Resistance exercise lowers blood pressure and improves vascular endothelial function in individuals with elevated blood pressure or stage-1 hypertension

Abstract

Lifestyle modifications are the first-line treatment recommendation for elevated blood pressure (BP) or stage-1 hypertension (E/S1H) and include resistance exercise training (RET). The purpose of the current study was to examine the effect of a 9-wk RET intervention in line with the current exercise guidelines for individuals with E/S1H on resting peripheral and central BP, vascular endothelial function, central arterial stiffness, autonomic function, and inflammation in middle-aged and older adults (MA/O) with untreated E/S1H. Twenty-six MA/O adults (54 ± 6 yr; 16 females/10 males) with E/S1H engaged in either 9 wk of 3 days/wk RET (n = 13) or a nonexercise control (Con; n = 13). Pre- and postintervention measures included peripheral and central systolic (SBP and cSBP) and diastolic BP (DBP and cDBP), flow-mediated dilation (FMD), carotid-femoral pulse wave velocity (cfPWV), cardiovagal baroreflex sensitivity (BRS), cardiac output (CO), total peripheral resistance (TPR), heart rate variability (HRV), and C-reactive protein (CRP). RET caused significant reductions in SBP {mean change ± 95% CI = [−7.9 (−12.1, −3.6) mmHg; P < 0.001]}, cSBP [6.8 (−10.8, −2.7) mmHg; P < 0.001)], DBP [4.8 (−10.3, −1.2) mmHg; P < 0.001], and cDBP [−5.1 (−8.9, −1.3) mmHg; P < 0.001]; increases in FMD [+2.37 (0.61, 4.14)%; P = 0.004] and CO [+1.21 (0.26, 2.15) L/min; P = 0.006]; and a reduction in TPR [−398 (−778, −19) mmHg·s/L; P = 0.028]. RET had no effect on cfPWV, BRS, HRV, or CRP relative to Con (P ≥ 0.20). These data suggest that RET reduces BP in MA/O adults with E/S1H alongside increased peripheral vascular function and decreased TPR without affecting cardiovagal function or central arterial stiffness.

NEW & NOTEWORTHY This is among the first studies to investigate the effects of chronic resistance exercise training on blood pressure (BP) and putative BP regulating mechanisms in middle-aged and older adults with untreated elevated BP or stage-1 hypertension in a randomized, nonexercise-controlled trial. Nine weeks of resistance exercise training elicits 4- to 8-mmHg improvements in systolic and diastolic BP alongside improvements in vascular endothelial function and total peripheral resistance without influencing central arterial stiffness or cardiovagal function.

INTRODUCTION

It is estimated that roughly 1.39 billion people have hypertension worldwide and that nearly 10 million deaths are attributed to hypertension-related complications annually (1, 2). There are multiple categories that define a higher than optimal blood pressure (BP), including elevated BP [systolic BP (SBP) of 120–129 mmHg and a diastolic BP (DBP) of less than 80 mmHg], stage-1 hypertension (SBP of 130–139 mmHg and/or a DBP of 80–89 mmHg), and stage-2 hypertension (SBP over 140 mmHg or a DBP over 90 mmHg) (2). Lifestyle interventions are the recommended first-line treatment for individuals who have elevated BP and stage-1 hypertension (E/S1H) in an attempt to prevent or delay the need for future pharmaceutical intervention (3). Resistance exercise training (RET) effectively reduces BP and is one of the recommended lifestyle interventions for individuals with E/S1H (4). Notably, however, the majority of studies examining the impact of RET on BP include BP only as a secondary outcome, and the mechanism by which RET lowers BP is unclear (5), particularly in middle-aged (MA/O) adults with untreated E/S1H. Therefore, there is a clear and urgent need for high-quality investigations examining the effects of RET on BP and putative BP-regulating mechanisms in individuals with E/S1H (5).

The limited body of evidence available on the mechanism by which RET lowers BP suggests that improvements in peripheral vascular endothelial function may be a major contributing factor (5, 6). However, the only study examining RET’s impact on vascular function in untreated MA/O adults with hypertension did not have a true nonexercise control group, used a short 4-wk RET program, did not measure FMD, and included individuals with stage-2 hypertension under the updated BP categories (7). In addition, although endothelial dysfunction and hypertension are related, the directionality and causality of this association is less clear (8). Interestingly, RET has been shown to worsen central arterial stiffness and cause alterations in autonomic nervous system function that result in impaired BP control in some (7, 9–17), but not all studies (18–24). These effects would be particularly worrisome in MA/O adults with E/S1H and would also seemingly be at odds with the potential beneficial effects of RET on BP. However, the effects of RET on central arterial stiffness and autonomic function in this population are, as of yet, not well described.

Therefore, the purpose of the current study was to examine the effect of a 9-wk RET intervention in accordance with current exercise guidelines provided by the American College of Sports Medicine (ACSM) for individuals with high BP (25) on resting peripheral and central BP, vascular endothelial function, central arterial stiffness, cardiovagal function, and inflammation in MA/O with E/S1H.

METHODS

Participants

Thirty-six individuals completed a screening visit for the current study, however, six of these individuals had BPs below the inclusion cutoff and thus were not eligible. Therefore, 30 physically inactive, male and female MA/O adults (aged, 45–64 yr) were determined to be eligible for the current study and were randomized into either a RET or a nonexercise control (Con) group (Table 1). Out of the enrolled participants, there were four individuals in RET and six individuals in Con who met ES1H criteria based on their SBP alone, whereas the remaining participants were classified as ES1H because of both their SBP and DBP. Following screening but before the first experimental visit, two participants in Con and one in RET dropped out of the study for the following reasons: scheduling conflicts (n = 1), unrelated injury (n = 1), and unresponsiveness to study-related communication (n = 1), whereas one participant in Con dropped out following the first experimental visit because of scheduling conflicts. As a result, 26 participants completed this investigation (Table 1). To determine eligibility, participants arrived at the laboratory for a screening visit after a 6 h fast, abstaining from caffeine consumption for at least 12 h, and having not engaged in moderate-to-vigorous intensity exercise for at least 24 h. During the screening visit, participants completed an informed consent form, health history questionnaire, and the Physical Activity Readiness Questionnaire (PAR-Q+), had their menopause status determined via Stages of Reproductive Aging Workshop +10 principal criteria (26), and had a resting brachial BP measured. During the screening visit, BP was recorded as the average of duplicate recordings following a 5-min seated rest period with a 1 min interval between measurements. For all measurements, participants were seated with an appropriately sized cuff placed on the upper right arm, which was supported and at heart level. The screening BP was collected using an automatic oscillometric BP device (OMRON Platinum Model BP5450, OMRON Healthcare Co., JPN) that has been validated for clinical accuracy and is on the US BP Validated Device Listing. If the first and second SBP or DBP measurements differed by ≥ 5 mmHg or ≥ 4 mmHg, respectively, a third measurement was completed, and the average of the two closest measurements was recorded. To be eligible, participants must have been 45–64 yr old, had a body mass index of 18.5–39.9 kg/m², have not been meeting the physical activity guidelines for at least 6 mo, have been determined to have no known cardiovascular, metabolic, or musculoskeletal disease (excluding hypertension), nor to be taking any medications treating such disease, according to self-reported health history, determined to be ready to begin an exercise program according to the PAR-Q+, and had a SBP between 120 and 139 mmHg and/or DBP between 80 and 89 mmHg as measured during the in-person screening visit. Participants were recruited using Institutional Review Board-approved emails via the university mass email system and by word of mouth. All study procedures and documents complied with the Declaration of Helsinki, except for preregistration in a publicly accessible database, and were approved by the University’s Institutional Review Board for the protection of human subjects (IRB Approval No.: 202201319). All participants consented to participate by signing an informed consent form explaining the nature, benefits, and risks of the study before participation.

