Secondary Logo

Journal Logo


Hybrid External Fixation for Arthrodesis in Knee Sepsis

Salem, Khaled Hamed, MD*,†; Keppler, Peter, MD*; Kinzl, Lothar, MD*; Schmelz, Andreas, MD*

Author Information
Clinical Orthopaedics and Related Research®: October 2006 - Volume 451 - Issue - p 113-120
doi: 10.1097/01.blo.0000223979.36831.70
  • Free


Arthrodesis of the knee was first performed by Albert in 1878 for a paralytic flail knee after poliomyelitis.6 Hibbs14 performed arthrodesis on a tuberculous knee in 1911, and Key15 described arthrodesis of the knee with external fixation in 1932. In 1960, Charnley5 reported a series of patients treated with uniplanar external fixators. His technique serves as the foundation for numerous external fixation approaches still used for arthrodesis.

In the era of total joint replacement, knee fusion as a primary reconstructive procedure seldom is indicated; however, it is still an important salvage tool in the surgeon's armamentarium. The most frequent indication for knee arthrodesis is salvage of a failed total knee arthroplasty (TKA). It is still a satisfactory treatment for infection, tumors, trauma, paralytic conditions, and some cases of inflammatory or degenerative arthritis.1,5-12,17-22

Knee sepsis is one of the most challenging orthopaedic problems.1,18 It often is complicated by poor bone and soft tissue conditions, limb shortening, and malalignment.18 Infection may occur after open fractures, gunshot injuries, internal fixation surgery, joint arthroscopy, or arthroplasty. Numerous techniques have been described for knee fusion with success rates ranging from 29% to 100%,2,3,7-13,17-22,24-30 with worse results in patients with joint sepsis. The lowest fusion rates were obtained with single-plane external fixation.11,27 Plate fixation is not recommended in grossly infected knees.6,18,21 Intramedullary nailing allows early weightbearing with a high fusion rate.7,24,27,28,30 The disadvantages include technical difficulty, long operative time, large blood loss, and lack of compression.10,18,29 It also is limited in its ability to correct limb shortening or malalignment and may lead to intramedullary spread of infection.18 The Ilizarov technique offers a reliable solution for such complications. We reviewed the results obtained with knee arthrodesis using hybrid Ilizarov frames for treating complicated septic knee infections.


We retrospectively reviewed 21 patients with persistent septic knee infections treated with knee fusion from May 1991 to October 2003. The patients were treated at our institution using hybrid Ilizarov frames (Smith and Nephew, Memphis, TN). There were 13 men and eight women with a mean age of 49.7 years (range, 21-75 years). Arthrodesis was performed to treat septic arthritis after open reduction and internal fixation (ORIF) for fractures around the knee in eight patients, after knee arthros-copy in four patients, after TKA in three patients, after spreading from a nearby soft tissue infection in two patients, and as a complication of cortisone therapy for rheumatoid arthritis (RA) in one patient. Three patients had the frame used for rearthrodesis after a fracture in a previously fused septic knee. Several patients had chronic osteomyelitis (n = 16), earlier attempts at fusion (n = 5), and were older (Table 1). The operative indication and the decision regarding union, complications, and frame removal were made by the senior author (LK). All authors were involved in the surgical treatment. Data were collected by three (KHS, PK, and AS) authors during outpatient visits. Data analysis was performed by the first author (KHS).

Patient Demographics, Results, and Complications

Patients were positioned supine under general (endotracheal) or regional (spinal or epidural) anesthesia on a radiolucent operating table with biplanar fluoroscopic control. The entire lower limb up to the hip was prepared and draped in a sterile fashion. The foot was left partly exposed to control rotation and evaluate circulation throughout surgery. No tourniquet or cell-saver was used. We preoperatively sized the frame on the limb to avoid problems with postoperative limb edema. Antibiotics were administered intravenously with the induction of anesthesia according to the last culture and sensitivity testing of infected materials. Patients with quiescent infection were given a second-generation cephalosporin (cefuroxime) as a preoperative, single-dose antibiotic prophylaxis. Postoperatively, this was coupled with an aminoglycoside and/or penicillin, dictated by the culture results of intraoperative swabs or specimens. Antibiotics were continued until dry, clean wounds with negative culture results were achieved.

