STAGED ARTHRODESIS FOR SALVAGE OF THE SEPTIC HALLUX METATARSOPHALANGEAL JOINT
June 22nd, 1994
Mark S. Myerson, MD; Stuart D. Miller, MD;
Michael R. Henderson, M.D; and Terence Saxby, FRACS
Abstract
Failed surgery at the hallux metatarsophalangeal joint may present substantial difficulties in treatment, especially when complicated by infection. This retrospective study reviews the staged treatment of five patients with complications of hallux valgus surgery associated with sepsis of the metatarsophalangeal joint. The initial salvage treatment included debridement with placement of an antibiotic cement spacer and either an intramedullary Kirschner wire (four patients) or an external fixator (one patient). At the second stage procedure, the spacer and fixation were removed and an autogenous tricortical iliac crest graft was inserted into the joint. All patients had clinical control of joint sepsis. Pseudoarthrosis occurred at the proximal end of the tricortical graft in two patients. Patients were evaluated at an average of 28 months after the fusion. Staged arthrodesis appears to be a satisfactory approach to postoperative sepsis of the hallux metatarsophalangeal joint.
Introduction
The salvage of failed surgery on the hallux metatarsophalangeal (MP) joint presents the surgeon with many challenges, including management of destructive changes in the hallux metatarsal head or proximal phalanx (which, uncorrected, can lead to a decrease in weight-bearing function of the first ray and subsequent lateral metatarsalgia5,8), restoration of length to the shortened hallux or first metatarsal, and/or eradication of any infection.
Arthrodesis of the MP joint14-16,20 has been used as treatment for avascular necrosis of the metatarsal head4 and salvage for failed resection arthroplasty5 or failed silastic arthroplasty.19 When failure of silastic prosthetic arthroplasty of the hallux MP joint is associated with bone loss and synovitis, salvage can be difficult, and treatment recommendations include arthrodesis of the first MP joint with incorporation of an autogenous bone block graft,19 removal of the implant with synovectomy,12 or bone-block arthrodesis.1 The incidence of infection after bunion surgery is low.10,20 Recommendations for treating this infection includes open drainage, debridement, antibiotics, and rest.10 However, arthrodesis has not been used for chronic infection of the hallux MP joint, except in one report of a patient with tuberculosis.20
We present a retrospective review of five patients treated with staged arthrodesis of the hallux MP joint for salvage of bunion surgery complicated by sepsis.
Materials and Methods
Patient Population
We treated five patients (four females, one male) with an average age of 45 (range, 28 to 54 years) for complications of hallux valgus surgery associated with infections of the MP joint. Prior surgeries included bunionectomy and Chevron osteotomy in two, bunionectomy and silastic arthroplasty in two, and bunionectomy and total joint replacement in one. All patients had septic arthritis of the MP joint. Four patients had osteomyelitis, and in two there were findings of avascular necrosis on plain radiographs. Diagnosis of infection was made on the results of the sedimentation rate, white blood cell count, plain radiographs, and positive bone or joint cultures at the time of surgery. The average time to salvage surgery was 6.4 months (range, 2 to 11 months) after the initial surgical procedure.
Technique
Reconstruction was carried out in two stages, each through the same dorsal longitudinal approach with elevation of full-thickness subperiosteal flaps. A tourniquet was not utilized.
The initial procedure consisted of an aggressive debridement: necrotic bone was removed using currettes and rongeurs, and inflammatory cysts were curetted, but unnecessary bone resection was avoided. A high-speed 4-mm oval burr was used for bone debridement, with irrigation to minimize further bone necrosis. Two or three threaded 0.62-inch Kirschner wires were inserted in a retrograde fashion from the base of the proximal phalanx out the distal end of the toe, but they did not protrude through the base of the proximal phalanx at this stage. An antibiotic-impregnated polymethylmethacrylate (PMMA) spacer was prepared by mixing 2 g of tobramycin powder with a 40-g packet of cement powder (Simplex P).
