Heart and blood vessels

Why heart disease comes first#

Heart attacks and strokes kill more people than anything else. Worldwide, ischemic heart disease caused 9.5 million deaths in 2023, and stroke was second at 6.8 million. In the US, cardiovascular disease killed 919,032 people that year.

The disease is quiet for decades and then abrupt. In the 26-year Framingham study, half of sudden deaths in men and 70% in women happened in people with no known heart disease. Waiting for symptoms means waiting for damage.

It is also the most preventable of the major killers. In a 2025 analysis of 2 million people, having none of five classic risk factors at 50 (high blood pressure, high cholesterol, abnormal weight, diabetes, smoking) meant more than a decade of extra life free of cardiovascular disease.

How plaque builds over decades#

Atherosclerosis starts with particles that carry cholesterol. LDL and its relatives (VLDL remnants and Lp(a)) slip through an artery's lining and get stuck in the wall. Immune cells swallow them and, over years, build a plaque. Most heart attacks happen when a plaque cracks or erodes and a clot forms on it. European Atherosclerosis Society (EAS) consensus statements describe this retention of ApoB-containing particles as the starting event.

The causal link is as solid as anything in medicine. A 2017 EAS consensus pooled more than 200 studies and 2 million people from genetics, cohorts, and trials and found that risk rises with both how high LDL is and how long the arteries are exposed to it. Plaque also starts early: in the PESA study, 49.7% of 1,779 adults aged about 45 with no conventional risk factors already had plaque or coronary calcium, and LDL predicted it even at levels labeled normal.

ApoB, LDL, and non-HDL cholesterol#

A standard cholesterol panel reports LDL cholesterol (LDL-C), the mass of cholesterol inside LDL particles. Two other numbers capture more of the problem. Non-HDL cholesterol (total minus HDL) counts the cholesterol in every plaque-causing particle and comes free with any panel. ApoB (apolipoprotein B) is a protein with exactly one copy on each of those particles, so it counts them directly; it costs an extra test.

When the numbers disagree#

LDL-C, non-HDL, and ApoB usually move together. They split in people with high triglycerides, insulin resistance, diabetes, obesity, or very low LDL on treatment, who carry many small, cholesterol-poor particles: LDL-C looks fine while the particle count is high. In 13,015 statin-treated adults in Copenhagen, above-median ApoB with below-median LDL-C meant 49% more heart attacks; the reverse carried no extra risk.

Since you cannot predict whose numbers will split, the optimal plan is ApoB on every lipid panel. Guideline minimum: the 2026 ACC/AHA guideline recommends ApoB selectively, in diabetes, cardiovascular-kidney-metabolic syndrome, high triglycerides, or established disease, once LDL and non-HDL goals are met. HDL, the "good cholesterol," is a marker, not a target: raising it with niacin did not reduce events, and very high HDL is linked to higher mortality.

How often to test#

Optimal: a lipid panel with ApoB every year from your 20s. It is cheap, needs no fasting, and catches the rise that comes with age, weight gain, and menopause, when the odds of an LDL of 130 or more roughly double. While you are starting or changing a drug or a major diet, retest about every 3 months, 4 to 12 weeks after each change as the 2018 guideline advised, until you reach your goal; once stable, yearly is enough. No trial compares intervals, so the yearly schedule is expert practice. Weak

Do not chase single values: short-term noise is about 15 mg/dL, so repeat a surprising result before changing treatment.

Guideline minimum: the 2026 guideline recommends a check at ages 9 to 11 and at 19, then at least every five years, more often if levels are high or treated.

Targets: guidelines vs prevention-minded practice#

Optimal targets. Prevention-minded practice aims lower and starts earlier, because of plaque-years and because trials found no floor: in FOURIER, an LDL of 30 mg/dL (ApoB 38) was safe and cut events. Values are mg/dL.

No trial has tested this in low-risk 35-year-olds, so the first row is expert practice built on genetics and trials. Moderate

Guideline minimum. The 2026 ACC/AHA multisociety guideline, published in March 2026, replaced the 2018 guideline and restored numeric goals tied to 10-year PREVENT risk (see risk calculators). European rows line up only roughly.

