Gut Protocol — Complete History

Four phases, eleven months, one root-cause thread: a Sep 2025 GI-MAP showing H. pylori and pathogenic EPEC/EHEC E. coli triggered a structured Clear → Repair → Repopulate protocol; the Jun 2026 retest confirmed that work succeeded on its original targets but surfaced a new problem — two keystone species, Akkermansia muciniphila and Faecalibacterium prausnitzii, both undetectable. Phase 4 is the answer: feed and replete, don't kill.

Started Nov 18, 2025

3 phases complete

Phase 4 active since Jul 8, 2026

Next GI-MAP: ~Jan 2027

🧭 The Short Version

Phases 1–3 (Nov 2025 – early 2026) worked exactly as designed. The pathogenic EPEC/EHEC E. coli that flagged on the Sep 2025 GI-MAP is now fully cleared, calprotectin sits at 0 with no active mucosal inflammation, and pancreatic/digestive function (elastase-1) actually improved. That's the direct evidence the clearing → repair → repopulate sequence hit its targets.

But the Jun 2026 retest changed the objective. Akkermansia muciniphila — which was actually elevated on the Sep 2025 draw — and Faecalibacterium prausnitzii — undetectable on both draws — are now both undetectable. These are the two most metabolically important species in the gut: Akkermansia for mucosal/metabolic health, F. prausnitzii as the primary butyrate producer. Overall bacterial biomass across both major phyla also came in below reference floor. Phase 4 is not a fourth antimicrobial round — it's ecological reconstruction: direct Akkermansia repletion, prebiotic substrate for F. prausnitzii (no standalone probiotic exists for it), and a butyrate bridge while that population re-establishes. Dr. Nguyen independently reviewed the same data and endorsed the same three additions by name.

Timeline at a Glance

Phase 1 · CLEARCOMPLETE

Nov 18 – ~Jan 2026Candibactin-BR + GastroMend-HP. Eradicate H. pylori load, pathogenic E. coli, opportunistic overgrowth.

Phase 2 · REPAIRCOMPLETE

Dec 28, 2025 →BPC-157 peptide restart. Rebuild mucosal barrier & tight-junction integrity after the clearing phase.

Phase 3 · REPOPULATECOMPLETE

Early 2026MegaSporeBiotic 8-week titration → Seed DS-01 maintenance. Rebuild general microbial diversity.

Phase 4 · REPLETEACTIVE

Jul 8, 2026 →Akkermansia + Sunfiber + Tri-Butyrin + MegaPre. Directly restore the two keystone species the retest showed missing.

Phase 1 · Clearing

Eliminate the pathogen load

Nov 18, 2025 → completed ~Jan 2026 · triggered by the Sep 18, 2025 GI-MAP

Why

The Sep 2025 GI-MAP showed H. pylori positive at 5,340 genome eq/g and pathogenic EPEC/EHEC E. coli at 2,050 genome eq/g (reference: <1,000) — an active infection, not a subclinical finding. Streptococcus was also elevated. Before any repair or repopulation work could be durable, the active pathogen burden had to come down first — repairing a lining that's still under attack, or repopulating a niche pathogens are actively occupying, doesn't hold.

What was taken

Agent	Dose	Duration	Mechanism
Candibactin-BR	2 tabs, once daily, between meals	45 days	Broad-spectrum botanical antimicrobial (berberine-based) — H. pylori & pathogenic E. coli
GastroMend-HP	2 caps, twice daily, between meals	30 days	Mastic gum / zinc-carnosine / DGL complex — gastric mucosal support alongside H. pylori suppression

Perfect Amino was paused during this window to keep focus on clearing rather than splitting attention across supplements.

Result (confirmed on the Jun 2026 retest): pathogenic EPEC/EHEC E. coli is now fully below detection — the clearest single win of the entire protocol. H. pylori dropped 5,340 → 1,710 genome eq/g (−68%), still technically positive but trending down, with every virulence-gene marker tested (babA, cagA, dupA, iceA, oipA, vacA, virB, virD) negative on both draws — a low-virulence strain with zero resistance genes on the panel.

