Journal Logo


Impact of High-Intensity Interval Training, Moderate-Intensity Continuous Training, and Resistance Training on Endothelial Function in Older Adults


Author Information
Medicine & Science in Sports & Exercise: May 2020 - Volume 52 - Issue 5 - p 1057-1067
doi: 10.1249/MSS.0000000000002226


Older adults are at an elevated risk of cardiovascular disease due to the development of atherosclerosis, which is initially characterized by an impaired endothelial-dependent dilation in response to physical or chemical stimuli (1). The noninvasive, flow-mediated dilation (FMD) assessment provides clinically relevant information regarding the ability of the endothelium to produce and release nitric oxide (NO) after a reactive hyperemia elicited by a prior period of distal ischemia. Most FMD studies have been performed in the brachial artery (BA), which provides an indication of coronary artery endothelial health (2) and is predictive of future cardiovascular events (3). However, lower limb arteries such as the popliteal artery (POP) are more susceptible to the development of atherosclerosis and peripheral vascular disease than upper-limb arteries (4). In addition to providing a clinically relevant measure of endothelial vasodilator function, the decline in conduit artery diameter observed during the distal ischemic period of the FMD test, termed low flow-mediated constriction (L-FMC), provides an index of endothelial-dependent vasoconstrictor function. Low flow-mediated constriction provides information regarding how reduced shear stress influences vascular tone (5) and is lower in older adults and patients with coronary artery disease (6). We have reported that the POP exhibits smaller L-FMC responses than the BA in older adults (7), which highlights the importance of determining whether exercise training impacts vascular endothelial function differently between upper- and lower-limb arteries.

It is well established that aerobic exercise training augments vasodilator function in older adults with chronic disease (e.g., hypertension, heart failure, type 2 diabetes), with high-intensity interval training (HIIT) eliciting superior improvements in BA-FMD than moderate-intensity continuous training (MICT) (8,9). The superior effects of HIIT have been attributed to larger and more sustained blood flow responses and/or greater reductions in oxidative stress, which together may enhance endothelial cell sensitivity to shear stress and increase NO bioavailability (as reviewed in Ramos et al. ([9])). Most previous reports that have compared the effects of HIIT and MICT on vasodilator function have been conducted solely in the BA of older adults with chronic disease and of a longer duration (i.e., >10 wk) (as reviewed in Ramos et al. ([9])). To date, the only study to directly evaluate the impact of aerobic exercise intensity in older adults free of chronic disease observed no improvement in BA-FMD after 2 wk of HIIT (10 × 1 min at 100% peak power output [PPO]) or MICT (40 min at 65% PPO) in postmenopausal women (10). As such, 2 wk (6 total exercise sessions) may not provide a sufficient stimulus to elicit favorable BA-FMD changes in this population. It is unclear if 6 wk of aerobic exercise improves BA-FMD in healthy older adults, improves endothelial function in arteries responsible for supplying blood flow to the active limbs during traditional modes of aerobic exercise (e.g., POP), or if HIIT produces greater improvements in FMD than MICT in these arteries. Cross-sectionally, we have demonstrated that cardiorespiratory fitness is moderately related to BA L-FMC (11) and strongly related to POP L-FMC (7) in older adults. Whether or not short-term aerobic exercise training interventions enhances L-FMC in either artery in older adults is unclear, and if HIIT augments the endothelial-dependent vasoconstrictor response more than MICT is also unknown in this population.

It is recommended that older adults engage in resistance training (RT) a minimum of 2 d·wk−1 (12). Resistance exercise elicits larger postcontraction blood flow responses than aerobic exercise, but this augmented blood flow response is transient and under high pressure in comparison to the sustained low-pressure flow observed during aerobic exercise (13–15). Despite these differences in exercise-induced shear stress profiles, RT appears to be an effective stimulus for increasing BA-FMD, albeit less than aerobic training in older adults with chronic disease (as reviewed in Ashor et al. ([8])). To date, the influence of short-term RT-induced adaptations on lower-limb endothelial-dependent vasodilatory or vasoconstrictor responses remains uncertain in healthy older adults.

We investigated the effects of short-term (6 wk) HIIT, MICT, and RT on upper- and lower-limb endothelial function in older adults. We hypothesized that HIIT would elicit superior improvements in BA-FMD and POP-FMD than MICT, and that these aerobic protocols will elicit greater endothelial-dependent vasodilator responses than RT. Based on our previous observations that greater aerobic fitness was associated with larger L-FMC responses in both the BA and POP (7,11), we anticipated that 6 wk of HIIT or MICT, but not RT, will augment endothelial-dependent vasoconstrictor function in this population.



Thirty-eight older adults (23 females; age, 56–83 yr) were recruited from the Active Aging program at Acadia University (Table 1). Participants had no physical limitations to exercise and a resting blood pressure <140/90 mm Hg. Three participants were on Synthroid for hypothyroidism (n = 1 in each group). Four participants were prescribed medications to treat high blood pressure. Specifically, participants were taking Teveten® (angiotensin-receptor blocker; n = 1; HIIT), Adalat® (calcium channel blocker; n = 1; RT), Diuril® (diuretic; n = 1; HIIT), and Coversyl Plus® (angiotensin converting enzyme inhibitor + diuretic; n = 1; RT). One person was asthmatic (RT). During the study, participants were requested to continue taking all prescribed medications. Participants were informed of the methods and study design verbally and in writing before providing written informed consent. Participants were randomized to HIIT (n = 12), MICT (n = 12), or RT (n = 14) after the pretraining determination of aerobic fitness and initial vascular testing day. The pretraining vascular endothelial data have previously been presented in a cross-sectional study investigating L-FMC responses between the BA and POP (7). All protocols and procedures conformed to the Declaration of Helsinki and were approved by the Dalhousie University Health Sciences and Acadia University Research Ethics Boards.

