Red cell distribution width and risk of peripheral artery disease: Analysis of National Health and Nutrition Examination Survey 1999

Abstract

Red cell distribution width (RDW) is an independent predictor of the 10-year estimated risk of coronary heart disease (CHD) events. However, RDW’s association with peripheral artery disease (PAD) – a CHD risk equivalent – has not been evaluated to date. In this cross-sectional study, we examined 6950 participants of the National Health and Nutrition Examination Survey, 1999–2004. PAD was defined as an ankle–brachial index below 0.9 (n = 618). RDW was divided into quartiles (Q) (Q1: ≤ 12.2; Q2: 12.3–12.5; Q3: 12.6–13.0; Q4: ≥ 13.1) and PAD risk was compared across these quartiles using adjusted multivariate logistic regression. A graded increase in prevalent PAD with increasing RDW quartiles was observed (4.2% in Q1 vs 13.9% in Q4; test of trend p < 0.001). Risk of PAD was significantly higher (odds ratio (OR) 1.19, 95% confidence interval (CI): 1.06–1.34; p = 0.003) after adjusting for age, sex, race, body mass index, hypertension, hyperlipidemia, diabetes, smoking, estimated glomerular filtration rate, C-reactive protein, hemoglobin, mean corpuscular volume, and nutritional factors (folate, iron and vitamin B₁₂) deficiencies with each unit (0.1) increase in RDW. Upon receiver-operating characteristics analysis, the predictive accuracy of the American College of Cardiology / American Heart Association (ACC/AHA)-defined PAD screening criteria (for a high-risk population) was 0.657 at best, but improved significantly (0.727) after addition of RDW (p < 0.0001). In conclusion, higher levels of RDW are independently associated with a higher risk of PAD and can significantly improve the risk prediction beyond that estimated by ACC/AHA-defined PAD screening criteria.

Keywords

peripheral artery disease RDW

Introduction

Red cell distribution width (RDW) is routinely reported in automated complete blood counts (CBC) and has traditionally been used to differentiate various types of anemia.¹ The evidence associating RDW with a higher risk of mortality has been expanding since the initial report of its prognostic utility in heart failure patients.² RDW has also been shown to independently predict overall and cardiovascular (CV) mortality in the general population and various high-risk populations.^3–7 We recently reported RDW to be strongly associated with a higher 10-year estimated risk of coronary heart disease (CHD) events in a predominantly healthy, low-risk general population, free of prior history of CV events and diabetes.⁸

From the data evaluating the association between RDW and mortality in heart failure patients, Felker et al.² suggested the possibility of RDW being an integrated marker of several underlying pathophysiological processes such as nutritional deficiencies, renal dysfunction, hepatic congestion and inflammatory stress. Subsequent research showed RDW to be strongly associated with markers of chronic subclinical inflammation, higher oxidative stress, under-nutrition, and beta-natriuretic peptide (BNP).^9,10 Our group also demonstrated that RDW was strongly and independently associated with microalbuminuria, a marker of generalized endothelial dysfunction, in a nationally representative cohort.¹¹

RDW is easily measured, standardized and typically reported with complete blood count values at no additional cost; its independent predictive value for various CV events makes further research of other ‘at-risk populations’ for CV events imperative. Peripheral artery disease (PAD), now widely considered a CHD risk equivalent, is one such condition.

PAD is preceded by higher levels of chronic subclinical inflammation, endothelial dysfunction and oxidative stress.^12–14 There is an increasing effort to identify a novel biomarker or set of biomarkers to effectively characterize individuals with or without PAD. Markers of inflammation (C-reactive protein (CRP), interleukin-6 (IL-6), etc.), endothelial dysfunction (soluble cell adhesion molecules, von Willebrand factor, etc.) and oxidative stress (homocysteine (Hcy), isoprostanes, etc.) have been shown to be associated with PAD.^12,15–17

To our knowledge, no prior study has explored the association between RDW and PAD. Therefore, in this cross-sectional cohort of the National Health and Nutrition Examination Survey (NHANES), we sought to evaluate the association between RDW and PAD. Contingent upon the results, we further sought to compare the clinical utility of RDW in identifying those at risk of PAD apropos the currently recommended American College of Cardiology / American Heart Association (ACC/AHA) PAD screening guidelines.¹⁸

Methods

Study population

NHANES is a cross-sectional survey of a non-institutionalized civilian cohort of nationally representative individuals. We analyzed participants of the NHANES, 1999–2004 who underwent standardized interviews, physical examination and laboratory testing along with ankle–brachial index (ABI) measurement. The study protocols for NHANES received the National Center for Health Statistics ethics/institutional review board approval.¹⁹

