Physical activity in relation to indices of endothelial function and angiogenesis factors in hypertension: a substudy of the Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT)

Authors


Abstract

Abstract.  Felmeden DC, Spencer CGC, Blann AD, Beevers DG, Lip GYH (University Department of Medicine, City Hospital, Birmingham, UK). Physical activity in relation to indices of endothelial function and angiogenesis factors in hypertension: a substudy of the Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT). J Intern Med 2003; 253: 81–91.

Background.  Hypertension is an important risk factor for cardiovascular disease. The latest guidelines recommend regular physical exercise as initial step or adjunct in the treatment of hypertension. We investigated the association between physical activity and the degree of hypertension, as well as the relation to indices of endothelial damage/dysfunction and angiogenesis.

Methods.  We studied 234 patients with hypertension (198 males; mean age 64 years; mean blood pressure 166/90 mmHg), who were compared with 60 age and sex-matched healthy normotensive controls. We assessed the patient's physical activity using the validated Baecke physical activity questionnaire and measured flow-mediated dilatation (FMD) of the brachial artery and von Willebrand factor (vWf) as indices of endothelial damage/dysfunction, whilst angiogenesis was assessed by measurement of plasma vascular endothelial growth factor (VEGF) and its soluble receptor (sFlt-1) both by ELISA.

Results.  When hypertensive patients were compared with the controls, there was no statistically significant difference in total physical activity score using the Baecke questionnaire although plasma VEGF and vWf levels were higher, but sFlt-1 levels and FMD lower (all P < 0.001). Patients with high physical activity were younger, and had lower mean diastolic blood pressure and 10-year Framingham stroke risk, when compared with those with low physical activity; but indices of endothelial damage/dysfunction and angiogenesis were not significantly different.

Conclusion.  Physical activity scores in hypertensive patients are not significantly different from healthy normotensive controls, and there appears to be no relation to the abnormal processes of endothelial damage/dysfunction and angiogenesis seen in hypertensives.

Introduction

Hypertension is an important risk factor for cardiovascular disease with major impact on health resources. Antihypertensive management has been shown to reduce morbidity and mortality. In the recent guidelines of the Joint National Committee of the Detection, Evaluation and Treatment of High Blood Pressure nonpharmacological measures have been recommended as initial step for persons with stage one or two hypertension [1]. This nonpharmacological management can be divided into three categories: nutrition, change of life style and aerobic exercise. Changes in the nutrition, such reduced sodium intake, weight reduction in the obese and limiting alcohol consumption to recommended levels, have been shown to reduce blood pressure [2–4]. Several epidemiological studies have also indicated that a high level of physical activity is associated with lower morbidity and mortality [5, 6]. The exact underlying pathophysiological mechanism(s) for these beneficial effects are as yet unclear.

There is evidence to suggest that aerobic exercise may prevent hypertension and reduce blood pressure [7, 8]. However, other trials demonstrated failure of exercise to reduce blood pressure in patients with mild hypertension [9]. Acute exercise has also been shown to affect haemorheological markers. Nevertheless, some studies demonstrated procoagulant effects [10], whereas others showed a reduction in fibrinogen and other procoagulant markers [11]. Other studies have suggested that the favourable cardiovascular outcome associated with increased physical exercise may be mediated via improved endothelial function [12, 13] and several studies have reported an increase in angiogenic growth factors in animal models and healthy humans following exercise [14–16]. The processes of endothelial damage/dysfunction and angiogenesis are evident in hypertension, and have been thought to contribute to the pathogenesis and vascular effects of hypertension [17, 18].

We hypothesized that physical activity may be abnormal in hypertensives at ‘high risk’ of cardiovascular (CHD) or stroke (CVA) according to the Framingham equation, and abnormal physical activity may relate to the processes of endothelial damage/dysfunction and angiogenesis in hypertension. We therefore investigated usual daily physical activity in patients with hypertension, as assessed by the validated Baecke questionnaire [19–21] in relation to: (i) endothelial damage/dysfunction, as indicated by plasma von Willebrand factor levels (vWf, an established index of endothelial damage) and flow mediated dilatation of the brachial artery (FMD, as a marker of endothelium-dependent vasodilatation) [22]; (ii) indices of angiogenesis, by measurement of plasma Vascular Endothelial Growth factor (VEGF) and its soluble receptor, sFlt-1 [23].

