Background

A common belief among researchers and clinicians is that a ruptured anterior cruciate ligament (ACL) has limited healing capacity. This belief has shaped current management strategies for ACL rupture. However, anatomical studies have demonstrated that the ACL has a rich vascular supply1 2 and histological studies describe ruptured ACLs passing though the typical phases of healing after injury, despite a slower rate of healing and reduced healing capacity compared with medial collateral ligament rupture.35 An absence of tissue bridging the gap between ligament remnants has been observed, which may inhibit healing of ACL rupture.6 The distance between the ACL origin to its insertion is shortest at 90135 of knee flexion.7 We have developed the novel Cross Bracing Protocol (CBP) that aims to reduce the gap distance between the ligament remnants by immobilising the knee at 90 of flexion for 4 weeks after acute ACL rupture in attempt to facilitate bridging of tissue and healing between the ruptured ACL remnants. After 4 weeks, knee range-of-motion is increased at weekly increments and the CBP is coupled with physiotherapist-supervised rehabilitation targeting lower limb neuromuscular control, muscle strengthening and power, and functional training to enable return-to-sport and recreational activities.

A 2021 systematic review identified only six studies that evaluated ACL healing on MRI after ACL rupture.8 Studies were of low methodological quality and five studies included 50 participants.8 More recently, an analysis of the KANON Trial observed MRI evidence of ACL healing at a 2-year follow-up in 30% of participants who were randomised to initial rehabilitation and optional delayed ACL reconstruction (ACLR).9 Those with MRI evidence of ACL healing reported better 2-year knee function and quality of life (QOL), compared with participants with no MRI evidence of ACL healing, and participants managed with early or delayed ACLR.9 Many ACL injured people experience poor long-term outcomes, including sport and activity limitations, persistent pain, an early onset of osteoarthritis and poor long-term QOL.1013 Considering the suboptimal outcomes with current management strategies, and the potential for ACL healing to result in favourable outcomes, new strategies to preserve and heal the native ACL should be explored. The objective of this study was to investigate MRI evidence of ACL healing, patient-reported outcomes and knee laxity in the first 80 individuals with acute ACL rupture managed non-surgically with the CBP.

This case series investigates outcomes from 80 consecutive patients with acute ACL rupture who were managed with the CBP. Data were collected in the course of clinical practice, and all participants provided informed consent for their data to be included in this study.

Eighty patients between the ages of 10 years and 58 years (mean (SD): 26 years (10 years)), who presented to a private sport and exercise medicine physician in Sydney, Australia (TC), with MRI confirmed acute ACL rupture between March 2016 and September 2021, were managed with the CBP (figure 1). Twelve out of 80 patients were residing outside of Sydney or impacted by COVID-19 restrictions and underwent virtual specialist consultations (TC), supplemented by in-person management and assessment from an experienced sports and exercise physician and physiotherapists trained in the CBP.

Participant flow chart. ACL, anterior cruciate ligament; ACLR, anterior cruciate ligament reconstruction; CBP, Cross Bracing Protocol; DVT, deep vein thrombosis; MCL, medial collateral ligament.

Patients of all ages, were considered eligible for the CBP if they presented within 1 month of acute ACL rupture, confirmed on MRI (ie, an ACL OsteoArthritis Score (ACLOAS) grade 3 representing full discontinuity of the ACL). To be considered for the CBP, patients needed to be functionally independent and capable of managing a period of knee immobilisation. Patients were considered ineligible if they had clinical or MRI evidence of structural concomitant injuries that necessitated surgical intervention (eg, an unstable bucket-handle meniscal tear) or a history of deep vein thrombosis (DVT) or pulmonary embolism. After the 10th participant, DVT screening was added to the eligibility criteria, whereby all patients underwent Doppler ultrasound to exclude DVT. The flow of participants through the study, including reasons for not offering the CBP and reasons for choosing ACLR (no participants chose rehabilitation alone), is presented in figure 1.

Patients were informed about treatment options: early ACLR, non-operative rehabilitation with optional delayed ACLR or trialling the CBP. Patients received information on the rationale and theoretical justification for the CBP, and they were aware that this was an experimental treatment with a chance of failure and possible need for ACLR in the future. The decision to trial the CBP was based on patient preference.

