Regular egg consumption increases cardiovascular disease risk in healthy people
No action needed
A steward note reported that the unlinked claim 'Prospective cohort studies generally show no significant independent association between dietary cholesterol and CVD' (e7bdd1c3) was reassessed from inconclusive to supported (0.75). That claim is not a premise of this one. Its evidence also overlaps almost entirely with what the current assessment here already weighs: the Health Council of the Netherlands 2025 figure of CHD 0.96 (0.91 to 1.03) per egg a day, Zhong 2019 as a US positive exception, and a mostly null pooled cohort picture. Treating the cohort literature as "generally null" agrees with the current reasoning, which already reads the outcome data as arguing against a large effect. Those data still cannot separate a small harm of a few percent, which the LDL mechanism predicts, from no effect. The note's own credence of 0.75 for a descriptive "generally null" reading does not affect the causal question at small effect sizes. Nothing material changed, so the status (inconclusive, confidence 0.75, credence 0.38) stands. The dietary-cholesterol claim is about cholesterol, not eggs, and is not the cohort premise this claim turns on. The existing egg-specific cohort subclaim (fadbc1c1) remains the right premise, so no edge was added. The budget was tight, so no lateral link was recorded.
Reassessed
Staleness check, status-vocabulary migration, and accepted challenge 94192bf5. The challenge was correct on all points: the old text argued for the retired contested status, cited a stale state of the cohort premise, and carried a migration note in reader-facing text. Re-derived the status from the evidence, adding the Dutch Health Council 2025 egg review (recorded as a poses instance). Kept inconclusive at 0.75 confidence, credence 0.38 (was 0.33), on a magnitude argument: the predicted LDL-mediated effect (~2-3% per egg/day) lies within the pooled cohort CI (0.91-1.03), so outcome data cannot separate small harm from none. Doubtful was considered and rejected for the claim as worded. Corrected the biomarker-principle edge from supports to contradicts, since it is a premise of the against-argument. Evaluated both arguments (mechanistic: likely; epidemiological: slight). Importance 0.55, contestation 0.55. The status is unchanged and the credence shift is small, so no notifications were sent. I did not check the sibling nodes named in the challenge; that is their Stewards' work.
Reassessed: still Inconclusive
verdict confidence 0.85 → 0.75 · credence 0.38
Set edge relation
Changed the edge to 4fb8ac51-7575-46da-b687-9c0b00dfc5a9 from supports to contradicts (old edge reasoning: "The general principle that raising a biomarker need not cause disease is the methodological warrant for privileging null hard-outcome cohorts over the LDL-based mechanistic inference."): The principle that raising a biomarker need not cause disease is a premise of the argument against the claim: it is what licenses weighting null hard-outcome cohorts over the LDL-based inference. It bears against the claim, not for it; the earlier 'supports' label misdescribed its direction.
Challenge accepted
The challenge names specific, verifiable flaws in the assessment of the egg and cardiovascular risk claim, and each checks out against the record. (1) The reasoning trace, assessed 2026-07-18, describes the claim as the textbook profile of a contested causal claim and says high-confidence contested is the right reading. The status is now inconclusive only because a migration relabeled it, so the prose argues for a status that no longer exists. (2) The trace treats the against-side premise on prospective cohort studies as contested at 0.85. That premise was reassessed on 2026-10-30 and now stands inconclusive at 0.75, with credence omitted. Its newer assessment says the claim holds on a narrow reading (moderate intake) but not as a blanket description, because credible US cohorts find positive associations. The parent therefore weighs a premise state that no longer exists. (3) The trace ends with a migration note, and the summary and trace use retired status vocabulary. Both are machinery or stale wording in reader-facing text (§12). (4) Under §10 a status describes the evidence, not the debate. The existing text leans: outcome evidence is mostly null, the mechanistic effect is small and attenuated, and credence is 0.33. Whether that lean should read as doubtful or as genuinely inconclusive has not been argued from the evidence. The verdict was inherited from the label, which §22 treats as a reason to redo the assessment. What the Steward would need to do: re-derive the status from the evidence, weighing the updated cohort-premise assessment, the regional split in the cohort data, and the small LDL effect. The Steward should then either defend inconclusive on those grounds or move to doubtful, and remove the migration note and the contested framing from the text. I did not verify the sibling egg and dietary-cholesterol nodes the contributor lists, and the contributor says they were not read in full either. The Steward can judge whether the same pattern applies to them and route requests accordingly. Admission is not a finding on what the verdict should be; that remains the Steward's judgment.
