Does Size Really Matter? Incidence of Periprosthetic Fractures After Femoral Osteotomy in Patients With Cerebral Palsy.

OBJECTIVE: Of children, 30% to 35% with cerebral palsy (CP) develop hip subluxation or dislocation and often require reconstructive hip surgery, including varus derotation osteotomy (VDRO). A recent literature review identified postoperative fractures as the most common complication (9.4%) of VDROs. This study aimed to assess risk factors for periprosthetic fracture after VDRO in children with CP.

METHODS: A total of 347 patients (644 hips, 526 bilateral hips) with CP and hip subluxation or dislocation (129 females; mean age at index VDRO: 8.6 y, SD 3.4, range: 1.5 to 17.7; 2 Gross Motor Function Classification System (GMFCS) I, 35 GMFCS II, 39 GMFCS III, 119 GMFCS IV, 133 GMFCS V, 21 unavailable) were included in this retrospective, single-group intervention (VDRO) study at a tertiary referral center. Imaging and clinical documentation for patients age 18 years or younger at index surgery, treated with VDRO were reviewed to determine demographic data, GMFCS level, surgeon, type of hardware implanted, use of anticonvulsants and steroids, type of postoperative immobilization, presence of periprosthetic fractures, fracture location and mechanism, and time from surgery to fracture. Potential determinants of periprosthetic fractures were assessed using mixed effects logistic regression.

RESULTS: Of 644 hips, 14 (2.2%, 95% CI: 1.3%, 3.6%) sustained a periprosthetic fracture, at a median of 2.1 years postoperatively (interquartile range: 4.6 y, range: 1.2 mo to 7.8 y). Patients with a fracture had a median age at index surgery of 7.3 years (interquartile range: 4.3, range: 2.8 to 17.8; 1 GMFCS II, 6 GMFCS IV, 7 GMFCS V). Periprosthetic fractures were not significantly related to age at index surgery (P= 0.18), sex (P= 0.30), body mass index percentile (P= 0.87), surgery side (P= 0.16), anticonvulsant use (P= 0.35), type of postoperative immobilization (P= 0.40), GMFCS level (P= 0.31), or blade plate size (P= 0.17). Only surgeon volume significantly related to periprosthetic fracture (odds ratio = 5.03, 95% CI: 1.53, 16.56, P= 0.008), with the highest-volume surgeon also using smaller blade plates (P< 0.01).

CONCLUSIONS: Periprosthetic fractures after VDRO surgery in children with CP are uncommon, and routine hardware removal appears unnecessary. The data suggest that the common dogma of putting in the largest blade plate possible to maximize fixation may increase the risk of periprosthetic fracture. Due to the overall low fracture rate, especially when contextualized relative to the risk of hardware removal, a reactive approach to hardware removal appears warranted.

LEVEL OF EVIDENCE: Level III-retrospective study (targeting varus derotational osteotomies in children with cerebral palsy).