Continuous Glucose Monitoring (CGM) Initiation Is Associated With Reduced Mortality in Older-Onset Type 1 Diabetes Patients: A Target Trial Emulation Study Within the Veterans Health Administration
Peter D. Reaven, Sharon A. Macwan, Kevin McConeghy, Gregory J. Norman, Paul R. Conlin, Donald R. Miller, Jin J. ZhouOBJECTIVE
Use of continuous glucose monitors (CGM) improves glucose control and reduces hypoglycemia, but data are lacking for its possible role in reducing other serious clinical events.
RESEARCH DESIGN AND METHODS
We conducted a target trial emulation (TTE) analysis in patients with type 1 diabetes (T1D) comparing all-cause mortality between CGM users and non–CGM users using observational health records from the Veterans Health Administration. Participants included adult T1D patients who had their second endocrine visit (time zero) during years 2017–2020. Each patient was cloned to both treatment strategies but was censored if observed care deviated from the assigned strategy (artificial censoring) during the 6-month grace period. Inverse probability weights accounted for artificial censoring, and negative outcomes examined residual confounding.
RESULTS
Of the 8,423 individuals initially assigned to both treatment groups, 1,039 were prescribed CGM devices, while 7,399 were not censored or assigned CGM during the grace period. Mortality was lower with CGM initiation, yielding adjusted risk ratios of 0.90 (95% CI 0.71–0.97) to 0.84 (CI 0.72–0.97) over 1–4 years of follow-up. Similar risk ratios were seen with different grace periods (3 or 9 months). Those age >65 years or not on insulin pumps appeared to have greater benefit, but effects did not vary by HbA1c, race or ethnicity, or frailty. Risk ratios did not differ between groups for incident nondiabetes outcomes, including outpatient or inpatient musculoskeletal or gastrointestinal conditions.
CONCLUSIONS
In this large TTE of CGM initiation in older T1D patients, CGM use was associated with reduced risk for all-cause mortality.