Abstract
AIM: To evaluate pregnancy and post-delivery outcomes with non-automated insulin pumps or multiple daily injections (MDI) in women with type 1 diabetes (T1D). METHODS: A preplanned secondary analysis of the CopenFast trial including women with T1D using continuous glucose monitoring (CGM), primarily intermittently scanned CGM, routinely. Pregnancy and post-delivery outcomes were compared between insulin pump users and MDI users. Insulin pump settings during pregnancy and post-delivery were compared to pre-pregnancy. RESULTS: Of 141 women, 39 used non-automated insulin pumps and 102 used MDI. Median diabetes duration was 17 (IQR 12-20) vs. 14 (8-21) years (P = 0.12). HbA(1c) was 48 (44-53) mmol/mol vs. 47 (42-53) at 9 weeks (P = 0.65) and 43 (40-46) vs. 43 (39-46) at 35 weeks (P = 0.53). Mean sensor glucose decreased from ~ 7.0 mmol/l at 9 weeks to 6.3 mmol/l at 33 weeks in both groups. Preterm delivery (< 37 weeks) was more common with insulin pumps (25.9% vs. 16.7%, P = 0.01), as was caesarean section (59% vs. 40%, P = 0.04), which was independent of diabetes duration, age and microvascular complications. At 1 and 3 months post-delivery, maternal and infant outcomes, including breastfeeding, were similar between groups. In insulin pump users, basal insulin rates were ~ 37% higher, and carbohydrate-to-insulin ratios ~ 61% lower at 33 weeks while basal insulin rates were ~ 20% lower, and carbohydrate-to-insulin ratios were similar 3 months post-delivery, compared to pre-pregnancy. CONCLUSION: Despite routine use of CGM and similar glycaemic control, non-automated insulin pump users did not achieve improved pregnancy outcomes compared to MDI users.