The aim of this study was to test the hypothesis that blood pressure (BP) increase before exercise stress testing is associated with the incidence of hypertension in a prospective study of 3,805 normotensive men without hypertension at baseline. Changes in BP were defined as the difference between seated BP at rest and BP measured immediately before exercise stress testing. Hypertension was defined as systolic and diastolic BP ≥140/90 mm Hg or hypertension diagnosed by a physician at the second examination. During 18,923 patient-years of follow-up, 371 new cases of hypertension developed (incidence rate 19.6 per 1,000 patient-years). Men with systolic BP changes >0 mm Hg and diastolic BP changes >7 mm Hg had 1.70 times (95% confidence interval [CI] 1.37 to 2.12) and 2.23 times (95% CI 1.76 to 2.82) increased relative risk for incident hypertension compared with men whose systolic BP changes were <0 mm Hg and diastolic BP changes were <7 mm Hg after adjustment for confounders. Men in the highest quartile of mean BP change (>10 mm Hg) had a higher incidence of hypertension (relative risk 2.98, 95% CI 2.19 to 4.06) compared with those in the lowest quartile (<0 mm Hg), and each 1 mm Hg increment in mean BP was associated with a 6% (95% CI 1.05 to 1.09) higher incidence of hypertension after adjustment for risk factors. In conclusion, BP increase before exercise stress testing is associated with incident hypertension, independent of risk factors in normotensive men. The assessment of BP immediately before exercise testing may be a useful addition to the standard exercise stress testing procedures.
Highlights
- •
BP increase before exercise stress testing is associated with incident hypertension, independent of risk factors in normotensive men.
- •
These associations persist even after adjusting for maximal SBP during exercise testing and levels of cardiorespiratory fitness.
- •
The assessment of BP immediately before exercise testing may be a useful addition to the standard exercise testing procedures.
The cardiovascular reactivity hypothesis states that heightened hemodynamic reactions to acute psychological or physiologic stress exposures increase cardiovascular disease risks and the development of hypertension. The time before exercise testing can be characterized as a time of psychological and emotional stress that could be used as a potential research model for real-life cardiovascular reactivity. Although the associations between exercise test–related variables obtained during and immediately after stress testing (electrocardiographic change, maximal oxygen uptake, maximal heart rate [HR] and blood pressure [BP], and recovery BP and HR) and cardiovascular risk have been extensively examined, it remains unclear if BP measured immediately before exercise testing is a useful risk factor for predicting cardiovascular health. A few studies have reported the association between greater increases in BP reactivity using anticipation of exercise stress testing and cardiovascular risk factors, but whether anticipatory BP response to exercise stress testing predicts the incidence of hypertension across ethnic groups is not well known in initially normotensive populations. We tested the hypothesis that BP increase before exercise stress testing is associated with the incidence of hypertension in middle-aged normotensive men. Furthermore, we investigated whether these associations persist with adjustment for potential confounders, including maximal systolic BP during exercise testing and cardiorespiratory fitness.
Methods
Participants were 5,616 men who took part in 2 general health examinations as part of a comprehensive health screening program from 1998 to 2009 at Samsung Medical Center (Seoul, South Korea). Women (n = 113) from a total of 5,616 participants were excluded because of too small a sample size for further analysis. Participants (n = 453) were also excluded for the following reasons: diagnosis of hypertension or type 2 diabetes by a physician, history of cardiovascular disease, and the use of antihypertensive or oral hypoglycemic medications. We further excluded participants who had systolic BP (SBP) and diastolic BP (DBP) ≥140 and ≥90 mm Hg, respectively, at rest (n = 936) and fasting glucose levels >126 mg/dl (n = 309). Therefore, the remaining 3,805 men (mean age 48 ± 6 years) who did not have hypertension, diabetes, or cardiovascular disease at baseline examination were included in this analysis. Incident hypertension was identified as SBP ≥140 mm Hg and/or DBP ≥90 mm Hg or hypertension diagnosed by a physician during follow-up health examination. Written informed consent was obtained from all participants before the health screening program, and the study was approved by the medical center institutional review board.