Subjects and methods

Subjects

Patients were recruited from amongst those screened as part of the Anglo Scandinavian Cardiac Outcome Trial (ASCOT) at our dedicated research clinic at the City Hospital, Birmingham, UK. The methods for the ASCOT study have previously been described [24]. The inclusion criteria were patients aged between 40 and 80 years with either newly diagnosed untreated hypertension with systolic blood pressure (BP) > 160 mmHg and/or diastolic BP > 100 mmHg or treated hypertension with systolic BP > 140 mmHg and/or diastolic BP > 90 mmHg. The blood pressure was measured after 10 min rest in quiet room, in the sitting position. Three consecutive blood pressure readings were taken and the average of the last two readings was used. Patients with secondary hypertension or malignant phase hypertension were excluded. The following clinical and laboratory assessments were also performed: signs of left ventricular hypertrophy (LVH) on the ECG according to Cornell voltage duration product (>2440) or Sokolow Lyon criteria (>38 mV); other ECG abnormalities (left ventricular strain pattern, abnormal Q waves, LBBB, ST-T changes compatible with IHD); diabetes mellitus according to WHO criteria; past medical history of cerebrovascular event, including transient ischaemic attack; male sex; age > 55 years; microalbuminuria/proteinuria; current smoking; increased plasma total cholesterol/HDL cholesterol ratio; family history of coronary artery disease in first degree relative before the age of 55 (males) or 60 (females), and peripheral vascular disease according to Edinburgh claudication questionnaire [25]. Using these data, the Framingham risk calculations of CHD and CVA risk score were derived [26].

The healthy control group consisted of 60 healthy normotensive (<140/80 mmHg) controls (46 men, 14 women; mean age 62 years, SD 7.8) recruited from healthy hospital staff, relatives of the patients and those attending the hospital of routine cataract surgery. The subjects had no clinical evidence of vascular, metabolic, neoplastic, or inflammatory disease, by careful history, examination and routine laboratory tests. The control group included 10 smokers (16.7%) representing a similar proportion than in the hypertensive group. These subjects were normotensive and in sinus rhythm, and were not taking aspirin, warfarin, lipid-lowering or antihypertensive drugs, NSAIDs or antibiotics. The reason for including this healthy control group was not to emphasize a case/control comparison of the research indices, but to indicate approximate ‘normal’ levels of the various markers for comparisons with the hypertension patient group.

Assessment of physical activity

We interviewed all participants using a standard Baecke questionnaire to quantify physical activity [22]. This validated method assesses 16 different areas concerning habitual activities. The responses were coded on a 5-point scale, which provided a guide of general daily activities. The Baecke score is based on three components: (i) an occupational/work activity score, (ii) sports score, and (iii) leisure time activity score. The total activity score of routine physical activity was calculated from the individual scores. We have previously applied this score in post myocardial infarction patients, from different ethnic groups [27, 28].

Flow-mediated dilatation technique

High-resolution ultrasound was used to assess changes in brachial artery diameter as previously described [29]. Measurements were taken in the morning after fasting patients had rested in a supine position for 20 min in a quiet room. Good-quality images were obtained by a single dedicated operator using 10 MHz vascular ultrasound probe (GE Vingmed Ultrasound, System V, Slough, Berkshire, UK). The brachial artery was scanned in longitudinal section approximately 5 cm above the elbow. The transducer remained in a fixed position relative to the patients' arm throughout the procedure. Vessel diameter was assessed at end diastole, with measurement of leading edge to leading edge. After baseline scan a pneumatic cuff was placed at the level of the mid forearm and inflated to 250–300 mmHg for 4.5 min. The second scan was performed 30–90 s after cuff release (peak changes during reactive hyperaemia), and 15 min were allowed for vessel recovery before performing the next scan. The scans of the brachial artery were repeated before and at 3.5 min after sublingual glyceryl trinitrate (400 μg GTN spray) administration, to assess endothelium-independent vasodilatation. Scans were recorded on super-VHS videotape for later analysis by an independent investigator blinded to treatment assignment. Flow-mediated (endothelium-dependent) dilatation and GTN-induced (endothelium-independent) dilatation were estimated as per cent change in diameter relative to their respective baseline measurements.

Laboratory methods

Blood was drawn after 8-h fasting period with minimal trauma from the antecubital vein. Samples were put on ice for 5 min and then centrifuged at 1500 g for 20 min. The plasma was stored at −70°C until assayed. In the routine hospital pathology laboratory, plasma was analysed by standard techniques for total cholesterol, triglycerides and high-density lipoprotein (HDL) cholesterol.

