Osteochondritis dissecans (OCD) is an idiopathic-acquired

Transcription

1 ORIGINAL ARTICLE Results of Surgical Management of Osteochondritis Dissecans of the Ankle in the Pediatric and Adolescent Population Dennis E. Kramer, MD, Michael P. Glotzbecker, MD, Benjamin J. Shore, MD, MPH, FRCSC, David Zurakowski, PhD, Yi-Meng Yen, MD, PhD, Mininder S. Kocher, MD, MPH, and Lyle J. Micheli, MD Background: There is a paucity of published literature on operative management of osteochondritis dissecans (OCD) in the ankle in adolescents. This study seeks to elucidate patient and lesion characteristics associated with surgical success and failure as well as reporting functional outcomes. Methods: Retrospective chart review identified all patients aged 18 years old or younger surgically treated for OCD of the ankle at our institution from 2001 to This included 109 ankles in 100 patients (75 female, 25 male), mean age 14.3 ± 2.3 years (range, 7 to 18 y), with a median follow-up of 3.3 years (range, 1 to 10.8 y). Patient and lesion data, surgical procedure, clinical results, and complications were recorded. Postoperative radiographs were reviewed in 80 ankles. A return to sport survey and a Foot and Ankle Outcome Score (FAOS) was sent to all patients. Multivariate statistical analysis evaluated predictors of reoperation rate, Berndt and Harty clinical grade, and FAOS scores. Kaplan-Meier analysis was applied to determine freedom from reoperation. Results: The OCD lesion was most commonly found on the medial talus (80, 73%). The most common procedures performed included transarticular drilling (59, 54%), fixation (22, 20%), and excision microfracture (27, 26%). The overall rate of reoperation was 27% (29/109). Berndt and Harty clinical grade was poor (33, 30%), fair (23, 21%), and good (53, 49%). Reoperation rates were significantly higher for OCD lesions in which postoperative radiographs had no change or looked worse (10/16, 63%) (P = 0.002). Thirty-six of 44 survey respondents (82%) were satisfied and 37 (84%) returned to sports at a median time of 6 months. Average FAOS score was 77 ± 18. Multiple linear regression confirmed that female sex and elevated body mass index were significant negative predictors for FAOS score. From the Division of Sports Medicine, Department of Orthopaedic Surgery, Boston Children s Hospital, Harvard Medical School, Boston, MA. The authors declare no conflicts of interest. Reprints: Dennis E. Kramer, MD, Department of Orthopaedic Surgery, Division of Sports Medicine, Boston Children s Hospital, Harvard Medical School, 300 Longwood Ave, Boston, MA dennis.kramer@childrens.harvard.edu. Copyright r 2014 by Lippincott Williams & Wilkins Conclusions: The reoperation rate following surgical intervention for OCD of the ankle is high. Females and those with a higher body mass index may have worse subjective functional ankle outcomes. Level of Evidence: Level IV retrospective case series. Key Words: ankle, articular cartilage, osteochondritis dessicans, pediatric sports medicine (J Pediatr Orthop 2014;00: ) Osteochondritis dissecans (OCD) is an idiopathic-acquired focal injury to subchondral bone that may secondarily progress to involve the overlying cartilage. The ankle is the third most common location for OCD lesions after the knee and elbow. 1 OCD of the ankle most commonly occurs in the medial or lateral talus and presents in adolescents at age 13 to 14 years with vague chronic ankle pain. 2 There may be a variable history of ankle sprains but there is typically not a single traumatic event. These lesions have been preoperatively classified into stages by their radiographic or magnetic resonance imaging appearance. 3 5 Nonoperative modalities such as immobilization, protected weight bearing, and/or activity modification are typically indicated for all stable lesions and some partially stable lesions in skeletally immature patients. Surgical intervention is generally reserved for symptomatic stable lesions that have failed 6 months of nonoperative management or for unstable lesions. Many surgical options exist for treatment of OCD of the ankle based on the size and stability of the lesion. These include attempts to heal the OCD lesion and preserve native cartilage through retrograde or transarticular drilling with or without internal fixation. 6 9 Nonsalvageable OCD lesions are commonly treated with excision followed by microfracture/marrow stimulation or osteochondral autograft (or allograft) transplantation. 10,11 There is a paucity of published literature on the results of operative management of OCD of the ankle in the pediatric population. We present a case series of patients who underwent surgery for OCD of the ankle at our institution over the last 10 years. This study seeks to review J Pediatr Orthop Volume 00, Number 00,

