International diabetes registries repeatedly show that adolescents and young adults with type 1 diabetes (T1D) have worse A1C outcomes compared with other age groups, with direct consequence on their future complication risks. Real-time continuous glucose monitoring (RT-CGM) use has been associated with lower A1C levels, although previous studies have reported that the reduction was attenuated in younger patients.

My colleagues and I speculated that device burden (eg, finger stick testing for calibration, inaccurate sensors, and false alarms) associated with older RT-CGM technologies may have been a contributory factor and led to the suboptimal user adherence and outcomes observed in prior research. People with T1D from diverse ethnic and social backgrounds have also tended to be under-represented in device and technology studies. There is still a lack of evidence on the efficacy and usability of diabetes technologies in these under studied cohorts.

Sensor-based glucose outcomes—specifically time in range (TIR)—have recently gained traction as a measure of glycemic control beyond A1C, as A1C alone does not provide information relevant to hypoglycemia risk or glucose variability, factors that are known to impact glycemic control and quality of life in people living with T1D. TIR of 70-180 mg/dL is now accepted as an outcome of therapeutic efficacy in T1D clinical studies, and improvement in TIR has been associated with reduced risk of microvascular complications.

Calibration-Free RT-CGM vs Capillary Glucose Testing

On this background, we published a randomized cross-over design study in Diabetes Care that compared the efficacy of a RT-CGM system that is calibration-free and linked to a smartphone app with that of conventional self-monitoring of blood glucose (SMBG). We included adolescents and young adults aged 16–25 with poorly controlled T1D from various ethnic and social backgrounds to represent usual clinical practice.

Participants were randomly assigned to 8 weeks of RT-CGM intervention and SMBG control periods, with a 4-week wash out between these two periods. During the intervention, they installed the Dexcom G6 app on their own smartphone and were trained on using RT-CGM sensor data for decision making and insulin dose adjustments. During the SMBG period, they performed finger stick capillary glucose measurements as per usual clinical practice. The main study endpoint was the difference in TIR (70-180 mg/dL) between RT-CGM and SMBG, with secondary outcomes that included A1C and usability (% of sensor wear).

For latest news and updates

Reduced Technology Burden & Improved Glycemic Outcomes

Our study had a higher proportion of participants from diverse ethnic backgrounds compared with other T1D technology studies, with more than one-half living in areas constituting the most deprived areas of England. At baseline, they had high A1C levels (78 mmol/mol, or 9.3%) and were predominantly hyperglycemic, consistent with observations in clinical practice. We found that RT-CGM users spent an additional 2.6 hours per day with glucose levels in the normal range, compared with SMBG users. A1C was reduced by 8.5 mmol/mol, in favor of RT-CGM (Table).

The proportion of time spent hyperglycemic (glucose levels >180 mg/dL) was substantially reduced with RT-CGM by nearly 3 hours per day when compared with SMBG. This cohort had minimal hypoglycemia (<1.5% of the whole period), and thus, no statistically significant difference was found between the two groups. Use of RT-CGM in our study was relatively high compared with previous studies in this age group (sensor use was 84% of the whole study period).

We speculate that the reduced daily burden of using a calibration-free RT-CGM that is licensed for insulin dosing and linked to a smartphone app may have attributed to the higher adherence of sensor use in this age group. This higher adherence was also likely associated with the A1C reduction observed with RT-CGM use when compared with SMBG use. Further investigation is required to evaluate longer-term outcomes in this patient population.