Levels VEGF and sFlt-1 in citrated plasma were measured by ELISA using commercially available reagents and recombinant standards (R&D Systems, Abingdon, UK) [18]. The VEGF assay has a minimum sensitivity of 0.5 IU dL−1 with an intra-assay coefficient of 4.4% and interassay coefficient of 9.1%. The sFlt-1 assay has a minimum sensitivity of 50 pg dL−1 with an intra-assay coefficient of 3.7% and interassay coefficient of 8.8%. Levels of vWF were analysed with a sandwich ELISA using commercially available reagents and standards (Dako Ltd, Ely, Cambs, UK). The assay has a minimum sensitivity of 0.5 IU dL−1 with an intra-assay coefficient of <5% and interassay coefficient of <10%. All laboratory work was performed in blinded fashion with respect to the identity of the samples.

Statistical analysis

The participants' characteristics were compared using Chi squared (χ2) test. Continuous data were subjected to the Ryan–Joiner test to assess distribution. Age, blood pressure, Baecke questionnaire scores and vWF levels were normally distributed and are expressed as mean and standard deviation (SD). VEGF, sFlt-1 levels, CHD and CVA risk were not normally distributed and are therefore shown as median and interquartile range (IQR). Data between cases and controls were analysed using independent t-test or Kruskal–Wallis test, as appropriate. Correlations between Baecke scores and the measured parameters were assessed according to Spearman's method. The stepwise regression analyses were used with Baecke scores as dependent variables and clinical indices (age, sex, blood pressure, past medical history, etc.) as predictors. All statistical calculations were performed on a microcomputer using a commercially available statistical package (SPSS 10.0 for Windows, Chicago, IL, USA). A P-value <0.05 was considered statistically significant.

Results

In our study the hypertensive patients and control group were comparable for age, gender, ethnic groups and smoking status. Mean systolic and diastolic blood pressure, body mass index and total cholesterol levels were significantly elevated in the hypertensive patients compared with controls. As expected, the patients also had a worse overall risk factor profile than the controls, with higher 10 years CHD and CVA risk according to the Framingham equation (Table 1).

Table 1.  Cross-sectional data
 ControlsHypertensivesP-value
  1. SBP, systolic blood pressure; DBP, diastolic blood pressure; LVH, left ventricular hypertrophy; PVD, peripheral vascular disease; CVA, cerebrovascular event; TIA, transient ischaemic event.

  2. Values are expressed as mean and SD, except CHD + CVA risk, VEGF and sFlt-1 as median and IQR. Statistical analysis: unpaired t-test, except for CHD + CVA risk, VEGF and sFlt-1: Mann–Whitney. Values in bold indicate statistically significant differences.

n60234 
Age (years)62 (7.8)64 (8.0)0.098
Male:female46/14198/360.144
BMI (kg m−2)25.8 (3.4)29.2 (4.4)<0.001
White:Black:Indo-Asian:other55:4:1:0198:24:11:10.318
Past medical history (n)
 LVH on ECG021 
 CVA/TIA022 
 Diabetes mellitus058 
 PVD018 
 Smoker10510.382
Total cholesterol (mmol L−1)5.6 (1.0)6.1 (1.0)0.001
HDL cholesterol (mmol L−1)1.4 (0.5)1.3 (0.4)0.568
SBP (mmHg)134 (14)166 (17)<0.001
DBP (mmHg)80 (7.4)90 (9.5)<0.001
CHD risk (10 years) (%)10.4 (6.1–16.4)23.6 (17.8–30.3)<0.001
CVA risk (10 years) (%)2.4 (1.5–4.0)7.9 (5.5–12.4)<0.001
Physical activity scores (Baecke questionnaire)
 Work activity score1.57 (0.91)1.66 (0.91)0.527
 Sports score2.24 (0.78)2.40 (0.660.118
 Leisure time activity score2.69 (0.50)2.71 (0.59)0.890
 Total activity score6.57 (1.58)6.77 (1.46)0.353
Indices of Endothelial damage/dysfunction
 FMD (%)8.80 (2.3)4.89 (1.62)<0.001
 GTN-MD (%)20.7 (2.2)17.4 (6.4)<0.001
 vWf (IU dL−1)95 (20)134 (30)<0.001
Indices of angiogenesis
 VEGF (pg mL−1)125 (40–213)340 (190–1300)<0.001
 SFlt-1 (ng mL−1)17.0 (10.0–33.0)4.3 (1.5–14.0)<0.001

Physical activity scores in hypertensive were not significantly different when compared with controls. Indices of endothelial damage/dysfunction and angiogenesis were different in hypertensives, compared with controls (Table 1). In particular, FMD and GTN-medicated dilatation, as well as sFlt-1 levels were reduced in patients with hypertension (Table 2).

