Controlled-feeding dose-response study in 20 healthy young men randomised to diets differing in polyunsaturated:saturated (P:S) fat ratio, each tested with dietary cholesterol added via eggs at 0, 750, or 1500 mg/day. On the lower-P:S-ratio (more typical Western) background diet, +750 mg/day cholesterol raised LDL-C by 16±14 mg/dL and +1500 mg/day raised it by 25±19 mg/dL; on a higher-P:S-ratio background diet neither dose produced a significant LDL change. Demonstrates a genuine dose-response of dietary cholesterol on LDL that is nevertheless conditional on background dietary fat quality. relevance_note: Controlled dose-response feeding trial quantifying the average LDL-raising effect of egg-derived cholesterol and showing it is modulated by background fat quality — core mechanism data point for the surrogate-lipid slice.
Correction (2a read): Ginsberg 1994 is a 20-man 8-week crossover feeding 0/1/2/4 eggs/day (128-858 mg cholesterol/d), not a ‘P:S-ratio 0/750/1500 mg’ design.
Step-3 note: On the true design, HDL showed NO significant dose response (contra the node title’s “LDL and HDL both rise”); the effect fell on LDL/total-C and apoB. Extraction below follows the corrected design and abstract (PMID 8148356).
Methodology
Controlled-feeding dose-response study: 20 healthy young men in a randomized four-way crossover, each eating 0, 1, 2, or 4 eggs/day (total dietary cholesterol 128 to 858 mg/day) for 8 weeks per diet, with a break between diets. Fasting and postprandial (postlunch and post-fat-formula) plasma lipids, lipoproteins, and apolipoproteins were measured. NIH/Columbia academic trial. Full text was paywalled (AHA journals 403); extraction rests on the verbatim PubMed/Europe PMC abstract (which carries the key slopes).
Results
O-75 - Fasting total cholesterol rose 1.47 mg-dL per 100 mg dietary cholesterol across a 0-4 egg-day crossover (linear, no threshold)
Corrected/true design (per step-2a note): a 20-man 8-week crossover of 0/1/2/4 eggs (128-858 mg/d), not the ‘0/750/1500 mg P:S-ratio’ design in the original node summary.
Link to originalMethodology
Controlled-feeding dose-response study: 20 healthy young men in a randomized four-way crossover, each eating 0, 1, 2, or 4 eggs/day (total dietary cholesterol 128 to 858 mg/day) for 8 weeks per diet, with a break between diets. Fasting and postprandial (postlunch and post-fat-formula) plasma lipids, lipoproteins, and apolipoproteins were measured. NIH/Columbia academic trial. Full text was paywalled (AHA journals 403); extraction rests on the verbatim PubMed/Europe PMC abstract (which carries the key slopes).
Link to original
O-76 - Fasting LDL cholesterol rose 1.38 mg-dL per 100 mg dietary cholesterol (P-0.001)
Link to original
O-77 - No significant dose response of HDL cholesterol or triglycerides to dietary cholesterol
Corrects the step-2 node title (‘LDL and HDL both rise’): HDL did NOT show a significant dose response in this trial. Because HDL was flat while LDL rose, the total/HDL and LDL/HDL ratios worsened with added dietary cholesterol.
Link to original
O-78 - Fasting apolipoprotein B rose ~10% between 0- and 4-egg diets, correlated with total-LDL cholesterol changes
apoB tracks the number of LDL/VLDL particles (roughly one apoB per particle).
Link to original
O-79 - No difference in postprandial lipid responses across dietary-cholesterol levels
Egg cholesterol shifted the fasting/steady-state LDL level up but did not add a postprandial-lipemia excursion.
Link to original
O-80 - Individual cholesterol responsiveness varied widely, appeared normally distributed, and was mostly positive; ApoE4 only a non-significant trend
The spread of responses in this cohort is continuous rather than obviously bimodal.
Link to original
Interpretation
H-34 - Egg-derived dietary cholesterol raises fasting LDL-total cholesterol linearly, roughly +1.4-1.5 mg-dL per 100 mg
The quantified average causal effect from controlled feeding. Modest per 100 mg but monotonic; magnitude in other populations (older, hyper-responders, different background fat) is left open.
Link to original
H-35 - Cholesterol responsiveness to dietary cholesterol is a continuous (normally distributed) trait, not a discrete responder-non-responder split
Reading of the observed normal spread of responses. Uncertain: n=20 has limited power to distinguish a unimodal from a subtly bimodal distribution, and the ApoE4 trend hints at genetic structure underneath the spread.
Link to original
Arguments
A-27 - The parallel ~10% apoB rise means added egg cholesterol raises atherogenic LDL-particle number, not just cholesterol per particle
Each LDL (and VLDL/IDL) particle carries exactly one apolipoprotein B molecule, so plasma apoB concentration is a direct count of atherogenic particle number, whereas LDL-cholesterol can rise either because there are more particles or because each particle is more cholesterol-enriched. Those two routes differ in atherogenicity: particle number (apoB) is the more consistent predictor of cardiovascular risk, since it is the particle count entering and being retained in the arterial wall that drives atherogenesis. In this trial LDL-cholesterol and apoB rose together (apoB +~10% from 0 to 4 eggs, correlated with the cholesterol changes), so the dietary-cholesterol effect is carried by an increase in particle number, not by cholesterol-enrichment of a fixed particle pool. This upgrades the bare LDL-cholesterol slope into a claim that egg cholesterol adds genuinely atherogenic particles, strengthening the dose-response hypothesis’s clinical relevance. The step is not automatic — LDL-C can move without apoB — which is exactly why the observed apoB co-movement is informative rather than trivial.
Validity (step 6)
status: approved — reason_if_not_false: checked. Traced the step; it is near-definitional given standard biochemistry. Premise: each apoB-lipoprotein particle (VLDL/IDL/LDL) carries exactly one apoB molecule, so plasma apoB concentration IS a count of atherogenic particles. Therefore a measured ~10% apoB rise entails a ~10% rise in particle number — the “raises particle number” clause follows directly. The “not merely cholesterol per particle” clause also holds arithmetically: cholesterol-per-particle ∝ LDL-C/apoB, and since apoB rose ~in step with LDL-C, that ratio is roughly unchanged, so the LDL-C rise is carried by particle number rather than enrichment. Probed for undercutting defeaters conditional on premises: (i) the apoB rise could reflect VLDL/IDL rather than LDL specifically — but the conclusion is about apoB-particle number broadly, and the reported correlation with LDL-C changes ties it to LDL; either way particle number rose. (ii) “atherogenically meaningful rather than cosmetic” is a mild evaluative addendum, but it rests on the stated premise that particle number (apoB) is the more consistent CVD predictor, which is granted here (its truth is priced downstream, not as validity). No defeater breaks the reason→conclusion link. Approved as stated.
Link to original