Endoscopic Versus Open Repair for Craniosynostosis in Infants Using Propensity Score Matching to Compare Outcomes: A Multicenter Study from the Pediatric Craniofacial Collaborative Group

BACKGROUND:The North American Pediatric Craniofacial Collaborative Group (PCCG) established the Pediatric Craniofacial Surgery Perioperative Registry to evaluate outcomes in infants and children undergoing craniosynostosis repair. The goal of this multicenter study was to utilize this registry to assess differences in blood utilization, intensive care unit (ICU) utilization, duration of hospitalization, and perioperative complications between endoscopic-assisted (ESC) and open repair in infants with craniosynostosis. We hypothesized that advantages of ESC from single-center studies would be validated based on combined data from a large multicenter registry. METHODS:Thirty-one institutions contributed data from June 2012 to September 2015. We analyzed 1382 infants younger than 12 months undergoing open (anterior and/or posterior cranial vault reconstruction, modified-Pi procedure, or strip craniectomy) or endoscopic craniectomy. The primary outcomes included transfusion data, ICU utilization, hospital length of stay, and perioperative complications; secondary outcomes included anesthesia and surgical duration. Comparison of unmatched groups (ESC: N = 311, open repair: N = 1071) and propensity score 2:1 matched groups (ESC: N = 311, open repair: N = 622) were performed by conditional logistic regression analysis. RESULTS:Imbalances in baseline age and weight are inherent due to surgical selection criteria for ESC. Quality of propensity score matching in balancing age and weight between ESC and open groups was assessed by quintiles of the propensity scores. Analysis of matched groups confirmed significantly reduced utilization of blood (26% vs 81%, P < .001) and coagulation (3% vs 16%, P < .001) products in the ESC group compared to the open group. Median blood donor exposure (0 vs 1), anesthesia (168 vs 248 minutes) and surgical duration (70 vs 130 minutes), days in ICU (0 vs 2), and hospital length of stay (2 vs 4) were all significantly lower in the ESC group (all P < .001). Median volume of red blood cell administered was significantly lower in ESC (19.6 vs 26.9 mL/kg, P = .035), with a difference of approximately 7 mL/kg less for the ESC (95% confidence interval for the difference, 3-12 mL/kg), whereas the median volume of coagulation products was not significantly different between the 2 groups (21.2 vs 24.6 mL/kg, P = .73). Incidence of complications including hypotension requiring treatment with vasoactive agents (3% vs 4%), venous air embolism (1%), and hypothermia, defined as <35°C (22% vs 26%), was similar between the 2 groups, whereas postoperative intubation was significantly higher in the open group (2% vs 10%, P < .001). CONCLUSIONS:This multicenter study of ESC versus open craniosynostosis repair represents the largest comparison to date. It demonstrates striking advantages of ESC for young infants that may result in improved clinical outcomes, as well as increased safety.