A Clinician's Guide to Using CGM to Optimize GLP-1 Therapy

A GLP-1 does reliable pharmacological work, but it doesn't tell you how a given patient is responding between visits. Continuous glucose monitoring fills that gap, turning a quarterly A1C into a continuous picture you can actually titrate against. Here's a practical framing for using CGM data to guide GLP-1 therapy.

Why CGM adds signal a GLP-1 alone doesn't give you

The effect of a GLP-1 concentrates on postprandial excursions, which is precisely what A1C averages away. In type 2 diabetes, the class produces the largest glucose-variability reduction of any glucose-lowering group, and it does so with minimal hypoglycemia because insulin release stays glucose-dependent. CGM is what makes that visible at the individual level: a patient's average can look unchanged while their post-meal curves and variability improve, and only the sensor shows it.

Reading the AGP with a GLP-1 in mind

Anchor on the ambulatory glucose profile and a few metrics. Time in range (70–180 mg/dL), with a general target above 70% in type 2 diabetes (Battelino 2019). Time below range as the safety metric, ideally under 4%, and the first thing to check in any patient also on insulin or a sulfonylurea. Coefficient of variation at or below 36% as the stability threshold. On a GLP-1, the postprandial windows of the AGP are where you'll see the drug working first, so read the daily overlay for peak height and return-to-baseline, not just the summary bar.

Using CGM to titrate

CGM gives titration an objective signal. As you step the dose up, watch whether post-meal excursions flatten and time in range trends up. Worth keeping in mind: incretin variability benefit isn't purely dose-dependent, so the highest tolerated dose isn't automatically the steadiest (Yoshikawa 2021). Let the AGP and tolerability, rather than dose alone, guide where a patient settles. In patients on background insulin or secretagogues, use rising time-below-range as your prompt to de-intensify those agents as the GLP-1 ramps.

The evidence base, and its edges

The CGM data supporting all of this is strongest in type 2 diabetes. Tirzepatide reached time in range well above target versus basal insulin in the SURPASS-3 CGM substudy (Battelino 2022), and the GRADE CGM substudy showed incretin-based agents delivering low variability and minimal time-below-range against active comparators (Bergenstal 2026). For non-diabetic and weight-management patients, CGM use is observational monitoring rather than a validated efficacy target, so frame it to those patients as watching their own response, not as a promised variability effect.

When to act on the data

A few patterns should prompt a change. Persistent time-above-range despite an adequate dose suggests revisiting diet timing or the broader regimen. Any pattern of lows warrants adjusting concomitant insulin or sulfonylurea. A climbing variability trend, even with an acceptable average, is a signal worth investigating rather than ignoring. The point of continuous data is earlier, smaller adjustments instead of waiting a quarter for the next A1C.