Blood flow restriction, effects on SmO2, DFA a1 and HR

Many years ago, at the initial stages of this blog, my intent was to provide some much needed factual perspective to the muscle O₂ desaturation topic. As readers can see, things have diverged markedly since that time, with the focus turning to dynamic HRV and DFA a1 in particular. But if we look back at how this all started, my interest was using NIRS to guide low load strength training (LLST) in a rehab setting. The field of LLST has since exploded, primarily around the use of blood flow restriction (BFR) with very low load lifting. I was well aware of BFR well before I was injured, but from a practical standpoint, it just was not feasible due to its cost and unclear benefits to muscles proximal to the occlusion cuff. For that reason, I chose the LLST approach with muscle O₂ monitoring as a proxy for BFR effect (see my first posts).

However, now there are many relatively affordable consumer cuff options (some even Bluetooth controlled), and studies have shown benefit to muscle groups proximal and distal to the cuff. In addition, a growing body of literature suggests that BFR training during "aerobics" could benefit fitness and performance, including VO2 max. I was recently asked by a colleague about the potential effects on a1 from BFR during aerobics. Unfortunately, I was unable to provide any helpful information, since there has never been any study looking at that. As this question focused on my interest in dynamic HRV/DFA a1 and my initial blog posts about "simulated" BFR, I decided to "come out of retirement" and do some N=1 testing once again. I do want to emphasize that this is a single experience with consumer cuff equipment, so please take that into consideration. This post will be an initial look at setup, basic science, and some practical considerations. Hopefully, after a couple of months of BFR training, I will have some "outcome" metrics of fitness change (VO2 max, Wingate 30s, thresholds). 

How could BFR improve endurance?

A) Larger metabolic stress at an easy pace
By trapping blood and metabolites, the cuffed muscle becomes locally hypoxic and acidic sooner. That ramps motor-unit recruitment and metabolic signaling—at low external workloads—which helps explain the V̇O₂-kinetics improvements seen with easy cycling + BFR. PubMed   

Where this becomes more important is in cases of injury or chronic disease. For instance, post-knee surgery or even chronic arthritic conditions, where the individual just can't train at very high loads/intensity

B) Better oxygen transport & delivery
Longer-term, BFR can remodel the vascular system. In human training studies, BFR increased femoral artery diameter and thigh O₂ delivery; reviews also highlight endothelial function (FMD) and VEGF-linked angiogenesis improvements—plausible routes to better endurance economy and fatigue resistance. PubMed+2The Physiological Society+2

C) Microvascular remodeling (more/better capillaries)
Biopsy and imaging work show capillary neoformation and altered microvascular morphology after short-term, high-frequency low-load BFRE. More capillaries improve diffusion distance for oxygen and metabolite clearance—great for sustained aerobic work. Frontiers+1

D) Ion handling and high-intensity tolerance
That one-leg cycling study didn’t just move performance—it improved K⁺ regulation (and related buffering/transporter adaptations), which supports repeated surges and end-of-race kicks. PubMed

E) Mitochondrial signaling
Reviews suggest BFR’s hypoxia + ROS add to classic endurance signals (think PGC-1α pathways). Not every study finds superior molecular responses vs. normal training, but the functional outcome—faster V̇O₂ kinetics—shows up reliably when programs are dosed well. The Physiological Society+1

What the strongest endurance-relevant studies show

1) Cycling intervals with one leg cuffed (6 weeks)
A clever within-athlete design: bilateral pedaling, but only one thigh is cuffed. After 6 weeks, the BFR leg showed greater performance gains and better potassium (K⁺) regulation than the uncuffed leg—adaptations that support repeated high-intensity efforts. PubMed+1

2) Low-intensity BFR can speed V̇O₂ on-kinetics (like HIIT)
In young adults, low-intensity interval cycling + BFR improved how fast oxygen uptake adjusts when work starts (phase-II V̇O₂ kinetics)—not different from classic high-intensity interval training, despite much lower mechanical load. PubMed+1

3) Continuous moderate cycling + intermittent BFR (8 weeks)
A Scientific Reports trial added 5-min ON / 5-min OFF BFR to 30 min at ~45–60% V̇O₂-reserve, 3×/week for 8 weeks, and found faster V̇O₂ kinetics adaptation than training without BFR. Nature+1

4) Systematic views are mixed
A meta-analysis in trained athletes reported no added aerobic benefit of BFR vs. matched training, reminding us that dose (pressure, time under restriction) and intensity are pivotal. PubMed

This is key - if you are physically able, this type of training may offer no benefit beyond conventional. 

Use caution/avoid: uncontrolled hypertension, thrombotic history, established vascular disease, pregnancy, or unexplained numbness/paresthesia on inflation. 

