Operative treatment of high-grade medial collateral ligament injury and chronic medial knee instability: a systematic review

Abstract

High-grade medial collateral ligament (MCL) injury and chronic medial knee instability can compromise both valgus and rotational control, especially in cruciate or multiligament injury. This systematic review evaluated operative treatment according to clinical phenotype rather than MCL grade alone. PubMed/MEDLINE and Web of Science Core Collection were searched from January 1, 2000, to August 10, 2026; a Cochrane CENTRAL update was completed on September 7, 2026. Eligible clinical reports included operative treatment of acute high-grade MCL/medial-complex injury or chronic symptomatic medial instability. Risk of bias was assessed with RoB 2, ROBINS-I, or the Joanna Briggs Institute case-series checklist. Thirty-five reports were included: 5 randomized or quasi-randomized trials, 9 nonrandomized comparative studies, and 21 case series. Routine acute medial surgery during ACL reconstruction was not consistently superior across heterogeneous grade III populations. In contrast, selected displaced distal or Stener-like lesions were plausible repair targets because interposition can prevent healing in apposition. Chronic symptomatic valgus and/or anteromedial rotatory insufficiency was most often treated with reconstruction, with directional improvement in objective stability and patient-reported outcomes in case series. Evidence for augmentation, the optimal reconstruction construct, and timing remains low certainty. Decisions should integrate repairability, residual valgus and rotational instability, associated ligament injury, alignment, and tissue quality. Prospective phenotype-stratified comparative studies using standardized stability, return-to-sport, complication, and long-term joint-preservation outcomes are required.

Level of evidence: Level IV, systematic review of Level I-IV therapeutic studies

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s00402-026-06517-8.

Introduction

The medial side of the knee functions as a stabilizing complex. The superficial MCL is the principal restraint to valgus loading, whereas the deep MCL, posterior oblique ligament, posteromedial corner, and capsular tissues contribute to external rotation and coupled anteromedial translation [1–3]. An isolated superficial MCL deficiency may therefore predominantly manifest as valgus opening, while a combined medial injury may produce anteromedial rotatory instability (AMRI), particularly when the ACL, PCL, meniscus, or other ligamentous structures are injured.

Experience with low-grade MCL injury should not be generalized to complete, displaced, avulsed, combined, or chronically insufficient lesions. Distal tibial avulsions can be displaced around or beneath the pes anserinus; in a Stener-like pattern, interposed tissue prevents direct apposition to the tibial attachment. Such structural features have an immediate influence on healing potential and may justify a different strategy from that used for a nondisplaced injury [7, 8, 12].

Repair, repair augmentation, and reconstruction are consequently not interchangeable treatments. Repair aims to preserve recognizable, viable acute tissue. Reconstruction is usually considered for chronic, irreparable, failed, or combined medial/posteromedial functional insufficiency. The conventional grade I-III classification remains useful for communicating injury severity, but it does not specify the injured layer, attachment site, displacement, tissue apposition, rotational component, chronicity, or mechanical environment. This review therefore synthesizes the literature using clinical phenotypes and distinguishes evidence-derived findings from evidence-informed clinical interpretation.

Materials and methods

Protocol, eligibility, and search strategy

This review followed PRISMA 2020 [13]. The protocol was specified before final selection but was not prospectively registered. Eligible reports included skeletally mature patients with acute high-grade MCL/medial-complex injury or chronic symptomatic medial instability treated with repair, repair augmentation, superficial MCL reconstruction, or medial-complex reconstruction. Randomized trials, comparative cohorts, and case series including at least five treated knees were eligible. Cadaveric, technical-only, imaging-only, nonoperative-only low-grade, duplicate, and nonseparable reports were excluded.

The original search of PubMed/MEDLINE and Web of Science Core Collection covered January 1, 2000, through August 10, 2026. Search terms combined medial collateral ligament, medial knee instability, medial ligament complex, and posteromedial corner with repair, reconstruction, augmentation, internal brace, and suture tape concepts; the PubMed strategy additionally used the MeSH term Knee Injuries. The original searches retrieved 868 and 1,040 records, respectively. In response to editorial review, CENTRAL was searched on September 7, 2026 using database-specific controlled vocabulary/free-text combinations and years 2000–2026, yielding 44 records. Embase could not be searched because institutional access was unavailable; this limitation is retained explicitly in the manuscript and flow diagram.

