CGM in type 2 diabetes without insulin: the CONNECT trial in the NEJM

For 26 weeks, 283 adults with type 2 diabetes who were not using insulin, and whose HbA1c was at least 7.5 percent, received either real-time CGM or usual care. The trial was run in US primary care. In Austria, Germany and Switzerland, insurers cover CGM only for people on insulin, so this group is usually not covered.

Results

CONNECT was first presented at the ADA meeting on 6 June 2026. The full paper appeared in the New England Journal of Medicine on 14 October. From 22 US primary-care practices, 283 adults were randomised to real-time CGM (Dexcom G7) or usual care with self-monitoring of blood glucose. The trial was open-label.

With CGM, mean HbA1c fell from 8.8 to 7.2 percent, and with usual care from 8.7 to 8.0 percent. The adjusted difference was 0.9 % (95% CI 0.6 to 1.1). Time in range at the end was 62 versus 41 percent, about five more hours a day. Forty-six percent of the CGM group reached an HbA1c below 7 percent.

Limits

Hard outcomes were not studied. An ongoing six-month extension is meant to show whether the effect lasts. Dexcom funded the trial, and employees of the manufacturer are co-authors.

Reimbursement in Switzerland, Germany and Austria

In Switzerland, basic insurance pays for CGM in intensive insulin therapy, or in insulin treatment when HbA1c is at least 8 percent. In Germany, intensive insulin therapy is the condition, including for FreeStyle Libre, which is listed as real-time CGM in the aids directory. In Austria, the ÖGK ties coverage to functional insulin therapy or pump therapy. The CONNECT population, type 2 diabetes without insulin, therefore sits outside usual reimbursement in all three countries.

Results

Limits

Reimbursement in Switzerland, Germany and Austria

Sources (7)