Table 2.  Ethnic subgroups
Ethnic groupaWhite peopleAfro-CaribbeanAsianP-value
  • a

    one patient was Chinese, therefore not included in analysis.

  • b

    White people versus Asians: P = 0.005.

  • c

    White people versus Afro-Caribbeans: P = 0.001.

  • Values are expressed as mean and SD, except VEGF, sFlt-1, CHD and CVA risk as median and IQR. Statistical analysis for comparison: One-way anova with post hoc Tukey and Kruskal–Wallis as appropriate. Values in bold indicate statistically significant differences.

n1982411 
Age (years)64 (7.6)61 (8.2)57 (7.9)b0.001
Male:female167:3720:410:10.809
Past medical history (n)
 LVH on ECG14610.044
 CVA/TIA21100.178
 Diabetes mellitus401350.001
 PVD17010.130
 Smoker43610.500
Total cholesterol (mmol L−1)6.2 (1.0)5.4 (1.0)c6.1 (1.1)0.002
HDL cholesterol (mmol L−1)1.3 (0.4)1.5 (0.3)1.3 (0.4)0.238
SBP (mmHg)166 (17)162 (12)165 (21)0.551
DBP (mmHg)90 (10)91 (11)90 (10)0.932
CHD risk (10 years) (%)23.6 (18.0–30.2)25.6 (17.6–33.6)17.0 (10.6–21.8)0.043
CVA risk (10 years) (%)7.9 (5.5–12.5)8.0 (6.2–13.3)7.1 (2.8–8.9)0.250
Physical activity scores (Baecke questionnaire)
 Work activity score1.65 (0.94)1.75 (1.05)1.5 (0.88)0.768
 Sports score2.42 (0.67)2.28 (0.66)2.25 (0.46)0.465
 Leisure time activity score2.70 (0.59)2.71 (0.57)2.86 (0.69)0.661
 Total activity score6.77 (1.47)6.74 (1.40)6.61 (1.26)0.938
Indices of endothelial damage/dysfunction
 FMD (%)4.9 (1.6)4.6 (0.9)5.6 (2.2)0.640
 GTN-MD17.2 (6.5)23.9 (1.4)16.6 (4.6)0.345
 vWF (IU dL−1)136 (29)124 (40)116 (30)0.098
Indices of angiogenesis
 VEGF (pg mL−1)325 (190–1405)460 (320–2950)250 (125–790)0.087
 sFlt-1 (ng mL−1)6 (1.5–14)7.5 (1.9–18.2)1.5 (0.7–4)0.089

High physical activity, body mass index, pulse pressure and antihypertensive therapy use

After dividing the hypertensive patients into two groups based on physical activity, that is, those with ‘low’ (= median Baecke score) and ‘high’ (>median Baecke score) physical activity, the ‘high’ physical activity group were younger (P < 0.001) and had lower mean 10-year CVA risk (P = 0.010) and diastolic blood pressure (P = 0.041). All other parameters, including the indices of endothelial damage/dysfunction and angiogenesis were not significantly different between ‘low’ and ‘high’ physical activity groups (Table 3).

Table 3.  Physical activity, endothelial damage/dysfunction and angiogenesis in relation to high and low physical activity
 Low Baecke scoreHigh Baecke scoreP-value
  1. SBP, systolic blood pressure; DBP, diastolic blood pressure; LVH, left ventricular hypertrophy; PVD, peripheral vascular disease; CVA, cerebrovascular event; TIA, transient ischaemic event.

  2. Values are expressed as mean and SD, except CHD + CVA risk, VEGF and sFlt-1 as median and IQR. Statistical analysis: unpaired t-test, except for CHD + CVA risk, VEGF and sFlt-1: Mann–Whitney. Values in bold indicate statistically significant differences.

