African American Ethnicity is Associated with Higher Plasma Levels of D-dimer in Adults with Hypertension

Summary

Background

African Americans with hypertension have higher cardiovascular morbidity and mortality than other ethnic groups. Plasma D-dimer, a fragment generated from fibrin during lysis of mature clot in vivo, is a predictor of adverse cardiovascular events.

Objective

We investigated whether plasma levels of D-dimer differ between African American (AA) and non-Hispanic white (NHW) adults with hypertension.

Methods

Participants included 933 AA (65±9 years, 72% women) and 821 NHW (61±9 years, 56% women) from the community. D-dimer was measured by an immunoturbidimetric assay. Multivariable regression analyses, stratified by sex, were performed to assess whether AA ethnicity was associated with D-dimer levels after adjustment for age, body mass index, total and high-density lipoprotein cholesterol, systolic blood pressure, diabetes, history of smoking, medication (statin and aspirin) use, lifestyle variables (physical activity, alcohol intake, and education), estimated glomerular filtration rate (eGFR), and a marker of inflammation, C-reactive protein (CRP).

Results

D-dimer levels were higher in AA men and women than in their NHW counterparts (mean±SD; men 255±199 vs. 190±182 ng/mL, P <0.001; women, 289±233 vs. 225±195 ng/mL, P <0.001). In both sexes, after adjustment for age, conventional risk factors, medication use, and lifestyle variables, AA ethnicity remained associated with higher D-dimer levels (P = 0.002 in men, P =0.006 in women). These associations remained significant after additional adjustment for eGFR and plasma CRP (P =0.003 in men, P <0.0001 in women).

Conclusions

Among adults with hypertension, AA ethnicity was independently associated with higher plasma levels of D-dimer.

Introduction

Disparity in cardiovascular health is among the most serious public health problems in the United States today. The basis for racial/ethnic disparities in cardiovascular disease appears to be complex and multifactorial and is not completely understood, although data from the Centers for Disease Control and Prevention [1] indicate a significantly higher prevalence of conventional cardiovascular risk factors in minority populations. African Americans with hypertension, especially, have higher cardiovascular morbidity and mortality than other ethnic groups [2]. There is considerable interest in indentifying biomarkers that may explain ethnic variation in susceptibility to cardiovascular diseases.

Plasma D-dimer, a fragment generated from fibrin during lysis of mature clot in vivo, is a sensitive marker of thrombus formation and its subsequent fibrinolysis [3, 4]. Higher baseline levels of D-dimer may be indicative of a procoagulant state favoring progression of atherosclerosis, and are associated with increased risk of myocardial infarction (MI), coronary death, ischemic stroke, and peripheral arterial disease (PAD) [5–17]. For example, in a large prospective cohort of 5201 men and women age 65 years and older, the relative risk of MI or coronary death was 2.5 fold higher for D-dimer values above the median. This relationship was independent of hypertension, diabetes, smoking status, ethnicity, body mass index (BMI), levels of plasminogen-activator inhibitor 1, and C-reactive protein (CRP) [17].

Although ethnic differences in plasma D-dimer levels have been previously described [18, 19], adjustment has been variably reported for potential confounders, particularly renal function and systemic inflammation. The aim of the present study was to investigate ethnic differences in plasma D-dimer levels between hypertensive African American and non-Hispanic white adults, and, if present, whether such a difference could be explained by age, conventional risk factors, medication use, lifestyle variables, and renal function. In addition, because D-dimer levels are associated with an acute phase response, we also adjusted for plasma levels for an established marker of inflammation, CRP.

Methods

Study population

The study was part of the Proteomic Markers of Arteriosclerosis Study which is investigating the association of multiple markers in various etiologic pathway of vascular disease with several phenotypes of arteriosclerosis [20, 21]. Participants included hypertensive adults participating in a multicenter community-based study of hypertensive sibships that aims to identify genes influencing blood pressure (BP) levels and development of target organ damage due to hypertension. Sibships with at least two members diagnosed with hypertension before age 60 years were included. Recruitment and subject characteristics in the initial phase of the study (September 1995 to June 2001) have been described previously. For the present study, we excluded normotensive participants (n = 311 in Jackson and n = 305 in Rochester) and 15 participants with D-dimer levels >1500 ng/mL, 55 participants who were on warfarin and 1 participant who had estimated glomerular filtration rate (eGFR) >300 mg/dL. We also excluded 125 participants with missing D-dimer level, leaving a final study sample of 1749 subjects. The study was approved by the Institutional Review Boards of the Mayo Clinic, Rochester, MN and University of Mississippi Medical Center, Jackson, MS. Written informed consent was obtained from each participant.

