Multiple Daily Injections of Insulin Versus Continuous Subcutaneous Insulin Infusion for Pregnant Women With Type 1 Diabetes

Abstract

Perinatal outcomes similar to those in nondiabetic parturients are the aim for pregnancies complicated by type 1 diabetes mellitus (T1DM). The inability to achieve stable daily maternal glycemic control is a main cause for poor pregnancy outcomes in diabetic women. Currently, the methods to enhance glycemic control in T1DM are continuous subcutaneous insulin infusion through a pump and multiple daily injections. This retrospective observational study was designed to assess and compare insulin injections and the insulin pump for their ability to improve maternal glycemic control. This study involved 128 patients with T1DM treated between 2004 and 2009. Data from 64 T1DM women treated with insulin pumps and 64 treated with insulin injections during pregnancy were obtained and retrospectively matched for age, beginning and duration of diabetes, body mass index before gestation, and HbA1c in the first trimester. A basalbolus protocol in patients utilizing insulin injections involved 4 or 5 daily injections of short- and long-lasting human insulin (or short-acting insulin analogues). Women received 20% to 30%, 15%, and 20% of total daily requirement of short-acting insulin before breakfast, lunch, and supper, respectively. Neutral protamine Hagedorn (NPH) insulin was administered once daily at bedtime (∼40% of the total daily dosage) or twice a day for a 5-injection protocol (morning and bedtime injections). Women on insulin pumps received short-acting insulin analogs (50% in the basal flow and remaining 50% of the daily requirements divided as boluses before meals). All parturients self-controlled their blood glucose levels 4 times daily and completed a diurnal glucose profile once a week. The target fasting glucose level was less than 5.0 mmol/L, and the 2-hour postprandial glucose target was less than 6.7 mmol/L. Episodes of hypoglycemia (glucose level G3.3mmol/L) and hyperglycemia (glucose level Q7.8mmol/L) and insulin requirements in each trimester were recorded. Neonatal data included gestational age at delivery, birth weight, 5-minute Apgar score, arterial umbilical pH value, numbers of preterm deliveries, and large- and small-for-gestational age infants. Infant hypoglycemia was diagnosed if the glucose level was less than 2.2 mmol/L, hypocalcemia if the calcium level was less than 1.8 mmol/L, and hypomagnesemia if the magnesium level was less than 0.6mmol/L. Hyperbilirubinemia and perinatal acidosis were also determined. The study groups were similar in demographic characteristics. The reduction in HbA1c level during pregnancy was significant only in the insulin pump group (P < 0.0004). In addition, there was a statistically significant decline in hypoglycemic episodes only in the insulin pump group. In both groups, perinatal outcomes and neonatal characteristics were similar, except for the incidences of neonatal hypocalcemia (injections, n = 25 [44.6%]; pump, n = 13 [22%]; P < 0.01). This study showed that the insulin pump can be safe as intensive insulin therapy with similar glycemic effects as subcutaneous injections, with a reduced rate of hypoglycemia and decreased insulin requirements.