Sources: ACC press release, ACC practice summary, a 2026 comparison of the two guidelines, the 2019 ESC/EAS guideline, and its 2025 update, which left the goals unchanged.

When to start treatment#

Optimal: decide on lifetime exposure, not only 10-year risk. Age dominates 10-year calculators, so most people under 50 score "low" while plaque-years pile up. Prevention-minded physicians discuss medication when, after 3 to 6 months of serious lifestyle change:

• ApoB stays above about 80 mg/dL (roughly an LDL of 100) by your late 30s or 40s, about the 20th percentile that Attia uses.
• You have plaque or high lifetime risk: any coronary calcium, a high 30-year PREVENT risk, high Lp(a), diabetes, or a strong family history.
• LDL is 160 or more at any adult age, or 190 or more, which suggests familial hypercholesterolemia.

In 438,952 UK Biobank participants, lifelong 39 mg/dL lower LDL plus 10 mmHg lower systolic pressure was linked to 78% lower odds of major coronary events. No trial has followed treated 35-year-olds for 30 years, so this is a reasoned bet. Moderate

Guideline minimum: the 2026 guideline ties treatment to 10-year PREVENT risk, and calls for early consideration of medication in young adults with an LDL of 160 or more or a strong family history of premature heart disease. The USPSTF threshold is in statins.

Lipoprotein(a) and inherited risk#

Lipoprotein(a), or Lp(a), is an LDL-like particle with an extra protein attached. Your level is set mostly by your genes, stays roughly constant for life, and barely moves with lifestyle. About 1 in 5 people have a high level, which causes both atherosclerosis and narrowing of the aortic valve.

Measure it once, with your first adult lipid panel, as the EAS advises ("preferably in the first lipid profile"); the 2026 ACC/AHA guideline, the NLA, and the ESC/EAS all say at least once in adulthood, so here the optimal plan and the guideline agree. Results 4 years apart correlated at 0.96 in 16,017 people. Recheck only if your result was near a threshold, after menopause (levels tend to rise), or with new kidney or thyroid disease, which raise it.

Labs report Lp(a) in nmol/L (preferred) or mg/dL. If yours is high, nothing yet lowers Lp(a) in a way proven to prevent heart attacks, so lower everything else harder:

• Push ApoB and LDL lower and earlier, as the NLA recommends, and control blood pressure, glucose, and smoking tightly. PCSK9 antibodies also lower Lp(a) by about 20–30%.
• Get a coronary calcium scan at the earlier ages in who should get one to see whether plaque is already present.
• Test your parents, siblings, and children (cascade screening).
• Skip niacin, which lowers Lp(a) a little but did not prevent events. Lipoprotein apheresis, a blood filtering, is approved only for a narrow group with FH, established disease, and Lp(a) of 60 mg/dL or more.

The first big Lp(a) drug trial was a setback: on 4 September 2026, Novartis announced that pelacarsen lowered Lp(a) in 8,323 people with heart disease but did not reduce major events in Lp(a)HORIZON. As of October 2026, two trials of more potent drugs continue: olpasiran's OCEAN(a)-Outcomes (results estimated 2028) and lepodisiran's ACCLAIM-Lp(a) (estimated 2029). The genetics still say high Lp(a) causes disease; what is unproven is that lowering it helps.

Blood pressure#

High blood pressure is the most common and most fixable cardiovascular risk factor. Nearly half of US adults, 48.1%, have it, and only about 1 in 4 of them have it under control. The optimal plan is to know your home average from your 20s, checked every 3 months (measuring at home). Guideline minimum: the USPSTF recommends an office check every year from age 40 or if you are at higher risk, and every three to five years from 18 to 39.

Categories and goals#

The 2025 AHA/ACC guideline kept the 2017 categories, adopted the PREVENT calculator, and moved toward earlier treatment (review). Categories describe your average, and if your two numbers fall in different categories, the higher one counts.

