[2026-08-01] Fitness extension A — cardiorespiratory fitness (HIIT/zone-2, RHR, HRV)

  • Searched: WebSearch — HIIT vs moderate continuous training VO2max meta-analysis older adults; resting heart rate all-cause cardiovascular mortality dose-response meta-analysis; heart rate variability all-cause mortality prediction systematic review general population; exercise interval vs continuous training mortality cardiovascular events clinical outcomes; zone 2 lactate threshold clinical outcomes mortality; Poon 2024 umbrella.
  • Filters: none; design limit = umbrella/SR/MA; deliberately thin (attention-is-an-anti-signal binds hard).
  • Found / screened: ~30 / 4
  • Included (staged): Poon - HIIT Cardiorespiratory Fitness Umbrella 2024 (gold); Aune - Resting Heart Rate Mortality 2017 (gold); Jarczok - Heart Rate Variability Mortality 2022 (gold).
  • Excluded (notable): Zhang 2016 CMAJ RHR (superseded by Aune); Milanovic 2015 staged (superseded by Poon for the keystone slot); Storoschuk 2025 zone-2 (narrative < gold); HIIT cardiac-rehab MAs (surrogate + disease-specific — they document the modality->outcome gap, don’t fill it); Wu/Chu 2026, young fat-loss MAs (surrogate/narrower).
  • Notes: CRF->mortality anchor already HELD (Kodama/Mandsager/Ross), so these answer how to raise it + are RHR/HRV targets. Poon/Oliveira reach only VO2max (surrogate). Aune + Jarczok reach mortality as PREDICTORS. Two type-G gaps confirmed, NOT filled: modality->patient-important-outcome (only surrogate exists); zone-2 / HRV-guided training as decision tools (coaching-lore).

[2026-08-01] Fitness extension B — resistance-training programming (load/volume/failure/injury)

  • Searched: WebSearch — resistance training dose-response network meta-analysis strength hypertrophy Currier; resistance training injury rates systematic review older adults; Pelland resistance training dose response meta-regression volume frequency; high vs low load resistance training strength hypertrophy meta-analysis; training to failure vs non-failure meta-analysis; resistance training adverse events injury systematic review safety; resistance training older adults function falls umbrella; protein timing leucine threshold strength function meta-analysis; ACSM position stand overview of reviews.
  • Filters: none; design limit = umbrella/NMA/SR/MA.
  • Found / screened: ~40 / 7
  • Included (staged): ACSM - Resistance Training Position Stand 2026 (gold, overview of 137 SRs); Currier - Resistance Training Prescription NMA 2023 (gold); Pelland - Resistance Training Dose Response 2025 (gold); El-Kotob - Resistance Training and Health Overview 2020 (gold); Serafim - Resistance Training Safety 2023 (moderate — injury, override logged).
  • Excluded (notable): Lopez 2021 load NMA + Grgic 2022 failure MA — offered as +2 granularity, HELD (not taken; fold into the ceiling finding); Schoenfeld 2017 low-vs-high-load (superseded by Lopez); Refalo 2022 proximity-to-failure (hypertrophy surrogate only); circuit/BFR/machine older-adult MAs (surrogate/niche, absorbed by the ACSM overview); protein-timing MA (kept as a finding = total protein, not a page).
  • Notes: anchor HELD (Momma/Morton/Cruz-Jentoft); the one patient-important programming outcome is INJURY (Serafim + El-Kotob adverse-events). ACSM overview and Currier NMA share their author team -> NOT independent (E-firewall). Pelland gives a located diminishing-returns knee (dose-response prior). Type-G gaps NOT filled: RPE / reps-in-reserve / time-under-tension / CNS-fatigue / rest-days (coaching-lore, surrogate-only); heavy-load injury in older adults specifically; RT-isolated falls.

[2026-08-05] acquisition | Fish + Parkinson’s gold umbrellas (OpenAlex primary; WebSearch/WebFetch complement)

Tools: OpenAlex API (title.search + DOI/type/OA/cite ranking — polite pool, throttled mid-session after rapid unauthenticated bursts; going-forward via bin/openalex.py authenticated). WebSearch = discovery; WebFetch = PMC record for the one OpenAlex throttled on. Acquisition rule applied (MA/SR/umbrella, gold; ask-on-failure).

[2026-08-05] SELECTION RECORD — Alcohol deliverable gaps (BACKFILL)

Backfilled after the four-field policy was written — the acquisition ran earlier the same day and had only auto-logged query lines + the spec acquisition list; this restores fields (3)+(4) to the ledger. Tool: bin/openalex.py. Gaps targeted: B1 drinking pattern, B2 beverage type, B3 matrix/resveratrol, plus the GBD population anchor.

  • B1 pattern — SELECTED Roerecke 2010 (10.1093/aje/kwp451, gold SR+MA): the binge/irregular-heavy vs regular-moderate at fixed volume MA — the pattern gate. Chosen over the older individual studies surfaced by “binge drinking mortality” (Britton 2002 / Mukamal 2005 / Korean cohorts, 2002-2012) and by “drinking pattern coronary heart disease” (all pre-MA singles) — passed over for the MA.
  • Population anchor — SELECTED GBD 2018 (10.1016/S0140-6736(18)31310-2, high/report): the “no safe level” landmark; the co-hit 10.1016/s2215-0366(18)30337-7 (alcohol+drug GBD) was off-question (substance-use disorder burden, not consumption-outcome) -> rejected.
  • B3 matrix — SELECTED Semba 2014 (10.1001/jamainternmed.2014.1582, moderate cohort): resveratrol null; found via “resveratrol levels all-cause mortality older adults” after “resveratrol cardiovascular outcomes randomized” returned 0 (title strictness). Tier-override lodged at ingest (null counterweight).
  • B2 beverage type — DEFERRED, no gold MA located. “wine beer spirits mortality”, “type of alcoholic beverage cardiovascular risk”, “alcohol consumption pattern type coronary” all returned 0 — title.search strictness, NOT absence; Mukamal 2003 (10.1056/nejmoa022095, NEJM, pattern+type) surfaced but is a single cohort and second-order to ethanol grams -> left as the named B2 gap in the deliverable.
  • Fetch: GBD / Roerecke / Semba all BLOCKED (Lancet/OUP/JAMA anti-bot) -> maintainer-drop; all three were later dropped, ingested, and the deliverable compiled. Rumgay 2021 (cancer burden) was already staged, not re-searched.

[2026-08-05] SELECTION RECORD — lit-research for Sleep / Fermented Foods / Soy specs

Curated companion to the auto-logged query lines above (which record terms + hits; this records the reasoning, the rejections, and the negative space). Tool: bin/openalex.py. Standard: gold/high SR/MA/ umbrella/MR/landmark-RCT, cite-count as a relevance prior, symmetric standards. Selection targets the SPEC GAP each source closes. Fetch note: publisher anti-bot blocks the headless downloader for Cell / OUP / ScienceDirect / NEJM / iiarjournals; PLoS + BMJ Open allowed it. Blocked = maintainer-drop, NOT downgraded.

Sleep

  • B2 regularity gate — SELECTED Windred 2023 (10.1093/sleep/zsad253, SLEEP, 232 cites): title is literally the finding the spec predicted (“regularity a stronger predictor of mortality than duration”). Fetch BLOCKED (OUP). The narrower query “sleep regularity index mortality” returned 0 — a title.search strictness artifact (the words aren’t all in the title), NOT literature absence; “sleep regularity predicts mortality risk” surfaced it.
  • B3 hypnotic-harm — SELECTED Kripke 2012 (10.1136/bmjopen-2012-000850, 405 cites): the canonical hypnotic-mortality cohort; matched-cohort design, confounding-by-indication is the live caveat to carry. Fetched to inbox/.
  • B3 melatonin — SELECTED Ferracioli-Oda 2013 (10.1371/journal.pone.0063773, 477 cites) over the older Buscemi 2005 (10.1111/j.1525-1497.2005.0243.x): newer, same question, both OA. Fetched.
  • B3 CBT-I — SELECTED van Straten 2018 (10.1016/j.smrv.2017.02.001, 685 cites) from a 4-hit set; the general CBT-I efficacy MA (chose over the internet-CBT-I-only Zachariae 2015 as broader). BLOCKED.
  • DEFERRED (lower priority / not locked): chronotype MR — the only hit (“morning chronotype and digestive tract cancers” 10.1002/ijc.34284) is too narrow; the Daghlas chronotype->depression MR is the better target, not yet located. B1 quality/architecture + B4 upstream not searched (Part-A + B2/B3 carry the compile).

Fermented Foods

  • RCT anchor — SELECTED Wastyk 2021 (10.1016/j.cell.2021.06.019, Cell, 1320 cites): the interventional
    • mechanistic landmark; found via the actual title (“gut-microbiota-targeted diets…”) after the topical query “fermented foods microbiota immune status” returned 0 (title-strictness, not absence). Preferred the published Cell version over its bioRxiv preprint (10.1101/2020.09.30.321448). BLOCKED (Cell).
  • Fermented-dairy CVD — SELECTED Zhang 2019 (10.1080/10408398.2018.1564019, 87 cites): the fermented-dairy->CVD MA (the full-fat-dairy-paradox leg). BLOCKED.
  • NEGATIVE SPACE: “yogurt consumption type 2 diabetes meta-analysis” and “dairy products incident type 2 diabetes dose-response” both 0 — title.search strictness; a yoghurt/dairy->T2D MA (Gijsbers 2016 AJCN candidate) certainly exists. Left as a secondary target, not a claimed absence.

Soy

  • Hormonal-fear gate — SELECTED Reed 2021 (10.1016/j.reprotox.2020.12.019, 70 cites): the expanded/ updated MA, the definitive null on male reproductive hormones; found via “male reproductive hormones” after “testosterone estrogen men meta-analysis” returned 0. Chose over the 2025 SR (10.1002/fft2.70090, 1 cite — too new/unestablished). BLOCKED (ScienceDirect).
  • Breast-cancer fear — SELECTED Chen 2014 (10.1371/journal.pone.0089288, 211 cites) from a 3-hit set, over Dong 2011 (10.1007/s10549-010-1270-8, 357 cites, not OA) and a 2022 MA (10.21873/invivo.12737, paywalled): Chen is OA + menopausal-status-stratified (the decision-relevant cut). Fetched.
  • Soy-protein->LDL — SELECTED Anderson 1995 (10.1056/NEJM199508033330502, 2058 cites): the landmark; noted a newer MA (Blanco Mejia 2019) updates the magnitude. Chose over the 2006 AmJCardiol MA. BLOCKED (NEJM).
  • DEFERRED (thinner sections): soy->thyroid review; natto/K2->bone — not searched, lower priority.

Fetch outcome (9 selected)

IN inbox/ (3): Kripke, Ferracioli-Oda, Chen. BLOCKED -> maintainer-drop (6): Windred, van Straten, Wastyk, Zhang, Reed, Anderson. All 9 DOIs recorded in the three deliverable specs.

[2026-08-06] SELECTION RECORD — cluster-A acquisition gaps (BMI-MR · apoB-MR · FMT-SR · non-Willett protein)

Research pass for the top acquire targets opened by the cluster-A batch (logged in ingest-queue OPEN ACQUIRE-THREADS 2026-08-06). openalex primary; gold/high gold-gate; fit-to-question = does it close the specific thread.

  • BMI -> mortality Mendelian randomization (Global BMI low-arm needs a genetic instrument):
    • SELECTED 10.1002/oby.22313 — “BMI and Mortality in UK Biobank: Revised Estimates Using Mendelian Randomization” (2018, 94 cites, OA-flagged). The landmark BMI->mortality MR; directly adjudicates whether the U-curve’s low/underweight arm is causal (the exact gap). ACQUISITION: BLOCKED (Wiley 403) -> MAINTAINER-DROP (DOI recorded; NOT downgraded).
    • REJECTED the prostate-cancer-specific MR (10.1007/s10552-015-0654-9, off-question) and the 2024 figshare dataset (not a paper).
  • Independent-lineage LDL/apoB causality MR (the nucleus is lineage-capped: Ference->Marston->CTT):
    • SELECTED + FETCHED 10.1371/journal.pmed.1003062 — Richardson et al, “Evaluating the relationship between circulating lipoprotein lipids and apolipoprotein B and coronary heart disease risk” (PLoS Med 2020, 941 cites, OA). The Bristol/Davey-Smith multivariable MR establishing apoB as the causal lipid trait over LDL-C/TG — genuinely OFF the Ference lineage -> the type-E independence the nucleus lacked. ACQUISITION: FETCHED to inbox/.
    • DEFERRED 10.1093/ije/dyaa216 (Zuber high-throughput MV-MR, same finding/era — a second-copy echo of Richardson; hold unless Richardson proves insufficient).
  • Non-Willett animal-vs-plant protein cohort (Naghshi + Seidelmann both Willett/NHS-HPFS):
    • SELECTED 10.1001/jamainternmed.2019.2806 — Budhathoki et al, JPHC Japanese cohort (2019, 203 cites, OA-flagged). A genuinely independent (non-Willett, non-Western) cohort on animal/plant protein -> mortality. ACQUISITION: BLOCKED (JAMA anti-bot) -> MAINTAINER-DROP (DOI recorded).
    • REJECTED bmj.m2412 (= Naghshi, ALREADY HELD); DEFERRED 10.1093/ije/dyy030 (Western pattern-cohort — Budhathoki is the cleaner independent lineage).
  • Cochrane FMT-CDI systematic review (pool van Nood’s wide early-stopped CI; MINOR — RCT held):
    • DEFERRED all — the well-cited MAs (10.1007/s10620-020-06185-7 95 cites; 10.1007/s10096-021-04163-z) are paywalled (oa=False); the only OA option (10.7759/cureus.90614, Cureus, 2 cites) is too weak to clear the ratchet on a gap this minor. Not worth a maintainer-drop now; revisit if the microbiome deliverable needs the pooled CI.

Net: 1 fetched (Richardson apoB-MR -> inbox/, hand off to process inbox); 2 maintainer-drops (BMI-MR oby.22313, Budhathoki jamainternmed.2019.2806 — both DOIs recorded); 1 gap deferred (FMT-SR, weak OA). No source downgraded on a fetch block.

[2026-08-06] SELECTION RECORD — “anything else to acquire?” scan (standing acquire-threads)

Probed the two strongest standing OPEN ACQUIRE-THREADS for gold sources.

  • Sarcopenia / low muscle mass -> mortality (elderly-protein cluster’s missing hard-outcome leg):
    • SELECTED 10.1016/j.exger.2021.111461 — “Low muscle mass in older adults and mortality: A systematic review and meta-analysis” (Exp Gerontol 2021, 59 cites, gold SR-MA). Closes a real decision-question (does low muscle mass predict mortality, magnitude). ACQUISITION: PAYWALLED (no OA pdf) -> MAINTAINER-DROP (DOI recorded; NOT downgraded).
  • Nutrigenetics / gene-diet interaction SR-MA (the twice-confirmed [G] responder gap):
    • NO GOLD SOURCE FOUND — a gene-diet-interaction / personalized-nutrition-by-genotype gold SR-MA does not surface (0 hits on the systematic queries; only an obscure low-cite article). This is a LITERATURE gap (insufficient-evidence at the gold level), NOT a fabric gap to fill by acquiring a weak source — symmetric standards forbid lowering the bar for a gap we want filled. Keep as a named insufficient-evidence gap; re-probe as the field matures, or route specific modifiers (FADS/APOE/CYP1A2) through Challenge first. Net: 1 maintainer-drop (sarcopenia MA); nutrigenetics confirmed thin (no acquire). Other standing threads (GLP-1 primary-prevention CVOT, MASLD-AASLD, acrylamide/HCA, HIIT-adherence) left for a later targeted pass.
  • [2026-08-06] openalex search “egg consumption cardiovascular disease mortality meta-analysis” -> 2 hits: https://doi.org/10.1093/jn/nxac105; https://doi.org/10.51745/najfnr.10.21.22-34
  • [2026-08-06] openalex search “egg consumption dose response cardiovascular” -> 2 hits: https://doi.org/10.1007/s00394-020-02345-7; https://doi.org/10.1093/jn/nxac105
  • [2026-08-06] openalex fetch “10.1007/s00394-020-02345-7” -> 0 hits: (none)

[2026-08-06] SELECTION RECORD — eggs / dietary cholesterol -> CVD (the source-review keystone gap)

The one clean genuinely-absent keystone from the 2026-08-06 source-portfolio review (high public-interest, decision-relevant, nothing held). Promoted from the [G]-route-through-Challenge list — a gold MA exists, so it is a real acquire, not insufficient-evidence.

  • SELECTED 10.1007/s00394-020-02345-7 — “Egg consumption and cardiovascular risk: a dose-response meta-analysis of prospective cohort studies” (Eur J Nutr 2020, 64 cites, gold dose-response MA — the right lighthouse; pools the cohorts). ACQUISITION: PAYWALLED (Springer, no OA pdf) -> MAINTAINER-DROP (DOI recorded; NOT downgraded). Author TBC on drop.
  • REJECTED 10.51745/najfnr.10.21.22-34 (2026, 0 cites, obscure journal — too weak); DEFERRED 10.1093/jn/nxac105 (2022 J Nutr MA, oa=False, smaller — the dose-response MA supersedes).
  • NOTE the contested-headline reference: Zhong 2019 (JAMA, 10.1001/jama.2019.1572) is the single pooled-cohort analysis that reignited the eggs/dietary-cholesterol debate (found a POSITIVE association) — a potential TENSION source vs the mostly-null MAs. Not acquired now (single analysis, the MA is the lighthouse); flag for maintainer-drop only if a tension needs the both-sides sources. Net: 1 maintainer-drop (Godos egg-CVD dose-response MA). No source downgraded.
  • [2026-08-06] openalex fetch “10.1056/NEJM199704173361601” -> 0 hits: (none)
  • [2026-08-06] openalex fetch “10.1056/NEJM200101043440101” -> 0 hits: (none)
  • [2026-08-06] openalex search “dairy consumption cardiometabolic umbrella review umbrella review” -> 0 hits: (none)
  • [2026-08-06] openalex search “dairy type 2 diabetes cardiovascular meta-analysis meta-analysis” -> 2 hits: https://doi.org/10.1093/nutrit/nuz006; https://doi.org/10.1016/j.idairyj.2019.02.019
  • [2026-08-06] openalex search “milk consumption all-cause mortality meta-analysis meta-analysis” -> 1 hits: https://doi.org/10.1007/s10654-017-0243-1
  • [2026-08-06] openalex search “milk dairy fracture risk meta-analysis meta-analysis” -> 1 hits: https://doi.org/10.1080/10408398.2019.1590800
  • [2026-08-06] openalex search “dairy milk prostate cancer meta-analysis meta-analysis” -> 0 hits: (none)
  • [2026-08-06] openalex search “dairy calcium prostate cancer risk dose-response” -> 0 hits: (none)
  • [2026-08-06] openalex search “fructose nonalcoholic fatty liver disease meta-analysis meta-analysis” -> 0 hits: (none)
  • [2026-08-06] openalex search “sugar sweetened beverages type 2 diabetes meta-analysis meta-analysis” -> 4 hits: https://doi.org/10.1007/s10654-020-00655-y; https://doi.org/10.17615/rpf0-7566; https://doi.org/10.3390/nu13082636; https://doi.org/10.17615/jf7b-nq36
  • [2026-08-06] openalex search “sodium intake cardiovascular mortality meta-analysis cohort meta-analysis” -> 0 hits: (none)
  • [2026-08-06] openalex search “high protein intake kidney function chronic kidney disease meta-analysis meta-analysis” -> 0 hits: (none)
  • [2026-08-06] openalex fetch “10.1007/s10654-017-0243-1” -> 0 hits: (none)
  • [2026-08-06] openalex fetch “10.3390/nu13082636” -> 0 hits: (none)

[2026-08-06] SELECTION RECORD — value-add scan over the (new) acquisition-gated deliverables

Research pass to assess which deliverable-gap acquisition targets are sufficiently value-add (gold/high + ratchet-advancer + fit-to-question), NOT to acquire indiscriminately. Per gated deliverable:

DAIRY (BLOCKED; 4 targets) — 3 of 4 are genuine value-add; the 4th stays a named gap

  • SELECTED + FETCHED 10.1007/s10654-017-0243-1 — Guo et al, milk/dairy -> CVD + all-cause mortality dose-response MA (Eur J Epidemiol 2017, 398 cites). KEYSTONE, and value-add REGARDLESS of the Dairy decision (milk-mortality is a standalone high-interest question; feeds the dairy-matrix claim pages). -> inbox/.
  • SELECTED -> MAINTAINER-DROP 10.1093/nutrit/nuz006 — dairy -> T2D SR-MA (Nutr Rev 2019, 42 cites, OUP-walled). The cardiometabolic-by-category backbone. Value-add IF Dairy is compiled.
  • SELECTED -> MAINTAINER-DROP 10.1080/10408398.2019.1590800 — milk/dairy -> osteoporosis/hip-fracture MA (Crit Rev Food Sci Nutr 2019, 90 cites, paywalled). The milk-fracture-paradox source. Value-add IF Dairy is compiled.
  • DEFER (LOW value-add) dairy -> prostate: no clean MA surfaced, and WCRF 2018 (HELD) already grades dairy probable-increased for prostate — a dedicated MA would add only a magnitude to a held direction. Keep as a named gap, do not acquire.
  • Verdict: Dairy is UNBLOCKABLE with 3 acquisitions (1 fetched + 2 maintainer-drop). But the value-add is CONDITIONAL on compiling Dairy (maintainer’s earlier ambivalence) — only the milk-mortality MA is unconditional.

ADDED SUGAR / SSB (2 targets) — 1 value-add, 1 thin

  • SELECTED -> MAINTAINER-DROP 10.1007/s10654-020-00655-y — sugar/artificially-sweetened beverages -> obesity/T2D dose-response MA (Eur J Epidemiol 2020, 338 cites; paywalled). The SSB->T2D backbone. OA alternative 10.3390/nu13082636 (MDPI, 136 cites) fetch-failed today too.
  • DEFER (thin) fructose -> MASLD/NAFLD: no gold SR surfaces; the leg is mechanism-forward + small controlled-feeding trials -> INSUFFICIENT-EVIDENCE, name as gap, do not acquire weak (symmetric standards).

SODIUM (2 targets, SECOND-ORDER) — DEFER both

  • The deliverable is already compilable (He/Huang/WHO/SSaSS/Aburto + SPRINT/BPLTTC held). Sodium-alone hard-outcome MA (contested PURE-vs-WHO field) and a J-curve MR are second-order — DEFER; name as gaps.

PROTEIN (1, MINOR) — optional

  • A protein -> kidney-function MA (Devries-type: no harm to healthy kidneys) would firm the renal contraindication vs safety; MODERATE value-add, near-compilable without it. Optional acquire.

DIETARY FAT TYPES — no acquisition needed (mostly covered).

Net: 1 fetched (Guo milk-mortality -> inbox); 3 maintainer-drops if the deliverables proceed (dairy-cardiometabolic, dairy-fracture, SSB-T2D); the rest named-gaps (dairy-prostate via WCRF, fructose-MASLD thin, sodium second-order). No weak source acquired.

  • [2026-08-06] openalex search “fructose liver fat controlled feeding systematic review” -> 0 hits: (none)
  • [2026-08-06] openalex search “fructose nonalcoholic fatty liver disease systematic review” -> 1 hits: https://doi.org/10.3945/ajcn.114.086314
  • [2026-08-06] openalex search “sugar sweetened beverages fructose fatty liver meta-analysis” -> 0 hits: (none)
  • [2026-08-06] openalex fetch “10.3945/ajcn.114.086314” -> 0 hits: (none)

[2026-08-06] SELECTION RECORD — fructose -> MASLD RE-SEARCH (corrects the earlier “thin” call)

Re-searched at maintainer request. The earlier value-add scan’s “no gold SR / insufficient-evidence” verdict was WRONG — an over-strict openalex query (double “meta-analysis” + title-word strictness) missed the landmark. Corrected:

  • SELECTED 10.3945/ajcn.114.086314 — Chung et al, “Fructose, high-fructose corn syrup, sucrose, and nonalcoholic fatty liver disease or indexes of liver health: a systematic review and meta-analysis” (AJCN 2014, 254 cites, gold SR-MA of controlled-feeding trials). The decision-relevant finding (verify at ingest): the fructose->liver-fat effect is largely a HYPERcaloric/overfeeding effect — ISOCALORIC fructose exchange does not clearly raise intrahepatic lipid — which DEFUSES the “fructose is uniquely hepatotoxic” claim rather than confirming it. Genuine value-add, not insufficient-evidence. ACQUISITION: PAYWALLED (AJCN/OUP) -> MAINTAINER-DROP (DOI recorded).
  • Companion (not needed unless the isocaloric/hypercaloric split wants a second source): Chiu et al 2014 (Eur J Clin Nutr) controlled-feeding SR-MA — same distinction. Lesson logged: a single over-strict openalex 0-hit is a query-tuning signal, NOT a literature-absence claim (the standing rule); the earlier scan should have loosened before calling it thin. The Sugar spec’s fructose leg is upgraded from named-gap to identified-acquire.
  • [2026-08-06] openalex search “protein intake older adults muscle mass strength meta-analysis meta-analysis” -> 0 hits: (none)
  • [2026-08-06] openalex search “sunscreen skin cancer prevention randomized” -> 0 hits: (none)
  • [2026-08-06] openalex search “sun exposure avoidance all-cause mortality cohort” -> 2 hits: https://doi.org/10.1111/joim.12251; https://doi.org/10.3410/f.718338950.793496044
  • [2026-08-06] openalex search “sunscreen melanoma randomized trial Nambour” -> 0 hits: (none)
  • [2026-08-06] openalex search “regular sunscreen use cutaneous melanoma trial” -> 0 hits: (none)
  • [2026-08-06] openalex search “ultraviolet radiation melanoma systematic review meta-analysis meta-analysis” -> 0 hits: (none)
  • [2026-08-06] openalex fetch “10.1111/joim.12251” -> 0 hits: (none)
  • [2026-08-06] openalex search “regular sunscreen melanoma reduced follow-up” -> 3 hits: https://doi.org/10.1200/jco.2010.28.7078; https://doi.org/10.3410/f.7222956.9563054; https://doi.org/10.3410/f.7222956.14688056
  • [2026-08-06] openalex fetch “10.1200/jco.2010.28.7078” -> 0 hits: (none)

[2026-08-06] SELECTION RECORD — four maintainer research asks (mid-ingest-seq)

  1. Older-adult protein (per-meal leucine / ~1.2-1.6 g/kg) — the Big Rocks (Elderly) “top acquirable gap”: ALREADY STAGED, NOT ACQUIRABLE. Bauer PROT-AGE 2013 (the g/kg guidance), Katsanos 2006 (leucine threshold), Moore 2014 (per-meal dose) are all staged. The spec’s gap statement is STALE (predates Bauer staging) — like the DASH-already-cut catch. Action: ingest the staged trio; no new source.
  2. UV / sun exposure / sunscreen / skin cancer — a NEW peripheral-lifestyle domain (currently OPEN-thin): a benefit-risk question (sun: vit-D/BP/mood + a mortality signal vs skin-cancer harm). Two landmarks identified:
    • SELECTED -> MAINTAINER-DROP 10.1111/joim.12251 — Lindqvist, “Avoidance of sun exposure is a risk factor for all-cause mortality” (J Intern Med 2014/2016, 111 cites; Swedish MISS cohort). The contrarian benefit-of-sun finding — CONFOUNDED single high-latitude cohort, needs the artifact-caveat framing.
    • SELECTED -> MAINTAINER-DROP 10.1200/jco.2010.28.7078 — Green, “Reduced Melanoma After Regular Sunscreen Use: Randomized Trial Follow-Up” (JCO 2010/2011, 799 cites; the Nambour RCT). THE sunscreen-> melanoma RCT (the only one). ASCO-blocked.
    • GAP still open: a UV/sunbed->melanoma SR-MA (the harm quantification; IARC UV is Group 1) — looser search owed. This is a new-domain cluster; if pursued it would seed a “Sun exposure and skin” question/deliverable.
  3. Per-factor dementia weights (Livingston Table 1): NOT an acquisition — Livingston 2024 is INGESTED; the 14 per-factor PAF values live in its Table 1 (a figure/table not yet extracted; the [AWAITS (Livingston)] on Dementia Prevention and Modifiable Risk Factors). This is an EXTRACTION/Revisit gap on the HELD source, not a new source. Action: a Revisit / re-read of Livingston Table 1, not Research. Net: 0 new acquirable for asks 1+3 (already-held/staged); 2 maintainer-drops for the UV domain (Lindqvist, Green).
  • [2026-08-06] openalex fetch “10.1136/bmj.e4757” -> 0 hits: (none)

[2026-08-06] SELECTION RECORD — sun / UV / sunscreen / skin cancer (new benefit-risk domain)

Assembled the sun/UV cluster (previously OPEN-thin in the lighthouse-index). A benefit-risk question: sun’s mortality/vit-D/BP signal vs melanoma/NMSC harm, stratified by intermittent-vs-chronic exposure + skin type.

  • STAGED Green - Sunscreen Melanoma Randomized Trial Nambour 2011 (10.1200/JCO.2010.28.7078, high·rct) — the ONLY sunscreen->melanoma RCT; the sunscreen-efficacy leg. (DOI was correct all along; earlier fetch failed on ASCO’s bot-block, not a wrong DOI.)
  • STAGED Lindqvist - Sun Exposure Major Causes of Death 2016 (10.1111/joim.12496, high·cohort) — MISS-cohort 20-yr competing-risk; sun-AVOIDANCE -> higher mortality (the contrarian benefit-of-sun signal). CONFOUNDED single high-latitude cohort — carries the artifact caveat. Supersedes the 2014 original (joim.12251).
  • IDENTIFIED -> MAINTAINER-DROP 10.1016/j.ejca.2004.10.016 (Gandini 2005, Eur J Cancer, 1257 cites) — the definitive sun-exposure->melanoma harm MA; carries the decision-critical INTERMITTENT (risk) vs CHRONIC (no increase) distinction. Paywalled.
  • IDENTIFIED -> MAINTAINER-DROP 10.1136/bmj.e4757 (Boniol 2012, BMJ, 638 cites) — sunbed/indoor-tanning -> melanoma dose-response; the artificial-UV harm signal. BMJ full-text bot-blocked despite OA. Net: 2 staged (Green, Lindqvist-2016); 2 maintainer-drops (Gandini harm-MA, Boniol sunbed-MA). Cluster is a candidate new “Sun exposure and skin” question/deliverable (not yet specced).

[2026-08-06] openalex queries run —no-log during the book-2 ingest (hand-logged per the four-field policy)

These ran with —no-log to avoid writing to the branch while the Malmir ingest held it; recorded here so the query/results audit half is not lost. Channel: openalex.

  • search “reduced melanoma regular sunscreen randomized trial follow-up” -> 1: 10.1200/jco.2010.28.7078 (Green, 799c)
  • search “avoidance sun exposure risk factor major causes death competing risk” -> 3: 10.1111/joim.12496 (Lindqvist 2016, 148c); 10.1111/joim.12613; 10.1111/joim.12538 (responses)
  • resolve 10.1111/joim.12251 -> Lindqvist 2014 (all-cause, 111c) [the original; superseded by 12496]
  • search “cutaneous melanoma sun exposure” -> 4: 10.1016/j.ejca.2004.10.016 (Gandini II sun-exposure, 1257c); + 3 non-MA
  • search “risk factors cutaneous melanoma meta-analysis” -> 4: ejca.2004.10.016 (II); 10.1016/j.ejca.2004.10.015 (I naevi); 10.1016/j.ejca.2005.03.034 (III); 10.1016/j.ejca.2009.05.009 (site)
  • search “melanoma attributable sunbed use” -> 3: 10.1136/bmj.e4757 (Boniol sunbed MA, 638c, OA); + 2 non-MA
  • ZERO-HIT (title.search strictness, not literature absence): “Lindqvist sun exposure mortality Melanoma Southern Sweden”; “sun exposure melanoma risk meta-analysis”; “indoor tanning sunbed melanoma systematic review meta-analysis”; “ultraviolet cutaneous melanoma risk pooled analysis”; “farmed wild fish nutritional contaminant health”; “farmed versus wild salmon fatty acid contaminant” (wild-vs-farmed confirmed: no gold MA exists).

[2026-08-06] sun/UV cluster — UPDATE (maintainer scoping)

  • STAGED Gandini - Cutaneous Melanoma Sun Exposure Meta-Analysis 2005 (10.1016/j.ejca.2004.10.016, gold MA, 1257c) — the sun-exposure->melanoma harm leg (intermittent-vs-chronic; the key nuance).
  • NOT WANTED 10.1136/bmj.e4757 (Boniol sunbed->melanoma) — maintainer: “not interested in sunbed risk per se”; the question is NATURAL sun benefit-risk, not artificial-UV. Routed to inbox/_hold, un-staged. (Registry note: leave un-registered.) Sun/UV cluster now STAGED (3): Green (sunscreen RCT), Lindqvist-2016 (sun-avoidance mortality), Gandini (sun->melanoma harm). Enough to seed a natural-sun benefit-risk question; sunbed excluded per scope.
  • [2026-08-06] openalex search “fish consumption chronic disease umbrella review” -> 1 hits: https://doi.org/10.1093/advances/nmaa029
  • [2026-08-06] openalex search “fish intake health outcomes umbrella review” -> 0 hits: (none)
  • [2026-08-06] openalex fetch “10.1093/advances/nmaa029” -> 0 hits: (none)
  • [2026-08-07] openalex search “protein energy restriction fat free mass meta-analysis” -> 0 hits: (none)
  • [2026-08-07] openalex search “dietary protein weight loss lean mass preservation systematic review” -> 0 hits: (none)
  • [2026-08-07] openalex search “protein caloric restriction resistance trained” -> 1 hits: https://doi.org/10.1123/ijsnem.2013-0054
  • [2026-08-07] openalex search “protein weight loss lean mass” -> 4 hits: https://doi.org/10.3945/ajcn.111.026328; https://doi.org/10.4016/39369.01; https://doi.org/10.1055/s-0034-1398551; ?
  • [2026-08-07] openalex search “protein hypocaloric fat-free mass” -> 0 hits: (none)
  • [2026-08-07] openalex search “protein energy deficit muscle mass” -> 2 hits: ?; ?
  • [2026-08-07] openalex search “protein resistance training energy restriction body composition meta-analysis” -> 0 hits: (none)
  • [2026-08-07] openalex search “high protein diet fat free mass meta-analysis meta-analysis” -> 0 hits: (none)
  • [2026-08-07] openalex search “protein intake weight loss resistance exercise lean meta-analysis” -> 0 hits: (none)
  • [2026-08-07] openalex resolve “10.1123/ijsnem.2013-0054” -> 1 hits: https://doi.org/10.1123/ijsnem.2013-0054
  • [2026-08-07] openalex search “protein resistance training weight loss body composition” -> 1 hits: https://doi.org/10.3945/ajcn.117.158246
  • [2026-08-07] openalex search “dietary protein energy restriction fat free mass” -> 0 hits: (none)
  • [2026-08-07] openalex search “protein overfeeding energy restriction resistance training lean” -> 0 hits: (none)
  • [2026-08-07] openalex search “protein caloric restriction athletes” -> 0 hits: (none)
  • [2026-08-07] openalex search “protein intake muscle mass energy restriction older” -> 0 hits: (none)
  • [2026-08-07] openalex search “protein requirements weight loss” -> 3 hits: https://doi.org/10.7759/cureus.83983; https://doi.org/10.26420/austinjnutrmetab.2021.1106; https://doi.org/10.1002/hep.29801
  • [2026-08-07] openalex resolve “10.1519/SSC.0000000000000888” -> 1 hits: https://doi.org/10.1519/ssc.0000000000000888
  • [2026-08-07] openalex search “protein supplements fat-free mass strength resistance training network meta-analysis” -> 1 hits: https://doi.org/10.1155/tsm2/5557511
  • [2026-08-07] openalex fetch “10.1519/SSC.0000000000000888” -> 0 hits: (none)

[2026-08-07] SELECTION RECORD — protein target during CALORIC DEFICIT / lean-mass preservation

Gap opened by the protein-target Revisit (Morton 1.62 is young-derived, non-obese, non-deficit, and a non-significant knee); caloric-deficit is the genuinely-unheld stratum with real synthesis potential (intersects GLP-1 and Lean Mass + Rapid Weight Loss). Channels: openalex (auto-logged above) + one web-search (hand-logged here).

Query (web, hand-logged): “meta-analysis dietary protein intake resistance training energy restriction fat-free mass retention” — to identify the landmark/newest SR-MA (openalex title-search too strict: multiple 0-hits from strictness not absence — “protein energy restriction fat free mass”, “dietary protein weight loss lean mass preservation”, “protein hypocaloric fat-free mass” all 0 on title-match).

Results (candidates):

  • Refalo et al. 2025 — «Effect of Dietary Protein on Fat-Free Mass in Energy Restricted, Resistance-Trained Individuals: An Updated Systematic Review With Meta-Regression» (Strength & Conditioning Journal). DOI 10.1519/SSC.0000000000000888 · SR + meta-regression · gold/high · oa=True (flag) but fetch PAYWALLED · cites=3.
  • Helms et al. 2014 — «A Systematic Review of Dietary Protein During Caloric Restriction in Resistance…» (IJSNEM). DOI 10.1123/ijsnem.2013-0054 · SR · gold/high · oa=False · cites=124. The landmark.
  • Network MA 2026 — «Which Protein-Based Dietary Supplements Most Effectively Enhance Fat-Free Mass and Strength … A Network Meta-Analysis» (Transl Sports Med). DOI 10.1155/tsm2/5557511 · network-MA · oa=True · cites=2.
  • Primary studies (AJCN 2012 «4 weight-loss diets …»; AJCN 2017 within-day protein distribution) — single RCTs.

