59.57 versus 61.37 minutes: how to use a carb mouth rinse

I carry gels for a 100 km. For a 10 km or a 40-minute tempo there is nothing to swallow, and the fluid would sit in my stomach for no return. In 2004 a group of trained cyclists finished a 1-hour time trial in 59.57 minutes after rinsing a carbohydrate solution and spitting every mouthful out, against 61.37 minutes on a placebo rinse ¹. The meta-analyses published since have shrunk that gap without erasing it.

What the research shows.

Carter, Jeukendrup and Jones ran the trial that opened the field ¹. Cyclists completed the same 1-hour time trial twice, rinsing a 6.4% maltodextrin solution or a taste-matched placebo for about 5 seconds at each 12.5% of the distance, then spitting. Finish time: 59.57 ± 1.50 minutes against 61.37 ± 1.56 (p = 0.011). Mean power: 259 ± 16 W against 252 ± 16 W (p = 0.003). Heart rate and rating of perceived exertion matched across trials. Since nothing was swallowed, no metabolic account of the gain survives.

Chambers, Bridge and Jones traced where the signal goes ². Eight endurance-trained cyclists rinsing 6.4% glucose finished a time trial in about 60.4 minutes against 61.6 on a saccharin placebo (p = 0.007). A separate group of eight rinsing 6.4% maltodextrin finished in about 62.6 against 64.6 (p = 0.012). Maltodextrin tastes of almost nothing, which removes sweetness as the channel. Their fMRI arm showed glucose in the mouth activating the anterior cingulate cortex and the striatum, reward regions that saccharin left quiet. Receptors in the mouth report incoming fuel, and the brain loosens its pacing before a gram reaches the gut.

Running carries less data than cycling. Rollo, Williams, Gant and Nute had 10 endurance-trained men run 30 minutes at a self-selected speed matching a perceived exertion of 15, rinsing either a 6% carbohydrate solution or a taste-matched placebo ³. Total distance covered was greater on the carbohydrate trial (p < 0.05). The brief was to hold a feeling rather than to race, so the result reads as free speed at a fixed effort.

The pooled effect is a fraction of the first trials.

Brietzke and colleagues pooled 13 studies and 16 cycling time-trial arms ⁴. Mean power output rose, with a standardised mean difference of 0.25 (95% CI 0.04 to 0.46). Time to complete the trial showed no advantage over placebo. Riders push harder on the rinse, and the clock does not always register it.

Deng and six co-authors ran a three-level meta-analysis over 35 articles ⁵. Pooled effect on exercise performance: Hedges' g = 0.18 (95% CI 0.09 to 0.28, p < 0.01). They graded the certainty of that evidence as low.

Zalagkitis and colleagues went wider, 90 controlled crossover trials ⁶. Workload improved by a small margin (Hedges' g = 0.25, 95% CI 0.08 to 0.41, p < 0.01), as did distance covered, performance time and speed. Muscular endurance, maximum power output and VO2max moved nowhere. Men made up 96.45% of the participants, at an average age near 25.

Three independent syntheses put the effect in the same band: small, dependent on which outcome you measure, and drawn almost entirely from young men. Plan around a fraction of a percent on the clock, and treat the 1.8 minutes of 2004 as the upper edge of what one well-run trial found.

The fasted rule flipped.

Early reviews told athletes to rinse on an empty stomach, on the logic that a brain already sensing fuel has nothing to learn from a mouthful. Lane and colleagues tested it directly ⁷. Twelve competitive male cyclists each rode four time trials: two starting 2 hours after a meal carrying 2.5 g of carbohydrate per kilo of body mass, two after an overnight fast, each paired with a 10-second rinse of 10% maltodextrin or a placebo. Pre-ride state and rinse both changed power output (p < 0.01 each), and the two interacted (p < 0.05). The rinse did more for fasted riders. The fastest condition overall was fed riders who also rinsed.

Deng's moderator analysis points the same way from a much larger base ⁵. Effects ran larger during aerobic exercise, with maltodextrin solutions, with rinses of 10 seconds or less, and after a high-carbohydrate meal. Skipping your pre-race breakfast to sharpen the rinse would trade a large certain loss against a small uncertain gain, so eat the meal and rinse on top of it.

