Ankle arthroscopy was first introduced in the 1930s, but it was not widely adopted until the 1990s1. The introduction of arthroscopic instruments for small joints broadened the indications for ankle arthroscopy and allowed a greater number of surgeons to perform these procedures.
Early studies of complications were sparse, and were limited to reviews and surgeon surveys. Reported complication rates ranged from 0% to 17%, with neurologic injuries being the most common2-10. Ferkel et al. emphasized this with their report of ankle arthroscopies that were performed between 1983 and 19926,11, which, to our knowledge, is the last large series to specifically report on the complications of ankle arthroscopy.
As the practice of ankle arthroscopy has matured, surgical technique has likewise evolved, with a movement away from the use of invasive joint distraction. We sought to evaluate the types of complications and the complication rate of contemporary ankle arthroscopy in which noninvasive ankle distraction with a padded foot strap is used. We hypothesized that the overall complication rate would be lower, that the types of complications would change as compared with the types reported in prior reports, and that certain complications would be eliminated.
Retrospective Analysis
We retrospectively reviewed all ankle arthroscopies performed by the senior author (B.F.D.) between September 1999 and December 2007. Institutional review-board approval was obtained prior to chart review. The operative note and all postoperative clinical notes were analyzed. Complications were identified as being any problem that resulted from the ankle arthroscopy itself, the intraoperative positioning, or the associated postoperative management.
In addition to complications, the following were recorded for each patient: age and sex, preoperative and postoperative diagnoses, procedure(s) performed, duration of follow-up, weight-bearing restriction, tourniquet location and duration, and an active Workers’ Compensation claim.
Standard evaluation methods were used to determine the mean patient age, mean duration of follow-up, and complication rates by the various types. With use of the two-sided Fisher exact test, the complication rate for patients who had an active Workers’ Compensation claim was compared with that of patients who did not. The prevalence of complications relative to distraction time was evaluated with the Wilcoxon-Mann-Whitney test.
Operative Procedure
All surgical procedures were performed by a single surgeon (B.F.D.). The patients were placed supine on the operating table and received either general or spinal anesthesia with sedation to allow lower-extremity muscle relaxation. All patients received intravenous prophylactic antibiotics. A well-padded genitourinary-type limb holder was placed under the thigh of the limb that was to be operated on, taking care to ensure that the popliteal fossa was free from compression. Depending on the anticipated duration of surgery, a thigh or calf tourniquet was placed. For shorter-length operations, a calf tourniquet was used. After being prepared and draped, the leg was exsanguinated and the tourniquet was inflated. The Acufex noninvasive ankle distractor (Smith & Nephew, Andover, Massachusetts) was attached to the operating table, and the foot strap was placed as demonstrated in Figure 1. Tension was then applied across the ankle, with care taken to avoid skin injury. Gross and fine tension was adjusted until adequate visualization was achieved while allowing ankle dorsiflexion and plantar flexion. A scalpel was used to make a standard longitudinal anteromedial incision through the skin, medial to the anterior tibialis tendon, after which blunt dissection was used to penetrate the deeper layers down to the joint capsule, with care taken to avoid injuring the sensory nerves. The joint was entered with a 2.9-mm conical obturator and cannula. The 2.7-mm arthroscope was then introduced into the joint. Prior to making the incision for the anterolateral portal, we palpated the ankle and foot to identify and avoid the intermediate cutaneous branch of the superficial peroneal nerve. Similar to the incision made for the anteromedial portal, the incision for the anterolateral portal was also limited to skin only, and blunt dissection was used to penetrate the deeper layers down to the joint capsule. Posterior and accessory portals were not routinely created. Muscle relaxation, ankle plantar flexion, and appropriate positioning of the distractor typically allowed for satisfactory visualization. Throughout the procedure, various 2.9-mm and 3.5-mm arthroscopic instruments were utilized.
At the completion of the procedure, traction was released, 10 mL of 0.5% bupivacaine was injected around the incisions, and the portals were closed with 3-0 nylon sutures. A sterile dressing was applied, and then the tourniquet was deflated. A posterior plaster splint with ankle stirrup was then applied. Postoperatively, patients were discharged from the hospital on the day of surgery. Aspirin (325 mg, twice a day), unless contraindicated for medical reasons, was used as prophylaxis against deep venous thrombosis. Partial weight-bearing (approximately 10% of body weight, for balance) was allowed until the first postoperative visit. Patients were typically allowed to bear weight as tolerated at five days after surgery with use of a walking boot. An exception to this was made when a cartilage procedure, such as drilling of an osteochondral defect, was performed, necessitating a longer period of non-weight-bearing. Regardless of the procedure performed, the patients were encouraged to begin range-of-motion exercises on their own. At two to three weeks after surgery, the sutures were removed and the patient began wearing an ankle brace rather than a walking boot. The specific rehabilitation protocol was unique to each patient, depending on the pathology involved and the procedure performed. Typically this included early range-of-motion and strengthening exercises at home. If prolonged non-weight-bearing was required, formal physical therapy was initiated.
