Bumetanide for neonatal seizures: Based on evidence or enthusiasm?

Seizures in the neonatal period are serious events associated with high morbidity and mortality. A major clinical problem is that neonatal epileptic seizure activity shows only limited response to the commonly used anticonvulsant drugs. Consequently, the optimal treatment of neonatal seizures (NSs) is currently a highly prioritized issue (Silverstein et al., 2008). There is not only a lack of well-controlled clinical trials of available antiepileptic medications, but there are also high expectations among clinicians that newly developed drugs would turn out to have excellent antiepileptic properties. It has been, therefore, obvious that every light in this tunnel is greeted with great enthusiasm. One example of this is bumetanide, which is regarded as a relatively safe drug because it has been used as a diuretic in newborn infants. In immature cortical structures (hippocampus and neocortex), bumetanide blocks the depolarizing actions of γ-aminobutyric acid (GABA), and thereby suppresses the normal endogenous activity (Yamada et al., 2004; Dzhala et al., 2005; Sipilä et al., 2006) that is generally thought to be crucial for early brain maturation (Pallas, 2001; Owens & Kriegstein, 2002). Subsequent to some recent studies where bumetanide is also reported to suppress seizures in animal models of epilepsy, there is a recent initiative in the United States to perform clinical trials involving bumetanide in human infants (Investigational New Drug in FDA, #101690). In addition, it has been suggested that this medication be posted on the priority list by the European Medicines Agency (EMEA), as one of few antiepileptic medications for neonatal seizures. In the present letter, we want to call attention to some key aspects in recently published data, which imply a therapeutic action of bumetanide on neonatal seizures. We feel that the most crucial aspects of the current evidence have not been evaluated in detail, leading to a situation where the clinical audience might prefer to look at the optimistic interpretations rather than to the actual findings underlying them. Most importantly, the interpretation of findings from animal models should be treated with special caution, not least when such data show diverging results. The effects of bumetanide on epileptiform activity have been studied in one in vivo and six in vitro models of epilepsy, published in five reports from four distinguished research groups (Dzhala et al., 2005; Huberfeld et al., 2007; Kilb et al., 2007; Dzhala et al., 2008; Rheims et al., 2008). Determining the relevance of experimental models to NS is a major challenge in epileptology. All animal studies evaluating bumetanide effects were based on models with electrical or pharmacologic induction of ictal-like or interictal-like activity, whereas most human neonatal seizures are caused by hypoxic–ischemic insults, hypoglycemia, or infections. It is hence obvious, that the brain mechanisms in the models studied so far are different from human NS. The available in vivo evidence for the antiepileptic properties of bumetanide is based on data from six rats that were given bumetanide together with kainic acid induction of epileptic brain activity (Dzhala et al., 2005). Twelve seizures in three of these rats were analyzed in more detail, and the main finding was that electroencephalography (EEG) power was reduced during seizures in the bumetanide-treated animals. However, a slight change in the mere positions of invasive electrode can have dramatic effects on the EEG power, which was not considered in the study. Furthermore, EEG power reduction is clearly not equal to an anticonvulsant effect, and may not even have anything to do with seizure suppression. Most importantly, changes in the spectral power of ictal EEG have little relevance to clinical seizure treatment, which aims at blocking–not modifying—the electrographic seizure activity. Consequently, it is difficult to understand how this in vivo evidence would support the concept that bumetanide is an effective anticonvulsant. The in vitro data on bumetanide come from five different studies (Dzhala et al., 2005; Huberfeld et al., 2007; Kilb et al., 2007; Dzhala et al., 2008; Rheims et al., 2008). Two of these studies (Kilb et al., 2007; Dzhala et al., 2008) used a model wherein neonatal brain slices were incubated in a low-Mg (magnesium) solution, which is a widely used in vitro model to induce synchronized, epileptic-like activity in brain tissue in all age groups. Notably, this model is also known for its pharmacoresistancy to phenobarbital in adult brain specimens (Dreier et al., 1998). The reported low responsivity of neonatal brain slices to phenobarbital in this model does not match with the relatively better (but still poor; reviewed in Booth & Evans, 2004) anticonvulsant action of phenobarbital for treatment of NSs in the clinical situation (cf. Discussion in Dzhala et al., 2008). Induction of epileptic-like activity in the other in vitro models is also mechanistically distinct from human NS conditions, with no particular relevance to human NS mechanisms. The ambiguity in these in vitro models was demonstrated in a recent study by Luhmann and coworkers (Kilb et al., 2007), which compared bumetanide's effects on the epileptic activity in a number of different in vitro epilepsy models. They demonstrate that bumetanide's action is entirely dependent on the experimental model, that is, bumetanide may enhance, suppress, or have no effect on paroxysmal activity in vitro, and its actions may be different on ictal-like versus interictal-like activity. Because of the multitude of models and effects studied in experimental epileptology, there is a possibility that several physiologically distinct mechanisms and concepts become confused with each other. For instance, effects of drugs on interictal spiking in humans, on spike-like events in the in vivo animal models, or on ictal-like events in the in vitro models may erroneously be thought to have direct clinical relevance. The concept "antiepileptic or anticonvulsant action" means, of course, that the given drug is able to reduce epileptic seizures. However, hundreds of substances (including diuretics; see refs Hochman et al., 1995; Haglund & Hochman, 2005; Dzhala et al., 2005; Huberfeld et al., 2007; Kilb et al., 2007; Dzhala et al., 2008; Rheims et al., 2008) are known to suppress these interictal, spike-like, or ictal-like events, but only very few of them are established to have a clinically relevant anticonvulsant (or antiepileptic) action. Notably, the experimental designs and data in basic neuroscience papers have become so complex that they might be difficult to evaluate by those who make decisions in the clinic. There is hence a risk that clinicians with limited experimental experience may only consider interpretations and concluding statements in the original papers. Despite the shortage of clinically relevant evidence, it is obvious that the ultimate clinical utility of any treatment strategy can be evaluated on the basis of clinical trials and/or errors only. In the current request to conduct evidence-based medicine, it is of outmost importance that clinicians understand the nature of available "evidence" before deciding to proceed with individual treatments or to participate in clinical trials. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. Disclosure: The authors report no conflicts of interest.