Age-adjusted deaths from heart disease in the United States fell 72.6% between 1950 and 2019, from 588.8 to 161.5 per 100,000 [1]. That decline did not come from one breakthrough. It came from identifying the causes of coronary heart disease one at a time across seven decades, and then acting on each of them.
This article walks through what the evidence actually shows for each of those risk factors, what the trials measured, and how far the numbers can still fall.
Table of Contents
- What causes a heart attack?
- Does quitting smoking lower cardiovascular risk?
- What blood pressure target does the evidence support?
- Which diet lowers cardiovascular events?
- How did obesity and diabetes change the picture?
- Does LDL cholesterol cause heart attacks?
- Why do people with normal cholesterol have heart attacks?
- Do statins cause dementia?
- What is ApoB and why does it matter?
- What lowers LDL, and how far can it go?
- Is a coronary artery calcium scan worth it?
- References
What causes a heart attack?
A heart attack happens when a coronary artery, one of the vessels feeding the heart muscle, becomes blocked. Blood flow stops, and the muscle downstream of the blockage starves. People who survive one are often left with a heart that no longer pumps properly.
For most of the twentieth century there was no answer to this. When Dwight Eisenhower had a heart attack in Denver on 23 September 1955, the best medicine available to a sitting US president amounted to an oxygen tent, morphine, four other drugs (papaverine, atropine, heparin and warfarin), and nearly seven weeks of bed rest [2]. Prevention was not on the menu because nobody yet knew what to prevent.
The turning point was not the treatment but the press conference. The cardiologist brought in from Boston, Paul Dudley White, stood in front of roughly 150 reporters and explained what a coronary thrombosis was: a clot starving the heart muscle [4]. He then named the factors that might lie behind it. Diet. Alcohol. Tobacco. Exercise. Family history [3].
The idea that the disease had identifiable causes, and that something could be done about them, was radical in 1955. It did not reverse the trend immediately. Overall heart disease mortality was already drifting downward, but the coronary form kept climbing through the 1950s, flattened during the 1960s, and only turned into a genuine decline at the end of that decade [5].
Working out which of those suspected factors actually mattered took the remainder of the century, and each one was settled separately. The first to be resolved was tobacco.
Does quitting smoking lower cardiovascular risk?
Yes, and the population-level data is unusually clean. The US Surgeon General's report linking smoking to disease was published in January 1964. Per-capita cigarette consumption fell that year and never returned to its 1963 peak. By 2011 the average American adult was smoking 1,236 cigarettes a year, down from 4,345, a drop of 72% [6].
At the individual level, a 2019 analysis quantified the benefit of stopping. Against people who continued smoking, those who had quit within the previous five years recorded 6.94 cardiovascular events per 1,000 person-years compared with 11.56, giving a hazard ratio of 0.61. Measured against people who had never smoked at all, the residual excess risk ceased to be statistically significant somewhere between 10 and 15 years after quitting [7].
Smoking cessation is therefore a real and large lever. It is not, however, a complete explanation. Plenty of people who never smoked still had heart attacks, which meant something else was driving the disease.
What blood pressure target does the evidence support?
Blood pressure was not on White's 1955 list. The case for investigating it had actually been made a decade earlier by a death that was impossible to ignore.
Franklin Roosevelt died of a cerebral haemorrhage on 12 April 1945 with a blood pressure of 300/190. For years his physicians had recorded readings of that magnitude as roughly what one would expect at his age. Three years later Congress funded a study of a single ordinary town, and on 11 October 1948 the Framingham Heart Study began collecting data on 5,209 residents [8].
Framingham established the association. Hypertension carried 2.6 times the coronary heart disease risk in men aged 40 to 59, and six times the risk in women of the same age group. The 1961 paper reporting this is widely credited with coining the term "risk factor" [9].
An observational study can show that pressure travels alongside risk. It cannot show that lowering pressure lowers risk. That required a randomised trial, and the trial came from the Veterans Administration in 1967. It enrolled 143 men whose diastolic pressure averaged between 115 and 129 mmHg, at a time when the current standard would call anything above 80 abnormal. Participants were randomly assigned to antihypertensive drugs or placebo [10].
