The Role of Morphine in the Pharmacologic Management of Neonatal Opioid Withdrawal Syndrome: A Systematic Review & Meta-analysis
By Jennifer J. Lee, MD, Jerri Chen, MD, PhD, Lena S. Sun, MD
Columbia University Medical Center
(Summarized and submitted by Jennifer J. Lee, MD)
With the ongoing opioid epidemic, there has been a dramatic rise in the incidence of neonatal opioid withdrawal syndrome (NOWS) in the U.S. over the last decade1. Nonpharmacologic interventions are the first line of treatment for affected infants; unfortunately, 60-80% of infants are not responsive to these measures alone and require pharmacotherapy2. While current guidelines recommend opioid replacement, most commonly with morphine or methadone, there has been insufficient evidence to advocate for one specific opioid over another3-4. We sought to systematically review the literature on the role of morphine in NOWS pharmacologic management and meta-analyze whether morphine vs methadone, and other novel agents such as buprenorphine and clonidine, is associated with improved outcomes. Primary outcomes were length of treatment (LOT) and hospital stay (LOS). A secondary outcome was rate of adjunct drug use (RAD).
We searched for human primary literature in the English language including randomized clinical trials (RCTs) and cohort studies in multiple databases from their inception through February 20, 2018: MEDLINE, Embase, PsycINFO, The Cochrane Library. Search strategies utilized keywords, medical subject headings, and Boolean operators. We looked for grey references on clinicaltrials.gov and used a snowballing technique of reviewing references of relevant studies. Two independent reviewers assessed studies for eligibility and performed quality assessments with the Cochrane Risk of Bias Tool and data extraction on the Covidence platform. We used a software program (RevMan, version 5.3) to conduct the meta-analysis using a random-effects model. Weighted mean differences (WMD), 95% CI were used for continuous outcomes, and risk ratios (RR), 95% CI for dichotomous outcomes.
Of 669 publications, 10 studies met inclusion criteria (n=758 infants) for quantitative analysis: five morphine vs methadone (1 RCT, four cohort studies), three morphine vs buprenorphine (RCTs), two morphine vs clonidine (cohort studies).  Morphine and methadone did not differ in LOT (WMD -0.20 [-0.85 to 0.45] days, I2 80%), LOS (WMD -0.33 [-0.89 to 0.24] days, I2 77%), or RAD (RR 1.30 [0.74-2.29], I2 51%). Sensitivity analyses separating RCT from cohort studies demonstrated similar results, and excluding one cohort study with a disparate study protocol resolved heterogeneity with no significant difference in results.  Buprenorphine was preferable to morphine in reducing both LOT (WMD 0.73 [0.35 to 1.12] days, I2 0%) and LOS (WMD 0.59 [0.21 to 0.97] days, I2 0%), however there was no difference in RAD (RR 0.76 [0.24-2.38], I2 31%). There was one retrospective cohort study comparing buprenorphine vs methadone treatment in which buprenorphine was associated with shorter LOT and LOS.  Clonidine was also favorable to morphine in terms of LOT (WMD 1.04 [0.15 to 1.93] days, I2 76%) and LOS (WMD 0.94 [0.52 to 1.36] days, I2 10%); there were insufficient data to compare RAD.
NOWS infants treated with morphine or methadone appear to have similar lengths of treatment and hospital stay with no difference in rates of adjunct drug use. Novel agents such as buprenorphine and clonidine seem to be associated with shorter treatment and hospitalization durations compared to conventional opioids and warrant further investigation.
- MMWR Morb Mortal Wkly Rep 2016;65(31):799-802.
- Pediatrics. 2013;134:e547-61.
- Cochrane Data Syst Rev 2010;10:CD002059.
- Pediatrics 2012;129:e650.