2 Kramer et al J Pediatr Orthop Volume 00, Number 00, 2014 the outcomes of surgical treatment of ankle OCD in the pediatric population, specifically investigating patient and lesion factors associated with surgical success from a radiographic and functional perspective. METHODS After institutional review board approval, computerized medical records were queried to identify all patients aged 18 years old or younger who were surgically treated for OCD of the ankle at our institution from 2001 to This search identified 126 pediatric patients of whom 26 had <1-year follow-up and were excluded. This resulted in a study population of 100 patients (75 female, 25 male), which included 9 patients who had OCD surgery on both ankles (109 ankles). Six of the 91 patients who had surgery for OCD on 1 ankle also had an OCD on the contralateral ankle that was treated conservatively. Mean age of the cohort was 14.3 ± 2.3 years (range, 7 to 18 y), and median follow-up was 3.3 years (range, 1 to 10.8 y; interquartile range, 2.0 to 5.1 y). In most patients, the indication for surgery was a symptomatic OCD lesion that failed to heal with conservative management as per prior studies. 12 In 12 cases surgical intervention was recommended at the initial clinic visit due to lesion instability. The cohort included 1 patient with neurofibromatosis-1, 1 patient with pauciarticular juvenile rheumatoid arthritis, and 1 patient with Ehlers-Danlos syndrome. For each patient, review of clinical and radiographic records was performed to record patient and lesion data, surgical procedure, clinical results, and complications. Patient data included age at surgery, sex, history of trauma, duration of symptoms, weight, height, and body mass index (BMI). Lesion data recorded included size, location, and status of the distal tibial physis. Preoperative radiographs were available on 83/109 patients and the OCD lesion was staged for these using standard Berndt and Harty criteria. 3 Preoperative magnetic resonance images were available on 61 patients and used to confirm the diagnosis, assess lesion size, location, and stability. Postoperative radiographs were reviewed by 1 of 3 different authors (all attending orthopaedic surgeons) in 80 ankles at a median time of 1.1 years postoperatively (range, 3 mo to 9.6 y) and were classified as healed if complete osseous union was observed, improved if osseous union was improved compared with preoperative radiographs, unchanged if the lesion looked the same, or worse if signs of osteoarthritis were observed (Fig. 1). Chart review recorded surgical complications, reoperation rate, and the Berndt and Harty clinical grade 3 for all patients. The Berndt and Harty clinical grade is defined as good if patients are pain free or have mild nondisabling symptoms, fair if symptoms are improved from before surgery but some measure of disability persists, and poor if symptoms are unchanged or if a reoperation was necessary. Other authors have also used this grading system. 2,7,9,13,14 Attempts were made to contact all patients and have them complete a return to sport survey and Foot and Ankle Outcome Score (FAOS). 15 The return to sport survey focused on their sport participation before and after surgery, satisfaction with surgery, and residual symptoms. The 9 patients who underwent surgery on both ankles were given modified return to sport survey (specific to each ankle) and were asked to complete the FAOS based on their more symptomatic ankle. Statistics Statistical analysis was performed to identify predictors of reoperation, Berndt and Harty clinical grade, and FAOS score. Risk factors assessed included: sex, age, status of physis, BMI, duration of symptoms, lesion size, location, Berndt and Harty stage, and procedure performed. Univariate analysis included the Fisher exact test and the w 2 test with respect to reoperation and return to sports. FAOS scores were compared between procedures using the Student t test and radiographic appearance using analysis of variance (ANOVA). Freedom from reoperation was determined using the Kaplan-Meier time to event method with Greenwood s formula applied to calculate 95% confidence intervals around the curve. Multivariate Cox proportional hazards regression was utilized to test for possible independent factors predictive of time to reoperation. Independent predictors of FAOS score were tested using multiple linear regression. Two-tailed values of P < 0.05 were considered statistically significant. Power analysis indicated that the sample size of 100 patients (109 ankles) provided 80% power to estimate the reoperation rate to within a precision of 10% using a Kaplan-Meier survivorship model. OPERATIVE TECHNIQUE All procedures were performed by 1 of 4 different fellowship-trained orthopaedic surgeons with the majority (103/109) performed by 3 of the surgeons (Fig. 2). All surgeons performed both lesion salvage and lesion excision procedures. In most cases, the lesion was initially visualized by standard 2-portal anterior ankle arthroscopy. Fluoroscopy was utilized to aid in lesion localization when necessary