Repair versus reconstruction for proximal anterior cruciate ligament tears: a study protocol for a prospective multicenter randomized controlled trial

Background For active patients with a tear of the anterior cruciate ligament (ACL) who would like to return to active level of sports, the current surgical gold standard is reconstruction of the ACL. Recently, there has been renewed interest in repairing the ACL in selected patients with a proximally torn ligament. Repair of the ligament has (potential) advantages over reconstruction of the ligament such as decreased surgical morbidity, faster return of range of motion, and potentially decreased awareness of the knee. Studies comparing both treatments in a prospective randomized method are currently lacking. Methods This study is a multicenter prospective block randomized controlled trial. A total of 74 patients with acute proximal isolated ACL tears will be assigned in a 1:1 allocation ratio to either (I) ACL repair using cortical button fixation and additional suture augmentation or (II) ACL reconstruction using an all-inside autologous hamstring graft technique. The primary objective is to assess if ACL repair is non-inferior to ACL reconstruction regarding the subjective International Knee Documentation Committee (IKDC) score at two-years postoperatively. The secondary objectives are to assess if ACL repair is non-inferior with regards to (I) other patient-reported outcomes measures (i.e. Knee Injury and Osteoarthritis Outcome Score, Lysholm score, Forgotten Joint Score, patient satisfaction and pain), (II) objective outcome measures (i.e. failure of repair or graft defined as rerupture or symptomatic instability, reoperation, contralateral injury, and stability using the objective IKDC score and Rollimeter/KT-2000), (III) return to sports assessed by Tegner activity score and the ACL-Return to Sports Index at two-year follow-up, and (IV) long-term osteoarthritis at 10-year follow-up. Discussion Over the last decade there has been a resurgence of interest in repair of proximally torn ACLs. Several cohort studies have shown encouraging short-term and mid-term results using these techniques, but prospective randomized studies are lacking. Therefore, this randomized controlled trial has been designed to assess whether ACL repair is at least equivalent to the current gold standard of ACL reconstruction in both subjective and objective outcome scores. Trial registration Registered at Netherlands Trial Register (NL9072) on 25th of November 2020.


Background
Historical overview of ACL repair The first documented surgical treatment of an anterior cruciate ligament (ACL) injury consisted of open repair in 1895 when Mayo Robson repaired a proximally avulsed ACL and posterior cruciate ligament back to the femur in a 41-year old male with good outcomes at sixyear follow-up [1]. In the twentieth century, Ivar Palmar [2,3] and Don O'Donoghue [4,5] reported on open primary repair as a treatment of ACL injuries, and in the early 1970s open primary repair became a popular treatment for ACL injuries [6][7][8][9].
Feagin and Curl were the first to present the outcomes of open repair in 1972 and noted good outcomes at short-term follow-up [8]. A few years later in 1976, however, they noted a deterioration of outcomes at midterm follow-up in their cohort [10]. Similarly, several other surgeons and researchers noted good short-term [11][12][13][14][15][16] but disappointing mid-term outcomes [17][18][19][20][21]. With these disappointing results and the promising outcomes of ACL reconstruction, several (randomized) prospective studies were started in the 1980s comparing open ACL repair with open ACL reconstruction [19,[22][23][24]. These prospective studies noted more reliable outcomes with ACL reconstruction when compared to ACL repair, which ultimately led to an abandonment of open ACL repair and to the current gold standard of ACL reconstruction for all patients [9].
In 1991, Sherman et al. were the first analyzing the disappointing mid-term outcomes of open ACL repair by performing an extensive subgroup analysis [21]. The authors found that a trend towards better outcomes in patients with proximal avulsion type tears and good tissue quality when compared to patients with midsubstance tears and/or tears with poor tissue quality. Unfortunately, the inclusion of the aforementioned prospective trials was already completed before the study by Sherman et al. was published, and thus the prospective trials contained all tear types including patients that might not have been ideal candidates for ACL repair (i.e., those with midsubstance tears or tears with poor tissue quality).
When critically reviewing the historical literature, and bearing in mind these findings by Sherman [25]. These findings indicate that ACL repair may have been prematurely abandoned for all tear types and perhaps may be a good treatment option for patients with proximal tears. Furthermore, outcomes of ACL repair can be expected to improve when benefiting from modern development, such as arthroscopy (instead of open repair) and modern rehabilitation (instead of casting and immobilization).

