Midazolam and amplitude-integrated EEG in asphyxiated full-term neonates

In the present, prospective study, the relation between the levels of midazolam, its two active metabolites--1-hydroxy-midazolam (OH-midazolam) and 1-hydroxy-midazolam-glucuronide (glu-midazolam)--and the aEEG were examined.

Fifteen full-term neonates with seizures due to hypoxic-ischaemic encephalopathy admitted to our NICU were included. Midazolam (loading dose 0.05 mg/kg in 10 min, maintenance dose 0.15 mg/kg/h) was used as an add-on anti-convulsant after phenobarbital and lidocaine because of continuing seizures. Amplitude-integrated EEG background pattern was scored at the start of midazolam and at the time of blood sampling as continuous normal voltage (CNV), discontinuous normal voltage (DNV), burst suppression (BS), continuous low voltage (CLV) or flat trace (FT). Serum levels of midazolam, OH-midazolam and glu-midazolam were measured at least 8 h after the start with HPLC.

In 11/15 patients, seizures were abolished with the addition of midazolam. In the remaining patients, seizure frequency was reduced in one and unchanged in three. Amplitude-integrated EEG background pattern at the start of midazolam was CNV in two, DNV in six, BS in five and CLV in two. Moderate, temporary suppression of the aEEG background pattern lasting less than 2 h was seen in four neonates. Amplitude-integrated EEG at midazolam sampling was CNV in two, DNV in seven, CLV in two and FT in four. Serum levels of midazolam ranged from 0.10 to 1.76 mg/l, OH-midazolam from 0.05 to 0.28 mg/l and glu-midazolam from 0.85 to 4.36 mg/l.

A brief and moderate suppression of the aEEG background pattern immediately after midazolam was seen in four neonates for less than 2 h. Suppression at a later time point, i.e. after more than 8 h of midazolam infusion, was demonstrated almost exclusively in neonates with a poor background pattern, and therefore these patterns appear to be determined mainly by the severity of hypoxic-ischaemic encephalopathy.