SBP and DBP at rest were measured in the seated position by a nurse after ≥5 minutes of quiet rest using an automated BP monitor (Dinamap PRO 100; GE Healthcare, Milwaukee, Wisconsin). The lowest value of 2 measurements was used as the rest BP. Mean baseline pressure (MBP) was calculated as [(2 × DBP) + SBP]/3. One hour after measurements at rest, participants performed symptom-limited exercise stress testing. After the attachment of a BP cuff on the upper arm and electrocardiographic electrodes on the participant’s chest, BP was measured in the seated position on the stool on the treadmill using an automated exercise BP monitor (STBP-680; Colin Corporation, Komaki, Japan) by an exercise physiologist. This BP measurement was used as the anticipatory BP response to the beginning of a treadmill exercise stress testing, as it may be affected by anxiety and emotional arousal anticipation to the impending challenge. Therefore, increases in BP in anticipation to exercise may reflect cardiovascular reactivity to stress. Changes in BP were defined as the difference between BP at rest and BP measured immediately before exercise stress testing.
All participants performed symptom-limited exercise stress testing using the Bruce protocol. A 12-lead electrocardiogram was obtained throughout the test (Quinton Q-4500; Philips Medical Systems, Bothell, Washington). Breath-by-breath expired gas was collected using a 1-way valve and analyzed using a metabolic cart (Jaeger Oxycon Delta; Erich Jaeger, Hoechberg, Germany). Oxygen uptake data were measured using 20-second intervals, and peak oxygen uptake (VO 2 ) was defined as the highest value recorded during peak exercise or immediate recovery. Exercise BPs were measured during the last minute of each 3-minute stage and at the moment of maximal effort, with the arm relaxed at the side without holding onto the side bar of the treadmill, using an automatic BP monitor designed for exercise testing (STBP-680). The maximal SBP was defined as the highest value achieved during the test. Exercise tests were terminated if 1 of the following criteria was present: a rating of perceived exertion >17, if the participant achieved >90% of age-predicted maximal HR, if the participant was too fatigued to safely continue walking on the treadmill, an increase in SBP >250 mm Hg, chest discomfort, severe dizziness, and/or >1 mm of horizontal or downsloping ST-segment depression.
HR at rest was measured in the supine position using an electrocardiograph (Hewlett-Packard ECG M 1700A; Hewlett-Packard, Palo Alto, California). Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. Blood samples were collected in the morning after a 12-hour overnight fast and analyzed by a technician at the hospital clinical laboratory. Total cholesterol (TC), triglycerides and high-density lipoprotein cholesterol were analyzed enzymatically using a Hitachi 747 (Tokyo, Japan) analyzer. Low-density lipoprotein cholesterol was calculated using the Friedewald formula. Fasting glucose concentrations were determined using the glucose oxidase method (Hitachi 747). White blood cell count was determined using a quantitative automated hematology analyzer (Sysmex Corporation, Tokyo, Japan). Inter- and intra-assay coefficients of variation were <5% for all blood variables. Smoking habits (never, past, or current), alcohol consumption (none or ≥3 times/week), and other health information was evaluated using a questionnaire.
Data are expressed as mean ± SD for continuous variables and as proportions for categorical variables. Baseline variable comparisons between men with and without the development of hypertension were performed using independent Student’s t test for continuous variables and chi-square tests for categorical variables. Changes in BP as a continuous variable were calculated as the difference between seated BP at rest and BP measured immediately before exercise stress testing. Then, changes in SBP and DBP were divided into 2 groups on the basis of median values, and changes in MBP were divided into quartiles. Group comparisons of incident hypertension by BP changes were performed using the chi-square test. Cox proportional-hazards regression after adjustment for confounding factors was used to determine the effect of BP changes on the incidence of hypertension using categorical and continuous variables. Our multivariate modeling strategy was to initially adjust for age, BMI, and BP at rest in model 1. Then, we adjusted for model 1 plus additional potential confounding factors (TC, high-density lipoprotein cholesterol, triglycerides, glucose, HR at rest, white blood cell count, smoking, and alcohol intake) in model 2. Finally, we adjusted for model 2 plus additional potential confounding factors (peak VO 2 and maximal SBP) in model 3 because we expected a strong association between these variables and the risk for developing hypertension. Statistical significance was set at p <0.05. All tests for statistical significance were 2 sided. Statistical analyses were performed using SPSS version 20 (IBM, Armonk, New York).