n117117 
Age (years)67 (7.2)61 (7.8)<0.001
Male:female97/20101/160.469
BMI (kg m−2)29.2 (4.0)29.2 (4.8)0.961
White:Black:Indo-Asian:other103:8 : 6 : 095 : 16 : 5 : 10.215
Past medical history (n)
 LVH on ECG7140.075
 CVA/TIA1390.370
 Diabetes mellitus27310.573
 PVD1260.141
 Smoker24270.655
Total cholesterol (mmol L−1)6.1 (1.1)6.0 (1.0)0.326
HDL cholesterol (mmol L−1)1.4 (0.4)1.3 (0.4)0.470
SBP (mmHg)166 (18)166 (16)0.828
DBP (mmHg)91 (10.0)89 (9.0)0.041
CHD risk (10 years) (%)25.1 (18.5–31.7)22.4 (16.8–30.1)0.075
CVA risk (10 years) (%)8.8 (6.2–13.9)7.5 (4.9–11.3)0.010
Physical activity scores (Baecke questionnaire)
 Work activity score1.13 (0.40)2.19 (1.05)<0.001
 Sports score2.03 (0.57)2.78 (0.52)<0.001
 Leisure time activity score2.43 (0.52)2.98 (0.53)<0.001
 Total activity score5.61 (0.91)7.92 (0.81)<0.001
Indices of Endothelial damage/dysfunction
 FMD (%)4.87 (1.85)4.89 (1.38)0.974
 GTN-MD (%)17.6 (6.6)17.2 (5.9)0.728
 vWf (IU dL−1)138 (30)131 (30)0.162
Indices of angiogenesis
 VEGF (pg mL−1)400 (190–1500)300 (170–1300)0.275
 sFlt-1 (ng mL−1)5.5 (1.4–12.5)4.0 (1.5–13.5)0.871

Subgroup analysis of patients in relation to median body mass index (BMI) demonstrated a higher ratio of females and diabetics as well as lower incidence of LVH on ECG in the high BMI group. The HDL cholesterol was also lower in this group. The remaining parameters were similar in both BMI groups (Table 4).

Table 4.  Physical activity, endothelial damage/dysfunction and angiogenesis in relation to body mass index
 Low (<median) BMIHigh (= median) BMIP-value
  1. SBP, systolic blood pressure; DBP, diastolic blood pressure; LVH, left ventricular hypertrophy; PVD, peripheral vascular disease; CVA, cerebrovascular event; TIA, transient ischaemic event.

  2. Values are expressed as mean and SD, except CHD + CVA risk, VEGF and sFlt-1 as median and IQR. Statistical analysis: unpaired t-test, except for CHD + CVA risk, VEGF and sFlt-1: Mann–Whitney. Values in bold indicate statistically significant differences.

n117117 
Age (years)65 (8.0)63 (7.7)0.070
Male:female13:10423 : 940.025
BMI (kg m−2)25.9 (2.0)32.6 (3.6)<0.001
White:Black:Indo-Asian101:10:697:14:50.834
Past medical history (n)
 LVH on ECG1740.003
 CVA/TIA1390.370
 Diabetes mellitus19390.002
 PVD1260.141
 Smoker28230.429
Total cholesterol (mmol L−1)6.1 (1.1)6.0 (1.0)0.373
HDL cholesterol (mmol L−1)1.4 (0.4)1.3 (0.3)0.038
SBP (mmHg)165 (16)167 (18)0.418
DBP (mmHg)89 (9.6)91 (9.3)0.075
CHD risk (10 years) (%)24.4 (18.1–32.7)23.4 (17.3–28.1)0.171
CVA risk (10 years) (%)8.3 (5.6–12.9)7.7 (5.1–12.0)0.285
Physical activity scores (Baecke questionnaire)
 Work activity score1.59 (0.92)1.73 (0.98)0.265
 Sports score2.45 (0.66)2.35 (0.66)0.279
 Leisure time activity score2.80 (0.61)2.61 (0.56)0.017
 Total activity score6.84 (1.42)6.70 (1.50)0.468
Indices of endothelial damage/dysfunction
 FMD (%)4.74 (1.61)5.07 (1.65)0.346
 GTN-MD (%)17.26 (6.50)17.58 (6.07)0.817
 vWf (IU dL−1)132 (29)135 (32)0.583
Indices of angiogenesis
 VEGF (pg mL−1)365 (488–1413)340 (190–1400)0.853
 sFlt-1 (ng mL−1)5.0 (1.7–14.0)3.9 (1.5–11.5)0.582

After subdividing the hypertensive patients into a below and above median pulse pressure, the group with lower pulse pressure was younger (P = 0.003), had lower systolic and higher diastolic blood pressure as well as decreased CHD and CVA risk (all P < 0.001). Levels of vWf were also significantly reduced (P = 0.033). The remaining parameters were not significantly different between the two groups (Table 5).

