This study showed no significant difference between length of stay or length of treatment between infants treated with methadone or morphine for neonatal opioid withdrawal syndrome due to in-utero heroin, methadone, or illicit opioid exposure. There was a trend towards a shorter length of stay and length of treatment for the methadone group, which may have showed statistical difference with a higher sample size, as was found in other studies [21,22,23]. This trend may be explained by a higher rate of rooming in for infants in the methadone treatment group, which is known in the literature to decrease withdrawal symptoms and need for treatment [6,7,8]. Use of methadone, because it is long acting, requires less nursing interaction for medication administration and therefore may be more compatible with a rooming in floor model. It is also worth mentioning that the weight based nature of the methadone protocol may contribute to our findings rather than the drug itself. Infants in the morphine treatment group had a trend toward increased requirement for feeding interventions, though this was not statistically significant. This accounted for some of the wider variance seen in the length of stay for the morphine treatment group. While our sample size was too small to detect any significant differences between groups, it is possible the heightened withdrawal experienced by infants in the morphine group, perhaps secondary to lower morphine equivalents, resulted in poorer feeding. This issue is likely protocol specific and does not preclude a different morphine weaning model that would prevent heightened withdrawal. Many institutions use weight based morphine protocols whereas our morphine protocol was score based in the model of past trials [26].
Baseline characteristics of polysubstance use in pregnancy in the study were different but additional stratified analysis was not possible due to small sample size. There was a trend toward more polysubstance use in the morphine group, which likely impacted length of stay and treatment, as polysubstance use is known to complicate opioid withdrawal [1, 3, 29]. Of note, the rate of tobacco use was lower than previously described studies [29], likely due to regional and temporal variation, with less tobacco used in the West and in the past 5 years during pregnancy [30]. Future studies exploring the impact of polysubstance exposure in utero and the impact on neonatal opioid withdrawal are warranted.
There was a statistically significant difference in morphine equivalents received with infants treated with methadone receiving three times the morphine equivalents of opioid medication due to the weight based loading/taper methadone protocol. Infants in the methadone treatment group also experienced 3 adverse events of over sedation requiring transfer to higher level of care (NICU). These two factors are likely related and may represent an important area of modification of the existing loading/taper weight based methadone protocol currently in use at our institution. The ideal amount of opioid may be a balance between these two extremes, or increasing the initial morphine dose to 0.05 mg/kg per dose as is used in several ESC protocols may also potentially allow for improved treatment of NOWS [31, 32].
In our study, both of our treatment groups had a lower pharmacological treatment rate (46%) than many previously published studies. This rate was also lower than our own institutional historical average of approximately 65% from previous internal analyses. These changes are likely a result of an increase in nonpharmacological measures include rooming in, skin to skin care, and encouragement of breastfeeding among our opioid exposed infants. Given the lower pharmacological treatment rate, a larger multi-center trial would be required to achieve a sample size to detect small differences in length of stay between treatment protocols.
Conclusions from our study also are also limited by the change in inclusion criteria to include babies with a < 24 h NICU stay after several months. This decision was made to increase possible participation after it was noted that several infants were missed for enrollment due to brief post-birth transitional observation that was not thought to relate to or impact opioid withdrawal. The criteria for admission to a routine newborn floor may differ between institutions, especially in the late preterm period, and thus is an important consideration in future study design. In addition, we chose to end our study without reaching stopping criteria, primarily due to the pilot study nature of our protocol and desire from a clinical standpoint to pursue more modern assessment and treatment for neonatal opioid withdrawal.
After completion of our study, researchers developed a Core Outcome Set for Neonatal Opioid Withdrawal Syndrome to guide future research efforts [33]. Our investigation included several of these core outcomes including need for pharmacologic treatment, total dose of opioid treatment, duration of treatment, feeding difficulties, parent-infant bonding (rooming-in), length of stay, breastmilk at discharge, weight gain at discharge. Our study did not include measurements of consolability, time to adequate symptom control, readmission rates for withdrawal, or developmental outcomes. Future research should include those measures.