Phase 2 · Repair

Rebuild the barrier

From Dec 28, 2025

Why

Pathogen clearance doesn't repair the tissue damage an active infection leaves behind — the intestinal lining, tight-junction proteins, and mucosal integrity needed their own targeted intervention once the pathogen load was down. This is the structural-repair phase between eliminating the problem and repopulating the ecosystem.

What was taken

Agent	Dose	Mechanism
BPC-157	1 cap/day, 4-weeks-on/2-weeks-off cycling	Mucosal repair peptide — promotes angiogenesis, upregulates EGF/VEGF/FGF, supports tight-junction protein expression

BPC-157 is the one Phase 1–3 agent still running today (now on ProHealth Longevity's Pure BPC-157, switched from Apeiron Elements Aug 18, 2026 — same mechanism, same cycling schedule, just a single-ingredient product going forward). It has stayed in the stack continuously as general tissue-repair support, independent of the specific gut-repletion work in Phase 4.

Result: calprotectin at 0 on the Jun 2026 retest (no active mucosal inflammation or bleeding), elastase-1 improved from 268 to 510 µg/g (well above the >200 threshold — stronger pancreatic/digestive enzyme output), and secretory IgA rose from 928 to 1,288 µg/g, still mid-range — mucosal immune tone intact, not exhausted.

Phase 3 · Repopulate

Rebuild general microbial diversity

Early 2026

Why

With the pathogen burden cleared and the lining repairing, the gut needed broad-spectrum reinoculation — a general-purpose diversity rebuild — before the more surgical, species-specific work in Phase 4 could even be assessed. You can't target two missing keystone species meaningfully against a background of low overall diversity; this phase established that background first.

What was taken

Agent	Dose	Role
MegaSporeBiotic	8-week titration protocol	Spore-based probiotic (Bacillus species) — survives stomach acid, transiently colonizes to recondition the gut environment
Seed DS-01	2 caps/day (ongoing)	24-strain synbiotic, ViaCap® delayed-release delivery — long-term diversity + barrier maintenance, still active today

Result: the general repopulation groundwork held — no autoimmune-associated bacteria, no parasites, calm inflammatory markers throughout. But the Jun 2026 GI-MAP showed this phase didn't reach two specific species: Akkermansia and F. prausnitzii remained undetectable, and total Firmicutes/Bacteroidetes biomass came in below reference floor on both phyla. That gap is exactly what Phase 4 was built to close.

Phase 4 · Microbiome Repletion ACTIVE

Restore the two missing keystone species

Started Jul 8, 2026 · full dose reached Jul 29, 2026 · physician-endorsed

Why — and why this is a philosophy shift, not phase 1 again

The Jun 19, 2026 GI-MAP retest (completed Jun 30) is the trigger. Two findings stand out: Akkermansia muciniphila — which was actually elevated at 1.7×10⁷ org/g on the Sep 2025 draw — is now undetectable, and Faecalibacterium prausnitzii, already undetectable in Sep 2025, still is. These are the two most metabolically relevant species in the gut: Akkermansia maintains the mucin layer and is tied to insulin sensitivity; F. prausnitzii is the primary butyrate producer and the anti-inflammatory backbone of a healthy colon. Secondary flags also appeared — Pseudomonas jumped from undetectable to 7.64×10⁵ org/g (well above the 1×10⁴ reference), Streptococcus stayed persistently elevated (~10,200 org/g), and Candida crossed into flagged territory (5,500 vs. <5,000 genome eq/g).

The deliberate decision: don't run a fourth antimicrobial round. Phases 1–3 asked "what needs to be killed or healed?" Phase 4 asks "what needs to be fed or repleted?" Adding antimicrobials now would also confound the read on whether repletion alone — without further pathogen suppression — is what moves these markers. Candibactin, oregano oil, garlic, neem, caprylic acid, and grapefruit seed extract are all explicitly excluded this round. Dr. Nguyen independently reviewed the same GI-MAP and reached the identical conclusion, naming Pendulum Akkermansia, PHGG, and Tri-Butyrin Supreme specifically, and confirmed the rest of the existing stack "looks great" underneath these additions.

