Sex Differences in the Endothelial Function of Untreated Hypertension


Andrew Sherwood, PhD, Box 3119, Duke University Medical Center, Durham, NC 27710


J Clin Hypertens (Greenwich). 2012;14:228–235. ©2012 Wiley Periodicals, Inc.

Vascular endothelial dysfunction is associated with increased risk for adverse cardiovascular (CV) events. However, less is known about sex differences in the endothelial function of untreated hypertensive individuals. The purpose of this study was to assess endothelial function in women and men with untreated hypertension. Ninety participants (35 women, 55 men), aged 40 to 60 years (mean age, 46.1±8.2 years), with untreated stage 1 hypertension (systolic blood pressure 140–159 mm Hg and/or diastolic blood pressure 90–99 mm Hg) underwent brachial artery endothelial-dependent flow-mediated dilation and endothelial-independent glyceryl trinitrate dilation. Women had a smaller flow-mediated dilation response than men (adjusted mean±standard error of the mean [SEM]; 1.8±0.6% vs 3.9±0.4%, P=.036), adjusting for baseline arterial diameter (P=.004), age (P=.596), ethnicity (P=.496), log shear stress ratio (P<.001), body mass index (P=.009), 24-hour diastolic blood pressure (P=.169), high-density lipoprotein (P=.225), log creatinine (P=.927), and log physical activity (P=.682). Glyceryl trinitrate dilation did not differ by sex in adjusted models. Women between the ages of 40 and 60 years with untreated stage 1 hypertension exhibited a greater impairment of endothelial function compared with their male counterparts. These findings raise the possibility that female sex may impart a greater risk of CV events in patients with untreated stage 1 hypertension potentially due to poorer endothelial function.

Sex differences are often observed in the prevalence, presentation, and outcome of the various cardiovascular (CV) conditions.1–4 Hypertension is an established CV risk factor. The Sibrafiban vs Aspirin to Yield Maximum Protection From Ischemic Heart Events Post-Acute Coronary Syndromes (SYMPHONY)4 and the Effect of Potentially Modifiable Risk Factors Associated With Myocardial Infarction in 52 Countries (INTERHEART)5,6 studies have reported a higher prevalence of hypertension in women compared with men presenting with acute coronary syndromes4 and a stronger association between hypertension and risk of first myocardial infarction in women than in men,5 especially among women younger than 60 years.6 Taken together, these findings suggest that hypertension may confer greater CV risk for women than for men. The reasons for these sex differences are not entirely clear.

Endothelial dysfunction, as indicated by impaired flow-mediated dilation (FMD), is characteristic of hypertension7–9 and associated with poor outcomes in hypertensive patients.10 The effect of sex on FMD has been examined in various populations and has been reported inconsistently in the literature.11–15 To our knowledge, sex differences in FMD in untreated hypertensive individuals have not been examined. Given that hypertension may be a stronger CV risk factor for women, especially among women younger than 60 years,6 the purpose of this study was to compare FMD in women and men with untreated hypertension, in order to explore the possibility that hypertension in women may be accompanied by a more marked impairment of FMD than for men.



The study included 90 participants (35 women, 55 men), aged 40 and 60 years (mean age, 46.1±8.2 years), with stage 1 hypertension (systolic blood pressure [SBP] 140–159 mm Hg and/or diastolic blood pressure [DBP] 90–99 mm Hg) who had not been treated with antihypertensive medications in the preceding 12 months. The present study sample was drawn from a larger study population that also included participants with prehypertension that was conducted at Duke University Medical Center from 2004 through 2007.14 As previously described,14 exclusion criteria were body mass index (BMI) >35 kg/m2; diabetes mellitus; pacemaker; atrial fibrillation; myocardial infarction, percutaneous coronary intervention, or coronary artery bypass graft surgery within 6 months of enrollment; heart failure; severe uncorrected primary valvular disease; uncorrected thyroid heart disease; oral contraceptive use; pregnancy; hormone replacement therapy; alcohol or drug abuse within 12 months; renal or hepatic dysfunction; dementia; inability to comply with the assessment procedures; or unwillingness to provide informed consent. Individuals with previously diagnosed obstructive sleep apnea or identified by the Berlin Questionnaire15 as being at high risk for sleep apnea syndrome were excluded given the association between obstructive sleep apnea and impaired endothelial function.16

Participants were recruited by advertisements in the Piedmont region of North Carolina, which includes a general population of more than 1 million who reside within a 30-mile radius of Duke University Medical Center. The study protocol was approved by the institutional review board at Duke University Medical Center. All eligible individuals provided written informed consent prior to participation in the study.

