➢ Chronic lateral ankle instability can occur in a subset of patients following ankle inversion sprains. Operative treatment to restore stability in the ankle and hindfoot and to prevent further degenerative changes may be indicated in cases in which nonoperative treatment has failed.
➢ Anatomical direct repair with use of native ligament remnants with or without reinforcement of the inferior retinaculum is the so-called gold standard operative strategy for the treatment of lateral ankle instability. The procedure has shown promising short and long-term outcomes. Candidates for the procedure have ligament remnants of sufficient quality that are amendable to direct repair.
➢ Anatomical reconstruction with use of autograft or allograft is reserved for patients with insufficient ligament remnants to fashion a direct repair, failed previous lateral ankle repair, high body mass index, or generalized ligamentous laxity. A wide variety of autografts have been described, each with potential advantages and disadvantages. These procedures can provide good-to-excellent short-term outcomes. However, there is no available information on their long-term clinical results.
➢ Non-anatomical lateral ligament reconstruction typically involves the use of the adjacent peroneal brevis tendon and has been applied in cases in which only poor-quality ligament remains. The procedure can provide good to excellent short-term outcomes, although reported long-term outcomes have differed among studies. The particular tendon used for the graft should be carefully considered given the potential alterations in the kinematics of the ankle and hindfoot.
➢ Arthroscopic ligament repair is becoming increasingly popular as it is minimally invasive. This procedure is restricted to patients who have good-quality ligament remnants. Good-to-excellent clinical outcomes have been reported after short and long-term follow-up, although a relatively high rate of complications—including nerve damage—has been reported following the procedure and therefore warrants further investigation before widespread adoption can be advocated.
Ankle sprains are among the most common sports-related injuries, with 23,000 occurring each day in the United States1. Most involve the lateral ankle ligament complex during forced inversion while the ankle joint is in plantar flexion2. Although a recent review by Petersen et al. suggested that most acute lateral ankle sprains can be treated nonoperatively3, as many as 20% of these injuries ultimately result in chronic lateral ankle instability (LAI)4.
Some kinematic studies have suggested that lateral ankle ligament deficiency alters both talar position and contact area, thereby inciting degenerative changes5-8. Thus, patients with persistent pain and recurrent sprains lasting greater than six months may be candidates for operative treatment. Nevertheless, a dearth of studies comparing operative strategies and a lack of consensus as to the preferred treatment modality have made the treatment of LAI challenging. A Cochrane review by de Vries et al. concluded that there was insufficient evidence to determine which operative strategy was the most beneficial to treat LAI9. The purpose of the present report is to provide an evidence-based review of available operative treatments for LAI.
Anatomical procedures for the treatment of LAI are broadly divided into two categories: direct repair of the injured ligament and anatomical reconstruction with use of a graft. Direct repair is indicated for patients with sufficient ligament remnants amendable to sutures, whereas reconstruction is indicated for those with insufficient or attenuated ligament remnants10-13. However, scoring criteria to evaluate ligament remnants have not been established and deserve future characterization.
Direct repair using the native ligament remnant with or without reinforcement of local tissue is generally considered to be the first-line operative treatment of LAI (Figs. 1-A, 1-B, and 1-C)10-16. The procedure carries inherent advantages such as low cost, procedural simplicity, minimal invasiveness, and low complication rates. Contraindications include longstanding ankle instability with insufficient ligamentous tissue, previous failed stabilization procedures, high body mass index, and generalized ligamentous laxity10-13.
Outcomes of Direct Repair
Direct repair has shown promising functional outcomes in the literature, with the majority of patients having good-to-excellent results. To our knowledge, the longest follow-up after the Broström procedure14 was reported by Bell et al., who followed thirty-two ankles for a mean of 26.3 years17. Those authors reported that the mean score for overall ankle function was 91.2 of 100 and the mean Foot and Ankle Outcome Score was 92 when averaged over five functional areas.
Lee et al. reported the long-term results of direct repair with reinforcement of local tissue in a study of thirty patients who underwent anterior talofibular ligament (ATFL) repair with inferior extensor retinacular reinforcement18. Twenty-eight patients returned to the preinjury level of activity, and, after a mean duration of follow-up of 10.6 years, the mean reported American Foot & Ankle Society (AOFAS) score was 91 and the mean visual analog scale (VAS) score was 87. Tourné et al. evaluated the long-term results of ligamentous retensioning and reinforcement with use of the extensor retinaculum in a study of 150 patients19. After a mean duration of follow-up of eleven years, 93% of the patients had satisfactory results and the mean overall Karlsson score20 was 94.8, with no deterioration of the articular surface detected on radiographs.
