Real-Time CGM vs Flash CGM: Type 1 Diabetes Evidence

Real-time CGM appears to improve the day-to-day shape of glucose control more than flash CGM in adults with type 1 diabetes. It produced better time in range, less time below 70 mg/dL, lower average glucose, and lower glucose variability in the only published head-to-head meta-analysis. Yet it did not produce a statistically significant HbA1c advantage, and no study has established whether the extra benefit is worth the extra cost.

That split matters. HbA1c remains a familiar scorecard for clinicians and insurers, while continuous glucose monitoring reveals how much turbulence hides behind that average. A person can have the same HbA1c with fewer lows, fewer highs, and less glucose whiplash. The available evidence suggests rt-CGM may help with those details—but it does not yet justify a universal upgrade recommendation.

What the head-to-head evidence found

Chen et al.'s 2024 systematic review and meta-analysis pooled six studies—three randomized trials and three observational studies—with 1,516 adults with type 1 diabetes. The included trials lasted 8 to 24 weeks. Compared with flash, or intermittently scanned, CGM, rt-CGM performed better on seven CGM-derived outcomes.

Time in range, defined as 70 to 180 mg/dL, favored rt-CGM with a standardized mean difference (SMD) of 0.59 (95% CI, 0.37 to 0.81; p<0.001). Time above range was lower at both greater than 10.0 mmol/L (SMD −0.38, 95% CI −0.71 to −0.04; p=0.03) and greater than 13.9 mmol/L (SMD −0.42, 95% CI −0.79 to −0.04; p=0.03). Mean glucose was also lower (SMD −0.18, 95% CI −0.31 to −0.06; p=0.003).

The largest differences involved variability. Glucose standard deviation favored rt-CGM by an SMD of −0.70 (95% CI −1.09 to −0.31; p<0.001), while coefficient of variation favored rt-CGM by −0.76 (95% CI −1.05 to −0.47; p<0.001). These are standardized effects, not percentage points or minutes per day. The meta-analysis does not provide the raw data needed to translate the time-in-range result into an absolute daily difference.

That missing translation limits clinical interpretation. A standardized effect can show a consistent directional advantage, but it cannot tell a reader whether the typical benefit was five extra minutes in range or an amount large enough to change treatment decisions. The study supports superiority on these measured CGM metrics; it does not establish a minimum clinically important time-in-range gain.

The HbA1c result keeps the conclusion narrow

Despite those advantages, Chen et al. found no statistically significant difference in HbA1c between rt-CGM and flash CGM. The study populations had baseline HbA1c values of roughly 7.5% to 8.5%, leaving limited room to separate two monitoring approaches on an average-glucose marker. The shorter studies also ran for less than, or only about, the 90-to-120-day period that HbA1c broadly reflects.

Those factors make the null result compatible with a longer-term difference that the studies could not detect. They do not prove the devices are equivalent for HbA1c. Still, the result is a real constraint: current head-to-head evidence does not show that rt-CGM lowers HbA1c more than flash CGM in these adults.

rt-CGM may redistribute glucose time—less variability, less moderate hypoglycemia, and more time in range—without shifting the average enough to separate HbA1c over 8 to 24 weeks.

That distinction is clinically sensible. HbA1c estimates average glucose exposure; standard deviation and coefficient of variation describe how widely glucose swings around that average. They answer different questions.

Hypoglycemia: the alarm may explain part of the gap

rt-CGM substantially reduced time below 3.9 mmol/L, or 70 mg/dL, versus flash CGM (SMD −1.45, 95% CI −2.33 to −0.57; p=0.001). At below 3.0 mmol/L, or 54 mg/dL, the meta-analysis found no statistically significant difference. The first result concerns mild-to-moderate biochemical hypoglycemia exposure; the second means the evidence does not show an advantage for the lower biochemical threshold.

It also does not answer the question many people care about most: severe hypoglycemia requiring another person's help or emergency treatment. None of the identified head-to-head studies reported that clinical outcome as a primary endpoint.

Device generations further complicate the comparison. In a real-world observational study of middle-aged adults with type 1 diabetes, Oriot et al. found that intermittently scanned CGM with low-glucose alarms reduced hypoglycemia incidence compared with non-alarm isCGM. This does not directly test rt-CGM against modern alarm-equipped flash devices, but it suggests that alerts—not continuous streaming alone—may account for part of the hypoglycemia advantage.

That is a meaningful limitation because newer flash systems have added alarms and, in some cases, real-time data transmission. The head-to-head meta-analysis may therefore describe an older comparison: rt-CGM versus largely non-alarm flash monitoring. No identified trial directly compares current-generation rt-CGM with alarm-equipped second- or third-generation flash CGM in adults with type 1 diabetes.

Variability is meaningful, but the benchmark is imperfect

Coefficient of variation has a practical target: 36% or lower is widely used in type 1 diabetes care. A 2024 study by Urakami et al. examined that threshold in 66 children and adolescents with type 1 diabetes and linked higher CV with worse glycemic outcomes. It offers useful context for rt-CGM's lower CV in the meta-analysis, especially for someone hovering near 36%.

But it cannot prove that the adult participants in Chen et al. crossed that threshold after switching devices. Urakami's population was pediatric, and Chen et al. reported standardized rather than absolute CV differences. The evidence supports lower variability with rt-CGM; it does not show how many adults reached a target because of it.

Cost remains an unanswered question

There is no published cost-effectiveness analysis of rt-CGM versus flash CGM specifically in adults with type 1 diabetes. That absence matters more than it may sound. Without an incremental cost-effectiveness ratio, an absolute time-in-range gain, or a cost per hypoglycemic event avoided, claims that an upgrade "pays for itself" are guesses dressed in a lab coat.

Related studies do not fill that gap. A 2024 Brazilian analysis of FreeStyle Libre found flash CGM cost-effective versus self-monitoring of blood glucose in insulin-treated people with type 1 or type 2 diabetes. A 2026 Australian analysis evaluated rt-CGM versus self-monitoring of blood glucose in adults with insulin-treated type 2 diabetes. Both comparisons differ from an rt-CGM-versus-flash decision in adult type 1 diabetes.

Broader evidence does support CGM over finger-stick monitoring in other populations. For example, a 2024 systematic review in adults with type 2 diabetes assessed both rt-CGM and intermittently scanned CGM against self-monitoring of blood glucose. That evidence helps explain why CGM has become common; it does not settle which CGM modality offers better value for a person with type 1 diabetes.

Who should take the evidence seriously?

Someone using first-generation, alarm-free flash monitoring who has frequent glucose swings or recurrent lows has the strongest evidence-based reason to consider rt-CGM. The meta-analysis found statistically significant advantages in time in range, variability, and time below 70 mg/dL.

For someone already using alarm-equipped flash CGM with stable variability, the evidence is less direct. Modern alarms may narrow the hypoglycemia gap, and no current-device trial quantifies how much. For someone whose main concern is severe hypoglycemia requiring assistance, neither modality has demonstrated superiority on that clinical endpoint in the identified evidence.

The practical verdict is restrained: rt-CGM looks better for several daily glucose metrics than older, non-alarm flash CGM in adult type 1 diabetes. It has not shown a statistically significant HbA1c advantage, a lower rate of the most severe biochemical lows, or a proven cost advantage over flash CGM. The upgrade may be valuable for the right person, but the published evidence cannot price that value yet.