Table 1.

Baseline participant characteristics

	RET	Control
Participants; females/males, n	13; 8/5	13; 8/5
Age, yr	52 (6)	55 (6)
Height, cm	171.5 (10.5)	171.5 (7.4)
Weight, kg	84.6 (9.7)	81.3 (15.5)
BMI, kg/m2	29.9 (4.2)	28.0 (4.3)
Postmenopausal, n	6	6
Years postmenopause, yr	6 (7)	8 (4)
Race, n
White	11	11
Asian	1	1
Black	1	1
Elevated blood pressure
Participants, n	4	5
SBP, mmHg	122 (1)	125 (3)
DBP, mmHg	73 (9)	74 (9)
Stage-1 hypertension
Participants, n	9	8
SBP, mmHg	130 (5)	133 (6)
DBP, mmHg	84 (3)	85 (8)

Values are means (SD); n, number of participants. RET, resistance training group; Control, nonexercised control group; BMI, body mass index; DBP, diastolic blood pressure; SBP, systolic blood pressure.

Experimental Design

An overview of the experimental design can be viewed in Fig. 1. All eligible participants visited the laboratory for two experimental visits occurring 3–7 days before (T0) and 5–7 days following (T1) the 9-wk intervention period. For all experimental visits, participants abstained from exercise for at least 5 days, caffeine for 12 h, and food for 6 h before arrival, and were studied at the same time of day during pre- and posttesting. During the experimental visits, participants had their body composition measured before laying supine for at least 10 min before a venipuncture. An additional 10-min supine rest was then provided before vascular endothelial function and central arterial stiffness was measured via FMD and carotid-femoral pulse wave velocity (cfPWV), respectively. Participants then transitioned into a seated position and sat quietly for 10 min before having their resting BP collected and their beat-by-beat BP and heart rate recorded for 5 min. At the end of each visit, participants engaged in strength testing, where their 10-repetition maximum (RM) was determined using a cable-loaded bench press and plate-loaded hack squat machine. Before strength testing, participants were provided with a 180-kcal hypoallergenic snack (Organic Strawberry Crispy Squares, MadeGood Foods). There were two premenopausal participants in each group. Two premenopausal women had an intrauterine device (RET, n = 1; Con, n = 1), whereas the other two completed their experimental visits in the follicular phase (RET, n = 1; Con, n = 1) to control for changes in circulating sex hormones.

Figure 1.

Overview of the experimental design. In week 1, load (10RM) for the prescribed repetitions was determined based on baseline strength testing. During weeks 2 and 3, loads (12RM) were initially determined by baseline strength testing and adjusted (+5–10%) whenever participants felt they could complete 2 repetitions more than prescribed on the final set of each exercise. During weeks 4–9, participants completed as many repetitions as possible during their final set of each exercise. Whenever a participant completed 2 or more repetitions than prescribed for 2 consecutive training sessions, the load was increased by 5–10% for the next exercise training session (e.g., 2 + 2 rule). Images created with a licensed version of BioRender.com.

A priori sample size determination.

The estimated sample size required to observe mixed-factors interaction effect for changes in BP, cfPWV, and FMD, was determined in G*Power (Autenzell, Germany). Collier et al. (7) observed an effect size of RET on BP and cfPWV of d = 0.65 and d = 0.67, whereas Ramirez-Valez et al. (23) observed an effect size of RET on FMD of 0.51. Combining these effect sizes with a standard power (1-β) of 0.8, two groups (RET and Con), and two measurements (T0 vs. T1) with a conservative correlation between measurements of 0.5, it was determined that 8, 8, and 10 participants would be needed per group to achieve adequate power for BP, cfPWV, and FMD outcomes, respectively. Therefore, assuming a 20% dropout rate, we aimed to recruit at least 12 participants for each group.

Intervention period.

During the 9-wk intervention period, individuals in the RET group came to the laboratory every Monday, Wednesday, and Friday to complete a ∼40-min RET session. Each RET session was led and supervised by one of three trained laboratory members and not more than four participants were trained in the same sessions to maintain adequate supervision while allowing sufficient flexibility to accommodate participants’ schedules. During each RET session, participants completed (in order) bench press, hack squat, latissimus dorsi pulldown, leg extension, seated row, leg curl, and plank resistance exercises. Specific set and repetition schemes for the RET program are shown in Fig. 1. All exercises were cable-loaded except for the hack squat, which was plate-loaded, and the plank, which was completed using only body weight. The initial weight was estimated based on strength testing at the end of T0 and used for week 1. During weeks 2 and 3, participants were queried and monitored to determine if they would have been able to complete two or more repetitions beyond the prescribed 12 repetitions on the final set of each lift. If this was the case, the weight was increased by 5–10% for the following workout. From week 4 and onward, participants were instructed to complete as many repetitions as possible on the final set of each lift (excluding plank) but were stopped if they completed three repetitions more than the prescribed amount. Weight was then increased by 5–10% if participants completed more than the prescribed repetitions on the final set for two workouts in a row for a particular lift. For planks, participants completed the same number of sets as the other exercises, but did so for time instead of repetitions, with a prescription of 15 s for week 1 and 20 s for week 2, followed by a 5 s increase in time for each workout thereafter. Participants rested for at least 1 min, but not longer than 3 min, between sets and were instructed to allow sufficient rest to ensure lingering fatigue did not affect subsequent set performance. All sets and repetitions were completed for a given exercise before moving to the next exercise in the session during weeks 1–4. During weeks 5–9, participants completed supersets with an upper- and lower-body exercise paired (e.g., bench press and hack squat). At the end of each session, participants were asked for their rating of perceived exertion (RPE) between 0 and 10, with 0 representing no exertion (i.e., rest) and 10 representing maximal exertion. Those in Con were asked to maintain their current lifestyle and dietary habits throughout the 9 wk period and did not come to the laboratory outside of the screening and experimental visits.