Demographic, Anthropometric, and Biochemical Assessment

Data were collected on age, sex, height, weight, BMI, ethnicity, cigarette smoking, and alcohol consumption. Blood samples were collected in the morning after an overnight fast. Specimens were analyzed by Labcorp using automated enzymatic assays for glucose, total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), and triglycerides; immunochemiluminometric assay for high-sensitivity C-reactive protein (CRP); and electrochemiluminescence immunoassay for follicle-stimulating hormone (FSH) among women. The self-reported date of last menstrual period was used to determine premenopausal women’s menstrual cycle phase, with 50% of women determined to be in the luteal phase, 33% in the follicular phase, and 17% in the menstrual phase.

BP Assessment

Clinic blood pressure (BP) was determined on 3 visits, each approximately 1 week apart. On each occasion, BP was measured after 5 minutes seated in a quiet, temperature-controlled room. Four BP readings, each 2 minutes apart, were taken using a mercury sphygmomanometer and stethoscope. The last 3 readings for each visit, a total of 9 readings, were used to represent mean clinic SBP and DBP.

On 3 weekdays, each approximately 1 week apart. 24-hour ambulatory BP (ABP) was assessed with the Oscar 2 ambulatory BP monitor (Suntech Medical, Raleigh, NC).17 ABP readings were reviewed and artifact edited as previously described.18

Physical Activity Assessment

Daytime physical activity was assessed with the Mini-Mitter Actiwatch wristwatch-style actigraphy (Mini-Mitter, Sunriver, OR). The actigraph was worn on the wrist for each of the 3 ABP monitoring days and averaged to provide a mean physical activity score.

Vascular Endothelial Function Assessments

Vascular studies of the brachial artery were assessed in the morning, following an overnight fast. Longitudinal B-mode images of the brachial artery, in the region 4 to 6 cm anterior to the antecubital fossa, were recorded and stored digitally using a 7- to 11-MHz linear-array transducer and Aspen ultrasound system. Images were captured: (1) after 10 minutes of supine rest; (2) during the first 120 seconds of reactive hyperemia, achieved by inflation of a pneumatic occlusion forearm cuff to supra-systolic pressure (∼200 mm Hg) for 5 minutes; (3) after a second rest period of 15 minutes; and (4) during the 3- to 6-minute period after administration of 400 μg sublingual glyceryl trinitrate (GTN), used to induce nonendothelial-dependent arterial dilation. End-diastolic images were stored and arterial diameters measured as the distance between the proximal and distal arterial wall intima-media interfaces using PC-based software (Brachial Analyzer, version 5.0; Medical Imaging Applications LLC, Iowa City, IA). Peak FMD response was assessed from 10 to 120 seconds post-deflation of the cuff, with peak arterial diameter quantified using polynomial curve fitting. Endothelial-dependent FMD and endothelial-independent GTN dilation (GTND) were expressed as absolute diameter change (maximum arterial diameter – baseline arterial diameter) and percent increase in arterial diameter ([maximum arterial diameter – baseline arterial diameter/baseline arterial diameter] × 100%), which is the index that has been adopted most widely.19 The percent change index may result in bias towards greater FMD in smaller arteries and in arteries experiencing greater shear stress.19,20 Therefore, FMD was examined using baseline arterial diameter and shear stress as covariates. Shear stress was calculated as using the formula described by Mitchell and colleagues21 (shear stress = 8 ×μ× V/baseline arterial diameter) at baseline and during reactive hyperemia, with μ representing assumed blood viscosity (0.035 dyne × s/cm2), V representing velocity (cm/s), and baseline arterial diameter in cm. Shear stress ratio was calculated by dividing hyperemic shear stress/baseline shear stress. Shear stress ratio was used as an index of the hyperemic shear stress stimulus for FMD.21

