Abstracts

Rationale: Treatment of neonatal seizures has hardly changed in decades due largely to a lack of rigorous, controlled trials.  Bumetanide blocks the chloride cotransporter NKCC1 that is highly expressed in neonatal neurons, thereby facilitating the inhibitory action of GABAergic drugs. We tested bumetanide (BTN) as add-on therapy for refractory neonatal seizures in the first trial to include a standard-therapy control group.   Methods: We employed a randomized, double-blind, dose-escalation design. Newborns with postmenstrual age 33-44 weeks at risk of or with clinical or EEG seizures were screened for enrollment. If refractory EEG-proven seizures occurred after 20-35mg/kg of phenobarbital (PB), subjects were randomized to receive their next dose of PB with either placebo (control) or BTN at a dose of 0.1, 0.2 or 0.3mg/kg (treatment). EEG monitoring continued for ≥48 hours after study drug administration (SDA) and was analyzed for all seizures. We obtained up to 6 BTN levels/subject for pharmacokinetic (PK) analysis. Results: We enrolled 111 newborns at 4 hospitals, and randomized 27 to treatment and 16 to control groups, and obtained 115 BTN levels. Notably, total seizure burden varied highly (Fig.1, median 81±139min, range 1-676), with >50% higher seizure burden in bumetanide (median 114min) vs. control groups (median 33 min, p<0.01),and a strong correlation between total and pre-SDA seizure burden (r=0.52, p<0.001). Seizure burden was reduced by a median of 41-75% in the period 0-4h post-SDA compared with 2h pre-SDA in all BTN dose groups and controls (Fig.2). With higher total seizure burden, there was a significantly greater reduction in seizure burden 2-4h (p=0.01) and 0-4h post-SDA (p=0.04) compared with 2h pre-SDA in BTN vs. control groups. This relationship remained significant for subjects with acute symptomatic etiologies. PK analysis showed similar clearance, distribution to published data, but high intersubject variability in clearance and volume of distribution (78%, 34%). Increasing BTN dose produced expected higher dose exposure (mean AUC0-4=1728, 3305, & 4812ug*h/L). There was no significant difference in measures of adverse events between groups, except diuresis was more common in treated subjects(p=0.02). Two treated subjects and one non-randomized subject had hearing impairment, and 1 BTN and 3 control subjects died. All serious adverse events were related to severe HIE with multiorgan dysfunction and/or withdrawal of care for poor prognosis.  Conclusions: Data from this Phase I/II randomized, controlled trial of BTN for neonatal seizures demonstrate an additional reduction in seizure burden attributable to BTN over PB. Notably, this is the first use of an ethical ‘placebo’ control group with which to compare drug response and safety.  Despite multicenter design, the relatively small number of randomized subjects was insufficient to prevent substantial differences in seizure burden between groups, requiring adjustment for seizure burden. However, seizure reduction was greater with higher seizure burden, different from the one randomized trial comparing PB to phenytoin. Adverse events were not significantly different in this small trial. Definitive trials of BTN will require substantially higher numbers of subjects to obtain accurate measures of efficacy and safety. Funding: National Institute of Neurological Disorders and Stroke (NINDS) - 5R01NS066929-05 Citizens United for Research in Epilepsy Harvard Catalyst- Harvard Clinical and Translational Science Center Translational Research Program, Boston Children's Hospital Charles H. Hood Foundation Mooney Family Initiative for Translational Studies in Rare Diseases, Boston Children's Hospital