Participant descriptive characteristics, resting hemodynamics and aerobic fitness.

Experimental design

Participants underwent four separate laboratory visits in total. Days 1 and 2 were conducted pretraining, whereas days 3 and 4 were completed after training. Days 1 and 3 involved measurements of height and body mass, which were followed by a graded, maximal cycling exercise test to determine aerobic fitness (V˙O2peak). Peak power output was also recorded and used to establish exercise intensities for the two aerobic training protocols. Days 2 and 4 were dedicated to the assessments of BA and POP vascular function and conducted either before, or a minimum of 48 h (or maximum 1 wk) after the graded exercise tests (see below for details). To minimize known confounding influences on endothelial-dependent dilation, vascular assessments were performed 6 h postprandial, whereas participants avoided strenuous physical activity, as well as the consumption of products known to acutely influence endothelial responses (e.g., caffeine, chocolate, kiwi, saturated fats, folic acid supplements, antioxidant and multivitamin supplements) for 24 h, consistent with FMD guidelines (16). Participants replicated their diet before both the pretraining and posttraining vascular measurements. Upon arriving to the laboratory, participants rested in the supine position for a minimum of 20 min before the vascular assessments. Vascular function was assessed first in the BA, followed by at least 10 min of rest before the POP measures. All study visits were performed in a thermoneutral environment (21°C). Days 2 and 4 were performed at the same time of day within each participant to control for diurnal variations in blood pressure and vascular function.

Anthropometrics and peak aerobic fitness

Height and weight were measured using a calibrated stadiometer (Health-O-Meter, McCook IL) to the nearest 0.5 cm and 0.1 kg, respectively. An incremental and maximal exercise test on a cycle ergometer (Lode Excalibur Sport, Groningen, The Netherlands) was administered to determine V˙O2peak via a mixing chamber-based commercial metabolic system (TrueOne 2400®; Parvomedics Inc., Sandy, UT). After a 5-min warm-up period of light-intensity cycling (30–50 W), the workload was set at 1 W·kg−1 body mass and gradually increased by 15 W·min−1 until voluntary exhaustion. Strong verbal encouragement was provided throughout the test. Upon completion of the test, the workload was immediately reduced to the warm-up level for a 5-min cool-down period. The workload of the last completed stage was considered as the PPO, measured in watts.

Training protocols

For the three training protocols, all sessions were supervised and conducted 3 d·wk−1 (Mondays, Wednesdays, and Fridays) for 6 wk. As outlined in Figure 1, all participants completed 18 supervised sessions with no participants dropping out of the study. Warm-up and cool-down periods consisted of 5 min at 25% PPO for both the HIIT and MICT protocols. The HIIT protocol was based on a previous study that compared the time to exhaustion, safety, participant preference, and time spent near V˙O2max in older adults (17). Specifically, the protocol implemented in the present study resulted in the longest time to exhaustion and a similar amount of time spent above 80% V˙O2max and 90% V˙O2max versus 3 other HIIT protocols varying in work/rest ratios and recovery intensity (active vs passive) (17). Furthermore, previous research has demonstrated that this HIIT protocol elicits greater mean V˙O2 responses than an isocaloric bout of moderate-intensity continuous exercise performed at 70% PPO (18). For the first 2 wk, the HIIT protocol consisted of forty 15-s intervals at 100% PPO interspersed with 15 s of passive recovery. After a 5-min passive recovery period, a second set of 40 intervals was completed (i.e., 40 min total time). To adjust for anticipated improvements in aerobic fitness and exercise tolerance, the duration of the HIIT protocol was increased to 45 min (2 × 22.5 min; 2 sets of 45 intervals) for the remaining 4 wk and workload increased by 15 W for the final 2 wk. Training volume was decreased by 10 min (from two sets of 45 intervals to two sets of 35 intervals) for the final two training sessions to ensure appropriate recovery before the posttraining assessment of aerobic fitness.

Schematic of the experimental design and each of the HIIT, MICT, and whole-body RT protocols. PPO (measured in watts). *The RT protocol consisted of eight exercises targeting all major muscle groups for two sets of 10 repetitions.

The MICT protocol was based on the American College of Sports Medicine physical activity guidelines that recommends at least 30 min of daily moderate aerobic physical activity (19). Continuous cycling at 60% PPO for approximately 34 min was initially prescribed. This duration was adjusted to ensure that the MICT protocol was isoenergetic to the HIIT protocol based on the assumption that mechanical efficiency, aerobic fitness, and PPO were similar between groups; in that 20 min at 100% PPO expends the same energy as 34 min at 60% PPO. To support this assumption, the HIIT and MICT groups demonstrated similar pretraining V˙O2peak and PPO (both, P = 1.00, Table 1). To accommodate the matched increase in energy expenditure, total exercise time was prolonged to 39 min for the remaining 4 wk and power output increased by 15 W for the final 2 wk. For the final two MICT sessions, participants decreased their cycling time from 39 to 30 min in preparation for the posttraining determination of maximal aerobic fitness.

Each RT session began and ended with 3 min of light cycling at 25% PPO. Thereafter, participants completed a total of eight strength exercises, alternating between muscle groups. The exercises were primarily isokinetic machine-based and included leg press, bench press, hamstrings curl, shoulder press, and leg extensions. Cable exercises included seated row and latissimus pull-down, and bird-dogs (i.e., a core exercise that involves kneeling on the floor and simultaneously extending the hip while flexing the contralateral shoulder). Each participant performed 2 × 10 repetitions at 70% of perceived one-repetition maximum (1RM) for the first 2 wk. Both sets of each exercise were performed before starting the next exercise with 30 to 60 s of rest between sets. Upper- and lower-body exercises were alternated. Participants were instructed to increase the number of repetitions to 12 ad libitum. Once participants were able to perform 2 × 12 repetitions, the resistance was proportionally increased to a weight that equated to 10 repetitions of their new estimated 70% of 1RM with the assistance of the supervising Canadian Society for Exercise Physiology Certified Exercise Physiologist.


