Reproducible improvement in endothelial function following two separate periods of high-intensity interval training in young men

Abstract

High-intensity interval training (HIIT) can improve vascular function, as assessed by brachial artery flow-mediated dilation (FMD). However, when separated by a period of detraining, the reproducibility of FMD responses to repeated periods of HIIT is unknown. The purpose of this study was to determine the group mean and intraindividual reproducibility of FMD responses to two 4-wk periods of HIIT, separated by 3 mo of detraining. Thirteen healthy, recreationally active men (21 ± 2 yr) completed the study. Each 4-wk HIIT period included 40 min of treadmill training four times/week. Each training session included four 7-min intervals: 4 min at 90%–95% heart rate maximum (HR_max) and 3 min at 70%–75% HR_max. Vascular (FMD) and cardiorespiratory fitness (maximal oxygen consumption [V̇o_2max]) assessments were conducted before and following each 4-wk training period. Training resulted in significant improvements in V̇o_2max (P < 0.001). Training also improved FMD (P < 0.001), with no differences between periods (P = 0.394), even after controlling for changes in baseline diameter and the shear rate stimulus. There was a significant, moderate relationship between the change in FMD in HIIT period 1 versus period 2 [R² = 0.493, P = 0.011, intraclass correlation coefficient: 0.600, coefficient of variation: 17.3%]. Consecutive periods of HIIT separated by detraining resulted in similar improvements in FMD at the group level, and individual FMD changes in period 1 of HIIT predicted FMD changes in response to period 2. Considered alongside substantial between-participant variability in magnitude of FMD improvement, this suggests that there are reproducible, interindividual differences in the potential to improve vascular function with HIIT.

NEW & NOTEWORTHY This is the first study examining endothelial function [flow-mediated dilation (FMD)] following repeated periods of high-intensity interval training (HIIT). Two periods of HIIT separated by detraining resulted in reproducible group-level improvements in FMD. Despite considerable between-subject variability in FMD adaptation, individual FMD changes with the first HIIT period predicted FMD changes in the second period. This indicates the existence of reproducible between-subject differences in susceptibility to FMD improvement with HIIT.

INTRODUCTION

Several studies have demonstrated that high-intensity interval training (HIIT) can improve endothelial function, as assessed by reactive hyperemia flow-mediated dilation (FMD) in the upper (brachial artery) and lower (popliteal, femoral artery) limbs (3, 21, 27, 28, 30–32, 34). A recent systematic review and meta-analysis by Ashor et al. (1) found a dose-response relationship between aerobic exercise intensity and improvements in endothelial function. Similarly, a systematic review and meta-analysis by Ramos et al. (28) found that HIIT resulted in greater improvements in FMD than moderate-intensity continuous training. Taken together, these reviews support the notion that when seeking to improve vascular function, HIIT is a viable alternative to traditional forms of moderate-intensity endurance training (11, 12).

In vivo, periods of training are often interspersed with less active periods where detraining occurs; this could occur with injury and illness, lack of time, or other barriers to continuing training (24, 32, 36). It is reasonable to speculate that a previous period of training could influence responses to subsequent periods (e.g., potentiating or attenuating responses over multiple training exposures). However, no studies to date have examined FMD changes in response to multiple periods of training separated by periods of detraining. Thus, the consistency of FMD changes in response to repeated periods of HIIT is unknown. In addition, although the evidence suggests that a single period of HIIT can result in improvements in FMD, there is considerable interindividual variability in the reported magnitude of improvement (21, 30). Given the absence of studies examining responses to multiple periods of training, it is currently unknown whether between-subject variability in the magnitude of FMD improvement reflects reproducible, and therefore potentially meaningful, between-subject differences in susceptibility to training-induced improvement in FMD.

There is considerable evidence that training-induced FMD adaptations are a result of repeated exposures of the endothelium to elevated conduit artery shear stress during training (2, 6, 22, 35). Given that acute exposure to increased shear stress has been shown to result in reproducible acute endothelial responses as measured by FMD over multiple trials (15, 18), it is reasonable to hypothesize that repeated exposure to elevated shear stress via HIIT will result in reproducible improvements in FMD at the group and individual levels.