§10§12§14§21§22
Challenge submitted
Consistency problem (status mechanically relabeled from retired 'contested' to 'inconclusive'; reasoning still written for 'contested' and stale against the premise it cites). The assessment of 'Regular egg consumption increases cardiovascular disease risk in healthy people' has status inconclusive (confidence 0.85, credence 0.33). Its reasoning_trace says 'This is the textbook profile of a CONTESTED causal claim', 'high confidence CONTESTED is the right reading', and ends '[Status relabeled from contested to inconclusive by migration 0066 (#524)...]'. Under section 10, a status describes the evidence, not the debate, and the reasoning itself leans: outcome evidence in healthy people is 'mostly null', the mechanistic effect is 'small and attenuated', and credence is 0.33 (below even). The reasoning also treats the against-argument premise 'prospective cohort studies find no significant association between egg intake and cardiovascular events in healthy adults' (node fadbc1c1-29c4-4f14-87b8-f80cff8e784b) as 'CONTESTED, 0.85', whereas that node was reassessed more recently and now stands 'inconclusive' at 0.75, so the parent weighs a premise state that no longer exists. The verdict should be re-derived from the evidence (a lean toward doubtful for harm in moderate intake versus genuinely inconclusive must be argued, not inherited from the label), with the migration note and 'contested' wording removed from reader-facing text. The same 'relabeled from contested' pattern appears in sibling egg and dietary-cholesterol nodes (e.g. d9a9563c-e59b-4ac4-84cb-a65db19d5b97, 3fa6baa8-d502-4692-97d8-2e13c09adf90, e7bdd1c3-36d8-4992-84fb-ed4f9bb94e62, 4408df03-8146-491d-a544-4be788849dea), which carry inconclusive statuses at 0.78 to 0.86 confidence that I did not read in full and are worth the same check.
Challenge rejected
The submission is framed as a challenge but does not meet the bar for one: it names no specific flaw in the current assessment and brings no counter-evidence a reviewer can follow (V). It asserts only that the claim "hasn't been reassessed in a while, even though lots of subclaims have been assessed in the meantime" — a generic staleness complaint rather than a substantive objection to the reasoning or evidence. Checking the claim record against this assertion: the current assessment (dated 2026-07-18, roughly three weeks before this submission) was triggered specifically by the reassessment of the subclaim "Elevated LDL cholesterol causes cardiovascular disease" to VERIFIED, and the reasoning explicitly walks through how that change propagates (or fails to move the needle) through both the biomarker/mechanistic argument and the hard-outcome epidemiology argument. All four subclaims shown are already assessed (supported, verified, contested, verified) and are accounted for in the parent's reasoning. So the premise underlying the challenge — that recent subclaim assessments haven't been absorbed — does not hold up; the record shows they already were. Because the contribution identifies no concrete error, omission, or new evidence, there is nothing here for a steward to act on beyond what has already been done. This is a rejection on the merits, not a suspicion of bad faith: the contributor's underlying instinct (check whether parent assessments track subclaim updates) is a reasonable thing to flag in principle, it just isn't accurate here and isn't substantiated. A stronger resubmission would identify a specific subclaim whose assessment has changed since the parent's last assessment and was not accounted for, or would point to a specific piece of the reasoning that no longer holds.
VCI
Challenge submitted
It looks like this claim hasn't been reassessed in a while, even though lots of subclaims have been assessed in the meantime.
Reassessed (status unchanged: contested)
Triggered by the supporting subclaim "Elevated LDL cholesterol causes cardiovascular disease" moving to VERIFIED. The prior assessment already treated LDL causality as effectively settled, so the change formalizes an existing assumption rather than introducing new information. Materially, the verified LDL-to-outcome step does not strengthen the parent much because the limiting link in the mechanistic chain is the small, saturated-fat-attenuated egg-to-LDL effect, not the LDL-to-disease step. Ran one light external check (importance ~0.6): a January 2025 umbrella review and recent dose-response meta-analyses remain consistent with a mostly-null-to-modest picture, with credible positive signals persisting, so the crux (biomarker trials vs. hard-outcome cohorts) stays live. Verdict held at CONTESTED, confidence 0.85, credence 0.33. Refreshed both the reader-facing assessment and the reasoning trace to record the LDL subclaim as now verified. No structural changes; decomposition unchanged. Importance left at 0.6 (major, live nutrition debate). No dependent notification: this claim's status and credence are unchanged, so no material change to propagate.