The trial was halted after roughly 18 months because the difference had become impossible to justify continuing [11]. Twenty-seven severe complicating events occurred in the placebo group against two in the treated group. Four deaths occurred on placebo and none on treatment [10].
Targets have moved downward since, and the reason is trial evidence rather than commercial pressure. SPRINT randomised 9,361 people to a systolic target below 120 or below 140. The intensive group had roughly a quarter fewer cardiovascular events (hazard ratio 0.75) and 27% fewer deaths from any cause (hazard ratio 0.73) [12]. A companion analysis found the same direction of benefit in adults aged 75 and older, including participants classified as frail [13].
Two practical qualifications apply to that target:
- Measurement technique matters. Readings should be taken properly, ideally at home, seated and rested.
- Very frail patients, and anyone who becomes dizzy on treatment, may need a higher target.
Reaching a lower pressure does not automatically mean medication. The DASH trial fed 459 adults for eight weeks on a diet rich in fruit, vegetables, fibre, potassium and low-fat dairy, holding sodium intake and body weight constant. Systolic pressure fell 5.5 mmHg further than on the control diet, and 11.4 mmHg further among the 133 participants who actually had hypertension [14]. A follow-up trial layered sodium restriction on top, and the low-sodium DASH combination beat a high-sodium typical American diet by 8.9 mmHg systolic [15].
Which diet lowers cardiovascular events?
The most informative trial on dietary fat and heart disease is CORDIOPREV, which enrolled 1,002 people who already had established coronary heart disease and randomised them to either a Mediterranean diet rich in unsaturated fats or a low-fat diet, with seven years of follow-up.
The primary endpoint occurred in 198 participants: 87 in the Mediterranean group and 111 in the low-fat group. Adjusted hazard ratios ranged from 0.719 to 0.753 in favour of the Mediterranean diet, or roughly 25% fewer events. The effect was more evident in men; among the 175 women in the trial no difference was found between groups [16].
That points toward unsaturated fats from extra-virgin olive oil, avocado, fish, nuts and seeds.
Saturated fat runs the other way. A 2020 Cochrane review of long-term randomised trials found that reducing dietary saturated fat cut combined cardiovascular events by 17% (risk ratio 0.83, 95% CI 0.70 to 0.98) across 12 trials and 53,758 participants, graded as moderate-quality evidence [17].
A useful diet on this evidence is rich in potassium, fibre, unsaturated fats, and protein sources such as chickpeas, lentils, beans and fish, alongside regular exercise. One caveat on fibre: people with irritable bowel syndrome or inflammatory bowel disease need a more nuanced approach than simply increasing intake.
From the MicroVitamin range
Psyllium husk is a soluble fibre. Each daily scoop of MicroVitamin+ Powder includes 2.5 g of it, alongside 26 core vitamins and minerals.
Even so, diet cannot be the whole answer either. Lean people who do not smoke, whose blood pressure sits below 120 systolic, and who eat well still have heart attacks.
How did obesity and diabetes change the picture?
While smoking, blood pressure and diet were improving, two other factors moved in the opposite direction. Adult obesity prevalence rose from 22.9% in 1988 to 1994 to 34.9% in 2011 to 2012 [18], and reached 40.3% between August 2021 and August 2023 [19]. Diagnosed diabetes nearly tripled, from 2.5% in 1990 to 7.2% in 2013 [18].
An attribution analysis of the 1980 to 2000 fall in US coronary deaths makes the cost visible. Rising body-mass index accounted for 25,905 deaths in the wrong direction, and rising diabetes prevalence for 33,465. Every other factor in the model pushed deaths down; these two pushed them up [20].
The drivers are food abundance and sedentary work. A controlled feeding study is instructive here: when meals were matched for presented calories, energy density, macronutrients, sugar, sodium and fibre, participants ate 508 kcal per day more on the ultra-processed diet, gaining 0.9 kg, and lost 0.9 kg on the unprocessed diet [21].
For fifty years there was no effective pharmacological answer to obesity. The eventual answer came from an unlikely direction.
In 1980 a gastroenterologist named Jean-Pierre Raufman began work at the National Institutes of Health on a question about venom and the pancreas. He later described it as the kind of "fishing expedition" a modern grant committee would reject, and said he "would never have conceived that there was any potential therapeutic benefit for anything" [22]. Of all the venoms his team tested, the strongest reaction came from a Gila monster.
