SYNTH-08 · Guide
Gut problems in training and racing
Expect them past two hours, in heat and on the run; ask about history, rehearse the race intake on long sessions and change one thing at a time; nothing in the base shows that a probiotic or a "gut-health" diet prevents symptoms or makes an athlete faster
The question it answers
Default (what to prescribe first)
1
Say that it is expected. A systematic review found that markers of intestinal injury, permeability and endotoxaemia rise, and gastric emptying, transit and absorption worsen, as intensity and duration increase; ≥2 h at 60% VO2max appears to be the threshold irrespective of fitness, heat stress and running make the markers worse, and the changes are reversible in healthy people (REF-525, A-; markers, not pooled; the step from permeability to symptoms is not established).
2
Ask about history first. In 221 athletes surveyed after Ironman, half-Ironman, marathon and cycling races, symptom scores correlated with a reported history of GI distress (r = 0.37 upper, r = 0.51 lower); serious symptoms were reported by about 31% in the Ironman races, 14% in the 70.3, 4% in the marathon and 4% in a 100/150-km cycling race (REF-787, B+, self-report after the race).
3
Rehearse the race intake; do not meet it on race day. Gastric emptying and stomach comfort can be trained and a high-carbohydrate diet raises intestinal SGLT1 density and activity, but the most effective method and the timeline are not established (REF-008, B+, narrative review).
Two trials tested one method: in 25 endurance runners, 2 weeks of repeated carbohydrate intake while running (gel-disc or food) cut gut symptoms in a 2 h run at 60% VO2max with 30 g every 20 min by 60% and 63%, more than placebo (REF-979, B+); in 18 runners, 90 g per hour during a daily 1-hour run for two weeks reduced symptoms, nausea and malabsorption while placebo did not (REF-980, B).
Same laboratory and protocol, so the cohorts may overlap; no race, heat or cycling. In 16 trained cyclists and triathletes, 28 days of carbohydrate in daily training raised oxidation of a glucose drink (54.6 to 63.6 g), with no symptom measure and the same approximately 6% time-trial gain as the low-carbohydrate group (REF-978, B).
The target and the weekly ramp are SYNTH-02.
4
Change one variable per long session. Not from a study; the base's own practice: rate, then product or sugar blend, then form, then timing — so the athlete learns which one mattered.
5
Eating less is not the first fix. In both Ironman races more carbohydrate per hour went with more nausea and flatulence and with a faster finish (r = −0.55 and −0.48; REF-787, correlational).
Decision (who → what)
| What the athlete reports | What to change | Source and how sure |
|---|---|---|
| Fullness, sloshing or nausea at a rate above what they have practised | Practice: go back to the last rate that sat well and rebuild by the SYNTH-02 ramp | Gut training at 30-90 g/h reduced discomfort in the interventions a 2026 perspective review covers (REF-3490, B+; full-text line of the entry); the two trials used the test rate daily for two weeks, not a ramp (REF-979, REF-980) — the ramp is practice |
| Symptoms above ~60 g/h on glucose or maltodextrin alone | Switch to a glucose:fructose blend; ratio per SYNTH-02 | Above ~60 g/h a fructose-to-glucose ratio of 0.6-1.0 (REF-3490, full text). A 1:1 blend was oxidised at 1.75 vs 1.06 g/min for glucose alone in 8 trained male cyclists (REF-527, B; oxidation only, no GI data). Inference: unabsorbed carbohydrate is the usual explanation for symptoms |
| Cannot stomach gels, or drinks, or chews | The form the athlete tolerates | At 120 g/h (1:0.8), drink, gel, chew or a mix gave the same oxidation with minimal GI distress in 9 trained male cyclists, after a high-carbohydrate day and a pre-exercise meal (REF-333, B; 3 h only). Fluid with every gel: practice |
| Symptoms only when it is hot | Expect them; pace by RPE and heart rate, cool, drink per SYNTH-68; do not cut carbohydrate "to protect the gut" | Heat worsens the markers (REF-525). At 32 vs 19 °C with airflow, 10 trained men oxidised 14-15% less of a 72 g/h glucose drink while GI comfort and I-FABP did not differ (REF-461, B; the no-cut reading is that entry's inference) |
| Fine on the bike, symptoms on the run | Not from a study; the base's own practice: rehearse eating while running at race pace, and take more of the plan on the bike | Running exacerbates the markers (REF-525) |