Table 5.  Physical activity, endothelial damage/dysfunction and angiogenesis in relation to pulse pressure
 Low (<median) PPHigh (= median) PPP-value
  1. SBP, systolic blood pressure; DBP, diastolic blood pressure; LVH, left ventricular hypertrophy; PVD, peripheral vascular disease; CVA, cerebrovascular event; TIA, transient ischaemic event.

  2. Values are expressed as mean and SD, except CHD + CVA risk, VEGF and sFlt-1 as median and IQR. Statistical analysis: unpaired t-test, except for CHD + CVA risk, VEGF and sFlt-1: Mann–Whitney. Values in bold indicate statistically significant differences.

n112122 
Age (years)62 (7.3)65 (8.2)0.004
Male:female17 : 9519 : 1030.933
BMI (kg m−2]29.3 (4.5)29.2 (4.4)0.899
White:Black:Indo-Asian92:13:6106:11:50.730
Past medical history (n)
 LVH on ECG8130.348
 CVA/TIA12100.510
 Diabetes mellitus26320.594
 PVD6120.199
 Smoker24270.897
Total cholesterol (mmol L−1)6.0 (1.1)6.2 (1.0)0.134
HDL cholesterol (mmol L−1)1.3 (0.4)1.3 (0.4)0.638
SBP (mmHg)155 (11)176 (15)<0.001
DBP (mmHg)92 (7.9)88 (10)<0.001
CHD risk (10 years) (%)19.8 (15.8–24.6)27.6 (21.1–34.0)<0.001
CVA risk (10 years) (%)6.3 (4.8–8.0)11.1 (7.5–16.4)<0.001
Physical activity scores (Baecke questionnaire)
 Work activity score1.71 (0.99)1.61 (0.92)0.477
 Sports score2.35 (0.68)2.45 (0.64)0.259
 Leisure time activity score2.67 (0.61)2.74 (0.57)0.322
 Total activity score6.73 (1.53)6.81 (1.40)0.654
Indices of endothelial damage/dysfunction
 FMD (%)5.09 (1.62)4.72 (1.62)0.283
 GTN-MD (%)18.66 (7.37)16.42 (5.11)0.101
 vWf (IU dL−1)128 (30)139 (30)0.033
Indices of angiogenesis
VEGF [pg/mL]305 (190–1425)360 (186–1400)0.834
sFlt-1 [ng/mL]5.0 (1.6–12.0)4.0 (1.4–14.0)0.829

Subgroup analysis of patients on hypertensive treatment at study entry compared with previously untreated (virgin) hypertensives showed that the latter were younger, had higher HDL cholesterol, systolic and diastolic blood pressure. Other baseline parameters were similar. Despite these differences, the markers of angiogenesis, endothelial dysfunction and thrombogenesis were not significantly different (Table 6).

Table 6.  Physical activity, endothelial damage/dysfunction and angiogenesis in relation to antihypertensive drug treatment
 No drugsTreated with antihypertensive RxP-value
  1. SBP, systolic blood pressure; DBP, diastolic blood pressure; LVH, left ventricular hypertrophy; PVD, peripheral vascular disease; CVA, cerebrovascular event; TIA, transient ischaemic event.

  2. Values are expressed as mean and SD, except CHD + CVA risk, VEGF and sFlt-1 as median and IQR. Statistical analysis: unpaired t-test, except for CHD + CVA risk, VEGF and sFlt-1: Mann–Whitney. Values in bold indicate statistically significant differences.

n20214 
Age (years)59 (8.1)64 (7.7)0.002
Male:female5:1531:1830.213
BMI (kg m−2)30.6 (6.0)29.1 (4.3)0.137
White:Black:Indo-Asian:other18:1:1180:23:100.673
Past medical history (n)
 LVH on ECG1200.809
 CVA/TIA1210.761
 Diabetes mellitus3550.430
 PVD3150.399
 Smoker7440.135
Total cholesterol (mmol L−1)6.4 (1.0)6.0 (1.0)0.201
HDL cholesterol (mmol L−1)1.5 (0.4)1.3 (0.4)0.028
SBP (mmHg)178 (16)165 (16)0.001
DBP [mmHg]100 (9.3)89 (9.0)<0.001
CHD risk (10 years) [%]20.1 (14.3–29.3)23.9 (18.0–30.7)0.157
CVA risk (10 years) [%]7.4 (4.9–11.2)8.0 (5.6–12.5)0.573
Physical activity scores (Baecke questionnaire)
 Work activity score1.71 (1.01)1.66 (0.95)0.778
 Sports score2.45 (0.59)2.40 (0.67)0.746
 Leisure time activity score2.63 (0.73)2.72 (0.57)0.514
 Total activity score6.79 (1.47)6.77 (1.46)0.946
Indices of Endothelial damage/dysfunction
 FMD (%)4.75 (1.86)4.91 (1.61)0.788
 GTN-MD (%)18.96 (6.39)17.22 (6.28)0.435
 vWf (IU dL−1)134 (25)134 (31)0.974
Indices of angiogenesis
 VEGF (pg mL−1)380 (190–1300)340 (189–1400)0.958
 sFlt-1 (ng mL−1)5.7 (1.4–15.3)4.0 (1.5–13.5)0.880