HR was determined via cardiac intervals obtained from lead II of a standard bipolar limb lead ECG. Beat-by-beat systolic (SBP) and diastolic (DBP) blood pressures were measured using finger photoplethysmography (Portapres®; Finapres Medical Systems, Amsterdam, Netherlands). The finger used for recording blood pressure was maintained at heart level throughout the protocol and minor deviations in height between the heart and finger were corrected using the Portapres® height correction unit. Left BA measurements of SBP and DBP were also recorded by an automated patient vital signs monitor (Carescape v100®; General Electric Healthcare) and used to perform a “physiological calibration” of the Portapres® waveform. All data were sampled continuously at 200 Hz using a PowerLab (PL3508 PowerLab 8/53; ADInstruments, Sydney, Australia) data acquisition system with the exception of the ECG waveform, which was sampled at 1000 Hz. Recordings were displayed in real time and analyzed offline using LabChart software (ADInstruments, Sydney, Australia).

Vascular measures

As described in O’Brien et al. (20), the right BA and left POP were imaged with the participants in the supine and prone positions, respectively. The BA was imaged 3 to 5 cm proximal to the antecubital fossa and the POP was imaged proximal to the bifurcation at or slightly above the popliteal fossa. A pressure cuff attached to a rapid inflation system (E20 and AG101; Hokanson®, Bellevue, WA) was positioned around the largest circumference of the forearm (BA; ~3 cm distal to the antecubital fossa) or lower leg (POP; ~10 cm distal to the popliteal fossa). All images were obtained using a 12-MHz multifrequency linear array probe attached to a high-resolution ultrasound system (Vivid i, General Electric Healthcare). Blood velocity signals were recorded in duplex mode at a pulsed frequency of 5-MHz and corrected with an insonation angle of 60° that remained constant throughout the study. The sample volume was adjusted for each participant such that the anterior to the posterior intima were included, as recommended in published guidelines (16). Artery lumen diameter and blood velocity were measured for a minimum of 2 min before inflation of the pneumatic cuff. The pressure cuff was then rapidly inflated to 250 mm Hg for 5 min. Continuous arterial lumen diameter and blood velocity recordings were collected throughout the cuff inflation period. Upon release of cuff pressure, lumen diameter and blood velocity recordings continued for an additional 5 min.

After a minimum 10 min of rest after the FMD test, the POP was imaged for 1 min before and 10 min after a sublingual administration of nitroglycerin spray (0.4 mg). This nitroglycerin test provides a measure of endothelial-independent vasodilation (21).

Data analysis

Relative V˙O2 data were averaged over 15-s intervals for the duration of the graded exercise protocol. Maximum or peak V˙O2 were considered as the greatest 30 s averaged V˙O2. The SBP and DBP were determined from the Portapres® waveform as the maximum and minimum waveform values, respectively. These pressures were then used to calculate mean arterial pressure (MAP) using the equation ⅓ SBP + ⅔ DBP. In addition, stroke volume (SV) was derived from the raw finger blood pressure waveforms using the ModelFlow® method incorporated into the noninvasive cardiac output add-on for LabChart® (ADInstruments, Sydney, Australia). Cardiac output (Q) was calculated as the product of HR and SV and total vascular conductance (TVC) as Q/MAP. Portapres® data (SV, Q, and TVC) were not recorded from two participants in the MICT group (1♂, 1♀) due to equipment malfunctions. Hemodynamic data were averaged over at least 5 min of beat-by-beat data recoded immediately before the POP-FMD protocol.

Video signals from the ultrasound were exported to a laptop via a video graphics array converter (Epiphan Systems Inc., VGA 2 USB, Ottawa) for offline analysis. Analysis of artery diameter, blood velocity, and shear rate (i.e., frictional force of blood flow on the endothelium) were performed using automated commercial edge-detection and wall-tracking software combined with simultaneous Doppler waveform envelope analysis (FMD Studio, Cardiovascular Suite, Quipu, Pisa, Italy). This software was used to measure baseline diameter, nadir diameter (i.e., L-FMC) and peak diameter (i.e., FMD). All vascular measurements were blindly analyzed by a single investigator from the laboratory, who has demonstrated coefficients of variation of 2.2%, 2.7%, 3.8%, and 4.2% for baseline diameter, nadir diameter, L-FMC% and FMD%, respectively (7,11,20,22).

Absolute FMD was calculated as the difference (in mm) between the postcuff deflation peak and baseline diameters. Relative FMD was calculated using the equation: FMD (%) = [(postcuff deflation peak diameter − baseline diameter)/ baseline diameter × 100%]. Blood flow (mL·min−1) was calculated as mean blood velocity (cm·s−1) × π × lumen radius (cm)2 × 60 (s·min−1). Popliteal and brachial vascular conductances were calculated by dividing POP and BA blood flow by MAP, respectively. All blood flow and MAP data from the corresponding FMD-baseline periods were used for these calculations. Shear rate (SR, s−1) was defined as [4 × mean blood velocity (cm·s−1)]/diameter (cm). Subsequently, the SR area under the curve (SRAUC) was calculated between the start of cuff deflation to the time that peak dilation occurred. Whether or not, and how, FMD responses should be normalized to SRAUC is debated in the literature (16,23). Our data did not meet the statistical assumptions required to normalize the FMD responses to SRAUC (23,24) in either the BA or POP [see Figures, Supplemental Digital Content 1 and 2, which contains scatterplots demonstrating the relationship between brachial ( and popliteal ( shear rate area under the curve vs FMD]). However, SRAUC responses and the relationships between FMD and SRAUC are presented for both arteries, as per recommendations (16,23).

