Use of professional-mode flash glucose monitoring, at 3-month intervals, in adults with type 2 diabetes in general practice (GP-OSMOTIC): a pragmatic, open-label, 12-month, randomised controlled trial

Summary Background Continuous glucose monitoring, either real-time (personal) or retrospective (professional mode), can identify day-to-day glucose profiles to guide management decisions for people with type 2 diabetes. We aimed to examine the effects of professional-mode flash glucose monitoring, done at 3-month intervals, in adults with type 2 diabetes in general practice. Methods We did a pragmatic, two-arm, open label, 12-month, individually randomised controlled trial (GP-OSMOTIC) in 25 general practices in Victoria, Australia. Eligible participants were adults aged 18–80 years, with type 2 diabetes diagnosed for at least 1 year and HbA1c at least 5·5 mmol/mol (0·5%) above their target in the past month despite being prescribed at least two non-insulin glucose-lowering drugs, insulin, or both (with therapy stable for at least 4 months). We randomly assigned participants (1:1) to either use of a professional-mode flash glucose monitoring system or usual clinical care (control). All participants wore the flash glucose monitoring sensor at baseline, and electronic randomisation (using permuted block sizes of four and six, and stratified by clinic) was done after the sensor was attached. Masking of participants and treating clinicians to group allocation was not possible, but the study statistician was masked to allocation when analysing the data. At baseline, and 3, 6, 9, and 12 months, participants in the flash glucose monitoring group wore the professional-mode flash glucose monitoring sensor for 5–14 days before their general practice visit. The sensor recorded interstitial glucose concentrations every 15 min, but the glucose data were not available to the participant until their general practice visit, where the sensor output would be uploaded to a computer by the health professional and discussed. Control group participants wore the sensor at baseline and at 12 months for data analysis only, and had usual care visits every 3 months. The primary outcome was the between-group difference in mean HbA1c at 12 months. Secondary outcomes were the between-group differences in: mean percentage time in target glucose range (4–10 mmol/L), based on ambulatory glucose profile data at 12 months; mean diabetes-specific distress (assessed with the Problem Areas In Diabetes [PAID] scale) at 12 months; and mean HbA1c at 6 months. Analysis was done by intention to treat. This trial is registered at the Australian and New Zealand Clinical Trials Registry, ACTRN12616001372471. Findings Between Oct 4, 2016, and Nov 17, 2017, we randomly assigned 299 adults: 149 to flash glucose monitoring and 150 to usual care. At 6 months, HbA1c was lower in the flash glucose monitoring group than in the usual care group (difference −0·5%, 95% CI −0·8% to −0·3%; p=0·0001). However, at 12 months (primary outcome), there was no significant between-group difference in estimated mean HbA1c (8·2% [95% CI 8·0 to 8·4] for flash glucose monitoring vs 8·5% [8·3 to 8·7] for usual care; between-group difference −0·3%, 95% CI −0·5 to 0·01; [66 mmol/mol, 95% CI 64 to 68 vs 69 mmol/mol, 67 to 72; between-group difference −3·0, 95% CI −5·0 to 0·1]; p=0·059). Mean percentage time spent in target glucose range at 12 months was 7·9% (95% CI 2·3 to 13·5) higher in the flash glucose monitoring group than in the usual care group (p=0·0060). Diabetes-specific distress PAID scores were unchanged at 12 months (between-group difference −0·7, 95% CI −3·3 to 1·9; p=0·61). No episodes of severe hypoglycaemia or treatment-related deaths were reported. One participant died during the study from causes unrelated to the intervention (following complications post-myocardial infarction with multiple comorbidities). Interpretation Professional-mode flash glucose monitoring in adults with type 2 diabetes in general practice did not improve the primary outcome of HbA1c at 12 months or diabetes-specific distress compared with usual care, but did improve time in target glucose range at 12 months and HbA1c at 6 months. Our findings suggest that professional-mode flash glucose monitoring can be implemented in a pragmatic primary care environment. Although there was no change in HbA1c at 12 months, the improved time in target range might reflect the potential of the technology to support personalised clinical care by providing insights into glycaemic profiles for some people with type 2 diabetes. Funding National Health and Medical Research Council of Australia, Sanofi Australia, and Abbott Diabetes Care.