Salvage Options in the Cerebral Palsy Hip: A Systematic Review.

BACKGROUND: No preferred procedure exists for the chronically painful, unreconstructable subluxated or dislocated hip in cerebral palsy. The purpose of this study was to compare pain relief and complication rates of salvage procedures in cerebral palsy for ambulatory and nonambulatory populations.
METHODS: We searched Medline, Embase, and Cochrane databases using the search terms "cerebral palsy" and "hip dislocation." Inclusion and exclusion criteria were established to maintain data quality for analysis. A systematic review yielded 28 studies. Relevant information for postoperative pain and complications were extracted from each study and described. Our initial search identified 721 articles. Two hundred twenty duplications were excluded. Five hundred one were screened by title and abstract. One hundred articles underwent further full text and reference evaluation, yielding 25 studies. An additional 3 studies were then identified from the list of 25, yielding a total of 28 studies, which met our inclusion criteria.
RESULTS: Among nonambulators, femoral head resection (FHR), valgus osteotomy (VO), and total hip arthroplasty (THA) were found to relieve pain better than arthrodesis [odds ratio (OR) 7.3, 95% confidence interval (CI), 2.2-24.8; OR 5.9, 95% CI, 1.6-22.8; OR 11.7, 95% CI, 1.1-297.5, respectively]. Arthrodesis had a significantly higher complication rate than FHR, VO, THA, and shoulder prosthetic interposition. No significant differences in complication rate were found between FHR and VO. Pain relief rates among nonambulators for FHR, VO, THA, shoulder prosthetic interposition, and arthrodesis were 90.4%, 88.4%, 93.8%, 90.9%, and 56.3%, respectively. Complication rates among nonambulators were 24.0%, 33.3%, 35.3%, 28.6%, and 106.3%, respectively. Comparison of pain relief and complication rates among ambulatory cerebral palsy patients in all procedures except THA was not possible because the populations could not be separated from nonambulators in numbers sufficient to perform statistical analysis. Data were available for 32 confirmed cases of THA in ambulators and was associated with a 93.3% pain relief rate and a 38.2% complication rate.
CONCLUSIONS: Among nonambulators, the available literature suggests that FHR, VO, and THA may be superior at relieving pain than arthrodesis. FHR had the lowest absolute percentage of complications; however, no significant differences in complication rate or pain relief were found in nonambulators undergoing FHR or VO. Most of the complications for VO were implant related, and potentially amenable to hardware removal versus complications in FHR, which were related to the procedure itself such as proximal migration and heterotopic bone formation. THA in nonambulators was associated with complications such as dislocation and revision. Arthrodesis in nonambulators was associated with >100% complication rate and inferior pain relief compared with other procedures. Ambulatory patients had excellent pain relief with THA; however, the complication rate is higher than can be expected with non-neurological populations. Insufficient data exist to support use of other salvage procedures in ambulators. These conclusions should be interpreted with caution as all studies involved level IV evidence. LEVEL OF EVIDENCE: IV (systematic review of level IV studies).