References

Heart attack, angina and rhythm

What the /heart-attack/ topic rests on: the definitions, the UK and US acute coronary syndrome guidelines, and the trials that moved the evidence ahead of one or both of them.

1. Fourth Universal Definition of Myocardial Infarction (2018)

   Thygesen K et al.; ESC/ACC/AHA/WHF; Eur Heart J 2019;40:237–269 · 2018

   SupportsThe spectrum chapter's definitions: infarction as a troponin rise and fall with clinical evidence of ischaemia; types 1 to 5; MINOCA as infarction with no artery over 50% narrowed; the distinction between myocardial injury and infarction that the clock chapter uses to say a raised troponin is not the same as a heart attack.

   Read it asThe operative definition in both the UK and the US in September 2026.

   Strong

2. 2025 ACC/AHA/ACEP/NAEMSP/SCAI Guideline for the Management of Patients With Acute Coronary Syndromes

   Rao SV et al.; Circulation 2025;151:e771–e862 and JACC 2025; with the JACC Guideline-at-a-Glance and the AHA Top Things to Know · February 2025

   SupportsThe US column throughout: DAPT loaded on arrival with prasugrel or ticagrelor preferred with PCI; DAPT 12 months, shortened to 1–3 months of monotherapy where bleeding risk is high with aspirin the agent dropped; radial access preferred; intracoronary imaging for complex lesions; complete revascularisation in STEMI and NSTE-ACS; culprit-only PCI in cardiogenic shock; the microaxial flow pump "reasonable" in selected AMI shock with attention to bleeding, limb ischaemia and renal failure; high-intensity statin for all with concurrent ezetimibe as an option; a non-statin agent recommended at LDL ≥1.8 mmol/L and reasonable at 1.4–1.8; no routine oxygen with saturations ≥90%; no routine GP IIb/IIIa pretreatment; invasive approach during the admission for intermediate- and high-risk NSTE-ACS; fasting lipids at 4–8 weeks; cardiac rehabilitation referral.

   Read it asReplaces the 2013 STEMI and 2014 NSTE-ACS guidelines and the 2016 DAPT update. Read from the at-a-glance, the AHA summary and two secondary reports; the class labels in the topic should be checked against the full text. Literature search closed April 2024, so it predates the 2025 beta-blocker trials.

   Strong

3. NICE NG185 — Acute coronary syndromes; NICE DG40 — high-sensitivity troponin for early rule-out; NICE CG95 — chest pain of recent onset

   National Institute for Health and Care Excellence; NG185 November 2020; DG40 August 2020; CG95 2010, updated 2016 · 2010–2020

   SupportsThe UK column: primary PCI within 120 minutes of the time fibrinolysis could have been given, for presentation within 12 hours; fibrinolysis otherwise with angiography within 24 hours; radial access; complete revascularisation for multivessel STEMI without shock during the admission; GRACE-based risk for NSTEMI with angiography within 72 hours for intermediate risk or above and immediately for instability; fondaparinux unless angiography within 24 hours or high bleeding risk; ticagrelor or prasugrel with PCI, clopidogrel with high bleeding risk or anticoagulation; DAPT 12 months; beta-blocker 12 months with preserved EF and indefinitely with reduced; ACE inhibitor indefinitely; echocardiography for LV function; cardiac rehabilitation. DG40: assay-specific early rule-out with a second sample at 1–3 hours. CG95: CT coronary angiography first line for stable chest pain.

   Read it asWritten from memory of NG185 and should be checked against the current version before sign-off, including whether NICE has amended any of it since 2020. The beta-blocker recommendation is the one the 2025 evidence has overtaken; the topic says so.

   Strong

4. 2023 ESC Guidelines for the management of acute coronary syndromes

   Byrne RA, Rossello X, Coughlan JJ et al.; Eur Heart J 2023;44:3720–3826 · 2023

   SupportsThe European positions cited where NICE is silent: the 0/1-hour high-sensitivity troponin algorithm; the LDL target below 1.4 mmol/L; beta-blockers "may be considered" with preserved EF; the recognition that women and older patients receive guideline therapy less often.

   Read it asCited at summary level; the topic's UK column is NICE.

   Strong

5. Beta-blockers after myocardial infarction with preserved or mildly reduced ejection fraction: REDUCE-AMI, ABYSS, REBOOT, BETAMI–DANBLOCK, and the individual-patient meta-analysis

   Yndigegn T et al., NEJM 2024;390:1372–1381 (REDUCE-AMI); Silvain J et al., NEJM 2024;391:1277–1286 (ABYSS); Ibáñez B et al., NEJM 2025 (REBOOT); Munkhaugen J et al., NEJM 2025 (BETAMI–DANBLOCK); Rossello X et al., NEJM, published 9 November 2025 (NEJMoa2512686, the pooled analysis); Kristensen AMD et al., Lancet 2025 (the mildly reduced EF pooled analysis) · 2024–2025

   SupportsThe clock chapter's beta-blocker note: REDUCE-AMI (EF ≥50%) and REBOOT (EF >40%) found no reduction in death, reinfarction or heart-failure admission; BETAMI–DANBLOCK found a composite benefit driven by fewer reinfarctions (5.0% v 6.7%); the pooled analysis of all four plus CAPITAL-RCT, 17,801 patients, found no benefit with EF ≥50% and a separate pooled analysis found benefit with EF 40–49%. ABYSS tested stopping a beta-blocker years after the event and could not show non-inferiority.

   Read it asVerified from the NEJM abstracts and two editorial commentaries this round. The trials were open-label and pragmatic with imperfect adherence, which the commentaries note. The topic's position is: continue with EF below 50%; a conversation, not a self-stop, with a preserved pump. NG185 (12 months for all) is behind this evidence; ESC 2023 and AHA/ACC 2025 are "may be considered".

   Strong

6. Atrial fibrillation: NICE NG196 (2021), the 2024 ESC/EACTS guideline, and the 2023 ACC/AHA/ACCP/HRS guideline

   NICE NG196, April 2021; Van Gelder IC, Rienstra M et al., Eur Heart J 2024;45:3314–3414; Joglar JA et al., Circulation 2024;149:e1–e156 · 2021–2024

   SupportsThe rhythm chapter's AF section and the clinical AF table. NG196: CHA2DS2-VASc ≥2 offer anticoagulation, 1 in men consider; ORBIT for bleeding; DOAC first line; no aspirin; rate control first; rhythm control when indicated; ablation after drug failure; LAA occlusion if anticoagulation contraindicated. ESC 2024 (verified): CHA2DS2-VA with the same cut-off regardless of sex; catheter ablation upgraded to a first-line option in paroxysmal AF; cardioversion window reduced from 48 to 24 h with a wait-and-see approach; comorbidity control as the C of AF-CARE; guidance on device-detected AF; surgical LAA exclusion at cardiac surgery as an adjunct. AHA 2023: the AF stages; annual-risk threshold; early rhythm control and first-line ablation reasonable.

   Read it asNG196 from memory; check the recommendation text. ESC 2024 verified from the guideline spotlight paper and the ACC key-points summary.

   Strong

7. AF and arrhythmia trials cited in the rhythm chapter: EAST-AFNET 4, CASTLE-AF, the alcohol-abstinence trial, NOAH-AFNET 6 and ARTESIA, REVERT, the ICD trials

   Kirchhof P et al., NEJM 2020;383:1305–1316; Marrouche NF et al., NEJM 2018;378:417–427; Voskoboinik A et al., NEJM 2020;382:20–28; Kirchhof P et al., NEJM 2023;389:1167–1179; Healey JS et al., NEJM 2024;390:107–119; Appelboam A et al., Lancet 2015;386:1747–1753; MADIT-II, SCD-HeFT, AVID, DINAMIT, IRIS · 1997–2024

   SupportsEarly rhythm control within a year reduced the composite of cardiovascular death, stroke and hospitalisation (EAST-AFNET 4); ablation in HFrEF with AF reduced death and HF admission (CASTLE-AF); abstinence roughly halved AF recurrence in regular drinkers; anticoagulating device-detected AF gave a small stroke reduction and more bleeding (NOAH, ARTESIA); the modified Valsalva terminated 43% of SVT episodes against 17%; the ICD primary-prevention criteria (EF ≤35%, ≥40 days post-MI, ≥90 days post-revascularisation) and the negative early-implant trials (DINAMIT, IRIS).

   Read it asAll figures from memory; check each. The ablation success and complication rates quoted (70–80% paroxysmal; 2–4% serious complications) are literature ranges.

   Strong

8. The unusual ones: InterTAK takotsubo consensus and registry; the AHA 2018 SCAD scientific statement and the DISCO registry; MINOCA statements; the ESC 2025 myocarditis and pericarditis guideline; COPE and CORP; RHAPSODY

   Ghadri JR et al., Eur Heart J 2018;39:2032–2046 and Templin C et al., NEJM 2015;373:929–938; Hayes SN et al., Circulation 2018;137:e523–e557; Cerrato E et al., Eur Heart J 2021;42:3161–3171 (DISCO); Tamis-Holland JE et al., Circulation 2019;139:e891–e908 (MINOCA); ESC 2025 guideline on myocarditis and pericarditis; Imazio M et al., Circulation 2005 (COPE) and Ann Intern Med 2011 (CORP); Klein AL et al., NEJM 2021;384:31–41 (RHAPSODY) · 2005–2025

   SupportsThe unusual-ones chapter: takotsubo criteria, nine in ten women, in-hospital death about 4%, recurrence about 1–2% a year, ACE inhibitors associated with fewer recurrences and beta-blockers not, phaeochromocytoma excluded; SCAD conservative management with >90% healing, up to a quarter of MIs in women under 50, FMD in about half, recurrence ~10% at three years, beta-blocker associated with fewer recurrences, DAPT associated with worse outcomes in DISCO; MINOCA definition and CMR; myocarditis exercise restriction 3–6 months and the vaccine statement; pericarditis colchicine halving recurrence (COPE, CORP) and IL-1 blockade for refractory cases (RHAPSODY).

   Read it asRegistry and consensus figures from memory; check each before sign-off. Graded Moderate because none of these conditions has trial evidence for its drug treatment except pericarditis.

   Moderate

9. Secondary prevention after MI: cardiac rehabilitation (Cochrane 2021 and NACR), NICE NG238 lipid targets, IAMI, CORDIOPREV, SELECT, COLCOT / LoDoCo2 / CLEAR SYNERGY, and NICE CG71 on familial hypercholesterolaemia

   Dibben G et al., Cochrane 2021; National Audit of Cardiac Rehabilitation annual report; NICE NG238 (2023); Fröbert O et al., Circulation 2021;144:1476–1484 (IAMI); Delgado-Lista J et al., Lancet 2022;399:1876–1885 (CORDIOPREV); Lincoff AM et al., NEJM 2023;389:2221–2232 (SELECT); Tardif JC et al., NEJM 2019; Nidorf SM et al., NEJM 2020; Jolly SS et al., NEJM 2025 (CLEAR SYNERGY); NICE CG71 (2008, updated 2019) · 2008–2025

   SupportsThe after-hospital and every-test chapters: rehabilitation reduces cardiovascular mortality and readmission and UK uptake is about half; the NG238 target LDL ≤2.0 / non-HDL ≤2.6; influenza vaccination after MI reduced death (IAMI); a Mediterranean diet beat a low-fat diet for events in established coronary disease (CORDIOPREV); semaglutide reduced MACE in obesity with cardiovascular disease (SELECT); low-dose colchicine positive in COLCOT and LoDoCo2 and neutral in CLEAR SYNERGY; FH criteria, genetic and cascade testing (CG71); lipoprotein(a) once (HEART UK; AHA 2018).

   Read it asThe rehabilitation effect is stated as "around a quarter", the site's rounding of the Cochrane range; figures from memory; check. The DVLA table in the clinical chapter is from the gov.uk cardiovascular chapter as read in September 2026.

   Strong

10. NICE TA314 — implantable cardioverter defibrillators and cardiac resynchronisation therapy

    National Institute for Health and Care Excellence, technology appraisal 314; June 2014 · 2014

    SupportsThe UK device criteria in the rhythm and clinical chapters and on the checklist: ICD for primary prevention with LVEF ≤35% on optimal medical therapy, assessed at least 4 weeks after a myocardial infarction and 3 months after revascularisation; secondary prevention after VF or haemodynamically compromising VT without a treatable cause; CRT by QRS duration, morphology and NYHA class.

    Read it asStill current in September 2026. The 40-day and 90-day figures in the US column are the AHA/ACC criteria from the same trials.

    Strong

11. DanGer Shock — microaxial flow pump in infarct-related cardiogenic shock; and IABP-SHOCK II

    Møller JE et al., NEJM 2024;390:1382–1393; Thiele H et al., NEJM 2012;367:1287–1296 · 2012, 2024

    SupportsThe cardiogenic shock section: DanGer Shock, 360 patients with STEMI and shock, death at 180 days 45.8% v 58.5% with the microaxial pump, with more major bleeding, limb ischaemia and renal replacement therapy; IABP-SHOCK II, 600 patients, no mortality benefit from the intra-aortic balloon pump, the basis for both guidelines not recommending it routinely.

    Read it asDanGer Shock figures from memory; check. The topic states UK availability as centre-dependent, which is observation rather than guidance.

    Strong

Aspirin for prevention

What /heart/aspirin/ rests on: the secondary-prevention trials, the three 2018 primary-prevention trials, the guidelines, and the specific uses outside the heart.

1. Secondary prevention: ISIS-2 and the Antithrombotic Trialists' Collaboration

   ISIS-2 Collaborative Group, Lancet 1988;2:349–360; Antithrombotic Trialists' Collaboration, Lancet 2009;373:1849–1860 · 1988, 2009

   SupportsISIS-2: 17,187 patients with suspected MI, aspirin 160 mg daily; five-week vascular mortality 9.4% v 11.8%, about 24 lives per 1,000, without a significant rise in intracranial or transfusion-requiring bleeds. ATT 2009: in secondary prevention, serious vascular events about 6.7% v 8.2% a year.

   Read it asFigures as confirmed in the owner's September 2026 fact-check of a UK emergency doctor's video, which also corrected the video's implication that ISIS-2 used a chewed 300 mg dose.

   Strong

2. Primary prevention: ASPREE, ASCEND and ARRIVE (2018)

   McNeil JJ et al., NEJM 2018;379:1499–1508, 1509–1518, 1519–1528 (ASPREE); ASCEND Study Collaborative Group, NEJM 2018;379:1529–1539; Gaziano JM et al., Lancet 2018;392:1036–1046 (ARRIVE) · 2018

   SupportsASPREE: about 19,000, mostly 70 and over, 100 mg for about 4.7 years; disability-free survival unchanged, major bleeding 2.8% to 3.8%, an unexplained excess of deaths mainly from cancer. ASCEND: about 15,500 with diabetes; serious vascular events 9.6% to 8.5%, major bleeding 3.2% to 4.1%, mostly non-fatal GI. ARRIVE: about 12,500 at moderate risk; no significant benefit, more mostly mild GI bleeding.

   Read it asVerified in the owner's fact-check. The ASPREE all-cause mortality figures (5.9% v 5.2%) are from memory and not quoted on the page.

   Strong

3. Guidelines on aspirin for primary prevention: USPSTF 2022, NICE, ACC/AHA 2019

   US Preventive Services Task Force, JAMA 2022;327:1577–1584; NICE NG238 (and CG181 before it); Arnett DK et al., 2019 ACC/AHA primary prevention guideline, Circulation 2019;140:e596–e646 · 2019–2023

   SupportsUSPSTF: individual decision at 40–59 with 10-year CVD risk of 10% or more; recommends against initiating at 60 or over; the colorectal-cancer benefit dropped from the recommendation. NICE: aspirin not recommended for primary prevention. ACC/AHA: may be considered in selected adults 40–70 at higher risk without raised bleeding risk; not routinely over 70.

   Read it asThe USPSTF position is as confirmed in the fact-check; the ACC/AHA wording and the NICE position are from memory and should be checked.

   Strong

4. The coronary calcium exception: MESA-based analyses of aspirin by CAC score

   Miedema MD et al., Circ Cardiovasc Qual Outcomes 2014;7:453–460; Cainzos-Achirica M et al., Circulation 2020;141:1541–1553 · 2014, 2020

   SupportsEstimated net benefit from aspirin in people with a CAC of 100 or more and low bleeding risk, and net harm below that; the basis for the CAC chapter's "aspirin discussion".

   Read it asObservational modelling, not a randomised trial; from memory, check. The page says so.

   Moderate

5. Aspirin outside the heart: pre-eclampsia (ASPRE, NICE NG133) and Lynch syndrome (CAPP2, NICE NG151)

   Rolnik DL et al., NEJM 2017;377:613–622 (ASPRE); NICE NG133 (2019); Burn J et al., Lancet 2011 and 2020 (CAPP2); NICE NG151 (2020) · 2011–2020

   SupportsNICE NG133: 75–150 mg daily from 12 weeks until birth for women at high risk of pre-eclampsia or with two or more moderate risk factors; ASPRE: 150 mg reduced preterm pre-eclampsia from 4.3% to 1.6%. NICE NG151: consider daily aspirin for at least two years in Lynch syndrome to reduce colorectal cancer risk.

   Read it asFrom memory; check the recommendation numbers and wording.

   Strong

6. Stopping aspirin, dose-tailoring, Reye's syndrome, and the lifestyle comparison

   Sundström J et al., Circulation 2017;136:1183–1192; Price MJ et al., JAMA 2011 (GRAVITAS); Collet JP et al., NEJM 2012 (ARCTIC); MHRA/NHS advice on aspirin in under-16s; Estruch R et al., NEJM 2018 (PREDIMED, republished); Ettehad D et al., Lancet 2016; Diabetes Prevention Program, NEJM 2002 · 2002–2018

   SupportsStopping long-term low-dose aspirin was associated with about a third more cardiovascular events (Swedish register); platelet-function-guided dosing showed no clear benefit; aspirin is not given to under-16s except on specialist advice because of Reye's syndrome; PREDIMED's ~30% relative reduction was about 1–1.5 percentage points absolute over ~5 years and the trial was retracted and republished in 2018; about 20% fewer major cardiovascular events per 10 mmHg systolic reduction; DPP's 58% was for incident diabetes, not cardiovascular events.

   Read it asEttehad, DPP and the PREDIMED framing are as confirmed in the fact-check; Sundström, GRAVITAS and ARCTIC are from memory; check.

   Moderate

Pulmonary embolism

What the /pe/ topic rests on. Two guidelines a year apart that disagree in the middle of the severity range, the trials that caused it, and the follow-up positions NG158 leaves to European guidance.

1. NICE NG158 — Venous thromboembolic diseases: diagnosis, management and thrombophilia testing

   National Institute for Health and Care Excellence; published March 2020, updated August 2023, reviewed unchanged May 2026 · 2020–2026

   SupportsEvery UK position in the topic, by recommendation number: PERC (1.1.16), two-level Wells and the 4-point cut (1.1.17), CTPA or V/Q with the CrCl-under-30 and radiation criteria (1.1.18), the four-hour rules and interim anticoagulation (1.1.18, 1.1.21, 1.3.2–1.3.4), age-adjusted D-dimer over 50 (1.1.14), outpatient care with the three mandatory pieces of information (1.2), apixaban or rivaroxaban first line (1.3.8), the renal, cancer, weight and antiphospholipid rules (1.3.11–1.3.20), no thrombolysis in stable PE (1.6.3), the filter rules (1.7), no compression stockings for prevention (1.7.5), the three-month review and annual review (1.4), cancer investigation limited to history, examination and baseline bloods (1.8), and the thrombophilia testing rules (1.9).

   Read it asRead in full from the NCBI Bookshelf copy for this topic. The guideline excludes pregnancy, does not address catheter therapies except by cross-reference to IPG524, and says nothing about CTEPH follow-up; those gaps are labelled as such wherever the topic fills them from the ESC or AHA documents.

   Strong

2. 2026 AHA/ACC/ACCP/ACEP/CHEST/SCAI/SHM/SIR/SVM/SVN Guideline for the Evaluation and Management of Acute Pulmonary Embolism in Adults

   Creager MA, Barnes GD et al.; Circulation 2026;153(12):e977–e1051, and JACC; with the JACC Guideline-at-a-Glance (2026;87:1620–1625) · March 2026

   SupportsThe A–E clinical categories and the respiratory modifier; the severity-score thresholds (PESI ≤85, sPESI 0, Bova ≤4 for low); the D1/D2 definitions of transient hypotension and normotensive shock; DOAC over VKA for A–B; LMWH over UFH for C–E except where VA-ECMO is contemplated; systemic lysis for E2, all reperfusion options equal for D and E1, catheter therapy "may be considered" for C3 and not recommended for C1–C2; PERT recommended; the 3–6-month follow-up.

   Read it asThe category-C description, the D2 perfusion markers and the treatment-by-category statements were taken from the guideline text and the CHEST Physician summaries; category A is described in the topic as incidental or asymptomatic on the basis of the at-a-glance table and should be checked against the full text before sign-off. Written before HI-PEITHO reported; its authors have said since that HI-PEITHO is the first hard evidence for the C3 recommendation.

   Strong

3. ESC 2019 Guidelines for the diagnosis and management of acute pulmonary embolism

   Konstantinides SV et al.; Eur Heart J 2020;41:543–603 · 2019

   SupportsThe low / intermediate-low / intermediate-high / high tiers used in UK practice, the RV/LV thresholds, the sPESI items, and the structured 3–6-month follow-up with echocardiography if symptomatic and V/Q if the echo suggests pulmonary hypertension.

   Read it asStill the operative European document in September 2026. Predates every catheter trial in this group.

   Strong

4. PEITHO — fibrinolysis for patients with intermediate-risk pulmonary embolism

   Meyer G et al.; N Engl J Med 2014;370:1402–1411 · 2014

   SupportsWhy neither guideline gives full-dose systemic thrombolysis to stable patients: 1,006 intermediate-high-risk patients, tenecteplase against placebo; death or haemodynamic decompensation at 7 days 2.6% v 5.6%, decompensation alone 1.6% v 5.0%; haemorrhagic stroke 2.0% v 0.2%; major extracranial bleeding 6.3% v 1.2%; no difference in death.

   Read it asThe bleeding penalty is the whole reason the field moved to catheter delivery of smaller doses. The 2017 long-term follow-up found no difference in mortality or residual RV dysfunction.

   Strong

5. HI-PEITHO — ultrasound-facilitated, catheter-directed fibrinolysis for acute pulmonary embolism

   Rosenfield K, Klok FA, Piazza G et al.; N Engl J Med, published 28 March 2026 (NEJMoa2516567); presented at ACC.26 · March 2026

   SupportsThe first randomised comparison of catheter therapy with anticoagulation alone: 544 intermediate-high-risk patients with additional severity criteria, 59 sites in the US and eight European countries including the UK; 7-day PE-related death, cardiorespiratory decompensation or recurrence 4.0% v 10.3%, a 61% relative reduction driven by decompensation; no intracranial haemorrhage; no significant difference in all-cause death or major bleeding.

   Read it asOpen-label, funded by the device manufacturer (Boston Scientific), and around 87% of screened patients were not randomised. The endpoint is short-term deterioration, not mortality or long-term function. Graded Emerging for those reasons and because no guideline has yet incorporated it.

   Emerging

6. PRAGUE-26 — catheter-directed thrombolysis in intermediate-high-risk pulmonary embolism

   N Engl J Med 2026 (NEJMoa2608012); 11 Czech tertiary centres · September 2026

   Supports558 patients, intermediate-high risk by ESC criteria, conventional catheter-directed alteplase plus anticoagulation against anticoagulation alone; 7-day death, recurrence or decompensation 0.7% v 6.8% (RR 0.10, 95% CI 0.02–0.44); bleeding 4.6% v 5.0%; 7-day deaths 0 v 4; median RV/LV ratio at 24 h 0.94 v 1.07. The ordinary infusion catheter produced the same result as the proprietary ultrasound device in HI-PEITHO.

   Read it asOpen-label with an unblinded events committee, which matters because decompensation criteria involved judgement (ventilation, NEWS ≥9). Both intracranial haemorrhages in the trial occurred in the catheter arm, and the single treated-arm death followed one of them. Too few deaths to say anything about mortality.

   Emerging

7. PEERLESS — large-bore mechanical thrombectomy versus catheter-directed thrombolysis in intermediate-risk pulmonary embolism

   Jaber WA, Gonsalves CF, Stortecky S et al.; Circulation 2025;151:260–273 · 2025

   Supports550 patients randomised between the two catheter approaches; similar safety, and a hierarchical composite favouring thrombectomy on ICU use and deterioration. Cited in the topic only to say the two catheter techniques look comparable; it has no anticoagulation-only arm and cannot support either over standard care.

   Read it asIndustry-funded (Inari). PEERLESS II and PE-TRACT, which compare thrombectomy with anticoagulation alone, were still recruiting in September 2026.

   Emerging

8. Extended-phase reduced-dose anticoagulation: AMPLIFY-EXT and EINSTEIN CHOICE, and the RENOVE trial

   Agnelli G et al., N Engl J Med 2013;368:699–708; Weitz JI et al., N Engl J Med 2017;376:1211–1222; Couturaud F et al., Lancet 2025 · 2013, 2017, 2025

   SupportsThe licensed half doses for extended prevention: apixaban 2.5 mg twice daily matched 5 mg twice daily for recurrence (about 1.7% a year against 8.8% on placebo) with major bleeding under 0.5%; rivaroxaban 10 mg once daily matched 20 mg (1.2% v 1.5%) and both beat aspirin (4.4%) with similar bleeding. Both trials enrolled patients in whom the case for continuing was equivocal.

   Read it asThe RENOVE sentence in the after-hospital chapter is written cautiously and must be checked against the paper before sign-off: the topic states only that reduced dose could not be confirmed non-inferior in higher-risk patients and bled less. NG158 does not name an extended dose; it says continue the current anticoagulant if tolerated.

   Strong

9. SOME — screening for occult cancer in unprovoked venous thromboembolism

   Carrier M et al.; N Engl J Med 2015;373:697–704 · 2015

   SupportsThe NG158 position (1.8) and the US position that cancer investigation after unprovoked VTE stops at history, examination, basic bloods and age-appropriate screening: 854 patients randomised to limited screening with or without CT of the abdomen and pelvis; occult cancer found in about 4% within a year in both arms, with no difference in cancers missed, time to diagnosis or cancer-related death.

   Read it asA subsequent meta-analysis of extended screening reached the same conclusion. The topic states the 4% figure and does not claim CT is harmful, only that it finds nothing extra.

   Strong

10. NICE NG89 — Venous thromboembolism in over 16s: reducing the risk of hospital-acquired DVT or PE

    National Institute for Health and Care Excellence; published March 2018, updated August 2019 · 2018–2019

    SupportsThe single line in the topic that a previous VTE must be flagged at any future admission or operation so that pharmacological and mechanical prophylaxis are given, and that hip and knee replacement always carry a post-operative course.

    Read it asCited for that purpose only; the topic does not restate its risk-assessment tool.

    Strong

11. BSH 2022 — Thrombophilia testing: a British Society for Haematology guideline; and the ASH 2023 guideline

    Arachchillage DJ, Mackillop L, Chandratheva A et al.; Br J Haematol 2022;198:443–458; Middeldorp S et al., ASH 2023 guidelines for management of VTE: thrombophilia testing, Blood Adv 2023;7:7101–7138 · 2022, 2023

    SupportsThe every-test chapter: no testing for MTHFR or PAI-1 (SERPINE1) variants; no routine testing of first-degree relatives, with selective testing of relatives of protein C, protein S and antithrombin probands where it would change life choices; the Genomics England monogenic thrombophilia panel criteria (VTE under 40, spontaneous or weakly provoked, in at least one first-degree relative, requested by a haematologist or geneticist, and only where it changes management); the factors that lower measured protein C, S and antithrombin (the guideline's table 1); clotting-based tests four to six weeks after anticoagulation stops, genetic tests at any time; antiphospholipid tests repeated at 12 weeks. ASH 2023 reaches the same positions and is slightly more open to testing after hormone-provoked VTE.