Several authors have compared the modalities of ligamentous reattachment. In this regard, Karlsson et al. compared the modified Broström-Gould procedure15 with the Karlsson technique16 in a prospective randomized study of sixty patients who were followed for a mean of 3.1 years21. Those authors found no significant differences between the treatment groups in terms of functional outcome or mechanical instability and concluded that both surgical techniques offered good-to-excellent functional results. In the study by Cho et al., fifty patients were prospectively randomized to bone-tunnel and suture-anchor techniques and were followed for two years postoperatively22. The authors reported no significant differences between the techniques when the results were evaluated with the Karlsson scale, the Sefton grading system23, and stress radiographs.
Long-term degenerative osteoarthritic changes have been evaluated in several case series. In the study by Krips et al., one of twenty-five patients who underwent two different variations of direct repair had medially located osteophytes without joint-space narrowing after a mean duration of follow-up of 12.3 years24. Muijs et al. assessed the development of osteoarthritic changes in a study of twenty-one patients who were managed with the technique of Duquennoy et al.25 and were followed for a mean of thirteen years26. The authors found that seven of the fifteen patients without preexisting osteoarthritis had developed grade-1 osteoarthritis (osteophytes without joint-space narrowing) after a mean duration of follow-up of more than nineteen years, although five of those seven patients had developed grade-1 osteoarthritis in the contralateral ankle as well26. One of five patients with preexisting grade-1 osteoarthritis had progression to grade 2 (joint-space narrowing with or without osteophytes) at a mean of twelve years of follow-up. Maffulli et al., in a study of thirty-eight patients who were followed for a mean of 8.7 years after an original Broström procedure, reported that five patients had grade-1 degenerative changes and three had grade-2 degenerative changes12.
On the basis of this evidence, direct repair has the potential to provide good-to-excellent clinical outcomes in both the short term and the long term. It also appears that long-term deterioration of the ankle joint is unavoidable in a subset of patients.
Anatomical reconstruction with use of a graft is necessary for patients with poor-quality ligament remnants, a failed previous lateral ankle repair, high body mass index, or generalized ligamentous laxity10-13. Several modalities of grafts, drill-hole configurations, and anchoring methods have been described.
Autograft options include local grafts (e.g., peroneal longus11,27, extensor digitorum longus28,29) and free grafts (e.g., Achilles tendon30,31, plantaris23,32, palmaris longus33, bone-patellar tendon34, and hamstrings35-38). The advantage of using autografts is tissue quality, whereas an inherent disadvantage is the possibility of donor-site morbidity. For allografts, toe extensor and/or flexor39, fascia lata40, hamstrings41-43, plantaris44, anterior tibialis45, and peroneus longus46 tendons have been used to treat lateral ankle instability. Allografts do not carry the risk of donor-site morbidity and provide shorter operative times and less postoperative pain42,43, but they carry the inherent yet low risk of disease transmission, slower biological healing, subclinical immune response, and high cost41,47. Of note, basic-science studies have suggested that viscoelastic behavior is different between tendon and ligament48, which might affect clinical outcome.
Variations in drill-hole configurations in the fibula also have been reported. Several authors have recommended either that two drill-holes should be placed close together in the fibula for reconstructions of the ATFL and calcaneofibular ligament (CFL) or that a single drill-hole, located 1 cm proximal to the tip of the lateral malleolus, should be used for both ligaments11,38,41,49,50. Paradoxically, most procedures involving two or three drill-holes have been associated with excellent short-term clinical outcomes27,28,33,35,44,51.
Outcomes of Anatomical Reconstruction with Autograft or Allograft
Several authors have reported good short-term clinical results following anatomical reconstruction11,34,36-38,42,43,45. However, to our knowledge, no study has described long-term outcomes and we are not aware of any Level-I studies comparing anatomical reconstruction with other treatment methods.