Low flow-mediated constriction was represented in absolute (mm) and relative (%) terms using the nadir diameter obtained during the final 30 s of the 5-min distal cuff occlusion period. Relative L-FMC was calculated using the equation: L-FMC (%) = ([baseline diameter − nadir diameter]/baseline diameter × 100%). The vasoactive range (VAR) was calculated as ([postcuff deflation peak diameter − nadir diameter]/baseline diameter × 100%) and expressed in both absolute (mm) and relative (%) terms.

The POP nitroglycerin-mediated vasodilation was calculated as a percentage change from baseline to the peak lumen diameter obtained during the 10-min period after sublingual administration of nitroglycerin. The POP nitroglycerin-mediated vasodilation was not conducted in two participants (total participants: HIIT, n = 11 and MICT, n = 11).

As described by Atkinson and Batterham (25), allometric scaling of FMD has been recommended to account for differences in baseline arterial diameter. Although based on FMD analysis, we previously demonstrated that baseline diameter is negatively correlated to the L-FMC responses in older adults (7). Allometric scaling is recommended if the relationship between the natural log of peak FMD diameter (or nadir diameter for L-FMC) and resting diameter yield an unstandardized β-coefficient that deviates from 1 and/or have an upper 95% confidence interval <1. These assumptions were met for L-FMC measurements in the BA and POP with β ± standard error (95% confidence intervals) of 0.979 ± 0.01 (0.97–0.99) and 0.989 ± 0.01 (0.98–1.00), respectively. However, the allometric assumptions were not met for BA-FMD (1.036 ± 0.01 [1.01–1.06]) and POP-FMD (1.008 ± 0.01 [0.99–1.03]), suggesting allometric scaling of FMD to be unnecessary in this study. These assumptions remained when checked for each group individually. Allometrically scaled L-FMC were examined using an ANCOVA model with the natural log of the difference (ln[nadir diameter] − ln[baseline diameter]) as the dependent variable and ln(baseline diameter) as the covariate. Group (HIIT, MICT, RT) and time (pretraining, posttraining) were used as a fixed-factors. Statistically significant ANCOVA were followed up with pairwise comparisons using Fisher least squares difference post hoc testing (25). For each instance, overall group (not individual) allometrically scaled arterial diameter changes were back transformed and presented as a percent change from the baseline diameter, as described in more detail previously (25). Separate ANCOVA models were conducted for each artery.

Statistical analysis

Participant descriptive characteristics were compared using one-way ANOVA. The effects of exercise training on resting hemodynamic and vascular measurements were compared using a between-subjects (group × time) repeated-measures ANOVA. The variance of differences was assessed using Mauchly’s test of sphericity and when violated, the Greenhouse–Geisser correction to the degrees of freedom was applied. Bonferroni post hoc testing was conducted on statistically significant ANOVA. All data were assessed for normality using a Shapiro–Wilk test, and nonnormalized data were appropriately transformed (e.g., log transformation) before statistical analysis. Based on the previously observed relationship between aerobic fitness and L-FMC (7,11), correlational analysis were conducted to determine if the change in aerobic fitness (ΔV˙O2peak) was related to the change (ΔL-FMC) in either artery. Although not the purpose of the present study, correlational analysis comparing the exercise training induced changes in L-FMC (ΔL-FMC) and FMD (ΔFMD) between the BA and POP are presented in Supplemental Digital Content 3 [see Figure, Supplemental Digital Content 3, Scatterplot showing the relationship between the change in relative flow-mediated dilation (FMD; A) responses and changes in relative L-FMC (B) between the brachial and popliteal arteries,]. Some participants were taking antihypertensive medication (n = 4), which may have influenced their vascular function. Removing these individuals from analyses did not alter the significant findings determined from the full sample (see below). All statistics were completed in SPSS Version 25.0 (IBM, NY). Statistical significance was accepted as P < 0.05. All data are presented as means ± SD.


Participant characteristics, systemic hemodynamics and aerobic fitness were similar between the groups at baseline (all, P > 0.29) and are summarized in Table 1. No pre–post differences were observed for body mass index (RT: P = 0.09), SV (HIIT: P = 0.08), CO, TVC (HIIT: P = 0.06), peak respiratory exchange ratio (HIIT: P = 0.08), or peak HR (all, P > 0.11 unless specified) in any of the training groups. Resting HR decreased (P = 0.04) in the MICT group only. The HIIT group had lower posttraining resting SBP, DBP, and MAP (all, P < 0.03), but MICT only decreased DBP (P = 0.02). RT did not change any systemic resting hemodynamic measurements (MAP: P = 0.06; rest of variables, P > 0.11). Compared with pretraining, all groups had a greater V˙O2peak and PPO at posttraining (both, P < 0.04).

BA hemodynamics

The BA lumen diameter was larger after RT (P = 0.02), but not HIIT or MICT (both, P > 0.79; see Table 2). The increase in resting BA diameter lead to a lower resting BA-SR in the RT group (P = 0.048). Resting BA blood flow velocity and blood flow were increased after HIIT only (both, P < 0.003), which resulted in a higher BA-SR (P = 0.01). The increases in BA blood flow and corresponding decreases in MAP after HIIT resulted in a larger resting BA vascular conductance in this group (P < 0.001). The MICT did not change any resting BA variables (all, P > 0.38).

Comparison of BA parameters across the HIIT, MICT and RT protocols.

BA vasoreactivity

As shown in Table 2, BA peak diameter was greater after training in all three groups (all, P < 0.03). BA-SRAUC was larger after HIIT and RT (both, P < 0.01), but not MICT (P = 0.12). Training did not change the time-to-peak diameter in any group (all, P > 0.37). Relative BA-FMD was enhanced after HIIT (4.8% ± 1.8% to 6.7% ± 1.3%; P < 0.001) and MICT (4.7% ± 1.9% to 6.8% ± 1.7%; P < 0.001), but not RT (4.7% ± 1.4% to 5.0% ± 1.4%; P = 0.52; Fig. 2A).

