High-pressure blood flow restriction with very low load resistance training results in peripheral vascular adaptations similar to heavy resistance training

During the first visit, participants were informed of the purpose of the study as well as possible risks related to resistance exercise and BFR. Exclusion criteria were reviewed with participants and included the following: younger than 18 or older than 35 years of age, a BMI of  ⩾30 kg m⁻², regular resistance exercise performed during the prior six months, or taking medication for hypertension. If participants did not meet any of the exclusion criteria and still wished to participate, they signed a written informed consent document. Height and body mass were measured prior to vascular measurements.

Forearm and calf vascular conductance and venous compliance were measured and calculated using strain gauge plethysmography. Prior to measurements, participants were instructed to avoid caffeine for 8 h, food for 2 h, and any exercise or alcohol consumption for 24 h. Participants lay supine on an exam table in a quiet, temperature-controlled room for 10 min prior to any measurement. Brachial systolic (SBP) and diastolic (DBP) blood pressure was measured using an automated blood pressure device (Omron Healthcare Inc., Vernon Hills, IL) with an appropriately sized cuff at 1 min intervals until two consecutive measures were recorded within 5 mmHg. MAP was calculated from the average blood pressure of the two consecutive measurements using the formula MAP  =  (2/3) DBP  +  (1/3) SBP. In order to control for the short-term BFR caused by the blood pressure measurement, blood pressure was measured in the arm opposite the one in which plethysmograph measures were occurring. Blood pressure measurements were performed immediately before plethysmography measurements.

The circumference was measured at the widest portion of the forearm distal to the olecranon. A mercury-in-silastic strain gauge at least 2 cm smaller than the forearm circumference and connected to the EC-6 Plethysmograph (Hokanson Inc., Belleville, WA) was placed around this portion of the forearm and held in place with a thin strip of athletic tape. A 5 cm cuff was placed around the wrist while a 10 cm cuff was placed around the arm proximal to the elbow. The participant's arm was abducted and supported approximately 90° from the midline of the body. A small foam pad was placed under the 10 cm cuff to support the arm while a foam pad supported the wrist and hand, elevating the forearm and the strain gauge to heart level without impeding or moving the strain gauge. The 5 cm cuff was inflated to 220 mmHg using a hand pump and remained inflated until the measurement protocol was completed in order to exclude hand blood flow from the measurement and capture flow only through the forearm. One minute following inflation of the wrist cuff, the 10 cm cuff was inflated to 50 mmHg via a rapid cuff inflator (Hokanson Inc., Belleville, WA). The cuff was inflated for 7 s and then deflated for 8 s. This inflation/deflation was repeated while the strain gauge recorded arterial inflow until five measurements were recorded using the Noninvasive Vascular Program 3 (NIVP3) software (Hokanson Inc., Belleville, WA). The slope of the change in the strain gauge circumference across the measurement duration was calculated in NIVP3, and the average of these five measurements was recorded as forearm blood flow (FBF). Forearm vascular conductance (FVC) was calculated according to the formula FVC  =  (FBF/MAP)  ×  1000.

The 5 cm cuff at the wrist was deflated and removed. Five minutes later, the 10 cm cuff was inflated to 20 mmHg for 1 min while strain gauge measurements of venous volume variation were recorded in NIVP3. The cuff was deflated for 1 min, followed by a 2 min inflation at 40 mmHg. One minute deflation intervals followed this and the next measurements of 3 min at 60 mmHg and 4 min at 80 mmHg. Venous compliance (C_V) was calculated as the slope of the regression line of the venous volume variations at each of the four measurement pressures (Bleeker et al 2004).

Following five further minutes of quiet rest, vascular conductance and venous compliance measures were performed on each of the participant's calves in the same manner and order as described above.

Four exercise conditions were compared using unilateral biceps curls and unilateral knee extensions: (1) 15% of one-repetition maximum (1RM) with no BFR (15/0), (2) 15% 1RM at 40% of arterial occlusion pressure (AOP) (15/40), (3) 15% 1RM at 80% AOP (15/80), and (4) 70% 1RM with no BFR (70/0). These conditions were chosen based upon prior research performed in our laboratory (Dankel et al 2017, Mouser et al 2017b). Participants were randomly assigned two of the four possible exercise conditions in the upper and lower body (i.e. one condition per limb) in a counter-balanced fashion such that no participant received the same exercise condition in the upper and lower limbs.

Prior to an exercise session, if a participant was performing blood flow restricted exercise, AOP was measured using a bidirectional Doppler probe (MD-6, Hokanson Inc., Belleville, WA). Briefly, the 5 cm cuff being used during BFR (SC5, Hokanson Inc., Belleville, WA, USA) was placed over the proximal-most portion of the arm or leg while the Doppler was placed over the radial artery or posterior tibial artery, respectively. The cuff was inflated to 50 mmHg and the pressure was slowly increased until no pulse could be detected aurally. The pressure at which occlusion occurred was recorded as AOP and used to calculate the relative (submaximal AOP) pressures used during training.

All participants in the 15% 1RM conditions performed four sets of exercise to a metronome cadence of 1 s concentric and 1 s eccentric with 30 s of rest between each set. Sets were limited to 90 repetitions each for time considerations and to limit as much as possible oxidative-based adaptations (Burd et al 2012). Participants in the high-load condition performed four sets of exercise to muscular failure with 90 s of rest between sets. Participants performed one set of exercise in each arm and each leg on the first training visit, two sets of the second visit, three sets on the third and fourth visit, and then four sets on the fifth through final visits. All exercise training and BFR was overseen by qualified laboratory personnel.

Separate statistical analyses were performed for the upper and lower body. Covariance pattern models were used to perform two-factor (2 by 4; time by condition) analysis of variance to examine the vascular responses to eight weeks of resistance exercise. This approach allows for repeated measures on time and also accounts for the dependency created because each participant contributed observations in two of the four possible exercise conditions, but not the same condition in both limbs of the upper or lower body. Using a person-period dataset (i.e. long format) with four observations per participant (one for each limb at pre, one for each limb at post), two different models of the error covariance structure, compound symmetry and unstructured, were initially estimated and compared. Akaike's information criterion (AIC) and Bayesian information criterion (BIC), both of which are statistical analyses examining the fit of the data to the model (Burnham and Anderson 2004), were calculated and compared. The model with the lower value of both AIC and BIC was chosen as this represented a better fit of the data to the model. If a significant condition by time interaction was found, simple effects were examined, otherwise main effects were investigated. Statistical significance was set a priori at α  =  .05, and all analyses were performed using SPSS 24 (IBM Corporation, Armonk, NY, USA).

In the upper body, a statistically significant condition by time interaction was found (p   =  .032) showing that the FVC response was lower in the 15/40 condition than in the 15/80 (−8.70 (−16.07, −1.33) ml · mmHg⁻¹, p   =  .021) and 70/0 (−9.01 (−16.29, −1.73) ml · mmHg⁻¹, p   =  .016) conditions. There was a training effect only in the 15/80 (+8.29 (3.01, 13.57) ml · mmHg⁻¹, p   =  .002) and 70/0 (+8.595 (3.45, 13.74) ml · mmHg⁻¹, p   =  .001) conditions.

A statistically significant condition by time interaction was found in the lower body (p   =  .004). The calf vascular conductance response was lower in the 15/0 condition than in the 15/80 (−9.07 (−15.42, −2.73) ml · mmHg⁻¹, p   =  .005) and 70/0 (−8.42 (−14.77, −2.07 ml · mmHg⁻¹, p   =  .010) conditions. Similarly, the 15/40 condition had a vascular conductance response that was lower than the 15/80 (−8.72 (−15.24, −2.20) ml · mmHg⁻¹, p   =  .009) and 70/0 (−8.06 (−14.59, −1.54) ml · mmHg⁻¹, p   =  .016) conditions (table 2).

In the upper body there was no interaction (p   =  .713) or main effect of condition (p   =  .813). There was a statistically significant main effect of time (p   =  .044), with venous compliance increasing from pretesting values across the duration of the study (+0.003 (.000 08, 0.006) ml · 100 ml⁻¹ · mmHg⁻¹). In the lower body there was no interaction (p   =  .335), nor were there main effects of time (p   =  .204) or condition (p   =  .684) (table 2).

A unique finding of our study was that both forearm and calf vascular conductance increased in the very-low-load, high-pressure condition as well as the high-load condition, and these were not statistically significantly different from each other. There are several possibilities for this finding. Two studies examining the angiogenic signaling response to blood flow restricted exercise found increases in endothelial nitric oxide synthase (eNOS) and vascular endothelial growth factor (VEGF) mRNA expression (Larkin et al 2012, Ferguson et al 2018). Increased eNOS activity leads to a more robust vasodilation, while VEGF is a potent angiogenic signal found in high concentrations in the muscle interstitium (Hoier and Hellsten 2014) following exercise, and is released from vesicles in the muscle fibers via repeated contractions. Similarly, hypoxia-inducible factor 1  −  α (HIF1  −  α) mRNA is increased following blood flow restricted exercise (Drummond et al 2008) as well as high-load exercise (Ribeiro et al 2017), and is implicated in an increase in capillarization as well (Krock et al 2011). Taken together, these acute increases in mRNA expression could signal long-term adaptations with repeated exposure.

In high-load resistance exercise, mechanical compression of the vessels is much greater than during low-load/very-low-load resistance exercise; mechanical compression and tissue stretch are also known to produce an angiogenic environment. It has been hypothesized that venous occlusion could increase the circumferential strain on the microvasculature (Price and Skalak 1994), leading to the release of angiogenic factors and a greater proliferation of arterioles and capillaries. That vascular adaptations occurred only in the high-pressure group (80% AOP) and not the moderate (40% AOP) or no pressure (0% AOP) groups tells us the following: 15% of 1RM is itself not mechanically able to cause an increase in angiogenic signaling sufficient to bring about adaptation, and 15% of 1RM combined with the moderate 40% of AOP is not able to increase circumferential strain to a degree that would bring about adaptation. These conclusions hold for both the upper and lower body. Past research investigating the ability of intermittent venous occlusion to treat peripheral vascular disease (Collens and Wilensky 1937) as seen with uncontrolled diabetes may need to be reexamined, and future research could focus on the application of high pressures alone to determine its contribution to capillary growth.

Venous compliance, which contributes to orthostatic tolerance, increased in the upper body only. Other research examining the venous response to BFR has been performed, primarily in the lower body. One study examined walk training while under BFR and found that calf venous compliance increased (Iida et al 2011). Another study examined calf venous compliance following knee extensor training and saw no change in compliance (Fahs et al 2014). This disparity in results can be explained by the different modes of exercise as well as the location of the measurement. The calf muscles are activated during walking, and so the calf muscle pump is activating with each step. Combined with an increased pressure load from the application of the pneumatic cuffs, the pressure in the veins will increase to a greater degree during walking. With knee extension exercise, the muscles of the lower leg are not always activated, and differences in knee extension technique below the quadriceps (i.e. isometric dorsiflexion versus plantarflexion versus relaxed) could alter the muscle pump in the calf.

The location of the measurement also plays a role in the results observed here. The veins of the legs are under higher pressures than those of the arms during everyday activities (Rowell 1993), and so it is possible that any amount of activity, from very-low-load resistance exercise to high-load resistance exercise places the arm veins under greater pressures than they would be subject to otherwise. That argument notwithstanding, the increase in venous compliance in the forearms was quite small, and caution should be taken when interpreting these results.

It should be noted that we employed the use of strain gauge plethysmography, which is an indirect measure of vascular proliferation and function. Although there is good agreement between ultrasound measures and plethysmography (Green et al 2011), plethysmography is unable to measure blood flow oscillatory patterns. Vascular function has been shown to decrease (Credeur et al 2010) following handgrip exercise performed under BFR, likely due to disturbances in these patterns, shown previously by Paiva and colleagues (Paiva et al 2016). Future research could directly measure responses to this type of training via muscle biopsy and capillary quantification, as well as measure blood flow patterns during exercise. We also elected to study adaptations to both the upper and lower body to provide a more comprehensive understanding of the vascular adaptations to different forms of training. Although adaptations are thought to be local, it is unknown if training all four limbs in a given session impacted the changes. Future research could explore this further. Lastly, we did not have a non-exercise control group, which would have allowed us to better account the error of the tester as well as the error associated with random biological variability over an 8 week intervention. However, the similar responses observed with vascular conductance in the upper and lower body gives some confidence that this is an effect of the treatment.