Before you read: where I stand
I rate Malcolm Kendrick’s work on blood-vessel damage and clotting. It’s made me look harder at causes that aren’t cholesterol.
My own Lp(a) is raised, so I have skin in this.
I sell blood testing, including the heart panel at the end of this post.
None of that changes the numbers below. Where the evidence is weak, I’ve said so, including where it’s weak for things I like.
The grades: Strong means trials or genetic studies agree. Moderate means real but less certain. Observational or Mechanistic means measured in populations or the lab, not tested as a treatment. Guideline means a guideline threshold or goal.
1. Why “cholesterol” is the wrong headline
LDL cholesterol measures the cholesterol carried inside LDL particles: the cargo. What drives plaque is the number of particles getting into the artery wall. Each of those particles carries one molecule of ApoB, so ApoB is a direct count.
Most of the time LDL-C and ApoB tell the same story. When they disagree (often with high triglycerides, insulin resistance or type 2 diabetes), ApoB is the better guide. That’s why I’d rather see it than not.
Is LDL actually causal, or just a marker? Genetic studies, more than 200 cohorts, Mendelian randomisation and trials of drugs that lower LDL without being statins all point the same way: more LDL particles for longer means more heart disease, in a dose-dependent way (Ference 2017, European Atherosclerosis Society consensus). Strong The full statin argument is in LDL, Statins and Heart Disease.
2. The markers
| Marker | What it tells you | Bands (both units) | Source | Grade |
|---|---|---|---|---|
| ApoB | Number of particles that can enter the artery wall | No single “normal”. The goal depends on your overall risk: below 100 mg/dL (1.0 g/L) moderate risk; below 80 mg/dL (0.8 g/L) high; below 65 mg/dL (0.65 g/L) very high. ESC calls these “secondary goals”, after LDL | ESC/EAS 2019, Table 7 (not changed by the 2025 update) | Guideline |
| LDL cholesterol | Cholesterol carried in LDL (usually calculated, not measured) | Goal depends on risk: 3.0 mmol/L (116 mg/dL) or below low risk; 2.6 mmol/L (100 mg/dL) moderate; 1.8 mmol/L (70 mg/dL) high; 1.4 mmol/L (55 mg/dL) very high | ESC/EAS 2019 (Mach) | Guideline Trial-based |
| Lp(a) | A particle more than 90% set by your genes; raises risk even when LDL is low | Read it in the unit your lab used. In mg/dL (the unit on the test I use): risk rises slightly from 30, is a risk-enhancing factor above 50, and is very high above 180. If your result is in nmol/L: risk rises slightly from 62, above 105 is the risk-enhancing level and above 430 is very high. At least 1 in 5 people are above the risk-enhancing level. Never convert between the units yourself: the conversion depends on the assay and on your own particle size | ESC/EAS 2025 Focused Update; ESC/EAS 2019 (180 mg/dL / 430 nmol/L) | Strong for risk Lowering it not yet shown to help |
| Blood pressure | The pressure your artery walls take, all day | Clinic 140/90 mmHg or above, confirmed by a home average of 135/85 mmHg or above, is how NICE diagnoses high blood pressure | NICE NG136 | Guideline |
| HbA1c | Average blood sugar over 2–3 months | 42–47 mmol/mol (6.0–6.4%) raised risk; 48 mmol/mol (6.5%) or above diabetes range | NICE / WHO | Guideline |
| Fasting glucose | Blood sugar this morning | Below 5.6 mmol/L (100 mg/dL) usual; 5.6–6.9 mmol/L (100–125 mg/dL) impaired; 7.0 mmol/L (126 mg/dL) or above diabetes range | ADA 2025 | Guideline |
| Triglycerides (fasting) | Fat carried in the blood; high levels often travel with insulin resistance | Below 1.7 mmol/L (150 mg/dL) lower risk | ESC/EAS 2019 | Risk marker Not a treatment target |
| HDL | Context only | Low: below 1.04 mmol/L (40 mg/dL) men, 1.29 mmol/L (50 mg/dL) women. ESC sets no HDL target; raising it with drugs hasn’t helped | Alberti 2009 | Context Not a target |
| hs-CRP | Low-grade inflammation | Below 1 mg/L lower; 1–3 mg/L average; above 3 mg/L higher (0.1 / 0.3 mg/dL). ESC counts persistently above 2 mg/L (0.2 mg/dL) as a risk modifier. Above 10 mg/L usually means an infection or injury: repeat when well | CDC/AHA 2003 (Pearson); ESC/EAS 2025, Box 1 | Guideline A modifier, not a target |
One unit trap worth knowing: at least one lab’s own guide prints the CRP bands in mg/dL where it means mg/L. Check the unit on your report before comparing with anything.