The purpose of this study was to determine the group mean and intraindividual consistency of FMD adaptation to two periods of HIIT, separated by a 3-mo detraining period. Exercise-induced improvements in endothelial function contribute to the cardiovascular health benefits of exercise (19, 23, 29). Examining the impact of intermittent periods of training is ecologically relevant and essential to begin to elucidate the potential importance of interindividual differences in the impact of training on endothelial function.

MATERIALS AND METHODS

Participants

Twenty-two healthy, young (18–30 yr), nonsmoking, recreationally active men were recruited from the Queen’s University and Kingston community as part of a larger study. The primary outcome of this larger study by Del Giudice and colleagues (8) was cardiorespiratory fitness in healthy young men. The present study recruited 17 participants from the larger study, and all of them met the inclusion/exclusion criteria. However, only 13 participants completed the additional vascular testing component owing to scheduling difficulties (n = 3) or an injury unrelated to the training (n = 1). Two of the 13 participants who completed the vascular testing sessions did not complete the cardiorespiratory fitness testing. As a result, in the present study, the n is 13 for all measures, apart from an n of 11 for the maximal oxygen consumption (V̇o_2max) measure. The study was submitted to, and approved by, the Queen’s University Health Sciences and Affiliated Hospitals Research Ethics Board. Prior to participation, volunteers were required to provide written informed consent on a form approved by the same Board.

Screening and Familiarization Visits

Prior to commencing the study, volunteers attended a screening visit to become familiar with the laboratory environment and to assess their eligibility for participation, as previously described (8). Briefly, volunteers who reported cardiovascular or metabolic diseases, contraindications to exercise participation, taking prescription medication, current involvement in a training program or participation in >3 h of structured moderate to vigorous activity per week were excluded. Participants also underwent a vascular screening, where a brachial artery image and a velocity signal were acquired to ensure that a strong signal and a clear image could be detected. Finally, the participants took part in a familiarization visit, where a cardiorespiratory fitness (V̇o_2max) test was completed in the week before the first experimental visit, in each period of training, as previously published (8).

Experimental Protocol

All participants attended four experimental visits for vascular testing (pre and post each of two 4-wk HIIT training periods) as described in Fig. 1. Vascular testing posttraining occurred 4–6 days following the last HIIT session (within subjects the number of days between the last HIIT session and vascular testing was the same in periods 1 and 2). After each vascular testing visit, V̇o_2max was assessed (Fig. 1).

Fig. 1.

Experimental protocol timeline. The posttraining vascular visit following each training period occurred 4–6 days following the last high-intensity interval training (HIIT) session. Each maximal O₂ consumption (V̇o_2max) test was separated by ~24–48 h.

Vascular testing.

Prior to the vascular testing visits, the participants were instructed to abstain from food, alcohol, and caffeine for 12 h and to avoid moderate-to-vigorous physical activity for 24 h. The participants came to the laboratory in the morning (5 AM–11 AM), at the same time for each visit (within subject) to control for diurnal variation in FMD (9). Upon arrival, the participants were asked to lie supine with their arms extended for a 20-min rest period and were instrumented for heart rate and right-arm blood pressure assessments. At the end of the rest period, one standard brachial artery reactive hyperemia FMD test was performed to assess endothelial function. All FMD tests were performed on the left arm.

High-intensity interval training and detraining.

As previously described (8), the participants took part in two 4-wk periods of HIIT, with a 3-mo detraining period in between. Each period of HIIT consisted of 4 wk of treadmill running training sessions, four times/ week, 40 min/session. Each training session included a 10-min warm-up to 70%–75% heart rate maximum (HR_max), followed by four intervals at the following levels: 4 min of higher-intensity exercise at 90%–95% HR_max with 3 min of lower-intensity exercise at 70%–75% HR_max between the higher-intensity intervals. The training sessions concluded with a 5-min cool-down at 70%–75% HR_max. If target HR_max was not achieved within 2 min of each interval, speed or incline was increased depending on the participant’s preference and recorded for future intervals. The participants’ HR_max was determined based on their highest heart rate (HR; Polar Team2 Pro, Kempele, Finland) achieved during any of the three V̇o_2max tests (familiarization and two experimental V̇o_2max tests) before each training period. During the detraining period, the participants were instructed to return to their previous activity levels and to discontinue HIIT training; however, activity levels were not assessed during this period.

