Salvage of infected total knee arthroplasty with Ilizarov external fixator.

BACKGROUND: Knee arthrodesis may be the only option of treatment in cases of chronic infected total knee arthroplasty (TKA) with concomitant irreparable extensor mechanism disruption, extensive bone loss or severe systemic morbidities. Circular external fixation offers possible progressive adjustment to stimulate the bony fusion and to make corrections in alignment. We evaluated the results of knee arthrodesis with one or two stage circular external fixator for infected TKA.
MATERIALS AND METHODS: 16 cases of femoro-tibial fusion were retrospectively evaluated. Male-to-female ratio was 10:6. Mean age of the patients was 62.2 years. Cierney-Mader classification was used for anatomical and physiological evaluation while the bone stock deficiency was classified into mild, moderate and severe. Surgical technique involved either single or two stage arthrodesis using circular external fixator.
RESULTS: Union was achieved in 15 patients (93.75%). The mean duration for union (frame application time) in these patients was 28.33 weeks (range 22 to 36 weeks). Analysis showed that in the group with frame application time of less than 28 weeks, the incidence of mild to moderate bone deficiency was 83.33%, while in the frame application time more than 28 weeks group the incidence was 20% (P-value 0.034). Similarly the incidence of Cierney-Mader 4B (Bl, Bs, Bls) was found to be 33.33% in the group of frame application time of less than 28 weeks, while it was 90% in the group with frame application time more than 28 weeks (P-value 0.035).
CONCLUSION: Circular external fixator is a safe and reliable method to achieve knee arthrodesis in cases of deep infection following TKA. Severe bone stock deficiency and Cierney- Mader type B host are likely risk factors for prolonged frame application time. We recommend a two-stage procedure especially when there is compromised host or severe bone loss.

INTRODUCTION

Knee arthrodesis may be the only option of treatment in cases of chronic infected total knee arthroplasty (TKA) with concomitant irreparable extensor mechanism disruption, extensive bone loss or severe systemic morbidities. Various surgical techniques have been used for knee arthrodesis for infected TKA. The three most common methods to achieve knee arthrodesis are internal plate fixation1 intramedullary rod fixation2–4 and external fixation.5–13 Internal fixation with nail or plates in patients with deep joint infection carries the risk of dissemination, reactivation, or maintenance of a latent joint infection. On the other hand, circular external fixation offers possible progressive adjustment to stimulate the bony fusion and to make corrections in alignment, and there is a considerably lower risk of intramedullary dissemination of the infection, and the easy removal of the hardware is possible. Disadvantages of circular external fixators are: frame maintenance, cosmetic discomfort, need for fixator removal, pin tract infections, pin loosening, risk of neurovascular damage during pin and wire insertion. Reported union rates with circular external fixator ranged from 64% to 100%.6–13 The mean fusion time ranged from 4.4 to 6 months.6–13

In the current report, we evaluated the results of knee arthrodesis with one or two stage circular external fixator for infected TKA in 16 patients and also analyzed predictive factors for prolonged frame application time and pseudoarthrosis.

MATERIALS AND METHODS

All cases of infected TKA treated for knee arthrodesis with Ilizarov external fixator from 2003-2010 were evaluated retrospectively after institutional ethical committee clearance. The cases whose details were fully available and who had completed follow up till union (or final outcome), were included, and those who died or were lost to follow up were excluded. During this period a total of 19 cases of infected TKA got operated, out of which 1 died after 3 months due to unassociated reasons, in 1 patient the apparatus was removed after 2 weeks of application because of cosmetic discomfort, and 1 case was lost to follow up. Thus the present analysis include 16 cases. Ten cases (62.5%) were male patients while the other 6 were females (37.5%). The mean age of the patients was 62.2 years (range 55-70 yrs). Primary diagnosis for which the patients underwent TKA was primary osteoarthritis in 10 (62.5%), inflammatory arthritis in 5 (31.25%) and post traumatic arthritis in one case (6.25%). Of these 16 cases, 10 were operated for TKA in our center while the other 6 got initially operated for TKA elsewhere. The mean number of surgeries undergone by each patient before the arthrodesis surgery was 4.25. Culture patterns of the knees showed Staphylococcus aureus (n=6), Staphylococcus epidermidis (n=4), Escerichia coli (n=2), Pseudomonas aureginosa (n=1) and in the other 3 cases culture was negative. However, based on the clinical findings of warmth, discharging sinus, and elevated erythrocyte sedimentation rate (ESR), C-Reactive protien (CRP) and radiological evidence of loosening, a diagnosis of infected TKA was made in the 3 culture negative cases [Table 1].

