The 30-day experiment ran from 2026-04-26 14:00 to 2026-05-26 14:00 Cambodia time, with baseline Biomed day-0 testing, a non-fasting 2026-05-12 midpoint checkpoint, zero alcohol/smoking, consistent Tracker logging, stool photos, Apple Watch capture, and a final-week coffee-reduction/tea-substitution attempt with logged coffee exceptions. The end results are now paired: 2026-05-27 blood improved (ferritin 61.60, TSAT about 33%, WBC normal, platelets 444 borderline/high), and 2026-05-28 stool follow-up was FOB negative with stool RBC absent. This strengthens exposure-cleanup as a useful signal and cools the April occult-blood branch, while repeat persistence rules still apply if bleeding/iron signals recur. Private archives preserve the 2025-08-20 visible-bleeding stool-photo comparator and a pre-bleed OTC topical mupirocin genital-exposure record for later Tracker/research review.
Status: completed; interpretation phase. The experiment ran 2026-04-26 at 14:00 -> 2026-05-26 at 14:00 Cambodia time. The planned no-coffee, tea-only final-week phase became a coffee-reduction / tea-substitution attempt with logged coffee exceptions from 2026-05-20 through the final labs; the earlier May 3-6 attempt remains a partial/interrupted exposure period.
Live cleanly and consistently for one month, log enough data to interpret the change, then compare end labs/stool results against the April baseline. This article is the clinical protocol and interpretation ledger; the Next Biomed Visit Plan remains the lab-logistics ledger.
The logging goal is meaningful long-term trends, not perfect symptom capture. Meals, exercise/walks, and stool photos are intended to be registered every time. Weak localized pains may be logged sporadically; stronger, new, persistent, or unusual pains matter more and should be logged when possible.
The experiment is not trying to prove one narrow mechanism. It is designed to separate four tracks that currently overlap: iron availability, occult blood, post-meal bloating/motility, and reactive inflammation/coagulation.
Day-0 state from the first Tracker observation, 2026-04-26 14:28 Cambodia:
| Signal | Baseline detail |
|---|---|
| Start conditions | After breakfast and two coffees; bloating already present on waking. |
| Bloating | 5/10 at start. |
| Localized pain | 1/10; one sudden far-left abdominal pain episode around belly-button height, lasting seconds. |
| Recent confounders | Beer three nights in a row; roughly 30 cigarettes in the prior 24 hours. |
| Exposure reset | Last cigarette at experiment start; smoking stops cold turkey. |
| Working hypothesis | Beer/alcohol may be a major bloating amplifier, but the month should show the pattern rather than assume it. |
The 2026-04-19/20 blood and stool results remain the practical baseline for most markers: ferritin 54.01 with low serum iron 11.64, derived TSAT about 19%, normal hemoglobin 14.4, platelets 494, WBC 13.1, positive FOB/stool RBCs, and calprotectin normalized to 13.3.
Baseline Biomed day-0 testing is now logged across 2026-04-26/27. Key additions: reticulocyte 1.0% normal; fibrinogen 3.7 g/L high-normal; fasting glucose 99 mg/dL with insulin 4.4 µIU/mL gives HOMA-IR about 1.08; urine clean; IgG 1299 and IgM 72.5 normal; hemoglobin electrophoresis normal. The IgG/IgM/urine block is reassuring but is an IgA side quest, not the core experiment. Hemoglobin electrophoresis does not replace serum protein electrophoresis for the IgA band-pattern question.
A 2026-05-12 midpoint blood checkpoint has been added as directional context, not a clean comparator. Unlike the first experiment-day labs and the planned final labs, it was improvised around 14:00, roughly 1 hour after a meal and coffee. It is still useful: WBC normalized to 8.2, CRP stayed normal at 2.71, platelets improved to 452 but remain high, hemoglobin eased to 13.6, ferritin fell to 43.28, and TSAT improved to 27.7%. Interpret it as mid-experiment signal/noise.