Comparison of BA relative FMD (%) (A), relative L-FMC (%) (B), and allometrically scaled L-FMC (%; C) from pretraining to posttraining after HIIT, MICT, and whole-body RT. Data are presented as means ± SD. Individual data are presented in panels A and B. *P < 0.05 vs pretraining within same training group. Data were analyzed via a between-subjects repeated-measures (training group × time) ANOVA with Bonferroni post hoc testing. HIIT, n = 12; MICT, n = 12; RT, n = 14.

BA L-FMC was greater after HIIT (−0.9% ± 1.7% to −1.7% ± 1.2%; P < 0.001), but not MICT (−0.9% ± 1.0% to −1.2% ± 0.9%; P = 0.14) or RT (−0.8% ± 0.6% to −0.9% ± 0.7%; P = 0.68; Fig. 2B). This observation was unchanged after allometrically scaling values for baseline diameter (HIIT, P = 0.03; MICT and RT, P > 0.42; Fig. 2C). The change in BA L-FMC was moderately correlated with the change in aerobic fitness (r = −0.51, P = 0.001; ΔV˙O2peak vs ΔBA L-FMC; see Supplemental Digital Content 4, The greater FMD (HIIT and MICT) and L-FMC (HIIT only) resulted in a greater VAR after both HIIT and MICT (both, P < 0.001).

Popliteal artery hemodynamics

As shown in Table 3, POP lumen diameter was larger after RT (P = 0.02) but not HIIT or MICT (both, P > 0.62). Mean POP blood flow velocity (MICT: P = 0.08), blood flow, resting SR (MICT: P = 0.08), and POP vascular conductance were unchanged after any of the three-training protocol (all, P > 0.10 unless specified).

Comparison of POP parameters across the HIIT, MICT, and RT protocols.

Popliteal artery vasoreactivity

A greater POP peak diameter in response to reactive hyperemia was observed after MICT and RT (both, P < 0.02), but not HIIT (P = 0.052) (Table 3). All groups had a posttraining increase in the SRAUC (all, P < 0.001) with no changes in the time-to-peak diameter (P > 0.52). Popliteal artery FMD was enhanced after HIIT (3.6% ± 1.9% to 4.9% ± 1.5%; P < 0.001) and MICT (2.6% ± 1.7% to 4.0% ± 1.9%; P < 0.001), but not RT (3.1% ± 1.9% to 3.4% ± 1.8%; P = 0.20; see Fig. 3A).

Comparison of POP relative FMD (%); (A), relative L-FMC (%) (B), and allometrically scaled L-FMC (%; C) from pretraining to posttraining after HIIT, MICT, and whole-body RT. Data are presented as means ± SD. Individual data are presented in panels A and B. *, P < 0.05 to pretraining within same training group. Data were analyzed via a between-subjects repeated-measures (training group × time) ANOVA with Bonferroni post hoc testing. HIIT, n = 12; MICT, n = 12; RT, n = 14.

Similarly, POP L-FMC was increased after HIIT (−0.9% ± 1.1% to −1.8% ± 0.9%; P = 0.01) and MICT (−0.8% ± 1.6% to −1.5% ± 0.7%; P = 0.045), but was unchanged after RT (−1.0% ± 1.7% to −1.0% ± 1.1%; P = 1.00; Fig. 2B). However, allometrically scaled POP L-FMC was not statistically greater in the MICT group (−0.4% ± 0.5% to −0.7% ± 0.6%; P = 0.16). In contrast, the greater L-FMC response after HIIT remained significant after scaling for baseline diameter (−0.4% ± 0.6% to −0.8% ± 0.6%; P = 0.047; Fig. 3C). The change in POP L-FMC was not correlated to the change in aerobic fitness (r = −0.27, P = 0.10; ΔV˙O2peak vs ΔPOP L-FMC; see Supplemental Digital Content 4,

The larger POP-FMD and POP L-FMC responses after HIIT and MICT resulted in a greater VAR in these groups (both, P < 0.001), that was not observed after RT (P = 0.44). Popliteal artery nitroglycerin-mediated vasodilation was unchanged after any of the training protocols (all, P > 0.33).


The purpose of this study was to compare endothelial-dependent vasodilator and vasoconstrictor responses in the brachial and popliteal arteries after short-term HIIT, MICT, and RT in healthy older adults. Contrasting with our hypothesis, 6 wk of HIIT and MICT similarly improved FMD responses in the brachial and popliteal arteries. Adding to the current literature, HIIT-enhanced endothelial vasoconstrictor responses in both arteries, but MICT only increased L-FMC in the POP (i.e., to active limbs). Whole-body RT did not improve either measure of vascular endothelial function in the BA nor POP. Favorable increases in resting blood flow and vascular conductance were observed in the BA after HIIT, but not MICT or RT. Similar responses were not identified in the POP. This is the first study to compare the effects of different exercise modalities on both endothelial vasodilator and vasoconstrictor function in healthy older adults. Furthermore, we report the first data to document vascular endothelial responses to training in the POP of older persons, which experiences different shear stress profiles during cycling exercise training than the BA.

