Association of Lateral Extra-Articular Tenodesis With Graft Failure and Return to Sport in Pivoting Athletes: A Multicenter Comparative Study of Augmented and Isolated ACL Reconstructions

Abstract

Background:

Anterior cruciate ligament (ACL) rupture is common in high-risk pivoting sports, with rerupture threatening athletic careers. Lateral extra-articular tenodesis (LET) has been proposed to reduce graft failure; nonetheless, comparative multicenter data across graft types remain limited.

Purpose:

To compare graft failure, return to sport, and functional outcomes after primary anterior cruciate ligament reconstruction using bone-patellar tendon -bone or semitendinosus autografts, with or without lateral extra-articular tenodesis, in competitive pivoting athletes.

Study Design:

Cohort study; Level of evidence, 3.

Methods:

The authors conducted a multicenter, prospective, intention-to-treat cohort study of 1260 competitive football and rugby athletes who underwent primary ACL reconstruction between 2018 and 2019. Participants were nonrandomly allocated to 4 groups according to graft choice and LET use, as determined by 16 individual treating surgeons: bone-patellar tendon-bone (BPTB) with modified Lemaire, semitendinosus tendon autograft (STRI) with modified Lemaire, isolated BPTB, and isolated STRI. At 5 years, evaluable follow-up data were available for 1186 of the 1260 athletes (94.1%). The primary endpoint was clinically relevant symptomatic knee instability consistent with graft failure, confirmed by imaging or revision surgery, defined as symptomatic knee instability confirmed by imaging or revision surgery. Follow-up included in-person medical examination with clinical stability testing when feasible and structured patient-reported questionnaires assessing knee stability and functional outcomes.

Results:

Graft failure rates were similarly low in both augmented groups (0.8% for BPTB + Lemaire vs 0.9% for STRI + Lemaire groups), with no statistically significant difference between graft types and significantly lower than in nonaugmented reconstructions (P < .001). Return to sport (RTS) at the preinjury level at 18 months was achieved in 88.1% of athletes in the BPTB + Lemaire group and 82.7% in the STRI + Lemaire group, compared with 79.8% and 75.9% in isolated BPTB and STRI reconstruction groups, respectively (P < .001). Functional outcomes were similarly high in augmented groups, with mean International Knee Documentation Committee scores of 90.8 ± 6.8 for the BPTB + Lemaire group and 89.6 ± 7.2 for the STRI + Lemaire group, with no statistically significant differences between augmented graft types (P > .05).

Conclusion:

LET was associated with significantly reduced clinically relevant graft failure and reliable RTS in pivoting athletes, independent of graft type.

Anterior cruciate ligament (ACL) rupture remains one of the most devastating injuries encountered in high-demand pivoting sports, such as football and rugby. Despite decades of progress in surgical technique and rehabilitation, young athletes engaged in these disciplines continue to experience disproportionately high rates of graft failure and rerupture, particularly within the first few years after returning to competition. The mechanical demands of cutting, pivoting, and tackling expose reconstructed knees to repetitive rotational stress, which often exceeds the physiological thresholds of both autograft tissue and fixation methods. These challenges underscore the persistent need for optimized surgical strategies capable of ensuring both mechanical durability and reliable restoration of function under elite-level athletic conditions.^9,12,17,18

Graft selection has historically been at the center of this debate. Bone-patellar tendon-bone (BPTB) autografts have long been considered the benchmark for contact athletes, largely because of their predictable bone-to-bone healing, superior initial fixation strength, and resistance to elongation under cyclic loading. Clinical series and registry data have shown that BPTB grafts are associated with lower revision rates than soft-tissue grafts, especially in younger athletes. Nevertheless, this benefit comes at the expense of donor-site morbidity, including anterior knee pain, kneeling discomfort, and potential extensor mechanism complications, which can impair both performance and long-term joint health. In contrast, hamstring tendon autografts using the semitendinosus tendon (STRI) are widely favored for their reduced anterior knee symptoms and preservation of quadriceps function.^12,18,22

However, concerns remain about the biomechanical resilience of hamstring tendon grafts under high rotational demands. Tendon-to-bone healing is slower and biologically less robust than bone-to-bone incorporation, leaving hamstring tendon grafts vulnerable to tunnel widening, micromotion at the interface, and subsequent laxity. While some meta-analyses suggest equivalence between BPTB and hamstring tendon grafts in general populations, evidence in elite pivoting athletes is inconsistent, with several reports showing higher rerupture rates with hamstring-based reconstructions.

Beyond graft selection, attention has increasingly shifted to the potential role of anterolateral augmentation. The modified Lemaire tenodesis, along with other lateral extra-articular procedures, has reemerged as a critical adjunct in ACL surgery.^10,25-28

Biomechanical studies consistently demonstrate that these procedures reduce residual pivot shift, constrain excessive rotational laxity, and unload the intra-articular graft during pivoting maneuvers. Clinical reports have confirmed lower graft failure rates with Lemaire augmentation, particularly in high-risk groups, such as athletes <25 years, patients with high-grade pivot shift, or those returning to contact-pivoting sports. Despite this growing evidence, most available data derive from single-center studies with relatively short-term follow-up, and very few have compared the effectiveness of Lemaire augmentation across different autograft types.^1,16,20,22

This knowledge gap is particularly striking in football and rugby, sports with the highest global incidence of ACL injuries, recurrent ruptures, and career-threatening sequelae. While some studies suggest that combining intra-articular reconstruction with extra-articular reinforcement may significantly reduce graft failure, the lack of large-scale, multicenter data in professional or competitive pivoting athletes has limited the establishment of clear, evidence-based surgical hierarchies. Moreover, no prospective series has demonstrated whether rerupture rates can be consistently reduced to levels substantially lower than those typically reported in high-risk athletic populations at mid-term follow-up.^16,20,22

The present multicenter study was designed to directly compare 4 widely used strategies for ACL reconstruction (ACLR) in pivoting athletes: BPTB with modified Lemaire tenodesis, STRI with modified Lemaire tenodesis, isolated BPTB, and isolated STRI. We hypothesized that BPTB combined with modified Lemaire tenodesis would provide the most robust protection against graft failure and the highest rates of return to preinjury competitive sport, with rerupture rates <1% at 5 years. Additionally, we expected STRI combined with Lemaire augmentation to represent the second-best option, yielding superior outcomes compared with STRI alone but not equaling the durability of BPTB-based reconstructions. By systematically evaluating these 4 surgical strategies in football and rugby athletes, our goal was to clarify the comparative effectiveness of graft-augmentation combinations and to provide actionable evidence for optimizing surgical decision-making in one of the most demanding athletic populations.

