{
  "slug": "akkermansia-metabolic-health",
  "file": "akkermansia-metabolic-health.review.md (shipped review)",
  "authoring_model": "Sol 6.1",
  "adversarial_reviewer": "Claude Opus 5.5 (Anthropic)",
  "date": "2026-11-02",
  "level_before": "Γ",
  "level_after": "Γ",
  "confidence_before": "Moderate",
  "confidence_after": "Moderate",
  "pmids_checked": 59,
  "pmids_failed": 0,
  "claims_spot_checked": 15,
  "findings": [
    {
      "id": "A5",
      "severity": "Major",
      "location": "reader-summary.json, limitations (reader-facing)",
      "original": "59 primary reports with 2 selected partial-body examinations; density, heterogeneous controls and incomplete access set the confidence, not correctness.",
      "corrected": "59 primary reports with 2 selected partial-body examinations; evidence density, heterogeneous controls and reproducibility set the confidence, not correctness or access depth.",
      "evidence": "Exact boilerplate struck from the score card by the 2026-10-14 pass (finding A4) under lead policy that access depth never justifies a band. final_score.confidence_rationale now reads 'Moderate reflects evidence density and reproducibility, not access depth', so the lay-facing reader summary states the opposite basis for the same band."
    },
    {
      "id": "A6",
      "severity": "Minor",
      "location": "reader-summary.json, redteamReviewed",
      "original": "\"redteamReviewed\": false",
      "corrected": "\"redteamReviewed\": true",
      "evidence": "The review banner states it was adversarially reviewed by Claude Opus 5.5 on 2026-10-14 with 4 corrections applied, and carries an ADVERSARIAL REVIEW AND APPROVAL section with findings A1-A4. The flag contradicts the shipped document."
    },
    {
      "id": "A7",
      "severity": "Minor",
      "location": "COMPETING MECHANISTIC MODELS closing sentence; AT A GLANCE 'Best-supported rival'",
      "original": "The tournament does not alter the nominal level. Compatible pathways and flexible alternatives are not automatically proved by explaining observations.",
      "corrected": "Same, plus: ranks measure how well each model accounts for the evidence bearing on the registered claim, not each model's own HSEF level; rank 3 carries Δ for its own dietary scope while ranking below two Γ models on the registered claim.",
      "evidence": "Score card competing_models: M3 'Diet and baseline host/community state drive response' has rank_in_tournament 3 with hsef_level_for_this_model Δ, above the Γ of the nominal M1 (rank 2) and of the rank-1 named rival M2. Nothing printed distinguishes rank from level, so the tournament reads as internally inconsistent to a reader."
    },
    {
      "id": "A8",
      "severity": "Minor",
      "location": "Upward threshold ladder, SC5 Δ rung",
      "original": "SC5 Δ: “Requires at least one intervention changing lifespan or healthspan in a way consistent with the hypothesis.” Satisfied: yes. ",
      "corrected": "SC5 Δ: “Requires at least one intervention changing lifespan or healthspan in a way consistent with the hypothesis.” Satisfied: yes. The 90-person randomized maintenance trial (PMID42120725) reports 1.2 versus 3.2 kg regain over 24 weeks after an eight-week diet lead-in; the crossing is scoped to that maintenance setting and is unreplicated.",
      "evidence": "Every other rung prints a next-step or evidence sentence after the verdict; this line terminates after 'Satisfied: yes.' with a trailing space. The one rung in the review where a Δ is actually crossed names no study, effect size or scope."
    },
    {
      "id": "A9",
      "severity": "Minor",
      "location": "COMPETING MECHANISTIC MODELS rank 4; SC3 cap audit rationale",
      "original": "Early heat-killed failure, later pasteurized improvements, novel-species live preference and direct yogurt contrasts map functional differences. / early heat-killed cells fail",
      "corrected": "Scope the failure to the single preparation tested in 2013 and note the later non-viable positives, e.g. 'the one heat-killed preparation tested in 2013 failed, while later non-viable preparations succeeded in other species'.",
      "evidence": "Contradicted by the review's own inventory: [7] heat-killed EB-AMDK19 suppressed weight gain, fat deposition and triglyceride rise in beagles over 12 weeks, its abstract stating the heat-killed strain 'exerts similar protective effects as the live bacterium in mice' (PMID 39264284, DOI 10.1080/1745039X.2024.2397221), and [44] heat-killed Akkermansia improved metabolic and muscle outcomes in STZ mice (PMID 39429872). Both source cards state this correctly; the generalisation is in the model summary and cap rationale above them. SC3 remains Γ either way."