Experimental Procedures

Heart rate and mean arterial pressure.

HR was measured continuously via a single-lead, three-electrode electrocardiogram throughout the vascular testing visit. HR signals were recorded in the program LabChart (AD Instruments, Colorado Springs, CO) for later analysis. Mean arterial pressure (MAP) was assessed with an automated sphygmomanometer and was characterized by an average of five automated measurements (BpTRU BPM-100, BpTRU Medical Devices, Coquitlam, BC, Canada).

Brachial arterial diameter and velocity measurements.

Brachial artery diameter was measured using two-dimensional ultrasound in B-mode (12 MHz, Vivid I, Vivid i2 Medical Systems, Mississauga, Canada), using an insonation angle of 68° following calibration at the angle, as previously described (25). Ultrasound images were recorded with a video graphics array (VGA) to USB frame grabber (Epiphan Systems Inc., Ottawa, Canada) and saved as .avi files on a separate computer using Camtasia Studio (TechSmith, Okemos, MI), as previously described (18). Brachial artery blood velocity was measured using Doppler ultrasound (4 MHz, Vivid i2 GE Medical Systems, Mississauga, Canada). The Doppler shift frequency spectrum was analyzed using a Multigon 500P TCD (Multigon Industries, Yonkers, NY) and was sampled continuously using Powerlab and stored via LabChart (AD Instruments, Colorado Springs, CO).

Reactive hyperemia FMD.

The FMD assessment was performed according to recent guidelines (33) and as previously described (18). Briefly, a pneumatic cuff was placed around the left forearm, distal to the location of the ultrasound probe. Brachial artery diameter and blood velocity measurements were recorded during 1 min of baseline before cuff inflation to 250 mmHg for 5 min. Diameter and velocity measurements were recorded in the last minute of cuff inflation and for 3 min following cuff release. The ultrasound probe location was measured from the antecubital fossa and anatomical landmarks were located, to ensure that assessments were taken at the same artery location during each visit.

Cardiorespiratory fitness (V̇o_2max) testing.

As previously described (8), the participants completed an incremental treadmill V̇o_2max test, and gas exchange variables were measured throughout the testing using a metabolic cart (Moxus, AEI Technologys, Pittsburgh, PA). Briefly, the participants started the test with 3 min of standing rest, followed by a 5-min warm-up at a slow walking speed (2.5 mi/h and 2% incline). Participants then took part in incremental ramp stages, with each 2-min stage increasing either the speed or incline until the participant reached volitional fatigue. A supramaximal confirmation was then performed following 10 min of rest, with the same incline and a speed of 0.5 mi/h faster than that achieved at the end of their ramp test, until volitional fatigue. Participants completed this protocol twice (separated by 24–48 h) at each measurement time point shown in Fig. 1.

Data Analysis

Heart rate and mean arterial pressure.

Resting HR was reported as the 1-min average recorded during the baseline minute of the FMD test, before cuff inflation. Resting MAP was reported using the average of five automatic blood pressure measures at the start of the vascular testing visit, using the following formula: MAP = [systolic blood pressure + 2 (diastolic blood pressure)]/3.

Brachial artery diameter and blood velocity analysis.

Diameter analysis was completed by a single researcher (J.S.W.) blinded to time (pretraining vs. posttraining) for each training period. Artery diameter was analyzed using automated edge-detection software (Encoder FMD and Bloodflow v3.0.3, Reed Electronics, UK), as described previously (18). Diameter values were compiled into 3-s time bins for FMD and shear rate analysis. Blood velocity was analyzed in 3-s average time bins using LabChart software (AD Instruments, CO), as described previously (25).

Flow-mediated dilation.