| Ultra: sweet food becomes impossible, late nausea | Practice: a lower hourly target and mixed, calorie-dense foods, set by rehearsal (SYNTH-02 ultra note) | The ISSN ultramarathon stand keeps its in-race carbohydrate figure below marathon-style fuelling and adds protein and calorie-dense foods (REF-528, A); ~90 g/h may not be tolerable in ultra competition (REF-663, B+, narrative) |
| Repeated bloating, urgency or diarrhoea despite rehearsal | Trial a short low-FODMAP period, with a dietitian if possible — a trial for day-to-day lower-gut symptoms; not shown to protect the run itself | The ISSN ultramarathon stand supports gut training and/or low-FODMAP diets for GI distress during racing (REF-528, A, running — transfer to other sports); in 11 recreational runners with a history of exercise-related gut symptoms, 6 d of low-FODMAP versus high-FODMAP eating lowered daily symptoms (flatulence, urge to defecate, loose stool, diarrhea) but not symptoms during the hard running sessions (REF-981, B, single-blinded crossover; the authors call it preliminary). How many days before is the base's own practice |
| Takes ibuprofen before or during long events | Practice built on the trial: stop | In 89 ultramarathon runners over 80 km, ibuprofen 400 mg every 4 h was followed by acute kidney injury in 52% vs 34% on placebo; the interval (−4% to 41%) includes zero (REF-892, A-). For the gut: in nine trained men, 400 mg of ibuprofen taken twice before cycling gave a higher marker of small-intestinal injury than cycling alone (peak I-FABP 875 vs 474 pg/mL) and the highest permeability of the four conditions (REF-983, C+, crossover; markers only, no symptoms measured) |
| Gut complaints at rest plus signs of under-eating | Practice: screen energy availability (SYNTH-54) before changing race nutrition | The IOC REDs consensus lists abdominal pain, cramps, bloating and altered bowel movements among the GI impairments of low energy availability (REF-341, A; full-text Table 2). Inference, not tested: eating enough may resolve them |
How to progress / periodise
Train the gut on the sessions that resemble the race. Rehearse the planned intake on long sessions, in the conditions the race will bring (REF-008). Not from a study; the base's own practice: that means race intensity, the heat, the run leg, the race products and the race breakfast.
The endpoint is the athlete's own cell in [SYNTH-02], not a fixed top rate for everyone. The elite mountain runners given 120 g/h had gut-trained beforehand (REF-331, B+, 20 men; its abstract reports no GI outcome).
Race week. Not from a study; the base's own practice: lower-fibre, familiar food for the last 1-3 days, nothing new on race morning.
When NOT to use it / red flags
Blood in stool or vomit, black stools, symptoms at rest or at night, fever, unexplained weight loss → a clinician. Not from a study; the base's own practice.
A diagnosed gut condition: follow the clinician's plan; the safety of strenuous exercise in chronic GI disease is not established (REF-525).
Do not stack food restrictions on an under-fuelled athlete (SYNTH-54). Not from a study; the base's own practice.
What the evidence does and does not show
Does not show: that gut training works in a race, in heat, on the bike or beyond two weeks; that a low-FODMAP diet reduces symptoms during exercise (REF-981: no difference); the base has no controlled trial with a symptom outcome for drink concentration, lower fibre before a race or meal timing.
Does not support treating the microbiome. Veillonella rose in runners after a marathon, and a V. atypica strain isolated from them lengthened treadmill run time when given to mice (magnitude not in the abstract); the human part is observational, and Veillonella probiotics have no proven effect in humans (REF-519, B). No entry supports a stool test, a "diversity" target or fermented-food quotas.
Probiotics: the ISSN position stand says specific strains can reduce the number of episodes, severity and duration of upper respiratory tract infections and can improve the integrity of the gut-barrier function in athletes, benefits being strain- and dose-dependent (REF-982, A, position stand — narrative, no strain, dose or effect size in its abstract, many authors with probiotic-industry ties).Its abstract makes no statement on gut symptoms during exercise or on endurance performance, so do not recommend a probiotic for race-day gut symptoms from this base; it is not among the few supplements SYNTH-67 finds a case for.
Gut-brain claims (gut inflammation driving fatigue or mood) are not measured in any entry.