The 2026-05-27/28 end summary is directionally better after the completed zero-alcohol/zero-smoking month and a final-week coffee-reduction/tea-substitution attempt. Blood-side: ferritin rose to 61.60, TSAT recovered to 33.4%, WBC stayed normalized at 7.7, CRP stayed normal at 2.52, and platelets improved to 444 — just below the common 450 thrombocytosis threshold but still above Biomed's reference range. Stool-side: FOB became negative and stool RBCs were absent on 2026-05-28. This strengthens the practical hypothesis that alcohol-plus-smoking periods destabilize gut symptoms, iron recovery, and reactive/inflammatory markers, and it cools the April occult-blood branch. It does not prove alcohol alone caused the April positive stool test, and recurrence rules still apply if visible blood, iron decline, or repeat stool positivity returns.
Tracker context before the final-week phase: coffee was still being logged through 2026-05-19, and 2026-05-12/13/15 observations linked coffee temporally with bloating or brief abdominal pain. The final-week log then shows tea substitution plus coffee exceptions on 2026-05-21, 2026-05-23, 2026-05-24, 2026-05-25, and 2026-05-26. That does not prove coffee is causal, and it means the final-week phase must be analyzed as reduced coffee with exceptions — not as a clean coffee-free trial.
| Question | Main evidence after 30 days | Why it matters |
|---|---|---|
| Did iron availability improve? | Ferritin, iron, TIBC/TSAT, hemoglobin, reticulocyte | Tests cleaner intake, reduced coffee/tea timing, alcohol abstinence, and occult-loss pressure. |
| Did the occult-blood signal clear? | FOB, stool RBC/direct exam, ferritin/TSAT trend | Keeps bleeding-source logic separate from bloating. |
| Did bloating improve? | Tracker bloating score, stool photos, meal gaps, abdomen-circumference deltas | Tests meal timing, alcohol removal, reduced-coffee/coffee-exception days, gut load, MMC/SIBO/SUDD branch. |
| Did reactive markers fall? | WBC, platelets, fibrinogen, CRP/ESR | Tests whether smoking/alcohol/gut irritation explained part of the inflammatory/thrombotic pattern; blood-side markers improved by the final summary, but platelet persistence remains its own branch. |
| Did passive physiology improve? | Apple Health sleep, resting HR, HRV, steps/exercise | Adds context without turning the month into manual data entry. |
| Rule | Practical definition | Interpretation value |
|---|---|---|
| No alcohol | Zero beer/wine/spirits for the 30-day window. | Removes a major gut-barrier, sleep, inflammation, and bleeding-risk confounder. |
| No smoking | Zero cigarettes; log any slip immediately. | Removes a major vascular and gut-motility confounder. |
| Log every non-water intake | Meals, snacks, tea, coffee slips, juice, soda, supplements with caloric drinks, and alcohol/smoking slips. | Water-only intervals are inferred from gaps between logged non-water/caloric events. |
| Meal spacing | Aim for 4-5 hour water-only windows when practical. | Tests the MMC/meal-spacing hypothesis for bloating. |
| 24-hour fast | Deferred during the final week and before day-30/31 labs. | Kept the final labs interpretable and avoided mixing the coffee-reduction attempt with fasting/re-feeding effects. |
| No 48-72 hour fast during clean month | Longer fasts were deferred until after day-30/31 labs unless a clinician specifically supervised a separate protocol. | Longer fasting adds avoidable lean-mass, sleep, mood, uric-acid, lipid/ApoB, inflammatory/platelet, bowel-output, and refeeding noise, and would blur the main no-smoking/no-alcohol/meal-spacing experiment. |
| Final-week coffee-reduction phase | Planned from 2026-05-20 through the final labs as no coffee with tea allowed, but Tracker shows coffee exceptions on 2026-05-21, 2026-05-23, 2026-05-24, 2026-05-25, and 2026-05-26. | Preserves a useful reduced-coffee/tea-substitution signal, but it is not a clean coffee-free trial for bloating, sleep/HR, pain timing, or iron-absorption analysis. |
| Tea/iron timing | Log each tea type and rough size; keep tea away from iron-focused meals when practical, ideally 1-2 hours after meals. | Tea polyphenols can still inhibit non-heme iron absorption; stable tea use reduces caffeine-withdrawal noise while preserving the iron branch. |