The hallux was then manually and longitudinally distracted, and the PMMA bead was inserted to bridge the gap between the proximal phalanx and the distal metatarsal (Fig. 1B). Tension on the soft tissues should be minimal. To facilitate later removal, extravasation of cement down the metatarsal was avoided. While the cement still had a doughy texture, the previously placed Kirschner wires were passed across the cement spacer and into the cortex of the metatarsal shaft (Fig. 1C). This technique was used in four patients. One patient, with severe sepsis and an open draining wound (Fig. 2, B and C), was treated with a mini-external fixation device with two pins in the metatarsal and two in the proximal phalanx. Both methods provided stability to the first ray during this first phase of reconstruction. The wound was then closed with interrupted monofilament sutures. Although primary wound closure is usually not a problem with these patients, significant tension of wound edges after distraction could preclude wound closure, in which case an external fixator could be used to provide initial stability and wound access.
A 3- to 5-week interval between the first and second procedures allowed elution of antibiotic and subsidence of the inflammatory process. During this interval, outpatient intravenous antimicrobial therapy was continued for 3 weeks, after which the need for additional antibiotic therapy was determined by the nature of the organism and extent of sepsis. In addition, patients wore a short-leg cast or a stiff-soled surgical shoe with a raised heel.
The second procedure utilized the same dorsal incision. The wires were removed before withdrawing the PMMA block. During this stage, we experienced no wire breakage secondary to fixation in the cement. An autogenous tricortical iliac crest graft was taken and sized according to the longitudinal measurements with the toe maximally distracted. (If needed, additional stability may be obtained by creating a football-shaped graft 4 to 6 mm longer than the gap between the proximal phalanx and distal metatarsal.) The high-speed burr was used to taper the graft at each end to ensure a snug fit within the medullary canals of both the metatarsal and proximal phalanx.
Additional cancellous bone graft was placed laterally to fill the space around the strut graft. In four patients, we used internal fixation to stabilize the graft with a dorsally placed one-third or one-quarter tubular plate. A longitudinally inserted Kirschner wire was used provisionally to maintain the position of the hallux while the plate was contoured. (Fully threaded 4.0-mm cancellous screws may be used to maximize purchase in generally osteopenic bone.) Longitudinal compression was then applied manually to the hallux while the screw holes were filled (Fig. 1D). In the fifth patient, we used threaded 7/64-inch Steinmann pins since the size of the bone block precluded the use of a plate.
After the second procedure, patients were allowed to bear weight on the heel only, either in a short leg cast with a heel or with a stiff surgical shoe with elevated heel wedge. Full plantigrade weight-bearing began only after radiographic evidence of graft incorporation, usually at 3 to 4 months after surgery (Fig. 1E). Organism-specific antibiotics were continued for 1 week after the second procedure.
Results
Fixation consisted of a dorsal plate and screws in four patients, and in one of these (#4) an obliquely inserted threaded 0.062-inch Kirschner wire was used for additional stability. One graft was fixed with threaded 7/64-inch Steinmann pins and went on to arthrodesis at 16 weeks. The average bone graft size was 15.2 mm (range, 11 to 22 mm). None of the four patients with Staphylococcus infections nor the patient with a Pseudomonas infection demonstrated signs of a joint sepsis at final follow-up.
Three patients achieved initial arthrodesis in an average of 13.8 weeks (range, 11 to 16 weeks). Two patients demonstrated nonunion at the proximal end of the autogenous bone graft block. One of these (patient #3) underwent revision surgery at 8 weeks for nonunion and plate failure. The broken plate was removed and a 3-mm burr was used to create bleeding bone surfaces dorsally and laterally. A small amount of cancellous bone graft was obtained from the calcaneus and packed around the lateral joint margin. Fixation was performed with a cortical screw inserted obliquely from proximal and plantar to distal and lateral. Fusion was achieved in 8 weeks. The second patient (#5) with a pseudoarthrosis is asymptomatic 2.5 years after attempted fusion.
Patients were evaluated at a mean of 28 months (range, 14 to 42 months) after salvage surgery with weight-bearing radiographs and clinical examination. The recently developed clinical rating system for the foot and ankle (Kitaoka, H., Alexander, I.J., Adelaar, R.S., Nunley, J.A., Myerson, M.S., and Sanders, M.: Clinical rating systems for the ankle/hindfoot, midfoot, hallux, and lesser toes. Presented at the Annual AOFAS Meeting, Ashville, NC, June 21, 1993) was used to evaluate these patients. This 100-point scale assigns 40 points for pain, 45 for function, and 15 for alignment. Since this study was not performed prospectively, valid preoperative scores are not available.
Discussion
The reconstructive technique of staged arthrodesis for sepsis of the hallux MP joint is performed to provide a stable first ray that can participate in weight-bearing and to eradicate infection. The method is modeled after salvage techniques in total knee arthroplasty complicated by infection.22,23 In the initial stage, the focus of the inflammatory and infective process is removed. Debridement should be extensive; adequate retraction and sharp tools should be available. A temporary spacer of bone cement is placed, preserving soft-tissue tension and length, while leaving a well-prepared bed of granulation tissue for later bone-graft incorporation.11,22,23 PMMA provides a convenient vehicle for antibiotic treatment in cases of osteomyelitis and it also has sufficient strength for a spacer. The integration of tobramycin powder into the PMMA at time of mixing has been well documented from the total joint literature and the ratio of antibiotic to cement has ranged from 1.0 to 2.0 g/40-g packet of cement.6,7,18,21,22 Alternatively, in patients allergic to tobramycin, vancomycin has been found to elute well from PMMA.13 The use of the antibiotic spacer in the first stage of reconstruction resulted in a rapid reduction in the symptoms of inflammation at the MP joint. It is quite possible that the antibiotic-impregnated cement spacer may not need any temporary internal fixation until the second stage is performed. We believed, however, that the added rigidity would enhance healing and decrease inflammation. Since we did not experience any complications associated with these wires, we would continue to recommend their use.
The staged method of arthrodesis quieted the inflammatory process well in these patients. Interposition iliac crest graft has been recommended as a salvage method after failed Keller procedure and after failed silicone arthroplasty.19 A potential problem with these single-stage restorations of bone length may occur in the presence of severe focal sepsis or synovitis. The local inflammation and possible infection could interfere with process of arthrodesis. The maintenance and reconstitution of first metatarsal length is needed to restore proper weight distribution in the foot and thus avoid lateral metatarsalgia.5,14 Corticocancellous graft appears to restore length and function of the first ray.5,19 The methods of fixation of the metatarsal and graft are variable; both a dorsal plate and intramedullary wires have been used.2,3,5,9 Although a plate may be used dorsally, it is applied to the compression and not the tension side of the joint and is biomechanically a poor construct.17 Fixation of the large graft is difficult to perform with crossed screws, and larger intramedullary wires may be preferred as previously recommended.5 In the light of the two failures (patients #3 and 5), we would recommend the use of intramedullary threaded Steinmann pins. These pins cross the hallux interphalangeal joint and can potentially cause hallux interphalangeal arthritis. From a standpoint of fixation, however, they are the easiest to insert and are associated with a satisfactory rate of arthrodesis.5
These results should be compared with an alternative approach to salvage with simple removal of infected material by creating a resection arthroplasty. A recent study investigated implant removal and synovectomy in 10 patients with failed first MP arthroplasties.12 The seven excellent and one good result demonstrate indications for simple removal without arthrodesis. The authors suggest possible need for a later fusion procedure and caution that their goal was not to compare resection arthroplasty to arthrodesis.
The results of this staged arthrodeses for sepsis of the hallux MP joint in this study supports a careful approach to these difficult reconstructive problems. While the current follow-up of 28 months is less than optimal to discern any weakness in the incorporated graft segment, the patients appeared to be doing well without a failure of the graft. Further long-term studies will be necessary to better understand the ultimate success of this technique.
References
1. Alexander, I.J.: Arthrodesis of the metatarsophalangeal and interphalangeal joints of the hallux. In Myerson, M. (ed.): Current Therapy in Foot and Ankle Surgery, St. Louis, Mosby-Year Book, Inc, 1993, pp. 81-90.
2. Coughlin, M.J.: Arthrodesis of the first metatarsophalangeal joint with mini-fragment plate fixation. Orthopedics 13(9):1037-1044, 1990.
3. Coughlin, M.J.: Arthrodesis of the first metatarsophalangeal joint. Orthop. Rev. 19(2):177-186, 1990.
4. Coughlin, M.J., Baxter, D.E., Bordelon, R.L., Cracchiolo, A., Johnson, K.A., and Mann, R.A.: Symposium: complications and salvage of bunion surgery: case discussion. Contemp. Orthop. 23(2):150-172, 1991.
5. Coughlin, M.J. and Mann, R.A.: Arthrodesis of the first metatarsophalangeal joint as salvage for the failed Keller procedure. J. Bone Joint Surg. 69A(1):68-75, 1987.
