THE OPERATIVE TREATMENT OF ACQUIRED HALLUX VARUS
June 24th, 1993
Thomas C. Skalley, M.D. and Mark S. Myerson, M.D.
The deformity of hallux varus is often associated with functional impairment, pain, difficulty with shoewear, and cosmetic dissatisfaction.3,4,34,50 Hallux varus occurs as a congenital3,8,10-12,17,21,22,35,41,45,47,49 or an acquired deformity.1,7,9,10,13-25,27,30,33,34,36,40,41,50-53 The etiology of acquired hallux varus is quite varied, and includes hallux valgus surgery,1,7,9,10,13-20,23-25,27,30,33,34,40,41,50-53 trauma,15,17,21,30,36,40,51 inflammatory arthritis,16,17,21,22,40 infection,51 neuromuscular imbalance,16,17,40,51 Charcot disease,16 idiopathy in the unshod foot,22,24,27 and burn contractures.21 The most commonly reported cause of acquired hallux varus is attempted correction of hallux valgus3,7,10,17,20,23,27,30,51 and the method most frequently associated with postoperative acquired hallux varus is the McBride bunionectomy.1,3,7,9,10,17,20,23,30
Although the literature describes both conservative10,40,52 and operative1,4,7,10,14-17,19,20,23-25,30,34,41,50-53 treatment methods for the management of acquired hallux varus, to our knowledge there is no report of a large series of patients treated via a comprehensive approach to the surgical management of this problem. We present an algorithm for the operative management of acquired hallux varus.
Materials and Methods
We retrospectively reviewed 54 patients whom we treated for acquired hallux varus deformity between 1986 and 1991. A variety of conservative treatment alternatives, including splinting and taping of the hallux, shoe modifications, and nonsteroidal antiinflammatory medications, were attempted in 26 (48%) of those patients. The other 28 patients had been treated nonoperatively elsewhere or had attempted some of these remedies themselves. Nonoperative treatment was moderately successful in 12 patients. The 42 remaining patients (45 feet) with symptomatic acquired hallux varus who were treated operatively formed the study group.
The etiology of the hallux varus was determined with particular attention to past surgical procedures, trauma, and the presence of degenerative or inflammatory arthritis. Of the 45 feet, 36 (80%) had undergone hallux valgus surgery [two of which initial procedures were performed by the senior author (MM); the other initial procedures were performed elsewhere], including 13 McBride procedures, 10 distal first metatarsal osteotomies, six proximal first metatarsal osteotomies, five simple bunionectomies (exostectomies), and two resection arthroplasties (Keller procedures). The etiology of hallux varus in the remaining nine feet was trauma in three, rheumatoid arthritis in three, idiopathy in two, and osteoarthritis in one.
The foot symptoms reported were multiple and varied, including hallucal pain (43), shoewear difficulty from medial rubbing or a poor fit (25), weakness with push-off (seven), a sensation of instability at the metatarsophalangeal (MP) joint (three), and metatarsalgia (three). The locations of the one or more sites of pain in each foot were the medial aspect of the hallux (32), the MP joint (15), the interphalangeal (IP) joint (six), the dorsomedial hallux (one), and the plantar-medial hallux (one).
On physical examination, a claw deformity of the hallux characterized by extension of the MP joint and flexion of the IP joint was noted in three feet. Painful medial hallucal callosities were present in 19 feet.
A weight-bearing preoperative anteroposterior roentgenogram was used to determine the MP joint angle, measured from the angle of intersection of a line drawn through the longitudinal axis of the first metatarsal with a line through the longitudinal axis of the proximal phalanx. The angle of the hallux MP joint averaged 23o of varus (range, 10--45o of varus) on preoperative roentgenographic evaluation.
Surgery was performed in the 45 feet after conservative treatment failed to alleviate symptoms. The algorithm of surgical decision making is presented in Figure 1. A tendon transfer was performed if the MP joint was congruent and mobile. A congruent joint was defined as one which maintained its range of motion when the hallux was passively corrected into neutral alignment. The joint was incongruent if crepitus was present or the range of motion decreased upon passive correction of the hallux from a position of varus to one of neutral. The extensor hallucis longus (EHL)19,20,30 or the extensor hallucis brevis (EHB)37,38 were used as tendon transfers. As originally described by Johnson and Spiegl,20 the entire EHL tendon was used in conjunction with an IP joint arthrodesis in the presence of a fixed and painful deformity of the hallux IP joint. If no fixed claw deformity of the hallux was present, a split EHL tendon transfer19,30 was performed in the earlier study patients; later in the series, the EHB tendon transfer37,38 was used.
