Guides
BFR Knee OA Clinician Guide
The Ultimate Guide · 2026 Evidence Edition · Clinician's Guide
Blood Flow Restriction
for Knee Osteoarthritis
The evidence, the protocol, patient selection, and how to talk to surgeons. Built from two randomized trials, Jacobs and Sørensen, and the complete 2020–2026 literature.
at one year
injections
per week
Owens Recovery Science · Personalized BFR education and the Delfi Personalized Tourniquet System · owensrecoveryscience.com
Share this with your patients
The companion patient handout covers the same evidence in plain language, with a printable PDF they can take home or bring to a referring surgeon.
Executive summary
The paradox, and the trials that solved it
The single strongest lever in knee osteoarthritis is not a drug, an injection or a muscle. It is how much the patient moves. In the Osteoarthritis Initiative, about 55 minutes a week of moderate activity cut the onset of mobility disability from 24% to 3% (RR 0.14; Dunlop 2019). Every major guideline puts exercise first for exactly that reason. The problem is the paradox: the quadriceps needs load to get strong enough to carry that activity, and the load that builds a quadriceps, roughly 70% of one-repetition maximum, is precisely what an arthritic knee will not tolerate.
Blood flow restriction (BFR) resolves the paradox. A personalized cuff creates the metabolic stimulus of heavy training at 20–30% of 1RM, loads a knee OA patient can perform. For a decade the BFR-for-knee-OA literature was small and underdosed, and the 2021 meta-analyses fairly concluded "no difference," measuring, as the field then did, quadriceps size and strength in isolation. Two randomized trials published in 2025 changed the question. A 120-patient trial from Ghent University (Annals of the Rheumatic Diseases; British Journal of Sports Medicine) showed a large quadriceps-strength effect and better pain at 12 weeks, and a year later the BFR group still had a clinically important pain advantage, were doing 1.5 more hours of self-directed activity each week, and needed 63% fewer intra-articular injections. A 96-patient trial from Copenhagen (Sørensen, Scand J Med Sci Sports) ran BFR against the Danish standard of care and found large functional advantages on every performance test and reduced pain sensitization. Read these trials not as muscle-building protocols but as the first proven clinical gateway back into the activity volume that changes the trajectory of the disease.
What this guide gives you
- The knee OA outcome hierarchy: what actually moves pain, function and disability, and where BFR fits
- The Jacobs trial and its one-year follow-up, in the detail needed to reproduce it
- The Sørensen trial and the complete 2020–2026 evidence arc, including the skeptics
- Patient selection, screening and contraindications
- A protocol card you can run Monday morning
- How to talk to surgeons, and what to tell patients to expect
- What the cost conversation can and cannot claim
- Two tools at the back: a one-page surgeon referral sheet you can put your logo on (Appendix A) and a medical-necessity documentation template (Appendix B)
The one-sentence version
If a 12-week program still shows a 15-point pain advantage and 63% fewer injections nine months after it ended, what is the argument for not offering it to knee OA patients who cannot tolerate heavy load?
Section 1
What actually improves knee OA outcomes
It is tempting to frame knee OA rehabilitation as "strengthen the quad, control the pain." The literature says the levers are broader, and that framing is better for BFR, not worse. Ranked by verified magnitude of effect on outcomes that matter:
| # | Lever | Verified magnitude | Where BFR fits |
|---|---|---|---|
| 1 | Physical activity volume | Adults with knee OA who managed about 55 minutes a week of moderate-to-vigorous activity, roughly eight minutes a day, developed mobility disability at 3% over four years, against 24% in those who did not. That is a relative risk of 0.14 (Dunlop 2019, Osteoarthritis Initiative, n=1,564). No treatment in the field comes close to that on a hard endpoint. | THE ON-RAMP Jacobs is the only knee OA exercise trial with a measured activity dividend a year later: 5.2 vs 3.7 hrs/week, unsupervised. |
| 2 | Weight, in the obese | The benefit scales with how much body weight comes off. Over 18 months, patients who lost 20% or more of their body weight had 25% less pain and better function than those who lost 10 to 20% (Messier 2018). Diet-based weight loss cut knee compressive force by about 200 N per step and lowered IL-6; exercise alone did neither (Messier 2013). | ADJUNCT, AND NO LONGER "NOTHING" BFR is not a weight-loss intervention. But in type 2 diabetes, 12 weeks of BFR training preferentially reduced visceral adipose tissue volume and waist circumference versus load-matched conventional resistance training (Trinks 2026, Cell Metab). Not yet shown in knee OA. Pair with weight management per guideline. |
| 3 | Expectations & self-efficacy | Unmet expectations are the top predictor of TKA dissatisfaction (10.7×; Bourne 2010). Catastrophizing and poor sleep predict central sensitization and post-surgical opioid use (Campbell 2015; Rhon 2022). | PAIR, DON'T REPLACE The hypoalgesic response to BFR is itself modulated by catastrophizing and self-reported health (Ogrezeanu 2023). Pressure selection matters: higher occlusive pressures can provoke hyperalgesia in persistent-pain populations (Gray 2026, narrative review). |
| 4 | Exercise itself | Pain −6.4/100 at 12 weeks, decaying to −3.4 at one year (Holden 2023, IPD meta, 31 RCTs). Quad-specific beats general lower-limb work (SMD 0.85 vs 0.39; Juhl 2014). Statistically equivalent to open-label saline injections for symptoms (Bandak 2022). | THE TOLERABLE VERSION Adverse events RR 0.26–0.45 vs high load; Ferraz: 4 of 16 high-load patients dropped out for knee pain, none with BFR. |
| 5 | Quadriceps strength | Weakness predicts who goes on to develop symptomatic OA (OR 1.85 women, 1.43 men; n=46,819; Øiestad 2022). But strength gain explains only part of why pain and function improve (Hall 2018); patients get better for more reasons than the quad alone. And its apparent protection against future knee replacement disappears once you account for how bad the radiographs already were (Skou 2016). Treat strength as a necessary input, not the outcome you are selling. | CLEAR SUPERIORITY vs load-matched low-load training: strength SMD 0.75, hypertrophy 0.81; vs high load: equivalent hypertrophy (Machado 2026; Centner 2019). |
| 6 | Analgesia | An enabler, not a treatment. Corticosteroid injection: SMD −0.48 at 2 weeks, nil by 6 months (Jüni 2015). Repeated triamcinolone: more cartilage loss than saline over 2 years, no pain benefit (McAlindon 2017). | HYPOALGESIA BFR raises pressure-pain threshold about 48% locally and the effect persists 24 hours (Hughes 2020). |
The reframe
If activity volume is the biggest lever and exercise is the on-ramp, the clinical question for a painful knee is: what version of exercise can this patient actually do enough of? BFR's value is not "a stronger quad." It is the tolerable on-ramp back into the activity volume that changes the trajectory, and the injection reduction that follows.