The CENTRAL update yielded 19 duplicates of records already held in the review library and 25 unique records that were ineligible at title/abstract screening. No new CENTRAL record met the eligibility criteria. Citation/reference searching identified one eligible clinical report [35]. During reconciliation, one previously retained database record was found to meet a transparently refined functional-medial-complex scope: an ACL plus grade II MCL cohort with an arthroscopic floating-meniscus sign, indicating a high-risk deep medial/meniscotibial phenotype [36]. This report is analyzed separately as indirect evidence rather than as conventional grade III MCL evidence.

Selection, extraction, risk of bias, and certainty

Two reviewers independently screened all database and citation-search records, completed full-text consensus, extracted prespecified data, and assessed risk of bias; disagreements were resolved by discussion. The CENTRAL update and citation-search records underwent the same independent eligibility assessment and data-field verification as the original records. The extraction form recorded chronicity, location and special pattern, associated injury, intervention and comparator, operative details, rehabilitation, follow-up, objective valgus and rotational assessment, patient-reported outcomes, return to activity, complications, reoperation, and the original authors’ failure definition. Manual valgus findings, stress-radiographic opening, and AMRI assessments were retained separately rather than converted into a pooled stability endpoint.

Randomized/quasi-randomized trials were assessed using RoB 2, nonrandomized comparative studies using ROBINS-I, and case series using the Joanna Briggs Institute checklist. Meta-analysis was not performed because indications, intervention constructs, co-interventions, stability measurement protocols, outcome definitions, and follow-up were clinically and methodologically heterogeneous. For this revision, evidence was additionally organized into six overlapping clinical phenotypes: acute potentially repairable MCL injury; displaced distal/Stener-like lesion; acute ACL-MCL pattern; persistent/chronic valgus instability; combined valgus plus AMRI; and revision ACL or multiligament injury. Acute versus chronic presentation, operative strategy, associated injury, and outcome domains were prespecified; the six-phenotype presentation and the floating-meniscus functional-medial-complex subgroup were post hoc synthesis refinements.

Certainty was assessed qualitatively using GRADE principles for the six phenotype-linked clinical questions. Certainty was downgraded for risk of bias, inconsistency, indirectness, and imprecision. A pooled effect estimate was not calculated, and the certainty labels should not be interpreted as formal comparative effect estimates.

Results

Study selection and characteristics

A total of 1,952 database records and one record from citation/reference searching were identified. After 638 duplicates were removed, 1,314 records were screened and 1,273 were excluded at title/abstract review. Forty-two reports were sought; seven were not retrieved or could not be reliably assessed (one unavailable report and six CAJ-format reports). Thirty-five reports met the eligibility criteria and were included: five randomized or quasi-randomized trials, nine nonrandomized comparative studies, and twenty-one case series. The evidence base included acute avulsions and Stener-like lesions, combined ACL-MCL injury, multiligament injury, and chronic medial/posteromedial insufficiency. Figure 1 shows the updated study selection process; Tables S1 and S2 provide study-level stability and adverse-event reporting.

Fig. 1.

PRISMA 2020 study-selection flow diagram. CENTRAL was searched on September 7, 2026. Embase was not searched because institutional access was unavailable

Risk of bias and certainty of evidence

Randomized evidence had low risk or some concerns in most domains, although incomplete outcome data, outcome measurement, and selective reporting commonly limited certainty. The quasi-randomized study was at high risk of bias because allocation did not ensure true randomization. Nonrandomized comparative studies were chiefly vulnerable to confounding by indication, as chronicity, tissue quality, injury location, baseline valgus/rotational laxity, associated injury, alignment, and surgeon preference influenced both treatment allocation and outcome. Case series were limited by small samples, incomplete follow-up, heterogeneous co-interventions, and inconsistent outcome definition; the consensus JBI appraisal for the 21 case-series reports is shown in Supplementary Figure S1.

Table 1.