FMD was reported as a percent change (%FMD) and an absolute change (AbsFMD) in diameter. %FMD and AbsFMD were calculated as the percent difference and absolute difference, respectively, between the baseline minute and the peak 3-s average diameter bin, after cuff release.

Shear rate.

Shear rate (SR) was calculated using the time-aligned 3-s average blood velocity and diameter time bins, using the following equation: 4 × mean blood velocity/artery diameter. SR was evaluated as an area under the curve to peak diameter following cuff release (SRAUC to peak diameter) (26). Baseline SR was reported as the average shear rate in the 1 min before cuff inflation.

Cardiorespiratory fitness (V̇o_2max).

V̇o_2max analysis has been described in detail previously (8). Briefly, V̇o_2max was calculated as the average of the highest 30-s interval V̇o₂ during the final stage of the incremental ramp protocol and the corresponding supramaximal confirmation for each of the two fitness testing visits at each time point (pretraining and posttraining). This resulted in a V̇o_2max that was the average of four values for each time point (pretraining and posttraining).

Statistical Analysis

All data are reported as mean ± standard deviation (SD), and statistical significance was set at P ≤ 0.05. All analyses were performed using IBM SPSS, Version 24 (SPSS Inc., Chicago, IL). To examine group-level effects, a linear mixed model, with factors time (pretraining, posttraining) and HIIT period (HIIT period 1, HIIT period 2), was used to assess all variables: body mass index (BMI), V̇o_2max, baseline diameter, peak diameter, time to peak diameter, baseline SR, baseline HR, baseline MAP, the SR stimulus (SRAUC to peak diameter), AbsFMD, and %FMD. FMD analysis was repeated with the following as covariates: the SR stimulus (SRAUC to peak diameter) and the baseline diameter. A linear regression between individual changes in FMD during HIIT period 1 (post-HIIT FMD – pre-HIIT FMD) and HIIT period 2 was performed. We also performed a linear regression between FMD during pre-HIIT period 1 and pre-HIIT period 2. Cook’s distance analysis was performed to identify outliers (16). To further quantify the agreement between the two measurements, Bland–Altman plots were created for both delta%FMD and %FMD premeasurements. Bland–Altman plots represent the mean of the two measurements plotted against the difference between the measurements. Reproducibility was also examined using intraclass correlation coefficients (ICCs: two-way random effects, absolute agreement, single-rater/measurement), with 95% confidence intervals (CIs), and within-subject coefficient of variation [CV = (standard deviation/mean) × 100%]. ICC values were interpreted as indicators of reproducibility, using the following criteria: poor (<0.50), moderate (0.50–0.75), good (0.75–0.9), or excellent (>0.90) (20). Mean absolute bias was also determined for the change in %FMD between the two HIIT periods (|delta%FMD period 1| − |delta%FMDperiod2|).

RESULTS

Participant Characteristics (Age and BMI) and V̇o_2max

Participants were young men with an average age of 21 ± 2 yr. BMI remained unchanged across the testing sessions (Table 1). V̇o_2max was significantly elevated following the training periods (main effect of time: P < 0.001; Table 1). V̇o_2max was also significantly lower during period 2 than period 1 (main effect of period: P = 0.002; Table 1). There was no interaction between period and time.

Table 1.