Bottom line

BFR isn’t magical—and it won’t replace your intervals or long days—but it can make easy and moderate sessions “count” more aerobically by speeding V̇O₂ kinetics, nudging vascular/microvascular adaptations, and improving ion handling, all at lower mechanical cost. Dose the pressure (60–80% LOP) and time under restriction thoughtfully, pair it with conservative running volumes, and you’ve got a smart tool for aerobic gains when load must be managed. PubMed+2

A word about protocols:

Ok, so you want to try it - where do you start? There are numerous studies (most short term) that use different pressures, cuffs, protocols with various results that, given the above heterogeneity of methodology, can't be compared. Having perused the literature, my first thought was to slowly ramp up through LT1 and examine muscle O₂ kinetics (looking for hypoxia) and DFA a1/HR. That is what we will look at in this post today.

Cuff choice and the "AOP"

Arterial Occlusion Pressure (AOP)—also called Limb Occlusion Pressure (LOP)—is the minimum external cuff pressure required to stop arterial inflow into a limb at rest. It’s a property of the person + limb + cuff + position at a given moment. In BFR training we usually measure AOP first, then set training pressure as a percentage of AOP (e.g., ~40–60% for arms, ~60–80% for legs), so we restrict venous return while still allowing some arterial inflow.

What determines AOP?

AOP rises or falls with factors that change how effectively cuff pressure is transmitted to the deep arteries:

• Cuff width & material: Wider cuffs transmit pressure deeper and need lower surface pressure to occlude; narrow cuffs need higher pressure. (This is what the graphics illustrated—wider cuffs have slower pressure decay with depth.)

• Limb size & tissue composition: Larger circumference or thicker, more compliant soft tissue → higher AOP.

• Blood pressure & autonomic state: Higher systolic/mean arterial pressure → higher AOP; relaxation, exhale, or lower BP → lower AOP.

• Body position & limb elevation (see below): Sitting vs. supine and limb above/below heart level change hydrostatic pressure; dependent limbs typically show higher AOP.

• Cuff placement & fit: Proximal placement, full contact, and no slack reduce variability.

• Device specifics: Elastic vs. nylon, bladder geometry, single vs. dual bladder, etc.

How is AOP measured?

Common approaches (ranked by fidelity):

1. Doppler ultrasound (gold standard): Place Doppler on distal artery (e.g., posterior tibial/radial); inflate until the arterial pulse disappears → that cuff pressure is AOP.

2. Photoplethysmography (PPG) / pulse oximeter distal to cuff: Inflate until the pulsatile waveform (AC component) vanishes.

3. Automated tourniquet systems: Built-in sensors/algorithms estimate AOP from oscillations/PPG and limb size.

4. Estimations from models: Use limb circumference, systolic BP, cuff width/material. Useful when direct sensing isn’t available, but verify if possible. 

I decided to use the Saga leg cuff device since they use Bluetooth control of pressure, can self calibrate and are relatively "wide." (No, those are not my legs.)

 

Wider cuffs will be able to penetrate compressive force (for occlusion) better than narrow elastic bands. This is a critical point since we don't want to damage more superficial tissue in an effort to occlude the deep vessels. The following graph shows the differential in pressure needed for total arterial occlusion between narrow (Kaatsu) vs. wide (Delfi) cuffs:

 

• As mentioned above, the pressure needed by the narrow cuff is ridiculously high, could be damaging to local tissue and/or induce nerve compression injury.

Problems finding the AOP:

Furthermore, determining the arterial occlusion pressure (AOP) is position determinant. Here is an excerpt from the paper that studies this:

Background Total arterial occlusive pressure (AOP) is used to prescribe pressures for surgery, blood flow restriction exercise (BFRE) and ischemic preconditioning (IPC). AOP is often measured in a supine position; however, the influence of body position on AOP measurement is unknown and may influence level of occlusion in different positions during BFR and IPC. The aim of this study was therefore to investigate the influence of body position on AOP. Methods Fifty healthy individuals (age = 29 ± 6 y) underwent AOP measurements on the dominant lower-limb in supine, seated and standing positions in a randomised order. AOP was measured automatically using the Delfi Personalised Tourniquet System device, with each measurement separated by 5 min of rest. Results Arterial occlusive pressure was significantly lower in the supine position compared to the seated position (187.00 ± 32.5 vs 204.00 ± 28.5 mmHg, p < 0.001) and standing position (187.00 ± 32.5 vs 241.50 ± 49.3 mmHg, p < 0.001). AOP was significantly higher in the standing position compared to the seated position (241.50 ± 49.3 vs 204.00 ± 28.5 mmHg, p < 0.001). Discussion Arterial occlusive pressure measurement is body position dependent, thus for accurate prescription of occlusion pressure during surgery, BFR and IPC, AOP should be measured in the position intended for subsequent application of occlusion. 

