Researchers Develop Risk Model to Predict Brain Injury, Stroke in Neonates
Researchers have developed a risk prediction model for the early detection of brain injury and stroke in term neonates, which could improve prevention and treatment efforts.
Source: Getty Images
By - A new study published this week in JAMA Network Open shows that a recently developed risk prediction model can identify term neonates at risk of perinatal arterial ischemic stroke (PAIS) using common clinical factors.
PAIS is a type of brain injury that disrupts cerebral blood flow between 20 weeks of gestation and postnatal day 28. Arterial ischemic stroke is a major cause of neurological disability in children, with reported annual incidence rates of 1.2 to 8 per 100,000 children, and one in 1,100 live term births. Some of these strokes will be symptomatic during the neonatal period, but others can go unrecognized for months or years. A subset of these never receive a diagnosis.
According to the study, prediction models designed to detect and prevent stroke in adults have had success in the past, but no such models have been developed for perinatal stroke. To bridge this gap in the research, the study authors sought to develop and validate a PAIS prediction model that relied on routinely documented clinical factors.
The researchers developed this model using multivariate logistic regression, and utilized data from the Alberta Perinatal Stroke Project, Canadian Cerebral Palsy Registry, International Pediatric Stroke Study, and Alberta Pregnancy Outcomes and Nutrition study. Included data were gathered between March 2003 and March 2020, and criterion for inclusion were term birth and no underlying medical conditions associated with stroke diagnosis. Records with more than 20 percent missing data were excluded.
Clinical variables associated with PAIS risk were sourced from perinatal literature, and variables were defined using the National Institutes of Health’s (NIH) National Institute of Neurological Disorders and Stroke (NINDS) Common Data Elements. Variables that were consistent with these definitions and comparable across datasets were then pulled from the neonates’ medical records.
Selected variables included maternal, pregnancy, obstetric, fetal, and neonatal factors. Of these, nine factors were chosen for the prediction model based on their strong association with PAIS in term neonates: maternal age, tobacco exposure, recreational drug exposure, preeclampsia, chorioamnionitis, intrapartum maternal fever, emergency cesarean delivery, low 5-minute Apgar score, and male sex.
A total of 1,924 term neonates, split into 321 positive cases and 1603 controls, were included for the development of the final model. The model demonstrated good discrimination between cases and controls and high predictive performance. To internally validate the model, researchers used a smaller subset of neonates from the Alberta Pregnancy Outcomes and Nutrition dataset. This analysis included 479 participants, 87 cases and 392 controls, and achieved similar high performance.
These findings, the researchers state, indicate that common clinical variables can be used to develop and validate models for prediction of PAIS risk in term neonates. Models that possess good predictive performance and strong validity have the potential to successfully identify neonates with a high risk of PAIS in a clinical setting, which could improve early diagnosis and health outcomes.
