Updated May 2026. Written by the Upwell Health Collective clinical team. Clinically reviewed May 2026. Next review due November 2026.
Reading time: approximately 28–32 minutes.
Disclaimer: This article is for educational purposes only. It does not constitute clinical advice. ACL management decisions should be made in conjunction with a qualified physiotherapist, sports physician, and/or orthopaedic surgeon with experience in ACL healing protocols. If you have recently sustained a knee injury, please seek assessment promptly as timing is critical.
You've just ruptured your ACL — or someone you know has — and somewhere in the fog of the first few days you read something about knees that can heal without surgery. A protocol. An Australian surgeon. A 90-degree brace. Ninety percent healing on MRI.
And you're wondering: is this real? Is my ACL actually healable?
The careful, honest answer, based on the best available evidence in 2026, is: in carefully selected acute injuries with favourable MRI features, the probability may be meaningfully higher than we once believed — but it is not guaranteed, and the definitive trial data is not yet in.
This is the complete, unvarnished guide to ACL healing — the biology behind why it wasn't supposed to work, the research that suggests it can, the clinical protocol that has changed the conversation, who it's for, who it isn't for, what the contrarian evidence says, and what the critical uncertainties still are in 2026.
Let's go deep.
For most of modern sports medicine, the ACL was considered the ligament that couldn't heal itself. Tear your medial collateral ligament (MCL)? It'll usually heal on its own with a brace and time. Tear your ACL? Surgery, reconstruction, 9–12 months of rehab. That was the orthodoxy, backed by decades of clinical experience.
The ACL is an intra-articular structure — it sits inside the knee joint, bathed in synovial fluid. This is fundamentally different from extra-articular ligaments like the MCL, which sit outside the joint capsule.
When any tissue tears, the body's first healing response is to form a fibrin-platelet clot at the injury site. This provisional scaffold is what allows cells to migrate in, lay down new collagen, and begin rebuilding. The MCL does this beautifully: blood pools around the torn ends, a clot forms, tissue grows back into it, healing proceeds over weeks.
The ACL can't do this. Synovial fluid contains high concentrations of plasmin — a proteolytic enzyme that breaks down fibrin clots. The moment any blood tries to coagulate around the torn ACL ends, the plasmin dissolves it (Murray & Fleming, 2013; Gagliardi et al., 2019). No clot means no provisional scaffold. No scaffold means no pathway for cells to bridge the gap.
The torn ends of the ACL retract, become encased in synovial membrane, and involute. Without intervention, this process is largely complete within 3–4 weeks of injury — which is precisely why the CBP timing window is so strict.
There are also additional biological challenges:
CBP borrows the orthopaedic logic of reduction and protection: bring the torn ends closer, protect the tissue early, then gradually reload. But ligament healing inside a synovial joint is not the same as bone healing, and the quality of the healed ACL still matters enormously.
The fact that healing can occur was hiding in plain sight. A 2021 systematic review identified just six studies examining ACL healing on MRI after rupture — all of low methodological quality and mostly with fewer than 50 participants. Spontaneous healing was noted in some, but the evidence was weak and the question largely unresolved.
Then came the secondary KANON trial analyses.
The KANON trial (Frobell et al., 2010) randomised 121 young, active adults with acute ACL ruptures to either early reconstruction plus rehab, or structured rehabilitation with optional delayed surgery.
Secondary analyses (Filbay et al., 2023) showed MRI evidence of ACL healing in 16 of 54 (30%, 95% CI 19–43%) participants randomised to optional delayed ACLR at 2-year follow-up. Excluding those who had eventually chosen delayed surgery, 53% of the rehabilitation-alone participants showed evidence of healing. Those with healed ACLs had significantly better 2-year outcomes: KOOS Sport/Rec higher by 25.1 points (95% CI 8.6–41.5) versus non-healed; KOOS QOL higher by 27.5 points (95% CI 13.2–41.8).
Critical nuance — the 11-year picture: The 2025 KANON 11-year secondary analysis (Filbay et al., 2025) complicates the healing narrative. ACL continuity on 5-year MRI was not clearly associated with superior 11-year outcomes, and in some analyses was associated with worse 11-year patient-reported outcomes compared with ACL reconstruction. This is an important corrective. MRI continuity at 2 years does not automatically translate into long-term superiority. The question is not simply whether an ACL heals, but whether it heals with sufficient quality, tension, and stability to protect the joint over a decade or more.