What's being taken

Agent	Brand	Dose	Job
Akkermansia 500 PRO	Pendulum	1 cap/day, with first meal, refrigerated	Direct live-strain repletion of the species confirmed undetectable — this is a confirmed-zero case, not a general-population guess
Sunfiber (PHGG)	Tomorrow's Nutrition	Titrated 2g → 4g → 7g/day (full dose reached Jul 26)	Primary substrate for F. prausnitzii — an obligate anaerobe with no standalone probiotic product; feeding it is the only lever available
Tri-Butyrin Supreme	Designs for Health	1 cap 2×/day (AM + PM)	Butyrate bridge — supplies the downstream metabolite F. prausnitzii would normally produce, while the population re-establishes. 4-month course, reassess
MegaPre (capsules)	Microbiome Labs	6 caps/day (full dose reached Jul 29)	FOS/XOS/GOS oligosaccharide blend — distinct substrate class from PHGG, feeds Akkermansia/F. prausnitzii/Bifidobacterium in parallel. Held back 2 weeks per Dr. Nguyen for tolerance-first sequencing

Everything else in the stack — Seed DS-01, BPC-157, FiberMend, the rest of the 21-item protocol — stays exactly as-is underneath these four additions. This is an addition, not a replacement.

Where things stand, Aug 28, 2026 (7 weeks in): too early for a lab-confirmed result — published repletion windows for these mechanisms run 4–12+ weeks, and colonization of an obligate-anaerobe niche like F. prausnitzii's is on the slower end of that. The working hypothesis is that this dysbiosis plausibly contributes to three current plateaus — fasting glucose, ApoB/LDL, and a slow ferritin decline — and the Jul CGM's steady glucose decline through the ramp is suggestive but correlational, not proof. The 6-month GI-MAP retest is the real checkpoint, not subjective symptoms or interim labs.

GI-MAP Before & After — Sep 2025 vs. Jun 2026

The two data points bracketing the first three phases. This is the evidence base for everything above — pulled directly from gimap_master.csv, not eyeballed from prior report copy.

Marker	Sep 18, 2025	Jun 19, 2026	Reference	Change
EPEC/EHEC E. coli (pathogen)	2,050 genome eq/g (H)	<dl (undetectable)	<1,000	Cleared
H. pylori	5,340 genome eq/g (H)	1,710 genome eq/g (H)	—	−68%, still positive
H. pylori virulence factors (8 tested)	All negative	All negative	Negative	Unchanged — low-risk strain
Akkermansia muciniphila	1.70×10⁷ org/g (H)	<dl (undetectable) (L)	10 – 8.2×10⁶	High → undetectable
Faecalibacterium prausnitzii	<dl (undetectable) (L)	<dl (undetectable) (L)	1,000 – 5×10⁸	Unchanged — still zero
Firmicutes (total biomass)	4.22×10¹¹ org/g (H)	1.07×10¹⁰ org/g (L)	5.7×10¹⁰ – 3.0×10¹¹	Sharp decline, below floor
Bacteroidetes (total biomass)	7.31×10¹² org/g (H)	2.36×10¹¹ org/g (L)	8.6×10¹¹ – 3.3×10¹²	Sharp decline, below floor
Pseudomonas spp.	<dl (undetectable)	7.64×10⁵ org/g (H)	<1.0×10⁴	New elevation, ~76× ref ceiling
Streptococcus spp.	10,600 org/g (H)	10,200 org/g (H)	<1,000	Persistently elevated, flat
Candida spp.	<dl (undetectable)	5,500 genome eq/g (H)	<5,000	New, just over threshold
Calprotectin (mucosal inflammation)	0 µg/g	0 µg/g	<173	Clean both draws
Elastase-1 (pancreatic function)	268 µg/g	510 µg/g	>200	Improved, well above floor
Secretory IgA (mucosal immunity)	928 µg/g	1,288 µg/g	510 – 2,010	Rose, mid-range

Reading the whole table together: this is a calm-immune-system, colonization-and-substrate problem, not an inflammation problem — a meaningfully easier situation to correct than if active inflammation were present alongside the missing species. Calprotectin at 0 and negative occult blood on both draws mean there's no IBD-type process underneath any of this. The honest complication is that Akkermansia's swing from elevated to undetectable, combined with both major phyla falling below their reference floors, suggests something disrupted overall biomass between the two draws — not just a failure to build up the two target species. Phase 4's four additions address the specific keystone-species gap directly; the broader biomass question is being tracked via the same 6-month retest rather than treated as a separate, unexplained problem right now.