Our results support the notion that the vasculature may be favorably modified by short-term aerobic exercise training in older adults known to experience the deteriorating effects of advanced aging (26), such as chronic low-grade inflammation and increased oxidative stress (27). Adding to the current body of literature, our short-term HIIT and MICT models (but not RT) elicited a sufficient stimulus for modifying both BA-FMD (ΔHIIT, 1.9%; ΔMICT, 2.1%; Fig. 2) and POP-FMD (ΔHIIT, 1.3%; ΔMICT, 1.4%; Fig. 3). Similar to observations in younger adults (28), short-term (i.e., 6–8 wk) aerobic training augmented vasodilator function in two arteries that experience very different shear rate stimuli during cycling exercise in our population of older adults. Future research should aim to uncover the time course of these changes and if longer training periods continue to increase vasodilator function. Of particular importance, a 1% increase in relative BA-FMD equates to a 10% reduction in the relative risk of a future cardiovascular event and all-cause mortality (3). Furthermore, BA-FMD has been shown to be predictive of coronary artery function; dysfunction of the coronary arteries precedes or is responsible for most cardiovascular disease (2). In addition, endothelial dysfunction is an early marker of atherosclerosis (1), and the POP is a common site for plaque development (4). Herein, both HIIT and MICT resulted in a clinically significant reduction in cardiovascular disease risk in a relatively short period of exercise training among healthy, older adults.

Of relevance, we did not observe any changes in POP nitroglycerin-mediated vasodilation after any training regimen, suggesting that the improved POP-FMD responses in the HIIT and MICT groups are not attributed to enhanced vascular smooth muscle sensitivity. This finding is consistent with previous literature conducted in the BA, in that a meta-analysis of 26 exercise training studies concluded that BA nitroglycerin-mediated vasodilation was only marginally greater or was unchanged after exercise training (8). Perhaps longer training periods may be needed to combat the diminished POP nitroglycerin-mediated vasodilation associated with advancing age (29). Certainly, the present study was not designed to test the mechanisms responsible for our observed findings, but possible mechanisms behind the improved relative FMD responses is a reduction in oxidative stress (i.e., reactive oxygen species) or resting levels of the potent vasoconstrictor endothelin-1. Aging is associated with exacerbated oxidative stress that promotes the uncoupling of endothelial NO synthase (30) and endothelin-1 expression (31), which has been shown to be favorably modulated via exercise training (32,33). Future studies are needed to confirm or refute this hypothesis.

We have previously demonstrated that higher cardiorespiratory fitness is associated with a greater L-FMC response in healthy older adults (7,11). The present study is the first to determine if an aerobic or resistance exercise training intervention can augment L-FMC in this population. Interestingly, HIIT augmented endothelial-dependent vasoconstrictor function in the BA and POP, but MICT only increased L-FMC in the POP (Fig. 2 and Fig. 3). Considering that POP L-FMC similarly increased after HIIT and MICT, but not after RT, the increased endothelial sensitivity to low-flow in the popliteal artery may be attributed to the increased shear stress in the active limb arteries during aerobic exercise, which translates to more effective signaling of vasodilatory/vasoconstrictor pathways. With that, our results suggest that the higher shear stress during HIIT versus MICT may be needed to elicit favorable adaptations in the vasoconstrictor side of endothelial function in the BA. L-FMC has been shown to be mediated through the endothelium (34), via the inhibition of vasodilatory signaling via endothelial-derived hyperpolarizing factor and prostaglandins (5), as well as enhanced vasoconstrictor signaling via endothelin-1 (35). Our results suggest that HIIT may be more effective at promoting the inhibition of these vasodilatory pathways and/or stimulation of the endothelin-1 pathway in response to reductions in shear stress to a greater extent than MICT. The importance of enhancing these pathways and the L-FMC response to overall vascular health and risk of adverse cardiovascular events is currently unknown. Contrary to our previous cross-sectional observations equating cardiorespiratory fitness and vasoconstrictor function (7,11), aerobic fitness but not L-FMC was improved in all three groups. It appears that other factors independent of increased aerobic fitness are responsible for subsequent enhancements to endothelial-dependent vasoconstrictor function. Of importance, these observations differ between arteries, with the relationship between the change in aerobic fitness and change in L-FMC being moderate-strength in the BA but nonexistent in the POP [see Figure, Supplemental Digital Content 4, Scatterplot showing the relationship between the change in aerobic fitness (peak oxygen consumption) with the change in BA relative L-FMC (A) and the change in popliteal artery L-FMC (B) in the pooled sample,]. Furthermore, no relationship was observed for the training-induced changes in L-FMC between the brachial and popliteal arteries [see Figure, Supplemental Digital Content 3, Scatterplot showing the relationship between the change in relative FMD (A) responses and changes in relative L-FMC (B) between the brachial and popliteal arteries,]. As such, it is plausible that different mechanisms may be responsible for the improvements in L-FMC between the BA and POP.

With advancing age, there is a reduction in BA blood flow (36). Interestingly, HIIT, but not MICT or RT, resulted in a greater resting BA blood flow and shear stress through an increase in blood flow velocity with no corresponding change in resting BA diameter. Higher anterograde and lower retrograde BA blood flow velocity is associated with greater physical function in older persons (36) and a higher resting shear stress may promote antiatherogenic gene expression (37). This suggests our HIIT protocol may be a more effective training program than MICT or RT alone at combatting the decline in physical function and progression of atherosclerosis in an aged population.

Few studies have investigated if RT influences vascular function in older adults (8). Our results show that short-term RT did not alter endothelial-dependent vasodilator or vasoconstrictor function in either artery in this population. It may be possible that a longer duration of RT, or more intense RT, may be needed to elicit these functional adaptations. Interestingly, the whole-body RT did increase resting lumen diameter. Although the mechanisms behind this are unclear, our observations are consistent in both the BA and POP. This warrants future investigations regarding the influence of transient, high-pressure shear stress patterns on local resting diameter in older adults, who are often encouraged to engage in regular RT (12). Of relevance, resting diameter is influenced by numerous competitive vasodilator and constrictor influences, limiting its use as an index of vascular structure (38). With that, the maximal dilation in response to nitroglycerin, which represents near maximal diameter, has been used as a preferred index of arterial structure (38) and was not influenced by RT in the POP. This suggests that RT likely did not elicit structural changes in this short-term intervention.

