Study design and setting

We will conduct a single-site, prospective, single blind, randomised controlled trial with two parallel treatment arms. The study will be conducted in the Queensland Children’s Hospital (QCH), the only paediatric focused quaternary hospital covering a population of 5 million people. We will offer recruitment to patients enrolled in the Australian Paediatric ACL Injury Registry at the QCH site. Inclusion will start in July 2021 and is expected to finalise in July 2023, which will allow for read-out of the primary endpoint in July 2024. The proposed flow of patients thought the trial is displayed in figure 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Experimental design flowchart according to theConsolidated Standards Of Reporting Trials (CONSORT) statement guidelines. ACL, anterior cruciate ligament; ACLR, ACL reconstruction.

Recruitment strategy

All consecutive patients between the age of 10 and 18 who present to the orthopaedic outpatient’s clinic at the QCH with an isolated ACL±meniscal injury will be provided with a recruitment package during their outpatient appointment. A follow-up phone call will be made to ensure that potential participants received the study information.

Inclusion and exclusion criteria

Potential participants will be excluded if they have:

1. Concomitant posterior cruciate ligament injury.

2. Collateral ligament instabilities of grade I or greater (2–5 mm).

3. Bilateral ACL deficiency.

4. Combined knee surgery with high tibial osteotomy or medial patellofemoral ligament.

5. Evidence of early knee osteoarthritis on MRI.

6. Previous knee surgery on the affected side.

7. Chronic musculoskeletal conditions.

8. BMI>35.

9. Surgeon recommendation for an extraphyseal technique for ACLR.

Study procedure

Potential participants will first be screened against the inclusion criteria in the orthopaedic outpatient clinic at the QCH. Following baseline testing (T1 MRI), eligible participants will be randomised into one of the two surgical groups. The surgeons performing the surgeries will be well versed in both ACLR techniques and will be advised of the patient’s randomised surgical technique just prior to the surgery and will proceed accordingly.

Patient and public involvement

Patients were not involved

• In the development of the research question.

• In the decision of the outcome measures.

• In the study design.

• In the recruitment to and conduct of the study.

• In how the results will be disseminated.

• In patient burden assessment.

Surgical techniques

Two senior surgeons will perform all the randomised ACLR surgeries as detailed below. Both have been allowed a surgical learning curve of 20 cases (minimum) before the start of the trial. So that there is no variability among equipment used, Smith & Nephew (S&N) will be used for the two techniques: single hamstring graft technique will use UltraButton+round ExtendoButton; double hamstring graft technique will use Endobutton+Biosure Regenesorb screw.35 36 Patients who exhibit a pivot-shift of grade 2 or greater following ACLR will be considered for an additional lateral-tenodesis.37

Tendon harvest and graft preparation

The semitendinosus hamstring tendon alone will be harvested for the single graft technique and the semitendinosus +gracilis hamstring tendons will be harvested for the double graft technique. A tendon stripper will be used to release the tendons from their proximal attachment via a standard push technique to an appropriate length.

Graft preparation for the single hamstring technique will involve tripling the single semitendinosus tendon over two S&N UltraButton loops (with a round ExtendoButton being attached to the tibial UltraButton). Graft preparation for the double hamstring technique will involve doubling the semitendinosus+gracilis tendons over one S&N Endobutton loop. For both methods, the ends of the tendon complex are sutured via a whip-stitch technique using a #1 Ethibond suture on a J-needle. The prepared tendon complex will be set aside and soaked in a Vancomycin laden Raytec sponge and placed into the soft tissue tunnel where the hamstring tendons were harvested.

Once prepared, the graft is measured for both length (to ensure that the graft will adequately span the knee joint and sufficiently pass through both tunnels) and diameter (to determine the required femoral and tibial tunnel diameters for reaming).

Knee arthroscopy

The knee arthroscopy is performed using standard lateral and medial portal sites. All compartments of the knee are identified and examined for any additional pathology (chondral injuries, meniscal tears, etc), the notch is identified, and femoral and tibial attachments are identified. The pathological ACL is identified, and its stump or scarred remnant is debrided sufficiently. As much of the tibial stump/footprint is left to allow for optimal tibial tunnel placement and to aid in revascularisation of the graft;38 39 however, enough is debrided away, so that it does not interfere with passing of the graft or cause impingement. The lateral wall of the notch is cleared with either a shaver or via radiofrequency ablation and prepped with a curette. The fat pad is debrided to allow for adequate visualisation. If any meniscal pathology is identified, it is addressed prior to the ACLR.