Conduit artery vascular function.

The brachial artery FMD technique was used to assess vascular endothelial function and reactive hyperemia (RH) in accordance with the most recent guidelines (27). Before data collection, participants laid in a supine position in a dark, temperature-controlled room for 10 min. With the participant’s left arm laterally extended, a segmental cuff (TMC7, Hokanson) was placed just distal to the medial epicondyle of the humerus, and a 12-MHz ultrasound probe (12 L-RS, General Electric) was used to visualize the brachial artery and measure blood flow, whereas a screen capture device (AV.io HD, Epiphan Systems) was used to record the ultrasound screen. Positioning of the segmental cuff distally to the ultrasound probe was chosen because the increase in postocclusive artery diameter is largely NO-mediated using this technique (28). Blood flow velocity was collected using an insonation angle of 60° to the axis of the vessel and a sample volume encompassing the entire width of the artery (29). FMD testing included a 2-min baseline period, a 5-min cuff occlusion period at 240 mmHg using a rapid cuff inflation system (E20, Hokanson), and a 3-min postocclusive period. A previously validated (30), continuous, semiautomated edge detection software (FMD Studio, Quipu srl, Italy) was used to continuously measure brachial artery diameter and blood flow velocity throughout the protocol, which were used to calculate the shear rate [4 × blood flow velocity (cm/s)/brachial diameter (cm)] (27). Baseline diameter (D_base) and shear rate (SR_base) were calculated as the average value during the 2-min baseline period, whereas peak diameter, shear rate (SR_peak), and shear rate area under the curve (SR_AUC) were calculated following cuff release up until peak diameter was observed using FMD Studio software, as previously described (31–33). RH was calculated as the difference in AUC of blood flow between baseline (BF_base) and the first 90 s of the postocclusive period. Relative and absolute FMD (FMD_abs) were calculated as relative (%) and absolute (mm) the change from D_base to maximal diameter, whereas FMD normalized to SR (FMD_SR) was calculated as FMD/SR_AUC. Probe location was measured from the superior border of the antecubital fossa during pretesting to ensure a similar placement of the probe during posttesting.

Carotid-femoral pulse wave velocity.

Central arterial stiffness was assessed using cfPWV (SphygmoCor XCEL, AtCor Medical, Inc.). While remaining in a supine position following the FMD test, participants had their carotid pulse palpated and marked on the left side of the neck and had a cuff placed on their upper left thigh to acquire the femoral pulse wave via volumetric displacement. The pulse waves of the carotid and femoral arteries were then recorded simultaneously by a tonometer and the femoral cuff, respectively. The distance between the site of the carotid and femoral pulse was then divided by the difference in pulse wave transit time between the two arteries (e.g., distance/time) to determine cfPWV. To correct for the known impact of distending pressure on cfPWV and the hypothesized reduction in BP expected in the RET group, a change in mean arterial pressure (MAP) was added as a covariate in cfPWV analyses.

Resting blood pressure.

During experimental visits, resting BP was collected in a seated position following a 10-min resting period in accordance with the American Heart Association guidelines (34) using a SphygmoCor XCEL cuff device (SphygmoCor XCEL, AtCor Medical, Inc.). SBP, DBP, MAP, and pulse pressure (PP) were determined automatically by standard oscillometric brachial BP measurement using an appropriately sized BP cuff. Immediately after, the cuff was inflated and held at a subdiastolic pressure level for 5 s, during which cuff displacement waveforms were measured and calibrated to the brachial SBP and DBP. Next, a generalized transfer function was applied to estimate the central BP waveform, from which central SBP (cSBP), and central DBP (cDBP) were determined using the device’s proprietary software (35).

Hemodynamic monitoring.

While seated, participants had a finger photoplethysmograph placed on the middle finger of the right hand, which was used to collect beat-by-beat BP (NOVA Finometer, Finapres Medical Systems, The Netherlands). The participants held their right hand over their heart during all hemodynamic testing, with their arm supported. Model flow technology was used to calculate cardiac output (CO) and total peripheral resistance (TPR). In addition, heart rate was collected using a three-lead electrocardiogram, and respiratory rate was collected using a respiratory belt with participants instructed to breathe at a normal rate during all testing (TN1132/ST; ADInstruments). Data were collected at 1,000 Hz using a data acquisition system (PowerLab Series 26; ADInstruments) and stored offline.

Cardiovagal baroreflex sensitivity and heart rate variability.

Raw beat-by-beat BP waveforms and ECG data were uploaded to Ensemble-R software, and the sequence method was used to assess cardiovagal BRS and HRV. BRS_pooled was assessed by averaging the slope between three sequences of either increasing (BRS_up) or decreasing (BRS_down) pulse waveform peak pressures with subsequent decreases or increases in R-R interval length, respectively, with a minimum correlation of r = 0.8, increase in SBP of 1 mmHg, and an R-R interval length of 4 ms. The log-transformed root mean square of successive differences (lnRMSSD), high-frequency power (lnHF), and low-frequency power (lnLF) were calculated to represent both time and frequency domain HRV.

C-reactive protein.

Whole venous blood was collected in a lithium heparin plasma separator tube (BD Vacutainer, Becton Dickinson) before being spun for 15 min at 1,000 g. Plasma was then transferred to 1.7-mL microcentrifuge tubes for storage at −80°C. Samples were later thawed, and high-sensitivity CRP was assessed using a commercially available enzyme-linked immunosorbent assay (CRP ELISA, Immundiagnostik, Germany). The detection range of the CRP ELISA kit was 1.8–150 ng/mL, with a sensitivity of 0.124 ng/mL, and an interassay coefficient of variation of <10%. All assays were performed in accordance with the manufacturer’s instructions and read using a microplate photometer (Multiskan FC Microplate Photometer, ThermoFisher Scientific).

Body composition.

