Continuous Glucose Monitoring versus Standard of Care in Adolescents With Type 2 Diabetes: A Pilot Randomized Cross-Over Trial

Youth with type 2 diabetes (T2D) develop early treatment failure and are at risk for diabetes-related comorbidities. Continuous glucose monitoring (CGM) has the potential to decrease the burden of diabetes care, improve consistent medication administration, and achieve in-target HbA1c in this patient population. ¹ We aimed to determine the feasibility of CGM use and its effect on glycemic outcomes in youth with T2D.

Adolescents 15 to 19 years old from a pediatric hospital in Los Angeles (CA) prescribed insulin and with HbA1c ≥8% were randomized to 3 months of Dexcom G6 CGM or blood glucose monitoring (BGM), followed by 3 months of washout period and then crossed over to the other intervention arm for 3 months. A nurse practitioner provided coaching for the first 30 days of CGM use. The CGM-SAT ² satisfaction survey was administered at the end of CGM wear. Descriptive statistics were used for group comparison. Mixed-effect models with a 3-month change in HbA1c as the dependent variable tested for treatment differences across groups and period effect, with p<.05 considered significant.

Adolescents (N=9) were female (55%), 17.4±2.0 years old, with a T2D duration of 4.0±2.8 years, and HbA1c 11.5±2.5%. Majority (89%) of the participants were Latinx and were on metformin and basal/bolus insulin regimen. All patients were publicly insured. Most patients had daily contact with the coach during the first week of CGM wear and contact intensity decreased in most cases to once a week after the second week. Figure 1 shows CGM use lowered HbA1c by 2.8% for participants in the CGM group (BGM: −0.1%, CGM: −2.9%, p=.003) and by 3.8% for those who wore the CGM >85% of the time (p=0.01). Frequency of CGM wear was unchanged over the 3 months of the CGM use (p=.5). On average, % time in range (sensor glucose 70-180 mg/dL) was 46.3±24.8%. About 26.7±32.4% of the time was above 250 mg/dL and 26.1±10.4% was between 181 and 250 mg/dL. Time below range was below 1%. There was no statistically significant difference in time above or in range across the 3 months of CGM wear. The CGM device was favorably perceived by participants, although participant CGM satisfaction survey score did not correlate with changes in HbA1c (correlation coefficient 0.223, 95% confidence interval [CI]: −0.164 to 0.097).

Figure 1.

Changes in HbA1c over time. Shaded area indicates washout period.

We demonstrate in this pilot that CGM wear in youth with T2D is feasible and effective. Similar to a recent report, ³ participants in our pilot study also expressed high satisfaction with the CGM device. In our cohort, however, we identified an improvement in HbA1c after 3 months of CGM wear, despite comparable mean HbA1c. Our pilot findings provide evidence that CGM use in the context of coaching has the potential to improve glycemia in this socioeconomically disadvantaged population. It supports public policies that reduce barriers to CGM access and expand insurance coverage of CGM in insulin-requiring youth with T2D. Future studies are needed to determine the role of coaching in improving glycemic outcome in CGM users.