Adjudication:

  • Refalo 2025 — SELECTED (the acquire). Ratchet-ADVANCER: first gold on the unheld caloric-deficit stratum, and an updated SR+meta-regression that supersedes Helms 2014. Fit-to-question exact (energy-restricted, resistance-trained, FFM outcome). NOTE it independently corroborates the fabric’s own thread: a LINEAR dose-response (no knee — Route 2), STRONGER when protein is expressed per FAT-FREE MASS (the denominator point), with stratum modifiers (duration, body-fat %, sex). ~1.9 g/kg-BM or ~2.5 g/kg-FFM to minimize FFM loss; higher for FFM gain.
  • Helms 2014 — DEFERRED (superseded). Kept as the landmark for provenance; Refalo is the updated version. Acquire only if the maintainer wants the original for citation depth (oa=False -> would also be maintainer-drop).
  • Network MA 2026 — DEFERRED (different gap). On the GENERAL protein-supplement-type x RT question (a network-MA that could ratchet-advance the Morton-2018 general question), NOT the caloric-deficit gap. Flag for a future general-protein-RET ratchet check; not this acquisition.
  • AJCN 2012/2017 primaries — REJECTED. Single RCTs superseded by the SR.

Acquisition-requested: Refalo 2025 (DOI 10.1519/SSC.0000000000000888) — fetch attempted, PAYWALLED despite OA flag (Strength & Conditioning Journal = LWW/Wolters-Kluwer, anti-bot class). -> MAINTAINER-DROP requested (do NOT downgrade to a weaker source; the source is identified, not lost). Nothing fetched to inbox/. On maintainer drop -> process inbox -> registry row (tier gold, type systematic-review) -> ingest-queue as the caloric-deficit + obese-denominator opener.

[2026-08-07] SELECTION RECORD — “genuinely not held” acquisition run (post lighthouse re-sync)

Seven acquire-targets + 2 named-AWAITS from the 08-07 lighthouse pass’s gap list. Channels: openalex (auto-logged above) + 4 web-searches (hand-logged here). Standard: gold/high, ratchet-ADVANCER, fit-to-question.

Web searches (hand-logged, channel=web):

  • “AASLD MASLD MASH clinical practice guideline 2023 2024” -> AASLD 2023 Practice Guidance (Rinella), the US family vs the held EASL-EASD-EASO 2024; + a Nov-2025 semaglutide-for-MASH update (not the core doc).
  • “GLP-1 receptor agonist primary prevention cardiovascular outcome trial low-risk without diabetes” -> SELECT is SECONDARY-prevention (established CVD, BMI>=27, no diabetes); no dedicated LOW-RISK primary-prevention CVOT exists. SURMOUNT-MMO (pending) is the nearest, at-risk not low-risk.
  • “SURMOUNT-MMO tirzepatide cardiovascular outcomes trial results published” -> results NOT yet published (ongoing, event-driven, ~15,000, primary MACE composite). Only the design/rationale paper is out.
  • “FINGER trial extended follow-up dementia incidence long-term results” -> the 11-year FINGER results exist (Alzheimer’s & Dementia 2025); the extended-FINGER assesses dementia/AD incidence + CV events.

Adjudication + acquisition status:

  • Dietary Acrylamide -> Site-Specific Cancer SR 2022 (10.3389/fnut.2022.875607, gold·SR, Frontiers, OA) — SELECTED + FETCHED to inbox/. The cooking/Maillard-carcinogen axis distinct from the held WCRF red-meat mechanism; a new-aspect opener (no acrylamide-cancer incumbent).
  • Devries - Higher Protein & Kidney Function SR-MA 2018 (10.1093/jn/nxy197, gold·MA, J Nutr/OUP, OA) — SELECTED -> MAINTAINER-DROP (OUP anti-bot 403). The landmark on the actual decision-question — high vs lower protein does NOT change kidney function in healthy adults — the protein->kidney G-gap’s best evidence. (The 1998 AJKD MA is protein restriction in existing CKD = a different question; rejected.)
  • AASLD - MASLD Practice Guidance 2023 (10.1097/hep.0000000000000323, gold·guideline, Hepatology, OA on PMC) — SELECTED -> MAINTAINER-DROP (PMC/doi.org both returned a landing page, not the PDF; browser-fetch also blocked). PDF: pmc.ncbi.nlm.nih.gov/articles/PMC10735173/pdf/nihms-1948413.pdf. The 2nd MASLD guidance family (guidance-null vs held EASL) — a genuine new aspect (US vs EU management divergence).
  • FINGER 11-Year Extended Follow-up 2025 (10.1002/alz70860_106542, high·rct-followup, Alz&Dem/Wiley, OA) — SELECTED -> MAINTAINER-DROP (Wiley pdfdirect 403). Cashes the registered [AWAITS (Ngandu - FINGER Extended Follow-up)]. CAVEAT for process-inbox: verify it reports dementia/AD INCIDENCE (the wanted datum), not only cognitive-change/adherence — FINGER was underpowered for incidence. UPDATE 08-07: the maintainer dropped it; the Wiley pdfdirect download is TRUNCATED (~1078t, abstract-only), NOT a wrong DOI — the DOI 10.1002/alz70860_106542 IS the full Ngandu 2025 Alz&Dem paper; the full text is on PMC12726459. Deferred to inbox/_hold pending the full PMC PDF.
  • Acrylamide/HCA cooking-meat MA (Nutrition Journal 2015, 10.1186/s12937-015-0111-3) — DEFERRED. The HCA/cooking-method leg, but 2015 and substantially a red/processed-meat->CRC MA (that question is CLOSED via WCRF) — low marginal value. The acrylamide SR (SELECTED above) already covers the higher-interest Maillard axis.
  • Sniderman apoB-vs-LDL — REJECTED (echo). apoB-over-LDL is now CLOSED (Marston 2022 + Richardson MR 2020 held); a Sniderman particle-number source is same-school advocacy, not a ratchet-advancer.
  • GLP-1 primary-prevention (low-risk) CVOT — NOT ACQUIRABLE (does not exist). SELECT is secondary- prevention; the low-risk primary-prevention trial is not yet run. Insufficient-evidence, not a fetch target.
  • SURMOUNT-MMO — NOT ACQUIRABLE (unpublished). Stays a named AWAITS; re-check on publication.
  • grip-strength-as-TARGET RCT — NOT ACQUIRABLE (does not exist as posed). Raising grip -> hard outcomes ~ the RT->mortality question (Momma, held); grip is a MARKER not an intervention lever. Insufficient-evidence.
  • fermented-food live-vs-pasteurized viability RCT — 0 hits (title-strictness + likely thin literature). DEFERRED; the interventional viability leg Wastyk does not close. Not confidently a literature-absence yet — re-search with looser terms next pass before declaring the null.

Net: 1 FETCHED (acrylamide SR) · 3 MAINTAINER-DROP (Devries protein-kidney · AASLD MASLD guidance · FINGER 11-yr) · 1 DEFERRED-echo (HCA 2015) · 1 REJECTED-echo (Sniderman) · 3 NOT-ACQUIRABLE/insufficient (GLP-1 primary-prevention · SURMOUNT-MMO · grip-target RCT) · 1 DEFERRED-thin (fermented viability). Next: on the maintainer drops -> process inbox -> register + stage; then the acrylamide SR can Ingest.

  • [2026-08-07] openalex resolve “10.1016/j.advnut.2022.04.045” -> 0 hits: (none)
  • [2026-08-07] openalex search “long-term adherence lifestyle cognitive change 11-year FINGER randomized controlled trial” -> 1 hits: https://doi.org/10.1002/alz70860_106542
  • [2026-08-07] openalex search “effectiveness of a multidomain intervention to reduce dementia preDIVA” -> 0 hits: (none)
  • [2026-08-07] openalex search “multidomain intervention omega-3 cognitive decline MAPT randomised” -> 0 hits: (none)
  • [2026-08-07] openalex search “multidomain dementia prevention pooled individual participant MAPT preDIVA” -> 1 hits: https://doi.org/10.1002/alz.14472
  • [2026-08-07] openalex search “FINGER two-year multidomain intervention seven-year extended follow-up cognition” -> 0 hits: (none)

[2026-08-07] ADDENDUM — FINGER 11-yr not OA -> FALLBACK staged (Coley MAPT+preDIVA pooled)

Resolving the deferred FINGER 11-yr drop: verified across Wiley pdfdirect, PMC (PMC12726459) and Europe-PMC render — all three serve the SAME 574-word, 1-page abstract; NCBI/Europe-PMC classify it article-type= abstract (an AAIC-2025 conference supplement under DOI 10.1002/alz70860_106542). The FULL 11-yr FINGER paper is not OA-available (likely not yet published as a full article). Not a wrong-DOI or a fetch-route problem.

  • Reframed the gap to what it actually is: multidomain lifestyle -> dementia INCIDENCE (hard outcome) — FINGER 2015 (held) is a 2-yr COGNITIVE surrogate; the incidence-outcome multidomain trials are MAPT + preDIVA.
  • SELECTED + FETCHED (Europe-PMC render, full 13pp/7942w, verified Methods/Results/HRs) Coley - Multidomain Dementia MAPT preDIVA Pooled 2025 (10.1002/alz.14472, high·pooled-IPD, OA) — pooled IPD of the MAPT + preDIVA RCTs; responder (route-b) analysis on the incidence axis. Registered + staged to raw/binary/. The better UNIT than the single FINGER follow-up (acquisition-rule: pooled > single trial).
  • IDENTIFIED (not fetched): the preDIVA (Moll van Charante, Lancet 2016) + MAPT (Andrieu, Lancet Neurol 2017) primaries are the incidence anchors but Lancet-family (anti-bot) -> maintainer-drop if the pooled analysis is insufficient. FINGER 11-yr full paper: re-check OA status later; keep the abstract in _hold.
  • [2026-08-07] openalex search “effectiveness multidomain vascular care intervention prevent dementia cluster randomised” -> 2 hits: https://doi.org/10.1016/s0140-6736(16)30950-3; https://doi.org/10.3410/f.726588451.793577000
  • [2026-08-07] openalex search “omega-3 multidomain intervention cognitive function elderly memory complaints randomised placebo” -> 1 hits: https://doi.org/10.1016/s1474-4422(17)30040-6
  • [2026-08-07] openalex fetch “10.1016/S0140-6736(16)30950-3” -> 0 hits: (none)
  • [2026-08-07] openalex search “dietary protein supplementation resistance training muscle hypertrophy dose-response meta-analysis meta-analysis” -> 0 hits: (none)
  • [2026-08-07] openalex search “dose-response relationship protein intake muscle mass systematic review” -> 2 hits: https://doi.org/10.1093/nutrit/nuaa104; https://doi.org/10.1101/2020.02.21.20026252
  • [2026-08-07] openalex search “protein supplementation resistance exercise training muscle mass strength meta-analysis” -> 1 hits: https://doi.org/10.3390/nu15153424
  • [2026-08-07] openalex resolve “10.1093/nutrit/nuaa104” -> 1 hits: https://doi.org/10.1093/nutrit/nuaa104
  • [2026-08-07] openalex search “apolipoprotein B LDL triglyceride multivariable Mendelian randomization coronary heart disease” -> 0 hits: (none)
  • [2026-08-07] openalex search “body mass index all-cause mortality Mendelian randomization” -> 1 hits: https://doi.org/10.1101/2023.10.31.23297612
  • [2026-08-07] openalex fetch “10.1093/nutrit/nuaa104” -> 0 hits: (none)
  • [2026-08-07] openalex search “apolipoprotein B Mendelian randomization coronary artery disease” -> 2 hits: https://doi.org/10.1093/ije/dyaa216; https://doi.org/10.1101/2020.02.10.20021691
  • [2026-08-07] openalex search “testosterone Mendelian randomization cardiovascular fracture outcomes” -> 0 hits: (none)
  • [2026-08-07] openalex search “calcium vitamin D supplementation fracture prevention randomised RECORD trial” -> 0 hits: (none)
  • [2026-08-07] openalex search “endogenous testosterone health outcomes Mendelian randomization” -> 0 hits: (none)
  • [2026-08-07] openalex search “sun exposure melanoma risk prospective cohort meta-analysis” -> 0 hits: (none)

[2026-08-07] SELECTION RECORD — independent “second opinions” for lineage-capped big-rock numbers

Targets: the decision-driving big-rock quantities that rest on ONE author-lineage (the fabric’s own caps). Vet = author-INDEPENDENCE (candidate must share no author with the capped lineage), on MERIT not OA.

  • Protein ~1.62 g/kg (Phillips/McMaster lineage: Morton + Devries) → SELECTED: Tagawa 2020 «Dose–response relationship between protein intake and muscle mass increase: a SR-MA» (Nutr Rev, 10.1093/nutrit/nuaa104, gold·MA, 116c). Japanese group, NO Phillips — a genuine independent second opinion, and it lands a different breakpoint (~1.3 g/kg vs Morton’s 1.62) → the informative disagreement case. OUP-blocked (403) → MAINTAINER-DROP. OA interim if wanted: medRxiv preprint 10.1101/2020.02.21.20026252 (same study, pre-peer-review — offered, NOT silently substituted).
  • LDL/apoB causal supremacy (Ference lineage: CTT + Marston + Richardson) → SELECTED: Zuber 2021 «High-throughput multivariable MR prioritizes apolipoprotein B as key lipid risk factor for CAD» (IJE, 10.1093/ije/dyaa216, high·MR, 91c). Authors Zuber/Gill/Ala-Korpela/Langenberg/Butterworth/Bottolo/Burgess — the Cambridge/Burgess MR group, NO Ference; independently reproduces apoB-supremacy. CAVEAT (verify at ingest): likely shares UKBB/CARDIoGRAM DATA + one methods author (Ala-Korpela) with Richardson → independent GROUP, not fully-independent DATA; a partial E-lift (medium→ toward high), not a clean one. OUP-blocked → MAINTAINER-DROP. OA interim: medRxiv 10.1101/2020.02.10.20021691.
  • BMI low-arm-artifact (Wade ∩ Global BMI: Davey Smith/Sattar) → IDENTIFIED, LOW-PRIORITY: a Burgess-group HUNT+UKBB BMI-mortality MR (medRxiv 10.1101/2023.10.31.23297612; authors Burgess/Sun/Zhou/Buck/Mason/Mai — no Davey-Smith/Sattar). Lineage-independent but a 2023 PREPRINT, and the artifact is already well-supported → DEFER (not worth a drop yet; re-check for the published version).
  • Testosterone → outcomes (one TRAVERSE family) · Vit-D/Ca → fractures (WHI-dominated pools) · Sun/UV harm arm (Green inside Gandini) → NO clean gold hit on the independence-breaking query, and each is peripheral / already-TENSION-covered. NOT pursued (honest low-value, not an OA dodge). RECORD trial + a testosterone MR could be maintainer-dropped later if these rows are ever promoted.

Net: 2 SELECTED big-rock second opinions → BOTH MAINTAINER-DROP (Tagawa protein · Zuber apoB; OUP-blocked, OA preprints noted) · 1 DEFERRED (BMI preprint) · 3 NOT-PURSUED (peripheral). Nothing downgraded-for-OA.

[2026-08-07 #3] SELECTION RECORD — post-adiposity-seq acquires: DPPOS hard-outcome + weekend-warrior/accumulation

Two Gate-21 acquire-targets: (1) convert Knowler-DPP’s incidence win to a HARD outcome; (2) close the PA accumulation-pattern gap (daily vs concentrated — the maintainer’s PA challenge). Standard: gold/high, best paper regardless of OA, blocked -> maintainer-drop.

  • DPPOS -> CV events (Diabetes Prevention Program + Outcomes Study) -> SELECTED: DPPOS 2022 Circulation «Effects of Long-term Metformin and Lifestyle Interventions on Cardiovascular Events in the DPP and its Outcome Study» (10.1161/circulationaha.121.056756, high·rct-followup, 170c). The hard-outcome (CV-events) follow-up that closes the DPP [AWAITS] — the incidence win alone does not establish CV/mortality benefit. MAINTAINER-DROP (AHA anti-bot 403; Europe-PMC render 500’d). On PMC: PMC9179081 -> pmc.ncbi.nlm.nih.gov/articles/PMC9179081/pdf/. NOTE: the raw/binary/cut/ DPPOS-2015 is a DIFFERENT paper (microvascular/durability, cut as Knowler-echo) — it STAYS cut; this 2022 CV-events paper is the resurface.
  • PA accumulation pattern (daily vs weekend-warrior) -> SELECTED + FETCHED: BJSM 2024 «Associations of the ‘weekend warrior’ physical activity pattern with all-cause, CV and cancer mortality» (10.1136/bjsports-2023-107612, high·cohort, OA, fetched 19pp). ACCELEROMETER-measured — the best fit for the gap: it is measurement-error-robust (the fabric’s binding constraint) and objective, and it directly tests concentrated-vs-spread. This is the primary answer to the challenge.
    • Also SELECTED -> MAINTAINER-DROP (the preferred pooled UNIT): Weekend-warrior MA 2022 (EurJPC 10.1093/eurjpc/zwac246, gold/high·meta-analysis, 65c) — «‘weekend warrior’ and regularly active … confer SIMILAR» risk reductions; the synthesized unit over the self-report cohorts. OUP paywalled.
    • IDENTIFIED, optional: O’Donovan 2017 (JAMA IM 10.1001/jamainternmed.2016.8014, 463c) — the landmark self-report pooled cohort; likely subsumed by the 2022 MA, so lower priority. JAMA paywalled -> drop only if the MA is insufficient.

Net: 1 FETCHED (BJSM-2024 accelerometer weekend-warrior — closes the PA-accumulation gap, objective) · 2 MAINTAINER-DROP (DPPOS-2022 CV-events [PMC9179081]; weekend-warrior MA-2022 [preferred unit]) · 1 optional (O’Donovan-2017). Nothing downgraded-for-OA — the fetched BJSM is objective/best-fit, not a convenience substitute. Hand off the fetched one to process inbox; drop the two for the hard-outcome + pooled-unit legs.

[2026-08-07 #4] SELECTION RECORD — Zhong 2019 (the eggs->CVD tension-gold)

[2026-08-07 #5] SELECTION RECORD — split-arm acquires (vit-D→cancer, omega-3→cognition) + elderly-protein-target gold

Three Pass-3 gap targets from the question-taxonomy diff. Standard: gold/high MA-SR, best paper regardless of OA, blocked→maintainer-drop, mental-health-as-end kept PERIPHERAL (telos). openalex queries auto-logged above.

Cluster 1 — vitamin D → CANCER mortality / all-cause mortality (the non-CV arm; VITAL held for CV only)

  • SELECTED → MAINTAINER-DROP: Keum 2019 «Vitamin D supplementation and total cancer incidence and mortality: a meta-analysis of RCTs» (Ann Oncol 10.1093/annonc/mdz059, gold·MA, 367c). The decision-relevant split: no cancer-INCIDENCE reduction but ~13% cancer-MORTALITY reduction (RR ~0.87) — the arm VITAL doesn’t foreground. Elsevier anti-bot 403; PMC6821324 available → drop from PMC.
  • REJECTED (echo of Keum / thinner): 2022 Crit-Rev-Food-Sci MA (10.1080/10408398.2022.2056574, 34c) · 2019 Biosci-Rep RCT-MA (bsr20190369, 33c) · 2021 INPLASY protocol (not a completed study). All concur; keep the landmark. NOTE all-cause-mortality arm is partly held via VITAL (null) — Keum adds the cancer-mortality signal.

Cluster 2 — omega-3 → COGNITION / dementia (prioritized over depression: mental-health-as-end is peripheral)

  • SELECTED + FETCHED (Europe-PMC, PMC10447496, full 10.6k-w): AJCN 2023 «The Relationship of Omega-3 Fatty Acids with Dementia and Cognitive Decline: Evidence from Prospective Cohort Studies…» (10.1016/j.ajcnut.2023.04.001, high·evidence-review, 154c). Recent, comprehensive, OA — the current synthesis of the (largely null-to-modest) omega-3→cognition evidence.
  • DEFERRED (optional definitive-null): Cochrane 2012 «Omega-3 for prevention of cognitive decline and dementia» (10.1002/14651858.cd005379.pub3, gold, 220c, OA) — the canonical RCT null, but 2012/dated; the AJCN 2023 updates it. Drop only if a Cochrane-grade null is specifically wanted.
  • REJECTED (out of scope-priority): omega-3→DEPRESSION MAs (2020 Nord-J-Psych 10.1080/08039488.2020.1843710; 2025 JAD) — depression as an END is PERIPHERAL (telos: mental health admitted mainly via its physical intersection); cognition/dementia is the physical hard-outcome arm we open. Recorded as a deliberate call, not an oversight — revisit if the physical-intersection (e.g. omega-3→depression→CV via adherence) is ever wanted.

Cluster 3 — older-adult protein TARGET → physical function / sarcopenia (the flagged TOP gap; Bauer only moderate)

  • SELECTED + FETCHED (Europe-PMC, PMC6163569, full 9k-w): Mendonca 2018 «Relative Protein Intake and Physical Function in Older Adults: A Systematic Review and Meta-Analysis of Observational Studies» (Nutrients 10.3390/nu10091330, gold·MA, 131c). Directly the TARGET question (g/kg → function), OA. Ratchets the elderly-protein-target row from moderate (Bauer consensus) toward gold.
  • DEFERRED → MAINTAINER-DROP (the newer upgrade): Coelho-Júnior 2022 «Protein intake and physical function in older adults: a SR-MA» (Ageing Res Rev 10.1016/j.arr.2022.101731, gold·SR-MA, 82c) — higher-impact journal, 4 yr newer; the ratchet-UPGRADE over Mendonca. Paywalled → drop if the strongest incumbent is wanted.
  • REJECTED (different question): 2017 AJCN protein+RT→body-composition (10.3945/ajcn.116.143594, 317c) — that is the protein×RT question already held (Morton), not the older-adult intake TARGET · 2023 plant-protein-intervention MA (nu15184060) — a sub-question (plant-source), not the target.

Net: 2 FETCHED (AJCN-2023 omega-3→cognition · Mendonca-2018 protein→function-older) · 1 MAINTAINER-DROP (Keum-2019 vit-D→cancer-mortality, PMC6821324) · 2 optional-drops (Cochrane-2012 omega-3-null · Coelho-2022 protein-upgrade) · omega-3→depression REJECTED on telos-scope. Hand the 2 fetched to process inbox.

[2026-08-07 #6] SELECTION RECORD — the one clean remaining lever: magnesium (an unopened [G] topic)

[2026-08-08 #1] SELECTION RECORD — coverage-expansion round (9 telos-edge topics)

Maintainer-requested sweep of 9 gap topics with an AGGRESSIVE gold-gate cut. Channel: openalex only (auto-logged above); no web search. Gold/high required; DEFER any topic with no gold reaching a patient-important / in-scope outcome. Held-not-resurfaced: Chen mindfulness->BP 2024 (MBSR partial), Richardson CGM + CGM framework (wearables partial), Livingston dementia commission (social-isolation PAF), Martineau vit-D->respiratory-infection, Valdes/Wastyk/WGO (microbiome).

SELECTED — gold, ratchet-advancing, in-scope (7):

  • Noetel 2024 BMJ — exercise->depression network MA (581c). KEYSTONE; patient-important (clinical depression), NMA is best design. [maintainer-drop] BMJ 403 (plain + headless). DOI 10.1136/bmj-2023-075847. Slug Noetel - Exercise Depression Network Meta-Analysis 2024.
  • Molendijk 2017 JAD — diet-quality->depression dose-response MA (552c). [maintainer-drop] Elsevier paywall. DOI 10.1016/j.jad.2017.09.022. Pairs w/ Noetel (diet vs exercise lever for the mental-health axis).
  • Wang 2023 Nat Hum Behav — social isolation/loneliness->mortality, 90 cohorts (422c). KEYSTONE for the social-connection axis. [maintainer-drop] Nature paywall (oa=False). DOI 10.1038/s41562-023-01617-6. Frame: large association, hard-to-dose lever (limited decision-change) + U/artifact + confounding check.
  • Namazi 2018 Nutrients — Dietary Inflammatory Index->CVD/mortality MA (318c). Anchors the chronic-inflammation “independent lever vs mediator that reduces to known patterns (MedDiet/fibre)” verdict. [maintainer-drop] MDPI 403 (plain + headless, versioned + clean URL both). DOI 10.3390/nu10020200.
  • Brickwood 2019 JMIR mHealth — consumer wearable trackers->physical activity MA (673c). FETCHED -> inbox/ (20pp PDF, clean). CAVEAT at ingest: outcome is PA level/steps (SURROGATE), not a hard endpoint; the honest verdict is adherence-tool, weak durable-outcome evidence. DOI 10.2196/11819.
  • Milajerdi 2020 Adv Nutr — dietary fibre/fruit/veg->IBD risk MA (123c). The IN-SCOPE autoimmune slice (prevention/incidence, not treatment). [maintainer-drop] OUP paywall. DOI 10.1093/advances/nmaa145.
  • GeroScience 2024 — periodontal disease->cognitive disorders/dementia/depression MA (49c). The single gold-ish anchor for the oral->brain topic; SELECTED low-priority to support a “mostly confounded observational association” verdict (streetlight/U-artifact framing), NOT “floss to save your brain”. [maintainer-drop] PMC NCBI hard-wall (20KB landing page only). DOI 10.1007/s11357-024-01243-8.

DEFERRED — no gold reaching a patient-important, in-scope outcome (zeros attributed):

  • Mindfulness/MBSR beyond held Chen — only paywalled Gal 2021 JAD meditation-apps->wellbeing MA (305c), which is out of telos scope (wellbeing-as-end, not a physical/longevity outcome). Held Chen already covers MBSR->BP (low confidence). No gold on meditation->hard health/longevity. DEFER.
  • Asthma / allergy / respiratory (diet) — MA-level/high-cite = 0; loosened surfaces only child- and pregnancy-focused low-cite SRs (MedDiet->childhood asthma, ~60c). No ADULT gold. Zero attributed to thin adult-gold literature, not query strictness. DEFER.
  • Hydration / water intake — 0 across water->mortality, hydration->chronic-disease, AND fluid->kidney-stone recurrence (the one hard endpoint), even with type-filter dropped and low cite floor. Genuine literature-absence at gold level. Confirms the 2026-08-07 CEILING NOTE. DEFER.
  • Anti-inflammatory diet -> RA (Schonenberger 2021 Nutrients, RA pain, 134c OA) — REJECTED: scope-edge (treatment of diagnosed disease + symptom/surrogate outcome). Kept the IBD-risk autoimmune slice instead.

Acquisition summary: 1 fetched (Brickwood); 6 maintainer-drop (DOIs above — 3 truly paywalled [Molendijk/Milajerdi/Wang], 3 OA-but-blocked [Noetel BMJ, Namazi MDPI, GeroScience PMC]). 3 topics deferred (mindfulness-hard-outcome, asthma, hydration) + 1 reject (RA-treatment). Hand off to process inbox.

[2026-08-08 #2] SELECTION RECORD — inflammation + autoimmune RE-SEARCH (maintainer challenge; asymmetric-search catch)

BIAS-MACHINERY NOTE: the #1 round searched inflammation ONLY from the diet angle (DII) and autoimmune from ONE narrow slice (fibre->IBD-risk), then cut on “scope”. Maintainer challenged both. Re-search under symmetric standards found the decisive gold the first pass never queried -> the first cut was asymmetric scrutiny, not a true gold-absence. Recorded per the record-the-expectation discipline.

INFLAMMATION — is it a CAUSAL modifiable lever? (the RCT + MR proof #1 missed):

  • SELECTED Ridker 2017 CANTOS — canakinumab (anti-IL-1beta) cut MACE with hsCRP-lowering and NO lipid change = the RCT proof inflammation is CAUSAL for CVD. Net-effect nuance: NO all-cause-mortality benefit (fatal infections offset it); expensive; not lifestyle. DOI 10.1056/nejmoa1707914 (9306c). [maintainer-drop] NEJM 403.
  • SELECTED Nidorf 2020 LoDoCo2 — colchicine (CHEAP, generic) reduced CV events in chronic coronary disease. DOI 10.1056/nejmoa2021372 (2161c). [maintainer-drop] NEJM 403.
  • SELECTED Emerging Risk Factors Collaboration 2010 Lancet — CRP->CHD/stroke/vascular+nonvascular mortality IPD-MA (the marker->outcome landmark). DOI 10.1016/s0140-6736(09)61717-7 (2541c). [maintainer-drop] Lancet.
  • SELECTED CRP-CHD Mendelian randomization 2011 BMJ — genetic CRP does NOT raise CHD => CRP is a MARKER, not the causal lever (target is upstream IL-1->IL-6). The surrogate-vs-target + U-artifact jewel. DOI 10.1136/bmj.d548 (658c). [maintainer-drop] BMJ 403 (genuinely OA).
  • SYNTHESIS (owed at ingest): inflammation is a VALIDATED causal node (CANTOS/colchicine/IL-6R-MR), but (i) CRP is marker-not-lever, (ii) drug levers apply to established-disease strata, (iii) LIFESTYLE anti-inflammatory levers (adiposity/diet/activity/smoking/sleep) largely OVERLAP big rocks already held, and (iv) DII (Namazi) is an observational diet-proxy. Both true: causal node AND mostly-known lifestyle levers.

AUTOIMMUNE — modifiable risk (prevention, IN-SCOPE) + longevity impact:

  • SELECTED + FETCHED DiGiuseppe 2014 — smoking->RA dose-response MA (strong modifiable risk). DOI 10.1186/ar4498 (274c, OA). -> inbox/ (547KB, clean).
  • SELECTED Qin 2015 — BMI/adiposity->RA dose-response MA. DOI 10.1186/s13075-015-0601-x (243c). [maintainer-drop] paywalled.
  • CONTEXT (belief confirmed): RA carries excess mortality that is largely CARDIOVASCULAR -> connects autoimmune -> systemic inflammation -> CVD -> longevity into one mechanism (the through-line for both beliefs).
  • DEFERRED still: omega-3->RA (0 at threshold, thin); vitamin-D->MS (top hit was VDR polymorphism, not the modifiable status/intake angle — needs a targeted re-search: Munger cohort / MR); general RA-excess-mortality MA (hits were ILD-subtype-specific) — flag, don’t force.

Acquisition this round: 1 fetched (DiGiuseppe); 5 maintainer-drop (CANTOS, LoDoCo2, ERFC-CRP, CRP-MR, Qin). CORRECTION TO #1: inflammation + autoimmune are NOT thin — they carry landmark gold; the #1 DEFER/REJECT was under-searched. Standing DEFERs from #1 (hydration, asthma-adult-gold, mindfulness->hard-outcome) still hold.

[2026-08-08 #3] SELECTION RECORD — polypharmacy / prescribing cascade / deprescribing (maintainer angle)

Maintainer belief: pharmacological intervention is a slippery slope (one drug -> a second -> …). In scope: the telos treats a drug as an exposure whose NET effect includes cascade/ADE burden (net-effect-not-intended; judge complete realistic strategies). Channel: openalex (auto-logged). Gold-gated.

SELECTED (3):

  • Deprescribing RCT SR 2023 (Int J Pharm Pract) — the DECISIVE test: does REVERSING polypharmacy improve patient-important outcomes? SR of RCTs. DOI 10.1093/ijpp/riad025 (91c, OA). [maintainer-drop] OUP 403. KEYSTONE — the honest verdict lives here (deprescribing is broadly SAFE; large mortality benefit not proven).
  • Rochon & Gurwitz “The prescribing cascade revisited” Lancet 2017 — the canonical NAMED-concept source for the slippery slope. DOI 10.1016/s0140-6736(17)31188-1 (231c). Tier: narrative/viewpoint = MODERATE (concept-definer, needs source-tier-override at ingest as the field’s naming source). [maintainer-drop] Lancet.
  • Savage 2020 JAMA Intern Med — empirical instance (CCB -> loop-diuretic cascade, prescription-sequence- symmetry). DOI 10.1001/jamainternmed.2019.7087 (117c, OA). Concrete worked example. [maintainer-drop] JAMA 403.
  • SYNTHESIS (owed): the cascade is REAL and named; polypharmacy->falls/ADE/hospitalization/mortality associations are real but heavily CONFOUNDED BY INDICATION (sicker take more drugs — reverse-causation family, same lens as the U-artifact); deprescribing RCTs are the clean test and show SAFETY, not a proven longevity lever. Both true: slope is real (net-effect discipline validated) AND the mortality signal is confounded. Bears directly on the held statin / BP-drug / GLP-1 pages (net-effect vs realistic alternative).

Acquisition: 0 fetched (all 3 OA-blocked: OUP/Lancet/JAMA) -> all maintainer-drop. Gold EXISTS; topic is NOT thin — confirms polypharmacy as a real coverage gap.

[2026-08-08 #4] SELECTION RECORD — occupation / commute / shift work / vocation -> health & longevity (maintainer angle)

Maintainer angle: NOT occupational-hazard/danger, but do some occupations/vocations/hobbies TEND to higher health & longevity? In scope: job, commute-mode, and shift schedule are modifiable structural exposures with patient-important outcomes. Channel: openalex (auto-logged). NOTE: openalex title.search needs EVERY query word IN the title -> 4 clusters returned 0 from STRICTNESS not absence; the landmarks were then confirmed by resolve <DOI>. A silent 0 here would have been a false absence.

SELECTED (5 gold keystones):

  • Kivimaki 2012 Lancet — job strain (high demand + low control) -> CHD, collaborative IPD-MA. The work-stress landmark (modest ~1.2x). DOI 10.1016/s0140-6736(12)60994-5 (991c). [maintainer-drop] Lancet.
  • Coenen 2018 BJSM — “Do highly physically active workers die early?” SR+MA: high OCCUPATIONAL PA -> HIGHER mortality in men (~1.18x) = the PHYSICAL-ACTIVITY PARADOX (occupational != leisure PA; net-effect / compensation lens; contradicts naive more-movement-is-better). DOI 10.1136/bjsports-2017-098540 (370c, gold, OA). [maintainer-drop] BMJ.
  • Celis-Morales 2017 BMJ — active commuting (esp. cycling) -> lower CVD/cancer/all-cause mortality, UK-Biobank cohort (cycling all-cause HR ~0.59). Commute-mode as a structural PA-delivery lever. DOI 10.1136/bmj.j1456 (527c, OA). [maintainer-drop] BMJ. (Partial overlap w/ held PA->mortality; new angle = structural/commute.)
  • Travis 2016 JNCI — night-shift work -> breast cancer: 3 prospective cohorts + MA, a notable NULL / no-association (counterweight to the IARC 2A classification; symmetric-standards + U/confounding lens). DOI 10.1093/jnci/djw169 (193c, OA). [maintainer-drop] OUP.
  • Cohen 2015 Psychosom Med — purpose in life -> lower all-cause mortality + CV events, MA. The vocation/hobby -> meaning channel (physiological intersection; overlaps the social-connection + mental-health axes). DOI 10.1097/psy.0000000000000274 (440c). [maintainer-drop] paywalled.

DEFERRED: retirement -> mortality (mixed/confounded, no clean gold — healthy-worker + reverse-causation); specific hobbies (gardening/music/volunteering) route through purpose + social + PA channels already captured; shift-work -> CHD (job-strain landmark + circadian via held sleep pages cover the decision; a dedicated shift->CHD MA is a lower-priority add).

Acquisition: 0 fetched (all 5 OA-blocked BMJ/OUP/Lancet or paywalled) -> all maintainer-drop. Gold EXISTS; occupation is a real, previously-uncovered cluster. The PA-paradox (Coenen) is the highest-value single finding.

  • [2026-08-08] openalex search “sedentary behaviour all-cause mortality dose-response meta-analysis” -> 1 hits: https://doi.org/10.1007/s10654-018-0380-1
  • [2026-08-08] openalex search “television viewing time all-cause mortality meta-analysis” -> 0 hits: (none)
  • [2026-08-08] openalex search “screen time cardiovascular mortality adults meta-analysis” -> 0 hits: (none)
  • [2026-08-08] openalex fetch “10.1007/s10654-018-0380-1” -> 0 hits: (none)
  • [2026-08-08] openalex search “television viewing mortality” -> 3 hits: https://doi.org/10.1093/aje/kwv164; https://doi.org/10.1016/j.atherosclerosis.2017.06.024; https://doi.org/10.1016/j.hlc.2016.03.006
  • [2026-08-08] openalex search “eating alone metabolic syndrome meta-analysis” -> 0 hits: (none)
  • [2026-08-08] openalex search “family meals dietary meta-analysis” -> 0 hits: (none)
  • [2026-08-08] openalex search “living alone all-cause mortality meta-analysis” -> 1 hits: https://doi.org/10.1016/j.eclinm.2022.101677
  • [2026-08-08] openalex search “occupational complexity dementia cognitive meta-analysis” -> 0 hits: (none)
  • [2026-08-08] openalex search “mentally stimulating work cognitive decline” -> 0 hits: (none)
  • [2026-08-08] openalex search “commensality eating together health” -> 0 hits: (none)
  • [2026-08-08] openalex search “eating alone health outcomes adults” -> 1 hits: https://doi.org/10.14373/jkda.2020.26.2.85
  • [2026-08-08] openalex resolve “10.1136/bmj.n1804 10.1017/S1041610215001349” -> 2 hits: https://doi.org/10.1136/bmj.n1804; https://doi.org/10.1017/s1041610215001349
  • [2026-08-08] openalex search “occupational mental demands dementia risk” -> 0 hits: (none)
  • [2026-08-08] openalex fetch “10.1093/aje/kwv164” -> 0 hits: (none)
  • [2026-08-08] openalex fetch “10.1016/j.eclinm.2022.101677” -> 0 hits: (none)

[2026-08-08 #5] SELECTION RECORD — communal eating / multigenerational living / physical-vs-mental work / sitting / screens

Maintainer angle: household + social + sedentary structural exposures -> health/longevity. Channel: openalex (auto-logged); zeros re-checked loose + landmarks confirmed via resolve-by-DOI (title.search strictness). Held already: Ekelund sitting-time 2016 + accelerometer-PA 2019 (SITTING covered); Coenen PA-paradox + Kivimaki job-strain (PHYSICAL occupation, round #4); Wang social-isolation->mortality (round #1).