Table 1

Clinical details of patients

Clinical classification was done using the Cierney-Mader classification for chronic osteomyelitis.14 All the 16 cases were stage IV (diffuse) osteomyelitis. As per the host physiology classification, 5 were class A, 2 were Bl, 6 were classified as Bs and the other 3 as Bls. For anatomical classification of the bone stock deficiency, we classified as mild, moderate, based on classification by Oostenbroek et al.10 Mild deficiency was encountered in 3 cases while 4 and 9 cases had moderate or severe deficiency respectively, [Table 1]. We considered a case with mild bone deficiency and Cierney-Mader classification 4A for single stage surgery. A case with mild bone deficiency but Cierney-Mader classification 4Bl or 4Bs or 4Bls or a case with moderate bone deficiency irrespective of Cierney-Mader classification 4A was considered as a relative contraindication for a single stage surgery, and a case with severe bone deficiency irrespective of Cierney-Mader classification was considered an absolute contra- indication for one stage surgery [Table 2].

Table 2

One or two stage arthrodesis based on Cierney-Mader classification and bone stock deficiency

All the cases were operated after obtaining the written informed consent with prior explanation of the procedure. All cases were opened through the previous surgical scar, and excision of sinus and scar margin was done along with a medial parapatellar arthrotomy. The implants were removed. Thorough debridement and wash was done. In one stage procedure the bone edges were freshened with a saw or bone file, and approximation of the bone ends was achieved by primary docking. Defect, if any, was filled with patella which was resected at this stage and used as an autogenous bone graft. The femoral frame construct included a proximal Ilizarov arch fixed to the femur, perpendicular to the axis of femur with two 5 mm percutaneous half pins. This arch was connected to a ring of appropriate size anchored in the distal femur with 2 olive/K-wires. This was connected to the tibial frame which included 2-3 full rings anchored to tibia perpendicular to its axis with 2 K-wires in each ring. One ring was in the proximal tibia and the second ring in the proximal diaphysis. An additional 3rd ring was used in cases of poor bone stock or cases where future lengthening was planned, which was then anchored in the distal tibia. The frame is adjusted to achieve a final position of neutral to 5 degrees of knee flexion with tibia externally rotated by 5-10 degrees in relation to femur. In cases with severe bone deficiency 2 stage procedure was performed. After removal of the implants and thorough debridement, antibiotic cement spacer was applied. The type of the antibiotic used in the spacer was based on the culture/sensitivity reports of the previous swabs, and was selected from one of the four heat stable antibiotics (Vancomycin, Tobramycin, Gentamycin, Erythromycin). Wounds were closed in layers over a suction drain. Stage 2 included removal of the cement spacer and application of the frame, as described above, after an interval of 4-6 weeks. In all cases operated by 2 stage procedure, the external fixator was applied only during the second stage surgery [Figures 1 and 2].

Figure 1

(a) Clinical picture of a patient with infected total knee arthroplasty of the right side with a discharging sinus over the anterior aspect of knee. (b) X-ray after stage I procedure where an antibiotic cement spacer was placed after thorough debridement. (c-d) X-ray (anteroposterior and lateral views) of the same patient after stage II surgery. (e-f) Clinical picture showing marginal wound necrosis after stage II surgery which successfully healed with secondary intention on conservative management. (g-h) X-ray of the right knee (anteroposterior and lateral views) at follow up of 5 years showing solid consolidation

Figure 2

(a) Anteroposterior X-ray of right knee after stage I procedure where antibiotic cement spacer was placed after thorough debridement. (b-c) X-ray (anteroposterior and lateral views) of the same patient after stage II surgery. (d) Clinical picture of the same patient partially bearing weight with the fixator in situ. (e) X-ray of right knee (anteroposterior and lateral views) at follow up of 2.5 years showing solid consolidation