Coughlin et al., in a study of twenty-nine ankles that were treated with autologous gracilis grafts for anatomical reconstruction of the ATFL and CFL, reported that all patients had good or excellent outcomes according to the AOFAS and Karlsson scores, with no restriction of motion in the ankle and subtalar joints, after an average duration of follow-up of twenty-three months36. Sugimoto et al., in a study of thirteen patients who were managed with bone-patellar tendon grafts for the treatment of chronic ankle instability, reported that all patients had a grade-1 condition according to the evaluation system of Good et al.52 after a mean duration of follow-up of 26.5 months34. Jeys and Harris described an anatomical reconstruction technique involving interference screws37. Takao et al. reported on an anatomical reconstruction with use of autologous gracilis tendon and an interference fit anchoring system in a study of seventeen patients with chronic ATFL injury and four patients with chronic ATFL and CFL injury38. The authors found that all patients achieved mechanical stability on stress radiographs. Kennedy et al. reported on a hybrid anatomical reconstruction technique involving a split peroneus longus autograft with interference screw fixation in a study of fifty-seven patients who were followed for a mean of thirty-two months11 (Figs. 2-A through 2-D). The authors reported that the Foot and Ankle Outcome Score and Short Form-12 score improved significantly at the time of the latest follow-up (p < 0.01) and that all patients achieved mechanical stability.
Compared with anatomical reconstruction with use of autografts, there is a paucity of literature describing anatomical reconstruction with use of allografts, although allografts are becoming increasingly utilized in practice39-46. Jung et al. evaluated the clinical outcomes for twenty-eight ankles that had been treated with anatomical reconstruction with use of semitendinosus tendon allograft and interference screws and followed for a median of nineteen months postoperatively42. The authors reported that the median AOFAS score increased from 63 to 91 and that the median Karlsson score20 increased from 55 to 80. Hua et al., in a study of thirty-six ankles that had been treated with anatomical reconstruction with use of semitendinosus allograft and followed for a mean of 37.9 months, reported that the mean AOFAS score increased from 42.3 to 90.4 and that the mean Karlsson score improved from 38.5 to 90.143. In a case series, Ellis et al. used the anterior tibial tendon for anatomical reconstruction and reported that ten of eleven patients had good-to-excellent results45.
Anatomical reconstruction with use of autograft or allograft has substantial potential to provide good short-term clinical results. Nevertheless, despite the multitude of operative techniques described in literature, it remains unclear which procedures are most beneficial in the long term. Insufficient data exist to evaluate long-term degenerative changes following these procedures.
Non-anatomical procedures typically involve tenodesis of the adjacent peroneus brevis tendon to stabilize the talocrural joint. A variety of techniques are currently employed, including the Watson-Jones (Fig. 3-A)53, Chrisman-Snook (Fig. 3-B)54, Evans55, and modified Evans procedures. The checkrein procedures have traditionally been utilized only in cases of failed ligament repair or reconstruction, insufficient ligament tissue for direct imbrication, or collagen insufficiency10-13. Disadvantages of tenodesis include overtightening, which can increase the risk of arthritis in the subtalar and talocrural joints56,57; wound complications and nerve damage due to an extensive skin incision; and residual eversion weakness and disruption of the peroneus brevis, a dynamic stabilizer58,59.
Outcomes of Non-Anatomical Procedures
Van der Rijt and Evans retrospectively evaluated the cases of nine patients who were followed for an average of twenty-two years after the Watson-Jones procedure (Fig. 3-A)66. The authors found that three patients had complete relief of symptoms and that two patients with favorable early results had deterioration seven to ten years after the procedure. Unfavorable long-term results were also reported by Karlsson et al., who found that twenty-one of forty-two patients who had been managed with the Evans procedure had unsatisfactory results after an average duration of follow-up of fourteen years and that twelve patients with satisfactory early results had deterioration within three to six years after the procedure67. Krips et al. compared the long-term outcomes of the Evans procedure with those of anatomical reconstruction after a minimum duration of follow-up of fifteen years68. The authors found that the Evans procedure was associated with poorer functional results, less effective restoration of stability, and more osteoarthritis of the ankle after fifteen to thirty years compared with the anatomical procedure. Nimon et al., in a study of ninety-one ankles that were followed for a mean of 9.7 years after the modified Evans procedure, found that only 65% of the ankles had mild or no pain and minimal instability and that 12% showed degenerative changes69.