Cardiorespiratory, hemodynamic, and vascular results

	HIIT Period 1		HIIT Period 2		Main Effects and Interactions P Values
	Pretraining	Posttraining	Pretraining	Posttraining
Body mass index (kg/m2)	23.1 ± 1.9	23.0 ± 1.9	21.2 ± 2.2	23.1 ± 1.8	T P = 0.331 P P = 0.873 T × P P = 0.197
Cardiorespiratory fitness (V̇o2max), mL·kg·−1min−1	59.5 ± 7.7	63.5 ± 9.3	57.9 ± 8.1	61.2 ± 9.2	T P < 0.001 P P = 0.002 T × P P = 0.565
Baseline resting HR, beats/min	57 ± 8	52 ± 7	60 ± 7	52 ± 6	T P < 0.001 P P = 0.256 T × P P = 0.398
Baseline resting MAP, mmHg	76 ± 6	76 ± 6	77 ± 5	77 ± 5	T P = 0.908 P P = 0.459 T × P P = 0.624
Baseline SR, s−1	76.6 ± 45.4	55.7 ± 15.9	89.5 ± 51.0	80.9 ± 46.9	T P = 0.138 P P = 0.057 T × P P = 0.528
SRAUC to peak diameter, ×104 s−1	1.58 ± 0.85	1.43 ± 0.53	1.97 ± 0.79	1.56 ± 0.49	T P = 0.062 P P = 0.085 T × P P = 0.384
Baseline diameter, mm	3.90 ± 0.55	3.92 ± 0.49	3.80 ± 0.37	3.77 ± 0.31	T P = 0.960 P P = 0.042 T × P P = 0.723
Peak diameter, mm	4.10 ± 0.56	4.17 ± 0.48	4.00 ± 0.35	4.03 ± 0.33	T P = 0.452 P P = 0.050 T × P P = 0.671
Time to peak diameter, s	45.9 ± 22.3	53.5 ± 37.9	57.9 ± 34.7	44.5 ± 6.8	T P = 0.702 P P = 0.842 T × P P = 0.169
AbsFMD, mm	0.20 ± 0.09	0.25 ± 0.10	0.21 ± 0.07	0.26 ± 0.09	T P < 0.001 P P = 0.626 T × P P = 0.770

Data are means ± SD, n = 11 for V̇o_2max, n = 13 for all other data. AbsFMD, absolute flow-mediated dilation; AUC, area under the curve; HR, heart rate; MAP, mean arterial pressure; P, training period; SR, shear rate; T, time; Significant main effects are bolded.

Baseline Hemodynamic Variables (HR and MAP)

Resting HR was significantly lower following training (main effect of time: P < 0.001; Table 1) with no effect of period or interaction between period and time. MAP was unaltered by training (Table 1).

Baseline SR and SR Stimulus (SRAUC to Peak Diameter)

Baseline SR and the shear rate stimulus (SRAUC to peak diameter) were unaltered by training (Table 1).

Baseline Diameter, Peak Diameter, and FMD

Baseline diameter and peak diameter were significantly lower during the second training period (main effect of period: P = 0.042 and P = 0.050, respectively; Table 1). Time to peak diameter was not impacted by training (Table 1). Training resulted in a significant improvement in FMD with no differences between periods and no interaction between period and time (main effect of time: P < 0.001; Fig. 2), and the same result was found for AbsFMD (main effect of time: P < 0.001; Table 1). FMD analysis with the addition of the SR stimulus (SRAUC to peak diameter) or baseline diameter as a covariate yielded similar results (main effects of time only: P = 0.001 and P < 0.001, respectively).

Fig. 2.

Percent change flow-mediated dilation before (pre) and after (post) 4 wk of high-intensity interval training, during two periods of training (period 1, period 2); ● and lines reflect individual responses, and the bars depict the group means. *Significantly different from pretraining. T, time; P, training period. %FMD: Percent change flow-mediated dilation

Assessments of Reproducibility

After the removal of an outlier, a significant relationship between FMD changes in period 1 and period 2 was observed [R² = 0.493, P = 0.011, ICC: 0.600 (95% CI: 0.054, 0.867), CV: 17.3%, Fig. 3A; before outlier removal: R² = 0.145, P = 0.199, ICC: 0.374 (95% CI: −0.234, 0.760), CV: 57.6%)]. The mean absolute bias between FMD changes in period 1 and period 2 was 1.13 ± 1.31% without outlier removal and 0.88 ± 1.00% with outlier removal. The Bland–Altman plot for delta %FMD shows a bias near 0 (0.161) and all but two datapoints inside the 95% limits of agreement (+3.60, −3.28; Fig. 3B). A significant relationship between FMD pre-period 1 and pre-period 2 was also observed, with no outliers identified [R² = 0.641, P = 0.001, ICC: 0.793 (95% CI: 0.462, 0.931), CV: 16.6%; Fig. 3C]. The Bland–Altman plot for pre-FMD shows a bias near 0 (−0.362) and all datapoints inside the 95% limits of agreement (+2.62, −3.34; Fig. 3D).

