CGM in Non-Insulin Type 2 Diabetes Reduces A1c by 0.5% — But Who Benefits Most Remains Unclear

Continuous glucose monitoring — the subcutaneous sensor that measures interstitial glucose every 5 minutes and displays readings on a smartphone or reader — transformed type 1 diabetes management over the past decade. In type 1 diabetes and insulin-using type 2 diabetes, CGM reduces A1c, reduces hypoglycemia events, and improves quality of life. The FDA cleared CGM for non-insulin type 2 diabetes, and payers including Medicare have expanded coverage. The question is whether CGM produces meaningful glycemic and clinical benefit in the much larger population of patients with type 2 diabetes managed with oral agents or GLP-1 receptor agonists — where there is no insulin to titrate in response to glucose readings.[1, 2]

The JAMA Internal Medicine systematic review identified 15 RCTs with 2,847 participants comparing CGM versus usual care (SMBG or clinical follow-up alone) in adults with type 2 diabetes not using insulin. The pooled effect on A1c was −0.50% (95% CI −0.64% to −0.36%), a statistically significant and clinically modest improvement. Time in range (70–180 mg/dL) improved by approximately 7–8 percentage points. Hypoglycemia events were numerically reduced, though most included trials were not powered to detect hypoglycemia differences in the non-insulin population.[1]

Critical findings from subgroup and quality analyses: The A1c benefit was larger in trials where participants had higher baseline A1c (≥8.5%) — suggesting that CGM has more room to help patients with genuinely poor glycemic control than those already near target. Trials lasting beyond 6 months showed attenuating effect sizes, raising the question of whether initial behavior-change from seeing real-time glucose data is sustained over time. Studies using intermittently scanned CGM (Libre-type devices) and real-time CGM showed similar magnitudes of benefit. No trial was powered to assess cardiovascular events, hospitalizations, or microvascular outcomes — the endpoints that matter most.[1, 3]

Clinical Context

The evidence base for CGM in type 2 diabetes has accumulated rapidly since the DIAMOND trial (2017, Ann Intern Med), which demonstrated that CGM reduced A1c by 1.0% (vs 0.5% with SMBG) in insulin-using type 2 diabetes patients over 24 weeks.[2] The MOBILE trial (2021, JAMA), crucially, extended CGM benefit to type 2 diabetes patients using only basal insulin — a much larger population — showing a 0.4% greater A1c reduction and fewer hypoglycemic episodes over 8 months.[3] These two trials anchored Medicare's 2023 coverage expansion to non-insulin type 2 diabetes. What remained uncertain after MOBILE was whether the benefit extends to the largest tier of type 2 diabetes patients: those managed with oral agents or GLP-1 agonists alone, without any insulin.

The plausible mechanism for CGM benefit in non-insulin users is behavioral: continuous feedback on how specific meals, activity patterns, and stress affect glucose may reinforce lifestyle choices in a way that quarterly A1c measurements cannot. Several smaller trials had reported modest A1c reductions in this population, but they were underpowered and often unblinded, making placebo effects from increased self-monitoring attention impossible to separate from true glucose feedback effects. A rigorous systematic review and meta-analysis that pools these studies — with attention to bias and heterogeneity — is the appropriate instrument to resolve whether the A1c reduction observed is large enough, consistent enough, and produced through a mechanism specific enough to justify CGM in non-insulin type 2 diabetes at scale.[2]

Why It Matters Clinically

The mechanism by which CGM improves glycemia in non-insulin users differs from the insulin-titration mechanism. In patients not using insulin, there is no rapid dose to adjust in response to a high reading. The benefit appears to be behavioral: seeing the real-time glucose response to meals, physical activity, sleep, and stress gives patients a feedback loop that supports dietary and lifestyle modification. This is a different proposition than insulin-using diabetes management. Whether it produces durable behavior change or only transient improvement — the "Hawthorne effect" of being observed, which resolves when novelty wears off — remains the key unanswered question after this meta-analysis.[2, 3]

Limitations

Most trials lasted only 3–6 months — too short to assess durability of glycemic benefit or impact on microvascular complications. The 0.5% A1c reduction, while statistically significant, is modest and may not justify the cost of CGM devices in resource-limited settings when compared to medication intensification. Trials enrolled predominantly White populations in high-income countries; generalizability is limited. No trial assessed CGM combined with structured dietary counseling versus CGM alone — the behavioral support needed to actualize CGM's potential.