Methods

Study Design and Setting

We conducted a multicenter, prospective cohort study with a planned minimum follow-up of 5 years. The study was conducted at 3 high-volume sports medicine centers in South America: Clínica Arthros (Quito, Ecuador), Hospital Universitario Austral (Buenos Aires, Argentina), and Centro de Rehabilitación y Traumatología Deportiva (Bogotá, Colombia). All participating institutions are tertiary referral centers for elite athletes, with subspecialty-trained orthopaedic surgeons dedicated to knee surgery. A common study protocol was prospectively registered and approved by the local ethics committees of each institution. Written informed consent was obtained from all participants before enrollment.

Eligibility Criteria

Athletes were eligible if they (1) sustained a primary complete ACL rupture confirmed by clinical examination and magnetic resonance imaging (MRI), (2) were aged between 18 and 35 years, (3) participated at a competitive or professional level in pivoting sports (eg, football or rugby), and (4) underwent ACLR within 6 months of injury.

The mechanism of injury (contact vs noncontact) was recorded at baseline based on reports from the athlete and medical staff. However, this variable was not incorporated into the primary comparative models because the proportion of clearly documented contact versus noncontact injuries varied between centers and did not allow reliable stratified analysis across all treatment groups.

The exclusion criteria were as follows: (1) revision ACLR; (2) concomitant multiligamentous injuries requiring staged or combined reconstructions (posterior cruciate ligament, posterolateral corner, medial complex); (3) previous ipsilateral osteotomy or major realignment procedure; (4) grade 4 chondral defects requiring cartilage restoration surgery; and (5) systemic or neuromuscular conditions limiting rehabilitation adherence.

Preoperative pivot-shift grading was assessed clinically in all athletes as part of routine evaluation; however, standardized quantitative grading was not uniformly recorded across centers and was therefore not included in the baseline comparative analysis.

Sample and Allocation

Patients were prospectively enrolled and stratified into 4 treatment groups according to graft type and the use of lateral extra-articular augmentation:

• Group A: BPTB + modified Lemaire tenodesis.

• Group B: STRI + modified Lemaire tenodesis.

• Group C: Isolated BPTB autograft.

• Group D: Isolated STRI autograft.

The choice of graft and the use of augmentation were determined by the treating surgeon, in consultation with the athlete, in accordance with institutional protocols.

Treatment allocation was not randomized and was determined by the treating surgeon based on clinical judgment. Preoperative pivot-shift and Lachman testing were performed in all athletes as part of routine clinical evaluation. However, standardized grading of pivot-shift severity (eg, grade 2 vs grade 3) was not uniformly recorded across centers and therefore could not be incorporated into subgroup analysis. Surgeons frequently consider increased rotational laxity when deciding whether to perform lateral extra-articular augmentation, potentially introducing selection bias.

Although preoperative pivot-shift and Lachman tests were performed in all athletes as part of routine evaluation, standardized grading (eg, pivot-shift grade 2 vs 3) was not uniformly recorded across centers. It therefore could not be incorporated into formal stratification or subgroup analysis.

This introduces potential selection bias, particularly if athletes with higher-grade rotational instability were preferentially selected for lateral extra-articular augmentation. Indications for lateral extra-articular augmentation were based on commonly accepted high-risk features, including participation in pivoting contact sports (eg, football or rugby), anticipated return to high-level competition, young age, and clinical evidence of increased rotational laxity on examination. These indications included the surgeon's assessment of increased rotational laxity on clinical examination, frequently identified through a high-grade pivot-shift test. However, because pivot-shift grading was not uniformly documented in all centers, it could not be included as a baseline comparative variable between treatment groups.

Final decision-making remained at the treating surgeon's discretion, reflecting real-world practice across centers.

Although treatment allocation was not randomized, the decision to perform lateral extra-articular augmentation was guided by predefined high-risk criteria shared across participating centers. These criteria included participation in pivoting contact sports (eg, football or rugby), age ≤25 years, high-grade pivot shift on clinical examination, anticipated early return to competition, and surgeon assessment of rotational laxity. While individual surgeon judgment played a role, these criteria were prospectively defined to minimize selection bias.

Surgical Technique

All procedures were performed arthroscopically by fellowship-trained knee surgeons using a standardized single-bundle anatomic technique. Femoral tunnels were created through an anteromedial portal, aiming for placement within the native ACL footprint behind the resident's ridge. Tibial tunnels were drilled under fluoroscopic and arthroscopic guidance to ensure optimal sagittal and coronal orientation.

In the BPTB groups, a central third of the patellar tendon with bone plugs (20-25 mm) was harvested, shaped, and fixed with interference screws (titanium or bioabsorbable). Patellar and tibial harvest-site bone defects were managed according to institutional protocols, typically involving bone grafting with autologous cancellous bone from tunnel reaming or local bone fragments to promote healing and minimize postoperative kneeling discomfort.

In BPTB reconstructions, patellar and tibial bone defects resulting from graft harvest were routinely managed according to institutional protocols, including autologous cancellous bone grafting from tunnel reaming or local bone chips, when deemed necessary by the surgeon, to promote defect healing and minimize anterior knee pain and kneeling discomfort.

In the STRI groups, the STRI was harvested and quadrupled to a minimum diameter of 8 mm, then fixed with a cortical suspensory device at the femur and an interference screw at the tibia.

For athletes assigned to extra-articular augmentation, a modified Lemaire tenodesis was performed. A strip of iliotibial band approximately 8 cm in length was harvested proximally and passed deep to the lateral collateral ligament. The graft was fixed to the femur just proximal and posterior to the lateral epicondyle with an interference screw, at 30° of flexion and neutral rotation, avoiding overconstraint.

Meniscal tears were addressed concomitantly with repair whenever feasible. Routine notchplasty was not mandated by the study protocol. Notchplasty, including A-shaped or anteromedial bundle-directed decompression, was selectively performed at the surgeon's discretion if graft impingement during extension was suspected intraoperatively.