The non-lipid risks, which the “it’s all cholesterol” camp tends to skip:
- Smoking. Still the biggest one you can change.
- Lead. In a large US national cohort, going from low to high blood lead (10th to 90th percentile) went with a 70% higher risk of dying from heart disease (HR 1.70), and the authors attributed about 256,000 US heart deaths a year to it (Lanphear 2018). Strong observational
- Air pollution. The American Heart Association judged the evidence “consistent with a causal relationship” between fine particulate matter (PM2.5) and heart disease and death, both short-term and over years (Brook 2010). Strong observational With mechanisms.
- Kidney function (eGFR, cystatin C). A falling eGFR raises heart risk; it’s on most panels already.
3. How often to check
No guideline sets a schedule for most of this in healthy people, so this is my practice, not a guideline, except where a guideline is named:
- Lp(a): once. It’s mostly genetic and barely moves. ESC and EAS recommend measuring it at least once in every adult’s lifetime (ESC/EAS 2025; EAS 2022). Women: it can rise after menopause, so I’d consider a second test if the first was borderline. If yours is high, tell your family: it’s inherited.
- ApoB and lipids: about 3 months after you change something (diet, training, weight, a medicine), then yearly when stable.
- Blood pressure: NICE says at least every 5 years if it’s normal, more often if it’s close to 140/90. At home, for tracking (my practice), once a week, same time of day, seated, two readings a minute apart. If you’re checking for a diagnosis, NICE’s method is different: two readings a minute apart, morning and evening, for at least 4 days and ideally 7, ignoring day one and averaging the rest (NICE NG136).
- HbA1c: yearly, or 3 months after a change.
- hs-CRP: only alongside the others, and repeat any single high result when you’re well.
Before you decide a number has changed, read Did Anything Actually Change?. Many markers wobble by 10–30% on their own between tests. A retest “going up” may be noise.
4. What moves each number
| Number | What moves it | Size of effect | Source | Grade |
|---|---|---|---|---|
| ApoB | A Mediterranean diet with extra virgin olive oil, vs a low-fat diet | About −3 mg/dL ApoB at 3 months (−2.9; CI −5.6 to −0.1) in older adults at high risk. Small. | PREDIMED substudy, Solá 2011 | RCT, modest Short |
| LDL / cardiovascular events | Eating less saturated fat, replaced with unsaturated fat or carbohydrate | Cardiovascular events about 21% lower over at least 2 years (RR 0.79; 0.66–0.93); about 1 event avoided per 56 people over 4 years in primary prevention. No clear change in deaths (RR 0.96; 0.90–1.03) | Cochrane, Hooper 2020 | Moderate GRADE |
| But: in the Minnesota Coronary Experiment, replacing saturated fat with corn oil | Lowered cholesterol and didn’t lower deaths. Each 30 mg/dL (0.78 mmol/L) fall in cholesterol went with a 22% higher risk of death. That result sat unpublished for decades | Ramsden 2016 | One RCT One diet swap, high dropout | |
| LDL | Statins and other medicines | Large and dose-dependent | Ference 2017; CTT | Strong |
| Blood pressure | Eating less salt | About 4.4 g/day less salt for at least 4 weeks: systolic −4.2 mmHg on average; −5.4 with high blood pressure, −2.4 without. The more you cut, the bigger the fall, with no floor found; a bigger effect in older people and with higher blood pressure | Cochrane, He 2013 (34 RCTs); Huang 2020 (133 trials); Filippini 2021 (85 trials) | Strong |
| Blood pressure | Losing weight | About 1 mmHg lower systolic per kg lost (−1.05); about −4.4 mmHg for 5 kg | Neter 2003 (25 RCTs) | Strong |