    Read it asThe prevalence and relative-risk ranges in the chapter's tier-three table (factor V Leiden about 1 in 20 of European ancestry, 3–5 fold; prothrombin about 1 in 50, 2–3 fold; antithrombin 1 in 2,000–5,000, 10–20 fold; protein C about 1 in 300–500) are textbook ranges rather than figures from this guideline, and should be checked. The chapter does not recommend any test; it states what each result changes.

    Strong

12. HSIB — clinical decision making: diagnosis and treatment of pulmonary embolism in emergency departments

    Healthcare Safety Investigation Branch (now HSSIB), national investigation report · March 2022

    SupportsThe framing of the recognising-it chapter: that missed and delayed PE diagnoses in English emergency departments follow the pattern of a plausible alternative accepted too early; and the 2023 rewording of the NG158 PERC recommendation, which the report prompted.

    Read it asA safety investigation, not a study; it identifies mechanisms rather than measuring rates.

    Moderate

13. Travel-related venous thrombosis: the WRIGHT project, the MEGA study, the BSH guideline, NaTHNaC and the CHEST position on aspirin

    World Health Organization WRIGHT project, phase I results (June 2007); Cannegieter SC et al., PLoS Med 2006;3:e307 (MEGA); British Society for Haematology guidance on travel-related VTE; NaTHNaC (TravelHealthPro) factsheet on venous thromboembolism; CHEST antithrombotic guideline on long-distance travellers · 2006–2026

    SupportsThe lifestyle chapter's sitting-still section, and the mirrored rows in the clots prevention chapter and the two risk-factor lists. WRIGHT: VTE risk approximately doubles after four hours or more seated and immobile, in a plane, train, bus or car alike; absolute risk about 1 in 6,000; multiple flights raise it; the risk stays raised for about four weeks. MEGA (1,906 patients, 233 who had travelled over four hours in the previous eight weeks): risk about doubled for all modes, highest in the first week; after overland travel the risk was 8.1× with factor V Leiden, 9.9× with BMI over 30, 4.7× over 1.90 m, and over 20× on oral contraceptives; short stature added risk for air travel. BSH and NaTHNaC: measures by risk grade; stockings and single-dose LMWH for high-risk travellers; aspirin is not recommended for travel-related venous thrombosis. CHEST: suggests against aspirin or anticoagulants for ordinary long-distance travellers; stockings for those at increased risk; the Cochrane review (Scurr 2021) on stockings reducing symptomless DVT.

    Read it asThe absolute figures vary between studies by an order of magnitude depending on how DVT was ascertained; the topic gives a range. The desk and gaming ("e-thrombosis") statement rests on case series and is labelled as thinner evidence. Regraded Strong in round 151 once the MEGA and WRIGHT figures were verified from the papers.

    Strong

Blood clots and DVT

What the /clots/ topic rests on beyond NG158 and NG89, which sit in the Pulmonary embolism group: the unusual-site guidance, the superficial thrombosis trial, the catheter and stocking trials, and the cerebral venous thrombosis positions.

1. BCSH 2012 — Guidelines on the investigation and management of venous thrombosis at unusual sites

   Tait C, Baglin T, Watson H et al.; British Committee for Standards in Haematology; Br J Haematol 2012;159:28–38 · 2012

   SupportsThe UK positions on upper-limb DVT (anticoagulate at least 3 months; a needed, functioning catheter may stay), splanchnic vein thrombosis (anticoagulate; test JAK2 and PNH where no local cause), cerebral venous thrombosis (anticoagulate including with haemorrhage), and superficial vein thrombosis: within 3 cm of the saphenofemoral junction treated as DVT; risk factors for extension treated with prophylactic-dose LMWH for 30 days or fondaparinux 2.5 mg for 30–45 days.

   Read it asThe current British guidance in September 2026 despite its age; the BSH has not replaced it. The 3-cm and 5-cm figures in the topic are stated as BCSH positions, and the CHEST 2021 difference (≥5 cm without the 3-cm rule) is noted.

   Strong

2. CALISTO — fondaparinux for superficial-vein thrombosis in the legs; and SURPRISE

   Decousus H et al.; N Engl J Med 2010;363:1222–1232; Beyer-Westendorf J et al., Lancet Haematol 2017;4:e105–e113 (SURPRISE) · 2010, 2017

   Supports3,002 patients with lower-limb SVT of at least 5 cm, not near the junction, randomised to fondaparinux 2.5 mg daily or placebo for 45 days: death, symptomatic PE or DVT, extension to the junction or recurrence 0.9% v 5.9%, no excess bleeding. SURPRISE: rivaroxaban 10 mg daily non-inferior to fondaparinux (3% v 2% for the composite).

   Read it asCALISTO excluded SVT within 3 cm of the junction, recent surgery, prior VTE and cancer, so the trial population was lower-risk than the clinic population; the topic says so through the BCSH risk-factor framing.

   Strong

3. Catheter-directed thrombolysis for DVT: CaVenT and ATTRACT; and compression stockings for post-thrombotic syndrome: SOX

   Enden T et al., Lancet 2012;379:31–38 (CaVenT); Vedantham S et al., N Engl J Med 2017;377:2240–2252 (ATTRACT); Kahn SR et al., Lancet 2014;383:880–888 (SOX) · 2012–2017

   SupportsCaVenT: post-thrombotic syndrome at 2 years 41% v 56% with catheter-directed thrombolysis for iliofemoral DVT. ATTRACT (692 patients, pharmacomechanical CDT): 47% v 48% overall, more major bleeding, less moderate-to-severe PTS in the iliofemoral subgroup. SOX (806 patients): active against placebo stockings for 2 years, PTS identical, the basis for NG158 1.7.5 and the CHEST position.

   Read it asFigures from memory of the papers; check before sign-off. The topic presents CaVenT and ATTRACT as disagreeing and frames CDT as a case-by-case decision under the NG158 1.6.1 criteria.

   Strong

4. Distal DVT: the CACTUS trial and the CHEST position

   Righini M et al.; Lancet Haematol 2016;3:e556–e562; Stevens SM et al., CHEST 2021 antithrombotic guideline · 2016, 2021

   SupportsCACTUS: nadroparin against placebo for isolated calf DVT in low-risk outpatients found no reduction in extension or embolism and more bleeding. CHEST 2021: serial imaging for two weeks in low-risk distal DVT, anticoagulation for higher-risk or symptomatic, 3 months if treated. Basis for the topic saying NICE scans proximal veins only, UK practice varies, and US guidance allows either approach.

   Read it asCACTUS was underpowered (stopped early at 259 of 572 planned) and the topic states its result qualitatively. Graded Moderate for that reason.

   Moderate

5. Cerebral venous thrombosis: ESO 2017 guideline, RE-SPECT CVT, ACTION-CVT and the AHA 2024 scientific statement

   Ferro JM et al., Eur J Neurol 2017;24:1203–1213; Ferro JM et al., JAMA Neurol 2019;76:1457–1465 (RE-SPECT CVT); Yaghi S et al., Stroke 2022;53:728–738 (ACTION-CVT); Saposnik G et al., AHA scientific statement, Stroke 2024 · 2017–2024

   SupportsAnticoagulation with LMWH acutely, including in the presence of intracranial haemorrhage; VKA or DOAC for 3–12 months (dabigatran matched warfarin in RE-SPECT CVT; ACTION-CVT observational comparison of DOACs and warfarin found similar recurrence and less major bleeding); endovascular treatment only for deterioration despite anticoagulation; decompressive surgery for large lesions with mass effect; permanent cessation of oestrogen. Incidence a few per 100,000; headache in about nine in ten; female predominance about 3:1.

   Read it asThe trial and statement details are from memory and should be checked; the presentation figures are approximate. The topic states the "anticoagulate through haemorrhage" position as shared by both guidelines, which it is.

   Strong

Guidelines and national guidance

These are syntheses rather than evidence in themselves, and they disagree with one another in ways that matter. Where the guidelines stand sets out how and why.

1. NICE NG238 — Cardiovascular disease: risk assessment and reduction, including lipid modification

   National Institute for Health and Care Excellence · Published 14 December 2023; last reviewed 2 September 2025

   SupportsUK lipid targets, the QRISK3 treatment threshold of 10%, and the statin intensities quoted throughout the heart topic.

   Read it asThe 2025 review added technology-appraisal links without changing recommendations. This is the document UK practice actually follows; the US and European numbers on this site do not override it.

   Strong

2. Summary of national guidance for lipid management

   NHS England / Accelerated Access Collaborative · Operational summary, updated periodically

   SupportsThe pathway most UK trusts work from in practice.

   Read it asHistorically carried the Joint British Societies targets of 1.8 and 2.5 mmol/L. A local protocol quoting those figures may predate alignment with NG238.

   Strong

3. 2026 ACC/AHA multisociety Guideline on the Management of Dyslipidemia

   Circulation; co-published in JACC 2026;87(19):2624–2757 · Online 13 March 2026

   SupportsPREVENT-ASCVD replacing the Pooled Cohort Equations, restored LDL-C and non-HDL-C goals, universal once-in-a-lifetime Lp(a) measurement, and the CAC-tiered targets.

   Read it asUS risk engines are calibrated on US populations. The targets are not transferable to UK practice, and NICE carries a cost-effectiveness constraint this document does not.

   Strong

4. 2025 Focused Update of the 2019 ESC/EAS Guidelines for the Management of Dyslipidaemias

   European Heart Journal 2025;46(42):4359–4378 · 7 November 2025; correction published February 2026

   SupportsThe European column of the target comparison table, and the Lp(a) risk threshold used in Europe.

   Strong

5. QRISK3

   ClinRisk / University of Nottingham · In current UK use

   SupportsThe 10-year risk figure the calculator asks for and the escalation thresholds built on it.

   Read it asValidated for ages 25–84 and calibrated on a UK primary-care population. It is the reason this site does not ship a second, cruder risk estimate alongside it. Note for clinicians: in October 2025 NICE removed the reference to the online QRISK3 from NG238, on the grounds that it should not be used for direct patient care — the version embedded in the clinical system is the one to use.

   Strong

6. NHS screening programmes in England — ages and intervals

   NHS.uk, NHS screening; NHS England public health commissioning pages; NHS England statements on the completed bowel screening age extension (2026) and the Targeted Lung Health Check programme · checked August 2026

   SupportsThe entitlement table in life health checks: cervical screening 25–64, five-yearly, HPV-first; NHS Health Check 40–74, five-yearly; bowel screening 50–74 by FIT every two years, the age extension from 60 down to 50 now complete, with kits available on request from 75 via 0800 707 6060; breast screening 50–70 every three years with self-referral thereafter; Targeted Lung Health Checks for ever-smokers aged 55–74 in parts of England; and a one-off abdominal aortic aneurysm ultrasound for men at 65.

   Read it asAges and intervals differ between the four UK nations and are revised periodically, and the topic says on every page that the invitation letter is the authority rather than the table. Three entitlements are given prominence because uptake data and common experience suggest they are widely unknown: the Health Check generally has to be booked rather than arriving unprompted, breast screening can be self-referred after 70, and bowel kits can be requested after 74. Overall bowel screening uptake was reported at 65.2% for 2024–25, and only about 56% among 54-year-olds — which is the specific gap the chapter is written against.

   Strong

7. Approved elsewhere, not available in the UK — three 2025–26 notices

   Tonmya: Tonix Pharmaceuticals and FDA approval announcements, 15 August 2025; RELIEF and RESILIENT Phase 3 trials, the latter published in Pain Medicine. Baxfendy: AstraZeneca and FDA announcements, 18 May 2026; BaxHTN, New England Journal of Medicine 2025, NCT06034743; Bax24, The Lancet March 2026. Foundayo: Eli Lilly and FDA announcements, 1 April 2026; ATTAIN-1 (NCT05869903) and ATTAIN-2 · checked 15 August 2026

   SupportsThree “what exists elsewhere” notices. Tonmya (cyclobenzaprine HCl sublingual 5.6 mg) in the fibromyalgia treatment chapter: FDA-approved 15 August 2025, once nightly, targeting non-restorative sleep, the first new fibromyalgia approval in over fifteen years and the fourth ever after pregabalin (2007), duloxetine (2008) and milnacipran (2009); two positive Phase 3 trials in about 1,000 people (RELIEF, NCT04172831; RESILIENT, NCT05273749), both significant on daily pain at 14 weeks — and a third, RALLY (NCT04508621), which randomised 514 people and failed its primary endpoint, stopped early on interim analysis in July 2021. Dosing is 2.8 mg tablets, one nightly for two weeks then two (5.6 mg), with a 2.8 mg maximum in older adults and mild hepatic impairment. Baxfendy (baxdrostat) in the blood pressure chapter: FDA-approved 18 May 2026, first-in-class oral aldosterone synthase inhibitor; BaxHTN randomised 796 patients on standard care, placebo-adjusted seated systolic reduction at 12 weeks of 9.8 mmHg (2 mg) and 8.7 mmHg (1 mg), with about 14 mmHg on 24-hour ambulatory monitoring in resistant hypertension. Foundayo (orforglipron) in the exercise and body composition chapter: FDA-approved 1 April 2026, first small-molecule oral GLP-1 for chronic weight management, no food or water restriction; ATTAIN-1 randomised 3,127 adults without diabetes, losing 7.5–11.2% of body weight at 72 weeks against 2.1% on placebo.

   Read it asAll three are reported under the rule set out in the ezetimibe chapter — a genuine advance gets reported with its status attached — and all three are date-stamped, because each will move. Each carries a deflation that the promotional coverage does not. Tonmya is a reformulation of a 1977 muscle relaxant rather than a new molecule, its effect sizes sit in the same range as the existing options, and “first-line treatment” is the manufacturer’s phrase, attributed and not adopted, since NG193 is unchanged; the null trial is named, per the practice established for JELIS, because a drug’s evidence base is the trials that were run rather than the trials that read well; cyclobenzaprine has never been UK-marketed in any form, and neither was milnacipran, so American and British options for this condition have differed for fifteen years rather than one. Baxdrostat is approved on blood pressure, a surrogate, with no outcome trial, hyperkalaemia in up to about one in ten at the higher dose, and no head-to-head against spironolactone — so the page states that spironolactone remains the UK step-4 answer. Orforglipron is framed as an access advance rather than an efficacy one: BaxHTN excluded people already taking a mineralocorticoid receptor antagonist or potassium-sparing diuretic, which is why it cannot speak to how baxdrostat compares with or adds to spironolactone. And orforglipron’s weight loss sits below semaglutide and tirzepatide, both already UK-available, on a cross-trial comparison labelled as such, and the manufacturer concedes the gap.

   Strong

Lipids, particles and causality

The claim doing the most work on this site is that lowering atherogenic particle count reduces events, and that it is causal rather than correlational. This is what that rests on.

1. Prevalence of LDL-C / apoB discordance

   Analyses of the Very Large Database of Lipids and of NHANES · 2015–2025

   SupportsThe “roughly one patient in five” figure in the clinical primer — discordance reported in about 20% of US adults, with a large real-world analysis finding around 21%.

   Read it asThe proportion moves substantially with how discordance is defined. In one recent cohort a median-based classification called 47.2% discordant, the clinically concerning high-apoB / low-LDL-C phenotype was 23.8%, and a ratio threshold gave 28.5%. One in five is the modal estimate and the useful one for a clinician; it is not a fixed property of the population, and the primer says so.

   Moderate

2. Low-density lipoproteins cause atherosclerotic cardiovascular disease. 1. Evidence from genetic, epidemiologic, and clinical studies

   European Atherosclerosis Society Consensus Panel — Ference BA et al., European Heart Journal 2017;38(32):2459–2472 · 2017

   SupportsThat LDL/apoB is causal, and the Mendelian randomisation argument summarised on the misconceptions chapter.

   Read it asThe single most useful citation for a sceptical reader, because it is explicitly about whether the causal criteria are met rather than about a single trial.

   Strong

Measurement, variation and the LDL-C equations

The evidence behind chapter 2. Most misinterpretation of a lipid panel happens here rather than in the biology.

1. Biological variation of direct and calculated LDL-C

   Scandinavian Journal of Clinical and Laboratory Investigation, and a meta-analysis of 30 studies by Smith et al. covering 1970–92 · 2025 and earlier

   SupportsWithin-person variation of roughly 8.7% for direct LDL-C, 9.3% by Friedewald and 9.0% by Martin–Hopkins; about 6% for total cholesterol, 7% for HDL-C and 22–28% for triglycerides.

   Read it asThese are healthy-subject estimates under standardised protocols. Individual people vary in how variable they are — published series find some subjects under 4% and others over 12% for the same analyte. The reference change values quoted on this site are derived from these figures and are approximate by design.

   Strong

2. Variability of measured and calculated LDL-C in statin-treated patients

   Kilpatrick ES, Kallner A, Atkin SL, Sathyapalan T, Annals of Clinical Biochemistry · 2025

   SupportsThat the equations differ in how much variation they add on top of the biology: Friedewald 12.9%, Sampson–NIH 10.4%, Martin–Hopkins 9.3% against direct measurement, in patients on atorvastatin.

   Read it asSmall crossover study, 26 patients with type 2 diabetes. Directionally useful, not a precise population estimate.

   Moderate

3. Comparison of methods to estimate LDL-C in patients with high triglyceride levels

   JAMA Cardiology, from the Very Large Database of Lipids (111,939 patients) · 2021

   SupportsThat at triglycerides of 400–799 mg/dL (4.5–9.0 mmol/L) the extended Martin–Hopkins equation was classified as accurate against ultracentrifugation in 62.1% of cases by the study’s own criterion, versus 40.4% for Sampson and 19.3% for Friedewald. What counts as “accurate” is a predefined band in that paper, not a self-evident quantity, and the figure should not be quoted without it.

   Read it asThe authors advise caution with any estimation in this triglyceride range regardless of method. The site reflects that: above about 4.5 mmol/L, treat a calculated LDL-C as unusable rather than as approximately right.

   Strong

4. Discordance between the Friedewald, Sampson and Martin–Hopkins equations at clinical cut-points

   Analyses of US electronic health record cohorts, including 146,106 patients with established ASCVD · 2020–2024

   SupportsThat Friedewald and Martin–Hopkins disagree about whether a patient is below LDL-C 1.8 mmol/L in about 15% of cases at triglycerides under 4.5 mmol/L, with much higher discordance in the 2.0–4.5 mmol/L triglyceride band. Martin–Hopkins consistently returns the higher value.

   Read it asUS cohorts and mg/dL cut-points. The direction of the bias transfers to UK practice; the exact percentages should not be quoted as UK figures.

   Strong

5. Impact of fasting duration on LDL-C estimated by the Friedewald, Martin-Hopkins and Sampson/NIH equations

   Biochemia Medica 2025;35(2) · 2025

   SupportsThat non-fasting sampling lowers calculated LDL-C by roughly 0.26–0.32 mmol/L depending on the equation, which is the basis for the site advising consistency of fasting state when comparing two results.

   Moderate

Cardiovascular risk in women

1. 2024 AHA/ASA Guideline for the Primary Prevention of Stroke, and the 2026 multisociety dyslipidaemia guideline

   American Heart Association / American Stroke Association; and the 2026 ACC/AHA multisociety guideline · 2024 and March 2026

   SupportsThat pregnancy complications, premature and early menopause and endometriosis are now formally named as risk factors, and that early menopause, preeclampsia and gestational diabetes are named among the risk-enhancing factors that should refine cardiovascular risk assessment.

   Read it asThese are US guidelines. The recognition of obstetric history as cardiovascular data is not controversial, but no widely used UK risk calculator currently ingests it — which is the practical gap the chapter is about.

   Strong

2. Hypertensive disorders of pregnancy and later cardiovascular disease

   Systematic reviews and meta-analyses, including sex-specific stroke risk factor reviews summarising the 2024 guideline · 2022–2026

   SupportsRoughly two- to four-fold higher later hypertension and cardiovascular disease after preeclampsia, and around 81% higher stroke risk in meta-analysis; with early-onset and severe disease carrying more risk.

   Read it asObservational throughout, and it remains unsettled whether preeclampsia causes later disease or reveals a pre-existing susceptibility — the “failed stress test” reading. That distinction matters for research and not for what a patient should do.

   Strong

Omega-3 and the comparator controversy

Two large trials, opposite results, and a design difference nobody can rule out. Listed in full because the calculator page previously cited only the positive one.

1. REDUCE-IT — icosapent ethyl 4 g/day against mineral oil

   Bhatt DL et al., New England Journal of Medicine 2019;380:11–22 · 2019

   Supportsn=8,179, median 4.9 years. Primary composite 17.2% against 22.0%, HR 0.75 (0.68–0.83).

   Read it asThe mineral oil comparator did not behave inertly: LDL-C rose about 10%, apoB about 8% and hs-CRP about 32% in the placebo arm, with later analyses reporting rises in IL-6, Lp(a) and oxidised LDL. The placebo question was reviewed by the FDA, Health Canada and the EMA, all of which approved the drug; the trial investigators (Steg and Bhatt, Eur Heart J 2021) report that all three concluded any effect of placebo choice on the 25% risk reduction would have been very small — a characterisation that comes from the investigators rather than from a regulatory document quoted directly, which is worth knowing when weighing it. Critics including STRENGTH’s executive committee chair regard the result as a false positive. The strongest argument on the other side is JELIS, an open-label trial of EPA with no mineral oil, which reported a 19% reduction in coronary events.

   Contested

2. NICE TA805 — Icosapent ethyl with statin therapy

   National Institute for Health and Care Excellence · Published 13 July 2022

   SupportsWhat the UK actually funds: icosapent ethyl for secondary prevention only — established cardiovascular disease, on a statin, fasting triglycerides 1.7 mmol/L or above, LDL-C above 1.04 and at or below 2.60 mmol/L. Not recommended for primary prevention, where NICE judged it unlikely to be cost effective, and not applicable in familial hypercholesterolaemia.

   Read it asNICE recommended it on the REDUCE-IT evidence while the comparator debate was already public, and set eligibility to mirror the trial population. Worth knowing that NHS England otherwise advises against omega-3 for cardiovascular prevention with this as the single stated exception — so the everyday question is usually eligibility rather than efficacy.

   Strong

3. STRENGTH — omega-3 carboxylic acid (EPA+DHA) against corn oil

   Nicholls SJ, Lincoff AM, et al., JAMA 2020;324:2268–2280 · 2020

   Supportsn>13,000, stopped for futility in January 2020. The corn oil comparator did not raise LDL-C or inflammatory markers.

   Read it asBoth trials found increased atrial fibrillation on omega-3. The two explanations for the discrepancy — EPA alone differing from EPA+DHA, or the REDUCE-IT comparator inflating the effect — cannot be separated by either design. The trial that would settle it, icosapent ethyl against corn oil, has not been run.

   Strong

Supplements and diet for lipids

Every entry here is graded on the LDL or vascular evidence and ungraded for events, because no supplement or dietary pattern on this list has a cardiovascular-outcome trial. The neutral results sit beside the positive ones on purpose.

1. Blueberries: the vascular evidence, and the LDL claim that does not hold

   Zhu et al., Frontiers in Physiology, 3 June 2024 (11 studies in the meta-analysis, endothelial function); Huang et al., Scientific Reports 2016 (22 trials, 1,251 participants, berries generally); Carvalho et al., Nutrition Research 2021;91:67-80 (18 trials, metabolic syndrome); Azari et al., Food & Function 2022 (25 studies); Miraghajani et al., Complementary Therapies in Medicine 2020 (11 trials); Wood et al., AJCN 2023 (n=61) · 2016 to 2024; checked September 2026

   SupportsThe blueberry card in the lipid tool, graded on the vascular evidence only. Flow-mediated dilation +1.50% (95% CI 0.81-2.20, I² 87%) and diastolic blood pressure -2.20 mmHg (95% CI -4.13 to -0.27, I² 11%) pooled across 11 studies. Systolic was null: -1.43 (95% CI -3.11 to 0.26). Consistent with Wood et al. 2023, already cited in the brain supplements chapter, where FMD rose at P<0.001. Trial dose 26 g freeze-dried powder = 178 g fresh berries.

   Read it asThis entry exists mainly to record what is NOT supported. The widely circulated claim that blueberries lower LDL by 5-15% does not survive checking. The -0.21 mmol/L figure quoted for it is Huang 2016, which studied berries generally, and whose own subgroup analysis placed the LDL signal in bilberry, whortleberry and black raspberry rather than blueberry; its authors further note lipids were not the primary outcome in most included trials, so null results may be unpublished. The blueberry-specific literature is split: Carvalho 2021 and Azari 2022 report LDL falls in metabolic syndrome, while Zhu 2024 found LDL-C +0.05 mmol/L (95% CI -0.14 to 0.24, I² 0%), null with zero heterogeneity, and said so explicitly against Huang. Miraghajani 2020 found only a marginal triglyceride change. A 2025 trial (BEACTIVE) found the placebo arm's LDL and apoB fell while the blueberry arm's did not. A 2024 review judged one of seventeen trials to be at low risk of bias. The soluble-fibre mechanism usually given is also wrong at food doses - blueberries carry about 2.4 g fibre per 100 g, so an 80 g portion is about 1.9 g, mostly insoluble; that mechanism belongs to psyllium and oat beta-glucan. No outcome data exists for blueberries.

   Moderate

2. Bergamot extract for LDL lowering, and the two guideline positions

   Mollace 2011 (RCT, n=237, 30 days); Gliozzi 2013 (open-label, n=77); Toth, Frontiers in Pharmacology 2015 and widely cited as 2016 (open-label, n=80, 6 months); Cicero and Fogacci 2023, Archives of Medical Science (double-blind RCT, n=90, 700mg, 12 weeks); ILEP position paper, Banach et al., JACC 2018;72:96-118; 2025 ESC/EAS focused update · Checked August 2026

   SupportsAn LDL reduction of about 15-20% from the best-designed independent trial (-17.7% at 12 weeks), and the observation that effect size in this literature falls as blinding and independence rise: -40.8% open-label, -38.6% at 1000mg in the 30-day RCT, about -18% at six months open-label, -17.7% double-blind. Also the two guideline positions: ESC/EAS 2025 Class III / Level B against supplements for lowering LDL and reducing risk, and ILEP Class IIa / Level B for bergamot in statin-intolerant patients specifically.

   Read it asWhich statins the CYP3A4 interaction actually affects, added round 136: atorvastatin (UK first-line), simvastatin and lovastatin are cleared by CYP3A4 and are affected \u2014 grapefruit raises atorvastatin exposure roughly 2.5-fold. Rosuvastatin and pravastatin are barely handled by that pathway and fluvastatin goes via CYP2C9, so the statin concern largely falls away on those three. The furanocoumarin behind the classic grapefruit\u2013statin effect is bergamottin, named for this fruit. SCOPE, and it matters more here than usual. There are NO cardiovascular outcome trials, so nothing here supports bergamot as a statin replacement. The 40% figure comes from an open-label 30-day study and should not be quoted as the expected effect. The reported 39% HDL rise exceeds what any approved lipid drug achieves and is treated on this site as a methodological signal rather than a finding. Most trials calculate LDL by Friedewald while triglycerides fall 30% or more, which inflates the apparent LDL reduction - see the measurement chapter. The direct-HMG-CoA-reductase-inhibition claim is an in-silico modelling result that did NOT hold when tested in cells. And the ILEP itself notes that most of the clinical literature comes from a single research unit and has not been independently confirmed.