Height was measured by stadiometer, weight by electronic balance, and BMI was calculated as weight in kilograms divided by the square of height in meters. Resting systolic BP and diastolic BP were measured by trained study coordinators using random zero sphygmomanometer and cuffs appropriate for arm size. Three readings were taken in the right arm after the subject had rested in the sitting position for at least 5 minutes; the last two readings were averaged for the analyses. Diagnosis of hypertension was established based on BP levels measured at the study visit (≥140/90 mmHg) or a prior diagnosis of hypertension and current treatment with antihypertensive medications. Diabetes was considered present if the subject was being treated with insulin or oral agents or had a fasting glucose level ≥126 mg/dL. “Ever” smoking was defined as having smoked more than 100 cigarettes. Each prescription drug recorded at the study visit was assigned a code number corresponding to the first six digits of the Medi-Span Generic Product Identifier. This number identifies pharmacologically equivalent drug products and was used to categorize agents with a similar therapeutic action.

Blood was drawn by venipuncture after an overnight fast. Serum total cholesterol and high-density lipoprotein (HDL) cholesterol were measured by standard enzymatic methods. eGFR was calculated using the Modification of Diet in Renal Disease (MDRD) equation as previously described [22].

Lifestyle variables

We constructed a physical activity scale by considering the responses to questions on how many hours per day of heavy activity, moderate activity, slight activity, and sedentary activity the participant engaged in. Specifically, the physical activity score was derived as follows: 3*heavy + 2*moderate + light (hours). Alcohol intake was quantified as ounces of alcohol per month and was estimated from responses to questions on the frequency (servings per week or servings per month), and ounces per serving of beer, wine, hard liquor, wine cooler, and sake. The ounces per month of each of these categories were then multiplied by 5%, 12%, 40%, 8%, and 15%, respectively, and summed to give ounces of alcohol per month. This variable was highly skewed and was transformed to a log scale after adding 1, for analysis purposes.

Plasma levels of D-dimer and CRP

D-dimer was measured immunoturbidimetrically using STA-Liatest D-Dimer reagents (Diagnostica Stago, Asnieres, France) on a STA Compact (Diagnostica Stago, Asnieres, France). Intra-assay coefficients of variation (CVs) were 24% and 3% at 100 and 1071 ng/mL, and inter-assay CVs were 15 and 4% at 245 and 2060 ng/mL. Plasma CRP levels were measured by a highly sensitive immunoturbidimetric assay. Inter-assay CVs were: 14%, 3.2%, 3.4%, and 3.6% at 0.33, 1.05, 9.07, and 23.8 mg/dL, respectively.

Statistical methods

Statistical analyses were carried out using SAS v 9.1 (SAS Institute, Cary NC) software package. Descriptive statistics are given as mean ± standard deviation or frequency and percent, separately in men and women in each ethnic group. Alcohol intake, eGFR, CRP, and D-dimer levels were log transformed due to their skewed distribution. In each sex and ethnic group, we assessed the correlation of D-dimer with the following: 1) age and conventional risk factors (total and HDL cholesterol, systolic BP, smoking, diabetes, and BMI); 2) use of medications (statins and aspirin); 3) lifestyle variables (physical activity, alcohol intake, and education); 4) a measure of renal function, eGFR, and a marker of systemic inflammation, CRP.

In each sex, we constructed stepwise multivariable linear regression models that adjusted for conventional risk factors and other potential confounding variables to assess whether African American ethnicity was independently associated with plasma D-dimer levels. Adjustments were performed for a) age; b) + BMI, systolic BP, smoking history, diabetes, total and HDL cholesterol, medications (statins and aspirin) use, previous history of MI or stroke, and lifestyle variables (physical activity, alcohol intake, and education); c) + renal function (eGFR); and finally d) + plasma CRP. Because of the presence of sibships in the sample, population-averaged generalized estimating equations were used to assess the possible impact of familial correlations on the relationships between predictor and outcome variables [23]. A two-sided P-value of <0.05 was deemed statistically significant.

Results

Subject characteristics are shown in Table 1. Significant ethnic differences were present in both men and women for conventional risk factors, medication use, and lifestyle variables. Plasma D-dimer levels were higher in African American men and women than in their non-Hispanic white counterparts (mean±SD; men 255±199 vs. 190±182 ng/mL, P <0.001; women, 289±233 vs. 225±195 ng/mL, P <0.001) (Fig 1) as the distribution of D-dimer levels was different with a significant shift in African Americans (Fig 2).

Table 1.