6. Curtis, M., Myerson, M., Jinnah, R.H., Cox, Q.G.N., and Alexander, I.: Arthrodesis of the first metatarsophalangeal joint: a biomechanical study of internal fixation techniques. Foot Ankle 14(7):395-399, 1993.
7. Goodell, J.A., Flick, A.B., Hebert, J.C., and Howe, J.G.: Preparation and release characteristics of tobramycin-impregnated polymethylmethacrylate beads. Am. J. Hosp. Pharm. 43(6):1454-1461, 1986.
8. Harrison, M.H.M. and Harvey, F.J.: Arthrodesis of the first metatarsophalangeal joint for hallux valgus and rigidus. J. Bone Joint Surg. 45A(3):471-480, 1963.
9. Henry, A.P, Waugh, W., and Wood, H.: The use of footprints in assessing the results of operations for hallux valgus. A comparison of Keller's operation and arthrodesis. J. Bone Joint Surg. 57B(4):478-481, 1975.
10. Holmes, G.B., Jr.: Arthrodesis of the first metatarsophalangeal joint using interfragmentary screw and plate. Foot Ankle 13(6):333-335, 1992.
11. Kenzora, J.E.: Complications in bunion surgery. Complicat. Orthop. May/June:79-89, 1990.
12. Kitaoka, H.B., Holiday, A.D.,Jr., Chao, E.Y.S., and Cahalan, T.D.: Salvage of failed first metatarsophalangeal joint implant arthroplasty by implant removal and synovectomy: clinical and biomechanical evaluation. Foot Ankle 13(5):243-250, 1992.
13. Lawson, K.J., Marks, K.E., Brems, J., and Rehm, S.: Vancomycin vs tobramycin elution from polymethylmethacrylate: an in vitro study. Orthopedics 13(5):521-524, 1990.
14. Mann, R.A. and Oates, J.C.: Arthrodesis of the first metatarsophalangeal joint. Foot Ankle 1(3):159-166, 1980.
15. Mann, R.A. and Thompson, F.M.: Arthrodesis of the first metatarsophalangeal joint for hallux valgus in rheumatoid arthritis. J. Bone Joint Surg. 66A(5):687-692, 1984.
16. McKeever, D.C.: Arthrodesis of the first metatarsophalangeal joint for hallux valgus, hallux rigidus, and metatarsus primus varus. J. Bone Joint Surg. 34A:129-134, 1952.
17. Moynihan, F.J.: Arthrodesis of the metatarso-phalangeal joint of the great toe. J. Bone Joint Surg. 49B(3):544-551, 1967.
18. Pritchett, J.W. and Bortel, D.T.: Tobramycin-impregnated cement in total hip replacements. Orthop. Rev. 21(5):577-579, 1992.
19. Sammarco, G.J. and Tabatowski, K.: Silicone lymphadenopathy associated with failed prosthesis of the hallux: a case report and literature review. Foot Ankle 13(5):273-276, 1992.
20. Thompson, F.R. and McElvenny, R.T.: Arthrodesis of the first metatarsophalangeal joint. J. Bone Joint Surg. 22(3):555-558, 1940.
21. Trippel, S.B.: Antibiotic-impregnated cement in total joint arthroplasty. J. Bone Joint Surg. 68A(8):1297-1302, 1986.
22. Wilde, A.H. and Ruth, J.T.: Two-stage reimplantation in infected total knee arthroplasty. Clin. Orthop. 236(Nov):23-35, 1988.
23. Windsor, R.E., Insall, J.N., Urs, W.K., Miller, D.V., and Brause, B.D.: Two-stage reimplantation for the salvage of total knee arthroplasty complicated by infection. J. Bone Joint Surg. 72A(2):272-278, 1990.
Figure Legends
Fig. 1. Patient #1. A, preoperative radiographic appearance of septic arthritis 6 months after bunionectomy. B, PMMA spacer in place. C, radiograph of Kirschner wire holding spacer in place. D, anteroposterior radiograph of graft and dorsal plate. E, lateral radiograph with good healing.
Fig. 2. Patient #2. A, preoperative radiographic appearance of osteomyelitis, avascular necrosis, and septic arthritis 4 months after bunionectomy. B, postoperative appearance with external fixator. C, radiographic appearance with fixator in place and PMMA spacer. D, graft in place after Steinmann pin removal.
|
 |