| Coke Zero / diet-soda avoidance | Avoid Coke Zero, diet cola, and other artificial-sweetened carbonated drinks while the symptom signal is live. If accidentally consumed, log exact drink, volume, timing, and symptom timing. | Removes a plausible same-day bloating/ache trigger and keeps coffee-reduction analysis cleaner. |
| Vegetarian iron focus | Lentils/beans/tofu/greens/seeds plus vitamin-C pairing when practical. | Gives the iron branch a fair test without jumping straight to iron supplements. |
| Apple Watch coverage | Keep charged, worn, synced, and configured for Sleep/Vitals/heart data; record real workouts deliberately. | Adds sleep, HR, HRV, respiratory rate, SpO2/wrist-temperature if available, exercise and steps context. |
| Stool photos | Photograph bowel movements before flushing when practical. | Creates a stool-appearance record for consistency/color review. |
| Trend-consistent logging | Register every meal, exercise/walk, and stool photo when possible; weak pains may be sporadic. | Preserves useful denominators for trend analysis without pretending pain capture is complete. |
A slip is data. Log it and continue; do not restart the clock unless the month becomes too noisy to interpret.
Use the Tracker app as the experiment log, not as a nutrition calculator: https://spinningowl.cloud/meal-tracker/.
| What to log | When / how |
|---|---|
| Meals and caloric drinks | Every time, at intake or as soon as remembered. |
| Diet soda / carbonated sweet drinks | For now: avoid. If there is a slip/rechallenge later, log product, volume, time, carbonation level if relevant, and pain/bloating/stool response. |
| Water | Do not log plain water; analysis uses gaps between logged non-water events. |
| Coffee / tea | Final week: planned no coffee, but actual Tracker data show a reduced-coffee phase with coffee exceptions; log every coffee/tea by type and rough size (coffee, green tea, oolong tea, rooibos tea, etc.). Label the 2026-05-03 to 2026-05-06 period as a partial/interrupted coffee-free attempt. |
| Alcohol/smoking slips | Log as ordinary events: beer, cigarette. |
| Bloating/pain observations | Bloating at least daily and ideally at symptom peaks. Pain means localized pain, not general bloating pressure; weak pains may be sporadic, while stronger/new/persistent/unusual pains should be logged when possible. |
| Maximum-width abdomen circumference | Morning baseline when practical; again at symptom peak or about 2 hours after a main meal when useful. |
| Stool photo | Intended for every bowel movement when possible, before flushing; if missed, do not backfill from memory. |
| Exercise/walks | Register every exercise/walk when possible, even if brief; intensity and notes can stay rough. |
| Supplements/medications | Log once when timing or dose changes. |
The app already handles timestamps, offline queueing, photo upload, experiment counter, and export. Those implementation details belong to the app, not this clinical protocol.
The useful signal is repeatable overnight and walking data, not more metrics. Assume default settings are fine unless data are missing.
Do not spend time auditing default privacy/heart-rate settings unless resting HR, walking HR, workout HR, or sleep data are missing. Apple does not offer a useful setting to increase passive measurement frequency. The practical frequency boost is workout tracking: Apple measures HR continuously during Workout app sessions; Runkeeper is acceptable if its workout in Health includes HR. Avoid Low Power Mode during sleep/workouts because it disables background HR/blood-oxygen measurements and heart notifications.
Manual sensors: take one calm baseline ECG if the app is already set up, then use ECG only for palpitations, skipped/rapid heartbeat, chest discomfort, dizziness, unusual shortness of breath, or an irregular-rhythm notification. ECG is a 30-second single-lead rhythm snapshot and does not detect heart attack, stroke, clots, cholesterol, or blood pressure. Optional blood-oxygen/current-HR checks are symptom-timed only. Skip sleep-stage micromanagement and daily HRV rituals unless effortless.
