Philip J Weston, Deborah L Harris, Malcolm Battin, Julie Brown, Joanne E Hegarty, Jane E Harding

Neonatal hypoglycaemia, a common condition, can be associated with brain injury. It is frequently managed by providing infants with an alternative source of glucose, given enterally with formula or intravenously with dextrose solution. This often requires that mother and baby are cared for in separate environments and may inhibit breast feeding. Dextrose gel is simple and inexpensive and can be administered directly to the buccal mucosa for rapid correction of hypoglycaemia, in association with continued breast feeding and maternal care. To assess the effectiveness of dextrose gel in correcting hypoglycaemia and in reducing long-term neurodevelopmental impairment. We searched MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials (CENTRAL), the Cumulative Index to Nursing and Allied Health Literature (CINAHL) and Web of Science from inception of the database to February 2016. We also searched international clinical trials networks and handsearched proceedings of specific scientific meetings. Randomised and quasi-randomised studies comparing dextrose gel versus placebo, no treatment or other therapies for treatment of neonatal hypoglycaemia. Two review authors independently assessed trial quality and extracted data and did not assess publications for which they themselves were study authors. We included two trials involving 312 infants. No data were available for correction of hypoglycaemia for each hypoglycaemic event. We found no evidence of a difference between dextrose gel and placebo gel for major neurosensory disability at two-year follow-up (risk ratio (RR) 6.27, 95% confidence interval (CI) 0.77 to 51.03; one trial, n = 184; quality of evidence very low). Dextrose gel compared with placebo gel or no gel did not alter the need for intravenous treatment for hypoglycaemia (typical RR 0.78, 95% CI 0.46 to 1.32; two trials, 312 infants; quality of evidence very low). Infants treated with dextrose gel were less likely to be separated from their mothers for treatment of hypoglycaemia (RR 0.54, 95% CI 0.31 to 0.93; one trial, 237 infants; quality of evidence moderate) and were more likely to be exclusively breast fed after discharge (RR 1.10, 95% CI 1.01 to 1.18; one trial, 237 infants; quality of evidence moderate). Estimated rise in blood glucose concentration following dextrose gel was 0.4 mmol/L (95% CI -0.14 to 0.94; one trial, 75 infants). Investigators in one trial reported no adverse outcomes (n = 237 infants). Treatment of infants with neonatal hypoglycaemia with 40% dextrose gel reduces the incidence of mother-infant separation for treatment and increases the likelihood of full breast feeding after discharge compared with placebo gel. No evidence suggests occurrence of adverse effects during the neonatal period or at two years' corrected age. Oral dextrose gel should be considered first-line treatment for infants with neonatal hypoglycaemia.