At both experimental visits, participants’ body composition was assessed via BodPod (COSMED) to assess body fat percentage (BF%), fat mass (FM), and fat-free mass (FFM).

Lifestyle controls.

All participants were asked to refrain from any other forms of exercise outside of the study and maintain their current dietary habits throughout the study period. Calories, protein, fat, and carbohydrate intake, along with physical activity in metabolic equivalent of task (MET) minutes per week were collected via self-report using 3-day dietary food logs and the Short Last 7 Days International Physical Activity Questionnaire (IPAQ), which were completed during both pre- and posttesting.

Statistical Analysis

Residual normality was assessed using Shapiro–Wilk tests, whereas homoscedasticity was assessed using Levene’s test. In the case of violation of either normality or homoscedasticity, data were transformed with a natural logarithm before being reverted to their original scale for reporting. Multiple independent two-way mixed-linear models [group (RET vs. Con) × visit (pre- vs. postintervention)] were run to determine the impact of the intervention period on all resting BP, vascular endothelial function, RH, cfPWV, HRV, BRS, dietary, and physical activity variables with sex included as a covariate. cfPWV was analyzed using a two-way mixed-linear model with ΔMAP and sex included as covariates. To decompose significant group × visit interactions, Tukey-adjusted post hoc comparisons were performed. Between-group effect sizes were determined using Cohen’s d. Pearson’s product correlations (r) or Spearman rank correlation coefficients (ρ) were used to explore the relationship between the changes in SBP, DBP, CO, TPR, FMD, and selected secondary variables of interest where residuals were normally or nonnormally distributed. Partial correlations (r_xy,z or ρ_xy,z) were also performed to remove the effect of sex and are reported in Fig. 5. Within- and between-group differences are reported in the text as mean differences with (95% CI Lower Bound, 95% CI Upper Bound) unless denoted otherwise. Confidence intervals were adjusted using the Bonferroni method. Significance was set at P ≤ 0.05. Statistical analyses were performed using JASP (JASP Team 2020, v. 0.13.1) or jamovi (v. 2.3.21.0) and figures were created using GraphPad Prism (v. 9.5.1).

Figure 5.

Relations between the changes (Δ) in SBP and flow-mediated dilation (FMD; A); SBP and FMD corrected for shear rate stimulus (cFMD_SR; B), systolic blood pressure (SBP) and resting brachial artery diameter (D_base) (C); SBP and resting blood flow (BF_base; D); total peripheral resistance (TPR) and D_base (E); TPR and cardiovagal baroreflex sensitivity down (BRS_down; F); cardiac output (CO) and fat-free mass (FFM) (G); and pulse wave velocity (PWV) and pulse pressure (PP) following a 9-wk RET program (RET; yellow-filled circles) or nonexercise control period (Con; dark gray-filled circles) (H). Note that relations between CO vs. FFM and PWV vs. PP are depicted as rank correlations because of nonnormality of residuals. Inset text boxes also display partial correlation coefficients (r_xy,z or ρ_xy,z) for the relation with the effect of sex removed. Total n = 26 for all correlation analyses presented.

RESULTS

Body Composition

There were no significant group × visit interactions, group main effects, or visit main effects for weight, FFM, FM, or BF% (Table 2).

Table 2.

Effect of resistance exercise vs. control on body composition and muscle strength

	RET		Control		Group × Visit		Group		Visit
	T0	T1	T0	T1	P	ηp2	P	ηp2	P	ηp2
Participants; males/females, n	13; 8/5		13; 8/5
Weight, kg	86.3 (6.1)	87.0 (6.1)	83.0 (6.1)	82.9 (6.1)	0.21	0.07	0.37	0.03	0.99	<0.01
FFM, kg	55.6 (3.2)	56.2 (3.2)	51.8 (3.2)	52.2 (3.2)	0.66	<0.02	0.08	0.13	0.65	<0.01
FM, kg	30.6 (4.7)	30.8 (4.7)	31.1 (4.7)	30.7 (4.7)	0.41	0.03	0.95	<0.01	0.66	<0.01
BF, %	35.6 (3.6)	37.4 (3.6)	37.4 (3.6)	36.9 (3.6)	0.61	0.01	0.51	0.02	0.39	0.03
Squat 1RM, kg	175.3 (21)	255.2 (21)#†‡	169.1 (21)	175.5 (21)	<0.001*	0.81	–	–	–	–
Bench 1RM, kg	34.2 (7.9)	56.4 (7.9)	25.2 (7.9)	25.5 (7.9)	<0.001*	0.85	–	–	–	–

Values are estimated means (±model 95% confidence intervals); n, number of participants. RET, resistance exercise training group; Control, nonexercise control group; FFM, fat-free mass; FM, fat mass; BF, body fat; 1RM, estimated one repetition maximum. P ≤ 0.05, *significant interaction effect, #significant within-group increase from T0 to T1, †significantly greater than in control at T0, or ‡significantly greater than in control at T1.

Strength

There was a significant group × visit interaction for both squat 1RM and bench 1RM. RET significantly increased both squat 1RM [+79.6 (64.1, 95.1) kg; P < 0.001] and bench 1RM [+22.7 (18.8, 26.6) kg; P < 0.001] from T0 to T1. In addition, both squat 1RM [+79.7 (39.8, 119.5) kg; P < 0.001; d = 2.33] and bench 1RM [+30.9 (15.6, 46.2) kg; P < 0.001; d = 2.37] were significantly greater at T1 in RET compared with Con. There were no significant changes in Con from T0 to T1 (P ≥ 0.05) (Table 2).

Blood Pressure

There were significant group × visit interactions for SBP, DBP, cSBP, cDBP, MAP, and PP (Fig. 2). Specifically, RET experienced a significant reduction in SBP [−7.9 (−12.1, −3.6) mmHg; P < 0.001], DBP [−4.8 (−10.3, −1.2) mmHg; P < 0.001], cSBP [−6.8 (−10.8, −2.7) mmHg; P < 0.001], cDBP [−5.1 (−8.9, −1.3) mmHg; P < 0.001], and MAP [−5.7 (−9.9, −2.0) mmHg; P < 0.001] from T0 to T1; however, the decrease in PP from T0 to T1 [−3.1 (−6.6, 0.5) mmHg; P = 0.089] in RET was not significant. There were no significant changes in any BP variable from T0 to T1 in Con (P ≥ 0.05). Accordingly, RET had significantly lower SBP [−9.8 (−17.3, −2.4) mmHg; P = 0.004; d = 1.52], cSBP [−9.7 (−17.0, −2.3) mmHg; P < 0.001; d = 1.52], and MAP [−6.1 (−11.9, −0.4) mmHg; P = 0.026; d = 1.22] than Con at T1. Although there were large effect size differences between RET and Con at T1 for DBP [−4.5 (−10.3, 1.2) mmHg; P = 0.142; d = 0.90], cDBP [−4.3 (−10.1, 1.4) mmHg; P = 0.173; d = 0.86], and PP [−5.3 (−11.3, 0.7) mmHg; P = 0.081; d = 1.02], these differences were not statistically significant.

