EpiStack

heterogeneity-subgroup-dose

5 min read

Orientation — slice: heterogeneity-subgroup-dose

Slice question: for whom, and at what level, is habitual egg consumption net beneficial/harmful/neutral? Covers (i) type-2-diabetic RCTs and dedicated diabetic-cohort analyses, (ii) dose-response above ~1 egg/day, (iii) inter-individual variation (hypo-/hyperresponders, ApoE genotype). 6 source nodes minted, S-1..S-6.

Type-2 diabetes: RCTs and dedicated diabetic-cohort analyses

  1. S-1 - DIABEGG high-egg diet and cardiovascular risk factors in type 2 diabetes, 3-month RCT — RCT, n=140, 2 eggs/day vs <2/week for 3 months; no adverse lipid change. Pro-egg. Surrogate. Egg-industry funded (Australian Egg Corporation).
  2. S-3 - Dietary fat and cholesterol and CVD risk among women with type 2 diabetes, Nurses’ Health Study — dedicated diabetic-cohort analysis (n=5,672 T2D women), dietary cholesterol → CVD risk gradient. Anti-egg. Hard endpoint. Not industry-funded. Also the slice’s secondary dose-response source (continuous per-200mg-cholesterol RR).
  3. S-2 - One egg per day vs oatmeal breakfast — inflammation and cardiometabolic markers in type 2 diabetes RCT — small crossover RCT, n=29 T2D patients, egg period lower TNF-α/AST, no other adverse change. Pro/neutral-egg. Surrogate. Heaviest COI in the slice (Egg Nutrition Center via Fernandez lab).

Dose-response above ~1 egg/day

  1. S-4 - Egg consumption and risk of type 2 diabetes in men and women, dose-response in Physicians’ and Women’s Health Studies — best dose-response primary: explicit intake categories, ≥1/day vs none gives RR 1.58 (men)/1.77 (women) for incident T2D, graded across categories. Anti-egg. Hard endpoint. Not industry-funded.
  2. S-3 - Dietary fat and cholesterol and CVD risk among women with type 2 diabetes, Nurses’ Health Study — cross-listed from above; models a continuous per-200mg-dietary-cholesterol dose-response for CVD risk in T2D women.

Inter-individual variation — responders and ApoE genotype

  1. S-5 - Characteristics of human hypo- and hyperresponders to dietary cholesterol — foundational: stable hypo-/hyperresponder classification in controlled feeding, uncorrelated with habitual diet or body weight (i.e., not predictable from lifestyle). Pro-egg-for-most argument. Surrogate.
  2. S-6 - Apolipoprotein E polymorphism and serum lipid response to dietary fat and dietary cholesterol — ApoE 4/4 carriers show ~10% total-cholesterol rise on a moderate dietary-cholesterol increase, vs minimal response in 3/3. Anti-egg-for-ApoE4 argument. Surrogate. Study manipulates dietary cholesterol generally (era-typical proxy for egg intake), not an egg-branded intervention — closest primary I could verify as non-review within budget.

search_scope

Named seeds seeded by the brief (DIABEGG/Fuller, Ballesteros, Katan 1987) were confirmed directly via targeted WebSearch + WebFetch against PubMed/AJCN/Nutrients/PMC pages, pulling exact citation, design, sample size, and funding/COI off the article page or abstract. For dose-response (T2D-outcome specifically, to avoid duplicating slice 2’s CVD/mortality dose-response ground) and for the ApoE-genotype sub-topic, ran additional targeted keyword searches (e.g. “Djoussé Gaziano egg consumption type 2 diabetes,” “ApoE genotype dietary cholesterol egg LDL response trial,” “Sarkkinen apolipoprotein E polymorphism dietary cholesterol”) rather than opening the named discovery hubs (Tamez 2016, Wallin, Fernandez responder reviews, Ordovas ApoE reviews, Godos 2021) and walking their reference lists — a narrower, faster method than full citation-snowballing, chosen given the test-run time budget. Cross-checked candidate anti-egg diabetic cohorts against slice 2’s named cohort-family list (NHS/HPFS, Zhong 2019 pooled set, PURE, China Kadoorie, EPIC, UK Biobank, Physicians’/Women’s Health, Japan JACC/JPHC) to keep S-3 and S-4 clearly diabetes-dedicated rather than general-population-with-a-subgroup.

exclusions

  1. Fuller et al. 2018 AJCN, “DIABEGG — randomized weight-loss and follow-up phase” (PMID 29741558) — same trial/cohort as S-1 - DIABEGG high-egg diet and cardiovascular risk factors in type 2 diabetes, 3-month RCT, extending it through a weight-loss phase and 6-month follow-up. Not minted separately: would be a near-duplicate at a 6-note budget; kept the original 3-month RCT as the representative primary.
  2. Djoussé/Physicians’ Health Study general-population CVD-mortality and heart-failure egg papers — out of scope, belong to slice 2 (general-population hard-endpoint cohorts), not diabetes-specific or dose-response-focused enough to be mine.
  3. Hu 1999 (NHS/HPFS) and Zhong 2019 (US pooled cohorts) — explicitly reserved for slice 2 per the search plan (general-population papers that merely report a diabetic subgroup); not re-minted.
  4. Weggemans/Zock/Ordovas/Katan 2001 (Atherosclerosis), “Apoprotein E genotype and the response of serum cholesterol to dietary fat, cholesterol and cafestol” — a pooled analysis across multiple feeding studies (n=395), i.e. a review-like discovery hub, not a primary; not minted. Its own primaries were not individually traced within budget.
  5. Herron/Fernandez hypo-/hyperresponder egg-feeding studies (e.g. Herron et al., “Men classified as hypo- or hyperresponders…,” J Nutr 2003, and related) — found in search results but not minted; S-5 - Characteristics of human hypo- and hyperresponders to dietary cholesterol already covers the foundational classification concept and the 6-note budget was tight. A fuller run should add at least one Herron/Fernandez primary to directly cover the egg-specific (not just generic-dietary-cholesterol) responder literature the brief names.

Slice summary

6 primaries, split 3 T2D-RCT-or-cohort / cross-listed dose-response / 2 responder-genotype (S-3 and S-4 double as the dose-response pair). Endpoint types mixed as instructed: S-1, S-2, S-5, S-6 surrogate; S-3, S-4 hard. Balance held: pro-egg-leaning = S-1, S-2, and the “unpredictable hypo-responders” reading of S-5; anti-egg-leaning = S-3, S-4, and the ApoE4-carrier reading of S-6. Motivatedness flagged on 2/6 (S-1 Australian Egg Corporation, S-2 Egg Nutrition Center/Fernandez); S-3/S-4 are independent (Harvard NIH-funded cohorts), which strengthens the anti-egg heterogeneity anchors’ credibility. Gaps: no egg-specific (as opposed to generic-dietary-cholesterol-vehicle) ApoE primary found within budget — S-6 is the closest verified non-review primary; and I did not snowball the named discovery-hub reference lists directly (see search_scope), so this pool likely misses some primaries a full citation-chase would surface, particularly more Herron/Fernandez-lineage responder studies and any dedicated diabetic-cohort analyses outside the Harvard system (e.g. Asian or European diabetic cohorts).

Graph View