We used the anterior midline approach except when contra-indicated by previous incisions. We performed copious wound irrigation with saline or antiseptic solutions throughout the open procedure. Careful soft tissue dissection and establishment of full-thickness flaps with minimal undermining were important to preserve local vascularity and prevent wound complications. The distal femur and proximal tibia were exposed, osteosynthesis or prosthetic implants were removed, and the joint was débrided thoroughly to remove all loose, necrotic, and infected tissues. The bone was prepared using an oscillating saw to create broad, flat, bleeding surfaces to maximize bone contact and ensure the best possible chance for a solid arthrodesis. Minimal bone re-section and preservation of the available bone stock were important to enhance fusion and minimize postoperative leg-length discrepancy. The cuts were planned to result in a final alignment of 5° to 10° valgus and 0° to 5° antecurvatum. The arthrodesis surfaces were fixed temporarily using two strong crossing Kirschner (K) wires or Steinmann pins, and the wound then was closed over a suction drain. When possible, the patella articular surface was resected, and the bone was fixed over the fusion site anteriorly using large fragment lag screws to enhance union and add local bone stock.

The standard frame in most procedures consisted of a 90°-femoral arch and a full ring mounted on the femur and two full rings mounted on the tibia with various pin and wire combinations using the hybrid advanced technique according to Catagni et al.3 Components were joined by a minimum of four threaded rods. The temporary fixation pins then were removed. Telescopic rods were used if lengthening was planned. Three patients required correction of concurrent malalignment. This was facilitated by addition of conical washers or hinges or both. Two patients required bone transport using the same frame with distal femoral osteotomies to overcome large bony defects from previous trauma or surgery.

If needed, wounds were débrided every 48 hours, usually with active drainage using vacuum sealing and temporary skin replacement with polyvinyl-alcohol foam (Coldex®, Velo Medizinprodukte GmbH, Sengenthal, Germany) until secondary closure was possible or definitive coverage was obtained. Six patients required a split-thickness (Thiersch) skin graft for wound closure. No patients required fasciocutaneous or free flaps. Three patients required postoperative evacuation of wound hematomas. We modified the frame after 10 days to correct postoperative malalignment in one patient.

Patients were instructed in pin care during hospitalization, and the pin sites were checked during the routine followups. Early partial or full weightbearing was allowed according to the pain tolerance in most patients postoperatively. Union was checked clinically by removing the connecting rods at the fusion site before frame removal and manually stressing the arthrodesis. Painless weightbearing with a loose frame also indicated solid fusion. This was confirmed radiographically using biplanar or computed tomography (CT) scans.


Union of the arthrodesis site was achieved in all but one patient (Figs 1-3). The average duration of external fixation was 22.7 weeks (range, 11-47 weeks). Eight patients required temporary immobilization using a knee orthosis after frame removal. No patients required secondary bone grafting to achieve union. Reresection and frame application was necessary in three patients: early (after 3 weeks) in one patient to treat persistent infection, and late in two patients with delayed union.

Fig 1A
Fig 1A:
F. A 53-year-old patient with diabetes had knee empyema complicating a nearby subcutaneous abscess. (A) Anteroposterior and (B) lateral radiographs show joint destruction with loss of cartilage despite repeated débridement. (C) Anteroposterior and (D) lateral radiographs show the patient after knee arthrodesis using a hybrid Ilizarov frame. These (E) AP and (F) lateral radiographs show solid bony fusion with good limb alignment.
Fig 2A
Fig 2A:
D. A salvage knee fusion was performed in a patient who had an infected TKA. The patient had 15 previous surgical procedures. (A) Anteroposterior and (B) lateral radiographs show the patient after knee arthrodesis using an Ilizarov frame after prosthetic removal. (C) Anteroposterior and (D) lateral radiographs show full consolidation of the fusion site despite associated bone loss.
Fig 3A
Fig 3A:
F. Infection after internal fixation of a complex patellar fracture. (A) Anteroposterior and (B) lateral radiographs show persistent knee sepsis despite having the patellectomy. (C) Anteroposterior and (D) lateral radiographs show minimal joint resection after knee arthrodesis using an Ilizarov frame. After fixator removal, the (E) AP and (F) lateral radiographs show solid fusion

Patient followup averaged 75 months (range, 13 months-13 years). All patients were fully ambulant at the final outpatient visit. Four patients reported mild occasional discomfort. The rest were totally pain-free. Only one young female patient was dissatisfied with the treatment result. She had resistant chronic osteomyelitis despite repeated surgical débridement before and after arthrodesis. All patients had a leg-length discrepancy after fusion with an average shortening of 2.8 cm (range, 1.5-5 cm) in the fused limb. This was partly or fully accommodated by shoe elevation.