Figure 2.

Peripheral systolic blood pressure (SBP), peripheral diastolic blood pressure (DBP), central SBP (cSBP), central DBP (cDBP), mean arterial pressure (MAP), and pulse pressure (PP) were collected before and following either 9 wk of resistance exercise training (RET; n = 13) or a nonexercise control period (Con; n = 13). All data are displayed as estimated marginal means (±95% CI). P ≤ 0.05, *significant within-group decrease from T0 to T1, †significantly lower in RET at T1 than in Con at T0; or #significantly lower in RET at T1 than in Con at T1.

Resting Hemodynamics

There were significant group × visit interactions for both CO and TPR (Fig. 3). CO significantly increased in RET [+1.21 (0.26, 2.15) L/min; P = 0.006], but there was no difference in CO between RET and Con at T1 [+0.9 (−0.6, 2.3); P = 0.33; d = 0.70]. TPR significantly decreased from T0 to T1 in RET [−398 (−778, −19) mmHg·s/L; P = 0.028], and while there was a large effect size difference between RET and Con at T1 [−369 (−849, 110) mmHg·s/L; P = 0.158; d = 0.87], this difference was not significant. There were no significant changes in CO or TPR in Con from T0 to T1 (P ≥ 0.64) (Fig. 3). There was no significant group × visit interaction, group main effect, or visit main effect for RHR (Table 3).

Figure 3.

Resting cardiac output (CO) and total peripheral resistance (TPR) collected before and following either 9 wk of resistance exercise training (RET; n = 13) or a nonexercise control period (Con; n = 13). All data are displayed as estimated marginal means (±95% CI). *P ≤ 0.05, significant within-group increase from T0 to T1.

Table 3.

Effect of resistance exercise vs. control on central arterial stiffness, vascular function and reactive hyperemia, inflammation, and autonomic function

	RET		Control		Group × Visit		Group		Visit
	T0	T1	T0	T1	P	ηp2	P	ηp2	P	ηp2
Participants; males/females, n	13; 8/5		13; 8/5
cfPWV, m/s	7.0 (0.5)	6.8 (0.6)	7.3 (0.6)	7.2 (0.5)	0.20	0.07	0.72	<0.01	0.23	0.07
Dbase, mm	3.53 (0.2)	3.64 (0.2)	3.76 (0.2)	3.73 (0.2)	0.07	0.14	0.22	0.06	0.39	0.03
FMDabs, mm	0.24 (0.1)	0.33 (0.1)#†‡	0.20 (0.1)	0.21 (0.1)	0.006*	0.29	–	–	–	–
SRbase, s−1	164.9 (40.2)	202.0 (40.2)	165.9 (40.2)	192.3 (40.2)	0.73	<0.01	0.81	<0.01	0.71	<0.01
SRpeak, s−1	1044.8 (140)	1031.3 (140)	937.2 (140)	1003.9 (140)	0.32	0.04	0.46	0.03	0.69	<0.01
SRAUC, AU·10−3	2.04 (0.5)	2.17 (0.5)	2.11 (0.5)	2.11 (0.5)	0.58	0.01	0.99	<0.01	0.93	<0.01
CRP, mg/L	3.16 (1.4)	2.17 (1.3)	2.23 (1.4)	2.25 (1.3)	0.33	0.05	0.39	0.04	0.75	<0.01
BRSpooled, ms/mmHg	5.83 (1.1)	5.87 (1.1)	5.22 (1.1)	5.14 (1.1)	0.81	<0.01	0.37	0.03	0.12	0.10
BRSup, ms/mmHg	5.67 (1.2)	5.60 (1.2)	4.72 (1.2)	5.24 (1.2)	0.39	0.03	0.41	0.03	0.47	0.02
BRSdown, ms/mmHg	6.19 (1.2)	5.98 (1.2)	5.41 (1.2)	5.01 (1.2)	0.78	<0.01	0.25	0.06	0.28	0.05
lnRMSSD, ms	3.18 (0.2)	3.09 (0.2)	3.08 (0.2)	2.97 (0.2)	0.92	<0.01	0.43	0.03	0.22	0.07
lnLF, ms2	5.21 (0.7)	4.97 (0.7)	5.09 (0.7)	4.65 (0.7)	0.63	0.01	0.60	0.01	0.37	0.04
lnHF, ms2	4.78 (0.6)	4.32 (0.6)	4.24 (0.6)	4.10 (0.6)	0.32	0.04	0.28	0.05	0.11	0.11
RHR, beats/min	68.9 (5.1)	70.2 (5.1)	71.8 (5.1)	72.1 (5.1)	0.65	<0.01	0.48	0.02	0.66	<0.01

Values are means (±95% confidence intervals); n, number of participants. AU, arbitrary units; AUC, area under the curve; BRS, cardiovagal baroreflex sensitivity; Control, nonexercise control group; RET, resistance exercise training group; D_base, baseline diameter; D_max, maximal diameter after cuff release; lnRMSSD, log-transformed root mean square of successive differences; lnHF, log-transformed high-frequency power; lnLF, log-transformed low-frequency power; RHR, resting heart rate; SR, shear rate. P ≤ 0.05, *significant interaction effect, #significant within-group increase from T0 to T1, †significantly greater than in Con at T0, or ‡significantly greater than in Con at T1.