The CBP and accompanying rehabilitation protocol is described in online supplemental appendix 1. In patients presenting in the first week post-injury, the use of cryotherapy and anti-inflammatory medications was discouraged to minimise impairment of the acute inflammatory response.5 14 Paracetamol was prescribed as needed for pain. Fourteen patients who presented 7 days post-injury with minimal or no hemarthrosis/effusion underwent a platelet-rich plasma injection.

The injured knee was then secured at 90 flexion in a standard limited range-of-motion brace as early as convenient following injury (range: 031 days post-injury, median (IQR): 5 days (411 days)), by a physiotherapist trained in the CBP. Patients were advised to keep the knee fixed in the brace at all times for the first 4 weeks, including during sleep and showering. Patients were educated regarding safe use of crutches during the first 8 weeks (and use of additional mobility aids if desired, such as a knee scooter, i-Walker or wheelchair), while unable to sufficiently extend the knee to walk unaided. Patients were given advice regarding self-care, comfort and DVT risk mitigation strategies, including hydration and calf pump exercises. Prophylactic Clexane injections were introduced after the 10th patient. From patient 20 onwards, Rivaroxiban 10 mg was prescribed (for the first 8 weeks of the CBP) instead of Clexane.

After 4 weeks, the range-of-motion brace was adjusted at regular increments to allow progressive increases in range-of-motion (see online supplemental appendix 1). At week 10, unrestricted range-of-motion was allowed, and the brace was removed at 12 weeks. Weight-bearing was encouraged within the available range and patients completed standardised goal-oriented exercise-based rehabilitation while in the brace, and after brace removal until the point of return-to-sport (online supplemental appendix 1). Patients had weekly physiotherapist consults to check/adjust the brace and progress exercise-based rehabilitation (online supplemental appendix 1). Return-to-sport was not recommended until 912 months post-injury, and was dependent on patient and clinical factors, including desire to return-to-sport, completion of required rehabilitation and passing functional return-to-sport criteria.15

The first 4 patients had the brace removed and their first follow-up assessment at 9 weeks. After this, a decision was made to extend the CBP from 9 weeks to 12 weeks to protect the ACL for longer and enable more accurate interpretation of the MRI at the time of brace removal. Additionally, two patients for personal reasons (work demands/to care for young children) removed the brace at the end of week 4 and week 6. Both patients were compliant with the CBP before brace removal and completed rehabilitation post brace removal.

Details of all outcomes, including measurement method, time of measurement and interpretation, are presented in table 1.

Outcome measurement and interpretation

Patients have been involved in the development and refinement of the CBP, and the authorship team includes a patient who was managed with the CBP (MD).

All patients with acute ACL rupture managed with the CBP prior to October 2021 participated in the study, including 31 (39%) females, people aged 1058 years at the time of injury, and both private (69%) and publicly (31%) funded patients. Although only 3 women are included in the authorship team, the lead researcher is a woman and we include authors from a variety of career stages and clinical disciplines.

All continuous variables were assessed for normality and mean (SD) or median (IQR) reported, as appropriate. Participant characteristics and outcomes are presented for all participants, and based on 3-month ACLOAS. Mann-Whitney U tests were used to compare continuous outcomes (Lysholm Scale and ACLQOL scores) and Pearsons 2 tests were used to compare categorical outcomes (Lachmans test, Pivot-shift test and return-to-sport) between groups with lower versus higher ACLOAS grades on 3-month MRI (ACLOAS grades 01 vs ACLOAS grades 23). Since seven participants completed the Lysholm Scale and ACLQOL score after ACL re-rupture, a subgroup analysis was performed to present data and compare groups after excluding these seven individuals from the analysis (online supplemental appendix 2). For the two participants with missing MRI data at 3-month follow-up (decided not to undergo MRI), the ACLOAS from 6-month MRI was used to classify 3-month ACLOAS for analysis (95% of participants had the same ACLOAS at 3 months and 6 months, only 1 participant had a worse ACLOAS grade due to re-injury). Six people were missing 6-month MRI data (due to ACL re-rupture (n=3), pregnancy (n=1) or decided not to undergo MRI (n=2)). Since only one participant had missing data for the Lysholm Scale and ACLQOL scores, a complete case analysis was performed.

