[Neonatal neurology decision-making starting from systematic reviews of Cochrane Collaboration]

Cochrane Collaboration (CC) provides growing and readily accessible resource to help that decision-making care is based on detailed, critical, and current reviews of the best available evidence. There are many clinical questions for which there is no good evidence on which to base clinical practice.

To analyse the bibliometric characteristics of the systematic reviews (SR) about neonatal neurology published in Neonatal CC.

Bibliometric analysis of The Cochrane Database Systematic Reviews in Neonatal CC, Issue 1, 2004 (n = 169 SR). The dependent variable registered in each SR was subject area of study (mainly neurology subject area) and the rest of variables were considered independent: authors (number and country), dates (late review and update), characteristics of included clinical trials (number and type), characteristics of the newborns included (number and gestational age), reviewer's conclusions and potential conflicts of interest.

Neurology was the third most important subject area in Neonatal CC (13 SR), after the subject area of respiratory (73 SR) and gastroenterology-nutrition (26 RS). We identified three selective clusters of secondary investigation in neonatal neurology SR: intraventricular haemorrhage (4 SR), perinatal asphyxia (4 SR) and opiate withdrawal syndrome (2 SR). The number of clinical trials (median 4), patients (median 193), update (46%), sufficient conclusion (54%) and potential conflicts of interest (8 %) in each SR is similar to the rest of SR in Neonatal CC; the only difference we found is less preterm infants (15%) in this SR. All the SR are about interventions for the treatment or prevention of diseases, and we don't found any review about diagnostic tests.

At the moment, the neonatal neurology SR published in Neonatal CC are infrequent and almost half of them the reviewer's conclusions are insufficient for inferring probable effects in clinical practice. Many therapies in neonatal neurology persist without supportive evidence, and some common therapies may actually be harmful, and these are the conclusions found in SR about intraventricular haemorrhage. We detected no SR about important neuropediatric themes in neonatal period: hypoxic-ischemic encephalopathy, periventricular leukomalacia, neonatal seizures, hypotonia, etc.