   Moderate

3. Berberine: lipid and glucose effects, the drug interactions, and the 2026 genetic study

   Blais, Huang and Zhao, Drugs 2023;83:403-427 (placebo-controlled meta-analysis); Yin, Xing and Ye, Metabolism 2008;57:712-717 (against metformin); Xie et al., Frontiers in Pharmacology 2022;13:1015045 (37 trials, 3,048 patients); Guo et al., Eur J Clin Pharmacol 2012;68:213-217 (CYP probe study); Gurley et al., Clin Pharmacol Ther 2005;77:415-426 (goldenseal, CYP2D6 and CYP3A4/5); Goetz et al., CPIC guideline for CYP2D6 and tamoxifen, Clin Pharmacol Ther 2018;103:770-777; Li et al., Drug Des Devel Ther 2019;13:129-139 (with simvastatin); Wu et al., Eur J Clin Pharmacol 2005;61:567-572 (ciclosporin); Zhao et al., npj Cardiovascular Health 2026;3:15 · Checked September 2026

   SupportsThe berberine card in the lipid tool, the myth card on the misconceptions chapter and the screening note on the clinical pathway. Placebo-controlled pooling: LDL -0.46 mmol/L (95% CI -0.62 to -0.30) and triglycerides -0.34 (-0.46 to -0.23); gut side effects 2-23% on berberine against 2-15% on placebo. Glucose-lowering similar to metformin in a 36-person pilot. No significant hypoglycaemia excess alone or with oral agents (RR 0.48, 0.21-1.08). After two weeks at 300mg three times daily: midazolam AUC +40% (CYP3A4), CYP2D6 probe ratio ninefold, CYP2C9 probe ratio doubled; goldenseal cut CYP2D6 and CYP3A4/5 phenotypic activity by about 40% over 28 days. Hence the codeine, tramadol and tamoxifen line: CYP2D6 activates them, and tamoxifen guidance advises avoiding strong and moderate CYP2D6 inhibitors. Ciclosporin trough levels markedly raised in renal-transplant recipients. The 2026 study: a proteomic signature from a 12-week trial in 80 Hong Kong men, applied by Mendelian randomisation in UK Biobank men, OR 0.85 (0.79-0.91) for ischaemic heart disease and 0.88 (0.80-0.96) for diabetes.

   Read it asNeutral result listed alongside the positive ones. The only direct human test of berberine with a statin (Li 2019: 300mg every 8 hours with simvastatin 40mg for seven days, 60 healthy volunteers across five arms, open-label) found no clinically obvious pharmacokinetic interaction. It is small and short, so the CYP3A4 concern is stated on this site as plausible, not demonstrated, and no claim of excess myopathy is made; by the 1.25-2-fold convention berberine is a WEAK CYP3A4 inhibitor. The site does not advise switching a tolerated statin for it, and follows NG238 and the NHS England pathway on CK: symptoms first, no routine measurement. The tamoxifen recurrence data are conflicting (a 2021 systematic review of about 100,000 patients found no consistent effect), which is why the site states the pharmacology and the guidance rather than an outcome. P-glycoprotein effects come from animal and cell work only, in opposite directions by model, and are not stated as a human finding. What the 2026 study does NOT support: any claim that taking berberine prevents heart disease. The signature separated berberine from placebo with a cross-validated AUC of only 0.61; it was built in East Asian men and applied to a mostly European cohort; it covers men and ischaemic heart disease only, not stroke; and the authors call the results exploratory and hypothesis-generating and state that no randomised trial of berberine on these outcomes has been run. The 34.5% gut-effect figure from the 2008 pilot is uncontrolled and is deliberately not quoted. UK food supplements are regulated as foods and can be marketed without evidence of quality (NHS Specialist Pharmacy Service), which is the basis for the purity wording.

   Moderate

4. The Portfolio Diet: the JAMA 2011 trial, the metabolic-ward studies, and the pooled analysis

   Jenkins et al., JAMA 2011;306:831-9 (351 participants, four Canadian centres, six months); Jenkins et al. 2003, metabolically controlled, against 20mg lovastatin; Chiavaroli et al., Prog Cardiovasc Dis 2018, pooled analysis · Checked August 2026

   SupportsLDL reduction of 13.8% on intensive portfolio advice and 13.1% on routine advice against 3.0% on a low-saturated-fat control, in free-living people over six months, with NO significant difference between the two portfolio arms (P = 0.66); about 17% pooled across controlled trials; 28-35% under metabolically controlled conditions, and -28.6% against -30.9% for 20mg lovastatin in the head-to-head.

   Read it asSCOPE. The 28-35% figure comes from feeding studies where every meal was provided, and is a ceiling rather than an expectation; the chapter leads with 13-17% for that reason. There is NO trial of this dietary pattern with cardiovascular events as an endpoint, and the site grades it Strong for LDL and Moderate for events, the latter inferred from the LDL evidence rather than demonstrated. The absorber-versus-producer explanation for variable response is stated as plausible mechanism and has NOT been verified against a source.

   Moderate

Inflammation and residual risk

Listed together and in order, because the direction of travel is the point. hs-CRP predicts risk well; whether lowering inflammation reduces events is a separate question and the recent large trials have gone the other way.

1. CANTOS — canakinumab, IL-1β inhibition

   Ridker PM et al., New England Journal of Medicine · 2017

   SupportsThat reducing a specific inflammatory pathway can reduce cardiovascular events with no change in lipids — the original proof of concept.

   Read it asExpensive injectable with an increased fatal-infection signal. Never adopted for cardiovascular prevention in practice.

   Strong

2. CIRT — low-dose methotrexate

   New England Journal of Medicine · 2019

   SupportsThat broad, non-specific immunosuppression does not reduce cardiovascular events.

   Strong

3. COLCOT (2019) and LoDoCo2 (NEJM 2020) — low-dose colchicine

   Two secondary-prevention outcome trials · 2019 and 2020

   SupportsThe positive colchicine case, and the regulatory approvals that followed it.

   Moderate

4. CLEAR SYNERGY / OASIS-9 — colchicine after myocardial infarction

   Jolly SS, d’Entremont MA, Lee SF, et al., New England Journal of Medicine 2025;392(7):633–642 · 2025

   SupportsThat the largest colchicine trial run — 7,062 patients, median 3 years — found no reduction in the composite of cardiovascular death, recurrent infarction, stroke or ischaemia-driven revascularisation, despite lowering CRP. More diarrhoea in the colchicine arm.

   Read it asThis is why the therapeutic claim on this site is graded Contested rather than Moderate.

   Strong

5. ZEUS — ziltivekimab, direct IL-6 inhibition

   Novo Nordisk headline results; trial design in JAMA Cardiology 2026;11(1):89–97 · Announced 31 July 2026

   SupportsThat in over 6,300 people with established cardiovascular disease, chronic kidney disease and hs-CRP ≥ 2 mg/L, hitting the IL-6 target as designed produced a hazard ratio for major adverse cardiovascular events of 0.99 (95% CI 0.88–1.11).

   Read it asCurrency warning. Headline results only at the time of writing; full publication was expected at a scientific meeting later in 2026, and HERMES and ARTEMIS read out in the first half of 2027. Re-check before relying on this.

   Strong

Statin tolerability and the nocebo effect

Listed as a set because no single one of these settles it, and because the clinical pathway lists a long range of atypical presentations that would read as alarming without these alongside.

1. Effect of statin therapy on muscle symptoms: an individual participant data meta-analysis

   Cholesterol Treatment Trialists’ Collaboration, Lancet 2022;400(10355):832–845 · 2022

   Supports19 double-blind placebo-controlled trials, 123,940 participants, weighted median 4.3 years. Muscle pain or weakness in 27.1% on statin versus 26.6% on placebo (RR 1.03, 1.01–1.06). Year one RR 1.07 with an absolute excess of 11 events per 1,000 person-years, giving roughly one in fifteen reports attributable to the drug. No significant excess after year one.

   Read it asThis measures muscle symptoms in trial populations, which under-represent the very elderly, the frail, and people with multiple interacting medicines. It does not address rare severe myopathy, which is real and separately documented.

   Strong

2. SAMSON — Self-Assessment Method for Statin side-effects Or Nocebo

   Howard JP et al., NEJM 2020;383:2182–2184 and JACC 2021;78(12):1210–22 · 2020–21

   SupportsN-of-1 crossover in 60 patients who had already stopped statins for side effects. Symptom scores 16.3 on atorvastatin, 15.4 on placebo, substantially lower with no tablet at all. Nocebo ratio 90%. Half restarted a statin successfully afterwards.

   Read it asThe no-tablet periods are what distinguish this trial: they show placebo symptoms were actively produced by taking a tablet rather than being background symptoms. Small, and participants self-selected by agreeing to rechallenge.

   Strong

3. StatinWISE — statin treatment and muscle symptoms: series of randomised placebo-controlled n-of-1 trials

   Herrett E et al., BMJ 2021;372:n135 · 2021

   SupportsThe UK primary-care replication. Mean difference in symptom score between atorvastatin 20 mg and placebo −0.11 (95% CI −0.36 to 0.14). Two-thirds resumed the statin or intended to.

   Strong

4. ASCOT-LLA blinded versus non-blind phases

   Lancet 2017;389:2473–81 · 2017

   SupportsThat excess muscle-related adverse events appeared in the non-blind extension phase and not in the blinded randomised phase of the same trial, in the same patients — the cleanest natural experiment on this question.

   Strong

5. Statin-associated muscle symptoms: EAS Consensus Panel statement

   Stroes ES et al., European Heart Journal 2015;36:1012–1022 · 2015

   SupportsThe working definition of statin-associated muscle symptoms used to distinguish myalgia from myopathy, and the structured assessment and rechallenge approach the clinical pathway follows.

   Moderate

6. The objection to reading the N-of-1 trials too far

   Commentary literature, including Thompson PD · 2021–2026

   SupportsThat one-month treatment blocks may be too short for symptoms with delayed onset, that patients willing to rechallenge are self-selected toward milder cases, and that group means can conceal a genuinely affected minority.

   Read it asIncluded because the site does not present one side of a live methodological argument as settled. It does not change the recommended action — structured rechallenge is the right next step under either reading.

   Contested

Lipid-lowering beyond statins: ezetimibe, PCSK9 and gene editing

Where the ezetimibe and PCSK9 chapters rest, and the one gene-editing trial the site mentions.

1. PCSK9 — genetics, outcome trials and UK eligibility

   Abifadel M et al., Nature Genetics 2003;34:154–156; Cohen JC, Boerwinkle E, Mosley TH, Hobbs HH, New England Journal of Medicine 2006;354:1264–1272; Sabatine MS et al. (FOURIER), NEJM 2017;376:1713–1722; Giugliano RP et al. (EBBINGHAUS), NEJM 2017;377:633–643; Schwartz GG et al. (ODYSSEY OUTCOMES), NEJM 2018;379:2097–2107; Ridker PM et al. (SPIRE, bococizumab), NEJM 2017;376:1527–1539; Ference BA et al., NEJM 2016;375:2144–2153; O’Donoghue ML et al. (FOURIER-OLE), Circulation 2022;146:1109–1119; Navar AM et al., NEJM 2026;394:529–539; NICE TA393, TA394, TA733 · 2003–2026, criteria checked August 2026

   SupportsThe PCSK9 chapter: the loss-of-function genetics — roughly 1 in 40 Black participants in the Cohen cohort carrying a nonsense variant, with LDL about 28% lower and coronary heart disease 88% lower, and a milder variant in white participants giving about 15% lower LDL and 47% lower risk; FOURIER, 27,564 patients, LDL 2.4 to about 0.8 mmol/L and major events down 15% (HR 0.85) over a median 2.2 years; ODYSSEY OUTCOMES, 18,924 patients, events down 15%; EBBINGHAUS finding no cognitive difference including at very low LDL; the bococizumab failure through anti-drug antibodies; and the eligibility thresholds — antibodies above 4.0 mmol/L at high risk or 3.5 at very high risk (recurrent events or more than one vascular bed), 5.0 for heterozygous FH without cardiovascular disease, against inclisiran’s 2.6 with prior events and primary care initiation.

   Read it asThis page was drafted by the reviewing AI and integrated rather than written here, and it is not yet clinically signed. Several figures the drafter flagged as unverified have been hedged rather than stated precisely — lerodalcibep’s percentage, the Lp(a) reduction range, and the plaque-imaging trial figures — and list prices were dropped entirely rather than guessed. Two internal links were corrected during integration: the gene-editing cross-reference points at section 8 of the ezetimibe chapter, where that content lives, not at the genetics page. The ezetimibe ladder was shortened to point here rather than repeat the eligibility detail, so the two pages cannot drift.

   Strong

2. PCSK9 inhibition in the UK — who may prescribe, and what exists elsewhere

   NICE TA393 (alirocumab), TA394 (evolocumab), TA733 (inclisiran); NHS England lipid management and inclisiran access guidance; ICB primary care prescribing guidance for inclisiran; Navar AM et al., New England Journal of Medicine 2026;394:529–539; FDA press announcement and Merck release on the approval of Lipfendra (enlicitide), 16 July 2026; NICE appraisal in progress for obicetrapib (ID6519, final scope January 2026, publication expected 14 October 2026) · checked August 2026

   SupportsThe step-4 row of the ladder in the ezetimibe chapter and the pathway line in the clinical depth: that the PCSK9 monoclonal antibodies alirocumab and evolocumab are specialist-initiated only, while inclisiran was appraised and is commissioned for initiation in primary care — for adults with a history of cardiovascular events whose LDL-C remains at or above 2.6 mmol/L on maximally tolerated statins — given twice yearly after loading doses. And section 7: that enlicitide (Lipfendra), a once-daily 20 mg oral macrocyclic peptide, was approved by the FDA on 16 July 2026, lowering LDL-C by a placebo-adjusted 56% and 59% at 24 weeks in its two registration trials, with no UK licence and no NICE appraisal as at 15 August 2026, and its cardiovascular outcome trial CORALreef Outcomes unreported.

   Read it asThis corrected a real error on the site. The clinical page previously described everything beyond bempedoic acid as specialist-led, which is right for the antibodies and wrong for inclisiran — the one agent in the class a GP can and is expected to start. That distinction is poorly known and is a plausible contributor to underuse, so the chapter now states it explicitly rather than leaving the class described as a single thing — with a visible dated correction note on the page, per the practice established for the VitaK-CAC error, because the old wording had a practical consequence rather than being merely imprecise. For scale, Greater Manchester reported roughly 1,900 people receiving inclisiran against an estimated 3,700 eligible. The correction was applied across three pages — the ladder, the clinical referral criterion and the calculator drug card — after the first pass fixed only one of them, which is the same cross-page failure mode as the round-36 choking correction. On enlicitide, the editorial rule is stated on the page: a genuine advance is reported with its status attached, because a UK reference that described only UK availability would leave a clinician unable to answer what a well-read patient asks. And the surrogate discipline is applied unchanged — enlicitide is approved on LDL reduction with its cardiovascular outcome trial unreported, which is the same position as inclisiran; a second example in the class does not strengthen the surrogate, it clarifies the pattern.

   Strong

3. In vivo base editing of PCSK9 with VERVE-102 — the Heart-2 trial

   Vafai SB, Täubel J, Ashdown T, et al. New England Journal of Medicine 2026, published online 25 May 2026; presented at the European Atherosclerosis Society Congress. ClinicalTrials.gov NCT06164730. Funded by Verve Therapeutics, a wholly owned subsidiary of Eli Lilly · 2026

   SupportsSection 7 of the ezetimibe chapter: that VERVE-102 is an adenine base editor plus guide RNA delivered in a GalNAc lipid nanoparticle as a single intravenous infusion of about four hours, permanently inactivating PCSK9 in hepatocytes. Interim Phase 1b analysis of 35 participants with heterozygous familial hypercholesterolaemia or premature coronary artery disease: dose-dependent mean PCSK9 reductions of 51% to 88%, LDL cholesterol reduced by up to 62% at the highest dose — an absolute fall of about 78 mg/dL — with durability reported as at least one year in 15 participants. Around 20% had grade 1 or 2 infusion-related reactions; one serious adverse event occurred. FDA Fast Track designation granted.

   Read it asPhase 1b, 35 people, no cardiovascular outcome data, and none for years. The chapter grades it Emerging and states explicitly that nothing here has prevented a single heart attack. Two points the chapter makes that the trial reports do not emphasise: this is somatic editing of liver cells and is not heritable, which is the confusion that makes the technology sound more alarming than it is; and permanence is the whole question in both directions — it removes the adherence problem that defeats about half of lipid prescriptions within a year, and it removes the ability to stop if a late signal appears. The predecessor VERVE-101 encountered liver enzyme and platelet signals; the delivery particle was redesigned for VERVE-102, which has so far avoided them. The LDL endpoint is a surrogate, but an unusually well-validated one — lowering LDL has cardiovascular outcome evidence across four pharmacologically distinct mechanisms — statins (HMG-CoA reductase), ezetimibe (NPC1L1), PCSK9 monoclonal antibodies, and bempedoic acid (ATP-citrate lyase, CLEAR Outcomes) — with Mendelian randomisation pointing the same way across the corresponding genes. That is why the chapter treats it differently from the epicardial fat and IL-6 surrogates elsewhere on the site. Inclisiran is deliberately excluded and named on the page as the counter-example: it lowers LDL substantially through the same PCSK9 pathway as the antibodies, so it is not even a distinct mechanism, and its outcome trials — ORION-4 and VICTORION-2 Prevent — have not reported. It is a drug licensed on the strength of the surrogate rather than evidence for it.

   Emerging

4. Ezetimibe — NPC1L1 genetics, IMPROVE-IT, and the ENHANCE surrogate failure

   Myocardial Infarction Genetics Consortium Investigators, NEJM 2014;371:2072–2082 (NPC1L1 inactivating mutations); Cannon CP et al., IMPROVE-IT, NEJM 2015;372:2387–2397; Kastelein JJP et al., ENHANCE, NEJM 2008;358:1431–1443; Garcia-Calvo M et al., PNAS 2005;102:8132–8137 (target identification) · 2005–2015

   SupportsThe ezetimibe chapter: that ezetimibe inhibits NPC1L1 in the intestine, blocking uptake of both dietary and biliary cholesterol, and that the target was identified in 2005, three years after approval. That carriers of naturally inactivating NPC1L1 mutations have LDL lower by about 12 mg/dL (0.3 mmol/L) from birth and about 53% less coronary heart disease. That IMPROVE-IT randomised 18,144 patients after acute coronary syndrome and found a 7-year event rate of 32.7% against 34.7% (absolute difference 2.0 points, hazard ratio 0.936, 95% CI 0.89–0.99, p = 0.016), with time-weighted average LDL of 53.7 against 69.5 mg/dL, and no difference in all-cause mortality (15.3% vs 15.4%) or cardiovascular death (6.8% vs 6.9%), and no excess of muscle, gallbladder, hepatic events or cancer. And that ENHANCE, in familial hypercholesterolaemia, found no difference in carotid intima-media thickness despite a larger LDL reduction.

   Read it asIMPROVE-IT is frequently described as having “definitively proven” ezetimibe. It did not. A 2-point absolute difference over seven years is a number needed to treat of about 50, with no mortality signal in either direction, and the chapter says so. Its real importance is what it established about LDL rather than about the drug: lowering LDL further by a non-statin mechanism produced approximately the benefit the LDL difference predicted, which is strong evidence that the LDL reduction is doing the work. ENHANCE is the site’s clearest worked example of a surrogate failure — carotid wall thickness, measured over two years in long-treated FH patients whose arteries were already thickened, was close to the least sensitive endpoint available.

   Strong

5. Ez-PAVE — intensive versus conventional LDL targeting in established ASCVD

   Kim B-K et al., Yonsei University College of Medicine, presented at ACC 2026 and published in the New England Journal of Medicine; NCT04626973 · 2026

   SupportsThe target comparison in the ezetimibe chapter: 3,048 patients with established atherosclerotic disease at 17 South Korean sites, randomised to an LDL target below 55 mg/dL or below 70 mg/dL, with therapy at physician discretion. At 3 years the primary composite — cardiovascular death, non-fatal myocardial infarction, non-fatal stroke, any revascularisation or hospitalisation for unstable angina — occurred in 6.6% against 9.7% (hazard ratio 0.67, 95% CI 0.52–0.86, p = 0.002; number needed to treat 32). Achieved median LDL 56 against 66 mg/dL. Safety was similar between groups.

   Read it asThree limits, all on the page. The trial was open-label, necessarily, because clinicians had to know the target — though outcomes were adjudicated blind. The benefit was driven mainly by revascularisation (4.8% against 7.5%), which is the softest component of the composite and the one most exposed to open-label design; the harder composite of cardiovascular death, myocardial infarction or stroke moved in the same direction (2.3% against 3.6%) on smaller numbers. And it was conducted entirely in an East Asian population, without access to inclisiran or bempedoic acid. Published commentary has argued on the revascularisation point that Ez-PAVE does not validate the <55 target, and the chapter reports that dissent rather than resolving it. Note also that the achieved LDL values differed by less than the target labels imply — 1.45 against 1.7 mmol/L.

   Moderate

Lipoprotein(a)

Mostly covered by the guidelines above, with three points worth sourcing separately.

1. The Lp(a)-lowering outcome trials

   Lp(a)HORIZON (pelacarsen, NCT04023552); OCEAN(a)-Outcomes (olpasiran, NCT05581303); ACCLAIM-Lp(a) (lepodisiran, NCT06292013); MOVE-Lp(a) (muvalaplin) · All phase 3. Lp(a)HORIZON reported 4 September 2026 and missed its primary endpoint; the other three had not reported as of 6 September 2026

   SupportsThe pipeline table in the genetics chapter. Phase 2 lowering of roughly 80% for pelacarsen and above 90% for the RNA-silencing agents. Lp(a)HORIZON reported on 4 September 2026 and was negative: 8,323 patients with established cardiovascular disease and Lp(a) at or above 70 mg/dL, pelacarsen substantially lowered Lp(a) and did not reduce four-point major adverse cardiovascular events. Full data are due at a medical congress and had not been presented as of 6 September 2026. OCEAN(a)-Outcomes expected to complete at the end of 2026.

   Read it asCurrency warning — the fastest-moving item on this site. None of the four is approved, and the one that has reported did not reduce events. Phase 2 percentage reductions are surrogate endpoints and should be read as such: inflammation had the same shape of evidence before CLEAR SYNERGY and ZEUS, and Lp(a) has now joined that list rather than escaped it. Whether the remaining three fare differently is an open question, not an expectation. Check the trial registries before relying on anything here.

   Emerging

2. Residential radon, Lp(a) heritability and the once-in-a-lifetime position

   2019 ESC/EAS guidelines (Class IIa, Level C) and the 2026 ACC/AHA guideline · 2019 and 2026

   SupportsThat Lp(a) is roughly 70–90% genetically determined and should be measured at least once in adult life.

   Read it asRecent real-world cohorts report more within-person variation than the “once only” framing implies, particularly around menopause and with changing kidney function. Treat a single borderline result with less confidence than a single high or low one.

   Strong

Plaque imaging and coronary calcium

The calcium score, what it does and does not predict, and the one vitamin trial that tested changing it.

1. VitaK-CAC — menaquinone-7 and coronary artery calcification

   Vossen LM, de Leeuw PW, Schurgers LJ, Heuts S, Adriaans BP, de Haan C, van Varik BJ, Kroon AA, et al. Two Years of Menaquinone-7 Supplementation and Coronary Artery Calcification: A Randomized Clinical Trial. JAMA Cardiology, published online 10 June 2026; doi:10.1001/jamacardio.2026.1279. Protocol: Nutrients 2015;7(11):8905–8915 · 2026

   SupportsThe K2 section of the CAC chapter: a double-blind, placebo-controlled trial at two Dutch hospitals randomising 180 symptomatic patients (90 per group) with a baseline CAC of 50–400 Agatston units to 360 µg MK-7 daily or placebo for two years. 150 completed, 75 per arm. Median CAC rose from 145 to 214 AU on placebo and 135 to 184 AU on MK-7 (p = 0.02, holding after adjustment for covariates) — approximately 29% less progression by Agatston score and 42% less by calcium mass. Median age 59–61, 42% women, 78% on statins, no high-risk plaque at baseline, 80% adherence and a 17% dropout rate, which reconciles with 180 rather than 167. Secondary finding: the rise in CAC score correlated with the number of non-calcified plaques becoming partially calcified (p = 0.04). Stenosis severity progressed in about 33% of the MK-7 arm and 41% of placebo, with no significant between-group difference — a null comparison rather than an absence of progression. Plasma MK-7 rose from 0.50 to 6.56 µg/L, confirming the intervention was taken.

   Read it asNote on the participant count: the ACC journal scan reports “167 randomised”, which does not reconcile with a 17% dropout and 150 completing; 180 randomised does. The figures asserted on the page are therefore 180 randomised and 150 completing only. A treated-population split has been seen in secondary coverage but is not quotable from a source checked here, so it is deliberately not stated. This entry also corrected an error on this site. An earlier version of the CAC chapter stated there was no trial showing K2 slows calcification progression on repeat scanning; VitaK-CAC is that trial, and the page now says so and says it was wrong. What has not changed is the more important claim: the endpoint is a surrogate, the trial was neither designed nor powered for cardiovascular events, and the authors state directly that whether slower calcification means fewer events is unanswered. Three further limits are on the page: recruitment ran from 2012 to 2022 with interruptions in a trial of 180 people; the population had symptomatic disease, was 78% statin-treated and had no high-risk plaque at entry; and the trial used a branded MK-7 preparation with the developing group inside the same university, which is the norm for supplement trials and worth seeing — the preparation was MenaQ7, originally Nattopharma, now part of Gnosis by Lesaffre. The dose gap is practically important: 360 µg was studied, 100–200 µg is typical on the shelf.

   Moderate

2. Spotty versus dense coronary calcification and plaque vulnerability

   Nakahara T et al., Coronary Artery Calcification: From Mechanism to Molecular Imaging, JACC: Cardiovascular Imaging; with the intravascular imaging literature on microcalcification and plaque rupture · current

   SupportsThat coronary calcium is not uniform: spotty or micro-calcification occurs within young, inflamed, lipid-rich plaque and is associated with greater inflammation and higher rupture risk, while dense or macro-calcification occurs in older plaque and is associated with lower inflammation and greater stability. And that a CT calcium score measures total calcium and cannot distinguish the two.

   Read it asThe chapter is careful to label the downstream inference as an inference. The observation that a calcium score can rise after lifestyle improvement or statin initiation is real; the explanation usually offered — that soft plaque is calcifying and stabilising as inflammation falls — is extrapolated from this literature rather than demonstrated by a trial, and the page says so rather than presenting it as established. The practical conclusion holds either way: a rising score alone is not evidence of deterioration, and a falling score is not a treatment target.

   Moderate

3. Coronary artery calcium, PREVENT risk and statin eligibility — MESA

   Rikhi R et al., Wake Forest University School of Medicine. Journal of the American College of Cardiology 2026; doi:10.1016/j.jacc.2026.05.039, published online 15 July 2026 · 2026

   SupportsThe MESA subsection in the CAC chapter: 5,698 adults without clinical ASCVD, mean age 61.5, 52.8% women, followed 10 years. Event rates per 1,000 person-years by 2026 ACC/AHA statin-eligibility band, CAC = 0 against CAC > 0 — not recommended 1.2 against 4.5; considered 2.7 against 5.2; recommended 5.8 against 16.6. In participants with LDL 70–189 mg/dL and no diabetes, CAC > 0 was associated with approximately twofold or higher event rates across all PREVENT risk tiers. Categorical net reclassification improvement 0.124, driven by appropriate reclassification of both events and non-events.