| Blood pressure | Regular exercise | Endurance training: −3.5/−2.5 mmHg overall; −8.3/−5.2 with high blood pressure; no clear change with normal blood pressure. Isometric (static-hold) training showed the largest fall (−10.9 systolic) but from only 5 study groups | Cornelissen 2013 (93 RCTs) | Strong For endurance; promising but thin for isometric |
| Blood pressure | More potassium | Supplement trials: −4.7/−3.5 mmHg; −6.8/−4.6 with high blood pressure, in people not on medication. Food first: potassium supplements can be dangerous with kidney disease and some blood pressure medicines | Binia 2015 (15 RCTs; authors from Nestlé Research) | Moderate |
| Lp(a) | Almost nothing in diet or lifestyle | Statins don’t change it (pooled data from 7 trials). New injected drugs lower it by 80–98% in trials; none yet shown to cut heart attacks | ESC/EAS 2025 | — |
| A high Lp(a) | Manage everything else harder and earlier | Matched to your overall risk. ESC says people with high Lp(a) whose overall risk is high enough should be “strongly encouraged” to take or stay on a high-intensity statin | ESC/EAS 2025; EAS 2022 | Guideline |
| LDL | Plant sterols | About 10% lower LDL; no trials showing fewer heart attacks | ESC/EAS 2025 | LDL: strong Outcomes: none |
Cutting down on alcohol also lowers blood pressure; I haven’t reviewed the size of the effect for this post.
Why I’ve included Minnesota. Both readings of it are fair. It’s a real case of an awkward result going unpublished. It also tested one swap (corn oil, high in linoleic acid) in institutions, with high dropout and short exposure for most participants. It doesn’t show that LDL doesn’t matter. It shows that how you lower it may matter, and that a cholesterol number isn’t the outcome.
Red yeast rice
Its active compound, monacolin K, is chemically the same molecule as the statin lovastatin. So I judge it as a statin in a supplement bottle, by the same standard as the prescription version.
- In 53 trials (about 8,500 people), it wasn’t linked to more muscle problems (OR 0.94; 0.53–1.65) (Fogacci 2019).
- Symmetry check: most of those trials were small and short, and the amount of monacolin varies between products in a way it doesn’t between prescriptions.
- The 2025 European guideline update: some purified preparations lower cholesterol, but there’s “no convincing evidence of a clinical benefit”. In a trial of 199 people (single-centre, single-blind, 28 days), none of red yeast rice, fish oil, cinnamon, garlic, turmeric or plant sterols lowered LDL more than placebo; low-dose rosuvastatin did (the SPORT trial).
- The EU banned higher-strength red yeast rice supplements in 2022 and puts warnings on the rest.
- If you’re taking it, your prescriber and pharmacist need to know, as for any statin.
The same rule for every supplement. The 2025 guideline update recommends against supplements or vitamins for heart protection unless they have documented safety and a real LDL-lowering effect (Class III). That’s the bar I hold my own recommendations to.
5. Where the critics are right, and where they go too far
Where they’re right:
- Publication bias is real. Minnesota, above.
- Industry money shaped the science. The sugar industry funded 1960s research that steered blame towards fat (Kearns 2016).
- The average benefit is small in low-risk people. Averaged over everyone in the trials, statins added a median 3.2 days of life in primary prevention and 4.1 days in secondary prevention, over 2–6 years (Kristensen 2015). That’s an average over everyone, most of whom never had an event in the trial, not a lifetime figure. But it’s a fair point.
- Statins slightly raise diabetes risk: about 1 extra case per 255 people treated for 4 years (Sattar 2010).