All studies were completed by the same sonographer who has more than 10 years of experience in performing vascular assessments in our laboratory. We have previously reported22 that our laboratory’s repeated FMD assessments on 20 healthy women and men during 2 consecutive days showed a correlation of r=−0.81 (P<.001), a mean difference of 0.6%±2.7%, and a coefficient of variation of 26.7%.

Statistical Analysis

Normality statistics (skew, kurtosis, Kolmogorov–Smirnov), histograms, boxplots, and normal probability plots were examined to assess normality. Non-normal variables underwent log base-e transformation (triglycerides, serum creatinine, creatinine clearance, baseline shear stress, hyperemic shear stress, shear stress ratio, and physical activity), square root transformation (GTND), or trimming at the 95th percentile to prevent excessive influence of outliers (glucose); CRP remained non-normally distributed and nonparametric testing was used. A total of 54% of participants did not consume alcohol, therefore the variable for the number of alcoholic beverages consumed in the past week was dichotomized into 0 alcoholic beverages and ≥1 alcoholic beverages.

Data are expressed as mean±standard deviation for continuous normally distributed variables or median (interquartile range) for non-normally distributed continuous variables. t tests (for normal distributions) or the Wilcoxon rank sum test (for non-normal distributions) were used to compare sex differences in continuous variables. Chi-square tests were used to assess differences in categorical variables between female and male participants. Analysis of covariance (ANCOVA) tests were used to evaluate the effect of sex on FMD and GTND. Based on previous research, covariates included in the planned model were baseline arterial diameter,7 age,23 ethnicity,7 and shear stress (for FMD analysis).21 In secondary confirmatory analysis, the presence of a univariate correlation between the variable of interest and FMD and GTND was performed. Variables were entered into the planned model if the univariate correlation between each variable and outcome variable (FMD or GTND) was P≤.2. The equal variance of analysis of variance assumption was confirmed with nonsignificant Levene’s tests. Women were classified as postmenopausal if they had amenorrhea for at least 1 year and serum FSH levels >40 IU/L. Statistical analyses were conducted using the SAS 9.2 system (SAS Institute, Cary, NC) with significance set at P=.05.


Sample Characteristics

Table I describes the participant characteristics. HDL cholesterol was higher among female participants (n=35) when compared with male participants (n=55). Men exhibited higher DBP, triglycerides, and glucose. There were 15 current smokers (6 women and 9 men) and 18 former smokers (7 women and 11 men). Women and men did not differ in the number of years they currently smoked (22.3±14.2 years vs 18.0±13.4 years, P=.535), formerly smoked (14.8±8.2 years vs 10.2±5.5 years, P=.169), or the number of years since they quit smoking (12.2±8.1 years vs 15.7±7.3 years, P=.353). Two men and 2 women were taking cholesterol-lowering medications. Ten women (28.6%) were classified as postmenopausal. Postmenopausal women were older (53.8±4.1 years vs 44.5±6.6 years, P<.001) and less likely to be African American (30% vs 80%, P=.015) than premenopausal women. Unadjusted vascular assessments of the brachial artery are presented in Table II.

Table I.   Participant Characteristics
 Sample (N=90)Women (n=35)Men (n=55)P Value
  1. Abbreviations: BMI, body mass index; CRP, C-reactive protein; DBP, diastolic blood pressure; HDL, high-density lipoprotein; LDL, low-density lipoprotein; SBP, systolic blood pressure. For continuous variables, data are expressed as mean±standard deviation or median (interquartile range) for normally distributed and non-normally distributed data, respectively. For categorical variables, data are expressed as relative frequencies (percentages). Bold P values indicate statistical significance. aTrimmed 95%.