Conduit Artery Vascular Function

There were significant group × visit interactions for FMD (P = 0.011; η_p² = 0.25), cFMD_SR (P = 0.017; η_p² = 0.22), FMD_abs (P = 0.006; η_p² = 0.29), as well as BF_base (P = 0.023; η_p² = 0.20). There was no significant interaction for D_base (P = 0.67; η_p² = 0.14). RET experienced significant increases from T0 to T1 in FMD [+2.37 (0.61, 4.14)%; P = 0.004], cFMD_SR [+1.25 (0.23, 2.27)%; P = 0.009], and FMD_abs [+0.09 (0.03, 0.16) mm; P = 0.002]. There were no significant changes in any of these variables from T0 to T1 in Con (all P ≥ 0.99). Consequently, ΔcFMD_SR [+2.1 (0.04, 4.09)%; P = 0.035; d = 1.18] and FMD_abs [+0.14 (0.005, 0.283) mm; P = 0.048; d = 1.12] were greater in RET than Con at T1, whereas there was a large, but nonsignificant difference in FMD [+3.8 (−0.7, 8.4)%; P = 0.104; d = 0.98]. RET also experienced significant increases from T0 to T1 in BF_base [+28.6 (5.1, 52.1); P = 0.009]. Although there was a large effect size difference between RET and Con at T1 for BF_base [+26.4 (−3.6, 56.4); P = 0.085; d = 1.00], this difference was not significantly different. There were no significant group × visit interactions, group main effects, or visit main effects for RH, SR_base, SR_peak, or SR_AUC (Fig. 4 and Table 3).

Figure 4.

Percent flow-mediated dilation (FMD), FMD corrected to shear rate (cFMD_SR), reactive hyperemia (RH), and baseline blood flow (BF_base) collected before and following either 9 wk of resistance exercise training (RET; n = 13) or a nonexercise control period (Con; n = 13). All data are displayed as estimated marginal means (±95% CI). P ≤ 0.05, *significant within-group increase from T0 to T1, or #significant difference between RET and Con at T1.

Central Arterial Stiffness

There was no significant group × visit interaction, group main effect, or visit main effect for cfPWV (Table 3).

Cardiovagal Baroreflex Sensitivity and Heart Rate Variability

An average of 26.9 ± 10.7 valid sequences were acquired per participant at the pre- and postintervention visits. There was no significant group × visit interaction, group main effect, or visit main effect for BRS_pooled, BRS_up, or BRS_down (Table 3). There was also no significant group × visit interaction, group main effect, or time main effect for lnRMSSD, lnLF, or lnHF (Table 3).

C-Reactive Protein

There was no significant group × visit interaction, group main effect, or time main effect for CRP (Table 3).

Correlations

Relations among the changes in hemodynamic, vascular, cardiovagal, and body composition variables are depicted in Fig. 5. ΔSBP was significantly correlated with ΔFMD (r = −0.48; P = 0.012), ΔcFMD_SR (r = −0.54; P = 0.005), ΔBF_base (r = −0.47; P = 0.016), and ΔD_base (r = −0.45; P = 0.021), as well as with Δsquat (ρ = −0.52; P = 0.007), and Δbench (r = −0.67; P < 0.001). ΔDBP was significantly correlated with Δbench (r = −0.49; P = 0.012). ΔCO was significantly correlated with ΔFFM (ρ = 0.58; P = 0.002), as well as with ΔTPR (r = −0.69, P < 0.001). ΔTPR was also significantly correlated with ΔD_base (r = −0.44; P = 0.023) and ΔBRS_down (r = −0.47; P = 0.016). ΔTPR was not related to ΔSBP (r = 0.26; P = 0.20) or ΔDBP (r = 0.30; P = 0.13). ΔPP was significantly related to ΔPWV (ρ = −0.51; P = 0.008).

Lifestyle Controls

The average total exercise session attendance was (means ± SD) 96 ± 6%. There were no significant group $\times$ visit interactions, group main effects, or time main effects for physical activity, nor the consumption of calories, fat, carbohydrates, or protein (Table 4).

Table 4.

Dietary and physical activity control data pre- and postintervention in the resistance exercise and control groups

	RET		Control		Group × Visit		Group		Visit
	T0	T1	T0	T1	P	ηp2	P	ηp2	P	ηp2
Participants; Males/Females, n	13; 8/5		13; 8/5
Physical activity, MET·min/wk	395.7 (148)	406.2 (148)	445.2 (148)	472.3 (149)	0.58	0.01	0.65	<0.01	0.46	0.02
Calories, kcal	2,135.7 (272)	2,161.8 (272)	1,995.9 (272)	2,027.5 (272)	0.95	<0.01	0.45	0.03	0.62	0.01
Carbohydrate, g	224.4 (34.3)	226.7 (34.3)	185.6 (34.3)	198.3 (34.3)	0.53	0.02	0.14	0.09	0.42	0.03
Protein, g	101.5 (15.7)	98.8 (15.7)	97.6 (15.7)	104.0 (15.7)	0.22	0.06	0.95	<0.01	0.35	0.04
Fat, g	92.6 (17.1)	96.1 (17.1)	97.6 (17.1)	94.5 (17.1)	0.44	0.03	0.87	<0.01	0.67	<0.01

Values are estimated marginal means (±model 95% confidence intervals); n, number of participants. Control, nonexercise control group; RET, resistance exercise training group; MET, metabolic equivalent of task.

DISCUSSION

To our knowledge, this was the first study to explore the putative vascular mechanisms driving RET-induced improvements in BP in MA/O adults with untreated E/S1H. The main finding of the current study was that a 9-wk RET program reduced peripheral and central BP, which was accompanied by an increase in FMD and a decrease in TPR. In addition, we reported increases in BF_base and CO, with no changes in either RH or cfPWV. Finally, the RET intervention did not cause any changes in autonomic function as measured by cardiovagal BRS or HRV, or changes in systemic inflammation as reflected by CRP. Furthermore, our data indicate that chronic RET does not positively or negatively affect central arterial stiffness, autonomic function, or inflammation in MA/O adults with untreated E/S1H.

To our knowledge, this is the first study to examine the impact of a RET intervention on BP in untreated MA/O adults with untreated E/S1H. Our data indicate that RET is effective at reducing both peripheral and central BP. RET reduced SBP and DBP by 8 and 5 mmHg, respectively, which is similar to reductions reported following other RET interventions (20, 36–39), aerobic exercise interventions (40), and slightly greater than the average pharmacological reductions reported over 6 mo (41). Thus, these reductions in BP are clinically relevant, and every 5 mmHg reduction in SBP is associated with a 10% decrease in CVD risk (42). Collier et al. (7), previously reported a 4 and 4 mmHg reduction in SBP and DBP, respectively, following just 4 wk of RET in MA/O adults with hypertension. In addition, middle-aged men with untreated stage-2 hypertension experienced 16 and 12 mmHg reductions in SBP and DBP, respectively, following a 12-wk RET program that was similar to the current study (39). In younger adults with untreated E/S1H who engaged in an 8-wk RET intervention, Beck et al. (20) reported a 10 and 8 mmHg reduction in peripheral SBP and DBP, and 9 and 8 mmHg reductions in cSBP and cDBP. Our data agree with and extend this previous work and indicate that RET lowered cSBP and cDBP by 7 and 5 mmHg in MA/O with E/S1H. Furthermore, RET promoted a 3 mmHg decrease in PP, which provides important prognostic information above and beyond SBP and DBP (43–45) and is largely determined by a mismatch of distal (e.g., conduit) to proximal (e.g., abdominal aorta) arterial diameters (45). Overall, our data strengthen prior evidence regarding the effect of RET on BP in MA/O with hypertension (7) by 1) experimentally isolating the effects of RET via inclusion of a nonexercise control group, and also by 2) specifically studying MA/O adults with untreated E/S1H, for whom lifestyle interventions such as RET are explicitly recommended as a first-line strategy for BP control. Accordingly, we build upon these prior studies and present important evidence that supports RET as a viable lifestyle intervention to improve BP, showing for the first time that just 9 wk of RET in MA/O adults with untreated E/S1H exhibit reductions that are slightly greater than those experienced following 6 mo of pharmacological treatment (41).