Log only context Apple/Runkeeper cannot know: tea type/size and any coffee slip, smoking/alcohol slips, illness, gut symptoms, and unusual stress/sleep.
Use the tape as an objective distension add-on. The 0-10 bloating score captures pressure/discomfort; maximum-width abdomen circumference captures physical expansion.
Protocol:
max-width abdomen 92.5 cm, standing relaxed.Best comparison points: morning baseline after toilet and before food/tea; symptom peak; about 2 hours after a main meal. Interpret the delta more than the absolute number. A +5 cm symptom-peak rise supports physical distension; a high bloating score without much circumference change points more toward visceral sensitivity/pressure.
This stays in the symptom/motility branch. It does not replace the bleeding dashboard: visible blood, FOB/stool RBC, ferritin/TSAT/Hb, and red-flag symptoms.
Classification: WATCH / optional later self-test, not a core intervention. Fasting was reviewed as a queued 24h/48h/72h idea for the active 30-day window. It was correctly not added before day-30/31 labs. The practical conclusion remains conservative: protect the clean signals from smoking abstinence, alcohol abstinence, meal timing, stool/iron follow-up, Apple Watch/BP context, and the weaker final-week coffee-reduction signal. A later 24-hour fast can answer a gut-motility question, but it should not become a major health-optimization project during this loaded protocol.
| Duration | Current role | What it could teach | Why not make it central now |
|---|---|---|---|
| 24 hours | Defer until after final labs | Whether bloating/distension/walking discomfort drops with complete food absence; whether refeeding triggers a reproducible symptom flare; hunger/craving/sleep/HR response | Would have mixed with final-lab interpretation and coffee-reduction noise. Lowest-disruption fasting option later, but not needed for the main endpoints. |
| 48 hours | Defer | Stronger ketosis/fat-oxidation and gut-rest signal | More sleep, mood, fatigue, exercise, bowel-output, and refeed noise; can stress nicotine-abstinence stability. Do only later if a 24h fast gives a clearly useful signal. |
| 72 hours | Skip during this experiment | Deep metabolic-switch experiment, not a decision-grade clinical test | Too much confounding and low personal yield: current insulin sensitivity is already good, and longer fasts can acutely alter lipids/ApoB, uric acid, inflammatory markers, platelet/coagulation signals, lean mass, and refeeding response. |
What fasting can plausibly help interpret: meal-driven bloating, MMC/no-snacking response, walking-related distension when fasted, refeed sensitivity, resting HR/HRV/sleep, and behavioral stability under food absence.
What it should not be used to interpret: occult-blood clearance, iron-loss source, diverticular rebleeding prevention, cancer screening, Lp(a) risk, aspirin/antithrombotic safety, or whether WBC/platelets are reactive versus clonal.
If a 24h fast is done, log start/end time, water/electrolytes/salt use, hunger/cravings, bloating and maximum-width abdomen circumference at baseline/pre-refeed/2h post-refeed, fasted walk response, stool/photo if any, sleep, resting HR/HRV, optional BP, and the first refeed meal. Refeed with a moderate familiar vegetarian meal; avoid a huge spicy/oily/raw-fiber challenge. Stop early for dizziness/fainting, chest pain, palpitations, unusual shortness of breath, severe weakness/confusion, severe abdominal pain, visible red/maroon/black stool, vomiting, or inability to hydrate.
Evidence anchors: prolonged fasting review (PMID: 37377031); intermittent-fasting RCT meta-analysis (PMID: 40533200); 48h cognition/mood studies (PMIDs: 28025637, 32504694); 72h metabolic studies (PMIDs: 30183740, 12388154); supervised fasting safety cohorts (PMIDs: 29458369, 30601864); prolonged fasting inflammation/platelet activation signal (PMID: 40268190).
Goal: make images comparable enough for later stool consistency/color review. Consistency matters more than perfect images.