All individuals managed with the CBP provided consent for their data to be included in this study. Participants were aged a mean (SD) 26 (10) years at injury, 39% were female and 49% had concomitant meniscal injury (38 stable vertical tears in posterior horn of medial and/or lateral meniscus and 1 displaced medial meniscus ramp lesion). Participant characteristics are reported in table 2 for all participants and by ACLOAS grade on 3-month MRI (grade 1 vs grades 23). Participant characteristics are presented separately for participants with an ACLOAS grades 2 and 3, in online supplemental appendix 3.

Participant characteristics

At 3-month follow-up, n=72 (90%) had a continuous ACL (n=40 (50%) ACLOAS grade 1, n=32 (40%) ACLOAS grade 2). Of the 8 patients with ACLOAS grade 3 on 3-month MRI, 6 ACLs had attached to the lateral wall (n=3) or lateral wall and posterior cruciate ligament (n=3). Between 3-month and 6-month MRI, 4 participants changed from ACLOAS grade 1 to grade 0 and 1 participant changed from ACLOAS grade 2 to grade 3 due to subsequent knee injury. ACLOAS grades from 3-month and 6-month MRIs (complete case analysis) are presented in online supplemental appendix 4. Other participants sustained the same ACLOAS grade at 3 months and 6 months. MRI examples of ACL healing for five participants are presented in figure 2.

MRI images demonstrating MRI evidence of ACL healing for five participants. ACL, anterior cruciate ligament; ACLOAS, Anterior Cruciate Ligament OsteoArthritis Score.

Participants with an ACLOAS grade 1 on 3 month MRI reported better patient-reported outcomes on the Lysholm Scale and ACLQOL (including all ACLQOL subscales) compared with participants with an ACLOAS grades 23 (table 3). Online supplemental appendix 2 presents outcomes separately for participants with an ACLOAS grades 2 and 3. Online supplemental appendix 5 depicts participant scores based on time post-injury and 3-month healing status. Participants with an ACLOAS grade 1 had reduced knee laxity and a higher proportion returned to pre-injury sport (92% vs 62%) compared with participants with an ACLOAS grades 23 (table 3).

Participant outcomes

Eleven (14%) participants re-injured their ACL (mean (SD): 10 months (4 months), range: 518 months), 4 had ACLOAS grade 1, and 7 had ACLOAS grade 2 on 3-month MRI. After re-injury, 9 of 11 participants underwent ACLR (mean: 2 months after re-injury, range: 06 months), 1 participant decided to undergo the CBP again (resulting in evidence of ACL healing on MRI, ACLOAS grade 1). Mechanisms of re-injury included AFL/rugby (n=3), basketball (n=1), skiing (n=1), cycling accident (n=1), netball (n=1), Oz-tag (n=1), wrestling (n=1), dancing (n=1) and climbing (n=1). The four participants who re-injured their ACL despite ACLOAS grade 1 on 3-month MRI, did so during high-speed skiing/cycling accidents (5 months and 18 months post-injury), rugby (contact injury 10 months post-injury) and AFL (contact injury 17 months post-injury).

Two of 80 (2.5%) participants underwent an arthroscopic knee surgery, one participant for cyclops lesion removal 7 months post-injury and another underwent a partial lateral meniscectomy 5 months post-injury. Thirty eight of 39 (97%) meniscal tears were asymptomatic following the CBP, including one displaced medial meniscus ramp lesion.

Two patients were diagnosed with a below knee DVT (before DVT prophylaxis was added to the protocol), which were identified in the second week of the CBP, and were successfully managed with therapeutic dosing of Clexane. Both patients completed the CBP. Follow-up Doppler ultrasound demonstrated complete resolution of the DVT for both patients. Thereafter (11th patient onwards), DVT risk mitigation strategies were deployed as described in the Methods section.

Most patients reported mild and transient discomfort while adapting to the brace during the first week, often citing an awkward or uncomfortable sleeping position with the knee fixed at 90. This discomfort resolved for all patients without intervention. No patients opted to exit the programme due to discomfort or complication. At the time of unrestricted knee flexion in the brace, a flexion contracture (typically 515) was observed in 11 patients (14%). This resolved in all patients with physiotherapy exercises within 3 weeks. Contralateral lower limb overuse injuries, including pes anserine bursitis (n=1), insertional hamstrings tendinopathy (n=1) and patellofemoral pain (n=3), were observed at the time of brace removal.

