Continuous Glucose Monitoring Initiation Within First Year of Type 1 Diabetes Diagnosis Is Associated With Improved Glycemic Outcomes: 7-Year Follow-Up Study

Continuous glucose monitoring (CGM) improves A1c and reduces hypoglycemia regardless of age, sex, and insulin delivery methods in patients with type 1 diabetes (1–4). Observational studies by our group and others have suggested that initiation of CGM soon after type 1 diabetes diagnosis is feasible and results in short-term improvements in A1c (5–8). In a previously published study by our group, we evaluated glycemic control up to 3 years among patients with type 1 diabetes who initiated CGM within the first year of type 1 diabetes diagnosis (n = 81) relative to patients who did not initiate CGM (n = 225) (6). Irrespective of insulin delivery methods, CGM users had a significantly greater improvement in glycemic control than did non-CGM users over 2.5 years of follow-up (6). To evaluate long-term (7 years) glycemic outcomes of early CGM initiation in patients recently diagnosed with type 1 diabetes, we followed this previously published cohort (6) for up to 7 years.

This is an extension of a previously published study (6). In brief, we searched electronic medical records of patients with type 1 diabetes, between 1 and 35 years of age, diagnosed between January 2013 and December 2015. Only nonpregnant patients who had at least two clinic visits per year for 3 years and had 70% of device use during the study period were included in the parent study. The details of the original study with first 3 years of follow-up are published (6) and provided in the Supplementary Appendix.

For the extension study, we collected baseline information on age, sex, ethnicity, insurance type, BMI, presence of other autoimmune disease, A1c, and CGM use (yes/no) at each visit from the onset of type 1 diabetes through a 7 year follow-up period by CGM groups. Patients who initiated CGM within the first year of diagnosis and continued using CGM through the study period were included in the CGM group, and patients who never used CGM during the study period were included in the no-CGM group. With increasing use of CGM, there were many patients who started CGM later during this ongoing observational study, so we defined the new-CGM group as patients who were not using CGM during the original study but started CGM during the extension phase. The Colorado Multiple Institutional Review Board approved the protocol under the exempt category.

The primary outcome of the study was the change in A1c over time between the CGM and no-CGM groups. The secondary outcome was the difference in A1c between the new-CGM group and the no-CGM group. A1c values at diagnosis were determined from the electronic medical record. A1c values at subsequent visits were averaged within each 6-month period through the first 3 years of the study and then annually through 7 years. Continuous data were presented as mean and SD, and categorical data were presented as a percentage. Linear mixed models were used to examine A1c levels by time points and by the three CGM groups adjusted for age, sex, insulin delivery methods, race or ethnicity, and insurance status.

A total of 396 patients with type 1 diabetes (n = 372 children aged <18 years; n = 24 adults aged ≥18 years) were included in this analysis and followed up to 7 years. There were no significant differences in patients’ age at diagnosis (10.4 ± 7 years vs. 10.2 ± 4.7 years) and in A1c at diagnosis (11.5% ± 2.3% vs. 11.6% ± 2.3%) between the CGM and no-CGM groups. Other baseline characteristics by CGM use are shown in Supplementary Table 1. The number of participants followed over the 7 years, by the three CGM groups and insulin pump use among these participants, is provided in Supplementary Tables 2 and 3.

Mean A1c at diagnosis did not differ between groups. There was significant improvement in A1c in the CGM group at 6 months, and this was maintained throughout the 7-year follow-up period, adjusting for age at diagnosis, sex, and insulin delivery method (Fig. 1). Among patients who did not initiate CGM within the first year of type 1 diabetes diagnosis but started CGM later (i.e., the new-CGM group), there was a significant reduction in A1c after CGM initiation. However, mean A1c was higher in the new-CGM group compared with those who initiated CGM within the first year of type 1 diabetes diagnosis (Fig. 1). Results of primary outcome did not differ after adjusting for age, sex, insulin delivery methods, race or ethnicity, and insurance status (Supplementary Table 4).