The Gila monster is one of the few venomous lizards on Earth. It lives underground in the deserts of the American Southwest and eats about four or five times a year, consuming as much as a third of its body weight in a single meal, yet its blood sugar stays remarkably stable between those meals [23].
At the Veterans Administration hospital in the Bronx, endocrinologist John Eng recognised the significance. He had trained under Nobel laureate Rosalyn Yalow, who developed a method for detecting minute quantities of hormone in blood, and he applied that technique to the lizard venom [25]. In 1992 he isolated a previously undescribed compound, exendin-4, about 53% identical to the human hormone GLP-1, the signal that prompts the pancreas to release insulin when blood sugar rises [24]. The decisive difference: human GLP-1 is destroyed within minutes, while the lizard version persisted for hours.
Nothing then happened for years. The Veterans Administration declined to patent the discovery, and drug companies were not interested in a peptide from lizard venom. In the mid-1990s Eng submitted a poster at the American Diabetes Association annual meeting [25], where a researcher from a small company called Amylin stopped to read it. "It was the most effective antidiabetic agent anyone had ever seen," that researcher, Andrew Young, said afterwards [26].
The first GLP-1 drug, exenatide, was approved by the FDA in late April 2005. It was modest: twice-daily injections, reasonable blood-sugar control, a few kilograms of weight loss. What it proved was that the target worked. Liraglutide followed in January 2010 as a once-daily agent, and semaglutide in December 2017 as a once-weekly one, each more potent and longer-acting than the last.
The surprise arrived at the higher dose tested for obesity rather than diabetes. At 2.4 mg weekly, mean body weight fell 14.9% from baseline at week 68, against 2.4% on placebo, a treatment difference of 12.4 percentage points [27]. The FDA approved that dose for obesity in June 2021.
Weight loss and cardiovascular protection are different endpoints, and no weight-loss drug had ever been shown to prevent a heart attack until 2023. SELECT randomised 17,604 people with established heart disease and a BMI above 27, but without diabetes, to weekly semaglutide or placebo for about three years. A primary cardiovascular endpoint event occurred in 6.5% on semaglutide against 8.0% on placebo, a hazard ratio of 0.80 [28].
Tirzepatide, which combines GLP-1 activity with a second gut hormone, GIP, followed. In SURMOUNT-1 the 15 mg weekly dose produced a mean weight change of 20.9% at week 72, against 3.1% on placebo [29]. Retatrutide adds a third hormone, glucagon. It remains investigational and legally available only to trial participants, but in a phase 3 trial of 2,339 people over 80 weeks the 12 mg dose produced a 28.3% weight reduction against 2.2% on placebo [30].
Blood pressure falls alongside the weight. A meta-analysis of 25 randomised trials covering 4,874 people put the relationship at roughly 1.05 mmHg systolic and 0.92 mmHg diastolic per kilogram lost [31].
This also explains the reversal after 2019. Risk-factor control deteriorated during the pandemic. The national cardiovascular age-adjusted mortality rate had declined 8.9% from 2010 to 2019, then rose 9.3% between 2019 and 2022, producing 228,524 excess cardiovascular deaths and erasing almost a decade of progress [32]. Statin dispensing dipped in the second quarter of 2020, by between 0.8% for high-intensity statins and 4.1% for low-intensity ones, before resuming its climb [33].
Does LDL cholesterol cause heart attacks?
This is the question that separates correlation from causation, and the cleanest answer came from genetics rather than from a drug trial.
In the 1990s a Paris team led by geneticist Catherine Boileau was studying French families showing every hallmark of familial hypercholesterolaemia, the inherited condition that drives cholesterol very high and causes heart attacks before age fifty. The known causative genes were normal in these families. Working with other laboratories, the team identified mutations in a gene called PCSK9 that, when overactive, sends blood cholesterol soaring [34].
The more informative finding was the mirror image. When researchers sequenced PCSK9 in people with unusually low LDL, they found two nonsense mutations that break the gene. These turned out to be common, carried by about 2% of African Americans and fewer than 0.1% of European Americans, and carriers ran LDL cholesterol roughly 40% lower than everyone else [35].