SELECTED (4):

  • Patterson 2018 Eur J Epidemiol — sedentary behaviour -> all-cause/CV/cancer mortality + T2D, SR + DOSE-RESPONSE MA (1450c, OA). FETCHED -> inbox/. UPGRADE to the held Ekelund sitting page: adds the dose-response SHAPE + T2D/cancer outcomes Ekelund’s PA-interaction pooling lacked. DOI 10.1007/s10654-018-0380-1.
  • Living Alone -> all-cause mortality 2022 (eClinicalMedicine) — SR+MA, community-dwelling adults (60c). The multigenerational/living-arrangement keystone (living-alone = inverse of multigen household). Distinct structural exposure though it overlaps social-isolation (Wang). DOI 10.1016/j.eclinm.2022.101677. [maintainer-drop] paywalled.
  • Kivimaki 2021 BMJ — cognitive stimulation at work -> LOWER dementia risk, IPD-Work consortium + plasma proteins (98c, OA). The MENTAL-occupation keystone (cognitive reserve; connects held Livingston/dementia). DOI 10.1136/bmj.n1804. FETCHED -> inbox/ (Steel cloud-browser; BMJ 403’d plain+headless, Europe-PMC PMC8372196 render endpoint failed, Steel on the BMJ pdf succeeded — 13pp verified). Pairs with Coenen (physical) to answer physical-vs-mental.
  • Sun 2015 Am J Epidemiol — TV viewing time -> all-cause mortality MA (67c, OA). The SCREEN keystone, but LOW-priority: adult screen-time is operationalized as TV = leisure sedentary, so it largely OVERLAPS Patterson (“is the category doing any work?” — screens ~ sitting, not an independent lever). DOI 10.1093/aje/kwv164. [maintainer-drop] OUP.

DEFERRED:

  • Communal eating / commensality / eating-alone — NO gold. Loosened only to a Korean-language low-cite article (eating-alone->health, 10c) + child/adolescent family-meal->diet-quality literature (not adult longevity). Plausible benefit is MEDIATED by social connection (Wang) + dietary pattern (held) -> not an independent lever. Genuine gold-absence for the adult-longevity question, not query strictness.
  • Physical occupation — already covered (Coenen PA-paradox, #4).

Acquisition: 2 fetched (Patterson; Kivimaki-2021 via Steel [2026-08-08 follow-up]); 2 maintainer-drop (Living-alone, Sun-TV). Physical-vs- mental work now answerable (Coenen physical + Kivimaki mental); sitting upgraded; screens ~ sitting; communal eating thin.

[2026-08-08 #6] SELECTION RECORD — QoL-extension gaps (frailty · chronic pain · OA · HPA/stress · circadian)

The telos QoL extension (telos-review 2026-08-08) opened these gaps; this fills them. Channel: openalex (auto-logged); title-strictness zeros re-checked loose + landmarks confirmed via resolve-by-DOI. Gold-gated. Held/covered already: Kivimaki job-strain + night-shift MA (shift-work = the circadian hard-outcome edge); grip/sarcopenia/falls (function); Stress-Mgmt->BP stub.

SELECTED (6):

  • Vermeiren 2016 (JAMDA) — frailty -> negative health outcomes (mortality + disability + hospitalization) MA, the frailty->OUTCOME keystone. DOI 10.1016/j.jamda.2016.09.010 (1055c). [maintainer-drop] Elsevier — GENUINELY paywalled (Europe-PMC checked 2026-08-08: no PMCID, isOA=N, inEPMC=N, ‘Subscription required’); not an anti-bot block, so Steel/EPMC cannot help — needs institutional access or an OA-alternative substitution decision.
  • CMAJ Open 2021 (Racey/Giangregorio team) — physical-activity interventions for frailty/pre-frailty: the “can you REVERSE it?” keystone. DOI 10.9778/cmajo.20200222 (70c, OA). FETCHED -> inbox/ (5pp). Companion available (nutrition + PA combined, 10.9778/cmajo.20200248) if the protein link is wanted.
  • Geneen 2017 Cochrane OVERVIEW — physical activity & exercise for chronic pain in adults, umbrella of Cochrane reviews (best design for the node). DOI 10.1002/14651858.cd011279.pub3 (1184c, OA). [maintainer-drop] Cochrane 403 (Steel candidate). The chronic-pain keystone.
  • Rausch Osthoff 2018 (RMD Open) — exercise/PA -> function in hip/knee OA + inflammatory arthritis, MA informing the EULAR recommendations. DOI 10.1136/rmdopen-2018-000713 (163c, OA). FETCHED -> inbox/ (exercise->OA keystone). Goh 2019 (10.1016/j.rehab.2019.04.006, 465c) DROPPED as an echo — ScienceDirect 403 and rmdopen already answers exercise->OA (no-echo ratchet).
  • Messier IDEA 2013 (JAMA) — intensive diet + exercise weight loss -> knee-OA load/inflammation/pain/function RCT, the weight-loss->OA keystone (landmark RCT, fit-to-question). DOI 10.1001/jama.2013.277669 (852c). [maintainer-drop] JAMA.
  • Allostatic Load -> Mortality MA 2022 (Am J Prev Med) — cumulative stress-physiology -> all-cause mortality, the HPA/psychosocial hard-outcome keystone that connects the non-collapsing psychosocial cluster (isolation · loneliness · purpose · job strain · living-alone) to a patient-important outcome. DOI 10.1016/j.amepre.2022.02.003 (166c). [maintainer-drop] Elsevier.

DEFERRED / not-acquired:

  • Circadian disruption — NO human-outcome gold beyond shift-work (held via Kivimaki + night-shift MA). Loosened only to mechanism/animal/microbiome/steatosis low-cite 2024-26 work = the “not-yet” bucket. DEFER; the hard-outcome edge is already held. Attributed to genuine human-outcome thinness, not query strictness.
  • Chronic-stress-specific (perceived stress -> CHD) — folded into Allostatic-Load MA + held Kivimaki job-strain; a separate acquire would be an echo. (Guessed DOI 10.1016/j.amjcard.2012.07.006 for Richardson-2012 resolved to an UNRELATED albuminuria paper -> NOT used; verified-not-fabricated.)
  • Outcome-first intervention hierarchy — NOT an acquire: the owed Layer-1 SYNTHESIS deliverable (outcome-anchored, absolute-magnitude, age-stratified ranking). Needs no new source; feeds Big-Rocks.

Acquisition: 2 fetched (CMAJ frailty-PA; rmdopen exercise-OA); 4 maintainer-drop (Vermeiren, Geneen-Cochrane, Messier-IDEA, Allostatic-Load). 1 echo-dropped (Goh). Circadian + chronic-stress-specific deferred; hierarchy = synthesis.

[2026-08-08 #7] SELECTION RECORD — exercise-deliverable gap round (RESUMED from the interrupted scan)

The reframe’s 3 gaps (mobility/flexibility · functional/balance · exercise injury/adverse-events), gold-gated. Resumed the interrupted scan (loosened the gap-1/2 strictness zeros). Fetched to inbox/; hand to process-inbox.

  • SELECTED + FETCHED: Lauersen 2013 BJSM — «The effectiveness of exercise interventions to prevent sports injuries» MA (565c, gold, OA). DUAL-PURPOSE: answers gap-3 (strength training cuts injury ~1/3; proprioception/balance helps) AND gap-1 (STRETCHING does NOT prevent injury -> mobility/flexibility’s patient-important verdict is null; stretching->ROM is a SURROGATE). DOI 10.1136/bjsports-2013-092538.
  • SELECTED + FETCHED: Kunutsor 2018 (J Frailty Sarcopenia Falls) — «Adverse events and safety issues associated with physical activity and exercise for adults with osteoporosis and osteopenia» SR (26c, gold-ish, OA). The dim-6 downsides/safety keystone, but SCOPED to osteoporosis/osteopenia (a fragile stratum), NOT general. Low cite = adverse events are systematically UNDER-reported in exercise trials (a streetlight finding worth stating). DOI 10.22540/jfsf-03-155. 2nd Kunutsor work (vs weekend-warrior 2022) -> slug disambiguates.
  • DEFERRED gap-2 (balance/functional/multicomponent) — ALREADY COVERED: held Exercise for Preventing Falls in Older Adults (Sherrington, balance+functional->falls) + incoming Racey (frailty/multicomponent PA). A fresh multicomponent MA (e.g. exger.2024.112604) would ECHO them -> no acquire (no-echo ratchet).
  • DEFERRED gap-1 residual (mobility beyond injury) — Behm 2016 APNM (stretching acute effects/ROM, 749c) is mostly SURROGATE (ROM/performance); the patient-important verdict (no injury benefit) is carried by Lauersen. Optional-only; not acquired.
  • NAMED GAPS (not acquired — thin/out-of-scope): overtraining / CNS-fatigue (mostly consensus/narrative, per telos «named gap not a page»); general exercise adverse-events beyond osteoporosis (under-reported); functional- fitness AS A DISTINCT CONSTRUCT (folds into falls/frailty/sarcopenia). Contraindication SCREENING = OUT OF SCOPE (prescriber act) — the deliverable appraises risks, does not screen a person. Acquisition: 2 fetched (Lauersen, Kunutsor); gap-2 + gap-1-residual deferred as echo/surrogate; rest named-gap/out-of-scope.
  • [2026-08-08] openalex search “weight cycling all-cause mortality meta-analysis” -> 0 hits: (none)
  • [2026-08-08] openalex resolve “10.1042/CS20140553” -> 1 hits: https://doi.org/10.1042/cs20140553
  • [2026-08-08] openalex search “metabolically healthy obesity cardiovascular events mortality meta-analysis” -> 0 hits: (none)
  • [2026-08-08] openalex resolve “10.7326/0003-4819-159-11-201312030-00008” -> 1 hits: https://doi.org/10.7326/0003-4819-159-11-201312030-00008
  • [2026-08-08] openalex search “metabolically healthy obese benign systematic review” -> 2 hits: https://doi.org/10.4103/ijpvm.ijpvm_392_20; https://doi.org/10.3410/f.718196353.793488213
  • [2026-08-08] openalex fetch “10.1042/cs20140553” -> 0 hits: (none)
  • [2026-08-08] openalex fetch “10.7326/0003-4819-159-11-201312030-00008” -> 0 hits: (none)

[2026-08-08 #8] SELECTION RECORD — the 2 fat-threshold targeted acquires (Obesity reframe, aspect 2)

Per maintainer “do the two targeted acquisitions (fat threshold).” Both SELECTED; both PAYWALLED (openalex + Europe-PMC checked: no PMCID, isOA=N, inEPMC=N) -> [maintainer-drop] (Steel cannot help a paywall; not downgraded).

  • Taylor & Holman 2015 — personal fat threshold «Normal weight individuals who develop Type 2 diabetes: the personal fat threshold concept…» (Clin Sci, 10.1042/cs20140553, 190c). The concept-DEFINER for the individual-storage-capacity phenotype (ectopic vs subcutaneous). mechanism/perspective tier -> needs a source-tier-overrides.tsv entry at ingest. Reversibility mechanism (twin-cycle) already held via DiRECT; this adds the individual-variation piece.
  • Kramer 2013 — MHO benign? MA «Are Metabolically Healthy Overweight and Obesity Benign Conditions? A SR+MA» (Ann Intern Med, 10.7326/0003-4819-159-11-201312030-00008, 961c, gold). The metabolically-healthy-obesity -> CV/mortality magnitude (MHO NOT benign over long follow-up) — the outcome leg of aspect 2.
  • NOT this batch: weight-cycling MA (candidate #3) stays deferred. Acquisition: 0 fetched (both genuinely paywalled); 2 maintainer-drop. DOIs above.
  • [2026-08-08] openalex search “weight cycling mortality meta-analysis” -> 0 hits: (none)
  • [2026-08-08] openalex search “body weight variability cardiovascular outcomes meta-analysis” -> 0 hits: (none)
  • [2026-08-08] openalex search “nonalcoholic fatty liver disease type 2 diabetes mendelian randomization” -> 4 hits: https://doi.org/10.3389/fendo.2023.1156381; https://doi.org/10.3389/fpsyt.2023.1068756; https://doi.org/10.1016/j.diabres.2023.110993; https://doi.org/10.1097/md.0000000000038008
  • [2026-08-08] openalex search “weight cycling weight fluctuation all-cause mortality risk” -> 0 hits: (none)
  • [2026-08-08] openalex search “weight cycling cardiovascular disease diabetes systematic review” -> 0 hits: (none)
  • [2026-08-08] openalex search “weight cycling” -> 6 hits: https://doi.org/10.1111/obr.12251; https://doi.org/10.1021/acs.est.5b03192; https://doi.org/10.1002/14651858.cd006982.pub4; https://doi.org/10.1007/s10815-015-0434-1; https://doi.org/10.1021/acsaem.0c01621
  • [2026-08-08] openalex resolve “10.1111/jdi.13456” -> 1 hits: https://doi.org/10.1111/jdi.13456
  • [2026-08-08] openalex fetch “10.1111/obr.12251” -> 0 hits: (none)

[2026-08-08 #9] SELECTION RECORD — obesity acquires: fat-threshold pair STAGED (maintainer-dropped) + weight-cycling + MASLD-MR probe

  • Taylor & Holman 2015 personal-fat-threshold (10.1042/CS20140553) — maintainer-DROPPED -> STAGED raw/binary/papers/, registered moderate/mechanism, tier-override added. (Dropped as a PT-titled file; verified = Taylor.)
  • Kramer 2013 MHO-benign? MA (10.7326/0003-4819-159-11-201312030-00008) — maintainer-DROPPED -> STAGED, registered gold/meta-analysis. (Dropped under a Portuguese filename «Obesidade e Saude revisao»; verified = Kramer/Zinman/Retnakaran.)
  • Weight-cycling (3rd acquire): no gold MA exists (openalex term polluted; loosened searches 0-for-relevant). Best = Montani 2015 Obes Rev «Dieting and weight cycling … who is really at risk?» (10.1111/obr.12251, 252c) — NARRATIVE REVIEW (moderate), paywalled (EPMC no OA) -> [maintainer-drop] PENDING + override-needed. The confounding-prone nature of weight-cycling->hard-outcomes IS the finding. (Zou DOI guess 10.1111/jdi.13456 resolved to an UNRELATED PTH/retinopathy paper -> discarded, not fabricated.)
  • MASLD->MetS causal grounding (for challenge #2): probed NAFLD<->T2D MR — only thin single 2023 studies (fendo.2023.1156381 14c OA; diabres 2023 paywalled) -> NOT acquired (gold-gate; causal-direction left a [G] gap). Acquisition: 2 staged (Taylor, Kramer); 1 maintainer-drop pending (Montani); MASLD-MR not acquired.

[2026-08-08 #10] Montani 2015 (weight-cycling) — maintainer-DROPPED -> STAGED

Completes the 3rd obesity acquire. Montani - Dieting Weight Cycling Cardiometabolic Risk 2015 (10.1111/obr.12251, Obes Rev, 252c), registered moderate/narrative-review + tier-override added. All 3 targeted obesity acquires (Taylor, Kramer, Montani) now staged in raw/binary/papers/.

[2026-08-08] SELECTION RECORD — Geneen pub3 refresh (supersedes held pub2)

Targeted acquire of the newer version of the chronic-pain Cochrane overview Chronic Pain and Physical Activity (held copy = pub2; registry-noted pub3 supersedes). Maintainer-requested.

[2026-08-09 #1] SELECTION RECORD — coverage-gap-scan Deep Research (top-down blind-spot pass) [RETRO hand-log]

Ledger-gap heal. The 2026-08-09 coverage-gap scan (coverage-gap-scan-2026-08-09.md) ran through 5 parallel evidence-researcher subagents doing WEB search to ground effect sizes + name the landmark gold per candidate exposure. Web/subagent channel = NO auto-log, and it was not hand-logged at the time — this block closes fields (1)-(4) retroactively. The candidate DOIs the researchers returned were later verified via openalex (block #2); this block records what was searched and how it was adjudicated.

  • Field 1 — queries (web, via 5 subagents): (a) social connection / isolation / loneliness -> mortality / CVD / inflammation; (b) aspirin primary prevention -> CV events vs major bleeding (net); (c) vitamin D -> non-CV outcomes (cancer / all-cause mortality / respiratory / fracture); (d) periodontal treatment -> HbA1c + the CVD-causality question, and OSA / CPAP -> CV / mortality; (e) the minors — dietary nitrate -> BP, air pollution PM2.5 -> mortality, omega-3 EPA-vs-DHA -> depression, magnesium -> BP/T2D, hydration/serum-sodium -> aging.
  • Field 2 — results (landmark gold named per gap): social — Wang 2023 10.1038/s41562-023-01617-6 + Valtorta 2016 10.1136/heartjnl-2015-308790; aspirin — Zheng/Roddick 2019 10.1001/jama.2018.20578 + USPSTF 2022 10.1001/jama.2022.4983; vit-D — Zhang 2019 10.1136/bmj.l4673 (+ the already-held Martineau i6583 / LeBoff NEJMoa2202106); periodontal — Simpson Cochrane 2022 10.1002/14651858.CD004714.pub4; OSA — Yu 2017 10.1001/jama.2017.7967 + Malhotra SURMOUNT-OSA 2024 10.1056/NEJMoa2404881; nitrate — Siervo 2013 10.3945/jn.112.170233; air pollution — Chen&Hoek 2020 10.1016/j.envint.2020.105974 + Di 2017 10.1056/NEJMoa1702747 + JACC-HEPA 2025 10.1016/j.jacc.2025.06.037; omega-3 — Liao 2019 10.1038/s41398-019-0515-5; magnesium — Zhang 2016 10.1161/HYPERTENSIONAHA.116.07664 + Fang 2016 10.1186/s12916-016-0742-z; hydration — Dmitrieva eBioMedicine 2023 10.1016/j.ebiom.2022.104404.
  • Field 3 — adjudication:
    • SELECTED -> staged (Tier-1/Tier-2, folded to registry+lighthouse 1235cd9): social connection (structural blind spot; ceiling MODERATE — modifiability insufficient-evidence, no RCT), aspirin de-adopt, vit-D non-CV (Zhang only — see echo note), periodontal->HbA1c, OSA/CPAP reframe, nitrate->BP. Each a clean decision-change; effect sizes source-grounded; gold-gated.
    • DEFERRED (Tier-3, verified real gold, NOT staged — small individual lever or peripheral scope): air pollution (large exposure, individual lever small — HEPA -3 mmHg only in elevated-BP), omega-3 EPA/DHA depression (mental-health scope, type-B disambiguation), magnesium (SBP -2 mmHg; page mostly bounds intake-cohort over-reads), hydration/serum-sodium (weak-causal, attention-anti-signal). Staged only if breadth reaches them.
    • REJECTED / dead-end (symmetric-reporting negatives): nutrigenetics/gene-diet (no gold SR-MA, insufficient-evidence — re-confirmed); sauna/cold (already cut/, no gold hard-outcome); “CPAP prevents CV events” (disconfirmed — the reframe IS the value); vit-D/omega-3/magnesium as ENHANCEMENT in the replete (null — reinforces the supplements discipline, not a new lever).
  • Field 4 — acquisition: deferred to block #2 (verification-then-fetch). Tier-1/Tier-2 folded to 8 placeholder registry rows + 6 lighthouse rows; nothing ingested.

[2026-08-09 #2] SELECTION RECORD — coverage-gap gold-source verification + acquisition (openalex)

Purpose: verify the block-#1 web-returned DOIs are real gold (not subagent-fabricated) and establish OA-fetchability. Fields (1)+(2) auto-logged above (2 resolve + 3 fetch lines). All 18 DOIs resolved to real papers with matching titles — zero fabrications in the web-subagent output.

[2026-08-09 #3] SELECTION RECORD — drinking-water chlorine + fluoride -> health/longevity (maintainer request)

Fresh research episode: is drinking-water chlorination (disinfection byproducts) and/or fluoride a material modifiable exposure? Genuine gap — the fabric holds fluoride only in the caries/sugars context (Moynihan), no water-quality exposure page, no registry rows for either. Two distinct exposures, each with its own outcome ledger. Channel: openalex (fields 1+2 auto-logged above) + web discovery for the headline landmarks openalex title.search strictness missed.

  • Field 1 — queries: openalex — “fluoride children intelligence” / “fluoride neurodevelopment cognitive” (—type filters, 0-1 hits from strictness), “fluoride intelligence children” / “fluoride IQ meta-analysis” (broad), “trihalomethanes bladder cancer”, “chlorination byproducts bladder cancer”, “water fluoridation dental caries”, “fluoride bone fracture” (—type MA); web — Taylor 2025 JAMA-Peds DOI, NTP 2024 monograph citation, Villanueva/Costet THM-bladder pooled DOI; openalex resolve x2 (9 landmark DOIs).
  • Field 2 — 0-hit attribution: “fluoride children intelligence” + --type meta-analysis -> 0 = title.search STRICTNESS (needs “meta-analysis” + every word in the title), NOT literature absence — the broad query returned 8 hits and the landmark MAs exist. Lesson re-confirmed: --type over-narrows; drop it and use core nouns, then resolve the named landmark by DOI.
  • Field 3 — adjudication (per candidate):
    • SELECTED — FLUORIDE:
      • 10.1002/14651858.CD010856.pub3 Iheozor-Ejiofor, Cochrane 2024 water-fluoridation->caries (gold SR; the caries BENEFIT arm). Newer-supersedes the 2015 pub2 (cd010856.pub2, 474c) — pub3 is the current incumbent (finds a smaller modern-era benefit given ubiquitous fluoride toothpaste). FETCHED.
      • 10.1001/jamapediatrics.2024.5542 Taylor 2025 JAMA-Peds fluoride->child-IQ MA (74 studies; inverse assoc, beta ~ -1.63 IQ / +1 mg/L urinary F; the decision-critical caveat: uncertain / limited below 1.5 mg/L, i.e. at the 0.7 mg/L US fluoridation target). The peer-reviewed MA arm.
      • 10.22427/ntp-mgraph-8 NTP Monograph 08 2024 (gold government OHAT SR; “moderate confidence” higher F [>1.5 mg/L] -> lower child IQ). The authoritative report. INDEPENDENCE FLAG (anti laundered-E): NTP and Taylor are the SAME NIEHS/NTP evidence stream (Taylor is NTP) — count as ONE backing, not two, on the IQ question.
      • 10.1016/j.envres.2023.115239 Veneri 2023 fluoride->cognition dose-response SR-MA (gold; the GENUINELY INDEPENDENT second IQ stream — Modena/Vinceti group, not NTP). This is what makes the IQ finding type-E-eligible: (NTP/Taylor) vs (Veneri/Modena) = two independent streams.
      • 10.1186/s12940-025-01226-y Mazzoli 2025 fluoride->bone-density/fracture dose-response MA (gold; the skeletal-harm arm). FETCHED. Lineage note: same Modena group as Veneri (Vinceti/Filippini/ Veneri co-authors) — different outcome, so not a laundering risk, but the group produces multiple fluoride-harm papers; hold the halo-across-a-lab caution.
    • SELECTED — CHLORINE / DISINFECTION BYPRODUCTS:
      • 10.1136/oem.2010.062703 Costet 2011 THM->bladder-cancer pooled analysis (gold MA; 6 case-control studies, the exposure-response function — the effect estimate).
      • 10.1289/ehp4495 Evlampidou 2020 EHP THM->bladder-cancer EU attributable burden (the modern landmark; pairs Costet — Costet = effect, Evlampidou = burden/recency).
    • REJECTED / not-taken: the pile of single cross-sectional fluoride-IQ studies (Ding 2011, Sudhir, Saxena, etc. — moderate/weak, superseded by the MAs); the two 2025/2026 critique preprints of Taylor (osf.io/zhm54, 10.1038/s41415-026-9613-y) — DEFERRED, acquire ONLY if a tension needs the counter-pole (symmetric-standards: a critique is a candidate for the friction ledger, not a gold in its own right yet); JDR 2025 global fluoridation SR 10.1177/23800844251342804 (oa=False, 2c) — DEFERRED as a caries echo of Cochrane pub3 (newer but not clearly superseding; take only if pub3 has a gap). Bare 1998 Cantor/Villanueva bladder study — superseded by Costet’s pooled function.
  • Field 4 — acquisition status:
    • FETCHED to inbox/ (2): Mazzoli 2025 fracture MA (BMC, 2.8 MB) · Iheozor-Ejiofor Cochrane 2024 (Lancashire OA eprint, 3.7 MB). Both validated (real multi-page PDFs). -> process inbox to stage.
    • MAINTAINER-DROP (5, NOT downgraded — pause-and-ask):
      • 10.1001/jamapediatrics.2024.5542 Taylor 2025 — oa=True but JAMA anti-bot-403.
      • 10.22427/ntp-mgraph-8 NTP Monograph — free gov doc but no openalex OA-PDF link; source = ntp.niehs.nih.gov / NCBI Bookshelf NBK606073 (multi-part, large).
      • 10.1016/j.envres.2023.115239 Veneri 2023 — oa=True but ScienceDirect 403.
      • 10.1136/oem.2010.062703 Costet 2011 — oa=False (OEM/BMJ).
      • 10.1289/ehp4495 Evlampidou 2020 — oa=True (EHP gov) but URLError on fetch.
      • Priority (per the two exposures): chlorine arm is a 1-decision gap (Costet+Evlampidou); fluoride IQ is the contested/high-value arm (Taylor + NTP + Veneri) — drop those first.
  • Scope note (not acquired, flagged for the eventual page): the DECISION here is small at the individual level — THM->bladder is a chlorinated-vs-not / filter question (relative risk modest, absolute tiny), and the fluoride-IQ signal is at >1.5 mg/L, ABOVE the 0.7 mg/L fluoridation target, so the typical-exposure decision-change is uncertain. Attention-is-an-anti-signal applies hard (both are high-discourse, contested-therefore-generative topics). The pages will likely land as honest-uncertainty / bounds-the-scare entries, not new big-rock levers. Nothing ingested.
  • [2026-08-09] openalex search “how does exercise treatment compare antihypertensive medications systolic blood pressure” -> 2 hits: https://doi.org/10.1136/bjsports-2018-099921; ?
  • [2026-08-09] openalex search “alcohol reduction blood pressure systematic review meta-analysis” -> 1 hits: https://doi.org/10.1016/s2468-2667(17)30003-8
  • [2026-08-09] openalex fetch “10.1136/bjsports-2018-099921” -> 0 hits: (none)
  • [2026-08-09] openalex fetch “10.1016/s2468-2667(17)30003-8” -> 0 hits: (none)

[2026-08-09 #4] SELECTION RECORD — Blood Pressure spec acquire-gaps (Naci + Roerecke)

Resolving the two named acquire-gaps in the Blood Pressure deliverable spec (fields 1+2 auto-logged above: 2 search + 2 fetch lines).

  • Field 3 — adjudication:
    • SELECTED — Naci 2019 10.1136/bjsports-2018-099921 (BJSM network MA, exercise vs antihypertensive drugs on SBP; 328 cites). The crown-jewel acquire for the lifestyle-vs-drug counterfactual comparison in the BP spec. Gold.
    • SELECTED — Roerecke 2017 10.1016/s2468-2667(17)30003-8 (Lancet Public Health MA, alcohol reduction -> BP dose-response; 556 cites). The BP-specific alcohol magnitude the held alcohol page may lack. Gold; take only if the ranking needs the number.
  • Field 4 — acquisition:
    • FETCHED to inbox/ (1): Naci 2019 (LSE eprints OA mirror, 620 KB, validated). -> process inbox.
    • MAINTAINER-DROP (1): Roerecke 2017 — oa=False (Elsevier/Lancet Public Health paywall). NOT downgraded (pause-and-ask). Identified, not lost.
  • [2026-08-09] openalex search “environmental risk factors Parkinson umbrella review meta-analyses” -> 0 hits: (none)
  • [2026-08-09] openalex search “physical activity Parkinson disease risk meta-analysis” -> 0 hits: (none)
  • [2026-08-09] openalex search “modifiable lifestyle risk factors Parkinson disease systematic review” -> 0 hits: (none)
  • [2026-08-09] openalex search “Parkinson umbrella review” -> 3 hits: https://doi.org/10.3390/brainsci11060685; https://doi.org/10.1155/2020/9748091; https://doi.org/10.17605/osf.io/rc39j
  • [2026-08-09] openalex search “physical activity Parkinson meta-analysis” -> 3 hits: https://doi.org/10.1093/ptj/pzae114; https://doi.org/10.1097/md.0000000000040141; ?
  • [2026-08-09] openalex search “Parkinson risk factors meta-analysis” -> 2 hits: https://doi.org/10.3390/life14010032; https://doi.org/10.1007/s11010-025-05392-y
  • [2026-08-09] openalex search “coffee caffeine Parkinson meta-analysis” -> 0 hits: (none)
  • [2026-08-09] openalex search “Non-Genetic Risk Factors Parkinson Overview Systematic Reviews” -> 0 hits: (none)
  • [2026-08-09] openalex search “physical activity risk Parkinson disease systematic review” -> 0 hits: (none)
  • [2026-08-09] openalex search “levels physical activity risk Parkinson” -> 1 hits: https://doi.org/10.1001/jamanetworkopen.2018.2421
  • [2026-08-09] openalex search “caffeine risk progression Parkinson meta-analysis” -> 0 hits: (none)
  • [2026-08-09] openalex search “Parkinson overview 46 systematic reviews” -> 0 hits: (none)
  • [2026-08-09] openalex search “non-genetic risk factors Parkinson” -> 0 hits: (none)
  • [2026-08-09] openalex search “effect caffeine risk progression Parkinson” -> 0 hits: (none)
  • [2026-08-09] openalex resolve “10.3233/JPD-202521 10.1001/jamanetworkopen.2018.2421” -> 2 hits: https://doi.org/10.3233/jpd-202521; https://doi.org/10.1001/jamanetworkopen.2018.2421
  • [2026-08-09] openalex search “hallmarks of aging” -> 2 hits: https://doi.org/10.1016/j.cell.2013.05.039; https://doi.org/10.1016/j.cell.2022.11.001
  • [2026-08-09] openalex search “geroscience linking aging chronic disease” -> 2 hits: https://doi.org/10.1016/j.cell.2014.10.039; https://doi.org/10.3410/f.725274403.793502267
  • [2026-08-09] openalex search “multimorbidity shared risk factors ageing systematic review” -> 0 hits: (none)
  • [2026-08-09] openalex fetch “10.3233/JPD-202521” -> 0 hits: (none)
  • [2026-08-09] openalex fetch “10.1001/jamanetworkopen.2018.2421” -> 0 hits: (none)
  • [2026-08-09] openalex fetch “10.1016/j.cell.2014.10.039” -> 0 hits: (none)

Filling the two ## Gaps to ACQUIRE items in the Age-Related Diseases spec: (1) a Parkinson’s modifiable-risk-factor MA, (2) a geroscience/common-soil source for the mechanistic backbone. openalex auto-logged its queries above (fields 1+2); the web/PMC channel has no auto-log — hand-logged here.

  • Field 1 — hand-logged non-openalex queries:
    • WebSearch: “Parkinson’s disease modifiable risk factors umbrella review meta-analysis physical activity caffeine landmark” -> surfaced the Chen 2021 umbrella (46 SRs), a 2025 critical review, a caffeine MA (PMC7353179), the Fang PA MA.
    • WebFetch PMC8461677 -> extracted Chen 2021 DOI/title/journal (openalex title.search kept returning 0 from strictness — same --type over-narrowing lesson).
  • Field 2 — 0-hit attribution: the strict-title openalex queries (“environmental risk factors Parkinson umbrella review meta-analyses”, “modifiable lifestyle risk factors Parkinson”, “hallmarks of aging” narrow, “multimorbidity shared risk factors ageing” +SR) returned 0 from title.search strictness, NOT literature absence — the landmarks exist and were reached via broad nouns / web + DOI-resolve. Re-confirmed: drop --type, use core nouns, resolve the named landmark by DOI.
  • Field 3 — adjudication:
    • GAP 1 — Parkinson’s. SELECTED (2):
      • Chen 2021 10.3233/JPD-202521 — Non-Genetic Risk Factors for PD: An Overview of 46 Systematic Reviews (J Parkinson’s Dis; 55 cites; gold umbrella, 46 SRs / 80+ factors / 6 categories). The ranked modifiable-risk LANDSCAPE — exactly the thin-field survey the spec’s honest “PA is the one credible lever” verdict needs. PRIMARY.
      • Fang 2018 10.1001/jamanetworkopen.2018.2421 — Association of Levels of Physical Activity With Risk of PD (JAMA Netw Open; 169 cites; gold MA). The PA->PD magnitude (highest total/MVPA RR ~0.79/0.71) — the one credible modifiable lever’s effect size. PA is already a Big Rock, so this is the cross-link magnitude, not a new lever.
    • GAP 1 — DEFERRED: caffeine->PD MA (PMC7353179, 2020) — an ECHO of the Chen umbrella’s caffeine row; acquire only if the umbrella+Fang don’t carry the caffeine number. The 2025 critical review (PMC11922612) — narrative/moderate, superseded by the umbrella for the ranking.
    • GAP 2 — Geroscience/common-soil. SELECTED (1, mechanism-tier exception):
      • Kennedy 2014 10.1016/j.cell.2014.10.039 — Geroscience: Linking Aging to Chronic Disease (Cell; 2760 cites). This IS the common-soil thesis (aging biology as the shared root of the age-related diseases) — more ON-THESIS for a modifiable-levers deliverable than the molecular hallmarks enumeration. TIER: mechanism/narrative-review -> gold-gate exception, needs a source-tier-overrides.tsv entry at ingest: it informs the mechanistic BACKBONE directionally, NOT effect magnitudes (those stay with the held GBD/Afshin epidemiology). PRIMARY backbone pick.
    • GAP 2 — NOTED (already staged, not re-acquired): Lopez-Otin Hallmarks of Aging 2013 10.1016/j.cell.2013.05.039 (15134 cites, oa=True) + 2023 update An Expanding Universe 10.1016/j.cell.2022.11.001 (6084 cites, oa=False) — the molecular ENUMERATION (staged per spec). Prefer Kennedy as the common-soil spine; ingest a Hallmarks version only if the enumeration is wanted (take the 2023 update as current — but it’s paywalled). Fried Frailty Phenotype / Barzilai TAME / CALERIE also staged per spec — not re-searched here.
    • GAP 2 — 0 gold epidemiological common-soil MA found (“multimorbidity shared risk factors” +SR = 0). Not a literature-absence claim on one query, but the epidemiological common-soil leg is already HELD (GBD/Afshin), so the residual gap is the MECHANISTIC backbone only — Kennedy fills it. No further search owed.
  • Field 4 — acquisition status (ALL maintainer-drop; NONE fetched — pause-and-ask):
    • 10.3233/JPD-202521 Chen — oa=True (IOSPress) but plain-fetch HTTPError AND headless 403. maintainer-drop.
    • 10.1001/jamanetworkopen.2018.2421 Fang — oa=True but JAMA anti-bot-403 (measured). maintainer-drop.
    • 10.1016/j.cell.2014.10.039 Kennedy — oa=True but Cell/Elsevier anti-bot-403 (measured). maintainer-drop.
    • 10.1016/j.cell.2022.11.001 Lopez-Otin 2023 update — oa=False (Cell paywall). maintainer-drop (only if the enumeration is wanted).
    • NOT downgraded/substituted — all identified (DOI + OA status recorded), not lost.
    • Priority: Chen + Fang (Parkinson’s leg, self-contained) -> Kennedy (geroscience backbone, needs tier-override).
  • [2026-08-09] openalex search “triglyceride HDL ratio cardiovascular meta-analysis” -> 0 hits: (none)
  • [2026-08-09] openalex search “triglyceride glucose index cardiovascular meta-analysis” -> 8 hits: https://doi.org/10.1186/s12933-021-01268-9; https://doi.org/10.1186/s12933-022-01546-0; https://doi.org/10.3389/fcvm.2021.774781; https://doi.org/10.1016/j.numecd.2021.08.026; https://doi.org/10.1186/s12933-025-02584-0
  • [2026-08-09] openalex search “insulin resistance cardiovascular disease meta-analysis” -> 7 hits: https://doi.org/10.1002/dmrr.1201; https://doi.org/10.2337/dc24-0475; https://doi.org/10.3389/fendo.2025.1699985; https://doi.org/10.1093/eurheartj/ehae666.2920; https://doi.org/10.2337/figshare.26083753
  • [2026-08-09] openalex search “triglyceride HDL ratio meta-analysis” -> 3 hits: https://doi.org/10.1093/labmed/lmae052; https://doi.org/10.1002/edm2.70161; https://doi.org/10.1186/s12884-026-09647-4
  • [2026-08-09] openalex search “triglyceride HDL cholesterol coronary meta-analysis” -> 2 hits: https://doi.org/10.3945/jn.111.148973; https://doi.org/10.1016/j.yane.2012.02.028
  • [2026-08-09] openalex search “atherogenic index plasma cardiovascular meta-analysis” -> 3 hits: https://doi.org/10.1186/s12933-024-02198-y; https://doi.org/10.1016/j.jscai.2025.102782; https://doi.org/10.2174/0109298673412215251211122028
  • [2026-08-09] openalex resolve “10.1186/s12933-021-01268-9 10.2337/dc24-0475” -> 2 hits: https://doi.org/10.1186/s12933-021-01268-9; https://doi.org/10.2337/dc24-0475
  • [2026-08-09] openalex resolve “10.1186/s12933-024-02198-y 10.1002/edm2.70161 10.1186/s12933-022-01546-0” -> 3 hits: https://doi.org/10.1186/s12933-024-02198-y; https://doi.org/10.1002/edm2.70161; https://doi.org/10.1186/s12933-022-01546-0
  • [2026-08-09] openalex search “HOMA insulin resistance coronary meta-analysis” -> 0 hits: (none)
  • [2026-08-09] openalex search “homeostatic model insulin resistance cardiovascular meta-analysis” -> 0 hits: (none)
  • [2026-08-09] openalex search “triglyceride HDL ratio mortality meta-analysis” -> 0 hits: (none)
  • [2026-08-09] openalex search “triglyceride HDL ratio incident cardiovascular” -> 2 hits: https://doi.org/10.2215/cjn.08730816; https://doi.org/10.1016/j.cca.2023.117737
  • [2026-08-09] openalex fetch “10.1186/s12933-022-01546-0” -> 0 hits: (none)
  • [2026-08-09] openalex fetch “10.1186/s12933-021-01268-9” -> 0 hits: (none)

[2026-08-09 #6] SELECTION RECORD — TG/HDL ratio + insulin resistance -> CVD (follows the Bikman challenge [2026-08-09])

Scope. Two decision-questions the Bikman challenge surfaced as PREDICTION-magnitude gaps (the fabric holds the apoB causal resolution — Marston: TG/LDL-C flat once apoB held — but no dedicated source on TG/HDL or IR-index prediction of CVD): (Q1) is the TG/HDL ratio predictive of CVD? (Q2) is insulin resistance predictive of CVD? Dedup: registry holds NO TG/HDL / TyG / HOMA-IR / IR-index -> CVD row (confirmed gap).