    },
    {
      "id": "A10",
      "severity": "Editorial",
      "location": "REFERENCES, entry 25",
      "original": "25. Asemi Z, Zare Z, Shakeri H et al. ... Access: abstract_metadata. Study: PMID23899653.",
      "corrected": "Same, plus 'Editorial notice: expression of concern, PMID 37334997.'",
      "evidence": "The notice appears in REVIEWER NOTES, on source card [25] and in the score card status_flags, but not on the reference-list line. Confirmed against PubMed: Expression of Concern, PMID 37334997, Ann Nutr Metab 2023;79(2):107, DOI 10.1159/000529287, publication types Editorial and Expression of Concern. This also closes the 2026-10-14 pass's residual limitation that the notice could not be confirmed."
    },
    {
      "id": "A11",
      "severity": "Editorial",
      "location": "PROPOSED DECISIVE EXPERIMENTS; decisive-perturbation tally sentence",
      "original": "... secondary mechanistic leads.. Controls: ... responders.. Supporting: ... evidence.. Falsifying: ... results.. / Tally: 1 positive qualifying study units",
      "corrected": "Single full stops in all four places; 'Tally: 1 positive qualifying study unit'.",
      "evidence": "Renderer artefact and number/noun disagreement. No content added or removed."
    }
  ],
  "rejected_findings": [
    "Reference print/issue years differing from the PubMed date (refs 6, 12, 28, 33, 35, 37, 38, 42, 55): each checked against the journal citation and correct as printed (e.g. Gut Microbes 17(1):2438823 2025; J Acad Nutr Diet 120(4):650-659 2020; Nat Med 23(1):107-113 2017). Declared non-error.",
    "PubMed title fields dropping italicised species or strain names ('PasteurizedHB05' at ref 3, 'Muc' through the Mount abstract, ref 48's title reduced to a full stop). The review is more accurate than the metadata.",
    "Template bleed: none found. 'NAD', 'ROS', 'senescence', 'mitochondrial' and 'lifespan/healthspan' occur only inside quoted rubric 06 criteria with explicit 'Not triggered literally' dispositions, and the body records that lifespan is not imposed on the claim.",
    "Rubric 06's literal lifespan criterion quoted at the SC2 and SC5 Δ rungs: same disposition as the prior pass; rewriting a quoted framework criterion would be the larger error. Recorded as a residual limitation.",
    "Tally arithmetic: 1 positive and 42 screened discriminating contrasts match causal_tally_audit and the 42 inventory records flagged discriminating_contrast; the eight printed inventory entries ([1],[2],[3],[4],[6],[39],[40],[43]) are exactly the eight records flagged decisive_perturbation, a different set, and the review does not conflate them.",
    "SC4 at Α and the weakest-link sensitivity Α: SC4 is a stated unsupported generic-to-named-product inference, so weakest-link Α is arithmetically correct; the headline is set by the declared primary-subclaim rule and the Α sensitivity is not presented as the grade.",
    "Momentum 'Mechanistic only': correct; no named pending trial is asserted and the 2026 trials are described as completed.",
    "No collective-author defect, empty DOI field, glued-numeral or fused-word defect in the review's own prose; glued strings occur only inside PubMed-sourced titles.",
    "Missing contradictory literature: a targeted PubMed search returns only PMID42343233 and PMID42120725 as pasteurized-MucT human trials, both cited, and no bearing meta-analysis. Nothing contradictory is omitted.",
    "Stability 'Stable': argued both ways and left as published. Two pivotal 2026 trials entered the base inside the review's window, one discovered mid-session and forcing post-discovery SC5 at 16:18 against a 16:06 registration, but the dimension keys on a named pending result and none was identified.",
    "Levels: SC1-SC5 and the headline Γ follow from the quoted caps, boundary_check_BGD and the declared aggregation rule; M/S/D did not move a letter (role_of_mechanism.tournament_used_as_tiebreaker false)."