Eight patients had pin tract infections develop during treatment, but only four needed additional surgery for treatment (pin/wire exchange, removal, local débridement, and frame modification). Infections in the remaining four patients resolved using local pin-site care and oral antibiotics. None of the patients had compartment syndrome or any neurovascular injuries develop from wire insertion or bone distraction.

Complications occurred in nine patients according to Paley's definition.23 Aseptic nonunion occurred in one patient treated for an infected revision TKA with progressive instability after frame removal. Rearthrodesis was performed using a locked arthrodesis nail. Refracture of an arthrodesis occurred in two patients. One patient, who was an alcoholic, experienced an undisplaced fracture after a fall while inebriated. The fracture healed with conservative treatment with an above-knee cast, whereas the fracture in the other patient healed after plate fixation using a 14-hole LC-DCP plate. A patient who was a heavy smoker, who had a Grade III open complex knee fracture and a compartment syndrome, had an infected nonunion of an associated proximal tibial shaft fracture despite full consolidation of the fusion site. The patient was treated by an above-knee amputation.

Persistence of infection with chronic osteomyelitis occurred in three patients. One patient was treated with a reresection and renewed frame application, whereas the other two patients needed multiple débridements with application of antibiotic-impregnated polymethylmethacry-late beads (Septopal®, Biomet Merck B, Darmstadt, Germany).

Complications were more common in elderly patients. All cases of delayed union, nonunion, and refractures occurred in patients older than 50 years. Infection also was more common in this age group.


We reviewed the results of Ilizarov knee fusion in a group of patients with resistant infections complicated by poor bone and soft tissue quality after repeated débridements or failed fusion. Such problems cannot be addressed reliably by conventional techniques, and ultimately, they may require that the patient have limb amputation. The Ilizarov technique resulted in solid fusion in all but one patient. Patients were able to walk independently with mild or no pain and most were satisfied with the outcome.

Hagemann et al12 were the first to use compression external fixation for fusion after an infected knee prosthesis. Such uniplanar frames have numerous disadvantages including weak stability, pin loosening, infection, and stress fractures.7,9,10,13,24,25 Improvement of external fixation from single to double frames and from uniplanar to biplanar fixation increased fixation rigidity, but the overall fusion rate remained low, reported by Damron and McBeath7 as 64%. Nichols et al21 achieved knee fusion using dual plate fixation in 11 patients with failed TKAs. However, they advised not using this technique in patients with gross or acute infection.21 Intramedullary nailing for knee fusion, first described by Chapchal,4 allows rigid fixation with easier rehabilitation. Unfortunately, complications like nail migration, nail breakage, delayed union, nonunion, and fractures have been reported in as much as 55% of patients.17,29,30 Long nails require that the medullary canal be free of any obstruction or deformity, whereas shorter nails require creating a cortical window for insertion and cutting the nail in segments for its removal.29

The Ilizarov ring fixator offers several mechanical and biologic advantages compared with other internal or external fixation methods of knee fusion. These include an excellent stability that allows early full weightbearing, a dynamic axial fixation that beneficially loads the fusion site, and frame versatility that allows postoperative adjustments in alignment and dynamization.3,16,18,22 Biologic bone stimulation through focal compression and distraction helps obtaining solid union without the need for bone grafting. The accompanying hyperemia also could help resolution of infection. The method can be used in the presence of acute infection as a one-stage procedure. It is a treatment solution for patients with large bony defects, comminuted fractures, advanced osteoporosis, or in whom the medulla is obliterated (eg, hip prosthesis). Fusion also can be checked clinically before frame removal.

