Is Abbott Lingo Accurate? What Its FDA Filing Shows

What Does Lingo Actually Measure?

Lingo measures interstitial glucose continuously and reports it through two proprietary metrics: a "Lingo Score" and "Lingo Counts," rather than the clinical-standard Time in Range figure directly. That's a meaningfully different design choice from Dexcom's Stelo, which shows Time in Range unmodified.

Abbott's own materials describe a "typical" healthy-range glucose band of 70-140 mg/dL and state plainly that Lingo is "not intended for diagnosis of diseases, including diabetes." Abbott's own guidance recommends consulting a healthcare professional before making dietary or exercise changes based on the data.

Is Lingo FDA Cleared?

Yes. The FDA cleared the "Lingo Glucose System" (applicant: Abbott) on May 29, 2024, as a traditional 510(k) clearance, alongside a related over-the-counter CGM, Libre Rio, in the same clearance window.

As with Stelo, FDA clearance establishes the device met the agency's safety and performance requirements for its stated intended use. It doesn't independently validate the newer, separate question of what glucose patterns mean for a healthy person wearing it out of curiosity rather than managing a diagnosis, covered below.

How Accurate Is Lingo?

Abbott did not run a separate accuracy study for Lingo. Its FDA clearance rests on the clinical data of the FreeStyle Libre 2, collected in people with diabetes, on Abbott's reasoning that diabetic users are the harder test because their glucose swings more. Lingo uses the same sensor and the same glucose calculation as the Libre 2, so the Libre 2 data describes the measurement. What has much less evidence is how that measurement performs in people without diabetes, who are the people Lingo is sold to.

What Abbott's FDA filing actually says

The Lingo Glucose System was cleared by the FDA on May 29, 2024 under 510(k) number K233655, as an over-the-counter integrated continuous glucose monitor for adults 18 and over who are not on insulin. The filing states that the biosensor hardware and technology are based on the FreeStyle Libre 2 sensor, and lists the glucose calculation algorithm as the same as the Libre 2's.

On accuracy, Abbott did not submit new clinical data. It submitted a statistical analysis arguing that the Libre 2's existing clinical data could be applied to Lingo, because diabetic participants show more glucose variability and larger swings than healthy people, so the diabetic data "reflects a worst-case scenario for sensor performance." On that basis the filing states that system accuracy met the FDA's special controls for this device class.

Three differences from the Libre 2 that the filing documents

A narrower display range. Lingo shows glucose between 55 and 200 mg/dL. The Libre 2 covers 40 to 400 mg/dL. Lingo is the same sensor with its readout clipped, so it is not designed to show severe highs or lows as numbers.

No fingerstick confirmation. The Libre 2 calls for a fingerstick check when symptoms do not match the reading, when no value is shown, and during the first 12 hours of wear. The filing states Lingo does not require fingerstick testing, based on its intended use.

Not for medical decisions. The filing states the user is not intended to take medical action based on Lingo's output without consulting a qualified healthcare professional.

What published research says about CGM accuracy in people without diabetes

The worst-case argument is reasonable on its face, but it went untested on Lingo until 2026, and the limited evidence in healthy people, including that test, does not clearly confirm it. In one randomized study under simulated altitude, a FreeStyle Libre sensor's mean absolute relative difference was 21.2 percent in healthy volunteers and 11.8 percent in people with diabetes, so under that condition the sensor was less accurate in the healthy group (Diabetes Technology & Therapeutics, 2026 [conflict statement: two authors have received honoraria or other payments from Abbott]). Two of that study’s authors report honoraria or other payments from Abbott. That study did not state which Libre generation was used, and it describes one specific condition rather than everyday wear.

A separate randomized crossover trial in 15 healthy adults found a continuous glucose monitor read both fasting and after-meal glucose about 0.9 mmol/L higher than fingerstick capillary samples, overstated time spent above 7.8 mmol/L roughly fourfold, and varied in accuracy from person to person (American Journal of Clinical Nutrition, 2025). That abstract does not name the CGM model, so it is evidence about the device category rather than about Lingo.

Exercise is one condition that has been tested on an Abbott sensor, though on the FreeStyle Libre 3, the generation after the one Lingo is built on. In 9 young adults without diabetes doing an intense leg-strength session, the Libre 3 read about 8 to 18 mg/dL below fingerstick blood on average across the phases, with the largest gap after exercise, and only 50.5 to 64.3 percent of readings fell in the most accurate error-grid zone. The authors found it satisfactory for following glucose trends and short of clinical accuracy criteria (European Journal of Applied Physiology, 2024).

The nearest thing to an independent test of Lingo's own hardware used the FreeStyle Libre 2 in 44 healthy men in their mid-twenties during a glucose tolerance test. Against fingerstick blood, readings were off by 12.9 percent on average and nearly all fell inside clinically acceptable error-grid zones, but the sensor read lower than blood at every time point: 89.7 against 97.3 mg/dL fasting, and 144.3 against 159.9 mg/dL half an hour after the glucose drink (Sensors, 2024). That runs in the opposite direction from the crossover trial above, which found a CGM reading high. The two studies disagree on which way the error goes, and agree that it is large enough to matter for a single reading.

