REF-483 · Brietzke et al., 2019 · Sports Medicine, · Paper · PubMed

Effects of Carbohydrate Mouth Rinse on Cycling Time Trial Performance: A Systematic Review and Meta-Analysis

Meta of CHO mouth rinse

A-fairly strong evidenceMeasured on a mixed groupWritten before October 2026, being re-checked

Participants

n=175 across 16 trials for mean power output; n=151 for time to complete the trial (13 studies).

Key Findings

Meta-analysis of carbohydrate mouth rinse (rinsing with a CHO solution without swallowing) versus placebo rinse on cycling time-, work- and distance-based trials.

1

Mean power output improved with CHO mouth rinse: standardized mean difference 0.25 (95% CI 0.04-0.46; p = 0.02) — a small effect.

2

Time to complete the trial did NOT improve: SMD -0.13 (95% CI -0.36 to 0.10; p = 0.25) — no superiority over placebo.

3

Studies showed a low risk of bias and homogeneous results (I² = 0% for both outcomes).

4

The abstract reports no moderator analyses (fasted vs fed, trial duration, concentration, rinse duration) and no brain-imaging data; the orosensory/reward-pathway explanation is the field's working mechanism, not a result of this meta-analysis (Inference, not tested here).

Practical Application

Brietzke et al. 2019 (meta-analysis, 13 studies, n=175): a carbohydrate mouth rinse raised mean power output in cycling trials by a small margin (SMD 0.25) but did not shorten time to complete the trial. Treat it as a cheap, low-risk marginal tool for the end of a race when the gut will not take more, not as something that changes a result on its own.
Not from this paper; the base's own practice: pair it with the caffeine mouth rinse (REF-422) late in long events, and use normal fuelling (REF-066, SYNTH-02) for the bulk of the ride — the rinse adds no fuel.
Rinse protocol (concentration, volume, duration) is not specified in the abstract; follow the trial protocols summarised in REF-066/REF-422 rather than a number from this entry.

Limitations

Meta-analysis — heterogeneity in protocols (concentration, volume, frequency)
Participant tiers are not given in the abstract
Female cohorts underrepresented
The effect is small (SMD 0.25) and appears only in mean power output, not in time to complete the trial — an incremental tool
How this entry was checked and changed
Why this level

regraded 2026-09-27 from A: meta-analysis of small trials (pooled n=175 for power output), one of two outcomes null

Content correction 2026-09-27
Level A -> A-: meta-analysis of small trials, pooled n=175, with one of two outcomes null.
"ES ~0.20-0.35" -> SMD 0.25 (0.04-0.46) for mean power output; added the null time-to-complete result (SMD -0.13) the entry omitted.
Removed as findings: stronger effect when fasted, stronger on 30-60 min TTs, brain regions (insula, nucleus accumbens, DLPFC), "optimal protocol 5-10 mL of 6-8% for 5-10 s" — none is in the abstract.
Removed the athlete-specific race-day stack (named-race timings, 75 mg caffeine gel, 1.4% caffeine rinse): not from this paper and not knowledge-base content.