Rationale for ACL repair
The rationale behind better outcomes of ACL repair of proximal tears compared to midsubstance tears is that better vascularity is present at the proximal end of the ligament [26] and, as a result, proximal tears have healing potential for reattachment that is similar to medial collateral ligament (MCL) tears [27]. The reason for the continued pursuit of repair as a treatment of ACL injuries can also be explained by the potential advantages of repair over reconstruction. With ACL repair, the native tissue can be preserved along with proprioception which may provide patients with a more normal feeling of the knee compared to ACL reconstruction [28,29]. Also, ACL repair is a less invasive surgery when compared to ACL reconstruction as no (or only small) tunnels need to be drilled and no graft tissues need to be harvested, leading to lower surgical morbidity [30][31][32][33], faster return of range of motion and fewer complications [34]. Furthermore, in case of failure of both treatments, revision surgery following primary repair is expected to be similar to primary reconstruction (no or only small tunnels have been drilled or grafts harvested), whereas revision of reconstruction surgery can be complicated by tunnel malpositioning or widening and pre-existing hardware and is associated with inferior outcomes compared to primary ACL reconstruction [35][36][37].

Recent literature on ACL repair
With the recognized relevance of tear location in ACL repair and the potential advantages of this treatment, several surgeons and researchers have pursued the concept of ACL repair of proximal tears [38][39][40][41][42][43][44][45][46][47]. Most of these studies were retrospective small case series reporting good short-term outcomes with an overall reported failure rates of 6 to 9%, reoperation rates of 0 to 4% and patient-reported outcome measures (PROMs) > 85% of the maximum score [48]. Three studies have also shown that the good outcomes are maintained at mid-term follow-up [44,45,49]. One prospective study has compared the outcomes of repair (n = 20) versus reconstruction (n = 20) in patients with proximal tears and reported similar outcomes regarding functional outcomes, failure rates and laxity examination [46]. However, no randomized studies or studies with sufficient number of patients to assess differences between the treatments have been performed, and a recent systematic review also concluded higher-level evidence studies for ACL repair are currently lacking [48]. Recent studies have also suggested that primary repair with suture augmentation results in lower failure rates when compared to primary repair without suture augmentation [42,48].
The current surgical gold standard of treating ACL injuries is ACL reconstruction using autograft tissue of either hamstring tendons, patellar tendon or quadriceps tendon. As for all new surgical techniques, the outcomes of arthroscopic ACL repair need to be compared to the current gold standard in order to assess whether this treatment can be used for standard patient care. Therefore, a randomized controlled trial (RCT) comparing ACL repair with ACL reconstruction is needed. The ACL study group of the Dutch Arthroscopy Association also recently declared that "the application of ACL repair could be considered in a medial ethical committeeapproved study until there is high-grade and long-term evidence regarding the efficacy of modern-day ACL repair."

Goal and hypotheses
The goal of this multicenter non-inferior prospective randomized controlled trial is therefore to compare the outcomes of arthroscopic ACL repair with suture augmentation to ACL reconstruction for patients with proximal tears in a 1:1 allocation ratio. The primary outcome is the subjective International Knee Documentation Committee (IKDC) score and the secondary outcomes are other patient-reported outcomes, objective outcomes and return to sports. It is hypothesized that patients following ACL repair with suture augmentation have non-inferior primary and secondary outcomes when compared to ACL reconstruction due to the less invasive surgery.

Methods
This study and manuscript have been designed in accordance to the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) guidelines.

Study design
This study is a multicenter prospective RCT with randomization into two treatment arms: (I) arthroscopic ACL repair with suture augmentation and (II) arthroscopic ACL reconstruction surgery. This study is a noninferiority study with the hypothesis that arthroscopic ACL repair is non-inferior to (equivalent or better than) arthroscopic ACL reconstruction. All patients with proximal tears will be randomized during the operation into one of these treatment arms and will be followed up to 10-years postoperatively.

