Automated Quantification of Opioid Withdrawal Severity in Neonatal Rat Pups Using Ethovision® XT Software

Abstract

Chronic prenatal exposure to opioids often causes fetal opioid dependence, which leads to neonatal opioid withdrawal syndrome (NOWS) shortly after delivery when opioid exposure is abruptly discontinued. NOWS causes neurological, autonomic, and gastrointestinal dysfunction in the neonate, and the long‐term effects of experiencing opioid withdrawal as a neonate are poorly understood. NOWS has been studied using rat models of prenatal opioid exposure; these studies quantified neonatal withdrawal severity using time‐consuming, labor‐intensive manual scoring methods. The goal of the present study was to develop and validate an automated method of quantifying opioid withdrawal severity in neonatal rat pups. Withdrawal signs quantified by manual scoring are predominantly based on locomotor activity; therefore, we used the animal behavior software Ethovision® XT to quantify locomotor activity of rat pups undergoing precipitated opioid withdrawal testing on postnatal day 1 in videos we had previously recorded. We compared results obtained from Ethovision® XT with those obtained from manual scoring. We first determined that the software could reliably quantify movements in 79.0% (166/210) of the pups; the remainder were excluded due to poor lighting. Two endpoints reported by Ethovision® XT, distance moved (cm) and movement time (s), had strong positive linear relationships with manual scores (distance moved: p < 0.0001; R2 = 0.7716; movement time: p < 0.0001; R2 = 0.8236). We next determined the sensitivity and specificity of each method to discriminate prenatal exposure to 15 mg/kg/day morphine, a positive control known to cause severe fetal dependence and neonatal withdrawal, from vehicle exposure. Receiver operator characteristic curve analysis indicated that all three methods had excellent accuracy (AUC > 0.90). Last, we used two‐way ANOVAs to analyze main and interaction effects of prenatal treatment (with vehicle or the mu opioid receptor full agonists morphine and norbuprenorphine) and postnatal challenge (with saline or the opioid receptor antagonist naltrexone) on each endpoint (i.e., manual scores, distance moved, and movement time). Results were similar for each endpoint; interaction and main effects were significant (p < .05) and highly significant (p < .0001), respectively. The percent of variation explained by each model was 77.6%, 87.57%, and 88.31% for manual scoring, distance, and movement time, respectively. Altogether, these results provide evidence that Ethovision® XT software can reliably quantify opioid withdrawal severity in neonatal rat pups with non‐inferiority to manual scoring. Lighting conditions during testing should be optimized to prevent shadows that obscure pup movements. Nonetheless, the present study describes a faster, less labor‐intensive method of quantifying opioid withdrawal severity in neonatal rat pups and is expected to accelerate progress in preclinical studies of NOWS.Support or Funding InformationThe project described was supported by the Translational Research Institute (TRI), grants 1U54TR001629‐01A1 and KL2TR000063 (NCATS/NIH) and by T32DA022981‐11 (NIDA/NIH). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.