The logical question from the 2-year data remained: if the ACL can sometimes heal spontaneously, can we create conditions that make healing predictable?
Sydney orthopaedic surgeon Mervyn Cross OAM and his son, sports physician Dr Tom Cross, developed a structured bracing protocol based on a biomechanical hypothesis: if the torn ends of the ACL can be approximated and the knee kept still enough to allow a provisional scaffold to form, the body's own healing machinery might do the rest.
The CBP addresses both key biological barriers:
The CBP's first successful patient was a 19-year-old netballer in 2014. By the time the first 80 patients were written up for publication, the results attracted Dr Stephanie Filbay at the University of Melbourne, who led the formal 2023 publication.
Filbay SR, Dowsett M, Chaker Jomaa M, et al. Healing of acute anterior cruciate ligament rupture on MRI and outcomes following non-surgical management with the Cross Bracing Protocol. Br J Sports Med. 2023 Dec;57(23):1490–1497. DOI: 10.1136/bjsports-2023-106931. PMID: 37316199. PMC10715498.
Important context before the data: This study is a clinical case series with protocol evolution over time, not a randomised controlled trial. The protocol changed after early DVT events (anticoagulation was added after initial DVT occurrences, with later patients receiving rivaroxaban 10mg during the early bracing period). After patient 50, patients with unfavourable MRI features were actively discouraged from CBP, creating a possible selection effect. Laxity testing was performed by unblinded physicians. There was no surgical or rehab-alone control group. The authors themselves state that longer-term follow-up and randomised clinical trials are needed. This is hypothesis-generating, important, and promising — but not definitive.
The cohort: 80 consecutive patients within 4 weeks of ACL rupture. Mean age 26 ± 10 years, 39% female, 49% concomitant meniscal injury, two-thirds competitive or professional athletes.
90% had evidence of ACL continuity on 3-month MRI. But not all healing is equal.
Lysholm: Grade 1 median 98 (IQR 94–100) vs Grade 2–3 median 94 (IQR 85–100)
ACLQOL: Grade 1 median 89 (IQR 76–96) vs Grade 2–3 median 70 (IQR 64–82)
Normal laxity at 3 months: Grade 1 100% vs Grade 2–3 40%
Return to pre-injury sport at 12 months: Grade 1 92% vs Grade 2–3 64%
The 90% headline is real. The more important number is the 50% who achieved Grade 1 — and within that group, the outcomes were exceptional.
11 of 80 patients (14%) re-injured within follow-up. Three re-injuries in the Grade 1 group were high-force contact events (skiing accidents, rugby, AFL contact). After re-injury, 9 of 11 proceeded to reconstruction. CBP does not close the surgical door — it potentially defers or avoids it for suitable patients.
38 of 39 meniscal tears (97%) became asymptomatic following CBP — a striking finding. However, "asymptomatic" does not necessarily prove structural meniscal healing on MRI. Most meniscal injuries in this cohort were stable tear patterns (vertical posterior horn tears, with one displaced ramp lesion). This finding is promising but should not be read as evidence that CBP heals all meniscal tear types. The meniscus finding deserves cautious optimism, not overclaiming.
By 2025, the Stadium Clinic team had managed over 1,080 patients with CBP. This experience allowed identification of acute MRI features that appear to predict healing quality.
Filbay SR, Dowsett M, van Haeringen M, et al. ACL-ARCH MRI criteria. J Sci Med Sport. 2026 Feb;29(2):145–148. PMID: 40935695. Published online August 2025.
Four MRI features of acute ACL injury relevant to healing potential:
1. Tibial/femoral attachment integrity
Femoral avulsion is the strongest predictor of suboptimal healing. In the 80-patient cohort, 37 of 40 (93%) with Grade 2–3 healing had a partial femoral avulsion, versus only 7 of 40 (18%) with Grade 1.
2. ACL tissue displacement outside the intercondylar notch
When fibres are displaced outside the notch, the torn ends cannot be approximated by positioning. The protocol was updated after patient 50 to actively discourage CBP in these cases.