Since we have to start somewhere, I decided to just give it a go with my first session below:

My first session:

According to the Saga cuff test, my AOP was about 220 mmHg when standing, but only 170 supine. Being the enthusiast, I used the 220 AOP pressure on my initial day of use with a 60% AOP net, (set by the cuffs). Not surprisingly, leg pain after cuff inflation was severe, and I ended up messing around with 50% AOP (which still hurt) for a few brief intervals. Several days later I tried again but used an AOP of 170, which was both the pressure measured by the Saga cuffs in the supine position and that calculated by the equation used here.

 

My second (better) session:

For this day I used my old BSX Insight NIRS and a single lead (3 wire) ECG recording at 1000 Hz (Faros 180) for HR/DFA a1. As mentioned, pressure was set to 60% AOP (of 170 mmHg). The BSX was on the rectus femoris muscle throughout.

 

After an 8 minute warmup, a 12-minute incremental ramp with a slope of 5 watts increase per minute was planned (130 to 190 watts). Since my LT1 is about 175 to 180 watts, that usually means that my DFA a1 will be 0.6 to 0.8 at termination (depending on ANS recovery/stress/artifacts). Unfortunately, by midramp the leg pain was so severe I had to deflate (but continued the ramp). Here are some plots of power, HR, SmO2 with DFA a1 on a separate plot below:

Plot With NIRS:

 

Plot With DFA a1: 

 

• The ramp is relatively smooth, with the midpoint power having a temporary dip from the cuff deflation.
• HR sharply rises with cuff inflation and reaches almost LT2 HR by the time I deflated the cuff mid ramp (well below Lt1).
• O2 sat starts at near 40% and ends up at 10% at cuff deflation. The initial drop was abrupt, followed by a short plateau and then a low nadir.
• DFA a1 decline is rapid and drops below 0.5 at an intensity well below LT1 and (for me) an easy pace, sustainable for many hours.
• After cuff deflation, both O₂ sat and DFA a1 recover quickly along with a drop in HR despite the ramp intensity increase. The final DFA a1 at ramp termination is in line with what it usually has been.

My take on this was either my cuff pressure was again too high (even though AOP appeared valid) or I need better acclimation to BFR. Given that I used an AOP taken supine (technically, I should have been standing or on the bike), I was surprised my tolerance was so poor. On the other hand, the hypoxia seen was impressive. 

How does the hypoxia compare to the literature? Here is some data from a study looking at multiple intervals (with continued inflation) using either 60 or 80% AOP:

 

Cycling power/intensity during cuff use was 40% VO2 max, which is a no-no in my book (being an honorary disciple of Juan Murias). Was that zone 1 or 2? It probably varied per person and is a reminder that "domains" should be based on true boundaries. However, here is the wattage from their results: "Mean power outputs were 89 ± 18 W for the LL, BFR60, and BFR80 cycling conditions and 240 ± 36 W for the HL cycling condition." That looks pretty low to me, but as most folks realize these days, we need to be more exact in assessing intensity domains/zones.

What they found was the follow:

 

• Both 60 and 80% AOP led to prominent desaturation, with 80% more profound over the course of the multiple intervals. They did remark that the "discomfort" was much higher with the 80% group:

 

 

Having read that paper, I did perform a few 2-minute "intervals" at a power near LT1. The following was about 90 minutes into my session (that started with the ramp):

O₂ sat:

 

 DFA a1:

 

• During the 2 minute cuff inflation, O2 sat dropped over 50%.
• Pain level was high at termination.
• HR rose to levels near LT2 but at LT1 power (for only 2 minutes!).
• DFA a1 rapidly dropped into the 0.25 range, and stayed there even after cuff deflation for several minutes. 

 

Although many studies have been done with various training protocols, it appears that there are "many roads to Rome" regarding aerobic BFR. In other words, not getting the AOP or occlusion pressure "just right" with the perfect intensity may not be required. The following study partially illustrates why this may be true.

Training protocol (4 weeks, 3×/week = 12 sessions)

• Session format: Repeated 2-min treadmill running bouts with BFR, each separated by 1-min recovery without BFR. Cuffs were inflated during work and deflated in the breaks. Two 5 × 120 cm thigh cuffs placed proximally on both legs. Pressures were set relative to estimated arterial occlusion from thigh circumference (≈partial ≈50% AOP vs complete occlusion).

• Sets per session: Most groups performed 10 bouts each session; the IP-IE group reduced across weeks (10→8→6→5) to accommodate the higher difficulty at high intensity with near-complete occlusion. Frontiers

• Four randomized groups (pressure × intensity progression over 4 weeks):

  • IP-CE (Increasing Pressure–Constant Intensity): Pressure: 160→190→210→240 mmHg; Intensity: constant 60% vV̇O₂max

  • CP_P-IE (Constant Partial–Increasing Intensity): Pressure: constant 160 mmHg (partial occlusion); Intensity: 60→70→80→85% vV̇O₂max. 