The decisive study followed those carriers for fifteen years in a cohort of nearly 13,000 Americans. Among 3,363 black participants, the 2.6% carrying a nonsense mutation had 28% lower mean LDL cholesterol and an 88% lower risk of coronary heart disease (hazard ratio 0.11, 95% CI 0.02 to 0.81). Among 9,524 white participants, 3.2% carried a milder variant associated with 15% lower LDL and 47% less coronary heart disease (hazard ratio 0.50, 95% CI 0.32 to 0.79) [36].
Nothing else distinguished these people. They took no drug and changed no diet. They were born with a gene variant that gave them less cholesterol in their blood across an entire lifetime, and they developed dramatically less coronary disease. This is as close to a randomised lifetime experiment as human biology offers.
Researchers subsequently identified an individual with no functioning copy at all: a compound heterozygote with no detectable circulating PCSK9 and an LDL cholesterol of 14 mg/dL. The paper describes her as a healthy, fertile college graduate [37].
Pooling more than 200 studies covering some two million people, every mechanism of lowering LDL cholesterol, whether statin, ezetimibe, antibody therapy or an inherited variant, reduces heart attacks in proportion to the size of the LDL reduction. The reviewers concluded that the evidence "unequivocally establishes that LDL causes ASCVD" [38].
Why do people with normal cholesterol have heart attacks?
Because the reference range on a lab report is a population average, not a biological threshold.
A 2009 analysis measured admission cholesterol in 136,000 Americans hospitalised with coronary artery disease. Half had an admission LDL below 100 mg/dL, and 17.6% were below 70. Fewer than a quarter were above 130. Only 21.1% had been taking any lipid-lowering medication before admission. The authors read this as support for guideline revisions targeting even lower LDL goals [39].
The PESA study explains why. Investigators scanned the arteries of 1,779 middle-aged adults, aged 40 to 54, who had none of the conventional cardiovascular risk factors. Subclinical atherosclerosis was present in 49.7% of them. LDL cholesterol was independently associated with both the presence and the extent of plaque, and this held even within the subgroup of 740 participants whose blood pressure, fasting glucose, HbA1c and total cholesterol were all in the optimal range. The authors' reading of their own data is that atherosclerosis develops above an LDL threshold of approximately 50 to 60 mg/dL, which is about 1.3 to 1.5 mmol/L [40].
Guidelines had settled on an LDL below 70 mg/dL for preventing a second heart attack. PESA implies plaque can still form at that level, which raised the question of whether a lower target performs better.
Ez-PAVE tested exactly that in March 2026. It randomised 3,048 people with atherosclerotic cardiovascular disease to a target LDL below 55 mg/dL or below 70 mg/dL, reaching the lower target by adding ezetimibe, an inexpensive non-statin tablet, on top of a statin. Over three years a primary endpoint event occurred in 6.6% of the intensive-target group against 9.7% of the conventional-target group, a hazard ratio of 0.67 [41].
Three qualifications belong with that result. Ez-PAVE tested a target rather than a drug, it was open-label so everyone knew which target they were aiming for, and much of the margin came from fewer revascularisation procedures rather than fewer deaths.
Do statins cause dementia?
The concern rests on a reasonable premise: the brain does need cholesterol. What the premise misses is where that cholesterol comes from. Almost all brain cholesterol is synthesised locally, behind a blood-brain barrier that efficiently prevents exchange with lipoprotein cholesterol in the circulation [42]. Lowering the number in your blood does not starve the brain.
It has also been tested directly. EBBINGHAUS gave 1,204 people a PCSK9-blocking antibody or placebo and ran a formal battery of cognitive tests over roughly 18 months. The change in the spatial working memory strategy index was 0.21 in the treated group and 0.29 in the placebo group, meeting the prespecified noninferiority threshold at P<0.001, with no association between LDL cholesterol levels and cognitive change [43].
A large observational study published in 2026 points in the same direction. Among 132,585 statin-naive Danes with new type 2 diabetes followed for a median 7.1 years, 2.7% developed dementia, and those who started a statin early had a 15% lower relative risk of dementia over 10 years rather than a higher one. Because it is observational it cannot establish cause, and the absolute difference is small on a base rate under 3%. It is nevertheless the opposite of the feared direction [44].
What is ApoB and why does it matter?