  ],
  "verification_detail": {
    "pmids_resolved": "59/59 reference PMIDs resolve to the stated first author, journal and DOI; the excluded conceptual record PMID41503894 also resolves (Kim HS, Gut Microbes 2026, DOI 10.1080/19490976.2025.2612582).",
    "editorial_notices_found": [
      "Reference 25 (PMID23899653): Expression of Concern PMID 37334997, Ann Nutr Metab 2023;79(2):107, DOI 10.1159/000529287 - confirmed, already disclosed in REVIEWER NOTES, source card and score card."
    ],
    "retractions_found": [],
    "full_text_used": [
      "PMC6699990 (PMID31263284) - confirmed 40 enrolled / 32 completed, 11/12/9 arms, baseline insulinemia P=0.014 and insulin sensitivity P=0.007 versus placebo, 'Fasting glycemia was not affected', 'HbA1c was not modified by the supplementation'."
    ],
    "claims_checked": [
      "2019 MucT pilot enrolment, completion, baseline imbalance, insulin sensitivity +28.62% P=0.002, insulinemia -34.08%, cholesterol -8.68%, weight -2.27 kg P=0.091, unchanged fasting glycemia and HbA1c",
      "2026 maintenance trial n=90, 8-week low-energy diet >=8% loss, 24-week maintenance, 1.2 vs 3.2 kg P=0.012",
      "142-person multicenter trial, 4 months, ITT Matsuda primary null, exploratory low-baseline benefits, weight P=0.06, trunk fat P<0.05",
      "HB05 trial n=100 aged 60+, 12 weeks, left extensor peak torque P=0.0103 and per body weight P=0.0052, follistatin P=0.0063",
      "AKK-WST01 n=58, 12 weeks, no significant between-group HbA1c or weight difference, low-baseline benefit with colonization",
      "WBF-011 n=76, glucose total AUC -36.1 mg/dL/180 min P=0.0500, A1c -0.6",
      "Plovier: pasteurization enhanced effect, Amuc_1100 TLR2 interaction, partly recapitulates",
      "Everard 2013: live reversed HFD phenotype, heat-killed did not",
      "Inulin pilot: weight -2 kg, lower insulin units, HbA1c and CGM unchanged, no rise in Akkermansia",
      "Allin: prediabetic faecal transfer did not impair glucose regulation in recipient mice",
      "Van Hul: Subdoligranulum variabile supplementation produced no significant difference in any obesity/diabetes hallmark",
      "Postbiotic-yogurt trial: waist, waist-to-height, body fat per cent, AST, appetite versus control; no physical-fitness contrast reported",
      "Hong: heat-killed EB-AMDK19 suppressed weight and fat gain in beagles",
      "Kumar: novel Akkermansia species, live form more pronounced",
      "Choi: pasteurized more potent than live for glucose tolerance"
    ],
    "scoring_rederived": "Caps precede aggregation; boundary_check_BGD ceiling Γ for SC1-SC4 and Δ for SC5; declared primary-subclaim rule returns SC2 = Γ = printed headline; printed sensitivities (weakest-link Α, primary-subclaim Γ, scope-matched Γ) consistent with the matrix; tally and decisive-perturbation set match the score card; two falsification patterns met with no downgrade, as printed; M/S/D does not alter the level.",
    "tournament_fairness": "All four models carry named supporting and contradictory evidence; the dietary and strain/processing alternatives are genuine rivals, not strawmen. Presentation defect only, recorded as A7."
  },
  "validator": "Prior package validator: PASS - 0 fail, 0 warn, 30 pass. Not re-run in this pass; no file was changed.",
  "scorecard_changed": false,
  "review_file_changed": false,
  "residual_limitations": [
    "Two of 59 records examined at partial-body depth, as the banner states; this pass used the pilot's PMC full text for verification only and claims no complete-paper audit.",
    "Rubric 06's literal aging wording remains in the threshold-ladder quotations.",
    "Scite verification unavailable in this pass (provider monthly limit reached); notice screening rests on PubMed publication types and comments-corrections records."
  ]
}