Rehabilitation Protocol

All centers adhered to a criterion-based, 5-phase rehabilitation program supervised by specialized physical therapists.

• Phase 1 (0-2 weeks): Swelling and pain control, passive range of motion (ROM) 0° to 90°, partial weightbearing with crutches.

• Phase 2 (3-6 weeks): Full ROM recovery, progressive quadriceps and hamstring strengthening, proprioception drills.

• Phase 3 (7-12 weeks): Closed-chain strengthening, treadmill running, dynamic balance exercises.

• Phase 4 (13-20 weeks): Plyometric training, agility drills, and sport-specific activities. Progression to plyometric and high-demand athletic drills was criterion-based rather than strictly time-based and was permitted only for athletes demonstrating adequate neuromuscular control and strength recovery, according to the physical therapist's assessment.

• Phase 5 (>6 months): Return-to-sport (RTS) clearance was based on a combination of time from surgery and functional criteria, including absence of pain or effusion, full ROM, and satisfactory performance on strength and functional tests. Although a limb symmetry index ≥90% on strength and functional testing was targeted, not all athletes reached this threshold at the same time point, and progression through the rehabilitation phases was individualized.

Outcome Measures

The primary endpoint was graft failure, defined as clinically relevant symptomatic instability confirmed by clinical examination and MRI evidence of graft rupture or revision ACL surgery.

At the 5-year follow-up, knee stability was systematically assessed in all athletes with evaluable follow-up. Athletes attending in-person visits underwent standardized clinical examination, including Lachman and pivot-shift testing, performed by fellowship-trained knee surgeons (H.R., S.Z., P.R., G.A.). For athletes unable to attend in-person evaluation, structured patient-reported outcome questionnaires assessed perceived knee stability, episodes of giving way, functional performance, and RTS status. MRI evaluation was reserved for cases with suspected graft compromise or clinically relevant instability. MRI was obtained in athletes presenting with instability when revision surgery was not immediately indicated, whereas in athletes undergoing revision surgery based on clinical findings, MRI confirmation was not mandatory.

The study was powered based on assumptions related to graft failure rates. No separate a priori power calculation was performed for RTS outcomes, and analyses of RTS should therefore be interpreted as exploratory.

Symptomatic instability was defined as patient-reported episodes of giving way, functional deterioration, or inability to perform sporting activities, corroborated by clinical examination when available. At 5 years, knee stability assessment was available for 912 athletes through in-person examination and for an additional 274 through structured patient-reported evaluation. Graft failure was not inferred solely from surgical revision but from clinically meaningful instability affecting function.

Graft failure rates were calculated based on athletes with evaluable follow-up within each treatment group (per-protocol analysis). Failure was defined as graft rupture or revision ACLR confirmed clinically or surgically during follow-up. Sensitivity analyses were performed to evaluate the potential impact of loss to follow-up.

At final follow-up, patients were evaluated through a combination of in-person clinical examination and standardized patient-reported outcome questionnaires. Clinical assessment included stability testing (Lachman and pivot shift). In patients presenting with recurrent instability or symptoms suggestive of graft failure, an MRI was performed to evaluate graft integrity.

Graft failure was defined as (1) symptomatic instability confirmed clinically, (2) revision ACLR, or (3) MRI evidence of graft rupture. Patients with no instability symptoms or reoperation were considered clinically stable at follow-up.

Graft failure rates were calculated using an intention-to-treat framework, with the original number of enrolled athletes in each group as the denominator, regardless of follow-up modality.

The coprimary endpoint was RTS at the same or higher competitive level compared with preinjury, assessed between 12 and 24 months after surgery and confirmed at 5 years.

Secondary outcomes were as follows:

• International Knee Documentation Committee (IKDC) subjective score

• Tegner Activity Scale

• Knee injury and Osteoarthritis Outcome Score (KOOS)

• Donor-site morbidity (anterior knee pain, hamstring weakness)

• Surgical complications (arthrofibrosis, infection, or hardware failure).

Assessments were performed at baseline and at 6, 12, 24, 36, 48, and 60 months by blinded independent evaluators (F.E., M.M.).

At the 5-year follow-up, outcome assessment was available for 1186 athletes (94.1%), including clinical examination with stability testing for 912 athletes and structured questionnaires for 274 athletes. Questionnaire-based follow-up included assessment of perceived stability, RTS status, and functional outcomes. Graft failure rates were calculated among athletes with evaluable follow-up within each treatment group (per-protocol analysis). Failure was defined as graft rupture or revision ACLR confirmed clinically or surgically during follow-up. Although overall follow-up was high, graft status could not be confirmed in all athletes, and unrecognized failures among those lost to follow-up cannot be entirely excluded.

At the 5-year follow-up, knee stability was evaluated using a standardized 2-tier assessment strategy.

In athletes attending in-person follow-up visits, clinical stability was evaluated through standardized physical examination, including Lachman and pivot-shift testing performed by experienced knee surgeons (P.R., G.A., H.R., M.F., C.C.). For athletes unable to attend in-person evaluation, structured patient-reported outcome measures were used to assess perceived stability, episodes of giving way, RTS status, and functional performance. Objective imaging (MRI) was reserved for cases with symptomatic instability or clinical suspicion of graft failure. Routine instrumented laxity testing or MRI screening was not performed in asymptomatic athletes.

Sample Size and Power Calculation

Sample size was calculated to detect a reduction in 5-year rerupture rates from 8% in the STRI group to <1% in the STRI + Lemaire group, with α = 0.05 and β = 0.20 (80% power). This required 260 athletes per arm. For the BPTB comparison (5.5% vs 0.8%), 330 athletes per arm were required. Accounting for 15% attrition and intercenter clustering, the final target enrollment was 1260 athletes across groups A (n = 360), B (n = 300), C (n = 300), and D (n = 300).

Sample size calculations were performed assuming a conservative 15% attrition rate over the follow-up period, resulting in a target enrollment of approximately 330 patients per group for the BPTB comparison. Final group sizes of 300 patients reflected pragmatic enrollment limits and balanced recruitment across centers. Importantly, the observed 5-year attrition rate was substantially lower than anticipated (5.9%), preserving adequate statistical power for the primary endpoint. No additional adjustment for intercenter clustering was applied, as recruitment was evenly distributed and the center was included as a covariate in adjusted analyses.

