Developing prediction rules to safely decrease unnecessary head CT scans for pediatric patients with hemophilia presenting with head injuries

Abstract

Abstract BACKGROUND: Guidelines for managing head injury in pediatric patients with hemophilia (PwH) are limited. The National Bleeding Disorders Foundation (MASAC #257) recommends factor correction before CT scans but does not specify when scans are unnecessary. In pediatrics, the PECARN tool predicts low risk of clinically important traumatic brain injuries (ciTBI), minimizing unnecessary CT scans. Despite a decline in the incidence of intracranial hemorrhage (ICH) due to prophylactic treatments, pediatric PwH receive CT scans more frequently than the general population. Unnecessary head imaging carries the risk of ionizing radiation, which can have adverse consequences. While a few small studies have evaluated the use of the PECARN tool in this population, there is a lack of data on how to risk stratify these patients. AIM: This study aims to develop a set of prediction rules for identifying pediatric PwH at low risk of ciTBI or ICH after head injury, in cases where CT scans are not recommended. The study aims to collect data to identify more features contributing to risk stratification using the PECARN tool as determined in our pilot data. METHODS: This is preliminary data from a regional, multi-institutional study reporting on three of the eight partnered pediatric hemophilia treatment centers. Data from patients aged 0-25 diagnosed with hemophilia treated by Children's Hospital of Richmond, Johns Hopkins Medicine, and Children's National who presented to an emergency department with a head injury between 2017 and 2023. Exclusion criteria included presentation over 24 hours post-injury, penetrating trauma, known brain tumor, and pre-existing neurologic disorders. The following data were collected: age, hemophilia type and severity, inhibitor status, class of prophylaxis, mechanism of injury, clinical findings, vital signs, PECARN predictions, imaging results, and disposition. The PECARN tool prediction “Observation vs CT” was considered a positive outcome because the tool is designed to identify children at very low risk, for whom a head CT is not recommended. RESULTS: A total of 152 patient encounters were included, with an average age of 4.8 years and a median age of 3 years. Of these, 53 encounters involved patients under 2 years old, with an average age of 15 months and a median age of 16 months. Of the total, 103 of 152 (67.7%) encounters were on prophylaxis. Among those, 50 of 103 (48.5%) were on non-factor therapy (e.g., emicizumab or concizumab), 10 of 103 (9.7%) were on extended half-life (EHL) factor, and 43 of 103 (41.7%) were on standard half-life (SHL) factor. A head CT was not obtained in 8 of 152 (5.26%) encounters where the PECARN tool also predicted “CT not indicated.” The tool predicted “CT not indicated” for 102 of 152 (67.1%) encounters, “Observation vs CT” for 34 of 152 (22.3%), and “CT indicated” for 16 of 152 (10.5%). Four of 152 (2.63%) encounters were positive for ICH and met criteria for ciTBI. The tool predicted “CT indicated” for 3 of 4 (75%) ICHs and “Observation vs CT” for the remaining positive encounter. The sensitivity and negative predictive value were both 100%. 2 of 4 (50%) ICHs had a glasgow coma scale of <15 and 3 of 4 (75%) ICHs had focal neruological deficits. 2 of 4 (50%) ICHs had a history of inhibitors. 3 of 4 (75%) ICHs were not on prophylaxis at time of encounter while one ICH was on weekly EHL factor. No patients on SHL factor or non-factor prophylaxis experienced ICH DISCUSSION: Presenting symptoms and physical exam findings pertinent to PECARN prediction continue to be essential factors in risk stratification. However, due to the low incidence of ICH in pediatric PwH, we still plan to include retrospective data from other participating centers in the region to prevent falsely elevated sensitivities and negative predictive values. Preliminary data suggests class of prophylaxis and inhibitor status may play a role in risk stratification but due to the study size and low incidence of ICH and ciTBI, this study is not powered to evaluate the efficacy of prophylaxis class on risk of ICH or ciTBI after head injury. CONCLUSIONS: Risk stratification for ICH or ciTBI in pediatric PwH after head injury remains inadequately addressed. Our preliminary data suggest that the PECARN tool is a promising starting point to safely reduce unnecessary CT scans in this population. Additionally, prophylaxis class and inhibitor status may be important factors in future risk stratification models.