O3189: Osteogenesis Imperfecta Testing Among Young Children Presenting with Long-bone Fractures

Background: Clinicians evaluating young children with long-bone fracture(s) may sometimes question the need for osteogenesis imperfecta (OI) tests as mild forms may not be clinically obvious. While AAP guidance suggests clinicians consider OI tests as part of suspected physical abuse assessment there are currently no specific indications for testing. Evaluation of current OI testing practices could identify practice variability and inequities. We aimed to describe frequency and results of OI tests among children < 4 years with long bone fracture(s) and compare demographic and clinical characteristics between children with and without OI tests.

Methods: Sub-analysis of a multi-center, prospective, validation study of children with abusive and accidental fractures. We included children < 4 years presenting to two pediatric emergency departments with ≥1 long bone fracture. We excluded motor vehicle accident injuries and children already diagnosed with OI. Primary outcome defined as laboratory test for OI (COL1A1/COL1A2, IFITM5). We obtained demographic and clinical characteristics from direct interviews, physical exams, and medical records. Hospital evaluation for abuse was defined as assessment by child abuse pediatrician and/or social work and/or skeletal survey. We calculated frequency of OI tests and results. Chi-squared, Kruskal-Wallis, Fisher exact and t-tests were used to compare characteristics of children with and without OI testing.

Results: 418 children were included. Frequency of OI testing was 32/418 (7.7%). OI test results were negative for 26/32 (81%), variant of unknown significance for 5/32 (16%) and positive for 1 child with OI family history and clinical features.

Frequency of OI tests differed significantly across enrollment sites (Table 1). When children with and without testing were compared overall there were significant differences in race (Table 1). However, race differed significantly across enrollment sites (p < 0.001) and when sites were analyzed individually there were no significant differences in race between children with and without OI tests (site 1 p= 0.449, site 2 p= 0.152). Insurance type did not differ significantly between children with and without OI tests. Almost all OI tests (97%) occurred among children evaluated for abuse. Of note, 13% of children evaluated for abuse were tested for OI.

Table 2 outlines injuries sustained by children evaluated for abuse overall and by OI testing status. OI tested children had significantly more fractures than those who were not tested. Rib fractures and classic metaphyseal lesions were also more common in the former. Frequencies of extra-skeletal injuries were similar in children with and without OI testing.

Conclusion: Frequency of OI testing varied by practice site. Many tested children had extra-skeletal injuries not associated with OI such as skin injuries with high specificity for abuse (TEN-4-FACESp). Results suggest research to develop standardized guidelines for OI tests among young children with fractures could decrease practice variability and refine testing criteria.