Fig. 3.

A: linear regression for the change in percent change flow-mediated dilation (%FMD) from training period 1 and period 2, with removal of an outlier (circle). Intraclass correlation coefficient (ICCs) represented with 95% confidence intervals. B: Bland–Altman plot representing the mean of the delta change in %FMD across periods, plotted against the difference in delta change %FMD across periods. C: linear regression for the pretraining %FMD from training period 1 and period 2. ICCs represented with 95% confidence intervals. D: Bland–Altman plot representing the mean of the pre-FMD measurements across periods, plotted against the difference in pre-FMD across periods.

DISCUSSION

The purpose of this study was to determine, for the first time, the impact of repeated periods of training on brachial artery FMD. In agreement with our hypothesis, two consecutive periods of high-intensity interval training, separated by 3 mo of detraining, resulted in similar improvements in FMD at the group level. In addition, responses to training period 1 predicted responses to training period 2, with moderate reliability, suggesting that within an individual, the conduit artery endothelium response is fairly reproducible when repeatedly exposed to the same training stimulus. Considered alongside the substantial between-participant variability in the magnitude of FMD improvement, these results suggest that there are reproducible, interindividual differences in the potential to improve vascular function with HIIT.

Vascular Adaptations to Training and Detraining: Group Response

In the present study, we observed a significant improvement in FMD following HIIT training, which is consistent with previous studies (3, 21, 27, 28, 30–32, 34). However, the magnitude of improvement (1.29% FMD, averaged across periods) was less than improvements reported in two recent meta-analyses [2.79% FMD (weighted average) (1); 2.26% FMD (28)]. The discrepancy in the magnitude of improvement may be explained by differences in the duration of training or the health status of the participants. The majority of studies in the meta-analysis by Ashor et al. (1) examined training protocols that were 12 wk in length, whereas the present study protocol was only 4 wk in length (28). It is possible that a longer duration of training would have resulted in greater vascular adaptations (30). In addition, both meta-analyses identified (1, 28) examined primarily middle-aged to older patients with varying cardiovascular and metabolic disorders, which may have contributed to a greater potential for improvement in endothelial function with training, compared with the present study that examined healthy young men. Indeed, a previous study examining 8 wk of HIIT in healthy young men has reported a magnitude of improvement in brachial artery FMD similar to that observed in the present study [1.46% FMD (21)].

Following the 3-mo period of detraining, it appears that the vascular adaptations gained from the first period of training were lost (Fig. 2), alongside losses in cardiorespiratory fitness (Table 1). This loss of vascular adaptation is consistent with previous literature reporting detraining effects in as little as 1–2 wk following a period of HIIT (24, 32, 36). In addition, we showed for the first time that the magnitude of improvement in FMD did not differ between two consecutive periods of HIIT, suggesting that a previous period of training does not influence brachial artery adaptation to a second period. Thus, although positive adaptations in endothelial function with training appear to be fully reversed with detraining, these appear to be fully regained when training resumes.

Vascular Adaptations to Training: Individual Variability

There was considerable interindividual variability in the vascular adaptations to HIIT training, such that improvements ranged from 0.4% to 99.3%, and there were two instances where FMD was lower posttraining. Although some studies have demonstrated fairly consistent interindividual responses to HIIT (31, 37), in agreement with the present study, others have identified considerable interindividual variability (5, 30). For example, in a sample of obese young adults, Sawyer et al. (30) observed a group-level increase in FMD at 4 wk of HIIT; however, ~20% of participants experienced a decline in FMD posttraining. In contrast to the considerable interindividual variability, the significant relationship between FMD adaptation to period 1 and period 2 of HIIT indicates that adaptations to the same training stimulus demonstrate consistency at the individual level (Fig. 3A). The moderate ICC (0.600), Bland–Altman plot centered near 0, a small mean absolute bias, and CV of 17.3% also support relatively consistent individual responses to HIIT. This reproducible adaptation to training occurred in the context of reproducible baseline FMD across a 3-mo period [evidenced by the strong relationship between pre-FMD period 1 and pre-FMD period 2, the “good” ICC (0.793), Bland–Altman plot centered near 0, and the CV of 16.6%]. Taken altogether, this suggests that the interindividual variability reflects “trait-like” differences in endothelial cell sensitivity to the HIIT training stimulus.