Data collection

Data on ABI were available for a total of 7571 individuals, from which the study was conducted among participants aged 40 years of older (n = 6950), after excluding those with missing values on RDW (n = 218), those with ABI values > 1.4 (n = 70) and other study variables (n = 333). A standardized procedure protocol was followed at all study sites by trained personnel. Data were collected for demographic variables such as age, sex, and ethnicity (self-reported) and for medical co-morbidities such as hypertension (defined as blood pressure ≥ 140/90 or use of anti-hypertensive medications or physician diagnosis of hypertension) and diabetes mellitus (DM) (defined as physician diagnosis of DM, taking oral hypoglycemics or insulin for DM, or non-fasting plasma glucose ≥ 11.1 mmol/l (200 mg/dl) or fasting plasma glucose ≥ 7 mmol/l (126 mg/dl), or hemoglobin (Hgb) A1c ≥ 6.5). Data were also collected for ever smoking > 100 cigarettes during lifetime, hyperlipidemia (diagnosed by physician or serum total cholesterol ≥ 240 mg/dl or use of lipid-lowering therapy), body mass index (BMI; kg/m2) (normal (BMI < 25), overweight (BMI 25–29.9), and obese (BMI ≥ 30)), and comorbidities (self-reported) such as coronary artery disease (CAD), myocardial infarction (MI), stroke, angina, anti-hypertension medication use and family history of MI.²⁰

Ankle–brachial index measurement

PAD, as defined by ABI, was the primary outcome variable of the study. For participants with ≥ 1 arm and weight < 400 lb (181.5 kg), the systolic blood pressure (SBP) was measured using blood pressure cuffs on the right brachial artery and both posterior tibial arteries with an 8-MHz Doppler probe. Two measurements at each site were taken and averaged for participants aged 40–59 years, whereas one measurement was taken for participants aged ≥ 60 years. For those with conditions precluding measurement of the right arm, the left brachial artery SBP was taken. The ABI was calculated as the ratio of the average ankle SBP to arm SBP.²⁰ The lower of the two measurements was considered the ABI for the present study. PAD was defined as ABI < 0.9 and participants with an ABI greater than 1.4 were excluded.

ACC/AHA screening criteria for PAD

Current ACC/AHA guidelines recommend screening certain high-risk groups using ABI (class ІB ). Recommendations to perform ABI include those individuals aged less than 50 years with a history of diabetes and one other atherosclerosis risk factor (smoking, dyslipidemia, hypertension, or hyperhomocysteinemia); those aged between 50 and 69 years with a history of smoking or diabetes; those aged 70 years and older; those with leg symptoms upon exertion (suggestive of claudication) or ischemic rest pain; those with an abnormal lower-extremity pulse examination; and those with known atherosclerotic coronary, carotid, or renal artery disease¹⁸ (Table 1).

Table 1.

Receiver-operative characteristics (ROC) analysis comparing area under the curve (AUC) for individual risk categories with and without RDW for peripheral artery disease^a

Risk factor categories		AUC not including RDW	AUC including RDW	Improvement in ROC–AUC
A	Age < 50 years with diabetes and one other atherosclerosis risk factor (smoking, dyslipidemia, hypertension, or hyperhomocysteinemia)	0.505	0.643	< 0.001
B	Age 50–69 years and history of smoking or diabetes	0.516	0.643	< 0.001
C	Age 70 years and older	0.676	0.736	< 0.001
D	Known atherosclerotic coronary, carotid or renal artery disease	0.586	0.671	< 0.001
E	Combination of criteria ‘A’ through ‘D’	0.657	0.727	< 0.001

Peripheral artery disease is defined as ankle–brachial index < 0.9.