but not routinely. An arthrotomy was performed at surgeon discretion when deemed necessary for adequate access to the OCD lesion. Most lateral lesions (17/22) were well visualized and addressed through an arthroscopic approach. When required, an anterolateral arthrotomy was performed by extending the anterolateral arthroscopic portal proximally and dissecting lateral to the peroneus tertius. The ankle was held in maximum plantar flexion to expose the lateral lesion. For medial lesions which can be posterior or near the medial malleolus, the lesion was visualized and addressed either arthroscopically with the ankle in plantar flexion (42/80) or through an open-anterior approach with the ankle in plantar flexion (7/80) or an open-posteromedial approach with the ankle in dorsiflexion (31/80). The anterior arthrotomy was made medial to the tibialis anterior tendon. The posteromedial approach involved a small 3 cm posteromedial incision and dissection in the plane between the posterior tibialis tendon and the flexor digitorum longus. The neurovascular bundle was retracted posteriorly as a posteromedial arthrotomy was made. The lesion was then well visualized with the 2 r 2014 Lippincott Williams & Wilkins

3 J Pediatr Orthop Volume 00, Number 00, 2014 Surgical Management of OCD of the Ankle FIGURE 1. Preoperative (left column) and postoperative (right column) anteroposterior radiographs depicting an OCD lesion classified as healed (top row), improved (middle row), and worse (bottom row). OCD indicates osteochondritis dissecans. ankle in maximum dorsiflexion. This approach was often necessary for posteromedial lesions in skeletally immature patients in whom medial malleolar osteotomy 9 is contraindicated, 2,14 but was also found to be adequate in skeletally mature patients in whom osteotomy would have been an option. Once visualized the size, location, and stability (using an arthroscopic probe) of the OCD lesion was r 2014 Lippincott Williams & Wilkins 3

4 Kramer et al J Pediatr Orthop Volume 00, Number 00, 2014 FIGURE 2. A 14-year-old female with an OCD lesion of the medial talus. A, Preoperative AP radiograph. B, Preoperative T1- weighted sagittal MRI. C, Intraoperative arthroscopic appearance showing a hinged OCD lesion. D, Intraoperative lateral fluoroscopic image showing transarticular drilling through a posteromedial arthrotomy. E, Postoperative AP radiograph demonstrating improved appearance of the OCD lesion. AP indicates anteroposterior; MRI, magnetic resonance imaging; OCD, osteochondritis dissecans. determined. The OCD lesion then underwent 1 of 3 treatments (transarticular drilling, drilling with internal fixation, or excision and microfracture/marrow stimulation) depending on its intraoperative appearance. One patient underwent retrograde bone grafting. Stable lesions with intact overlying cartilage were treated with transarticular drilling of the lesion in situ (59 cases). 2,9 Drilling was performed on the entire lesion using small 1.1 mm Kirschner wires, with drill holes placed several millimeters apart. The goal of transarticular drilling is to violate the underlying subchondral bone in an attempt to restore vascularity to the fragment and stimulate healing. Larger lesions (> 11 cm) that were unstable to probing and had salvageable overlying cartilage were treated with fragment reduction and transarticular drilling with internal fixation (22 cases). A bioabsorbable implant was routinely used in these cases the most commonly utilized included 1.5 mm Smart Nails (ConMed Linvatec Ltd, Tampere, Finland): 15 cases, mean 1.5 Smart Nails/ lesion; or 2.7 mm Bio-Compression Screws (Arthrex Inc., Naples, FL): 5 cases, mean 2.0 screws/lesion; or a combination of both implants (1 case). Excision of the OCD lesion followed by microfracture was reserved for unstable lesions with damaged and nonviable overlying cartilage. In these cases the lesion was first debrided with a curette to remove the damaged tissue so that the rim of the defect consisted of viable cartilage. A microfracture awl was then used to perforate the subchondral bone at 3 to 4 mm intervals, followed by a thorough lavage and inspection to ensure that no loose fragments remained. Postoperatively, most patients were placed in a boot for immobilization with limited weight bearing; cast immobilization was reserved for patients who underwent internal fixation. Full weight bearing was allowed at 6 weeks postoperatively. Formal physical therapy was initiated at 6 weeks postoperatively focusing on ankle range of motion and gentle strengthening. Return to impact sports was allowed at surgeon discretion starting at 3 months after surgery based upon clinical symptoms, physical examination, and imaging. RESULTS The characteristics of patients in the study population and OCD lesion data are summarized in Table 1. Median follow-up was 3.3 years (range, 1 to 10.8 y; interquartile range, 2.0 to 5.1 y). Lesion stage was not related to physis status (stage II was most common for both open and closed physes) or lesion location. Lesion stage was correlated with intraoperative instability: stage I (1/ 11 lesions unstable, 9%), stage II (13/50, 26%), and stage III (10/16, 63%). Table 2 depicts a breakdown of the procedure performed by lesion location, distal tibial physis status, and lesion stage. Although all 4 surgeons 4 r 2014 Lippincott Williams & Wilkins

