CGM Lowered A1c by 0.9 Points More Than Routine Care in Type 2 Diabetes Without Insulin. Coverage Has Not Caught Up.
The Dexcom-funded CONNECT trial of 283 adults, published Oct. 14 in NEJM, ran 26 weeks with little hypoglycemia. Medicare has not yet decided whether to cover sensors for patients not on insulin.
On Oct. 14, the New England Journal of Medicine published CONNECT, a randomized trial of continuous glucose monitoring in adults with type 2 diabetes who do not take insulin. The trial was funded by Dexcom, which makes the sensor tested. In its announcement, the company called it the first NEJM report of clinical data on standalone real-time CGM in nearly two decades.
The device produced a larger drop in A1c than routine care over six months. Whether that becomes coverage, and a workable step in your clinic, is a separate question that Medicare has not yet answered.
What the trial tested
CONNECT ran for 26 weeks at 22 primary care practices in the United States. According to the published abstract, 145 adults were assigned to CGM and 138 to routine care with a glucose meter; ages ranged from 21 to 84. A Cleveland Clinic Journal of Medicine summary of the June presentation at the American Diabetes Association Scientific Sessions reports that 440 people were screened, 265 completed the trial and enrollment required an A1c of 7.5 percent or higher. The CGM arm used the Dexcom G7. All participants received diet and exercise education and stayed on their existing medications.
Mean age was 60 years and 31 percent began with an A1c of 9 percent or higher. About 40 percent were taking an incretin-based drug and 37 percent an SGLT-2 inhibitor. The primary outcome was A1c at 26 weeks.
What it showed
Mean A1c fell from 8.8 to 7.2 percent with CGM and from 8.7 to 8.0 percent with routine care, an adjusted difference of 0.9 percentage points (95% CI, 0.6 to 1.1; P<0.001), per the abstract. Time in range (70 to 180 mg/dL) was 21 percentage points higher with CGM (95% CI, 16 to 26), or 62 percent against 41 percent in the ADA presentation, about five more hours a day.
The Cleveland Clinic summary adds that 68 percent of CGM users finished below 7.5 percent and 46 percent below 7.0 percent, and that patients starting above 10 percent saw a 3.1-point drop with CGM, 2.1 points more than control. Dexcom says 97 percent of participants were still using the sensor at 26 weeks. Roy W. Beck, MD, PhD, medical director of the Jaeb Center for Health Research and the senior author, said in the Dexcom release that the results support "consideration of the use of a CGM much earlier in Type 2 diabetes management before insulin is needed."
A six-month, sponsor-funded trial shows that a sensor lowers A1c in patients already engaged in primary care. It does not yet show that the effect lasts.
Safety and limits
Hypoglycemia was rare in both arms. Time below 54 mg/dL was 0.17 percentage points lower with CGM and time below 70 mg/dL 0.20 points lower, meeting the trial's noninferiority test, and no severe hypoglycemia events were reported, per the abstract.
The limits are plain. Dexcom funded the trial and only its G7 was tested. Patients entered with an A1c of at least 7.5 percent, so the result does not speak to well-controlled patients. The control arm used a meter without CGM, the trial was open by design, and A1c is a surrogate endpoint. A six-month extension is underway to test whether the benefit holds at 12 months.
Coverage is the open question
MedTech Dive reports that Dexcom is waiting on the Centers for Medicare and Medicaid Services to decide whether to expand coverage of CGM to people with type 2 diabetes who do not use insulin, and that the timing is still unknown. Leerink Partners analyst Mike Kratky wrote that the company expects a decision before the end of the year, with implementation in mid-2027. That is an analyst's estimate, not a CMS timeline; we found no CMS statement on the matter.
What it means in clinic
Identify patients on oral agents or a GLP-1 receptor agonist with an A1c at or above 7.5 percent; they most resemble the trial population. Before ordering, ask your billing office what each major payer covers for patients not on insulin, since an uncovered sensor is not an option for many of them. If you start someone on one, decide in advance who reviews the data and how often, and how that work is billed, a question our remote monitoring coverage touches on. Set limits on how sensor alerts reach you, as with any patient portal inbox, so the device adds a clinical decision rather than another inbox item.
Reported, fact-checked and edited by The Script Pad's editorial staff.
This article is for professional education and does not replace clinical judgment. Treatment decisions should be based on the individual patient and current guidelines.