Laboratory measurements

RDW, Hgb and mean corpuscular volume (MCV) were measured using the Beckman automated Coulter counter method in combination with an automatic diluting and mixing device for sample processing, and a single-beam photometer for hemoglobinometry.²⁰ High-sensitivity C-reactive protein (hs-CRP) was measured using latex-enhanced nephelometry, HbA1c by high-performance liquid chromatography (HPLC) and plasma Hcy by a fully automated fluorescence polarization immunoassay (FPIA; Abbott Diagnostics).²⁰ Serum creatinine was standardized by the following formula: 1.013*serum creatinine (mg/dl) + 0.147. The estimated glomerular filtration rate (eGFR) (categorized as eGFR ≥ 60 vs eGFR < 60) was then measured using the formula suggested by the Modification of Diet in Renal Disease study equation.²¹ Iron deficiency was defined as the participants meeting at least two of the following three criteria: transferrin saturation < 15%, serum ferritin level < 12 ng/ml, and erythrocyte protoporphyrin level > 69.8 μm.²² Additionally, a vitamin B₁₂ deficiency was defined as a serum B₁₂ level < 200 pg/dl, whereas a folic acid deficiency was defined as serum folic acid < 2.6 ng/ml or a red blood cell folate level < 102.6 ng/ml. Urine creatinine was analyzed by Jaffe reaction and urinary albumin was measured using a solid-phase fluorescent immunoassay.²⁰ The urine albumin–creatinine ratio (mg/g) was calculated by dividing the urinary albumin value by the urinary creatinine concentration. Microalbuminuria was defined as a urine albumin/creatinine ratio (ACR) ≥ 30 mg/g. Detailed data on the NHANES operations are available on the NHANES website (www.cdc.gov/nchs/nhanes/nhanes_questionnaires.htm).

Statistical analysis

We analyzed RDW both as a continuous and a categorical variable (in quartiles; based on its distribution among participants without PAD). Baseline characteristics were compared across these quartiles using the chi-square test for categorical variables (%) and the analysis of variance test for continuous variables (mean ± SD). Hs-CRP and Hcy were log-transformed to normalize their distribution. The primary aim of the study was to assess whether RDW is independently associated with prevalent PAD beyond the risk factors proposed by ACC/AHA guidelines. Additionally, we sought to evaluate whether variables such as Hgb, MCV, a deficiency of nutritional factors or inflammatory markers like CRP and the presence of renal dysfunction served as confounders or effect modifiers for the association of RDW with PAD. Therefore, the following multivariable adjusted logistic regression models were generated: Model 1 – analysis of RDW and PAD adjusted for age, sex, race, BMI, hypertension, DM, hyperlipidemia, smoking, eGFR, hs-CRP, blood Hgb and MCV levels; Model 2 – adjusted for variables in Model 1 + history of CV disease, history of congestive heart failure, family history of CHD, mean platelet volume; and Model 3 – variables in Model 2 adjusted for deficiencies of serum iron, folate and vitamin B₁₂ (analysis excludes observations (n = 976) with missing information on above nutrients). Formal statistical analysis for the interaction between various independent variables with RDW was also performed, which was found to be statistically non-significant. A graphical plot was generated assessing the predicted probability and observed rate of prevalent PAD against RDW values. To further evaluate for the association between RDW and levels of ABI values, we performed Pearson’s correlation analysis after excluding those with the missing information on nutritional factors and those with extreme RDW elevations (RDW > 17; n = 80). Upon identifying RDW to be significantly associated with prevalent PAD, we sought to evaluate whether it added to the predictive accuracy of ACC/AHA recommended criteria for pursuing work-up for PAD. Therefore, we performed receiver-operating characteristics (ROC) analysis which tests the general discriminative power of a diagnostic test using sensitivity (Se) and specificity (Sp) parameters over the whole range of testing values and also provides area under the curve (AUC) which indicates diagnostic power of the test. We compared the diagnostic accuracy of RDW for PAD risk prediction to that of ACC/AHA-defined screening criteria. Six-year sample weights provided by the National Center for Health Statistics were used in all analyses to incorporate the complex survey design of NHANES with appropriate reflections of stratification and clustering. Statistical significance was defined as a p-value < 0.05 for the entire analysis. All analyses were performed using statistical software STATA, version 10 (STATA Corp LP, College Station, TX, USA).

Results

Significant differences were observed in baseline characteristics distribution across the RDW quartiles (Table 2). There was a graded increase in the proportion of patients with comorbidities such as CAD, hypertension, DM, MI, stroke and angina, and microalbuminuria with increasing quartiles of RDW (p < 0.001). Similarly, there was a graded increase in mean values for age, SBP, HbA1c, hs-CRP, Hcy and serum creatinine (p for trend < 0.001). In contradistinction, mean levels of Hgb, MCV and eGFR decreased with increasing quartiles of RDW (p for trend < 0.001). The proportion of African Americans increased and the proportion of Caucasian Americans decreased with increasing RDW quartiles (p < 0.001). Blood levels of lipids (e.g. total cholesterol and high-density lipoprotein) exhibited a non-graded difference in mean levels across RDW quartiles.

Table 2.