Optimal target: under 120/80. In about a million adults, the risk of dying from heart disease or stroke in middle age doubled with each 20 mmHg of systolic pressure, down to 115. Three trials in high-risk adults compared a systolic target under 120 with under 140, and all found fewer events: SPRINT in 9,361 adults without diabetes (2.19% to 1.65% a year, deaths down 27%), ESPRIT in 11,255 adults (11.1% to 9.7% over 3.4 years), and BPROAD in 12,821 with type 2 diabetes (2.09 to 1.65 events per 100 person-years). The price was more fainting, low blood pressure, and electrolyte problems. Across 48 trials, each 5 mmHg drop cut major events by about 10%, even at starting pressures under 130. Strong

When to treat (optimal). If your home average stays at 130/80 or above after about 3 months of serious lifestyle change, take medication even if your 10-year risk is low. At 120 to 129, keep working on lifestyle, recheck every 3 months, and discuss medication if you have plaque, diabetes, kidney disease, or high risk. Moderate

Guideline minimum: the 2025 guideline's goal is under 130/80, with a target under 120 systolic for people at increased risk, and medication for low-risk stage 1 after a 3- to 6-month lifestyle trial if pressure is not at goal.

Measuring at home#

Office readings mislead both ways ("white-coat" and "masked" hypertension), so the USPSTF says to confirm a diagnosis outside the clinic. Use an upper-arm cuff from the US Blood Pressure Validated Device Listing in the right size: a standard adult cuff fits arms of 27–34 cm, a large one 35–44 cm. Cuffless and smartwatch readings are not recommended for diagnosis or treatment; for devices, see wearables.

How often (optimal). The European Society of Hypertension (ESH) recommends a 7-day home series before visits, while adjusting treatment, and whenever pressure seems to change.

Once you are stable, more is not better: the ESH discourages daily readings for long-term follow-up and advises against adjusting your own dose unless your clinician has set up a plan for it.

What lowers blood pressure#

The optimal plan is a DASH-style diet, sodium near 1,500 mg a day, at least 5% weight loss if you carry extra weight, regular aerobic and resistance exercise, and no alcohol. The 2025 guideline recommends the same habits, with sodium under 2,300 mg as its floor and at most one or two drinks a day. Effects are largest in people who already have high blood pressure.

Potassium-enriched salt substitutes are an easy win: in 20,995 older adults in rural China, a 75% sodium, 25% potassium substitute cut strokes by 14% and deaths by 12%. Check with your doctor first if you have kidney disease or take drugs that raise potassium. See nutrition and exercise, and if you have high blood pressure, get a home sleep apnea test.

Estimating your risk#

Risk calculators#

The AHA's PREVENT calculator is now the US standard for adults 30 to 79 without heart disease. It estimates 10-year and 30-year risk of heart attack and stroke, heart failure, and total cardiovascular disease from age, sex, cholesterol, blood pressure, diabetes, smoking, kidney function, and medications, with optional urine albumin, HbA1c, and zip code. It was built in 6.6 million US adults and does not use race. The older equations overestimated risk by 40–50%, so PREVENT numbers look lower, and the new thresholds account for that: for lipids, low is under 3%, borderline 3% to under 5%, intermediate 5% to under 10%, and high 10% or more. Recalculate it each year with your lipid panel.

Calculators miss things, so the guideline adds risk enhancers: early heart disease in the family (classically a father or brother before 55, mother or sister before 65); psoriasis, rheumatoid arthritis, or HIV; overweight, diabetes, or kidney disease; South Asian or Filipino ancestry; pregnancy complications, early menopause, and polycystic ovary syndrome; and high Lp(a) or hs-CRP. Adults 40 to 75 with diabetes, stage 3 or worse kidney disease, or HIV should get cholesterol-lowering therapy regardless of score.

Family history and familial hypercholesterolemia#

Familial hypercholesterolemia (FH) is an inherited defect in clearing LDL. It affects about 1 in 250 US adults, causes very high LDL from birth, and leads to heart attacks in the 30s to 50s if untreated. It remains underdiagnosed. Suspect it if your LDL is 190 mg/dL or higher, the level European lab guidance flags as possible FH, or if you or a parent or sibling had heart disease young or very high cholesterol.