Participants characteristics and ethnic differences

	Women			Men
	AA (n =679)	NHW (n =455)	P	AA (n =254)	NHW (n =357)	P
Age, years	64.7±8.8	61.0±9.2	<.0001	65.4±8.5	60.8±9.3	<.0001
BMI, kg/m2	32.7±7.0	31.7±6.9	0.002	29.5±4.8	31.1±5.1	0.003
Total cholesterol, mg/dL	205.9±42.1	203.2±33.0	0.17	191.8±42.0	187.4±30.5	0.22
HDL cholesterol, mg/dL	60.6±17.6	55.7±14.8	<.0001	50.1±16.4	44.7±11.7	<.0001
Systolic BP, mm Hg	143.6±21.3	136.6±17.8	<.0001	140.4±20.4	133.4±15.6	<.0001
Diastolic BP, mm Hg	79.2±11.3	73.6±9.5	<.0001	82.5±11.7	76.0±9.6	<.0001
Ever smoker, n (%)	208 (30.6)	177 (38.9)	0.001	166 (65.4)	212 (59.4)	0.16
Diabetes mellitus, n (%)	231 (34.0)	80 (17.1)	<.0001	82 (32.3)	67 (18.8)	0.0005
Previous history of MI or stroke, n (%)	66 (9.7)	41 (9.0)	0.59	44 (17.3)	62 (17.4)	0.70
Statin use, n (%)	134 (19.7)	123 (27.0)	0.001	60 (23.6)	151 (42.3)	<.0001
Aspirin use, n (%)	227 (33.4)	200 (44.0)	0.0004	115 (45.3)	207 (58.0)	0.0009
Physical activity score	9.5±3.2	12.5±4.8	<.0001	10.0±4.2	13.8±5.3	<.0001
Alcohol intake, oz/month	0.8±4.0	2.9±6.3	<.0001	3.2±7.9	9.8±15.0	<.0001
Education, years	12.1±3.4	13.3±2.0	<.0001	11.5±4.0	13.4±2.6	<.0001
eGFR, mg/dL	100.4±32.5	84.4±24.3	<.0001	90.4±28.3	82.1±21.3	<.0001
C-reactive protein, mg/dL	7.4±13.3	5.8±7.7	<.0001	5.0±7.6	2.8±3.5	<.0001
D-dimer, ng/mL	289.7±233.0	224.7±195.3	<.0001	252.7±197.1	190.3±182.6	0.0004

AA, African American; NHW, non-Hispanic white; BMI, body mass index; HDL, high density lipoprotein; BP, blood pressure; MI, myocardial infarction; eGFR, estimated glomerular filtration rate.

Figure 1.

Mean±SD of plasma D-dimer levels in each gender/ethnic group.

*P value <0.001

Figure 2.

Distribution of plasma D-dimer levels in each gender/ethnic group. A. Women; B. Men.

Spearman rank correlations between D-dimer and select variables are shown in Table 2. Among women, older age, higher BMI, higher systolic BP, lower diastolic BP, lower physical activity score, lower level of education, lower eGFR, and higher CRP levels, correlated with higher D-dimer levels (Table 2). Higher levels of serum cholesterol in African American women and the presence of diabetes in non-Hispanic white women were also correlated with higher D-dimer levels. Among men, older age, higher systolic BP, lower diastolic BP, the presence of diabetes, Previous history of MI or stroke, use of aspirin, lower level of education, lower eGFR, and higher CRP levels, correlated with higher D-dimer levels (Table 2). Lower physical activity score in African American men and lower levels of serum cholesterol and smoking history in non-Hispanic white men were also correlated with higher D-dimer levels.

Table 2.

Spearman correlations between plasma D-dimer levels and select variables

	Women		Men
	AA (n =679)	NHW (n =455)	AA (n =254)	NHW (n =357)
Risk Factors	r P value	r P value	r P value	r P value
Age, years	0.32 <.0001	0.31 <.0001	0.45 <.0001	0.34 <.0001
BMI, kg/m2	0.08 0.040	0.23 <.0001	−0.06 0.30	0.04 0.43
Total cholesterol, mg/dL	0.13 0.0009	−0.06 0.24	0.03 0.68	−0.15 0.005
HDL cholesterol, mg/dL	0.04 0.31	−0.09 0.044	−0.03 0.65	−0.05 0.32
Systolic BP, mm Hg	0.08 0.028	0.13 0.007	0.15 0.016	0.18 0.0006
Diastolic BP, mm Hg	−0.08 0.043	−0.11 0.018	−0.12 0.047	−0.11 0.036
Ever smoker	−0.03 0.41	−0.04 0.35	0.03 0.61	0.12 0.028
Diabetes mellitus	−0.03 0.40	0.16 0.0009	0.13 0.037	0.17 0.001
Previous history of MI or stroke	0.05 0.20	0.05 0.25	0.13 0.046	0.16 0.002
Statin use	−0.01 0.89	0.03 0.46	−0.08 0.22	0.04 0.47
Aspirin use	0.04 0.36	0.09 0.047	0.15 0.018	0.15 0.004
Physical activity score	−0.22 <.0001	−0.10 0.037	−0.28 <.0001	−0.04 0.42
Alcohol intake, oz/month	−0.06 0.079	−0.06 0.18	−0.03 0.68	−0.08 0.10
Education, years	−0.16 <.0001	−0.11 0.019	−0.25 <.0001	−0.19 0.0002
eGFR, mg/dL	−0.23 <.0001	−0.27 <.0001	−0.16 0.010	−0.15 0.008
C-reactive protein, mg/dL	0.11 0.003	0.19 <.0001	0.28 <.0001	0.15 0.005