Source of Funding
There was no external funding source for this study.
Between September 1999 and December 2007, 307 ankle arthroscopies were performed on 294 patients. None of the bilateral cases were performed at the same time. For patients who underwent arthroscopy on both ankles during the study period, only the first procedure was included. Thus, 294 procedures were eligible for review. Of those included in the review, one hundred fifty-two patients were women (51.7%), one hundred forty-two patients were men (48.3%), and the average age was thirty-seven years (range, twelve to seventy-seven years).
While a large majority of the patients had joint synovitis and scarring, the most common diagnoses for which the arthroscopy was performed were osteochondral lesion (59.2%), loose body (36.1%), posttraumatic arthritis (14.3%), isolated synovitis (13.3%), and degenerative joint disease (12.6%). Thigh tourniquets were used in 254 cases and inflated for an average of sixty-four minutes (range, eighteen to 135 minutes). Calf tourniquets were used in forty cases and inflated for an average of fifty-five minutes (range, twenty-eight to 107 minutes).
There were a total of twenty complications, resulting in an overall complication rate of 6.8%. There were sixteen neurologic complications, which accounted for 80% of the total number of complications and resulted in a neurologic complication rate of 5.4% for all ankle arthroscopies. There were four non-neurologic complications, which included two superficial infections, one deep venous thrombosis, and one prolonged synovial-fluid drainage from a portal (Table I).
Details of Neurologic Complications
Portal-Associated Complications
Of the sixteen neurologic complications, six were related to the anterolateral portal. These patients characteristically experienced hypersensitivity or paresthesia over the anterolateral incision that was beyond what was considered to be normal incision-related pain. We believe that this represented an injury to the intermediate cutaneous branch of the superficial peroneal nerve. Although the specific timing of each symptom varied from patient to patient, in most cases, the hypersensitivity was present at the second office visit, which was typically two to three weeks after surgery, when the patients were returning to more normal activities. No specific therapy was initiated for these complaints, but the patient's symptoms were closely monitored. Five of the six patients with anterolateral portal nerve injury had complete resolution of signs and symptoms, typically within two to three months after surgery. One patient had hypersensitivity at the last follow-up visit, at 9.5 months after surgery. She first noticed the hypersensitivity at sixteen weeks after surgery, but, because it gradually began to decrease, she considered it to be insignificant.
One case was specifically related to the suture used to close the portal. The patient had shooting pain into her foot with plantar flexion of the ankle. After the removal of the suture from the anterolateral portal incision, her symptoms immediately improved. She continued to have dysesthesias but the symptoms had fully resolved by the six-week postoperative visit.
Nonportal-Associated Complications
Ten of the sixteen patients with neurologic complications experienced symptoms that were unrelated to an arthroscopic portal. On examination, the arthroscopic portal sites appeared benign in these patients, without any signs of hypersensitivity, numbness, or paresthesias.
Midfoot Dysesthesias
Eight of the sixteen patients experienced nerve hypersensitivity or dysesthesias localized to the dorsal part of the midfoot, where the foot strap had been applied over the sensory cutaneous nerve branches. These terminal nerve branches include the medial dorsal and intermediate dorsal cutaneous branches of the superficial peroneal nerve and the lateral dorsal cutaneous branch of the sural nerve. Six of these patients reported their symptoms at their second postoperative visit, typically two to three weeks after surgery. The symptoms in five of these six patients resolved by their next office visit, approximately two months after surgery. One of the eight patients, a thirty-six-year-old woman, had pain that persisted for one year. She was referred to a pain clinic, and the symptoms improved. Two of the eight patients experienced the onset of symptoms much later than was typically observed. One fifty-three-year-old man with a Workers’ Compensation claim underwent ankle arthroscopy for debridement of a full-thickness osteochondral lesion of the talus and removal of loose bodies. At twelve weeks after surgery, he complained of numbness and tingling on the dorsum of the foot. No formal intervention was initiated and his symptoms were resolved at the next office visit, twenty weeks following surgery. The other patient was a thirty-two-year-old man with a Workers’ Compensation claim who underwent ankle arthroscopy for debridement of isolated anterolateral synovitis. Eight months after surgery, he began complaining of nerve irritation on the dorsum of the foot. Symptomatic treatment was employed, including the use of nerve desensitization methods. At his latest follow-up visit at 2.5 years postoperatively, he continued to complain of these symptoms.