Statistical Analysis

Continuous variables were summarized as means with standard deviations or medians with interquartile ranges (IQR), and categorical variables as frequencies and percentages. Between-group comparisons were made using analysis of variance (ANOVA) or Kruskal-Wallis tests for continuous variables and the chi-square or the Fisher exact tests for categorical variables.

Survivorship free from graft failure was analyzed using Kaplan-Meier curves and the log-rank test. Multivariable Cox proportional hazards models were used to adjust for confounders (age, sport type, meniscal status, and center), with robust standard errors to account for clustering at the site level. For continuous outcomes compared across the 4 groups, ANOVA was performed, followed by post hoc pairwise comparisons using the Tukey honest significant difference test with adjustment for multiple comparisons.

RTS was analyzed using logistic regression for binary outcomes and Cox regression for time-to-event analysis. Propensity score matching and inverse probability of treatment weighting (IPTW) were performed to compare augmented versus nonaugmented reconstructions within each graft type. These methods were specifically applied to mitigate potential selection bias related to nonrandomized treatment allocation.

Predefined subgroup analyses evaluated outcomes by sport (football vs rugby) and age (<20 vs ≥20 years). Sensitivity analyses were performed by restricting the analysis to athletes with complete follow-up data and by excluding patients lost to follow-up before 24 months. Additional competing risk models were conducted to account for contralateral ACL rupture as a competing event.

All analyses were conducted using R Version 4.3.2 (R Foundation for Statistical Computing). A 2-tailed P < .05 was considered statistically significant.

IPTW was used to reduce confounding related to nonrandom treatment allocation by creating a weighted pseudopopulation in which baseline covariates were balanced between groups. Propensity scores were calculated using age, sport type, meniscal status, and treatment center. Missing data were minimal (<3%) and handled through complete-case analysis. IPTW methodology followed previously published recommendations. Weights were derived from propensity scores estimated using age, sport type, meniscal status, and treating center. Missing data for these covariates were minimal (2.3%) and were handled using complete-case analysis. Sensitivity analyses demonstrated that the direction and magnitude of the main findings remained unchanged after weighting. IPTW methodology was applied in accordance with established recommendations.

Results

Between January 2018 and December 2019, a total of 1456 athletes with primary ACL rupture were assessed for eligibility across the 3 participating centers. After applying the inclusion and exclusion criteria, 1260 athletes were enrolled in the study. They were subsequently allocated to 4 groups according to graft type and augmentation strategy: 360 patients underwent BPTB reconstruction + modified Lemaire tenodesis; 300 underwent STRI reconstruction + Lemaire augmentation; 300 received isolated BPTB autografts; and 300 underwent isolated STRI reconstruction. At the 5-year follow-up, 1186 athletes (94.1% of the cohort) completed at least 1 follow-up evaluation, either through in-person clinical assessment or structured patient-reported questionnaires. Follow-up rates were balanced across all 4 treatment arms, with no significant differences in loss to follow-up. The proportion of athletes with evaluable follow-up data was similar across the 4 treatment groups, with approximately 94% of athletes in each group contributing follow-up information (Table 1). No statistically significant differences in follow-up rates were observed between groups.

Table 1.

Follow-up Completeness and Mode of Assessment at 5-Year Follow-up ^a

Group	Total Enrolled, N	Completed Follow-up	In-Person Clinical Examination	PROMs Only	Lost to Follow-up
BPTB + Lemaire	360	339 (94.2)	262 (72.8)	77 (21.4)	21 (5.8)
STRI + Lemaire	300	282 (94)	216 (72)	66 (22)	18 (6)
BPTB	300	282 (94)	219 (73)	63 (21)	18 (6)
STRI	300	283 (94.3)	215 (71.7)	68 (22.6)	17 (5.7)
Total	126	1186 (94.1)	912 (72.4)	274 (21.7)	74 (5.9)

^aData are presented as n (%). BPTB, bone-patellar tendon-bone; PROMs, patient-reported outcome measures; STRI, semitendinosus tendon autograft.

Baseline patient and surgical characteristics were comparable between groups. Because standardized pivot-shift grading was not uniformly documented across centers, the distribution of baseline rotational instability could not be formally compared between groups. Although the mechanism of injury (contact vs noncontact) was recorded during baseline clinical assessment, incomplete documentation in a subset of athletes prevented reliable comparative analysis between treatment groups.

The overall mean age of the population was 22.6 ± 4.8 years, and nearly 4 out of 5 participants were men (79.3%). Football was the most common sport represented (72.4%), while rugby accounted for 27.6% of cases. Meniscal injury was identified in 44.8% of athletes, with a similar proportion across treatment groups. As expected, mean graft diameter differed between techniques: 9.3 ± 0.6 mm in BPTB reconstructions versus 8.4 ± 0.5 mm in STRI grafts. The full distribution of baseline characteristics is presented in Table 2.

Table 2.

Baseline Characteristics of the Study Population ^a

Variable	BPTB + Lemaire, n = 360	STRI + Lemaire, n = 300	BPTB, n = 300	STRI, n = 300	P
Age, years	22.5 ± 4.7	22.9 ± 4.9	22.4 ± 4.6	22.8 ± 5	.64
Male sex	287 (79.7)	235 (78.3)	241 (80.3)	237 (79)	.91
Football	262 (72.8)	215 (71.7)	219 (73)	212 (70.7)	.88
Rugby	98 (27.2)	85 (28.3)	81 (27)	88 (29.3)
Meniscal injury	159 (44.2)	139 (46.3)	131 (43.7)	135 (45)	.92
Graft diameter, mm	9.3 ± 0.5	8.4 ± 0.4	9.2 ± 0.6	8.3 ± 0.5	<.001 b

^aData are presented as mean ± SD or n (%). BPTB, bone-patellar tendon-bone; STRI, semitendinosus tendon autograft.

^bDifference expected by graft type.

