CGM in Running and Endurance Sports: Promising Tool or Overhyped Fuel Sensor?

Continuous glucose monitors were not designed for trail runners. They were designed for people with diabetes, for whom real-time glucose data is a clinical necessity rather than a performance curiosity. But over the last several years, CGMs have migrated aggressively into the endurance sports market, propelled by a wave of commercial interest, athlete testimonials, and a genuinely appealing idea: what if you could see your glucose in real time and use it to make better fuelling decisions mid-run?

The idea is appealing enough that it deserves honest scrutiny. The evidence on CGM in non-diabetic endurance athletes is now substantial enough to draw some clear conclusions - about what CGMs can legitimately tell you, where the data breaks down, and where the marketing has outrun the science considerably.

What a CGM Actually Measures

A CGM sensor is inserted into the subcutaneous tissue, typically on the back of the upper arm or abdomen, where it measures glucose concentration in the interstitial fluid - the fluid surrounding the body's cells - rather than in the blood directly. The device generates a reading every few minutes and transmits it to an app or receiver, creating a continuous trace of glucose over hours and days.

The distinction between interstitial glucose and blood glucose is important and frequently glossed over in consumer-facing CGM content. These two values are related but not identical. After carbohydrate intake, interstitial glucose increases with a delay of up to 15 minutes compared with concentrations in the blood, and during exercise this lag becomes more complex and less predictable. Studies on newer CGM models have found that the drop in interstitial glucose lags behind the drop in blood glucose during prolonged aerobic exercise by 12 minutes on average, with mean absolute relative difference increasing to 13% during exercise - and separate work has found lag times of anywhere from 5 to 25 minutes depending on physiological changes during exercise, alterations in blood flow rate, body temperature, and acidity.

This is not a minor calibration issue. It means that what your CGM is showing you during a hard climb or a race-pace effort is not your current glucose - it is your glucose from some minutes earlier, at an uncertain delay. A comparison of CGM against lab-based blood glucose analysis under resting and exercising conditions in athletes found that accuracy during higher intensity exercise was poorest during the earlier stages of a bout, even in newer models. The faster glucose is changing in the blood, the less accurately the interstitial reading reflects it.

Where CGMs Are Genuinely Useful for Endurance Athletes

Despite those limitations, there are contexts where CGM data is meaningfully informative for endurance runners, and it is worth being specific about what they are.

Understanding food-glucose responses in daily life. Outside of exercise, and particularly in the hours before and after training, CGM data is reasonably accurate and can reveal a great deal about how your body responds to specific foods, meal timing, and pre-session nutrition choices. The individual variability in postprandial glucose responses is substantial - research using CGM has demonstrated that individual variability of postprandial glucose responses to identical meals was as large as the variability between responses to different meals, indicating that factors beyond food composition itself drive glucose response. This means that the pre-race breakfast that keeps one runner's glucose stable through the first two hours of an event may cause a spike-and-crash pattern in another runner's trace. CGM can show you which category you fall into, and allow you to adjust meal composition and timing accordingly.

Monitoring recovery after hard efforts. CGM studies in elite endurance athletes have shown that glucose patterns during high training load - including transient excursions well above 140 mg/dL and below 70 mg/dL - are better understood as adaptive responses to training rather than signs of metabolic dysfunction. Tracking how quickly your glucose trace normalises after a hard training block can serve as a marker of recovery status that subjective feel alone cannot provide. If variability is still elevated 48 hours after a hard long run, the data is telling you something about systemic recovery.

Identifying unintentional hypoglycaemia during long training runs. For runners training at low carbohydrate intake on long efforts, CGM can flag episodes of falling glucose that the runner may not reliably perceive, since the subjective experience of early-stage hypoglycaemia is easily confused with general fatigue. This is less about race-day real-time decisions and more about understanding how your body responds to specific fuelling protocols during long training runs in controlled conditions.

Energy availability awareness. Endurance athletes, particularly those managing body composition alongside high training load, are at meaningful risk of low energy availability - consuming insufficient calories to support both training demand and basic physiological function. CGM patterns, particularly chronically suppressed resting glucose and unusual glycaemic responses to training, can be an early signal of energy availability issues that might otherwise go undetected until the athlete is already underfuelled enough to affect performance and health.

Where the Hype Significantly Outpaces the Evidence

This is where the post necessarily diverges from most CGM content aimed at athletes, because the commercial narrative and the research literature are currently pointing in meaningfully different directions.