Other Neurologic Presentations
Two of ten nerve complications not associated with the portals were not localized to the dorsal part of the midfoot. One patient had findings consistent with a stretch injury as described by Dowdy et al12. He was a fifty-year-old man with a Workers’ Compensation claim who underwent ankle arthroscopy for debridement of a partial-thickness osteochondral lesion of the talus and hypertrophic synovium. Four months after surgery he began complaining of shooting pain in the dorsal and plantar aspects of his foot. The anatomic distribution was consistent with injury to the superficial peroneal nerve and posterior tibial nerve. There was no intervening trauma or injury that could explain the sudden onset of the symptoms after four months of being asymptomatic. He was referred to a pain clinic and, at his latest follow-up at 8.5 months, he remained symptomatic.
A second patient experienced postoperative dysesthesia of unclear etiology. The nerve-dysfunction symptoms were noted on follow-up evaluation, but the location was not documented in the medical record. The source of the symptoms was not determined. These symptoms resolved by the patient's next office visit.
Details of Non-Neurologic Complications
There were four non-neurologic complications, which indicated a non-neurologic complication rate of 1.4% for all ankles in the study. One healthy twenty-three-year-old man developed a deep venous thrombosis in the calf, diagnosed on postoperative day 14, and was treated with warfarin followed by aspirin. His symptoms resolved without complication. Another patient experienced prolonged wound drainage from the anteromedial portal. At three and one-half weeks postoperatively, the anteromedial wound had gapped open and was producing a small amount of synovial fluid drainage. Steri-Strips (3M, St. Paul, Minnesota) were applied, and the wound healed without difficulty. One patient developed a superficial portal infection that was treated with oral antibiotics. Another patient developed a superficial stitch abscess that was decompressed in the office and treated with topical bacitracin ointment.
Of the 294 total cases reviewed, there were thirty-eight Workers’ Compensation-associated cases. There was a 21% complication rate in this group of patients (eight of thirty-eight) versus a 4.7% complication rate for the group of patients who were not involved in an active Workers’ Compensation claim (twelve of 256). These two rates are significantly different (p < 0.005).
This study is the first large series, to our knowledge, to evaluate the complications of ankle arthroscopy in which contemporary noninvasive techniques are used. Until now, reports of complications were based on fifteen to twenty-year-old techniques that combined both invasive and noninvasive distraction1,6,13,14. Our results demonstrate that many non-neurologic complications have been eliminated, resulting in a decrease in the overall complication rate as compared with the rates reported in prior studies. While the neurologic complication rates did not decrease, the pattern of neurologic injury raises the awareness of potential mechanisms of nerve injury, including the contributions attributed to the foot strap of the distraction device.
Ferkel et al. published a comprehensive analysis of complications of ankle arthroscopy performed from 1983 to1992 and specifically reported on the propensity for neurologic complications6. In that series, the authors noted an overall complication rate of 9.0%, with neurologic complications accounting for 49% of these. Utilizing both invasive and noninvasive distraction techniques, they noted that all of the neurologic complications were the direct result of portal placement or distractor-pin placement. Although neurologic complications were the most common, the rate of non-neurologic complications was 4.6% and most commonly involved superficial and deep infection, pain at the site of the distractor pin, and fracture at the site of the distractor pin. Invasive distraction directly contributed to 25% of non-neurologic complications.
We believe that joint distraction remains an important component of ankle arthroscopy, but in an attempt to lower the rate of complications, we exclusively utilize a noninvasive distraction technique. With this technique, the location of the foot strap is important. Ideal placement is just below the ankle in the hindfoot (Fig. 1), avoiding a more superficial midfoot location, which may have less soft tissue available to protect underlying nervous structures.