To our knowledge, this is the first study to demonstrate improvement in A1c over 7 years with CGM initiation within first year of diabetes diagnosis in patients with type 1 diabetes. Our study findings suggest improvement in A1c regardless of CGM initiation timing. However, sustained improvement in A1c was significantly better in those who initiated CGM within the first year of diabetes diagnosis compared with CGM initiation after 3 years of type 1 diabetes. Studies have suggested that glycemic control may settle into long-term patterns within the first 5 years after diabetes diagnosis (9,10); therefore, initiating CGM early may set a better glycemic trajectory. In addition, early glycemic control may offer protection against microvascular complications (11,12).

Disparity in diabetes care and in initiation of diabetes technologies such as insulin pumps and CGM is well known (13,14). There was significant disparity in CGM initiation in our study, with more non-Hispanic White patients than people of color initiating CGM within the first year of type 1 diabetes diagnosis. CGM coverage has been expanded in the last 5 years, and government insurance such as Medicaid and Medicare is now covering CGM for people with type 1 diabetes. Thus, differences in insurance coverage may not fully explain the disparity in CGM initiation we observed between non-Hispanic Whites and people of color. In regard to higher diabetes-related morbidity and mortality rates in ethnic and minority populations (15), early initiation of CGM may have a key role in optimizing glycemic control and decreasing long-term complications.

Long-term follow-up and rigorously collected data on a large number of patients with type 1 diabetes are major strengths of this study. The single-center, retrospective study design and small sample size during the end of the study are major limitations. The possibility of selection bias in initiating CGM by providers on the basis of insurance coverage, household income, access to health care, and/or status of patients’ or families’ diabetes knowledge cannot be excluded. We did not collect data on differences in health care use (e.g., emergency visits for hyperglycemia or hypoglycemia) among CGM users versus nonusers. Our previous study showed significantly reduced health care use among early CGM users with type 1 diabetes (6). We also did not collect the data on the use of automated insulin delivery systems. It is possible that patients using both an insulin pump and CGM may be using an automated insulin delivery system, which may confound A1c differences between two groups. We followed patients who initiated CGM during the 2013–2015 period; CGM during that period used older-generation devices. Newer CGM devices are more accurate, smaller, and do not require finger-stick blood glucose samples for CGM calibration and insulin dosing decisions. Therefore, we believe that findings of our study would be replicable and may show even better A1c reduction with early initiation of the newer-generation of CGM devices.

In conclusion, our study demonstrated significant improvements in A1c among CGM users compared with non-CGM users. The long-term improvement in A1c was significant with early initiation of CGM within first year of type 1 diabetes diagnosis compared with CGM initiation after 3 years since type 1 diabetes diagnosis. Therefore, CGM initiation soon after type 1 diabetes diagnosis may be a good strategy for improving glycemic outcome in people with type 1 diabetes, among whom one in every three patients does not meet the American Diabetes Association recommended glycemic targets (A1c <7%).

Acknowledgments. The authors thank Bing Wang, EPIC data manager at the Barbara Davis Center for Diabetes, for his help with EMR data collection.

Funding. Part of this work is funded by an investigator-initiated grant to the University of Colorado from Dexcom Inc.

Duality of Interest. H.K.A. has received research grants through the University of Colorado from Eli Lilly and Company, Senseonics, Dexcom, REMD Biotherapeutics, and IM Therapeutics. V.N.S. has received grants through the University of Colorado from Sanofi, NovoNordisk, Eli Lilly and Company, Insulet, and Dexcom; and honoraria from Dexcom, Insulet, and Medscape for consulting and/or speaking. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. A.C. reviewed medical charts and cleaned the data. G.T.A. and H.K.A. helped with data cleaning and electronic medical record data query. J.K.S.-B. helped with data cleaning and analysis. V.N.S. conceptualized the project, wrote the first draft of the manuscript, and secured funding. All authors reviewed and approved final version of the manuscript. V.N.S. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.