Historical comparator: two user-supplied stool photos timestamped 2025-08-20 have been archived privately under the local archive identifier stool-photos-2025-08-20. They may represent the actual late-August 2025 visible bleeding incident or an immediately adjacent stool. Do not publish the images; use the private originals and manifest only as a later comparator for visible blood/color patterns when reviewing 30-day Tracker stool photos or stool-blood research.
Historical exposure context: a separate private archive, topical-mupirocin-prebleed-2025-08, records the user-supplied photo of 康立邦 莫匹罗星软膏 (mupirocin ointment), reported as bought OTC and applied to penile/genital skin before the August 2025 bleeding incident. The front panel shows 5 g and external-use OTC markers; matching listings show 2% (20 mg/g). Use this later as exposure context when comparing the historical bleed with current stool-photo/stool-blood data; do not treat it as causal until researched.
| Factor | Guideline |
|---|---|
| Conditions | Keep conditions as similar as possible: same bathroom, same lights, same phone/camera mode when practical. |
| Flash | Always use flash. |
| Angle/size | Take the photo from straight above, focus on what is inside the toilet bowl, and maximize stool size in the frame while keeping it recognizable. |
| Focus/color | Retake blurry photos; no filters, night-mode color effects, or editing. |
| Notes | Add a note only for visible red/black area, unusual color, urgency, diarrhea, constipation, pain, or uncertainty. |
Practical default: same conditions, straight above, flash on, focused on the toilet bowl contents, stool as large as practical in the frame.
The deliberate coffee-removal plan started on 2026-05-20, but the actual Tracker record shows a coffee-reduction / tea-substitution phase with exceptions, not a clean coffee-free trial. Logged final-week coffee exceptions were 2026-05-21 (“Coffee”), 2026-05-23 (“Coffee”), 2026-05-24 (“Tuna bun and coffee”), 2026-05-25 (“Coffee”), and 2026-05-26 (“Daily coffee before working”). The original 2026-05-03 14:00 -> 2026-05-10 14:00 Cambodia time window was paused/postponed as of 2026-05-06 because nicotine/smoking abstinence was the higher-priority experiment load. Treat May 3-6 as logged context and May 20-26 as reduced coffee with exceptions, not as a clean no-coffee comparator.
| Rule | Protocol |
|---|---|
| Main rule | Planned rule was no coffee at all, but actual adherence became reduced coffee with logged exceptions. Analyze actual coffee days separately from tea-only days. |
| Caffeine rule | Caffeine is allowed from tea, but keep it moderate and fairly stable. The intended test was coffee removal rather than total caffeine withdrawal; actual interpretation is weaker because coffee exceptions occurred. |
| Default teas | Green tea and oolong tea, unsweetened/plain, are the preferred default. They keep some caffeine while removing coffee-specific compounds and preparation habits. |
| Caffeine-free options | Rooibos, ginger, chamomile, lemongrass, chrysanthemum, and mild mint teas are allowed. Use these especially late day. |
| Secondary options | White tea or black tea are allowed if wanted, but black tea is more caffeine/polyphenol-heavy, so keep it secondary rather than the main replacement. |
| Avoid | Yerba mate, guayusa, yaupon, energy drinks, cola-as-caffeine replacement, Coke Zero/diet colas, bottled sweet tea, bubble/milk tea, and green-tea extract supplements. They add avoidable stimulant, carbonation, sweetener, sugar, calorie, or supplement confounding. |
| Iron timing | Tea polyphenols, not just caffeine, can inhibit non-heme iron absorption. Keep tea away from iron-focused meals when practical: ideally about 1-2 hours after meals, and use vitamin-C pairing with vegetarian iron meals. |
| Logging | Log every tea with type and rough size: green tea, oolong tea, ginger tea, etc. If headache/fatigue/irritability appears in the first 2-3 days, mark it as possible caffeine reduction/withdrawal rather than assuming gut worsening. |
Final-phase interpretation default: not a clean no-coffee trial. It can still inform whether reducing coffee and substituting tea was behaviorally feasible and whether clearly tea-only days differed from coffee-exception days. Recommended default remains useful if the habit continues: green tea or oolong earlier in the day; rooibos/ginger/chamomile/lemongrass/chrysanthemum later, with tea kept away from iron-focused meals when practical. Keep meals, alcohol/smoking abstinence, exercise, and other routines stable when evaluating any future coffee-removal signal.
