EpiStack

mechanism-surrogate

5 min read

Orientation — mechanism-surrogate (slice 1b-A)

Slice: controlled-feeding / metabolic-ward studies and short RCTs in general or healthy(-ish) adults measuring blood-lipid response (LDL-C, ApoB, HDL, lipoprotein particles) to dietary cholesterol / egg intake. Endpoint = surrogate throughout. 7/7 sources written (budget exactly met).

Classic dose-response magnitude (how much does LDL/cholesterol rise per unit dietary cholesterol?)

  1. S-10 - A dose-response study of the effects of dietary cholesterol on fasting and postprandial lipoproteins in healthy young men — modern (1994), rigorous 4-way within-subject crossover (0/1/2/4 egg yolks/day), directly reports LDL-C; the best single “rise, with a plateau” primary in the pool.
  2. S-8 - Serum cholesterol response to changes in the diet - the effect of dietary cholesterol (metabolic-ward dose-response series) — foundational (1965) Keys metabolic-ward series behind the classic cholesterol-response equation; older methodology (total serum cholesterol, pre-LDL-fractionation era; no COI disclosure norms existed yet), so ranked below Ginsberg despite its historical primacy.

Real-world whole-egg feeding (ecological validity, not purified cholesterol)

  1. S-12 - Ingestion of egg raises plasma low density lipoproteins in free-living subjects — clean single-food add-on in free-living lacto-vegetarians; LDL +12%, ApoB +9%, HDL unchanged. Note: the search plan’s seed list mis-cited this as “Sacks NEJM” grouped with the null-finding side — verification shows it’s Lancet 1984 and it is in fact a rise finding; recorded accurately here, not forced into the plan’s assumed bucket.
  2. S-20 - Daily egg consumption in hyperlipidemic adults - effects on endothelial function and cardiovascular risk — higher-baseline-risk (hyperlipidemic) population, 6-week sustained phase vs. a real cholesterol-free egg-substitute comparator (not just egg-free); LDL essentially unchanged.

Mechanisms for attenuated/null responses (why some studies find no rise)

  1. S-19 - Dietary cholesterol contained in whole eggs is not well absorbed and does not acutely affect plasma total cholesterol concentration — absorption/bioavailability mechanism: intact-egg cholesterol is poorly absorbed, distinguishing whole-egg feeding from purified-cholesterol feeding (e.g. Keys).
  2. S-15 - Dietary cholesterol from eggs increases plasma HDL cholesterol in overweight men consuming a carbohydrate-restricted diet — diet-context mechanism: on a low-carbohydrate background, egg cholesterol raises HDL and preserves the LDL/HDL ratio.
  3. S-17 - Intake of up to 3 eggs per day is associated with changes in HDL function and increased plasma antioxidants in healthy young adults — particle-quality/functional mechanism: reframes the outcome from LDL-C mass to particle size and HDL function.

search_scope

Queries run via WebSearch/WebFetch against the brief’s named seeds (Ginsberg 1994/1995, Sacks 1984, Keys/Anderson/Grande 1965, Fernandez-lab whole-egg RCTs [Mutungi 2008, DiMarco 2017], Njike 2010, Kim & Campbell 2018) plus snowball searches for exact author lists, DOIs/PMIDs, and funding/COI disclosures (PubMed, journal publisher pages — AHA journals, Lancet/ScienceDirect, J Nutr, Nutrients/MDPI, Nutrition Journal/PMC; Semantic Scholar API attempted for citation counts but rate-limited (429) on every call, so citation_count is unknown on all 7 — a genuinely cheap fill I could not get, not a lazy skip). Did not independently open the named discovery hubs (Clarke 1997 BMJ, Weggemans 2001 AJCN, Berger 2015 AJCN, McNamara/Vincent reviews) — the brief’s seed list already resolved to 7 well-balanced, verified, distinct-angle primaries within budget, so hub reference-list mining was skipped as unnecessary for this pass; a deeper pass should still mine them for primaries this slice missed.

exclusions

  1. Kim JE et al., “Effects of Egg Consumption on Cardiovascular Disease Risk Factors in Adults with Prediabetes and Type 2 Diabetes” (a different Kim/Campbell paper from the one minted as S-19) — out of scope, diabetic study population → slice 3 (heterogeneity-subgroup-dose). Left un-minted per the search plan’s cross-slice rule.
  2. Dawber et al. 1982 (Framingham cohort, within-population correlation of egg intake vs. serum cholesterol) — structural gap, not a rejection: it’s a surrogate-lipid outcome (slice 1’s endpoint) but an observational/cross-sectional design (not controlled-feeding/RCT, which slice 1’s brief requires), so it doesn’t cleanly fit any of the 3 slices as scoped. Left un-minted; flagged here rather than forced in.
  3. In-scope but cut for budget only (verified as real, findable primaries, not minted): Ginsberg 1995 ATVB (healthy young women, dose-response — both LDL and HDL rise modestly, a nice symmetric companion to S-10); Harman/Leeds/Griffin 2008 Eur J Nutr (parallel-group, energy-restricted diet, 2 eggs/day, LDL does not rise); Greene et al. 2006 Nutr Metab (Fernandez lab, elderly, egg intake positively influences LDL/HDL particle characteristics). Any of these could swap in for a curated slot if step 2 wants another angle.

Slice paragraph

The mechanism-surrogate pool splits cleanly 3/4 rise-vs-null-or-favorable across 7 primaries spanning 1965–2018: two purified/dose-response classics (rise, with diminishing returns), one ecological single-egg-addition study (rise), and four whole-egg RCTs supplying three distinct mechanistic explanations for why other studies see no adverse change (poor cholesterol absorption from intact eggs; low-carbohydrate dietary context raising HDL; LDL particle-size/HDL-function reframing) plus one hyperlipidemic/higher-risk-population replication. Egg Nutrition Center (industry) funding is confirmed for 2 of 7 (Mutungi 2008, Njike 2010) and inferred-but-unconfirmed for 2 more (DiMarco 2017 via a same-trial companion paper; Kim & Campbell 2018 via lab pattern) — worth step 2 re-checking full-text funding statements directly rather than relying on my inference. Went looking for and did not get: citation counts for any source (Semantic Scholar rate-limited throughout); a PMID-verified exact page range for the 1965 Keys paper (used the DOI instead, high confidence on venue/year/finding, lower confidence on exact pagination). No genotype/responder-classification papers here by design — that’s slice 3’s territory even though the same labs (Fernandez) publish both kinds of studies.

Graph View