Compression was started from day 1 at the rate 0.5 mm/day in two sittings for 2 weeks and till radiological appearance of complete compression. Weight bearing was allowed as tolerated. Sutures were removed from day 14 to day 18 depending on the wound healing followed by oral antibiotics for another 3 weeks based on the culture reports. Antibiotics were given till suture removal in cases where the culture reports were negative. Radiographic fusion was defined as trabecular bridging along the whole fusion area in both the anteroposterior and lateral planes. Once the radiographic fusion was observed the connecting rods, were removed to clinically assess the fusion by varus/valgus stress test. A clinically stable arthrodesis was the criterion for successful fusion. After the removal of the fixator, the patient was put on knee brace for 2-3 weeks. Based on Paley's description, we classified all our complications as problems, obstacles and complications. In rest of the manuscript by the word “complication” we mean complication as defined by Paley.15 Problem is defined as difficulties that require no operative intervention to resolve, while obstacles are defined as difficulties that require an operative intervention. All difficulties that are not resolved before the end of treatment are considered as true complications.

All statistical analysis was done using statistical package for the Social Sciences (SPSS) 12.0 software and nonparametric tests were used for comparison. Fisher 2 × 2 test was used for comparison of categorical data while Mann Whitney U test was used for comparison of numerical data.

RESULTS

Union

Out of 16 patients, the sound arthrodesis was achieved in 15 patients (93.75%). The mean duration for union (frame application time) in these patients was 28.33 weeks (22 to 36 weeks). The mean number of surgical procedures needed till union in these 15 patients was 2.2. Two of these 15 patients needed additional bone grafting (other than the patella) for consolidation. Out of these 15 patients, single stage arthrodesis was performed in 3 patients, out of which 2 were classified as mild bone deficiency with Cierney-Mader classification 4A. The other patient in whom single stage surgery was done was a case of moderate bone defeciency and Cierney-Mader classification 4Bs (case No.1). However, that patient needed re-debridement after 2 weeks for persistence of infection. In all the 3 cases where one stage surgery was done, consolidation was achieved. In the other 13 cases where 2 stage surgery was done consolidation is achieved in 12 cases. Out of these remaining 12 cases as per bone deficiency classification, 8 were severe, 3 were moderate while one was mild. As per Cierney-Mader classification 5 were 4Bs, 3 were 4A, 2 were 4Bls and the other 2 were 4Bl. The mean follow up period in these cases is 3.83 yrs and mean shortening is 4.4 cm [Table 3].

Table 3

Details and outcome of the cases

A total of 7 (43.75%) problems occurred, out of which 5 were superficial pin tract infections, 1 case of marginal necrosis of the wound which healed by secondary intention on conservative management and 1 case of contact dermatitis (allergy to betadine). A total of 6 obstacles (40%) occurred, out of which 2 are incidents of loosening and pin breakage, 1 was a case of pin loosening only, and 1 was a case of persistence of infection. One of the cases with pin breakage was also bone grafted in the same sitting. In another case of delayed union, autogenous bone grafting was done which eventually consolidated. In all these 5 cases re surgery has solved the issue by replacement of pins in 3 cases and redebridement in the other case. One case (6.25%) developed complication in the form of pseudoarthrosis [Table 3] One case of fixator removal due to poor cosmetics was also considered as a true complication as per Paley's definition increasing the complication rate to 12.5%.

A comparative analysis was done between the cases where consolidation was achieved within 28 weeks to those where the frame application time (FAT) was more than 28 weeks. The duration of 28 weeks was taken as a cut off, based on the mean frame application time in cases of union documented in the previous studies (28.45 weeks)6–13 as well as our current study (28.33 weeks). The analysis showed that in the group of FAT of ≤ 28 weeks, the incidence of mild to moderate bone deficiency was 83.33%, while in the FAT ≥ 28 weeks group the incidence is 20% (P-value 0.034). Similarly the incidence of Cierney-Mader 4B(Bl, Bs, Bls) was found to be 33.33% in the group of FAT of ≤ 28 weeks, while it is 90% in the group of FAT > 28 weeks (P-value 0.035) [Table 4].