With the exception of arthrodesis, surgical correction of the varus deformity required proper balancing of the forces across the MP joint. Meticulous attention was therefore given to the medial and lateral soft-tissue balance of the hallux MP joint. A medial soft-tissue release, including capsulotomy and lengthening or tenotomy of the abductor hallucis contracture, was performed in each case. If the varus was caused by a tight capsulorrhaphy only, correction was obtained with the medial capsulotomy and abductor lengthening alone in mild deformities, and with an additional tendon transfer in moderate to severe deformities. These methods of soft-tissue reconstruction were used when the overall alignment of the first metatarsal was satisfactory. In patients with arthritis, either an arthrodesis or a resection arthroplasty was performed, depending on the age and activity level of the patient. Inflammatory arthritis was treated with an arthrodesis. If the varus was caused by overcorrection of a metatarsal osteotomy, the malunion was corrected if the joint was congruent. The osteotomy was performed at the apex of the metatarsal deformity according to the original procedure. The revision metatarsal osteotomy was combined with a tendon transfer if the MP joint was unstable, or if the varus was not fully corrected after the osteotomy.
Treatment for the feet in this study group included: arthrodesis, 17; Keller resection arthroplasty, seven; EHL transfer with IP arthrodesis, seven; split EHL transfer, six; EHB transfer, four; medial soft-tissue release alone, two; and distal first metatarsal osteotomy, two (one with and one without additional tendon transfer).
All patients were examined at follow-up averaging 3.8 years (range, 1--7 years). Subjective complaints, physical findings, MP joint angles on standing roentgenograms, and patient satisfaction were evaluated.
No pain was present preoperatively or at follow-up in the two patients treated by medial soft-tissue release alone. Pain was improved in every other treatment group. In the five patients reporting pain at follow-up, four had pain due to the position of an arthrodesis, and one had pain at the site of the resection arthroplasty. Shoewear problems were also improved in each treatment group. Weakness at push-off, if present, was improved after all three types of tendon transfers and arthrodesis. No improvement was found in the one patient with preoperative weakness at push-off who underwent a resection arthroplasty. The three cases of MP instability were treated successfully with a split EHL tendon transfer in two and an EHL tendon transfer with IP arthrodesis in one. Metatarsalgia, present in three patients preoperatively, was unchanged by arthrodesis, and developed postoperatively in one patient after a resection arthroplasty. Additional complaints of MP joint stiffness were recorded at follow-up after split EHL transfer (three) and EHL transfer with IP arthrodesis (three).
On physical examination, painful medial hallucal callosities, if present preoperatively, were improved with each type of treatment. In the three patients with a preoperative claw deformity of the hallux, two were treated successfully with EHL tendon transfer and IP arthrodesis, and one continued to have an asymptomatic claw deformity after resection arthroplasty.
Limitations in MP joint motion were noted after both types of EHL tendon transfer techniques. Dorsiflexion averaged 38o (range, 20--70o) and 35o (range, 20--45o), and plantar flexion averaged 5o (range, 0--20o) and 6o (range, 0--10o), respectively, for the split EHL transfer and EHL transfer with IP arthrodesis. Dorsiflexion (average, 70o; range, 60--80o) and plantar flexion (average, 21o; range, 10--30o) motion was well preserved by EHB tendon transfer technique alone and, in one patient, by EHB tendon transfer combined with a distal osteotomy (dorsiflexion, 65o; plantar flexion; 30o). No significant limitations in dorsiflexion (mean, 71o; range, 60--80o) or plantar flexion (mean, 11o; range, 10--20o) motion were present after resection arthroplasty.
The MP angle improved in all feet from a mean 23o of varus to a mean 3o of valgus (range, 5o of varus to 15o of valgus). The mean correction in each foot was 26o (range, 5--50o). Roentgenographic union was noted in 16 of 17 arthrodesed MP joints. The single nonunion which occurred was asymptomatic.
Thirty-nine patients reported good or excellent satisfaction (Table 5) after medial soft-tissue release alone (two), split EHL tendon transfer (four), EHL tendon transfer with IP joint arthrodesis (four), EHB tendon transfer (four), distal metatarsal osteotomy (two), arthrodesis (16), and resection arthroplasty (seven). The one patient who reported poor satisfaction developed degenerative joint disease after an EHL transfer and IP joint arthrodesis, which required an eventual arthrodesis. Four patients reported fair satisfaction after EHL tendon transfer, all due to symptomatic MP joint stiffness. One patient reported fair satisfaction after arthrodesis because of painful rubbing of the hallux on the second toe.
