The 36-Hour Marathon Carb Load for 10 to 12 Grams Per Kilogram

Traditional carb-loading used hard glycogen-depleting exercise, several days of low-carbohydrate eating, then several days of high-carbohydrate eating. That model proved that muscle glycogen could supercompensate, but later work showed that runners can load effectively without the low-carbohydrate phase. The practical question is whether you can eat enough carbohydrate without creating gastrointestinal problems before the race.

The current evidence supports a narrower protocol. Bussau et al. showed that trained men reached maximal measured muscle glycogen after 1 day at 10 g/kg/day while physically inactive, and the 2016 nutrition and athletic performance position statement recommends 10 to 12 g/kg/day for 36 to 48 hours before events lasting more than 90 minutes.¹² What matters is the dose, the foods you choose, and the absence of a hard depletion run close to race day.

01Why modern marathon carb loading takes 36 hours

Bergstrom, Hermansen, Hultman, and Saltin published the original supercompensation paper in 1967.⁷ They studied 9 healthy subjects, varied diet after hard exercise depletion, and measured muscle glycogen from quadriceps biopsies. Glycogen ranged from 0.6 to 4.7 g per 100 g wet muscle, and work to exhaustion at 75% VO2max averaged 59, 126, and 189 minutes after the protein, mixed, and carbohydrate diets.

The follow-up evidence narrowed the window in two important steps.

Sherman, Costill, Fink, and Miller ran the cleaner comparison in 1981.⁸ They tested three regimens in trained runners: 15% carbohydrate for 3 days then 70% carbohydrate for 3 days, 50% carbohydrate for 3 days then 70% carbohydrate for 3 days, or 6 days at 50% carbohydrate. All groups used a 5-day exercise taper. Muscle glycogen at the end of loading was 207, 203, and 159 mmol glucosyl units/kg wet tissue. The two higher-carbohydrate finishes loaded better than the control condition without requiring several low-carbohydrate days.

Bussau, Fairchild, Rao, Steele, and Fournier closed the gap further in 2002.¹ They studied 8 endurance-trained male athletes who ate 10 g/kg/day of high-carbohydrate, high-glycemic-index foods for 3 days while physically inactive. Muscle glycogen increased from 95 ± 5 to 180 ± 15 mmol/kg wet mass after 1 day and did not increase further after 3 days. That supports short loading in trained men, while the 36- to 48-hour version fits the later position-statement range and leaves a practical buffer for travel, missed meals, and under-eaten snacks.

None of these protocols is magic. The carbohydrate intake is what fills the tank. The separate low-carbohydrate depletion phase is not necessary for most marathoners. If you can hold 10 to 12 g/kg/day for 36 to 48 hours, you are inside the modern race-week target.

02Calculate your carb-loading target

A 70 kg runner targeting 10 g/kg/day needs 700 g of carbohydrate. A 60 kg runner needs 600 g. An 85 kg runner needs 850 g. Those numbers feel abstract until you start counting them on real food. A medium bagel is 50 g of carbohydrate. A cup of cooked white rice is 45 g. A medium banana is 27 g. A 20-ounce sports drink is 35 to 40 g. To reach 700 g you need roughly fifteen of those food units in a day, on top of the protein and fat you should still be eating.

The trap is that most runners stay on the same training-day food template (high-volume, mixed macros, vegetable-heavy) and try to brute-force the carbohydrate target on top. Their plates are too big, their guts revolt, and they end up at 450 g of carbohydrate, 1,000 extra kcal from fat and oversized portions, and the worst sleep of the week.

The actual move is to do three things together for the final 36 hours.