3. Gap distance between torn ends
Gap <6mm: associated with higher Grade 1 healing probability. Gap ≥7mm: 88% of non-healers had a gap of ≥7mm — a major red flag for candidacy.
4. Retraction and involution of ACL stumps
Once the torn ends retract into rounded stumps, end-to-end healing potential is significantly reduced. Involution is substantially complete by 3–4 weeks post-injury.
An important accuracy note: The ACL-ARCH criteria are based on clinical experience with 1,080 patients and represent the most sophisticated current framework for assessing CBP candidacy. However, they are not yet a prospectively validated clinical prediction rule. They require further validation before they can be used with formal probability estimates. They shift the conversation from binary eligibility to nuanced MRI risk stratification — but they are observational guidance, not validated criteria.
Eligibility window: Within 4 weeks of confirmed acute ACL rupture. Ideally 4–10 days. The EMBRACE trial targets brace fitting within 21 days — the earlier the better.
Weeks 1–4: Knee locked at 90° flexion, 24/7. Non-weight bearing. Crutches or mobility scooter required because the knee cannot extend for normal walking. The specific mobility aid and weight-bearing status should be directed by the prescribing CBP clinician based on pain, safety, DVT risk, and individual protocol setup — this will vary.
DVT management: Prolonged bracing and reduced mobility increase DVT risk. In the original study, DVT prophylaxis was added after early DVT events, with later patients receiving rivaroxaban 10mg during the early bracing period. CBP requires medical oversight and prescribed anticoagulation. This is non-negotiable.
Weeks 5–6: Brace opened 90° to 60° (Week 5), then 90° to 45° (Week 6). Non-weight bearing.
Weeks 7–8: Partial weight bearing. 30° to full flexion (Week 7), 20° to full flexion (Week 8).
Weeks 9–12: Full weight bearing. 10° to full flexion (Week 9), full ROM and full weight bearing (Weeks 10–12).
Post-week 12: Brace removed. Criteria-based rehabilitation begins in full — same framework as post-surgical patients: strength, neuromuscular control, running, plyometrics, return to sport.
Throughout: manual therapy, strengthening, and gait retraining under physiotherapist supervision.
Acute injury (ideally within 4–10 days), small gap distance (<6mm), femoral attachment largely intact, no tissue displacement, no absolute contraindications, full protocol compliance possible, strong home support.
If you've just injured your knee, time is your most precious resource. Contact us urgently — every day matters.
Porter MD, Shadbolt B. Cross Bracing Protocol for ACL Rupture Has Unacceptably High Failure Rate Relative to Surgical Stabilization: A 2-year Controlled Cohort Study. Clin J Sport Med. 2026 Feb 2. doi:10.1097/JSM.0000000000001416. PMID: 41622530.
This study compared 40 CBP patients vs 40 surgical stabilisation patients, all with isolated non-contact ACL ruptures, skeletally mature (16–40 years), active in pivoting sports, intending to return. Results: CBP group had 70% recurrent instability vs 2.5% surgical, 62% vs 2.5% medial meniscal tears, and inferior scores on all PROMs at 2 years. These are serious findings that cannot be dismissed.
Context that matters:
The honest summary: For high-level pivoting athletes intending aggressive return to competitive sport, the evidence for CBP is substantially less compelling than for recreational athletes, older adults, or patients with ideal injury anatomy and timing. The Porter & Shadbolt data is a critical signal for patient selection — not evidence that CBP is universally ineffective, but evidence that it should not be offered broadly to all acute ACL ruptures.
The Australian EMBRACE trial is the randomised controlled trial comparing CBP directly with early ACL reconstruction. It is currently recruiting participants aged 16–40 with primary acute ACL rupture within the last 15–20 days, MRI confirmation, and at least moderate pre-injury activity. Brace fitting occurs within 21 days of injury. Primary outcomes include KOOS4 at 6, 12, and 18 months.
Until EMBRACE reports, CBP remains a promising but emerging pathway — not a proven replacement for reconstruction. The trial is also practically useful as a current eligibility guide: if you have an acute injury within ~15 days, you may be EMBRACE-eligible, which means you're almost certainly within the CBP assessment window.