Baseline characteristics of the study cohort (National Health and Nutrition Examination Survey, 1999–2004)

	RDW quartiles (n = 6950)				p-value
	Q1	Q2	Q3	Q4
	≤ 12.2	12.3–12.5	12.6–13.0	≥ 13.1
	n = 1896	n = 1417	n = 1522	n = 2115
Age, years	56.5 ± 11.9	59.2 ± 12.8	61.2 ± 12.7	63.0 ± 13.4	< 0.001
Males, %	50	55	51	49	0.007
Race, %
African American	9.3	10.3	17.3	29.5	< 0.001
Hispanic American	24.5	27.0	27.0	22.3
Caucasian American	62.6	60.2	52.8	45.2
Other American	3.5	2.4	2.9	2.9
Coronary artery disease, %	4.2	5.8	5.9	8.8	< 0.001
Diabetes, %	13.6	13.9	17.3	19.5	< 0.001
MI, %	4.2	4.9	5.8	9.4	< 0.001
Stroke, %	2.0	3.7	4.1	6.3	< 0.001
Angina, %	3.9	4.3	6.8	6.4	< 0.001
Smoking (> 100 cigarettes in lifetime)	53.2	52.1	54.0	55.4	0.24
Family history of MI	12.7	12.3	12.5	11.6	0.72
Systolic blood pressure, mmHg	128 ± 19	130 ± 20	133 ± 21	134 ± 22	< 0.001
Anti-hypertension medication use, %	23.4	27.9	35.2	41.1	< 0.001
Body mass index
Overweight	43.6	41.5	39.2	37.0	< 0.001
Obese	25.4	30.8	36.4	35.9
Blood hemoglobin, g/dl	14.7 ± 1.2	14.7 ± 1.3	14.4 ± 1.3	13.8 ± 1.6	< 0.001
Mean corpuscular volume, μm³	92.4 ± 3.8	91.8 ± 3.9	91.3 ± 4.3	88.6 ± 7.2	< 0.001
Total cholesterol, mg %	210 ± 42	211 ± 38	211 ± 42	207 ± 43	0.016
Serum high-density lipoprotein, mg %	53 ± 16	52 ± 16	53 ± 16	53 ± 17	0.05
Hemoglobin A1c, %	5.6 ± 1.2	5.7 ± 1.0	5.8 ± 1.1	5.9 ± 1.2	< 0.001
Standardized serum creatinine, mg %	1.0 ± 0.2	1.02 ± 0.2	1.04 ± 0.4	1.13 ± 0.7	< 0.001
Estimated GFR, ml/min/1.73 m²	73 ± 17	71 ± 18	70 ± 18	68 ± 20	< 0.001
High-sensitivity C-reactive protein, mg/dl	0.36 ± 0.6	0.40 ± 0.64	0.46 ± 0.72	0.64 ± 1.19	< 0.001
Homocysteine, μmol/l	8.8 ± 3.2	9.2 ± 3.4	9.6 ± 3.8	10.8 ± 6.4	< 0.001
Ankle–brachial index (< 0.9)	4.2	6.1	10.3	13.9	< 0.001
^aIron deficiency, %	7.5	7.6	9.3	17.3	< 0.001
^aFolate deficiency, %	0.31	0.35	0.52	1.49	< 0.001
^aVitamin B₁₂ deficiency, %	1.7	2.2	2.1	3.4	0.006
^bUrine microalbuminuria, %	19	22	26	36	< 0.001

Sample size excluded missing values on deficiencies of iron, folic acid and vitamin B₁₂ (n = 5974).

Sample size excluded individuals with macroalbuminuria (n = 2767).

RDW, red cell distribution width; MI, myocardial infarction; GFR, glomerular filtration rate.

The proportion of individuals with PAD increased significantly as the RDW quartile increased (4.22% in Q1 vs 13.90% in Q4; p < 0.001) (Table 3). The odds of having PAD increased by 19% with each unit (0.1) increase in RDW (95% confidence interval (CI): 1.06–1.34; p = 0.003) after adjustments shown in Model 3 (Table 3). Similar results were obtained when RDW was evaluated in categories (Table 3).

Table 3.

Red cell distribution width and overall risk of prevalent peripheral artery disease

A: RDW as a continuous variable (for each 0.1 increase)		OR [95% CI]; p-value
Model 1		1.23 [1.14–1.33]; < 0.001
Model 2		1.17 [1.06–1.30]; 0.004
Model 3^a		1.19 [1.06–1.34]; 0.003
B: RDW quartiles (n)	Q1 (1896)	Q2 (1417)	Q3 (1522)	Q4 (2115)
Range	≤ 12.2	12.3–12.5	12.6–13.0	≥ 13.1
Case (%)	80 (4.22%)	87 (6.14%)	157 (10.32%)	294 (13.90%)
		OR [95% CI]; p-value	OR [95% CI]; p-value	OR [95% CI]; p-value
Model 1	ref. 1	1.15 [0.76–1.75]; 0.48	1.54 [1.12–2.12]; 0.009	2.20 [1.48–3.27]; < 0.001
Model 2	ref. 1	1.10 [0.72–1.67]; 0.66	1.42 [1.01–2.02]; 0.047	1.73 [1.10–2.70]; 0.018
Model 3^a	ref. 1	1.01 [0.64–1.59]; 0.96	1.43 [0.93–2.19]; 0.09	1.65 [1.00–2.73]; 0.050

