Does Blood Flow Restriction Training Work? What a New Trial Found
Quick answer: A randomised trial by Hou, Zhang and colleagues, published 8 June 2026 in Frontiers in Physiology, put 60 trained men through 8 weeks of lifting at 30%, 50% or 70% of their max with cuffs on, or 70% with no cuff. The lightest cuff protocol lost on nearly every measure. Cuffs added nothing to heavy lifting.
Blood flow restriction training is the one gym idea that sounds like it should not be legal. Strap a cuff around the top of your arm, curl a weight you could lift fifty times, and grow.
The pitch is that you get heavy-lifting results from light weights. This trial tested that pitch directly, and the answer is more awkward than either camp wants.
What the researchers actually did
Sixty trained men were randomly split into four groups of fifteen and put on the same programme for eight weeks, three mornings a week.
Everyone did the same lifts: bent-over row, bench press, hang power clean, back squat, deadlift, hip thrust, weighted pull-up, weighted dip, incline dumbbell press. Everyone did 4 sets of 8 on every exercise. Every session was supervised.
The only things that changed between groups were the load and the cuffs.
- HL-BFR: 70% of 1RM with pneumatic cuffs on both limbs
- ML-BFR: 50% of 1RM with cuffs
- LL-BFR: 30% of 1RM with cuffs
- HL-CON: 70% of 1RM, no cuffs
Loads were recalculated every one to two weeks so the percentages stayed honest as people got stronger. Before and after, the team tested one-rep maxes on four lifts, a countermovement jump, pull-ups and dips for reps, a sprint, a jump, a throw, an agility run, body composition, and resting hormones.
Note what is missing: there is no 30% or 50% group without cuffs. That single design choice shapes everything you can conclude, and we come back to it below.
What they found
Start with the numbers people actually care about. Here is how much weight each group added to their one-rep max over the eight weeks.
Weight added to the one-rep max after 8 weeks
- 70% + cuffs
- 50% + cuffs
- 30% + cuffs
- 70%, no cuffs
View the data
| Group | Bench press | Squat | Deadlift | Biceps curl |
|---|---|---|---|---|
| 70% + cuffs | 4.8 kg | 4.87 kg | 5.13 kg | 3.37 kg |
| 50% + cuffs | 5.8 kg | 4.87 kg | 4.8 kg | 2.83 kg |
| 30% + cuffs | 2 kg | 2.87 kg | 2.87 kg | 0.97 kg |
| 70%, no cuffs | 5.87 kg | 4.93 kg | 4.93 kg | 3.2 kg |
The 30% group added 2 kg to their bench press. Everyone else added between 4.8 and 5.9 kg. On the biceps curl, the lift where light-load BFR is supposed to shine most, the 30% group added under a kilogram while the others added roughly three.
The gaps between 30% and the rest were big enough on curl and bench press that the researchers could rule out luck. On the squat, only the no-cuff heavy group separated from it cleanly.
The rep tests told the same story.
Extra reps gained on bodyweight tests after 8 weeks
- 70% + cuffs
- 50% + cuffs
- 30% + cuffs
- 70%, no cuffs
View the data
| Group | Pull-ups | Parallel bar dips |
|---|---|---|
| 70% + cuffs | 3.6 reps | 4 reps |
| 50% + cuffs | 3.2 reps | 3.8 reps |
| 30% + cuffs | 1.67 reps | 2 reps |
| 70%, no cuffs | 3.2 reps | 3.73 reps |
Two other results deserve a mention. The jump test improved by roughly the same amount in every group, with no meaningful separation. And resting testosterone, cortisol and growth hormone were flat everywhere after eight weeks, which quietly kills the old story that BFR works by flooding you with hormones.
The headline the researchers landed on: 50% with cuffs kept up with 70% without them, and 70% with cuffs was no better than 70% alone.
Why did the light-load group fall behind?
The obvious reading is that BFR does not work. The researchers argue something more specific, and it is worth understanding.
Every set in this study was fixed at 4 sets of 8. That is the design's biggest tell. Light-load BFR normally works by taking the muscle close to genuine failure, where the cuff has starved it of oxygen and forced the big motor units to take over. Eight reps at 30% of your max is nowhere near that point. It is a warm-up with a cuff on.
So the 30% group probably did not fail because the loads were light. They failed because the sets were easy. The authors say this directly: they think comparable results from low-load BFR would need either more volume or sets taken much closer to volitional fatigue.
That also explains why cuffs added nothing at 70%. At that load you already have plenty of mechanical tension and near-complete motor unit recruitment. There is no gap left for metabolic stress to fill. The same logic shows up in our piece on training to failure, where effort matters most when the load is doing the least.
This is the researchers' proposed explanation rather than something the trial measured. They did not manipulate proximity to failure, so it stays a hypothesis.
What this study doesn't prove
There was no light-load group without cuffs. The only no-cuff condition was 70%. So this study cannot tell you whether the 30% group did badly because of the load or despite the cuff. It compares four packages, not the isolated effect of BFR, and the authors say so in their own conclusions.
Nobody trained near failure. A fixed 4 sets of 8 at 30% of 1RM is a lenient version of low-load BFR. The protocols that produce good results in the wider literature take the first set to failure or close. This trial did not test that.