Li, Liu, Miao, Zhang and Lu pooled the trials of carbohydrate, caffeine and menthol rinses into one network, covering 609 participants ⁸. Carbohydrate lowered ratings of perceived exertion against placebo, the only one of the three to reach significance in that pairwise comparison. A rinse that makes 4:30 per kilometre feel like 4:40 is useful at kilometre 8 of a 10 km and dangerous at kilometre 2.

The protocol.

1. Reserve it for hard efforts of roughly 30 to 75 minutes. Below 30 minutes there is little pacing left to influence; above 75 you should be swallowing 60 to 90 g of carbohydrate an hour rather than spitting it out.
2. Mix maltodextrin at 6% to 10%, so 6 to 10 g in 100 ml of water. That range covers the solutions used in the trials with the clearest effects ¹²⁷. Plain glucose also works ². Non-caloric sweeteners do not ².
3. Rinse about 25 ml for 5 to 10 seconds, then spit. Deng found larger effects at 10 seconds or less ⁵, so a long swill buys nothing.
4. Repeat every 7 to 8 minutes of hard effort. Carter's riders rinsed at each 12.5% of a 1-hour trial, eight times in all ¹.
5. Keep your pre-race meal. The fed-plus-rinse condition beat fasted-plus-rinse on power ⁷, and pooled moderators favour rinsing after a high-carbohydrate meal ⁵.
6. Hold your planned pace by watch or by split through the first third of the effort. Perceived exertion drops on the rinse ⁸, which makes an over-fast opening kilometre easy to run and expensive to pay back.
7. Rehearse it on three hard sessions before you race it. Most of this literature was built on stationary bikes, and spitting mid-pack while holding threshold pace is a skill with its own learning curve.
8. Size the expectation to the evidence: g = 0.18 ⁵. On a 40-minute tempo that is a handful of seconds. Worth it when you already carry a bottle, not worth a hand on a technical trail descent.

Sources.

1. Carter, J. M., Jeukendrup, A. E., & Jones, D. A. (2004). The effect of carbohydrate mouth rinse on 1-h cycle time trial performance. Medicine & Science in Sports & Exercise, 36(12), 2107-2111. doi
2. Chambers, E. S., Bridge, M. W., & Jones, D. A. (2009). Carbohydrate sensing in the human mouth: effects on exercise performance and brain activity. The Journal of Physiology, 587(8), 1779-1794. doi
3. Rollo, I., Williams, C., Gant, N., & Nute, M. (2008). The influence of carbohydrate mouth rinse on self-selected speeds during a 30-min treadmill run. International Journal of Sport Nutrition and Exercise Metabolism, 18(6), 585-600. doi
4. Brietzke, C., Franco-Alvarenga, P. E., Coelho-Junior, H. J., Silveira, R., Asano, R. Y., & Pires, F. O. (2019). Effects of carbohydrate mouth rinse on cycling time trial performance: a systematic review and meta-analysis. Sports Medicine, 49(1), 57-66. doi
5. Deng, Y., Fan, Y., Liu, Y., Song, Y., Ahmad Fuaad, A. A. H., Bin Mohd Nasiruddin, N. J., & Bin Naharudin, M. N. (2025). Fed, not fasted: is carbohydrate mouth rinsing still ergogenic? A three-level meta-analysis. Journal of the International Society of Sports Nutrition, 22(1), 2579027. doi
6. Zalagkitis, C., Kraniotaki, A., Simitzis, N., Dallas, K., Moustogiannis, A., Chatzinikita, E., et al. (2026). Effects of carbohydrate mouth rinse on human capacity and performance: a systematic review and meta-analysis of 90 controlled trials. Nutrients, 18(18), 3099. doi
7. Lane, S. C., Bird, S. R., Burke, L. M., & Hawley, J. A. (2013). Effect of a carbohydrate mouth rinse on simulated cycling time-trial performance commenced in a fed or fasted state. Applied Physiology, Nutrition, and Metabolism, 38(2), 134-139. doi
8. Li, Q., Liu, S., Miao, C., Zhang, Y., & Lu, Y. (2026). Effects of carbohydrate, caffeine and menthol mouth rinses on exercise performance and ratings of perceived exertion: a systematic review and network meta-analysis. Frontiers in Nutrition. doi