- This year’s trials support the “it doesn’t make you live longer” half: in STAREE (about 10,000 people aged 70+ without heart disease), heart attacks and strokes fell 30%, but survival free of dementia or disability didn’t change (Zoungas 2026). In ten primary-prevention trials (85,829 people), deaths were 3.9% vs 3.9% (Morsell 2026). In people aged 75+ already on a statin, stopping wasn’t worse over 3 years (Bonnet 2026).
- The main trialists’ group doesn’t share its individual data. Mainstream researchers say so too.
Where they go too far:
- “Cholesterol doesn’t cause heart disease.” The genetic and trial evidence above says it does (Ference 2017).
- “Statins don’t work.” The same 2026 trials show heart attacks down by about a third and strokes by about a quarter (Morsell 2026; Zoungas 2026). Another 2026 analysis of 25 trials found fewer cardiovascular deaths too (Long 2026). Fewer heart attacks isn’t nothing.
- “One in five people are harmed.” In blinded trials, muscle aches were reported by 27.1% on a statin and 26.6% on a dummy pill. More than 90% of the aches in people on statins weren’t caused by the statin (CTT 2022). This year, only 4 of 66 label side effects held up beyond muscle and diabetes (CTT 2026).
- “Give CoQ10 with every statin.” The meta-analyses disagree. Read CoQ10 and Statins: What the Trials Show.
- The clotting model. Damage to the artery lining and clotting do play a part in plaque and in heart attacks. That fits alongside LDL causing disease; it isn’t a replacement for it. Mechanistic Not tested as a replacement.
The honest version: if you don’t have heart disease, a statin lowers your chance of a heart attack or stroke but hasn’t been shown to make you live longer. Whether that trade is worth it depends on your own risk and how you feel on it. That’s a decision for you and your prescriber, not a slogan. Please don’t stop a medicine on the strength of a blog post.
6. Red flags: don’t wait for a test
- Chest pain that feels tight or like squeezing, or chest pain spreading to your arms, neck or jaw, or severe difficulty breathing: call 999 (NHS). Feeling sick, sweating or breathlessness can come with it.
- Face drooping on one side, not being able to lift both arms and keep them there, or slurred or confused speech: call 999 (stroke, NHS “Act FAST”). Don’t drive yourself to A&E.
- Blood pressure of 180/120 mmHg or higher: contact your GP the same day. NICE asks for urgent assessment at this level, and same-day specialist review with life-threatening symptoms. With chest pain, breathlessness, confusion or sudden vision change, call 999.
- Total cholesterol above 7.5 mmol/L (290 mg/dL), or a heart attack or other coronary event before 60 in you or a parent, brother, sister or child: see your GP. NICE says this should raise the question of an inherited cause (familial hypercholesterolaemia), and your relatives may need testing too.
- LDL above 4.9 mmol/L (190 mg/dL) or total cholesterol above 8 mmol/L (310 mg/dL) on its own puts you in a high-risk group, whatever your other numbers (ESC/EAS 2025).
7. Testing
If you want ApoB and Lp(a) together: the Doctor’s Data CardioMetabolic Profile, through Regenerus, £288 all-in: the test, the blood draw and a results consultation with me. Fasting sample. It also carries hs-CRP, glucose, insulin, GlycoMark and kidney function (cystatin C, eGFR). It doesn’t include HbA1c. Details on blood testing options.
- It reports Lp(a) in mg/dL. If you’ve had Lp(a) done in nmol/L elsewhere, the two can’t be compared directly.
- Two markers on it (leptin and adiponectin) are marked by the lab itself as research use only. I won’t read them as a diagnosis.
I sell this test. A test tells you where you are; it doesn’t decide what you should do. That’s what the consultation is for.
