Objective
We hypothesized that bolus and basal insulin doses in women with type 1 diabetes mellitus who use insulin pumps would increase 2-fold to maintain hemoglobin A1c <6.5% across gestation.
Study Design
This was a retrospective study of 9 women with type 1 diabetes mellitus with preconceptional hemoglobin A1c ≤7.4% using insulin pumps. The primary outcome was absolute and percentage change of basal and bolus insulin from preconception to delivery.
Results
Total daily dose of insulin increased from 33.3 ± 7.8 U/d before conception to 93.5 ± 27.9 U/d at delivery. Basal rates rose modestly (50% increase, from 16.2 ± 6.5 U/d to 24.0 ± 9 U/d); bolus insulin doses quadrupled from 17.1 ± 6.1 U/d to 69.5 ± 29.6 U/d ( P = .0001). Bolus insulin increased from approximately 50% of total daily dose of insulin before conception to 75% of total daily dose of insulin at 36 weeks’ gestation.
Conclusion
In well-controlled type 1 diabetes mellitus, insulin requirements increased 3-fold from before conception to 36 weeks’ gestation. Most of this requirement was attributed to an increase in bolus rates that are required for control with meals.
Type 1 diabetes mellitus (T1DM) is characterized by the loss of pancreatic beta cells that leads to subsequent absolute insulin deficiency. This disease affects nearly 1 in 300 individuals by age 18 years, and approximately 19.7 in 100,000 children and adolescents are diagnosed with T1DM per year. In pregnancy, T1DM has been associated with an increased risk of congenital birth defects, miscarriage, fetal death, and preeclampsia; preconceptional glycemic control and rigorous medication adjustments during gestation are associated with reduced complications.
The achievement of tight glycemic control depends on an understanding of physiologic changes of pregnancy that lead to insulin resistance. This results from a combination of placental hormones that include prolactin, progesterone, human placental lactogen, placental growth hormone, cortisol, leptin, and adiponectin. Typically, insulin requirements follow a characteristic pattern in pregnancy, with a decrease in the first trimester and a rise in the second and third trimesters. Increases in insulin requirements of 36-114% from preconceptional baseline to the second and third trimesters have been reported. However, those past studies included patients with suboptimal glycemic control before conception and have had various gestational ages (GAs) at study entry.
More patients with T1DM are being treated with insulin pumps to achieve adequate glycemic control. Recent studies have demonstrated that patients can be treated effectively with insulin pumps when compared with multiple subcutaneous insulin injections. However, expected changes in basal insulin and bolus insulin requirements in pregnant women who use insulin pumps remain poorly defined. Previous studies have focused on patients with multiple daily insulin injections. One study included patients who used insulin pumps, but the changes that are unique to this group were not evaluated.
We hypothesized that bolus and basal insulin doses in women with T1DM who are treated with insulin pumps and who are in good control (hemoglobin A1c [HbA1c], <7.4%) would increase by ≥2-fold above preconceptional baseline to maintain an HbA1c level of <6.5% across gestation.
Materials and Methods
This retrospective study was conducted at the University of California San Diego Diabetes in Pregnancy Program among patients who delivered from 2010-2011. The study was approved by the University of California San Diego Human Research Protection Program before initiation. Inclusion criteria were singleton pregnancies with preexisting T1DM managed with a continuous insulin pump, preconceptional HbA1c level of ≤7.4%, preconceptional insulin dosing and self-monitored plasma glucose values available for review, and delivery at term. Exclusion criteria included patients who were not using an insulin pump preconceptionally or who were switched to subcutaneous injections after the initiation of care. The inclusion criterion of HbA1c level of ≤7.4% was chosen because it is associated with similar rates of congenital malformations and spontaneous abortion as in nondiabetic women whose data have been reported in the Diabetes Control and Complications Trial.
All patients were provided nutritional counseling that included recommended meal carbohydrate/fat/protein content and meal frequency (3 meals and 3-4 snacks daily). Patients were advised to monitor plasma glucose levels at the following times: in the morning after fasting; preprandially; 1 hour postprandially for breakfast, lunch, and dinner; at bedtime, and between 3 and 4 am . Goals for glycemic control were fasting values of 90 mg/dL, preprandial values of 80-120 mg/dL, and 1-hour postprandial values <130 mg/dL. Target values were chosen optimally to balance the risks of fetal macrosomia with maternal hypoglycemia. Patients recorded their capillary glucose measurements in glycemic control logs, and glucose meters were reviewed by clinical diabetic educators to verify accuracy. Glucose logs with dietary recall and actual insulin dosing were evaluated weekly in person or by fax/email by clinical diabetes mellitus educators and by the study investigators. Adjustments to insulin basal and bolus dosing were made by the 2 senior authors (T.R.M. and G.A.R.). Insulin requirements were calculated based on units per kilogram weight. Basal and bolus insulin dosing (absolute and percent change) was compared between preconception and weeks 9, 16, 20, 24, 28, 32, 36, and 38 of pregnancy. During the day, basal rates were generally divided into 5 segments: 0000-0500, 0500-0900, 0900-1600, 1600-2200, and 2200-2400. These segments correspond with typical times of day when insulin requirements increase or decrease; however, each patient had an individualized regimen that was based on her time of awakening, work/sleep schedule, and meals. Bolus insulin doses were subdivided by the breakfast, lunch, and dinner meals. HbA1c data were obtained every trimester.
Maternal medical records were reviewed to obtain demographic data, evidence of end-organ disease, glycemic control from preconception through delivery, HbA1c level, and delivery data. Body mass index was calculated with the documented weights during pregnancy. Height was based on maternal recall. Neonatal medical records were reviewed to obtain gestational age at delivery and birthweight.
The primary outcome measure was the absolute and percentage change of basal and bolus insulin requirements from preconception to delivery. Secondary outcomes included gestational age at peak basal and bolus insulin doses and changes in basal rates and bolus insulin dosing during specific time segments of the day.
Statistical analysis was performed using SPSS software (version 20; SPSS Inc, Chicago, IL). The Student t test was used for a comparison of continuous variables. Data are presented as mean ± SD.
Results
Thirty-one women with T1DM were identified during the time period; 9 of these women met inclusion criteria and were included in our analysis. Mean maternal age at diagnosis of T1DM was 13.5 ± 5.8 years ( Table 1 ). The mean preconception HbA1c level was 6.4% ± 0.5%, and the mean HbA1c level at delivery was 5.8% ± 0.6%. The mean preconception and first-, second-, and third-trimester body mass index measurements were 24.7 ± 3.3 kg/m 2 , 25.4 ± 3.6 kg/m 2 , 27.2 ± 3.1 kg/m 2 , and 29.5 ± 3.1 kg/m 2 , respectively. Subjects were all white; the mean maternal age was 31.4 ± 2.2 years, and 3 women were nulliparous. No patient had evidence of end-organ damage. Four women had vaginal deliveries; 3 women had primary cesarean deliveries, and 2 women had repeat cesarean delivery. Indications for primary cesarean delivery were macrosomic profile on ultrasound, a nonreassuring fetal heart rate tracing in the setting of severe preeclampsia, and a history of a shoulder dystocia in a previous delivery. Neonates of these subjects were born at 38 weeks 5 days’ (±6 days) gestation and weighed 3695 ± 440 g ( Table 2 ).
| Characteristic | Measure |
|---|---|
| Age, y a | 31.4 ± 2.2 |
| Parity, n (%) | |
| Nulliparous | 3 (33) |
| Multiparous | 6 (67) |
| Age at diagnosis, y a | 13.5 ± 5.8 |
| Ethnicity: white | 9 (100) |
| Hemoglobin A1c, % a | |
| Before conception | 6.4 ± 0.5 (range, 5.8–7.4) |
| Delivery | 5.8 ± 0.6 (range, 5.3–6.4) |
| Body mass index, kg/m 2 a | |
| Before conception | 24.7 ± 3.3 |
| First trimester | 25.4 ± 3.6 |
| Second trimester | 27.2 ± 3.1 |
| Third trimester | 29.5 ± 3.1 |
| Neonatal outcome | Measure |
|---|---|
| Gestational age at delivery, wk/d a | 38/5 ± 6 d |
| Birthweight, g a | 3695 ± 440 |
| Mode of delivery, n (%) | |
| Vaginal | 4 (44) |
| Primary cesarean | 3 (33) |
| Repeat cesarean | 3 (33) |
By the end of gestation, total daily dose of insulin (TDI) nearly tripled; at preconception evaluation, the mean dose was 33.3 ± 7.8 units/day (U/d), which rose to 93.5 ± 27.9 U/d at the time of delivery. This represents a 181% change in daily insulin requirement. Total units of insulin per kilogram of maternal body weight increased from 0.6 U/kg preconceptionally to a maximum of 1.3U/kg at 36 weeks’ gestation.
Across gestation, the basal TDI increased moderately by 48%, from 16.2 ± 6.5 U/d preconceptionally to 24.0 ± 9 U/d at delivery. The basal insulin dose peaked at 33.2 ± 2.7 weeks’ gestation, with a mean TDI increase of 5.8 ± 4.3 U/d from baseline. The basal nadir was at 9 weeks’ gestation when the TDI decreased by 1.9 U/d from the preconceptional baseline, which represents a decrease of 11.4% ( Figure ). Analysis of the specific time segments during the day demonstrated that the time of day that required the highest increase of basal insulin was the morning segment (0500-0900), with an 85% (range, 3–150%) increase that was noted from preconception to peak (33.2 ± 2.7 weeks’ gestation; Table 3 ). The time segment that required the smallest increase of basal insulin was the late morning-early afternoon (0900-1600) segment, with an increase of only 8% ( Table 3 ).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