At the 5-year follow-up, evaluable outcome data were available for 339 of 360 athletes (94.2%) in the BPTB + Lemaire group and 282 of 300 athletes (94%) in the STRI + Lemaire group. Comparable follow-up completeness was observed in the isolated BPTB and STRI groups. Follow-up included in-person clinical examination in approximately 72% of cases and structured patient-reported outcome assessment in the remaining athletes. All outcome analyses were conducted using the original enrolled cohort as the denominator within an intention-to-treat framework. In contrast, outcome ascertainment was based on the 1186 athletes (94.1%) with evaluable 5-year follow-up data. Among these athletes, 912 (72.4%) underwent direct clinical examination, including Lachman and pivot-shift testing, while 274 (21.7%) completed structured patient-reported assessments of stability and function.

Within a predefined subcohort of 330 knees selected for detailed stability assessment, 5-year follow-up data were available for 278 knees (84.2%). Among these athletes, 162 underwent in-person clinical examination, while 116 completed standardized follow-up questionnaires assessing knee stability and functional status.

Within this evaluated subgroup, 21 of the 278 athletes with evaluable follow-up presented with symptomatic instability, prompting further clinical evaluation and MRI assessment. Confirmed graft failure was identified in 18 knees, corresponding to a graft failure rate of 6.5% among athletes with evaluable follow-up in this subgroup.

At 5 years, evaluable follow-up data were available for 1186 of the 1260 enrolled athletes (94.1%). Of these 1186 athletes, 912 (76.9%) underwent in-person clinical evaluation, including physical examination with assessment of knee stability, whereas 274 (23.1%) completed structured patient-reported outcome assessments evaluating perceived stability, episodes of giving way, functional status, and RTS participation. The distribution of follow-up modalities was similar across the 4 treatment groups.

Importantly, stability assessment was not limited to symptomatic knees returning for treatment. Athletes with evaluable follow-up underwent systematic stability evaluation either through clinical examination or structured patient-reported assessment.

At 5 years, significant differences in graft survival emerged between groups. The cumulative rerupture rate was 0.8% in the BPTB + Lemaire group (3 failures among 360 enrolled athletes) and 0.9% in the STRI + Lemaire group (3 failures among 300 enrolled athletes). These rates were calculated using the original cohort size as the denominator, in accordance with the intention-to-treat framework. In contrast, isolated BPTB reconstructions demonstrated a rerupture rate of 5.5% (16/300), while isolated STRI reconstructions exhibited the highest rate at 8.2% (24/300). Kaplan-Meier survival analysis confirmed superior graft survivorship in the augmented groups compared with nonaugmented reconstructions (P < .001, log-rank test), with both augmented cohorts clustering closely together and well above the survival probability of their isolated counterparts (Figure 1). At the 5-year follow-up, evaluable outcome data were available for 1186 of the 1260 enrolled athletes (94.1%). Graft failure rates were calculated using the original cohort size as the denominator within an intention-to-treat framework, and failure ascertainment was based on clinical examination or structured patient-reported follow-up. Undetected failures among athletes lost to follow-up cannot be entirely excluded.

Figure 1.

Kaplan-Meier survival curves by group. BPTB, bone-patellar tendon-bone; STRI, semitendinosus tendon autograft.

Exploratory subgroup analyses based on available injury mechanism data suggested that the protective effect of lateral extra-articular augmentation was present in both contact and noncontact injury patterns. However, these analyses should be interpreted with caution due to incomplete documentation of the injury mechanism in some athletes and the lack of a significant interaction between injury mechanism and graft failure (P > .05).

Multivariable Cox regression analysis further validated the protective effect of lateral augmentation. This analysis included 1186 athletes (94.1% of the original cohort) with evaluable 5-year follow-up, comprising 912 who underwent in-person clinical examination and 274 who completed structured patient-reported follow-up assessments. After adjustment for age, sport type, and meniscal status, Lemaire augmentation was associated with an 85% reduction in the risk of graft failure (hazard ratio [HR], 0.15 [95% CI, 0.07-0.31]; P < .001). Neither graft type (STRI vs BPTB), patient age (<20 years), sport (rugby vs football), nor meniscal repair status reached statistical significance in the adjusted model (Table 2). The multivariable Cox regression analysis presented in Table 3 was based on 1186 athletes with evaluable 5-year follow-up data (94.1% of the original cohort), including both athletes who underwent in-person clinical examination and those assessed through structured patient-reported outcome questionnaires.

Table 3.

Multivariable Cox Regression for Graft Survival (5 Years)

Variable	HR	95% CI	P
STRI vs BPTB	1.42	0.78-2.60	.25
Lemaire augmentation, yes vs no	0.15	0.07-0.31	<.001
Age <20 years	1.27	0.73-2.19	.40
Rugby vs football	1.33	0.84-2.12	.22
Meniscal repair vs intact	0.82	0.51-1.32	.41

HRs represent the relative hazard of graft failure. Values >1 indicate an increased risk of graft failure, whereas values <1 indicate a reduced risk compared with the reference category.

Thus, an HR <1 for Lemaire augmentation indicates a protective effect against graft failure, whereas an HR >1 indicates increased risk relative to the reference group. BPTB, bone-patellar tendon-bone; HR, hazard ratio; STRI, semitendinosus tendon autograft.

In multivariable Cox regression analysis, lateral extra-articular augmentation emerged as the only independent factor significantly associated with improved graft survivorship. In contrast, graft type, age, sport, and meniscal repair were not significant predictors.

RTS analysis demonstrated similarly compelling findings. At 18 months, 88.1% of athletes in the BPTB + Lemaire group returned to their preinjury competitive level, compared with 82.7% in the STRI + Lemaire, 79.8% in isolated BPTB, and 75.9% in isolated STRI groups (P < .001). The median time to return was shortest in the BPTB + Lemaire group at 8 months (IQR, 7.2-8.8 months), whereas the isolated STRI reconstruction group required the longest recovery period, with a median of 9 months (IQR, 8.1-9.8 months). Both BPTB + Lemaire and STRI + Lemaire reconstruction groups demonstrated superior graft survival, RTS rates, and functional outcomes compared with the isolated reconstruction group. Direct pairwise comparison between the 2 augmented groups (BPTB + Lemaire vs STRI + Lemaire) using post hoc pairwise testing after ANOVA with the Tukey adjustment for multiple comparisons did not demonstrate statistically significant differences in graft failure rates or functional outcomes (P > .05 for both comparisons). Differences between the 2 augmented strategies were minimal and not statistically significant, indicating that lateral extra-articular augmentation is the dominant factor influencing outcomes in this cohort. Logistic regression, adjusted for sport, age, and meniscal treatment, confirmed the superiority of augmentation, with the BPTB + Lemaire group showing nearly a 2-fold higher likelihood of successful return compared with the isolated STRI reconstruction group (odds ratio [OR], 1.73 [95% CI, 1.40-2.13]).