We further extended existing evidence by examining the effect of RET on putative BP-regulating mechanisms in relation to RET-induced BP changes. BP is the product of CO and TPR and in addition to lowering BP, RET increased CO and decreased TPR in the present study. The measure that was most strongly correlated with ΔCO in the current study was ΔFFM (r = 0.56; P = 0.003). Thus, although it is tempting to speculate that these relations may be explained by a training-induced increase in blood flow demand to supply a greater volume of metabolically active tissue and/or an increase in venous capacity or return (46–49), we did not observe significant RET-induced changes in whole body FFM (50). However, we also observed a 40% RET-induced increase in resting blood flow (e.g., BF_base) and a nonsignificant 3.1% (+0.11 mm) increase in brachial artery diameter (e.g., D_base). Prior studies have suggested that RET promotes increases in resting arterial lumen size in the conduit arteries feeding active muscle beds (51), likely because of arterial remodeling and/or a decrease in resting arterial tone in response to large, repeated, chronic increases in blood flow and the resultant arterial shear stress (51, 52). Furthermore, changes in conduit artery blood flow reflect changes in the tone of the downstream resistance vessels, which are so named because they are the major arterial bed that modulates vascular resistance. Hypertension is characterized by increased peripheral resistance caused by decreased resistance vessel lumen diameters, decreased resistance vessel density due to rarefaction, and/or reduced vasomotor function (53, 54). Thus, it is likely that the increases in BF_base observed herein reflect a decreased resistance to flow in the resistance vessels, perhaps by reversal of microvascular rarefaction and decreased constrictor tone (54, 55). Notably and in support of this hypothesis, both changes in resting brachial artery diameter and blood flow were inversely associated with changes in SBP in the present study (Fig. 5). In addition, it is also plausible that RET-induced increases in conduit artery diameter contributed to decreases in PP by reducing the ratio between proximal and distal conduit artery diameters (45), although this is highly speculative and will require future investigation. Therefore, taken together, our data suggest that the BP-lowering effect of RET may be explained by decreases in TPR secondary to changes in peripheral vascular tone. However, due to the lack of significant relationships between TPR and both SBP and DBP, additional mechanistic research is necessary to confirm this hypothesis.

Multiple indices of vascular endothelial function examined in the current study also improved following the RET intervention. Our data indicated that RET elicited improvements in both FMD and cFMD_SR in MA/O with untreated E/S1H, a finding that is largely in agreement with prior work in individuals with high BP (56–58). RET causes dramatic increases in blood flow to the muscles active in each exercise (59), causing acute increases in shear stress on the vascular endothelium (60). Notably, chronic exposure to repeated, transient increases in shear stress derived from exercise promotes an endothelial phenotype that is characterized by increased endothelial NO synthase (eNOS) expression and greater NO bioavailability (61). In addition, the observed significant improvements in both FMD and cFMD_SR suggest that increases in FMD were not caused by an increase in the shear stimulus on the vascular endothelium, but rather by improvements in endothelium-dependent function (28, 62). Although macrovascular function improved, we did not see any improvements in microvascular function as measured by RH. Unlike FMD, which is endothelium-dependent (63), RH provides insight into the dilation of the downstream resistance vessels and is minimally dependent on NO (29, 64). Our data disagree with those of Heffernan et al. and Collier et al. (7, 22), who have previously reported that multiweek RET interventions improved RH. However, this difference could be due to the use of strain-gauge plethysmography versus the use of Doppler ultrasound to measure RH (7, 22) or due to a lack of power in the present study considering that RET improved RH in all but one participant. An improvement in RH may have been hypothesized given that TPR is primarily regulated at the level of the resistance vessels, but changes in resting tone rather than the response to ischemia are likely more important to resting blood pressure control. Accordingly, and as previously described, BF_base was elevated following RET, potentially suggesting greater dilation of downstream resistance vessels at rest, although we cannot rule out the possibility that the increase in BF_base was caused by elevated CO. Still, it is noteworthy that changes in SBP were inversely associated with increases in both FMD and BF_base (Fig. 5, A, B, and D). Overall, these data indicate that RET improves vascular endothelial function and resting peripheral blood flow that may contribute to RET-mediated improvements in BP, and which are also likely to improve long-term cardiovascular risk in individuals with E/S1H.

While there is consistent cross-sectional evidence suggesting that individuals who engage in RET have stiffer central arteries (65–68), the experimental evidence regarding the influence of RET on arterial stiffness is less clear. Notably, RET had no effect on cfPWV in the present study. Although the majority of studies do not report increases in central arterial stiffness following RET (18, 20–24), there are multiple reports indicating that RET may promote increased stiffness (7, 9–11). A potential methodological explanation for this discrepancy is the timing of posttest measurements, with all but one of the studies that have reported increases in arterial stiffness following a RET intervention having completed posttesting within 24 h of the final exercise session (7, 9–11, 19). However, acute RET may increase central arterial stiffness for up to 3 days alongside transiently increased SNS activity and inflammation, and decreased parasympathetic nervous system activity (69–71). Therefore, it is plausible that the increases in aortic stiffness reported in these studies are due to transient changes in response to acute exercise, but do not reflect chronic maladaptive structural changes. It is also possible that RT interventions lasting several weeks or months are either not long enough to persistently alter the stiffness of the aorta, or other uncontrolled factors may be confounding the cross-sectional findings. Nevertheless, our data, collected 5–7 days (133 ± 19 h) after the last bout of RET, indicate that short-term RET does not change central arterial stiffness as measured by cfPWV among MA/O adults with E/S1H.