Apolipoprotein B is a structural protein, and every lipoprotein particle capable of entering an artery wall carries exactly one copy of it. That makes an ApoB measurement a particle count rather than a cholesterol mass, and it is the particle count that drives the damage.
Across roughly 430,000 people, when ApoB, LDL cholesterol and triglycerides were assessed together in one model, only ApoB remained associated with myocardial infarction risk (adjusted hazard ratio per 1 standard deviation 1.27, 95% CI 1.15 to 1.40). The risk was best captured by the number of ApoB-containing lipoproteins [45].
For context on what a physiological level looks like, cord-blood ApoB in full-term newborns measures about 28 mg/dL [46]. Above roughly that range, ApoB particles begin depositing cholesterol into artery walls.
The public conversation around this has had measurable consequences. During the 2010s an online backlash against cholesterol lowering spread widely. In a Danish cohort of 674,900 adults started on a statin between 1995 and 2010, negative statin-related news stories raised the odds of early discontinuation (odds ratio 1.09, 95% CI 1.06 to 1.12), and early discontinuation carried a higher rate of myocardial infarction (hazard ratio 1.26, 95% CI 1.21 to 1.30) [47]. Over the same decade, US statin use nevertheless rose 79.8%, from 21.8 million adults in 2002 to 2003 to 39.2 million in 2012 to 2013 [48].
What lowers LDL, and how far can it go?
The cheapest options remain generic statins such as rosuvastatin, which are particularly effective when paired with ezetimibe, the combination Ez-PAVE used to reach its lower target.
Statin muscle symptoms are real but less common than widely assumed. The National Lipid Association estimates the prevalence of pharmacologic statin-associated muscle symptoms at about 1% to 2%, with a range of 0.5% to 4% [49]. The SAMSON trial separated drug effect from expectation by cycling participants through statin, placebo and no-tablet months: mean symptom intensity was 8.0 in no-tablet months, 15.4 in placebo months, and 16.3 in statin months [50]. Most of the symptom burden appeared on placebo.
For people who cannot tolerate statins, or who need more reduction than statins deliver, PCSK9 inhibitors replicate what the broken gene does naturally. The FDA approved two monoclonal antibodies in 2015: alirocumab on 24 July and evolocumab on 27 August [51].
The VESALIUS-CV trial extended that evidence into primary prevention. It randomised 12,257 patients with atherosclerosis or diabetes, but without a previous myocardial infarction or stroke, and with LDL cholesterol of at least 90 mg/dL, to evolocumab or placebo. Over a median 4.6 years, a 3-point major adverse cardiac event occurred in 6.2% of the evolocumab group against 8.0% on placebo (hazard ratio 0.75, 95% CI 0.65 to 0.86). No between-group difference was seen in safety events [52].
Timing multiplies the benefit. Mendelian randomisation across 312,000 people found that naturally randomised long-term exposure to lower LDL cholesterol was associated with a 54.5% reduction in coronary heart disease risk per mmol/L, roughly three times the reduction per unit of LDL seen when a statin is started later in life [53].
The obstacle with PCSK9 antibodies has been that they require injection and cost a great deal. That changed on 16 July 2026, when the FDA approved enlicitide, an oral PCSK9 inhibitor [54]. It is taken as one 20 mg tablet daily on an empty stomach, since food within 30 minutes roughly halves exposure, and it lowers LDL cholesterol by 56% to 59%, the same range as the injectable antibodies [55]. In a trial of adults with heterozygous familial hypercholesterolaemia, LDL fell 58.2% at week 24 against a 2.6% rise on placebo, with adverse events, serious adverse events and discontinuations similar between groups [56]. It is not inexpensive yet, with a US list price reported at about $315 for a 30-day supply [57], but the injection is no longer required.
Further out, a base-editing therapy aims to switch off PCSK9 permanently with a single treatment. The evidence is early, a phase 1b study in 35 people, but a single infusion at the top dose reduced LDL cholesterol by up to 62%, with durability observed up to 18 months and no treatment-related serious adverse events or dose-limiting toxicities reported [58].
Meanwhile the population trend has resumed its descent. The US heart disease death rate fell 3.1% in 2023, from 167.2 to 162.1 per 100,000 age-adjusted [59]. What remains is largely a matter of using tools that already exist. Among 600,000 Americans with established atherosclerotic cardiovascular disease, statin therapy was prescribed to only about half, and only about one in five received a high-intensity statin [60].