Study sample
Potential candidates will be selected from five participating orthopaedic surgery departments, of which one is an academic hospital, three are teaching hospitals and one is a private hospital. Inclusion and exclusion criteria for participation in the study are displayed in Table 1. In general, potential inclusion involves all patients with acute, isolated, complete, proximal ACL tears that have a desire to return to pre-injury activities and exclusion involves all concomitant ligamentous and osteoarthritic injuries and skeletally immature patients. A flowchart of the study is shown in Fig. 1. Patients can withdraw their participation in this study at any time point, at which their data will be deleted.

Randomization
All patients will be consented preoperatively for the study. Patients are taken into the operating room, general or epidural anesthesia is induced, and the leg is prepped and draped for standard arthroscopic knee surgery with a tourniquet high at the upper thigh. Then standard anteromedial and anterolateral portals are created, and the knee is assessed for cartilage, meniscus and ligamentous injuries. After cartilage and meniscus injuries are addressed, the tear type of the ACL and eligibility for this study is assessed. First, it should be confirmed whether a proximal tear is present (i.e., whether the distal remnant of the ACL is of sufficient length to be reattached to the anatomical femoral footprint of the ACL) and whether sufficient tissue quality is present (i.e., whether the ligament remnant is of sufficient quality to withhold suture passage and can be tensioned towards the femur).  If these conditions are present, patients are randomized between both treatment arms, and if these conditions are not present, the patient is excluded, and standard ACL reconstruction will be performed. A computer block randomization of 10 patients per block will be done digitally prior to the study, and the allocation concealment is performed by sequentially numbered, opaque, sealed envelopes containing the name of the procedure in a randomized order. The envelopes are placed in the operating room and opened when the surgeon deems the ACL tear eligible for the study. A participant timeline is shown in Fig. 2.

Surgical techniques
Prior to the start of the trial, a cadaver session will be held in order to standardize the technique of ACL repair and ACL reconstruction for all surgeons and to minimize the learning curve. All surgeons have extensive experience with ACL reconstruction and two out of five participating centers have experience with ACL repair.
The surgical technique of arthroscopic ACL repair has been more extensively described in the literature [39,43,55]. In brief, first the native torn ACL will be sutured with a loop using FiberWire sutures and advanced with one to two passes, so that the sutures exit the avulsed ligament towards the femur. Then, a small tunnel will be drilled from the native femoral insertion towards the lateral epicondyle using an ACL drill guide. The sutures will be passed through a TightRope button along with an additional FiberTape. The sutures and TightRope will be passed through the femoral tunnel and the button will be flipped. Then, a small tunnel will be drilled through the tibia from the anteromedial cortex towards the anterior part of the tibial footprint, and the Fiber-Tape will be channeled through the tibial tunnel and, after cycling the knee, the FiberTape is fixed into the anteromedial cortex using a suture anchor at full extension. Finally, the repair sutures will be tensioned and tied in order to reapproximate the ACL towards the femoral footprint at 90°flexion.
For ACL reconstruction, a standard all-inside autograft hamstring tendon anatomic reconstruction technique is used [56,57]. First, autologous hamstrings (semitendinosus and gracilis tendon) are harvested to the preference of the surgeon and will be prepared for graft usage with a minimum graft diameter of 8 mm [58,59]. Then, femoral and tibial sockets are independently drilled in retrograde fashion using a FlipCutter drill. The graft is placed into the sockets, the knee is cycled in order to achieve optimal tension of the graft, and the graft is then fixed at the femoral and tibial side using a cortical button.

Rehabilitation
Both treatment arms undergo the same rehabilitation program and consists of a milestone-based program according to the Dutch national guidelines for rehabilitation following ACL reconstruction and consists of three phases [60][61][62]. The first phase focuses on controlling swelling, restoration of range of motion and return of quadriceps muscle control, and generally takes 4 to 8 weeks. The second phase focuses on resuming light sporting activities and work without symptoms, and phase three focuses on full return to sports activities and heavy work. In case of meniscus repair, the first 6 weeks patients are partial weight bearing, range of motion is restricted to 0-90°and patients are not allowed deep bending or squatting for 4 months. Although the rehabilitation is milestone based and no strict time goals can be set, generally cycling on a stationary bike is allowed at 4-6 weeks, running at 10-12 weeks and return to sports and pivoting activities at a minimum of 9 months postoperatively.
Blinding Blinding for patients is not possible due to different scars, different postoperative radiographs and practical reasons. However, the data analysis will be performed in blinded fashion.
Primary outcomes/endpoint (Table 2) The primary outcome of this non-inferiority RCT is the subjective patient reported outcome (PROM) at twoyear follow-up consisting of the subjective IKDC score [63] (Dutch validation [64]), as to a recent RCT on a similar topic [65,66]. The primary endpoint is the subjective IKDC at two-years postoperatively. Patients will ultimately be followed for 10 years.