Sources
- Ference BA et al. Low-density lipoproteins cause atherosclerotic cardiovascular disease (EAS consensus). Eur Heart J 2017. PMID 28444290. doi:10.1093/eurheartj/ehx144
- Mach F et al. 2019 ESC/EAS Guidelines for the management of dyslipidaemias. Eur Heart J 2020. PMID 31504418. doi:10.1093/eurheartj/ehz455
- Mach F, Koskinas KC, Roeters van Lennep JE et al. 2025 Focused Update of the 2019 ESC/EAS Guidelines for the management of dyslipidaemias. Eur Heart J 2025;46:4359–4378. doi:10.1093/eurheartj/ehaf190
- Kronenberg F et al. Lipoprotein(a) in atherosclerotic cardiovascular disease and aortic stenosis (EAS consensus). Eur Heart J 2022. PMID 36036785. doi:10.1093/eurheartj/ehac361
- NICE CG71. Familial hypercholesterolaemia (updated 2019), 1.1.1. nice.org.uk/guidance/cg71
- NICE NG136. Hypertension in adults. nice.org.uk/guidance/ng136
- American Diabetes Association. 2. Diagnosis and Classification of Diabetes: Standards of Care in Diabetes—2025. Diabetes Care 2025;48(Suppl 1):S27–S49. PMID 39651986. doi:10.2337/dc25-S002
- Alberti KGMM et al. Harmonizing the metabolic syndrome. Circulation 2009. PMID 19805654.
- Pearson TA et al. Markers of inflammation and cardiovascular disease (CDC/AHA). Circulation 2003;107:499–511. PMID 12551878. doi:10.1161/01.cir.0000052939.59093.45
- Lanphear BP et al. Lancet Public Health 2018. PMID 29544878.
- Brook RD et al. Particulate matter air pollution and cardiovascular disease (AHA). Circulation 2010. PMID 20458016. doi:10.1161/CIR.0b013e3181dbece1
- Solá R et al. (PREDIMED). Atherosclerosis 2011. PMID 21640348. doi:10.1016/j.atherosclerosis.2011.04.026
- Hooper L et al. Reduction in saturated fat intake for cardiovascular disease. Cochrane 2020. PMID 32428300. doi:10.1002/14651858.CD011737.pub2
- Ramsden CE et al. Re-evaluation of the traditional diet-heart hypothesis. BMJ 2016. PMID 27071971. doi:10.1136/bmj.i1246
- He FJ et al. Effect of longer term modest salt reduction on blood pressure. BMJ 2013. PMID 23558162. doi:10.1136/bmj.f1325
- Huang L et al. BMJ 2020. PMID 32094151. doi:10.1136/bmj.m315
- Filippini T et al. Circulation 2021. PMID 33586450. doi:10.1161/CIRCULATIONAHA.120.050371
- Neter JE et al. Influence of weight reduction on blood pressure. Hypertension 2003. PMID 12975389. doi:10.1161/01.HYP.0000094221.86888.AE
- Cornelissen VA, Smart NA. Exercise training for blood pressure. J Am Heart Assoc 2013. PMID 23525435. doi:10.1161/JAHA.112.004473
- Binia A et al. J Hypertens 2015. PMID 26039623. doi:10.1097/HJH.0000000000000611
- Fogacci F et al. Pharmacol Res 2019. PMID 30844537. doi:10.1016/j.phrs.2019.02.028
- Kearns CE et al. Sugar industry and coronary heart disease research. JAMA Intern Med 2016. PMID 27617709.
- Kristensen ML et al. BMJ Open 2015. PMID 26408281.
- Sattar N et al. Statins and risk of incident diabetes. Lancet 2010. PMID 20167359.
- Zoungas S et al. (STAREE) 2026. PMID 42670961.
- Morsell 2026. PMID 41864911.
- Long 2026. PMID 42068528.
- Bonnet 2026. PMID 42580354.
- Cholesterol Treatment Trialists’ Collaboration. Statin therapy and muscle symptoms. Lancet 2022. PMID 36049498.
- Cholesterol Treatment Trialists’ Collaboration 2026. PMID 41655587.
- NHS. Heart attack: symptoms; Stroke: Act FAST. nhs.uk, read 8 October 2026.