Our findings also indicate that RET had no adverse effects on cardiovagal function as assessed by cardiovagal BRS and HRV. Prior evidence suggests that RET may reduce BRS in MA/O with untreated hypertension (72), and either reduces (73) or does not influence BRS in young healthy adults (74, 75). In the former study, Collier et al. (72) reported that aerobic exercise training and RET caused divergent changes in BRS in response to spontaneous decreases in BP, as well as in the low- to high-frequency HRV ratio suggesting that aerobic and resistance exercise training may have different effects on sympathovagal balance in MA/O with hypertension. The authors also reported increases in central arterial stiffness that may have resulted in a decreased ability of the aorta to return to smaller diameters during periods of decreasing BP, and thus smaller changes in baroreceptor firing and lower BRS_down (72). However, it is important to highlight that because a nonexercise control group was not included and because residual transient effects may have influenced the poststudy measurements, which were performed 24–48 h after the final exercise session (72), central questions remained regarding the effects of RET in this population. In the present study, RET did not result in chronic changes in central arterial stiffness or vagal modulation, which may explain the lack of change in cardiovagal BRS. It is also notable that while the slope of the association between changes in BP and heart rate did not change, BP did, suggesting that some degree of baroreceptor resetting to operate at lower arterial pressures may have occurred. Therefore, our data indicate that a 9-wk RET program does not adversely impact cardiovagal function in MA/O with E/S1H.

The RET intervention in the current study significantly improved both squat and bench press strength, but surprisingly did not significantly influence FFM, FM, or BF%. In contrast, Moraes et al. (39) reported that a 12-wk RET intervention promoted a 3 kg increase in FFM, a 4 kg decrease in FM, and a 4% reduction in BF% among middle-aged men with untreated stage-2 hypertension. On the other hand, Collier et al. (7) report no changes in BF% following 4 wk of RET in middle-aged to older adults with elevated BP or hypertension. RET programs typically result in immediate improvements in neuromuscular function during the first several weeks of training, with changes in FFM becoming increasingly detectable with longer durations of training (76, 77). In addition, it is likely that changes in FFM with RET are less robust among aging individuals with subclinical or clinical vascular dysfunction, which is associated with anabolic resistance (78). Thus, it is likely that neural adaptations explain the marked RET-induced improvements in strength, whereas the current program was not long enough to promote significant increases in FFM. However, this increase in strength without an increase in FFM is likely still to be of benefit relative to improved functional capacity, and possibly also given the independent relationship between strength and all-cause mortality (79). We also observed no changes in circulating CRP concentrations, which is in contrast to prior studies showing that chronic RET reduces CRP (80, 81) but may also be explained by the lack of change in body composition in this study (82, 83). Future studies may wish to directly examine whether E/S1H is associated with blunted skeletal muscle hypertrophy in response to RET in MA/O adults.

There were several limitations to our study. First, we studied a mixed sample of males and females, which included both pre- and postmenopausal women. Men may be more susceptible to potential RET-induced increases in central arterial stiffness than postmenopausal women (84), and since the current study was not powered to detect sex differences and comprised mostly females, we are unable to determine if this explained our lack of findings regarding cfPWV. However, biological sex served as a covariate in all of our analyses. In addition, the distribution of pre- and postmenopausal women was balanced between groups. Second, the short nature of the study limits our ability to understand the long-term impact of RET participation in this population. It is possible that the differences in cross-sectional and intervention data regarding the impact of RET on central arterial stiffness may result from the progressive nature of changes in stiffness, and interventions longer than 9 wk may be necessary to induce changes. In addition, participants did not have clinically diagnosed E/S1H in the current study because of the lack of an in-office clinical measurement (85), and participants were also not required to complete their screening visit at the same time as the first experimental visit. However, it should be noted that all participants had a BP consistent with E/S1H classification at both their screening and first experimental visit. There was also no change from pre- to postintervention among individuals in Con who all still had a BP consistent with E/S1H classification at posttesting. Together, these serve to validate the E/S1H classification of the individuals included in this study. Finally, we did not conduct regular check-ins with Con throughout the intervention period, which would have assisted with ensuring protocol compliance. However, all participants in Con confirmed compliance at the end of the study, and this is supported by a lack of significant changes from pre- to postintervention in the Con group in this study.

The current study indicates that 9 wk of RET performed in accordance with the current exercise guidelines for individuals with E/S1H is effective for lowering BP to a degree consistent with the effects that may be expected by prescription of BP-lowering medications (41). These improvements were observed alongside a decrease in TPR and an increase in vascular endothelial function, as measured by the FMD technique. Moreover, we observed no effects of RET on central arterial stiffness or cardiovagal function. Therefore, our findings suggest that habitual RET lowers BP and improves vascular endothelial function among MA/O adults with E/S1H. Future studies should continue to investigate the acute and long-term effects of RET on BP and vascular function, as well as the influence of training status to better understand the potential discrepancies in the literature regarding RET and central arterial stiffening.

DATA AVAILABILITY

Data will be made available upon reasonable request.

GRANTS

Within the last 2 years, N.F.B. and E.M.R. have received graduate assistant stipend funding from Woodbolt, LLC. N.F.B. and N.D.M.J. have received grant funding from the National Strength and Conditioning Association Foundation. E.M.R. has received grant funding from the American College of Sports Medicine. N.D.M.J. has received grant funding from the American Heart Association, the Center for Integrative Research on Childhood Adversity Grant P20GM109097 [through the National Institute of General Medical Sciences (NIGMS)], Injury Prevention Research Center Grant R49 CE003095 [through the National Center for Injury Prevention and Control (NCIPC)/CDC], the National Institute on Aging through the Research Network on Animal Models to Understand Social Dimensions of Aging, Woodbolt Distribution, LLC, and Applied Food Sciences, Inc., and has been the recipient of a National Institutes of Health Clinical Research Loan Repayment Award.

DISCLAIMERS

The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

N.F.B., A.E.S., K.M.W., and N.D.M.J. conceived and designed research; N.F.B. and E.M.R. performed experiments; N.F.B., E.M.R., and N.D.M.J. analyzed data; N.F.B., E.M.R., and N.D.M.J. interpreted results of experiments; N.F.B. prepared figures; N.F.B. and N.D.M.J. drafted manuscript; N.F.B., E.M.R., A.E.S., K.M.W., and N.D.M.J. edited and revised manuscript; N.F.B., E.M.R., A.E.S., K.M.W., and N.D.M.J. approved final version of manuscript.