Only a few reports on Ilizarov knee arthrodesis are available (Table 2). Manzotti et al18 used the Ilizarov method in six patients after infected TKA. Five patients completed treatment, and all achieved a stable arthrodesis. Garberina et al10 retrospectively studied 19 patients who had knee arthrodesis with circular external fixation. Fusion was successful in 13 patients after a mean fixation time of 18 weeks. Nonunion occurred in six patients, four of whom had infected TKA. Oostenbroek and van Roermund22 treated 15 patients with infected TKA by prosthesis removal and Ilizarov knee arthrodesis. All but one knee fused during the first attempt. David et al8 reported all of their 13 patients with failed TKA achieved solid fusion. The average healing time was 27.6 weeks, of which 18.8 weeks were spent wearing the frame and the rest were spent wearing a cylinder cast.8 The high fusion rate in our patients compared favorably with the rates from the other reports, and was much higher than obtained using traditional external fixators.7,13,25

Fusion Rates with Ilizarov Knee Arthrodesis

The Ilizarov frame has the unique ability to correct limb shortening and malalignment simultaneously. We performed bone transport to compensate for bone loss in two patients. Kitson and Eyres16 and Manzotti et al20 also published case reports on Ilizarov knee arthrodesis with limb lengthening using the same Ilizarov frame. Monofocal distraction was performed at the site of arthrodesis in the former, and trifocal treatment with simultaneous lengthening via proximal femoral and distal tibial osteotomies was performed in the latter. We performed distraction using distal femoral osteotomies.

The difficulties and complications we encountered were consistent with those reported in the literature, with pin-tract problems occurring most frequently.8,10,18,22 Bone-healing complications and refractures occurred in elderly patients, but second fusion attempts were successful. A principal challenge in our cohort was the persistence of infection despite fusion. The Ilizarov technique does not guarantee resolution of infection if débridement and antibiotic treatment are not adequate. Whether to fuse while infection is still present is controversial, although there are fewer hazards with external fixation than with internal fixation.6,18,21 Bengston and Knutson1 and Knutson et al17 reported an increased chance of a solid fusion if the infection was first treated successfully. In contrast, Hak et al13 and Garberina et al,10 using conventional and circular fixators, respectively, found no significant difference between one-stage and two-stage attempts at fusion.

Several factors play a vital role in obtaining a successful arthrodesis. Raothacker and Cabanela25 identified rigid fixation as the main requisite needed to achieve a solid fusion. Knutson et al17 and Brodersen et al2 highlighted the importance of bone loss after failed TKA. They reported a significantly greater rate of successful arthrodeses in patients who had failed minimally or partially constrained knee arthroplasties than in patients with failed hinged knee prostheses.2,17 We think maximizing bone contact through proper alignment and compression fixation are most important in knee arthrodesis. It also is important to limit bone resection to the extent needed to resolve infection without jeopardizing the available bone stock or adding to the difficulty of obtaining a stable fusion. The variety of causes for joint infection with subsequently varying degrees of bone loss was a major weakness that prevented comparisons among the different etiologies. A larger patient cohort would be useful.

Our results emphasize the clinical success achieved using the Ilizarov method in knee arthrodesis. The technique offers several advantages compared with other fusion modalities because of fixation stability and versatility. Most complications occurred in elderly patients; however, a high fusion rate was achieved despite long-standing infection, failed previous treatments, or associated comorbidity.