Qualitative certainty of evidence profile by clinical question

Clinical question	Evidence base	Evidence-derived finding	Certainty	Main limitations
Routine acute MCL surgery during ACL reconstruction	3 randomized/quasi-randomized studies plus comparative cohorts	No consistent advantage across unselected grade III populations; selected morphology/instability may matter	Low	Heterogeneity, indirectness, imprecision
Repair of displaced distal or Stener-like lesions	Selected cohorts	Favorable stability and return-to-activity reporting supports recognition of non-apposition phenotype	Very low	No controlled lesion-specific comparison
Repair versus reconstruction for acute medial injury	1 quasi-randomized study and cohorts	Both approaches improved outcomes; comparative superiority uncertain	Very low	High risk of bias and clinical heterogeneity
Repair plus augmentation versus repair alone	1 randomized equivalence trial and 1 cohort	No durable clinical/radiographic augmentation benefit demonstrated	Low	Limited samples and follow-up
Reconstruction for chronic valgus/rotational insufficiency	Case series and comparative cohorts	Directional improvement in stability and PROMs	Very low	Selection bias; co-interventions; no uniform comparator
Isolated sMCL versus combined medial/posteromedial reconstruction	Comparative cohorts and case series	Construct should match demonstrated rotational phenotype; superiority unproven	Very low	Indirect biomechanics and heterogeneous constructs

Evidence-derived findings by clinical phenotype

Acute potentially repairable MCL and displaced distal and stener-like lesions

Acute repair is biologically most persuasive when viable tissue can be restored to a recognizable attachment. In selected distal avulsion and Stener-like cohorts, repair was associated with restoration of stability and favorable return-to-sport or patient-reported outcomes [7, 8]. The evidence supports prompt recognition of a lesion that is unlikely to heal in apposition; it does not establish that every acute grade III MCL injury requires operation. A 2026 comparative cohort did not demonstrate a midterm clinical or radiographic advantage of augmentation over repair alone for Stener-like lesions [12].

Acute ACL-MCL phenotype

Randomized evidence does not establish a uniform benefit for routine medial surgery in every ACL-associated grade III MCL injury. A three-arm randomized trial found no consistent overall advantage of repair or isolated suture-tape bracing over no MCL repair during ACL reconstruction, although a morphology-defined subgroup had more residual opening without repair [9]. In an earlier randomized study, nonoperative MCL management was associated with faster early flexion and quadriceps recovery, without a clear long-term motion difference [4]. The high-risk grade II floating-meniscus trial is indirect to the review’s conventional high-grade population, but it reported lower residual medial laxity and fewer ACL reconstruction failures after medial compartment reconstruction than after bracing [36].

The Blanke cohort operationalized the clinical role of AMRI in an ACL-MCL pathway. Patients without AMRI were managed without MCL surgery, whereas patients with AMRI underwent MCL repair with posteromedial plication; the study reported no recurrent AMRI at final follow-up, but allocation was determined by the instability phenotype and no causal comparison is possible [35]. Thus, the clinical signal is hypothesis-generating: residual functional valgus/rotational instability, rather than ACL association alone, may identify patients in whom medial treatment warrants consideration.

Persistent and chronic valgus and combined valgus plus AMRI

Chronic medial or posteromedial reconstruction cohorts generally reported directional improvement in objective valgus stability and patient-reported outcomes [1–3, 10, 23, 29]. No reconstruction construct was conclusively superior. A comparative cohort of superficial MCL-posterior oblique ligament reconstruction versus superficial MCL reconstruction with posteromedial capsular advancement reported similar final stability/function but more flexion loss after capsular advancement [16]. These data support reconstructing deficient stabilizers in chronic symptomatic functional insufficiency, but not routine multibundle extension in every knee with coronal opening.

Revision ACL or multiligament phenotype

In revision ACL and multiligament settings, persistent medial opening or unrecognized AMRI may coexist with tunnel constraints, prior grafts, meniscal/root pathology, cartilage disease, and alignment abnormality. Observational studies have associated selected medial reconstruction strategies with lower ACL graft-failure rates, but treatment assignment is substantially confounded by severity and surgeon judgment [6, 18]. Alignment, range of motion, tissue quality, and the interaction among planned ACL/PCL/medial tunnels should therefore be evaluated before a procedure is selected.