An expert panel recommends genetic testing of the LDLR, APOB, and PCSK9 genes for anyone with probable FH, then testing relatives ("cascade testing") so they can start treatment young. Every child should have a cholesterol check at ages 9 to 11 (from age 2 if a close relative has FH or early heart disease), as the 2026 guideline recommends; adults should then follow the yearly schedule in how often to test. See genetic tests.

Coronary calcium scans#

A coronary artery calcium (CAC) scan is a quick, non-contrast CT that measures calcified plaque in the heart's arteries as an Agatston score. It takes under 20 minutes and delivers about 1.7 mSv, roughly six months of natural background radiation. In MESA (6,814 adults, 11 years), the 10-year rate of heart attack or stroke was 1.3% to 5.6% with a score of zero and 13.1% to 25.6% above 300, and each doubling of the score added about 14% to risk.

Sources: ACC press release, ACC practice summary, 2026 review, MESA. The guideline also weighs your percentile for age and sex, and the free MESA calculator gives a 10-year risk that includes your score. Under the optimal targets above, any calcium moves you to the plaque row.

Who should get one (optimal). If you are not already on cholesterol-lowering treatment, get a first scan around 40 for men and 45 for women, whatever your 10-year score, and about five years earlier with a strong family history, FH, high Lp(a), diabetes, or a long smoking history. Plaque appears earlier than calculators suggest: among 19,725 adults aged 30 to 45, 26% of White men and 10% of White women already had calcium, and in CARDIA any calcium at 32 to 46 meant five times the rate of coronary events. The Society of Cardiovascular Computed Tomography supports scanning lower-risk adults with a family history of premature heart disease. If you already treat ApoB to an optimal target, a scan adds less. Moderate (strong for prognosis; no trial shows that scanning younger adults prevents events)

Guideline minimum: the 2026 guideline recommends CAC, now at its strongest class, for men 40 and older and women 45 and older at borderline or intermediate risk when the result would change the statin decision.

Know the downsides. The scan includes part of the lungs: in 966 people having cardiac CT, 41.5% had a finding outside the heart, 7% needed follow-up, and 1.2% had one that mattered; one person was harmed by the workup. Ask in advance how lung findings will be handled.

A zero is not a lifetime pass. Under the 2018 guideline, a zero did not justify skipping a statin for smokers, people with diabetes, or those with a strong family history, and the scan cannot see soft, uncalcified plaque. In MESA, 19% of 10-year coronary events in people starting at zero happened before a repeat scan at three to five years, and the analysis suggests rescanning a zero after 3 to 7 years, sooner with diabetes. Once your score is above zero and you are treating to a goal, repeat scans rarely change the plan; statins shrink plaque while making it more calcified, so a rising score on treatment is not a failure.

CT angiography and plaque analysis#

Coronary CT angiography (CCTA) uses injected contrast to show calcified and soft plaque and any narrowing, at about 8.7 mSv; doses vary 37-fold between hospitals, so ask. Its outcome evidence is in people with symptoms: in SCOT-HEART, adding CCTA for 4,146 people with stable chest pain lowered five-year coronary death or heart attack from 3.9% to 2.3%, mostly because more people started prevention, and US guidelines place it in the evaluation of chest pain.

When it is worth it without symptoms. CCTA sees soft plaque that a calcium scan misses: in the Miami Heart Study of 2,359 adults without symptoms, 16% of those with a zero calcium score had plaque, though only 0.8% had a narrowing of 50% or more. Prevention clinics order it, often with AI plaque analysis such as Cleerly (which Aetna covers for known or suspected disease), when finding plaque would change treatment: a zero score with a strong family history, FH, very high Lp(a), or long smoking history. Weak (as screening)

Why it is not for everyone. No trial shows that screening healthy people with CCTA prevents heart attacks, and the Danish DANCAVAS program, which included calcium scans, did not significantly lower deaths among 46,611 men aged 65 to 74. Beyond radiation and contrast, the main risk is a cascade to catheterization and stents, yet in 5,179 people with stable coronary disease and moderate or severe ischemia, an invasive strategy did not reduce heart attacks or deaths. Agree beforehand that plaque without symptoms means stronger prevention, not an automatic stent.