Although improvements in BA-FMD were observed in the HIIT and MICT groups, endothelial-independent dilation was not assessed in the BA. However, it is unlikely that 6 wk is long enough to enhance vascular smooth muscle function in the BA (8). We appreciate that our study is unable to provide information regarding the mechanisms of our observed findings. Future research incorporating blockades of endothelial-dependent vasodilatory and vasoconstrictor pathways pretraining and posttraining are warranted. With that, we did not observe training induced changes in POP nitroglycerin-mediated dilation (i.e., endothelial-independent dilation), excluding alterations in vascular smooth cell sensitivity to endothelial-derived substances as a potential mechanism for the enhanced FMD and L-FMC responses. Sex differences in the vascular responses to exercise training may exist (39). However, the number of men and women in the exercise groups were similar. Future studies should examine the role that sex has on exercise mode and aerobic exercise intensity-induced vascular responses in older populations. We acknowledge our study did not include a control group, but our repeated-measures design allowed participants to serve as their own controls. As well, randomized controlled studies investigating exercise modalities and endothelial function consistently demonstrate no improvement in the control group, as reviewed in (8), which in conjunction with the high reproducibility of our ultrasound measures suggests the lack of a control group to be a minor limitation that does not take away from our primary findings.

The findings of this study demonstrate that 6 wk of HIIT and MICT are superior to RT at eliciting improvements in endothelial-dependent dilation in the brachial and popliteal arteries of older adults. Although HIIT and MICT enhanced endothelial-dependent vasoconstrictor function in the POP, only the HIIT group exhibited a greater endothelial sensitivity to low flow and resting blood flow in the BA. High-intensity interval training produces clinically meaningful increases in vascular endothelial function in the short-term, which translates to a reduced risk of cardiovascular disease of a greater magnitude than MICT or RT in older adults. Therefore, exercise training programs aimed at improving peripheral vascular function in healthy older adults should consider implementing HIIT.

Support for the work was provided by Canadian Foundation for Innovation: Leader’s Opportunity Fund (to D. S. Kimmerly), Faculty of Health Professions Research Development (to D. S. Kimmerly), Nova Scotia Health Research Foundation (NSHRF) Development/Innovation (to D. S. Kimmerly and S. Mekary) grants, as well as the Acadia University McCain Foundation Fund (to S. Mekary). M. W. O’Brien was supported by a Heart & Stroke Foundation of Canada BrightRed Scholarship, a Nova Scotia Graduate Scholarship, and a NSHRF Scotia Scholars Award.

The authors report no conflicts of interest. The results of the present study do not constitute endorsement by ACSM. The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.