5 J Pediatr Orthop Volume 00, Number 00, 2014 Surgical Management of OCD of the Ankle TABLE 1. Patient and OCD Lesion Data Patient Characteristics N (%) Total patients 100 Females 75 (75) Males 25 (25) Mean age (y) 14.3 ± 2.3 (range, 7-18) Right ankle 64/109 (59) Left ankle 45/109 (41) BMI (kg/m 2 ) 23.6 ± 4.5 (range, ) History of trauma 38 (35) Duration of symptoms (mo) < 6 31 (28) (21) > (39) Unknown 12 (11) Distal tibial physis Open 51 (47) Closed 58 (53) Lesion characteristics Location Medial talus 80 (73) Lateral talus 22 (20) Central talus 5 (5) Tibia 2 (2) Lesion size (mm) Coronal 8.7 (range, 3-16) Sagittal 12.3 (range, 5-26) Surface area 113 ± 62 (range, ) Berndt and Harty stage* I 14 (17) II 50 (60) III 16 (19) IV 3 (4) *Stage could not be determined on 26 ankles. BMI indicates body mass index; OCD, osteochondritis dissecans. performed all 3 procedures, procedure preference varied between surgeons (w 2 = 15.99, P = 0.001). Specifically, the excision rate was significantly higher for one surgeon as compared with the others. Associated procedures included 1 Brostrom lateral ligament repair, 1 Chrisman- Snook ankle reconstruction, and 1 os trigonum excision. There were no intraoperative complications. Clinical results after initial surgery were graded for all patients based on the Berndt and Harty clinical grade (in which reoperation for any reason is considered a poor result). For the cohort, the Berndt and Harty clinical grade was poor (33, 30%), fair (23, 21%), and good (53, 49%). Preoperative and postoperative radiographs were compared in 80 cases. On postoperative radiographs the OCD lesion was judged as healed (13/80, 16%), improved (51/80, 64%), unchanged (14/80, 18%), and worse (2/80, 3%). Berndt and Harty clinical grade analyzed by stage and procedure is depicted in Figure 3. Postoperative radiographic appearance analyzed by stage and procedure is depicted in Figure 4. Clinical grade did not correlate with radiographic appearance overall and especially in stage II lesions that had the lowest rate of good/excellent clinical grade (58%) and the highest rate of healed or improved postoperative radiographs (85%). There was no significant difference in radiographic healing/improvement between open (35/44, 80%) and closed (29/36, 81%) physes (P = 0.99) or in good/fair Berndt and Harty TABLE 2. Procedure Performed by Size, Location, Physis Status, Berndt and Harty Stage Transarticular Drilling Drilling With Internal Fixation Excision Microfracture All patients OCD size (mm 2 ) 110 ± ± ± 54 OCD location Lateral talus* Medial talus Central talus Distal tibia Physis status Open Closed* Berndt and Harty stage I II III* IV *One patient had retrograde bone grafting. OCD indicates osteochondritis dissecans. clinical grade between open (27/51, 53%) and closed (26/ 58, 45%) physes (P = 0.45). Univariate and multivariate Cox regression analysis was unable to identify any other predictors for Berndt and Harty clinical grade (including age, BMI, sex, history of trauma, duration of symptoms, physis status, lesion stage, location, surface area, and procedure performed including type of fixation). Twenty-nine ankles (27%) underwent a reoperation at a mean of 1.7 (range, 0.4 to 4.7 y) years postoperatively (Table 3 and Fig. 5). Reoperation rates were significantly higher for OCD lesions in which postoperative radiographs had no change or looked worse (10/16, 63%) compared with healed (2/15, 15%) or improved (15/51, 29%) (P = 0.002). There was a trend toward higher reoperation rates in the Bio-Compression screw group (3/5, 60%) as compared with the Smart Nail group (2/15, 13%) (P = 0.07, Fisher exact test). Multivariate Cox regression analysis was unable to identify other predictors for reoperation across all Berndt and Harty clinical grades. In the reoperation group the new Berndt and Harty clinical grade at latest follow-up after reoperation was poor (8, 28%), fair (13, 45%), and good (5, 17%) with 3 patients not graded due to inadequate time (< 6 mo) between reoperation and follow-up. For this reoperation group, the mean follow-up after the second operation was 1.9 (range, 0.2 to 6.2 y) years. A total of 44 patients filled out the return to sport survey and FAOS outcome measure (5 declined participation and the remainder could not be contacted). Survey responders (n = 44) were compared with survey nonresponders (n = 65) and no significant differences were noted with respect to age, sex, BMI, duration of symptoms, status of physis, location and size of lesion, and procedure performed. The survey responder group also included a similar percentage of patients who required reoperation (12/44, 27%) versus the survey nonresponder group (17/65, 26%). Of the survey responders, r 2014 Lippincott Williams & Wilkins 5