   Read it asFour limits, and the chapter states all of them. The design is observational, so it shows that CAC sorts people by risk rather than that acting on it improves outcomes. CAC results were not blinded to clinicians or participants, which biases toward the test looking useful. Risk was calculated with PREVENT against 2026 ACC/AHA thresholds — UK practice uses QRISK and NICE does not recommend calcium scoring for routine risk assessment, so the reclassification figure does not transfer, though the direction does. And the authors note MESA was a relatively healthy primary-prevention cohort, which will have lowered absolute event rates. An NRI of 0.124 is a measurable rather than dramatic gain, on a metric that is itself contested. The most useful single figure is the “not recommended” row, because it locates the benefit of scanning in the group a calculator has already dismissed.

   Moderate

Blood pressure and sleep

The two big rocks the site was missing until round 25.

1. Primary aldosteronism: prevalence, under-detection and screening

   Narrative reviews and cohort series 2015–2026, including the 2024 ESC hypertension guideline recommendation on aldosterone-to-renin ratio screening (Class IIa, Level B) · 2024–2026

   SupportsThat primary aldosteronism accounts for roughly 5–10% of all hypertension — a pooled figure near 9% in unselected hypertensive populations — rising to 20–30% in resistant hypertension and around 18% in young-onset disease; that only about 9–37% of confirmed cases have hypokalaemia; and that screening is heavily underused, with audits finding most eligible patients untested and a third or more of positive screens receiving no follow-up.

   Read it asPrevalence estimates vary widely with the diagnostic criteria used — one 2026 comparison found prevalence in the same cohort rising from 8.8% to 16.1% purely by moving from 2016 to 2025 Endocrine Society criteria. So the range quoted on this site is a range on purpose. Graded Moderate: the under-detection finding is consistent across settings, the exact prevalence is not.

   Moderate

2. NICE NG136 — Hypertension in adults: diagnosis and management

   National Institute for Health and Care Excellence · 2019, amended 2022; visual summary last updated July 2025

   SupportsThe staging table and both target tables in the blood pressure chapter: stage 1 at clinic 140/90 with ABPM or home average 135/85; stage 2 at 160/100 with 150/95; targets below 140/90 under 80 and below 150/90 at 80 and over, with the corresponding home figures.

   Read it asABPM is the reference standard for diagnosis, with home monitoring where ABPM is unsuitable — a raised clinic reading alone is explicitly not a diagnosis. NICE also instructs clinical judgement in frailty and multimorbidity, which is not a footnote.

   Strong

3. Arm Position and Blood Pressure Readings: the ARMS crossover randomised clinical trial

   Liu S, Zhao D, Sabit A, Brady TM, et al., JAMA Internal Medicine · 7 October 2024

   SupportsThat against the guideline position — arm supported at heart height — a hand resting in the lap overestimates systolic by 3.9 mmHg and diastolic by 4.0, an arm unsupported at the side by 6.5 and 4.4, and systolic by about 9 mmHg in people who are already hypertensive.

   Read it as133 participants, crossover, so each acted as their own control. The authors note it may apply only to automated devices. The variation around the mean was substantial, which matters: this is a systematic bias, not a fixed correction you can subtract.

   Strong

4. Sleep regularity is a stronger predictor of mortality risk than sleep duration

   Windred DP, Burns AC, Lane JM, Saxena R, Rutter MK, Cain SW, Phillips AJK, Sleep 2024;47(1):zsad253 · 2024

   SupportsThe regularity section of the sleep chapter: Sleep Regularity Index from over 10 million hours of accelerometry in 60,977 UK Biobank participants; the top four quintiles had 20–48% lower all-cause mortality, 16–39% lower cancer mortality and 22–57% lower cardiometabolic mortality against the least regular quintile, with regularity outperforming duration.

   Read it asObservational, one cohort, older and healthier and less ethnically diverse than the UK as a whole, with one week of accelerometry standing in for lifetime habit. Irregular sleep travels with shift work, illness, caring and poverty; adjustment does not fully remove those. Graded Moderate for that reason, and because the site does not upgrade a single cohort however large.

   Moderate

5. MHRA June 2026 — strengthened ACE inhibitor angioedema warning

   MHRA Safety Roundup, June 2026, GOV.UK · June 2026

   SupportsThe angioedema row in the blood pressure chapter and the note on the treatment review page: that the product information for all ACE inhibitors is being updated to strengthen warnings about delayed-onset angioedema, which may occur after weeks to years of treatment; and that bradykinin-mediated angioedema does not respond to standard anaphylaxis treatment, so healthcare professionals should consider it when that treatment is ineffective.

   Read it asThis confirms rather than changes what the chapter already said — the delayed-onset point and the bradykinin-versus-histamine distinction were both already there — but a regulator formally strengthening a warning is worth date-stamping, and it sharpens one clinical point: failure to respond to standard anaphylaxis treatment is itself a diagnostic clue rather than a reason to give more. The site deliberately does not carry batch recalls, and the review page says so and explains why: recalls are lot-specific, handled through manufacturers and pharmacies, and any list would be stale within weeks while worrying more people than it helped. Regulatory changes that outlast a batch are carried; a recall ticker is not.

   Strong

6. TIME trial — morning versus evening dosing of antihypertensives

   Mackenzie IS, Rogers A, Poulter NR, Williams B, Brown MJ, Webb DJ, et al. Cardiovascular outcomes in adults with hypertension with evening versus morning dosing of usual antihypertensives in the UK (TIME study), The Lancet 2022;400(10361):1417–1425; with the MAPEC and Hygia Chronotherapy trials and the published critiques of them · 2022

   SupportsThe dosing-time section of the treatment review tool: a prospective, randomised, open-label, blinded-endpoint trial in 21,104 UK adults randomised to evening or morning dosing of their usual antihypertensives, finding no difference in the primary composite of vascular death or hospitalisation for non-fatal myocardial infarction or stroke over a median 5.2 years — hazard ratio 0.95 (95% CI 0.83–1.10) — and no difference in all-cause mortality.

   Read it asThis settles a question the internet has not caught up with. The MAPEC and Hygia Chronotherapy trials, both from a single Spanish group, reported very large reductions in cardiovascular events with bedtime dosing — Hygia an adjusted hazard ratio of 0.55 — and those effect sizes were widely judged implausible, with a 2022 systematic review finding all eight relevant studies at high risk of bias. TIME is the independent, large, pragmatic test and it is null. The practical conclusion the page draws is about adherence rather than chronobiology: take them when you will actually remember, which for most people is the morning. The one stated exception is documented high night-time blood pressure on ambulatory monitoring, where evening dosing may still be considered — and the page is explicit that this is a specific finding rather than something to infer from home readings.

   Strong

7. Inter-arm blood pressure difference and the ankle-brachial index

   Clark CE, Warren FC, Boddy K, et al., Associations between systolic interarm differences in blood pressure and cardiovascular disease outcomes and mortality: individual participant data meta-analysis — the INTERPRESS-IPD collaboration, Hypertension 2021;77:650–661; NICE NG136 and ESC hypertension guidance on measuring both arms; standard ankle-brachial index thresholds as used in vascular practice · 2021 and current

   SupportsThe blood pressure measurement page and calculator: that a systolic inter-arm difference of 10 mmHg or more is generally taken as the upper limit of normal and is present in about 3.6% of the general adult population, 7.4% of people with diabetes and 11.2% of people with hypertension; that in pooled individual participant data from more than 50,000 people across 24 cohorts with 10 years of follow-up a difference at that level was associated with higher cardiovascular events, cardiovascular mortality and all-cause mortality independent of standard risk scores; and that NICE and the ESC treat 15 mmHg as the threshold of additional cardiovascular risk. Also the ABI bands: above 1.40 non-compressible, 1.00–1.40 normal, 0.91–0.99 borderline, 0.90 or below indicating peripheral arterial disease, below 0.40 severe.

   Read it asTwo points the calculator is built around because they are routinely missed. First, an ABI above 1.40 is not a good result — it usually means the ankle arteries are too calcified to compress, renders the test uninterpretable and carries raised risk of its own; treating a high number as reassuring is the commonest misreading of the test. Second, an automatic upper-arm monitor is not validated at the ankle, so a home ABI is a prompt to ask rather than a result, and specifically not an all-clear. The tool also requires symptoms as well as numbers before raising acute aortic dissection, because a large inter-arm difference alone is common and the alternative would generate alarm in people who mostly have measurement variation. Sequential rather than simultaneous measurement exaggerates inter-arm differences, which is why the page insists on confirmation before acting.

   Strong

8. NICE NG136 — antihypertensive choice, monitoring and adverse effects

   NICE NG136, Hypertension in adults: diagnosis and management, visual summary updated July 2025; NHS Specialist Pharmacy Service guidance on ACE inhibitor and ARB monitoring and on calcium channel blocker peripheral oedema; MHRA drug safety updates on ACE inhibitors and ARBs in pregnancy (2007) and breastfeeding (2009) · 2019, updated 2025

   SupportsSection 7 of the blood pressure chapter: the A/C/D sequence — ACE inhibitor or ARB at step 1 for adults under 55 not of Black African or African-Caribbean family origin and for type 2 diabetes at any age, calcium channel blocker for those 55 or over or of Black African or African-Caribbean family origin, thiazide-like diuretic where a CCB is not tolerated because of oedema; A+C or A+D at step 2; A+C+D at step 3; and spironolactone at step 4 where potassium is 4.5 mmol/L or below, otherwise an alpha or beta blocker. That beta blockers are not recommended as routine initial therapy for hypertension. That an ARB is preferred to an ACE inhibitor in adults of African and Caribbean family origin, the committee’s stated reason being an increased risk of angioedema. That an ACE inhibitor and an ARB are not combined. That NICE names cough as a reason to switch from an ACE inhibitor to an ARB, and oedema as a reason to switch from a CCB to a thiazide-like diuretic. And that kidney function and potassium are checked within 1–2 weeks of starting or increasing an ACE inhibitor or ARB, then periodically.

   Read it asTwo things in that section are framed more strongly than the source documents phrase them, deliberately. The triple whammy — ACE inhibitor or ARB plus diuretic plus NSAID — is a well-recognised cause of acute kidney injury in primary care and is presented here as something a patient should actively ask about, because the NSAID is frequently bought over the counter and never mentioned. And sick day rules are described as usual practice with an explicit instruction to confirm the plan with the person’s own prescriber, because local guidance varies and the page should not be the authority for holding a prescribed drug. The incidence figures — ACE inhibitor cough at roughly 1 in 10 with a range of 1 in 20 to 1 in 3, amlodipine oedema rising from about 1 in 50 at 2.5 mg to about 1 in 10 at 10 mg and roughly threefold over placebo, orthostatic hypotension in about one in five over-60s and around half of care home residents — are drawn from the meta-analytic and review literature cited in the further reading rather than from NG136 itself.

   Strong

Alcohol

The observational and genetic literatures disagree here, so both are listed. The alcohol chapter says which it weights and why.

1. UK Chief Medical Officers’ Low Risk Drinking Guidelines, and the Guidelines Development Group report

   UK Chief Medical Officers · 2016, current

   SupportsThe 14-unit weekly guideline for both sexes, spread over three or more days; the UK unit as 10 ml / 8 g of pure alcohol; and the removal of the cardioprotection claim.

   Read it asThe 14-unit figure is the modelled point at which lifetime risk of dying from an alcohol-related cause reaches about 1%. The GDG states explicitly that these are guidelines for low-risk rather than safe drinking, and that risk continues to fall below the guideline. Quoting 14 units as a safety threshold misrepresents the source.

   Strong

2. Alcohol consumption and the risk of all-cause and cause-specific mortality — a linear and nonlinear Mendelian randomization study

   Kassaw NA, Zhou A, Mulugeta A, Lee SH, Burgess S, Hyppönen E, International Journal of Epidemiology 2024;53(2):dyae046 · 2024

   SupportsThat per additional unit (8 g) per day of genetically predicted intake the odds ratio was 1.27 for all-cause mortality, 1.30 for cardiovascular disease and 1.20 for cancer, with a linear relationship and no protective benefit at modest intake. 278,093 white-British UK Biobank participants; non-linear analysis used the doubly-ranked method and found no support for a J-shape.

   Read it asMendelian randomisation removes the sick-quitter problem and reverse causation but has its own assumptions — principally that the variants act on outcomes only through drinking. Non-linear MR methods are also newer and less settled than the linear form.

   Strong

3. Association between alcohol and cardiovascular disease: Mendelian randomisation analysis based on individual participant data

   Holmes MV, Dale CE, Zuccolo L, et al., BMJ 2014;349:g4164 · 2014

   SupportsThe earlier and more specific finding for coronary heart disease: no evidence of reduced risk at low intake once the genetic design removes confounding.

   Strong

4. The observational case that a protective effect is real

   Carr et al. 2024 and reviews weighting the cohort literature, including the European Heart Network report · 2024–25

   SupportsThat among 95 cohort and 27 case-control studies, most judged to have handled reverse causation adequately, a protective association with ischaemic heart disease persists in men, maximal around 20 g of ethanol daily.

   Read it asListed because the site does not present a live disagreement as settled. The dispute is about how much weight observational data retain once genetic data point the other way. This site weights the genetic evidence more heavily, consistently with how it treats lipids, and the chapter says so.

   Contested

5. Alcohol consumption — IARC Group 1 carcinogen classification, and the CMO review’s absolute risk figures

   International Agency for Research on Cancer; UK CMO Alcohol Guidelines Review · IARC classification long-standing; CMO review 2016

   SupportsCausal links to cancers of the mouth, throat, oesophagus, larynx, liver, colon and rectum, and breast; and the oesophageal cancer figures for men of roughly 6 per 1,000 in abstainers, 13 per 1,000 within the guideline and 25 per 1,000 at 14–35 units weekly.

   Read it asGroup 1 describes the strength of evidence that something causes cancer in humans, not the size of the risk. Reading it as “as dangerous as asbestos” is the standard misinterpretation and the chapter guards against it.

   Strong

Meal timing and time-restricted eating

1. TREAT (2020) and the 2022 randomised trial of time-restricted eating with calorie restriction

   TREAT, JAMA Internal Medicine 2020; and a randomised trial in the New England Journal of Medicine 2022 · 2020 and 2022

   SupportsThe Contested grade on the claim that time-restricted eating improves metabolic markers beyond the effect of eating less. TREAT found no significant benefit against a control eating pattern; the 2022 trial found time restriction added nothing to calorie restriction, with improvements in both arms tracking weight change.

   Read it asThese do not show TRE is useless — a rule about when is easier to follow than a rule about how much for some people, which is a real benefit. They show that an effect of timing independent of intake is unproven, which is the claim usually made for it.

   Contested

Blood glucose and prediabetes

What the CGM and prediabetes chapters rest on, and the boundary those chapters keep: none of this supports a claim that wearing a sensor improves health.

1. Individual variation in glycaemic response to carbohydrate

   Wu Y, Ehlert B, Metwally AA, et al., McLaughlin T, Snyder MP. Stanford; NCT03919877. Nature Medicine 2025;31(7):2232-2243. doi:10.1038/s41591-025-03719-2. Open access · Published 4 June 2025; checked September 2026

   SupportsThe individual-variation section of the CGM chapter. 55 adults with no diabetes diagnosis (27 euglycaemic, 26 prediabetes, one T2D by HbA1c on the day), seven standardised 50 g-carbohydrate meals each eaten at least twice after a 10-12 hour fast, sedentary for three hours. Rice was the most glucose-elevating meal on average. By highest peak rise: rice-spikers 35%, bread-spikers 24%, grape-spikers 22%. No participant peaked on beans or mixed berries. Potato-spikers were more insulin resistant with lower disposition index; grape-spikers more insulin sensitive. Preloads of 10 g fibre, 10 g protein or 15 g fat 10 minutes before rice reduced the peak only weakly overall (Cohen's d 0.12, 0.19 and 0.05 respectively) and did little in insulin-resistant participants. Replicate correlation (ICC) 0.26 to 0.73 by meal.

   Read it asRead the design before quoting the numbers. Single foods eaten alone on an empty stomach under standardised conditions — not free-living eating. The authors state the preload results cannot be extrapolated to a mixed meal. The subgroup analyses that produce the headline — mitigators failing in insulin resistance — rest on about six or seven participants per group, and two in one beta cell cell. The ethnicity finding (Asian participants more often rice-spikers) did not survive false-discovery correction (pFDR 0.21) and the blood-pressure finding in bread-spikers is labelled exploratory by the authors; this site states neither. Participants were blinded to their own glucose and no clinical outcomes were measured, so it does NOT support any claim that using a CGM improves health. It supports that responses differ between people and that the differences track insulin resistance and beta cell function. 55 people in California, recruited 2018-2023.

   Moderate

2. Healthier You: the NHS Diabetes Prevention Programme

   NHS England; national service evaluation of the first 324,699 referrals; cohort analysis of the first 36,000 patients · Programme from 2016; universal English coverage by 2018; checked August 2026

   SupportsEligibility at HbA1c 42-47 mmol/mol or fasting plasma glucose 5.5-6.9 mmol/L within 12 months, 18 or over, not pregnant, no diabetes diagnosis; a history of gestational diabetes qualifying regardless of HbA1c; nine months, at least 13 group sessions and a minimum 16 hours, with a digital alternative. Completer outcomes of about -3.6 kg and -2.1 mmol/mol. And the funnel: of 324,699 referred, 53% attended initial assessment, 36% attended at least one session, 19% completed.

   Read it asThe completion figure is the load-bearing one and the chapter uses it as an argument for choosing a format you will finish, not as an argument against the programme. It does NOT support any claim about how many people in the band progress to diabetes, which the chapter deliberately gives as varying rather than as a figure.

   Moderate

Adiposity, exercise and the surrogate problem

Waist-to-height, epicardial fat, and why a change on a scan is not a change in outcome.

1. Epicardial adipose tissue — semaglutide, SGLT2 inhibitors and the surrogate problem

   STOP trial prespecified analysis, Journal of the American College of Cardiology 2024 (semaglutide, n=115 with type 2 diabetes and known coronary disease); Khanna S et al., International Journal of Obesity 2026 (multimodality meta-analysis, 11 studies, 284 patients); and a systematic review and meta-analysis of randomised trials of SGLT2 inhibitors on adipose distribution, Diabetol Metab Syndr 2023 · 2023–2026

   SupportsThe epicardial fat section of the exercise and body composition chapter: that epicardial adipose tissue lies in direct contact with myocardium without an intervening fascial barrier and shares the coronary microcirculation; that semaglutide reduced epicardial fat volume by approximately 9% against placebo with a 5% increase in fat density over a year in the STOP imaging analysis; and that the SGLT2 evidence is inconsistent — a 2026 multimodality meta-analysis reporting a significant reduction (Hedges g −0.62) against a randomised-trial meta-analysis finding no significant effect on epicardial fat (SMD 0.03, 95% CI −0.52 to 0.58) while showing clear reductions in visceral, subcutaneous and hepatic fat.

   Read it asGraded Contested, and the reason is the same one the site applies to inflammation generally. The causal chain offered for epicardial fat — it secretes IL-6, IL-6 drives coronary disease, therefore reducing the fat prevents events — has a broken middle link: ZEUS suppressed IL-6 directly in over 6,300 people and produced a hazard ratio of 0.99. Epicardial fat volume is therefore a surrogate, and no trial has tested whether reducing it reduces events. The semaglutide finding is best read as a candidate mechanism for a cardiovascular benefit already established by outcome trials, not as a reason to take the drug. Source material for this section claimed SGLT2 inhibitors produce “the largest reductions in heart fat of anything studied”; the randomised evidence does not support that, and the chapter says so.

   Contested

2. NICE NG246 — waist-to-height ratio and central adiposity

   National Institute for Health and Care Excellence, NG246, Overweight and obesity management (2025), consolidating and replacing CG189; recommendations dated 2022 · 2022 recommendations, republished in NG246 2025

   SupportsThe waist-to-height bands in the exercise and body composition chapter: 0.40–0.49 healthy central adiposity, 0.50–0.59 increased, 0.60 or above high, applicable to both sexes and all ethnicities including adults with high muscle mass, and to be measured in adults with a BMI below 35 alongside BMI. That NICE asks clinicians to explain it as keeping the waist to less than half the height. And that the guidance asks for permission before discussing overweight, obesity or central adiposity and before taking measurements, with a refusal respected without judgement.

   Read it asWhy the ratio beats the circumference tables above it on the same page. Raw waist circumference needs separate thresholds by sex and by ethnicity; the 0.5 cut-off has held across both in the underlying evidence, which is why NICE applies one number to everyone. It is also the measure that catches central adiposity at a BMI in the healthy range, which is where BMI fails hardest. Above BMI 35 it stops discriminating, because nearly everyone in that group exceeds it — hence the restriction. NICE separately advises lower BMI thresholds for people of South Asian, Chinese, other Asian, Middle Eastern, Black African or African-Caribbean family background.

   Strong

Fatty liver disease

The evidence behind the MASLD chapter.

1. EASL–EASD–EASO Clinical Practice Guidelines on the management of MASLD

   Journal of Hepatology · 2024

   SupportsThe two-step pathway: FIB-4 under 1.3 as a rule-out, 1.3–2.67 as indeterminate requiring elastography or ELF, 2.67 and above prompting referral; and the alternative of a year of intensified cardiometabolic management before retesting in the indeterminate band.

   Read it asFIB-4’s negative predictive value is around 90% and its positive predictive value is poor — which is the design, not a flaw, but it means a raised score is a trigger for a second test rather than a diagnosis.

   Strong

2. Age thresholds and FIB-4 performance

   McPherson et al. and subsequent validation work; summarised in AASLD and ADA guidance · 2017–2025

   SupportsThat in people over 65 the lower FIB-4 threshold should be around 2.0 rather than 1.3, and that below about 35 FIB-4 performs poorly enough that elastography is the better first test.

   Read it asAge is a term in the FIB-4 formula, so the score drifts upward with age independently of the liver. Meta-analyses also report meaningful heterogeneity in FIB-4 performance across populations, so these thresholds are conventions with good support rather than fixed constants.

   Moderate

3. Discordance between FIB-4 and liver stiffness measurement

   Systematic reviews and meta-analyses of non-invasive tests in MASLD · 2025–2026

   SupportsThat up to about 30% of patients have discordant FIB-4 and elastography results, with liver stiffness generally the better reflection of true fibrosis stage.

   Moderate

4. MAESTRO-NASH — resmetirom in MASH with liver fibrosis

   Harrison SA, Bedossa P, Guy CD, et al., New England Journal of Medicine 2024;390:497–509 · 2024

   SupportsThat roughly 26–30% achieved MASH resolution without worsening fibrosis at 12 months against about 10% on placebo, with fibrosis improvement also more frequent. Basis for FDA accelerated approval in March 2024 and subsequently MHRA authorisation.

   Read it asHistological surrogate endpoint. The trial measured what the biopsy looked like, not whether people avoided cirrhosis, transplant or death. Accelerated approval is explicitly conditional on confirmatory outcome data. Graded Moderate for that reason despite being a large positive phase 3.

   Moderate

5. ESSENCE — semaglutide in MASH with F2–F3 fibrosis

   Phase 3 trial, part 1 at 72 weeks · 2025–2026

   SupportsMASH resolution without worsening fibrosis in about 63% against roughly 34% on placebo, and fibrosis improvement in about 37% against 22.5%. Basis for FDA and then MHRA approval for MASH with moderate-to-advanced fibrosis.

   Read it asSame surrogate caveat. Part 2 of the trial, which carries the clinical outcome endpoints, runs to 2029. Both UK approvals also precede a NICE technology appraisal, so authorisation is not yet NHS availability.

   Moderate

6. Cardiovascular disease as the leading cause of death in MASLD

   Consistent finding across cohort literature and current review syntheses · ongoing

   SupportsThe framing of the chapter: that the commonest cause of death in fatty liver disease is cardiovascular rather than hepatic, and that a MASLD diagnosis should raise the estimate of cardiovascular risk.

   Strong

Inherited cardiac conditions

1. NICE CG71 — Familial hypercholesterolaemia: identification and management

   National Institute for Health and Care Excellence · 2008, updated 2017

   SupportsFH prevalence around 1 in 250, the case for cascade testing of first-degree relatives, and the Simon Broome criteria referenced in the clinical pathway.

   Read it asThe under-diagnosis figure is the point: most people with FH in the UK are not identified, and cascade testing from a known case is the highest-yield step available in preventive cardiology.

   Strong

2. Hypertrophic cardiomyopathy and the inherited arrhythmia syndromes

   Standard cardiology reference sources and UK inherited cardiac conditions service practice · current

   SupportsHCM affecting roughly 1 in 500 and being among the commonest causes of sudden cardiac death in the young; long QT at roughly 1 in 2,000; and that assessment is usually ECG and echocardiography with genetic testing following a finding rather than replacing it.

   Read it asPopulation screening is not recommended anywhere, and the site says so: these conditions are uncommon, ECG interpretation in young athletic hearts generates false positives, and the consequences of a wrong label are substantial. A negative genetic test also does not exclude the condition.

   Strong

Diet and food

1. Saturated fat as a share of energy: the UK limit, the NICE lipid target, and the AHA position

   Scientific Advisory Committee on Nutrition, Saturated fats and health (2019); NICE NG238 (2023), lifestyle recommendations carried from CG181; American Heart Association dietary guidance for people who need to lower LDL (Eckel RH et al., Circulation 2014;129:S76–S99) · 2014–2023

   SupportsThe plan's saturated-fat target and the audit rules: the UK population limit of 10% of energy from saturated fat (SACN 2019); NICE advice for people at raised cardiovascular risk of 7% or less; the AHA's 5–6% for people who need to lower LDL. The plan was changed in round 162 from about 10% to about 7% by switching to 0% Greek yoghurt, turkey breast mince, chicken breast and egg whites; the grams-per-day reference (20 g) was replaced on the page by the share of energy, because the plan runs below 2,000 kcal.

   Read it asThe NICE 7% figure is from memory of CG181, whose lifestyle recommendations NG238 carried over; check against NG238's text. The Portfolio elements and the plant sterol effect (roughly 7–10% LDL reduction at 1.5–2.4 g a day) are sourced in the heart diet chapter and the calculator's sterol card.

   Strong

2. The breakfast products: Linwoods Flaxseed, Almonds, Brazil & Q10; Linwoods Flaxseed, Bio Cultures & Vitamin D; Rude Health Bircher Muesli; 99% psyllium husk

   Waitrose product listings (ingredients, allergy advice and nutrition per 100 g or per serving), read September 2026; typical values for 99% organic psyllium husk from several listings; the owner could not supply a link for the NaturaleBio product itself · September 2026

   SupportsThe breakfast-bowl section and the foods table on /eating/plan/. Q10 mix: 69% flaxseed, 10% each almonds, Brazil nuts, walnuts; per 100 g 572 kcal, fat 48 g (saturates 6.1), carbohydrate 3.8 (sugars 1.8), fibre 17, protein 22, salt 0.13, selenium 270 µg, omega-3 14 g; 20 mg CoQ10 per 20 g serving; labelled gluten-free. Vitamin D mix: 99.7% flaxseed, Bacillus coagulans GBI-30 (culture contains soya), vitamin D3; per 100 g 530 kcal, fat 42 (saturates 4.7), carbohydrate 4.1 (sugars 1.4), fibre 21, protein 24, vitamin D 25.5 µg, omega-3 20 g; labelled gluten-free. Bircher: oats 84%, dried apple 10%, raisins 4%, dried banana 2%; per 50 g 178 kcal, fat 3.5 (saturates 0.75), carbohydrate 29 (sugars 6.0), fibre 6.0, protein 5.0; not labelled gluten-free. Psyllium: about 190 kcal, 85 g fibre, 1.7 g carbohydrate, 0.2 g sugars per 100 g.

