Key Takeaways
1. Consuming carbohydrates is not the same as using carbohydrates.
2. It takes time — up to 60 minutes — before the carbohydrates you consume in the session reach a steady share of fuel.
3. Start long runs topped-up by eating carbohydate-rich meals in the days before.
4. Fuel from the start of a long run and complete the plan to the finish.
5. Without fuel, the body would only have enough reserves to last approximately 80 minutes at target marathon pace.
6. Use the training period to experiment and refine a race day fuel strategy.
About
On 1 November, Andreas Almgren will run the New York City Marathon — his first attempt at the distance. World champion over the half marathon and a former European champion in the 10,000m, this build is more than an experiment. The fueling and training knowledge he acquires during this period will be important as he undertakes a purposeful transition into longer distance running.
In-race fueling is new for Andreas. This page is about the collaborative process that happened throughout the marathon build to develop a fuel strategy from scratch — to put fueling into practice.
A first attempt necessitates a period of getting to know what’s required of the body to meet the demands of the distance. Some training sessions have a greater influence over the learning process. Testing Andreas during his hardest sessions determined how much fuel his body actually utilizes and the implications for two hours of hard running.
His fuel plan has been built from the results of this testing.
Race
New York City Marathon
Date
Sunday 1.11.26
Distance
42.195 km
Total climb
165 m
With no elite pacing, the New York City marathon could be described as the racers’ race. It could favor the tactical awareness and ability to meet the demands of dynamic racing that have brought Andreas success on the road and track. But the rolling terrain of the course, with its bridges and long avenues, add geographical complexity to the mix.
“In New York, you get those rolling hills, and even some long hills as well, where you really need to push it and be able to handle the lactate that's probably going to accumulate.”
Andreas has split his marathon preparation block between Stockholm and a final five weeks training at altitude in Flagstaff.
Long runs 1 and 2 — 27.8.26 + 5.9.26, Stockholm, Sweden
Session overview
The first testing period involved two progressive long runs, nine days apart: easy start, steady build, hard finish. Pace and heart rate increase throughout.
Long run 1 — from 3:50 min/km at 120 bpm to under 3:00 min/km at 160 bpm.
Long run 2 — longer, faster and with no stops. From 3:35 min/km to 2:50 min/km.
The long runs are treated like a race simulation — where fueling matters. “You can't store enough carbohydrates to be able to utilize everything for the entire workout, you need extra fueling.”
Comparing the long runs is how Andreas tracks progress. “You can just compare over time how you feel when you get into that second hour. Are you able to maintain your mechanics? Are you able to fuel well? Does your heart rate spike?”
Testing — Exogenous Carbohydrate
The test is used to quantify how effectively an athlete can oxidize ingested carbohydrate during exercise. By tracing the appearance of ¹³CO₂, the test provides an estimate of the rate at which consumed (exogenous) carbohydrates are absorbed, delivered, and ultimately oxidized as a fuel source, while the body is under the metabolic demands of exercise. It is ultimately an analysis of efficiency: how well an athlete turns the carbohydrate they consume into fuel they can use, while the whole system is stressed by the demands of exertion.
Fueling is not simply about how much carbohydrate is consumed, but how much of that carbohydrate can actually be utilized during training and competition. Measured in grams per hour (g·h⁻¹), it shows whether an athlete’s fueling strategy is supporting the intended exercise demand.
Exogenous carbohydrates — fuel that comes from outside the body. Carbohydrates consumed and utilized during exercise.
Protocol
Andreas consumed Maurten Drink Mix 320 containing a naturally occurring, non-radioactive carbon-13 (¹³C) tracer. Before he drank the labeled carbohydrate, a baseline breath sample was collected to determine background ¹³C enrichment in expired CO₂.
Breath samples were collected in-session at the following intervals. Long run 1: before the start, at 5, 10, 16, 23, 30 and 37 km, and once after the finish. Long run 2: before the start, then approximately every 15 minutes, from minute 17 to the finish at minute 126.