The consolidating insight. The TyG index (triglycerides + glucose) is BOTH a triglyceride-based ratio AND the dominant validated insulin-resistance surrogate — so the TyG MAs answer Q1 and Q2 together with the strongest gold evidence. HOMA-IR (classic direct IR) and raw TG/HDL have been largely SUPERSEDED by TyG in the recent CVD-prediction MA literature.

SELECTED — fetched to inbox/ (2 gold OA MAs, Cardiovascular Diabetology):

  • 10.1186/s12933-022-01546-0 Liu 2022 (431 cites) — “TyG index and risk of CVD and mortality in the general population”: SR+MA. LEAD for weighting — best fit (general population, CVD incidence + mortality). -> inbox/Liu - Triglyceride-Glucose Index Cardiovascular Mortality Meta-Analysis 2022.pdf.
  • 10.1186/s12933-021-01268-9 Ding 2021 (514 cites) — “TyG index and incidence of ASCVD”: MA of cohorts. LANDMARK (most-cited, kept for provenance; likely shares cohorts with Liu -> ingest decides subsumption). -> inbox/Ding - Triglyceride-Glucose Index ASCVD Incidence Meta-Analysis 2021.pdf. Both answer Q1+Q2; observational-cohort evidence pooled = the right evidence-type for a prediction question, and it dovetails with the held apoB causation nucleus (prediction vs causation, the challenge’s split).

DEFERRED (identified, NOT fetched — different question / thinner):

  • 10.1186/s12933-024-02198-y AIP-in-CAD 2024 (116 cites, OA) — atherogenic index of plasma = log(TG/HDL), but population = established coronary artery disease => prognosis in CAD, not incidence prediction. The source to fetch IF a secondary-prevention prognosis angle is later wanted; wrong question for Q1 now.
  • 10.3389/fendo.2025.1699985 METS-IR 2025 (5 cites, OA) — a different IR score; newer/thinner, TyG MAs dominate.

REJECTED (fit-to-question over pedigree — the pedigree hid the mismatch):

  • 10.2337/dc24-0475 IR->CVD Diabetes Care 2024 (86 cites, top journal) — population = Type 1 Diabetes; Bikman’s claim is general/metabolic-syndrome IR, not T1D. Wrong population.
  • 10.1002/edm2.70161 TG/HDL SR 2026 (OA) — outcome = Chronic Kidney Disease, not CVD. Off-outcome.
  • 10.2215/cjn.08730816 TG/HDL 2017 (67 cites, OA) — single cohort (not MA), CJASN/nephrology population.
  • 10.1093/labmed/lmae052 TG/HDL 2024 — outcome = diabetes risk, not CVD. 10.1002/dmrr.1201 visfatin 2011 — off-question biomarker. 10.3945/jn.111.148973 — DHA/algal-oil intervention on TG (off-question).

ZERO-HIT attributions (all title-strictness artifacts, not literature absence):

  • “triglyceride HDL ratio cardiovascular meta-analysis” (0), “…mortality” (0) — every word forced into title; the ratio literature indexes under “TyG”, “atherogenic index of plasma”, or names TG/HDL mid-title. Looser “triglyceride HDL ratio” returned 3.
  • “HOMA insulin resistance coronary meta-analysis” (0) + “homeostatic model insulin resistance cardiovascular meta-analysis” (0) — strictness; ALSO a real signal that HOMA-IR->CVD gold MAs are sparse vs the TyG surge.

FINDING (a gap, honestly recorded): no dedicated general-population gold MA of the RAW TG/HDL ratio -> incident CVD was located — the field consolidated onto TyG (prediction) + AIP (CAD prognosis). Q1 is therefore best served via TyG, with the caveat that raw-TG/HDL-specific pooled evidence is thin.

Acquisition status: 2 fetched to inbox/ (OA, clean). 0 blocked / 0 maintainer-drop. Hand off to process inbox (dedup vs done/cut/queue, chunk/OCR check, register rows + tier, stage). Nothing ingested.

[2026-08-09 #7] SELECTION RECORD — HOMA-%S (insulin sensitivity) + HOMA-%B (beta-cell function) -> CVD (continues #6 / the Bikman thread)

Scope. HOMA2 outputs beyond HOMA-IR: %S = insulin sensitivity (~ 1/HOMA-IR), %B = beta-cell function (insulin secretion). Question: are %S / %B predictive of CVD / mortality? Includes the Bikman-relevant hyperinsulinemia framing (high fasting insulin = the compensatory-high-%B state he argues is causal), and %B’s natural outcome (T2D incidence/progression). Dedup: registry holds NO HOMA / fasting-insulin / beta-cell row (confirmed #6 + this pass).

OUTCOME — a GAP, nothing acquired (gold-gate not clearable; pause-and-ask, NOT downgrade). Searched %S->mortality/CVD, %B/insulin-secretion->CVD + T2D-incidence, fasting-insulin->CVD, hyperinsulinemia->CHD, beta-cell-decline->progression, across --type meta-analysis, --type systematic-review, and broad. No gold-tier pooled MA/SR/umbrella of HOMA-%S, HOMA-%B, fasting insulin, or hyperinsulinemia on hard CV outcomes was located. The literature is a scatter of single cohorts, several 1990s-2000s, and openly contested.

Why no acquire (symmetric standards, applied): the prior turn (#6) found gold pooled evidence for the TyG IR-surrogate (Liu 2022 / Ding 2021, now in inbox), and apoB causation is held gold. Acquiring a lone 1990s cohort for %B/hyperinsulinemia would hold that Bikman-load-bearing exposure to a WEAKER bar than the lipoprotein evidence it is pitched against — the exact asymmetric-scrutiny trap. The paucity is itself the decision-relevant finding: the hyperinsulinemia/secretion thesis lacks the pooled hard-outcome evidence that TyG and apoB have. (%S is in any case ~1/HOMA-IR, so it is substantively covered by the acquired TyG MAs; the genuinely un-pooled arm is %B / insulin secretion / hyperinsulinemia.)

Single cohorts IDENTIFIED (not lost; candidates IF a deliberate counterweight is later wanted) — NOT gold:

  • 10.1161/01.CIR.98.5.398 Hyperinsulinaemia predicts CHD, healthy middle-aged men, 1998 (351 cites, OA) — the landmark single-cohort hyperinsulinemia->CHD (the Bikman-adjacent finding). Single prospective cohort.
  • 10.1007/s00125-009-1588-0 HOMA insulin-sensitivity index -> all-cause + CV mortality, Diabetologia 2009 (56 cites, OA) — the closest DIRECT HOMA-%S->CVD/mortality cohort. Single cohort.
  • 10.1016/j.jacc.2005.02.093 Impaired insulin sensitivity as independent mortality risk factor, JACC 2005 (304 cites, OA) — single study. 10.1177/1479164119827604 baseline fasting insulin -> MACE 2019 (OA).
  • 10.1016/0002-8703(91)90434-J Reaven, compensatory-hyperinsulinemia narrative 1991 (254 cites) — the FOUNDING narrative, not evidence. 10.1159/000422658 HOMA/CIGMA methods chapter — a METHOD/context ref only.
  • Contested-signal marker: 10.1111/j.1532-5415.1997.tb05163.x “Elevated Fasting Plasma Insulin… An Innocent Companion?” 1997 — the literature disputes causality vs bystander; do not read the positive cohorts as settled.
  • Beta-cell RECOVERY (already held, different question): 10.1007/s00125-011-2204-7 Lim/Taylor 2011 (1160 cites) — VLCD normalises beta-cell function / reverses T2D -> already covered by Total Diet Replacement and Type 2 Diabetes Remission + the twin-cycle spine; a reversal finding, not a prediction MA.

ZERO-HIT attributions: --type meta-analysis/systematic-review on beta-cell/insulin-sensitivity/ fasting-insulin/hyperinsulinemia all 0 — PARTLY title-strictness, but the broad (no---type) reruns returned ONLY single-cohort articles, so the MA absence is a real corpus signal, not a query artifact.

Acquisition status: 0 fetched, 0 maintainer-drop. GAP recorded. PAUSE-AND-ASK: does the maintainer want a deliberate non-gold counterweight (the 1998 Circulation hyperinsulinemia->CHD cohort + the 2009 Diabetologia HOMA-%S->mortality cohort, both OA, each needing a source-tier-overrides.tsv entry at ingest), or leave the gap recorded? Default = leave as recorded gap (do not downgrade). Nothing ingested.

[2026-08-09 #8] SELECTION RECORD — HOMA/insulin -> CVD, maintainer-directed acquire ONE recent high/medium (corrects #7)

Directive. Maintainer: “acquire one (recent) high/medium source” for the HOMA/insulin -> CVD question. Re-searched hard-outcome + recent pooled evidence (queries auto-logged above). This CORRECTS #7’s over-broad claim that “no gold-tier pooled MA of HOMA/fasting-insulin -> CVD exists”: strict multiword + --type queries had masked several. Pooled MAs DO exist — the true residual gap is narrower (below).

Candidates surfaced (hard-outcome / recent pooled):

  • SELECTED -> inbox: 10.1042/bsr20170947 Zhang (Xiaohong) et al., Fasting insulin, insulin resistance, and risk of cardiovascular or all-cause mortality in NON-DIABETIC adults: a meta-analysis, Bioscience Reports 2017 (65 cites, gold OA via PMC6448479). Fit-to-question WINNER: exposure = fasting insulin AND HOMA-IR (fasting insulin = Bikman’s exact hyperinsulinemia claim), population = non-diabetic / general (right stratum), outcomes = hard (CV + all-cause mortality), 7 studies / 26,976 adults. Result is honestly HEDGED: all-cause RR 1.13 (1.00-1.27, P=0.058, borderline NS) for fasting insulin vs 1.34 (1.11-1.62) for HOMA-IR; CV-mortality RR 2.11 (1.01-4.41) HOMA-IR but only 2 studies (wide CI), fasting-insulin CV-mortality 1.40 (0.49-3.96, NS, 1 study). Exactly the adjudication the Bikman thread needs: IR modestly predicts hard outcomes; fasting-insulin-ALONE is weaker/borderline than HOMA-IR. Likely tier at ingest = moderate [downgrade: small study counts, wide CIs, 2017]; type = meta-analysis.
  • REJECTED 10.2337/dc24-0475 Diabetes Care 2024 (86 cites) — most recent/highest, BUT population = Type 1 Diabetes, exposure = eGDR (not fasting insulin/HOMA), and closed-access. Population + metric mismatch (the same T1D trap rejected in #6). Not acquired.
  • NOTED (not taken) 10.3389/fendo.2023.1271857 Li et al., HOMA-IR -> coronary artery calcification MA, Front Endocrinol 2023 (15 studies, 60,649; OA; GRADE self-rated LOW/very-low). More recent + OA, but endpoint is a SURROGATE (CAC) and certainty low -> weaker steel-man than a hard-outcome MA. Available if the maintainer later wants the recent-surrogate companion.

Recency tradeoff, stated: picked 2017-hard-outcome over 2023-surrogate/2024-T1D per the telos rule recency is an input to relevance, never a quality rank + fit-to-question not pedigree — the acquired MA uniquely matches exposure (fasting insulin) x population (non-diabetic) x endpoint (hard). Fetch: the registry OA url (bioscirep.org) is a DEAD host -> recovered via Europe PMC render (PMC6448479), no maintainer-drop needed.

RESIDUAL GAP (the true #7 finding, narrowed): still NO pooled MA of HOMA-%B / insulin SECRETION -> CVD or T2D-incidence — %B / beta-cell function remains single-cohort only. HOMA-%S is now covered in spirit (fasting insulin + HOMA-IR ~ 1/%S pooled here; TyG MAs Liu/Ding staged). The un-pooled arm is secretion/%B.

Acquisition status: 1 fetched (Zhang 2017 -> inbox/, gitignored), 0 maintainer-drop, 0 ingested. Hand off to process inbox (pairs with the 2 staged TyG MAs as the IR-surrogate -> CVD-prediction node).

[2026-08-13] SELECTION RECORD — “top-tier acquire” pass -> ALL FOUR ALREADY INGESTED; two gaps confirmed no-gold; maintainer-drop re-surfaced

Trigger: maintainer approved a research pass on the AWAITS-review “top-tier acquisition targets” (Ding TyG ASCVD 2021, Bougma iodine, Falkingham iron) + the queue’s “top next-up” Liu 2022, plus the gene x exposure gap; and to surface the maintainer-drop blocked list.

Field 3 — adjudication (the negative space is the finding):

  • Ding - Triglyceride-Glucose Index ASCVD 2021 — NOT an acquire. INGESTED 2026-08-09 (source page + real extraction, cited [EXTRACTED] on Insulin Resistance Surrogates and Cardiovascular Risk as a type-F extension onto Liu; own log ingest entry). The [AWAITS (Ding...)] tokens still visible are STALE narration inside the Liu ingest entry (written minutes before Ding landed), living only in log.md (historical, append-don’t-rewrite) + hot.md (cache) — never on a claim page, so reconcile.py awaits-arrived is correctly clean.
  • Bougma iodine 2013 · Falkingham iron 2010 — NOT acquires. INGESTED 2026-08-04 as the deficiency-repletion trio (items 2/3, 3/3), cited on Deficiency Repletion vs Enhancement; AWAITS cashed at ingest. Registered gold.
  • Liu TyG CV-mortality 2022 — NOT an acquire. INGESTED 2026-08-09 (minted the ir-surrogates nucleus).
  • => The AWAITS-review “acquisition gap” was a false alarm — it read historical log/hot AWAITS narration as live acquisition targets. Correct state: 0 of the 4 need acquiring.
  • gene x diet / polygenic x lifestyle interaction on hard outcomes (the over-personalization [PRIOR] route-(b) test) — RE-PROBED, confirms the 2026-08-06 ingest-queue finding: NO gold gene-diet SR/MA surfaces. Only hits: one non-OA weak 2024 obesity-genetics x lifestyle article (10.1016/j.ypmed.2024.107886, cites 9); zero for polygenic x diet. INSUFFICIENT-EVIDENCE literature gap, NOT no-effect. REJECTED acquire (gold-gate + do-not-acquire-weak). Landmark Khera 2016 NEJM is cohort-pooled, not SR/MA -> below the gold bar; gap stands.
  • TG/HDL-C ratio -> CV outcomes (the genuine open 2nd AWAITS on the ir-surrogates node — generalize beyond one IR surrogate) — PROBED, NO gold SR/MA. Best OA hit is a 2023 MDPI Diagnostics narrative review (10.3390/diagnostics13050929, weak/moderate); remaining hits single cohorts / cross-sectional. REJECTED acquire (gold-gate). The AWAITS stays OPEN but is a G-gap (no gold evidence exists to fill it), not an acquisition task.

Field 4 — acquisition status: 0 fetched, 0 ingested, 0 NEW maintainer-drop. Nothing to hand to process inbox. Maintainer-drop blocked list RE-SURFACED for decision (all established blocked in prior passes; not re-attempted). Deduplicated of since-resolved items (Chen Parkinson 2021, Messier IDEA 2013, Siervo DASH MA 2015 are now HELD):

  • Roerecke 2017 alcohol -> CVD/BP dose-response MA — 10.1016/S2468-2667(17)30003-8 (Lancet Public Health) — oa=False paywall. [distinct from held Roerecke 2010]
  • Fang 2018 physical activity -> Parkinson’s MA — 10.1001/jamanetworkopen.2018.2421 — oa=True but JAMA anti-bot 403. [PD leg]
  • Kennedy - Geroscience 2014 — 10.1016/j.cell.2014.10.039 — oa=True but Cell/Elsevier 403. [registry placeholder, not held; needs tier context]
  • Lopez-Otin - Hallmarks of Aging 2023 update — 10.1016/j.cell.2022.11.001 — oa=False Cell paywall. [only if the enumeration is wanted; not registered]
  • Siervo - nitrate/beetroot -> BP MA 2013 — 10.3945/jn.112.170233 — oa=True but ScienceDirect 403. [distinct from held Siervo DASH MA 2015]
  • Fried - Frailty Phenotype 2023 update — paywalled. [registry holds the 2001 original as placeholder]
  • [2026-08-13] openalex search “sleep duration all-cause mortality dose-response meta-analysis meta-analysis” -> 6 hits: https://doi.org/10.1161/jaha.117.005947; https://doi.org/10.1186/s12889-020-09275-3; https://doi.org/10.2196/preprints.17705; https://doi.org/10.21203/rs.2.22427/v3; https://doi.org/10.21203/rs.2.22427/v2
  • [2026-08-13] openalex fetch “10.1161/jaha.117.005947” -> 0 hits: (none)

[2026-08-13] SELECTION RECORD — sleep-mortality dose-response (challenge-driven)

Trigger: maintainer challenge — the fabric gives a single POOLED short-arm mortality RR (Cappuccio 2010, RR 1.12 for <7h) and cannot resolve <5h / 5-6h / 6-7h. Held source structurally bundles them (short-arm flat across cutoffs, P=0.89), so this is a genuine silent-gap, not a fabric error.

  • Query/channel: openalex search "sleep duration all-cause mortality dose-response meta-analysis" --type meta-analysis --from-year 2015 (auto-logged above; 6 hits).
  • Adjudication:
    • SELECTED — Yin 2017, JAHA 10.1161/JAHA.117.005947 (717 cites, gold, dose-response non-linear MA, 7h nadir; the landmark that gives the per-hour curve Cappuccio pools). The exact instrument the gap needs.
    • REJECTED — the three 2020 “older people / elderly” MAs (s12889-020-09275-3, rs.2.22427/v2,v3, 2196/preprints.17705): narrower (elderly-only subpopulation), lower-cited, and one is a preprint pair (same work). Yin dominates on scope + design + citation; no independent perspective added -> ratchet-reject.
    • DEFERRED — none.
  • Acquisition: Yin 2017 FETCHED. openalex direct fetch FAILED (AHA IP-wall) -> escalated: EuropePMC OA copy PMC5634263 via fetch_browser.py --download (the render URL fires a download event; the --download flag captures it — plain goto/curl do not). 6.4MB/82pp incl. supplement, text-clean. Staged to raw/binary/papers/, registry row added (yin2017, gold). NOT yet ingested (delegated).
  • [2026-08-13] openalex search “fruit vegetable intake cardiovascular cancer mortality dose-response meta-analysis” -> 2 hits: https://doi.org/10.1093/ije/dyw319; https://doi.org/10.3389/fnut.2023.1153165
  • [2026-08-13] openalex search “fruit vegetable consumption mortality dose response prospective meta-analysis” -> 1 hits: https://doi.org/10.1093/nutrit/nuaf013
  • [2026-08-13] openalex resolve “10.1093/ije/dyw319” -> 1 hits: https://doi.org/10.1093/ije/dyw319
  • [2026-08-13] openalex fetch “10.1093/ije/dyw319” -> 0 hits: (none)

[2026-08-13] SELECTION RECORD — fruit/veg -> mortality/stroke dose-response (challenge-driven)

Trigger: maintainer challenge on the Plant Foods deliverable (2026-08-13, sidecar R-c) — the deliverable quotes ESC 2021’s BORROWED, CI-less fruit/veg->stroke RRs (11%/26%/4%-per-serving), which are themselves an MA-summary (ESC refs 435,436) the wiki does not hold. = the Plant Foods spec’s #1 named gap. Goal: acquire the held effect+CI+shape MA to replace the borrowed ESC bare RR.

[2026-08-18 #9] SELECTION RECORD — HOMA-%S / HOMA-%B as an INDICATION OF INSULIN RESISTANCE (validity, not outcome)

Scope + how it differs from 7/8. Maintainer question: “are HOMA-%S and -%B useful as an indication of insulin resistance?” This is the measurement-validity question (does the readout track true IR / clamp), NOT the 7/8 hard-outcome question (does %S/%B predict CVD/mortality). Distinct search target: agreement vs the euglycemic-hyperinsulinemic clamp, and the correct-use/limits of the HOMA2 calibrated model. Split the two indices up front: %S = insulin SENSITIVITY (~ 1/HOMA-IR) — directly on-question as an IR readout; %B = beta-cell FUNCTION (secretion) — NOT insulin resistance itself, so “%B as an indication of IR” is a partial category mismatch (it indexes the secretion limb, useful for the beta-cell side of the T2D story, not for resistance). Dedup: fabric holds HOMA-IR (Zhang 2017 on Insulin Resistance Surrogates and Cardiovascular Risk)

  • TyG; holds NOTHING on the HOMA2 %S/%B calibrated model or its validity.

Finding — NO gold SR/MA on HOMA2-%S/%B validity exists; the evidence base is foundational method papers + scattered single-cohort clamp-agreement studies. --type systematic-review/meta-analysis on HOMA validity/clamp all 0; broad reruns returned only the two model papers + single-cohort index-vs-clamp comparisons. Every “insulin resistance indices” MA that surfaced is an INTERVENTION review (chromium, exercise, green-coffee, DHEA, erectile-dysfunction association) that merely USES an index as an outcome — off-question, rejected. “HOMA-IR diagnostic accuracy” MA = 0. The absence of a pooled validity SR is itself the reportable finding (a method-defined, thin-for-pooling field).

Candidates (fit-to-question, tier, adjudication):

  • SELECTED -> maintainer-drop (BLOCKED): 10.2337/diacare.27.6.1487 Wallace, Levy, Matthews, Use and Abuse of HOMA Modeling, Diabetes Care 2004 (5153 cites, OA-flagged). THE defining reference for exactly the two indices asked about: how HOMA2 yields %S and %B from paired fasting glucose+insulin (or C-peptide), and — the on-question payload — when the readouts are and are not valid. Ratchet-advancer (fabric holds nothing on HOMA2). Tier at ingest = method/moderate, deliberate gold-gate exception (foundational method reference in a field with no gold validity SR) -> will need a source-tier-overrides.tsv entry. FETCH BLOCKED: openalex direct 403 + local Playwright 403 (diabetesjournals.org, OUP/Diabetes-Care anti-bot class). PAUSE-AND-ASK: maintainer to drop the PDF into inbox/. DOI recorded. NOT downgraded.
  • DEFERRED 10.2337/diacare.21.12.2191 Levy, Matthews, Hermans, Correct HOMA Evaluation Uses the Computer Program, Diabetes Care 1998 (2048 cites, OA=False). Introduces the HOMA2 computer model (why the 1985 linear HOMA1 equations diverge from the properly-computed %S/%B). Substantively SUBSUMED by Wallace 2004 for our purpose; acquire only if the 2004 read shows a gap. Also OUP-blocked if wanted.
  • DEFERRED (single-cohort clamp-agreement, NOT gold; candidates only if a counterweight is wanted): 10.2337/diacare.22.9.1462 Matsuda-DeFronzo 1999 (OGTT indices vs clamp; landmark but OGTT-derived, not HOMA2); 10.1016/j.metabol.2005.11.005 (2006) + 10.1016/j.metabol.2007.04.002 (2007) simple-indices-vs-clamp comparisons; 10.1186/1472-6823-14-65 (BMC Endocr Disord 2014, OA) — the counter-signal “fasting insulin sensitivity indices are not better than routine clinical variables”, a symmetric-standards keep-in-view.
  • REJECTED: 10.1080/10641963.2026.2625124 / figshare (2026, non-insulin indices for pre-hypertension — TyG class, insulin-FREE, so off-question for %S/%B); all intervention-outcome “IR indices” MAs (see above).

Acquisition status: 0 fetched, 1 maintainer-drop (Wallace 2004, DOI 10.2337/diacare.27.6.1487), 0 ingested. Nothing woven — the answer to the maintainer is given as UNHELD/routed, not adopted into fabric (feedback rules no-definitive-claims-from-nonfabric / no-fabric-edit-ungrounded-in-held-fabric). If acquired + ingested, the natural home is a facet on Insulin Resistance Surrogates and Cardiovascular Risk (the %S~1/HOMA-IR arm) plus the beta-cell/%B secretion limb that ties to the twin-cycle spine.

[2026-08-18 #10] SELECTION RECORD — HOW POPULATION LIFESTYLE GUIDELINES ARE DEVELOPED (method groundwork for a fabric-grounded guideline)

Question served: how respectable bodies (WHO, NICE, USPSTF, GRADE) turn evidence into a population guideline — the process/method, as groundwork for writing a fabric-grounded, whole-population, lifestyle-focused guideline. Maintainer constraint: do NOT anchor on status quo; do NOT weigh feasibility / sustainability / carbon / ethics-of-meat. That constraint is a deliberate DEPARTURE from standard Evidence-to-Decision (EtD), which blends exactly those non-health criteria — it maps to the fabric’s “one axis, named — never blended” rule.

Already HELD (no acquisition needed):

ACQUIRED / PREPARED this episode:

  • CHUNKED (was in-hand, un-ingested): WHO - Handbook for Guideline Development 2e (main PDF, ch1-13) -> raw/sources/who---handbook-for-guideline-development-2e/ (9 chunks). Chunk 06 = ch10 “Developing recommendations” carries WHO’s native EtD-criteria enumeration (quality of evidence; values & preferences; balance of benefits/harms; resource implications; priority of the problem; equity & human rights; acceptability; feasibility -> strong/conditional). This is the criteria list the maintainer’s design keeps-or-strips, in a guideline body’s own voice. NOT yet ingested (delegated op, separate turn).
  • FETCHED to inbox/: Moberg et al. 2018, The GRADE Evidence to Decision (EtD) framework for health-system and public-health decisions (DOI 10.1186/s12961-018-0320-2, Health Res Policy Syst, OA) -> inbox/Group - The GRADE Evidence to Decision EtD 2018.pdf. SELECTED as the POPULATION/public-health EtD variant — most on-question for a whole-population guideline; enumerates the public-health EtD criteria (adds equity/acceptability/feasibility explicitly at population scale).

SELECTED but fetch-BLOCKED -> maintainer-drop (pause-and-ask, did NOT downgrade):

  • Alonso-Coello et al. 2016, GRADE EtD frameworks 2: from evidence to recommendations — DOI 10.1136/bmj.i2089 (BMJ, OA in principle; publisher 403-blocked to openalex + headless). The canonical EtD criteria/structure paper. Now LOWER priority — WHO ch10 (chunked) + Moberg 2018 (fetched) already cover the criteria enumeration; acquire only if the formal EtD table structure is wanted verbatim.
  • Alonso-Coello et al. 2016, GRADE EtD frameworks 1: introduction — DOI 10.1136/bmj.i2016 (BMJ). Same block; companion to the above; optional.

DEFERRED (identified, not taken):

  • AGREE II (Brouwers et al. 2010, 10.1503/cmaj.090449, CMAJ, OA) — guideline-QUALITY appraisal instrument (a checklist for whether a finished guideline is trustworthy), not a construction method. Useful as a self-audit rubric for OUR guideline later; not needed for the “how are they made” question.
  • NICE Developing NICE guidelines: the manual (PMG20) — NICE’s process manual; web PDF, no clean DOI. NICE’s process is GRADE-adjacent; the held GRADE Handbook + WHO handbook cover the shared machinery. Acquire only if NICE-specific process detail (e.g. its committee/consultation structure) is wanted.

Acquisition status: 1 fetched (Moberg 2018 -> inbox/), 1 source chunked (WHO handbook), 2 maintainer-drop (BMJ i2089/i2016, optional), 2 deferred (AGREE II, NICE PMG20), 0 ingested. Nothing woven — Research/discovery only. The guideline itself is a later Phase-2 Compile, NOT written this turn.

[2026-08-19] SELECTION RECORD — RATCHET / LIGHTHOUSE SURVEY (better sources + disagreeing gold)

Batched research op across the lighthouse-index.md ratchet gap map (2026-08-07): find a GOLD advancer for each sub-gold UPGRADE row (lens A = “better sources”) and a DIVERGENT gold for each CLOSED-gold TENSION-acquire (lens B = “disagreeing gold”). 5 parallel discovery agents; discovery + vetting only, symmetric standards, author-diff independence checks per the named excluded lineages. Acquisition centralized here (pause-and-ask on block, no downgrade). Auto-log (fields 1+2) above.

Lens A — UPGRADE (a gold that supersedes a sub-gold incumbent)

  • grip strength -> mortality (held Celis-Morales 2018 high cohort): SELECTED Lopez-Bueno 2022 10.1016/j.arr.2022.101778 (48-study dose-response MA, 3.1M, gives kg thresholds; independent Spain/Denmark group). Corroboration DEFERRED Wu 2017 10.1016/j.jamda.2017.03.011 (MA, paywalled). Marker-not-lever caveat retained (upgrades tier of the ASSOCIATION, not decision-status).
  • leucine threshold, elderly (held Katsanos 2006 weak): SELECTED Komar 2015 10.1007/s12603-014-0559-4 (SR-MA, independent Schwingshackl group) — weak->gold; caveat it measures supplementation EFFECT on strength, not a threshold NUMBER (partial fit).
  • organic food -> health (held Baranski/Baudry moderate, composition-only): SELECTED Theodoridis 2025 10.3390/life15020160 (21-study cancer-incidence MA, HR 0.93 [0.78-1.12] NULL) — the gold HARD-OUTCOME MA the incumbents lacked; resolves the gap to a null. Journal MDPI-Life -> register high not auto-gold (journal-tempered).
  • Ca +/- vit-D -> CV harm (held Bolland 2011 high, contested): SELECTED Myung 2021 10.3390/nu13020368 (13 double-blind RCTs, CVD RR 1.15 [1.06-1.25]) — newer, RCT-only, INDEPENDENT of Bolland/Reid (Korean group) -> both tier-upgrade AND type-E corroboration. 10.1007/s12603-020-1551-9 (vit-D-alone/combined arm) DEFERRED as an adjacent null.
  • MedDiet -> dementia (held Garcia-Casares 2021 high, 3 cross-sectional): SELECTED Nucci 2024 10.1007/s40520-024-02718-6 (21 studies, prospective-cohort-inclusive; dementia OR 0.89, AD OR 0.73). Caveat: mixed designs, observational, plausibly cardiometabolic-mediated -> upgrades certainty/tier, not causal identification.
  • carbohydrate %E -> mortality U-shape (held Seidelmann 2018 high, single cohort + 8-study MA): SELECTED Clinical Nutrition 2022 10.1016/j.clnu.2022.12.010 (41-study restricted-cubic-spline dose-response SR-MA; non-linear all-cause mortality, mid-carb nadir; independent of the Willett/Seidelmann lineage).
  • weight-loss maintenance / regain (held Franz 2007 high, dated ~1yr): SELECTED Machado 2022 10.1016/j.clnesp.2022.03.020 (27-trial regain-TRAJECTORY MA; independent Brazilian group). Distinct from regain-BIOLOGY (held). DEFERRED Flore 2022 10.3390/nu14061259 — different question (does intensive maintenance intervention beat control), not the descriptive trajectory.
  • elderly protein g/kg TARGET (index credits Bauer PROT-AGE 2013 high): NO independent gold advancer. The three “supplementation-efficacy” MAs are REJECTED off-question (Ten Haaf 2018 10.1093/ajcn/nqy192; 10.3390/nu16070941; 10.1186/s12877-025-05885-x — they answer does supplementation help, not what g/kg target). Only step is DEFERRED Coelho-Junior 2022 10.1016/j.arr.2022.101731 — a SAME-lineage F-update of the already-STAGED gold Coelho-Junior 2018 10.3390/nu10091330. Index correction owed: this row’s on-question incumbent is the staged Coelho-Junior 2018, not Bauer.
  • per-meal protein dose, older vs younger (held Moore 2014 moderate): NO gold advancer — MA/SR queries 0-hit (genuine literature absence; per-meal dose answered only at acute-MPS surrogate level). Accept Moore 2014; surrogate-only gap stands.

Lens B — DISAGREEING GOLD (a divergent gold that flips a CLOSED-gold to TENSION)

  • eggs -> CVD (held Godos 2020 gold, near-null; Zhong 2019 high cohort STAGED): SELECTED Ma 2022 10.1093/jn/nxac105 (dose-response MA; egg intake linearly assoc. CVD mortality in US / >60y / hyperlipidemia strata; independent Chinese group). Lifts the harm signal from the staged Zhong cohort to MA grade. Stratified/conditional tension, not a flat reversal — a near-null MA cluster also exists (Advances in Nutrition 2022, Godos-concurring). File as US/older-conditional.
  • low-carb/keto -> HARD outcomes (held Naude 2022 + Ge 2020 gold, CLOSED on weight/markers): SELECTED Ghorbani 2023 10.1016/j.arr.2023.101997 (dose-response MA; plant-LCD lower all-cause mortality, animal-LCD higher cancer death; independent Iranian group). Corroboration SELECTED Qin & Bo 2023 10.1039/d3fo01374j (LCD assoc. CHD; lower-impact journal -> corroboration not 2nd nucleus). DEDUP NOTE: the direction is ALREADY in the fabric via ingested Seidelmann 2018 -> these REINFORCE an existing tension (type-E/F), not a new one. Ghorbani also serves the Lens-A carb-U-shape row.
  • UPF -> health, is there an independent PROCESSING effect? (held Lane 2024 umbrella + Hall 2019 RCT gold): SELECTED Aramburu 2024 10.3389/fnut.2024.1421728 (SR of 4 RCTs; declines to attribute harm to processing per se — benefits plausibly reflect other diet/PA components). Tensions the umbrella’s causal framing with the RCT layer (insufficient-evidence position, not harm-reversal). DEFERRED medRxiv 2026 preprint 10.64898/2026.06.03.26354787 (energy-intake SMD 0.71 vs 0.02 when energy-density matched — strongest “energy density not processing” evidence, but PREPRINT, re-check on peer review).
  • coffee -> mortality causality (held Poole 2017 gold umbrella, observational U): SELECTED Nordestgaard 2016 10.1093/ije/dyw325 (same-paper observational-U + genetically-instrumented NULL; Danish CGPS/CCHS; clean independence from Poole). Genuine observational-benefit-vs-MR-null TENSION. DEFERRED Yuan/Larsson 2021 10.3390/nu13072218 (2nd independent MR-null; corroboration, no mortality headline).
  • BMI -> mortality, independent genetic witness (held Global BMI Collab 2016 + Wade 2018; EXCLUDE Davey Smith / Sattar / Timpson / Carslake): SELECTED Sun 2019 10.1136/bmj.l1042 (non-linear MR, Cambridge + Norwegian HUNT; NONE of the excluded authors; low-BMI arm largely a smoking artifact). DEFERRED Burgess 2023 medRxiv preprint (same-lineage refinement, unpublished).
  • apoB vs LDL-C, the flagged residual independence gap (held Marston 2022 + Zuber 2021; EXCLUDE BOTH Ference AND Ala-Korpela): SELECTED Helgadottir 2022 10.1093/eurjpc/zwac219 (deCODE + Danish data; NEITHER excluded author; finds cholesterol content not particle number mediates the risk). Fills the double-exclusion gap AND tensions the held apoB-particle-supremacy -> highest-value find of the survey. DEFERRED Li 2025 medRxiv preprint (ratio not head-to-head; shared-data risk).

Acquisition status (field 4) — pause-and-ask, NO downgrade

[2026-08-20] SELECTION RECORD — erythritol / allulose / polyol sweeteners -> platelet, thrombosis, MI, stroke

Question: is there GOLD-tier evidence (SR/MA/umbrella) that erythritol, allulose, and related rare-sugar/polyol sweeteners raise platelet reactivity / thrombosis / MI / stroke? Channel: openalex (primary), 9 queries. Maintainer request.

Landscape (the honest answer): NO mature gold SR/MA on the specific exposure->hard-outcome yet. The erythritol->thrombosis evidence is HIGH-quality-but-not-gold-by-design (a landmark cohort + human mechanism, not a pooled SR/MA); allulose->CV is an EMPTY literature (insufficient-evidence).

