Article Text
Statistics from Altmetric.com
Editor—The study by Stenninger and colleagues is another unacceptable contribution to the already confused scientific literature on hypoglycaemia in neonates.1
Leaving aside the obvious problems of selecting a matched socioeconomic control population for a group of babies known to have complex developmental problems, and who were exposed to insulin in utero, the authors do not provide us with adequate data on their patients.
They refer us to their original study,2 but neither there nor in the current study do they give us any idea of the severity or duration of the hypoglycaemia to which these babies were exposed, nor is it acceptable, nowadays, to discuss the likely effects of hypoglycaemia without providing data on plasma concentrations of other alternative brain fuels. It is perfectly possible, from the data supplied, that some of their hypoglycaemic babies had a blood glucose concentration no lower than 1.2 mmol/l for less than an hour. Surely not even the most fervent apologist for the deleterious effects of neonatal hypoglycaemia believes that this is likely to cause neurological sequelae, no matter how soft and uncontrolled the data?
The implication of the last paragraph of their paper is that transient asymptomatic postnatal hypoglycaemia of infants born to gestationally diabetic mothers should be treated with intravenous glucose. That must be wrong. The iatrogenic damage caused by this, rebound hypoglycaemia, and mother–child separation are likely to be considerably greater than the trivial, non-specific abnormalities they report. I believe, therefore, that this study is not only unsound, but potentially dangerous.
Dr Stenninger responds: We used the definition of neonatal hypoglycaemia as a blood glucose concentration of < 1.5 mmol/l according to criteria from the Swedish Paediatric Association (1978). This definition was in use when we started our 1988 study for neonatal hypoglycaemia in full term babies. We agree with Dr Koh that it would have been interesting to have set the level of hypoglycaemia at <2.6 mmol/l.
As the primary purpose of our investigation was not to determine a safe limit for neonates born to diabetic mothers, we cannot give a recommendation for such a limit in these circumstances. In Sweden we now use the lower limit of 2.2 mmol/l for the diagnosis of neonatal hypoglycaemia in term babies (Swedish Paediatric Association, 1997), and we have found no evidence to suggest that neonates born to diabetic mothers would tolerate lower blood glucose concentrations than healthy mothers.