Sources: PubMed; identifiers in References.
Section 2
The Jacobs trial: 12-week results
Jacobs E, Stroobant L, Victor J, et al. Vascular occlusion for optimising the functional improvement in patients with knee osteoarthritis: a randomised controlled trial. Ann Rheum Dis 2025;84(2):341–350. NCT04996680. Ghent University Hospital, Belgium.
Design
- 120 patients with symptomatic knee OA, Kellgren–Lawrence 1–4, aged 30–80, referred by orthopaedic surgeons and not currently surgical candidates. 87 women, 33 men; mean age 58; mean BMI 25; baseline pain 3.3/10; 47% KL grade 3.
- Excluded: BMI >30, cardiovascular/metabolic/neuromuscular disease, VTE history, knee injection in the prior year, pregnancy. The injection exclusion was a trial design choice to keep the cohort clean. It is not a clinical contraindication, and a recent injection does not disqualify a patient from BFR in practice.
- 1:1 randomization, stratified by sex and side; assessor-blinded; intention-to-treat. Pre-registered, and reported to CONSORT and TIDieR standards, which is rarer in exercise trials than it should be and is part of why surgeons take this paper seriously.
- Both arms: 24 supervised sessions over 12 weeks (2×/week), identical programs delivered by BFR-trained physiotherapists. The only variable was a cuff on the two quadriceps exercises.
- Primary outcome: KOOS at 3-month follow-up. Secondary: quadriceps and hamstring strength (dynamometry), 6-min walk, 40 m fast walk, 30-s chair stand, 30-s knee bend, stair climb, Pain Catastrophizing Scale.
Results
| Outcome (BFR vs exercise alone) | 12 weeks | 3 months post |
|---|---|---|
| Quadriceps strength | ES 0.81, p<0.0001 | ≈1 N/kg advantage held |
| KOOS pain | +8.8 pts, ES 0.58 | +9.4 pts, ES 0.55 |
| KOOS symptoms | ns | +9.0 pts, ES 0.59 |
| KOOS quality of life | ns | +13.2 pts, ES 0.66 |
| 6-minute walk | +57 m, ES 0.75 (significant at 6, 12 and 24 wk) | |
| 30-s chair stand | +4.7 reps, ES 0.78 | |
| Hamstring strength | No difference at any time point. Expected: the hamstrings were trained identically in both arms and were never cuffed. | |
| Dropout | Comparable: 6.7% BFR vs 10% control withdrew for pain | |
How strict the statistics were
Most trials call a result significant at p<0.05. Because this trial tested many outcomes at once, the authors applied a Dunn–Šidák correction and set the bar far higher: p<0.0019 for the primary outcome and p<0.0034 for secondary outcomes. That is roughly 25 times stricter than the conventional threshold. Several results that would count as "significant" in most papers, including KOOS activities of daily living (p=0.014) and stair climb (p=0.024), were reported as not meeting the threshold. Every positive finding in this guide cleared that higher bar, which is what makes them worth building a protocol on.
Reading the Kellgren–Lawrence grades
The trial enrolled KL 1–4 and 47% were grade 3, a moderately advanced cohort, not early disease. For reference:
| Grade | Radiographic definition | Practical read |
|---|---|---|
| KL 0 | No radiographic features of OA | Normal film |
| KL 1 | Doubtful joint space narrowing; possible osteophytic lipping | Doubtful |
| KL 2 | Definite osteophytes; possible joint space narrowing | Minimal; the usual threshold for "definite OA" |
| KL 3 | Multiple osteophytes, definite joint space narrowing, some sclerosis, possible bone-end deformity | Moderate; the largest single group in Jacobs |
| KL 4 | Large osteophytes, marked joint space narrowing, severe sclerosis, definite bone-end deformity | Severe, bone-on-bone; often referred for arthroplasty |
The end-stage patients stayed
Seventeen patients entered the trial at KL grade 4, bone-on-bone. Six patients across both arms left the study by mutual decision with their surgeon to proceed to knee surgery, and five of those six were KL 4. That leaves twelve end-stage knees that completed twelve weeks of training and the three-month follow-up. For a population many clinicians assume is only waiting for a replacement, that is a retention story worth telling a surgeon.
Nothing separated at 6 weeks
No KOOS subscale differed between groups at week 6. The advantage appeared at 12 weeks and grew after supervision ended. Set patient expectations accordingly: plan for 24 sessions.
The mechanism showed up in the data
Only the two quadriceps exercises were cuffed. Quadriceps strength separated at every time point. Hamstring strength never did, and that is the point: both groups trained the hamstrings the same way with no cuff, so there was no reason for them to differ. The cuffed muscle changed, the uncuffed muscle did not.
One serious adverse event
One DVT in the BFR arm after session 12 (no redness or swelling; confirmed by ultrasound and D-dimer). Full recovery. At intake the patient had not disclosed contraceptive use or a prior calf episode after ankle immobilisation. Screening is not optional. See the intake checklist and the session-day check in Sections 5 and 6.
Source: Jacobs et al., Ann Rheum Dis 2025 (doi 10.1136/ard-2024-226579), full text.
Section 3
One year later: durability and injections
Jacobs E, Stroobant L, Witvrouw E, et al. Sustained benefits of blood flow restriction therapy in knee osteoarthritis rehabilitation: 1-year follow-up of a randomised controlled trial. Br J Sports Med 2025;59(21):1481–1489.
Eighty of the 120 patients returned at one year, nine months after the last supervised session, with no restrictions on care after the 3-month follow-up. Patients could seek injections, surgery, or anything else. That is what makes the utilisation finding meaningful.
They kept getting better
The usual pattern after a course of exercise therapy is a peak at discharge and a slow slide back. The BFR arm did the opposite. KOOS pain separation was 8.8 points at 12 weeks, 9.4 points at 3 months, and 15.1 points at one year. Quality of life went from no difference at 12 weeks to a 13.2-point advantage at 3 months and 14.7 points at one year. Symptoms followed the same curve. The gap widened at every measurement, long after the last supervised session.
| Outcome at 1 year | BFR vs exercise | ES |
|---|---|---|
| KOOS pain | +15.1 points (p=0.0039) | 0.79 |
| KOOS quality of life | +14.7 (p=0.0032) | 0.61 |
| KOOS activities of daily living | +11.3 (p=0.0054) | 0.54 |
| KOOS symptoms | +10.5 (p=0.0074) | 0.51 |
| KOOS sport & recreation | +13.4 (ns at corrected threshold) | n/a |
| Quadriceps strength | +0.84 N/kg (p=0.0010) | 0.48 |
| 30-s chair stand | +4.9 reps (p=0.0005) | 0.75 |
| 40 m walk / stairs / 6MWT | −2.5 s / −0.8 s / +51 m | 0.4–0.5 |
| Self-initiated activity | 5.2 vs 3.7 hrs/week (+1.5 hrs, p=0.036) | n/a |
Intra-articular knee injections in the year after rehab
13 of 39 vs 5 of 40 patients. Odds ratio 3.6 (95% CI 1.14–11.35), p=0.024. Relative risk 0.366 (95% CI 0.14–0.93), a 63.4% reduction.
Why the injection number is the story
Pain scores persuade clinicians. Injection counts persuade surgeons, administrators and payers. Each injection is a recurring visit and a recurring cost, and repeated corticosteroid injection produced more cartilage loss than saline over two years with no pain benefit (McAlindon 2017, JAMA): two years of repeated triamcinolone cost 0.21 mm of cartilage thickness versus 0.10 mm with saline. Read the two findings together. Every injection avoided is a documented cartilage exposure avoided.
Frame it honestly. This is a secondary, non-preregistered outcome from 79 completers with a wide confidence interval. The lower bound (OR 1.14) means the true effect could be modest, but the direction is consistent with the mechanism, and the harm being avoided is documented. Say all of that to a surgeon. It respects their statistical literacy and keeps the argument intact.
Exercise alone fades
The control arm tracks the published long-term picture for traditional therapy: gains of roughly 3.5–5.2 points on a 0–100 scale, below the 10-point MCID. The BFR arm did not fade. It kept improving.
The virtuous cycle
Less pain, more strength, more confidence, and 1.5 more hours of unsupervised activity per week. The first BFR trial in any population to measure a downstream activity dividend, and the most likely explanation for why the gap kept widening.
What the follow-up tells you clinically
Discharge is not the finish line. The patients who did best were the ones who kept moving after the supervised block ended. Build the handoff to self-directed activity into week 12, not as an afterthought.
Source: Jacobs et al., Br J Sports Med 2025 (doi 10.1136/bjsports-2024-109524), full text.
Section 4
Sørensen and the wider literature, 2020–2026
Sørensen B, Magnusson SP, Aagaard P, et al. Effects of blood-flow restricted resistance exercise versus neuromuscular exercise on self-perceived knee pain, function, quality of life, and objective measures of functional performance and pain sensitization in adults with knee osteoarthritis: a randomized controlled trial. Scand J Med Sci Sports 2025;35(10):e70154. Copenhagen, Denmark.
The second trial
Ninety-six patients with symptomatic, radiographically confirmed knee OA (mean age 57; 49 women, 47 men) were randomized to 12 weeks of twice-weekly BFR resistance exercise or to the Danish standard of care, the GLA:D neuromuscular exercise program (NEMEX), extended to 12 weeks. Both groups received patient education. The BFR protocol was, in every practical respect, the Jacobs protocol: 24 supervised sessions, unilateral leg press and knee extension at 30% 1RM, 30-15-15-max, pressure at 60% AOP in week 1, 70% in week 2, then 80% (three patients stayed at 70%).
The primary outcome, KOOS pain at 12 weeks, favoured BFR by 6.4 points (95% CI −1.0 to 13.7, ES 0.35, p=0.07) and did not reach significance. Within groups, BFR improved pain by 15.4 points and NEMEX by 9.1. Two-thirds of BFR patients reached the clinically relevant change in pain, against 43% on standard care. Everything measured objectively separated. BFR was superior on every functional test: 30-second sit-to-stand, sit-to-stand power, 4×10 m fast walk and stair climb, with effect sizes from 0.89 to 1.56 (all p<0.0001). BFR also reduced spreading pain sensitization (ES 0.43–0.55), a marker of central pain processing that standard exercise did not move. No exercise-related adverse events in either group; nobody withdrew for pain.
The muscle underneath
A companion analysis in the same journal measured what changed in the quadriceps. At 12 weeks, BFR beat standard care on every measure of mechanical muscle function, all p<0.01: maximal isometric strength (ES 1.47), leg extensor power (ES 0.95), rate of force development at 50 and 200 ms (ES 1.15 and 1.07), and muscle cross-sectional area of rectus femoris (ES 2.09) and vastus lateralis (ES 3.33). Rate of force development, the ability to produce force quickly, improved only with BFR. Standard care did not move it at all.
Two things to say about it honestly. Baseline pain in this cohort was milder than in most knee OA trials (KOOS pain around 69 out of 100), which leaves less room for a pain score to move. And a 36-week KOOS follow-up found no between-group pain difference. What Sørensen adds is not durability; that is Jacobs' contribution. It adds a second, independent, harder-comparator trial in which the same protocol produced the same large, objective gains.
Why it matters alongside Jacobs
Same protocol, same population, same duration, personalized pressure. Different country, different research group, and a harder comparator: not exercise minus a cuff, but the established standard of care that Danish clinics already deliver and that several countries have adopted. That makes Sørensen the trial to cite when someone asks whether BFR beats what they are doing now. Read the two together and the picture is consistent. Both show large, objective functional gains that standard exercise does not produce. Jacobs shows those gains grow over a year and cut injection use. Sørensen shows they hold up against the best current comparator, reach into pain processing, and rest on measurable changes in the muscle. This is no longer a single-trial story.
How the evidence matured
- 2020–2021, the null years. Grantham et al. and Wang et al. each pool five small trials: no difference between BFR and regular resistance training. Grantham: "evidence does not support clinicians using BFR in knee OA." Wang notes fewer adverse events with BFR (RR 0.45). Read the trials underneath and the reason is plain. Most used a fixed cuff pressure rather than a measured LOP; one group ran the same absolute pressure first in women, then in men, and concluded BFR "works in women but not men" rather than that the men were underdosed. Most ran four to six weeks, when resistance exercise physiology takes nine to twelve to show meaningful change. And the Jacobs trial alone enrolled about as many patients as all of those pooled trials combined.
- 2024. Lin C-L network meta-analysis of all knee OA exercise (52 RCTs, 4,255 patients, 27 arms): low-load training plus BFR ranks first for muscle hypertrophy, SMD 1.28, SUCRA 0.94, GRADE high.
- 2025. The Jacobs RCT (n=120) and its one-year follow-up. Sørensen (n=96). Li (n=100, dose-finding): 80% AOP beats 30% on pain, proprioception and torque, no serious adverse events.
- 2025. Yang network meta (14 RCTs, 866 patients): BFR beats low-intensity training on pain, function and strength; matches high-intensity with fewer adverse events. Lin Q (10 RCTs): pain SMD 0.25, strength 0.46; over 65, pain relief dominates; 65 and under, strength and sit-to-stand dominate.
- 2026. Machado overview of 18 systematic reviews (57 RCTs, 5,656 participants): vs low load, strength SMD 0.75 and hypertrophy 0.81; vs high load, adverse events RR 0.26 (0.09–0.72).
- 2026, the comparison that settles it. A JOSPT dose-response meta-analysis of conventional resistance training in knee OA, no BFR (20 RCTs, n=791; Núñez-Cortés 2026), found that very light to moderate loads gave the best pain relief and were the loads patients could tolerate. The strength gain from that tolerable load: SMD 0.12, statistically significant and clinically invisible. The same review estimated the volume needed: about 6,000 repetitions for pain, 2,600 for function, 1,600 for strength. Set that beside the BFR pooled estimate of 0.75 to 0.81 for strength and hypertrophy at the same loads, and beside the roughly 3,600 cuffed repetitions in the Jacobs protocol, which clears the function and strength thresholds and most of the pain threshold in a single twice-weekly program.
The skeptics, and the honest reply
Zeitlin 2025 pooled BFR across all knee conditions (15 trials, only 3 in OA): pain SMD 0.47, very low certainty, "exercise caution." Two fair responses: the search closed in October 2023, before either Jacobs paper or Sørensen; and pooling ACL reconstruction, patellofemoral pain and cartilage repair with OA dilutes the population where the new evidence is strongest. Ferlito 2020 has been retracted and should not be cited.
Earlier trials worth knowing
| Trial | What it showed |
|---|---|
| Ferraz 2018 n=48, 12 wk |
Strength and quad CSA (+7%) matched 80% 1RM training. WOMAC pain improved with BFR (−45%) but not with high load. 4 of 16 high-load patients withdrew for knee pain; none with BFR. |
| Bryk 2016 n=34, 6 wk |
Equal gains to 70% 1RM; less anterior knee pain during BFR sessions. Used a fixed 200 mmHg, not the recommended individualized approach. |
| Harper 2019 n=35 pilot |
Older US veterans, 20% 1RM BFR vs 60% 1RM: numerically favoured moderate load; fewer knee pain reports with BFR (3 vs 14). Underpowered, but it exists. |
| Petersson 2022 n=14 feasibility |
BFR walking in elderly knee OA: 5 of 14 withdrew, the heaviest, most painful, slowest patients. Walking BFR is a tool for some, not a default. |
| Ogrezeanu 2023 n=26 crossover |
End-stage knee OA, sham-controlled: 80% AOP gave no more hypoalgesia than 40% and worsened acute pain. Higher pressure buys adaptation, not comfort. |
The pattern to recognise
Compared with load-matched low-load training, BFR is superior for strength and size. Compared with high-load training, it is equivalent with fewer adverse events. Compared with standard neuromuscular care, it produces larger functional gains. The question was never "does BFR beat heavy lifting." It is "what do you give the patient who cannot lift heavy," and two trials now answer that, one of them with a year of follow-up.
The loading spectrum
The whole literature in one table. Show it to a patient and ask which row they want.
| Approach | Patient tolerance & safety | Pain relief | Functional improvement | Strength & hypertrophy |
|---|---|---|---|---|
| Low load, no cuff 20–40% 1RM |
High | Good | Moderate | Negligible (SMD 0.12; Núñez-Cortés 2026) |
| High load 70–90% 1RM |
Low. Adverse events roughly 4× BFR (Machado 2026); 1 in 4 high-load patients withdrew for pain in Ferraz 2018 | Poor | Good | Moderate to high |
| Low load + BFR 20–30% 1RM at 60% LOP |
High | High | High (ES 0.75–1.56 across Jacobs and Sørensen) | High (SMD 0.75–0.81 vs low load; equal to high load) |
All reviews and trials verified on PubMed; identifiers in References.
Section 5
Patient selection and screening
The trial cohorts were deliberately clean. Real-world knee OA patients are heavier and carry more comorbidity, which makes structured screening the part of the protocol you cannot skip. The contraindications below are the ORS standard taught in PBFR certification.
Good candidates
- Symptomatic knee OA, KL grade 1–4, with pain or loss of function limiting loaded exercise
- Not currently a surgical candidate, or choosing to defer surgery
- Failed or plateaued on conventional low-intensity exercise therapy
- Patients whose pain makes 70% 1RM loading intolerable
- Older patients where pain relief is the primary goal; younger patients where strength and function are the goal (Lin 2025)
- Able to attend 2 supervised sessions per week for 12 weeks
Absolute contraindications
- Unstable and/or severe hypertension
- Sickle cell anemia
- Active or recent venous thromboembolism or DVT
- IV drug use in the limb intended for BFR
- Lymphedema
Relative: medical clearance first
- Remote history of VTE or DVT
- Active cancer
- Compromised vascular circulation
- Limb of lymphectomy
- Inherited thrombophilia
Any condition for which your institution's tourniquet policy would prohibit cuff use also applies.
Additional screening prompts
Not contraindications. Findings from the trial literature that should shape how you start.
- Recent immobilisation, long-haul travel, active malignancy, smoking: cumulative VTE risk. Ask.
- BMI >30: excluded from Jacobs; the dropouts in the BFR-walking feasibility study were the heaviest patients. A reason to start conservatively, not to decline.
- Metabolic disease: excluded from Jacobs; one multicentre RCT in knee OA with fatty liver disease (Hu 2023) was positive.
- Blood pressure: pooled data show no overall BP effect, but a +7.5 mmHg systolic signal in elderly subgroups (Yudiansyah 2026). Measure, monitor, use the lowest effective pressure.
Intake checklist
Safety, in numbers
Adverse events vs high-load training: RR 0.45 (knee OA pools), RR 0.26 (2026 overview of 18 reviews). Across 25,813 individuals in the broader literature: about 6.5% adverse events, overwhelmingly transient numbness, dizziness and bruising (Anderson 2022). A 2025 survey of 134 US physical therapists: zero major events, minor transient events in 8% (Weatherholt 2025). No knee OA trial has reported a VTE except the single Jacobs case.
Screening guidance draws on the ORS PBFR certification standard, Patterson 2019 (BFR position stand), da Cunha Nascimento 2026, and the Jacobs trial.
Section 6
The protocol card
Reproduced from the Jacobs trial, the only protocol with one-year outcomes, with the safety screen built into the card rather than filed elsewhere.
One session, in order
| Dose | 24 supervised sessions over 12 weeks, 2×/week, delivered by BFR-trained clinicians |
| Cuffed exercises | Leg press and leg extension only, single leg. All other exercises identical, no cuff. |
| Set scheme | 30 – 15 – 15 – max to fatigue. Cuff deflated between the two quadriceps exercises (reperfusion, not continuous occlusion). |
| Pressure | 60% of limb occlusion pressure, measured on the training limb, preferably in the training position. If the reading is unstable there, measure supine with the patient still. Mean LOP in the trial: 131 mmHg (SD 15). |
| Load | Start at 30% 1RM. Pain-free for two consecutive sessions: +5%. Pain persisting to the next day or next session: −5%, and a comfortable rep count on the last set. |
| Familiarisation | Weeks 1–2: cuff on the leg press only. |
| Measure LOP | Every session, before the first cuffed set. LOP changes with position, blood pressure, hydration and training status, and drifted down about 33 mmHg over 9 weeks in one study (Cerqueira 2026). On a system that automates it, this takes seconds. |
| Every session | Before the first cuffed set: resting BP; ask about new calf pain or swelling, new medication (including hormonal), recent travel or immobilisation; inspect the cuff site. Stop and reassess for calf pain, unusual swelling, or chest symptoms. |
| Expectation | No separation at 6 weeks. Advantage appears at 12 weeks and grows after discharge. |
Dosing notes from the literature
- Pressure. Individualised %LOP measured on the training limb is the defensible standard; fixed absolute pressures are not. The only head-to-head study in knee OA (Li 2025) found 80% AOP superior to 30% for adaptation, but in end-stage OA, 80% worsened acute pain without more hypoalgesia (Ogrezeanu 2023). Jacobs used 60%; Sørensen used 60–80%. 60% is a well-supported starting point.
- Load. 20–30% 1RM. Harper's 20% arm underperformed moderate load; Jacobs' 30% with a ±5% pain rule is the reference.
- Volume. A dose-response meta-analysis found an inverted U: roughly 2,000 total repetitions across a program maximised strength (Ogrezeanu 2025). More is not better.
- Exercise selection. Machine leg press and knee extension dominate every positive trial. Walking BFR is poorly tolerated in heavier, more painful patients.
- Pain, not just pressure. In persistent-pain populations, higher occlusive pressures can provoke hyperalgesia rather than relief, and catastrophizing and kinesiophobia amplify it (Gray 2026). If a patient's pain rises across sessions, drop pressure before dropping the program.
- Effort is the endpoint, not the number on the cuff. Both Jacobs arms trained to fatigue and progressed load the same way; only the cuffed arm changed. Fatigue is the great equalizer in resistance exercise, but a knee OA patient cannot reach true muscular fatigue at a load the joint tolerates without a cuff. The cuff is what gets them there in a modest number of reps. The practical corollary: a patient who does not tolerate the target pressure can train at a lower pressure and still get the effect, as long as the last set is taken to that effort endpoint.
- Weeks 1–2 are acclimation. A 12-week course is really ten weeks of training. That is an argument for the full 24 sessions, and, where you have to make the case to a payer, an argument for courses of care that match the physiology rather than the authorisation.
- Device. Whatever system you use must measure LOP on the limb being trained. The Jacobs trial used a validated LOP-measuring cuff system; the Delfi PTS measures LOP automatically and regulates pressure throughout the set.
Goal-based framing
Same mechanism, different bottleneck. Older patients start with more pain and less reserve, so hypoalgesia dominates what they feel; younger patients have more trainable muscle, so strength dominates what you measure (Lin 2025). Let the goal shape the conversation.
Protocol source: Jacobs et al. 2025, Table 1 and Methods (full text).
Section 7
Surgeons, referrals and patient expectations
Talking to an orthopaedic surgeon
Surgeons referred every patient in the Jacobs trial. The conversation works best when it is short, evidence-led, and about their problem: the patient who is not yet a surgical candidate, or who wants to defer, and who is cycling through injections.
A 60-second referral script
"For your knee OA patients who aren't ready for a replacement and can't tolerate loaded strengthening, we run the Ghent protocol, the 120-patient RCT in Annals of the Rheumatic Diseases and BJSM. Twelve weeks of supervised exercise with personalized blood flow restriction on the quad work. At one year the BFR group had a 15-point KOOS pain advantage and needed about two-thirds fewer intra-articular injections. A second trial out of Copenhagen found the same kind of functional gains against standard care. We screen for VTE risk, measure limb occlusion pressure on every patient, and send you baseline and discharge KOOS and function scores. Who would you like us to start with?"
What to send with the referral request
- The Jacobs and Sørensen citations. Surgeons respond to the journals.
- Your screening checklist (Section 5). It answers the safety question before it is asked.
- Your outcome set: KOOS, 30-s chair stand, 40 m walk at baseline, 12 weeks and 3 months
- A one-line note on the device: personalized LOP measurement on every patient
Or send all of it at once: the one-page referral sheet in Appendix A carries the evidence, the screen, the outcome set and a slot for your logo.
Positioning against the alternatives
- Versus repeat injections: steroid effect gone by 6 months; repeated triamcinolone accelerated cartilage loss (McAlindon 2017). BFR reduced injection need 63% at one year.
- Versus standard PT: same program, one added variable; exercise-only gains fade below MCID by a year, BFR gains kept growing. Against Danish standard neuromuscular care, BFR produced larger gains on every functional test.
- Versus waiting for surgery: a bridge that builds quadriceps strength, the variable most implicated in post-arthroplasty function, without loading the joint heavily. It works in both directions. Some patients defer or avoid the replacement. Others cannot get cleared for one because the quad is too weak to rehabilitate afterwards, and BFR is what gets them to the table. Either way the surgeon gets a stronger knee.
What to tell the patient
In plain words
- "Your knee needs a stronger thigh muscle, but heavy weights hurt. The cuff lets light weights do the job of heavy ones."
- "Ever notice stairs hurt more going down than up? Going down, your thigh muscle has to act as the brake. When it's weak, the knee lands stiff and the joint takes the hit. A stronger thigh is the brake you're missing."
- "It feels like hard work for a few minutes, a deep burn in the muscle, then it passes. You'll be more comfortable than you would be lifting heavy."
- "Two sessions a week for twelve weeks. Don't expect to feel a difference at six weeks; the study didn't either. The change comes at twelve, and keeps going."
- "In the biggest study, patients were doing an hour and a half more activity a week a year later, and needed far fewer injections."
- "Tell us about any history of blood clots, any hormone medication, and any unexplained calf pain, before we start."
Expectation timeline
| When | What to expect |
|---|---|
| Weeks 1–2 | Familiarisation; cuff on leg press only. Muscle burn, transient numbness or bruising under the cuff are normal. |
| Week 6 | Strength gains under way; patient-reported pain not yet different from standard exercise. Say so up front. |
| Week 12 | Discharge. Expect measurable strength (≈1 N/kg advantage), better pain, chair-stand and walking performance. |
| Month 3 after | Gains grow rather than fade. Encourage continued self-directed activity. |
| One year | Pain, function and quality-of-life advantages larger than at discharge; fewer injections; more weekly activity. |
The expectation lever
Unmet expectations are the strongest predictor of dissatisfaction after knee replacement. Setting them correctly here, including the six-week plateau, is not bedside manner. It is part of the treatment effect.
The patient handout that accompanies this guide is designed to be co-branded and handed across the desk.
Section 8
Cost, value and the payer conversation
Administrators and medical directors ask a different question than surgeons do. This section gives you what the evidence supports and, just as usefully, what it does not.
What a course of PT is worth against an injection
The only head-to-head economic comparison in knee OA is the Deyle trial and its cost-effectiveness analysis. In 156 randomized patients, physical therapy beat intra-articular glucocorticoid injection on WOMAC at one year by 18.8 points on the 0–240 scale (95% CI 5.0–32.6; Deyle 2020, NEJM).
The economics followed. Mean one-year knee-related costs were essentially equal, $2,113 in the injection arm versus $2,131 in the PT arm, while PT gained 0.076 more QALYs (95% CI 0.02–0.126, p=0.003). The incremental cost-effectiveness ratio for knee-related costs was $8,103 per QALY, with 99.2% of bootstrapped estimates below a $100,000 willingness-to-pay threshold (Rhon 2022, JAMA Netw Open).
The honest read: up front, an injection is much cheaper. Across a year, knee-related costs equalise and the therapy course buys meaningfully more health.
Where BFR sits in that picture
BFR is a delivery method layered onto therapeutic exercise, not a separate service. It does not generate additional billing. The same exercise codes apply whether or not a cuff is on the limb, so the incremental cost of adding BFR to a course of care is the device, amortised as practice overhead, and the clinician time to measure limb occlusion pressure.
Against that near-zero incremental cost, Jacobs found 12.5% of the BFR arm versus 33.3% of the exercise-alone arm received an intra-articular injection in the year after rehab: roughly one patient in five avoiding an injection episode, alongside pain and function advantages that were still growing twelve months out.
What this guide will not claim
There is no cost-effectiveness analysis of BFR versus injection, and none of BFR versus standard PT. Do not present one. In particular, do not build the argument on the dollar value of the avoided injections. A Medicare-allowed injection encounter is a modest sum, and the direct saving across a realistic cohort lands in the low thousands of dollars, a number that understates BFR rather than supporting it, and one a payer can argue back at you.
The defensible argument has three parts, in this order. BFR adds measurable effect at essentially no incremental billing cost. It sits inside a course of care already shown to be cost-effective against injection. And the harm it displaces, repeated corticosteroid with documented cartilage loss and no pain benefit at two years (McAlindon 2017), is a cost that shows up later, in a different budget line. Arthroplasty avoidance is the endpoint that would settle the economics, and no one has measured it.
Sources: Deyle 2020 NEJM (PMID 32268027); Rhon 2022 JAMA Netw Open (PMID 35072722); Jacobs 2025 BJSM. A separate ORS coding reference covers CPT usage for BFR-delivered exercise. It is a reference document, not billing advice, and does not form part of this clinical guide.
Section 9
What the evidence does not show
- No structural disease modification. No BFR trial in OA has measured cartilage, synovitis, effusion or joint space.
- No joint-loading measurement. "Lower external load" is measured; "less joint stress" is inferred.
- No arthroplasty-avoidance or cost-effectiveness data. Injection reduction is the closest proxy.
- Long-term durability rests on one trial. Jacobs is the only trial with a year of follow-up; Sørensen's 36-week KOOS follow-up found no between-group pain difference. Detraining after BFR in older adults is unstudied.
- Obese and comorbid patients are under-represented and were the dropouts when enrolled.
Why say it
Stating these plainly is what makes the rest of this guide credible to a surgeon, a medical director or a payer. Every claim in these pages is bounded by this list.
The right next study is a multicentre, sham-controlled, load-matched trial in a real-world cohort with arthroplasty incidence and cost as endpoints. ORS actively supports investigator-initiated research. Contact us.
Trained by ORS · delivered on Delfi
Owens Recovery Science pioneered BFR rehabilitation in the U.S. military at the Center for the Intrepid and has since trained more than 15,000 clinicians in personalized BFR.
The Delfi Personalized Tourniquet System measures each patient's limb occlusion pressure automatically and regulates cuff pressure throughout every set, the individualized approach the evidence base is built on.
Trusted by the Department of Defense, NASA, the U.S. Olympic & Paralympic Committee, teams across all five major U.S. professional sports leagues, more than 150 universities, and healthcare systems from the largest networks to independent clinics.
Get certified
ORS Personalized BFR certification, CME-accredited, live and online, teaches LOP-based dosing, screening and protocol design for orthopaedic populations. owensrecoveryscience.com/bfr-courses
Equip the clinic
The Delfi PTS Generation 3 measures limb occlusion pressure automatically and regulates pressure throughout the set. With Two Patient Mode, one Gen 3 now runs two independent treatment channels, so two knee OA patients can train on different protocols at different pressures, on one device, at the same time. Ask about clinic and health-system programs.
Hand this to a patient
The companion patient handout is designed to be co-branded with your clinic's logo and given across the desk. The surgeon referral sheet and the documentation template are Appendices A and B of this guide, fillable in the PDF; an editable Word version of the template is also available. Download at owensrecoveryscience.com/guides/bfr-knee-osteoarthritis
References
Sources
Every citation in this guide was verified on PubMed. Identifiers are given so any claim can be checked at its source.
Jacobs E et al. Ann Rheum Dis 2025;84:341-50. PMID 39919906
Jacobs E et al. Br J Sports Med 2025;59:1481-9. PMID 40846463
Sørensen B et al. Scand J Med Sci Sports 2025;35(10):e70154. PMID 41103113
Sørensen B et al. Mechanical muscle function, secondary analysis. Scand J Med Sci Sports 2025;35(5):e70069. PMID 40375440
Li et al. BMC Musculoskelet Disord 2025. PMID 40781308
Ferraz RB et al. Med Sci Sports Exerc 2018;50:897-905. PMID 29266093
Bryk FF et al. Knee Surg Sports Traumatol Arthrosc 2016. PMID 26971109
Harper SA et al. J Clin Med 2019;8:265. PMID 30795545
Petersson N et al. J Rehabil Med 2022;54. PMID 35266006
Hu et al. Front Endocrinol 2023. PMID 38155949
Cerqueira et al. Front Physiol 2026. PMID 41640842
Ogrezeanu DC et al. PM R 2023;15:1565-73. PMID 37796567
Ogrezeanu DC et al. Transl Sports Med 2025. PMID 40256207
Grantham B et al. Phys Ther Sport 2021;49:37-49. PMID 33582442
Wang HN et al. Arthritis Care Res 2022;74:89-98. PMID 34549541
Lin CL et al. Biomedicines 2024;12:1524. PMID 39062097
Lin Q et al. Front Physiol 2025;16:1524480. PMID 40166718
Yang H et al. Am J Phys Med Rehabil 2025. PMID 40856440
Zeitlin C et al. Phys Ther Sport 2025;74:65-74. PMID 40435680
Machado FA et al. Clin Med Insights Arthritis Musculoskelet Disord 2026. PMID 42367229
Centner C et al. Sports Med 2019;49:95-108. PMID 30306467
Núñez-Cortés R et al. Tailoring resistance exercise for knee osteoarthritis: dose-response meta-analysis. J Orthop Sports Phys Ther 2026;56(9):585-95. PMID 42676087
Hughes L, Patterson SD. J Appl Physiol 2020;128:914-24. PMID 32105522
Gray L, Hughes L, Kelly L, et al. Pressure point: blood flow restriction exercise and the pain paradox in musculoskeletal injury and persistent pain populations, a narrative review. Front Pain Res (Lausanne) 2026;7:1822981. PMID 42382319
Trinks N, et al., Pesta DH. Blood-flow restriction resistance training improves skeletal muscle mitochondrial capacity and cardiovascular risk factors in type 2 diabetes. Cell Metab 2026;38:812-823.e6. PMID 41610852
Minniti MC et al. Am J Sports Med 2020;48:1773-85. PMID 31710505
Anderson KD et al. Mil Med 2022;187:e1059-64. PMID 35284924
Weatherholt AM et al. Int J Exerc Sci 2025. PMID 40909299
Yudiansyah et al. J Sports Med Phys Fitness 2026. PMID 41757628
da Cunha Nascimento D et al. Semin Thromb Hemost 2026. PMID 42049170
Patterson SD et al. Front Physiol 2019;10:533. PMID 31156448
Dunlop DD et al. Am J Prev Med 2019;56:664-72. PMID 30902564
Messier SP et al. JAMA 2013;310:1263-73. PMID 24065013 · Arthritis Care Res 2018. PMID 29911741
Holden MA et al. Lancet Rheumatol 2023. PMID 38251550
Juhl C et al. Arthritis Rheumatol 2014;66:622-36. PMID 24574223
Bandak E et al. Ann Rheum Dis 2022;81:537-43. PMID 34844929
Øiestad BE et al. Br J Sports Med 2022;56:349-55. PMID 34916210
Hall M et al. Osteoarthritis Cartilage 2018;26:495-500. PMID 29427725
Skou ST et al. Osteoarthritis Cartilage 2016;24:1350-6. PMID 27066879
Bourne RB et al. Clin Orthop Relat Res 2010;468:57-63. PMID 19844772
Campbell CM et al. Arthritis Care Res 2015;67:1387-96. PMID 26041510
Rhon DI et al. J Bone Joint Surg Am 2022;104:1447-54. PMID 35700089
Rhon DI et al. Cost-effectiveness of physical therapy vs intra-articular glucocorticoid injection for knee osteoarthritis. JAMA Netw Open 2022;5:e2142709. PMID 35072722
Deyle GD et al. Physical therapy versus glucocorticoid injection for osteoarthritis of the knee. N Engl J Med 2020;382:1420-9. PMID 32268027
Jüni P et al. Cochrane 2015;CD005328. PMID 26490760
McAlindon TE et al. JAMA 2017;317:1967-75. PMID 28510679
Appendix A · Surgeon referral sheet
Supervised exercise with personalized blood flow restriction for knee OA
This page is laid out as a one-page handout in the PDF edition, with a slot for your clinic's logo. Print it on its own and hand it to a referring orthopaedic or primary care office.
For patients who are not yet surgical candidates, or who want to defer, and who cannot tolerate loaded quadriceps strengthening. We run the protocol from the two 2025 randomized trials below, screen for VTE risk, measure limb occlusion pressure on every patient, and send you baseline and discharge outcome scores.
The evidence, in brief
- Jacobs et al., Ghent University (n=120). Surgeon-referred knee OA, KL 1–4, 47% KL 3. Both arms did the same 24-session, 12-week program; the only variable was a cuff on the two quadriceps exercises. Pre-registered, assessor-blinded, significance thresholds about 25× stricter than p<0.05.
- At one year, with no restrictions on care after month 3: KOOS pain +15.1, quality of life +14.7, ADL +11.3 (all past MCID); 1.5 more hours of self-directed activity per week; 3.6× lower odds of a knee injection.
- Sørensen et al., Copenhagen (n=96). Same BFR protocol against the Danish standard of care (GLA:D). BFR superior on every functional test, quadriceps strength (ES 1.47), rate of force development and muscle cross-sectional area; reduced spreading pain sensitization. No exercise-related adverse events in either group.
- Safety: adverse events vs high-load training RR 0.26 (overview of 18 systematic reviews). One DVT in the Jacobs BFR arm, in a patient with two undisclosed risk factors; hence the screen.
What we do not claim
No structural disease modification measured; no arthroplasty-avoidance or cost-effectiveness data specific to BFR (injection reduction is the closest proxy); long-term durability rests on one trial.
Set expectations with the patient
No between-group difference at 6 weeks in either trial. The advantage appears at 12 weeks and, in the Jacobs follow-up, grew after discharge. We tell patients this on day one; it helps if they hear it from you too.
How we run it
- Two supervised sessions per week for 12 weeks; leg press and knee extension at 30% 1RM under the cuff, all other exercises uncuffed
- Cuff pressure set at 60% of the patient's limb occlusion pressure, measured at every session
- Load progressed by a pain rule: +5% after two pain-free sessions, −5% if pain persists to the next day
- Every session: resting BP, calf symptom check, medication and travel check, cuff-site inspection
Who we screen out
Absolute: unstable or severe hypertension; sickle cell anemia; active or recent VTE or DVT; IV drug use in the limb; lymphedema. With your clearance: remote history of VTE, active cancer, compromised vascular circulation, limb of lymphectomy, inherited thrombophilia. We also ask about hormonal therapy, recent immobilisation, family clotting history and unexplained calf symptoms.
What you get back
- KOOS, 30-second chair stand and 40 m fast walk at baseline, 12 weeks and 3 months
- Quadriceps strength and the pressure and load used
- Any adverse event, same day
- Whether the patient is on track to defer surgery, or is now a stronger surgical candidate
References: Jacobs E et al. Ann Rheum Dis 2025;84:341-350 (PMID 39919906). Jacobs E et al. Br J Sports Med 2025;59:1481-1489 (PMID 40846463). Sørensen B et al. Scand J Med Sci Sports 2025;35(10):e70154 (PMID 41103113) and 2025;35(5):e70069 (PMID 40375440). Machado FA et al. Clin Med Insights Arthritis Musculoskelet Disord 2026 (PMID 42367229).
Appendix B · Documentation template
Medical necessity and outcome documentation
A structured template to document medical necessity, establish objective baselines, and justify continued visits for a supervised exercise program, with or without blood flow restriction, in knee osteoarthritis. Complete at initial evaluation and at each reassessment. In the PDF edition every field is fillable; an editable Word version is available from the same download page.
Read this first
A documentation aid, not billing advice and not a substitute for your clinical judgment or your organisation's compliance policy. Coding rules, thresholds and payer requirements change and vary; verify them before relying on any line here. The evidence-based dose for BFR in knee OA is 24 supervised sessions over 12 weeks, twice weekly (Jacobs 2025; Sørensen 2025). Document the published dose and the outcomes it produced rather than an average visit count.
1. Diagnosis and severity
- Knee OA diagnosis and ICD-10 code (specify laterality)
- Radiographic grade (Kellgren-Lawrence 0–4) if available, and clinical severity
- Impairments driving the referral: quadriceps weakness or atrophy, loss of range of motion, gait deviation, functional instability, fall risk, pain limiting activity
- Relevant comorbidities and prior treatment (injections with dates, NSAIDs, prior PT)
2. Skilled-service justification
State why licensed therapist skill is required: individualised exercise prescription and progression, safety monitoring, manual therapy where indicated, contraindication screening, and, for BFR, individualised limb occlusion pressure measurement and titration. Physical therapy delivered as an individualised program outperformed intra-articular glucocorticoid injection on WOMAC at one year in a randomized trial (Deyle 2020, NEJM) and was cost-effective at $8,103 per QALY for knee-related costs (Rhon 2022, JAMA Netw Open).
3. Objective baseline
Measure each targeted deficit at baseline and repeat the same tools at reassessment. The performance tests below are the OARSI-recommended set for knee OA (Dobson 2013) and match the outcome battery used in the Jacobs and Sørensen trials.
| Domain | Tool | Baseline | Date |
|---|---|---|---|
| Pain and function | KOOS subscales, 0–100; MCID about 10 points. If using WOMAC, record the version and scale (0–96 or 0–240) so the MCID matches. | ||
| Pain intensity | NPRS (0–10) | ||
| Strength and endurance | 30-second chair stand (repetitions) | ||
| Gait speed | 40 m fast-paced walk (seconds or m/s) | ||
| Stair function | Stair climb test (seconds) | ||
| Walking capacity | 6-minute walk (m) or Timed Up and Go (s) | ||
| Range of motion | Knee flexion and extension (goniometry) | ||
| Quadriceps strength | Dynamometry (N/kg or Nm/kg) or estimated 1RM |
4. Measurable, functional, time-bound goals
Three goals, each tied to an activity (stairs, walking distance, sit-to-stand, work or recreation), not to the modality, with a target week.
5. Plan of care
- Interventions and CPT codes anticipated (confirm with your billing department): 97110 therapeutic exercise, 97112 neuromuscular re-education, 97530 therapeutic activities, 97116 gait training
- Frequency and duration, and estimated total visits. Published BFR dose: 2 sessions per week for 12 weeks, 24 supervised sessions; the separation between groups appeared at week 12, not week 6
- Weeks 1–2 of the published protocol are familiarisation (cuff on one exercise only), so a 12-week course is ten weeks of progressive training. A shorter course has no supporting evidence in knee OA
6. BFR-specific documentation (if used)
- Rationale for low-load training: joint pain precluding 65–70% 1RM loading despite a documented strength deficit
- Limb occlusion pressure measured (mmHg, limb), position (exercise position preferred, or supine and still), training pressure as a percentage of LOP, and the device and method (for example Delfi PTS automatic LOP; Doppler)
- Contraindication screen completed (ORS standard). Absolute: unstable or severe hypertension; sickle cell anemia; active or recent VTE or DVT; IV drug use in the limb; lymphedema. Relative, medical clearance obtained: remote history of VTE or DVT; active cancer; compromised vascular circulation; limb of lymphectomy; inherited thrombophilia. Screening prompts: hormonal contraception or HRT; recent immobilisation or long travel; family history of thrombophilia; unexplained calf pain or swelling; resting BP recorded
- Targeted quadriceps strength deficit and load prescription (start 30% 1RM; ±5% by pain rule)
- Session-day check documented each visit: BP, new calf symptoms, new medication, cuff-site inspection
7. Reassessment (repeat Section 3 at 8–12 weeks and at discharge)
Record the reassessment value, the change from baseline and the date for each tool; then progress toward goals, and, if goals are not met, the reasoned modification (type, intensity, duration, frequency). Set expectations in the note as well as with the patient: in the reference trial, no KOOS subscale separated between groups at week 6; the advantage appeared at week 12 and grew afterwards.
8. Continued-visit justification
- Remaining measurable deficits
- Demonstrated ongoing response (objective measures)
- Why skilled care, rather than an independent home program, is still required
- Medicare Part B: KX modifier applied where medical necessity is attested beyond the therapy threshold. Confirm the current threshold and payer policy
Anchor measures to a fixed schedule
Baseline, reassessment at 8–12 weeks, discharge, and where possible a follow-up at 3 months, using the same tools each time. Patients in the reference trial kept improving after discharge; a follow-up measurement is the documentation that shows it.
Additional sources for this template: Dobson F et al. OARSI recommended performance-based tests. Osteoarthritis Cartilage 2013;21:1042-1052. VA/DoD Clinical Practice Guideline for the Non-Surgical Management of Hip and Knee Osteoarthritis, 2020. Documentation aid for licensed clinicians; not billing, coding or legal advice.