Model 1: RDW adjusted for age, sex, race, body mass index, hypertension, diabetes, hyperlipidemia, smoking, estimated glomerular filtration rate and high-sensitivity C-reactive protein, blood hemoglobin and mean corpuscular volume levels.

Model 2: Model 1 + history of cardiovascular disease, history of congestive heart failure, family history of coronary heart disease, mean platelet volume.

Model 3^a: Model 2 + deficiencies of serum iron, folate and vitamin B₁₂.

Analysis is done for total sample size of 5974 individuals.

RDW, red cell distribution width; OR, odds ratio; CI, confidence interval.

The interaction term between RDW and co-variates for PAD risk prediction was found to be statistically non-significant (age (< 70 vs ≥ 70; p = 0.94), sex (p = 0.94), diabetes (p = 0.23), hypertension (p = 0.83), renal impairment (p = 0.052), history of atherosclerotic diseases (p = 0.27), nutritional factors deficiency (p = 0.509) and ethnicities (p > 0.05 for all)).

Figure 1 shows the relationship between the predicted and observed rate for probability of PAD with increasing RDW values; as shown, there appears to be a linear relationship between the two. Similarly, Figure 2 shows the association between RDW values and ABI values as continuous variables; as shown, a statistically significant linear negative association is apparent (Pearson’s rho = −0.18; p < 0.001). Figure 3A shows the results of multivariate logistic regression analysis evaluating the independent role of RDW in PAD risk prediction beyond that conferred by traditional risk factors outlined in the ACC/AHA guidelines. It is worth noting that we also performed analyses presented in Tables 1 and 2 using Hcy as a co-variate in Models 2 and 3; we did not observe any change in the strength of association between RDW and PAD upon adjustment for Hcy (data not shown).

Figure 1.

The relationship between RDW values and expected versus observed rate of having prevalent PAD among National Health and Nutrition Examination Survey 1999–2004 participants

Figure 2.

Pearson’s correlation between RDW and ankle–brachial index values among National Health and Nutrition Examination Survey 1999–2004 participants

Figure 3.

Value of red cell distribution width (RDW) for peripheral artery disease (PAD) risk prediction: (A) multivariate logistic regression model demonstrating significant risk factors for PAD; (B) receiver-operative characteristics analysis comparing the predictive accuracy (area under the curve (AUC)) of American College of Cardiology/American Heart Association-defined high-risk group and RDW alone and in combination, for the diagnosis of subjects with ankle–brachial index < 0.9 (p-value for improvement AUC by including RDW to high-risk groups < 0.001)

Table 1 and Figure 3B show results of the ROC analysis comparing the predictive accuracy (AUC) provided by RDW and the ACC/AHA-defined screening criteria for a high-risk population at risk of PAD (individual criteria and overall combination) alone and in combination. The overall predictive accuracy of the ACC/AHA-defined criteria (risk category E in Table 1) was 0.657, which improved significantly (to 0.727) after the addition of RDW. Similarly, individual risk categories for those less than 70 years of age (defined by ACC/AHA) provided poor predictive accuracy for identifying those at risk of PAD (AUC of 0.505 for category A and 0.516 for category B; Table 1), which improved significantly after addition of RDW to the model (AUC of 0.643 for both category A and B; p-value for improvement in AUC < 0.001) (Table 1). RDW improved the AUC for all the individual risk categories; although, the magnitude of increase was greater among those under the age of 70 years (Table 1).

Discussion

In this study, we systematically evaluated the potential association between RDW and prevalent PAD for the first time. The following conclusions may be drawn from our analysis: (1) RDW is strongly associated with PAD with a graded increase in prevalence across quartiles of increasing RDW; (2) higher RDW is associated with PAD independent of sex, DM, hypertension, and deficiency of nutritional factors; (3) RDW is independently associated with PAD in participants of all the ethnicities and participants with/without renal dysfunction, though the association was apparently stronger in African Americans, Caucasian Americans and participants without renal dysfunction; and (4) RDW significantly improved the risk prediction offered by current ACC/AHA-defined screening guidelines for a high-risk population.