6 Kramer et al J Pediatr Orthop Volume 00, Number 00, 2014 A 100% A 100% Berndt & Harty Clinical Grade Good or Fair Results (%) 90% 80% 70% 60% 40% 30% 20% 79% 58% 63% 67% Postoperative Radiographs Healed or Improved (%) 90% 80% 70% 60% 40% 30% 20% 75% 85% 67% 10% 10% 0% Stage I (n = 14) Stage II (n = 50) Stage III (n = 16) Stage IV (n = 3) 0% Stage I (n = 12) Stage II (n = 48) Stage III (n = 6) Stage IV (n = 3) B Berndt & Harty Clinical Grade Good or Fair Results (%) 100% 90% 80% 70% 60% 40% 30% 20% 10% 0% 66% Drilling (n = 59) 73% Fixation (n = 22) 78% Excision (n = 27) FIGURE 3. Berndt and Harty clinical grade by stage (A) and procedure (B). B Postoperative Radiographs Healed or Improved (%) 100% 90% 80% 70% 60% 40% 30% 20% 10% 0% 79% Drilling (n = 42) 83% Fixation (n = 18) 88% Excision (n = 16) FIGURE 4. Postoperative radiographic appearance by stage (A) and procedure (B). 36 of 44 (82%) were satisfied and 37 of 44 (84%) returned to sports at a median time of 6 months (interquartile range, 4 to 12 mo; full range, 2 to 30 mo). Average FAOS score was 77 ± 18. Subgroup analysis of the FAOS revealed the following scores by category: symptoms, 73 ± 21 points; pain, 81 ± 21; function/sports, 76 ± 19; ADLs, 91 ± 15; QOL, 64 ± 25; and total, 385 ± 90. Multiple linear regression indicated that independent of the other covariates, sex and BMI were significant predictors for FAOS score (Fig. 6). For total FAOS score, the mean ± SD was 444 ± 44 for males and 368 ± 93 for females (P < 0.01). This suggests that females and those with higher BMI tend to have significantly worse FAOS outcomes. No significant differences in FAOS scores were detected according to procedure performed or postoperative radiographic appearance. DISCUSSION Although the etiology of OCD lesions remains unclear, most authors believe that repetitive microtrauma devascularizes a focal vulnerable area of underlying subchondral bone which then secondarily progresses through stages to involve the overlying cartilage. 2,3,14,16,17 OCD lesions should be distinguished from acute osteochondral fractures (OCFs), which follow a single traumatic injury and are more often seen in adults. In our series, most OCD lesions occurred in females (75%) and on the medial talus (73%), which is similar to other pediatric reports. 2,9,14 In contrast, the adult literature (which often includes mostly OCFs) describes a 2:1 male predominance for ankle lesions evenly split between the medial and lateral talar dome. 2,7,14,18,19 The chronicity of symptoms and the questionable vascularity of the underlying subchondral bone are unique aspects of OCD lesions in the pediatric population. Prior case series have combined adults and pediatric patients and are difficult to interpret as they likely combine OCD lesions with traumatic osteochondral injuries and OCFs. 3,4,9,20 Although other studies have discussed constitutional factors such as short stature or endocrine abnormalities as they relate to OCD of the knee, 21,22 these factors were not identified in our series. In this report, 75% of patients underwent an initial surgical attempt to salvage the overlying cartilage through transarticular drilling with or without internal fixation. Although the salvage group had a slightly higher rate of reoperation (and second reoperation), it should be emphasized that excision was performed in our series only 6 r 2014 Lippincott Williams & Wilkins

7 J Pediatr Orthop Volume 00, Number 00, 2014 Surgical Management of OCD of the Ankle TABLE 3. Reoperations Grouped by Index Procedure Age/Sex OCD Stage Location on Talus Surface Area (mm 2 ) Time to Reoperation (y) Reoperation Procedure Second Reoperation Index procedure: OCD drilling (18/59, 31%) 15.5 F 2 Medial Fixation Yes 7.0 M 2 Medial Fixation No 12.2 F 4 Medial Excision Yes 13.7 F 3 Lateral * Yes 14.5 F 2 Medial Fixation Yes 14.8 F N/A Tibia Bone grafting Yes 16.9 M 2 Medial Fixation No 12.8 F 2 Medial Bone grafting No 10.4 F N/A Medial N/A 1.9 Drilling No 16.3 M 3 Medial Chondroplasty No 13.7 F 1 Central Chondroplasty No 12.7 F 3 Medial Drilling No 18.2 F 2 Medial Excision Yes 13.5 F 2 Medial Excision No 10.3 F 2 Medial Excision No 15.4 F 2 Medial Excision No 15.9 F 2 Medial N/A 0.6 Excision No 15.0 M 2 Medial Synovectomyw No Index procedure: OCD fixation (5/22, 23%) 11.7 F 2 Medial Drilling No 12.7 F 2 Medial Excision No 15.6 F 2 Medial Excision No 14.6 F 3 Lateral Excision No 10.8 F 3 Lateral Excision No Index procedure: OCD excision (5/27, 19%) 15.7 F 2 Medial Drilling Yes 14.7 M 2 Medial Fixation Yes 13.6 F 2 Medial Excision No 14.7 F 2 Medial Spur excisionw No 12.2 M 2 Medial Synovectomyw No Index procedure: retrograde bone grafting (1/1) 14.9 F 3 Lateral Excision No *Patient had