Tunnel drilling

The femoral tunnel is drilled first. Once the notch has been cleared, the optimal position is determined—this is usually 30° lateral from the roof of the notch (11 o’clock in the right knee, 1 o’clock in the left knee) at the point 5 mm anterior to the posterior cortex of the notch.36 The patient’s knee is positioned into full flexion (normally 120°) to allow for adequate placement of the guide pin at the desired position. The pin is drilled through the femur and pushed through the soft tissues and skin of the anterolateral thigh. The desired position is achieved using an Acufex offset drill guide and is reviewed on scope. In skeletally mature patients, the pin should be directed approximately 30° anteriorly and 30° laterally, with respect to the femoral long axis, however, in the younger patient with open physes the pin is directed as vertical as possible to ensure that the physis and subsequent growth is disrupted as little as possible.40 With the knee still in full flexion, the femoral tunnel is drilled first with a 4.5 mm drill through the lateral femoral cortex, this tunnel is measured to determine tunnel length, then the preliminary tunnel is reamed to desired diameter as determined by graft measurement.

The tibial tunnel is positioned using an S&N ACUFEXTM drill guide. The tip of the guide is passed through the anteromedial portal and positioned to create a tunnel that enters the joint through the posteromedial portion of the ACL footprint.13 41 In the skeletally immature patient, the length and angle of the tunnel may be more variable as the surgeon will target the drill tunnel perpendicular to the growth plate to minimise the likelihood of growth disturbance. A 2.4 mm drill-tipped guide wire is used to establish the line of the tibial tunnel using the jig. Using the positioned guidewire, this tunnel is reamed to desired diameter as determined by graft measurement earlier. The tunnel is again measured.

Both femoral and tibial tunnels are smoothed cleared of any debris using a shaver to allow for smooth graft passage in the subsequent steps. The graft is marked at either end using a sterile marking pen factoring in tunnel lengths and desired graft length spanning the knee joint. A nylon suture is threaded into the slot of a passing pin. The loop end of the suture is held externally, the passing pin is drawn into the joint and out through the femur, drawing the free ends of the suture out through the lateral thigh.35 36

Single tendon, single bundle, graft passing and fixation

The nylon suture is pulled through the tibial tunnel, so that the loop end is exiting out of the tibial tunnel. The graft is orientated and the Ethibond sutures are placed through the nylon loop. The Ethibond sutures are passed through the tibial and femoral tunnels and out of the skin of the anterolateral thigh. The arthroscope is placed into the knee joint and the Ultrabutton loop and graft are pulled into the femoral tunnel under vision to the line drawn on the graft from previous measurements. The Ultrabutton is toggled and flipped on the femoral side. The graft is held taught and the Femoral Ultraloop is reduced pulling the graft into the femoral tunnel to its desired length (~20 mm). The knee is then brought out to 30–40^o flexion, and the tibial Ultraloop is reduced to pull the graft into the tibial tunnel. Final tensioning of the graft is done with knee extended to 0^o flexion the ExtendoButton is secured down to the tibia.35

Double tendon, single bundle, graft passing and fixation

As above, however, in this case, an Endobutton loop is used as opposed to the Ultrabutton loop. The graft is held taught with the Ethibond suture tails on the tibial side of the graft and, the knee is cycled through flexion/extension a number of times and the knee is then brought out to 20–30^o flexion, the assistant is asked to secure the foot as well as place anterior directed force the distal femur. A wire is inserted behind the graft in the tibial tunnel and a Biosure Regenesorb screw is fully advanced and secured to fix the graft into the tibial tunnel.36

Postoperative care and rehabilitation

The knee is placed in a Richard splint in full extension. In-patient (oral analgesia with option for intramuscular opioid injection) and discharge medication (oral analgesia only) doses as per individual patient requirements. The patient is seen by the physiotherapist day 1 postoperatively for assessment and treatment as per the Children’s Health Queensland ACL Post-operative Rehabilitation Guidelines (online supplemental file 1).

In cases where meniscal repair was concomitantly performed weight-bearing and range of motion (ROM) restrictions are surgeon dependent based on extent and location of repair, but the majority will be non-weight bearing and have ROM restricted for 6 weeks postoperatively. These patients will all be discharged in an ROM brace with increasing ROM allowed over the 6-week period as per surgeon advice. Following 6 weeks, the brace is removed and patients can complete ROM and full weight bearing.

