In a prospective randomised study, 47 fasting women undergoing elective caesarean section received one of the following three glucose regimens for prevention of hypotension before administration of epidural bupivacaine to produce anaesthesia from the T4 sensory level: (i) 150 ml of 5% dextrose in water (D5W) and 1000 ml normal saline (7.5 g dextrose); (ii) 150 ml of D5W, 350 ml of 5% dextrose in normal saline (D5NS), and 650 ml of normal saline (25 g dextrose); (iii) 150 ml D5W and 1000 ml D5NS (57.5 g dextrose). The maternal blood sugar at delivery was significantly raised in patients receiving 25 and 57.5 g of dextrose. 57.5 g of dextrose also raised the mean umbilical cord venous levels of glucose to 11.7 mmol/l and of insulin to 70 +/- 7 microU/ml, while decreasing glucagon to 24 +/- 5.3 pg/ml and arterial pH to 7.19 +/- 0.015 compared with the levels with 7.5 g dextrose. The raised insulin level caused neonatal hypoglycaemia (1.8 +/- 0.22 mmol/l) and delayed release of glucagon at 2 hours of age. There was a higher incidence of neonatal jaundice in infants exposed to 57.5 g of dextrose than in those exposed to 7.5 g. The addition of large amounts of dextrose to intravenous fluids was harmful to the fetus and of no benefit to the mother. Until a safe rate of administration is established, it is recommended that dextrose infusions immediately before delivery be limited to no more than 6 g/h.