Contrary to those reports, several studies have demonstrated good long-term outcomes after non-anatomical tenodesis. Korkala et al., in a study of twenty-four patients (twenty-five ankles) who were followed for a mean of twenty years after the Evans procedure, reported that 80% of ankles had good-to-excellent results despite residual instability, although three ankles showed initial signs of osteoarthritis on radiographs70. Sugimoto et al., in a retrospective study of thirty-four ankles that were treated with the Watson-Jones procedure, found that thirty ankles had good-to-excellent results and that eighteen ankles had osteophyte formation without joint-space narrowing after an average duration of follow-up of fourteen years71. Snook et al., in a study of forty-eight ankles that were treated with the Chrisman-Snook procedure, reported that thirty-eight ankles had an excellent result after an average duration of follow-up of ten years (Fig. 3-B)72.
While non-anatomical procedures appear to provide good short-term results, long-term results have been inconsistent. As such, these procedures should be limited to patients who require a salvage procedure or who have insufficient ligament remnants or collagen. As the peroneal tendon is a dynamic stabilizer, its harvest should be considered carefully.
Studies Comparing Non-Anatomical with Anatomical Procedures
To our knowledge, only two studies have compared the clinical outcomes of anatomical and non-anatomical procedures in a prospective, randomized fashion59,73. Hennrikus et al. , in a study of forty-two ankles that were followed for an average of twenty-nine months, reported no significant differences between the Chrisman-Snook procedure and a modified Broström procedure in terms of clinical outcomes based on Sefton scores23, although the Chrisman-Snook group had significantly higher rates of wound complications and nerve damage (p < 0.001)59. Six patients who had been managed with the Chrisman-Snook procedure reported that the ankle felt “too tight,” compared with two patients who had been managed with the modified Broström procedure. Rosenbaum et al. compared the short-term clinical and functional outcomes of the Evans tenodesis (ten patients) with those of the modified Broström procedure (ten patients)73. After an average duration of follow-up of ten months, the authors found that all patients in the Evans group and nine patients in the modified-Broström group reported subjective improvement in terms of ankle stability. Three patients in the Evans group and one in the modified-Broström group complained of occasional pain and swelling during exercise. Finally, the authors reported that patients in the Evans group had significantly less talar tilt (p < 0.05). Several other retrospective studies comparing anatomical and non-anatomical procedures have demonstrated similar results68,74,75.
Sammarco, in a review of the complication rates following anatomical and non-anatomical LAI stabilization procedures58, reported that non-anatomical procedures were associated with a 9.7% rate of local nerve complications, a 4.0% rate of wound complications, and a 0.18% rate of reflex sympathetic dystrophy, whereas direct repair procedures were associated with rates of 3.8%, 1.6%, and 0.31%, respectively. Thus, although few comparisons are currently available in the literature, the results suggest that direct repair procedures for the treatment of LAI may be superior given the correct indications and patient expectations.
Arthroscopic repair for LAI is becoming increasingly popular. Primarily performed with suture anchors76-82, this minimally invasive procedure is expected to reduce postoperative pain and complications while promoting faster recovery. However, arthroscopic procedures have been associated with high complication rates and are technically challenging83. The indications for an arthroscopic procedure include a good-quality ligamentous remnant84.
Several studies have investigated the differences between arthroscopic and standard open procedures85-88. Giza et al. found no significant differences in terms of load to failure in a study of seven matched pairs of human cadaveric ankle specimens85. Lee et al. assessed the torque to failure, degrees to failure, and stiffness in a study of eleven matched pairs of human cadaveric specimens and reported no significant differences between an arthroscopic procedure and an open modified Broström procedure88. Drakos et al., in a study of twenty matched lower-extremity cadaveric specimens, found no significant differences between open and arthroscopic procedures86. On the basis of these results, arthroscopic repair appears to provide adequate ankle stabilization at time zero.
With respect to portal location, Drakos et al. examined the anatomy of the lateral aspect of the ankle after arthroscopic repair in ten lower-extremity cadaveric specimens and found that the proximity of the peroneus tertius and the extensor tendons to the ATFL and the intermediate branch of the superficial peroneal nerve posed a high risk of entrapment by the ATFL sutures86.
Outcomes of Arthroscopic Repair
To our knowledge, the only available reports on arthroscopic repair have been case series76-81,89. Nery et al., in what appears to represent the study with the longest-term follow-up, reported that 94.7% of patients had good-to-excellent clinical results at a mean of 9.8 years after an arthroscopic Broström-Gould repair76. No patient was reported to have delayed wound-healing, infection, or nerve damage76. Cottom and Rigby prospectively studied forty patients (forty ankles) who were managed with an all-inside Broström procedure77. After a mean duration of follow-up of 12.1 months, the VAS, Karlsson20, and AOFAS scores had improved significantly compared with the preoperative values (p < 0.05). Reported complications included deep-vein thrombosis (DVT), neuritis of the intermediate dorsal cutaneous nerve, and a distal fibular fracture (one patient each). Similarly, Vega et al., in a study of sixteen patients who were managed with all-inside arthroscopic repair, reported that the mean AOFAS score increased from 67 preoperatively to 97 points after a mean duration of follow-up of 22.3 months78. Complications included superficial infection and delayed wound-healing (one patient each).