Both augmented groups achieved higher RTS rates compared with isolated reconstruction groups. Although the BPTB + Lemaire group showed numerically higher RTS rates than the STRI + Lemaire group, this difference was not statistically significant, indicating that the benefit in RTS is primarily driven by lateral augmentation rather than graft type.

Lateral extra-articular augmentation was associated with a substantial reduction in graft failure. Rerupture rates were similarly low in both augmented groups, indicating that the protective effect is primarily attributable to the addition of the Lemaire procedure rather than to the choice of intra-articular graft.

RTS outcomes presented in Figure 2 reflect initial return to preinjury level of competition within the first 24 months after surgery. These analyses capture the time to first successful return and do not distinguish between sustained and transient returns.

Figure 2.

Cumulative RTS curves by group. BPTB, bone tendon bone autograft; RTS, return to sport; STRI, hamstring tendon autograft.

At the 5-year follow-up, RTS status was reassessed through clinical evaluation or structured questionnaires. While the primary time-to-event analysis focused on initial return within 24 months, most athletes who returned to sport remained active at follow-up. However, the present study was not specifically designed to quantify sustained versus transient return over time, and this distinction should be interpreted with caution (Figure 3).

Figure 3.

Forest plot of adjusted HR (graft survival) and OR (RTS).

The plot displays effect estimates with 95% CIs. ORs are shown for RTS outcomes, and HRs are shown for graft survival (risk of graft failure). The vertical dashed line represents the null value (effect size = 1). For graft survival analyses, HR values <1 indicate a reduced risk of graft failure relative to the reference category, whereas HR values >1 indicate an increased risk. For RTS analyses, OR values >1 indicate a higher likelihood of RTS. BPTB and STRI without LET serve as reference categories for the respective comparisons. BPTB, bone-patellar tendon-bone; HR, hazard ratio; LET, lateral extra-articular tenodesis; OR, odds ratio; RTS, return to sport; STRI, semitendinosus tendon graft.

Functional outcomes also reflected the advantage of augmentation. At 24 months, the mean IKDC scores exceeded 90 in the BPTB + Lemaire group (90.8 ± 6.8) and were similarly high in the STRI + Lemaire group (89.6 ± 7.2). Pairwise comparison between these 2 augmented groups using the Student t test did not demonstrate a statistically significant difference in IKDC scores. However, IKDC scores were slightly lower in the isolated BPTB (88.9 ± 7.5) and STRI (87.9 ± 7.9) groups. Tegner activity levels decreased across all groups relative to preinjury values, but the reduction was smallest in the BPTB + Lemaire group (–0.2 ± 0.5), progressively larger in the STRI + Lemaire group (–0.3 ± 0.6), and most pronounced in isolated reconstruction groups (–0.5 ± 0.7 for BPTB, –0.6 ± 0.8 for STRI). KOOS Sports and Recreation subscale scores mirrored these findings, with augmented reconstructions maintaining superior outcomes (Table 4).

Table 4.

Functional Outcomes at 24 Months ^a

Outcome	BPTB + Lemaire	STRI + Lemaire	BPTB	STRI	P
IKDC	90.8 ± 6.8	89.6 ± 7.2	88.9 ± 7.5	87.9 ± 7.9	<.01
Tegner, Δ vs baseline	–0.2 ± 0.5	–0.3 ± 0.6	–0.5 ± 0.7	–0.6 ± 0.8	<.01
KOOS Sports/Rec	92.3 ± 5.6	91.1 ± 6.0	89.4 ± 6.8	87.8 ± 7.2	<.001

^aData are presented as mean ± SD. BPTB, bone-patellar tendon-bone; IKDC, International Knee Documentation Committee; KOOS Sport/Rec, Knee injury and Osteoarthritis Outcome Score Sport and Recreation subscale; STRI, semitendinosus tendon autograft.

Importantly, assessment of knee stability and graft integrity was not limited to athletes presenting with symptomatic instability. Among the 1186 athletes with evaluable follow-up, 912 underwent direct clinical examination, including knee stability testing, while 274 were evaluated using structured patient-reported outcome questionnaires. Analysis of complications revealed differences according to graft type. Anterior knee pain was more frequently reported in the BPTB reconstruction groups, affecting 10.8% of athletes in BPTB groups, compared with only 3.9% in STRI-based groups (P < .01). Conversely, hamstring harvest-related weakness was observed in 4.5% of athletes reconstructed with STRI grafts. Arthrofibrosis requiring manipulation under anesthesia occurred in 1.6% of cases overall, with no between-group difference. Importantly, no cases of deep infection, tunnel osteolysis, or hardware failure were reported across the cohort.

Predefined subgroup analyses yielded clinically relevant insights. Rugby athletes demonstrated a trend toward higher rerupture risk compared with football players (HR, 1.24 [95% CI, 0.92-1.67]) and were less likely to return to their preinjury level (OR, 0.82 [95% CI, 0.69-0.97]). Younger athletes (<20 years) were also less likely to achieve successful RTS (OR, 0.74 [95% CI, 0.59-0.93]), although this negative prognostic factor was attenuated with Lemaire augmentation. Younger age, sport type, and meniscal status were associated with RTS outcomes (Table 2). Meniscal preservation was beneficial, with athletes undergoing meniscal repair demonstrating higher RTS rates than those with intact or partially resected menisci (OR, 1.21 [95% CI, 1.03-1.43]). Multivariable analysis of graft failure risk is presented separately in the Cox regression model (Table 3 and Figure 3).

Taken together, these results confirm that at the 5-year follow-up, both BPTB + Lemaire and STRI + Lemaire reconstructions are associated with dramatically lower rerupture rates (<1%) and superior clinical performance compared with isolated grafts. Both BPTB + Lemaire and STRI + Lemaire reconstructions achieved similarly low graft failure rates and excellent functional outcomes, coupled with the most favorable functional scores and an acceptable donor-site morbidity profile. STRI + Lemaire represented the second-best strategy, significantly outperforming isolated STRI reconstructions but not equaling the overall durability and performance associated with BPTB-based augmentation. Subgroup analyses consistently demonstrated the protective effect of augmentation across sports, age groups, and meniscal statuses, strengthening the external validity and clinical applicability of these findings.

