Treatment of gestational diabetes with short‐ or long‐acting insulin and neonatal outcome: a pilot study

Major problems with gestational diabetes (GDM) have been the high frequency of fetal macrosomia, and perinatal complications, especially among mothers needing insulin in addition to diet therapy (1). Both long-acting insulin (LAI) and short-acting insulin (SAI) are used, the former to reduce the basal glucose level and the latter to control postprandial glucose peaks after meals. Despite many improvements, insulin therapy has not solved the above-mentioned problems. We compared the effects of SAI and LAI on perinatal outcome in insulin-requiring GDM. We hypothesized that SAI therapy and postprandial glucose monitoring would lead to better glycemic control than LAI and therefore reduce fetal macrosomia. After obtaining approval by the Ethics Committee of the Helsinki University Central Hospital, department of Obstetrics and Gynecology and given information to all study participants, we included 23 consecutive women with insulin-requiring GDM in a prospective randomized trial. A 2-hr oral glucose tolerance test (OGTT) with 75 g of glucose was performed at 24–28 weeks of gestation in women with risk factors for GDM (BMI > 25 kg/m2, age > 40 years, previous GDM, previous child with birth weight > 4500 g, glucosuria, macrosomia in current pregnancy). If the mother had two or more abnormal values in the OGTT (fasting 4.8 mmol/l, 10.0 mmol/l at 1 hr and 8.7 mmol/l at 2 hr, representing the 97.5th percentile values of 1004 Finnish pregnant women) a 24-hr glucose profile (glucose measured every 4 hr) was performed. Insulin therapy was started when blood glucose was 5.5 mmol/l repeatedly or when, with one abnormal preprandial value, one postprandial value was 7.8 mmol/l. Diet was planned individually with a dietician, and the daily energy intake was 1800 kcal, of which carbohydrates represented 50%. Patients were randomized to SAI (Actrapid, Novo Nordisk, Bagsvaerd, Denmark) and LAI (Protaphan, Novo Nordisk, Denmark) treatment groups. Eleven patients were treated with three daily injections of short-acting human insulin at initial doses of 4 + 6 + 4 IU before breakfast, lunch and dinner, respectively. Twelve patients were treated with one daily injection of long-acting human insulin at an initial dose of 14 IU in the morning. Insulin dose was adjusted individually after a second 24-hr blood glucose profile the day after beginning insulin treatment. Patients were followed in the maternity outpatient clinic every 2–4 weeks. Home monitoring of blood glucose included a fasting value before breakfast and preprandial (0.5 hr before) and postprandial (2 hr after) values at lunch and dinner at least twice a week. In the LAI group, the dose was increased if two of the daily preprandial blood glucose values exceeded 5.5 mmol/l. In the SAI group, the dose was increased when one of the postprandial glucose values exceeded 7.8 mmol/l. HbA1c was determined at the beginning of insulin therapy and at delivery. Results are expressed as mean ± SD. The statistical significance of differences between means was assessed by the Student's t-test. A P-value < 0.05 and a χ2-test in case of categorial variables was considered statistically significant. The groups were similar in terms of age, BMI and parity. There were no differences in the daily total insulin dose, duration of insulin treatment or criteria for initiation of insulin treatment between the groups. Insulin dose was raised in 6 of 11 pregnancies in the SAI group and in 10 of 12 pregnancies in the LAI group. There was no difference in the duration of pregnancy between the groups. The relative birth weight was significantly higher (mean ± SD) in the LAI group (1.33 ± 1.29 SD units) than in the SAI group (−0.28 ± 1.11 SD units, P = 0.007). The absolute birth weights were 3943 ± 492 g and 3079 ± 722 g, respectively (P = 0.005). There were no differences in the mode of delivery, neonatal complications, malformations, number of cases of Apgar score under 7 at 1 min of age, or maternal HbA1c between the groups (Table 1). Our results suggest that postprandial hyperglycemia is of importance in the development of fetal macrosomia in insulin-treated GDM. Our results are partly in concordance with those of De Veciana et al. (2), who showed that a GDM postprandial-monitoring group received more insulin than those in a preprandial-monitoring group. Hence, glycemic control improved and neonatal complications including fetal macrosomia decreased (2). In our small prospective study the overall diabetic control was similar, and the daily insulin doses did not differ between the groups. This suggests that the different action profile of the two insulins most probably played an important role. Our results suggest that GDM should be treated with short-acting insulin. Further controlled studies are now needed to confirm this finding.