RDW is an emerging marker, with published data supporting its role in predicting survival in various cardiovascular conditions.^2–6 However, no prior study has evaluated the association between RDW and PAD in a nationally representative sample of multi-ethnic individuals. We observed an RDW ≥ 13% (Q4) to be associated with a 79% higher odds of having PAD (95% CI: 1.37–2.34; p < 0.001) after adjustments for renal function, hs-CRP and Hgb level in addition to demographic variables, co-morbidities and nutritional deficiencies. Additionally, upon stratifying the cohort based on deficiency status of nutritional factors, renal function, hypertension, diabetes, sex and ethnicity, RDW emerged as an independent predictor of PAD (Table 1). Comparable results were reported by Patel et al.³ and Perlstein et al.⁵ who showed RDW to be associated with all-cause and cardiovascular mortality upon similar adjustments in a cohort of NHANES-ІІІ study participants. The novelty of our study findings is that we observed higher RDW to be positively associated with a higher risk of prevalent PAD in a graded manner, raising the potentially significant utility of RDW as a novel biomarker for PAD.

The precise pathobiologic mechanisms for the association between RDW and various cardiovascular diseases are yet to be elucidated, although the potential role of oxidative stress, chronic inflammation and endothelial dysfunction warrants consideration. During their average lifespan of 120 days, red blood cells (RBCs) are exposed to a variety of harmful insults that range from mechanical and oxidative to hyperosmotic stress.²³ They are directly involved in the regulation of reactive oxygen and nitrogen species delivery to peripheral tissue, thereby modulating the cycle of oxidative damage.²⁴ It has been hypothesized from experimental studies that RBCs have free radical scavenging (the presence of an abundant amount of powerful anti-oxidant enzymes which help control oxidative stress under normal physiological conditions) and pro-oxidant (redox imbalance in RBCs resulting from overwhelming oxidative damage and inflammation in pathological tissues which alters and overcomes the anti-oxidant mechanisms and, therefore, RBCs function as pro-oxidant cells modifying and regulating distant vascular beds and endothelial function) and signaling properties (oxidative stress causing alteration in RBC membrane properties and the interaction among RBCs and between RBCs and other cells which leads to perpetuation of oxidative damage).^25,26 Oxidative stress can lead to cytoskeleton rearrangement and loss of lipid asymmetry in RBC membranes leading RBCs to become more rigid and asymmetric in size – ‘anisocytosis’.²⁷ It has been shown that ‘oxidized RBCs’ have higher aggregability and adhesiveness to the endothelial tissue leading to a vicious cycle of oxidative damage-induced endothelial dysfunction.²⁸ In this context, we observed an independent association between RDW and microalbuminuria – a well-established marker of generalized endothelial dysfunction in the general US population free of prior history of CV disease.¹¹ Similarly, in the current study, we again noted that the proportion of participants with microalbuminuria increased significantly with higher RDW quartiles (19% in Q1 vs 36% in Q4; p for trend < 0.001). The association of higher RDW with markers of oxidative stress,¹⁰ endothelial dysfunction¹¹ and chronic inflammation⁹ underscore the importance of RDW as a real-time integrative biomarker of generalized atherosclerotic burden. In this context, the incremental association of higher RDW with a higher risk of PAD found in our study emphasizes the clinical relevance of this biomarker. Whether RDW has a differential association with pathophysiological processes going on in different vascular beds, is a subject of future research.

RDW and ACC/AHA screening guidelines for a high-risk population

According to the current ACC/AHA guidelines, it is deemed appropriate to consider measuring the ABI in high-risk subgroups (Table 1) at risk of PAD.¹⁸ Although the current recommendations are based on epidemiological studies only, to date no prior study has analyzed the predictive utility of these screening criteria in the general population, especially one representing a multi-ethnic US general population. As our analysis revealed RDW to have an independent predictive value beyond the established ACC/AHA-defined screening criteria, we further attempted to explore the incremental value of RDW over these criteria. We observed a significant improvement in the predictive accuracy of currently recommended screening criteria with the addition of RDW (values > 13) for the identification of individuals with PAD (Table 1). Since RDW is readily available as part of the routine complete blood count, and is easily reproducible and cost-effective, it could be potentially utilized alongside recommended screening criteria to finesse PAD risk prediction.

