Mendelian randomization studies show genetically lower LDL cholesterol causally reduces cardiovascular disease risk
7 events · 3 assessments · 3 decisions
Reassessed (status unchanged), backfilled argument evaluation, confirmed importance
Triggered by subclaim_change: the genetic-association leg (db81a342) was assessed verified (0.93/0.97). This confirms a premise already load-bearing in the prior verdict, so no destabilizing change. Kept status verified; nudged confidence 0.92→0.93 now that both the association leg and the RCT-convergence leg (97aca08f) are verified nodes rather than inferred solely from external evidence. Updated the reader-facing assessment and reasoning trace accordingly. Supplied the missing evaluation for the sole named argument ("Genetic instrument evidence for causal LDL effect": holds), noting it lives on the no-pleiotropy assumption (05c19437, supported). Recorded importance 0.4 / contestation 0.15: consequential in cardiovascular medicine but settled consensus, not a live crux. No structural changes. Did not notify the single dependent ("Elevated LDL cholesterol causes cardiovascular disease", already verified): this pass introduced no material change to propagate, only coherence confirmation of an existing verdict.
Reassessed: still Verified
verdict confidence 0.92 → 0.93 · credence 0.97
Reassessed (status unchanged)
Triggered by the RCT corroboration subclaim ("Randomized trials show lowering LDL reduces cardiovascular events proportionally") moving from unassessed to verified. That corroboration was already load-bearing in the prior verdict, so the change is confirmatory and absorbed without a status change (§22). Status remains verified (confidence 0.92, credence 0.97); folded the now-verified RCT leg into the reasoning trace. No structural change: the claim's decomposition (association leg, no-pleiotropy presupposition, RCT convergence under a single 'for' argument) still captures what the claim turns on. Importance left at 0.4: a consequential clinical/genetic finding but a mature, convergent, largely uncontested consensus, so effort is bounded. No dependent notification, since this is a strengthening of an already-verified claim with no material change to propagate.
Reassessed: still Verified
verdict confidence 0.92 · credence 0.97
Structured and assessed
First pass. Decomposed into two new subclaims under a single 'for' argument: the empirical MR association (variants lowering LDL track with proportionally lower CVD risk, requires, importance 0.3) and the load-bearing methodological presupposition that the instruments act on CVD only via LDL and not confounding pleiotropy (presupposes, importance 0.4, the genuine MR crux). Linked the existing RCT-proportionality claim (97aca08f) as a supporting corroboration edge; Matcher confirmed both new subclaims novel and that the RCT claim is distinct (different evidence methodology). Wrote the argument's form. Set importance 0.4: consequential pillar of the verified 'LDL causes CVD' claim but scientifically settled with little credible dissent. Assessed verified (confidence 0.92, credence 0.97) after an adversarial pass on pleiotropy: multi-gene concordance (HMGCR/NPC1L1/PCSK9/LDLR giving equal per-unit effects) and drug-trial convergence defeat the main threat to MR inference. Confirmed against MR literature via web search (Ference reviews, 2x2 factorial NPC1L1/HMGCR MR).
Assessed Verified
verdict confidence 0.92 · credence 0.97
Mendelian randomization uses genetic variants that lower LDL cholesterol as natural, lifelong "instruments": because genotype is fixed at conception and precedes disease, associations between these variants and later cardiovascular outcomes are largely free of the confounding and reverse causation that bias ordinary observational studies. Across many such studies, variants that lower LDL are associated with proportionally lower risk of coronary heart disease, and the genetic evidence supports a causal interpretation. The strongest reason to read the genetic association causally is convergence. Independent variants in different genes, each lowering LDL through a distinct biological route (for example HMGCR, the statin target; NPC1L1, the ezetimibe target; PCSK9; and LDLR), yield approximately the same reduction in coronary risk per unit of LDL lowered, and the effect is log-linear in cumulative exposure. Pleiotropy, the main threat to this kind of inference, would be expected to differ from gene to gene; the concordance across unrelated genes argues against it. The genetic estimates also align with, and prospectively anticipated, the results of randomized trials of LDL-lowering drugs acting through those same targets, an independent evidence stream pointing the same way. The finding is a settled consensus in the specialist literature rather than a live scientific dispute. What would unsettle it would be genetic instruments shown to lower coronary risk through pathways other than LDL, or a reversal of the genetically predicted association, neither of which the current evidence shows.
Claim entered the graph