AA, African American; NHW, non-Hispanic white; BMI, body mass index; HDL, high density lipoprotein; BP, blood pressure; MI, myocardial infarction; eGFR, estimated glomerular filtration rate.

After adjustment for age, conventional risk factors, medication use, and lifestyle variables, African American ethnicity was associated with higher D-dimer levels in both men (P =0.002) and women (P =0.006). Additional adjustment for eGFR, accentuated this association in both men (P =0.0002) and women (P <0.0001) (Table 3). Further adjustment for CRP levels slightly attenuated the association in men (P =0.003), but not in women (P <0.0001). In addition, in men, older age, higher systolic BP, the presence of diabetes, lower eGFR, and higher CRP were associated with higher plasma D-dimer levels. In women, older age, higher BMI, higher systolic BP, lower eGFR, and higher CRP were associated with a higher plasma D-dimer (Table 4). Neither lifestyle variables nor smoking were independently associated with plasma levels of D-dimer in men or women, in the multiple regression models.

Table 3.

Association between African American ethnicity with plasma D-dimer levels

	Women, n = 1134			Men, n = 611
	Estimate	SE	P	Estimate	SE	P
Model 1	0.18	0.04	<.0001	0.20	0.05	0.0002
Model 2	0.12	0.04	0.006	0.17	0.05	0.002
Model 3	0.22	0.05	<.0001	0.20	0.05	0.0002
Model 4	0.21	0.05	<.0001	0.16	0.05	0.003

eGFR, CRP, and D-dimer were natural log-transformed.

The co-efficients represent the change in log D-dimer for each unit change in the continuous predictor variable or for the presence of the specified condition of the categorical predictor variable.

Model 1: Adjusted for age

Model 2: Adjusted for age, BMI, total and HDL cholesterol, systolic BP, diabetes, history of smoking, medication (statin and aspirin) use, previous history of MI or stroke, and lifestyle variables (physical activity, alcohol intake, and education)

Model 3: Adjusted for above variables and eGFR

Model 4: Adjusted for above variables and CRP

BMI, body mass index; BP, blood pressure; HDL, high-density lipoprotein; MI, myocardial infarction; eGFR, estimated glomerular filtration rate; CRP, C-reactive protein.

Table 4.

Variables associated with plasma D-dimer levels: multi-variable regression models.

	Women, n =1134			Men, n =611
	Estimate	SE	P	Estimate	SE	P
Age, years	0.02	0.002	<.0001	0.02	0.003	<.0001
BMI, kg/m2	0.01	0.003	<.0001	-	-	-
Systolic BP, mm Hg	0.002	0.001	0.059	0.003	0.001	0.040
Diabetes	−0.09	0.04	0.041	0.16	0.06	0.008
Physical activity score	−0.01	0.005	0.087	-	-	-
Log eGFR	−0.45	0.07	<.0001	−0.22	0.10	0.031
Log C-reactive protein	0.08	0.02	0.0001	0.12	0.03	<.0001
African American ethnicity	0.21	0.05	<.0001	0.16	0.05	0.003

eGFR, C-reactive protein, and D-dimer were natural log-transformed.

The co-efficients represent the change in log D-dimer for each unit change in the continuous predictor variable or for the presence of the specified condition of the categorical predictor variable.

BMI, body mass index; BP, blood pressure; eGFR, estimated glomerular filtration rate.

Discussion

The main finding of this study is that among adults with hypertension, plasma D-dimer levels are significantly higher in African American men and women than in their non-Hispanic white counterparts. The differences could not be explained by age, conventional risk factors, medication use, lifestyle factors, renal function, and systemic inflammation, suggesting that unmeasured environmental and genetic factors may contribute. These results suggest that fibrin turnover is higher in hypertensive African Americans than in their non-Hispanic white counterparts and may be a possible explanation for increased cardiovascular morbidity and mortality among African Americans with hypertension.