   Read it asLabels change; the page states them as read. The Linwoods cholesterol claim is the EU/UK authorised ALA claim, which applies at 2 g ALA a day.

   Strong

3. What the breakfast add-ons do: psyllium and LDL, CoQ10, selenium limits, vitamin D and cardiovascular events

   Jovanovski E et al., Am J Clin Nutr 2018;108:922–932 (psyllium meta-analysis); Mortensen SA et al., JACC Heart Fail 2014;2:641–649 (Q-SYMBIO); EFSA 2023 scientific opinion on the tolerable upper intake level for selenium; SACN, Vitamin D and Health (2016); Manson JE et al., NEJM 2019;380:33–44 (VITAL); Thompson B et al., BMJ 2023;381:e075230 (D-Health) · 2014–2023

   SupportsPsyllium at about 10 g a day lowers LDL by roughly 0.3 mmol/L; the page states "modestly and reliably" and does not quote the figure. CoQ10 evidence is for heart failure (Q-SYMBIO, 300 mg a day) and statin muscle symptoms; the calculator uses 100–200 mg; 20 mg is a fifth to a tenth of that. Selenium upper limit 255 µg a day (EFSA 2023). UK advice: 10 µg vitamin D a day in autumn and winter for everyone; VITAL and D-Health found no reduction in major cardiovascular events.

   Read it asFigures from memory; check before sign-off. The page is written qualitatively where the numbers are uncertain.

   Moderate

4. Buying fish: UK labelling, salmon pigment, raw-fish freezing, and mackerel sourcing

   Retained Regulation (EU) 1379/2013 (fishery product labelling); EFSA opinions on astaxanthin in salmonid feed; retained Regulation (EC) 853/2004 Annex III section VIII (freezing of fish for raw consumption) and FSA guidance; MSC (North-East Atlantic mackerel certificates suspended 2019) · 2004–2026

   SupportsThe "Buying fish you can trust" list: labels must show species, production method and catch area; farmed salmon colour comes from astaxanthin in feed, the pigment wild salmon obtain from crustaceans; fish for raw consumption must be frozen first unless farmed under conditions shown to be parasite-free, which is why sushi salmon is usually farmed; most UK mackerel is North-East Atlantic (FAO 27); Atlantic mackerel is low in mercury and king mackerel is a different species.

   Read it asWritten from memory of the regulations and statements; check before sign-off. The mackerel certification point is a sustainability matter, not a safety one, and the page does not present it as a safety issue.

   Moderate

5. The 14-day plan: omega-3 content of fish, the UK fish advice, and the saturated-fat limits

   Mozaffarian D, Wu JHY, J Am Coll Cardiol 2011;58:2047–2067 (EPA+DHA per 100 g, from USDA data); USDA National Nutrient Database SR27 as tabulated in Nutrients 2016 (PMC4967672); the AHA 2002 fish table; Scientific Advisory Committee on Nutrition, Advice on fish consumption: benefits and risks (2004); NHS guidance on fish and shellfish; NHS guidance on saturated fat · 2002–2026

   SupportsThe planning figures on /eating/plan/: farmed salmon about 1.97 g EPA+DHA per 100 g against wild about 1.84 (Mozaffarian and Wu), with sockeye listed from 0.68 g per 3 oz (AHA 2002) to much higher in the raw SR27 values; mackerel 1.2 g per 100 g (Mozaffarian and Wu) to 2.3 raw; sardines about 1.0; trout about 0.9–1.0. The plan counts wild Alaskan salmon at 1.0 g per 100 g, the conservative end, so its weekly totals (about 5.4 and 5.0 g) understate rather than overstate. SACN: two portions of fish a week, one oily, to provide about 450 mg a day of long-chain omega-3, roughly 3 g a week. NHS: at most four oily-fish portions a week for most adults, two for anyone pregnant, planning a pregnancy or breastfeeding. NHS saturated-fat limits: 30 g a day for men, 20 g for women.

   Read it asThe fish tables disagree by up to threefold for the same species, depending on season, feed, raw against cooked, and edition; the page says so and uses conservative figures. The farmed-salmon-is-higher point is consistent across the tables verified this round. The Greek-yoghurt saturated-fat ranges (roughly 5–10 g per 150 g of 5–10% yoghurt) are calculated from typical label values, not a source. The SACN and NHS limits were written from memory and should be checked against the current pages.

   Moderate

6. Re-evaluation of the risks to public health from bisphenol A in foodstuffs

   EFSA Panel on Food Contact Materials, Enzymes and Processing Aids · April 2023

   SupportsThe tolerable daily intake for BPA cut from 4 µg/kg bw/day to 0.2 ng/kg bw/day, about a 20,000-fold reduction, and the subsequent EU ban on BPA in food contact materials (Regulation 2024/3190) in force from 20 January 2025.

   Read it asGraded Contested deliberately. The European Medicines Agency and Germany’s BfR filed diverging opinions, and the UK Committee on Toxicity did not adopt the new TDI, judging the weight-of-evidence analysis insufficient. So UK and EU positions differ and the EU ban does not cover UK-market products. Two competent bodies, the same toxicology, different conclusions.

   Contested

7. Ultra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses

   Lane MM et al., BMJ 2024;384:e077310 · 28 February 2024

   SupportsThe ultra-processed food card on the exposures chapter and the UPF framing in the eating guide. 45 pooled analyses from 14 reviews, 9,888,373 participants, 32 adverse outcomes.

   Read it asObservational throughout. The authors graded only a subset as convincing, and the exposure measure — food frequency questionnaires and recall — is the weak link. Strong for “this association is real and consistent”, not for any specific mechanism.

   Strong

Obstetric emergencies and tocolysis

1. NICE NG25 — Preterm labour and birth

   National Institute for Health and Care Excellence · 2015, updated

   SupportsThat nifedipine is first-line tocolysis in the UK with atosiban where nifedipine is contraindicated, and that the purpose of tocolysis is to buy time for antenatal corticosteroids, magnesium sulfate where indicated, and transfer — not to prevent preterm birth.

   Read it asTransdermal glyceryl trinitrate does not appear in NG25. The chapter says so explicitly rather than implying guideline status.

   Strong

2. Tocolysis with nifedipine versus atosiban and perinatal outcome: an individual participant data meta-analysis

   Individual participant data meta-analysis of randomised trials comparing nifedipine with atosiban for 48 hours of tocolysis at 24+0 to 34+0 weeks · 2022

   SupportsThe comparable pair in the prolongation table on the preterm labour chapter: nifedipine 18 days to delivery against atosiban 10 days (HR 0.83, 96% CI 0.69–0.99), with NICU admission less frequent after nifedipine (46% against 59%).

   Read it asAnd the finding the chapter leans on hardest. Despite nearly twice the prolongation, there was no difference in the composite of neonatal morbidity and perinatal mortality, and neonatal mortality was numerically higher in the nifedipine arm (OR 1.4, 95% CI 0.60–3.4) — not significant, and flagged by the authors as warranting further investigation. This is the clearest available demonstration that days gained is a surrogate rather than the outcome, and it is why the chapter is organised around what the days are spent on. Individual participant data is the strongest design in this literature, which is why these two figures are the only directly comparable pair in the table.

   Strong

3. Pregnancy prolongation figures for nifedipine and magnesium sulfate

   Cochrane maintenance-tocolysis estimate; Sayin et al. 2004; a randomised maintenance trial reporting median 20 days (IQR 2.5–51); and a head-to-head against magnesium sulfate reporting 6.1 against 4.6 days · 2004–2017

   SupportsThe lowest and highest reported averages in the table: 5.4 days from Cochrane maintenance data and 26.7 days from Sayin, with an interquartile range of 2.5 to 51 days from one randomised trial giving a better sense of the spread than any average.

   Read it asThese figures are not measuring the same quantity. Some report days from trial entry to delivery, some prolongation from the start of treatment, some maintenance rather than acute tocolysis. The chapter shows the span as a range of study averages and states that comparing one agent’s bar against another is weaker than it looks — which is why indomethacin and betamimetics are left without a bar rather than given an invented number.

   Moderate

4. How long transdermal GTN prolongs pregnancy — the duration literature

   Lees C et al., Lancet 1994;343:1325–1326; Lees CC et al., Obstet Gynecol 1999;94:403–408; Bisits A et al., Randomized Nitric Oxide Tocolysis Trial (RNOTT), Am J Obstet Gynecol 2004;191:683–690; systematic review and meta-analysis, Am J Obstet Gynecol 2013; Gaikwad V, Gaikwad S, Tiwari P, Cureus 2024 · 1994–2024

   SupportsThe duration panel in the preterm labour chapter: a mean prolongation of 34 days in the original 1994 consecutive series of 13 women and 20 episodes, all of which responded; a mean of about 9.5 days in the GTN arm of RNOTT; a mean of 28.6 days in a 100-woman retrospective cohort at 27–35 weeks, with 90% reaching 48 hours and contractions inhibited in 92%; and gestation prolonged by 74% of the time remaining to 37 weeks in the 1999 multicentre trial, identical to ritodrine but with fewer preterm deliveries and fewer side effects.

   Read it asThe apparent contradiction is real and resolvable. Continuous measures — mean days gained, mean gestational age at delivery — consistently favour GTN. Binary endpoints against placebo — delivery within 48 hours, delivery before 28, 34 or 37 weeks — do not reach significance in the 2013 pooled analysis. An agent can move an average substantially while failing to shift a threshold most participants in both arms were going to cross or avoid regardless, and small trials are underpowered for the harder endpoint. The 2013 review also found GTN superior to β2-agonists for birth before 34 and 37 weeks, NICU admission and ventilation, and no different from nifedipine or magnesium. The retrospective 28.6-day figure is the weakest design in the group and should not be read as the expected result. On the application site: published series specify the anterior abdominal wall — for example Parveen et al. 2012, 65 women at 28–34 weeks with a 5 mg/12 h patch applied there — which is why the protocol uses the abdomen. That is a matter of following the studied route and of practicality, not of local drug action: transdermal GTN is a systemic delivery system, and the myometrium is centimetres deep behind fat, fascia and peritoneum even at term, well beyond the depth at which topical application produces meaningful tissue concentrations. GTN’s genuine local effects — venodilation for cannulation, anal fissure, Raynaud’s — all involve targets millimetres below the skin.

   Moderate

5. Tocolytics for delaying preterm birth: a network meta-analysis

   Wilson A, Hodgetts-Morton VA, Marson EJ, et al. Cochrane Database of Systematic Reviews 2022;8(8):CD014978 · 10 August 2022

   SupportsThe comparison in the preterm labour chapter: that all tocolytic classes — betamimetics, calcium channel blockers, magnesium sulfate, oxytocin receptor antagonists and nitric oxide donors — and their combinations are probably or possibly effective at delaying birth by 48 hours and 7 days against placebo or no treatment; that effects on neonatal and perinatal mortality and on maternal and neonatal infection were uncertain; and that betamimetics and combinations were the most likely to have treatment stopped for adverse effects. Betamimetic maternal harms against placebo: dyspnoea RR 12.09, palpitations RR 7.39, tachycardia RR 3.01, vomiting RR 1.91, headache RR 1.91. Nitric oxide donors: birthweight higher by a mean 425.53 g, fewer neonates under 2500 g at RR 0.40, gestational age 1.35 weeks more advanced — all low certainty. Combinations: fewer under 2500 g at RR 0.74, low certainty.

   Read it asThe load-bearing finding is the negative one. Delay is a surrogate; the outcome that matters was not established for any class. That is why the chapter frames tocolysis as instrumental — it creates the window for corticosteroids, magnesium neuroprotection and transfer, which are the interventions with outcome evidence. The nitric oxide donor signals are the most striking in the review and are low certainty on small numbers; they are reported on the page with that caveat attached rather than as a reason to prefer GTN.

   Strong

6. Comparison between nitroglycerin dermal patch and nifedipine for treatment of preterm labor

   Kashanian M, Zamen Z, Sheikhansari N, Journal of Perinatology 2014;34:683–687 · 2014

   SupportsThe adverse-effect comparison in the GTN protocol: headache comparable between GTN and nifedipine at around 28% each, hypotension roughly 25% against 28%, and nausea and tachycardia more common with GTN. Direction supported by a 2025 meta-analysis and the NICHD tocolytic synthesis.

   Read it asSingle randomised comparison, modest size. The figures are quoted as the source of a comparison rather than as population rates, and the protocol names the study on the page so the numbers can be weighed and superseded.

   Moderate

7. Head injury observation, delayed bleeding, and what detects it

   NICE NG232 on head injury assessment and early management; UK and Australian mild head injury pathways; case series and reviews of delayed subdural haematoma in anticoagulated patients; and the 2024 multidisciplinary chronic subdural haematoma guidelines · 2023–2026

   SupportsThe monitoring section in the head injury chapter: a responsible adult staying for 24 hours with the ability to rouse them; that a normal scan with a Glasgow Coma Scale below 15 at 24 hours prompts further imaging; the two-to-three-week window for a chronic subdural haematoma with its higher risk in older, anticoagulated and heavy-drinking patients; the subtle late signs including personality change, slowed thinking, unsteadiness and new incontinence; and that abnormal alertness, behaviour and cognition detect subtle brain injury better than the Glasgow Coma Scale.

   Read it asTwo points the chapter leans on. Anticoagulated patients can develop a significant haematoma after a normal initial scan — described at 48 hours in reported cases — which is why any head impact on an anticoagulant warrants assessment regardless of appearance. And pupil asymmetry is a late sign reflecting brainstem pressure, with roughly one person in five having benign anisocoria at baseline; the chapter therefore treats pupils as confirmatory rather than as a detection tool, and says that if pupils are what finally prompts the 999 call, the call is late.

   Strong

8. Aminoglycoside ototoxicity and the m.1555A>G variant

   NICE NG195 on neonatal infection; the PALOH implementation trial and subsequent UK national survey of neonatal first-line antibiotics; CPIC 2021 guidance; and NICE early value assessment of the Genedrive MT-RNR1 test · 2019–2026

   SupportsThe gentamicin section in the preterm labour chapter: that NICE recommends benzylpenicillin plus gentamicin first-line for suspected early-onset neonatal sepsis, chosen for its narrow spectrum rather than superior efficacy; that roughly 1 in 500 people carry m.1555A>G and can suffer profound irreversible sensorineural deafness from a standard exposure, equating to about 1,200 affected infants born annually in England and Wales and an estimated 120 a year exposed unnecessarily; that a point-of-care test detects the variant from a buccal swab in about 26 minutes and was implemented without delaying antibiotics, with only 3.3% of admissions untested; and that a cephalosporin-based regimen is the alternative in carriers.

   Read it asThe framing matters in both directions. Gentamicin is not a drug to avoid by default — it is UK first-line therapy and the alternative regimens drive more resistance. But for carriers the risk is not small, it is a single exposure away from permanent deafness, and CPIC advises avoidance unless infection severity and the absence of a safe alternative outweigh it. Because inheritance is mitochondrial, maternal genotype predicts the infant’s: one series found the variant in every mother of an affected child and in none of nearly 3,000 mothers of unaffected infants, which is the basis for the proposal to test mothers during preterm labour. Not routine UK practice, and a maternal-line family history of hearing loss is the available clinical flag.

   Strong

9. Newborn resuscitation and support of transition of infants at birth — thermal care

   Resuscitation Council UK Guidelines 2025; with UK regional prehospital neonatal frameworks · 2025

   SupportsThe gestation split in birth before arrival: at or near term, dry, hat, skin to skin. Preterm, and roughly below 32 weeks, dry the head only and put the body into a clear polyethylene bag or wrap without drying, head out, because evaporative heat loss outpaces towelling. Plus the fallbacks — if no bag can be found quickly, dry gently and wrap in a warm towel rather than delay, and a bag is still worth using on a baby already dried.

   Read it asThe plastic-wrap technique is supported by randomised trials and Cochrane review for preventing hypothermia in preterm and low-birth-weight infants, and hypothermia at admission is independently associated with mortality. Note the interaction: combining a bag with an exothermic mattress has produced admission hyperthermia in some series, so it is one measure, not a stack.

   Strong

10. Transdermal glyceryl trinitrate as a tocolytic

    Lees C et al., Lancet 1994;343:1325–1326; Lees et al. 1999 comparison with ritodrine; Smith GN et al., Canadian Preterm Labour Nitroglycerin Trial, Am J Obstet Gynecol 2007;196:37.e1–8; systematic review and meta-analysis, Am J Obstet Gynecol 2013; Duckitt K, Thornton S, Cochrane review of nitric oxide donors · 1994–2024

    SupportsThat GTN measurably reduces uterine contractility, that the original series arrested preterm labour in a small consecutive cohort, and that against ritodrine it achieved similar acute tocolysis with a lower overall preterm delivery rate and fewer side effects.

    Read it asGraded Contested and not Moderate. The effect on contractility is not in dispute; the clinical trial evidence is small, mixed, and has not persuaded UK guideline committees. More recent comparisons have found oral nifedipine superior on both efficacy and side-effect profile. The chapter presents the counterweight alongside the supportive trials rather than after them. Headache is very common and often severe; hypotension and transient tachycardia occur; contraindications include hypotension, nitrate sensitivity, glaucoma, raised intracranial pressure and obstructive hypertrophic cardiomyopathy. This is a prescribing decision, and the chapter frames any use as pre-arranged with an obstetric team and as a bridge to hospital rather than an alternative to it.

    Contested

Longevity, healthspan and environmental exposures

1. Healthy life expectancy, UK: between 2011 to 2013 and 2022 to 2024

   Office for National Statistics · Released 19 February 2026

   SupportsThe lifespan-versus-healthspan table on the longevity Essentials: 60.7 years in good health for men and 60.9 for women, decreases of 1.8 and 2.5 years against 2019–21, and the lowest since the series began.

   Read it as“Good health” here is self-reported in a survey, not clinically assessed. It measures something real but subjective, and the deprivation gap of roughly twenty years dwarfs anything individual behaviour explains.

   Strong

2. National life tables, UK

   Office for National Statistics · 2022–24 data

   SupportsLife expectancy at birth of 79.1 years for men and 83.0 for women.

   Strong

3. UK PM2.5 sources and targets — domestic burning as the largest single source

   Defra National Atmospheric Emissions Inventory, 2023 emissions; House of Commons Library briefing CBP-10543, Controls on domestic wood-burning and solid fuel use in England; Environment Act 2021 statutory targets and the Environmental Improvement Plan; WHO Global Air Quality Guidelines 2021 · 2021–2026

   SupportsThe air pollution chapter: that domestic burning of wood and solid fuel accounted for roughly a fifth of UK PM2.5 emissions in 2023, making it one of the two largest sources — similar in size to road transport once brake, tyre and road wear are included, and more than twice road transport's exhaust component, which is the like-for-like combustion comparison. Emissions rose steeply through the 2010s and have fallen somewhat since 2020, leaving it the only major source substantially higher than a decade ago; that England’s statutory target is an annual mean of 10 µg/m³ by 2040 with an interim target of 12 by 2028 and a 35% reduction in population exposure against 2018, against a WHO guideline of 5 µg/m³; that average exposure in England fell about 54% between 2003 and 2023 with two-fifths of that fall occurring in 2020; and that the most deprived fifth of areas experienced about 8% higher PM2.5 than the least deprived in 2023, with no clear narrowing over two decades.

   Read it asThe chapter is deliberately blunter about wood burners than about anything else on the register, and the justification is arithmetic rather than taste: it is one of the two largest sources, it is the only one substantially higher than a decade ago, survey evidence indicates it is supplementary rather than primary heating for most owners, and the exposure falls substantially on neighbours. It also states plainly that Ecodesign and Defra-exempt labels mean lower emissions than an open fire, not low emissions. The chapter separately insists that the cardiovascular benefit of outdoor exercise exceeds the harm of the pollution breathed during it at realistic UK exposures, because discouraging activity is the commonest way this subject causes net harm. Local shares vary widely and city-level figures differ from the national inventory; the chapter uses the national figure and says shares are higher in some places.

   Strong

4. Ultra-processed food — the umbrella review and the metabolic ward trial

   Lane MM et al., Ultra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses, BMJ 2024;384:e077310; Hall KD et al., Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain, Cell Metabolism 2019;30(1):67–77 · 2019 and 2024

   SupportsThe ultra-processed food chapter: the umbrella review of 45 meta-analyses covering approximately 9.9 million participants, reporting associations with 32 adverse health outcomes including roughly 50% higher cardiovascular mortality and about 21% higher all-cause mortality. And the Hall randomised crossover trial: 20 adults resident on a metabolic ward for four weeks, two weeks on each of an ultra-processed and an unprocessed diet matched for presented calories, sugar, fat, fibre and macronutrients, eating ad libitum — consuming approximately 500 kcal per day more on the ultra-processed arm, with weight gained on that arm and lost on the other.

   Read it asThe chapter argues that the category is weak and that the site should say so. NOVA sorts by production method rather than composition, so wholemeal supermarket bread, baked beans, plain yoghurt and most infant formula sit alongside cola and crisps; independent coders reproduce the classification imperfectly; and ultra-processed intake correlates strongly with free sugar, salt, energy density, low fibre and deprivation, making the processing effect hard to isolate. UK advisory opinion is now named on the page: the Scientific Advisory Committee on Nutrition's position statement on processed foods and health (July 2023) and its rapid evidence update (2025), which describe the associations as concerning while setting out NOVA's limitations and the likelihood of residual confounding, and make no ultra-processed-specific dietary recommendation. SACN's own critique of the Hall trial matches the caveats already on the page. The chapter resolves the tension by replacing the category with its underlying properties — eating rate, energy density, fibre and protein content, liquid calories — which sort the borderline foods correctly where NOVA does not.

   Strong

5. Bisphenol A in food contact materials — the EU ban and the Great Britain position

   Commission Regulation (EU) 2024/3190, published 31 December 2024, in force 20 January 2025, amending Regulation (EU) No 10/2011 and repealing Regulation (EU) 2018/213; European Commission implementation guidance, December 2025; and the UK Government publication BPA in food contact materials: EU rules and UK implications, with the FSA and Food Standards Scotland consultation response · 2024–2026

   SupportsSection 3 of the non-stick and food contact chapter: that the EU has prohibited BPA in food contact plastics, coatings, varnishes, printing inks, adhesives, silicones, rubber and ion-exchange resins, with the previous specific migration limit of 0.05 mg/kg repealed on entry into force and an 18-month transition, so the prohibition has applied since 20 July 2026, with later dates for repeat-use articles (2027) and professional equipment (2028); that the ban also covers bisphenol S, bisphenol AF and other harmonised-classified bisphenols, restricting the substitution route, though with conditional derogations for some existing uses; that paper and board are outside the ban unless combined with regulated materials; that it applies in Northern Ireland under the Windsor Framework; and that in Great Britain the FSA and FSS have consulted and stated an intention to introduce restrictions subject to the policy and legislative process, but had not legislated as at August 2026, with reporting suggesting domestic legislation is unlikely before mid-2027 and UK–EU sanitary and phytosanitary negotiations a wildcard. The divergence is therefore live rather than prospective.

   Read it asThis is the second Great Britain / Northern Ireland divergence recorded on this site, the other being fragrance allergen labelling, and the chapter draws the connection explicitly: “UK rules” is no longer a single answer for consumer chemistry. The practical consequence stated is narrow and checkable — a “BPA-free” claim is a weaker claim in GB than in the EU, because the substitute bisphenols are restricted there and not here. The chapter deliberately declines to turn this into alarm about tinned food: exposure has been falling through voluntary reformulation, the tolerable intake is a matter of genuine expert disagreement, and the nutritional cost of avoiding tinned pulses and fish would be real. The GB position is actively moving and should be rechecked.

   Strong

6. Where microplastics come from: the measurement studies behind the fifty-source ranking

   Yakovenko N et al., PLOS One 2025;20:e0328011 (indoor air); Catarino AI et al., Environ Pollut 2018;237:675–684 (dust fallout v mussels); Hussain KA et al., Environ Sci Technol 2023 (microwaving containers); Hernandez LM et al., Environ Sci Technol 2019;53:12300–12310 (tea bags), with the BfR re-assessment and the 2020 published comment; Qian N et al., PNAS 2024;121:e2300582121 (bottled water); Yadav H et al., Environ Sci Technol 2023;57:8225–8235 (chopping boards); Chaib I, Duflos G et al. (ANSES), J Food Compos Anal 2025 (drinks by container) · 2018–2025

   SupportsThe ranked sources page. Verified from the papers or their abstracts: indoor inhalation of 1–10 µm particles about 68,000 a day for adults and 47,000 for children, from 16 samples, car cabins about four times homes on median; dust fibres falling on a meal 13,731–68,415 particles a person a year against 123 from UK mussel consumption; microwaving some baby-food containers for three minutes released up to 4.22 million microplastics and 2.11 billion nanoplastics per cm²; a plastic tea bag at 95 °C about 11.6 billion microplastics and 3.1 billion nanoplastics per cup; bottled water about 240,000 particles a litre (110,000–370,000; about 90% nanoplastic; three US brands); chopping boards 14.5–71.9 million polyethylene or 79.4 million polypropylene particles a year on stated assumptions, with wooden boards shedding more (wood) particles; glass-bottled soft drinks, lemonade, iced tea and beer about 100 particles a litre, five to 50 times plastic bottles or cans, traced to cap paint, with water 4.5 v 1.6 and wine low.

   Read it asFrom memory and flagged in the review queue: Ranjan 2021 (lined paper cups, about 25,000 per 100 mL), Li 2020 (polypropylene baby bottles, up to about 16 million per litre), the Australian rice study, the non-stick coating study, and the 2025 chewing-gum conference data (not peer-reviewed). Dropped: a "17 to 68 quadrillion particles" figure for household paint in the supplied list, which could not be traced. The tea-bag count is stated as disputed. The ranking itself is a judgement, and the page says so; the counts cannot be compared across studies because the detection limits differ by orders of magnitude, which the page explains.

   Moderate

7. Microplastics and nanoplastics in atheroma — and the contamination objection

   Marfella R, Prattichizzo F, Sardu C, et al. Microplastics and Nanoplastics in Atheromas and Cardiovascular Events, New England Journal of Medicine 2024;390(10):900–910, doi:10.1056/NEJMoa2309822; with the published correspondence on sample contamination, NEJM doi:10.1056/NEJMc2404154 · 2024

   SupportsSection 1 of the microplastics chapter: a prospective multicentre observational study of 257 patients undergoing carotid endarterectomy for asymptomatic disease, detecting micro- and nanoplastics — principally polyethylene and polyvinyl chloride — in the excised plaque of 150 (58%) by pyrolysis gas chromatography–mass spectrometry, stable isotope analysis and electron microscopy. Over a median 34 months, the composite of myocardial infarction, stroke or death from any cause occurred in 30 of 150 with particles against 8 of 107 without — hazard ratio 4.53, 95% CI 2.00 to 10.27, p < 0.001. Plaque containing particles showed higher IL-18, IL-1β and CRP.

   Read it asThe published objection is given equal weight on the page, because it is not a quibble. Correspondence in the same journal notes that no pre-analytical anti-contamination procedures were described beyond the use of glass tubes, and no blank samples were taken from the operating theatre — and that a surgical environment contains abundant polyethylene and PVC, precisely the two polymers reported. If the particles originated there rather than in the patient, the finding does not stand. The authors dispute this and it is unresolved. The chapter also declines to repeat the widely quoted “credit card a week” ingestion figure, which derives from a modelling exercise rather than a measurement and has been criticised sharply, and treats organ-burden mass estimates as provisional given that quantification methods are still being validated and contamination is a pervasive problem across the field. The argument it lands on is that the additives plastics carry — bisphenols, phthalates, PFAS — have decades of human epidemiology where the particles have one contested study, and that the same practical actions reduce both, so nobody needs to resolve the science to act.