history of neurofibromatosis-1 and underwent hemiepiphysiodesis and distal tibial-fibular fusion for ankle valgus. windication for reoperation was unrelated to OCD and lesion appeared healed at reoperation. F indicates female; M, male; N/A, data not available; OCD, osteochondritis dissecans. in situations where the surgeon felt the lesion was damaged beyond repair. Despite these findings we continue to support lesion salvage whenever possible in these cases, Freedom from Reoperation (%) Kaplan-Meier Freedom from Reoperation (95% CI) 12 Months: 91% (86-96%) 24 Months: 81% (75-87%) 36 Months: 75% (67-83%) 48 Months: 68% (58-78%) 60 Months: 65% (55-75%) Time Since Surgery (months) FIGURE 5. Kaplan-Meier survivorship curve for reoperation. Error bars denote 95% confidence intervals around the curve as calculated by Greenwood s formula. given the potential importance of native cartilage preservation for long-term function in our young patient population. Because OCD of the ankle is a primary problem in the subchondral bone that secondarily involves the overlying cartilage, in many cases the overlying cartilage is intact and of good quality. We believe that antegrade transarticular drilling provides the most accurate method of subchondral bone penetration to address the questionable vascularity of the subchondral fragment. Retrograde drilling has also been described, 23,24 which has the theoretical advantage of not violating the overlying chondral surface but is technically difficult and can result in drill misplacement even with fluoroscopic 2-dimensional navigation. 23 We utilized bioabsorbable implants in all cases of internal fixation in our series, which obviates the need for later hardware removal. We did not note any direct complications from these implants. To date, few studies have evaluated the results of surgery for OCD of the ankle in the pediatric population and those published are limited by small sample size, limited follow-up, and heterogenous nonoperative and operative treatment protocols. 2,16 Letts et al 2 recently r 2014 Lippincott Williams & Wilkins 7

8 Kramer et al J Pediatr Orthop Volume 00, Number 00, 2014 Total FAOS Score * Females Males >30 BMI (kg/m 2 ) * Drilling Fixation Excision FIGURE 6. The 2 multivariate predictors of FAOS: sex and BMI. Sex (*P < 0.01). BMI (*P < 0.01) for >30 kg/m 2 compared with each of the other categories (no significant difference between BMI 16 to 25 vs. 25 to 30, P = 0.73). For procedure performed, there was no significant effect on total FAOS (P = 0.29). BMI indicates body mass index; FAOS, Foot and Ankle Outcome Score. reported their experience with talar OCD lesions in 24 patients, mean age 13 years (range, 6 to 17 y). Nonoperative management was successful in only 9 of 24 children (37.5%), including all 5 children under the age of 12 years. The other 15 patients underwent transarticular drilling (12), fragment excision (5), and internal fixation (1). Fair to good clinical results (based on the same Berndt and Harty clinical grade used in our study) were reported in 71% of patients treated surgically (including 81% of stage II lesions as compared with 38% of stage III lesions). 2 Higuera et al 14 reported results on an 11-year series of 18 consecutive cases of talar OCD in patients, mean age 12.5 years old (range, 8 to 16 y). Only 5 patients required surgery, all were skeletally mature, and 4 of 5 required malleolar osteotomy for exposure. Kumai et al 9 included 11 children (age below 16 y) in their series of 17 patients with OCD of the talus, all of whom were treated operatively. These pediatric patients all had medial lesions and were treated with percutaneous transmalleolar arthroscopic drilling with 10 patients (91%) having fair to good results (based on the same Berndt and Harty clinical grade). 9 The clinical results in our series were similar with 70% of patients having a fair to good result by strict Berndt and Harty clinical grading. Twenty-nine of the 33 patients who were graded as poor received this grade due to their reoperation. Following reoperation, an additional 18 of these patients would have been reclassified as having a fair or good clinical grade. Our series focused on the operative management of ankle OCD in a mostly adolescent patient population. Patients managed successfully without surgery were not identified. The success rate for nonoperative management of OCD lesions in the ankle is not known but may be low. A recent pediatric study noted a 16% rate of clinical and radiographic healing after 6 months of nonoperative management for an exclusively skeletally immature population with ankle OCD. 12 Other authors have reported success with nonoperative treatment in 37% to 45% of cases. 2,25 For OCD of the knee, the pediatric population is thought to have greater healing potential than adults Knee OCD studies have also shown poorer healing rates for larger OCD lesions. 