High complication rates have been reported following arthroscopic repair in other studies as well. Wang et al., in a review of 178 patients (179 ankles) who were followed for a mean of 38.9 months after arthroscopic repair, found that thirty-one patients had complications, including nerve damage, DVT, delayed wound-healing, superficial infection, and distal fibular fracture83.
On the basis of the current evidence, it appears that arthroscopic repair can provide good clinical outcomes, although it is associated with a higher risk of complications, possibly because of a technical learning curve. Additional conclusions cannot be reached at this early junction, and further research, including comparative studies, is warranted.
Arthroscopic Thermal Shrinkage
The controlled heat of the arthroscopic thermal capsular and ligament shrinkage procedure results in reduction of the lengths of the capsule and ligament, theoretically stabilizing the ankle90,91. This procedure includes arthroscopic examination and treatment as well as shrinkage of the ATFL and adjacent capsule, typically via radiofrequency energy with approximately 70°C. Potential advantages include minimal invasiveness, shorter operative time, reduced morbidity of the ankle, and decreased complications compared with other procedures.
Candidates for the procedure include patients with an attenuated, but not avulsed, ATFL. However, it is difficult to describe the ideal candidate because of discrepancies in the inclusion criteria among the available reports, all of which represent low-quality evidence.
Outcomes of Arthroscopic Thermal Shrinkage
All of the available reports were case series without control groups, limiting the strength of the evidence1,92-98. All studies showed good short-term clinical results on the basis of AOFAS or Karlsson scores, although there were some conflicts among the reports with regard to ankle stability after the procedure.
In the largest case series, Ventura et al. retrospectively reviewed eighty-eight patients with chronic lateral ankle instability characterized by a positive anterior drawer sign95. At an average of four years postoperatively, the average AOFAS and Karlsson scores had significantly improved from 63.5 to 92.3 and from 61.8 to 88.4, respectively, and the number of patients still exhibiting a positive anterior drawer sign had decreased to five.
De Vries et al., in a prospective study of thirty-nine patients with positive anterior talar translation or talar tilt on stress radiographs, showed that talar tilt did not improve significantly but that the Karlsson score increased significantly after a mean duration of follow-up of nine months (p < 0.001)96. De Vries et al. postulated that the discrepancy between their study and the study by Ventura et al. in terms of ankle stability was the result of the different postoperative regimens that had been used. The patients in the study by de Vries et al. were allowed functional ambulation with full weight-bearing as early as five to seven days postoperatively as tolerated, whereas the patients in the study by Ventura et al. were managed with cast immobilization for two to six weeks.
To our knowledge, no studies have evaluated the long-term outcomes of the arthroscopic thermal shrinkage procedure and no comment has been made on the presence of degenerative arthritic changes. Therefore, caution should be exercised in utilizing this procedure without further controlled investigations to determine whether it is efficacious over the long term and does not provide simply a transient, short-term effect. The historical use of thermal capsular shrinkage in the shoulder ultimately failed, resulting in residual instability and requiring further surgery99.
The operative treatment of LAI remains a challenging, multifactorial topic of debate. Depending on the quality of the ligament remnant, various operative strategies can provide promising clinical outcomes. Arthroscopic procedures are increasing in popularity because they are expected to decrease complications and complexity. However, standard open procedures remain the standard of care. The majority of comparative studies performed to date have demonstrated no differences between open and arthroscopic procedures. Therefore, well-designed clinical trials with patient-specific, validated, and objective outcome measurements are required to develop evidence-based clinical guidelines for the future treatment of lateral ankle instability.
Investigation performed at the Hospital for Special Surgery, New York, NY
Disclosure: No external funding was received for this study. On the Disclosure of Potential Conflicts of Interest forms, which are provided with the online version of the article, one or more of the authors checked “yes” to indicate that the author (or the author’s institution) had a relevant financial relationship in the biomedical arena outside the submitted work.
- Copyright © 2016 by The Journal of Bone and Joint Surgery, Incorporated