Samples were analyzed for ¹³CO₂ enrichment, and when combined with respiratory gas exchange measurements, used to calculate the rate of exogenous carbohydrate oxidation — how much ingested fuel contributed to fueling the session.
Fueling in both runs: Target 36 g of carbohydrate (225 ml Drink Mix 320) every 15 minutes.
Long run 1
It takes time for the body to start utilizing exogenous fuel sources.
Peak utilization rate from Drink Mix 320: 1.4 g per minute (83 grams per hour), at 30 km. Total carbohydrates utilized from Drink Mix 320: 92 grams. Estimated total carbohydrate utilization: 222 grams per hour.
Run
Long run 1
Date
27.8.26
Run
Long run 1
Date
27.8.26
Run
Long run 1
Date
27.8.26
Figure 1. Heart rate and pace over distance in long run 1 (27 August 2026, 37.0 km). Lines are 30-second watch means with a 2-minute moving average. Markers are the mean over ±0.5 km at every 5 km. The dashed line marks a 2:53 min pause at 9.5 km.
Long run 2
A similar pattern. After 17 minutes, Drink Mix 320 represented 2% of the carbohydrate Andreas utilized. At 34 minutes it rose to 10%. From 52 minutes it was 27%.
Peak utilization rate from Drink Mix 320: 1.45 g per minute (87 grams per hour), at 96 minutes.
Total carbohydrates utilized from Drink Mix 320: 110 grams.
Estimated total carbohydrate utilization: 238 grams per hour.
Run
Long run 2
Date
5.9.26
Distance
40.3 km
Time
2:05:25
Avg pace
3:07 /km
Avg heart rate
146 bpm
Figure 2. Heart rate and pace over distance in long run 2 (5 September 2026, 40.3 km, no pauses). Lines are 30-second watch means with a 2-minute moving average. Markers are the mean over ±0.5 km at every 5 km.
At its peak, Drink Mix 320 accounted for approximately a third of his carbohydrate utilization. Most of the work is fueled by what he stores before the start, so loading glycogen before a long run matters.
Two runs made the numbers robust enough to build a race fueling strategy. Allowing for variation in location, route profile and weather, the race fueling model uses the average of these two runs.
“I really want to make sure that I'm not going to be in a calorie deficit.”
Two runs — same conclusion. It’s not about how much carbohydrate is consumed. It’s about analyzing how much of Andreas’ ingested carbohydrates are ultimately utilized as fuel in the session. Simply, drinking is not the same as using. It was approximately one hour of running before the carbohydrates from Drink Mix 320 reached a steady share of his fuel. The fuel has to be consumed early — long before any need is felt.
Exogenous fuel supports endogenous fuel sources, enabling longer availability of glycogen.
Session 2 — Double threshold
Session overview
Morning
50 min at LT1 · 15.7 km
Morning
50 min at LT1 · 15.7 km
Two threshold sessions — one day. Sessions were separated by several hours of recovery. The same breath bag testing method as the long runs.
Double threshold purpose: accumulate a large volume of high-quality training at or around the lactate threshold, without creating the fatigue associated with a single, intense workout. Maximize the amount of useful physiological work Andreas can perform, while managing the cost of recovering from it.
Morning: 50 minutes at LT1. 15.7 km. A controlled session in which running feels steady and manageable. Testing protocol: six breath bags — before the warm-up, after the warm-up, 15, 30, and 45 minutes into LT1, and after the cool-down.
Evening: 10 × 1000 m and 10 × 400 m. 16.1 km. Testing protocol: five breath bags — before the warm-up, before the start, after 5 × 1000 m, after 10 × 1000 m, and after the session.
Session 3 — Long run 3
Session overview
The third long run.
Location: Flagstaff.
Altitude: 2,100 meters.
The same distance as long run 2, with a higher average heart rate: 154 bpm against 146.
“Your legs and body can feel quite fine, but your breathing and your pulse — your heart rate just spikes instantly.”