30 For ankle OCD, no studies to date have compared clinical, radiographic, and function results in patients stratified by physis status or lesion size. Both nonoperative management 2 and transarticular drilling 6 of talar OCD lesions have been shown by others to be effective in skeletally immature patients with open physes without comparison with a closed physis group. Bruns and Rosenbach 7 compared surgical results between adults and children reporting better long-term outcomes and less osteoarthritic changes in children regardless of the severity of the lesion. However, this study was limited in that it only involved 38 patients separated into 2 groups by chronologic age and no mention was made of the physis status in the adolescent group. It is notable that in our study neither physis status nor lesion size had an effect on clinical, radiographic, or functional results in a statistically significant manner. Although this could be related to limitations in study design, there may also be less clinical, radiographic, and functional healing potential overall for OCD lesions in the ankle as compared with OCD lesions in the knee even for smaller lesions and for skeletally immature patients. In addition, although prior studies have suggested a different natural history between medial and lateral talar lesions, in our study lesion location did not influence the above results in a statistically significant manner. Complete radiographic healing of the lesion was rare in our series. This is in agreement with other studies on OCD of the ankle, which have shown low radiographic healing rates despite improved clinical results. 9,12,14 These findings are concerning as partially healed OCD lesions could potentially recur as chondral lesions in adults. Although clinical grade did not correlate with radiographic results, poor radiographic results were predictive of reoperation. This discrepancy may be related to limitations in the Berndt and Harty clinical grading system and to lack of long-term follow-up. Patients with poor radiographic results may develop worsening clinical symptoms over time. Longer follow-up is necessary to determine the implications of radiographic healing in this population. The advantages of this study are that it provides demographic and lesion data along with clinical and functional results for a large number of young patients with OCD of the ankle treated surgically with 1 of 3 different procedures. Study limitations include its retrospective nature, lack of long-term follow-up, lack of preoperative functional scores, and limited postoperative subjective functional data. The procedure preference for salvage versus excision varied between surgeons, which introduces potential bias. The sample size was too small to make conclusions regarding implant type for fixation and comparison of implants may be biased by surgeon preference or lesion appearance. The Berndt and Harty clinical grade, 8 r 2014 Lippincott Williams & Wilkins

9 J Pediatr Orthop Volume 00, Number 00, 2014 Surgical Management of OCD of the Ankle while utilized in multiple similar studies, offers a limited assessment of the surgical result. In addition, although its terminology is simple, the FAOS score has not been validated in the pediatric population. Although radiographs were graded for healing by 1 of 3 different fellowshiptrained attending orthopaedic surgeons using agreed-upon guidelines, this is also a study limitation as prior studies have documented a significant lack of interrater reliability for judging healing of OCD lesions on knee radiographs. 31 A high rate of reoperation is notable within our study group, with a nonsignificant positive trend for lesion salvage versus lesion excision. The overall high rate of reoperation may be related to the younger and highdemand nature of our study population. We continue to believe the benefit of retaining native cartilage in this active population outweighs the potentially higher risk of reoperation and that an attempt at lesion salvage is preferable if the overlying chondral surface is viable. Many of the reoperations occurred at 2 and 3 years postoperatively, which emphasizes the need for long-term follow-up in these patients. The optimum treatment algorithm for OCD of the ankle in the pediatric population is not known but will likely differ from adult recommendations for osteochondral injuries in the ankle. 