Age, y46.1±8.247.2±7.345.5±8.8.343
BMI, kg/m228.8±3.528.1±4.129.2±3.1.178
Clinic SBP, mm Hg143±7143±7143±7.860
Clinic DBP, mm Hg92±490±593±4.005
24-h SBP, mm Hg134±11134±12135±10.600
24-h DBP, mm Hg81±881±982±7.785
24-h HR, beats per min76±979±974±8.010
Total cholesterol, mg/dL196.9±34.0194.4±33.7198.6±34.4.572
HDL, mg/dL54.7±16.562.8±19.149.5±12.3<.001
LDL, mg/dL117.5±27.4112.7±29.4120.6±25.8.189
Triglycerides, mg/dL99.5 (71.0–150.0)76.0 (55.0–117.0)116.0 (85.0–183.0)<.001
Glucosea, mg/dL94.1±9.990.2±9.996.6±9.1.002
CRP, mg/L1.3 (0.6–3.5)1.4 (0.6–3.5)1.2 (0.6–3.7).928
Serum creatinine, mg/dL0.9 (0.8–1.0)0.7 (0.7–0.8)1.0 (0.9–1.1)<.001
Creatinine clearance, mL/min113.9 (100.4–137.4)104.0 (88.9–131.9)119.8 (104.3–143.1).036
Ethnicity (% African American)55.665.749.1.122
Current smoker, %17.217.617.0.936
Former smoker, %20.720.620.8.985
Current or former smoker, %37.938.237.7.963
≥1 Alcoholic beverage(s) in the past week, %46.142.948.2.645
Physical activity, units94209.4 (72997.3–127092.1)104019.4 (81308.0–141331.4)84710.2 (62760.9–111016.4).011
Table II.   Mean Values of Vascular Assessments
 Sample (N=90)Women (n=35)Men (n=55)P Value
  1. Abbreviations: FMD, flow-mediated dilation; GTND, glyceryl trinitrate dilation. Data are expressed as mean±standard deviation or median (interquartile range) for normally distributed and non-normally distributed data, respectively. Bold P values indicate statistical significance.

FMD assessments
 Baseline artery diameter, mm4.5±0.73.9±0.44.9±0.6<.001
 Absolute FMD, mm0.13±0.130.09±0.10.16±0.1.014
 FMD, %3.0±2.92.4±2.73.4±3.0.128
 Baseline shear stress, dyne/cm27.5 (6.0–9.7)8.1 (6.6–9.8)7.1 (5.7–9.5).129
 Hyperemic shear stress, dyne/cm239.7 (28.6–66.7)45.0 (27.2–74.9)39.0 (28.6–56.8).350
 Shear stress ratio6.1 (3.9–8.3)6.1 (3.7–9.3)6.1 (3.9–7.9).983
GTND assessments
 Baseline artery diameter, mm4.4±0.73.8±0.54.8±0.5<.001
 Absolute GTND, mm0.7±0.20.8±0.20.7±0.2.252
 GTND, %15.7 (13.0–19.9)18.7 (15.0–25.4)15.5 (11.3–17.4)<.001

Flow-Mediated Dilation by Sex

The effect of sex on FMD was evaluated with ANCOVA (Table IIIa). FMD was significantly impaired in female participants compared with male participants (adjusted mean±SEM; 1.4±0.6% vs 4.0±0.4%, P=.002). A sex-by-race interaction was not significant when included in the model.

Table III.   The Effect of Sex on FMD in (a) Planned and (b) Confirmatory ANCOVA Models
 dfSSMean SquareFP ValuePartial η2
  1. Abbreviations: BMI, body mass index; DBP, diastolic blood pressure; FMD, flow-mediated dilation; HDL, high-density lipoprotein; Log, log base-e; SS, sum of squares.