Results
Baseline characteristics of men with and without developed hypertension during follow-up are listed in Table 1 . BMI, BP, TC, low-density lipoprotein cholesterol, HR at rest, and maximal SBP were higher, whereas peak VO 2 was lower in men with hypertension. Men with hypertension were more likely to be smokers than men without hypertension.
| Variable | Hypertension | p-Value | |
|---|---|---|---|
| NO (n = 3,434) | YES (n = 371) | ||
| Age (years) | 48 ± 6 | 48 ± 7 | 0.06 |
| Follow-up (years) | 4.9 ± 2.8 | 5.3 ± 2.7 | 0.015 |
| Body mass index (kg/m 2 ) | 24.2 ± 2.4 | 24.9 ± 2.4 | <0.001 |
| Smoker | 25.2% | 31.1% | 0.003 |
| Alcohol intake (≥3 drinks/week) | 5.2% | 5.4% | 0.972 |
| Systolic blood pressure (mmHg) | 116.1 ± 11 | 123.3 ± 10 | <0.001 |
| Diastolic blood pressure (mmHg) | 74.5 ± 8 | 79.3 ± 7 | <0.001 |
| Mean blood pressure (mmHg) | 88.3 ± 9 | 93.9 ± 7 | <0.001 |
| Total cholesterol (mg/dl) | 200.7 ± 33 | 205.3 ± 34 | 0.012 |
| High-density lipoprotein cholesterol (mg/dl) | 49.4 ± 11 | 49.6 ± 12 | 0.760 |
| Low-density lipoprotein cholesterol (mg/dl) | 126.0 ± 31 | 130.8 ± 32 | 0.004 |
| Triglycerides (mg/dl) | 148.4 ± 82 | 147.1 ± 74 | 0.774 |
| Glucose (mg/dl) | 95.3 ± 9 | 96.3 ± 10 | 0.073 |
| White blood cell count (×10 9 cells/l) | 6.0 ± 1.6 | 6.1 ± 1.6 | 0.633 |
| Resting heart rate (beats/min) | 62.5 ± 8 | 63.8 ± 8 | 0.006 |
| Maximal systolic blood pressure (mmHg) | 172.1 ± 22 | 182.7 ± 21 | <0.001 |
| Peak oxygen consumption (ml/kg/min) | 35.0 ± 5.1 | 34.4 ± 4.9 | 0.017 |
During 18,923 patient-years of follow-up, 371 men developed hypertension (incidence rate 19.6 per 1,000 patient-years). The incidence rates of hypertension per 1,000 patient-years were higher in men with SBP changes (>0 mm Hg, 23.3 per 1,000 patient-years) and DBP changes (>7 mm Hg, 24.7 per 1,000 patient-years) compared with men who had SBP changes (≤0 mm Hg, 16.8 per 1,000 patient-years) and DBP changes (≤7 mm Hg, 15.6 per 1,000 patient-years), respectively.
Table 2 lists the relative risks and 95% confidence intervals (CIs) of the incidence of hypertension according to changes in SBP and DBP categories. The relative risk for incident hypertension in men with greater changes in SBP and DBP versus men with less changes in SBP and DBP were 1.70 (95% CI, 1.37 to 2.12) and 2.23 (95% CI, 1.76 to 2.82), respectively, after adjustment for confounding factors. Each 1 mm Hg increment in SBP and DBP was associated with 4% (95% CI 1.03 to 1.05) and 6% (95% CI 1.05 to 1.09) higher incidence of hypertension after adjustment for confounding factors.