1. Choose lower-FODMAP carbohydrate sources if high-FODMAP foods have triggered GI symptoms. A 48-hour crossover trial in endurance athletes found lower pre-exercise total gastrointestinal symptoms and less gut discomfort and upper-GI symptoms during exercise on a high-carbohydrate low-FODMAP diet versus a high-FODMAP version, with no performance difference and higher recovery gastrointestinal symptoms on the low-FODMAP arm.⁹ Lis et al. (2016) reported a similar symptom-reduction pattern in a low-FODMAP case study.¹⁰ This evidence supports changing FODMAP content for symptom-prone runners, not a blanket reduction in total fiber.
2. Cut fat to make room for carbohydrate. A creamy pasta sauce can double the kcal of a plate without adding glycogen. Pair refined starch with lean protein (chicken breast, tofu, eggs, fish) and limit oil. Keep the low-fat version contained to the final 36 hours.
3. Use drinks for the last 300 g. Once solid food has filled half the target, add liquid carbohydrate. A 32-ounce sports drink is 55 to 65 g. A maltodextrin powder mixed into juice is 60 to 90 g per serving. Drinkable carbohydrate avoids the volume ceiling that stops most attempts at 600 g.

03Follow a 36-hour carb-loading plan

Start the 36-hour window around noon two days before the race (for example, Friday for a Sunday marathon) and finish at lights-out the night before race morning. The race-morning meal is separate and uses race-morning fueling targets.

The table below shows a working 36-hour template for three body weights. Carbohydrate columns are in grams. Target 5 g/kg over the first 12 hours. Then target 11 to 12 g/kg over the final 24 hours.

The food itself does not have to be exotic. A breakfast of one large white bagel with jam and honey, a glass of orange juice, and a sliced banana is roughly 130 g of carbohydrate. A lunch of three cups of plain white rice with a chicken breast and a salt-heavy broth is 130 g. A dinner of two cups of plain pasta with a simple tomato sauce, a side of white bread, and a sherbet for dessert lands near 170 g. The remaining 270 g over the day comes from a sports drink between lunch and dinner, a granola or rice-based bar mid-afternoon, and a small bowl of dry cereal or rice pudding before bed.

04Use drinks to reach your carb target

The single most common failure pattern is a runner who can hit 500 g of solid carbohydrate and then physically cannot eat more. The fix is to plan in 250 g of liquid carbohydrate across the day from the start. Select a 32-ounce sports drink labeled 60 g of carbohydrate and drink two bottles (120 g total). Mix 90 g maltodextrin (90 g carbohydrate) with enough fruit juice to provide 40 g carbohydrate (use the label to determine the volume), then dilute to at least 1 L total volume; that mixture supplies another 130 g. The two bottles and the maltodextrin-and-juice mixture total 250 g of liquid carbohydrate. The two-decade trend in elite marathon and triathlon prep has been to push more of the load into liquid form because the gut tolerates it better than another plate of pasta. The same intra-race math from the 90 to 120 g/h fueling protocol applies in reverse on the pre-race side. Concentrated drinks are an efficient delivery system.

Athletes who train gut tolerance for race-day fueling usually find the pre-race load easier as a side effect. The same intestinal capacity that handles 90 g/h on the run also handles 600 to 800 g over a day at rest.

05Carb-load effectively as a woman

James et al. (2001) showed no sex difference when men and women were fed 12 g carbohydrate/kg lean body mass/day for 3 days with training cessation.³ Walker et al. (2000) showed that well-trained women increased preexercise glycogen from 625.2 to 709.0 mmol/kg dry muscle and extended cycling time to exhaustion from 106:35 to 115:31 on a high-carbohydrate diet, but the study was women-only and reported a smaller glycogen gain than previously observed in men.⁴ Tarnopolsky et al. (1995) showed that a 4-day rise from 55 to 60% carbohydrate to 75% of energy intake increased glycogen and performance in men. It did not increase glycogen or performance in women in that study design.⁵

The practical reading is that female endurance athletes should still plan around the same race-week carbohydrate range, while recognizing that studies differ depending on whether carbohydrate is prescribed as percent of energy, body mass, lean body mass, or absolute intake. The risk profile to watch is low energy availability in female endurance athletes, where chronic under-fueling makes the acute load both more necessary and harder to execute.

Menstrual cycle phase has a smaller effect than habitual fueling. The luteal phase shifts substrate utilization slightly toward fat at submaximal intensities, but the loading response itself is preserved across phases when carbohydrate intake is adequate.⁵

06Expect water-weight gain during carb loading

Body weight often rises during carb loading because glycogen storage increases total body water. Olsson and Saltin (1970) reported that carbohydrate enrichment increased body weight by 2.4 kg and total body water by 2.2 L.⁶ Costill et al. (1981) showed that higher carbohydrate intake after strenuous running increased muscle glycogen resynthesis.¹¹ A higher race-morning scale number can reflect stored glycogen and associated water rather than fat gain.

The most reliable own-goal of race week is the runner who steps on the scale on Sunday morning, panics at 2 kg above Friday, decides to "play it safe" with breakfast, and starts the race with a partially empty fuel tank. Do not weigh yourself in the final 48 hours. The number tells you nothing useful about fitness or fat.

07Avoid these marathon carb-loading mistakes

The classic errors are well-defined and almost always preventable.

• The "depletion" long run on Tuesday. Bussau et al. (2002) and the 2016 position statement on nutrition and athletic performance support loading without a low-carbohydrate depletion phase, while Sherman et al. (1981) showed the low-carbohydrate phase was not required even though all groups still completed a 5-day depletion-taper exercise sequence.¹²⁸ Replace the hard depletion run with a normal taper run at an easy intensity and call the carb load on Friday morning.
• A bulky training-day template through Saturday. Whole-wheat pasta, bran cereal, and big salads can make the load physically bulky. Switch to familiar carbohydrate sources for 36 hours; if high-FODMAP foods have triggered symptoms, choose lower-FODMAP alternatives. This may reduce GI symptoms but will not necessarily eliminate them.
• Pasta sauce that doubles the fat. Cream sauces, sausage ragu, and oil-heavy pestos crowd out the carbohydrate target. A simple marinara, a tablespoon of olive oil, lean protein, and salt is the better template.
• Alcohol with the load. Burke et al. (2003) used 1.5 g/kg alcohol after glycogen-depleting exercise and found lower glycogen storage when alcohol displaced carbohydrate. The authors concluded the direct effect of alcohol on glycogen synthesis was unclear and the main practical problem was replacing carbohydrate with alcohol.¹² Save it for after the race.
• Novel foods. A new restaurant, a new gel flavor, a new bar, or a new protein source in the final 48 hours is the wrong test environment. Use foods you have already eaten through training. The race-week nutrition plan is the broader frame for this discipline.
• Dropping protein to make room. Protein still belongs in the load. Aim for 1.6 to 2.0 g/kg/day even during the carbohydrate push. The carbohydrate expansion comes from removing fat and reducing meal bulk, not from cutting protein. See protein distribution for the per-meal targets.

08Plan your race-morning meal

The final piece of the load is the race-morning meal. The 2016 position statement targets 1 to 4 g/kg of carbohydrate 1 to 4 hours before the start.² For a marathon, the high end of that range can work when rehearsed. A 4 g/kg breakfast eaten 3 to 4 hours before the gun gives a 70 kg runner 280 g of additional carbohydrate at the right time, on top of an already loaded muscle. A bagel with jam and honey, a banana, a glass of juice, and a sports drink with a bottle of water sipped through warm-up is a practical default for runners who tolerate that volume.

That breakfast lives inside the broader race-week structure covered in How to Set Up a Race-Week Nutrition Plan, and the start-line sodium and fluid math belongs in sodium loading for endurance racing. The 36- to 48-hour load works best when the plan is simple: no depletion run, familiar foods, symptom-guided lower-FODMAP choices, enough liquid carbohydrate, expected water-weight gain, and a rehearsed breakfast.