Abstract
Background:
In the setting of chronic osteomyelitis following fractures about the ankle, reconstruction through bony arthrodesis may be used as a reconstructive alternative to amputation. During these cases, surgeons often avoid using internal fixation in an attempt to avoid reinfection or premature hardware failure. In this retrospective review, we analyzed the outcomes of chronic osteomyelitic patients who had an arthrodesis of the ankle using either internal or external fixation, focusing on salvage rates, infection clearance, union rates, and functional outcomes. No device was implanted into a known active infection.
Methods:
We performed a retrospective chart review of adult patients undergoing arthrodesis in the setting of a previously septic ankle following a traumatic injury. In each case, multiple irrigation and debridement procedures and local and systemic antibiotics were used. Infection status was determined by clinical exam, MRI, nuclear medicine studies, and ultimately bone biopsies. No fixation device was implanted in ankles with known active infections. Patients were divided into 2 cohorts: those fused with internal devices and those fused with external fixators. Thirty patients underwent a total of 32 arthrodesis procedures. Mean follow up time was 27 months (range, 6 to 144).
Results:
Nineteen fusions were performed using internal fixation; only 2 required amputations, therefore limb salvage was 90%. Fifteen were able to ambulate with or without the assistance of an orthosis (79%). Four patients experienced recurrent infection (21%) and 5 developed nonunion (26%). Of the 13 fusions performed with external fixators, only 1 required an amputation, putting limb salvage at 92%. Ten patients were able to walk with or without the assistance of an orthosis as their final functional status (77%). Two patients experienced recurrent infection (15%), and 4 went on to nonunion (31%).
Conclusion:
When analyzing these 2 fusion methods in posttraumatic patients with previously septic ankles, with the numbers available both methods achieved similar rates of limb salvage and final functional status in these patients, as well as similar rates of infection clearance and bony union. As internal fixation is often less labor-intensive for the surgeon and more palatable for the patient postoperatively, we encourage surgeons to consider arthrodesis with internal fixation once the infection is successfully eradicated, especially in a noncompliant patient population.
Level of Evidence:
Level III, retrospective comparative series.
Keywords
In the setting of chronic osteomyelitis following an ankle fracture, limb salvage by eradication of infection and arthrodesis may be used as an alternative to amputation.4,11,14,15,18,20 This treatment protocol serves several purposes. First, it restores the stability that may have been compromised by major debridement of the distal tibia and fibula. Second, it eliminates motion that may contribute to recalcitrant or recurrent infection. Third, it alleviates pain from degeneration of an ankle joint that may be present as a sequela of previous septic or posttraumatic arthritis. Finally, it reduces deformity that may have occurred during the patient’s postinfectious course due to hardware failure, bony collapse, or soft tissue contracture.5,14,19,20
When performing fusions in osteomyelitic patients, surgeons often prefer the use of external fixation over internal fixation.5,14,17 The arguments against internal fixation are that implanted hardware may result in recurrent infection and premature hardware failure due to poor bone stock at the site of injury, as a result of prior infection and multiple debridement procedures.19,20 However, external fixation is not without complications as it can be labor-intensive for the surgeon, and it requires consistent maintenance, which can be problematic in a noncompliant patient population.4,7,9
Surgeons must consider the potential complications of each method when choosing the appropriate way to perform an arthrodesis in a joint with chronic osteomyelitis. There have been many studies comparing internal fixation to external fixation in the fusion of noninfected joints.1-3,8,10,21,22 In addition, there have been several case series published on the fusion of septic ankles.5,12,14,18,20 We conducted a retrospective chart review of patients fused at our institution with either internal fixation or external fixation and hypothesized that there would be no difference between these 2 treatment methods in terms of salvage and secondary outcomes such as recurrent infection, nonunion rates, or functional status, when performed in patients who had already experienced resolution of active infection through the use of intravenous antibiotics and multiple debridements.
Methods
After receiving approval from our institutional review board, we conducted a retrospective chart review of 30 consecutive patients who had undergone a total of 32 arthrodeses in the setting of a previously septic ankle following a traumatic injury. All patients underwent multiple procedures for irrigation and debridement along with administration of local and intravenous antibiotics. All patients were treated with a staged protocol to eradicate any active infection prior to implantation of any device. The protocol was as follows: First, the patient underwent aggressive irrigation and debridement of the infected ankle. This always involved removal of hardware. In some cases, provisional external fixators were placed to retain stability until fusion was ultimately attempted. In other cases, either no gross instability resulted from the aggressive debridement, or the placement of an antibiotic cement rod and spacer, along with splinting, provided sufficient inherent stability. The antibiotic spacers or beads consisted of 6 g of vancomycin antibiotic powder and 1.2 g of tobramycin in 1 package of Palacos cement. At least 1 spacer and bead exchange was performed at the 6-week period, along with a bone biopsy and cultures. Positive biopsies or cultures led to mandated further debridement and exchange. Intravenous antibiotics were administered to all patients for 3 months, targeted at the specific bacteria cultured from the initial irrigation and debridement. Patients who received internal fixation were treated with at least 3 months of antibiotics prior to implantation of any device. Finally, following a negative bone biopsy, the spacer or beads were removed and the internal fixation device was implanted. This protocol ensured that no patient with a known active infection underwent arthrodesis. Patients treated with external fixators who showed no outward signs of infection at 6 weeks underwent arthrodesis by debriding the joint surfaces and compressing their frame, and then received another 6 weeks of intravenous antibiotics in the postoperative period.
Patients fell into 2 cohorts based on the arthrodesis method utilized. One group was fused using internal fixation devices such as intramedullary nails, locking plates, or screws, while the other group consisted of patients who were fused using multiplane compressive external fixators. In these patients, routine pin care included removal of nonviable tissue with a handheld curette as necessitated, as well as flushing the sites with normal saline through a 16-gauge intravenous catheter. This study describes the treatment protocol of the senior author with respect to chronic osteomyelitis of the ankle that underwent arthrodesis. Eight of the first 10 arthrodeses that were performed in these situations were done with external fixation, as the senior author was reluctant to place any internal fixation device in a previously septic ankle, even when there was no evidence of ongoing active infection. Over time, it was felt that the negative aspects of external fixators in our noncompliant patient population might outweigh the potential benefits when compared to internal fixation. Therefore, a transition was made toward the use of internal fixation in the very same situations that had previously utilized only external fixation. Ultimately, 9 of the final 11 patients in this series were fused using internal fixation. The remaining 2 patients who were treated with external fixators had undergone removal of infected total ankle arthroplasties, and it was felt that distraction was needed across the arthrodesis site to maintain length during healing. In this series, external or internal fixation was not chosen as the treatment modality based on the relative difficulty of each specific case; rather, it was primarily a temporally driven decision based on a change in the standard treatment protocol of the senior author.
Inpatient and outpatient notes as well as operative records were reviewed to determine demographic information, mechanism of injury, type of fixation, operative details, medical comorbidities, tobacco use, physical examination findings, and clinical course. Relevant information regarding patients’ history and clinical course were collected and organized for further analysis.
This information was used to highlight 1 surgeon’s experiences with ankle salvage in the setting of chronic osteomyelitis following fractures about the ankle, while secondarily comparing fixation methods with the outcome measures of infection recurrence, union rates, and functional status. Initial infection was diagnosed using clinical exam, MRI, nuclear medicine studies, and wound cultures. All diagnoses were confirmed with a bone biopsy examined by a pathologist. A biopsy with necrotic bone or chronic inflammatory changes was considered to be consistent with osteomyelitis. The presence of reinfection was confirmed using the same criteria. An infection was considered to be completely cleared and nonrecurrent if there were no clinical signs of infection at the last recorded follow-up visit. Of note, baseline values of C-reactive protein (CRP), sedimentation rate (ESR), and white blood cell (WBC) counts were taken for all patients upon initial presentation. These values were then trended throughout their treatment regimen as well as following definitive treatment by arthrodesis. They were used to loosely guide the author toward suspicion of reinfection, although clinical, radiological, and biopsy data were always used as confirmatory studies. Fusion status was based primarily on radiographs taken at interval follow-up visits. Radiographs were reviewed by both an orthopedic foot and ankle attending and a musculoskeletal radiologist. Fusion was considered complete when all sites of fusion in the ankle showed bony consolidation across the joint line. Final functional status was determined by interval follow-ups.
Thirty patients who underwent a total of 32 arthrodesis procedures were included in the study. One patient had bilateral fusions performed (case 22). Another patient initially treated with an external fixator had a failure and was then treated with internal fixation (cases 12a and 12b). Of the 32 fusions, 19 were performed using internal fixation while 13 were performed using external fixation. Average follow-up time was 27 months (range, 6-144), excluding 6 patients who either underwent an amputation or died of nonrelated causes. All patient specific information is available in Table 1.
All Patient Data.
Abbreviations: CAM, controlled ankle walker; CLL, chronic lymphocytic leukemia; CVA, cerebral vascular accident; DM, diabetes mellitus; HCV, hepatitis C virus; IM, intramedullary; MRSA, methicillin-resistant Staphylococcus aureus; MVA, motor vehicle accident; WC, workers’ compensation. Patient 12 failed arthrodesis with external fixation (12a) before undergoing subsequent internal fixation (12b); patient 22 underwent bilateral fusions: L and R designate location.
Number of previous operations refers to the number of surgeries needed to eradicate the infection.
Of the 19 patients fused with internal fixation, 15 were male and 4 were female. The mean age of this group was 46 years old (range, 27 to 69). The patients had varying mechanisms of initial injury, which included 7 falls, 5 motor vehicle accidents, 3 falls from height, 2 pedestrians struck, and 2 motorcycle accidents. Prior to fusion procedures, all patients had been diagnosed with bacterial osteomyelitis: 11 (58%) with Staphylococcus aureus, 3 (16%) of which were methicillin-resistant Staphylococcus aureus (MRSA), and the other 8 (42) with various other types of bacterial osteomyelitis (Table 1). Each patient underwent an average of 5.7 (range, 2-13) operative procedures attempting to eradicate this infection before attempting fusion. The 19 ankle fusions involved the following joints: 12 tibiotalocalcaneal, 3 tibiocalcaneonavicular, 3 tibiotalar, and 1 tibiocalcaneal. Hardware used included 13 intramedullary nails, 4 sets of lag screws, 1 intramedullary nail with locking plate, and 1 locking plate. The average BMI of these patients was 33.7 (Table 2). Patients had the following relevant comorbidities: 13 were obese, 7 were smokers, 6 were diabetic, 4 received workers’ compensation, and 2 had neuropathy. Four patients in the internal fixation group had wounds severe enough to require reconstruction prior to arthrodesis: case 4, a local advancement flap; case 6, a serratus anterior free flap; case 7, a latissimus dorsi free flap; and case 18, a sural fasciocutaneous rotation flap.
Demographics.
Of the 13 fusions performed with external fixation, 9 were on males and 4 were on females. The mean age of this group was 48 years old (range, 16 to 81). Mechanisms of initial injury included 8 falls, 2 falls from height, 2 motor vehicle accidents, and 1 motorcycle accident. Prior to fusion procedures, all patients had been diagnosed with bacterial osteomyelitis: 8 (62%) with Staphylococcus aureus, 6 (46%) of which were MRSA, and 5 (38%) patients with various other types of bacterial osteomyelitis (Table 1). Each patient underwent an average of 4.4 (range, 1 to 11) operative procedures attempting to eradicate this infection before attempting fusion. The 13 ankle fusions involved the following joints: 10 tibiotalar, 2 tibiotalocalcaneal, and 1 tibiocalcaneal. All 13 patients were treated with multiplane compressive external fixators. The average BMI for these patients was 29.1 (Table 2). These patients had the following relevant comorbidities: 4 were diabetic, 4 were smokers, 2 were obese, 1 received workers’ compensation, and 1 had neuropathy. Two patients in the external fixation group had extensive reconstruction requirements: case 22, free tissue transfer of rectus abdominis muscle to both ankles; case 29, a serratus anterior free flap.
A power analysis was performed to determine minimum enrollment necessary to detect significant differences between the 2 groups with regard to the outcome measures. These numbers varied, however the minimal number was approximately 40 patients in each arm and as high as 300 in each arm, depending upon the delta we sought between the two groups’ outcome measures. Therefore, we chose direct comparison between the two groups to not be the main focus of our paper. Correlation analysis was done using a chi-square table to assess any relationship between patient characteristics and the outcome measures of infection recurrence and bony union. The outcome measures of infection recurrence and nonunion rates were directly compared between the 2 treatment methods using a z test of 2 proportions. Statistical analysis was conducted with the SPSS software package (SPSS, Chicago, IL).
Results
In the internal fixation group, 4 patients experienced recurrent infection (21%) with 2 of them requiring amputation (cases 11 and 19), another failing multiple irrigation and debridement procedures before being lost to follow-up for 7 years (case 4), and another who died 28 days postoperatively of preexisting liver failure (case 6). Five patients went on to nonunion (26%), including the 2 patients who later required amputation (cases 11 and 19), and 1 patient who was a brittle diabetic who developed a stable fibrous union (case 14) died of a cerebrovascular accident 182 days postoperatively. The final 2 developed nonunions but were fitted for rocker bottom shoes (cases 3 and 16). The average time to union was 8.5 months (range, 4 to 97). This number does not include 1 patient who was lost to follow-up for 7 years, only to return with a fused ankle, making his known time to fusion approximately 97 months (case 4). Mean total follow-up time for the group was 27 months (range, 6 to 144).
In the external fixation group, 2 patients experienced recurrent infection (15%) at last known follow-up (Table 3). The first patient experienced recurrence at 187 days postoperatively and was brought to the operating room for irrigation and debridement (case 21). He died shortly after of noninfectious trauma-related causes. The other patient who experienced infection recurrence underwent many irrigation and debridement procedures along with multiple courses of IV antibiotics before deciding on a second fusion with internal fixation as treatment (cases 12a and 12b). Four patients went on to nonunion (31%), including 1 who experienced no recurrence but went on to an unstable nonunion and opted for elective amputation rather than continued attempts at limb salvage (case 27). The other 3 patients walked with the assistance of an orthosis, including 2 rocker bottom shoes (cases 12b and 23) and a Charcot restraint orthotic walker (case 26). The average time to union was 8.1 months (range, 2-21). Mean total follow-up time for the group was 21 months (range, 7 to 61).
Outcome Measures.
Functional status means able to ambulate with or without an orthosis.
Therefore salvage was similar between the 2 groups as only 2 patients in the internal fixation group required amputation due to reinfection (cases 11 and 19), while 1 patient in the external fixation group underwent amputation for a nonunion (case 27). Functionally, 15 of 19 patients (79%) in the internal fixation group were walking with a rocker bottom shoe while 10 of 13 patients (77%) in the external fixation group were walking with a rocker bottom shoe, a custom-molded shoe, or a Charcot restraint orthotic walker (Table 3).
Discussion
The trend in the current literature is to avoid internal fixation in favor of external fixation when fusing a septic or previously septic joint.5,8,14,16,17 We reviewed the senior author’s experience regarding the treatment of chronic osteomyelitis through arthrodesis with either internal or external fixation, focusing on limb salvage, infection recurrence, bony union, and final functional status.
When performing an arthrodesis in a noninfected joint, similar rates of fusion have been achieved by some surgeons utilizing either internal or external fixation.2,10,21,23 Internal fixation with either screws or intramedullary nails has become the most common method. However, in clinical scenarios involving infection, internal fixation has been generally avoided.7,14,19 It is thought that introducing foreign material into the infected joint may lead to an inability to eradicate the chronic osteomyelitis.14,19 In addition, patients may have poor local bone stock at the site of infection as a consequence of prolonged non-weight-bearing, necrotic bone, and multiple debridement procedures. It is felt that this may lead to inadequate screw purchase, thus resulting in a failed attempt at fusion.16,19,20 In fact, several patients in our study were referred to us by other surgeons who had been unsuccessful in fixation or infection eradication (Figure 1).
(A) This patient (patient 10) presented with an extruded talus after being propelled out of a delivery truck following a motor vehicle accident. (B) The patient’s ankle was irrigated and debrided, followed by a pinning of the ankle through the calcaneus. (C) He returned several months later with active pus exuding from his ankle. Following infection eradication with 2 months of antibiotics and the placement of an antibiotic spacer, a tibiotalocalcaneal arthrodesis was performed using an ankle fusion nail. (D) Six months postarthrodesis, the joints show complete bony fusion on radiograph.
Therefore, nearly all of the literature on fusion of an osteomyelitic ankle involves standard or hybrid external fixation. Using these methods, infection clearance ranged from 84% to 100%, while fusions rates ranged from 69% to 96%.5,12-14,20 In the internal fixation arm of our study, we achieved similar rates of infection clearance (79%) and fusion (74%) (Figure 2). Utilizing external fixation we achieved an infection clearance rate of 85% and a union rate of 69%. Our review of this series shows that the efficacy of internal fixation was similar to that of external fixation with regard to limb salvage as well as infection clearance, bony fusion, and final functional status. Notably, the average BMI was approximately 4 points higher in the internal fixation group.
(A) Sixteen-year-old male (patient 29) was referred to us from another institution 3 months after an attempted fixation of a pilon fracture sustained sliding down a cement aqueduct. He presented with an active infection, a severe equinovarus contracture, and a dislocated tibiotalar joint. (B) The external fixator has been removed, and a pin from the device can be seen penetrating directly through an infected eschar. The patient’s resulting severe equinovarus contracture can be seen to the right of the picture. (C) Pus was exuding from the wound, leading us to perform aggressive irrigation and debridement seen in the left side of the picture. Antibiotic bead placement and free tissue transfer were then employed to assist in eradication of the infection. (D) After undergoing a course of intravenous antibiotics, a tibiocalcaneal arthrodesis was performed using a Taylor spatial frame. (E) One year postarthrodesis, the fusion is complete and the patient is able to walk with a custom-molded rocker bottom shoe.
One vital but often unemphasized aspect of these surgeries relates to patient compliance. External fixators can be cumbersome and difficult to care for, leading patients to ignore precautions and end up with undesirable results. For example, pin sites and fresh wounds should be protected to avoid further complications.4,6,13,17 However, 1 patient in this study admitted that she went swimming in a public pool with the apparatus uncovered shortly after surgery. She later presented to the clinic with pus exuding from her wound. Another patient admitted that the bulky nature of the frame led him to use his own garage tools to dismantle it, which undoubtedly disrupted any potential bony healing occurring at the site of fusion.
It is important to note the final functional status of these patients. Only 3 patients in each group were unable to utilize a functional brace/rocker bottom shoe in their final functional outcome. Therefore, regardless of the fusion method, it appears that infection eradication and proper joint alignment, either through fibrous or bony union, provided the patient with a high likelihood of achieving a functional outcome with the assistance of various orthoses. We believe this finding should encourage surgeons to consider limb salvage by fusion, specifically with internal fixation, instead of performing early amputations.
Our study is not without limitations. First, the sample size of our study is relatively small, making it difficult to demonstrate statistical significance or achieve adequate power when comparing our 2 cohorts. Our power analysis revealed that significantly more patients were needed in both groups to avoid a beta error. However, utilizing ankle arthrodesis for limb salvage in the setting of chronic infection is a relatively uncommon procedure. Our study of 32 ankles is one of the larger series of chronic ankle osteomyelitis in the literature, and this study is unique in that it contrasts 2 different fusion methods utilized by the same surgeon.7,20 Another unique point about this study is that each of the cases of osteomyelitis initially resulted from a fracture about the ankle.
Second, our study contains all the flaws associated with a retrospective chart review, lacking in both group homogeneity and consistent follow-up times. In the internal fixation group, patients were treated with varying fixation devices, ranging from screws to fusion nails. This was based on the specific variables available for each patient’s unique clinical situation. However, all patients met the inclusion criteria and therefore were included in our retrospective review. With regard to follow-up time, many patients were indigent at our urban level 1 trauma center and had difficulty following up in an optimal fashion. It would be unlikely that a well-controlled prospective study could be successfully implemented in this setting. It is not uncommon for patients to be lost to follow-up for a variety of reasons such as social issues, drug use, and homelessness.
In conclusion, when analyzing these 2 fusion methods in posttraumatic patients with previous osteomyelitis of the ankle, both internal and external fixation provided sufficient limb salvage and functional outcomes for patients. Secondarily, there was no difference in union rates or infection clearance between these 2 methods. However, internal fixation is often more palatable for the patient during the postoperative period. This is especially pertinent in a noncompliant patient population. Based on these findings, we encourage surgeons to consider limb salvage through arthrodesis with internal fixation in the setting of chronic osteomyelitis, but only after all evidence of active infection has been eradicated.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.