Our study suffers from the inherent limitations of a cross-sectional analysis, restricting our efforts to evaluation of prevalent rather than incident PAD. Further studies with a prospective follow-up will be needed to examine the temporal relationship between RDW with incident PAD. The observed weaker association between RDW and PAD in patients with renal impairment appears to be secondary to the relatively small sample and event size in that subgroup. This database also lacked serial measurements of RDW, which might have provided further insight into its association with PAD severity. Misclassification of PAD may also have occurred, as NHANES did not follow current recommended guidelines for ABI measurements.¹⁸ However, prior studies utilizing ABI measurements from this database have contributed critical information leading to significant advances in the understanding of PAD.^29–31 We lacked data on cardiovascular mortality and thus were unable to explore the effect modification of RDW, if any, on the association between PAD and cardiovascular outcomes. Additionally, the cross-sectional nature of NHANES precludes conclusions about direction or causality of associations observed in the present study. Despite these limitations, our study does have the strengths of a large, multi-ethnic sample of nationally representative individuals, in addition to a wide array of data collected by standardized protocols adhering to rigorous quality control. Pending additional research, our findings should be viewed as preliminary and serve as a basis for future studies evaluating the causal and temporal relationships between RDW, PAD and cardiovascular outcomes.

Conclusion

Our novel observations suggest a strong and independent relationship between RDW with PAD. These findings are particularly noteworthy, as RDW appears to be an emerging, cost-effective biomarker, which could potentially help optimize the predictive accuracy of the ACC/AHA currently recommended screening guidelines aimed at identifying individuals at risk for PAD.

Footnotes

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

The contents are solely the responsibility of the authors and do not represent the official view of any organizations. The authors have no conflicts of interest or financial disclosures to report.

References

Tefferi

Hanson

Inwards

. How to interpret and pursue an abnormal complete blood cell count in adults. Mayo Clin Proc 2005; 80: 923–936.

Felker

Allen

Pocock

. Red cell distribution width as a novel prognostic marker in heart failure: data from the CHARM Program and the Duke Databank. J Am Coll Cardiol 2007; 50: 40–47.

Patel

Ferrucci

Ershler

Longo

Guralnik

. Red blood cell distribution width and the risk of death in middle-aged and older adults. Arch Intern Med 2009; 169: 515–523.

Patel

Semba

Ferrucci

. Red cell distribution width and mortality in older adults: a meta-analysis. J Gerontol A Biol Sci Med Sci 2010; 65: 258–265.

Perlstein

Weuve

Pfeffer

Beckman

. Red blood cell distribution width and mortality risk in a community-based prospective cohort. Arch Intern Med 2009; 169: 588–594.

Tonelli

Sacks

Arnold

Moye

Davis

Pfeffer

. Relation between red blood cell distribution width and cardiovascular event rate in people with coronary disease. Circulation 2008; 117: 163–168.

Zalawadiya

Zmily

Farah

Daifallah

Ali

Ghali

. Red cell distribution width and mortality in predominantly African-American population with decompensated heart failure. J Card Fail 2011; 17: 292–298.

Zalawadiya

Veeranna

Niraj

Pradhan

Afonso

. Red cell distribution width and risk of coronary heart disease events. Am J Cardiol 2010; 106: 988–993.

Forhecz

Gombos

Borgulya

Pozsonyi

Prohaszka

Janoskuti

. 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: 659–666.

10.

Allen

Felker

Mehra

. Validation and potential mechanisms of red cell distribution width as a prognostic marker in heart failure. J Card Fail 2010; 16: 230–238.

11.

Afonso

Zalawadiya

Veeranna

Panaich

Niraj

Jacob

. Relationship between red cell distribution width and microalbuminuria: a population-based study of multiethnic representative US adults. Nephron Clin Pract 2011; 119: c277–c282.

12.

Loffredo

Marcoccia

Pignatelli

. Oxidative-stress-mediated arterial dysfunction in patients with peripheral arterial disease. Eur Heart J 2007; 28: 608–612.

13.

Beckman

Preis

Ridker

Gerhard-Herman

. Comparison of usefulness of inflammatory markers in patients with versus without peripheral arterial disease in predicting adverse cardiovascular outcomes (myocardial infarction, stroke, and death). Am J Cardiol 2005; 96: 1374–1378.

14.

Wildman

Muntner

Chen

Sutton-Tyrrell

. Relation of inflammation to peripheral arterial disease in the national health and nutrition examination survey, 1999–2002. Am J Cardiol 2005; 96: 1579–1583.

15.

Ridker

Cushman

Stampfer

Tracy

Hennekens

. Plasma concentration of C-reactive protein and risk of developing peripheral vascular disease. Circulation 1998; 97: 425–428.

16.

