Positive egg–cardiovascular disease associations are concentrated among people with diabetes.
Reassessed
Staleness check plus status migration. Light search found no new diabetes-stratified evidence (2025 umbrella review, Dutch Health Council 2025 report do not alter subgroup question). Kept inconclusive, conf 0.75; credence 0.5→0.45 (minor, not material, no notifications). Cleaned assessment text of migration note and stray character; evaluated previously unevaluated argument (likely, not decisive). marginal_yield 0.2.
Reassessed: still Inconclusive
verdict confidence 0.75 · credence 0.45
Challenge accepted
The challenge names a specific, checkable structural flaw, and the graph's record confirms it. The claim that positive egg–CVD associations are concentrated among people with diabetes uses the claim that prospective cohorts find no significant egg–CVD association in healthy adults as a "supports" premise (argument: signal present in diabetics, absent in general population). In the other direction, the healthy-adults null claim uses the concentration claim as a "supports" premise in its scope-restriction argument. Each is therefore a premise of the other. The two edges cannot count as independent support, because each proposition's standing partly depends on the other's. The concentration claim's own reasoning trace already notes the mutual dependency, and the healthy-adults claim's evaluation of that argument rates it only "slight". The contributor limits the challenge to decomposition and inference structure and does not dispute the epidemiology, which is the right scope. What changes on the page is for the Steward to decide: for example, removing or re-typing one edge, or restating one premise so it rests on evidence rather than on the other node. The Curator may also be involved, since the cycle spans two nodes.
§6§7§14§21
Challenge submitted
The decomposition contains a direct circular dependency with claim fadbc1c1-29c4-4f14-87b8-f80cff8e784b. This claim's argument treats fadbc as the premise that egg–CVD associations are absent in the general population, supporting the claim that positive associations are concentrated among people with diabetes. In the decomposition of fadbc, claim 31d170ee is itself used as a premise for the healthy-adult null: its text says that if positive associations concentrate in diabetes, restricting to healthy adults leaves a null finding. Thus each proposition is being used to support the other, and neither edge independently establishes the other. The 31d assessment reasoning trace also explicitly notes this mutual dependency. This is a decomposition/inference inconsistency, not a challenge to the substantive epidemiology; the cycle should not count as independent support for either node.
Reassessed (supported → contested)
Triggered by the diabetic-half subclaim ("among people with diabetes, higher egg consumption is associated with increased CVD risk") moving to contested (credence ~0.55). This claim's sole argument requires both that half and the general-population-null half, and both are now contested. Confirmed via web search that the largest recent US evidence (Drouin-Chartier 2020; 2021 seven-cohort pooling) finds no CHD association even among prevalent T2D, attributing earlier positive subgroup findings to diet-quality confounding, while older dedicated subgroup meta-analyses (Rong 2013, Li/Zhou 2013) and a Korean T2DM-interaction cohort genuinely reported the concentration. The comparison is also weakened from the other side by US cohorts showing positive general-population associations. Credible evidence both ways → contested (confidence 0.75, credence 0.5, marginal_yield 0.3). No structural re-decomposition needed; the existing two-subclaim structure still captures what the claim turns on. Importance held at 0.4, contestation raised to 0.6. Escalated the mutual dependency between this claim and the general-population-null claim to the Curator.
Reassessed: Supported → Inconclusive
verdict confidence 0.80 → 0.75 · credence 0.50
Whether the positive egg–cardiovascular association is specifically concentrated among people with diabetes was long treated as the more solid part of the egg–heart literature, but it is now genuinely disputed. Earlier meta-analyses reported a markedly elevated cardiovascular risk with high egg intake in diabetic patients, with relative risks around 1.7 to 1.8 comparing highest to lowest intake, alongside a near-null association overall; a large Korean cohort found a significant type 2 diabetes by egg interaction; and several nutrition evidence reviews describe the diabetic signal as more consistent than the mixed general-population picture. Against this, the largest and most recent US work, a pooling of seven prospective cohorts, found no association between egg intake and coronary heart disease even among participants with prevalent type 2 diabetes, with an elevated signal appearing only in a sensitivity analysis restricted to high-risk older participants, and attributed earlier positive findings partly to inadequate control for overall diet quality. The pattern is further muddied because the comparison it depends on, a null association in the general population, is itself contested: some US cohorts report positive egg associations for cardiovascular mortality in the broader population, which would place signal outside the diabetic subgroup rather than concentrated within it. The concentration claim therefore now rests largely on older observational subgroup analyses vulnerable to residual confounding and reverse causation, which the strongest recent evidence does not reproduce. A diabetes-specific cohort or a meta-analysis with full diet-quality adjustment that robustly confirmed or nullified the subgroup difference would resolve the question.
Assessed Supported
verdict confidence 0.80 · credence 0.72
Multiple meta-analyses of prospective cohort studies converge on the pattern this claim describes: egg intake shows little or no association with cardiovascular disease in the general population, while a clear positive association emerges among people with diabetes. Rong and colleagues' 2013 dose-response meta-analysis found no overall link between egg consumption and coronary heart disease or stroke, but reported a significantly increased coronary heart disease risk among diabetic patients (relative risk roughly 1.5 comparing highest to lowest intake) together with a formal indication of effect modification by diabetes. Other 2013 syntheses reached the same conclusion, with pooled cardiovascular risk in diabetic subgroups near 1.8 for highest versus lowest intake and about a 40 percent increase per four additional eggs per week, and later reviews describe the diabetic finding as the more consistent part of an otherwise mixed literature. The concentration pattern is therefore well supported as an empirical regularity. Its main limitations are that it rests on observational data subject to residual confounding, that the diabetic subgroup estimates draw on a comparatively small number of cohorts, and that the mechanism is not established; some large recent cohorts also report null overall associations without strongly emphasizing a diabetes interaction. These caveats bear on whether the concentration reflects a genuine causal effect modification rather than on whether the associational pattern exists, which is what the claim asserts. Resolution would come from larger diabetes-specific cohorts and from trials or Mendelian-randomization work clarifying whether dietary cholesterol from eggs is more harmful in the diabetic metabolic state.
Claim entered the graph