Table 4

Comparison of various parameters between group of patients with frame application time less than 28 weeks to group of patients with frame application time more than 28 weeks

DISCUSSION

Deep infection is a dreaded complication of TKA and usually needs implant removal and arthrodesis. The various available options for arthrodesis include arthrodesis over an intramedullary, plate arthrodesis or external fixator arthrodesis. Intra medullary nailing is considered to be the most reliable method to achieve arthrodesis with a reported union rates of 67-100%.16–18 Similar rates of fusion are also documented using dual compression plates.119 However, both these techniques have an inherent disadvantage of being able to be used only after complete eradication of the infection. Also they need a second surgery for implant removal. Reported complications of both these techniques include recurrent infection, nail breakage and migration and stress fractures at plate ends.116–19 The technique of knee arthrodesis with a monoaxial or biaxial external fixator is also well documented.52021 These fixators, however, do not allow significant changes in the axis, and are rigid and not entirely reliable for complete load-bearing.21

Circular external fixators offer the advantage of frame adjustment, and if required change of axis of compression during treatment without a separate surgery, permitting maximum compression and stability at the arthrodesis, and also allows early weight bearing, usually within a few days after surgery.6 The disadvantages are frame maintenance, cosmetic discomfort, fixator removal, less predictable fusion rates, pin tract infections, pin loosening, risk of neurovascular damage during pin and wire insertion.6 A meta-analysis of techniques used for knee arthrodesis after failed TKA, most of which were infected, reported overall fusion rates of 95% in patients treated with intramedullary nail (IMN), compared with 64% for patients treated with external fixator.22 In a comparative study of nail versus external fixator for knee arthrodesis, Mabry et al. recommended that intra medullary nailing appears to have a higher rate of successful union but a higher risk of recurrent infection when compared with external fixation knee arthrodesis.7

Single staged versus two staged technique for knee arthrodesis in infected TKA patients is debatable. It is unclear whether there is a significant difference in fusion rates between one- or two-stage external fixator because of the variability of selection criteria for knee arthrodesis between studies.23 Some authors suggest that staged external fixator should always be used in an attempt to eradicate multimicrobial or virulent infection where there is extensive bone loss.23 We are also a strong believer of two staged technique especially in patients with moderate to severe bone stock deficiency and 4B type of hosts according to Cierney-Mader classification. This technique of 2 stage fusion also gives a revised option of decision making for both the surgeon and patient to consider revision TKA after making sure that the infection is controlled. In the current series, irrespective of the bone stock deficiency, no effort was made to primarily graft the arthrodesis site other than the use of the patellar remnant. In spite of this, we needed to do secondary bone grafting in only 2 of the 9 cases with severe bone stock deficiency, however these cases needed prolonged frame application time as compared to others.

The existing literature on knee arthrodesis with an Ilizarov circular external fixator reports success rates that vary in different studies from 64% to 100%6–13 [Table 5]. In our current series also, the union rate was 93.75%. We attribute this high union rate with low incidence of persistence of infection to 2 stage surgery. One case developed pseudoarthrosis and was ambulatory with a brace in situ. In one patient, we had to remove the external fixator just 2 weeks after the application due to poor tolerance by the patient. Hence it is important to counsel the patient before undertaking circular frame application and such similar problems are well documented in the literature.6 In the current study, we included preoperative evaluation parameters like bone stock deficiency and Cierney-Mader classification for deep infection. Based on the parameters, we retrospectively evaluated and found that Cierney-Mader 4B (Bl, Bs, Bls) and severe bone stock deficiency correlated with a longer frame application time.

Table 5

Review of literature of knee arthrodesis with circular external fixator for infected total knee arthroplasty

The retrospective nature of the study is the limitation of our study. The strengths of the study include that it involved a significant number of cases and we were able to identify a few prognostic factors which can predict a longer duration of frame application time. However, further prospective trials with large sample size are needed. Also whether application of auto/allo graft or substitutes at the time of surgery will enhance the consolidation or not needs further studies.

We conclude that Ilizarov circular external fixator is a safe and reliable method for knee arthrodesis for cases of deep infection arising in cases of TKA. Our choice of the modality for arthrodesis is supported by its cost effectivity, versatility and possibility to perform modifications during treatment. Severe bone stock deficiency and local or systemic compromise are predictors for longer duration of frame application time. We recommend a two-stage procedure especially when there is a compromised host immunity or severe bone loss.