses were performed. An estimated 10-year CHD risk of 10% served as the reference category for the analysis. The following models were generated: model 1, unadjusted (RDW alone); model 2, adjusted for race, body mass index, history of dyslipidemia, estimated glomerular filtration rate, and CRP; model 3, the variables in model 2 plus blood hemoglobin and mean corpuscular volume; and model 4, the variables in model 3 plus the serum vitamin B 12, folic acid, and iron levels. Furthermore, to evaluate the effect of anemia on the association between RDW and greater 10-year CHD risk, all the study participants were divided into 2 categories according to their anemia status (hemoglobin 13 g/dl in men and 12 g/dl in women, as defined by the World Health Organization criteria). 9 Subsequently, the subjects were also divided into 2 categories according to the distribution of RDW in both anemic and nonanemic subjects (RDW at the seventy-fifth percentile or less vs greater than the seventy-fifth percentile). An RDW level at the seventy-fifth percentile or less in those without anemia served as the reference category for the analysis. Adjustments were applied similar to those for models 3 and 4, as described earlier. For all the analyses, p 0.05 was considered statistically significant. All statistical analyses was performed using Stata/SE statistical software, version 10 (StataCorp, College Station, Texas). Results Table 1 lists the baseline characteristics distribution for the entire cohort stratified according to the 3 studied FRS risk categories. The participants in the high 10-year CHD risk category were significantly older and more likely to be men, obese, hypertensive, and dyslipidemic compared to the reference category. A linear increase was seen in the RDW value with a greater CHD risk category. Blood hemoglobin, mean corpuscular volume, total cholesterol, hemoglobin A1c, serum folic acid, serum creatinine, and blood urea nitrogen levels were also greater in the higher risk CHD risk categories (p 0.001). Similarly, serum CRP levels were high and serum high-density lipoprotein levels low in those in the high-risk category. We performed Spearman s correlation analysis between RDW and FRS (Figure 1), which showed a statistically significant linear positivity between them (rho 0.14, p 0.001). Subsequently, unadjusted and adjusted multivariate logistic regression models (Table 2) were generated. These essentially represented the odds of being in higher risk categories, with higher RDW values for 10-year CHD risk compared to the reference category ( 10%). The risk of being in the intermediate-risk category (FRS 10% to 20%) compared to the reference category increased significantly by 1.16 times (95% confidence interval [CI] 1.09 to 1.22) after the adjustments shown in model 2. Additional adjustments with blood hemoglobin and mean corpuscular volume levels resulted in increase in the odds ratio (OR) to 1.35 times for classification in the intermediate-risk category (95% CI 1.27 to 1.45, p 0.001). With adjustment for nutritional factors, as shown in model 4, the adjusted risk of being in the intermediate-risk category was 1.53 times (95%

4 Preventive Cardiology/RDW and CHD Risk 991 Table 3 Red cell distribution width (RDW) and 10-year coronary heart disease (CHD) risk role of anemia Anemia RDW Category Subjects (n 7,556) Model A Model B* 10% vs 10 20% 10% vs 20% 10% vs 10 20% 10% vs 20% No 75th percentile 5,277 OR 95% CI p Value 1 (referent) 1 (referent) 1 (referent) 1 (referent) 75th percentile 1,652 OR % CI p Value Yes 75th percentile 464 OR % CI p Value th percentile 163 OR % CI p Value * Model B represents subgroup analysis after exclusion of those with missing data on serum iron and vitamin levels. Model A, RDW plus race, body mass index, dyslipidemia, estimated glomerular filtration rate, mean corpuscular volume, hemoglobin A1c, and C-reactive protein; model B, variables included in model A plus serum iron serum folic acid and vitamin B 12 levels. Figure 2. Estimated 10-year risk of hard CHD events and RDW role of anemia. (A) Adjusted for race, body mass index, dyslipidemia, estimated glomerular filtration rate, mean corpuscular volume, hemoglobin A1c, and