Corrected total89747.2    
 dfType III SSMean SquareFP ValuePartial η2
Baseline artery diameter141.441.47.8.0070.08
Log shear stress ratio1226.2226.242.4<.0010.34
 dfSSMean SquareFP ValuePartial η2
Corrected total85740.2    
 dfType III SSMean SquareFP ValuePartial η2
Baseline artery diameter143.
Log shear stress ratio1252.3252.352.3<.0010.41
24-h DBP19.
Log creatinine10.
Log physical activity10.<0.01

In secondary confirmatory analysis (Table IIIb), BMI, HDL cholesterol, 24-hour DBP, log serum creatinine, and log physical activity correlated with FMD at P≤.2 (Table IV) and were entered into the planned ANCOVA model. In the final model (Figure), women continued to display a smaller FMD response than men (adjusted mean±SEM; 1.8±0.6% vs 3.9±0.4%, P=.036). A sex-by-race interaction was not significant when included in the model. FMD did not differ by menopausal status in the planned and fully adjusted models.

Table IV.   Correlates of FMD and GTND in Women and Men With Untreated Stage 1 Hypertension
 rP ValuerP Value
  1. Abbreviations: ANCOVA, analysis of covariance; BMI, body mass index; CRP, C-reactive protein; DBP, diastolic blood pressure; FMD, flow-mediated dilation; GTND, glyceryl trinitrate dilation; HDL, high-density lipoprotein; HR, heart rate; LDL, low-density lipoprotein; Log, log base-e; n/a, not applicable; SBP, systolic blood pressure; sqrt, square root. Pearson’s r correlations are presented for continuous variables normally distributed variables. For categorical and non-normally distributed variables, Spearman’s r correlations are presented. aTrimmed 95%.

Female/Male (0/1)0.16.140−0.32.002
Age, y−0.12.263−0.03.750
BMI, kg/m20.17.108−0.26.012
Clinic SBP, mm Hg−0.12.279−0.13.220
Clinic DBP, mm Hg−0.16.1310.02.864
24-h SBP, mm Hg−0.05.619−0.18.091
24-h DBP, mm Hg−0.16.142−0.08.458
24-h HR, beats per min−0.05.6720.17.109
Total cholesterol, mg/dL−0.07.528−0.13.214
HDL, mg/dL−0.16.1230.29.007
LDL, mg/dL0.02.825−0.14.188
Log triglycerides, mg/dL0.03.791−0.37<.001
Glucose, mg/dLa0.10.351−0.27.009
CRP, mg/L0.004.973−0.12.243
Log serum creatinine0.14.185−0.20.066
Log creatinine clearance0.06.550−0.23.030
White/African American ethnicity (0/1)−0.01.8980.11.286
Current smoker (no/yes, 0/1)−0.05.6640.02.888
Former smoker (no/yes, 0/1)0.03.814−0.01.926
Current or former smoker (no/yes, 0/1)−0.02.835−0.002.983
≥1 Alcoholic beverage(s) in the past week−0.06.5500.16.146
Log physical activity, units−0.15.166−.01.962
Baseline artery diameter, mm−0.08.457−.60<.001
Log baseline shear stress0.20.064n/an/a
Log hyperemic shear stress0.61<.0001n/an/a
Log shear stress ratio0.56<.0001n/an/a
Figure FIGURE.

 Flow-mediated dilation (FMD) responses in women and men with untreated stage 1 hypertension. FMD means adjusted baseline arterial diameter, age, ethnicity, log shear stress ratio, body mass index, 24-hour diastolic blood pressure, high-density lipoprotein, log serum creatinine, and log physical activity.

Glyceryl Trinitrate Dilation by Sex

Baseline arterial diameter and ethnicity were significant determinants of sqrtGTND (Table Va). SqrtGTND was not significantly different between women and men (adjusted mean±SEM; 3.8±0.2% vs 4.1±0.1%, P=.121). A sex-by-race interaction was not significant when included in the model.

Table V.   The Effect of Sex on GTND in (a) Planned and (b) Confirmatory ANCOVA Models
 dfSSMean SquareFP ValuePartial η2
  1. Abbreviations: ANCOVA, analysis of covariance; BMI, body mass index; GTND, glyceryl trinitrate dilation; HDL, high-density lipoprotein; HR, heart rate; LDL, low-density lipoprotein; Log, log base-e; SBP, systolic blood pressure; SS, sum of squares; sqrt, square root.