  • Witkowski/Hazen 2023, Nature Medicine 10.1038/s41591-023-02223-9 (cites=221) — the landmark: large discovery+validation cohorts (erythritol -> incident MACE) + human/animal platelet & thrombosis mechanism. Tier ~high (single-group cohort + mechanism, NOT SR/MA). SELECTED (first source on a no-gold question = ratchet-advancer). Fetch BLOCKED (escholarship HTTPError) -> maintainer-drop.
  • Witkowski/Hazen 2024, ATVBAHA 10.1161/atvbaha.124.321019 (cites=22) — prospective human ingestion (erythritol, not glucose) -> enhanced platelet reactivity; the interventional platelet arm the maintainer named. SELECTED (mechanism complement). Fetch BLOCKED (landing-page stub, then PMC 131-byte anti-bot block) -> maintainer-drop (PMC11338701).
  • Artificial Sweeteners and CVD: SR + Meta-analysis 2025, Nutrition Reviews 10.1093/nutrit/nuaf298 (cites=1) — the ONLY gold SR+MA at the umbrella level, BUT BLOCKED 2026-08-20 (maintainer): cannot acquire (OUP paywall, oa=False, no reachable copy). Not downgraded to a weaker substitute (acquisition rule). Scope also uncertain: erythritol is a POLYOL/sugar-alcohol and many “artificial sweetener” MAs cover aspartame/sucralose/saccharin and EXCLUDE polyols, so it might not even answer the erythritol question. Left un-acquired (no claim-page hold planted — scope-uncertain and not erythritol-specific); re-attempt only if an OA copy or a maintainer PDF appears.
  • Circulating erythritol and related polyols and risk of cardiometabolic diseases 2026 (MA) 10.17605/osf.io/xrky8 (cites=0, oa=True but no PDF link) — REJECTED 2026-08-20 (maintainer): the OSF DOI is NOT a paper (registration/OSF artifact, no completed results MA behind it). Do not re-surface as a candidate; the on-question gold MA this question wants remains un-found (a real [G]).
  • allulose -> cardiovascular / thrombosis: 0 hits (allulose cardiovascular=0; only postprandial GLP-1 / gut-microbiota / metabolic studies exist). Attributed to genuine literature-absence (insufficient-evidence), not a title.search artifact -> a real [G]/no-acquire (nothing gold to get).
  • xylitol -> CV: 0 on xylitol cardiovascular thrombosis (title.search strictness — the Hazen 2024 xylitol prothrombotic paper exists under a different title; SAME evidence class as erythritol, not a separate gold). Not separately pursued.
  • 0-hit attributions: non-sugar sweeteners cardiovascular(—type meta-analysis)=0 and non-sugar sweeteners cardiovascular disease=0 are title.search STRICTNESS artifacts (hyphenated multi-word, every word ANDed into title), NOT literature-absence — the SR/MA nuaf298 was found via a looser artificial sweeteners cardiovascular + --type systematic-review.

Acquisition-requested: Nat Med 2023 + ATVBAHA 2024 acquired + ingested (see the erythritol ingest, 2026-08-20). xrky8 REJECTED (not a paper, maintainer); nuaf298 BLOCKED (OUP paywall, un-acquirable). Maintainer-drop list now EMPTY — nothing left to fetch. The broad non-sugar-sweetener/polyol -> CVD gold SR/MA remains un-held: a real [G] gap, not a pending acquisition. Allulose = the same gap.

  • [2026-08-20] openalex fetch “10.2903/j.efsa.2012.2985” (EFSA Mercury) -> FETCH-FAIL (Wiley HTTPError)
  • [2026-08-20] openalex fetch “10.1186/s12916-017-0791-y” (Jacka SMILES) -> inbox/ (770 KB, 10pp PDF OK)
  • [2026-08-20] openalex fetch “10.1016/S1470-2045(07)70373-X” (Straif shift-work) -> PAYWALLED (oa=False)

[2026-08-20] SELECTION RECORD — three maintainer-approved OA acquisitions (Straif · EFSA Hg · Jacka SMILES)

Context. Follow-up to the AWAITS/G acquisition-gap survey: the maintainer approved acquiring the three registry-held [NOT YET HELD] placeholders judged likely open-access. Acquisition only (Research op) — nothing ingested here; each hands off to process inbox when its PDF is in hand.

  1. Query — three direct DOI fetches via openalex.py fetch (primary tool, clean registry DOIs); EFSA escalated down the ladder (openalex fetch -> curl UA -> fetch_browser.py headless -> EFSA europa.eu + epdf mirror), all four hosts tried. Channel: openalex + headless browser. No title-search (DOIs already registered), so no title.search-strictness 0-hits to attribute.
  2. Results —
    • Jacka - SMILES Trial Diet Depression 2017 10.1186/s12916-017-0791-y (BMC Medicine, OA) -> FETCHED to inbox/ (770 KB, valid 10-page PDF v1.4).
    • EFSA - Mercury and Methylmercury in Food 2012 10.2903/j.efsa.2012.2985 (EFSA Journal, oa=True, cites=896) -> BLOCKED. Genuinely OA but every known host anti-bot-blocks: Wiley pdfdirect (openalex HTTPError + browser HTTP 403), Wiley epdf (403), EFSA europa.eu main_documents mirror (403). Wiley is in the measured anti-bot family (Lancet/OUP/JAMA/Cell/SD/NEJM).
    • Straif - Carcinogenicity of Shift Work 2007 10.1016/S1470-2045(07)70373-X (Lancet Oncology commentary, pp.1065-1066) -> PAYWALLED (oa=False, no OA PDF anywhere) — as predicted (Elsevier short statement; the underlying IARC Monograph Vol 98 is OA on the IARC site but is a different document than this Lancet Oncol summary the row points at).
  3. Adjudication — all three SELECTED (maintainer-approved, ratchet-advancers on questions with no held source: circadian/shift-work carcinogenicity; methylmercury dose/benefit-risk in fish; whole-diet RCT -> depression). Symmetric bar met at intake; tiers per registry (Straif high/report, EFSA gold/report, Jacka moderate/rct [downgrade: tiny-n underpowered single RCT]). No downgrade-substitution attempted for the two blocked ones (acquisition pause-and-ask rule).
  4. Acquisition-requested — Jacka -> inbox/ (ready for process inbox). Straif shift-work 2007 10.1016/S1470-2045(07)70373-X -> maintainer-dropped 2026-08-20 to inbox/ (Lancet Oncol Policy Watch, Vol 8 Dec 2007 p.1065, 2-page PDF verified) — so the row stays pointed at the Lancet Oncol summary, not the IARC Monograph Vol 98 (that re-point option is now moot). Maintainer-drop list (1): EFSA Hg 2012 10.2903/j.efsa.2012.2985 — OA but Wiley-anti-bot-blocked (403 on every host); a maintainer browser can save the PDF, or authorize the paid Steel rung. Identified, not lost — DOI above.

[2026-08-21] SELECTION RECORD — protein target for overweight/obese + caloric-deficit adults

Gap (maintainer-raised): a more robust protein-target RANGE for (a) overweight/obese adults and (b) adults in caloric deficit — beyond the held «~1.6-2.2 g/kg, doesn’t transport to obesity». Held on this question: Morton 2018 (healthy trained, per total BM, ~1.62 g/kg RT breakpoint), Tagawa 2020 (independent JP lineage, ~1.3 g/kg), Refalo 2025 (caloric-DEFICIT FFM retention but NON-obese + not independent of Helms), Bauer/PROT-AGE (older adults). So the ratchet wants: a DEFICIT synthesis independent of Refalo/Helms/Phillips, and anything gold on the OBESITY-specific target / which-kg basis.

SELECTED (gold, independent, on-gap) — both OA-flagged but fetch-BLOCKED -> maintainer-drop:

  • Wycherley 2012 10.3945/ajcn.112.044321 — AJCN, gold META-ANALYSIS, 602 cites, OA. Energy-restricted HIGH-protein vs standard-protein diets -> body composition (fat mass, LEAN mass) + cardiometabolic. Clifton/Brinkworth (Adelaide/CSIRO) lineage — INDEPENDENT of Phillips/Morton AND of Refalo/Helms. Directly answers the caloric-deficit stratum with hard-ish body-comp outcomes. Fetch BLOCKED (ScienceDirect 403) -> maintainer-drop.
  • Kim 2016 10.1093/nutrit/nuv065 — Nutrition Reviews, gold META-ANALYSIS, 230 cites, OA. Dietary protein -> body-composition change AFTER weight loss in older adults (lean-mass preservation in deficit). Campbell/Purdue lineage — INDEPENDENT of Wycherley AND of the held protein cluster. Pairs with Wycherley as a potential type-E robustness leg (two independent lineages -> “higher protein preserves lean mass in a deficit”). Fetch BLOCKED (OUP 403) -> maintainer-drop.

DEFERRED (candidates, not acquired now):

  • Nunes 2022 10.1002/jcsm.12922 — J Cachexia Sarcopenia Muscle, gold SR+MA, 262 cites, OA. Broad protein -> muscle mass & function (health + disease), dose-response. BUT Phillips senior author -> NOT type-E independent of held Morton/Devries, and NOT obese/deficit-specific. Candidate as a type-F UPDATE to the RT-target cluster at a later pass, not this gap’s priority.
  • Verreijen 2015 10.1016/j.clnu.2015.02.016 — Clin Nutr, single nitrogen-balance study, non-OA, moderate/weak. DIRECT on the obese-older-adult protein requirement during weight loss, but not gold. Weijs group. Maintainer-drop ONLY if the which-kg-in-obesity question is prioritized (labelled-weak counterweight in a thin field).
  • Calculation of protein requirements (bodyweight vs FFM) 2022 10.1016/j.clnesp.2022.01.014 — Clin Nutr ESPEN, methods-comparison article, OA, moderate. DIRECT on the which-basis (total BW vs FFM) question the maintainer raised. Not a SR/MA. Candidate to acquire as the labelled-moderate direct evidence on the basis question.

FINDING (G-gap, honestly reported): the which-kg-BASIS-in-obesity question (scale protein by total vs lean vs adjusted body weight) has NO gold SR/MA — the direct evidence is single studies (Verreijen) + a methods-comparison paper (Clin Nutr ESPEN 2022). Zero-hit queries Q14/Q15 (“protein adjusted body weight”, “protein intake obesity guideline”) are title.search STRICTNESS artifacts (every word ANDed into the title), NOT literature-absence proof. The absence of a gold SR IS the honest state: the field has not resolved the obesity basis-choice with systematic evidence.

Acquisition-requested: Wycherley 2012 + Kim 2016 -> BLOCKED at headless fetch (both genuinely OA; ScienceDirect/OUP anti-bot 403) -> PAUSE + ASK maintainer to drop the 2 PDFs into inbox/ (DOIs above). Nothing downgraded; nothing ingested. Hand-off to process inbox once the PDFs land.

[2026-08-21] SELECTION RECORD — within-category heterogeneity of processed red meat (attribute decomposition)

Gap (maintainer-raised). The fabric treats “processed meat” as one IARC/WCRF category. It holds category-level associations (Red and Processed Meat and Cancer: WCRF processed-meat RR 1.16 [1.08-1.26] per 50 g/day; the WCRF/NutriRECS clash on Should Adults Reduce Red and Processed Meat) and the abstract within-category-variance argument (Is the Food Category Doing Any Work), but does NOT decompose within processed meat by the attributes that actually vary across bacon/sausage/deli/ prosciutto: N-nitroso compounds (nitrite/nitrate curing), heme iron, HCA/PAH (high-temp cooking + smoking), salt. That decomposition is the decision-relevant object — it is what determines whether a genuinely nitrite-free product is a different exposure (nitroso-driven harm -> reformulation helps) or the same (heme-driven harm -> intrinsic, reformulation futile). Analogous to the UPF/NOVA within-category work already held (Lane/Aramburu/Hall).

Key structural finding — the gold literature does NOT organize by product name. Subtype/product-name title searches (processed-meat-subtypes / -type CRC) returned 0 under --type meta-analysis strictness; the pooled evidence works at the ATTRIBUTE/MECHANISM level, so “differences between bacon vs sausage vs prosciutto” is answered mechanistically (heme / nitroso / HCA content), NOT by product-name MAs. The product names are proxies for attribute bundles (bacon = nitrite-cured + high-temp fried -> nitroso AND HCA; dry-cured prosciutto = low residual nitrite, not high-temp cooked; etc.).

SELECTED (2 gold MAs — both ratchet-ADVANCERS: new-aspect openers on the attribute decomposition the held category-level sources do not touch):

  1. Bastide - Heme Iron Colorectal Cancer Meta-Analysis 2011 (DOI 10.1158/1940-6207.capr-10-0113; gold/meta-analysis; OA; 389 cites) — the INTRINSIC-attribute channel. Heme iron is present in red meat regardless of curing, so this bounds how much nitrite-removal can help. Caveat: 2011 — check for supersession at ingest (no newer heme-CRC MA found; the 2023 OA hits are re-deposits of this same paper). REASON: opens the heme-vs-nitroso mechanistic split; complements #2 as the two competing explanations (a future D-tension / A-synthesis).
  2. Said Abasse - Dietary Nitrate Nitrite Site-Specific Cancer 2022 (corrected on acquisition from the “Seyyedsalehi” pre-fetch placeholder — actual first author Kassim Said Abasse; registered slug/citekey abasse2022nitrate) (DOI 10.3390/nu14030666; gold/meta-analysis; OA; 94 cites) — the CURING/NITROSO channel; separates dietary NITRATE (largely vegetable, often neutral/protective) from NITRITE (the meat-preservative concern) across cancer sites. REASON: directly informs whether nitrite-free reformulation is a real lever — the maintainer’s sharpest hook (“sausages genuinely free of nitrates and nitrites”).

DEFERRED (named gaps, NOT acquired):

  • HCA/PAH channel — only moderate/narrative-tier reviews found (Genes&Environment 2021, 29 cites, OA); the high-temp-cooking carcinogen channel is already partly held (Dietary AGEs and Cooking Method, Filippini acrylamide). Below the gold-gate; a named gap, not a page. Revisit if a gold HCA dose-response MA surfaces.
  • EFSA N-nitrosamines risk assessment (2023 opinion) — the major evidence-bearing report on this, but not cleanly in OpenAlex (only a 2024 SKLM commentary, moderate) and a multi-hundred-page risk assessment (heavy to ingest). Optional maintainer-drop if the two MAs leave the nitroso magnitude under-specified; the 2022 MA is the tractable first source.

GAP DOCUMENTED (expectancy test) — no direct gold OUTCOME evidence on nitrite-free/uncured meat. Searches “nitrite-free processed meat health” and “uncured meat celery nitrate cancer” both returned 0 hits. This is a genuine G-gap, not a tool artifact: a nitrite-free-vs-conventional-meat hard-outcome RCT is not a study anyone can run, and “uncured / no-nitrite-added” products commonly use celery-powder nitrate (bacterially converted to nitrite), so the label may not deliver the exposure it implies. The attribute-level MAs (#1/#2) are how the fabric can reason about this indirectly.

ACQUISITION STATUS — BOTH BLOCKED -> maintainer-drop requested. Plain openalex fetch AND the rung-3 headless browser both returned HTTP 403 (AACR + MDPI anti-bot-walled here despite OA flags). Per acquisition-pause-and-ask: NOT downgraded. Clean DOIs recorded above; awaiting maintainer PDF-drop to inbox/, then -> process inbox -> Ingest (pair the two on the heme-vs-nitroso split).

[2026-08-22] SELECTION RECORD — obesity-as-disease framing candidates (maintainer-offered)

(1) Query / channel. No new search run; maintainer offered four candidate sources during the build of The Disease Model of Obesity, flagged “low GRADE?”. Adjudicated against the gold-gate; no acquisition attempted.

(3) Adjudication + reason (all REJECTED for ingest; DEFERRED as optional framing-layer):

  • Jastreboff et al. 2019 — “Obesity as a Disease: The Obesity Society 2018 Position Statement” — REJECT (ingest). A position statement = consensus/advocacy tier, below the gold-gate. It asserts the disease label; it does not supply systematic evidence. The wiki’s rule is to INDUCE the framing from held mechanistic/interventional evidence, not borrow a body’s advocacy label (received-wisdom-prior risk in reverse). Legitimate ONLY as a later guidance-null / framing-layer source, held explicitly as recommendation/consensus tier used for FRAMING not effect — a separate deliberate acquisition with a documented tier override, not needed for this page.
  • Allison et al. 2008 — “Obesity as a Disease: A White Paper on Evidence and Arguments” — REJECT. A white paper = opinion/advocacy (weak). Same reason; older, and superseded by the held RCT pharmacology as the actual evidentiary base.
  • Vallis et al. 2017 — “Philosophical determinants of obesity as a disease” — REJECT. Philosophy of the disease definition, not a falsifiable/quantifiable claim on a patient-important outcome — outside the evidence scope (context/method at most). The disease label is framing-grade by the page’s own Evidence-status section; this source argues the label, does not evidence the mechanism.
  • Rosenbaum & Leibel 2016 — “Models of energy homeostasis in response to maintenance of reduced body weight” — DEFER (the strongest of the four). A narrative review (moderate) by the reduced-weight-state / leptin authorities; it is the mechanistic backbone under Sumithran/Hall. BUT it CORROBORATES what the held sources already establish (type-E/F, not new), and the page is well-grounded without it. Not a gold need. Optional later acquisition to strengthen the metabolic-defense arm — pause-and-ask, don’t downgrade.

(4) Acquisition-requested. NONE. Page built on held higher-grade sources (Sumithran 2011 cohort, Hall 2022 review, Wilding STEP-1 RCT, Jastreboff SURMOUNT-1 RCT). The four candidates are the FRAMING layer, held apart from the EVIDENCE layer per the page’s Evidence-status section.

[2026-08-22] SELECTION RECORD — belief-harvest run 6 (depth + named-figure heterodox lane)

Channel: WebSearch (discovery of widely-held beliefs, NOT acquisition). Hand-logged (no auto-log for web channel). Run 6 = depth within categories after the runs 1-4 wide-dispersion pass; ~1/5 heterodox (low-carb pole this run — R17 cap; plant-based pole owed next run).

(1) Queries. (a) intermittent fasting 16:8 metabolic benefits beyond calorie restriction 2026; (b) carbohydrate insulin model Bikman Ludwig "insulin not calories" fat storage claim; (c) blue light screens before bed ruin sleep evidence; (d) red and processed meat causes cancer public belief; (e) omega-3 fish oil supplements protect heart belief.

(2) Results (belief confirmed widely-held + representative source). (a) MedicalNewsToday / NMCD-journal 2024 isocaloric-IF-vs-CR SR-MA (IF not superior to CR); (b) AJCN Ludwig 2021 CIM + Wikipedia CIM (Taubes/Ludwig/Bikman proponents; contested); (c) Healthline / ClinicBarcelona / SleepFoundation (blue-light concern “overstated”, ~10-min onset delay, arousal matters more); (d) WHO/IARC/ACS (processed=Group1, red=Group2A colorectal); (e) Cochrane / UTSW (fish-oil supplements no CV benefit). No zero-hit queries.

(3) Adjudication (which beliefs entered the run + why). SELECTED (a),(b),(c),(d) into run 6. (b) is the heterodox/named-figure low-carb-pole slot (Bikman/Ludwig CIM), adjudicated as a CLAIM not a persona. DEDUP-SKIP (e) omega-3-fish-oil->heart: already adjudicated in the challenge ledger (row 1196, deliverable-critique #11 — general-population null held; high-dose EPA REDUCE-IT contested, queued) -> true ledger skip (anti-flip-flop), REPLACED by (f) spot reduction (targeted fat loss — exercise- programming depth, 0 fabric/0 ledger coverage). Final run-6 five: IF-beyond-CR, CIM-insulin-not-calories, blue-light-sleep, red/processed-meat-cancer, spot-reduction.

(4) Acquisition-requested. NONE fetched — a belief-harvest routes GAPs as acquire-targets, not fetches. Two GAP rows registered in ingest-queue.md ## BELIEF-HARVEST GAPS: [belief-harvest: sleep-light-screens] (SR/MA on evening screen/blue-light exposure -> sleep onset/quality) and [belief-harvest: spot-reduction] (SR/MA/RCT on localized resistance training -> regional adiposity). Both are named gaps, unregistered to a specific source (no [AWAITS] handle yet). The 2024 NMCD isocaloric SR-MA was NOT acquired: Semnani-Azad 2025 (99-RCT NMA, held) already covers the question more comprehensively -> no thin-spot.

[2026-08-22] SELECTION RECORD — belief-harvest run 7 (depth + plant-based heterodox pole)

Channel: WebSearch (discovery of widely-held beliefs, NOT acquisition). Hand-logged (no auto-log for web channel). Run 7 = depth within under-sampled categories; ~1/5 heterodox = the PLANT-BASED pole (owed from run 6’s low-carb pole — R17 requires both poles sampled).

(1) Queries. (a) whole food plant-based diet reverses heart disease Ornish Esselstyn claim; (b) is moderate drinking red wine good for your heart belief 2026; (c) is coffee bad for your heart common belief; (d) does fasted cardio burn more fat; (e) magnesium supplement for better sleep evidence.

(2) Results (belief confirmed widely-held + representative source). (a) PLANTSTRONG / NutritionFacts (Esselstyn) / T.Colin Campbell Center / PCRM — WFPB “reverses” CAD (Esselstyn 8/11 regression case series; Ornish 1990 Lifestyle Heart Trial); (b) Mayo Clinic / MassGeneralBrigham / Johns Hopkins / a 2026 cohort (21% lower CV death in moderate wine drinkers) vs “don’t start drinking for your heart”; (c) AHA 2026 / GoodRx / Heart&Stroke — “coffee gets a bad rap”; 2-4 cups linked to LOWER CHD/HF/stroke; (d) GoodRx / UNSW / UCLA Health — fasted burns more fat DURING session, no long-term body-fat advantage (2017 SR), more muscle protein breakdown; (e) NCCIH / Cleveland Clinic / SleepFoundation + a 2021 SR-MA (sleep-onset -17.36 min, LOW quality). No zero-hit queries.

(3) Adjudication (which beliefs entered + why). SELECTED all five. (a) is the heterodox/named-figure PLANT-BASED-pole slot (Ornish/Esselstyn/Greger), adjudicated as a CLAIM not a persona/movement. Dedup: alcohol + coffee are heavily HELD in the fabric (Alcohol and Mortality and Vascular Disease, Coffee Consumption and Health) but have NO prior belief-harvest ledger row -> per step 3 a fabric hit is a CONFIRM/REFINE/REFUTE candidate to RECORD, not a true skip. No target ratio curated.

(4) Acquisition-requested. NONE fetched — GAPs routed as acquire-targets, not fetches. THREE queue rows registered in ingest-queue.md ## BELIEF-HARVEST GAPS: [belief-harvest: wfpb-cvd-regression] (the Ornish Lifestyle Heart Trial + Esselstyn + any WFPB angiographic-regression SR — the decision-critical “reversal” evidence the fabric does not hold), [belief-harvest: fasted-cardio-fat-loss] (thin-spot: a direct fasted-vs-fed exercise body-composition SR — the net-loss inference is already bounded by the held energy-balance spine), [belief-harvest: magnesium-sleep] (a magnesium -> sleep-onset/quality SR/MA; public evidence exists but thin/low-quality). All named gaps, no [AWAITS] handle yet.

[2026-08-22] SELECTION RECORD — belief-harvest run 8 (depth: diet/macro/single-food + marketing-heterodox)

Channel: WebSearch (discovery of widely-held beliefs, NOT acquisition). Hand-logged. Run 8 = depth across under-sampled categories (breakfast, dietary-fat, single-food dairy, sugar-behaviour) + a marketing- heterodox slot (collagen). Both R17 named-figure poles already sampled (low-carb run 6, plant-based run 7).

(1) Queries. (a) is breakfast the most important meal of the day belief skipping breakfast weight gain; (b) does eating fat make you fat low-fat diet belief; (c) is sitting the new smoking sedentary behaviour independent risk; (d) is sugar addictive like a drug belief evidence; (e) do collagen supplements improve skin joints hair evidence; (f) is dairy inflammatory bad for you belief milk.

(2) Results (belief confirmed widely-held + representative source). (a) Harvard Health / Cleveland Clinic — “skipping causes weight gain exceeds the scientific evidence”, RCTs null; (b) Harvard Chan / Examine — “fat doesn’t make you fat, excess calories do” (search also surfaced the low-carb rebuttal — already refuted run 6); (c) NewsMedical / TheConversation — “sitting is the new smoking” (DEDUP-SKIP, see 3); (d) NCBI/DSM-5 reviews — contested, sugar reward != clinical addiction, rat bingeing only under intermittent access; (e) Harvard Health / NatGeo — early evidence (skin hydration/elasticity, OA pain), hair thin, “hype ahead of science”; (f) IFIC / Arthritis Foundation — not inflammatory for most healthy people (2017 review of 52 studies: some anti-inflammatory; yogurt best). No zero-hit queries.

(3) Adjudication (which beliefs entered + why). SELECTED (a),(b),(d),(e),(f). DEDUP-SKIP (c) sitting- is-the-new-smoking: a COMPLETED ledger analysis already covers it (maintainer challenge 2026-08-18 — sitting IS a distinct near-independent risk, magnitude calibrated against the “new smoking” overstatement)

  • held page Sedentary Behaviour and Chronic Disease Risk -> true skip (anti-flip-flop), REPLACED by (f) dairy-is-inflammatory/bad (Dairy and Cardiometabolic Health hit -> a REFUTE candidate to record). No target ratio curated. Selection-phase note: runs 6-8 (DEPTH) target contested/heterodox/marketing beliefs, so they skew refute/gap; runs 1-5 (DISPERSION) hit more mainstream-confirmed beliefs (several CONFIRMs) — a property of the SELECTION phase, transparently flagged, NOT a fabric-quality claim.

(4) Acquisition-requested. NONE fetched. TWO queue rows: [belief-harvest: sugar-addiction] (a sugar/sweet-taste -> addiction-criteria / neuro-reward review — the fabric holds sugar’s dental+SSB harms but no addiction position) and [belief-harvest: collagen-supplements] (a hydrolysed-collagen -> skin elasticity/hydration + OA joint-pain SR/MA; emerging, low-quality). Named gaps, no [AWAITS] handle yet.

[2026-08-22] SELECTION RECORD — belief-harvest run 9 (DISPERSION — under-sampled categories)

Run 9 (belief-harvest). Category spread targeted the R17-flagged under-sampled lanes: hormonal/endocrine, drugs-as-alternative, exercise-programming, plus a popular-diet and a sleep belief. No named-figure heterodox slot this run (both poles sampled runs 6-7; the ~1/5 lane is a cap, not a quota).

(1) Queries. (a) menopause HRT dangerous causes breast cancer belief women avoiding hormone therapy; (b) Ozempic Wegovy you regain all the weight when you stop dangerous side effects belief; (c) you must lift heavy train to failure to build muscle bodybuilding belief; (d) keto diet burns more fat metabolic advantage than other diets belief low carb; (e) some people only need five hours of sleep thrive short sleeper genetic belief. Five swapped-out candidates were dedup-caught BEFORE search (see 3).

(2) Results (belief confirmed widely-held + representative source). (a) Maven/BCRF/CancerResearchUK — HRT-breast-cancer fear widespread; FDA removed the breast-cancer black-box warning early 2026 as over-stated; WHI tested one synthetic formula in older women, rarely prescribed today; progestin > body- identical progesterone for risk. (b) ScientificAmerican/Sutter/HenryFord — “regain all the weight on stopping” widely held; STEP-1 extension regained ~two-thirds, but a 2026 real-world study found ~0.5% regain (many restart/transition); drugs designed as lifelong; robust harm is GI. (c) StrengthLog/Jefit/ Equinox — “train to failure / lift heavy or no gains” is the gym-culture default; failure NOT required, close-to-failure suffices, failure’s edge is load-specific (<60% 1RM). (d) IncheckFit/Healthline — “keto burns more fat / metabolic advantage” is a keto-evangelist staple; Hall metabolic-ward isocaloric refutes it (EE +57 kcal/d but fat-loss RATE slowed); real-world edge is modest + satiety-mediated. (e) UCSF/CNN/ DocParsley — “some people thrive on 5h” widespread; DEC2/ADRB1/NPSR1 true short sleepers are REAL but “exceptionally rare… almost never the reason you are running on five hours.” No zero-hit queries.

(3) Adjudication (which beliefs entered + why). SELECTED (a)-(e), all clean-novel vs the ledger + fabric + Processed list. DEDUP-SWAPPED-OUT (caught pre-search, TRUE ledger/Processed skips, anti-flip- flop): sauna (row 1178, GAP), multivitamin (row 1163, REFUTE), cold-plunge (row 1179, GAP), weekend catch-up sleep (Processed, run 3), TRT/testosterone (adjacent to row 1172 + HRT belief 1 this run) — each replaced by a clean-novel belief. Belief 2 (GLP-1) is a DISTINCT sub-claim of the earlier GLP-1 adjudication (row 1175, “GLP-1 is cheating”): row 1175 was the moral/agency framing; belief 2 is the regain-on-stopping + danger framing, backed by GLP-1 Non-Cardiometabolic Effects and Safety (Yang 2026 umbrella, ingested since) + the Wilding-withdrawal fabric — legitimately re-testable, distinction noted. No target ratio curated.

(4) Acquisition-requested. NONE fetched. ONE queue row: [belief-harvest: short-sleeper-genetics] (a DEC2/ADRB1/NPSR1 natural-short-sleeper phenotype source — the fabric holds the population sleep- duration dose-response but takes no position on the rare genetic-short-sleeper subgroup, a route-(b) effect-modification gap). Named gap, no [AWAITS] handle yet.

[2026-08-22] SELECTION RECORD — belief-harvest run 10 (DISPERSION — 5 novel categories)

Run 10 (belief-harvest). Category spread targeted five distinct lanes (diet-pattern · exercise- programming · sleep · supplement · metabolism) to avoid re-fishing the now-crowded macro/GLP-1 pools. No named-figure heterodox slot this run (the ~1/5 lane is a CAP not a quota; spent runs 6-9). The run is GAP-heavy by honest selection — after the mainstream-confirmable beliefs deduped out as already-adjudicated, the novel-vs-fabric residue is dominated by unheld corners. Ratio DESCRIPTIVE, never curated: 0 CONFIRM : 1 REFINE : 0 REFUTE : 4 GAP : 0 OUT.

(1) Queries. (a) gluten free healthier without celiac disease belief avoid gluten; (b) stretching before exercise prevents injury warm up belief; (c) alcohol nightcap helps you sleep belief drink before bed; (d) apple cider vinegar lowers blood sugar weight loss belief; (e) small frequent meals grazing boost metabolism eat more often belief. Two candidates were dedup-caught BEFORE search (see 3).

(2) Results (belief confirmed widely-held + representative source). (a) Healthline/BeyondCeliac/ Harvard — gluten-free-is-healthier is widespread (~20-30% have tried it undiagnosed); public evidence (UNHELD here) is no non-celiac benefit + a fibre/micronutrient downside of the diet. (b) Mayo/NASM/ Healthline — pre-exercise static-stretch-prevents-injury is the default coaching belief; the SR signal (UNHELD) is that static stretching does not reduce injury and may transiently cut power, dynamic warm-up favoured. (c) SleepFoundation/Cleveland/Healthline — the nightcap belief is common; alcohol speeds onset but suppresses REM/deep sleep and fragments the back half (UNHELD, not asserted). (d) Healthline/ MedicalNewsToday — ACV-lowers-blood-sugar/aids-weight is a wellness staple; evidence is small-n/short- horizon, a modest T2D fasting-glucose signal + weak short-lived appetite effect (UNHELD -> INSUFFICIENT, not “works”). (e) Healthline/BritishHeartFoundation/Examine — grazing-boosts-metabolism is a persistent diet-culture belief; the HELD isocaloric principle (TRE 12-month null, Liu n=139) adjudicates it — meal pattern is not a weight-independent metabolic lever. No zero-hit queries.

(3) Adjudication (which beliefs entered + why). SELECTED (a)-(e), all clean-novel vs the ledger + fabric + Processed list. DEDUP-SWAPPED-OUT (caught pre-search, TRUE ledger/Processed skips, anti-flip- flop): salt-raises-blood-pressure (the fabric holds the WHO sodium->BP dose-response — a mainstream CONFIRM already covered, not novel) and 5-a-day / fruit-and-veg (ESC fruit/veg->mortality is held, the plateau exemplar) — each swapped for a clean-novel belief. NOTE the first dedup grep used grep -iE with \| (a literal backslash-pipe under -E, false-zero on all 5 including salt/BP which the fabric plainly holds); re-run with bare | alternation + positive control (4 known BP pages surfaced) exposed the two true skips. No target ratio curated — the 4:1 GAP:REFINE split is what honest novel selection returned, not a tuned figure.

(4) Acquisition-requested. NONE fetched (all four GAPs are named acquire-targets, gold-gated, routed not pulled). FOUR queue rows in ingest-queue.md BELIEF-HARVEST GAPS: gluten-nonceliac, stretching-injury-prevention, alcohol-sleep-quality, acv-glycemia-weight — each a public belief the fabric cannot yet adjudicate. ONE weave signal (WS-005, grazing meal-frequency refinement UNSURFACED -> pending-intersections.md). No [AWAITS] handles yet (no candidate source registered).

[2026-08-22] SELECTION RECORD — belief-harvest run 11 (DISPERSION — under-covered categories)

Run 11 (belief-harvest). Channel = WebSearch (no auto-log; hand-logged). Category spread targeted five under-covered lanes (supplement · single-food/marketing · gut · exercise-programming · metabolism-PA), avoiding the crowded macro/GLP-1/sleep pools. No named-figure heterodox slot (the ~1/5 lane is a CAP not a quota; spent runs 6-9). REFINE-heavy by honest selection — this run’s novel beliefs happened to land on WELL-HELD fabric (4 of 5 have a claim page), the honest contrast to run 10’s GAP-heavy thin-area batch. Ratio DESCRIPTIVE, never curated: 1 CONFIRM : 3 REFINE : 0 REFUTE : 1 GAP : 0 OUT.

(1) Queries. (a) everyone should take vitamin D supplement prevent disease belief; (b) organic food healthier more nutritious than conventional belief; (c) probiotic supplement improves gut health general health belief; (d) muscle soreness DOMS caused by lactic acid belief no pain no gain; (e) you can't outrun a bad diet exercise alone weight loss belief.

(2) Results (belief confirmed widely-held + representative source). (a) TIME/Mayo Clinic Proceedings/ Columbia/Endocrine Society 2024 — universal vitamin-D-for-disease-prevention is a default wellness belief; the deficiency-vs-repletion distinction is decisive (held: VITAL null in the replete). (b) Stanford ~250-study review/Mayo/Baranski/NutritionFacts — organic-is-more-nutritious is widespread; measured nutrient differences small and of unproven outcome relevance, pesticide-residue the real difference (held: Organic vs Conventional Food). (c) Harvard Nutrition Source/AGA/PMC RCT SR-MA — probiotics-for- general-gut-health is a supplement-aisle staple; benefit is strain-and-indication-specific (held: Gut Microbiome and Health, WGO 2023). (d) Pfizer/Boston Sports Med/kingsportstraining — DOMS-is- lactic-acid + soreness-means-good-workout is default gym lore; lactic-acid theory long refuted (UNHELD by the fabric — a genuine gap). (e) AHA/BJSM editorial/Healthline — can’t-outrun-a-bad-diet is a widely repeated (and correct) slogan; held with better warrant (compensation mechanism, Exercise Energy Compensation). No zero-hit queries.

(3) Adjudication (which beliefs entered + why). SELECTED (a)-(e), all clean-novel vs the ledger + fabric + Processed list. Dedup found FABRIC hits for (a)(b)(c)(e) pre-adjudication — these are CONFIRM/ REFINE candidates, NOT silent skips (only a ledger/Processed hit is an anti-flip-flop skip). The one adjacent ledger row on organic (deliverable-critique #29) is a welfare-axis labelling point, not the core health claim, so the core belief is freshly adjudicated. No belief swapped to manufacture a ratio — the 1:3:0:1:0 split is what honest novel selection on under-covered categories returned.

(4) Acquisition-requested. NONE fetched (the one GAP is a named acquire-target, gold-gated, routed not pulled). ONE queue row in ingest-queue.md BELIEF-HARVEST GAPS: doms-soreness-marker (DOMS-etiology

  • soreness-as-adaptation-marker SR — the fabric holds no exercise-soreness position). ONE weave signal (WS-006, outrun-a-bad-diet findability — add lay-searchable aliases to Exercise Energy Compensation -> pending-intersections.md). No [AWAITS] handles yet (no candidate source registered).

[2026-08-22] SELECTION RECORD — belief-harvest Tier-A acquisition (6 gold targets)

Acquisition episode (NOT ingest). Maintainer reviewed the 24 open belief-harvest GAP acquire-targets and authorized fetching Tier A (6): the fear-unblocks-a-big-rock set + cheap canonical gold. Channel = openalex.py (queries auto-logged above this block). All 6 gold-tier, DOIs vetted below.

(3) Adjudication — SELECTED (6), each the canonical gold source for its gap:

  • statin-nocebo -> SAMSON 10.1056/nejmc2031173 (Wood, NEJM 2020; n-of-1 RCT, the landmark statin- nocebo trial). Unblocks LDL-lowering (big rock). Feeds WS-004 + the held USPSTF cognition statement.
  • protein-renal -> Devries 10.1093/jn/nxy197 (J Nutr 2018; SR-MA, kidney function does not differ by protein level in healthy adults). Unblocks adequate protein (sarcopenia/function).
  • rt-women-hypertrophy -> Roberts 10.1519/jsc.0000000000003521 (JSCR 2020; SR-MA, sex differences in RT — women gain similar RELATIVE but less ABSOLUTE size). Unblocks resistance training. OA fallbacks noted: 10.1007/s40279-020-01388-4 (Sports Med 2020 MA) + 10.7717/peerj.19042 (PeerJ 2025 MA, OA PDF).
  • vitamin-c-colds -> Hemila Cochrane 10.1002/14651858.cd000980.pub4 (2013; the definitive prophylaxis-vs-treatment / regular-vs-therapeutic split). Cheap canonical null.
  • omega6-inflammation -> Su 10.1039/c7fo00433h (Food Funct 2017; SR-MA, dietary linoleic acid -> blood inflammatory markers). NB: the CLA hits (10.1038/s41430-017-0048-z etc.) are a DIFFERENT compound (conjugated LA) — rejected, not the seed-oil question.
  • creatine-safety -> Kreider/ISSN 10.1186/s12970-017-0173-z (JISSN 2017 position stand; cites=828). The rejected consumer-explainer (inbox/_rejected/) was a weak echo of THIS; this is the primary.