Evidence gaps
Low-FODMAP against a habitual diet, and with a during-exercise benefit (REF-981 found none) — no entry.
A named probiotic strain and dose with a gut-symptom outcome in athletes (REF-982 names none in its abstract) — no entry.
NSAIDs and gut symptoms or bleeding, not markers (REF-983) — no entry.
Drink concentration; pre-race fibre, fat and caffeine; women — no entry.
How this entry was checked and changed
Why this level
regraded 2026-10-01 from B: what provokes symptoms rests on an A- systematic review of markers and a B+ post-race survey; trainability on a B+ narrative review and two small laboratory RCTs in runners (B+ and B, one laboratory); low-FODMAP on one line of an A position stand and one B crossover of 11 runners; the NSAID gut signal on one C+ marker study; the troubleshooting order — the core decision — is house practice
Content correction 2026-09-27
"Trainable with repeated CHO exposure (REF-333 Hearris 2022)" and "Gut training progression (REF-333)" -> REF-333 is a delivery-form study at 120 g/h and no longer says "gut-trained"; trainability re-sourced to REF-008 (which names the timeline unknown), the 4-6 week ladder labelled practice with REF-334's narrative estimate.
"Veillonella (Scheiman 2019)" -> cited as REF-519, with the facts as the entry has them: enriched after a marathon, mice +13% TTE, human part observational. Prevotella / Bacteroides / diversity and the gut-brain axis labelled practice / inference.
"Jäger et al. 2019 ISSN position: specific strains show small benefits for GI symptoms and URTI" -> not an entry in the base; abstract-checked and restated (strain- and dose-dependent; gut and immune health most studied); CFU / week rules labelled practice.
"Protein 1.6-2.2 g/kg/day (REF-348)" -> REF-348 is the older-adult entry; 1.6-2.0 g/kg/day traced to REF-578 (ISSN), 1.6-2.2 kept for masters via REF-348.
Hydration "400-800 mL/h; avoid dehydration > 2-3%" -> REF-061's 500-800 mL/h with sodium and < 2% body mass; "30-70% GI distress", "≤ 7-8% concentration", NSAIDs labelled practice.
"LEA + gut (REF-341, REF-343): dysmotility, altered microbiome, slower transit" -> REF-341 carries constipation / gastroparesis / reduced iron absorption; REF-343 has no gut outcome and is cited for prevalence only; the rest labelled inference.
"Chronic gut inflammation → IL-6, TNF-α → mental fatigue (REF-370)" -> REF-370 measures mental fatigue (TTE −12%, RPE up), not gut cytokines; the link is labelled inference.
Cross-reference "REF-357 (omega-3 — anti-inflammatory gut benefit)" -> that entry has no gut outcome; wording corrected. Scope "Section 16 (nutrition)" -> Section 22.
(sync 2026-09-27, after the full-base pass) REF-519 "+13% running time to exhaustion" -> the corrected entry (now B) says only "significantly increased" run time in mice; figure removed.
REF-334 "4-6 week gut-adaptation estimate" -> removed from REF-334 (not in its abstract; review now C+ and finds no clear evidence ≥100 g/h beats 60-90 g/h); the 4-6 week ladder stays as base practice without that attribution.
REF-331 "1:0.8 ratio used" and "trail runners" -> its abstract gives no CHO composition; restated as 120 g/h vs 60/90 g/h damage markers in elite mountain runners; transport-half level parenthetical updated (REF-331 now B+).
REF-061 "< 2% body mass, 500-800 mL/h" -> the consensus sets no threshold; 500-800 mL/h is labelled base practice.
REF-371 "pooled ES −0.30 to −0.55" -> systematic review of 11 experiments with no meta-analysis (now A-); restated as direction only.
Resync 2026-09-28: REF-370 "time to exhaustion cut by 12% in recreational participants, physiology unchanged" -> 754 → 640 s, about 15% by arithmetic, in 16 subjects whose training status the abstract does not give; physiology "largely unaffected". The earlier note in this section quoting "TTE −12%" is superseded.
Resync 2026-09-28: REF-519 "enrichment in Veillonella atypica" -> the genus Veillonella was enriched post-marathon; V. atypica is the strain isolated and gavaged into mice.