Corrected total8864.9    
 dfType III SSMean SquareFP ValuePartial η2
Baseline artery diameter116.416.435.0<.0010.05
 dfSSMean SquareFP ValuePartial η2
Corrected total8159.0    
 dfType III SSMean SquareFP ValuePartial η2
Baseline artery diameter19.89.822.6<.0010.01
Log triglycerides10.
24-h SBP10.<0.01
Log creatinine clearance10.
Drinker (yes/no)
24-h HR11.

In secondary confirmatory analysis (Table Vb), BMI, HDL cholesterol, LDL cholesterol, log triglycerides, glucose, 24-hour SBP, log creatinine clearance, alcohol drinker, and 24-hour heart rate correlated with sqrtGTND at P≤.2 (Table IV) and were entered into the planned ANCOVA model. Women and men did not differ on sqrtGTND (adjusted means±SEM; 3.4±0.3% vs 3.9±0.2%, P=.072). A sex-by-race interaction was not significant when included in the model. Results did not differ based on menopausal status in the planned and fully adjusted models.


In this study, women between the ages of 40 and 60 years with untreated stage 1 hypertension exhibited poorer endothelial function, as indicated by smaller brachial artery FMD responses, compared with their male counterparts. In contrast, nonendothelium-dependent vasodilation, assessed by the GTND response, was comparable between sexes, supporting the interpretation that women with hypertension exhibit an impairment of vascular endothelial function rather than a generalized vascular smooth muscle dysfunction.24 These findings may be particularly relevant given that women have a higher sensitivity threshold than men for impaired FMD being indicative of the presence of coronary artery disease (CAD)11 and that reduced endothelial function has been found to be associated with an increased risk of CV events in women without CAD but not men.12

In accordance with the results of the current study, FMD was lower in women compared with men after adjustment for baseline artery diameter in 2265 adults (aged 24–39 years) participating in the Cardiovascular Risk in Young Finns Study.25 In a subgroup analysis of 224 pairs of women and men matched for identical baseline artery diameter, men continued to show larger FMD responses than women (8.2±4.7% vs 7.1±4.7%, P=.004) even after adjustment for traditional CV risk factors.25 However, not all studies have reported similar sex differences in FMD responses. In healthy men and women, FMD has been reported to be inversely related to age,23 with men experiencing a progressive decline in endothelial function after age 40, while women exhibited a more dramatic age-related decline in their early 50s.26 In a community sample (mean age 61±9 years), women demonstrated greater FMD responses than men up until 70 years of age23; a study of older adults (mean age 74.5±13.1 years) reported relatively impaired endothelial responses in women compared with men27; and no sex differences in FMD responses were observed among individuals (mean age 62±12 years) undergoing angiographic evaluation of CAD.12 These inconsistent findings suggest that age, population under study (clinical vs community), treatment with CV medications, and adjustment for baseline artery diameter may play a role in the inconsistencies observed in the literature regarding sex-based differences in endothelial function. Our present findings in a middle-aged (mean age 46±8 years) sample of untreated hypertensive women and men suggest that hypertension may also moderate the association between endothelial function and sex.

Anand and colleagues6 explored sex differences in CV risk factors using the INTERHEART database and found that hypertension was more strongly associated with first myocardial infarction in women (odds ratio [OR], 2.95; 95% confidence interval [CI], 2.66–3.28) than in men (OR, 2.32; 95% CI, 2.16–2.48), and the relationship was stronger for women younger than 60 years (OR, 4.00; 95% CI, 3.31–4.84) compared with older women (OR, 2.84; 95% CI, 2.49–3.23). The reasons for this sex disparity are not known but our present observations raise the possibility that hypertension may impart an increased CV risk in women because of its association with poorer endothelial function. In addition, because hypertension is less prevalent in younger (<45 years) women compared with men,28 the presence of hypertension in younger women may be indicative of more extensive impairment of endothelial function.