Secondary outcomes (Table 2)
The secondary outcomes of this RCT are fourfold and consist of (I) other subjective outcomes, (II) objective outcomes, (III) return to sports, and (IV) long-term osteoarthritis.
The objective outcomes consist of failure of ACL repair/graft, reoperation, contralateral injury, and laxity. Failure is defined as a (traumatic) rerupture or symptomatic instability with activities. Reoperation is defined as any new operation on the same knee for any other reason than revision (e.g., symptomatic meniscus tear, hardware irritation, infection or stiffness/arthrofibrosis). Contralateral injury was defined as a complete ACL rupture of the contralateral ACL. Stability is defined as the laxity found with physical examination using the IKDC objective score form [72], which includes the Lachman, anterior drawer and pivot shift test, and side-to-side differences is assessed using KT-2000 or Rollimeter.
Return to sports is defined as (I) returning to sports, (II) returning to the same sport, and (III) returning to the preinjury level of sport. The preinjury and postoperative Tegner activity scale are also collected, which enables comparison with other studies [73] (Dutch validation [70]). Finally, confidence of return to sports and fear of reinjury are assessed using the ACL-Return to Sports Index (ACL-RSI) score [74] (Dutch validation [75]).
Osteoarthritis will be reviewed at 10-year follow-up. Radiographs of both knees will be performed, and the operated knee will be compared to (I) the contralateral knee if no operation occurred in that knee, and (II) the ipsilateral knee radiograph preoperatively. The Kellgren-Lawrence (KL) grade will be used to assess the incidence and grades of osteoarthritis.

Sample size
The sample size calculation was based on the primary outcome of this study (subjective IKDC score), similar to another RCT design on this topic [65]. It has been Table 2 This chart provides an overview of which outcomes are collected at the different follow-up visits Pre shown that a difference of 8.8 points in the subjective IKDC score is the minimal clinically important difference (MCID) [76]. Using this non-inferiority limit of 8.8 points, and a standard deviation of 11 points [42,65,77] along with a two-sided alpha of 0.05, a power of 90%, and a lost-to-follow-up rate of 10%, a total of 37 patients in each group (74 patients in total) are needed to assess the primary outcome of this non-inferiority RCT. This sample size is also sufficient for the MCID of KOOS [78] and Lysholm score [79]. Given the recent studies that showed that 30-40% of the acute tears will have repairable proximal ACL tears [50,80], we estimate that approximately 200 patients will be needed to be screened preoperatively to achieve the sample size of 74 patients [81].

Statistical analysis
Both an intention to treat analysis and per protocol analysis will be performed for this non-inferiority study. Comparison of nominal variables between ACL repair and ACL reconstruction will be performed using twoby-two tables with Pearson's Chi-square test or Fisher's exact test (in case one of the cells is < 5). For comparison of continuous variables, first tests for normal distribution of values are performed and independent t-tests are used of normal distributed values and nonparametric t-tests are used for not-normally distributed values.
A mixed model analysis for repeated measures will be performed to assess differences between both groups. Furthermore, a multivariate regression analysis will be performed for the primary endpoint of IKDC at twoyears follow-up in order to correct for potential confounders. Statistical analysis will be performed using SPSS version 25.0 (IBM Software, Armonk, NY, USA). All tests are two-sided and a p-value of < 0.05 is considered statistically significant.