CBP is not a shortcut. 12 weeks of bracing, followed by the same criteria-based rehab as post-surgical patients. Total sport return timeline is similar to reconstruction — typically 9–12 months.
MRI continuity ≠ stability. A Grade 2 heal (thinned/elongated) may appear continuous on MRI but not provide the mechanical stability of a thick, taut Grade 1 ligament. Lachman's and pivot-shift at 3 and 6 months tell you what the healed tissue actually does functionally.
CBP doesn't close the surgical door. Suboptimal healing, re-injury, or instability may still require reconstruction. 9 of 11 re-injured patients in the Filbay cohort eventually had surgery.
CBP is not yet standard of care. It remains an emerging protocol with a compelling but early evidence base. The strongest study is a case series with protocol evolution. Informed consent must include honest discussion of what is known, what is uncertain, and what Porter & Shadbolt found in high-demand athletes.
CBP requires experienced clinical teams. MRI review with ACL-ARCH criteria, contraindication screening, DVT management and anticoagulation prescription, brace fitting, and supervised rehabilitation requires a team with specific CBP experience.
At full extension (0°), the ACL is under maximum tension and the femoral and tibial attachments are at their greatest distance. At 90° of flexion, the femoral condyle rolls and glides to a position that reduces tibial translation and brings the two torn ends closest together.
Immobilising at 90° achieves a "closed reduction" of the soft tissue injury — analogous to how a fracture is reduced before immobilisation. The gap distance data from ACL-ARCH confirms this mechanically: when the torn ends are within 6mm of each other and can be brought into contact by the 90° position, Grade 1 healing is far more likely than when there's a large gap or displaced tissue that positioning cannot approximate.
Healing in the ACL context is not the same as a normal ACL. A Grade 1 ACLOAS-scored ligament — thickened, continuous, with high intraligamentous signal — is not structurally identical to a pre-injury ACL. The collagen architecture, mechanoreceptor density, and precise biomechanical properties will differ.
What the evidence does say is that Grade 1 healing is associated with: normal laxity (100% in the cohort), 92% return to sport, Lysholm scores (median 98) among the highest reported in any ACL management cohort, and better 2-year KOOS outcomes than the surgically reconstructed KANON group.
The critical caveats: the KANON 11-year data shows that ACL continuity does not automatically translate to long-term superiority, and we don't yet have specific long-term OA or functional outcome data for CBP patients. These are the most important evidence gaps remaining.
At 11-year follow-up in KANON, 44% of all participants had radiographic osteoarthritis in the injured knee, regardless of surgical pathway. Early surgery did not protect against arthritis.
Does CBP — preserving native tissue and avoiding a surgical insult — offer better OA protection? Theoretically plausible: no donor site morbidity, no second inflammatory hit from surgery, potential preservation of mechanoreceptors, KANON data suggesting continuity is associated with less OA at 11 years in some analyses. But we do not yet have OA data specifically for CBP vs reconstructed patients. This is a critical evidence gap that will become answerable as the CBP cohort ages.
If you've passed the 4-week window, you still have options:
If you're still within the window, don't delay. Contact our team at Upwell for urgent assessment, or book directly online.
EMBRACE trial results will be the most important data in this space over the next 2–3 years. Watch for KOOS4 outcomes comparing CBP with early ACLR in a properly randomised, controlled design.
BEAR (Bridge-Enhanced ACL Restoration): FDA-approved 2020 for adults, expanded 2025 to children/adolescents and partial tears, updated FDA-cleared labelling in 2026 around radiographic PTOA risk vs hamstring reconstruction. BEAR is a surgical complement to CBP's non-surgical approach — using a collagen scaffold soaked in the patient's own blood to bridge the torn ends. Not yet widely available in Australia — do not factor into an immediate Australian decision without confirming local availability.
Biologic augmentation: PRP, stem cell therapies, and novel fibrin-collagen scaffolds under investigation. None has yet achieved the evidence base for standard recommendation, but the rationale — providing a clot-stabilising scaffold that resists plasmin degradation — is exactly what both BEAR and CBP attempt by different means.
Validated prediction rules: The ACL-ARCH criteria represent the beginning of MRI-guided patient selection. As the dataset grows, validated prediction rules with quantified probability estimates will eventually be possible.
If you think you've torn your ACL and want to explore CBP:
The EMBRACE trial makes this urgency concrete: eligibility requires acute ACL rupture within the last 15 days. If you're reading this within two weeks of a suspected ACL tear, this is time-sensitive.
CBP should be considered an emerging, evidence-informed option — not a guaranteed surgery replacement. Before proceeding, patients should understand:
| Pathway | Best suited to | Main upside | Main risk |
|---|---|---|---|
| CBP | Acute (<4 weeks, ideally <10 days), favourable MRI (small gap, intact femoral attachment, no displacement), highly compliant patient, no contraindications | Possible native ACL healing; avoids surgery and graft morbidity | Suboptimal healing; higher instability risk in high-demand athletes; medical oversight required; emerging evidence base |
| Rehab-first (non-operative) | Lower-demand athletes, older adults, patients willing to test function first, missed CBP window | Avoids surgery if knee proves stable; 30–53% spontaneous healing possible | Persistent instability; risk of secondary meniscal/cartilage damage if unstable |
| Early ACL reconstruction | High-demand pivoting athlete, significant instability, locked knee, repair-demanding meniscal injury, professional sport timeline, unfavourable CBP MRI features | Predictable stabilisation; proven long-term evidence base | Surgical/graft morbidity; no OA guarantee; graft ligamentisation 12–18 months |
Step 1: Clinical assessment — Senior musculoskeletal physiotherapist: mechanism, laxity testing (Lachman's, pivot-shift), associated injury screening, absolute contraindication review.
Step 2: MRI review — Specific ACL-ARCH feature assessment: gap distance, femoral attachment integrity, tissue displacement, involution status. We tell you explicitly what your MRI shows and what it means.
Step 3: Shared decision-making — Full evidence picture: CBP outcomes by healing grade, Porter & Shadbolt data, rehab-first evidence, surgical options. We educate. You decide.
Step 4: Referral if CBP is chosen — CBP requires a prescribing physician with DVT management experience, brace fitting, and supervised rehabilitation. We have established referral pathways to Melbourne clinicians experienced in CBP. Your rehabilitation is managed through our physiotherapy and exercise physiology teams using criteria-based frameworks identical to post-surgical patients.
Step 5: Outcome monitoring — 3-month and 6-month MRI, clinical laxity testing, MRSS 2.0 progression. If healing is insufficient, the surgical pathway remains open.
What we know: ACL healing is biologically possible. CBP achieves 90% MRI continuity at 3 months in a selected consecutive series. Grade 1 healing (50% of the cohort) is associated with 92% return to sport, 100% normal laxity, and exceptional Lysholm scores. The ACL-ARCH criteria provide increasingly sophisticated MRI-based candidacy guidance. The KANON trial confirms healed ACLs produce better 2-year outcomes than non-healed. Contrarian data (Porter & Shadbolt, 2026) shows 70% recurrent instability in high-level pivoting sport athletes at 2 years.
What we don't know: Long-term (5+ year) outcomes specifically for CBP patients. OA risk vs reconstruction at 10+ years. Whether CBP outcomes hold in RCT conditions. Precise validated candidacy prediction rules. Whether the KANON 11-year nuance (continuity ≠ long-term superiority) applies to CBP-managed patients specifically.
What it means for you: CBP deserves serious consideration for the right patient, with the right injury anatomy, assessed promptly by an experienced team. For recreational athletes, older adults, and patients with favourable MRI features and full protocol compliance ability, the evidence is genuinely promising. For elite athletes needing aggressive return timelines, or patients with unfavourable MRI features, surgery may remain the more reliable path. That's not anti-CBP. That's honest medicine.
The most important thing: this decision needs to be made quickly, by an experienced team, with a high-quality acute MRI in hand. The window closes fast.
If you're within the window, contact us today. If you've passed it, we can still help navigate the full range of options. Book online here.
This article is for educational purposes only and does not constitute clinical advice. ACL management decisions should be made with a qualified physiotherapist, sports physician, and/or orthopaedic surgeon experienced in ACL healing protocols. If you have recently sustained a knee injury, please seek assessment promptly. This information was last reviewed May 2026.