1. Davignon J, Ganz P. Role of endothelial dysfunction in artherosclerosis. Circulation. 2004;109(23):27–32.
2. Broxterman RM, Witman MA, Trinity JD, et al. Strong relationship between vascular function in the coronary and brachial arteries. Hypertension. 2019;74(1):208–15.
3. Xu Y, Arora RC, Hiebert BM, et al. Non-invasive endothelial function testing and the risk of adverse outcomes: a systematic review and meta-analysis. Eur Heart J Cardiovasc Imaging. 2014;15(7):736–46.
4. Debasso R, Åstrand H, Bjarnegård N, Rydén Ahlgren Å, Sandgren T, Länne T. The popliteal artery, an unusual muscular artery with wall properties similar to the aorta: implications for susceptibility to aneurysm formation? J Vasc Surg. 2004;39(4):836–42.
5. Gori T, Dragoni S, Lisi M, et al. Conduit artery constriction mediated by low flow a novel noninvasive method for the assessment of vascular function. J Am Coll Cardiol. 2008;51(20):1953–8.
6. Gori T, Muxel S, Damaske A, et al. Endothelial function assessment: flow-mediated dilation and constriction provide different and complementary information on the presence of coronary artery disease. Eur Heart J. 2012;33(3):363–71.
7. O’Brien MW, Johns JA, Robinson SA, Mekary S, Kimmerly DS. Relationship between brachial and popliteal artery low-flow-mediated constriction in older adults: impact of aerobic fitness on vascular endothelial function. J Appl Physiol. 2019;127(1):134–42.
8. Ashor AW, Lara J, Siervo M, et al. Exercise modalities and endothelial function: a systematic review and dose-response meta-analysis of randomized controlled trials. Sports Med. 2015;45(2):279–96.
9. Ramos JS, Dalleck LC, Tjonna AE, Beetham KS, Coombes JS. The impact of high-intensity interval training versus moderate-intensity continuous training on vascular function: a systematic review and meta-analysis. Sports Med. 2015;45(5):679–92.
10. Klonizakis M, Moss J, Gilbert S, Broom D, Foster J, Tew GA. Low-volume high-intensity interval training rapidly improves cardiopulmonary function in postmenopausal women. Menopause. 2014;21(10):1099–105.
11. O’Brien MW, Mekary S, Robinson SA, Johns JA, Kimmerly DS. The relationship between aerobic fitness and low-flow-mediated constriction in older adults. Eur J Appl Physiol. 2019;119(2):351–9.
12. Chodzko-Zajko WJ, Proctor DN, Fiatarone Singh MA, et al. Exercise and physical activity for older adults. Med Sci Sports Exerc. 2009;41(7):1510–30.
13. Donato AJ, Uberoi A, Wray DW, Nishiyama S, Lawrenson L, Richardson RS. Differential effects of aging on limb blood flow in humans. Am J Physiol Heart Circ Physiol. 2006;290(1):H272–8.
14. Thijssen DH, Dawson EA, Black MA, Hopman MT, Cable NT, Green DJ. Brachial artery blood flow responses to different modalities of lower limb exercise. Med Sci Sports Exerc. 2009;41(5):1072–9.
15. Gaenzer H, Neumayr G, Marschang P, Sturm W, Kirchmair R, Patsch JR. Flow-mediated vasodilation of the femoral and brachial artery induced by exercise in healthy nonsmoking and smoking men. J Am Coll Cardiol. 2001;38(5):1313–9.
16. Thijssen DHJ, Bruno RM, van Mil ACCM, et al. Expert consensus and evidence-based recommendations for the assessment of flow-mediated dilation in humans. Eur Heart J. 2019;40(30):2534–47.
17. Guiraud T, Juneau M, Nigam A, et al. Optimization of high intensity interval exercise in coronary heart disease. Eur J Appl Physiol. 2010;108:733–40.
18. Guiraud T, Nigam A, Juneau M, Meyer P, Gayda M, Bosquet L. Acute responses to high-intensity intermittent exercise in CHD patients. Med Sci Sports Exerc. 2011;43(2):211–7.
19. Garber CE, Blissmer B, Deschenes MR, et al. American College of Sports Medicine position stand. Quantity and quality of exercise for developing and maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness in apparently healthy adults: guidance for prescribing exercise. Med Sci Sports Exerc. 2011;43(7):1334–59.
20. O’Brien MW, Robinson SA, Frayne RJ, Mekary S, Fowles JR, Kimmerly DS. Achieving Canadian physical activity guidelines is associated with better vascular function independent of aerobic fitness and sedentary time in older adults. Appl Physiol Nutr Metab. 2018;43(10):1003–9.
21. Chen Z, Zhang J, Stamler JS. Identification of the enzymatic mechanism of nitroglycerin bioactivation. Proc Natl Acad Sci. 2002;99(12):8306–11.
22. O’Brien MW, Johns JA, Williams TD, Kimmerly DS. Sex does not influence impairments in popliteal endothelial-dependent vasodilator or vasoconstrictor responses following prolonged sitting. J Appl Physiol. 2019;127(3):679–87.
23. Atkinson G, Batterham AM, Black MA, et al. Is the ratio of flow-mediated dilation and shear rate a statistically sound approach to normalization in cross-sectional studies on endothelial function? J Appl Physiol (1985). 2009;107(6):1893–9.
24. Tschakovsky ME, Pyke KE, Fergus S. Reply to Drs. Harris and Padilla. J Appl Physiol. 2007;103(3):1109–9.
25. Atkinson G, Batterham AM. Allometric scaling of diameter change in the original flow-mediated dilation protocol. Atherosclerosis. 2013;226(2):425–7.
26. Early KS, Stewart A, Johannsen N, Lavie CJ, Thomas JR, Welsch M. The effects of exercise training on brachial artery flow-mediated dilation: a meta-analysis. J Cardiopulm Rehabil Prev. 2017;37(2):77–89.
27. Ungvari Z, Kaley G, de Cabo R, Sonntag WE, Csiszar A. Mechanisms of vascular aging: new perspectives. J Gerontol A Biol Sci Med Sci. 2010;65(10):1028–41.
28. Tinken TM, Thijssen DH, Black MA, Cable NT, Green DJ. Time course of change in vasodilator function and capacity in response to exercise training in humans. J Physiol. 2008;586(20):5003–12.
29. Parker BA, Ridout SJ, Proctor DN. Age and flow-mediated dilation: a comparison of dilatory responsiveness in the brachial and popliteal arteries. Am J Physiol Heart Circ Physiol. 2006;291:3043–9.
30. Donato AJ, Eskurza I, Silver AE, et al. Direct evidence of endothelial oxidative stress with aging in humans: relation to impaired endothelium-dependent dilation and upregulation of nuclear factor-B. Circ Res. 2007;100(11):1659–66.
31. Donato AJ, Gano LB, Eskurza I, et al. Vascular endothelial dysfunction with aging: endothelin-1 and endothelial nitric oxide synthase. Am J Physiol Heart Circ Physiol. 2009;297(1):H425–32.
32. Sallam N, Laher I. Exercise modulates oxidative stress and inflammation in aging and cardiovascular diseases. Oxid Med Cell Longev. 2016;2016:239639.
33. Nyberg M, Mortensen SP, Hellsten Y. Physical activity opposes the age-related increase in skeletal muscle and plasma endothelin-1 levels and normalizes plasma endothelin-1 levels in individuals with essential hypertension. Acta Physiol (Oxf). 2013;207(3):524–35.
34. Dawson EA, Alkarmi A, Thijssen DH, et al. Low-flow mediated constriction is endothelium-dependent: effects of exercise training after radial artery catheterization. Circ Cardiovasc Interv. 2012;5(5):713–9.
35. Spieker LE, Lüscher TF, Noll G. ETA receptors mediate vasoconstriction of large conduit arteries during reduced flow in humans. J Cardiovasc Pharmacol. 2003;42(3):315–8.
36. Credeur DP, Dobrosielski DA, Arce-Esquivel AA, Welsch MA. Brachial artery retrograde flow increases with age: relationship to physical function. Eur J Appl Physiol. 2009;107(2):219–25.
37. Laughlin MH, Newcomer SC, Bender SB. Importance of hemodynamic forces as signals for exercise-induced changes in endothelial cell phenotype. J Appl Physiol. 2008;104(3):588–600.
38. Naylor LH, Weisbrod CJ, O’Driscoll G, Green DJ. Measuring peripheral resistance and conduit arterial structure in humans using Doppler ultrasound. J Appl Physiol (1985). 2005;98(6):2311–5.
39. Parker BA, Kalasky MJ, Proctor DN. Evidence for sex differences in cardiovascular aging and adaptive responses to physical activity. Eur J Appl Physiol. 2010;110(2):235–46.


Supplemental Digital Content

Copyright © 2019 by the American College of Sports Medicine