   Emerging

8. Fragrance allergen labelling — the EU expansion and the Great Britain divergence

   Commission Regulation (EU) 2023/1545 of 26 July 2023, amending Regulation (EC) No 1223/2009 as regards labelling of fragrance allergens, in force 16 August 2023, based on SCCS Opinion 1459/11; UK Cosmetics Regulation as retained and enforced by the Office for Product Safety and Standards; Windsor Framework arrangements for Northern Ireland · 2023, first deadline 31 July 2026

   SupportsThe labelling section of the fragrance and cleaning products chapter: that a perfume may be declared as the single word parfum, with only recognised contact allergens named separately; that Regulation (EU) 2023/1545 expanded that list from 24–26 substances to about 82, including several essential oils named as whole substances; that the declaration thresholds are 0.001% in leave-on and 0.01% in rinse-off products; and that the transition dates are 31 July 2026 for placing new products on the market and 31 July 2028 for selling through existing stock. That Great Britain has not adopted the expansion and still runs the original list of 26, while Northern Ireland follows EU cosmetics law.

   Read it asThis is a live divergence inside the United Kingdom, not a historical Brexit point, and it has practical consequences a shopper can see: the same product may carry a longer ingredients list in Belfast than in Birmingham, and the longer one is more informative. The chapter also makes the point that none of this covers household cleaning products at all — detergents, surface sprays, fabric conditioners and air fresheners fall under detergents and chemicals law rather than cosmetics law, where the labelling duty is narrower, so the cupboard containing the most fragrance carries the least disclosure. Any change to the GB position would come through OPSS; the position stated is as at August 2026 and should be rechecked.

   Strong

9. Flame retardants in UK furniture, and the 2025 exclusion of children's products

   Furniture and Furnishings (Fire) (Safety) Regulations 1988; Furniture and Furnishings (Fire) (Safety) (Amendment) Regulations 2025, in force 30 October 2025; House of Lords Library briefing on fire safety regulation reform, citing a UK Research and Innovation convened scientific group; House of Commons Environmental Audit Committee, Toxic chemicals in everyday life (2019); OPSS and Department for Business and Trade consultation on further reform · 1988–2026

   SupportsThe flame retardant section of the fragrance and cleaning chapter: that the 1988 regulations require an open-flame ignition test which manufacturers have largely satisfied by adding chemical flame retardants, and that a UKRI-convened scientific group has observed the UK has some of the highest usage of flame retardants in the world, in large part because of that test; that additive flame retardants migrate from foam and textile into household dust, which is the exposure route rather than skin contact with fabric; and that the Amendment Regulations 2025 removed a list of baby and children's products — cot mattresses under 170 cm, pram sacks, car seat inserts and similar — from the scope of the 1988 regulations, on the explicit basis that for those products the chemical exposure risk is greater than the fire risk.

   Read it asThis entry corrects a common claim rather than supporting it. The advice to wash new clothes “because of fire retardants” is aimed at the wrong product: ordinary UK clothing is not flame-retarded, nightwear excepted. Washing new clothes remains sensible for a different reason — residual disperse dyes, a recognised cause of contact dermatitis, and formaldehyde-based crease-resist finishes. The substantial domestic flame retardant exposure is upholstered furniture, and it travels in dust, which makes small children on floors the most exposed group and makes damp-dusting and HEPA vacuuming the effective response. Further reform is under consultation, including a smoulder test and a requirement that manufacturers justify using chemical flame retardants at all, so this position is expected to move.

   Strong

10. Cleaning products, indoor air chemistry and respiratory outcomes

    Svanes Ø et al., Cleaning at Home and at Work in Relation to Lung Function Decline and Airway Obstruction, American Journal of Respiratory and Critical Care Medicine 2018;197(9):1157–1163 (ECRHS cohort); with the indoor air literature on terpene–ozone reaction chemistry and secondary organic aerosol formation, and occupational asthma surveillance data on quaternary ammonium compounds · 2018 and current

    SupportsSection 4 of the chapter: 6,230 adults across 22 European centres with lung function measured three times over more than 20 years, finding FEV₁ declines of 22.1 mL/year in women responsible for cleaning at home and 22.4 in occupational cleaners against 18.5 in non-cleaners — an excess of about 3.6 and 3.9 mL/year, with FVC excesses of about 4.3 and 7.1. The authors compared the occupational effect to somewhat less than 20 pack-years of smoking, a cumulative exposure. And section 3: that limonene and other terpenes react with indoor ozone to form formaldehyde and secondary organic aerosol, so exposure peaks after cleaning stops rather than during it.

    Read it asThe caveat the chapter insists on, because almost every repetition of this study omits it: Svanes found no faster decline in the FEV₁/FVC ratio and no accompanying excess of obstructive airway disease. It is a measurable loss of lung volume over two decades, not a demonstration that domestic cleaning causes COPD, and the smoking comparison is to 20 pack-years — a cumulative lifetime exposure — rather than to twenty a day, which is the form it is almost universally repeated in and a materially larger claim. An earlier version of this site reported the 22 mL/year total decline as though it were the excess attributable to cleaning, overstating the effect roughly sixfold; that is corrected. No effect was seen in men, though the number of male occupational cleaners was small. The design is observational. The chapter grades airway irritation and occupational asthma Strong and the systemic claims made about the same products at domestic exposure Emerging, because those rest largely on mechanism.

    Strong

11. Sunscreen — application quantity, UVA rating, and UK filter regulation

    ISO 24444 SPF test methodology (2 mg/cm² application); the UK and EU requirement that UVA-PF be at least one-third of the labelled SPF for the UVA circle; MHRA and SAG-CS reviews of homosalate and oxybenzone; and the behavioural literature on higher-SPF products and time spent in sun · current

    SupportsThe sunscreen chapter: that SPF is tested at 2 mg/cm² while real-world application is typically a quarter to a half of that, and that because the protection relationship is exponential rather than linear, applying half the tested thickness gives approximately the square root of the labelled SPF — so an SPF 50 applied at half thickness behaves like roughly SPF 7. That about 35 ml covers an adult body, and the two-finger measure covers face and neck. That the UVA circle is a floor (UVA-PF at least one-third of SPF) rather than a grade, while the star rating expresses UVA protection relative to SPF. That the EU restricted oxybenzone to 2.2% in body products in 2022 and the UK followed in January 2026, and that the UK reviewed homosalate separately and retained 10% across all product types where the EU cut it to 7.34% in face products. And that people given higher-SPF products stay out longer, which is the mechanism of the false-security effect.

    Read it asThe square-root approximation is a rule of thumb, not a physical law, and real products deviate from it — but it is directionally right and it is the single most useful thing a reader can take from the chapter, because it reframes “which sunscreen” as the less important question. The chapter also states plainly that demonstrated systemic absorption of filters is not demonstrated harm: the FDA maximal-usage studies crossed a screening threshold that triggers further toxicology rather than establishing toxicity, and no human harm signal has emerged. The restrictions are precautionary; the skin cancer risk from non-use is not.

    Strong

12. Bemotrizinol, bisoctrizole and the modern UV filters

    US FDA final order adding bemotrizinol to permitted sunscreen actives, 9 June 2026 (effective 9 August 2026, up to 6%); with BASF and DSM-Firmenich product information for Tinosorb S / Parsol Shield, Tinosorb M and Uvinul A Plus, and L’Oréal for Mexoryl 400 · 1999–2026

    SupportsThat bemotrizinol (bis-ethylhexyloxyphenol methoxyphenyl triazine, Tinosorb S, Parsol Shield) is a broad-spectrum, highly photostable filter that also stabilises avobenzone, in European use since about 1999, and was approved in the United States on 9 June 2026 — its first new sunscreen active in around 25 years, with products expected from August 2026. That bisoctrizole (Tinosorb M) is a different compound and remains unapproved in the United States, as do Uvinul A Plus (DHHB) and the Mexoryls.

    Read it asThe chapter frames this as American news rather than British news, because that is what it is: UK and EU consumers have had these filters for a quarter of a century, and the practical implication here is only that they are worth choosing. Bemotrizinol and bisoctrizole are routinely conflated — including in the source material this section was written from — and the chapter separates them explicitly. The FDA data supported bemotrizinol at concentrations up to 6% with absorption below the agency’s systemic-exposure threshold.

    Strong

13. Sun exposure beyond vitamin D — nitric oxide, and the MISS cohort mortality data

    Liu D, Feelisch M, Weller R and colleagues on cutaneous nitric oxide mobilisation; Lindqvist PG et al., Melanoma in Southern Sweden (MISS) cohort; Chowdhury R et al. pooled 25(OH)D analysis (~849,000 participants); Alfredsson L et al., Int J Environ Res Public Health 2020;17:5014 · 2013–2020

    SupportsThe extended sun card on the exposures chapter: that UVA mobilises nitrite and nitrate stored in skin into circulating nitric oxide, lowering blood pressure independently of vitamin D, with warmth and dietary nitrate excluded as explanations; that low 25(OH)D is strongly associated with cardiovascular, cancer and all-cause mortality while supplementation trials show no corresponding benefit — the signature of a marker rather than a mechanism; and that in the MISS cohort the lowest sun-exposure group had the highest mortality, with sun avoidance estimated as a risk factor of magnitude similar to smoking.

    Read it asRead the smoking comparison with care. It is observational, from a fair-skinned Nordic cohort in which sun-seeking also tracks affluence, holidays, outdoor activity and better baseline health, and reverse causation is plausible because unwell people go out less. The 2020 review that popularised the attributable-death figures is a position paper arguing a case, and its numbers derive from vitamin D association data whose own authors state that supplementation does not help. The nitric oxide mechanism is the better-supported part. None of it changes the melanoma evidence, and the chapter says so.

    Moderate

14. Lung cancer deaths from indoor radon and the cost effectiveness and potential of policies to reduce them

    Gray A, Read S, McGale P, Darby S, BMJ 2009;338:a3110 · 2009

    SupportsThat radon is associated with around 1,100 UK lung cancer deaths a year, 3.3% of the total, and that over 85% of them arise from concentrations below the 100 Bq/m³ action level. Mean UK home: 21 Bq/m³.

    Read it asThe sub-action-level finding is the one that changes behaviour, and it is why the advice on this site is to test rather than to check whether your area is designated.

    Strong

15. Residential radon and lung cancer: collaborative analysis of individual data from 13 European case-control studies

    Darby S et al., BMJ 2005 · 2005

    SupportsThe dose-response relationship itself: linear, no threshold, roughly 16% excess relative risk per 100 Bq/m³ after correction for measurement error, and multiplicative with smoking.

    Strong

16. Excess winter deaths and illness and the health risks associated with cold homes (NG6)

    National Institute for Health and Care Excellence, with the Marmot Review Team’s The Health Impacts of Cold Homes and Fuel Poverty · NG6 published 2015; Marmot review 2011

    SupportsThat England and Wales record on the order of twenty thousand excess winter deaths in an average year, that these are predominantly cardiovascular and respiratory rather than hypothermia, and the Marmot finding that excess winter deaths in the coldest quarter of housing ran at close to three times the rate of the warmest quarter.

    Read it asExcess winter mortality is a noisy measure: England and Wales recorded 18,200 in 2013/14 and around 49,000 in 2017/18, so any single year is a poor guide and the site quotes an order of magnitude rather than a figure. Attribution to cold housing specifically, as opposed to winter generally, is modelled rather than measured — which is why the exposure is graded Moderate.

    Moderate

17. WHO Housing and Health Guidelines — systematic review on the effect of indoor cold on health

    World Health Organization, incorporating Public Health England’s minimum home temperature review (Wookey, 2014) · 2018

    SupportsThe 18 °C minimum indoor temperature used on this site, and the finding that the certainty of evidence for warming a cold home to 18 °C reducing cardiovascular mortality and morbidity is moderate.

    Read it asThe site takes WHO’s own grading rather than upgrading it. The blood-pressure mechanism is well evidenced; randomised evidence on deaths does not exist and is unlikely to.

    Moderate

18. Updated mortality burden estimates attributable to air pollution

    Evangelopoulos D, Walton H — UK Health Security Agency, Chemical Hazards and Poisons Report issue 28 · 2022, for 2019 exposure

    SupportsThe estimate of 29,000–43,000 attributable deaths a year in UK adults aged 30 and over, from long-term PM2.5 and NO₂.

    Read it as“Attributable deaths” is a statistical construct — an equivalent number spread across a population, not a count of identifiable people. The range is wide because PM2.5 and NO₂ effects overlap by an amount that is still being quantified. The widely quoted single figure of 40,000 is a midpoint judgement, not a measurement.

    Moderate

Bone, fracture and osteoporosis

Calcium and vitamin D, mechanical loading, and the drug risk figure most often misquoted.

1. Calcium and vitamin D supplementation and fracture incidence

   Zhao J-G, Zeng X-T, Wang J, Liu L. JAMA 2017;318(24):2466–2482; with Bolland MJ, Reid IR and colleagues on the cardiovascular signal · 2017

   SupportsThe calcium section of the bone chapter: a meta-analysis of 33 randomised trials and 51,145 community-dwelling adults over 50 finding no significant association between calcium, vitamin D, or the combination and the incidence of hip fracture, vertebral fracture, non-vertebral fracture or total fracture.

   Read it asThe population matters and the chapter says so. This is community-dwelling older adults, and it does not overturn the older institutional trials in deficient, sun-deprived care-home populations where benefit was shown, nor does it apply to people who cannot meet the requirement from diet or who are starting a bone-protecting drug, which is generally given alongside calcium and vitamin D replacement. The cardiovascular signal reported by Bolland and Reid is contested — not reproduced in all analyses, mechanism debated — and the chapter presents it as a reason not to assume a supplement without fracture benefit is free, rather than as established harm.

   Strong

2. Mechanical loading and bone — spaceflight countermeasures and the LIFTMOR trial

   NASA International Space Station bone countermeasure programme, comparing the interim resistive exercise device with the Advanced Resistive Exercise Device (ARED); and Watson SL, Weeks BK, Weis LJ, Harding AT, Horan SA, Beck BR, LIFTMOR randomised controlled trial, Journal of Bone and Mineral Research 2018;33(2):211–220 · 2013–2018

   SupportsThat astronauts lose hip and spine bone at approximately 1–1.5% per month in microgravity despite controlled nutrition and calcium intake; that cardiovascular exercise and resistance loading limited to around 300 lb did not prevent that loss, while ARED at up to around 600 lb did. And LIFTMOR: post-menopausal women with a T-score below −1.0, randomised to 8 months of twice-weekly, 30-minute, supervised high-intensity resistance and impact training at above 85% of one-repetition maximum against a low-intensity home programme, with bone density improving at spine and hip in the training group against decline in controls — femoral neck +1.1% against −0.4% — alongside improved function and height, and good tolerability.

   Read it asThe supervision is load-bearing in the evidence, not a disclaimer. LIFTMOR participants were screened and every session was coached; the finding is that heavy loading is safe when taught and supervised in this population, which is a narrower claim than the one usually repeated from it. The spaceflight comparison is observational and uncontrolled in the formal sense, but the strength of the inference comes from the design of the setting rather than randomisation: same people, same controlled diet, same duration, with load magnitude as the variable that changed.

   Moderate

3. Osteonecrosis of the jaw — incidence at osteoporosis versus oncology doses

   Pooled estimates from post-marketing surveillance and cohort studies of oral nitrogen-containing bisphosphonates in osteoporosis; with the American Association of Oral and Maxillofacial Surgeons position papers and national cohort data on tooth extraction as a risk factor · current

   SupportsThe proportionality argument in the bone chapter: that osteonecrosis of the jaw with oral bisphosphonates at osteoporosis doses is commonly estimated at around 1 case per 100,000 patient-years, was not reported as an adverse event across tens of thousands of patient-years in the randomised osteoporosis trials, and is a different proposition from the condition described in oncology patients receiving intravenous bisphosphonates at roughly ten times the dose. That tooth extraction during treatment is the clearest risk factor, raising risk substantially, which is the basis for completing dental work before starting where practical.

   Read it asThis entry exists because the risk is routinely misrepresented in the direction that causes harm. Source material for this chapter described jaw osteonecrosis as “the major potential side effect” of bisphosphonates; at osteoporosis doses that overstates it by orders of magnitude, and the fracture being prevented is common — hip fracture carries roughly 20–30% one-year mortality with a large proportion of survivors losing independent living. Reported incidence varies considerably between cohorts, with some national datasets giving higher figures than 1 per 100,000, particularly with longer duration and after extraction; the chapter gives the commonly cited estimate and names extraction as the modifiable factor rather than presenting a single precise rate. Atypical femoral fracture is covered separately as the usual reason for a treatment break.

   Strong

Chronic pain and fibromyalgia

What NG193 says about drugs, and the genetics.

1. NICE NG193 — chronic primary pain, and what it says about drugs

   NICE NG193, Chronic pain (primary and secondary) in over 16s: assessment of all chronic pain and management of chronic primary pain, April 2021 · 2021

   SupportsThe fibromyalgia treatment chapter, which is anchored to it: fibromyalgia falls within chronic primary pain, and the guideline recommends a supervised exercise programme and psychological therapy — CBT or acceptance and commitment therapy — as core treatment rather than as adjuncts; offers antidepressants as an option, all off-label for this indication; and advises against starting opioids, NSAIDs, paracetamol, benzodiazepines, corticosteroids and antiepileptics including the gabapentinoids.

   Read it asThis corrected the chapter's gabapentinoid row, and the correction matters because UK and US practice genuinely diverge here. Pregabalin is licensed for fibromyalgia in the United States and trial evidence exists — but a UK GP following NG193 will not usually initiate it, so describing it as “reasonable to try” misdescribed the position a reader would actually meet. The page now separates starting from already taking: for someone established on one the guidance is to review whether it helps and discuss reducing or stopping if not, never to stop abruptly. The opioid row already matched the guideline, including the slow-taper language, and is unchanged. The shape of NG193 is itself worth noticing: a pain guideline whose central recommendations are not medicines.

   Strong

2. The genetic architecture of fibromyalgia across 2.5 million individuals

   Kerrebijn I, Bjornsdottir G, Arbabi K, et al. (Chronic Pain Genomics Consortium, FinnGen, DBDS Genomic Consortium, Estonian Biobank Research Team, Genes & Health Research Team). Nature Medicine, published online 28 July 2026, doi:10.1038/s41591-026-04492-6; PMID 41001472 · 2026

   SupportsThe fibromyalgia topic: a multi-ancestry GWAS meta-analysis across 2,563,755 individuals — 54,629 cases and 2,509,126 controls — from 11 cohorts, about 90% of European ancestry, identifying the first 26 risk loci for fibromyalgia. The strongest association was rs149109767-A, an in-frame glutamic acid deletion in exon 58 of the 67-exon HTT gene (odds ratio 1.09, p = 2.2 × 10⁻¹²) — part of the Huntington-disease-associated ‘A1’ haplotype in Europeans but far from the exon 1 repeat expansion that causes the disease. Gene prioritisation implicated the HTT regulator GPR52 plus CAMKV, DCC, DRD2/NCAM1, MDGA2 and CELF4. Heritability was exclusively enriched within brain tissues and neural cell types, peaking in hippocampal dentate gyrus and enteric neurons. Genetic correlation above 0.7 with chronic low back pain, post-traumatic stress disorder and irritable bowel syndrome.

   Read it asTwo things are commonly reported loosely and the topic corrects both. First, the HTT signal is not the Huntington mutation, does not raise anyone’s risk of Huntington’s disease, and confers about 9% higher odds — a population shift, not a diagnosis. Second, “not autoimmune” overstates it: the study found heritability enriched exclusively in neural tissue and positive genetic correlation with autoimmune as well as psychiatric and somatic conditions. Those are compatible — where the heritable risk sits is the brain; what it shares genetic ground with is broad. The authors’ own limitations are stated on the page: the cohort is predominantly European, limiting generalisability, and cases were ascertained from diagnostic codes, so people not meeting full syndromic criteria may be included and undiagnosed people are necessarily excluded — which, given the diagnostic history, is not a random exclusion. A GWAS establishes where the heritable signal lies, not mechanism, and the topic says so and states plainly that no treatment changes as a result.

   Strong

Urinary infection and urinalysis

1. NICE NG109 — Urinary tract infection (lower): antimicrobial prescribing, with NG111, NG112 and NG113

   National Institute for Health and Care Excellence · 2018, updated 2025

   SupportsThe prescribing tables in the urinary infection chapter: nitrofurantoin as first choice with a 3-day course in non-pregnant women under 65 and 7 days in men; second choices of nitrofurantoin if not used first, pivmecillinam or fosfomycin; cefalexin in pregnancy at term or where nitrofurantoin is unsuitable; and the exclusions — nitrofurantoin avoided below an eGFR of 45 and at term in pregnancy because of neonatal haemolysis, trimethoprim contraindicated in the first trimester as a folate antagonist, and amoxicillin only on a sensitive culture. Also the nitrite and leukocyte decision table, and that a dipstick should not be used to diagnose urinary infection in people over 65 or with a catheter.

   Read it asThe point of the penicillin-allergy table is pivmecillinam. NICE offers it as a second choice, it is unfamiliar to many prescribers, and it is a beta-lactam — so it is the option most likely to be reached for by mistake in a penicillin-allergic patient. Nitrofurantoin, trimethoprim and fosfomycin are all structurally unrelated to penicillin and are the ones that survive. NICE also notes trimethoprim resistance is more likely with use in the previous 3 months and in older people in residential care, which is the group most likely to be given it blind.

   Strong

2. Multistix 10 SG reagent strips — pack insert, read times and interference

   Siemens Healthineers Multistix 10 SG product information and equivalent 10-parameter strip inserts; with standard urinalysis reference texts · current

   SupportsThe dipstick chapter: read times of 30 s for glucose and bilirubin, 40 s ketones, 45 s specific gravity, 60 s blood, pH, protein, urobilinogen and nitrite, and 120 s for leukocytes; that colour change on the leukocyte pad after 2 minutes is disregarded; the runover phenomenon in which acid buffer from the protein pad falsely lowers the pH reading if excess urine is not blotted; that any bilirubin is abnormal and the normal urobilinogen range is 0.2–1.0 mg/dL with 2.0 the transition; that the ketone pad measures acetoacetate rather than beta-hydroxybutyrate; that the blood pad detects peroxidase activity and so responds to myoglobin and to bacterial peroxidases; that leukocyte esterase detects neutrophils, eosinophils, basophils and monocytes but not lymphocytes; and the interference list — ascorbic acid above about 25 mg/dL causing false-negative nitrite, high specific gravity causing false-positive protein and false-negative glucose, blood and nitrite, cefalexin, tetracycline and high glucose lowering leukocytes, formaldehyde, oxidising agents, and azo dyes masking the pads. Optimum reading temperature 22–26 °C.

   Read it asThe manufacturer's multicentre data reports that 93% of samples negative for both leukocytes and nitrite were culture-negative, which is the basis for treating the strip as a rule-out rather than a rule-in test. The nitrite pad's two structural blind spots are the ones worth remembering: organisms that do not reduce nitrate (Enterococcus, Staph. saprophyticus, Pseudomonas) and insufficient bladder dwell time in someone voiding frequently because it hurts.

   Strong

Stroke

What the /stroke/ topic and the emergency stroke chapter rest on: the guidelines, the thrombolysis and thrombectomy trials, haemorrhage, ward care, and the secondary-prevention trials behind the targets. The two surfaces are held to the same figures by tools/audit-stroke-drift.php.

1. 2026 Guideline for the Early Management of Patients With Acute Ischemic Stroke (AHA/ASA)

   Prabhakaran S, Gonzalez NR, Zachrison KS et al.; Stroke, published 26 January 2026 (STR.0000000000000513); with the AHA "Top Things to Know" summary · January 2026

   SupportsThe US column throughout: tenecteplase 0.25 mg/kg or alteplase 0.9 mg/kg within 4.5 h (class 1) for disabling deficits regardless of NIHSS and without advanced imaging; extended-window thrombolysis 4.5–9 h or wake-up with imaging (2a) and 4.5–24 h with LVO and no thrombectomy (2b); DAPT rather than thrombolysis for non-disabling deficits; thrombectomy for selected large-core strokes and for basilar occlusion to 24 h with NIHSS ≥10; mobile stroke units and direct transport to thrombectomy centres; no adjuvant antithrombotics with thrombolysis; glucose 80–130 mg/dL not recommended; intensive BP lowering after EVT harmful and no target below 140; first paediatric recommendations.

   Read it asReplaces the 2018 guideline and its 2019 update. The class-of-recommendation labels in the topic (1, 2a, 2b) come from the guideline summaries and the TCTMD and UIC pharmacotherapy reports; check the full text before sign-off.

   Strong

2. National Clinical Guideline for Stroke for the UK and Ireland, 2023 edition

   Intercollegiate Stroke Working Party; Royal College of Physicians, SIGN and RCPI · April 2023

   SupportsThe UK column: alteplase or tenecteplase within 4.5 h regardless of age or severity; 4.5–9 h and wake-up thrombolysis on CT/MR perfusion or DWI-FLAIR mismatch; thrombectomy within 6 h for proximal occlusion and NIHSS ≥6, 6–24 h on ASPECTS plus mismatch, posterior circulation within 12 h; thrombolysis before thrombectomy irrespective of presenting hospital; ICH systolic 130–139 within 1 h for 7 days if 150–220 within 6 h of onset; idarucizumab for dabigatran, 4-factor PCC for factor Xa inhibitors with andexanet in trials only; DAPT 21 days clopidogrel or 30 days ticagrelor after TIA or minor stroke; carotid endarterectomy within 7 days; anticoagulation for AF within 5 days (mild) or 5–14 days; PFO closure under 60 within 6 months; clinic systolic below 130 and LDL below 1.8 for secondary prevention; three hours a day, five days a week of therapy.

   Read it asRead from the RCP announcement, the Clinical Medicine editorial and the Medscape summary rather than the full 2023 PDF; the acute-care chapter PDF was consulted for the INTERACT2 discussion. The 12-hour basilar window is stated in the topic alongside NICE's 2025 amendment to 24 h.

   Strong

3. NICE NG128 — Stroke and transient ischaemic attack in over 16s (2019, amended 2025), and TA990 tenecteplase (2024)

   National Institute for Health and Care Excellence · 2019–2025

   SupportsThrombolysis under TA990 (tenecteplase) and TA264 (alteplase) within 4.5 h; emergency-department staff may administer within an organised stroke service; thrombectomy within 6 h with thrombolysis for confirmed proximal anterior occlusion; and the 2025 amendment: consider thrombectomy up to 24 h, including wake-up strokes, for confirmed basilar or posterior cerebral occlusion with salvageable tissue on imaging.

   Read it asTA990 (July 2024) recommends tenecteplase as an option within 4.5 h; MHRA licensed it for stroke in April 2024. The topic says "NICE approved tenecteplase for the NHS in 2024" on that basis.