Three patients developed minor postsurgical wound infections, which were treated with oral antibiotics. Other than the one asymptomatic nonunion after MP arthrodesis, there were no other specific complications in this patient group.
Congruent motion at the MP joint and mediolateral soft-tissue balance are integral to function of the hallux after bunionectomy procedures. Several factors have been attributed to hallux varus after forefoot surgery, including removal of the fibular sesamoid,1,7,14-18,20,23-25,30,34,50,52,53 excessive medial capsular reefing,1,7,14,15,28,30,51 removal of an excessive amount of the medial eminence,1,7,15,18,25,30,41,51 overcorrection of the intermetatarsal angle,7,15,18,30,51 medial tibial sesamoid subluxation,1,7,14,18,25,28,30 excessive plantar-lateral release,1,10,15,17 and excessive postoperative bandaging.1,7,30,51 In our patients, these and other causes were identified, including avascular necrosis of the first metatarsal head with subsequent collapse into varus, osteoarthritis, and malunion of a distal first metatarsal osteotomy.
As originally described, the McBride procedure32,33 included a sesamoidectomy as part of the forefoot correction, which has since been identified as the most common cause of subsequent hallux varus.1,3,7,9,10,17,20,23,30 Detachment of the adductor hallucis and lateral head of the flexor hallucis brevis tendon in combination with a tight medial capsular reefing is also a common cause of hallux varus when the sesamoid is left behind.10,14
Inflammatory arthritis typically causes hallux valgus,46 but hallux varus has also been reported in rheumatoid arthritis.16,17,21,40 Less commonly, varus is associated with degenerative osteoarthritis of the hallux MP joint, as seen in one of our patients. Trauma to the hallux, present in three of our patients, is an unusual cause of acquired hallux varus.16,17 One of these patients, a gymnast, dislocated the MP joint, resulting in recurrent varus instability on push-off. In the other two patients, varus occurred after a stubbing injury to the hallux; one of these patients developed posttraumatic degenerative joint disease.
The surgical approach to correction can be simplistically divided into procedures that primarily address the soft tissues (e.g. release and tendon transfer), the bone (e.g. metatarsal osteotomy), or the joint (e.g. resection arthroplasty and arthrodesis). Common to all of these procedures is correction of the underlying deformity and restoration of soft-tissue balance to the MP joint. Although correction with soft-tissue releases alone has been described,1,7,30,51 medial capsulotomy and abductor hallucis lengthening alone are rarely successful. This approach may be useful in a mild deformity, particularly when the primary etiology is an excessively tight previous capsulorraphy,19 but it does not adequately restore soft-tissue balance in moderate to severe hallux varus deformities. In such situations, the addition of a tendon transfer is necessary for successful correction of the deformity. Unlike what others have reported,1,7,25,52,53 we did not find it necessary to perform a tibial sesamoidectomy in any patient for correction of the deformity.
The addition of a tendon transfer to the medial soft-tissue release is recommended in most cases of hallux varus in young, active patients with no arthritis and good range of motion at the MP joint. The tendon transfers used initially in this series were the EHL transfer with IP arthrodesis described by Johnson and Spiegl20 and a split EHL transfer without IP arthrodesis described by Johnson19 and Mann and Coughlin.30 Symptomatic MP joint stiffness was frequently reported in our patients (43 and 50%, respectively, in the EHL transfer with IP arthrodesis and split EHL transfer groups). Additionally, one patient complained of weak active hallux dorsiflexion after EHL transfer and IP arthrodesis. Although described as providing dynamic balance to the MP joint,20 we believe that in most cases a static rather than a dynamic correction was obtained in our patients with these techniques.
In an attempt to decrease some of the morbidity associated with the EHL transfer, yet still achieve the same functional correction, the senior author developed the EHB transfer.37,38 Despite the shorter follow-up period, we have identified a decreased incidence of problems with dorsiflexion and MP joint stiffness after this technique.