DEFERRED at this gate (not fetched): Tier B (wfpb-regression, sauna-Laukkanen, sleep-light, alcohol-sleep, fasted-cardio; anabolic-window + stretching-injury already in the pipeline — Lauersen 2013 is staged) and Tier C (fad/small-effect nulls — acv, collagen, cold-plunge, magnesium-sleep, sugar-addiction, hydration, gluten, spot-reduction, catch-up-sleep, short-sleeper-genetics, doms). Attention-anti-signal held the line.

(4) Acquisition-requested — ALL 6 FETCH-BLOCKED, routed to maintainer (pause-and-ask, NOT downgraded):

  • openalex.py fetch: SAMSON = FETCH-FAIL (NEJM HTTPError, blocks automated pdfdirect); the other 5 = PAYWALLED (no OA PDF resolved by OpenAlex, incl. JISSN/Cochrane which are nominally OA).
  • Local Playwright (fetch_browser.py) tried on the clearest OA one (Kreider JISSN BMC counter/pdf) -> returned a 3 KB HTML block page (IP-wall), not the PDF. Deleted; nothing weaker substituted.
  • Next rung = Steel (cloud browser for hard IP-walls) OR maintainer-drop to inbox/. Awaiting maintainer decision on channel — no ingest until a real PDF lands. DOIs above are clean + re-fetch-cheap.
  • [2026-08-22] openalex fetch “10.7717/peerj.19042” -> 0 hits: (none)
  • [2026-08-22] openalex resolve “10.1007/s40279-020-01388-4” -> 1 hits: https://doi.org/10.1007/s40279-020-01388-4

UPDATE 2026-08-22 (Steel escalation attempted, failed — all channels exhausted). Maintainer chose “Steel OA + drop paywalls”. Steel connected (sessions released) but fetch_browser.py --service steel --download errored inside the in-page fetch->base64 step on every direct-PDF URL tried (Kreider JISSN BMC counter/pdf; Kreider PMC5469049 pdf) — a fetcher limitation on binary responses, NOT a paywall. Net: all three automated channels (openalex OA-resolve, local Playwright, Steel) cannot retrieve a PDF in this environment. ALL 6 -> maintainer-drop (DOIs above). Hand off to process inbox once dropped; nothing downgraded, nothing ingested. Re-fetch is cheap — the DOIs are clean and the adjudication stands.

[2026-08-22] PROCESS-INBOX INTAKE RECORD (maintainer-drop landed)

The maintainer manually dropped 6 PDFs into inbox/ after all automated fetch channels failed (the 2026-08-22 Tier-A acquisition + a couple of extras). Mechanical intake — nothing ingested.

(2) Results — 6 PDFs identified (all clean text, no OCR needed):

  • NEJMc2031173.pdf -> Wood - SAMSON Statin Nocebo 2020 (10.1056/NEJMc2031173).
  • linoleic-acid/inflammation -> Su - Linoleic Acid Inflammatory Markers 2017 (10.1039/c7fo00433h).
  • ISSN creatine position stand -> Kreider - Creatine Safety and Efficacy ISSN Position Stand 2017 (10.1186/s12970-017-0173-z).
  • Robertsetal.2020.Sex-Differences-Meta.pdf -> Roberts - Sex Differences Resistance Training Meta-Analysis 2020 (JSCR; DOI TBD).
  • document2.pdf -> Douglas - Vitamin C Common Cold 2005 (PLoS Med; 10.1371/journal.pmed.0020168).
  • nxy197.pdf -> Devries - Higher Protein Kidney Function Meta-Analysis 2018 (10.1093/jn/nxy197).

(3) Adjudication:

  • STAGED (5) to raw/binary/papers/, registry rows added (Su/Wood/Roberts/Douglas; Kreider row pre-existed):
    • Su (gold MA) -> serves [belief-harvest: omega6-inflammation]; RCT-MA upgrade of held Johnson 2012 SR (not an echo).
    • Wood/SAMSON (high, rct) -> serves [belief-harvest: statin-nocebo] + WS-004 (nocebo half).
    • Kreider (moderate, consensus — sponsor COI, needs tier-override at ingest) -> serves [belief-harvest: creatine-safety].
    • Roberts (high MA) -> candidate for [belief-harvest: rt-women-hypertrophy]; peripheral/exercise-programming, route-(b).
    • Douglas (moderate, PLoS Med summary of the Cochrane) -> PARTIALLY serves [belief-harvest: vitamin-c-colds]; gold Hemila Cochrane still the upgrade target.
  • DUPLICATE (1) -> inbox/_duplicates/: Devries protein-kidney MA already ingested (in done/papers/, source page + Protein Intake and Kidney Function exist).

(4) Acquisition-requested: none — all already fetched (maintainer-drop). The Tier-A maintainer-drop loop is now CLOSED (SAMSON/Su/Kreider landed). Ingest is downstream (plan-queue -> ingest-seq); nothing ingested here.

  • [2026-08-22] openalex search “Vitamin C for preventing and treating the common cold systematic review” -> 0 hits: (none)
  • [2026-08-22] openalex resolve “10.1002/14651858.CD000980.pub4” -> 1 hits: https://doi.org/10.1002/14651858.cd000980.pub4

[2026-08-22] PROCESS-INBOX INTAKE (2nd drop — vitamin-C Cochrane SR)

Maintainer dropped CD000980.pdf after the DOI capture. It is the 2004 Douglas/Chalker/Treacy EDITION of the Cochrane vitamin-C/common-cold review (base DOI 10.1002/14651858.CD000980), NOT the 2013 Hemila & Chalker .pub4 (10.1002/14651858.CD000980.pub4) whose DOI I resolved. Still the gold primary SR (37pp, clean text, 2 chunks) — a real upgrade over the PLoS Med summary.

VERIFY 2026-08-22 (Hemila .pub4 double-check): maintainer could not find the 2013 Cochrane. Confirmed existent across 3 sources (OpenAlex work, NCBI idconv, PubMed esummary): Hemila H & Chalker E, “Vitamin C for preventing and treating the common cold”, Cochrane Database Syst Rev 2013 Jan 31; DOI 10.1002/14651858.CD000980.pub4; PMID 23440782; PMCID PMC8078152. Wiley/Cochrane DOI page is PAYWALLED; free green-OA full text is PMC8078152. DOI given earlier was correct.

PROCESS-INBOX 2026-08-22 (3rd drop — the .pub4 arrived): maintainer fetched CD000980-2.pdf = the CURRENT Hemila & Chalker 2013 Cochrane (.pub4, 95pp, clean text, 4 chunks). Staged as Hemila - Vitamin C Common Cold Cochrane 2013 (gold). The 2004 Douglas edition -> moved to inbox/_hold/ (untracked, retained, SUPERSEDED); the Douglas 2005 PLoS summary stays the secondary echo. One vitamin-C page owed from the Hemila 2013 SR; acquire-target satisfied.

[2026-08-25] SELECTION RECORD — 3(a) telos-coverage acquisition (post-Gate-28 lighthouse review)

Triggered by the lighthouse/ratchet/gap review: acquire the genuine NOT-YET-HELD gold/high landmarks. Scope-tightened against ground truth FIRST — several report-listed targets were already held or superseded and were NOT searched:

  • Gardner DIETFITS 2018 — already HELD (done/papers/); dropped.
  • Goldberg DPPOS 2022 — already HELD (done/papers/); dropped.
  • Gijsbers dairy->T2D dose-response 2016 — REJECTED as echo: the held Mishali 2019 (dairy->T2D+CV MA, newer) supersedes it; a newer MA supersedes an older on the same question.
  • EFSA methylmercury 2012 (efsacontam2012 citekey) — REJECTED: it is the mercury/fish report, NOT acrylamide; the fish/methylmercury route-c is already held (Jayedi + EFSA-Seafood), and the acrylamide axis is held (Filippini, just freshened). Off-question echo.
  • EASD/DNSG 2023 — already STAGED (guidelines/, + a dup in inbox/_duplicates/); no acquire.

SELECTED (genuine gap; gold/high; ratchet-advancer):

  • Tuomilehto Finnish DPS 2001 10.1056/NEJM200105033441801 (NEJM) — independent lifestyle->T2D-prevention replication of the held Knowler DPP 2002 (58% RRR); a genuine type-E independent-backing leg for the diabetes-prevention-by-lifestyle claim. oa=True flag but NEJM anti-bot-blocks -> FETCH-FAIL, maintainer-drop.
  • Pan Da Qing 1997 10.2337/diacare.20.4.537 (Diabetes Care) — second independent (Chinese IGT) diet/exercise prevention replication; type-E. oa=False -> maintainer-drop.
  • Hamman DPP weight-loss mediation 2006 10.2337/dc06-0560 (Diabetes Care) — per-kg weight-loss mediation of DPP prevention (dose-response mechanism, F-refinement of Knowler). Landing OA-flag optimistic; PAYWALLED no OA pdf -> maintainer-drop.
  • Blanco Mejia soy-protein->serum lipids 2019 10.1093/jn/nxz020 (J Nutr; 46-study FDA-identified MA; 144 cites) — soy->LDL/TC, a DISTINCT question from the held soy->hormones/breast-cancer row; surrogate-only (LDL), so bounded marginal value but a clean gold. oa=True flag but OUP anti-bot-blocks -> FETCH-FAIL, maintainer-drop.
  • SCORE2-Diabetes 2023 10.1093/eurheartj/ehad260 (Eur Heart J) — the diabetes-stratum CV-risk instrument (base SCORE2 excludes diabetes); staged only as .xml in papers/, need the PDF. oa=True flag but OUP anti-bot-blocks -> FETCH-FAIL, maintainer-drop for the PDF.

Drop-pending trio carried from the 2026-08-19 survey (already adjudicated; pure fetch attempt): all four resolve gold/high meta-analyses, all oa=False (Elsevier ScienceDirect + RSC, which 403-block):

  • carb-U spline: Clinical Nutrition 2022 10.1016/j.clnu.2022.12.010 -> maintainer-drop.
  • weight-regain trajectory: Machado 2022 10.1016/j.clnesp.2022.03.020 -> maintainer-drop.
  • plant/animal LCD->mortality: Ghorbani 2023 10.1016/j.arr.2023.101997 -> maintainer-drop.
  • LCD->CVD: Qin&Bo 2023 10.1039/d3fo01374j -> maintainer-drop.

Acquisition-requested / status: 0 fetched to inbox/ (every target paywalled or anti-bot-blocked); 9 DOIs recorded for maintainer-drop (per the pause-and-ask rule — NO downgrade to a weaker fetchable source). Registry rows already exist (with DOIs) for the DPP-family + Blanco + SCORE2-Diabetes; the trio are logged here. Nothing ingested. Hand-off: maintainer drops the 9 PDFs into inbox/, then process inbox stages them.

[2026-08-26] SELECTION RECORD — discovery-generator INDUCTIVE-IMPORT (GBD + landscape umbrellas)

Triggered by the discovery-method work: the four-generator rotation (deductive taxonomy · inductive umbrella-import · discourse harvest · structural scan). The structural scan (same day) surfaced the also-rans already in the fabric but is BLIND to zero-toehold areas (massage/sauna/cold/light — 0 fabric hits); this record is the INDUCTIVE import that reaches them, plus the master cross-domain enumeration.

SELECTED:

  • GBD 2019 Risk Factors 10.1016/S0140-6736(20)30752-2 (Lancet 2020;396:1223; 9825 cites) — the master cross-domain hierarchy: 87 risk factors ranked by attributable DALYs (behavioural + metabolic
    • environmental). The single highest-leverage enumeration-ordered-BY-Layer-1 (big rocks first) source the fabric can hold; the held Afshin GBD is DIET-risks only, so this is the whole-telos superset, not an echo. oa=True flag but Lancet anti-bot-blocks -> PAYWALLED, maintainer-drop. DOI recorded.
  • Mind-body exercise -> cognition umbrella 10.1186/s11556-023-00325-4 (Eur Rev Aging Phys Act 2023; cites=40; umbrella) — enumerates AND head-to-head ranks the mind-body-movement modalities the structural scan surfaced (tai chi / yoga / qigong vs aerobic + resistance) on a patient-important outcome (cognition). Closes the tai-chi/yoga/qigong cluster (taxonomy axis 2L) with a gold enumeration. oa=True (BMC) -> FETCHED to inbox/; ready for process inbox.
  • Massage therapy on function in pain populations SR+MA 10.1093/pm/pnw099 (Pain Medicine 2016; cites=180; SR+MA) — reaches the #1 confirmed ZERO-TOEHOLD area (massage: 0 fabric hits, the flagship unknown-unknown the structural net cannot see). On-telos: massage -> function/pain (patient-important QoL outcome). oa=True flag but OUP anti-bot-blocks -> FETCH-FAIL, maintainer-drop. DOI recorded.

DEFERRED:

  • Sauna -> CV/mortality — landmark is Laukkanen 10.1001/jamainternmed.2014.8187 (JAMA IM 2015, paywalled) and the OA hits (10.1186/s12916-018-1198-0 etc.) are the SAME single Finnish KIHD cohort lineage -> volume-is-not-independence bites; NO gold SR/MA surfaced. Recorded as a “not yet” / insufficient area (a gold sauna SR may not exist); do NOT acquire a single cohort as if gold. Revisit if an independent SR/MA appears.

REJECTED (0-hit or no gold enumeration):

  • Complementary/integrative umbrellas (s41562-025-02256-9 autism, 1516-4446-2024-3705 depression, hsr2.2118 male-infertility) — all CONDITION-SPECIFIC, not a broad enumeration of the peripheral space; off-question for an import. REJECTED.
  • Lifestyle-medicine “pillars” hits (jabs.v16i2.20876, gengpi.5.0824.2215, two zenodo datasets) — all narrative / dataset / weak; no gold enumeration. REJECTED.
  • The two mind-body interventions health outcomes umbrella review / initial complementary integrative queries returned 0 — title.search-strictness artifacts (every word must hit the title), not literature absence; re-run looser recovered the gold hits above.
  • [2026-08-26] openalex resolve “10.1080/00031305.2016.1154108 10.1038/d41586-019-00857-9 10.1007/s10654-016-0149-3” -> 3 hits: https://doi.org/10.1080/00031305.2016.1154108; https://doi.org/10.1038/d41586-019-00857-9; https://doi.org/10.1007/s10654-016-0149-3
  • [2026-08-26] openalex fetch “10.1007/s10654-016-0149-3” -> 0 hits: (none)
  • [2026-08-26] openalex fetch “10.1080/00031305.2016.1154108” -> 0 hits: (none)
  • [2026-08-26] openalex fetch “10.1038/d41586-019-00857-9” -> 0 hits: (none)

[2026-08-26] SELECTION RECORD — CI-interpretation / compatibility-interval (WS-011, challenge 2026-08-26b)

Gap: challenge-2026-08-26b Facet A routed a WS-011 concept page (CI as compatibility interval; null-inclusion is not null-evidence). Target sources named by the maintainer + the detailed OA companion. All three are method/context tier (statistics epistemology, not domain evidence — gold-gate exempt), mutually reinforcing not echoes: institutional consensus / detailed misinterpretation catalog / the compatibility-interval reform manifesto.

  • SELECTED + FETCHED Greenland - P Values Confidence Intervals Guide 2016 (DOI 10.1007/s10654-016-0149-3; Eur J Epidemiol; Greenland/Senn/Rothman/Carlin/Poole/Goodman/Altman; cites 3220; OA Springer) -> inbox/ (452 KB, 14 pp, clean text). The load-bearing one for WS-011 — the 25-misinterpretations catalog directly covers reading a CI and the significance-dichotomization error.
  • SELECTED + FETCHED Amrhein - Retire Statistical Significance 2019 (DOI 10.1038/d41586-019-00857-9; Nature Comment “Scientists rise up against statistical significance”; cites 3080; OA) -> inbox/ (2728 KB, 3 pp). The rename-to-compatibility-interval + read-the-whole-interval argument the user invoked. (Short Comment, not a research article — cite for the reform framing, not for a magnitude.)
  • SELECTED, BLOCKED -> MAINTAINER-DROP Wasserstein - ASA Statement P Values 2016 (DOI 10.1080/00031305.2016.1154108; The American Statistician; Wasserstein & Lazar; cites 6252; OA flag set but Taylor & Francis anti-bot-blocks the fetch — HTTPError, as the fetch-reality note predicts). Identified, not lost. ACTION: maintainer drop the PDF into inbox/ (free at tandfonline.com/doi/full/10.1080/00031305.2016.1154108). NOT downgraded to a fetchable substitute (acquisition-pause-and-ask rule).

Nothing ingested. Next: process inbox stages the two fetched + registers rows; WS-011 then compilable once the CI-interpretation evidence is held (do NOT assert Greenland/Amrhein/ASA as findings until ingested).

[2026-08-27] SELECTION RECORD — sauna / heat exposure (refill; verified-unheld domain)

Refill pass after the staged queue drained. Sauna confirmed genuinely unheld (grep -ic sauna wiki/meta/sources.md = 0; 0 fabric pages). Verdict: a real unheld domain, but a LOW-YIELD refill target — DO NOT downgrade to fetch something. The domain’s structure, established this pass:

  • Hard-outcome gold evidence = a single Finnish cohort (Laukkanen / KIHD, Kuopio) — a built-in independence + transportability caveat (one team, one Nordic sauna-culture population); the JAMA mortality landmark and the Age&Ageing dementia paper are both this cohort.
  • The only sauna meta-analysis is on an ACUTE SURROGATE (CV function, not events/mortality), paywalled, and barely cited (zvab050, cites=1).
  • Everything acquirable-now is weak — mechanism / acute-physiology / pilot-feasibility / same-group narrative review. No gold hard-outcome SR exists that is also fetchable.
  • Every gold-ish source is fetch-blocked (JAMA paywall, OUP anti-bot, dead Hindawi link) -> the two worth having are maintainer-drops, not downgrades.

Layer-1 note: sauna is a peripheral lever, low expected rank (heavily discussed, small-to-uncertain effect on hard outcomes, single-cohort backing). Recorded so a future pass does not re-discover it cold.

Candidates + adjudication:

  • SELECTED, BLOCKED -> MAINTAINER-DROP Laukkanen - Sauna Bathing Fatal CV All-Cause Mortality 2015 (DOI 10.1001/jamainternmed.2014.8187; JAMA Intern Med; Laukkanen/KIHD; cites 288; oa=False). The landmark hard-outcome cohort (2315 men, frequency-graded sauna -> fatal CVD + all-cause mortality). The ONE source worth holding if the domain is opened. Paywalled — identified, not lost. ACTION: maintainer drop the PDF into inbox/ if this domain is wanted. NOT downgraded (pause-and-ask rule).
  • SELECTED, BLOCKED -> MAINTAINER-DROP (weak tier — optional) Hussain - Regular Dry Sauna Bathing Clinical Effects SR 2018 (DOI 10.1155/2018/1857413; Evid-Based Complement Alternat Med / Hindawi; OA-flagged but the downloads.hindawi.com PDF URL is dead — HTTPError). Gold-by-design (SR) but a demoted venue (ECAM/Hindawi — quality-flagged, many retractions), so a lower-tier SR at best. Maintainer-drop only if the domain is opened AND no better SR surfaces; otherwise DEFER.
  • DEFERRED Laukkanen - Sauna Bathing Dementia Alzheimer Cohort 2016 (DOI 10.1093/ageing/afw212; Age & Ageing; OUP anti-bot HTTPError). Same KIHD cohort as the mortality landmark -> not independent; attaches to the held dementia fabric only as a facet, not a gold opener. Drop later if dementia coverage is deepened.
  • DEFERRED Acute short-term efficacy of sauna treatment on cardiovascular function MA 2021 (DOI 10.1093/eurjcn/zvab050; Eur J Cardiovasc Nurs; MA; oa=False; cites=1). The only sauna MA, but on an acute surrogate (CV function), not a patient-important outcome — off the Layer-1 target. Surrogate + paywalled + uncited -> not worth a maintainer-drop now.
  • REJECTED Mayo Clin Proc 2018 narrative review (DOI 10.1016/j.mayocp.2018.04.008) — same Finnish group, narrative (moderate tier), not independent; would echo the cohort, not corroborate it.
  • REJECTED the “passive heat therapy” reframing hits (phy2.15699, ajpregu.00012.2025, ep093511, s40814-023-01314-1, the CU dissertation) — all mechanism / acute-physiology / pilot-feasibility, none a hard-outcome gold SR. Off-tier for a fabric opener.
  • ZERO-HIT queries (attributed): sauna ... systematic review, sauna health systematic review, sauna mortality meta-analysis all returned 0 -> title.search strictness (every word required in the title), NOT literature absence; the looser sauna bathing / sauna cardiovascular meta-analysis recovered the real hits above.

Nothing fetched, nothing ingested (research does not ingest). Net: this refill pass produced no gold acquirable-now source; the domain’s one worth-having source (JAMA mortality) is a maintainer-drop decision. Awaiting the maintainer’s call on whether to open a peripheral, single-cohort, block-walled domain vs. spend the refill elsewhere.

[2026-08-27] SELECTION RECORD — disagreeing-golds acquisition (coverage-gap discovery pass)

Targets from the coverage-gap discovery fan-out (AWAITS-harvest + GBD-diff + disagreeing-gold streams; all three subagent reports parent-verified). These three complete live type-D tensions the fabric already named by name — the highest-value new acquisitions the pass surfaced. All THREE confirmed oa=True in OpenAlex, but all three fetch-blocked at the built-in + local-Playwright rungs (publisher anti-bot / entitlement walls, NOT literature absence). Steel (last rung) not burned — these are OA and trivially maintainer-fetchable; pause-and-ask is the clean path. NOT downgraded.

  • SELECTED, BLOCKED -> MAINTAINER-DROP Nicholls - STRENGTH Omega-3 Cardiovascular 2020 (DOI 10.1001/jama.2020.22258; JAMA; Nicholls first author / Nissen senior; cites 1013; oa=True but jamanetwork.com articlepdf HTTP 403 to both requests + headless Chromium). The clean-comparator (corn-oil placebo) omega-3 NULL that adjudicates REDUCE-IT’s mineral-oil-control controversy — completes the type-D on Vitamin and Mineral Supplements for Disease Prevention (which names STRENGTH explicitly). Hard CV outcome (MACE). Highest value. Free PDF at jamanetwork.com/journals/jama/fullarticle/2773120.
  • SELECTED, BLOCKED -> MAINTAINER-DROP Graudal - Low Versus High Sodium Diet Cochrane 2020 (DOI 10.1002/14651858.cd004022.pub5; Cochrane Database Syst Rev; cites 130; oa=True but the cochranelibrary.com /pdf/full endpoint returns a Cloudflare HTML interstitial, not the PDF, to both request paths). The GOLD sodium-J-curve / harm-of-low pole (renin-aldosterone-catecholamine-lipid responses + modest BP effect) that Sodium Intake and Blood Pressure flags in its ”## The J-curve — named as the reason for the review, then never adjudicated” section (names Graudal). Big-rock (sodium->BP). Free at cochranelibrary.com (DOI above).
  • SELECTED, BLOCKED -> MAINTAINER-DROP Mente - PURE Sodium Cardiovascular Events 2016 (DOI 10.1016/s0140-6736(16)30467-6; Lancet; cites 489; oa=True via green-OA handle hdl.handle.net/10379/16625, which did not serve to requests or headless Chromium — redirect/ click-through). The landmark J-curve hard-outcome cohort (urinary sodium -> CV events, the low-arm harm signal); complements Graudal (BP/biomarker-weighted) on the events arm. Sodium Intake and Blood Pressure names PURE directly. NOTE: we hold OTHER PURE papers (Mente diet-score 2023, Dehghan fats-carbs 2017) but NOT this sodium-specific one — same cohort, distinct work, distinct slug.

ACTION: maintainer drop all three PDFs into inbox/ (all OA; DOIs above). Then process inbox stages them; a later sodium/omega-3 weave cashes the two flagged type-D gaps. Do NOT assert STRENGTH/Graudal/PURE findings until ingested (no-definitive-claims-from-nonfabric).

Zero-hit note: none — all three resolved. Nothing ingested (research does not ingest).

[2026-08-27] SELECTION RECORD — kidney gap (modifiable-exposure -> CKD; GBD-diff big-rock hole)

Discovery pass on the one genuine big-rock coverage HOLE the GBD lighthouse-diff found: no nucleus on modifiable exposure -> CKD incidence/progression (GBD #9 kidney dysfunction, ~2.4M deaths). Scope-edge but IN (risk-factor split): diet / activity / weight / smoking -> incident CKD or eGFR-decline slowing is a lifestyle-prevention question; OUT is prescriber-zone (dialysis, agent selection, drug titration in established CKD). Target = a gold SR/MA per modifiable lever to anchor a new kidney nucleus.

QUERIES (openalex, —sort cites; auto-logged above): diet/dietary-pattern -> CKD; healthy-lifestyle -> CKD; physical-activity -> CKD (MA); obesity -> CKD (MA); modifiable-risk-factors CKD; umbrella variants.

  • SELECTED + FETCHED He - Dietary Patterns Chronic Kidney Disease Risk SR 2020 (DOI 10.1186/s12937-020-00661-6; Nutrition Journal / BMC; He, Ling-Qiong … Shu, Long; gold SR+MA of 17 observational studies, 149,958 participants; oa=True) -> inbox/ (1490 KB, real PDF). The DIET lever. NOTE collision: first author He, Ling-Qiong is NOT the Feng-J He of He - Salt Reduction Blood Pressure 2013 — carry the given name in authors: at staging.
  • SELECTED + FETCHED Seidu - Physical Activity Chronic Kidney Disease Risk Meta-Analysis 2023 (DOI 10.1007/s10654-022-00961-7; Eur J Epidemiol; Seidu … Kunutsor … Khunti, Yates; gold MA of 12 cohorts, 1,281,727 participants; oa=True) -> inbox/ (708 KB, real PDF). The PHYSICAL-ACTIVITY lever. NOTE: Setor K. Kunutsor co-authors (we hold Kunutsor’s osteoporosis exercise-safety SR) — a shared-author flag: NOT type-E-independent of Kunutsor’s other work if ever cross-linked.
  • SELECTED, BLOCKED -> MAINTAINER-DROP (optional) obesity landmark 2017 Kidney Int — obesity predicts onset of CKD (DOI 10.1016/j.kint.2016.12.013; MA; cites 371; oa=False, Elsevier). THE landmark obesity->CKD MA. Lower priority: obesity is already a nucleus (GBD BMI); the OA Brazilian substitute (10.1590/1414-431x202010022) is a weaker journal/fewer cites, so NOT taken (no-downgrade). Drop only if a dedicated obesity->CKD-incidence facet is wanted.
  • DEFERRED the low-protein-diet-in-CKD umbrella (10.1093/nutrit/nuae178, oa=False) — progression-in- ESTABLISHED-CKD = prescriber-zone-adjacent, off the incidence question; and Protein Intake and Kidney Function already holds the protein->kidney facet.
  • DEFERRED the MDPI/Nutrients lifestyle review (10.3390/nu14183787) — narrative/moderate-tier, not a gold SR; the He + Seidu golds cover diet + activity better.

Two gold OA MAs fetched (diet + activity), one landmark maintainer-drop optional (obesity). Enough to anchor a modifiable exposure -> CKD nucleus at ingest. Nothing ingested (research does not ingest); hand off to process inbox.

[2026-08-27] SELECTION RECORD — lighthouse-bibliography-diff (method pilot + candidate harvest)

New discovery method mechanized: openalex.py refs <DOI> (BACKWARD = a work’s own bibliography, via filter=cited_by:) + cites <DOI> (FORWARD = works citing it, via filter=cites:), each anti-joined against the 326 registry DOIs (hits already held -> known, hidden) and --goldish-filtered to umbrella/SR/MA/RCT titles. Pilot seeds = our two DOI-bearing broad umbrellas (NNR/WCRF carry no DOI, not walkable). Nothing acquired in this pass — candidates only; acquisition is the user’s pick.

METHOD FINDING (corrects the prior). Predicted “backward-walking umbrellas is high-yield”; the pilot REVERSES it. Direction matters more than seed-type:

  • BACKWARD (refs) = LOW net-new. An umbrella’s bibliography is dominated by (i) methods-section citations (Egger’s test 57k cites, DerSimonian-Laird, AMSTAR, NutriGrade — pure noise) and (ii) its own SUBSUMED constituent MAs (fibre->cancer-site; fish->CHD/stroke/prostate). We deliberately hold the umbrella not its parts, so known=0 throughout — the anti-join does little work AND the constituents are mostly ratchet-echoes.
  • FORWARD (cites) = HIGHER, better-targeted net-new. Newer works citing our gold are either SUPERSEDING umbrellas (re-nucleus / F-refinement — “is our incumbent dated?”) or ADJACENT-question umbrellas. This is the direction that pays.

CANDIDATE SHORTLIST (all NEW, gold-tier, none held), classified by action-state:

  • RE-NUCLEUS / staleness-check (newer gold on a question we already nucleus): sugar umbrella 2023 (10.1136/bmj-2022-071609, 331 cites) vs held sugar gold; fibre umbrella 2025 (10.1016/j.clnu.2025.06.021) + updated fibre->mortality MA 2023 (10.1016/j.clnu.2023.11.005) vs our Veronese 2018 fibre incumbent (now 7 yr old — genuine staleness flag).
  • NEW-NUCLEUS (thin coverage, highest interest): the Schwingshackl food-groups dose-response MA series — T2D (10.1007/s10654-017-0246-y, 814), CHD/stroke/HF (10.1080/10408398.2017.1392288, 713), hypertension (10.3945/an.117.017178), all-cause mortality (10.3945/ajcn.117.153148), obesity (10.1093/advances/nmy092). Only 2 registry rows mention food-groups; a comparative multi-food-group -> outcome dose family is a strong Layer-1-ranking input. Vegetarian-diets umbrella 2020 (10.1016/j.clnu.2020.02.037; only 1 vegetarian row held).
  • ADJACENT, likely SUBSUMED (check at plan-queue, do not auto-take): diet->T2D-incidence umbrella 2019 (10.1136/bmj.l2368 — 23 T2D-diet rows already); Dietary Inflammatory Index umbrella 2021; handgrip umbrella 2020 (we hold Grip Strength and Mortality); SSB umbrella 2024; BP-lowering dietary-patterns umbrella 2020.
  • OFF-TELOS noise (drop): loneliness / eating-disorders / suicide / eye-disease / social-isolation umbrellas surfaced by the forward walk (they cite Veronese as a methods exemplar).

RESIDUAL LIMITATION: goldish + anti-join yields a clean but ~30-60/seed pool still mixed with off-telos umbrellas + methods papers; the “does it add value” step stays a read (gold-gate passes mechanically; telos-fit + already-covered-at-nucleus + ratchet do not). Recommend acquiring the Schwingshackl food-group series + the fibre/sugar staleness pair first; defer the rest to plan-queue.

[2026-08-27] SELECTION RECORD — umbrella backward-walk, all 15 walkable umbrellas (batch aggregation)

Extended the lighthouse-bib-diff per the maintainer steer (“walk the highest-quality overviews first; would a MA/SR even cite other MA/SRs?”). MEASURED the synthesis-fraction of a doc’s own bibliography (scratch aggregator over openalex.py refs, cross-seed co-citation ranking):

DATA — % of a document’s OpenAlex references that are themselves umbrella/SR/MA (title-heuristic lower bound): primary MA (Devries protein-kidney) 12% · umbrella (Veronese fibre) 42% · umbrella (Jayedi fish) 54% · GBD diet (Afshin) 10% · GBD alcohol 8%. CONFIRMS the maintainer’s intuition: a PRIMARY MA cites mostly primary studies (its pooled trials), so backward-walking one yields sub-gold primaries. Only UMBRELLA REVIEWS (SRs-of-SRs, by construction) cite syntheses at a useful rate. GBD is a POOR backward target — sparse OpenAlex ref-parsing (35-39 captured vs its true hundreds) AND low-synthesis (cites cohorts + own methods).

WALKABILITY SPLIT: the report-PDF apex bodies the maintainer named (WHO, NICE, SACN, WCRF, NNR) carry NO DOI and OpenAlex under-parses report reference lists -> NOT backward-walkable via OpenAlex (their bibliographies would need PDF-mining from done/, a separate path). Walkable = the ~15 journal-published umbrella reviews (+ journal-published guidelines ESC/ACC-AHA/EASL as a mixed second tier).

BATCH: walked all 15 walkable umbrella DOIs -> 414 distinct NEW goldish candidates after registry-DOI anti-join + methods-noise stoplist. LIMITATION FOUND: cross-seed co-citation ranking is CONTAMINATED — SR-methods/tooling papers (RoB2, Rayyan, PROSPERO, funnel-plot, “Ten simple rules for umbrella reviews”, Egger, AMSTAR) co-cite MORE strongly than domain syntheses because every umbrella’s methods section cites them; needed a 30-term stoplist and still leaks a few. AND the anti-join is DOI-exact -> misses same-trial papers under a different DOI (surfaced the DiRECT 2-yr durability paper though Lean - DiRECT T2D Remission 2018 is held).

STANDOUT NET-NEW (genuinely unheld, telos-central): the Schwingshackl food-groups dose-response MA series — food-groups -> T2D (10.1007/s10654-017-0246-y, 814) / CHD-stroke-HF (10.1080/10408398.2017.1392288, 713) / all-cause mortality (10.3945/ajcn.117.153148, 605) / hypertension (10.3945/an.117.017178) / obesity (10.1093/advances/nmy092). A comparative multi-food-group -> outcome DOSE family = a direct Layer-1 intervention-ranking input; only 2 registry rows touch food-groups. RECOMMEND acquire. SECONDARY (staleness / plan-queue, not urgent): sugar umbrella 2023 (10.1136/bmj-2022-071609) + fibre umbrella 2025 vs held incumbents; diet->T2D umbrella 2019 (10.1136/bmj.l2368, likely subsumed by Szczerba/Churuangsuk); low-cigarette-dose->CHD MA (10.1136/bmj.j5855, possible Jha F-refinement). DROP (off-telos): MS/Parkinson’s environmental-risk umbrellas, Vitamin-D-multiple-outcomes, insufficient-PA prevalence trends. Nothing acquired this pass.

  • [2026-08-27] openalex fetch “10.1007/s10654-017-0246-y 10.1080/10408398.2017.1392288 10.3945/ajcn.117.153148 10.3945/an.117.017178 10.1093/advances/nmy092” -> 0 hits: (none)

[2026-08-27] ACQUISITION — Schwingshackl food-group dose-response series (from umbrella-walk standout)

Fetching the 5-paper Schwingshackl food-group -> outcome dose-response MA series (surfaced by the umbrella backward-walk; the Layer-1 ranking-input standout). Escalation ladder run (openalex fetch -> local Playwright -> Europe PMC / NCBI PMC OA); Steel NOT burned (paused per acquisition discipline).

  • FETCHED T2D: 10.1007/s10654-017-0246-y -> inbox/ (real PDF, 555 KB). SELECTED.
  • OA-in-PMC but bot-walled through all free rungs (Steel-or-maintainer-drop): hypertension 10.3945/an.117.017178 (PMC5683007, isOA=Y); obesity 10.1093/advances/nmy092 (PMC6416048, isOA=Y). NCBI /pdf/ path returns a 20 KB bot-wall HTML; PMC OA fcgi service 404s. Recoverable via Steel.
  • NOT OA anywhere (maintainer-drop; the OpenAlex OA flag was a publisher soft-wall): CHD/stroke/HF 10.1080/10408398.2017.1392288 (Tandfonline; EuropePMC isOA=N); all-cause mortality 10.3945/ajcn.117.153148 (ScienceDirect; EuropePMC isOA=N).

Status: 1/5 fetched, 2/5 Steel-or-drop (OA), 2/5 maintainer-drop (paywalled). Paused for maintainer call on Steel-vs-drop rather than downgrading or burning the gated rung unasked.

[2026-08-27] SELECTION RECORD — comprehensive-guideline benchmark acquisition (DGA / DGAC / AHA)

Goal: acquire the flagship general-population comprehensive guidelines the fabric lacks, as guidance-null BENCHMARK objects for the decomposition-diff (“which guideline is closest to the fabric”). Not new gold evidence — the DGA/AHA are recommendation/guideline-tier; the DGAC report is the evidence- bearing half. Channel: direct download (gov) + openalex resolve (AHA DOI) + headless-browser escalation.

Targets, all SELECTED (fit-to-question: the comprehensive-guideline gap):

  1. DGA 2020-2025 — Dietary Guidelines for Americans, 2020-2025 (USDA/HHS). No DOI. URL: https://www.dietaryguidelines.gov/sites/default/files/2020-12/Dietary_Guidelines_for_Americans_2020-2025.pdf -> BLOCKED (ETIMEDOUT) — dietaryguidelines.gov times out from this egress IP (curl + urllib + headless Playwright all ETIMEDOUT; PLoS control fetched fine, so it is gov-domain-specific, not broad egress). Maintainer-drop requested.
  2. 2020 DGAC Scientific Report — Scientific Report of the 2020 Dietary Guidelines Advisory Committee (the evidence-bearing half; NESR systematic reviews). No DOI. URL: https://www.dietaryguidelines.gov/sites/default/files/2020-07/ScientificReport_of_the_2020DietaryGuidelinesAdvisoryCommittee_first-print.pdf -> BLOCKED (ETIMEDOUT), same as #1. Maintainer-drop requested.
  3. AHA 2021 (Lichtenstein) — 2021 Dietary Guidance to Improve Cardiovascular Health: A Scientific Statement from the AHA. Circulation 2021;144:e472-e487. DOI 10.1161/CIR.0000000000001031 (openalex oa=False, cites 1038). Free at https://www.ahajournals.org/doi/10.1161/CIR.0000000000001031 -> BLOCKED (HTTP 403) — ahajournals.org anti-bot (measured OUP/JAMA/Lancet-class wall). Maintainer-drop requested.

Adjudication: 3/3 SELECTED, 0 rejected, 0 fetched, 3 blocked -> maintainer-drop (pause-and-ask, no downgrade). Escalation ladder exhausted through headless Playwright; only Steel (--service steel, cloud IP) remains untried — gated on maintainer request. All three are free/public docs.