After reviewing 294 ankle arthroscopies performed by a single surgeon using a single noninvasive technique, our overall 6.8% complication rate was lower than the benchmark of 9.0%6. When we excluded patients who were involved in Workers’ Compensation claims, our complication rate was reduced to 4.7%. Patients receiving Workers’ Compensation experienced complications at a rate of 21%. Consistent with previous reports3,4,6, neurologic injury was the most common complication in our study, accounting for 80% of all complications. This proportion of neurologic injuries is higher than previously reported, but the shift was largely due to the decrease in non-neurologic complications from a rate of 4.6% to 1.4%. This was a result of a lower infection rate and the elimination of complications associated with invasive distraction, such as pin-site pain and fracture. Having avoided the neurologic complications associated with invasive distraction, we suspected that the overwhelming majority of our nerve complications would be due to injury at the anterolateral portal. However, only 38% (six of sixteen) of our neurologic complications were attributable to the anterolateral portal, and eight of our sixteen neurologic injuries were characterized by dysesthesias or hypersensitivity in the midfoot in the distribution of the cutaneous nerves underlying the foot strap. Given the anatomic nature of the signs and symptoms, we postulate that the padded foot strap may have been responsible. It may be the superficial location of the cutaneous nerves that makes them vulnerable to injury from the foot distraction strap. After an extensive review of the literature, we were unable to identify any previously reported nerve injury that matches this description. These neurologic signs and symptoms in this subset of patients are distinct from the hypersensitivity or numbness often found around the anterolateral portal. However, due to the nature of this retrospective study, we cannot definitively conclude that the use of a distractor bar and padded foot strap is responsible for the dorsal midfoot neurologic symptoms. We continue to utilize the same padded foot strap and distraction technique because we believe that the overall intraoperative benefits outweigh the potential risks. This risk may be minimized by ensuring that the hip is flexed to 45°, as well as by adjusting the distraction strap to minimize excessive plantar flexion of the ankle.
In this study, patients who were involved in an active Workers’ Compensation claim had a significantly higher complication rate than those who were not. When specifically looking at neurologic injury, the complication rate was 16% for Workers’ Compensation patients and only 3.9% for patients not involved with Workers’ Compensation. The medical literature is replete with studies examining the effect of Workers’ Compensation status on outcomes of surgery. The majority of studies, including at least two meta-analyses, show a strong association between Workers’ Compensation status and poorer outcome15,16. This information should be kept in mind when counseling patients about the risks and benefits associated with ankle arthroscopy.
To clarify the potential contribution of tourniquet location to complications, we analyzed both tourniquet location (the proximal part of the thigh or calf) and duration as a reasonable approximation of total distraction time. In considering the location of the tourniquet, only one patient experienced dysesthesias following use of a calf tourniquet. With only one complication in this group, we are unable to draw meaningful conclusions regarding the rate of dysesthesias in relation to the location of the tourniquet. When neurologic complications were analyzed in relation to the duration of joint distraction, a nonsignificant trend toward increasing prevalence of complications was noted as distraction time increased. This information highlights the need to be aware of distraction time, and to make efforts to release this distraction at the earliest opportunity. In prolonged procedures, we would also suggest decreasing the force of ankle distraction to relax the soft tissue after approximately one hour.
Previous literature has examined the safety of noninvasive ankle distraction. Dowdy et al. elicited reversible nerve conduction changes to the deep and superficial peroneal cutaneous nerve branches when awake volunteers underwent noninvasive ankle distraction. The degree of nerve conduction changes was associated with the magnitude of the distraction force and the duration of traction12. Similarly, in our series, we identified one patient who sustained an apparent distraction injury to the superficial peroneal nerve and tibial nerve. In a series that made use of the same proprietary distraction technique as that used by Dowdy et al., Amendola et al. reported three neurologic complications in seventy-nine cases, but the etiology of the nerve injury was unclear in two of them. None of the injuries had resolved at the time of final follow-up17.
Despite the consistent risk of injury to the cutaneous branches of the superficial peroneal nerve, the placement and usage of the anterolateral portal is fundamental to ankle arthroscopy. Our 2% rate of neurologic injury at the anterolateral portal is similar to the 2.5% rate reported by Ferkel et al. Their anatomic descriptions of the key nerves and the technical pearls to prevent injury at the portal sites are very instructive and are recommended reading for anyone performing ankle arthroscopy6.
It is important to note the limitations of this study. As a retrospective study, it relied on chart documentation. There was, however, specific assessment of the sensation at the portals and throughout the distal cutaneous distribution at each clinic visit. Second, the force of traction placed across the ankle was not quantified. We do not have data to evaluate whether the patients who experienced complications were exposed to larger traction forces than were the patients who did not experience complications.
In conclusion, ankle arthroscopy with noninvasive distraction is a safe and reproducible procedure. Our findings reaffirm that nerve injury is the most common complication associated with ankle arthroscopy. By exclusively utilizing a noninvasive distraction technique, we noted a marked reduction in non-neurologic complications. It is important to note that the use of noninvasive distraction carries a risk of neurologic injury beyond the risk associated with the arthroscopic portals, with this risk being particularly high in the patients with Workers’ Compensation claims. We propose that the padded foot strap may contribute to these dysesthesias. These findings invite future investigation regarding the optimal design of the foot strap and the safest magnitude and duration of distraction force across the ankle joint.