These were trained men in a military setting. Sixty men, average age 24, all with a structured lifting background, all training under supervision at the same time of day. No women, no untrained beginners, no one training around a job and a bad night's sleep.
Eight weeks is short and fifteen per group is small. With that sample, real differences smaller than the ones reported here would simply not show up. The authors flag both.
Body composition came from a bioimpedance scale. The small lean mass changes should be read as noise until something better measures them. The authors add that caveat themselves.
Two separate outcomes are reported with identical numbers. The standing long jump and the medicine ball throw appear in the paper with the same change values, the same confidence intervals and the same effect sizes for all four groups. That looks like a transcription error, which is why neither is charted here.
The study was funded by an internal research grant at the Army Engineering University of the PLA. The authors declare no commercial conflicts.
How to test this in your own training
Group averages do not tell you what happens to your arms. Eight weeks of your own logged numbers do.
- Pick one exercise to test, not a whole programme. A single-joint lift with cuffs is where BFR makes most sense: biceps curl, leg extension, lateral raise.
- Use a load you could do about 20 to 30 times, then actually get close to failure. This trial's lesson is that the effort, not the cuff, is what was missing. Stop 1 to 2 reps short on the first set, not 12.
- Keep every other variable still. Same exercise order, same rest, same weekly frequency. If you change two things you learn nothing.
- Record the reps and the load every session, plus how hard the last set felt. A BFR set that felt like a 6 out of 10 is the set that explains a flat eight weeks.
- Retest a real number at week 8. Estimate your max from your best set with the one rep max calculator and compare it to where you started.
The four conditions side by side
| Condition | Load | Cuffs | 8-week bench press gain |
|---|---|---|---|
| HL-BFR | 70% of 1RM | Yes | 4.8 kg |
| ML-BFR | 50% of 1RM | Yes | 5.8 kg |
| LL-BFR | 30% of 1RM | Yes | 2.0 kg |
| HL-CON | 70% of 1RM | No | 5.9 kg |
Read the last two rows together. The cheapest condition in the table, plain heavy lifting with no equipment at all, produced the biggest bench press gain.
Logging a training experiment you can actually read later
Testing something like this only works if the log is honest about the details. Most gym apps have a field for weight and reps and nowhere to put "cuff at 7, first set close to failure, second set died".
Gym Note Plus takes that whole line as plain text, the way you would type it into your notes app, and still turns it into a chart.
- Write the set and the context in one line, no dropdowns to fight
- Per-exercise history, so week 1 and week 8 sit next to each other
- Estimated one-rep max tracked over time from the sets you logged
- Works offline, which matters when the gym has no signal
Paste an old session into the free notes to workout translator to see what your existing notes turn into.
A cuff is not a substitute for effort. This trial's light-load group had the cuffs and skipped the hard part, and their numbers say so.
Frequently Asked Questions
What did the 2026 blood flow restriction study find?
The trial by Hou and colleagues, published 8 June 2026 in Frontiers in Physiology, ran 60 trained men through 8 weeks of lifting at 30%, 50% or 70% of 1RM with cuffs, or 70% without. The 30% group gained about half as much strength as the others. Cuffs added nothing at 70%.
Does blood flow restriction training build muscle?
Blood flow restriction training can build muscle at light loads, but this 2026 trial suggests it only does so when sets are taken close to failure. Its 30% group performed a fixed 4 sets of 8, well short of fatigue, and gained roughly half the strength of the heavier groups.
Is BFR training better than lifting heavy?
No. In this trial, adding cuffs to 70% of 1RM produced no advantage over 70% without cuffs on any strength or performance test. BFR's real use case is when heavy loading is off the table, such as during rehabilitation or when a joint cannot tolerate the weight.
How heavy should you lift with blood flow restriction?
This trial found 50% of 1RM with cuffs matched plain heavy lifting on bench press, squat and deadlift strength, while 30% fell well behind. If you are running fixed sets rather than sets to failure, a moderate load looks like the safer choice.
Does BFR training raise growth hormone?
Resting growth hormone, testosterone and cortisol were unchanged in all four groups after eight weeks in this trial. Acute post-exercise spikes were not measured, but the flat resting values argue against systemic hormones being the mechanism behind blood flow restriction training.
Can beginners use blood flow restriction training?
This study used trained men in a supervised military setting, so it says nothing about beginners. A new lifter gains well from ordinary progressive loading and does not need cuffs. Our guide to progressive overload covers what to do first.
Final Takeaways
- The cuff is not the active ingredient. Effort is. The group with the lightest load and the easiest sets gained the least.
- 50% with cuffs kept up with 70% without. That is a real, useful result if heavy loading is a problem for you.
- 70% with cuffs beat nothing. Adding restriction to heavy lifting was pointless in this trial.
- Hormones did not move. The old growth hormone story does not survive eight weeks of resting bloodwork.
- This is one 8-week trial in 60 trained men. Treat it as a reason to run your own test, not as a verdict on BFR.
Josh Ibbotson
Josh is the creator of Gym Note Plus, building tools that make workout tracking as simple as taking notes.
Ready to try it out?
Download Gym Note Plus and start tracking your workouts the way you want.
Download on App Store