What it can answer now: whether reduced coffee plus more tea coincided with any obvious change, and whether tea-only days differ from coffee-exception days. What it cannot answer cleanly: whether strict coffee removal improves bloating, stool/urgency changes, sleep/HR signals, or iron-absorption noise. A future clean trial would need no coffee at all and no cola/diet-soda replacement, because carbonation, phosphoric acid/cola flavoring, caffeine, and non-caloric sweeteners are separate GI variables.
Evidence basis kept practical: brewed coffee is roughly 96 mg caffeine per 8 oz, green tea about 29 mg, and black tea about 48 mg, with wide preparation variability (Mayo Clinic, 2025). Caffeine withdrawal can begin 12-24 hours after stopping, peak around 20-51 hours, and last 2-9 days, sometimes from habitual intakes around 100 mg/day (PMID: 15448977); therefore a tea-caffeine allowance keeps withdrawal from drowning the coffee-specific signal. Iron timing still matters because polyphenol-containing drinks, including black tea, coffee, peppermint, and several herbal teas, can strongly inhibit non-heme iron absorption with a meal (PMID: 10999016). Coffee also has GI effects beyond caffeine, including gastric-acid/gastrin, bile/pancreatic secretion, microbiome, and colonic-motility effects, with individual symptom response still variable (Nutrients 2022 coffee-GI review, PMCID: PMC8778943).
At day 30/31 (2026-05-26/27 Cambodia), repeat the paired set if possible.
Blood:
Stool:
Optional blood add-ons only if the result would change interpretation: zinc ($16.25) and copper ($15.00, especially if adding zinc). CEA is not a routine stool-blood add-on; use it only if a clinician specifically wants tumor-marker follow-up in the context of prior CEA history, symptoms, imaging, or endoscopy planning.
If visible blood, melena, worsening pain, dizziness, unusual fatigue, or a major bowel-pattern change appears, do not wait for day 30. Repeat the stool/blood set sooner and treat it as a GI-source escalation question.
| Pattern after 30 days | Interpretation |
|---|---|
| Iron/ferritin improve and stool blood clears | This is the current end-experiment pattern: cleaner intake/absorption plus no obvious ongoing occult-loss signal on the repeat stool sample. |
| Iron/ferritin improve but stool blood is not repeated or still pending | Blood-side recovery improves the exposure/iron story, but does not answer the bleeding-source question. |
| Iron/ferritin improve but stool blood persists | Diet/absorption may help, but bleeding-source question remains unresolved. |
| Iron/ferritin fall and stool blood persists | Ongoing loss becomes the dominant concern. |
| Iron/ferritin fall but stool blood clears | Intake/absorption, small-bowel, or non-GI explanations remain possible. |
| Bloating improves but iron does not | Symptom branch and iron branch are probably not the same problem. |
| Bloating does not improve despite clean spacing | SIBO/MMC/SUDD branch stays active; meal-spacing hypothesis weakens. |
| Stool photos improve while FOB remains positive | Visible consistency improves, but occult-bleeding logic remains active. |
| Stool photos worsen with stable labs | More likely motility/diet/microbiome branch unless red flags appear. |
| WBC/platelets/fibrinogen normalize | Reactive exposure/inflammation explanation strengthens. |
| WBC/platelets/fibrinogen stay high | Reactive-only explanation weakens; revisit inflammatory/hematology logic. |
| Apple Health sleep/HRV/resting HR improves with symptoms | Systemic recovery signal strengthens. |
| Apple Health improves but GI symptoms do not | Gut-specific branch remains active despite general recovery. |
These remain separate tracks even if the month goes well: