Abstract
Background:
Wound complications following total ankle replacement (TAR) potentially lead to devastating consequences. The aim of this study was to compare the operative and demographic differences in patients with and without major wound problems which required operative management. We hypothesized that increased tourniquet and operative time would negatively influence wound healing.
Methods:
We identified a consecutive series of 762 primary TARs performed between December 1999 and April 2014 whose data were prospectively collected. We then identified the subset of patients who required a secondary surgery to treat major wound complications (ie, operative debridement, split-thickness skin grafting, and soft tissue reconstruction). All patients requiring a second surgery had operative wound debridement. We then compared the demographics, operative characteristics, and functional scores to see if any differences existed between patients with and without major wound complications. Clinical outcomes including secondary procedures and implant failure rates were recorded.
Results:
Twenty-six patients (3.4%) had a total of 49 operative procedures to treat major wound issues. Eighteen patients had flaps and 14 had split-thickness skin grafts. The median time to operatively treating the wound was 1.9 (range: 0.5-12.5) months after the index TAR. The median follow-up time from the wound procedure was 12.7 (range: 1.2-170.8) months. Compared to the control group, patients with major wounds had a significantly longer mean surgery (214.8 vs 189.3 minutes, P = .041) time and trended toward a longer median tourniquet time (151 vs 141 minutes, P = .060). Patients without wound complications were more likely to have posttraumatic arthritis, whereas those with wound complications were more likely to have primary osteoarthritis (P = .006). The control group trended toward having a higher mean BMI (29.5 vs 27.2, P = .056). There were 6 failures in the major wound complication cohort (23.1%), including 2 below the knee amputations.
Conclusion:
Ankle wounds that required operative management had high failure rates and some resulted in devastating outcomes. We did not find any increase in major wound complications in those with various risk factors as identified by other studies. Given our data, we recommend limiting operative time. While correcting hindfoot and midfoot alignment is important for improving patient functionality and survivorship of the implant, thought should be given to staging the TAR if multiple pathologies are to be addressed at the time of surgery to limit operative time.
Level of Evidence:
Level III, retrospective comparative series.
Introduction
The advent of modern total ankle replacements (TAR) has resulted in an exponential increase in utilization of this emerging technology over the past two decades in the United States and abroad.22,25 Though instrumentation and prosthesis designs have exhibited improving sophistication from a mechanical perspective,2-4,6,14,27 little has been examined concerning the management of the soft tissue at the time of index procedure. Although radiographic alignment, revision rate, disease or joint-specific patient-reported outcome measures and generic health related quality of life represent TAR’s most tangible and targeted metrics of success,13,21 wound complications remain its most devastating negative consequence, with amputation as the most feared sequela.
Soft tissue reconstruction options and incidence of their employment have previously been described as evaluated within a prospectively collected TAR database. 8 Details of enrollment and outcome measures in this registry at the authors’ academic medical center have been published prior.9-11,30 Previous studies from this database have evaluated patient-level characteristics for associations with wound complications and found that age, 7 diabetes mellitus, 9 and obesity 11 have demonstrated no measurable influence upon wound complication rates. To date, potential intraoperative predictors of wounds have yet to be analyzed within the context of a large TAR cohort.
The aim of the present study was to compare operative and demographic characteristics in patients with and without wound complications that required operative intervention. We hypothesized that increased tourniquet and operative times were controllable intraoperative factors that may negatively correlate with proficient wound healing secondary to prolonged tissue ischemia. Secondarily, we predicted that those with operative wound complication issues would have lower functional outcome scores.
Methods
After receiving Institutional Review Board approval, we retrospectively reviewed the prospectively collected data of all patients who underwent TAR at our institution. This included a consecutive series of 762 primary TARs performed between December 1999 and April 2014. Each patient’s medical chart was reviewed and all subsequent procedures were noted. We then identified a subset of patients who required a secondary surgery to treat major wound complications (ie, operative debridement, split-thickness skin grafting, and soft tissue reconstruction). The control group (736) consisted of consecutive patients who did not have major wound complications. All demographics including age, sex, race, American Society of Anesthesiologists (ASA) score, body mass index (BMI), smoking history, previous operative history, and etiology of arthritis and operative information including surgical implant, operative time, tourniquet time, and intraoperative complications were collected. Operative time was defined as the point from incision to when the cast was placed. Tourniquet time was defined as the time the thigh tourniquet was consecutively inflated.