Medications that prevent heart attacks#

Statins#

Statins block cholesterol production in the liver, which then pulls LDL from the blood. Across 26 trials of about 170,000 people, each 39 mg/dL (1 mmol/L) drop in LDL cut major vascular events by 22% and deaths from any cause by 10%, with no increase in cancer. High-intensity statins (atorvastatin 40–80 mg or rosuvastatin 20–40 mg) lower LDL by half or more. Strong

When to start is covered above. Guideline minimum: the USPSTF recommends a statin for adults 40 to 75 with at least one risk factor (high cholesterol, diabetes, high blood pressure, or smoking) and a 10-year risk of 10% or more (older calculator), and selectively at 7.5% to 10%. The 2026 guideline recommends treatment more strongly as PREVENT risk rises.

The real downside is small. Statins raise the chance of a diabetes diagnosis by 10% at moderate doses and 36% at high doses, mostly in people already near the threshold, and the cardiovascular benefit already accounts for it.

Other cholesterol-lowering drugs#

When a statin is not enough or not tolerated, the 2026 guideline adds ezetimibe, bempedoic acid, or a PCSK9 antibody, and favors early combinations; reaching optimal targets often takes one. Status as of October 2026:

Ezetimibe, bempedoic acid, and the PCSK9 antibodies have outcome trials Strong; inclisiran, enlicitide, and obicetrapib are still waiting on theirs Moderate. Icosapent ethyl's benefit was not repeated by a different omega-3 product in STRENGTH, and the 2026 guideline says it may be considered only for statin-treated people with established disease or diabetes, triglycerides of 150–499 mg/dL, and LDL under 100 Mixed. Ordinary fish oil is a different product; see supplements. Triglycerides under 100 mg/dL are optimal, but that is a lifestyle target (weight, alcohol, refined carbohydrates), not a reason for drugs on its own.

Blood pressure medicines#

The 2025 guideline names three first-line classes: an ACE inhibitor or angiotensin receptor blocker (ARB), a long-acting calcium channel blocker such as amlodipine, and a thiazide-type diuretic such as chlorthalidone. Stage 2 starts with two in one pill, which improves adherence; beta blockers are for specific reasons such as heart failure. If side effects bother you, switching classes usually solves them; stopping does not. Strong

Aspirin#

Aspirin prevents clots but causes bleeding, and without heart disease the balance has turned against it. The USPSTF calls it an individual decision at 40 to 59 with a 10-year risk of 10% or more and recommends against starting at 60 or older; the 2019 ACC/AHA guideline says it might be considered for select higher-risk adults 40 to 70 at low bleeding risk. Here the optimal choice and the guideline agree: do not take it routinely.

In ASPREE, daily aspirin in 19,114 healthy older adults did not prolong disability-free life, and deaths were slightly higher (12.7 vs 11.1 per 1,000 a year). Across earlier prevention trials, aspirin cut serious vascular events only from 0.57% to 0.51% a year while raising major bleeding from 0.07% to 0.10%. Negative (for healthy people)

Who should still take it. People who have had a heart attack, stent, bypass, or ischemic stroke get a much larger benefit and should never stop on their own, especially after a stent. Some prevention-minded physicians consider low-dose aspirin for a calcium score of 100 or more with low bleeding risk, where a MESA modeling study estimated benefit exceeded harm; that is observational reasoning and a decision for your doctor. Mixed Aspirin is not a substitute for anticoagulation in atrial fibrillation.

