This study builds upon our prior report describing the use of buprenorphine in the treatment of opioid NAS. In the initial cohort, an observation of an initial rapid up-titration, frequent attainment of maximum dose, lower than anticipated plasma buprenorphine levels, and lack of opioid toxicity led to a revision of dose that was employed in the currently described study. A strategy of dose optimization of increasing the 1) initial dose, 2) rate of dose up-titration, and 3) maximum daily dose was used. A formulation increase in drug concentration was used to maintain manageable drug volume for administration. Sublingual buprenorphine continued to maintain safety at these higher doses. In addition, there was continued advantage over morphine in length of treatment and length of stay, which was now statistically significant. It is arguable that many signs and symptoms observed in the infant with CMV infection did not reflect severity of NAS but instead were manifestations of the multiorgan viral process. Data from this infant was included in the primary analysis. However, removal of this infant causes the mean length of treatment to drop from 23 to 21 days, and length of hospital stay to drop from 32 to 26.
Three patients in the buprenorphine group required the use of adjunctive phenobarbital compared to one patient in the morphine group. Phenobarbital is often used as in cases of treatment failure. Specifically, this is a rescue therapy when maximum opioid replacement therapy dose is reached without adequate resolution of symptoms. It has also been used as an initial adjunct in combination therapy with an opioid (11
) or as initial monotherapy. (21
) Morphine and buprenorphine employ different up-titration rates and number of up-titrations until maximum dose is reached (6 for buprenorphine and 9 for morphine). Thus, it could be argued that the need for adjunctive phenobarbital is not necessarily a surrogate of “treatment failure” in infants who, independent of treatment allocation, have a more severe withdrawal symptom complex. Phenobarbital arguably has a therapeutic index similar to that of opioid treatment in this patient population. It is also not clear where on the dose response curve the present maximum buprenorphine dose lies. It is plausible that as a partial agonist, buprenorphine may not be able to induce the dense signal generation at the mu opioid receptor obtained with morphine. Alternatively, it could be hypothesized that a higher maximum dose of buprenorphine could reduce the frequency with which adjunctive phenobarbital would need to be used. However, as there are no definitive adverse events associated with short-term exposure to phenobarbital, it is possible that a short course of phenobarbital may be a useful adjunct to reduce total duration of treatment in children with more severe withdrawal symptoms.
A mechanistic basis for the improvement of buprenorphine over morphine is not immediately clear. The longer duration of action and residence at the mu opioid receptor of buprenorphine relative to morphine may reduce the sudden shifts in receptor agonism and withdrawal symptoms. A more prolonged persistence of drug effect following cessation likely reduces symptoms. Two features of buprenorphine dosing compared to morphine also merit noting. Buprenorphine has a higher up-titration of dose compared to morphine (25 vs. 10%), which may allow a quicker attainment of symptom control. Though both regimens employ a 10% weaning schedule, buprenorphine is stopped when the dose is within 10% of the starting dose. In contrast, the initial morphine dose is 0.4 mg/kg/day, while cessation of dosing takes place only after the drug has been weaned to 0.15 mg/kg/day. Finally, the buprenorphine group dosing was based upon birth weight, while the morphine group was dosed according to daily weight. This difference serves to decrease a relative dose of buprenorphine as the infant grows.
This study was of an open label, randomized, design with primary goals of demonstrating safety and feasibility of sublingual buprenorphine. It is possible that there was occult bias in the scoring of infants by the nurse evaluators, despite the use of structured training sessions. A double blind, double dummy study would be required to fully evaluate any differential efficacy of buprenorphine over morphine. Another limitation is that the study excluded infants with benzodiazepine exposure and preterm infants. While these populations were excluded for safety reasons, they do limit the generalizability of results. Infants with in utero benzodiazepine exposure have a longer length of stay and treatment compared to those without. (14
) Such patients represent a group in whom research should be directed. Premature infants typically have less severe NAS, (25
) but buprenorphine does not have an established safety record in this population. Finally, while the length of treatment and length of stay noted is this study is consistent with observational data at Thomas Jefferson University Hospital between 2000-2006, (14
) it is longer than has been noted at other institutions. (9
) It is not clear if the magnitude of advantage over standard therapy would be maintained at institutions with a shorter length of stay. While features of buprenorphine may make its use in the outpatient treatment of NAS attractive, this has not been evaluated.
In conclusion, buprenorphine at an optimized dose schema has been demonstrated to be safe and efficacious in the treatment of NAS. While indications of a therapeutic advantage over morphine again have been demonstrated, these need to be verified in a double blind clinical trial. Additionally, the unique characteristics of buprenorphine make exploration of outpatient use for treatment of NAS a ripe area for clinical research.