Participants.

We recruited subjects older than 40 yr through the Ryugasaki city's public relations magazine. In the advertisement for subjects, we specified that the inclusion criteria of this study included the absence of any disorders in which exercise is contraindicated (e.g., unstable ischemia, acute low back pain, etc.). In the cross-sectional study, subjects who had treated or untreated hypertension [systolic blood pressure (SBP)/diastolic blood pressure (DBP) ≥ 140/90 mmHg], were on hormone replacement therapy, or smoked tobacco were excluded, because these factors can affect associations between physical activity level and blood pressure. The formulas to calculate daily physical activity level used in this study were different between men and women. There are sex differences in maximal oxygen uptake (32). These sex differences might affect correlations between blood pressure and physical activity level or maximal oxygen uptake. Since the number of male subjects who met the inclusion criteria was only 18, the remaining 57 women between 45 and 74 yr of age were analyzed (Table 1). In the interventional study, 76 men and women chose the control or training group and volunteered to participate in this study. However, only 66 subjects between 46 and 74 yr of age enrolled and completed the study protocol (Table 2). Since the aim of interventional study was to compare blood pressure before and after the intervention period, men and women were included. Subjects refrained from alcohol consumption and intense physical activity starting on the day before testing and caffeine consumption on the day of testing.

A power calculation was performed for Pearson's correlation coefficients between daily physical activity level and SBP during exercise in the cross-sectional study and for repeated measures two-way ANOVA in SBP during exercise of the intervention study using G*Power 3 (9). The sample size of this study was enough to detect the exercise effects at 90% power and with an α of 5% when effect size was assumed as intermediate between the medium and large (4).

This study was approved by the Ethics Committee of Ryutsu Keizai University and conformed to the principles of the Helsinki Declaration. All participants gave their written, informed consent before study participation.

Resistance exercise test protocol.

Subjects performed two sets of 10 repetitions of the one-hand arm curl exercise at 20% of 1 RM using an arm curl bench and dumbbells. Each repetition was performed for 8 s (3 s of concentric contraction, 1 s of maintaining full flexion, 3 s of eccentric contraction, and 1 s of maintaining a slightly flexed position), according to the Exercise Guide 2006 published by the Ministry of Health, Labour and Welfare of Japan. The interset recovery period was 100 s. Subjects were instructed to continue breathing normally during the exercise. After completion of the 20% 1-RM session, 160 s elapsed before the 40% 1-RM session commenced in a similar fashion. In the exercise test after the intervention, postintervention 1 RM was used to determine exercise intensity.

Blood pressure and heart rate measurement.

Resting blood pressure was measured in quintuplicate by oscillometry after at least 30 min of rest (DINAMAP; GE Healthcare, Buckinghamshire, UK). A three-lead electrocardiogram (ECG) was recorded at 1,000 Hz to calculate the heart rate (HR; LRR-03; GMS, Tokyo, Japan). During resistance exercise, blood pressure measurements were taken by oscillometry 20 s after the onset of exercise until the end of exercise. Exercise was performed with the dominant arm, and blood pressure was measured in the nondominant arm. The mean value of each parameter over two sets at 20% and 40% of 1 RM, respectively, was calculated.

Daily physical activity level assessment.

Daily physical activity levels were measured using an accelerometer (Active Style Pro; Omron Healthcare, Kyoto, Japan). This device estimates the intensity of physical activity [metabolic equivalents (METs)] over 10-s intervals based on accelerations in three dimensions (sampling rate, 32 Hz). The estimation of METs for household and locomotive activities using this device has been validated (22). The device automatically calculates physical activity levels (kcal/day) based on all levels of METs, and basal metabolic rate, estimated from age, sex, height, and body weight, but the manufacturer has not disclosed how it does so in detail. All subjects wore an accelerometer on the waist continuously for 14 days, except during sleeping and bathing. However, subjects sometimes forgot to wear the accelerometer and spent a day as unusual (e.g., staying in bed due to poor health). In addition, especially in the first few days, wearing the accelerometer could have psychological influence on behavior of subjects. Therefore, data from a continuous 7-day period that began 3–4 days after the commencement of monitoring were used to assess the level of daily physical activity (34). The accelerometer measures nonwear time, defined as periods with no acceleration for longer than 20 min. All subjects wore the accelerometer over 10 h per day for at least 7 days.

Maximal oxygen uptake estimation.

Three-lead ECG (LRR-03; GMS) and breath-by-breath oxygen uptake (AE300S; Minato Medical Science, Osaka, Japan) were monitored during incremental cycling [4 min at 30 W, with a 20 W (male) or 15 W (female) increase every 2 min to 85% of age-predicted maximum HR: 208 − 0.7 × age (37)]. The linear regression line between HR and oxygen uptake was determined using values from the last 30 s of each workload. Oxygen uptake corresponding to the maximum HR was considered the maximal oxygen uptake.

1-RM assessment.

The one-hand arm curl 1 RM was assessed using an arm curl bench and dumbbells. Subjects performed a single repetition of the exercise using their dominant arm with progressively heavier weights; the heaviest weight that a subject could lift once through a complete range of movement was considered to be his or her individual 1 RM.

Blood chemical analysis.

Blood samples were collected from the antecubital vein by venipuncture after an overnight fast. Serum concentrations of cholesterol, triglycerides, and insulin, as well as plasma level of glucose, were determined according to our laboratory's previous study (23). In the intervention study, serum high-sensitivity C-reactive protein (hsCRP) concentrations and creatine kinase (CK) activity were measured by using a commercial immunonephelometry kit (CardioPhase hsCRP; Siemens Healthcare Diagnostics, Tokyo, Japan) and the Japan Society of Clinical Chemistry transferable method kit (L-type CK; Wako Pure Chemical Industries, Osaka, Japan). In addition, plasma levels of interleukin-6 (IL-6) were determined by using an enzymatic kit (QuantiGlo Human IL-6, R&D systems, Minneapolis, MN).

