Endurance Fueling with T1D: Part 4 — Monitoring Glucose Levels during Training and Races

For endurance athletes with type 1 diabetes, fueling is never just about carbohydrates. It’s about knowing where your blood glucose is, where it’s going, and how fast it’s changing — all while juggling training or racing.

Semi-pro T1D triathlete Riccardo Raso explains how he monitors glucose in training and racing, why he values trend data more than single numbers, and how he uses CGM alerts as a support tool — not a distraction.

The Full Series: 

1. The T1D Athlete Mindset 
2. Carb Loading with T1D 
3. Triathlon Fuel & Hydration Plan
4. Glucose Monitoring during Training and Races (this article)
5. Fuel Adjustments Based on Glucose Levels

The Reluctance to Use Sensors

I was initially very reluctant to use external sensors to monitor my glucose. At the time, it made me feel like a hybrid humanoid robot, and in 2015 people were still seen as “strange” if they had devices attached to their body. On top of that, medical-grade adhesives were not yet reliable enough to allow heavy sweating or swimming without constant concerns.

Those circumstances pushed me to develop a skill that, even today — despite knowing many people with type 1 diabetes — I consider fairly uncommon: the ability to perceive my blood glucose level with an error of roughly 4-6 mg/dL compared to capillary measurements, without sensors or devices. This premise is important because it allows me to make clearer, calmer decisions under race pressure.

Moving to CGM — Without Losing Awareness

That said, today I do use a Dexcom G7, which is widely regarded as one of the most accurate CGM systems currently available in diabetes management.

What matters most to me is not the absolute number at a single moment, but the direction and speed of change — whether glucose is stable, drifting down, or rising too quickly.

This helps me anticipate issues before they become performance-limiting, and my ability to internally perceive glucose trends with a very small margin of error allows me to make more targeted decisions.

During races, I deliberately keep monitoring as simple as possible and avoid overreacting to short-term fluctuations.

Practical Limits of Glucose Monitoring

On a practical and technical level, there are some limitations related to integrations and APIs. In practice, viewing Dexcom G7 data on Garmin watches or Edge bike computers requires the Dexcom G7 app to be running on a compatible smartphone, and the phone must remain within Bluetooth range (the phone acts as the authorized bridge).

An active mobile data connection is not required for the Bluetooth bridging itself, although internet connectivity may be used for cloud features depending on setup.

Direct smartwatch connectivity is currently available only with Apple Watch. However, it lacks some of the advanced features typically required for high-level endurance racing compared to dedicated sports watches.

Glucose Monitoring Without a Phone during Races

Because I generally don’t want to carry a phone during races, my solution is to use an external Dexcom G7 Receiver and keep it easily accessible.

• On the bike: mounted
• On the run: I move it into a race belt/pouch setup.

The pouch also carries insulin and, on the external hooks, two Nduranz Nrgy Gels 45 (45 g CHO each) as emergency carbohydrates. 

This backup approach is also useful for non-T1D athletes, because they can accidentally drop or lose their nutrition during the bike or run leg, so having a contingency plan always makes sense.

Using CGM Alerts in Training and Racing

I use Dexcom G7 alarms and predictive alerts. Alerts are configured as a decision-support system (early signal + trend), not as something that forces constant “micro-corrections”.

During the run, I rely heavily on perception and race feel. If the receiver sounds, it doesn’t necessarily mean I’m “in trouble.”

Often these alerts indicate that the trend is shifting, and that I should make a decision earlier than I might have otherwise — such as adding a small carbohydrate adjustment or applying my pre-planned insulin strategy.

Dexcom G7 gives you:

• Threshold alerts (High/Low)
• Predictive alerts
• Rate-of-change alerts 

The most important for endurance contexts is the predictive alert:

• Urgent Low Soon (predictive): Dexcom can warn you ~20 minutes before it predicts you may reach 3.1 mmol/L (55 mg/dL), so you have time to act early rather than react late.

This is separate from your “Low” threshold alert, and it’s the one that matters most for safety in endurance contexts.

• Rising Fast/Falling Fast (rate-of-change): optional alerts that trigger when glucose is changing quickly — useful because in training/races the direction and speed of change can matter more than the single absolute number. 

My Two Alert Profiles

Dexcom G7 also allows two alert profiles, which is what enables me to run a “daily-life” setup vs a “training/race” setup. 

1) Daily-life profile (tighter): My objective in normal life is to emulate a non-diabetic pattern as closely as practical. I’m comfortable sitting low-normal and aiming for relatively tight post-meal behavior. 

