High egg consumption is associated with increased cardiovascular disease risk in a dose-response manner.
Reassessed
Staleness check, retired 'contested' label, and a steward note pointing to the Health Council NL 2025 review. Rewrote the assessment from scratch, removing the reasoning written for 'contested' and the migration note. Read the HCN report directly (CHD RR 0.96, 0.91 to 1.03, per egg/day across 28 cohorts) and recorded it as a 'poses' instance. Status moved from inconclusive to doubtful (credence 0.33 to 0.22): the near-null pooled per-egg estimate with a tight interval weighs against a general dose-response, though a small LDL-mediated effect cannot be excluded. Evaluated both arguments (for: slight; against: likely). Set contestation to 0.5. No dependents found to notify via get_claim. Diabetic-subclaim edge left as is.
Reassessed: Inconclusive → Doubtful
verdict confidence 0.85 → 0.72 · credence 0.22
The claim that higher egg intake tracks a graded rise in cardiovascular risk has a plausible mechanism behind it but little support in the pooled observational record. On the mechanistic side, adding eggs to the diet raises LDL cholesterol (randomised trials put the rise at roughly 0.1 mmol/L per egg a day), and elevated LDL causes cardiovascular disease. On a linear LDL model, an increase that size implies a coronary risk excess of only about 2 to 3 percent per daily egg, which is too small for cohorts to detect reliably. Pooled US cohorts also find a dose-response for dietary cholesterol, and one pooling of six US cohorts reported a modest graded association for eggs themselves. The broader pooled evidence does not show a gradient. The 2020 meta-analysis of 28 cohorts in Europe, the United States and Asia (about 1.4 million participants and nearly 60,000 coronary events) found a relative risk of 0.96 (95% CI 0.91 to 1.03) per additional egg a day, and 0.97 for highest against lowest intake. Stroke meta-analyses are null or inverse, and the inverse results are strongest in Asian cohorts. Positive associations appear mainly in some US cohorts and at intakes above about one egg a day, where data are sparse. That pattern fits regional heterogeneity and confounding by the foods and lifestyles that come with eggs better than a universal dose-response. The Health Council of the Netherlands' 2025 review called the coronary evidence inconclusive and said too little research exists on stroke above 300 g a week. Its pooled estimate still sits close to null, with an upper bound that rules out much more than a 3 percent excess per egg. Overall the evidence weighs against a general dose-response association, but it cannot rule out a small effect of the size the LDL pathway predicts, or a real positive gradient at high intake or in US populations. A graded association that held up across well-adjusted cohorts in several regions, or showed clearly at intakes above two eggs a day, would revive the claim.
Reassessed (status unchanged)
Subclaim_change trigger: the diabetic-concentration subclaim moved supported→contested (Drouin-Chartier 2020/2021 US pooling finds no CHD association even in prevalent T2D; diet-quality confounding leading explanation; older subgroup meta-analyses still affirm). Assessed materiality as low. That subclaim is one scope-restriction leg of the 'against' argument; weakening it makes the field murkier but supplies no graded monotonic association, which is the parent's least-supported component. Parent remains contested (confidence 0.85, credence 0.33) — the verdict rested on the threshold/subgroup-vs-gradient distinction, which is intact. Refreshed both reader-facing assessment and reasoning trace to reflect the eroding diabetic-subgroup explanation. Structure unchanged (no missing dependency; do not re-decompose). Importance left at 0.45: a well-known but moderately consequential nutrition debate, single local dependent (itself a subclaim). No dependent notification: parent status did not change, so no material propagation.
Reassessed: still Inconclusive
verdict confidence 0.85 · credence 0.33
Reassessed (status unchanged: contested)
Triggered by subclaim a4173db8 (regional US-positive / Europe-Asia-null pattern) moving from supported to contested, since the two leading meta-analyses (BMJ 2020 incidence vs Circulation 2022 mortality) disagree on the specific regional shape. Judged materiality as low for the parent's status: the specific regional characterization weakens, but regional heterogeneity and the absence of a clean universal positive dose-response remain intact, so the change is absorbed rather than flipping the parent. If anything the meta-analytic disagreement reinforces the contested verdict. Confirmed the current evidentiary spread with a web search (Godos 2020 inverse at moderate intake; Drouin-Chartier/BMJ 2020 null RR≈0.98; positive signals only at high intake or in diabetic subgroups), which supports keeping status=contested. Refined the reader-facing assessment and reasoning trace to soften the now-disputed 'US positive / Europe & Asia null' summary and to foreground that the dose-response (monotonic gradient) component is the weakest part of the claim. Kept confidence 0.85 and set credence 0.33. Left importance at 0.45 (a live public-nutrition debate but niche in global claimspace). No structural changes: did not re-decompose; existing subclaims and arguments remain apt. Only local dependent is one of the claim's own subclaims (4408df03), and parent status did not change, so no dependent notification warranted.
Reassessed: still Inconclusive
verdict confidence 0.85 · credence 0.33
Structured and assessed
First pass. Decomposed into two lines of reasoning using existing graph claims (no new nodes needed, reflecting a mature neighborhood). FOR: pooled-cohort dose-response for dietary cholesterol plus the eggs-raise-LDL / LDL-causes-CVD mechanistic chain. AGAINST: null association in healthy adults, US-vs-Europe/Asia heterogeneity, concentration among diabetics, and a threshold (risk at high but not moderate intake) rather than monotonic gradient. Set importance to 0.45 (contested, publicly relevant but a narrow dose-response framing). Assessed CONTESTED, confidence 0.85, credence 0.35: credible evidence on both sides, but the specific dose-response framing is the weaker component; a universal monotonic gradient is not well established, a subgroup/context-dependent association is more defensible. Search budget for external verification was exhausted this run, but the graph's neighboring claims (independently assessed contested/supported) and established literature strongly corroborate the contested reading.
Assessed Inconclusive
verdict confidence 0.85 · credence 0.35
Whether high egg consumption is associated with cardiovascular disease in a graded, dose-response manner remains genuinely unsettled, and the evidence divides along recognizable lines. Some large pooled analyses of United States cohorts report that each additional increment of egg or dietary cholesterol intake tracks with modestly higher cardiovascular risk, consistent with a dose-response pattern, and a plausible mechanism supports this: eggs raise LDL cholesterol, and elevated LDL causes cardiovascular disease. Against a general dose-response stand several strong observations. Large meta-analyses and cohorts find no significant association in healthy adults, and the positive signal is not universal: it appears in US populations but is null or even inverse in European and Asian cohorts, and is concentrated among people with diabetes. Where an association does emerge, it tends to appear at high intake rather than rising smoothly from moderate intake, pointing to a threshold rather than a clean monotonic gradient. The disagreement is largely empirical and turns on residual confounding (egg intake correlates with other dietary and lifestyle risk factors), population and dietary-context differences, and how intake is modeled. It would be narrowed by better-adjusted analyses that separate eggs from accompanying foods and by consistent findings across diverse populations. As stated, a universal dose-response association is weakly supported at best; a subgroup- and context-dependent association is the more defensible reading.
Claim entered the graph