Although hypertension is an established CV risk factor that is associated with endothelial dysfunction,7–9 the mechanisms accounting for our observations of reduced endothelial function among hypertensive women remain to be clarified. Given that women have smaller conduit arteries than men,29,30 it is possible that over time, similar pressures in smaller arteries may have a more profound effect on the endothelium. On the other hand, high BP may be a manifestation of more severe endothelial dysfunction in women, requiring more vascular disease to “overcome” the tendency of women to have lower BP. Palatini and colleagues31 reported that untreated stage 1 hypertensive premenopausal women aged between 18 and 45 years had an increased risk of developing hypertensive end-organ damage (microalbuminuria and left ventricular hypertrophy) than men of the same age despite lower 24-hour SBP, BMI, triglycerides, and glucose. The authors hypothesized that the negative consequences of a hypertensive BP may be greater for younger women given their general tendency to have lower BP than men of similar ages. Donahue and colleagues32 also examined biomarkers of endothelial function (E-selectin and soluble intracellular adhesion molecule-1) in women and men who progressed from normoglycemia to prediabetes. The female participants who developed prediabetes exhibited poorer endothelial function than the normoglycemic controls while no differences were found among the male participants. The authors suggest that the prediabetic state may diminish the cardioprotective effect of being female, predisposing women to greater CV risk than their male counterparts. In the present study, untreated hypertension may also play a similar role in negating the cardioprotective advantage among women.

Women have often been viewed as being at lower CV risk due to the presence of cardioprotective reproductive hormones during the premenopausal years. However, sex-specific differences exist in CV disease (CVD).1–4,33 Hypertension is more prevalent in younger (<45 years) men and older (>65 years) women,28 and women more commonly have hypertension when diagnosed with acute coronary syndromes.4 The INTERHEART researchers reported that younger (<60 years) hypertensive women had a greater risk of myocardial infarction than their older counterparts.6 In addition, Gierach and colleagues34 found that among women with coronary risk factors undergoing coronary angiography for suspected myocardial ischemia, SBP, and pulse pressure were stronger risk factors for CAD in premenopausal women compared with postmenopausal women. The findings reinforce the need for enhanced CVD risk assessment and management in all hypertensive women.

Limitations and Strengths

The present study must be interpreted within the context of its potential limitations. Vascular function assessments of premenopausal women were not measured at the same menstrual cycle phase. FMD responses fluctuate throughout the menstrual cycle, being lowest during the menstrual phase.35 Only 17% of premenopausal women in this study were assessed in their menstrual phase and therefore it is unlikely to account for our findings. It should be noted that chronic periodontal disease is associated with impaired endothelial function.36,37 Periodontal disease was not assessed in the current investigation and it is possible that such recognized and other currently unrecognized confounding factors may have played a role in the study findings. Strengths of the study include the large percentage of African American participants and younger women with hypertension.


Hypertension is a major risk factor for CVD and CVD-related mortality in both men and women. Sex disparities in BP control and CVD risk management may exist, in part, due to an underestimation of women’s CVD risk. Our data suggest that women with high BP may exhibit poorer endothelial function compared with their male counterparts. This observation may reflect greater subclinical CVD among younger hypertensive women and support a more careful assessment and BP management for hypertensive women. Our present observations raise the possibility that female sex may impart an increased risk of CV events in those with untreated stage 1 hypertension because of its association with endothelial dysfunction. Additional studies are needed to confirm our findings and to examine potential mechanisms that contribute to the greater endothelial dysfunction in women with untreated stage 1 hypertension.

Acknowledgments and disclosures:  We thank Julie Bower, PhD, and Amy Franklin, PA, for coordinating the study while at Duke University Medical Center, and Michael Ellis, RDMS, RVT, for the acquisition of brachial artery images. We thank Melinda Higgins PhD, Associate Research Professor, Senior Biostatistician, Nell Hodgson Woodruff School of Nursing, Emory University for her statistical expertise and guidance in the revisions of this manuscript. This study was supported by grant HL072390 from the National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD, and grant M01-RR-30 from the General Clinical Research Center program, National Center for Research Resources, National Institutes of Health.