Range of Motion Improvement Following Surgical Management of Knee Arthrofibrosis in Children and Adolescents

Background: Arthrofibrosis of the knee is well-described in adults as a potentially debilitating postoperative complication following anterior cruciate ligament reconstruction, total knee arthroplasty, or fracture fixation. Knee arthrofibrosis in children and adolescents, however, has received little attention. The primary purpose of this study was to report improvements in range of motion (ROM) following lysis of adhesions and manipulation under anesthesia (LOA/MUA) in children and adolescents with knee arthrofibrosis, and, secondarily, to evaluate for any effect of preoperative dynamic splinting on ROM outcomes. Methods: Ninety patients aged 18 years and below (mean, 14.4 +/- 3.5) and 31% male who underwent LOA/MUA at an urban tertiary care hospital following prior knee surgery were evaluated. Demographic, clinical, ROM, and revision data were compiled. Primary outcome was absolute ROM. Secondarily, ROM was analyzed as a categorical variable with "Full ROM" defined to be -5 to 130 degrees or better, "functional" ROM was defined as unable to obtain -5 to 130 degrees but not requiring revision, and "failure" defined as resulting in revision arthrofibrosis surgery. t tests and chi(2) analyses were used to compare ROM and count variables between dynamic splinting subgroups. Results: Mean time from index surgery to LOA/MUA was 6.0 +/- 4.4 months, and follow-up was 42 +/- 56 months. Index procedures included anterior cruciate ligament reconstruction (N = 33), tibial spine arthroscopic reduction and internal fixation (N = 18), fracture fixation (N = 17), soft tissue repair (N = 17), and multiligament reconstruction (N = 5). In total, 68 subjects (76%) had any flexion loss, 57 subjects (63%) had any extension loss, and 40 subjects (44%) had both flexion and extension loss. Fifty-six subjects (62%) had full ROM at final follow-up, 25 subjects (28%) had functional ROM, and 9 subjects (10%) required revision. No demographic, clinical, or surgical variable was predictive of treatment failure. Patients who underwent dynamic splinting preoperatively (N = 46; 51%) had greater preoperative flexion (99 +/- 16 vs.77 +/- 34 degrees; P = 0.001), but no difference in flexion at final follow-up (121 +/- 20 vs.128 +/- 11 degrees; P = 0.08). Failure was not associated with time from index procedure to LOA/MUA, and the proportion who regained full ROM postoperatively was equivalent between those who had dynamic splinting and those who did not (65% vs. 59%; P = 0.70). Conclusions: LOA/MUA for children with arthrofibrotic knees results in significant improvements in ROM with 90% revisionfree success. Preoperative dynamic or static progressive splinting improves preoperative flexion but does not affect postoperative range of motion or failure rate.