Tzoulaki

Murray

Lee

Rumley

Lowe

Fowkes

. C-reactive protein, interleukin-6, and soluble adhesion molecules as predictors of progressive peripheral atherosclerosis in the general population: Edinburgh Artery Study. Circulation 2005; 112: 976–983.

17.

Krzyzanowska

Mittermayer

Krugluger

. Asymmetric dimethylarginine is associated with macrovascular disease and total homocysteine in patients with type 2 diabetes. Atherosclerosis 2006; 189: 236–240.

18.

Hirsch

Haskal

Hertzer

. ACC/AHA 2005 Practice Guidelines for the management of patients with peripheral arterial disease (lower extremity, renal, mesenteric, and abdominal aortic): a collaborative report from the American Association for Vascular Surgery/Society for Vascular Surgery, Society for Cardiovascular Angiography and Interventions, Society for Vascular Medicine and Biology, Society of Interventional Radiology, and the ACC/AHA Task Force on Practice Guidelines (Writing Committee to Develop Guidelines for the Management of Patients With Peripheral Arterial Disease): endorsed by the American Association of Cardiovascular and Pulmonary Rehabilitation; National Heart, Lung, and Blood Institute; Society for Vascular Nursing; TransAtlantic Inter-Society Consensus; and Vascular Disease Foundation. Circulation 2006; 113: e463–654.

19.

NCHS Research Ethics Review Board (ERB) approval. Centers for Disease Control and Prevention (CDC), National Center for Health Statistics (NCHS). Available at: www.cdc.gov/nchs/nhanes/irba98.htm (accessed 15 March 2010).

20.

Centers for Disease Control and Prevention (CDC). National Center for Health Statistics (NCHS). National Health and Nutrition Examination Survey Questionnaire (or Examination Protocol, or Laboratory Protocol). Hyattsville, MD: US Department of Health and Human Services, Centers for Disease Control and Prevention, 1999–2004 [ www.cdc.gov/nchs/nhanes/nhanes_questionnaires.htm ].

21.

Levey

Coresh

Greene

. 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: 247–254.

22.

Looker

Dallman

Carroll

Gunter

Johnson

. Prevalence of iron deficiency in the United States. JAMA 1997; 277: 973–976.

23.

Crawford

Chacko

Kevil

Patel

. The red blood cell and vascular function in health and disease. Antioxid Redox Signal 2004; 6: 992–999.

24.

Comporti

Signorini

Buonocore

Ciccoli

. Iron release, oxidative stress and erythrocyte ageing. Free Radic Biol Med 2002; 32: 568–576.

25.

Yelinova

Glazachev

Khramtsov

Kudryashova

Rykova

Salganik

. Studies of human and rat blood under oxidative stress: changes in plasma thiol level, antioxidant enzyme activity, protein carbonyl content, and fluidity of erythrocyte membrane. Biochem Biophys Res Commun 1996; 221: 300–303.

26.

Minetti

Agati

Malorni

. The microenvironment can shift erythrocytes from a friendly to a harmful behavior: pathogenetic implications for vascular diseases. Cardiovasc Res 2007; 75: 21–28.

27.

Minetti

Malorni

. Redox control of red blood cell biology: the red blood cell as a target and source of prooxidant species. Antioxid Redox Signal 2006; 8: 1165–1169.

28.

Berliner

Rogowski

Aharonov

. Erythrocyte adhesiveness/aggregation: a novel biomarker for the detection of low-grade internal inflammation in individuals with atherothrombotic risk factors and proven vascular disease. Am Heart J 2005; 149: 260–267.

29.

Pande

Perlstein

Beckman

Creager

. Association of insulin resistance and inflammation with peripheral arterial disease: the National Health and Nutrition Examination Survey, 1999 to 2004. Circulation 2008; 118: 33–41.

30.

O’Hare

Glidden

Fox

Hsu

. High prevalence of peripheral arterial disease in persons with renal insufficiency: results from the National Health and Nutrition Examination Survey 1999–2000. Circulation 2004; 109: 320–323.

31.

Selvin

Erlinger

. Prevalence of and risk factors for peripheral arterial disease in the United States: results from the National Health and Nutrition Examination Survey, 1999–2000. Circulation 2004; 110: 738–743.

Red cell distribution width and risk of peripheral artery disease: Analysis of National Health and Nutrition Examination Survey 1999–2004

Abstract

Keywords

Introduction

Methods

Study population

Data collection

Ankle–brachial index measurement

ACC/AHA screening criteria for PAD

Laboratory measurements

Statistical analysis

Results

Discussion

RDW and ACC/AHA screening guidelines for a high-risk population

Conclusion

Footnotes

References