   Strong

4. Thrombectomy: HERMES pooled analysis, DAWN, DEFUSE 3, the large-core trials and the basilar trials

   Goyal M et al., Lancet 2016;387:1723–1731 (HERMES); Nogueira RG et al., NEJM 2018;378:11–21 (DAWN); Albers GW et al., NEJM 2018;378:708–718 (DEFUSE 3); Sarraj A et al., NEJM 2023;388:1259–1271 (SELECT2); Huo X et al., NEJM 2023;388:1272–1283 (ANGEL-ASPECT); Yoshimura S et al., NEJM 2022;386:1303–1313 (RESCUE-Japan LIMIT); Tao C et al., NEJM 2022;387:1361–1372 (ATTENTION); Jovin TG et al., NEJM 2022;387:1373–1384 (BAOCHE) · 2016–2023

   SupportsHERMES: mRS 0–2 at 90 days 46.0% v 26.5%, number needed to treat about 2.6, no excess symptomatic haemorrhage. DAWN (6–24 h, clinical-core mismatch): functional independence 49% v 13%. DEFUSE 3 (6–16 h, perfusion): 45% v 17%. SELECT2: 20% v 7% independent; ANGEL-ASPECT: 30.0% v 11.6%; RESCUE-Japan LIMIT in the same direction. ATTENTION (≤12 h) and BAOCHE (6–24 h): good outcome (mRS 0–3) about 46% v 23–24%, symptomatic haemorrhage about 5% v under 1%, mortality lower.

   Read it asEvery figure in this entry was written from memory of the papers and must be checked against them before sign-off. The large-core trials used different core definitions (ASPECTS 3–5 or core volume ≥50 mL) and the benefit is a shift across the disability scale more than a jump to independence; the topic says so.

   Strong

5. Late-window thrombolysis: EXTEND and WAKE-UP

   Ma H et al., NEJM 2019;380:1795–1803 (EXTEND); Thomalla G et al., NEJM 2018;379:611–622 (WAKE-UP) · 2018–2019

   SupportsEXTEND (alteplase 4.5–9 h or wake-up, perfusion mismatch): mRS 0–1 35.4% v 29.5%, symptomatic haemorrhage 6.2% v 0.9%, no mortality difference. WAKE-UP (alteplase, DWI-FLAIR mismatch): mRS 0–1 53.3% v 41.8%, symptomatic haemorrhage 2.0% v 0.4%. TRACE-III, TIMELESS and HOPE are in the Stroke group below with the emergency chapter.

   Read it asBoth trials stopped early, EXTEND for loss of equipoise after WAKE-UP and WAKE-UP for funding; the effect sizes are therefore less certain than the confidence intervals suggest. Figures from memory; check.

   Strong

6. Intracerebral haemorrhage: INTERACT2, INTERACT3 and ANNEXA-I

   Anderson CS et al., NEJM 2013;368:2355–2365; Ma L et al., Lancet 2023;402:27–40; Connolly SJ et al., NEJM 2024;390:1745–1755 · 2013–2024

   SupportsINTERACT2: systolic target below 140 within 1 h in 2,839 patients did not change death or major disability but improved the ordinal outcome; the basis of the 130–140 targets in both guidelines. INTERACT3: a care bundle (BP, glucose, temperature, anticoagulant reversal, all within 1 h) improved functional outcome in 7,036 patients across 121 hospitals. ANNEXA-I: andexanet controlled haematoma expansion more often than usual care but with more thrombotic events, which is why the UK guideline confines it to trials and NICE's existing appraisal does not cover ICH.

   Read it asThe topic cites the American 2022 ICH guideline's 130–140 statement through the 2026 acute guideline's summary; the ANNEXA-I numbers are not quoted.

   Strong

7. Ward care: CLOTS 3, AVERT, the DAPT trials and stroke-unit care

   CLOTS 3 Collaboration, Lancet 2013;382:516–524; AVERT Collaboration, Lancet 2015;386:46–55; Wang Y et al., NEJM 2013;369:11–19 (CHANCE); Johnston SC et al., NEJM 2018;379:215–225 (POINT); Johnston SC et al., NEJM 2020;383:207–217 (THALES); Gao Y et al., NEJM 2023;389:2413–2424 (INSPIRES); Stroke Unit Trialists' Collaboration, Cochrane 2020 · 2013–2023

   SupportsCLOTS 3: intermittent pneumatic compression reduced proximal DVT from 12.1% to 8.5% and improved survival; CLOTS 1 and 2 found stockings ineffective. AVERT: very early, high-dose mobilisation produced a favourable outcome in 46% v 50% with usual care. CHANCE and POINT: clopidogrel plus aspirin for 21 days after minor stroke or high-risk TIA reduces 90-day stroke; THALES: ticagrelor plus aspirin for 30 days; INSPIRES: benefit when started up to 72 h. Cochrane: organised stroke-unit care reduces death or dependency by about a fifth.

   Read it asFigures from memory of the papers; check before sign-off. The three-hours-a-day rehabilitation figure is the NCGS 2023 standard, not a trial result.

   Strong

8. SSNAP — Sentinel Stroke National Audit Programme, State of the Nation 2025 (April 2024 to March 2025)

   King's College London for HQIP · 2025–2026

   SupportsMedian door-to-needle time in England 56 minutes in 2024–25 (54 the year before); the new key indicators aligned to NCGS 2023 including advanced imaging and rehabilitation intensity; the audit measure of anticoagulant reversal or antihypertensive treatment within 1 h of arrival for ICH; the GIRFT statement that thrombectomy delivery falls short of the eligible population and the 10–15% eligibility estimate.

   Read it asThe topic does not quote a national thrombectomy rate because the figure was not confirmed from the report; it says "a fraction of the 10–15% estimated to be eligible". If you want the rate stated, it needs the report page.

   Strong

9. Time to treatment with intravenous alteplase and outcome in stroke — pooled analysis

   Lees KR, Bluhmki E, von Kummer R, et al. Lancet 2010;375:1695–1703; with Emberson J, Lees KR, Lyden P, et al. individual patient data meta-analysis, Lancet 2014;384:1929–1935 · 2010 and 2014

   SupportsThe time-benefit table in the stroke chapter. Adjusted odds of a favourable 3-month outcome, modified Rankin 0–1: 2.55 (1.44–4.52) for 0–90 minutes, 1.64 (1.12–2.40) for 91–180, 1.34 (1.06–1.68) for 181–270, and 1.22 (0.92–1.61) for 271–360, the last no longer significant. Mortality odds rising with delay: 0.78, 1.13, 1.22, 1.49 across the same intervals. Number needed to treat about 14 in the 3-to-4.5-hour window.

   Read it asThe finding that makes the urgency argument, and it is easy to miss. Large parenchymal haemorrhage occurred in 5.2% of treated patients against 1.0% of controls with no relationship to how quickly treatment started. The benefit shrinks steeply with delay; the bleeding risk does not move. So the deteriorating balance is entirely a matter of giving up upside rather than accruing harm — which is why the chapter presents the two together.

   Strong

10. Tenecteplase against alteplase, and why the practical choice differs from the trial result

    AcT randomised trial (Canada); ATTEST-2, Lancet Neurology 2024; RAISE (reteplase); with the tenecteplase individual patient data meta-analysis, Int J Stroke 2016 · 2016–2024

    SupportsThat the two are equivalent on outcome — AcT: 36.7% against 35.9% free of significant disability at 90 days, odds ratio 1.03 (0.81–1.31), symptomatic haemorrhage 3.2% against 3.1%; ATTEST-2 confirming non-inferiority without superiority; reteplase also no different. And that tenecteplase is given as a single bolus of 0.25 mg/kg to a maximum of 25 mg, against a bolus plus one-hour infusion for alteplase.

    Read it asThe chapter argues for tenecteplase on logistics rather than on efficacy, and says so. Given that the odds of a good outcome fall from 2.55 to 1.34 across the treatment window, removing an infusion pump from the pathway — and with it the delay in starting and the difficulty of transferring a patient mid-infusion to a thrombectomy centre — converts into outcome even though the head-to-head trials show no difference. It is an unusual and instructive case: two equivalent drugs, one clearly better in a hospital.

    Strong

11. Thrombolysis beyond 4.5 hours — three trials, three answers

    TIMELESS, New England Journal of Medicine 2024 (NCT03785678); TRACE-III, New England Journal of Medicine 2024;391:203–212; HOPE, JAMA 2025;334(9); with a 2025 meta-analysis in Stroke · 2024–2025

    SupportsThe extended-window section of the stroke chapter. TIMELESS: tenecteplase 4.5–24 h in mostly thrombectomy-treated patients, no better than placebo; symptomatic intracranial haemorrhage 3.2% against 2.3%, mortality 19.7% against 18.2%. TRACE-III: tenecteplase 4.5–24 h in large-vessel occlusion without access to thrombectomy, 33.0% free of disability at 90 days against 24.2% (relative rate 1.37, 95% CI 1.04–1.81), symptomatic haemorrhage 3.0% against 0.8%, mortality unchanged at about 13%. HOPE: alteplase, more haemorrhage at 3.8% against 0.5%, no mortality difference.

    Read it asGraded Contested because the trials disagree, and the chapter presents them rather than a conclusion. On the TRACE-III figures the absolute benefit is about 8.8 percentage points, giving roughly 11 needing treatment for one extra disability-free outcome, against roughly 45 for one extra symptomatic bleed. That is a favourable trade for someone facing severe disability and it is a real trade. It applies only to the population studied: large-vessel occlusion, salvageable tissue on perfusion imaging, and no thrombectomy available. Within 4.5 hours the evidence remains Strong. Around 60% of patients arrive more than eight hours from onset, which is why the question matters so much.

    Contested

12. Standard-window thrombolysis, thrombectomy, and secondary prevention

    Pooled analyses of alteplase trials; thrombectomy trials with imaging selection to 24 hours; and UK secondary-prevention practice following NICE guidance · current

    SupportsThrombolysis within 4.5 hours of last known well with benefit falling steeply inside the window; symptomatic intracranial haemorrhage of a few per cent against about 1% untreated, with more early deaths but no difference in mortality at three to six months. Thrombectomy for large-vessel occlusion to 6 hours, extending to 24 in selected patients. Antiplatelet then clopidogrel, with a short dual course after minor stroke or high-risk TIA; anticoagulation replacing antiplatelet where atrial fibrillation is found; high-intensity statin; blood pressure; carotid surgery for symptomatic severe stenosis, most effective within two weeks.

    Read it asThe chapter states that a single ECG finds only a minority of paroxysmal atrial fibrillation and that prolonged monitoring finds substantially more — which matters because the finding changes treatment from an antiplatelet to an anticoagulant. It is the most consequential omission in an otherwise complete workup.

    Strong

13. 2021 Guideline for the Prevention of Stroke in Patients With Stroke and Transient Ischemic Attack (AHA/ASA)

    Kleindorfer DO et al.; Stroke 2021;52:e364–e467 · 2021

    SupportsThe US column of the prevention and cause chapters: BP below 130/80; high-intensity statin with LDL below 70 mg/dL and ezetimibe then PCSK9 inhibition; aspirin, clopidogrel or aspirin-dipyridamole long term with no long-term DAPT; long-term rhythm monitoring reasonable after cryptogenic stroke; no empirical anticoagulation for ESUS; PFO closure in the under-60s with PASCAL; carotid endarterectomy within two weeks; pioglitazone reasonable in insulin resistance; dissection managed with either antithrombotic; warfarin for antiphospholipid syndrome.

    Read it asThe current US secondary-prevention document in September 2026; the 2026 acute guideline defers to it for prevention. Read from the guideline summary and the AHA top-ten.

    Strong

14. Anticoagulation timing after AF-related stroke: ELAN, OPTIMAS, TIMING and the CATALYST pooled analysis

    Fischer U et al., NEJM 2023;388:2411–2421 (ELAN); Werring DJ et al., Lancet 2024;404:1731–1741 (OPTIMAS); Oldgren J et al., Circulation 2022;146:1056–1066 (TIMING); CATALYST collaboration, Lancet 2025 · 2022–2025

    SupportsELAN (2,013 patients): early DOAC (within 48 h for minor or moderate stroke, day 6–7 for major) against later; the 30-day composite was estimated between 2.8 points lower and 0.5 higher with early treatment, with no excess symptomatic haemorrhage in any infarct size. OPTIMAS (UK, 3,648 patients): DOAC within 4 days non-inferior to 7–14 days at 90 days. TIMING (888 of 3,000 planned): 6.89% v 8.68%. CATALYST individual-patient meta-analysis of all four: early initiation within 4 days reduced the 30-day composite of recurrent ischaemic stroke, symptomatic ICH or unclassified stroke.

    Read it asThe topic states that the evidence supports the early end of the NCGS 2023 windows (5 days mild, 5–14 moderate/severe) and that neither guideline has yet incorporated CATALYST. ELAN and OPTIMAS figures verified from the abstracts; the OPTIMAS patient count is from memory.

    Strong

15. Treat Stroke to Target, SPARCL, and the LDL target after stroke

    Amarenco P et al., NEJM 2020;382:9–19 (TST); Amarenco P et al., NEJM 2006;355:549–559 (SPARCL) · 2006, 2020

    SupportsSPARCL: atorvastatin 80 mg after stroke or TIA reduced stroke from 13.1% to 11.2% over about five years, with a small excess of haemorrhagic stroke. TST: an LDL target below 1.8 mmol/L (70 mg/dL) against 2.6–3.6 reduced major cardiovascular events from 10.9% to 8.5% over a median 3.5 years. Basis for the identical UK (NCGS: LDL <1.8, non-HDL <2.5) and US (<70 mg/dL) targets.

    Read it asFigures from memory of the papers; check. TST was stopped early for administrative reasons and the French cohort drove most of the result.

    Strong

16. Occult atrial fibrillation after stroke: CRYSTAL-AF and STROKE-AF; empirical anticoagulation for ESUS: NAVIGATE ESUS and RE-SPECT ESUS

    Sanna T et al., NEJM 2014;370:2478–2486; Bernstein RA et al., JAMA 2021;325:2169–2177; Hart RG et al., NEJM 2018;378:2191–2201; Diener HC et al., NEJM 2019;380:1906–1917 · 2014–2021

    SupportsCRYSTAL-AF: an implantable loop recorder after cryptogenic stroke detected AF in 8.9% v 1.4% at 6 months, 12.4% v 2.0% at 12 months and about 30% by 3 years. STROKE-AF: 12.1% v 1.8% at 12 months in strokes attributed to large- or small-vessel disease. NAVIGATE ESUS (rivaroxaban) and RE-SPECT ESUS (dabigatran): no reduction in recurrent stroke with empirical anticoagulation, and more bleeding with rivaroxaban. Together the basis for "monitor for AF, and anticoagulate the AF you find, not the AF you assume".

    Read it asFigures from memory of the papers; check. Whether device-detected brief AF warrants anticoagulation remains contested (ARTESIA, NOAH-AFNET 6) and the topic does not take a position on it.

    Strong

17. PFO closure: CLOSE, REDUCE and RESPECT; carotid endarterectomy: NASCET and the timing analysis

    Mas JL et al., NEJM 2017;377:1011–1021 (CLOSE); Søndergaard L et al., NEJM 2017;377:1033–1042 (REDUCE); Saver JL et al., NEJM 2017;377:1022–1032 (RESPECT); NASCET, NEJM 1991;325:445–453; Rothwell PM et al., Lancet 2004;363:915–924 · 1991–2017

    SupportsCLOSE: recurrent stroke 0% v 6% over about five years with closure in under-60s with a large shunt or atrial septal aneurysm; REDUCE: 1.4% v 5.4%; RESPECT long-term follow-up in the same direction; new AF after the device about 1 in 20, mostly transient. NASCET: for symptomatic 70–99% stenosis, ipsilateral stroke at two years 9% v 26%, about one stroke prevented per six operations. Rothwell 2004 pooled analysis: benefit greatest within two weeks of symptoms and largely gone after twelve; basis for the UK seven-day and US two-week targets.

    Read it asFigures from memory of the papers; check.

    Strong

18. Antithrombotics after intracerebral haemorrhage: RESTART, and the 2024–25 anticoagulation trials

    RESTART Collaboration, Lancet 2019;393:2613–2623; ENRICH-AF and PRESTIGE-AF, reported 2024–2025 · 2019–2025

    SupportsRESTART: restarting antiplatelet therapy after ICH did not increase recurrent haemorrhage and probably reduced vascular events. ENRICH-AF and PRESTIGE-AF: restarting anticoagulation for AF after ICH reduced ischaemic events and increased recurrent haemorrhage, with the balance depending on haematoma location and the presence of amyloid angiopathy. Basis for the "specialist decision by ICH location" position on both surfaces.

    Read it asThe ENRICH-AF and PRESTIGE-AF statements are written at summary level from memory of the presentations and must be checked against the publications before sign-off. Graded Moderate for that reason.

    Moderate

19. Young and cryptogenic stroke workup: Fabry disease prevalence, and the NCGS 2023 and AHA/ASA 2021 positions

    Rolfs A et al., Lancet 2005;366:1794–1796; Rolfs A et al. (SIFAP1), Stroke 2013;44:340–349; NCGS 2023 sections on stroke in young adults, CADASIL and cerebral amyloid angiopathy; AHA/ASA 2021 secondary prevention guideline · 2005–2023

    SupportsThe stroke every-test chapter's tier three: antiphospholipid antibodies with repeat at 12 weeks and warfarin if positive; the autoimmune and vasculitis screen on a clinical clue; haemoglobin electrophoresis; JAK2 on an abnormal count; homocysteine only for suspected homocystinuria; Fabry testing in unexplained stroke under 55 (Rolfs 2005 found Fabry in about 4.9% of men and 2.4% of women with cryptogenic stroke aged 18–55; SIFAP1, larger and stricter, found definite Fabry in about 0.5%, hence the chapter's "around 1%"); NOTCH3 for the CADASIL phenotype; the other monogenic genes on a phenotype; inherited thrombophilia only where a venous route is plausible; syphilis and HIV in young stroke; no MTHFR, no routine panels, lipoprotein(a) once.

    Read it asWritten from memory of the guideline sections and the two cohorts; check before sign-off. Graded Moderate because the Fabry prevalence figures vary tenfold between the two studies and the chapter chose a conservative middle.

    Moderate

20. DVLA — stroke and TIA, and the return to driving

    Driver and Vehicle Licensing Agency, "Assessing fitness to drive: a guide for medical professionals", neurological conditions chapter · Current edition, September 2026

    SupportsGroup 1 (car, motorcycle): must not drive for one month after a stroke or TIA; may resume without notifying the DVLA if clinical recovery is satisfactory and there is no other bar. Group 2 (bus, lorry): licence refused or revoked for one year, with medical assessment. Multiple TIAs, seizures, visual field loss and persistent limb weakness alter the rules.

    Read it asRules change; the page is the authority. Checked September 2026.

    Strong

Brain health, dementia and cognition

The dementia chapter and the menopause-and-cognition material. Prevention estimates here are population attributable fractions, not promises to an individual.

1. Lancet Commission on dementia prevention, intervention and care 2024

   Livingston G et al., Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission, The Lancet, published 31 July 2024 and presented at AAIC 2024 · 2024

   SupportsThe dementia chapter: 14 modifiable risk factors with an estimated 45% of dementia worldwide potentially preventable or delayed if all were eliminated. Two factors added in 2024: raised LDL cholesterol in midlife, at about 7% of cases, on evidence including a meta-analysis of three UK cohorts in which each 1 mmol/L higher LDL was associated with approximately 8% higher all-cause dementia incidence; and untreated vision loss in later life, at about 2%. Population attributable fractions recalculated using the Norwegian HUNT study. The two largest contributors globally are hearing impairment and raised LDL. And that addressing these factors matters even in people at high genetic risk.

   Read it asThe chapter spends its first section on what the 45% figure is not, because it is quoted everywhere as though it were a personal risk reduction. It is a population attributable fraction: it assumes causality, assumes complete elimination, and cannot have its components added together because the factors overlap heavily in the same people. Much of it — education, air pollution, deprivation — is structural rather than individual. What it does establish is direction and priority, which is genuinely new: dementia was widely presented as bad luck twenty years ago. The chapter is placed in the heart topic rather than longevity because the modifiable risk is largely vascular and every relevant factor already has a chapter there.

   Strong

2. Herpes zoster vaccination and dementia — the Wales natural experiment

   Eyting M, Xie M, Michalik F, Heß S, Chung S, Geldsetzer P. A natural experiment on the effect of herpes zoster vaccination on dementia, Nature 2025;641(8062):438–446, doi:10.1038/s41586-025-08800-x; with Xie M et al., Cell 2025;188(25):7049–7064, on effects at different stages of the disease course · 2025

   SupportsSection 5 of the dementia chapter: that Welsh zoster vaccination eligibility from 1 September 2013 was determined by exact date of birth, those born before 2 September 1933 being permanently ineligible; that vaccination uptake rose from about 0.01% among those one week too old to 47.2% among those one week younger; and that regression discontinuity analysis across that boundary found the live-attenuated vaccine reduced new dementia diagnoses over seven years by approximately one fifth. Negative-control analyses showed no effect on other causes of mortality or morbidity, and no increased uptake of other vaccinations or preventive measures.

   Read it asThe design is the point, and the chapter explains why. Ordinary vaccine research is defeated by the fact that people who get vaccinated are healthier and more engaged with healthcare than people who do not; a birth-date cutoff removes that entirely, and the negative controls behaving correctly is what a confounded result usually fails. Replicated in Australia by the same method, with concordant signals reported from England, New Zealand and Canada, and a separate US analysis of the Zostavax-to-Shingrix switch pointing the same way. Caveats stated on the page: not randomised; the effect appeared largely in women in both natural experiments and the reason is unknown; the mechanism is unsettled; and the product studied was the older live-attenuated vaccine rather than the recombinant one now in use. A randomised trial is being sought. The chapter draws a narrow practical conclusion — take up the vaccine you are offered rather than defer it — and explicitly does not present it as a dementia treatment.

   Moderate

3. Menopause: identification and management (NG23), and the randomised evidence on hormone therapy and cognition

   NICE NG23 (2024 update); KEEPS-Cog and ELITE, with the KEEPS continuation (2024); Andy et al., Frontiers in Endocrinology 2024; systematic review in Lancet Healthy Longevity, December 2025 · Checked August 2026

   SupportsThat NICE says do not offer hormone therapy to prevent dementia and describes the evidence on cognitive symptoms as uncertain; and that for cognitive PERFORMANCE started near the transition the randomised answer is neutral - KEEPS-Cog and ELITE both null, the KEEPS continuation null about a decade on.

   Read it asThe distinction is the point. The randomised question has been answered for cognitive performance and remains open for dementia INCIDENCE, where the observational cohorts conflict. An earlier version of the chapter said the timing hypothesis had never been tested, which was right for incidence and wrong for performance. None of this supports any statement about hormone therapy for SYMPTOMS, which this site does not cover.

   Moderate

Fainting

1. Collapse around exercise — postural collapse, heat illness and exercise-associated hyponatraemia

   Sports and exercise medicine consensus positions on exercise-associated collapse; heat illness guidance; and the exercise-associated hyponatraemia consensus statements · current

   SupportsThe exercise section of the fainting chapter: that collapse after stopping endurance effort is usually exercise-associated postural collapse, caused by loss of the leg muscle pump while vessels remain dilated, and resolves with recumbency and elevation of the legs and pelvis; that collapse during effort is treated as cardiac; that mental state rather than temperature separates heat exhaustion from heat stroke; and that exercise-associated hyponatraemia should be suspected where large volumes of plain water have been drunk with no weight loss or with weight gain, in which case further hypotonic fluid is harmful.

   Read it asThe during-versus-after distinction is the load-bearing one and is the reason the section exists. The hyponatraemia scenario is uncommon but is the one case in this chapter where the instinctive treatment — give them water — makes the patient worse, which is why it is named despite its rarity. Auditory prodromal symptoms, including muffling and popping, are a recognised part of presyncope and are described by patients far more often than they appear in warning-sign lists.

   Moderate

2. Syncope — assessment, red flags and physical counter-pressure

   ESC syncope guidelines and the NICE transient loss of consciousness guidance; with the Physical Counterpressure Manoeuvres Trial (PC-Trial) for the manoeuvres · current

   SupportsThe structure of the fainting chapter: the three high-risk features — syncope during exertion, syncope while supine, and syncope with no prodrome — as the features that reclassify a faint as a cardiac event; orthostatic hypotension defined as a fall of 20 mmHg systolic or 10 mmHg diastolic within three minutes of standing; the value of an ECG in unexplained syncope; and that leg crossing, handgrip and arm tensing raise blood pressure enough to abort a reflex faint if begun during the prodrome.

   Read it asThe chapter states that a normal pulse afterwards excludes nothing, because the arrhythmia responsible has usually terminated by the time anyone reaches the wrist — which is the argument for an ECG rather than a pulse check. The 30-minute observation period is offered as a working default and labelled as such; no guideline specifies a bystander observation time.

   Strong

Emergency and first aid guidance

Everything in the emergency topic traces to one of these. Where they have been superseded, the superseding version is what the pages use.

1. Post-cardiac arrest care — 2025 AHA and ERC-ESICM guidelines

   Hirsch KG et al., Part 11: Post–Cardiac Arrest Care, 2025 AHA Guidelines, Circulation 2025; Nolan JP, Sandroni C et al., ERC-ESICM Guidelines 2025: Post-resuscitation care, Intensive Care Medicine 2025; with TTM2 (NEJM 2021), BOX, TAME, COACT and TOMAHAWK · 2025

   SupportsSection 7 of the resuscitation chapter: actively prevent fever at or below 37.5 °C for at least 72 hours, with routine induced hypothermia at 33 °C no longer preferred and the legacy 32–36 °C option removed; emergency coronary angiography for ST elevation with a selective strategy otherwise, following COACT and TOMAHAWK; oxygen saturation 94–98% and normocapnia; mean arterial pressure supported; and neuroprognostication deferred at least 72 hours after normalising temperature and sedation, using several independent findings rather than one.

   Read it asThis is a genuine reversal and it is recent. Deliberate cooling to 33 °C was standard for two decades on the strength of two small 2002 trials; TTM2 found it no better than strict normothermia, and the 2025 guidelines removed the older target entirely. Fever prevention is the part that survived. Anyone taught the 33 °C protocol learned current practice at the time and is now out of date, which is why the chapter flags the change explicitly rather than simply stating the new target.

   Strong

2. Resuscitation Council UK Guidelines 2025

   Resuscitation Council UK · Published late 2025, current

   SupportsAdult and paediatric basic life support throughout the resuscitation chapter: 30:2 for adults at 5–6 cm and 100–120/min; paediatric 5 rescue breaths then 15:2 at a third of chest depth; and the 2025 changes — calling 999 for an unresponsive person before assessing breathing, AEDs usable at any age including infants, thumb-encircling compressions for infants, and the revised child pad placement.

   Read it asCurrency matters unusually here. The 2021 guidelines are still widely quoted in training material and in charts written before the update, and several points differ. Anything on this site that conflicts with a 2021-era source is deliberate.

   Strong

3. Emergency treatment of anaphylaxis: guidelines for healthcare providers

   Resuscitation Council UK · May 2021, with the 2021 evidence review in Resuscitation

   SupportsThe IM adrenaline doses by age, repeating after 5 minutes, the definition of refractory anaphylaxis as no improvement after two appropriate doses, the positioning guidance, and the risk-stratified observation periods.

   Read it asTwo changes in the 2021 update are still not universally known: corticosteroids are no longer recommended for routine emergency treatment, and the repeat interval is 5 minutes rather than the 10–15 some older material gives. The doses are chosen to be safely and quickly drawn up rather than precisely weight-derived, which the guideline states explicitly.

   Strong

4. NICE NG143 — Fever in under 5s, and the traffic light system

   National Institute for Health and Care Excellence, NG143 · 2019, updated 2025

   SupportsThe three-band assessment and the paediatric thresholds on the vital signs page: that the highest single feature present sets the band; the red features including pale, mottled, ashen or blue skin, no response to social cues, not waking or not staying awake, a weak high-pitched or continuous cry, grunting, a respiratory rate above 60, moderate or severe chest recession, reduced skin turgor, a bulging fontanelle, neck stiffness, a non-blanching rash and focal or prolonged seizures; the amber features including impaired social cues, nasal flaring, oxygen saturation of 95% or less, dry mucous membranes, poor feeding, capillary refill of 3 seconds or more and reduced urine output; the age-specific respiratory thresholds of 60, 50 and 40 per minute for 0–5 months, 6–12 months and over 12 months; that a temperature of 38 °C or above under 3 months is a red feature and 39 °C or above at 3–6 months is amber; that fever of 5 days or more prompts assessment for Kawasaki disease; and that neither the height nor the duration of fever predicts serious illness.