All patients will undergo rehabilitation as per the Children’s Health Queensland ACL Reconstruction Rehabilitation Guidelines for Physiotherapists will be advised to completely abstain from full return to sports for at least 12 months and prior to return, they will be required to meet phase 5 criteria of the Rehabilitation Protocol (online supplemental file 1).

Standard medical imaging protocol

All patients enrolled into the randomised controlled trial obtain the following standardised imaging in accordance with current clinical practices at the QCH.

Preoperative medical imaging:

• Plain X-rays of the affected knee, in both coronal and sagittal projections

• Bilateral anterior–posterior lower limb weight-bearing X-rays to assess leg length and coronal plane alignment (ie, mechanical axis deviation, lateral distal femoral angle, medial proximal tibial angle).

• Wrist or elbow anterior–posterior and lateral X-rays for assessment of patient’s bone age.

• MRI to assess chondral status, menisci status, ACL or graft morphology, growth plate status.

Day 1 postoperative medical imaging:

• Plain X-rays of the affected knee, in both coronal and sagittal projections

1, 2 and 5-year postoperative medical imaging:

• Plain X-rays of the affected knee, in both coronal and sagittal projections

• Bilateral anterior–posterior lower limb weight-bearing X-rays to assess leg length and coronal plane alignment (ie, mechanical axis deviation, lateral distal femoral angle, medial proximal tibial angle).

• MRI to assess chondral status, menisci status, ACL or graft morphology, growth plate disturbance, position of tunnels and hardware.

Study outcome measures

Sample size determination

Our sample size calculations are based on a three-level outcome variable comparing anterior-posterior tibiofemoral laxity in the reconstructed and unaffected knee (‘low‘ side-to-side difference (<3 mm), ‘moderate’ side-to-side difference (3 to 5 mm) or ‘severe’ side-to-side difference (>5 mm or ruptured) as differences of this magnitude were considered to be clinically important.44 50 Pilot data from our hospitals’ perspective ACL injury registry classified 74% as low, 7% as moderate and 19% as severe or rupture, and we assume that these percentages will hold in our single hamstring tendon graft ACLR group. We expect that in the experimental arm, the equivalent probabilities will be 93%, 2% and 5%. This is equivalent to specifying a proportional OR of 0.07. With alpha=0.05 and power=80%, we are required to record outcome data on 43 participants in each group to detect a between-group difference of this size or greater. To increase the power of the study and allow the maximal tolerated level of dropout at T2, 100 patients will be randomised.

Randomisation

Participants will be randomly allocated by the Griffith University randomisation service to the single hamstring tendon graft ACLR group or the double hamstring tendon graft ACLR intervention group. Participants will be stratified according to skeletal maturity and sex to minimise confounding bias and a randomly varied block size will be used to ensure participants are more evenly allocated throughout the entire trial. The Griffith University randomisation service will conceal group allocation from the study investigators until the participant has been enrolled and baseline data have been collected.

Blinding

To minimise ascertainment-bias, this trial is single-blinded, where patients are blinded to surgical technique. External entry points for surgery are equivalent for both surgical techniques and, therefore, patients will not be able to guess group allocation. Due to the nature of the intervention, the treating surgeons cannot be blinded to group allocation. Furthermore, it is not feasible to blind the postoperative management team as surgical notes will be reviewed prior to follow-up appointments.

Data management

The percentage of eligible participants successfully recruited, and numbers of eligible who choose not to participate will be recorded along with their age and sex. Participant retention will be recorded throughout the trial period. Paper documents and files will be deidentified, labelled with a participant identification code and stored in a locked filing cabinet. Consent forms and demographic information will be kept separately, also in a locked filing cabinet. The list of patient identification codes and all other electronic data will be stored securely through a Research Electronic Data Capture database (https://www.project-redcap.org/software/) on a secured network accessible only to the registered members of study team.