Correlations

The Baecke score as well as the subscores for work, leisure and sport activity were similar in both groups and correlated well with each other (all P < 0.001) (Table 7). The total Baecke physical activity score correlated negatively with age (r = −0.299, P < 0.001), BMI (r = −0.175, P = 0.006) and 10 years of CVA risk (r = −146, P = 0.025). There was also a correlation between CHD/CVA risk and the work score (r = −0.194, P = 0.003, r = −0.302, P < 0.001), but none of the other activity subscores. There was no other significant correlation between the Baecke score and any of the other parameters assessed (gender, total cholesterol, HDL, systolic, diastolic blood pressure vWf, VEGF, sFlt-1, CHD or CVA risk scores, according to the Framingham equation).

Table 7.  Correlation between Baecke activity subscores
 Work activityP-valueSports activityP-valueLeisure activityP-value
  1. Values in bold indicate statistically significant differences.

Sports activity0.1350.02110.334<0.001
Leisure activity−0.0320.5890.334<0.0011
Total activity0.669<0.0010.696<0.0010.527<0.001

Discussion

In the present study, we have demonstrated that physical activity scores in hypertensive patients were not significantly different from healthy normotensive controls, and there appeared to be no relation to the abnormal processes of endothelial damage/dysfunction and angiogenesis seen in hypertensives. The present study confirms previous observations of increased plasma vWf and VEGF, and decreased sFlt-1 levels and FMD, in hypertension and vascular disease [18, 22, 30, 31]. Physical activity scores did correlate negatively with age and the 10-year CVA risk according to the Framingham equation was significantly lower in the group with above median activity scores.

The beneficial effects of exercise on endothelial function has been demonstrated mainly in animal studies [32] and healthy men [33]. Recently Higashi et al. [34] investigated endothelium-dependent dilatation in mild untreated hypertensive subjects after 12 weeks of aerobic exercise. The endothelial function was assessed by measuring response of forearm vasculature to acetylcholine using strain-gauge plethysmography. Endothelial-dependent vasorelaxation improved following long-term exercise through an increase in nitric oxide (NO) as this effect could be abolished by infusion of NO synthase inhibitor. However, in our study we could not demonstrate an association between physical exercise and endothelial function in high-risk hypertensive patients. The Japanese study population was much younger with a much lower cardiovascular risk profile, with lower systolic blood pressure, total cholesterol and BMI. Hence, the beneficial effects of exercise may not have a large enough impact on the severely dysfunctional endothelium. Furthermore, aerobic exercise was used as an intervention, unlike the assessment of habitual exercise in our study. Because of our highly motivated hypertension clinic patient cohort, exercise levels were relatively high in comparison with healthy controls, and hence the difference between the two groups may not have been sufficiently large enough to result in significant differences.

Abnormal angiogenesis was demonstrated in high-risk hypertensive patients confirming the finding of Belgore et al. [18]. Animal and human data from studies investigating the effects of exercise on angiogenesis suggested increased expression of VEGF following exercise [14, 15]. Even in patients with moderate heart failure, exercise training amplified VEGF gene expression [15]. Thus, we would have expected to find elevated levels of VEGF in patients with higher physical activity, but in the present study, these were not apparent. This would suggest that there might be a difference between ‘habitual exercise’, which is usually stable over long periods of time and ‘exercise as part of an intervention’ with significant changes over a short period of time. Currently there are no data to suggest different VEGF levels depending on the degree of regular daily exercise. Another possible explanation may be the lack of any ‘additional’ effect from physical activity in this cohort, as the already markedly increased levels of VEGF in our hypertensive patients with several additional cardiovascular risk factors, which contribute to higher VEGF levels, might not be significantly further elevated in response to exercise.