ARRIVED 2026-08-27 (maintainer-drop -> process-inbox). All 3 dropped into inbox/, all digital (no OCR), dedup-clean, ratchet-clear. Staged + registered: DGA - Dietary Guidelines for Americans 2020 (high) + DGAC - Scientific Report 2020 (gold, top-level) + AHA - Dietary Guidance Cardiovascular Health 2021 (high, papers/). Ingest queued as NEXT-UP #4; they fill matrix G1.

[2026-08-27] SELECTION RECORD — two divergent-pole comprehensives for the proximity matrix (Canada · EAT-Lancet)

Maintainer: “acquire canada food guide and eat-lancet” — the two divergent-pole general comprehensives the matrix-column analysis ranked #1 (Canada, dropped-dairy / plant-forward) and #3 (EAT-Lancet, plant-forward extreme). Acquisition for the guidance-null matrix, NOT fabric ingest.

  1. Query — (channel: web / canada.ca-scoped) Canada's Dietary Guidelines 2019 health professionals policy makers PDF Health Canada; then Wayback availability API on the moved artifact URL. EAT-Lancet needed no query (already held).
  2. Results —
    • Canada: canonical doc = Canada's Dietary Guidelines for Health Professionals and Policy Makers (Health Canada, Jan 2019, ISBN 978-0-660-25310-7, Cat H164-231/2019E-PDF). Live URL food-guide.canada.ca/.../CanadasDietaryGuidelines.pdf now 302-redirects to a Cloudflare JS bounce (file relocated); direct WebFetch of the canada.ca guidelines page 403 (Cloudflare anti-bot). Recovered via Wayback id_ raw snapshot (20251206084214, status 200) — complete 62pp, 139,896 chars, coverage 0.992 OK, clean digital text (no OCR).
    • EAT-Lancet: 0 new query — raw/binary/Willett - EAT-Lancet Planetary Health Diet 2019.pdf already staged + registered (moderate; DOI 10.1016/S0140-6736(18)31788-4). openalex resolve on the DOI errored on an un-quoted shell glob (paren in DOI) — moot, PDF already in hand.
  3. Adjudication — Canada SELECTED (highest-value column add: a general comprehensive that DIVERGES from the NNR/DGA consensus — dropped the dairy food group, plant-forward protein, no serving sizes; tests whether the NNR≈DGA convergence is robust, informs the fabric’s open dairy row). EAT-Lancet SELECTED (already held; the plant-forward / planetary-diet extreme; env-entangled + moderate tier — a compromised health-first benchmark, valued as the pole not the clean column). Both are RECOMMENDATION summaries (warrant separately published) -> high/moderate, NOT gold.
  4. Acquisition — Canada fetched to inbox/ (Wayback), then staged raw/binary/Canada - Dietary Guidelines 2019.pdf + registered (high/guideline, citekey healthcanada2019, no DOI — gov report). EAT-Lancet already staged. Both queued for the matrix comparator-read (NOT fabric ingest), per the established DGA/AHA lightweight-read pattern. No fetch block hit that needed a maintainer-drop.
  • [2026-08-27] openalex resolve “10.1016/j.kint.2016.12.013” -> 1 hits: https://doi.org/10.1016/j.kint.2016.12.013
  • [2026-08-27] openalex search “Dietary Factors Risk Chronic Obstructive Pulmonary Disease Systematic Review Meta-Analysis meta-analysis” -> 0 hits: (none)
  • [2026-08-27] openalex resolve “10.1007/s00198-023-06682-1” -> 1 hits: https://doi.org/10.1007/s00198-023-06682-1
  • [2026-08-27] openalex resolve “10.1183/23120541.00168-2019” -> 1 hits: https://doi.org/10.1183/23120541.00168-2019
  • [2026-08-27] openalex resolve “10.1186/s12886-026-04721-z” -> 1 hits: https://doi.org/10.1186/s12886-026-04721-z
  • [2026-08-27] openalex resolve “10.3390/nu17061037” -> 1 hits: https://doi.org/10.3390/nu17061037
  • [2026-08-27] openalex resolve “10.1038/bjc.2014.579” -> 1 hits: https://doi.org/10.1038/bjc.2014.579
  • [2026-08-27] openalex resolve “10.1136/bmj-2022-071620” -> 1 hits: https://doi.org/10.1136/bmj-2022-071620
  • [2026-08-27] openalex fetch “10.1038/bjc.2014.579” -> 0 hits: (none)
  • [2026-08-27] openalex fetch “10.1136/bmj-2022-071620” -> 0 hits: (none)
  • [2026-08-27] openalex fetch “10.1007/s00198-023-06682-1” -> 0 hits: (none)
  • [2026-08-27] openalex fetch “10.1183/23120541.00168-2019” -> 0 hits: (none)
  • [2026-08-27] openalex fetch “10.1186/s12886-026-04721-z” -> 0 hits: (none)
  • [2026-08-27] openalex fetch “10.3390/nu17061037” -> 0 hits: (none)

Maintainer research request: modifiable/lifestyle causes across dementia/neurodegeneration, Parkinson’s, cancers, CVD, Alzheimer’s, frailty, musculoskeletal, metabolic, pulmonary, sensory, kidney. Discovery + vetting only; nothing ingested. Run as a 4-agent fan-out in ISOLATED worktrees (so each agent’s openalex.py auto-log stayed off the shared ledger); agents discovered + vetted but did NOT fetch — the parent adjudicated, fetched OA centrally, and hand-logs the agents’ queries here (their worktree auto-logs were discarded on teardown). Coverage baseline first: registry disease-term counts flagged the GAPS (pulmonary/COPD=0 rows, macular/AMD=0, Parkinson’s/ALS thin) vs the well-covered families (cancer ~49, CVD ~71, diabetes ~35) held to a STRICT ratchet.

1. QUERIES (channel=openalex unless noted; incl. zero-hits — the negative space):

  • Cluster A (pulmonary/kidney/vision): diet COPD MA=0 · diet COPD SR=0 · obesity COPD MA=0 · fruit vegetable COPD MA=0 · physical activity COPD risk MA=0 · physical activity COPD MA=6 (all rehab-in-existing-COPD) · dietary patterns COPD MA=1 (ERJ 2020) · dietary factors COPD =2 · lifestyle CKD umbrella=0 · obesity CKD MA=6 · sodium CKD MA=6 (all SGLT2/serum-Na, off-target) · dietary pattern kidney disease MA=2 · diet macular degeneration MA=2 · diet cataract MA=0 · cataract nutrition MA=0 · physical activity macular degeneration MA=6 · lutein zeaxanthin macular MA=1. WebSearch (hand-logged): “diet dietary pattern COPD incidence SR MA” + “physical activity COPD incidence MA cohort” -> confirmed NO PA->COPD-incidence MA exists (cohorts only).
  • Cluster B (neurodegeneration): Parkinson physical activity MA=3 (none incidence) · Parkinson Mediterranean diet=1 (proceeding) · Parkinson diet MA=1 · Parkinson caffeine=6 (no new MA) · Parkinson dairy=3 · Parkinson exercise MA=6 (all symptom/motor, treatment-zone) · Parkinson modifiable risk factors=0 (Chen 2021 umbrella is incumbent) · Parkinson dietary pattern MA=0 · air pollution dementia MA=6 · physical activity dementia risk MA=2 · ultra-processed food cognitive=6 · alcohol dementia MA=4 · ALS physical activity risk=4 (no MA) · ALS lifestyle risk=4 (cohorts + a dissertation, no MA).
  • Cluster C (cancer, STRICT ratchet): alcohol site-specific cancer dose-response MA=1 (Bagnardi 2014) · physical activity cancer umbrella=6 (Rezende 2018 landmark) · adiposity cancer anatomical sites umbrella=1 (Kyrgiou 2017) · diet cancer umbrella (from 2017)=4 (all single-site) · alcohol cancer umbrella (from 2018)=1 (predatory venue). WebSearch: WCRF CUP / Third Expert Report citations (confirmed all held).
  • Cluster D (musculoskeletal/frailty): exercise bone mineral density MA=6 · resistance training BMD postmenopausal MA=6 · protein intake bone fracture MA=1 (0 cites) · dietary protein BMD MA=0 · weight loss knee osteoarthritis MA=6 · multidomain intervention frailty MA=0 (term artifact) · nutritional intervention frailty MA=5 · protein supplementation frailty SR=6.

2. RESULTS + 3. ADJUDICATION (SELECT / DEFER / REJECT with reason):

SELECTED — acquired to inbox/ (2):

  • Bagnardi 2014 — alcohol -> site-specific cancer comprehensive dose-response MA (10.1038/bjc.2014.579, gold, OA, 1348 cites). ADVANCER: fabric holds Rumgay 2021 (attributable burden, modelled), NOT the per-drink dose-response by site (the “no safe threshold; light drinking already raises oral/oesophageal/ breast risk” shape). Different question, not an echo. Fetched (421 KB, 12pp).
  • Exercise -> BMD in postmenopausal women, updated MA 2023 (10.1007/s00198-023-06682-1, gold, OA, 118 cites). ADVANCER: fabric holds exercise->falls (Sherrington) + ->fall-fractures (Wang) but NO exercise->BMD efficacy link. Surrogate caveat at ingest (BMD is a target, not patient-important — carry transmission-to-fracture). Fetched (3410 KB, 26pp).

SELECTED — genuinely OA but fetch-blocked -> MAINTAINER-DROP requested (4):

  • Wilker 2023 — ambient air pollution -> clinical dementia MA (10.1136/bmj-2022-071620, BMJ, gold, OA, 183 cites). Supplies per-unit PM2.5/NO2 dose-response HRs that Livingston 2024 holds only as a qualitative PAF. Blocked: openalex found no OA pdf; BMJ main-site gated.
  • Dietary Patterns -> COPD MA 2020 (10.1183/23120541.00168-2019, ERS Open Research, gold, OA, 38 cites). Primary pulmonary opener — fills a ZERO-row cell (healthy pattern OR 0.88 [0.82-0.94]). Blocked: ERS openres HTTPError on plain fetch + 404 on the browser PDF path.
  • Mediterranean Diet -> AMD (development + progression) MA 2025 (10.3390/nu17061037, MDPI Nutrients, MA, OA, 12 cites). Diet-pattern opener for the AMD gap (macular=0 rows). Venue caveat (MDPI) at ingest. Blocked: plain fetch + browser both 403.
  • Physical Activity -> cataract + AMD MA 2026 (10.1186/s12886-026-04721-z, BMC Ophthalmology, MA, OA, 1 cite). Dual-outcome PA lever for the vision gap; supersedes the 2017 AMD-only AJO MA. Recency = lower prior caveat. Blocked: browser returned an anti-bot HTML stub, not the PDF (discarded).

SELECTED — not OA -> MAINTAINER-DROP (1):

  • Garofalo 2017 — obesity predicts onset of CKD MA (10.1016/j.kint.2016.12.013, Kidney Intl / Elsevier, gold, 371 cites, OA=false). New kidney lever (adiposity axis, distinct from staged He-diet + Seidu-PA). ScienceDirect anti-bots OA anyway; no fetch attempted per the pause-and-ask rule.

DEFERRED (recorded, not acquired — take only if the named cell opens): Kyrgiou 2017 adiposity-cancer umbrella + Rezende 2018 PA-cancer umbrella (credibility-grading complements; WCRF held gold carries the decision) · Dietary-Factors->COPD component MA 2020 (food-level pair for ERJ; DOI unresolved, over-specified query=0) · alcohol->dementia MA 2024 (10.1016/j.arr.2024.102419, not OA) · domain-specific PA->dementia 2026 (10.1016/s2468-2667(26)00142-8) · UPF->cognition SR 2026 (10.1136/bmjnph-2025-001325, obs SR, no pooled magnitude) · MIND-diet->PD Agarwal 2018 (10.1007/s12603-018-1094-5, single cohort — thin-field counterweight) · weight-loss->knee-OA MA 2018 (10.1016/j.semarthrit.2018.06.005) · nutrition+/-PA->frailty MA 2021 (10.9778/cmajo.20200248) · obesity->CKD MHO/MUO refinement 2023 (10.1111/obr.13656) · omega-3->AMD 2025 (10.1016/j.tjnut.2025.101289) · lutein->MPOD 2021 (surrogate outcome, fails the bar).

REJECTED (echo / superseded / out-of-scope): dairy->PD (echo of Chen 2021 umbrella) · PD exercise/motor MAs (treatment-zone, diagnosed patients) · diet->single-site cancer umbrellas (frontier deepening of held WCRF matrix) · older COPD MA 2015 (superseded by ERJ 2020) · protein->fracture SR 2017 (0 cites, weak) · dietary-pattern->CKD MA 2021 (appears to BE staged He or its twin) · older AMD-only PA MA 2017 (superseded).

4. ACQUISITION: 2 fetched to inbox/ (Bagnardi, Exercise-BMD). 5 blocked -> maintainer-drop list handed to the maintainer (Wilker/BMJ · ERS-COPD · MDPI-AMD · BMC-PA-vision · Garofalo/Elsevier-CKD), all with clean DOIs above. Nothing downgraded to a weaker source to get something fetchable.

GAPS (no gold exists; expectancy-tested — do NOT force-acquire): PA->COPD incidence (cohorts only, no MA) · dietary-salt->CKD incidence (only SGLT2/serum-Na hits) · diet/antioxidant->cataract incidence (0 MAs) · adiposity/diet->lung-function/FEV1 trajectory · diet-pattern (MIND/Med)->Parkinson’s incidence (single cohorts only) · lifestyle->ALS incidence (cohorts + a dissertation) · protein intake->fragility fracture endpoint (no well-cited gold MA). Each recorded as a named G-gap for the coverage ledger.

[2026-08-28] SELECTION RECORD — protein-target GAP sweep (non-athlete / obese / elderly / deficit)

Maintainer-raised: the 1.62 g/kg breakpoint is a non-significant CI-midpoint, not robust guidance — raise a consolidated gap and look (again) for gold/high sources on the protein target across the strata that depart from Morton’s lean/normal-BMI trained sample. Discovery + vetting only; nothing ingested.

1. QUERIES (channel=openalex, auto-logged above unless noted; incl. zero-hits — the negative space):

  • protein intake requirement older adults muscle MA=0 (title-strictness artifact, 5 words) · protein supplementation older adults MA=6 (mostly does-it-help, not what-target) · protein energy restriction lean mass MA=0 (strictness) · high protein diet weight loss SR=3 · protein intake dose response muscle MA=4 (top hit = HELD Tagawa 2020 nuaa104; + a 2025 “updated dose-response MA”, no DOI) · protein sarcopenia umbrella=0 · dietary protein requirement aging SR=0 (strictness) · protein weight loss fat free mass MA=0 (strictness) · enhanced protein intake muscle mass overweight obesity MA=1 (Kokura 2024, clnesp) · updated protein intakes muscle mass strength MA=1 (the DOI-less 2025 rec).
  • WebSearch (hand-logged, no auto-log): “updated systematic review dose-response meta-analysis protein intakes muscle mass strength 2025” -> surfaced the Kokura 2024 obesity MA + confirmed Tagawa/Morton as the held dose-response incumbents; “Enhanced protein intake … overweight/obesity Clin Nutr ESPEN 2024 authors” -> first author Yoji Kokura (+ Ueshima, Saino, Maeda), DOI 10.1016/j.clnesp.2024.06.030.

2. RESULTS adjudication (SELECTED / DEFERRED / REJECTED + why):

  • SELECTED (acquire, maintainer-drop): Kokura - Protein Muscle Overweight Obesity 2024 (10.1016/j.clnesp.2024.06.030, SR+MA RCTs, tier high) — directly targets the OBESE stratum, the thinnest cell of the gap (protein amount to preserve muscle during weight loss). oa=False -> maintainer-drop.
  • DEFERRED: the 2025 “updated dose-response MA of protein intakes” — no DOI / 0 cites in OpenAlex, identity unresolved; freshest general dose-response but not cleanly citable. Maintainer to locate before acquiring; NOT registered (would be an unstable placeholder slug).
  • REJECTED (already held / off-target): Tagawa 2020 (HELD); the elderly-supplementation MAs (leucine, whey/soy, sarcopenia-RCT pooling) answer does supplementation help, not what target on which basis — the elderly TARGET is already covered by Bauer/Coelho-Junior/Moore (held). The 0-hit queries are title.search strictness (5-word titles), NOT literature absence.

3. ACQUISITION-REQUESTED: Kokura 2024 -> registered [NOT YET HELD] in sources.md (row + DOI) + ingest-queue PROTEIN-TARGET GAP thread + [AWAITS (Kokura - Protein Muscle Overweight Obesity 2024)] on Protein and Resistance Training for Muscle and Strength. Fetch status: maintainer-drop pending (paywalled).

[2026-08-28] SELECTION RECORD — poultry -> outcome evidence (for the planned “Meat” deliverable reframe)

Trigger: maintainer wants Red Meat retargeted to a “Meat” umbrella with a poultry section, but the fabric holds NO poultry->outcome evidence (poultry appears only as a substitute/comparator in Red Meat + the protein/substitution pages; ZERO poultry rows in the Bechthold matrix). Chose acquire-first, reframe-after (maintainer decision) — this search fills the poultry cell before any rename.

1. QUERY: openalex search, poultry x {cardiovascular / mortality / type-2-diabetes / cancer}, —type meta-analysis|systematic-review, —from-year 2015/2018 (auto-logged lines above). The 5-word-title strict first query (“poultry white meat cardiovascular mortality”) returned 0 from title.search strictness, NOT absence — broadening to 2-noun queries surfaced the landmarks.

2. RESULTS adjudication (SELECTED / DEFERRED / REJECTED + why):

  • SELECTED (acquire, all maintainer-drop):
    • Baker - Poultry Cardiovascular Disease 2026 (10.1016/j.cdnut.2026.107736, SR prospective cohorts, gold) — poultry->CVD morbidity+mortality, lands NULL. The CVD/mortality poultry cell.
    • Connolly - Poultry Type 2 Diabetes 2024 (10.1016/j.cdnut.2024.102336, SR+MA, gold) — poultry->T2D, lands NULL. The T2D poultry cell. NOTE same Campbell/Reed lab as Baker -> the two are NOT type-E-independent of each other (one coordinated group).
    • Papp - Poultry Cardiovascular Mortality Meta-Analysis 2021 (10.1080/10408398.2021.1975092, dose-response SR+MA, gold, cites=36) — poultry->CVD + all-cause mortality, the most-cited landmark AND the INDEPENDENT witness (Vienna/Schwingshackl group) against the Campbell-lab series. Worth the drop for independence + dose-response shape.
  • DEFERRED (cancer — combined-exposure / pattern-level, weaker fit; Red Meat already establishes poultry carries no colorectal signal):
    • 2022 BMC Medicine (10.1186/s12916-022-02257-9, cites=97, OA) — fish-poultry diet pattern vs meat-eaters (EPIC-Oxford style), a dietary-PATTERN cancer MA, not a poultry dose-response. Off the poultry-as-food question. Reconsider only if a “Meat” cut needs a pattern-level cancer corroborator.
    • 2017 poultry+fish cancer mortality (10.1080/01635581.2018.1412474) + 2019 poultry+egg cancer (10.14715/cmb/2019.65.1.3) — combined-exposure, cannot isolate poultry. REJECTED for fit.
  • REJECTED (off-target): the 2026 avian alpha/recombinant veterinary MA (10.5455/jrvs…) — animal husbandry, not human nutrition; a title.search false-friend on “poultry”.

3. ACQUISITION-REQUESTED: all three SELECTED registered as [NOT YET HELD] rows in sources.md (+ DOIs in staged-source-dois.tsv). Fetch status: maintainer-drop pending for all three — Baker/Connolly (Curr Dev Nutr) 403 on both openalex fetch AND headless-browser; Papp (T&F) paywalled oa=False. Per the pause-and-ask rule, NOT downgraded to a weaker fetchable source. Handed to process inbox once dropped; ingest + the Red Meat->“Meat” reframe follow AFTER these land and are vetted.

[2026-08-28] SELECTION RECORD — poultry~longevity alternative sources (MEAT-POULTRY-1)

Channel: openalex.py (auto-logged queries above: “poultry mortality”, “poultry all-cause mortality”, “white meat mortality”, “chicken mortality” [off-topic veterinary hits], “poultry longevity” [0 hits]) + a web-search authorship/independence check.

[2026-08-28] SELECTION RECORD — research: white/non-red/poultry meat -> all-cause mortality

Channel: openalex.py (auto-logged: “poultry all-cause mortality”, “white meat mortality”, “poultry mortality meta-analysis”, “meat all-cause mortality”, “chicken consumption mortality” [0 hits], “poultry cardiovascular all-cause mortality”; umbrella-type “meat mortality”/“red meat mortality”; 2022+ “poultry mortality”/“white meat cardiovascular”/“poultry health outcomes”) + web/API authorship checks. Already held/staged on this question (deduped OUT): Lupoli 2021 (poultry->all-cause mortality + CV events, STAGED), Kim 2017 (poultry->stroke, STAGED), Baker/Connolly abstracts (PARKED), Papp 2021 (dose-response all-cause mortality, registered maintainer-drop).

  • SELECTED, maintainer-drop: Ramel - White Meat Cardiovascular Diabetes Meta-Analysis 2023 (10.29219/fnr.v67.9543, Food & Nutr Research, OA, cites=24). NNR2023 SR group; white meat(poultry) -> CVD + T2D; INDEPENDENT gold, newer than Lupoli, DISPLACES the NCC-funded Connolly T2D abstract. Fetch: openalex RemoteDisconnected + local browser HTTP2-error + Steel returned HTML-not-PDF -> maintainer-drop. (Endpoint is CVD+T2D, not all-cause mortality — complements Lupoli, does not duplicate.)
  • SELECTED, maintainer-drop: Hidayat - Red Meat Substitution CHD Mortality Meta-Analysis 2022 (10.1093/nutrit/nuac017, Nutrition Reviews, oa=False, cites=35). SUBSTITUTION MA: red meat -> other proteins (incl. poultry) -> CHD + all-cause mortality; INDEPENDENT (Soochow/Qin). Serves the Meat reframe’s substitution frame (red->poultry swap on all-cause mortality). OUP paywalled -> maintainer-drop.
  • REJECTED: hsr2.70990 2025 “Red vs White Meat on CVD: Global Cross-Sectional ECOLOGICAL study” — weakest design (ecological fallacy, country-level), not gold. 10.1080/01635581.2018.1412474 (poultry+FISH -> cancer mortality) — combined exposure + cancer (DEFERRED for this reframe). Connolly 2023 narrative review (nu15163550) — narrative, not gold. Veterinary “poultry mortality” hits (bird flu / farm / worker cancer) — off-topic.
  • DEFERRED (Red Meat side, not this question): processed-meat umbrella 2026 (10.3389/fpubh.2026.1763155, OA) + red+processed-meat cancer umbrella 2025 (10.1371/journal.pone.0315436, OA) — relevant to the Red Meat deliverable, not poultry/white-meat -> all-cause mortality.
  • [2026-08-29] openalex search “Parkinson disease modifiable risk factors umbrella review” -> 0 hits: (none)
  • [2026-08-29] openalex search “Parkinson disease modifiable risk factors umbrella review” -> 0 hits: (none)
  • [2026-08-29] openalex search “Parkinson risk factors meta-analysis” -> 2 hits: https://doi.org/10.3390/life14010032; https://doi.org/10.1007/s11010-025-05392-y
  • [2026-08-29] openalex search “Dementia prevention intervention and care 2024 report Lancet standing Commission” -> 3 hits: https://doi.org/10.1016/s0140-6736(24)01296-0; https://doi.org/10.1016/j.maturitas.2025.108217; https://doi.org/10.4103/jgmh.jgmh_4_25
  • [2026-08-29] openalex search “Modifiable risk factors for incident dementia and cognitive impairment umbrella review of evidence” -> 2 hits: https://doi.org/10.1016/j.jad.2022.07.008; https://doi.org/10.21203/rs.3.rs-1151683/v1
  • [2026-08-29] openalex search “Non-genetic risk factors Parkinson’s disease overview 46 systematic reviews” -> 0 hits: (none)
  • [2026-08-29] openalex search “Non-genetic risk factors Parkinson disease systematic reviews” -> 0 hits: (none)
  • [2026-08-29] openalex search “Environmental risk factors Parkinson’s disease umbrella review meta-analyses” -> 1 hits: https://doi.org/10.1016/j.parkreldis.2015.12.008
  • [2026-08-29] openalex search “Potentially modifiable risk factors dementia mild cognitive impairment umbrella review meta-analysis” -> 1 hits: https://doi.org/10.1159/000536643
  • [2026-08-29] openalex search “Umbrella review Delphi study modifiable factors dementia risk reduction” -> 1 hits: https://doi.org/10.1002/alz.13577
  • [2026-08-29] openalex search “overview systematic reviews risk factors Parkinson disease” -> 0 hits: (none)
  • [2026-08-29] openalex resolve “10.3233/JPD-202521” -> 1 hits: https://doi.org/10.3233/jpd-202521
  • [2026-08-29] openalex search “Dietary patterns risk multiple cancers umbrella review meta-analyses prospective cohort” -> 1 hits: https://doi.org/10.1016/j.ajcnut.2024.11.020
  • [2026-08-29] openalex search “Body Mass Index Cancer Risk Umbrella Review Meta-Analyses Observational” -> 2 hits: https://doi.org/10.1080/01635581.2023.2180824; https://doi.org/10.21203/rs.3.rs-1894100/v1
  • [2026-08-29] openalex search “Vegetarian dietary patterns cardiovascular risk factors disease prevention umbrella review” -> 1 hits: https://doi.org/10.1016/j.ajpc.2024.100868
  • [2026-08-29] openalex search “Modifiable Lifestyle Factors Primary Prevention CKD systematic review” -> 2 hits: https://doi.org/10.1681/asn.2020030384; https://doi.org/10.3410/f.738597328.793578488
  • [2026-08-29] openalex resolve “10.1038/s41531-023-00603-z” -> 1 hits: https://doi.org/10.1038/s41531-023-00603-z
  • [2026-08-29] openalex resolve “10.1016/j.jacc.2025.08.015” -> 1 hits: https://doi.org/10.1016/j.jacc.2025.08.015
  • [2026-08-29] openalex search “Lifestyle metabolic syndrome nonalcoholic fatty liver disease umbrella review observational randomized” -> 0 hits: (none)
  • [2026-08-29] openalex search “Dietary factors risk chronic obstructive pulmonary disease systematic review meta-analysis” -> 0 hits: (none)
  • [2026-08-29] openalex search “Fiber intake risk chronic obstructive pulmonary disease dose response meta-analysis” -> 1 hits: https://doi.org/10.1002/fsn3.3640
  • [2026-08-29] openalex resolve “10.1016/S0140-6736(20)30367-6” -> 1 hits: https://doi.org/10.1016/s0140-6736(20)30367-6
  • [2026-08-29] openalex resolve “10.1007/s00774-021-01230-5” -> 1 hits: https://doi.org/10.1007/s00774-021-01230-5
  • [2026-08-29] openalex resolve “10.3390/app16062951” -> 1 hits: https://doi.org/10.3390/app16062951
  • [2026-08-29] openalex resolve “10.3390/vision9010016” -> 1 hits: https://doi.org/10.3390/vision9010016
  • [2026-08-29] openalex search “Vitamin D Supplementation Fractures Adults Systematic Umbrella Review Meta-Analyses Controlled Trials” -> 2 hits: https://doi.org/10.1210/clinem/dgab742; https://doi.org/10.3410/f.741035012.793592203
  • [2026-08-29] openalex search “umbrella review lifestyle nonalcoholic fatty liver disease” -> 1 hits: https://doi.org/10.1111/obr.13464
  • [2026-08-29] openalex search “nutritional lifestyle metabolic factors nonalcoholic fatty liver disease umbrella review” -> 0 hits: (none)
  • [2026-08-29] openalex search “Risk factors age-related macular degeneration Updated systematic review meta-analysis” -> 1 hits: https://doi.org/10.1097/md.0000000000041599
  • [2026-08-29] openalex search “Body mass index influences age-related cataracts updated meta-analysis systematic review” -> 0 hits: (none)
  • [2026-08-29] openalex search “dietary factors chronic obstructive pulmonary disease meta-analysis fruit fish vitamin” -> 0 hits: (none)
  • [2026-08-29] openalex search “Prevalence Related Factors Hearing Loss Among Adults Systematic Review Meta-Analyses” -> 1 hits: https://doi.org/10.1177/00034894241293045
  • [2026-08-29] openalex search “impact lifestyle factors across life course sarcopenia physical frailty” -> 1 hits: https://doi.org/10.1097/mco.0000000000001111
  • [2026-08-29] openalex search “Risk factors frailty sarcopenia community older adults meta-analysis” -> 1 hits: https://doi.org/10.1515/med-2025-1259
  • [2026-08-29] openalex search “Risk factors development knee osteoarthritis across lifespan systematic review meta-analysis” -> 1 hits: https://doi.org/10.1016/j.joca.2025.03.003
  • [2026-08-29] openalex search “Narrative Review Lifestyle Risk Factors Oxidative Stress Age-Related Hearing Loss” -> 1 hits: https://doi.org/10.3390/antiox12040878
  • [2026-08-29] openalex search “Body mass index age-related cataracts meta-analysis” -> 2 hits: https://doi.org/10.1371/journal.pone.0089923; https://doi.org/10.5935/0004-2749.2021-0382
  • [2026-08-29] openalex search “age-related cataract modifiable risk factors systematic review meta-analysis smoking sunlight” -> 0 hits: (none)
  • [2026-08-29] openalex search “modifiable risk factors age-related macular degeneration umbrella review” -> 0 hits: (none)
  • [2026-08-29] openalex search “dietary factors risk chronic obstructive pulmonary disease” -> 2 hits: ?; https://doi.org/10.1183/13993003.congress-2021.pa1010
  • [2026-08-29] openalex search “diet quality chronic obstructive pulmonary disease incidence cohort” -> 0 hits: (none)
  • [2026-08-29] openalex search “MASLD lifestyle intervention diet physical activity systematic review 2024” -> 0 hits: (none)
  • [2026-08-29] openalex search “associations nutritional lifestyle metabolic factors nonalcoholic fatty liver disease” -> 0 hits: (none)
  • [2026-08-29] openalex resolve “10.3233/JPD-202521” -> 1 hits: https://doi.org/10.3233/jpd-202521
  • [2026-08-29] openalex resolve “10.1038/s41531-023-00603-z” -> 1 hits: https://doi.org/10.1038/s41531-023-00603-z
  • [2026-08-29] openalex resolve “10.1186/s12902-022-01015-5” -> 1 hits: https://doi.org/10.1186/s12902-022-01015-5
  • [2026-08-29] openalex resolve “10.1111/obr.13464” -> 1 hits: https://doi.org/10.1111/obr.13464
  • [2026-08-29] openalex resolve “10.1002/fsn3.3640” -> 1 hits: https://doi.org/10.1002/fsn3.3640
  • [2026-08-29] openalex resolve “10.1681/asn.2020030384” -> 1 hits: https://doi.org/10.1681/asn.2020030384
  • [2026-08-29] openalex fetch “10.3233/JPD-202521” -> 0 hits: (none)
  • [2026-08-29] openalex search “Population fraction Parkinson attributable preventable risk factors” -> 0 hits: (none)
  • [2026-08-29] openalex resolve “10.1038/s41531-023-00603-z 10.3390/app16062951 10.1097/MD.0000000000041599 10.1016/j.ajpc.2024.100868” -> 4 hits: https://doi.org/10.1038/s41531-023-00603-z; https://doi.org/10.3390/app16062951; https://doi.org/10.1097/md.0000000000041599; https://doi.org/10.1016/j.ajpc.2024.100868
  • [2026-08-29] openalex fetch “10.1681/asn.2020030384” -> 0 hits: (none)
  • [2026-08-29] openalex fetch “10.1038/s41531-023-00603-z” -> 0 hits: (none)
  • [2026-08-29] openalex fetch “10.3390/app16062951” -> 0 hits: (none)
  • [2026-08-29] openalex fetch “10.1097/MD.0000000000041599” -> 0 hits: (none)
  • [2026-08-29] openalex fetch “10.1002/fsn3.3640” -> 0 hits: (none)
  • [2026-08-29] openalex fetch “10.1016/j.ajpc.2024.100868” -> 0 hits: (none)
  • [2026-08-29] openalex fetch “10.1186/s12902-022-01015-5” -> 0 hits: (none)
  • [2026-08-29] openalex fetch “10.1515/med-2025-1259” -> 0 hits: (none)

Scope. Maintainer research request across 11 age-related disease domains (dementia, Parkinson’s, cancer, CVD, Alzheimer’s, frailty, musculoskeletal, metabolic, pulmonary, sensory, kidney). Discovery fanned out to 4 parallel scouts (web-discovery + openalex-verify); acquisition + this record centralized. Coverage survey first (registry keyword rows): CVD ~103, metabolic ~70, cancer ~52, dementia ~23, MSK ~23, frailty ~15, kidney ~9, sensory ~9, pulmonary ~8, Parkinson ~2. Held-lighthouse cross-check confirmed the OPEN domains: sensory (nothing held), pulmonary/COPD (nothing; Omenn is lung cancer), Parkinson (Fang 2018 activity-only), CKD (He 2020 dietary-patterns SR only). Acquisition prioritized the open domains + dietary-PATTERN-axis advancers for the well-covered ones (the single-nutrient WCRF/GRADE corpus under-covers patterns). ECHO-GUARD applied to dementia/cancer/CVD/metabolic.

Fetch reality: environment-level IP-wall — plain openalex fetch AND local Playwright both hit HTTP 403/HTTPError on ASN, MDPI, Wiley, IOS Press; only Nature (npj) and BMC succeeded. Blocked-but-genuinely-OA sources are maintainer-drops (NOT downgraded, per acquisition-pause-and-ask). Steel cloud-browser not invoked unilaterally (gated/costly — maintainer’s call).

SELECTED — fetched to inbox/ (2)

  • Payami et al. 2023 — Population fraction of Parkinson’s disease attributable to preventable risk factors (npj Parkinson’s Disease; DOI 10.1038/s41531-023-00603-z; PAF/high; OA). PD PAF rock-sizer, opens the attributable-fraction framing for an open domain. -> inbox/.
  • Peng et al. 2022 — Lifestyle, metabolic syndrome and NAFLD: umbrella review (BMC Endocrine Disorders; DOI 10.1186/s12902-022-01015-5; umbrella/gold; OA). Multi-lever NAFLD-outcome consolidation; component levers partly held, value is the MASLD-outcome umbrella. -> inbox/.

SELECTED — MAINTAINER-DROP (vetted advancers, publisher-blocked; drop PDF into inbox/)

Open-domain landmarks (highest priority):

  • Kelly et al. 2020 — Modifiable Lifestyle Factors for Primary Prevention of CKD: SR+MA (JASN; DOI 10.1681/asn.2020030384; SR+MA 104 studies/2.75M; gold; OA but ASN 403). ★ THE CKD-incidence opener.
  • Belvisi et al. 2021 — Non-Genetic Risk Factors for Parkinson’s: overview of 46 SRs (J Parkinsons Dis; DOI 10.3233/JPD-202521; overview-of-SRs/gold; OA but IOS 403). ★ broadest PD multi-lever opener.
  • Hearing CV-risk umbrella 2026 (Applied Sciences/MDPI; DOI 10.3390/app16062951; umbrella/gold; OA but MDPI 403). ★ opens sensory-hearing at gold-design tier (new, appraise rigor at ingest).
  • AMD risk factors updated SR+MA 2025 (Medicine/LWW; DOI 10.1097/MD.0000000000041599; SR+MA; paywalled). ★ opens sensory-vision (smoking-dominant lever + cardiometabolic).
  • Fiber intake & COPD dose-response MA 2023 (Food Sci Nutr/Wiley; DOI 10.1002/fsn3.3640; SR+dose-response MA; OA but Wiley 403). Cleanest fetchable COPD diet lever (per-10g/d fiber ~ -26%). Covered-domain new-axis advancers:
  • Zhang et al. 2024 — Dietary patterns & multiple cancers: umbrella (AJCN; DOI 10.1016/j.ajcnut.2024.11.020; umbrella/gold; Elsevier). Pattern-axis the single-nutrient WCRF corpus under-covers.
  • Vegetarian patterns & CVD: umbrella of SRs 2024 (Am J Prev Cardiol; DOI 10.1016/j.ajpc.2024.100868; umbrella/gold; OA-journal but no-OA-pdf located). Fresh incidence-level pattern lever (CVD incidence RR 0.85).
  • GBD 2023 CVD burden & risk factors (JACC; DOI 10.1016/j.jacc.2025.08.015; GBD/high; OA). Layer-1 rock-sizing update to held Afshin GBD 2019 (79.6% of CVD burden modifiable).
  • Yu et al. 2022 — Modifiable risk factors for dementia: umbrella (J Affective Disorders; DOI 10.1016/j.jad.2022.07.008; umbrella/gold; Elsevier). Graded per-lever magnitudes the held Livingston 2024 Commission borrows rather than shows.
  • WCRF/AICR CUP diet+lifestyle report, updated Sept 2025 (WCRF; no DOI, direct wcrf.org PDF; report/gold-eq). Newest CUP, postdates held WCRF 2018; pattern-level judgements.
  • Knee OA risk factors across lifespan: SR+MA 2025 (Osteoarthritis & Cartilage; DOI 10.1016/j.joca.2025.03.003; SR+MA/gold; Elsevier). Overweight + prior injury as dominant modifiable levers w/ attributable fractions.
  • Tanaffos diet-COPD MA 2019 (Seyedrezazadeh; no DOI, PMCID PMC7309892; MA/high; PMC captcha). Broader-but- messier COPD diet map (fruit/fiber/fish protective); acquire if the fiber MA leaves it thin.

