There is no current agreement on the definition of neonatal hypoglycemia. Three different bases for the definition are presented, with a review of evidence of relevant to each approach. (1) Not commonly encountered; requires evidence of normal distribution of blood glucose concentrations. (2) Increased risk of adverse sequelae; requires evidence of short-term and long-term sequelae of different blood glucose concentrations. (3) Benefits of intervention outweigh risks; requires evidence from randomized trials of intervention. The usual distribution of neonatal blood glucose concentrations varies with birthweight, postnatal age, nutritional intake and other factors. Clinical signs occur in some but not all babies with low blood glucose concentrations, but such signs lack specificity. Abnormality in sensory evoked potentials is associated with variations in blood glucose concentrations; these changes suggest that values less than 2.6 mmol/L have acute effects in areas of the brain with high glucose demand. Among low birthweight babies, case series suggest a high risk of impairment (50%) following symptomatic hypoglycemia, and controlled studies generally confirm an association between very low neonatal blood glucose concentrations and adverse neurodevelopment. However, these studies do not establish a level of blood glucose concentration or duration of hypoglycemia that is critical. There are no reports of randomized trials that have assessed the effect on neurodevelopment of alternative policies of glucose provision in the neonatal period. There is a need for a randomized controlled trial to assess reliably the short- and long-term clinical effects of alternative policies of the clinical management of blood glucose concentration in babies at high risk both of low neonatal blood glucose concentrations and adverse neurodevelopment.