122 - A Simple Scoring System for Prediction of IVH in Very-Low-Birth-Weight Infants

Sunday, April 30, 2023

3:30 PM – 6:00 PM ET

Poster Number: 122
Publication Number: 122.338

PRAVEEN KUMAR, University of Illinois College of Medicine, PEORIA, IL, United States; Mounika Polavarapu, The University of Toledo, Toledo, OH, United States

Presenting Author(s)

Praveen Kumar, MBBS, DCH, MD

Professor of Clinical Pediatrics
University of Illinois College of Medicine
PEORIA, Illinois, United States

Background: Early identification of infants at highest risk of developing IVH is important for to evaluate interventions for prevention of IVH.

Objective: To develop a simple scoring system to predict the likelihood of either any IVH (Grades 1-4) or severe IVH (Grades 3 and 4) in VLBW infants at birth.

Design/Methods: In this retrospective cohort study, 923 VLBW infants (GA ≥23 weeks and BW ≤1500 grams) were divided into derivation (n=431) and validation (n=492) cohorts. The discrimination performance of the models was measured using area under the curve (AUC). Stepwise, multivariable logistic regression analyses were done to identify variables associated with “any” and “severe IVH”. A variable with OR of 0-1.9 was assigned a risk score of 0, OR of 2-3.5 a score of 1, OR of 3.6-6.5 a score of 2, OR of 6.6-12.5 a score of 3, and OR of >12.6 received a score of 4.

Results:

The incidences of any and severe IVH were 22.9% and 9.9%. GA, maternal race, birth location and gender were independent predictors of any IVH (AUC 0.75). Birth weight, Apgar score at 1 minute, birth location and gender were independent predictors of severe IVH (AUC 0.84). For any IVH, 8.7% of infants with score of 1 or less (low risk), 26.5% with score of 2-3 (medium risk) and 56.5% with score of 4 or 5 (high risk) developed any IVH. For severe IVH, a score of 2 or less (low risk) was associated with zero likelihood of severe IVH, score of 3-5 (medium risk) was associated with 7.5% likelihood of severe IVH and a score 6 or more (high risk) had 30% likelihood of severe IVH. Only 62 (14.4%) of all infants were deemed to be at high risk.
A comparison of the performance of any IVH risk scores in derivation and validation cohorts showed that a risk score of 4 or more for any IVH had sensitivities of 30.2% and 30.8%, specificities of 94.1% and 92.7%, positive predictive values of 59.2% and 57.1%, and negative predictive values of 82.6% and 80.9%, respectively. For severe IVH, A risk score of six or more had sensitivities of 50.0% and 60.0%, specificities of 89.1% and 83.3%, positive predictive values of 30.7% and 26.7%, and negative predictive values of 94.8% and 94.8% for derivation and validation cohorts, respectively.

Conclusion(s): This easy to calculate, validated scoring system with a high NPV can help clinicians in employing prevention strategies only for infants at highest risk of IVH and thus, eliminate the risks of these interventions in a large majority of VLBW infants.