Taken together: the sensor's accuracy is documented in people with diabetes, Abbott's case for applying it to healthy users is an argument rather than a measurement, and the studies of CGMs in healthy people find individual readings can be off by a meaningful margin, in either direction. One independent study has now tested Lingo itself: in 39 adults without diabetes, Lingo differed from fingerstick blood by 12.9 percent on average, with 66.5 percent of readings within 15 percent, and agreement was worse after a sugary drink than after a placebo, 15.3 against 10.6 percent (Prakash et al., Journal of Diabetes Science and Technology, 2026). In the same study the readings ran about six minutes behind fingerstick blood, error grew when glucose was changing fast, and 13.8 percent of the trend arrows fell in the study's moderate-to-extreme risk categories. Another tested the Libre 2 sensor it is built on. KnowYourPrime's CGM desk rates CGM accuracy in people without diabetes as Uncertain.

What Is a Normal Lingo Reading?

In people without diabetes, a continuous glucose monitor spends most of the day between 70 and 140 mg/dL. In 153 healthy, non-obese volunteers aged 7 to 80, the median was 96 percent of the day in that range, with about 30 minutes a day above 140 and about 15 minutes below 70; average glucose was 98 to 99 mg/dL, and 104 in those over 60 (Journal of Clinical Endocrinology and Metabolism, 2019). In 560 adults with normal blood sugar from the Framingham Heart Study, average age 58.5, the figure was 87 percent, with roughly 3 hours a day above 140 (Journal of Clinical Endocrinology and Metabolism, 2025 [funding listed from Dexcom]). KnowYourPrime's CGM desk rates this normal range as Established.

Is going above 140 after a meal a problem?

Not on its own. In the healthy-volunteer group, meals took glucose from about 93 mg/dL to an average peak near 130, reached roughly an hour and a half after eating (Journal of Diabetes Science and Technology, 2021 [conflict statement: an author has Dexcom research grants and sits on its advisory board]). In the older Framingham group, a few hours a day above 140 was normal for people with normal blood sugar. What separated the groups was the share of the whole day in range: 87 percent with normal blood sugar, 77.1 percent with prediabetes and 46.2 percent with diabetes (Journal of Clinical Endocrinology and Metabolism, 2025 [funding listed from Dexcom]). Spikes are common in people whose blood tests are normal: across 1,521 days of everyday CGM wear in a Spanish community sample, 73 percent of people with normal fasting glucose and HbA1c had episodes above the level that defines impaired glucose tolerance, and 5 percent went above the diabetes threshold (Diabetes Research and Clinical Practice, 2018).

There is no agreed warning line for healthy people

When 18 CGM specialists independently reviewed 20 reports from people without diabetes, they agreed poorly on who needed follow-up. More than half flagged people who spent over 2 percent of the day above 180 mg/dL, even when their HbA1c and fasting glucose were normal, but there was no consensus (Journal of Diabetes Science and Technology, 2026 [conflict statement: authors have received fees or research support from Dexcom and Abbott]).

Reading these figures on a Lingo screen

The reference studies used Dexcom sensors, not the Libre 2 hardware inside Lingo. If the Libre 2 result above holds, a Lingo trace may sit slightly below these reference figures. Lingo also only displays values between 55 and 200 mg/dL, so it will not show the full range of a reading outside those limits.

What Does the Research Actually Say About Non-Diabetic CGM Use?

It depends on the question. Two things are well supported in people without diabetes: a CGM trace reflects what was eaten, and most of the day sits between 70 and 140 mg/dL. Whether wearing one changes behavior enough to improve glucose is less settled, and whether it catches prediabetes earlier than a blood test, or says anything about heart risk, is unsettled.

Glucose tracks diet: Established. In 677 adults without diabetes in the Framingham Heart Study, better overall diet quality and carbohydrate quality went with more favorable CGM glucose measures, and swapping 5 percent of energy from protein to carbohydrate went with a mean glucose about 1 mg/dL higher (American Journal of Clinical Nutrition, 2025 [funding statement: Dexcom provided the CGMs at a discounted rate]). That is the use Lingo is sold for, and it is the best-supported one.

Diet moves glucose swings more than exercise: Established. In 30 healthy women wearing the FreeStyle Libre 2, the sensor Lingo is built on, two weeks of calorie-tailored low-glycemic meals cut the time glucose spent rising or falling at 1 to 2 mg/dL a minute by 4.76 percent, while two weeks of added physical activity changed it little. The study was a small fixed-order pilot whose authors note major design limitations (Nutrients, 2026).

What normal looks like: Established. People without diabetes spend most of the day between 70 and 140 mg/dL, and short rises above 140 after meals are a normal part of the trace, which is why a single spike on its own is not a finding.

Wearing one improves glucose control: Emerging. The evidence leans toward a benefit for people without diabetes, but several of the studies it rests on do not cleanly show one.

Catching prediabetes, predicting heart risk: Uncertain. One study separated prediabetes from normal glucose on CGM metrics; another found the groups overlapped and the CGM added nothing beyond a blood test. Time in range tracked predicted cardiovascular risk, not observed heart events, in an analysis its authors call exploratory.