Risk for ketonaemia in type 1 diabetes pregnancies with sensor-augmented pump therapy with predictive low glucose suspend compared with low glucose suspend: a crossover RCT

Aims To determine the frequency of ketonaemia in pregnant women with type 1 diabetes treated with a sensor-augmented pump (SAP) in predictive low glucose suspend (PLGS) mode compared with low glucose suspend (LGS) mode. Methods An open-label crossover pilot RCT in ten women with type 1 diabetes treated with a 640 Medtronic insulin pump, with inclusion between 12-30 weeks of pregnancy. Participants were 1/1 randomly assigned (allocation by statistician using a permuted block size of 2) to either 2 weeks with an SAP in PLGS mode or 2 weeks in LGS mode. After the first 2 weeks, participants were switched to the other mode. Ketones in the participants' serum were measured three times daily (fasting, midday and evening) during the 4 weeks. The primary endpoint was the frequency of blood ketones > 0.6 mmol/l. Participants and healthcare providers were not blinded to group assignment for assessment of outcomes. Results The median gestational week at inclusion was 12.5 weeks (12.0-15.0), participants had a median age of 31.5 years (24.0-33.0), BMI of 26.6 kg/m(2) (24.5-31.8), baseline HbA(1c) of 41 mmol/mol (40-43; 5.9% [5.8-6.1]) and baseline time in range (TIR, 3.5-7.8 mmol/l) of 64.6% (55.6-68.7). Comparing the LGS mode with the PLGS mode, insulin suspension time per day was 2.0 h (1.3-2.3) vs 3.5 h (3.3-5.0; p = 0.002), ketonaemia > 0.6 mmol/l was 0% vs 0.5% (p = 1.000) and no participants had ketonaemia > 1 mmol/l. TIR on LGS was 64.7% (58.0-68.7) vs 61.1% (56.5-67.5) on PLGS (p = 0.492), time < 3.5 mmol/l was higher on LGS at 7.5% (4.6-8.3) vs 4.2% (2.4-6.9) on PLGS (p = 0.014). Treatment satisfaction and fear for hypoglycaemia were similar whether using LGS or PLGS mode. Conclusions/interpretation Despite longer time periods with suspended insulin delivery, pregnant women using an SAP in PLGS mode were not at higher risk of developing ketonaemia compared with those in LGS mode. Women with an SAP in PLGS mode had similar TIR with less time in hypoglycaemia.

Automated summary

This study aimed to compare the frequency of ketonaemia in pregnant women with type 1 diabetes treated with a sensor-augmented pump in predictive low glucose suspend (PLGS) mode versus low glucose suspend (LGS) mode. The results showed that despite longer periods of suspended insulin delivery, pregnant women using an SAP in PLGS mode did not have a higher risk of developing ketonaemia compared to those in LGS mode, while also showing similar time in range and less time in hypoglycemia.