Inflammation and colchicine#

Inflammation helps drive plaque, and hs-CRP (high-sensitivity C-reactive protein) is the usual marker: under 1 mg/L is low risk, 1 to 3 average, over 3 high, and a value above 10 usually means infection or injury, so repeat it. Optimal: measure it with your yearly blood work, as a 2025 ACC scientific statement urged (broad hs-CRP screening alongside LDL); in 27,939 healthy women followed for 30 years, hs-CRP, LDL, and Lp(a) each predicted events independently. Guideline minimum: the 2026 guideline uses hs-CRP selectively as a risk enhancer. Treating inflammation directly has a mixed record:

• Canakinumab cut events by 15% in CANTOS but raised fatal infections; methotrexate did nothing in CIRT; ziltivekimab lowered hs-CRP but not events in the ZEUS trial announced in July 2026.
• Low-dose colchicine (0.5 mg a day) cut events from 7.1% to 5.5% after a heart attack in COLCOT and from 9.6% to 6.8% in stable coronary disease in LoDoCo2, and the FDA approved it in June 2023. The largest trial, CLEAR, then found no benefit (9.1% vs 9.3%) in 7,062 people started right after a heart attack.

Colchicine is a reasonable add-on for some people with established coronary disease, decided with a cardiologist; it interacts with several common drugs and is not a prevention drug for healthy people. Mixed

Rhythm, stroke, and heart failure#

Atrial fibrillation#

Atrial fibrillation (AF) is a fast, chaotic rhythm in the heart's upper chambers that lets blood pool and clot. It carries about a fivefold higher risk of ischemic stroke and causes about 1 in 7 strokes. Lifetime risk is 30–40% for White Americans and about 20% for Black Americans, and in Framingham it rose from about 1 in 5 with optimal risk factors to more than 1 in 3 with one elevated factor.

The 2023 AF guideline targets high blood pressure, excess weight, diabetes, alcohol, smoking, and inactivity, and calls sleep apnea worth screening for; anyone with AF should get a home sleep test. Even one drink a day was linked to 16% higher AF risk, and quitting cut recurrences in regular drinkers with AF from 73% to 53%, and the guideline recommends losing at least 10% of body weight if your BMI is over 27. Exercise lowers AF risk overall, though athletes have about 2.5 times the odds of non-athletes (exercise).

Smartwatches can catch AF: in the Apple Heart Study of 419,297 people, 0.52% got an alert, and 34% of those tested with an ECG patch had AF. The USPSTF finds evidence insufficient to recommend screening adults 50 and older. If your watch alerts you, get a proper ECG.

With AF, the key decision is stroke prevention: the guideline uses the CHA2DS2-VASc score and recommends an anticoagulant when annual stroke risk is 2% or more, roughly a score of 2 or more in men and 3 or more in women, preferring newer direct oral anticoagulants over warfarin for most people. Strong

Stroke#

More than 795,000 Americans have a stroke each year, and about 87% are caused by a blocked artery. The steps that prevent heart attacks prevent most strokes: blood pressure above all, then cholesterol, smoking, diabetes, and anticoagulation for AF.

Learn the B.E. F.A.S.T. signs: sudden loss of Balance, Eye or vision changes, Face drooping, Arm weakness, Speech trouble, and Time to call 911. Note when symptoms started and do not drive yourself; the best treatments work in the first hours. For stroke and dementia, see brain health.

Heart failure#

Heart failure means the heart cannot pump enough to meet the body's needs, and about 6.7 million US adults have it. Its main causes are coronary disease, high blood pressure, diabetes, and obesity, and US guidelines treat people with those conditions as already at stage A. Two drug classes stand out:

• SGLT2 inhibitors (empagliflozin, dapagliflozin) cut heart failure hospitalization or cardiovascular death by 26% in reduced-pumping heart failure and by 21% and 18% in the preserved-pumping form, with or without diabetes. Strong
• GLP-1 drugs helped people with obesity and preserved-pumping heart failure: tirzepatide lowered cardiovascular death or worsening heart failure from 15.3% to 9.9%, and semaglutide improved symptoms with 13% weight loss and cut heart attacks and strokes by 20% in people with obesity and heart disease (metabolic health). Strong

Erectile dysfunction as an early warning#

Erectile dysfunction (ED) often reflects the same blood vessel disease that causes heart attacks and can show up earlier. In a meta-analysis of 92,757 men, ED meant 44% more cardiovascular events and 25% higher death from any cause, with the strongest link in younger men. New ED without an obvious cause is a reason for a full cardiovascular risk assessment, not just a prescription (hormones).

Symptoms you should never ignore#

Action plan#