Cushman et al [18] reported higher D-dimer levels in non-whites (most non-whites were African Americans) than in whites in a nested case control study of 923 participants (78% were whites) from the Longitudinal Investigation of Thrombosis Etiology (LITE) study. However the authors did not report whether the association was independent of conventional cardiovascular risk factors. In a random sample of 1727 community-dwelling elderly persons (age >72 years), Pieper et al [19] reported that average plasma D-dimer level was nearly 40% higher in African Americans than in white participants. This difference was not substantially attenuated in multivariable analyses adjusting for demographic and socioeconomic variables. Our study shows higher D-dimer levels in hypertensive African Americans than in their non-Hispanic white counterparts. In addition, we report that this association was independent of conventional risk factors, medication use, lifestyle factors, renal function, and systemic inflammation.

We found D-dimer levels increased with age likely due to higher fibrinogen concentrations in the elderly, reduced renal function, increased fibrin generation, occult disease, higher prevalence of risk factors, and increased inflammation [19, 24]. We also observed that female sex was independently associated with higher D-dimer levels (analyses not shown). This is consistent with the results of a prospective study of 1000 healthy adults by Hughes et al [25] that found age, female sex (independent of hormone therapy), the use of hormone therapy, thrombophilic state, presence of co-morbid conditions, and BMI, as the major predictors of plasma D-dimer levels. However, in the present study, we did not find estrogen use to be associated with D-dimer levels (analyses not shown). Higher D-dimer levels were independently associated with lower eGFR in both men and women, likely due to decreased clearance by the kidney [26, 27]. Renal insufficiency is associated with increased levels of inflammatory and procoagulant biomarkers, including CRP, interleukin-6 (IL-6), fibrinogen, and D-dimer [28]. In the present study, in spite of higher eGFR in African Americans than in non-Hispanic whites, D-dimer levels were higher in former.

The association between D-dimer with plasma CRP was significant and independent of potential confounders. Local fibrin formation and lysis are part of the inflammatory response and it has been shown that D-dimer levels increase during an acute phase response [29, 30]. Systemic inflammation by causing low-grade activation of coagulation cascade may therefore increase levels of D-dimer. Multiple inflammatory cytokines such as tumor necrosis factor (TNF)- α and IL-6 may directly activate the coagulation system [31, 32]. However, the association may be bi-directional as D-dimer has been shown to stimulate neutrophil and monocyte activation, secretion of cytokines, including IL-6 and interleukin-1, and thus indirectly stimulate the production of acute phase proteins [33–36].

A strength of the present study is the use of a uniform protocol in the two ethnic groups, including, questionnaires, anthropometric measurements, and D-dimer measurements that were made in a single core laboratory. The gold standard for the measurement of the plasma D-dimer tests is an enzyme-linked immunosorbent assay. In the present study, we used the STA Liatest D-Dimer reagents on a STA Compact. Waser et al [37] showed good agreement (kappa >0.81) and correlation (r >0.95, P <0.0001) between the STA Liatest and the classical enzyme-linked immunosorbent assays.

Clinical implications

Mortality rates in United States differ substantially according to ethnicity [2]. Hypertension mediated mortality from target organ damage and stroke is higher in African Americans and this is not fully explained by higher burden of conventional cardiovascular risk factors, including diabetes and hypertension in this ethnic group [38]. D-dimer levels have been consistently associated with presence and severity of cardiovascular diseases [39–42]. Whether plasma D-dimer contributes to the ethnic difference in cardiovascular disease needs further investigation. Measurement of D-dimer levels may be helpful in assessing the risk of cardiovascular disease, especially in African Americans. Whether lowering the elevated levels of D-dimer in susceptible population groups by lifestyle/pharmacologic means will lower cardiovascular morbidity and mortality will need confirmation by randomized controlled trials.

Conclusion

In conclusion, our results indicate that plasma levels of D-dimer are higher in hypertensive African Americans than in their non-Hispanic white counterparts. The ethnic differences are independent of age, conventional risk factors, medication use, lifestyle factors, renal function, and systemic inflammation. Unmeasured environmental and genetic factors may partially contribute to the ethnic difference in plasma D-dimer levels. Further studies are needed to elucidate the pathophysiological basis for this observation. The higher D-dimer in hypertensive African Americans than in their non-Hispanic white counterparts may be a potential mechanism of increased cardiovascular risk in the former. Further investigation is needed to confirm this hypothesis.