16 As most lesions are initially treated nonoperatively, future studies are necessary to determine the optimal method and duration of nonoperative management, and investigate which patient and lesion characteristics are associated with persistent symptoms that eventually require surgical intervention. In conclusion, there is a paucity of studies on surgical management of OCD lesions of the ankle in children and adolescents. Our study shows that surgical management is more common in females and more frequent on the medial side of the talus. Surgical management with lesion salvage can produce fair to good clinical results with a high rate of reoperation. Complete radiographic healing is rare and the lack of radiographic improvement postoperatively was predictive of reoperation. Females and patients with higher BMI may have lower functional outcomes. Longer term studies are necessary to determine whether ankle function degenerates over time especially in patients with partially healed OCD lesions and lesions treated with native cartilage excision. REFERENCES 1. Aichroth P. Osteochondral fractures and their relationship to osteochondritis dissecans of the knee. An experimental study in animals. J Bone Joint Surg Br. 1971;53: Letts M, Davidson D, Ahmer A. Osteochondritis dissecans of the talus in children. J Pediatr Orthop. 2003;23: Berndt AL, Harty M. Transchondral fractures (osteochondritis dissecans) of the talus. J Bone Joint Surg Am. 1959;41-A: De Smet AA, Fisher DR, Burnstein MI, et al. Value of MR imaging in staging osteochondral lesions of the talus (osteochondritis dissecans): results in 14 patients. AJR Am J Roentgenol. 1990;154: Dipaola JD, Nelson DW, Colville MR. Characterizing osteochondral lesions by magnetic resonance imaging. Arthroscopy. 1991;7: Bradley J, Dandy DJ. Results of drilling osteochondritis dissecans before skeletal maturity. J Bone Joint Surg Br. 1989;71: Bruns J, Rosenbach B. Osteochondrosis dissecans of the talus. Comparison of results of surgical treatment in adolescents and adults. Arch Orthop Trauma Surg. 1992;112: Gepstein R, Conforty B, Weiss RE, et al. Closed percutaneous drilling for osteochondritis dissecans of the talus. A report of two cases. Clin Orthop Relat Res. 1986;213: Kumai T, Takakura Y, Higashiyama I, et al. Arthroscopic drilling for the treatment of osteochondral lesions of the talus. J Bone Joint Surg Am. 1999;81: Imhoff AB, Paul J, Ottinger B, et al. Osteochondral transplantation of the talus: long-term clinical and magnetic resonance imaging evaluation. Am J Sports Med. 2011;39: Valderrabano V, Leumann A, Rasch H, et al. Knee-to-ankle mosaicplasty for the treatment of osteochondral lesions of the ankle joint. Am J Sports Med. 2009;37(suppl 1):105S 111SS. 12. Perumal V, Wall E, Babekir N. Juvenile osteochondritis dissecans of the talus. J Pediatr Orthop. 2007;27: Canale ST, Belding RH. Osteochondral lesions of the talus. J Bone Joint Surg Am. 1980;62: Higuera J, Laguna R, Peral M, et al. Osteochondritis dissecans of the talus during childhood and adolescence. J Pediatr Orthop. 1998; 18: Roos EM, Brandsson S, Karlsson J. Validation of the foot and ankle outcome score for ankle ligament reconstruction. Foot Ankle Int. 2001;22: Zwingmann J, Sudkamp NP, Schmal H, et al. Surgical treatment of osteochondritis dissecans of the talus: a systematic review. Arch Orthop Trauma Surg. 2012;132: Ogilvie-Harris DJ, Sarrosa EA. Arthroscopic treatment of osteochondritis dissecans of the talus. Arthroscopy. 1999;15: Loomer R, Fisher C, Lloyd-Smith R, et al. Osteochondral lesions of the talus. Am J Sports Med. 1993;21: McCullough CJ, Venugopal V. Osteochondritis dissecans of the talus: the natural history. Clin Orthop Relat Res. 1979;144: Zinman C, Wolfson N, Reis ND. Osteochondritis dissecans of the dome of the talus. Computed tomography scanning in diagnosis and follow-up. J Bone Joint Surg Am. 1988;70: Mubarak SJ, Carroll NC. Juvenile osteochondritis dissecans of the knee: etiology. Clin Orthop Relat Res. 1981;157: Mubarak SJ, Carroll NC. Familial osteochondritis dissecans of the knee. Clin Orthop Relat Res. 1979;140: Geerling J, Zech S, Kendoff D, et al. Initial outcomes of 3-dimensional imaging-based computer-assisted retrograde drilling of talar osteochondral lesions. Am J Sports Med. 2009;37: Hyer CF, Berlet GC, Philbin TM, et al. Retrograde drilling of osteochondral lesions of the talus. Foot Ankle Spec. 2008;1: Verhagen RA, Struijs PA, Bossuyt PM, et al. Systematic review of treatment strategies for osteochondral defects of the talar dome. Foot Ankle Clin. 2003;8: viii-ix. 26. Cahill BR. Osteochondritis dissecans of the knee: treatment of juvenile and adult forms. J Am Acad Orthop Surg. 1995;3: Hefti F, Beguiristain J, Krauspe R, et al. Osteochondritis dissecans: a multicenter study of the European Pediatric Orthopedic Society. J Pediatr Orthop B. 1999;8: Kocher MS, Czarnecki JJ, Andersen JS, et al. Internal fixation of juvenile osteochondritis dissecans lesions of the knee. Am J Sports Med. 2007;35: Kocher MS, Micheli LJ, Yaniv M, et al. Functional and radiographic outcome of juvenile osteochondritis dissecans of the knee treated with transarticular arthroscopic drilling. Am J Sports Med. 2001;29: Wall EJ, Vourazeris J, Myer GD, et al. The healing potential of stable juvenile osteochondritis dissecans knee lesions. J Bone Joint Surg Am. 2008;90: Parikh SN, Allen M, Wall EJ, et al. The reliability to determine healing in osteochondritis dissecans from radiographic assessment. J Pediatr Orthop. 2012;32:e35 e39. r 2014 Lippincott Williams & Wilkins 9