1. Bengston S, Knutson K. The infected knee arthroplasty: a 6-year follow-up of 357 cases. Acta Orthop Scand. 1991;62:301-311.
2. Brodersen MP, Fitzgerald RH Jr, Peterson LF, Coventry MB, Bryan RS. Arthrodesis of the knee following failed total knee arthroplasty. J Bone Joint Surg Am. 1979;61:181-185.
3. Catagni MA, Malzev V, Kirienko A. Arthrodesis. In: Maiocchi AB, ed. Advances in Ilizarov Apparatus Assembly. Milan, Italy: I1 Quadratino; 1998:123-125.
4. Chapchal G. Intramedullary pinning for arthrodesis of the knee joint. J Bone Joint Surg Am. 1948;30:728-734.
5. Charnley J. Arthrodesis of the knee. Clin Orthop Relat Res. 1960;18:37-42.
6. Christian CA, Donley BG. Arthrodesis of Ankle, Knee, and Hip. In: Canale ST, ed. Campbell's Operative Orthopaedics. 9th ed. St Louis, MO: Mosby; 1998:145-187.
7. Damron TA, McBeath AA. Arthrodesis following failed total knee arthroplasty: comprehensive review and meta-analysis of recent literature. Orthopedics. 1995;18:361-368.
8. David R, Shtarker H, Horesh Z, Tsur A, Soudry M. Arthrodesis with the Ilizarov device after failed knee arthroplasty. Orthopedics. 2001;24:33-36.
9. Fidler MW. Knee arthrodesis following prosthesis removal: use of the Wagner apparatus. J Bone Joint Surg Br. 1983;65:29-31.
10. Garberina MJ, Fitch RD, Hoffmann ED, Hardaker WT, Vail TP, Scully SP. Knee arthrodesis with circular external fixation. Clin Orthop Relat Res. 2001;382:168-178.
11. Gristina AG, Kolkin J. Current concepts review: total knee replacement and sepsis. J Bone Joint Surg Am. 1983;65:128-134.
12. Hagemann WF, Woods GW, Tullos HS. Arthrodesis in failed total knee replacement. J Bone Joint Surg Am. 1978;60:790-794.
13. Hak DJ, Lieberman JR, Finerman GA. Single plane and biplane external fixators for knee arthrodesis. Clin Orthop Relat Res. 1995;316:134-144.
14. Hibbs RA. An operation for stiffening the knee joint with report of four cases from the service of the New York Orthopedic Hospital. Ann Surg. 1911;53:404-407.
15. Key J. Positive pressure in arthrodesis for tuberculosis of the knee joint. South Med J. 1932;25:909-915.
16. Kitson J, Eyres KS. Simultaneous leg lengthening and knee arthrodesis using an external ring fixator: a case report. Acta Orthop Scand. 2002;73:712-714.
17. Knutson K, Lindstrand A, Lidgren L. Arthrodesis for failed knee arthroplasty: a nationwide multicenter investigation of 91 cases. Clin Orthop Relat Res. 1984;191:202-211.
18. Manzotti A, Pullen C, Deromedis B, Catagni MA. Knee arthrodesis after infected total knee arthroplasty using the Ilizarov method. Clin Orthop Relat Res. 2001;389:143-149.
19. Manzotti A, Pullen C, Guerreschi F, Catagni MA. The Ilizarov method for failed knee arthrodesis following septic TKR. Knee. 2001;8:135-138.
20. Manzotti A, Pullen C, Guerreschi F, Catagni MA. Knee arthrodesis and limb lengthening in the treatment of infected total knee arthroplasty: case report. J Trauma. 2002;2:359-363.
21. Nichols SJ, Landon GC, Tullos HS. Arthrodesis with dual plates after failed total knee arthroplasty. J Bone Joint Surg Am. 1991;3: 1020-1024.
22. Oostenbroek HJ, van Roermund PM. Arthrodesis of the knee after an infected arthroplasty using the Ilizarov method. J Bone Joint Surg Br. 2001;83:50-54.
23. Paley D. Problems, obstacles, and complications of limb lengthening by the Ilizarov technique. Clin Orthop Relat Res. 1990;250:81-104.
24. Rand JA. Alternatives to reimplantation for salvage of the total knee arthroplasty complicated by infection. J Bone Joint Surg Am. 1993;75:282-287.
25. Rothacker GW Jr, Cabanela ME. External fixation for arthrodesis of the knee and ankle. Clin Orthop Relat Res. 1983;180:101-108.
26. Stiehl JB, Hanel DP. Knee arthrodesis using combined intramedullary rod and plate fixation. Clin Orthop Relat Res. 1993;294:238-241.
27. Vlasak R, Gearen PF, Petty W. Knee arthrodesis in the treatment of failed total knee replacement. Clin Orthop Relat Res. 1995;321: 138-144.
28. Waldman BJ, Mont MA, Payman KR, Freiberg AA, Windsor RE, Sculco TP, Hungerford DS. Infected total knee arthroplasty treated with arthrodesis using a modular nail. Clin Orthop Relat Res. 1999;367:230-237.
29. Wiedel JD. Salvage of infected total knee fusion: the last option. Clin Orthop Relat Res. 2002;404:139-142.
30. Wilde AH, Stearns KL. Intramedullary fixation for arthrodesis of the knee after infected total knee arthroplasty. Clin Orthop Relat Res. 1990;248:87-92.
© 2006 Lippincott Williams & Wilkins, Inc.