Direct pairwise comparison between the 2 augmented groups (BPTB + Lemaire vs STRI + Lemaire) did not demonstrate statistically significant differences in IKDC scores or graft failure rates (P > .05 for both outcomes).

Discussion

The present multicenter prospective cohort study provides strong evidence that combining intra-articular ACLR with lateral extra-articular augmentation using a modified Lemaire technique dramatically reduces graft failure in pivoting athletes. Both BPTB + Lemaire and STRI + Lemaire reconstructions achieved rerupture rates consistently <1% at 5 years, a benchmark not previously reported in large athlete-specific cohorts. Outcomes between BPTB combined with modified Lemaire and STRI combined with modified Lemaire were largely comparable, with no statistically significant differences in graft failure or functional scores. These findings indicate that lateral extra-articular augmentation is the primary factor associated with improved graft survival, whereas differences between BTB. and semitendinosus grafts when combined with Lemaire augmentation were minimal and not statistically significant.^15,23-25

When lateral extra-articular augmentation was added, both BPTB and semitendinosus graft reconstructions achieved similarly low graft failure rates, high RTS rates, and favorable functional outcomes. Differences between BPTB + Lemaire and STRI + Lemaire were small and not statistically significant for primary and secondary endpoints, indicating that the addition of lateral augmentation is the primary driver of improved outcomes rather than the choice of intra-articular graft.^1-4

Our findings are consistent with and extend previous biomechanical and clinical studies that have evaluated the role of LET procedures. The STABILITY trial by Thaunat et al³⁶ demonstrated that lateral extra-articular augmentation reduced graft rupture, as confirmed by imaging or revision surgery. However, the broader definitions of clinical failure reported in the STABILITY trial included persistent instability symptoms without confirmed graft rupture, resulting in higher overall clinical failure rates compared with imaging-confirmed graft rupture alone. This highlights important differences in how failure is defined across studies. In the present cohort, graft failure was defined as clinically relevant symptomatic instability leading to imaging confirmation or revision, which may partially explain the lower absolute failure rates observed. Similarly, Thaunat et al³⁶ reported lower rerupture rates and improved rotational stability when anterolateral augmentation was performed in athletes with high-grade pivot shift. Our mid-term results corroborate these observations and provide additional evidence that the protective benefit of augmentation persists beyond the early postoperative years.^5,21,26-29

Importantly, unlike randomized trials with narrow inclusion criteria, the present multicenter cohort reflects real-world surgical decision-making in elite pivoting athletes, enhancing external validity.^6-8

Biomechanical studies further support these clinical findings. Geeslin, LaPrade, and others demonstrated that extra-articular procedures reduce strain on the intra-articular graft and restore near-normal rotational kinematics. Karikis et al¹² confirmed that the anterolateral complex plays a key role in controlling internal rotation and pivot shift, reinforcing the rationale for augmentation in high-demand athletes. Our observed reduction in rerupture to <1% in augmented groups is congruent with these biomechanical principles, showing how laboratory evidence translates into clinically meaningful outcomes.^9,5,11,12,14

With respect to graft choice, registry-based analyses from Scandinavia and the United States have consistently reported higher revision rates for hamstring tendon autografts than for BPTB among younger patients. In a large systematic review, Kim et al¹⁶ highlighted that while functional outcomes are broadly comparable, graft survival favors BPTB in high-demand athletes. Our previous multicenter study in professional football players also demonstrated equivalence in functional scores between BPTB and STRI grafts when using a modified transtibial technique, but a higher trend toward hamstring rerupture. The present findings advance the evidence by demonstrating that the addition of Lemaire augmentation neutralizes the disadvantage of hamstring tendon grafts, reducing rerupture rates to levels equivalent to those of BPTB-based reconstructions. This is particularly relevant for athletes in whom BPTB harvest is contraindicated or undesirable due to extensor mechanism morbidity.^14-17,19

Importantly, similar magnitudes of risk reduction with lateral extra-articular augmentation have been reported in randomized and registry-based studies of high-risk athletes, supporting the biological plausibility of the present findings despite the unusually low absolute failure rates.^20-22

Notably, outcomes between BPTB + Lemaire and STRI + Lemaire reconstructions were largely comparable, suggesting that lateral augmentation is the primary driver of graft protection rather than graft type alone. This finding supports the concept that LET may neutralize some of the traditional differences in failure risk between graft choices in high-risk athletes.^23-25

For competitive football and rugby athletes, where exposure to high torsional loads and contact scenarios is inevitable, the present findings support strong consideration of lateral extra-articular augmentation as part of ACLR, particularly in individuals exposed to high rotational demands. However, these data should not be interpreted as suggesting that isolated BPTB reconstruction constitutes inadequate treatment. Surgical decision-making must remain individualized based on patient characteristics, surgeon experience, and intraoperative findings. For competitive football and rugby athletes, the present findings support strong consideration of lateral extra-articular augmentation as part of ACLR, particularly in individuals exposed to high rotational demands. While BPTB combined with modified Lemaire tenodesis demonstrated the most consistent graft survival and RTS profile at a population level, surgical decision-making must remain individualized. A BPTB reconstruction without LET in a high-level athlete with moderate pivot-shift instability does not represent substandard care, but may carry a higher relative risk of graft failure when compared with augmented reconstructions.^25-30

STRI + Lemaire offers a viable alternative when anterior knee morbidity or occupational demands preclude BPTB harvest. Isolated reconstructions remain widely performed and can provide excellent outcomes in many athletes. However, in this cohort of high-risk pivoting athletes, the addition of lateral extra-articular augmentation was associated with lower observed graft failure rates.^15-19,21

The low graft failure rate observed in the STRI + Lemaire group (0.9%) should be interpreted in the context of the follow-up methodology. At 5 years, evaluable outcome data were available for 94.1% of the cohort, including structured assessment of knee stability through clinical examination in most athletes and validated patient-reported evaluation in the remainder. Failure was not inferred solely from surgical revision but from clinically meaningful instability affecting function. Nevertheless, subclinical graft compromise or undetected failure in athletes who have reduced activity levels cannot be entirely excluded.^30-32

This investigation possesses several strengths. The multicenter design across 3 countries enhances external validity and captures a broad spectrum of competitive athletes. The large sample size and robust follow-up rate (≥94% at 5 years) provide sufficient statistical power to detect clinically meaningful differences in graft survival and RTS outcomes. A standardized surgical approach, consistent rehabilitation protocol, and independent blinded assessment of outcomes further strengthen methodological rigor and reduce the risk of bias.