C-reactive protein (model A). (B) Adjusted for serum folic acid, vitamin B 12, and serum iron in addition to variables included in model A (model B). CI 1.38 to 1.69, p 0.001). The association was even stronger for those in the high-risk category (FRS 20%). Adjustments showed an OR of 1.23 (95% CI 1.12 to 1.35, p 0.001) using model 2, 1.38 (95% CI 1.25 to 1.53, p 0.001) using model 3, and 1.57 (95% CI 1.37 to 1.80, p 0.001) using model 4 for being classified in the high-risk category relative to the reference category. Compared to the reference category (RDW at the seventy-fifth percentile or less in nonanemic subjects), nonanemic subjects with an RDW value greater than the seventy-fifth percentile had significantly greater odds of being in the intermediate-risk (OR 1.72, 95% CI 1.37 to 2.18, p 0.001) or high-risk (OR 2.38, 95% CI 1.64 to 3.48, p 0.001) category using the adjustments shown in model A (Table 3). Additional adjustments for nutritional factors revealed a significant OR of 1.77 (95% CI 1.40 to 2.24, p 0.01) for the intermediate-risk category and 2.45 (95% CI 1.68 to 3.59, p 0.01) for the high-risk category. Similarly, for the anemic participants with an RDW greater than the seventy-fifth percentile, the risk of being in the intermediate- or high-risk categories was significantly greater using both model A and model B (Table 3). In contrast, anemic participants with RDW values at the seventy-fifth percentile or less had a nonsignificant association with greater risk categories of 10-year CHD risk. A graphic representation of this risk association between the RDW and 10-year CHD risk, stratified according to anemia status, is shown in Figure 2. Discussion Our observations suggest that RDW, a simple morphologic marker of red blood cell size heterogeneity, is a powerful predictor of future cardiovascular risk events. We found this association to be independent of multiple potential confounding factors, including anemia, nutritional factors, hemoglobin A1c, and renal function. To our knowledge, this is the first study to associate RDW with a greater

5 992 The American Journal of Cardiology ( risk of hard CHD events in a large nationally representative multiethnic cohort of subjects. Considerable effort has been expended at identifying patients purportedly at greater risk of CHD, in particular because traditional cardiovascular risk factors (included in the FRS) do not entirely explain the risk of incident hard CHD events From our observations, it appears that RDW holds considerable promise as a powerful predictor of CHD risk; particularly attractive is that these data are widely available (as a part of the complete blood count) to the clinician at no extra cost. The risk of being classified in the intermediaterisk category of CHD was 53% greater (95% CI 1.38 to 1.69, p 0.001; model 4), with each unit increase in RDW value after adjustments. Similarly, classification into the high-risk category was also significantly elevated with greater RDW values (OR 1.57%, 95% CI 1.37 to 1.80, p 0.001; model 4). Similar independent associations have been published by previous studies evaluating the role of RDW in the prediction of all-cause mortality and cardiovascular events in different cardiovascular conditions In a post hoc analysis, Tonelli et al 4 reported findings from a randomized controlled trial of patients with coronary artery disease (without heart failure) and indicated that mortality was twice as likely in patients with an RDW 13.8% (highest quartile) compared to those with an RDW 12.6% (lowest quartile). However, our findings conflicted with the observations of Chen et al, 13 who found RDW to be a nonsignificant predictor of CHD events in a communitybased, monoethnic cohort of Chinese (adjusted hazard ratio 1.05%, 95% CI 0.65 to 1.68, p 0.83; n 3,226 and 151 CHD events during a follow-up of 15.9 years). In contrast to the latter study that exclusively enrolled Chinese subjects, our analyses evaluated a cross-section of representative United States ethnicities (Hispanics, whites, blacks, and others); therefore, our findings might be more applicable to the general United States population. Although the precise reasons for the difference in the reported association remain speculative, it is plausible that ethnic differences in CHD risk prediction (using available risk metrics), as well as differences in population characteristics, sample size, and the variables adjusted for in the studied models, might have been responsible. 