This case series found that 72 out of 80 (90%) people with acute ACL rupture who were managed with a novel bracing protocol involving immobilisation of the knee at 90 flexion, had evidence of ACL healing (a continuous ACL) on 3-month MRI. An ACLOAS grade 1 on 3-month MRI was associated with better 12-month knee function and QOL, reduced passive knee laxity and a higher rate of return-to-sport, compared with an ACLOAS grades 23.

A recent analysis of the KANON trial found that 16 of 54 (30%) participants randomised to initial rehabilitation and optional delayed ACLR had signs of ACL healing on a 2-year MRI.9 Of the 30 participants who were managed with rehabilitation alone, 53% had MRI evidence of ACL healing at 2 years.9 In comparison, applying the same criteria we observed ACL healing in 72 of 80 (90%) participants on 3-month MRI. Six of 8 ACLs with discontinuous fibres had attached to the lateral wallposterior cruciate ligament. Although we graded these as discontinuous, it is possible that attachment to these structures could provide some function/stability, and it is not clear how this compares to the function/stability of an ACL graft. The high rate of healing observed on 3-month MRI suggests that the CBP could be conducive to ACL healing. To explore this potential, further research, including mechanistic studies, is required. Interestingly, patients had a range of concomitant injuries at baseline which became asymptomatic after the CBP. Only 1 of 39 patients with concomitant meniscal injuries had persistent symptoms after the CBP and underwent meniscal surgery. It is possible that the CBP could be beneficial for healing of concomitant injuries, this warrants further research.

Additionally, 37 of 40 participants (93%) with an ACLOAS grades 23 on 3-month MRI had an ACL rupture with a partial femoral avulsion, compared with only 7 (18%) participants with an ACLOAS grade 1. Although outside the scope of this study, it is possible that characteristics of ACL rupture observed on acute MRI (including partial/complete femoral avulsion, the displacement of ACL tissues outside of the intercondylar notch and gap distance between the ruptured ACL stumps) are associated with the likelihood of ACL healing. Further studies are needed to explore this possibility, with potential to inform ACL injury management decisions.

The favourable outcomes observed in patients with signs of ACL healing in our study are supported by findings from the KANON trial. In the KANON trial, participants with an ACLOAS of 02 on a 2-year MRI reported better knee function and QOL compared with participants with ACL discontinuity, and people who had delayed or early ACLR.9 Notably, only 8 (10%) patients in our study had ACL discontinuity on a 3-month MRI, and we used a different cut-off when comparing outcomes between groups. Collectively, results from the KANON trial and CBP suggest there may be a spectrum of ACL healing, whereby a more normal MRI appearance of the ACL may be associated with favourable patient outcomes.

Surprisingly, patients with an ACLOAS grades 23 reported excellent Lysholm Scale scores on average, even though scores were lower than patients with an ACLOAS grade 1. A Lysholm median score of 98 reported by people with an ACLOAS grade 1 is better than mean scores reported 2471 months after ACLR using autograft (mean scores range from 85 to 95).16 In contrast, the difference in ACLQOL scores was large between people with lower and higher grades of ACL healing. People with an ACLOAS grade 1 reported a median (IQR) ACLQOL score of 89 (7696). In comparison, people managed with ACLR or rehabilitation alone, report mean ACLQOL scores in the range of 5076, across a variety of time-points after ACL injury.17 The ACLQOL scores reported by people with an ACLOAS grades 23 may be more comparable with ACLQOL scores reported after ACLR and management with rehabilitation alone.17 Examining the ACLQOL domain scores suggests the greatest differences were within the recreational and sport participation, lifestyle and social emotional domains, with smaller differences observed in the symptoms and physical complaints and work-related concern domains.

The lower proportion who returned to sport with an ACLOAS grades 23 (64%) compared with an ACLOAS grade 1 (92%) could partly explain the lower QOL in this group considering return-to-sport is a key determinant of QOL after ACL injury.18 These return-to-sport rates are high compared with studies in ACL reconstructed individuals, where a pooled average of 55% of non-professional athletes returned to sport after ACLR.10 It should also be noted that patients in our study were aware of the degree of healing observed on MRI. It is possible that patients who received feedback that they had a suboptimal healing result on MRI had lower knee confidence and negative mental impacts compared with patients who received more positive feedback. This could also contribute to lower QOL scores in these domains.