Discussion
This study reports on the study design of the REPAIRtrial (Repair versus rEconstruction for Proximal Anterior cruciate lIgament teaRs). Few studies have examined the outcomes of repair versus reconstruction with favorable outcomes for ACL reconstruction [22][23][24]. However, these studies were performed over 30 years ago and are limited by the fact that all tear types were repaired rather than only proximal tears and that repair was performed using an arthrotomy [9,25,82]. Recently, four RCT studies have been designed to assess the outcomes of ACL repair [65,[83][84][85] but these are either performed in midsubstance tears [65,83], assess the outcomes of dynamic intraligamentary stabilization (DIS) versus ACL reconstruction [65,83], repair versus DIS [84] or Bridge-Enhanced ACL Repair (BEAR) with reconstruction [85]. Our current RCT differs from these studies as only proximal tears will be treated rather than all tear types and as the ligament will be reattached to the femoral footprint in a minimally invasive way.
The renewed interest of repair of proximal tears can be explained by improved understanding of patient selection. Research has shown that proximal tears have a better vascularity compared to midsubstance tears [26] and therefore have excellent healing capacity by reattachment to the femoral wall which is similar to the healing capacity of MCL tears [27]. Both historical studies on open ACL repair [9,25,82] and more recent studies on repair with DIS (also known as Ligamys) have shown that the clinical outcomes are indeed better when repairing proximal tears. Two studies have shown failure rates of repair with DIS in midsubstance tears of 24% in all patients and 36% in competitive athletes with midsubstance tears [86,87]. Our current study applies strict patient selection criteria of proximal tears and good tissue quality. As the length of distal remnant and possibility of repair can only be assessed intraoperatively, randomization in this study should perform during surgery after the surgeon has confirmed the possibility of repair. Consequently, patients will be consented that they might be excluded during surgery if a nonrepairable tear is present, and these patients will undergo standard ACL reconstruction.
It should be noted that there is also a potential disadvantage of ACL repair. By performing ACL surgery in the early phase (since early surgery prevents ligament retraction and preserves tissue quality that is both needed for repair [4,5,88]), it is likely that too many ACL surgeries will be performed. Current day standards recommend that patients following ACL injury will be treated conservatively first as approximately half of the patient may be copers and do not need surgical intervention [53,60,89]. By performing surgery on all ACL injured patients, patients will undergo surgery while they might be copers and do not need surgery. This risk is minimized in this study by only including patients aged 18 -50 and only patients that desire to return to sports. It would be best if it is known preoperatively which patients will not do well with conservative treatment and ultimately require ACL surgery, as this both increases the chance of performing ACL repair and as early reconstruction outcomes decreases the risk for meniscal and chondral damage [60] at longer follow-up when compared to delayed reconstruction.
Several studies have recently reported good short-term outcomes of arthroscopic ACL repair using different techniques: in some studies femoral fixation consisted of using two suture anchors [42,44], one suture anchor (for both bundles) [40,45,46] or transosseous tunnels with or without cortical button fixation [39,41,43,55,90], and some studies used ACL repair without [40,41,45,46] or with [39,43,55,90] additional suture augmentation. For this study, femoral fixation will consist of cortical button fixation with additional suture augmentation (FiberTape) in order to protect the repair in the early phases of rehabilitation, becuase it has been suggested that additional suture augmentation leads to lower rerupture rates [42,48].
This study has been designed to assess the outcomes following repair and reconstruction of proximal ACL tears. We hypothesize that the repair treatment is a good treatment for proximal tears as it has potential advantages over ACL reconstruction: the surgery is short and minimally invasive, it has a low complication rate, rehabilitation is easier, and in case ACL repair fails then primary reconstruction surgery can be performed. Noninferiority of arthroscopic ACL repair compared to arthroscopic ACL reconstruction may lead to a treatment algorithm in which patients with proximal avulsion tears can be repaired in the acute setting whereas patients with midsubstance tears will undergo ACL reconstruction in either the acute or delayed setting [91,92].

Funding
This study is funded by the Dutch Association of Arthroscopy (NVA) for which the authors are grateful and for which external peer review took place. This funding source had no role in the design of this study and will not have any role during its execution, analyses, interpretation of the data, or decision to submit results.

Availability of data and materials
Available at study publication.

Declarations
Ethics approval and consent to participate This study has been approved by the Medial Ethics Committee of the Amsterdam Medical Center on the 16th of February 2021 under the reference numbers 2020_322#B202163 and 2020_322#B202189 and approval is for all sites. The study has been registered at the Netherlands Trial Register (NL9072) on 25th of November 2020. All patients will be asked to give their written informed consent for this study and can stop participation of this study at any time point, at which their data will be deleted. The results of this study will be presented at national and international conferences and peer-reviewed journals in a de-identified manner.

Consent for publication
Not applicable.