Statistical analysis plan

Standard principles for RCTs will be followed, and primary analyses will be conducted using between-group comparisons on all participants on an intention-to-treat basis. There is a small risk that the surgeon may deem a harvested hamstring tendon inadequate for a single tendon graft (see safety consideration below) and, therefore, need to break randomisation. To account for this potential, a secondary, per-protocol analysis will be used to assess the effect of treatment received and models will be adjusted for potentially confounding variables if necessary. The primary timepoint will be after 1 year, and the primary comparison will be the quantitative anterior–posterior laxity results between the two surgical techniques. Effect of surgical technique will be assessed using ordinal logistic regression with technique (single/double) included as the main effect. To determine between-group differences at 2 and 5 years postsurgery, we will employ mixed effects ordinal logistic regression models with patient included as a random effect, and time (1, 2 and 5 years) and surgical technique included as main effects and a time-by-technique interaction included as fixed effects. For continuous outcomes, comparison will be by linear regression models. Where continuous data do not meet linearity assumptions, as assessed by inspection of boxplots and the Shapiro-Wilk test, it will be assessed using non-parametric methods such as median regression. For dichotomous outcomes, comparison will be by logistic regression models. For count outcomes, comparison will be by Poisson regression models. Significance will be accepted at p<0.05. Sensitivity analyses to assess the effect of missing data will be undertaken on an outcome-by-outcome basis using MAR (multiple imputation) or NMAR (using pattern-mixture models) according to the pattern of the missing data.

Cost-utility analysis

Utility values will be obtained from the EQ-50 (with a 1-week recall) at baseline (T1), and 1-year, 2-year and 5-year postsurgery and will be transformed into Quality of life-adjusted years with means and variances. The health economic evaluation will be determined using the incremental cost-utility ratio. Resource utilisation will be determined from hospital finance reports related to the initial stay in hospital and using a patient (or parent) administered case report form that documents information on patient and caregiver demographics, educational and employment information, used of health resources (ie, visits to general practitioners, physiotherapists, emergency department, patient expenses related to medication and out-of-pocket transportation coasts to receive additional medical care, patient or caregiver days off work) using a health resource utilisation questionnaire. Furthermore, the following health service utilisation data will be collected:

• Details of hospital admissions, outpatient episodes and emergency department presentations including episode, clinical, demographic and costing information (such as diagnosis, procedures, length of stay, cost of encounter, etc) for the duration of the study and 12 months prior to consenting—from routinely collected hospital and emergency department administrative data.

• Medicare Benefits Schedule claim details, costs and service provider information and Pharmaceutical Benefits Scheme item description, costs and prescribing details—from Services Australia.

Safety, adverse events and complications

Surgical treatment and clinical follow-up will be conducted in accordance with current clinical practices for the participating surgeons at the QCH. Intraoperatively, if a harvested hamstring tendon is deemed to be inadequate (ie,<6 mm in diameter and/or of poor quality) the decision may need to be made to convert from the single to the double hamstring graft technique. Presence of meniscus injuries, defined by the necessity to repair or partially resect tissue due to meniscus instability, will be determined during ACLR. Any adverse events occurring during preoperative and follow-up testing sessions will be recorded on an Excel spreadsheet. Minor adverse events are classified as muscle soreness, muscular fatigue or mild injuries that do not require medical attention. Major adverse events are conditions that require medical attention, such as a fracture, equipment failure or infection and would likely result in the child discontinuing the testing session. All adverse events, regardless of their severity, will be documented and reported to a senior study advisor and if serious, escalated to the ethics committee with information reported to the child’s treating physician as necessary. Risk assessments, including strategies to minimise adverse events, will be completed prior to participation in the testing session. For all onsite testing, participants will be directly supervised by an investigator who is trained to deliver first aid and CPR. Postoperative complications (ie, thrombosis, infection, rupture) will be recorded. A data monitoring committee will convene every 6 months to monitor patient safety and treatment efficacy during the surgical stage of the trial. The DSMB membership comprising a Chair, Medical Monitor and Secretary, will be independent to the trial and will not participate as investigators of the trial or have any financial, scientific or other conflict of interest with the trial.

Ethics and dissemination

Full ethical approval for this study has been obtained by the Children’s Health Queensland Human Research Ethics committee (HREC/21/QCHQ/73043, Protocol V.2.1 29072021). The study will be conducted in agreement with the Helsinki declaration. Written and informed parent/guardian consent will be obtained prior to study enrolment by the study investigator. Verbal assent will be obtained from children under the age of 12 years and written assent will be obtained from children who are 12 years and older. This trial is registered with the Australian New Zealand Clinical Trials Registry and the study protocol is reported according to the Standard Protocol Items: Recommendations for Intervention Trials statement51 (see Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) checklist online supplemental file 2). Changes to the study protocol will be communicated to the ethics committee and updated on the trial registry. The primary study results will be submitted for publication to an international, peer-reviewed journal and disseminated to participants and healthcare professionals via newsletters and conference presentations.