Randomized trials show lowering LDL cholesterol reduces cardiovascular events proportionally to the reduction achieved
7 events · 3 assessments · 3 decisions
Corrected assessment record
The first update_claim_assessment call was submitted with a placeholder reasoning_trace ("duplicate") by mistake. Re-submitted with the full audit reasoning trace; assessment text, status (verified), confidence (0.87), and credence (0.9) unchanged. No dependent notification issued: the parent verdict is unchanged and only marginally refined by the reinforcing non-statin subclaim, so no dependent could reasonably need to re-judge.
Reassessed: still Verified
verdict confidence 0.87 · credence 0.90
Reassessed
Triggered by the non-statin LDL-lowering subclaim moving to SUPPORTED (credence ~0.85) with a mechanism-scope caveat (niacin/CETP inhibitors lowering LDL by non-receptor routes did not show proportional benefit). This is a reinforcing refinement, not a reversal: it strengthens the causal/LDL-not-drug-class reading while sharpening the boundary condition. The change is not material to the parent verdict, which remains VERIFIED. I refreshed the reader-facing assessment and reasoning trace to fold in the sharpened mechanism-scope caveat and to name the non-statin evidence (ezetimibe, PCSK9 inhibitors, bempedoic acid) explicitly. Confidence nudged 0.86 to 0.87 as the non-statin evidence firmed; credence held at 0.9. Importance left at 0.55 (major, consequential in CV medicine, but the core dose-response relationship is settled among mainstream sources; live contestation is largely at the margins of magnitude and low-risk absolute benefit). No structural change: the existing decomposition (statin per-mmol/L benchmark + non-statin proportionality) still captures what the claim turns on. No dependent notification: verdict unchanged and only marginally refined.
Reassessed: still Verified
verdict confidence 0.86 → 0.87 · credence 0.90
Structured and assessed
First pass. Decomposed into two novel supporting subclaims (confirmed novel by Matcher): the CTT statin meta-analysis dose-response (~a fifth per 1 mmol/L) and the corroborating non-statin (ezetimibe/PCSK9) proportional benefit, both scored 0.4. Kept it lean as one coherent RCT-evidence line, so no named argument. Set importance 0.55 (major band): high consequence for lipid guidelines, moderate contestability from a statin-skeptic minority, below the broader "Elevated LDL causes CVD" claim (0.8). Assessed verified (confidence 0.86, credence 0.9) on direct reading of the CTT Lancet meta-analyses plus IMPROVE-IT/FOURIER/ODYSSEY; ran an adversarial pass on the skeptic critique (extrapolation, per-trial scatter, small low-risk absolute benefit), which qualifies magnitude but does not unseat the proportional meta-analytic finding.
Assessed Verified
verdict confidence 0.86 · credence 0.90
Pooled analyses of randomized lipid-lowering trials show that the reduction in major vascular events tracks the absolute reduction in LDL cholesterol achieved, and this proportional dose-response is among the best-replicated findings in cardiovascular medicine. The Cholesterol Treatment Trialists' Collaboration, drawing on individual-participant data from more than two dozen statin trials and roughly 170,000 participants, found that each 1 mmol/L (about 39 mg/dL) reduction in LDL cholesterol lowers the annual rate of major vascular events by roughly a fifth (about 20 to 22 percent), with more intensive lowering yielding correspondingly larger benefits. The relationship generalizes beyond statins: trials of ezetimibe (IMPROVE-IT) and PCSK9 inhibitors (FOURIER, ODYSSEY OUTCOMES) added to statin therapy produce incremental benefit broadly consistent with the same per-unit relationship, which supports the interpretation that the benefit follows from LDL lowering itself rather than from any drug-specific effect. A persistent minority of clinicians and researchers disputes not the existence of benefit but its size and reach, arguing that the pooled per-mmol/L estimates extrapolate beyond the LDL reductions actually achieved in trials, that individual trials show scatter that departs from a clean proportional line, and that the absolute benefit is small in low-risk primary prevention. These are genuine caveats about magnitude, individual-trial variability, and the primary- versus secondary-prevention distinction, but they do not overturn the central meta-analytic finding of a proportional relationship, which remains well supported. What would most sharpen the picture is further outcome data at very low achieved LDL levels and in lower-risk populations.
Claim entered the graph