CGM as a real-time fuelling guide. The most aggressively marketed CGM use case for endurance athletes is real-time glucose monitoring during races and long efforts, with the idea that the athlete can use their glucose trace to time gel intake, avoid bonking, and stay within an optimal "performance glucose zone." The most prominent commercial application of this concept, Supersapiens, shut down its endurance athlete-focused CGM business in early 2024, and the research that underpinned its "Glucose Performance Zone" concept has been characterised by researchers as hype until proven otherwise.

The fundamental problem is that CGM does not measure muscle glycogen concentrations or carbohydrate flux, thereby limiting its efficacy as a fuel sensor or as a guide to fuelling strategies. Muscle glycogen, not circulating blood glucose, is the primary fuel store that determines bonking in endurance events. Blood glucose - and by extension, interstitial glucose - can remain within a normal range while muscle glycogen is critically depleted, because the liver is still releasing glucose into the bloodstream. The CGM tells you what is in circulation. It does not tell you what is left in the tank.

Elite athletes showing apparently abnormal glucose patterns. CGM profiles recorded during training camps of elite cyclists have shown glucose patterns that, without additional context, can appear to resemble glucose intolerance - but these patterns are better understood as transient episodes driven by changes in exercise intensity and high carbohydrate intake rather than pathological glucose dysregulation, occurring in athletes who demonstrate normal glucose tolerance on formal testing. For recreational athletes using CGM and comparing their traces against diabetic reference ranges or non-athlete normals, this is a significant interpretive risk. Elite athletes may spend 10 to 20% of their waking day with glucose above 140 mg/dL and 5 to 7% below 70 mg/dL - ranges that would be flagged as problematic in a diabetic context but appear to be normal for athletes under high training load.

There are currently no validated athlete-specific reference ranges for CGM data that would allow a runner to meaningfully interpret whether their trace is concerning, optimal, or simply normal for their training state. Research attempting to establish athlete-specific reference ranges for glycaemic variability is ongoing, and what is considered healthy, suboptimal, or optimal for athletes without diabetes remains unclear.

CGM as a substitute for metabolic testing. Some CGM marketing positions the technology as a replacement for laboratory metabolic testing - the indirect calorimetry and substrate oxidation testing described elsewhere on this site. It is not. CGM shows you a glucose trace. It does not tell you your resting metabolic rate, your fat oxidation capacity at various intensities, your Fatmax, or your total daily energy expenditure. These remain laboratory measurements, and the insights they provide are not derivable from glucose data alone.

How to Use a CGM Intelligently

Given the above, CGM is most useful for endurance athletes as an educational and investigative tool rather than a real-time performance optimisation device. The appropriate mindset is n=1 experimentation - using a sensor for a defined period to answer specific questions about your glucose responses to food, training, and recovery, rather than wearing one indefinitely and treating every glucose fluctuation as actionable information.

Specific questions worth investigating with a CGM over a 4 to 6 week period include: How does my pre-session meal affect my glucose during the first hour of a long run? Do I show glucose instability in the days following a hard race or very long training effort, and if so, for how long? What does my overnight fasting glucose look like across a training block, and does it change at periods of high load versus recovery? Is there evidence of early-morning hypoglycaemia that might be affecting the quality of my morning training sessions?

These are questions where the CGM data is recorded in relatively stable, low-intensity conditions where the interstitial lag is minimal and accuracy is highest, and where the findings can directly change nutrition and training behaviour in a meaningful way.

What a CGM cannot tell you during exercise should be treated as a firm interpretive boundary rather than a limitation to work around. The glucose trace during a hard interval session or a race-pace long run is the least reliable data the device produces. Using it to make mid-effort fuelling decisions in real time is not currently supported by the evidence, and doing so risks both misinterpreting normal exercise-induced glucose variation and overlooking the fact that the number on screen is already several minutes out of date.

Conclusion

CGM is a legitimate and genuinely interesting tool for endurance athletes, but its value sits in specific, well-understood contexts that are quite different from the way it has been marketed to the athletic population. The data it produces outside of intense exercise - on food responses, recovery status, and energy availability - is reasonably reliable and can support meaningful n=1 experimentation. The data it produces during hard exercise is substantially less reliable, and the idea that CGM can function as a real-time fuel sensor guiding mid-race carbohydrate intake has not been supported by the research and has already resulted in the market's most prominent commercial application in this space closing its doors. Used with appropriate expectations and honest interpretation, a CGM is a useful addition to the toolkit. Treated as the glucose oracle that some marketing has suggested it might be, it will disappoint - and may actively mislead.