Fueling protocol: 36 g of carbohydrate (225 ml Drink Mix 320) every 15 minutes, as in long runs 1 and 2. [Confirm]
Testing protocol: breath bags before the start and at 4.7, 10, 14.9, 20, 25, 30, 34.2 and 40 km.
Results [¹³C results pending]
The breath samples from this run are still being analyzed. This section will show if altitude impacts the share of fuel from Drink Mix 320 (exogenous), the peak utilization rate, and the total utilized fuel — the same measures as long runs 1 and 2.
Estimated total carbohydrate utilization: 274 grams per hour. This is estimated from his heart rate and lab profile, not from ¹³C. For comparison: 222 grams per hour in long run 1 and 238 grams per hour in long run 2. Using Andreas’ physiological data it’s also possible to estimate that in the final stages of each long run, 93% of energy is from carbohydrate. In these closing moments, approximately 5.6 g/minute of total carbohydrate are utilized and 0.2 g/minute of fat.
Using Andreas’ physiological data it’s also possible to estimate that in the final stages of each long run, 93% of energy is from carbohydrate. In these closing moments, approximately 5.6 g/minute of total carbohydrate are utilized and 0.2 g/minute of fat.
Distance
40.0 km
Moving time
2:09:58
Distance
40.0 km
Moving time
2:09:58
Figure 3. Heart rate and pace over distance in long run 3 (6 October 2026, 40.0 km). Lines are 30-second watch means with a 2-minute moving average. Markers are the mean over ±0.5 km at every 5 km. Two pauses: 2:24 min at 5.3 km and 9:53 min at 8.0 km.
Key takeaways
The race model is built on long runs 1 and 2. Long run 3 shows whether those numbers hold with the additional demands of altitude.
Race fueling plan — New York City Marathon
The plan
One bottle every 5 km — the same frequency as in training, aligned to the on-course fuel stations.
The first limitation to overcome was a simple logistical issue. All in-race fueling bottles are positioned on the right side of the course. Andreas is left-handed, so he trained with tables on the track, to practice grabbing bottles at race pace.
2 min before the start
5, 10, 15, 20, 25, 30, 35 and 40 km
15 km
Testing the plan
Before race day, the fuel plan was simulated through a predictive model of his potential race day performance. This involved combining lab profile data, the exogenous carbohydrate oxidation results from long runs 1 and 2, and the New York course profile. It tracks the carbohydrate left in his muscles and liver, minute by minute, building a picture of fuel availability for the duration of the race. It doesn’t make any consideration of physiological durability.
The model shows that supply meets demand all the way to the finish — his muscle stores run low only in the final minutes. A second scenario — simulated by the model without exogenous fuel — shows his blood sugar dropping too low to sustain pace and performance after 80 minutes.
Bicarb
Andreas uses the Maurten Bicarb System. uses bicarb on his long runs and will use it in New York, for the hills: “You’ve got to be able to handle the lactate that's probably going to accumulate during those runs. I've used it for both my half marathons, and I felt I got a very good effect from it.”
During intense exercise, rapid ATP turnover is associated with increased production and accumulation of hydrogen ions (H+), contributing to a reduction in muscle pH. Sodium bicarbonate supplementation increases extracellular bicarbonate concentration and buffering capacity, helping facilitate the movement of H+ out of the working muscles.
Although sodium bicarbonate is not typically considered a primary ergogenic aid for marathon running, there is evidence that it can improve exercise capacity and performance during higher-intensity efforts. Some studies have also reported reductions in perceived exertion at a given workload. For a marathon runner, it may provide capacity to respond to the race when required — surges, hills, changes in pace, or tactical demands — with less disruption to their overall physiological state. Evidence for a direct performance benefit during continuous running is less consistent, so this should be viewed as a plausible race-specific application, rather than a marathon effect.
Strategy
Start topped up. Fuel from the start line — not from halfway — acknowledging, in line with the ¹³C exogenous carbohydrate oxidation test, it takes approximately 60 minutes for the Drink Mix to reach a steady share of the fuel supply.
Keep the rhythm: one bottle every 5 km. Caffeine at 15 km.