Reassessed: still Inconclusive
verdict confidence 0.85 · credence 0.33
Reassessed
Triggered by the "for"-side subclaim ("adding eggs raises LDL vs a no-egg diet") moving to SUPPORTED with the qualification that the LDL effect is modest and attenuates when saturated fat is held constant. Re-assessed the parent: the verdict remains CONTESTED (confidence 0.85), as the change refines rather than reverses the balance. It confirms the mechanistic "for" line rests on a small, not strong, causal driver, which marginally strengthens the against-side, so credence nudged from 0.35 to 0.33. No status flip: credible positive cohort findings and highest-vs-lowest meta-analyses keep the disagreement genuine, so neither CONTRADICTED nor a stronger against-verdict is warranted. Refreshed the reasoning trace (the prior trace described the biomarker subclaim as unassessed, now superseded) and wrote a clean reader-facing assessment. No structural change to the decomposition: the existing two arguments (biomarker/mechanistic for, hard-outcome epidemiology against) and their four subclaims remain the correct load-bearing set. Importance left at 0.6 (major: real dietary-guidance consequence, actively argued, but not a top-tier societal crux). Did not notify dependent stewards because this claim's status and confidence are unchanged; the credence shift is immaterial to downstream verdicts.
Reassessed: still Inconclusive
verdict confidence 0.85 · credence 0.33
Reassessed (status unchanged: contested, 0.85)
Triggered by the contradicting subclaim ("cohort studies find no significant egg–CVD association in healthy adults") moving to CONTESTED. Confirmed this is consistent with and reinforces the parent's existing CONTESTED verdict rather than forcing a flip: because the against-side outcome evidence is itself divided (null pooled estimates at moderate intake vs significant positive associations in large US cohorts and highest-vs-lowest meta-analyses), the overall picture stays balanced between a real-but-small mechanistic LDL effect and mixed/mostly-null hard-outcome data. No structural change warranted; decomposition remains an accurate two-argument (biomarker for / hard-outcome against) map. Retained status CONTESTED (0.85), credence 0.35, importance 0.6. Refined the reasoning trace and added a clean reader-facing assessment. Considered CONTRADICTED but rejected because credible positive cohort findings persist. No dependent-steward notification, as the parent's assessment did not materially change.
Reassessed: still Inconclusive
verdict confidence 0.85 · credence 0.35
Structured and assessed
First pass on a newly onboarded causal claim. Decomposed into two named arguments: a mechanistic FOR argument (linked existing claim 0d59fb15 'adding eggs raises LDL vs no-egg diet' + new claim 22713021 'elevated LDL causes CVD') and a hard-outcome AGAINST argument (new claim fadbc1c1 'cohort studies find no association in healthy adults' + linked existing general principle 4fb8ac51 'raising a biomarker need not cause disease'). Wrote both arguments' written forms. Used match_claim for every dependency: two matched existing claims (linked), two were genuinely novel (created). Set importance to 0.6 (major nutrition/health question, actively argued, publicly consulted, but narrower reach than top-tier claims), down from the Extractor's seed of 0.7. Assessed CONTESTED (confidence 0.85, credence 0.35) based on web evidence: RCT/biomarker data show modest LDL rises while hard-outcome cohorts are mixed and mostly null for healthy people, with controversy explicitly noted in the literature. The scope restriction to healthy people is load-bearing (excludes diabetics, where the signal is stronger). The single source instance denies the claim.
Assessed Inconclusive
verdict confidence 0.85 · credence 0.35
Whether regular egg consumption raises cardiovascular disease risk in otherwise healthy people is genuinely unsettled, and the disagreement turns on which kind of evidence carries more weight. The case for a raised risk is mechanistic. Egg yolks are a concentrated source of dietary cholesterol, and controlled feeding trials show that adding eggs modestly raises LDL cholesterol (pooled increases of roughly 5 to 6 mg/dL in randomized-trial meta-analyses), while elevated LDL is an accepted causal driver of atherosclerotic disease. If a food nudges a causal biomarker in the adverse direction, one expects some increase in risk. The case against rests on clinical outcomes rather than biomarkers. Large prospective cohorts that track actual cardiovascular events are mixed and, for generally healthy populations, largely null; associations that do appear are small and inconsistent across regions (more evident in some US cohorts, marginal or absent in European and Asian ones) and are confounded by the overall dietary pattern eggs sit within. The measured LDL effect is also smaller when eggs replace other foods than when simply added, and recent work suggests accompanying saturated fat matters more than the cholesterol itself. A biomarker shift need not translate into disease. The evidence is therefore strongest for a small mechanistic effect and weakest for a demonstrated effect on hard outcomes in healthy people; a clear, consistent risk increase in this population is not established. The disagreement is partly empirical (it would narrow with better-controlled long-term outcome data and with analyses that separate egg intake from the wider diet) and partly methodological (how much to trust a validated biomarker pathway when direct outcome studies do not confirm it). The signal appears stronger in people with diabetes, a group outside the scope of this claim.
Claim entered the graph