All index TAR were performed at our institution by foot and ankle surgeons with extensive TAR experience. The indication for the index TAR included incapacitating pain and dysfunction in an arthritic ankle joint that failed conservative treatment. Initial wound debridements were performed by the initial foot and ankle orthopedic surgeon (every patient included in this study had at least one debridement by one of the senior surgeons) per plastic surgery’s wishes; all local or free flaps were performed by the Division of Plastic and Reconstructive Surgery. Only primary arthroplasty procedures were included; patients were excluded if they underwent a revision TAR procedure. Patients were also excluded if treatment of their postoperative wound only required local wound care or an in-office minor wound debridement.
A standard 10-cm anterior ankle arthrotomy was used for each TAR while trying to incorporate any previous operative incisions. The ankle prosthesis was implanted according to the manufacturer’s guidelines. Prior to closure, the operative site was copiously irrigated with a dilute antibiotic solution. The operative site was also closed in layers including the deep ankle capsule, the tibialis anterior or extensor hallucis longus tendon sheath, the subcutaneous tissue layer, and the skin with 4-0 Nylon vertical mattress sutures.
Postoperatively, all patients were placed into a non-weight-bearing short leg cast and given nasal oxygen for 24 hours. Two weeks postoperatively, the cast was removed and the wound inspected. If the wound skin edges were clean, dry, and intact, the sutures were removed and the patient allowed to weight-bear as tolerated in a CAM walker boot. If the skin edges were macerated or evidence of drainage, erythema, or other signs of delayed healing were found, the sutures were left in place and the ankle recasted for an additional 7 days. If the operative site continued to be concerning (ie, ongoing drainage, erythema, or dehiscence) the patients were sent for an immediate evaluation by plastic surgery.
The decision to proceed with either local wound care versus immediate flap coverage was dictated by the presence of exposed TA/EHL tendon, neurovascular structures, or hardware as described by Cho and colleagues. 5 If these structures were not seen following a local in-office wound debridement, then a vacuum-assisted wound closure dressing, biologic matrix, or dressing change was employed. If the wound failed to heal after 2 weeks, then a rotational or free flap was performed. If vital structures were exposed and drainage was found, the patient’s wound was debrided to determine the extent and depth of the soft tissue failure. If the infection was superficial without communication with the prosthesis, all necrotic tissue was debrided, the area copiously irrigated, and the wound primarily closed. In situations where immediate primary closure was not possible, a vacuum-assisted wound closure dressing was applied until edema improved. If the vacuum-assisted wound closure failed to bring the skin edges together, the wound bed was covered with a split-thickness skin graft or dermal substitute. If the vital structures remained exposed and significant soft tissue defect was present, the patient underwent flap coverage.
Clinical outcomes including wound healing and implant failure rates were recorded preoperatively, at 6 months, and annually thereafter. The American Orthopaedic Foot & Ankle Society (AOFAS) hindfoot-ankle score, Short Musculoskeletal Function Assessment (SMFA), Foot and Ankle Disability Index (FADI), and Foot and Ankle Outcome Score (FAOS) were used to evaluate preoperative and final postoperative function as has been reported previously in TAR.17,29 The Short Form-36 (SF-36) was used to assess overall health. Pain was rated using a visual analog scale (VAS), ranging from 0 (no pain) to 100 (maximum pain). Functional and pain outcomes were compared between patients who had major operative wound management and those who did not. We also compared the baseline demographics, operative characteristics, and failure rates to determine if any differences existed between the wound complication and no wound complication groups.
Statistical Analysis
Continuous data were summarized using the mean, median, range, and standard deviation as appropriate for continuous variables and counts and percentages for categorical variables. An analysis of variance was used to compare normally distributed demographic data, whereas the Kruskal-Wallis test was used for skewed data. Categorical variables were compared using either a chi-squared test or Fisher exact test in the presence of small expected cell counts (< 5). All statistical tests were 2-sided, and the significance level for all tests was P <.05. All statistical analyses were performed using SAS, version 9.4 (Cary, NC).
Results
Demographics
We identified 26 patients after total ankle arthroplasty who had a total of 49 operative procedures (median= 1.5; range= 1 to 6) to treat wound healing complications (3.4% of all prostheses) (Table 1). Overall, 44.4% of patients had a concomitant procedure at the time of initial TAR (Table 2). Within patients having wound healing complications, the average age was 63.8 ± 11.0 (range: 37.0-83.0) years at the time of the initial TAR. The median follow-up time for these patients was 13.2 (range: 4.2-174.8) months after TAR and 12.7 (range: 1.2-170.8) months after the final wound procedure. Fourteen patients were male (53.8%) and the mean body mass index (BMI) was 27.1 ± 3.7. Comorbidities included diabetes (9.5%), coronary artery disease (7.7%), hypertension (34.6%), arrhythmia (3.8%), and history of smoking (46.2%). The average American Society of Anesthesiologists (ASA) score was 2.5 ± 0.6.
Demographics.
Abbreviations: ASA, American Society of Anesthesiologists INBONE I/II (Wright Medical, Arlington, TN); BMI, body mass index; OA, osteoarthritis; PT, posttraumatic; RA, rheumatoid arthritis; Scandinavian Total Ankle Replacement (STAR; Stryker, Kalamazoo, MI), and Salto-Talaris (Tornier, Bloomington, MN).
Concurrent Procedures at the Time of Total Ankle Arthroplasty.
Eight patients (30.8%) developed a deep space infection that required a total of 14 incision and drainage procedures with 5 polyethylene exchanges. Each patient had an initial incision and debridement. Unfortunately, there was no consistency as to when the polyethylene was exchanged. Two patients (7.7%) eventually needed a below-the-knee amputation after one flap failure, and 1 patient (3.8%) had a soft tissue defect after persistent deep wound infection. Each was still infected. Two additional patients needed a revision of both the tibial and talar component following antibiotic cement spacer placement for a deep infection.
Comparison to Control Group
When comparing the demographics of the control group to the patients with operative wound complications, mean age (63.8 vs 62.7) and percentage male (53.8 vs 46.5%) were both similar between the patients with and without wound complications, respectively. Patients without wound complications trended toward having a larger mean BMI (29.4 vs 27.2), but the difference did not reach statistical significance (P = .056). Both wound complication groups had similar preoperative health comorbidities, including diabetes, coronary artery disease, hypertension, smoking history, and ASA score. No implant type had a significantly higher rate of operative wound complications (P = .361). However, the control group had a higher preoperative posttraumatic etiology for ankle arthritis (70.4% vs 42.9%), and the wound group had a higher rate of primary osteoarthritis (42.9% vs 13.7%, P = .006). Mean operative time was significantly longer in patients who had postoperative wound complications (214.8 vs 189.3 minutes; P = .041). Similarly, median tourniquet time was trending longer in patients with wounds compared with the control group (151 vs 141 minutes; P = .060).
Preoperatively (Table 3), patients with major wound complications had a significantly lower AOFAS Hindfoot (32.3 vs 40.7; P = .018) and FADI index score (0.46 vs 0.55; P = .014), whereas the preoperative FAOS, SF-36, SMFA, and VAS scores were similar. Postoperatively, patients with wound complications had worse FAOS pain subset scores compared with the control (68.3 vs 81.6; P = .044), as well as worse SMFA function scores (67.0 vs 57.8; P = .047).
Average Outcome Scores for Wound and Control Group Patients Preoperative and at Most Recent Follow-up.
Abbreviations: AOFAS, American Orthopaedic Foot & Ankle Society; FADI, Foot and Ankle Disability Index; FAOS, Foot and Ankle Outcome Score; SMFA, Short Musculoskeletal Function Assessment; VAS, visual analog scale.
P value for pairwise comparison of wound to the control group.
Discussion
Wound complications following a total ankle arthroplasty can potentially lead to loss of limb. 8 Dorsal ankle skin can be unforgiving as it has limited mobility following trauma/surgery; it is also thin without significant muscle bulk. 5 Because of this, even minor wounds can expose underlying tendons and neurovascular structures, which can, in turn, lead to deep infection and implant failure. Preoperative planning regarding the need to perform time-consuming procedures, such as hindfoot realignment, may give surgeons a better idea if they are increasing the risk for postoperative major wound complications. The patient should be involved in this decision-making process. Additionally, any preventative strategies to reduce the incidence of major wound complications may potentially lead to limb salvage or improved outcomes. Such strategies include trying to limit operative and tourniquet time.
When engaged in the shared decision-making process with patients considering operative treatment options for end-stage ankle arthritis, the discussion of wound complications is nebulous at best. Though numerous publications have reported on wound complication rates, the inferences to be made from these heterogenous results are unclear. Mercer et al. conducted a systematic review of total ankle arthroplasty literature and found the reporting of wound complications and adverse events to be highly variable. In 117 studies evaluated by their group, 572 unique terms were used to describe undesirable outcomes of surgery. 18 The inconsistency of reporting and lack of standardization across our specialty leave foot and ankle orthopedic surgeons without reliable data to support or refute purported rates of wound complications following TAR.