As previously reported by Del Giudice et al. (7), there were similar group-level improvements in V̇o_2max over each training period. However, unlike the present FMD data, Del Giudice et al. (8) observed poor within-participant reproducibility for V̇o_2max (P = 0.15, r = 0.40, ICC = 0.369, CV = 74.4%). The authors highlight that the lack of reproducibility of V̇o_2max may be due in part to poor reproducibility of underlying mechanisms in skeletal muscle (8). This is supported by recent research by the same group demonstrating poor reproducibility of skeletal muscle mRNA expression in response to two acute bouts of aerobic exercise in a similar population of active young men (17). In comparison, the present data suggest that mechanisms underlying the FMD response to training may exhibit greater reproducibility.

Mechanistically, the improved endothelial function observed in this study may have been due to shear-stress-induced improvements in nitric oxide (NO) bioavailability, through upregulation of endothelial nitric oxide synthase (eNOS) expression and/or through improved antioxidant balance and ability to scavenge reactive oxygen species (4, 10, 13, 14, 37). For example, Cocks et al. (4) identified increased eNOS protein content following 6 wk of sprint interval training, and Wisloff et al. (37) found that FMD improvements with training correlated with increases in total antioxidant status of blood plasma. Taken altogether, it is possible that the participants in the present study who consistently experienced improved FMD may have had greater increases in eNOS content and/or total antioxidant status with training, and thus increased NO bioavailability. However, further research exploring the mechanisms underlying interindividual variability in vascular adaptations to HIIT is needed.

Limitations

This study did not include a control group. However, previous results from our laboratory in a similar sample of healthy, recreationally active young men have demonstrated that FMD remains unaltered following 4 wk with no training [n = 13, pre: 5.9 ± 1.8%, post: 5.3 ± 1.8%, P = 0.207 (unpublished observations)]. This, in combination with the observed consistency in FMD changes, suggests that it is unlikely that the results are due to a systematic effect of time rather than training. Owing to time constraints, we were only able to perform one FMD test per visit. This may have resulted in an underestimation of the consistency of responses because averaging multiple FMD trials each visit would have provided a superior characterization of endothelial function by reducing variability (7). Also, activity levels of participants were not assessed during the detraining period; as a result, there is no objective assessment of their return to pretraining levels of activity. However, given that both cardiorespiratory fitness and vascular function returned to pretraining period 1 levels, it is likely that the detraining stimulus was sufficient without monitoring. This study also involved a homogenous population of healthy young men, and thus, we cannot generalize the findings to women, older, or clinical populations. Given the paucity of research on women in exercise training studies, expanding this work to include women is a particularly important priority for future research. Finally, we were not able to assess variables related to NO bioavailability that could have provided additional insight regarding the mechanisms underlying the FMD adaptations.

CONCLUSIONS

This study demonstrated for the first time that improvements in FMD are reproducible at the group level across two identical periods of HIIT separated by detraining. In addition, although there was considerable interindividual variability, the magnitude of training-induced adaptations in FMD exhibited consistency across HIIT periods at the individual level. These results suggest that there are persistent, between-subject differences in the potential to improve vascular function with HIIT training. Further research is needed to investigate the impact of interindividual differences in the vascular adaptations to HIIT on long-term cardiovascular protective effects of exercise. In addition, further studies are needed to examine the mechanisms underlying individual differences and to examine the consistency of responses to different training modalities in a range of populations.

GRANTS

This research was funded by Natural Science and Engineering Research Council of Canada (NSERC) Discovery Grants to K.E. Pyke and B.J. Gurd. J.S. Williams was supported by funding through an NSERC Canadian Graduate Scholarship – Master’s.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

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