Non-statin LDL-lowering therapies reduce cardiovascular events proportionally to the LDL reduction achieved
Reaffirmed
Staleness check. No subclaim changes. Re-affirmed supported (confidence 0.85, credence 0.85, marginal yield 0.1). Both argument evaluations recorded. Importance stays at 0.6, contestation 0.35. No dependents notified because nothing changed materially. No new search because the budget was limited.
Reassessed: still Supported
verdict confidence 0.88 → 0.85 · credence 0.85
Reassessed (status unchanged) and brought argument evaluations current
Trigger: the supporting subclaim "Adding ezetimibe to statin therapy reduces cardiovascular events" moved to verified (0.82) on IMPROVE-IT. This was the last of the three support pillars not yet formally assessed, so the change is confirmatory in the "for" direction and was already anticipated. Status holds at supported: the three supporting premises (statin benchmark, PCSK9 inhibitors, ezetimibe) are now all verified and give a consistent per-unit slope for LDL-receptor-mediated agents, but the canonical wording quantifies over all non-statin LDL-lowering therapies without restriction, and the verified counterexample subclaim falsifies that strict universal reading. Nudged confidence to 0.88 and recorded credence ~0.85; refreshed the reader-facing prose and reasoning trace. The substantive gap this pass was that both named arguments carried stale evaluations, one still asserting ezetimibe was "not yet formally assessed"; both re-evaluated against current premise standing (both holds_with_caveats). Importance left at 0.6 (major, moderately live cardiology thesis). No dependent notification: status and direction unchanged, purely confirmatory.
Reassessed: still Supported
verdict confidence 0.88 · credence 0.85
Reassessed: still Supported
verdict confidence 0.86 → 0.88 · credence 0.86
Reassessed
Triggered by the contradicting subclaim "Some interventions that lower LDL cholesterol failed to reduce cardiovascular events" moving to VERIFIED (0.95). Judged the change confirmatory, not status-moving: the subclaim is a neutral existence statement whose own reasoning notes the failures are explained by off-target harms (torcetrapib) or insufficient on-target lowering (niacin, earlier CETP inhibitors), with anacetrapib later reducing events. This confirms, rather than undermines, the existing basis for holding the claim at "supported" (literal all-therapies scope falsified by counterexamples) rather than "verified" (receptor-mediated proportionality well backed by IMPROVE-IT, FOURIER, ODYSSEY OUTCOMES, and the statin benchmark). Ran one external evidence check confirming a broadly consistent ~0.85 RR per 1 mmol/L across statins/ezetimibe/PCSK9, with absolute benefit scaling to baseline risk. Held status supported, nudged confidence 0.85->0.86, credence held 0.85. Filled the two missing named-argument evaluations (both holds_with_caveats) and brought the reasoning current with the anacetrapib reconciliation. Confirmed importance at 0.6 (Major) with contestation 0.4. No canonical-form change: narrowing to "receptor-mediated" would be a different claim (individuation), left to the broad debated wording. No dependent notification: status unchanged and the change is not material upstream.
Reassessed: still Supported
verdict confidence 0.85 → 0.86 · credence 0.85
Reassessed
Triggered by the PCSK9-inhibitor subclaim moving to VERIFIED (FOURIER, ODYSSEY OUTCOMES). The change is confirmatory and strengthens the "for" argument (which also rests on the ezetimibe subclaim and the verified statin-benchmark subclaim). It does not flip the status: the reason this claim sits at "supported" rather than "verified" is the scope ambiguity in the literal wording (all non-statin LDL-lowering therapies), which is bounded by the niacin/CETP counterexamples captured in the contradicting subclaim, and strengthening PCSK9 does not resolve that. Kept status supported; nudged confidence 0.83 to 0.85 to reflect that one supporting mechanism is now formally verified; held credence at 0.85. No structural change: the existing decomposition (two supporting mechanisms, statin benchmark, and the failed-intervention counterexample) still captures what the claim turns on. Set importance to 0.6 (major) with contestation 0.3: consequential for lipid-lowering practice but the receptor-mediated core is now largely settled. No dependent notification: my assessment did not materially change (status and credence essentially stable), so downstream stewards need no re-judgment.
Reassessed: still Supported
verdict confidence 0.83 → 0.85 · credence 0.85
Structured and assessed
First pass. Decomposed into two arguments. For: created subclaims on ezetimibe (IMPROVE-IT) and PCSK9 inhibitors (FOURIER/ODYSSEY), and linked the existing statin benchmark claim (d81f89e3) as the reference proportionality slope. Against: created a contradicts subclaim capturing LDL-lowering interventions (niacin, CETP inhibitors) that failed to show proportional benefit. Matcher confirmed the ezetimibe, PCSK9, and counter subclaims were novel; the proposed "meta-analysis shows non-statin proportionality matches statins" framing matched this claim itself, so it was not minted (used as evidence in reasoning instead). Verified external evidence via web search: Silverman 2016 JAMA meta-analysis (similar RR per mmol/L across statin/non-statin receptor-mediated therapies), IMPROVE-IT, CLEAR Outcomes bempedoic acid. Set importance 0.5 (major clinical claim, moderate contestation), contestation 0.35. Assessed SUPPORTED (conf 0.83, credence 0.85): well-established for LDL-receptor-mediated therapies, but literal "all non-statin therapies" reading bounded by niacin/CETP counterexamples, hence supported not verified. This claim already supports the verified broad proportionality claim 97aca08f.
Assessed Supported
verdict confidence 0.83 · credence 0.85
The evidence favours this claim for the therapies that lower LDL cholesterol by increasing LDL-receptor activity, and this is the mainstream view in preventive cardiology. Randomized outcome trials of three distinct non-statin mechanisms point the same way: ezetimibe added to a statin (IMPROVE-IT) produced further LDL lowering and a matching reduction in cardiovascular events; PCSK9 inhibitors on top of statins (FOURIER, ODYSSEY OUTCOMES) achieved large LDL reductions with event reductions tracking the LDL change; and bempedoic acid in statin-intolerant patients (CLEAR Outcomes) reduced major events by about the amount its LDL lowering predicts. A meta-analysis pooling statin and non-statin trials found the relative risk of major vascular events falls by roughly a fifth per 1 mmol/L of LDL reduction regardless of which of these agents achieves it, the same per-unit slope established for statins. The important qualification is mechanistic rather than a genuine rival account. Several interventions that lower LDL by other routes, notably niacin and several CETP inhibitors, did not deliver the expected event reduction. These are generally explained by off-target harms or by the drugs acting largely outside the LDL-receptor pathway, so they bound the claim rather than overturn it: proportionality is robust for therapies that reduce circulating LDL particle number, and should not be assumed automatically for any compound that happens to lower measured LDL. Read as a statement about LDL-receptor-mediated LDL lowering, the proportional relationship is well supported; read as a claim about every non-statin agent that lowers LDL, it is too strong. What would sharpen the assessment further is longer-term and mechanism-diverse outcome data and formal tests of whether the per-unit slope is statistically indistinguishable from the statin benchmark.
Claim entered the graph