Whalen et al 32 looked at 57 consecutive total ankle arthroplasties with a 28% wound complication rate. Analysis of various risk factors included cardiovascular disease, peripheral vascular disease, and a smoking history of greater than 12 pack-years. Recently, Patton et al. examined risk factors for TAR infection in a large cohort, identifying 29 cases out of 966 operative patients spread out over a 17-year period. Though they did not specifically look at risk factors for wound complications, they found wound healing problems more than 14 days after surgery, along with diabetes and prior ankle surgery, to be predictive of postoperative infection. Interestingly, no significant difference was found between infected and uninfected groups with respect to smoking, body mass index and operative time. 24 Raikin et al. reviewed risk factors for post-operative TAR wound complications in a small series of patients. 26 Nine patients (8.5%) had major wound complications that required a return to the operative room for formal debridement. Eventually, 3 of those ankles required implant removal, whereas 1 patient required flap coverage and 5 were treated with a vacuum-assisted closure device without need for revision surgery. Two of the explanted ankles eventually fused after placement of a multiplanar external fixator and the other needed a below-the-knee amputation. Patients who had an inflammatory connective-tissue disease, females, and corticosteroid use had a higher risk to develop a major wound following surgery. A history of inflammatory connective-tissue disease was the only risk factor to show a statistically significant increase for major wound complications after multivariate logistic regression analysis (odds ratio = 14.03, 95% confidence interval = 2.35-83.8, P = .004). These patients had all immunomodulators stopped for at least 1 month prior to surgery. Diabetic patients tended to have a significantly higher risk of minor wound complications (treatable in the office with local wound care).
In our series, patients with rheumatoid disease or women were not associated with having a major wound complication. We did not, however, record the use of corticosteroids. Although Raikin and colleagues mentioned the recording of tourniquet time, their article does not record the time between the 2 groups, nor does it say if the power is strong enough with only 9 patients with major wound complications. Interestingly, our operative wound patients had a significantly higher pre-operative diagnosis of primary osteoarthritis compared to our control group. One would expect post-traumatic patients to have a scarred and less pliable soft tissue envelope which could lead to more wound complications. Perhaps patients with primary osteoarthritis had a higher rate of ankle corticosteroid injections prior to surgery, which could potentially lead to wound complications or deep infection. 31
Tourniquets are frequently used in foot and ankle surgery. A pneumatic tourniquet allows for greater operative precision and improved technical ease because of a decrease in intraoperative blood loss and better visualization. 33 However, tourniquet use is not without its limitations and consequences as argued widely in the total knee arthroplasty and arthroscopy literature. Adverse events can and do occur with tourniquet use, including: vascular injury, 15 neuropraxia,19,20 increased postoperative pain, 12 postoperative quadriceps weakness, 28 and joint stiffness. 1 In a prospective study of knee arthroplasty patients, tourniquet time (more than 100 minutes) was associated with an increase in superficial wound infections. 23 Furthermore, elevated levels of proteolytic activity are seen in the downstream muscle tissue following a thigh tourniquet. 16 Although tourniquet time has not been implicated in serious foot and ankle wound complications to date, it is feasible that local tissue ischemia in an already tenuous soft tissue envelope can adversely affect the healing of an ankle arthrotomy. A reperfusion injury may occur in the local soft tissues.16,20 Furthermore, increased operative time and the use of retractors may stifle local blood flow and delay healing. It is also possible that the longer tourniquet time reflects a more complex case with the possibility of increased soft tissue swelling that could increase the risk of a wound complication. Therefore, we advocate limiting tourniquet time.
This study has a number of limitations. It is a retrospective review of a nonrandomized patient population; thus, selection bias may have been an influencing factor. We were unable to retrospectively record who used corticosteroids or who had a recent ankle steroid injection. Our database did not record who had minor wound complications that were treated with in-office care. These wounds may have led to deeper wounds and larger soft tissue defects if the natural history played out. Therefore, we could be missing a different subset of patients who could have been analyzed. These local wound care patients were analyzed in the same group with the patients who had normal healing and could influence the differences between our control group and those with operative wound debridement. In terms of operative timing, we did not assess the need for coronal plane deformity correction, which could have added significant time to the procedure.
Conclusions
Ankle wounds that require operative management have high failure rates and may result in devastating outcomes. We did not find any increase in major wound complications in those with various risk factors as identified by other studies. Given our data, we recommend limiting tourniquet and operative time. Although correcting hindfoot and midfoot alignment is important for improving patient functionality and survivorship of the implant, thought should be given to staging the TAR if multiple pathologies are to be addressed at the time of surgery to limit operative time.
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.