DEFERRED / REJECTED (logged, not acquired — reason)

  • Livingston 2024 Lancet dementia Commission — ALREADY HELD (registry: Livingston - Dementia Prevention 2024). Echo.
  • Karger dementia umbrella 2024 (10.1159/000536643), Rosenau Delphi 2024 (10.1002/alz.13577) — convergence / gap-finder; take only if Yu leaves levers thin.
  • Bellou 2016 PD umbrella (10.1016/j.parkreldis.2015.12.008) — superseded by Belvisi 2021.
  • BMI-cancer umbrella 2023 (10.1080/01635581.2023.2180824) — partial-held (refreshes WCRF Body Fatness 2018).
  • Nat Commun 2021 diet-cancer umbrella (10.1038/s41467-021-24861-8) — superseded by Zhang 2024.
  • Nutrients cardiometabolic-patterns umbrella (10.3390/nu16223873) — surrogate endpoints; DOI unverified.
  • Hip-fracture lifestyle MA 2021 (10.1007/s00774-021-01230-5), Vit-D fractures umbrella 2021 (10.1210/clinem/dgab742) — MSK intervention side well-covered (vit-D/Ca/fracture heavily held); single-exposure.
  • Frailty risk MA 2025 (10.1515/med-2025-1259, paywalled) + lifestyle-sarcopenia narrative review 2025 (10.1097/MCO.0000000000001111) — frailty moderately covered; MA mixes non-modifiable, review sub-gold.
  • NAFLD treatment umbrella (10.1111/obr.13464) — prescriber-edge (treatment of established NAFLD).
  • Cataract BMI MA (10.5935/0004-2749.2021-0382) — single-exposure; cataract has NO comprehensive umbrella.
  • CKD progression+QoL SR 2023 (PMC10205767) — secondary to Kelly; DOI unverified.

G-GAPS surfaced (named, not acquisition targets)

ARRIVED 2026-08-29 — process-inbox reconciliation (10 staged, 1 deduped, 3 still-missing)

Maintainer dropped 9 PDFs; combined with the 2 openalex-fetched (Payami, Peng). All processed into raw/binary/papers/, registered in sources.md, DOI-logged, queue-blocked (ingest-queue.md INBOX DROP 2026-08-29). Chunk-clean (born-digital, no OCR). Scout-name corrections (verified from PDF title pages at staging):

  • PD overview “Belvisi et al. 2021” is actually Chen, Yancong et al. (10.3233/JPD-202521) — and it was ALREADY INGESTED (registry chen2021pd, woven into the Age-Related Diseases deliverable). The re-dropped copy DEDUPED to inbox/_duplicates/; not re-staged.
  • Cancer umbrella “Zhang et al. 2024” is actually Yin, Jia-Li et al. (AJCN, 10.1016/j.ajcnut.2024.11.020); registered yin2025cancer (2025 print vol).
  • Dementia umbrella “Yu et al. 2022” is the Zhang, Yaru et al. ResearchSquare PREPRINT (10.21203/rs.3.rs-1151683/v1); VoR = J Affect Disord 2022;314:160-167, 10.1016/j.jad.2022.07.008, cites=108 — registered tier moderate with a [PREPRINT] caveat; acquire the VoR before citing effect sizes.

Tier notes set at staging: Moreno-Herraiz hearing umbrella -> moderate (GRADE LOW + MDPI Applied Sciences off-field + cites=1); Payami PD-PAF -> moderate (single regional case-control, not an SR). The rest gold/high by design.

STILL-MISSING (3 of the maintainer-drop advancers — NOT provided in this drop):

  • GBD 2023 CVD burden & risk factors — JACC, 10.1016/j.jacc.2025.08.015 (Layer-1 rock-sizing refresh of held Afshin GBD 2019).
  • WCRF/AICR CUP diet+lifestyle report, updated Sept 2025 — direct wcrf.org PDF, no DOI.
  • Tanaffos diet-COPD MA 2019 (Seyedrezazadeh) — no DOI, PMCID PMC7309892 (PMC captcha-blocked). Re-drop any of these if wanted.

FOLLOW-UP DROP 2026-08-29b — 2 more files

  • WCRF CUP Global DLP report ARRIVED -> cashes the still-missing “WCRF/AICR CUP 2025” slot. Staged raw/binary/papers/WCRF - Dietary Lifestyle Patterns Cancer Prevention 2025.pdf; registered wcrf2025dlp (gold report, no DOI). Chunk-clean (2 chunks, paragraph-pack). Pattern-axis (breast 84 pubs / colorectal 86 pubs) postdating the held WCRF 2018 corpus.
  • Second file was NOT the Tanaffos MA. It is Turdaliyeva et al. 2023 - Cultural Food Habits & COPD in Kazakhstan (ResearchSquare PREPRINT 10.21203/rs.3.rs-3130446/v2), a single-population case-control (609 non-smoking men, red-meat -> COPD, self-report). Sub-gold + preprint + weaker than the held Valisoltani/Parvizian COPD MAs -> PARKED inbox/_hold/ (not active-queued, not registered).
  • Still-missing: GBD 2023 CVD (10.1016/j.jacc.2025.08.015) + the intended Tanaffos diet-COPD MA 2019 (Seyedrezazadeh, PMC7309892).
  • [2026-08-29] openalex search “placebo adherence mortality meta-analysis” -> 2 hits: https://doi.org/10.1007/s00392-013-0642-6; ?
  • [2026-08-29] openalex search “healthy adherer effect” -> 6 hits: https://doi.org/10.1111/jep.12095; https://doi.org/10.1136/bmj.333.7557.18; https://doi.org/10.1177/2047487314544084; https://doi.org/10.1007/s00198-014-2900-5; https://doi.org/10.1002/acr.21759
  • [2026-08-29] openalex search “dietary guideline adherence confounding” -> 0 hits: (none)
  • [2026-08-29] openalex search “secular changes coffee coronary heart disease confounding” -> 0 hits: (none)
  • [2026-08-29] openalex search “trans fat dietary recommendations margarine history” -> 0 hits: (none)
  • [2026-08-29] openalex fetch “10.1136/bmj.333.7557.18” -> 0 hits: (none)
  • [2026-08-29] openalex resolve “10.1007/s00392-013-0642-6” -> 1 hits: https://doi.org/10.1007/s00392-013-0642-6

[2026-08-29] SELECTION RECORD — “was ‘the guideline, not the food, was causal’ ever quantified?” (challenge-driven)

Question: quantify guideline-adherence confounding — the pure-bundle magnitude, and any guideline-specific- food natural experiment (margarine promotion / egg vilification). Channel: openalex (5 queries auto-logged).

  • SELECTED -> acquire (pure-form healthy-adherer magnitude, UNHELD) — DOI CORRECTED [2026-08-29]:
    • 10.1136/bmj.38875.675486.55 Simpson SH et al. 2006 BMJ 333:15 “A meta-analysis of the association between adherence to drug therapy and mortality” (oa=True, cites 1374) — THE magnitude paper (the pooled placebo/drug-adherence -> mortality OR). Acquisition BLOCKED (openalex fetch HTTPError on the BMJ PDF, both attempts) -> pause-and-ask maintainer (escalate fetch_browser/Steel, or manual drop of bmj.com/content/bmj/333/7557/15.full.pdf). NOT downgraded.
    • 10.1136/bmj.38865.675486.55 Chewning 2006 BMJ 333:18 commentary “The healthy adherer and the placebo effect” — IN HAND (maintainer drop) but it is a one-page NARRATIVE commentary ON Simpson, carries the qualitative statement («adherence … a surrogate marker for overall healthy behaviour») but NO pooled number; tier context/weak, NOT the anchor. My earlier DOI 10.1136/bmj.333.7557.18 pointed HERE (page 18) by mistake, not at the page-15 meta-analysis.
    • 10.1007/s00392-013-0642-6 2013 Clin Res Cardiol MA “placebo adherence on reducing CV mortality” (meta-analysis, cites 31, oa=False) — the pooled hard-outcome magnitude. Non-OA -> maintainer-drop.
    • 10.1111/jep.12095 2013 “Healthy adherer effect — the pitfall in interpretation” (oa=False) — the interpretive review; DEFER (secondary to the two above).
  • REJECTED / not-pursued: the drug-specific replications (bisphosphonate 10.1007/s00198-014-2900-5; cardiac-rehab 10.1177/2047487314544084; fracture 10.1002/acr.21759; Archintern 2009 10.1001/archinternmed.2009.324) — corroborate the effect in narrow drug contexts, add no independent magnitude beyond the MA; hold in reserve.
  • ZERO-HIT queries (Q3-Q5: guideline-adherence-confounding / secular coffee-CHD drift / trans-fat-margarine history) — attribute to TOOL, not corpus. openalex title.search is AND-strict on long multi-token phrases; Q1/Q2 fired, so the engine works. These zeros do NOT establish literature absence — a broad web/Scholar pass is the honest next step if pushed. Working inference (identification-argument-backed): the guideline-specific- food natural-experiment quantification is a genuine G-gap.
  • [2026-08-29] openalex search “meta-analysis association between adherence to drug therapy and mortality” -> 1 hits: https://doi.org/10.1136/bmj.38875.675486.55
  • [2026-08-29] openalex fetch “10.1136/bmj.38875.675486.55” -> 0 hits: (none)
  • [2026-08-30] openalex search “inorganic nitrate blood pressure meta-analysis” -> 6 hits: https://doi.org/10.3945/jn.112.170233; https://doi.org/10.1016/j.niox.2021.04.006; https://doi.org/10.1097/hjh.0000000000002524; https://doi.org/10.1016/j.freeradbiomed.2024.02.011; https://doi.org/10.32920/26521168.v1
  • [2026-08-30] openalex search “physical activity exercise chronic pain systematic review” -> 3 hits: https://doi.org/10.3390/healthcare4020022; https://doi.org/10.1177/1357633x17716576; https://doi.org/10.53841/bpshpu.2019.28.3.10
  • [2026-08-30] openalex search “physical activity exercise chronic pain overview Cochrane reviews” -> 3 hits: https://doi.org/10.1002/14651858.cd011279.pub3; https://doi.org/10.1002/14651858.cd011279.pub2; https://doi.org/10.1002/14651858.cd011279
  • [2026-08-30] openalex search “healthspan lifestyle intervention randomized” -> 0 hits: (none)
  • [2026-08-30] openalex search “falls prevention interventions network meta-analysis older” -> 0 hits: (none)
  • [2026-08-30] openalex search “healthspan randomized controlled trials systematic review” -> 1 hits: https://doi.org/10.1093/gerona/glag133
  • [2026-08-30] openalex search “diet exercise healthspan meta-analysis” -> 0 hits: (none)
  • [2026-08-30] openalex search “exercise fall prevention network meta-analysis” -> 2 hits: https://doi.org/10.3389/fmed.2026.1831444; https://doi.org/10.1016/j.puhe.2026.106316
  • [2026-08-30] openalex search “interventions preventing falls community older adults meta-analysis” -> 3 hits: https://doi.org/10.1111/jgs.17375; https://doi.org/10.1016/j.ijnurstu.2020.103564; https://doi.org/10.1016/j.maturitas.2022.12.003
  • [2026-08-30] openalex resolve “10.3945/jn.112.170233 10.1002/14651858.cd011279.pub3 10.1093/gerona/glag133” -> 3 hits: https://doi.org/10.3945/jn.112.170233; https://doi.org/10.1002/14651858.cd011279.pub3; https://doi.org/10.1093/gerona/glag133
  • [2026-08-30] openalex fetch “10.1002/14651858.cd011279.pub3” -> 0 hits: (none)
  • [2026-08-30] openalex fetch “10.1111/jgs.17375” -> 0 hits: (none)
  • [2026-08-30] openalex fetch “10.3945/jn.112.170233” -> 0 hits: (none)
  • [2026-08-30] openalex fetch “10.1093/gerona/glag133” -> 0 hits: (none)

[2026-08-30] SELECTION RECORD — breadth restock (research-first, per maintainer): nitrate->BP · PA->pain · falls-NMA · healthspan

Context: age-related-diseases breadth batch (8/8) fully ingested; maintainer chose “research first, then ingest” over the deferred in-hand domain gold (Brickwood/Franz defer-carded; Fried HELD). Target = the 4 “continue breadth” gold names from the Gate-31 queue-review (2026-08-28). One search cluster per gap.

SELECTED (4 — all identified with clean DOIs; NONE plain-fetchable -> maintainer-drop):

  1. Siervo - Inorganic Nitrate Beetroot Blood Pressure 2013 — DOI 10.3945/jn.112.170233 (J Nutr; gold/MA; cites 373; oa=True). The landmark nitrate/beetroot->BP MA; opens the dietary-nitrate->BP cell (unheld). Fetch: ScienceDirect HTTP 403 (plain + local Playwright). -> maintainer-drop.
  2. Geneen - Physical Activity Chronic Pain Cochrane Overview 2017 — DOI 10.1002/14651858.cd011279.pub3 (Cochrane overview-of-reviews; gold; cites 1186; oa=True). The PA->chronic-pain umbrella; extends Chronic Pain and Physical Activity. Fetch: Cochrane Library HTTP 403. -> maintainer-drop.
  3. Falls Prevention NMA Community Older Adults 2021 — DOI 10.1111/jgs.17375 (JAGS; SR + NETWORK meta-analysis; gold; cites 210; oa=True). Falls-prevention NMA (the queue’s “Pillay falls-NMA” cell; SELECTED on fit-to-question — a gold falls NMA — author to VERIFY at process-inbox; if not the exact intended paper it is a valid gold substitute for the same cell). Fetch: Wiley pdfdirect HTTP 403. -> maintainer-drop.
  4. Zheng - Interventions Multidimensional Healthspan 2026 — DOI 10.1093/gerona/glag133 (J Gerontol; systematic review; gold; cites 0 [new 2026]; title exact-match “Interventions that prolong multidimensional healthspan in humans: a systematic review”). Author = VERIFY at process-inbox (queue named “Zheng”). Fetch: OUP PAYWALLED (no OA pdf resolvable). -> maintainer-drop.

Zero-hit queries attributed (query-strictness, NOT literature-absence):

  • “healthspan lifestyle intervention randomized” (0), “diet exercise healthspan meta-analysis” (0) — openalex title.search AND-strictness on the rare word “healthspan”; the looser “healthspan randomized controlled trials systematic review” returned the target (glag133). Not a literature absence.
  • “falls prevention interventions network meta-analysis older” (0) — over-specified; “interventions preventing falls community older adults” (—type meta-analysis) returned jgs.17375. Not an absence.

REJECTED / not-selected:

  • Nitrate cluster: the 2021 niox.2021 (older-adults-only, cites 32), 2020 hjh (repeated-admin, oa=False), 2024 freeradbiomed (during/post-EXERCISE BP, off-question) — all narrower/lower-cited than Siervo 2013; Siervo is the broadest landmark. Held as fallbacks only if Siervo cannot be dropped.
  • Falls cluster: 2020 ijnurstu (multifactorial only, oa=False), 2023 maturitas (cost-effectiveness, off-question), 2026 Frontiers/PubHealth (brand-new cites=0, narrower exercise-only) — jgs.17375 is the well-cited comprehensive NMA. 2026 ones held as fallback.
  • PA-pain cluster: 2016 healthcare4020022 (non-specific LBP only), 2017 telemedicine (delivery-mode, not the exposure) — the Geneen Cochrane overview dominates as the umbrella.

Acquisition-requested: all 4 -> BLOCKED (403/paywall) on plain openalex fetch AND local Playwright. Per acquisition-pause-and-ask: NOT downgraded. PAUSED for maintainer PDF-drop into inbox/ (or authorize a Steel/—service steel attempt on the 3 OA ones). DOIs recorded above; nothing ingested.

[2026-08-30] CORRECTION — breadth-restock batch was ALREADY HELD (acquisition dedup failure)

The 2026-08-30 SELECTION RECORD above staged 4 “gold” sources (Geneen, Dautzenberg, Siervo, Zheng) as fresh acquisitions. ALL FOUR were already held: Geneen/Siervo/Zheng are exact-DOI duplicates of sources ingested 2026-08-28; Dautzenberg is an older falls NMA saturated by the ingested Pillay 2024 NMA + Sherrington 2019. The research pass deduped only against the active ingest-queue, not against ingested ground truth (done/ + wiki/sources/), so it re-discovered landmark papers already in the fabric. NONE ingested. Caught by the delegated author-!=-auditor agent (Geneen) + parent saturation check (the other 3). Rows removed from the registry; all 4 recorded in rejected-sources.md; process fix logged to backlog.md (dedup on DOI/md5 vs the held set at the acquisition step, not slug vs queue). Search-ledger lesson: a candidate’s novelty must be checked against INGESTED ground truth, not the queue.

[2026-08-30] SELECTION RECORD — consumption-cascade gaps (efficacy-vs-effectiveness challenge)

Filling the two G-gaps opened by the 2026-08-30 challenge (user-challenge-ledger.md; queue consumption-cascade section): SUGAR/palatability -> reward -> intake, and ALCOHOL -> food intake. Dedup done vs INGESTED ground truth (wiki/sources/ + done/), not just the queue — no held source covers either cascade (all held alcohol = cancer/vascular/BP/mortality; all held sugar = caries/weight/glucose). Gold-gate + symmetric standards (a source that would REFUTE the maintainer’s intuition clears the same bar as one that confirms it).

GAP 2 — ALCOHOL -> food/energy intake. FILLED.

  • SELECTED + FETCHED -> inbox/: «Effect of alcohol consumption on food energy intake: a systematic review and meta-analysis» — BJN 2019, doi 10.1017/s0007114518003677, OA, cites 84, gold (MA). Slug Kwok - Alcohol and Food Energy Intake Meta-Analysis 2019. Directly the aperitif-effect / passive over-consumption mechanism; the exact decision-question the gap names.
  • REJECTED: the 2003 “Meta-analysis and Pooled Analysis… Diet and Chronic Diseases: Alcohol” (no DOI, 0 cites, oa=False) — off-question (chronic-disease pooling, not acute intake).
  • 0-hit “alcohol energy intake appetite systematic review” = title.search strictness (the MA above carries “food energy intake” not “appetite”), NOT literature absence — proven by Q2 hitting it.

GAP 1 — SUGAR/palatability -> reward -> intake. PARTIALLY FILLED (2 golds, both BLOCKED -> maintainer- drop); the reward->intake OUTCOME MA appears NOT to exist (mechanism-dominated field).

  • SELECTED, BLOCKED -> maintainer-drop: «A systematic review and meta-analysis examining the effect of eating rate on energy intake» — AJCN 2014, doi 10.3945/ajcn.113.081745, cites 351, gold (SR/MA). Slug intent Robinson - Eating Rate and Energy Intake Meta-Analysis 2014 (Robinson E, confirmed). The RATE side of the rate-vs-reward question the fabric’s held Hall UPF RCT opened (Hall: +508 kcal driven by eating RATE, “not palatability”). AJCN OA flag is unreliable — plain fetch returned PAYWALLED. NOT downgraded.
  • SELECTED, BLOCKED -> maintainer-drop: «What Is the Evidence for “Food Addiction?” A Systematic Review» — Nutrients 2018, doi 10.3390/nu10040477, cites 317, gold (SR). Author to confirm at intake (not guessed). The REWARD-side evidence-appraisal — a symmetric-standards gold that tests, rather than assumes, the sugar-reward-drive claim. Genuinely OA (MDPI) but plain fetch AND fetch_browser both 403; likely on PMC — maintainer-drop or PMC route.
  • DEFERRED: «The Prevalence of Food Addiction as Assessed by the Yale Food Addiction Scale: A Systematic Review» (Nutrients 2014, doi 10.3390/nu6104552, cites 546, gold SR, OA) — measures how COMMON the YFAS construct is, not whether reward drives intake/weight; lower decision-value than the 2018 evidence SR. Reconsider if the 2018 SR proves unobtainable.
  • REJECTED (mechanism tier, would inform DIRECTION only — the gap wants a human OUTCOME source): the food-reward neuroimaging/dopamine landmarks (Volkow tics.2010.11.001; Stice a0013600) and the palatable-food animal-circuit papers (AgRP celrep.2024; CB1 s41386-021) — all mechanism/animal, not human ad-libitum outcome. Confirm the field’s shape: the reward cascade is MECHANISM-rich, human-outcome-SR-poor.
  • ADJACENT, not acquired (scope — a THIRD mechanism, and individual RCTs not SR): energy-density -> ad libitum intake (Rolls/Stubbs covert-manipulation lineage, ajcn/62.2.316 etc.; oa=False). Note for a future energy-density gap, distinct from reward and rate.

Honest sub-finding (records to the challenge verdict): the SUGAR-reward -> subsequent-intake channel has NO gold human-OUTCOME MA — it is served by mechanism/neuroimaging/animal work plus contested “food addiction” construct SRs. This VINDICATES the fabric’s pre-existing “insufficient-evidence” position and its Hall eating-rate emphasis: the closest gold human evidence is on eating RATE and energy DENSITY, not on reward/palatability drive. Acquire Robinson (rate) + the 2018 evidence SR (reward appraisal); do NOT expect a clean reward->intake magnitude to exist.

Hand-off: process inbox stages Kwok (fetched); the two maintainer-drop DOIs await PDF drop before staging. Nothing ingested here.

[2026-08-30] SELECTION RECORD — hyper-palatable foods (HPF) as a construct distinct from ultra-processing

Extension of the same-day consumption-cascade research (maintainer-raised): research the engineered-palatability / habit-forming channel as its OWN construct, separate from NOVA ultra-processing. This is the telos “operationalize before admitting” move — HPF (Fazzino) is the specified, measured version of the vague “palatability” label, and it separates the reward channel from the processing/eating-rate channel that the held Hall UPF trial confounded. Gold-gate + symmetric standards (an “engineered to be addictive” claim gets the same evidence bar — fad-bar applies).

Field shape (the honest finding): HPF is a NEWER field whose gold OUTCOME SR/MA does not yet exist. The real advance is the operational DEFINITION (Fazzino 2019) — it makes HPF admissible as a specified exposure at all. Outcome evidence is single-study / narrative-review dominated. This CONFIRMS and sharpens the sugar-food-reward gap: palatability can be OPERATIONALIZED even though the outcome magnitude is not yet gold — admit HPF as a specified candidate/gap under the transportability caveat, not as an established effect.

SELECTED + FETCHED -> inbox/:

  • «Which Foods May Be Addictive? The Roles of Processing, Fat Content, and Glycemic Load» — Schulte, PLoS ONE 2015, doi 10.1371/journal.pone.0117959, OA, cites 657. Slug Schulte - Which Foods May Be Addictive 2015. THE “which properties are habit-forming” landmark — the maintainer’s exact framing (processing level + fat + glycemic load as the addictive-load predictors). Tier: moderate (single cross-sectional YFAS + food-rating study), but a field-defining landmark -> tier-override at ingest.

SELECTED, BLOCKED -> maintainer-drop:

  • «Hyper-Palatable Foods: Development of a Quantitative Definition and Application to the US Food System Database» — Fazzino, Obesity 2019, doi 10.1002/oby.22639, cites 172. Slug intent Fazzino - Hyper-Palatable Foods Quantitative Definition 2019. The construct anchor / operational definition (fat+sugar, fat+sodium, carb+sodium threshold combinations) — the admissibility key without which HPF is just the vague label. Genuinely OA but Wiley pdfdirect 403 on plain + browser; on PMC (PMC6768746) — PMC route or maintainer-drop. Tier moderate (definition/food-system-DB paper) -> tier-override.
  • «Ultraprocessed, hyper-palatable, and high energy density foods: Prevalence and…» — Obesity 2023, doi 10.1002/oby.23897, cites 40, oa=False. The THREE-CONSTRUCT disambiguation (UPF vs HPF vs energy density prevalence/overlap) — the single most fabric-relevant piece for separating the rate / reward / processing / density channels. maintainer-drop. Tier moderate.

DEFERRED (moderate, noted not acquired — avoid moderate-tier bloat / retention ratchet):

  • «The Reinforcing Natures of Hyper-Palatable Foods: Behavioral Evidence» (Curr Addict Rep 2022, doi 10.1007/s40429-022-00417-8, cites 32, oa=False) — the behavioral-reinforcement narrative review; on-question but narrative-tier. Reconsider if the outcome channel needs a synthesis.
  • «Relationship between Emotional Eating, Consumption of Hyperpalatable Energy-Dense Foods…» (J Obesity 2022, Hindawi, doi 10.1155/2022/4243868, cites 101, OA, flagged systematic-review) — Hindawi provenance flag + narrower (emotional-eating) question. The one nominal “SR” hit; quality-gate it hard before trusting.

REJECTED: the neuronal-plasticity / animal HPF pieces (fnbeh.2017.00019 etc.) — mechanism/animal tier, direction-only, off the human-outcome question.

Routes to the queue (challenge builds no fabric): the [consumption-cascade: sugar-food-reward] gap is EXTENDED into an HPF-operationalization sub-gap — the specified-exposure handle that lets the reward channel be admitted distinct from UPF. Nothing ingested. Hand-off: process inbox stages Schulte + Kwok; Fazzino / 2023-Obesity / (the 2018 Nutrients + Robinson from the earlier block) await maintainer PDF drop.

[2026-08-30] SELECTION RECORD — breadth restock into GENUINELY-UNCOVERED domains (Gate-32 corrective #1)

Context. Gate 32 corrective #1: the 2026-08-30 restock died by re-discovering already-held landmarks without checking ingested ground truth. This pass ran an Explore coverage-map FIRST (registry + hubs + index + named-gap grep + staged/cut dedup) and targeted only cells confirmed THIN/ZERO. Dedup vs registry: flavonoid / antidepressant / levothyroxine / e-cigarette / cannabis all 0 hits — genuinely uncovered. Deliberate NON-targets (already appraised + CUT, do not re-search): sauna/heat, cold immersion (no gold hard-outcome; raw/binary/cut/). Channel: openalex only (queries auto-logged above). Nothing ingested; all SELECTED sources are fetch-blocked -> maintainer-drop (pause-and-ask, no downgrade).

Cluster 1 — Tea / dietary flavonoids -> CVD & mortality (diet beverage + isolated-component; parallels held coffee cluster). Opens BOTH a whole-food (tea) and an isolated-nutrient (flavonoid) cell -> sets up the food-vs-component / healthy-user-confound question the fabric prizes.

  • SELECTED Chung - Tea Consumption Cardiovascular Dose-Response 2020 (10.1093/advances/nmaa010; Adv Nutr; gold MA; dose-response SHAPE; cites=132; OA but OUP-blocked). WHY: the tea landmark with the curve.
  • SELECTED Mazidi - Flavonoid Intake and Mortality 2020 (10.3390/nu12082350; Nutrients; gold MA; dose-response; cites=28; OA but MDPI-blocked). WHY: the isolated-flavonoid cell (distinct exposure from tea). NOTE: Mazidi is prolific/sometimes-contested -> quality eye at ingest, design is gold MA.
  • DEFERRED Kim - Flavonoid Lignan Mortality 2016 (10.1093/aje/kww207; AJE; 287 cites, higher landmark) — fallback if Mazidi quality-fails. DEFERRED currency-checks: 2024 tea-mortality MA (10.4178/epih.e2024056), 2025 coffee+tea CVD (10.3389/fnut.2025.1570644) — Chung’s dose-response opens the cell; these add little at the margin. REJECTED: ...1570644.s001 (a supplement table, not a paper); 2014 Zhang tea MA (10.1007/s10654-014-9960-x; superseded by Chung dose-response).

Cluster 2 — Antidepressants as a standing drug (telos names them; currently only an exercise-comparator). Efficacy + limitations frame.

  • SELECTED Cipriani - 21 Antidepressants Efficacy Acceptability 2018 (10.1016/S0140-6736(17)32802-7; Lancet; gold NMA, 522 trials/21 drugs; cites=3555; OA but Lancet-blocked). WHY: THE efficacy/acceptability landmark.
  • SELECTED Henssler - Antidepressant Discontinuation Symptoms 2024 (10.1016/S2215-0366(24)00133-0; Lancet Psychiatry; gold SR+MA; cites=96; OA but Lancet-blocked). WHY: the LIMITATIONS half (withdrawal incidence — separates true discontinuation-symptoms from nocebo).
  • DEFERRED ... first-vs-second-gen 2016 (10.1136/bmjopen-2015-010919) — superseded by Cipriani. REJECTED: appi.focus reprint (10.1176/appi.focus.16407) + ResearchGate dup of Cipriani (both same paper).

Cluster 3 — Subclinical hypothyroidism / levothyroxine -> patient-important outcomes (endocrine + standing drug; sharp de-escalator question).

  • SELECTED Feller - Thyroid Hormone Therapy Quality of Life 2018 (10.1001/jama.2018.13770; JAMA; MA; cites=262; OA but JAMA-blocked). WHY: the RIGHT stratum — general adults, QoL/symptom outcome (pools TRUST); the landmark de-escalator (no symptom benefit).
  • REJECTED (off-stratum): ALL the pregnancy-outcome levothyroxine MAs (10.1093/humupd/dms052; 10.1186/s12902-021-00699-5; 10.1186/s12958-018-0410-6; 10.1136/bmjopen-2018-022837; 10.3389/fendo.2021.797423) — pregnancy management is a narrower, therapeutics-adjacent stratum, OFF the general-adult standing-drug question. DEFERRED companion RCT TRUST Stott 2017 NEJM (Feller’s MA already pools it; acquire only if the RCT-level detail is needed).

Cluster 4 — Cannabis / e-cigarettes -> CV (modern substance exposure; ZERO coverage; the analogue to held smoking+alcohol big-rocks). HONEST field-shape: NO gold hard-outcome (mortality) MA exists — cohorts too young; evidence is surrogate/moderate. These OPEN the cell at insufficient-hard-outcome-evidence (the “not yet” bucket, symmetric standards), NOT as a gold lever. Will need tier-overrides at ingest.

  • SELECTED Skotsimara - Electronic Cigarettes Cardiovascular Effects 2019 (10.1177/2047487319832975; Eur J Prev Cardiol; MA; cites=197; NOT OA -> paywalled). WHY: the e-cig CV-effects landmark (surrogate-heavy: HR/BP/autonomic — label as such).
  • SELECTED Sebastian - Cannabis Atherosclerotic Cardiovascular Disease 2025 (10.1016/j.disamonth.2024.101849; Dis Mon; MA; cites=5; NOT OA). WHY: the most recent cannabis->ASCVD MA; thin but the best held-able signal.
  • DEFERRED Tetrault - Marijuana Pulmonary Function 2007 (10.1001/archinte.167.3.221; the respiratory SR landmark, but 2007/dated). REJECTED: e-cig autonomic/surrogate SRs (10.1016/j.ypmed.2019.105770; 10.1007/s10286-020-00683-4; 10.1186/s12954-020-00440-w — narrower surrogate arms); HYP abstract P152 (10.1161/hyp.81.suppl_1.p152 — conference abstract, not a paper).

Acquisition status (fields 4). ALL 7 SELECTED = fetch-BLOCKED (openalex plain + headless browser both 403 on Chung/Mazidi; Lancet/JAMA/SAGE/Elsevier hard-walls) -> MAINTAINER-DROP pending (DOIs in staged-source-dois.tsv). Steel-escalation NOT run (pause-and-ask; awaiting maintainer go-ahead vs manual drop). Nothing downgraded. Hand-off: once dropped, process inbox stages + registers (canonical slugs + tier-overrides for the moderate cluster-4 pair) -> then Ingest.

  • [2026-08-31] openalex search “cocoa flavanol cardiovascular events randomized trial meta-analysis” -> 0 hits: (none)
  • [2026-08-31] openalex search “cocoa flavanol cardiovascular” -> 3 hits: https://doi.org/10.1093/ajcn/nqac055; https://doi.org/10.1186/s12937-025-01121-9; https://doi.org/10.1007/s10654-026-01382-6
  • [2026-08-31] openalex search “flavonoid supplementation cardiovascular meta-analysis” -> 0 hits: (none)
  • [2026-08-31] openalex resolve “10.1093/ajcn/nqac055” -> 1 hits: https://doi.org/10.1093/ajcn/nqac055
  • [2026-08-31] openalex search “flavonoid biomarker mortality cohort” -> 0 hits: (none)
  • [2026-08-31] openalex search “flavonoid Mendelian randomization cardiovascular” -> 0 hits: (none)
  • [2026-08-31] openalex fetch “10.1093/ajcn/nqac055” -> 0 hits: (none)

[2026-08-31] SELECTION RECORD — component-isolating source to close the food-vs-component thread

Gap (fields 1). Close the food-vs-component held thread (Gate-33 top corrective). The fabric holds only the FFQ-collinear OBSERVATIONAL side of the flavonoid question — Mazidi - Flavonoid Intake and Mortality 2020 (component computed FROM carrier foods, never isolated) + Chung - Tea Consumption Cardiovascular Dose-Response 2020 (the food). Is the Food Category Doing Any Work names the exact missing designs: “no biomarker, no isolated-flavonoid trial arm, and no Mendelian randomization — none of the designs that separate a food-borne component from its carrier.” Searched openalex (channel) for the three isolating designs: isolated-flavanol RCT on a hard outcome / flavonoid-biomarker cohort MA / flavonoid-MR.

Candidates + adjudication (fields 3).

  • SELECTED Sesso - Cocoa Flavanol Cardiovascular COSMOS 2022 (10.1093/ajcn/nqac055; AJCN; landmark RCT, ~21,442 participants, cocoa-flavanol EXTRACT 500 mg flavanols/d incl. 80 mg epicatechin vs placebo, hard CV events; cites=181). WHY: the single most decisive closer — an actual RANDOMIZED PHYSICAL ISOLATION of the flavanol on a patient-important hard outcome, the strongest of the three isolating designs (beats an observational biomarker cohort, which is still confounded). Tier high (landmark RCT), type rct; a genuine ratchet-ADVANCER (a different design on the flavonoid question, not an echo of the held FFQ observational). The flavanol analog of the held Semba/resveratrol-biomarker null on the wine/resveratrol case.
  • REJECTED 10.1186/s12937-025-01121-9 (2025 SR, cocoa flavanols -> peripheral/cerebral MICROVASCULAR function; oa=True, fetchable). WHY: outcome is a SURROGATE (microvascular function), not a patient-important hard outcome — fails fit-to-question. Fetchable, but taking it to dodge the COSMOS block would be exactly the downgrade the pause-and-ask rule forbids.
  • DEFERRED 10.1007/s10654-026-01382-6 (2026, “Cocoa flavanol supplementation and prevention of CVD: a novel [analysis]”; oa=False). WHY: likely a re-analysis of COSMOS; acquire only if the primary trial’s detail is insufficient after ingest.
  • 0-hit queries: “flavonoid biomarker mortality cohort”, “flavonoid Mendelian randomization cardiovascular”, “flavonoid supplementation cardiovascular meta-analysis” — all attributable to title.search STRICTNESS (the design word rarely sits in the title), NOT a literature-absence claim. Not chased: COSMOS is the stronger design and one isolating source closes the flavonoid instance; the biomarker/MR routes stay a named G-gap if COSMOS alone proves insufficient.

Acquisition status (fields 4). SELECTED COSMOS = fetch-BLOCKED across ALL sanctioned rungs: openalex plain fetch -> PAYWALLED (OUP/AJCN, no accessible OA PDF despite the oa=True flag); PMC free-to-read copy (PMC9170467) behind a Cloudflare browser-check AND not in the PMC OA bulk-download subset (oa.fcgi 404); headless Playwright + Steel both returned the 131-char Cloudflare interstitial. -> MAINTAINER-DROP pending (DOI 10.1093/ajcn/nqac055; PMC9170467 free-to-read for manual save). Nothing downgraded. Hand-off: once dropped to inbox/, process inbox stages + registers (slug Sesso - Cocoa Flavanol Cardiovascular COSMOS 2022, tier high, type rct) -> then a DELEGATED ingest subagent -> main-agent weave onto Is the Food Category Doing Any Work / Flavonoid Intake and Mortality / Tea Consumption and Cardiovascular Risk.

[2026-08-31] SELECTION RECORD — nitrate/beetroot -> BP (next-breadth acquisition; staged gold drained)

Trigger: user chose “tee up next breadth ingest”; staged pile holds no fresh domain gold (ingest-queue L336), so breadth = acquisition. Target cell: dietary nitrate/beetroot -> blood pressure (clean unsaturated cell, actionable lever, named in Gate-31 review L297). Channel: openalex.py search (auto-logged above) + browser/curl fetch attempts (manual-logged here).

Candidates + adjudication:

  • SELECTED (maintainer-drop owed): 10.1039/d2fo03140j — “Regulatory effect of dietary nitrate on BP: a meta-analysis of randomized…” (Food & Function, 2023, cites=12). RCT-only MA = the ideal design here (nitrate is a dosable, blindable isolate — the fibre-isolate logic: clean design beats mixed-evidence tier). oa=False -> fetch not attempted (paywalled venue).
  • SELECTED-alt (OA, maintainer one-click): 10.1186/s12937-025-01114-8 — Nitrate/BP/vascular biomarkers MA (Nutrition Journal / BMC, 2025, cites=11). BMC = genuinely OA, but the sandboxed network gated every fetch route (openalex “paywalled” artifact; browser download-start error; curl -> 3KB HTML consent gate). Maintainer can download the free PDF directly.
  • DEFER: 10.1016/j.jff.2024.106082 — nitrate/BP umbrella (Journal of Functional Foods, 2024, cites=17). Highest design tier BUT genuinely paywalled + JFF is a functional-foods venue (advocacy-lean watch). Take only if the two above are unavailable.

Acquisition-requested: d2fo03140j (primary) + s12937-025-01114-8 (OA alt) -> BLOCKED by sandboxed network on all automated routes -> maintainer-drop to inbox/. Per acquisition-pause-and-ask: did NOT downgrade to a weaker source, did NOT pivot cells silently. Fetch status: 0 fetched, 2 maintainer-drop.

  • [2026-09-01] openalex search “white meat poultry cardiovascular diabetes meta-analysis” -> 0 hits: (none)
  • [2026-09-01] openalex fetch “10.29219/fnr.v67.9543” -> 0 hits: (none)