2) Training/Race Profile (wider, earlier warnings): During sport I intentionally shift to a safer operating band: • Low alert: 130 mg/dL • High alert: 190 mg/dL

This is not because 130 is “low” in clinical terms — it’s an early warning threshold that gives me time to make small, controlled adjustments before I’m anywhere near a true hypo risk. Likewise, 190 mg/dL is not an automatic “panic number”; it’s a decision-point that tells me I should consider my pre-agreed correction strategy if the level is sustained.

One precision point (to avoid inconsistencies): the Dexcom Urgent Low Soon prediction is tied to a low endpoint around 55 mg/dL within ~20 minutes, not 65 mg/dL. My 130 mg/dL low alert is my personal, higher “early warning” threshold for sport; the predictive safety alert is a separate layer.

How I Use CGM Alerts in Training

In training, the setup is straightforward:

• The phone/app remains in range so it can act as the Bluetooth “bridge” to Garmin devices
• But I silence phone alerts to avoid duplicates and distractions
• Alerts appear on my Garmin equipment 

For shorter, easy sessions (e.g., an easy run <60 minutes), I often train without routine carbs (or I carry only a minimal emergency option). The goal is to behave as close as possible to a non-diabetic endurance athlete: stable glucose, minimal “over-management”, and preserving normal metabolic signaling where appropriate.

For long or demanding sessions, I keep the same logic but with higher execution discipline: alerts are there to prevent me from discovering a problem too late.

How I Use CGM Alerts in Races

In races, the physiology is the same, but the logistics change: devices are harder to access and distractions are costly. So the priority becomes:

•  Keep alerts meaningful, not noisy
• Use vibration or quiet modes when possible
• Treat alerts as cues to act earlier, not to panic 

Net effect: the system helps me do what I’m trying to do everywhere in this whole approach — stay predictable, stay safe, stay executable.

Example T1D Ironman CGM Data

Example 1: Ironman 140.6, Cozumel 

Timing/Context:

• Race Start: ~06:00 
• CGM View: ~12-hour window from pre-swim through post-finish
• Environmental Context: Hot and humid conditions throughout the race
• Official Race Splits:
  • Total ≈ 08:36:51
  • Swim: 00:52:29
  • T1: ~00:02:30
  • Bike: 04:30:40
  • T2: ~00:02:40
  • Run: 03:08:32

CGM interpretation aligned to race phases:

• Pre-Swim/Swim (06:00 → ~06:53): Glucose starts in a controlled, deliberately moderate range and remains stable during the swim, consistent with a short swim duration, full glycogen availability, and no need for external carbohydrate intake.
• T1 + Early bike (~06:53 → ~07:30): A mild, progressive rise in glucose appears during transition and early bike, aligning with catecholamine response, transition stress, and the initiation of structured bike fueling.
• Main Bike Segment (~07:30 → ~11:30): Glucose sits within a broad, stable working band. This corresponds to the phase with the highest and most regular fluid and carbohydrate intake, balanced by high sweat rates and sustained metabolic demand in hot, humid conditions.
• T2 + Early run (~11:30 → ~12:15): The curve begins a gentle downward trend as mechanical load increases and fuelling becomes more constrained compared to the bike.
• Main Run Segment (~12:15 → ~14:45): A gradual, predictable decline continues toward the lower end of the target range, reflecting accumulated fatigue, rising core temperature, and reduced intake. No abrupt drops or rebounds are visible.
• Post-Finish/Now: Glucose stabilises around ~88 mg/dL with a flat trend arrow, indicating controlled and safe glycaemic management at the end of a long, thermally demanding event.

Example 2: Half Ironman

Timing/Context: 

• Race Start: ~09:00 
• CGM View: ~6-hour window 
• Total Time: ≈ 04:12

Relative to the full Ironman, the timeline is compressed and race intensity is higher for the duration.

At the start, glucose is again stable and controlled. Through the swim and early bike, glucose rises more quickly than in the full-distance race, consistent with higher relative intensity and earlier fueling.

The mid-section shows a tighter, more compact working band, which aligns with a more aggressive execution profile in a shorter event. During the run, glucose trends downward over a much shorter window than in the full Ironman. With less cumulative thermal load and fatigue, the decline is typically shallower and resolves rapidly as the event concludes.

Taken together, the two traces reflect the same control logic applied to different formats: long-duration stability under heat exposure in a full Ironman versus tighter, higher-intensity control across a shorter half-distance race.

Up next: What happens when glucose levels are trending up or down? In the next part of the series, Riccardo walks us through how he makes adjustments to his fueling plan during races. Read it here.