In an article being early released this week in Pediatrics, Dr. Thomas Reese and colleagues from Vanderbilt University Medical Center and Emory University describe the development of a risk prediction tool for neonatal opioid withdrawal syndrome (NOWS) (10.1542/peds.2024-068673).
Approximately 6 in 1,000 liveborn infants in the US have NOWS (2020 data), yet identifying those infants who will experience severe withdrawal and need pharmacologic therapy is surprisingly difficult.
The American Academy of Pediatrics (AAP) guidelines recommend tailoring the duration of in-hospital monitoring to the infant’s prenatal opioid exposure, with monitoring for 3 days, 4-7 days, and 5-7 days for immediate-release opioids, buprenorphine and sustained-release opioids, and methadone, respectively. However, many opioid-exposed infants will have mild symptoms only and can be safely discharged home, with multiple unnecessary hospital days resulting in a heavy emotional, social, and financial burden for families, providers, and hospital systems. Conversely, discharging home an infant who subsequently develops severe withdrawal and needs hospital readmission is clearly unwanted. Thus, a better clinical prediction tool would have great value.
The authors suspected that additional factors related to maternal medical and social history and infant birth history could be harnessed to create a prediction tool more powerful than type of maternal opioid use alone. They sought to identify factors routinely gathered within the electronic medical record and available at birth to create a pragmatic and clinically useful tool.
Their cohort included all 33,991 live infants >33 weeks gestation born at their medical center between November 2017 and January 2024; 1243 (4%) had chronic opioid exposure and 108 (0.3%) were treated for NOWS with oral morphine (institutional standard care beyond non-pharmacologic interventions).
Of the 37 potential factors evaluated, the authors found 7 that accurately predicted NOWS and were significantly associated with severe NOWS:
► (Maternal) opioid use disorder
► Cigarette smoking
► 5-minute Apgar score
► Prescriptions for SSRI anti-depressive, anti-nausea, long-acting opioid, and gabapentin medications
In comparison to the AAP screening criteria (chronic opioid exposure), and at a risk threshold of 5% (i.e., accepting a probability of a “miss” occurrence up to 5%), the prediction tool identified 3 additional infants per 10,000 who would develop severe NOWS without resulting in any extended hospital observation days due to false positives.
The article walks the reader through model development, model testing, and performance comparisons in easy-to-understand language. The prediction tool crystallizes what many neonatologists and other providers have long suspected—that poly-substance exposure (including cigarette smoking and medications beyond opioids) is clinically very important and increases infant risk of NOWS. Kudos to the authors for this well-done undertaking that has great potential to benefit both families and hospital systems.