  • CP_C-IE (Constant Complete–Increasing Intensity): Pressure: constant 240 mmHg (complete occlusion); Intensity: 60→70→80→85% vV̇O₂max.

  • IP-IE (Increasing Pressure–Increasing Intensity): Pressure: 160→190→210→240 mmHg; Intensity: 60→70→80→85% vV̇O₂max (with the reduced number of bouts noted above).

Outcomes (pre–post, between-group highlights)

• All groups improved in V̇O₂max, vV̇O₂max, time-to-fatigue (TTF), running economy (RE), leg strength, and anaerobic power (p ≤ 0.05).

• V̇O₂max (% change): IP-CE +9.6 ± 2.0%, CP_P-IE +11.2 ± 5.5%, CP_C-IE +14.8 ± 4.9%, IP-IE +8.4 ± 2.4%; CP_C-IE was significantly greater than IP-IE (Bonferroni p = 0.01). 

• Running economy (RE): Best post-training RE (lower O₂ cost at 10 km·h⁻¹) in CP_C-IE; other groups were significantly worse vs CP_C-IE (ANCOVA p = 0.024; “#” denotes different from CP_C-IE in Table 3).

• Time-to-fatigue (TTF at vV̇O₂max): CP_C-IE achieved the largest TTF (post 310.5 ± 39.6 s) with a significant overall group effect (ANCOVA p = 0.042).

• Leg strength: Greatest gain in CP_C-IE (post 131.8 ± 7.4 kg); significant group effect (ANCOVA p = 0.001).

• Anaerobic power (peak/avg/min W·kg⁻¹): All groups improved; several variables showed significant group effects favoring higher pressure with rising intensity. 

 

• One should note that AOP was estimated by equation, and total occlusion was NOT confirmed by Doppler (from the article, “instead of the gold standard method, such as Doppler, we estimated arterial occlusion from the thigh circumference”). 
• Having done this myself, I highly doubt that these folks were running intervals with total occlusion at 85% VO2 max!
• Having said that, it's clear that all groups did well, even the constant pressure and constant 60% VO2 max intensity group. It would have been nice to see muscle O₂ desaturations in the various groups for comparison.

 

Observations and thoughts:

• BFR training in a cycling model is achievable with consumer equipment. The Saga cuffs are a reasonable choice, although I have nothing else to compare them to.
• While the concept of the AOP is important, the pressure needed for arterial occlusion will vary according to body position, thigh diameter, cuff width and motion (among other things). Therefore, trying to mimic a given study that used "xx%" of AOP is arguably futile. Remember that the "gold" standard for proving AOP is to see that no blood flow occurs via Doppler.
• If we have issues with the % AOP parameter, what other avenues do we have to adjust BFR "training levels"? The AOP % can be a rough guideline, but as seen above, it appears we have multiple other clues that tell us that an effective occlusion is occurring. Most studies I have seen show that HR rises (with aerobic BFR), above normal levels per intensity (watts, speed). As noted above, DFA a1 plummeted rapidly during BFR intervals, indicating marked ANS stress, metabolic perturbation, and sympathetic/parasympathetic alteration. This, coupled with SmO₂ decline, appears to herald a notable physiological response to BFR pressure applied. In short, O₂ desat, DFA a1 decline, and HR elevation above normal are all markers of a potentially successful BFR session and could be potential clues for users to strive for.
• In addition to the "objective" responses above (O₂ desat, DFA a1 decline, HR rise), we do have another important metric to follow, that of the pain and discomfort caused by the BFR. I have not delved much into this, but believe me, discomfort is too mild a description at certain times. On the other hand, if you are using cuffs and there is not much in the way of "pain," something may not be right. 
• There is probably a wide range comprising enough compression, appropriate intensity, and interval duration that will result in beneficial results. In other words, it may be that one can get BFR results by various combinations of the three components.
• Impacts on Training Load.. Many BFR users will be using cycling/running intensities in the moderate domain (zone 1). However, it is essential to note that successful BFR (by pain, O₂ desat, tachycardia, DFA a1 suppression) will most likely be associated with a higher than usual "internal" training load, as opposed to the traditional, wattage based external load framework. I plan to elaborate further on this in the next post, but here is a brief tidbit. The day after a BFR session (comprising all zone 1, below LT1 intensity, including 6 sets of 2 minute BFR at 60% AOP), my DFA a1 was markedly suppressed while riding 90 minutes in the zone 1 range (15 watts below LT1):

 

As noted, only 12% of the session was spent at an a1 above 0.75. The usual profile for this type of day would have been much different, with an a1 near 1 as noted below (same power, time, different week):

 

How much additional physiological/ANS stress BFR adds to the big picture of training load is not clear, but should be considered given potential overreaching. 

Blog index