   Read it asTwo things the guideline says that are worth quoting to anyone who suggests otherwise: a parent’s report of fever should be taken seriously without a thermometer reading, and the classic meningitis signs are often absent in infants with bacterial meningitis. A general practice cohort of 6,703 children — Kerr, Verbakel et al., Accuracy of the NICE traffic light system in children presenting to general practice, British Journal of General Practice 2022;72(719):e398 — found that impaired social cues alone triggered 97% of amber and 61% of red classifications, while abnormal vital signs contributed to only a small fraction — which is the evidence behind the page leading on behaviour rather than numbers. NG143 also carries its own warning that pulse oximeters can overestimate saturation in people with dark skin.

   Strong

5. 2024 ESC Guidelines for the management of elevated blood pressure and hypertension

   European Society of Cardiology, 2024; with commentary in Circulation and Hypertension on the reclassification · 2024

   SupportsThe divergence noted in the blood pressure chapter: three categories replacing the previous scheme — non-elevated below 120/70, elevated at 120–139 or 70–89, and hypertension at 140/90 or above — with the stages abandoned, and a treatment target of systolic 120–129 where tolerated.

   Read it asThe definition of hypertension did not move; the labels below it did. Published commentary notes that someone with an office reading of 118/72 was “optimal” under the 2018 and 2023 European schemes and is “elevated” under this one purely on a diastolic value in the seventies, and that the diastolic evidence for “lower is better” is weaker than the systolic evidence. UK practice follows NICE NG136, which retains 140/90 and the staging. The chapter presents both rather than picking, because a reader may be shown either table.

   Strong

6. Anti-choking suction devices — regulatory position and evidence

   MHRA Device Safety Information alert on counterfeit and unbranded anti-choking devices; FDA De Novo classification of LifeVac under 21 CFR 874.5400; Resuscitation Council UK position; AHA, American Red Cross, AAP and ILCOR positions; Ramaswamy et al. cadaver study, Laryngoscope 2023; Haupt et al. pilot cadaveric study, Resuscitation Plus 2025; and the 2020 systematic review and 2023 manikin study · 2020–2026

   SupportsThe device section in choking and drowning: that the FDA De Novo classification creates the device type “suction anti-choking device as a second-line treatment”; that LifeVac and Dechoker are the only two brands with valid UKCA or CE marking and MHRA registration in the UK, both registered for use only after back blows and abdominal thrusts have failed; that the MHRA estimates over 10,000 counterfeit or unbranded devices were bought by the UK public in two years and warns they may fail or worsen the obstruction because they lack a functioning one-way valve; that RCUK neither encourages nor discourages while declining to endorse them on insufficient evidence; and that cadaveric testing found Dechoker failed in every trial while LifeVac cleared only moistened crackers, with both applying significant pressure to the tongue.

   Read it asThe commonly repeated claim that these devices are “not FDA approved” is out of date — the De Novo classification changed that, and it changed it specifically by creating a second-line device type. The accurate criticism is not that they are unapproved but that the favourable evidence is manufacturer-held self-report and manikin work whose own systematic review rated certainty as very low, while independent cadaveric testing was unfavourable. The counterfeit market is the part that is unambiguously dangerous. The behavioural risk — a device substituting for training in back blows and thrusts — is raised by several of the bodies above and is not a small consideration.

   Moderate

7. Heat stroke — recognition and cooling

   European Resuscitation Council First Aid Guidelines 2025 (based on the 2025 ILCOR consensus); Society of Critical Care Medicine, Guideline for the Treatment of Heat Stroke, Crit Care Med 2025;53(2):e490–e500; with a 2025 Annals of Emergency Medicine description of an in-department cold-water immersion protocol · 2025

   SupportsThe heat chapter: heat stroke defined as core temperature at or above 40 °C with central nervous system dysfunction; whole-body cold water immersion from 1 to 26 °C continued until core temperature falls below 39 °C, with ice water at 1–5 °C giving the fastest rate; active cooling preferred over passive; the alternatives where immersion is impossible, including tarp-assisted cooling, ice sheets, packs to neck, axillae and groin, and evaporative cooling with fanning; cool first and transport second, targeting below 39 °C within 30 minutes of onset; and that rectal temperature is the only reliable measurement, with tympanic, temporal, oral, axillary and infrared readings inaccurate at the extremes.

   Read it asThe 60% intensive-care mortality and the roughly one-third of survivors with lasting cognitive or motor impairment are what justify the urgency of the cooling instruction. The published case in which a tympanic thermometer read 36.7 °C against a rectal 41.1 °C is quoted on the page because it is the clearest possible demonstration that a normal-looking ear reading excludes nothing. Paediatric immersion is a genuine evidence gap — there is little data below about 12 years, children cool faster for their size, and the chapter therefore recommends evaporative cooling with frequent checks rather than immersion.

   Strong

8. Accidental hypothermia, and electrical and crush injury

   European Resuscitation Council Guidelines 2025 on special circumstances and first aid; UK Search and Rescue and wilderness medicine consensus on rewarming and frostbite; disaster medicine literature on crush syndrome · 2025

   SupportsThe cold chapter: hypothermia staging, gentle handling and horizontal positioning because of myocardial irritability, core rather than peripheral rewarming, afterdrop and rewarming collapse, prolonged resuscitation with rewarming, and frostbite rewarming at 37–39 °C with the instruction not to thaw where refreezing is possible. The electrical chapter: isolating the supply before contact, standoff from overhead lines, entry and exit wounds with deep tissue injury, and reverse triage in lightning strike. The crush chapter: crush syndrome from potassium and myoglobin release on reperfusion, fluid resuscitation before release, and that tourniquets are not recommended to prevent it.

   Read it asTwo points on the page go slightly beyond guideline text and are flagged as judgements. The fifteen-minute crush threshold is a widely used working figure rather than a validated cut-off, and the chapter says so and instructs that unknown timing be treated as prolonged. And the chapter deliberately declines to tell a lone bystander to leave someone trapped — it states the preference for release alongside fluids while making clear that a person who lifts the load has not acted recklessly, because the alternative framing produces paralysis at a scene.

   Strong

9. Resuscitation Council UK First Aid Guidelines 2025

   Resuscitation Council UK · 2025, current

   SupportsThe escalating approach in choking — cough, back blows, abdominal thrusts, with no blind finger sweeps and mandatory clinical assessment for anyone who has received thrusts; the escalation in life-threatening bleeding from manual direct pressure to haemostatic dressing and tourniquet; and the drowning sequence of flotation, removal from water and prevention of hypothermia.

   Read it asThe bleeding escalation is the notable one: tourniquets sit in the guidance as a step in a sequence rather than as a desperate last resort, which reverses what many older first aid courses taught.

   Strong

10. Supplying take-home naloxone without a prescription

    UK Government guidance, following The Human Medicines (Amendments Relating to Naloxone and Transfers of Functions) Regulations 2024 · Regulations December 2024; guidance updated 2025

    SupportsThat anyone may administer naloxone in an emergency; that it remains a prescription-only medicine so pharmacies cannot sell it over the counter; and that since December 2024 a much wider group may supply take-home and carry kits without a prescription — drug and alcohol services, pharmacy professionals, registered nurses and midwives, paramedics, police forces, prison and probation staff, and armed forces medical services.

    Read it asThe clinically important point is not legal but pharmacological: naloxone has a shorter duration of action than most opioids, so a person can stop breathing again after apparently full reversal. Synthetic opioids in the current UK supply, including nitazenes and fentanyl, frequently need repeated doses.

    Strong

11. Normal ranges of heart rate and respiratory rate in children from birth to 18 years

    Fleming S, Thompson M, Stevens R, et al., Lancet 2011;377(9770):1011–1018 · 2011

    SupportsThe heart rate and respiratory rate bands in the school charts, alongside APLS reference ranges and the 2017 AAP paediatric blood pressure guideline.

    Read it asReference ranges genuinely disagree between sources. Comparisons of APLS, PALS and the Fleming and O’Leary centiles show the published limits sitting in different places, with PALS upper limits falling outside the 99th centile in some age bands. The chart on this site had to choose, and the alert bands are a judgement about when to escalate rather than a published boundary. That is stated on the page.

    Strong

12. Button battery ingestion — time-critical management and the honey protocol

    UK paediatric transport and regional network guidelines (KIDS, NWTS, East of England ODN) and NHS trust patient information · 2017–2025

    SupportsThat a battery lodged in the oesophagus is an emergency even in an asymptomatic child, that damage can begin within two hours and removal is targeted inside that window, that over half of fatalities involve delayed or missed diagnosis, and that life-threatening bleeding can occur up to 28 days after removal or passage. And the honey protocol: children over 12 months, 10 ml every 10 minutes, where ingestion is known or suspected within 12 hours, otherwise nil by mouth.

    Read it asHoney is excluded under 12 months because of infant botulism risk, and where perforation is suspected. Ordinary commercial honey rather than artisanal. Risk falls sharply once the battery is past the stomach, but that determination needs an x-ray and is not a decision to make at home.

    Strong

13. NICE NG95 Lyme disease, and NG184 on antimicrobial prescribing for bites and stings

    National Institute for Health and Care Excellence · NG95 2018 with updates; NG184 2020

    SupportsThe bites and stings chapter: that erythema migrans appears 3–30 days after a bite, expands, is often not a bullseye and is not usually itchy or painful; that it should be diagnosed clinically and treated without waiting for serology; that antibiotic prophylaxis after a tick bite is not routinely recommended in the UK; and that antibiotics are not indicated for a bite where the skin is unbroken.

    Read it asThe no-prophylaxis position differs from practice in some other countries, so patients ask for it. It rests on a benefit-against-harm judgement rather than on evidence that transmission does not occur, and the safety net is surveillance for the rash. Around 60–80% of UK Lyme cases develop erythema migrans, so its absence does not exclude the diagnosis.

    Strong

14. Button battery ingestion triage — National Capital Poison Center guideline and ESPGHAN position

    National Capital Poison Center Button Battery Ingestion Triage and Treatment Guideline; and the ESPGHAN position paper on foreign body ingestion in children · current

    SupportsThe honey protocol in poisoning: honey for children over 12 months where ingestion is known or suspected within the preceding 12 hours, at intervals while in transit, to coat the battery and slow alkaline injury; with exclusions under 12 months for infant botulism risk and where perforation is suspected. Both bodies also underpin the two-hour removal target for an oesophageal battery and the point that a battery past the stomach carries much lower risk.

    Read it asThe honey recommendation is derived from animal and ex-vivo work on pH neutralisation rather than from randomised human trials, and is a mitigation while in transit rather than a treatment. It never substitutes for going, which is what the page says.

    Strong

15. Rabies risk assessment, including bats in the UK

    UK Health Security Agency rabies risk assessment guidance · current

    SupportsThat there is no indigenous rabies in UK animals other than bats, and that bat contact anywhere — including in the UK — warrants urgent assessment for post-exposure treatment, as does any mammal bite or scratch abroad.

    Read it asThe bat exception is the part most often missed, because the general message that Britain is rabies-free is otherwise correct.

    Strong

16. Inter-arm blood pressure difference, and cuff sizing

    NICE NG136 on measuring both arms; AHA cuff-sizing standards; and cohort evidence including the Fasa Adults Cohort Study and pooled inter-arm analyses · 2019–2025

    SupportsThat blood pressure should be measured in both arms at diagnosis, that a difference above 15 mmHg should be repeated and, if it persists, the higher arm used thereafter; that bladder width should be about 40% and length at least 80% of arm circumference; and that an undersized cuff overestimates systolic pressure by roughly 10–15 mmHg and sometimes 20.

    Read it asThe association between a persistent inter-arm difference and cardiovascular risk is observational and the thresholds proposed in the literature vary from 5 to 15 mmHg. The action the difference triggers — use the higher arm, and think about dissection in the right clinical context — is what the page rests on rather than a precise risk estimate.

    Moderate

17. Water compared with other solutions for wound cleansing

    Fernandez R, Green HL, Griffiths R, Atkinson RA, Ellwood LJ. Water for wound cleansing. Cochrane Database of Systematic Reviews 2022, Issue 9, CD003861 · Fifth update, 14 September 2022

    SupportsThat potable tap water is a reasonable irrigant for uncomplicated acute wounds, with no demonstrated increase in infection against sterile saline.

    Read it asThe certainty fell when the review was updated. Earlier versions concluded tap water should be preferred on cost and availability; the 2022 fifth update, covering 13 trials, rates the evidence low to very low certainty, the trials being small and inconsistent in how infection was defined. The practical recommendation stands — there is no case for delaying irrigation to find saline — but as reasonable practice rather than proven superiority. The trials also excluded bites, punctures, heavily contaminated wounds, foreign bodies and immunocompromised patients, which are precisely the wounds that need a clinician.

    Moderate

18. Pulse oximetry limitations, including skin pigmentation

    Regulatory safety communications from the MHRA and FDA, and the paediatric pulse oximetry literature · 2021–2025

    SupportsThat pulse oximeters can overestimate true oxygen saturation in people with darker skin, that carboxyhaemoglobin is read as oxyhaemoglobin so readings are falsely high in carbon monoxide poisoning, that poor perfusion and movement make readings unreliable, and that cyanosis is not clinically evident until saturation falls below about 85%.

    Read it asThe pigmentation bias is the newest of these and the least widely known. None of it makes the device useless; all of it means the reading is evidence rather than an answer.

    Strong

19. Tick bites, Lyme disease and tick-borne encephalitis in the UK

    NICE NG95, Lyme disease: diagnosis and management (2018, with rationale and impact documents); UKHSA, Defra and APHA joint report on vector-borne disease, May 2026; Holding M et al. and colleagues, Tick-borne encephalitis: from tick surveillance to the first confirmed human cases, the United Kingdom, 2015 to 2023, Eurosurveillance 2025; UKHSA report of a probable UK-acquired TBE case, July 2019; IDSA/AAN/ACR Lyme disease guideline 2020; CDC, What to Do After a Tick Bite, reviewed 9 June 2026; Bryant KA, When Does a Tick Bite Need Antibiotics?, Medscape commentary, May 2026 · 2018–2026, checked August 2026

    SupportsThe tick chapter. From NG95: erythema migrans as a clinical diagnosis treated on sight without testing; doxycycline 100 mg twice daily for 21 days for adults and young people 12 and over, amoxicillin 1 g three times daily for 21 days as the alternative and in pregnancy, doxycycline 200 mg twice daily or intravenous ceftriaxone where the central nervous system is involved, and no azithromycin in cardiac involvement because of QT effects; specialist discussion for anyone under 18 except a child of 9 to 12 with a single erythema migrans and no other features; two-tier serology with ELISA then immunoblot and repeat testing at 4 to 6 weeks if negative with persisting symptoms. From the May 2026 UKHSA and Defra report: 1,168 laboratory-confirmed Lyme cases in England in the reporting year, up from 959 and comparable to 1,151 in 2023, and six UK-acquired tick-borne encephalitis cases in total with the virus limited to Thetford Forest, the New Forest, Dartmoor and Devon, North Yorkshire and parts of Scotland.

    Read it asThe chapter names the UK/US divergence in both directions rather than picking a side. The IDSA recommends a single 200 mg dose of doxycycline within 72 hours, but only where all its high-risk criteria are met — an Ixodes tick, a highly endemic area, and attachment for 36 hours or more — and states that monitoring is otherwise as effective; most UK bites do not meet that bar. It also flags that the UK treats erythema migrans for 21 days and the US for 10, and says plainly that both are defensible readings of thin evidence, so a reader comparing sources does not conclude one is wrong. Section 7 on persistent symptoms is written to hold two things together: prolonged antibiotics have been tested in randomised trials, do not help and do cause harm — and that is not the same as saying the symptoms are imagined. It warns specifically about non-validated private testing, and points at the treatable alternatives a Lyme label can close off. The TBE figures will date and are marked as a May 2026 position; louping ill virus cross-reactivity is noted as a diagnostic complication rather than a disease risk. Round 100 corrected one error and closed three gaps after reading the CDC page and the Medscape commentary in full. The removal list had said not to use fingers; that was stricter than CDC guidance and would cause the delay the guidance exists to prevent, so the section now leads with do not wait to see anyone and permits ordinary tweezers or fingers over waiting. The tick-testing argument went from one reason to four — commercial labs not held to clinical diagnostic standards, a positive not meaning infection, a negative being falsely reassuring because another tick may have been missed, and results arriving after symptoms would. And a travel section was added, because the American lone star tick causes illnesses absent from the UK: ehrlichiosis with a rash in up to 60% of infected children, tularemia, Bourbon and Heartland viruses, alpha-gal syndrome, and STARI — an expanding rash that mimics erythema migrans and is not Lyme disease. Rocky Mountain spotted fever is named as treated urgently on suspicion with no role for prophylaxis, and the European TBE belt as a travel-vaccine conversation rather than an afterthought.

    Strong

Consumer ECG devices

1. Clinical Validation of 5 Direct-to-Consumer Wearable Smart Devices to Detect Atrial Fibrillation — the BASEL Wearable Study

   JACC: Clinical Electrophysiology 2022 · 2022

   SupportsThe figures in the ECG chapter: against a simultaneous physician-read 12-lead ECG in 163 patients, sensitivity for atrial fibrillation of 84–96% and specificity of 91–98% across five devices once inconclusive tracings are excluded; inconclusive rates of 17–26%, with only 79% of recordings returning an automated diagnosis; and 8% of patients in sinus rhythm labelled as AF by at least one device.

   Read it asThe inconclusive rate is the part usually omitted when these sensitivities are quoted, and the chapter leads with it. Manual review resolved the rhythm in about 99% of tracings, so an inconclusive result is generally an algorithm limitation rather than a bad recording — which is why the advice is to keep the trace. Validation populations are adult and older, mean ages in the mid-sixties.

   Strong

2. Accuracy and clinical relevance of the single-lead Apple Watch electrocardiogram to identify atrial fibrillation

   Cardioversion cohort study, 74 patients · 2022

   SupportsThe repeat-once rule: repeating an unclassified recording once reduced the unclassified rate substantially, while a third attempt lowered accuracy and produced false positives.

   Moderate

Vaping

1. Carbonyl and metal emissions in e-cigarette aerosol

   Pooled analyses of aerosol chemistry, including a 2023 health risk assessment in Scientific Reports and the Farsalinos carbonyl systematic review in Frontiers in Physiology · 2017–2023

   SupportsThe per-puff yields behind the exposure estimator: mean formaldehyde around 0.86 µg/puff with a reported maximum near 28 µg/puff, acetaldehyde around 0.67 and acrolein around 0.37, with metals from nichrome heating elements.

   Read it asThe single most important caveat in this literature. The highest published carbonyl figures come from machine protocols producing “dry puff” conditions — the coil hot with insufficient wicking — which a person finds acutely unpleasant and stops doing. Quoting those as ordinary use is the commonest error in the field. Yields span roughly three orders of magnitude with power and device, which is why the estimator separates device power from puff frequency and why its high-power figures sit well below the published maxima.

   Moderate

2. Cancer and non-cancer risk from metals in e-cigarette liquids and aerosols

   International Journal of Environmental Research and Public Health 2020;17(6):2146 · 2020

   SupportsThat chromium and nickel from nichrome coils appear to be the main contributors to calculated cancer and non-cancer risk from vaping, and that those calculated risks exceeded those from formaldehyde and acetaldehyde in the same devices.

   Read it asUnder-appreciated relative to the aldehydes, which get the coverage. This is a modelled risk assessment from measured concentrations, not an outcome study — there is no cohort showing cancers caused by coil metals.

   Moderate

3. Relative cancer potency of e-cigarette emissions against tobacco smoke

   Stephens WE, modelling analysis of emissions data; and the 2018 National Academies consensus report on the public health consequences of e-cigarettes · 2017–2018

   SupportsThe denominator the estimator keeps visible: most e-cigarette emissions were modelled at cancer potencies below 1% of tobacco smoke, with exceptions; and the National Academies conclusion that cancer risk from long-term e-cigarette use, if any, is likely very low compared with combustible cigarettes.

   Read it asMass is not potency, and the estimator says so. Vape aerosol can approach a cigarette on formaldehyde by weight while remaining far below on total carcinogenic potency, because the nitrosamines and polycyclic aromatic hydrocarbons that carry most of tobacco smoke’s potency are largely absent rather than merely reduced. Modelling, not outcomes — there are no thirty-year cohorts, which is also why the estimator refuses to output a risk percentage.

   Moderate

4. Use of e-cigarettes among young people in Great Britain

   Action on Smoking and Health (ASH) · 2026 survey

   SupportsYouth prevalence figures, the fall in disposable use from 69% to 13% among young vapers, and the finding that 62% of young people believe vaping is as harmful as smoking or worse.

   Read it asSelf-reported survey data. The disposables collapse is well evidenced; the claim that youth vaping itself has fallen proportionally is not, and this site says so.

   Moderate

5. Disposable vapes banned from 1 June 2025

   UK Government · In force 1 June 2025

   SupportsThe regulatory timeline on what changed in UK law.

   Read it asMade under environmental rather than tobacco legislation, which is why the ban covers single-use devices rather than nicotine content.

   Strong

Eyes and sight

Glaucoma detection and the UK treatment pathway, macular degeneration, and the two things currently promoted as neuroprotection.

1. Glaucoma: diagnosis and management (NG81)

   NICE · Published 2017, updated 26 January 2022

   SupportsThe UK treatment pathway: 360-degree selective laser trabeculoplasty as a first-line option for open-angle glaucoma and ocular hypertension, rather than lifelong drops.

   Read it asThe January 2022 update is the operative change and the reason this topic leads with it. It does NOT support any claim about lifestyle prevention of glaucoma, and no such claim is made anywhere in the topic.

   Moderate

2. Selective laser trabeculoplasty versus eye drops for first-line treatment of ocular hypertension and glaucoma (LiGHT)

   Gazzard et al., Lancet 2019; six-year results Ophthalmology 2023;130(2):139-151 · 2019, extended 2023

   SupportsSLT as a reasonable first-line treatment. Treatment-naive patients, six UK hospitals, observer-masked.

   Read it asThe primary outcome was health-related quality of life, NOT sight preserved. That is why this site grades SLT as a reasonable first-line option and explicitly not as superior to drops for preserving vision.

   Moderate

3. Latanoprost-netarsudil for previously treated open-angle glaucoma or ocular hypertension (TA1009)

   NICE; netarsudil (Rhokiinsa) has held a UK marketing authorisation since November 2019 · Checked August 2026

   SupportsThat Rho kinase inhibitors ARE available in the UK, and that the latanoprost-netarsudil combination is NICE-recommended for adults already treated for open-angle glaucoma or ocular hypertension.

   Read it asAdded round 115 to correct the glaucoma chapter, which had placed netarsudil in its US-only section. That was wrong and it is the kind of error that sends a reader to ask for something abroad that their own optometrist can already discuss. It does NOT support any claim about where in the sequence the combination sits relative to SLT or a prostaglandin alone.

   Moderate

4. The nicotinamide phase 2 trials

   Hui et al., Australia (crossover, 57 participants); De Moraes et al., JAMA Ophthalmology (32 participants, nicotinamide plus calcium pyruvate) · 2020 and 2022

   SupportsThe Emerging grade on nicotinamide, and the caution about gram doses.

   Read it asBoth trials are small and short, and both reported research endpoints rather than vision: the photopic negative response on electroretinography, and sensitivity at eight visual field test locations. Neither showed that anyone saw better. A specialty-society position statement making the same point was cited here until 6 September 2026 and was removed at review, because the attribution could not be stood behind; the point does not depend on it, and is now made from what the trials measured. The phase 3 trials report from late 2027. This does not support recommending nicotinamide, and this site does not.

   Moderate

5. Ocular Hypertension Treatment Study (OHTS)

   Kass et al. and the OHTS Group; 22 US clinical centres, 1,636 participants · Primary outcome 2002; corneal thickness analyses 2001 and 2020; 20-year follow-up

   SupportsThe five-year conversion figures (9.5% untreated against 4.4% treated), the risk stratification from about 1-2% to 25-35% over five years, and the five baseline predictors: age, IOP, central corneal thickness, cup-to-disc ratio and visual field pattern standard deviation. Also the corneal thickness figures: mean 573 micrometres, 24% above 600, and African American participants averaging about 23 micrometres thinner.

   Read it asThe trial predates OCT, so early structural damage may have been under-detected — which if anything makes the conversion figures conservative. It does NOT support any statement about how fast an individual will progress; this site gives ranges and says explicitly that rate cannot be known at diagnosis.

   Moderate

6. Glaucoma pipeline and sustained-release delivery: Durysta, iDose TR and what follows them

   FDA approvals (Durysta March 2020, iDose TR December 2023); Glaucoma Research Foundation; Review of Ophthalmology 2026 pipeline; Ophthalmology Times 2026 · Availability position checked August 2026

   SupportsThe US-only treatment section and the pipeline section: the iDose phase 3 pressure reductions of 6.6-8.4 mmHg across 1,150 patients, the Durysta single-administration limit and its roughly 3.5% corneal endothelial cell loss at 18 months, and the biodegradable latanoprost implant reporting 26-35% reduction over 48 weeks with 94% needing no drops.

   Read it asAVAILABILITY WILL AGE. These are stated as approved in the United States and not available in the UK at the time of writing, and the page is date-stamped and says so. Anything here should be re-checked rather than relied on. Trade names appear because a reader will meet them in American coverage and needs to recognise what they are reading; no product is recommended.

   Moderate

7. Dry Eye Assessment and Management (DREAM) study, and its extension

   National Eye Institute; 535 patients, 27 centres · Primary result 2018; withdrawal extension 2019

   SupportsThat omega-3 supplementation at 3,000mg/day is no better than a refined olive oil placebo for moderate-to-severe dry eye, on symptoms or on any objective sign. Also the teaching point that BOTH arms improved substantially, by about 14 and 12.5 points, which is why uncontrolled dry eye testimonials are worth little.

   Read it asA meta-analysis pooling seventeen trials does report benefit. The site weights the single large well-controlled trial over the pooled smaller ones and says so on the page. The olive-oil placebo is the main published objection and is not settled.

   Moderate

8. Perfluorohexyloctane for dry eye: FDA approval, and the same molecule as a European medical device

   FDA approval May 2023 (Miebo, Bausch + Lomb / Novaliq); GOBI and MOJAVE phase 3 trials, >1,200 patients; sold in Europe and the UK over the counter as EvoTears and Hycosan Shield · Availability checked August 2026

   SupportsThat the first FDA-approved prescription drug targeting tear evaporation is the same molecule available here without prescription, and that it met both primary endpoints in its pivotal trials.

   Read it asAVAILABILITY AND REGULATORY ROUTE WILL AGE. A device authorisation is a lower evidential bar than a drug approval; the page says so rather than presenting the UK route as simply better. Product names appear because a reader meets them in American coverage, not as recommendations.

   Moderate

9. Age-Related Eye Disease Study 2 (AREDS2)

   National Eye Institute · 2013 onwards

   SupportsThe AREDS2 formulation, and the removal of beta-carotene on lung cancer risk in smokers and former smokers.

   Read it asCRITICAL SCOPE. Tested in people with intermediate AMD, or advanced AMD in one eye, for progression to advanced disease. It does NOT support taking the formula for prevention in someone with healthy maculae — which is how it is most often sold, and the chapter says so explicitly.

   Moderate