14 Our models were also adjusted for hemoglobin A1c and CRP, both of which have been shown to predict future cardiovascular events independent of various conventional and nonconventional risk factors. 15,16 RDW predicted CHD risk independent of these adjustments, suggesting that it is a robust predictor of CHD risk. Our observations complement recently published work by Patel et al, 3 who showed that RDW was a strong predictor of mortality (including cardiovascular mortality) in a similar, large representative national sample of 8,175 community-dwelling adults aged 45 years (participants in the 1988 to 1994 NHANES) after adjusting for nutritional factors, along with CRP, hemoglobin, mean corpuscular volume, and other co-morbidities. Although the putative pathophysiologic mechanisms explaining RDW s association with more CHD cardiovascular events are yet to be elucidated, chronic subclinical inflammation could play a role. 17 Chronic subclinical inflammation is a well-established entity preceding de novo cardiovascular events and could adversely influence erythropoiesis by a variety of mechanisms, including direct myelosuppression of erythroid precursors, reducing renal erythropoietin production and the bioavailability of iron, increasing erythropoietin resistance in erythroid precursor cell lines, and by promoting cell apoptosis. 19,20 Thus, inflammation could lead to anisocytosis from release of immature red blood cells into the peripheral circulation. Greater RDW values have been independently associated with greater CRP levels, a well-established surrogate marker of inflammation, as well as numerous other inflammatory markers such as interleukin-6 and soluble tumor necrosis factor receptors 1 and 2. 21,22 In our study, we observed that CRP levels increased linearly with greater RDW quartiles (p for trend 0.001, data not shown). Exposure to greater oxidative stress might be yet another potential contributing pathobiologic mechanism. High oxidative stress has been associated with greater risk of incident cardiovascular events. 23 High oxidative stress can also cause anisocytosis by reducing red blood cell survival, thereby escalating release of premature red blood cells into the peripheral circulation. 24 Thus, a greater RDW might portend a greater level of oxidative stress, a potential harbinger of future cardiovascular events. RDW is also commonly elevated in conditions associated with increased red blood cell destruction (hemolytic anemias) or defective erythropoiesis (e.g., nutritional deficiencies of iron, folic acid, vitamin B 12 ) or blood transfusion. Similarly, we also confirmed a linear increase in RDW with increasing age (data not shown). Thus, elevated RDW might be a surrogate composite measure of multiple pathophysiologic processes (i.e., chronic inflammation, greater oxidative stress, nutritional deficiencies, aging), which, to varying extents, might play a role in the etiopathogenesis of adverse cardiovascular events. The strengths of our study included its large sample size that included a nationally representative multiethnic cohort of participants. We performed a detailed assessment of potential confounding factors and effect modifiers that should make our results more generalizable; however, the possibility of residual confounding could not be entirely eliminated. Although our study did have limitations such as the unavailability of actual cardiovascular events during a set follow-up period, the FRS is a well-established risk assessment tool with a high degree of concordance between the expected and actual event frequencies, supporting the validity of our study findings. However, future research should focus on specifically exploring the use of RDW to reclassify those in the intermediate FRS category. Selection bias resulting from the exclusion of participants with missing data, although possible, appeared less likely. The results from the present large multiethnic sample of a representative population disregard any possible assumptions of differences in the risk prediction between the available sample and the excluded sample (data not shown). 1. van Kimmenade RR, Mohammed AA, Uthamalingam S, van der Meer P, Felker GM, Januzzi JL Jr. Red blood cell distribution width and 1-year mortality in acute heart failure. Eur J Heart Fail 2010;12:

6 Preventive Cardiology/RDW and CHD Risk Dabbah S, Hammerman H, Markiewicz W, Aronson D. Relation between red cell distribution width and clinical outcomes after acute myocardial infarction. Am J Cardiol 2010;105: Patel KV, Ferrucci L, Ershler WB, Longo DL, Guralnik JM. Red blood cell distribution width and the risk of death in middle-aged and older adults. Arch Intern Med 2009;169: Tonelli M, Sacks F, Arnold M, Moye L, Davis B, Pfeffer M. Relation between red blood cell distribution width and cardiovascular event rate in people with coronary disease. Circulation 2008;117: Centers for Disease Control and Prevention (CDC), National Center for Health Statistics. National Health and Nutrition Examination Survey data. Available at: Accessed on March 15, Centers for Disease Control and Prevention (CDC), National Center for Health Statistics (NCHS). NCHS research ethics review board (ERB) approval. Available at: Accessed on March 15, Levey AS, Coresh J, Greene T, Stevens LA, Zhang YL, Hendriksen S, Kusek JW, Van Lente F. Using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate. Ann Intern Med 2006;145: 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 2001;285: Nutritional anaemias: report of a WHO scientific group. World Health Organization Tech Rep Ser 1968;405: Helfand M, Buckley DI, Freeman M, Fu R, Rogers K, Fleming C, Humphrey LL. Emerging risk factors for coronary heart disease: a summary of systematic reviews conducted for the U.S. Preventive Services Task Force. Ann Intern Med 2009;151: de Lemos JA, Lloyd-Jones DM. Multiple biomarker panels for cardiovascular risk assessment. N Engl J Med 2008;358: Smith SC Jr. Current and future directions of cardiovascular risk prediction. Am J Cardiol 2006;97:28A 32A. 13. Chen PC, Sung FC, Chien KL, Hsu HC, Su TC, Lee YT. Red blood cell distribution width and risk of cardiovascular events and mortality in a community cohort in Taiwan. Am J Epidemiol 2010;171: Kurian AK, Cardarelli KM. Racial and ethnic differences in cardiovascular disease risk factors: a systematic review. Ethn Dis 2007;17: Khaw KT, Wareham N, Bingham S, Luben R, Welch A, Day N. Association of hemoglobin A1c with cardiovascular disease and mortality in adults: the European prospective investigation into cancer in Norfolk. Ann Intern Med 2004;141: Ridker PM, Cushman M, Stampfer MJ, Tracy RP, Hennekens CH. Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. N Engl J Med 1997;336: Libby P, Ridker PM, Maseri A. Inflammation and atherosclerosis. Circulation 2002;105: Danesh J, Wheeler JG, Hirschfield GM, Eda S, Eiriksdottir G, Rumley A, Lowe GD, Pepys MB, Gudnason V. C-reactive protein and other circulating markers of inflammation in the prediction of coronary heart disease. N Engl J Med 2004;350: Ghali JK. Anemia and heart failure. Curr Opin Cardiol 2009;24: Laftah AH, Sharma N, Brookes MJ, McKie AT, Simpson RJ, Iqbal TH, Tselepis C. Tumour necrosis factor alpha causes hypoferraemia and reduced intestinal iron absorption in mice. Biochem J 2006;397: Lippi G, Targher G, Montagnana M, Salvagno GL, Zoppini G, Guidi GC. Relation between red blood cell distribution width and inflammatory biomarkers in a large cohort of unselected outpatients. Arch Pathol Lab Med 2009;133: Forhecz Z, Gombos T, Borgulya G, Pozsonyi Z, Prohaszka Z, Janoskuti L. Red cell distribution width in heart failure: prediction of clinical events and relationship with markers of ineffective erythropoiesis, inflammation, renal function, and nutritional state. Am Heart J 2009;158: Berliner JA, Navab M, Fogelman AM, Frank JS, Demer LL, Edwards PA, Watson AD, Lusis AJ. Atherosclerosis: basic mechanisms. Oxidation, inflammation, and genetics. Circulation 1995;91: Ghaffari S. Oxidative stress in the regulation of normal and neoplastic hematopoiesis. Antioxid Redox Signal 2008;10: