Real time continuous glucose monitoring in high-risk people with insulin-requiring type 2 diabetes: A randomised controlled trial

Aims: To investigate the impact of real-time continuous glucose monitoring (rtCGM) on glycaemia in a predominantly indigenous (M & amacr;ori) population of adults with insulin-requiring type 2 diabetes (T2D) in New Zealand. Methods: Twelve-week, multicentre randomised controlled trial (RCT) of adults with T2D using >= 0.2 units/kg/day of insulin and elevated glycated haemoglobin (HbA1c) >= 64 mmol/mol (8.0%). Following a 2-week blinded CGM run-in phase, participants were randomised to rtCGM or control (self-monitoring blood glucose [SMBG]). The primary outcome was time in the target glucose range (3.9-10 mmol/L; TIR) during weeks 10-12, with data collected by blinded rtCGM in the control group. Results:Sixty-seven participants entered the RCT phase (54% M & amacr;ori, 57% female), median age 53 (range 16-70 years), HbA1c 85 (IQR 74, 94) mmol/mol (9.9 [IQR 8.9, 10.8]%), body mass index (36.7 +/- 7.7 kg/m(2)). Mean (+/- SD) TIR increased from 37 (24)% to 53 (24)% [Delta 13%; 95% CI 4.2 to 22; P = 0.007] in the rtCGM group but did not change in the SMBG group [45 (21)% to 45 (25)%, Delta 2.5%, 95% CI -6.1 to 11, P = 0.84]. Baseline-adjusted between-group difference in TIR was 10.4% [95% CI -0.9 to 21.7; P = 0.070]. Mean HbA1c (+/- SD) decreased in both groups from 85 (18) mmol/mol (10.0 [1.7]%) to 64 (16) mmol/mol (8.0 [1.4]%) in the rtCGM arm and from 81 (12) mmol/mol (9.6 [1.1]%) to 65 (13) mmol/mol (8.1 [1.2]%) in the SMBG arm (P < 0.001 for both). There were no severe hypoglycaemic or ketoacidosis events in either group. Conclusions: Real-time CGM use in a supportive treat-to-target model of care likely improves glycaemia in a population with insulin-treated T2D and elevated HbA1c.