Arterial stiffness assessment.

In the interventions study, brachial-ankle pulse-wave velocity (baPWV) was measured using air plethysmography, as in our laboratory's previous studies (24, 25). Briefly, in supine position, brachial and posttibial artery pulse waves were obtained in triplicate (BP-203RPE II; Omron Colin, Tokyo, Japan). Pulse wave transit time was automatically determined by the device, as the delay between the proximal and distal waveforms. The distance traveled by the pulse waves was calculated based on each subject's height. BaPWV was calculated as the distance divided by the transit time.

Exercise intervention.

Subjects in the training group underwent supervised walking (35–50 min) once per week for 6 wk. Initially, subjects walked at a relatively low intensity (60–65% maximal HR). As their exercise tolerance improved, the intensity of walking was increased to 75% maximal HR. In addition, subjects walked two to four times a week on their own at the same pace as the supervised walking and recorded the actual duration of their walks. Adherences to the prescribed walking frequency and time were calculated without distinction between supervised and unsupervised walks. Mean HR during walking was calculated based on HR during supervised walking. Subjects in the control group participated in a muscle stretching class (30–45 min) at local community centers every other week.

Statistical analysis.

Results are expressed as means ± SE. Relationships between two variables were investigated using Pearson's correlation coefficients and partial correlation analysis. Stepwise regression was used to identify independent predictors of SBP during exercise. In the cross-sectional study, changes in blood pressure and HR during exercise were evaluated using repeated-measures one-way ANOVA. Effects of intervention on blood pressure during exercise and other variables were tested using repeated-measures two-way ANOVA. If a significant F value was found, a Fisher's post hoc test was performed. The unpaired t-test or χ² test was used to detect intergroup differences in variables before the intervention, as appropriate. P values < 0.05 were considered statistically significant.

Cross-sectional study.

SBP (rest, 114 ± 2 mmHg; 20% 1 RM, 136 ± 3 mmHg; 40% 1 RM, 142 ± 3 mmHg), DBP (66 ± 1, 78 ± 1, and 81 ± 2 mmHg, respectively), HR (65 ± 1, 75 ± 1, and 78 ± 1 beats/min, respectively), and rating of perceived exertion (Borg's 6–20 scale; 20% 1 RM, 10.6 ± 0.2; 40% 1 RM, 13.0 ± 0.2) increased during resistance exercise in an exercise-intensity dependent manner.

Physical activity level (Fig. 1) and maximal oxygen uptake (Fig. 2) were correlated with SBP during resistance exercise at 20% and 40% 1 RM. In partial correlation analysis, physical activity level and maximal oxygen uptake were correlated with SBP at 20% and 40% 1 RM, independent of resting SBP (partial r = −0.318 to −0.464, P = 0.01-0.0002). Physical activity level or maximal oxygen uptake, age, and SBP at rest were identified as independent predictors of SBP during exercise in the stepwise regression analysis that included 1 RM, body mass index (BMI), DBP at rest, and blood levels of cholesterol, triglycerides, glucose, and insulin (Table 3).

Physical activity level was not associated with DBP and HR both at rest and during resistance exercise. Maximal oxygen uptake was correlated with DBP (rest, r = −0.389, P = 0.002; 20% 1 RM, r = −0.480, P = 0.0001; and 40% 1 RM, r = −0.432, P = 0.0007) and HR (r = −0.286, P = 0.03; r = −0.366, P = 0.004; and r = −0.370, P = 0.004; respectively). In partial correlation analysis, only DBP at 20% 1 RM was correlated with maximal oxygen uptake, independent of the resting value (partial r = −0.306, P = 0.02).

Interventional study.

There were no intergroup differences in male-to-female ratio, age, BMI, laboratory values, and exercise parameters before the intervention (Table 2). Subjects in the training group exercised for an average of 4.3 ± 0.3 days/wk and 55.6 ± 4.1 min/day at 70.7 ± 1.2% of maximal HR. Subjects in the control group participated in the stretching class for an average of 0.3 ± 0.3 day/wk. The subjects did not change their lifestyle during the experimental period, except for participation in this study. There was an interaction between group and training period in maximal oxygen uptake; maximal oxygen uptake increased in the training group after the intervention (Table 2).

An interaction between group and exercise intensity was identified in SBP (Fig. 3). In the training group, SBP at 20% and 40% 1 RM were lower after the intervention compared with before the intervention, whereas there were no differences in SBP in the control group before and after the intervention. After the intervention, SBP at 20% and 40% 1 RM were lower in the training group than in the control group, but there were no intergroup differences in SBP before the intervention. There were no differences in resting SBP between groups and before and after the intervention. ANOVA demonstrated that trends in DBP changes during exercise were affected by the exercise training, although these multiple comparisons did not reach statistical significance. Results in women were similar to results in all subjects, although the decrease in SBP at 40% 1 RM of the training group from before to after the intervention did not reach statistical significance (P = 0.05).

There was an interaction between group and training period in baPWV; baPWV in the training group decreased after the training period compared with baseline (Fig. 4). Changes in baPWV of the training group from before to after the intervention were correlated with the changes in SBP at 20% (r = 0.488, P = 0.003) and at 40% (r = 0.488, P = 0.003) 1 RM.