Seizures are a common problem in the neonatal intensive care unit (NICU). Surprisingly, there is scant evidence regarding the optimal anticonvulsant for treating neonatal seizures [1]. In the USA, of the neonatal seizures that are treated with a non-benzodiazepine drug, phenobarbitone accounts for 76% and phenytoin for 16% [2]. Until now, the only good quality randomised controlled trial (RCT) comparing phenobarbitone and phenytoin was by Painter, et al. [3] in 1999. Pathak, et al. [4] have conducted a much-needed randomized controlled trial comparing phenobarbitone and phenytoin. This RCT is commendable because it was conducted within the resource constraints of a level II neonatal unit in India. They were unable to measure serum phenobarbitone and phenytoin levels, presumably owing to financial constraints. The authors also acknowledge that their unit–like most units in developing countries has no access to EEG monitoring.

However, the fact remains that the non-availability of EEG raises several methodological issues. The correct clinical identification rate of clonic seizures is only 66%; of subtle seizures is 32% and of non-seizure movements is 47% [5]. The inter-observer agreement (kappa) among pediatricians for the identification of neonatal seizures is just 0.21. Thus, in clinical practice, there is widespread overdiagnosis and underdiagnosis of neonatal seizures and disagreement between observers [6]. This poses a problem in a research study, such as the one conducted by Pathak, et al. [4], where the inclusion of patients is solely on clinical grounds. The relationship between EEG and clinical seizures is particularly weak in the case of subtle seizures and tonic seizures; and in the current study, 87% of all seizures were either subtle or tonic [4,7]. In this open-label trial with fixed, small block sizes (27 blocks of 4 each), non-concealment of allocation towards the end of each block was inevitable. Thus, at least 27 (25%) patients would have their allocated intervention known beforehand. A selection bias in allocating 25% or more patients could easily swing the results of the study. Being open-label, there was an unavoidable risk of performance and ascertainment bias – further compounded by the fact that cessation of neonatal seizures (like the diagnosis of seizures) has a high rate of misclassification and inter-observer disagreement. Pathak, et al. [4] did not mention the time frame for measurement of the primary outcome nor the waiting time before giving the second anticonvulsant for persistent seizures- both of which could affect the measurement of the primary outcome.

Painter, et al. [3] had enrolled neonates on the basis of a risk for seizures and presence of electrographic evidence of seizures. They reported no evidence of the superiority of phenobarbitone or phenytoin in terms of electrographic seizure cessation. The small sample size precluded any change of practice based on Painter’s study alone. Pathak, et al. [4] claim a much higher success rate for phenobarbitone for clinical seizures and recommend its use. Previous EEG-based studies–both randomized and non-randomized – on the efficacy of phenobarbitone have reported efficacy rates ranging from 43% to 50% only [3,8,9]; hence the efficacy rate of 72% in the study by Pathak, et al. [4] is exceptional. They state that their study resembles real-life situations in developing countries and therefore has a high external validity. However, one must be cautious before accepting this conclusion because without robust internal validity, external validity is of limited relevance.

There is an urgent need for large, well-designed RCTs with a low-risk of bias and adequate follow-up, comparing phenobarbitone with phenytoin and newer anti-convulsants. The RCT by Pathak, et al. [4] has certainly brought back our focus on to this important area. It is high time that researchers in this field were seized of the matter.

Funding: None; Competing interests: None declared.

1. Booth D, Evans DJ. Anticonvulsants for neonates with seizures. Cochrane Database Syst Rev. 2004;CD004218.

2. Blume HK, Garrison MM, Christakis DA. Neonatal seizures: treatment and treatment variability in 31 United States pediatric hospitals. J Child Neurol. 2009;24:148-54.

3. Painter MJ, Scher MS, Stein AD, Armatti S, Wang Z, Gardiner JC, et al. Phenobarbital compared with phenytoin for the treatment of neonatal seizures. N Engl J Med. 1999;341:485-9.

4. Pathak G, Upadhyay A, Pathak U, Chawla D, Goel SP. Phenobarbitone and phenytoin for treatment of neonatal seizures: Open-label randomized controlled trial. Indian Pediatr. 2013;50:753-7.

5. Malone A, Ryan CA, Fitzgerald A, Burgoyne L, Connolly S, Boylan GB. Interobserver agreement in neonatal seizure identification. Epilepsia. 2009;50:2097-101.

6. Scher MS, Aso K, Beggarly ME, Hamid MY, Steppe DA, Painter MJ. Electrographic seizures in preterm and full-term neonates: clinical correlates, associated brain lesions, and risk for neurologic sequelae. Pediatrics. 1993;91:128-34.

7. Mizrahi EM, Kellaway P. Characterization and classification of neonatal seizures. Neurology. 1987;37:1837-44.

8. Boylan GB, Rennie JM, Chorley G, Pressler RM, Fox GF, Farrer K, et al. Second-line anticonvulsant treatment of neonatal seizures: a video-EEG monitoring study. Neurology. 2004;62:486-8.