Complications, failure, and outcome reporting

Stiffness, manipulation under anesthesia, arthrofibrosis, wound problems, recurrent laxity, graft failure, and reoperation were reported inconsistently. A pooled complication rate was not defensible because definitions and denominators differed. Residual valgus laxity, persistent AMRI, revision medial surgery, graft rupture, and ACL graft failure are clinically related but noninterchangeable endpoints. Similarly, return-to-sport reporting ranged from elite-athlete return to preinjury competition to activity scores without a defined return criterion. Table S2 separates the original studies’ reported events and failure definitions.

Discussion

Evidence-derived conclusions

The most defensible conclusion is that operative treatment should be phenotype-guided rather than grade-driven. Current randomized and comparative evidence does not support routine acute MCL surgery for all ACL-associated grade III injuries. Selected displaced distal/Stener-like lesions have a biologic rationale for acute repair because direct tissue apposition is impaired. For chronic symptomatic valgus and/or rotational insufficiency, reconstruction commonly improves objective stability and function in case series, but certainty remains very low because populations, constructs, co-interventions, and endpoints differ.

Augmentation is a technical adjunct rather than a demonstrated long-term clinical advantage. The available comparative evidence did not show a consistent benefit of augmentation over repair alone, and the literature cannot rank single-bundle, combined medial/posteromedial, or other reconstruction designs. The key limitation is not the absence of plausible techniques; it is the absence of phenotype-defined comparisons using the same stability protocol and rehabilitation pathway.

Evidence-informed clinical interpretation

Figure 2 is a proposed evidence-informed clinical framework, not a validated treatment algorithm. It organizes the findings around repairability, functional instability, associated injury, and alignment. The framework should not be interpreted as evidence that a specific operation is superior for each phenotype. Rather, it identifies the clinical factors that should be documented before treatment: injury location and apposition, valgus stress at extension and 20–30 degrees, rotational/AMRI examination, contralateral comparison, stress radiography when available, MRI and associated pathology, limb alignment, prior surgery, and rehabilitation response.

Fig. 2.

Proposed evidence-informed clinical framework for medial knee instability. This interpretive framework is designed to organize evidence synthesis and is not a validated treatment algorithm

A repairable lesion and a functional insufficiency are related but distinct problems. A viable avulsion that cannot heal in apposition may justify acute repair. A nondisplaced injury with improving objective stability and no AMRI may be managed with protection and structured rehabilitation. Persistent symptomatic valgus opening, rotational instability, failed prior treatment, or a complex revision/multiligament environment may shift the goal toward reconstruction of the demonstrated deficient restraints. In chronic cases, uncorrected coronal malalignment can undermine a soft-tissue reconstruction and should be addressed in the treatment plan.

Limitations

This review is limited by the absence of Embase access, one unavailable report, and six reports not reliably assessable in CAJ format. CENTRAL was added during revision, but did not yield an additional eligible CENTRAL record. Most evidence comprised small nonrandomized cohorts and case series with heterogeneous injury patterns, surgical constructs, co-interventions, outcome definitions, and follow-up. The six clinical phenotypes are a useful synthesis structure, but may overlap and have not been prospectively validated. The floating-meniscus grade II study was included as a transparent functional-medial-complex scope refinement and remains indirect to traditional high-grade MCL populations. These limitations precluded meta-analysis and warrant cautious interpretation.

Conclusions

Operative treatment of high-grade MCL injury and chronic medial knee instability should be individualized according to lesion repairability, objective residual valgus and/or rotational instability, concomitant injury, alignment, tissue quality, and prior treatment. Acute displaced distal and Stener-like lesions are reasonable repair targets; chronic symptomatic medial or posteromedial functional insufficiency is a reasonable indication for reconstruction. Current evidence remains insufficient to recommend a universal augmentation strategy, reconstruction construct, or timing pathway.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

ZYX, CL and HYZ contributed to the conceptualization; ZYX, CL and DXP were responsible for the methodology; HYZ and DXP undertook the software development; ZYX, YHC, WXY and SY performed the validation; YHC and WXY carried out the formal analysis; ZYX, YHC, WXY and SY conducted the investigation; ZYX provided the resources; ZYX, HYZ, CL and DXP completed the data curation; ZYX and DXP wrote the original draft; ZYX, HYZ and CL participated in the writing‑review and editing; ZYX, HYZ and CL were in charge of visualization; XYY and SXY provided supervision; ZYX handled the project administration and funding acquisition. All authors have verified their respective contributions.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Conflict of interest

The authors declare no competing interests.