Grading the level of evidence of neonatal pharmacotherapy: midazolam and phenobarbital as examples

BackgroundMany drugs are used off-label or unlicensed in neonates. This does not mean they are used without evidence or knowledge. We aimed to apply and evaluate the Grading and Assessment of Pharmacokinetic-Pharmacodynamic Studies (GAPPS) scoring system for the level of evidence of two commonly used anti-epileptic drugs.MethodsMidazolam and phenobarbital as anti-epileptics were evaluated with a systematic literature search on neonatal pharmacokinetic (PK) and/or pharmacodynamic [PD, (amplitude-integrated) electroencephalography effect] studies. With the GAPPS system, two evaluators graded the current level of evidence. Inter-rater agreement was assessed for dosing evidence score (DES), quality of evidence (QoE), and strength of recommendation (REC).ResultsSeventy-two studies were included. DES scores 4 and 9 were most frequently used for PK, and scores 0 and 1 for PD. Inter-rater agreements on DES, QoE, and REC ranged from moderate to very good. A final REC was provided for all PK studies, but only for 25% (midazolam) and 33% (phenobarbital) of PD studies.ConclusionsThere is a reasonable level of evidence concerning midazolam and phenobarbital PK in neonates, although using a predefined target without integrated PK/PD evaluation. Further research is needed on midazolam use in term neonates with therapeutic hypothermia, and phenobarbital treatment in preterms.ImpactThere is a reasonable level of evidence concerning pharmacotherapy of midazolam and phenobarbital in neonates. Most evidence is however based on PK studies, using a predefined target level or concentration range without integrated, combined PK/PD evaluation.Using the GAPPS system, final strength of recommendation could be provided for all PK studies, but only for 25% (midazolam) to 33% (phenobarbital) of PD studies.Due to the limited PK observations of midazolam in term neonates with therapeutic hypothermia, and of phenobarbital in preterm neonates these subgroups can be identified for further research.