Several authors have reported on the use of distal osteotomies to correct hallux varus, including a lateral metatarsal neck closing wedge (reverse Reverdin)1,3,7,10,52 and a reverse Austin1,7,52 (reverse Chevron) osteotomy. These osteotomies work, in principle, if the original procedure was a distal and not a proximal metatarsal osteotomy. Provided the deformity is in the distal metatarsal and the MP joint is congruent, any distal metatarsal osteotomy may conceivably correct the deformity. Because of potential further shortening of the metatarsal with a lateral closing wedge (Reverdin) or Chevron-type osteotomy,29,42 correction was obtained in our patients with a dome osteotomy at the apex of the distal deformity.
Although proximal metatarsal osteotomies have been described for the correction of hallux varus,1,52 none were performed in this study. In the six patients with hallux varus after a prior proximal metatarsal osteotomy, the etiology was overcorrection at the osteotomy site in two, and excessive tightening of the medial capsulorrhaphy in four. The two patients with overcorrection of the metatarsal osteotomy were noted to have concomitant degenerative joint disease of the MP joint and were treated with arthrodesis. The other four patients were treated successfully with a medial soft-tissue release, either alone (two) or with an additional EHB tendon transfer (two).
Correction of hallux varus may also be obtained at the MP joint using implant arthroplasty,1,3,7,10,50,52,53 resection arthroplasty,1,3,17,34,50,52 and arthrodesis.17,30,50 Implant arthroplasty was not used in our patients because of previously reported problems, including silicone synovitis,17 delayed wound healing with skin necrosis,43 infection,17 potential for lymphatic uptake,2,6,26,44 malignant lymphoma associated with intranodal silicate lymphadenopathy,2,6 and the risk of prosthetic fracture, erosion, and dislocation.17,43 Rheumatoid or inflammatory arthritis, avascular necrosis with collapse of the metatarsal head, and degenerative arthritis were all treated at the site of the deformity. The treatment of rheumatoid arthritis involving the MP joint has proponents for implant arthroplasty,5 resection arthroplasty,17 and arthrodesis.31,39,48 Rheumatoid patients in this series were treated with arthrodesis because of their severe varus deformity and significant medial soft-tissue contraction, and because of the increased stability31,39,48 afforded by arthrodesis.
Avascular necrosis of the metatarsal head with subsequent varus collapse at the joint was seen in one patient after a Chevron procedure. As advocated by other authors for the treatment of avascular necrosis,30,42 arthrodesis was performed, resulting in good patient satisfaction.
Patients with degenerative joint disease involving the MP joint were evaluated by age and activity level. Patients felt to be physiologically older than 65 years with sedentary lifestyles were treated with resection arthroplasty. For patients physiologically under 65 years or for active patients of any age, arthrodesis was recommended. In this study, patients undergoing resection arthroplasty averaged 66.3 years (range, 63--69 years); those undergoing arthrodesis averaged 49.5 years (range, 29--67 years). In this study, both procedures provided a stable outcome over time, with 16 (94%) arthrodesis and seven (100%) resection arthroplasty patients reporting good or excellent results, respectively, at follow up.
In conclusion, the etiology of acquired hallux varus is quite varied. We have determined surgical correction according to the underlying deformity, etiology, age, and activity level, and have provided an algorithm for treatment.
1. Banks, A. S., Ruch, J. A., and Kalish, S. R.: Surgical repair of hallux varus. J. Am. Podiatry Assoc. 78:339, 1988.
2. Benjamin, E., Ahmed, A., Rashid, A. T. M. F., and Wright, D. H.: Silicone lymphadenopathy: a report of two cases, one with concomittant malignant lymphoma. Diagn. Histopathol. 5:133, 1982.
3. Bilotti, M. A., Caprioli, R., Testa, J., Cournoyer, R., and Esposito, F. J.: Reverse Austin osteotomy for correction of hallux varus. J. Foot Surg. 26:51, 1987.
4. Clark, C. W. D.: Abductor hallucis tendon transfer for hallus varus. J. Foot Surg. 23(2):146, 1984.
5. Cracchiolo, A., III: Results of implant arthroplasty in the arthritic forefoot (abstr). Orthop. Trans. 4:150, 1980.
6. Digby, J. M.: Malignant lymphoma with intranodal silicone rubber particles following metacarpophalangeal joint replacements. Hand 14:326, 1982.
7. Edelman, R. D.: Iatrogenically induced hallux varus. Clin. Podiatry Med. Surg. 8(2):367, 1991.
8. Farmer, A. W.: Congenital hallux varus. Am. J. Surg. 95:274, 1958.
9. Feinstein, M. H. and Brown, H. N.: Hallux adductus as a surgical complication. J. Foot Surg. 19:207, 1980.
10. Greenfogel, S. I., Glubo, S., Werner, J., Sherman, M., and Lenet, M.: Hallux varus -- surgical correction and review of the literature. J. Foot Surg. 23:46, 1984.