Resync 2026-09-29: re-read against REF-341. No claim here depended on its relabelled caloric-restriction line, but REF-341's systems table (now checked against the full text, Table 2) no longer lists constipation or gastroparesis: the GI row reads abdominal pain, cramps, bloating and altered bowel movements. The LEA + gut bullet is corrected to that list; the 2026-09-27 correction log above is left as history.
2026-10-01: rewritten in the decision-framework format and narrowed to gut problems in training and racing plus gut training. Default changed from a four-topic overview to: expect symptoms, ask about history, rehearse the race intake, change one variable at a time.Contradictions resolved: the ladder ending at "90-120 g/h" for every athlete is removed — target and ramp are the athlete's SYNTH-02 cell; "500-800 mL/h with sodium" is removed — drinking is SYNTH-68.Removed as unsupported by any entry: fibre, fermented-food and probiotic rules (Jäger 2019 is now an evidence gap), microbiome-composition and gut-brain lines, protein targets (→ SYNTH-69), the "30-70%" prevalence and "≤ 7-8%" concentration. Folded in: REF-525, REF-787, REF-3490, REF-527, REF-461, REF-528, REF-663, REF-892, REF-776. Level B -> B-.
Verification 2026-10-01: all 14 cited REF entries opened in KB.md. Numbers that match: ≥2 h at 60 % VO2max, n=221, r=0.37/0.51, ~31/14/4/4 %, r=−0.55/−0.48, 30-90 g/h, ratio 0.6-1.0 above ~60 g/h, 1.75 vs 1.06 g/min (n=8), 120 g/h 1:0.8 (n=9, 3 h), 32 vs 19 °C / 72 g/h / 14-15 % (n=10), 52 % vs 34 % with interval −4 % to 41 % (n=89, 80 km), n=20 for REF-331. Levels/designs next to refIds match head blocks. No NUMBER, LEVEL, WRONG or SOURCE findings.
Verification 2026-10-01: Level B- is honest and arguably conservative: provocation and history rows are evidence, the troubleshooting order is labelled practice, and the 'does not show that any fix works' paragraph is accurate against the snapshot (no gut-training, low-FODMAP, probiotic or NSAID-gut-injury trial entry found by tag search). No LEVEL-OVERALL.
Verification 2026-10-01: claims of absence checked by tag search over KB.md — no probiotic entry other than REF-519 (mouse/observational), no low-FODMAP trial, no gut-training trial with symptom outcome: confirmed.
Verification 2026-10-01: formatting — the Sport and Tags fields share one line in the head block, and there is no Applicability line; not a content issue.
Verification 2026-10-01: consistent with SYNTH-02 (ratio, ramp, ultra start, no carbohydrate cut in heat) and SYNTH-68 (hyponatraemia red flag).
Revision 2026-10-01: Default 3 (rehearse the race intake) now rests on trials, not only on a narrative review: REF-979 (Costa 2017, B+, 25 runners) and REF-980 (Miall 2017, B, 18 runners) — two weeks of practising the intake cut symptoms; flagged as one laboratory with possibly overlapping cohorts, 60% VO2max, no race.REF-978 (Cox 2010, B) added as the oxidation trial with no symptom measure and no extra performance gain. The 'no controlled trial of any fix' statements in the lead and in 'Does not show' were narrowed accordingly.
Revision 2026-10-01: low-FODMAP row narrowed, not strengthened: REF-981 (Lis 2018, B, 11 runners) lowered day-to-day symptoms but NOT symptoms during the running sessions, so 'before long sessions and races' became 'a short trial for day-to-day lower-gut symptoms; not shown to protect the run itself'.
Revision 2026-10-01: probiotic line changed from 'the base has no probiotic entry; do not recommend one' to what the ISSN stand's abstract claims (strain-specific URTI and gut-barrier effects) and does not (no strain, dose, effect size; nothing on exercise gut symptoms or endurance performance; industry ties).The instruction is now 'do not recommend a probiotic for race-day gut symptoms'. Title claim (no probiotic shown to prevent symptoms or make an athlete faster) still holds.
Revision 2026-10-01: ibuprofen row: the 'gut-damage rationale is not from a study in this base' clause replaced by REF-983 (Van Wijck 2012, C+, nine men, markers only); the kidney sentence (REF-892) untouched. Evidence gaps rewritten to what is still missing. Level B- unchanged (two small trials from one laboratory; the troubleshooting order is still practice).