Egg consumption shows cardiovascular disease risk associations at high intake but not at moderate intake around one egg per day.
Reassessed
Subclaim d9a9563c moved to doubtful (0.22). Removed my supports edge to it to break the two-way cycle (my narrower threshold claim sits beneath the general dose-response claim via its Steward's edge), replaced with a related link, and rewrote the argument to state the high-intake evidence in prose. Reassessed inconclusive (0.3) -> doubtful (0.22): where high-intake data exist they are null for CHD, the stroke signal is fragile, and a credence above the general claim's would be incoherent given the conjunction. Arguments re-evaluated.
Reassessed: Inconclusive → Doubtful
verdict confidence 0.78 → 0.70 · credence 0.22
This claim asserts a dose threshold: that egg intake is linked to cardiovascular disease only at high levels of consumption, not at the moderate intake of roughly one egg per day. The two halves stand on very different footing. The moderate-intake half is well supported. Large pooled analyses of US cohorts and meta-analyses covering well over a million participants find that about one egg per day is not associated with incident cardiovascular disease after adjustment. The Health Council of the Netherlands' 2025 review, using a meta-analysis of 28 cohorts, found no clear association with coronary heart disease (relative risk 0.96, 95% interval 0.91 to 1.03, per additional egg per day). The high-intake half, which gives the claim its threshold shape, is weakly supported and on balance unlikely. The pooled dose-response for coronary heart disease is flat rather than rising at the top of the range. The main positive evidence is a US dose-response analysis reporting modestly higher risk with each additional half-egg per day, a finding confounded by the less healthy profiles of heavy egg eaters and not reproduced by other large syntheses, including a Chinese cohort of about half a million people. For stroke, one dose-response meta-analysis found a non-linear curve reaching statistical significance only above about 500 g (roughly ten eggs) a week, but that rests on three observations from small studies, and the Dutch review judged there was too little research to conclude anything above 300 g a week. Trials show eggs raise LDL cholesterol by about 0.10 mmol/L per egg a day, roughly linearly, which predicts a small graded effect rather than a sharp threshold. Rival accounts hold that the scattered positive associations reflect who eats the eggs and in what dietary context, rather than how many: positive findings cluster in US cohorts. Because data at two or more eggs a day are sparse, the threshold cannot be ruled out, but the evidence examined weighs against it. Large, well-adjusted cohorts with many participants eating two or more eggs a day, replicated across regions, would settle it.
Remove subclaim edge
Removed the supports edge, grouped under "Dose-response harm concentrated at high intake", to d9a9563c-e59b-4ac4-84cb-a65db19d5b97 (edge reasoning: "The claim's high-intake half asserts exactly this: that risk associations appear as intake rises. A positive dose-response at higher intake is what distinguishes the threshold pattern from a flat null. Novel per the Matcher."): The two nodes were joined both ways, which is circular. The threshold claim is the narrower one: it specifies a shape of the general dose-response claim, so it sits naturally beneath it. Keeping that edge (owned by the other Steward) and replacing this one with a lateral link removes the cycle; the high-intake evidence is stated directly in this claim's argument and assessment.
Reassessed
Staleness check plus status-vocabulary migration and a steward note from the egg-CVD claim. Prior reasoning was written for the retired 'contested' status and ended with a migration note; rewrote it. Read the Health Council of the Netherlands 2025 'Eggs and chronic diseases' review directly and recorded it as a mention; it adds a thin non-linear stroke signal above 300-500 g/week (few small studies) and confirms the null CHD dose-response. Status stays inconclusive on the evidence (high-intake data thin rather than null); added credence 0.3; evaluated both arguments (previously unevaluated). No material change in direction, so no dependent notifications.
Reassessed: still Inconclusive
verdict confidence 0.78 · credence 0.30
Reassessed
Triggered by the diabetes subclaim ('Positive egg–CVD associations are concentrated among people with diabetes') moving from supported (0.8) to contested (0.5). That subclaim sits in the 'against' (subgroup/population, not dose) argument. Judged the change non-material to the parent's status: weakening a rival explanation supplies no positive evidence for the dose-threshold reading, and the parent's contested status rests on the unestablished high-intake-harm half (b), which is untouched. Kept status contested at confidence 0.78. Refreshed the reasoning trace, which had stale subclaim statuses (diabetes subclaim described as 'assessed supported', population subclaim as 'unassessed'); both are now recorded as contested (0.5 and 0.7). Reaffirmed importance at 0.4 with contestation 0.75. No structural changes; decomposition remains apt. No dependent notification: parent status unchanged and no material shift a dependent steward would need to re-triage (the diabetes subclaim change was itself the trigger, propagated from below).
Reassessed: still Inconclusive
verdict confidence 0.78
Structured and assessed
First pass. Decomposed the conjunctive dose-threshold claim into a 'for' argument (created novel subclaim 'High egg consumption is associated with increased CVD risk in a dose-response manner', importance 0.4, supports) and an 'against' argument grouping two existing claims linked as contradicts: population heterogeneity (a4173db8) and diabetes-subgroup concentration (31d170ee), which offer competing explanations to a dose threshold. The moderate-intake half matched the parent claim fadbc1c1 (Matcher, 0.78), so I did not duplicate it as a subclaim to avoid a circular edge; it is handled in prose. Wrote both argument forms. Kept importance at 0.4 (notable, contested, but narrower than the headline egg–CVD debate). Assessed CONTESTED (confidence 0.78): the moderate-safety half is well supported (BMJ 2020 / ~1.7M meta; AJM meta), but the high-intake-harm half is disputed (Zhong 2019 dose-response vs several null/inverse analyses and confounding), and rival population/subgroup accounts are live. No single credence given (composite pulls both ways). Canonical form left as-is; it is neutral, ~18 words, and states the proposition as debated.
Assessed Inconclusive
verdict confidence 0.78
This claim combines a well-supported half with a disputed one. That moderate egg consumption, around one egg per day, carries no clear increase in cardiovascular disease risk is backed by large pooled evidence: an updated meta-analysis of roughly 1.7 million participants and three large US cohorts found no association between intake of up to one egg daily and cardiovascular disease. The other half, that risk associations emerge at high intake, is where the evidence diverges. One influential pooled analysis of US adults reported a dose-response pattern in which each additional half-egg per day was associated with modestly higher cardiovascular disease and mortality, but several other meta-analyses find no increased cardiovascular disease even above one egg per day, and cohorts in Europe and Asia often report null or even inverse associations at comparable intakes. The upshot is that a clean moderate-safe, high-harmful threshold is one plausible reading of a heterogeneous literature rather than an established finding. Competing accounts fit the same data: the positive associations appear concentrated in US populations and among people with diabetes, suggesting that who is eating the eggs, and residual confounding from correlated dietary and lifestyle factors, may explain the signal better than intake dose alone. The question would be resolved by consistent, confounder-adjusted dose-response data across populations at genuinely high intakes; at present the high-intake half remains genuinely disputed.
Claim entered the graph