It is not known whether a continuous ACL observed on MRI reflects restoration of pre-injury ACL function. Although, the high self-reported knee function and return-to-sport rate in people with an ACLOAS grade 1 suggests a positive association with knee function. It is important to note that 11 patients (14%) had re-ruptured their ACL at the time of follow-up. The 4 patients who re-injured their ACL despite an ACLOAS grade 1 on a 3-month MRI did so during competitive sport (rugby/AFL contact injuries) or high-speed skiing/cycling accidents. It is not clear whether re-injury of the ACL is a reflection of reduced tensile strength of the ACL fibres, considering the mechanisms of re-injury were similar to the mechanisms of initial ACL rupture. The rate of re-injury observed in our study may be comparable with re-injury rates following ACLR, whereby approximately 1-in-5 young athletes suffer a rupture of the ACL graft or contralateral ACL, and around 90% of these injuries occur after return to high-risk sports.19 Interestingly, one patient elected to undergo the CBP after re-rupturing their ACL and achieved an ACLOAS grade 1 on a 3-month MRI and returned to sport (rugby), after re-completion of the CBP. Longer-term follow-up is required to gain greater understanding of survivorship of the healed ACL and the risk of subsequent knee injury following management with the CBP.

In view of the promising outcomes of this case series and the potential advantages of preserving the native ACL following injury, further research in this area is warranted. Prognostic studies are needed to determine whether certain presentations are less likely to heal when managed with the CBP. In the future, the potential for the ACL to heal may be an important consideration when deciding on surgical or non-surgical management. We found that signs of ACL healing were apparent on MRI as early as 3 months after ACL rupture. Three months is the typical duration that people trial initial rehabilitation before considering delayed ACLR. It is possible that 3-month MRI findings could identify patients who would benefit from ACLR. Only 10% of patients with an ACLOAS grade 1 at 3 months had progressed to an ACLOAS grade 0 at 6 months. Further research is needed to understand the timeline and stages of ACL healing. Additionally, clinical trials are needed to compare outcomes for patients managed with the CBP, compared with those who are managed with ACLR or rehabilitation alone. Of particular importance will be the investigation of re-injury rates, return-to-sport, patient-reported outcomes, functional stability and the prevalence of knee osteoarthritis.

Bridge enhanced ACL repair (BEAR) is a surgical technique that augments repair of the ligament with a scaffold implant (a resorbable protein-based implant containing autologous blood) positioned between the torn ends of a midsubstance ACL tear in attempt to facilitate healing.20 An interesting area for future research is the comparison of ACL healing on MRI and clinical outcomes following management with BEAR compared with the CBP. Additionally, there may be specific ACL rupture presentations that benefit from early surgical intervention to assist with facilitating ACL healing. For example, acute ACL injuries with a large gap distance between torn remnants and acute ACL injuries with displaced ACL tissue outside the intercondylar notch might benefit from surgery to reduce and realign the ACL tissues. There could also be a role for bracing in knee flexion postoperatively to protect the repair and/or reduction of the injured ACL tissues (akin to management of displaced bone fractures with open reduction internal fixation followed by a period of postoperative immobilisation). Further research is needed to explore this potential.

This was a pragmatic study, whereby data were collected in the course of clinical practice rather than in a research setting. For this reason, some adaptations were made to the CBP overtime. Our study design did not allow comparison of outcomes with people managed with ACLR or rehabilitation alone. There is also potential for selection bias. After the 50th patient was braced, patients were discouraged from undertaking the CBP if they had a femoral avulsion and/or ACL tissue displaced outside the boundaries of the intercondylar notch. Although 14 of 29 patients underwent the CBP despite this advice, 15 of these individuals chose ACLR. The overall proportion with an ACLOAS grade 1 at 3 months may have been lower if these 15 individuals had not received this advice and underwent the CBP. Tests of passive knee laxity were performed by two unblinded physicians and may be subject to detection and observer bias, a clinical trial, including blinding of examiners, is needed. Further studies may also benefit from using a knee arthrometer to collect more objective measures of knee laxity. MRIs were graded by three radiologists who were aware that patients had undertaken the CBP. Although this case series highlights the potential for positive outcomes using the CBP, larger cohorts with longer-term follow-up and, in particular, randomised clinical trials are needed.

Read the original:
Healing of acute anterior cruciate ligament rupture on MRI and ... - British Journal of Sports Medicine