11. Gurin, J. and Szoke, G.: Management of hallux varus. J. Magyar Traumatol. Orthop. Helyreallito Sebescet. 333(2):93, 1990.
12. Haas, S. L.: An operation for the correction of hallux varus. J. Bone Joint Surg. 20:705, 1938.
13. Hansen, C. E.: Hallux valgus treated by the McBride operation. A follow up. Acta Orthop. Scand. 45:778, 1974.
14. Hawkins, F. B.: Acquired hallus varus: cause, prevention and correction. Clin. Orthop. 76(May):169, 1971.
15. Hunter, W. N. and Wasiak, G. A.: Traumatic hallux varus correction via split extensor tenodesis. J. Foot Surg. 23:321, 1984.
16. Jahss, M. H.: Spontaneous hallux varus: relation to poliomyelitis and congenital absence of the fibular sesamoid. Foot Ankle 3(4):224, 1983.
17. Jahss, M. H.: Disorders of the hallux and first ray. In Disorders of the Foot, edited by Jahss, M. H. Ed. 2. Philadelphia, W.B.Saunders, 1991, pp. 1078-1089.
18. Janis, L. R. and Donick, I. I.: The etiology of hallux varus: a review. J. Am. Podiatry Assoc. 65(3):233, 1975.
19. Johnson, K. A.: Hallux valgus surgery dissatisfaction. In Surgery of the Foot and Ankle, vol 1, edited by Mann, R. A. and Coughlin, M. J. Ed. 6. New York, Raven Press, 1989, pp. 50-59.
20. Johnson, K. A. and Spiegl, P. V.: Extensor hallucis longus transfer for hallux varus deformity. J. Bone Joint Surg. 66A(5):681, 1984.
21. Joseph, B., Chacko, V., Abraham, T., and Jacob, M.: Pathomechanics of congenital and acquired hallux varus: a clinical and anatomical study. Foot Ankle 8(3):137, 1987.
22. Joseph, B., Jacob, T., and Chako, V.: Hallux varus -- a study of 30 cases. J. Foot Surg. 23:392, 1984.
23. Katz, J. B.: Correction of hallux varus via split tendon transfer. J. Am. Podiatry Assoc. 80(9):498, 1990.
24. Kimizuka, M. and Miyanaga, Y.: The treatment of acquired hallux varus after the McBride procedure. J. Foot Surg. 19(3):135, 1980.
25. Langford, J. H. and Maxwell, J. R.: A treatment for post surgical hallux varus. J. Am. Podiatry Assoc. 77(3):142, 1982.
26. Lim, W. T., Landrum, K., and Weinberger, B.: Silicone lymphadenitis secondary to implant degeneration. J. Foot Surg. 22:243, 1983.
27. Makhani, J. S.: Hallux varus. J. Bone Joint Surg. 51B:781, 1969.
28. Mann, R. A.: Complications in surgery of the foot. Orthop. Clin. North Am. 7:851, 1976.
29. Mann, R. A.: Excessive shortentage III), no reactivation of this process occurred in the midfoot, regardless of the initial treatment. Seven patients, however, experienced a second focus of neuroarthropathy in the hindfoot and ankle an average of 3 years (range, 1 to 5 years) after initial treatment of the midfoot. There was no consistent pattern in these seven patients to conclude that either the initial presentation or treatment was the cause of the subsequent neuroarthropathy.
At the time of their most recent follow-up examination, 82 (95%) of 85 feet were considered to be stable. Three had not reached a stable state, and four patients had died without sufficient follow-up to evaluate their feet.
There were 24 instances in which a change in the treatment plan was necessary. Seventeen (29%) of the 58 feet initially managed with a total-contact cast subsequently underwent surgical treatment (six ostectomies, six arthrodeses, and five amputations). Six of the 22 feet initially treated in a brace ultimately required arthrodesis. One patient who was initially treated surgically for an abscess later required a revision procedure.
The treatment goal for patients with neuroarthropathy of the midfoot is to obtain a durable plantigrade foot for ambulation. Amputation remains a necessary management option in cases of infection or ischemia. However, by controlling the rate and severity of ulceration, a carefully designed treatment protocol can decrease.