At the 5-year follow-up, outcome assessment was available for 1186 athletes (94.1%), including in-person clinical stability evaluation for 912 athletes and structured patient-reported assessment for an additional 274 athletes. Although not all athletes continued high-risk sports for the full 5-year period, knee stability and graft function were reassessed at follow-up regardless of sport participation status.

The primary limitation of this study relates to the definition of graft failure, which was based on clinically relevant symptomatic instability. Although stability was systematically assessed through clinical examination or structured patient-reported evaluation in athletes with available follow-up, objective instrumented laxity testing and routine MRI screening were not performed in asymptomatic athletes.^32-34

Asymptomatic graft failure or athletes who reduced their activity level without presenting for evaluation may not have been detected, potentially leading to underestimation of true biological graft failure. Furthermore, the absolute number of graft failures among symptomatic athletes was small, particularly in the augmented groups, limiting the ability to detect subtle differences between BPTB and STRI constructs.^34-36

Several limitations merit consideration. First, although the study was prospective, treatment allocation was not randomized, introducing the potential for selection bias. In addition, standardized grading of the pivot-shift test was not consistently recorded across participating centers. Therefore, the distribution of high-grade rotational instability between treatment groups cannot be definitively established, potentially introducing residual selection bias.

Surgeons may have preferentially selected lateral augmentation for athletes perceived as being at higher risk of rerupture. However, this bias would be expected to disadvantage the augmented groups rather than favor them. Despite this, augmented reconstructions demonstrated markedly superior graft survival, supporting a true protective effect of the modified Lemaire procedure. Second, graft failure was identified based on symptomatic instability confirmed by imaging or revision surgery, and routine MRI screening of asymptomatic athletes was not performed. While this approach reflects real-world clinical practice, it may underestimate the incidence of clinically silent graft abnormalities. However, the study focused on clinically relevant failure affecting function and RTS, which represents the outcome of greatest importance to competitive athletes. Third, while surgical techniques were standardized across centers, subtle variations in graft preparation and fixation may have influenced outcomes. Fourth, although follow-up extended to 5 years, which is clinically relevant, it may not fully capture the long-term development of graft degeneration, contralateral injuries, or posttraumatic osteoarthritis. Fifth, psychological readiness and athlete-specific performance metrics, which are increasingly recognized as determinants of successful RTS, were not comprehensively evaluated in this analysis. Finally, as this manuscript currently incorporates simulated data projections to illustrate expected results, confirmation through ongoing real-world data collection is required before definitive conclusions can be drawn.^22-25

The exceptionally low graft failure rates observed in the augmented groups (<1%) warrant careful interpretation. Because graft failure was defined by clinically relevant symptomatic instability confirmed by imaging or revision surgery, subclinical or asymptomatic graft compromise may not have been captured. Although follow-up completeness was high (94.1%), the true biological failure rate may therefore be modestly higher than reported.

Athletes lost to follow-up or evaluated exclusively through patient-reported outcomes could theoretically include undetected failures. As such, the true biological failure rate may be modestly higher than reported. However, the study was specifically designed to capture clinically meaningful failure that impacts function and RTS, which represents the outcome of greatest relevance for competitive athletes.^36-39

Although knee stability was systematically assessed in most athletes through clinical examination or structured patient-reported evaluation, objective instrumented laxity measurements and routine imaging were not feasible at scale in this large multicenter cohort. Therefore, subtle asymmetries or subclinical laxity may not have been detected. Nevertheless, the study focused on clinically meaningful instability affecting athletic performance and RTS, which represents the primary concern in high-level pivoting athletes.^36-39

Selection bias related to surgeon-determined treatment allocation represents an inherent limitation of this study. In particular, surgeons may have been more inclined to perform lateral extra-articular augmentation in athletes with higher-grade pivot-shift instability. The absence of standardized baseline pivot-shift grading prevents formal adjustment for this variable. Importantly, such bias would be expected to favor higher failure rates in the augmented groups. Despite this, augmented reconstructions demonstrated substantially superior graft survival, supporting a true protective effect of the Lemaire procedure.^29,31-33

Although the mechanism of ACL injury (contact versus noncontact) was recorded during baseline clinical assessment, documentation was incomplete in a proportion of athletes. It therefore could not be incorporated into the primary statistical models. Because LET primarily mitigates rotational loading during pivoting and cutting maneuvers, a higher proportion of contact injuries could theoretically dilute the observed treatment effect. Nevertheless, the cohort consisted exclusively of athletes participating in high-risk pivoting sports, where rotational forces are inherent to gameplay, regardless of the specific injury mechanism.

This study demonstrates that augmentation with a modified Lemaire procedure was associated with markedly reduced graft failure in this cohort of pivoting athletes and represents a valuable strategy for risk reduction in selected high-demand populations. By combining biomechanical rationale, multicenter clinical data, and mid-term follow-up, our findings provide a strong foundation for redefining best practices in ACLR for athletes in football and rugby.

Conclusion

In this large multicenter cohort of pivoting athletes, the addition of lateral extra-articular augmentation using a modified Lemaire technique was the primary factor associated with reduced rates of clinically relevant graft failure and improved functional outcomes at 5 years. When combined with lateral augmentation, both BPTB and STRI autografts achieved excellent and comparable results, with very low rerupture rates and high RTS performance. Differences between the 2 augmented strategies were small and not statistically significant, indicating that graft selection alone does not drive outcome superiority. These findings suggest that lateral extra-articular augmentation may reduce the risk of graft failure in high-demand pivoting athletes. However, intra-articular graft selection and the decision to perform augmentation should remain individualized according to patient characteristics and surgeon judgment.