Our data confirm findings of the prospective Atherosclerosis Risk in Communities (ARIC) Study [35] of nearly 16 000 individuals, where physical activity was (weakly) associated with a reduced risk of ischaemic stroke, when comparing lowest and highest quartiles of the activity scores. In our study, physical exercise was also associated with lower diastolic blood pressure and younger age, which may explain the different CVA risk according to the Framingham equation. These integrate the risk factors age, gender, systolic/diastolic blood pressure, smoking status, history of diabetes mellitus, presence of left ventricular hypertrophy, total and HDL cholesterol [26]. After adjustment for age, neither diastolic blood pressure nor CVA risk were significantly associated with the habitual activity score. Similarly in the ARIC study [35], the inverse association between exercise and CVA risk was attenuated after correction for age. These slight dissimilarities may be related to differences in the two study groups: our group consisted purely of hypertensive patients (about one-third were hypertensives in ARIC) with a far greater proportion of men (84.6% vs. 43.2%) and less smokers (21.6% vs. 25.7%). The weak association between CVA risk and physical activity may therefore partially be explained by the influence of associated risk factors.

In a study analysing preadmission levels of physical activity in patients admitted with myocardial infarction, ethnic differences were apparent, with white people having higher sport, leisure and total activity scores when compared with Indo-Asians [27]. However, in our study the white people group had a significantly higher mean age than the Indo-Asians. Despite this fact, their activity scores were similar. In view of the negative correlation between age and activity score, the habitual activity was still similar. One could postulate that were the mean ages similar, white people would probably have higher the levels of physical activity than Asians similar to the ethnic differences noted by Lip et al. [27].

There has been increasing evidence of pulse pressure being a strong indicator of cardiovascular risk even amongst normotensive individuals [36]. Not surprisingly, the subgroup in the present study with low (below median) pulse pressure had a younger mean age, as pulse pressure rises with age because of loss of arterial compliance [37]. However, in a stepwise regression for CHD risk allowing for age, systolic/diastolic blood pressure and pulse pressure, pulse pressure still contributed significantly to CHD risk. Indeed, the contribution of systolic blood pressure was no longer significant when allowing for pulse pressure, suggesting that in this cohort of ‘high risk’ hypertensive patients, pulse pressure would be a better predictor of future cardiovascular events than systolic blood pressure. Indeed, recent data also suggests that pulse pressure is a particular useful marker for cardiovascular risk in the middle aged and elderly [38, 39]. The importance of pulse pressure is further supported by the fact that the indices for endothelial function were significantly decreased and angiogenesis were increased in the subgroup with high pulse pressure levels.

This study is limited by its cross-sectional nature and it is possible that the observed lack of difference in physical activity between the hypertensive patients and our control group may be because of the highly motivated patient cohort in our study. Whilst physical activity improves endothelial dysfunction in young individuals [12, 13], this may not necessarily hold true for our older study population, as vascular reactivity is poorer in the older population. The demonstrated absence of correlation between exercise and cardiovascular risk factors in high-risk hypertensive patients should not result in reduced efforts to motivate hypertensive patients to exercise. However, the beneficial effects of exercise in this particular group might be smaller than anticipated from previous observational studies. There is therefore a need for further interventional trials assessing this relationship. A questionnaire-based approach to quantifying physical activity does not account for differences in responses from actual practice, as well as variation in habitual physical activity. Finally, we used the Baecke questionnaire to give a measure of physical activity score (to at least allow a quantitative figure) but recognize the limitations of using a questionnaire based tool.

In conclusion, physical activity scores in hypertensive patients are not significantly different from healthy normotensive controls, and there appears to be no relation to the abnormal processes of endothelial damage/dysfunction and angiogenesis seen in hypertensives. Further studies are needed to assess this relationship, and the effects of intervention and/or drug treatment.

Acknowledgements

We acknowledge the support of the City Hospital Research and Development programme for the Haemostasis Thrombosis and Vascular Biology Unit. We would like to thank Dr F Belgore, Sisters Ranjit Dhillon, Maggie Lal, Zoë Townend and Ruth Watson for their help with data collection. Other ASCOT Investigators are listed in [24].

Received 2 May 2002; revision received 5 November 2002; accepted 8 November 2002.

 Dr Gregory YH Lip, Haemostasis, Thrombosis and Vascular Biology Unit, University Department of Medicine, City Hospital, Birmingham B18 7QH, UK (fax: +44 121 5544083; e-mail: g.y.h.lip@bham.ac.uk).

Ancillary