Adapting a Risk Prediction Tool for Neonatal Opioid Withdrawal Syndrome

Area of Science:

• Neonatal care
• Pharmacology
• Clinical prediction models

Background:

• Neonatal opioid withdrawal syndrome (NOWS) management guidelines recommend up to 7 days of observation for infants with chronic opioid exposure.
• However, not all exposed infants develop NOWS, and some with less apparent exposure may experience severe symptoms requiring pharmacotherapy.
• Current screening methods may not optimally identify infants at risk for severe NOWS.

Purpose of the Study:

• To adapt and validate a prediction model for identifying infants at birth who are likely to develop severe NOWS.
• To compare the performance of the prediction model against existing guideline screening criteria.

Main Methods:

• A prognostic study involving 33,991 births was conducted.
• Severe NOWS was defined by the administration of oral morphine.
• A logistic regression model with a least absolute shrinkage selection operator approach was developed using 37 predictors.
• Decision curve analysis was used to compare the model with chronic opioid exposure screening (maternal OUD diagnosis or long-acting opioid prescription).

Main Results:

• 108 infants received oral morphine for NOWS, and 1243 had chronic opioid exposure.
• The prediction model demonstrated high discriminative ability with an area under the receiver operating curve of 0.959.
• Maternal diagnosis of opioid use disorder (OUD) was the strongest predictor (adjusted odds ratio, 47.0).
• Decision curve analysis indicated superior clinical utility of the model across all risk levels compared to guideline criteria.

Conclusions:

• Risk prediction for severe NOWS at birth can enhance clinical decision-making.
• The model aids in tailoring non-pharmacologic interventions and determining the need for extended hospitalization.
• This approach offers a more precise method than solely screening for chronic opioid exposure.