🔮 Prediction — What the Next GI-MAP Should Show

Framed against published repletion timelines and the specific mechanism of each Phase 4 agent. These are directional expectations to check the retest against — not guarantees, and stated with the confidence level each one actually deserves.

Akkermansia muciniphila HIGH CONFIDENCE

Now: undetectable (<dl)

Should return detectable, plausibly back into or near the 10 – 8.2×10⁶ reference range. Direct live-strain dosing for 6 months is the most mechanically straightforward intervention in the whole protocol — this is what Pendulum's own colonization data is built on. If it's still undetectable at retest, that's the single most important finding of Phase 4 and would call for a formulation or dosing review, not a "just keep going."

Faecalibacterium prausnitzii MODERATE CONFIDENCE

Now: undetectable (<dl), unchanged since Sep 2025

Likely to move but may still read low or undetectable. There's no direct probiotic for this obligate anaerobe — Sunfiber, MegaPre, and resistant starch only feed a population that has to re-establish itself from a low starting colony count, which is slower and less certain than Akkermansia's direct-repletion path. Partial movement (a detectable-but-low value) would still count as real progress here.

Pseudomonas spp. MODERATE CONFIDENCE

Now: 7.64×10⁵ org/g (~76× reference ceiling)

Expected to decline via competitive exclusion as keystone species reclaim niche space — but this is the flag with the least direct intervention aimed at it, and the magnitude here is unusual enough that it's the one to revisit directly with a targeted agent if it hasn't meaningfully improved by retest.

H. pylori HIGH CONFIDENCE

Now: 1,710 genome eq/g, still positive

Likely continued gradual decline, consistent with the −68% trend already seen and the low-virulence, zero-resistance-gene profile. Not being actively treated this round by design — full eradication isn't the expectation here, just continued monitoring for a strain that hasn't shown any concerning virulence markers on either draw.

Firmicutes / Bacteroidetes biomass LOWER CONFIDENCE

Now: both phyla below reference floor

The least predictable line item. MegaPre and Sunfiber should support broad biomass recovery, but neither phase-4 agent targets total biomass directly, and the mechanism behind the Sep 2025 → Jun 2026 drop isn't fully explained yet. Worth flagging explicitly at the next draw rather than assuming it self-corrects alongside the two named species.

Streptococcus / Candida MODERATE CONFIDENCE

Now: Strep flat ~10,200; Candida newly flagged 5,500

Plausible modest improvement through the same competitive-exclusion mechanism as Pseudomonas, though Streptococcus has now stayed essentially flat across both draws despite the entire Phase 1–3 protocol running in between — the weakest secondary-flag response so far, worth watching rather than assuming it resolves passively.

📅 When to Schedule the Next GI-MAP

Target window: mid-December 2026 to mid-January 2027 — Phase 4 started Jul 8, 2026, and the adopted protocol calls for a minimum 6-month course before retesting (the same interval used to bracket Phases 1–3, and enough time for both the slow F. prausnitzii colonization curve and the Tri-Butyrin 4-month course's downstream effects to resolve). Six months from the Jul 8 start lands ~Jan 8, 2027; drawing anywhere from Dec 15, 2026 – Jan 15, 2027 keeps the interval honest without waiting so long that an underperforming agent goes unnoticed for longer than necessary.

Don't draw earlier than ~Nov 2026 — at the 7-week mark today (Aug 28), it's already stated above that published repletion windows run 4–12+ weeks past the point of full-dose (Jul 29), and F. prausnitzii's colonization curve specifically favors the longer end of that range. A retest before 4 months from full dose would risk a false "it isn't working" read on an agent that just needs more time.

What to draw alongside it: per the adopted protocol, retest fasting glucose, uric acid, ferritin, and ApoB/LDL at the 90-day mark independent of the full GI-MAP timeline (due ~Oct 2026) — these are the biomarkers most plausibly linked to the dysbiosis findings, and a standalone check there gives an earlier read than waiting for the full stool panel.