Glycaemic index was created to help diabetics manage postprandial glucose. Somewhere along the way, endurance athletes adopted it as a performance tool — and applied it to contexts for which it was never designed.
The glycaemic index (GI), developed by Jenkins and Wolever at the University of Toronto in 1981, ranks carbohydrate foods by the area under the 2-hour blood glucose curve following consumption, expressed as a percentage of a reference food (glucose = 100 or white bread = 100). It was validated in clinical nutrition for type 2 diabetes management. Its relevance to trained athletes — who have dramatically different insulin sensitivity, glycogen turnover and carbohydrate oxidation rates — has been largely assumed rather than empirically established.
---
The fundamental limitation is that GI describes a single food, consumed in isolation, in a fasted state. Athletes eat mixed meals. Wolever and colleagues (1991) demonstrated that the addition of fat, protein or fibre to a high-GI carbohydrate food dramatically reduces the glycaemic response — by 20–50% in some combinations. This means that a meal's composite GI cannot be reliably predicted from the GI values of its components.
More practically useful is glycaemic load (GL), which accounts for the actual carbohydrate content of a serving: GL = (GI × grams of available carbohydrate) ÷ 100. A food can have a high GI but a low GL if the serving size is small. Watermelon is the canonical example: GI of 76 (high), but GL of only 4 per 120g serving — because 90% of watermelon is water, delivering only 6g of carbohydrate per portion. Avoiding watermelon because of its GI would be an application of the index to a context for which it was never validated.
---
Where does GI genuinely matter in sport? Thomas and colleagues (1991, International Journal of Sports Nutrition) showed that a low-GI pre-exercise meal (3 hours before) resulted in higher fat oxidation rates during 2-hour cycling and greater endurance time to exhaustion compared to a high-GI equivalent — attributed to lower pre-exercise insulin, which permitted greater fat mobilisation during exercise. This effect was confirmed by Wee and colleagues (2005) in a systematic review of six controlled trials, with the proviso that it applies primarily to steady-state aerobic exercise at moderate intensity, not to high-intensity or interval-based sessions where carbohydrate is the dominant fuel regardless of pre-exercise insulin state.
Post-exercise, the GI calculus reverses. Burke and colleagues (1993, Journal of Applied Physiology) demonstrated that high-GI carbohydrate post-exercise produced significantly greater glycogen resynthesis at 6 and 24 hours compared to low-GI equivalents — because the insulin spike accelerates glucose uptake into muscle via GLUT4 translocation and glycogen synthase activation. Waiting for slow-release carbohydrate when glycogen repletion is the objective is counterproductive.
---
The training-day structure that emerges from this evidence is practical:
- Pre-exercise (3+ hours out): Low-GI carbohydrate to stabilise blood glucose and preserve fat mobilisation for steady-state efforts.
- Within 60 minutes pre-exercise: GI becomes less relevant because insulin will be suppressed by catecholamines during the session regardless.
- During exercise: Rapidly absorbable, typically high-GI carbohydrate to avoid gastric delay.
- Post-exercise (0–60 minutes): High-GI carbohydrate at 1–1.2 g/kg/hour for glycogen resynthesis, with protein co-ingestion to amplify insulin-mediated uptake.
---
Macro planning across training and rest days involves a more nuanced carbohydrate strategy than GI alone can provide. For athletes managing carbohydrate distribution across different session types, intensities and recovery periods, the tool at winsport.uk/tools/nutrition/macro-meal-generator calculates meal-by-meal macro targets based on your training schedule, body weight and goals — integrating timing logic that goes beyond single-food GI rankings.
Have you been applying glycaemic index as a pre-exercise or post-exercise tool — and does this change how you'll structure carbohydrate timing?