Is a coronary artery calcium scan worth it?
It depends entirely on whether the result would change the decision.
A coronary artery calcium score exists to resolve uncertainty about whether to start treatment. Where a patient is hesitant about beginning lipid-lowering therapy, scanning is reasonable if the result is likely to affect their acceptance of that therapy [61].
Once treatment is underway, repeat scanning is not indicated. Serial scoring is not recommended in patients with established coronary artery calcium above zero, nor as a way of assessing whether cholesterol-lowering therapy is working [61]. The 2026 ACC/AHA dyslipidemia guideline is explicit that repeat testing after lipid-lowering therapy has started is not indicated, because statin therapy often increases calcium scores even while stabilising or reducing plaque volume [62].
The practical implication is that someone already controlling every modifiable risk factor gains nothing from the scan. The risk factors that drive coronary heart disease have been identified, and to a large degree addressed: smoking, blood pressure, diet, exercise, obesity through GLP-1 therapies, and ApoB above the level we are born with.
References
1. https://www.cdc.gov/nchs/data/hus/2020-2021/SlctMort.pdf
2. https://doi.org/10.1056/NEJMp2031046
3. https://www.heart.org/en/news/2024/02/15/the-presidential-heart-attack-that-changed-america
4. https://protomag.com/cardiology/the-heart-of-a-president/
5. https://www.cdc.gov/nchs/nvss/mortality/hist293.htm
6. https://pmc.ncbi.nlm.nih.gov/articles/PMC3910051/
7. https://pubmed.ncbi.nlm.nih.gov/31429895/
8. https://pmc.ncbi.nlm.nih.gov/articles/PMC4159698/
9. https://pmc.ncbi.nlm.nih.gov/articles/PMC9316391/
10. https://jamanetwork.com/journals/jama/fullarticle/336799
11. https://profiles.nlm.nih.gov/spotlight/xf/feature/study
12. https://www.nejm.org/doi/full/10.1056/NEJMoa1511939
13. https://jamanetwork.com/journals/jama/fullarticle/2524266
14. https://pubmed.ncbi.nlm.nih.gov/9099655/
15. https://pubmed.ncbi.nlm.nih.gov/11136953/
16. https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(22)00122-2/abstract
17. https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD011737.pub3/full
18. https://jamanetwork.com/journals/jamacardiology/fullarticle/2530594
19. https://www.cdc.gov/nchs/products/databriefs/db508.htm
20. https://doi.org/10.1056/NEJMsa053935
21. https://doi.org/10.1016/j.cmet.2019.05.008
22. https://whyy.org/segments/ozempic-how-gila-monster-venom-led-to-weight-loss-drugs/
23. https://askabiologist.asu.edu/gila-monster-food
24. https://www.research.va.gov/research_in_action/Diabetes-drug-from-Gila-monster-venom.cfm
25. https://www.goldengooseaward.org/01awardees/diabetes-medication
27. https://www.nejm.org/doi/full/10.1056/NEJMoa2032183
28. https://www.nejm.org/doi/full/10.1056/NEJMoa2307563
29. https://doi.org/10.1056/NEJMoa2206038
31. https://doi.org/10.1161/01.HYP.0000094221.86888.AE
32. https://pubmed.ncbi.nlm.nih.gov/37972797/
33. https://pmc.ncbi.nlm.nih.gov/articles/PMC10033441/
34. https://pubmed.ncbi.nlm.nih.gov/12730697/
35. https://pubmed.ncbi.nlm.nih.gov/15654334/
36. https://www.nejm.org/doi/full/10.1056/NEJMoa054013
37. https://pubmed.ncbi.nlm.nih.gov/16909389/
38. https://academic.oup.com/eurheartj/article/38/32/2459/3745109
39. https://pubmed.ncbi.nlm.nih.gov/19081406/
40. https://www.jacc.org/doi/10.1016/j.jacc.2017.10.024
41. https://www.nejm.org/doi/full/10.1056/NEJMoa2600283
42. https://doi.org/10.1161/01.ATV.0000120374.59826.1b
43. https://doi.org/10.1056/NEJMoa1701131
45. https://doi.org/10.1001/jamacardio.2021.5083
46. https://pubmed.ncbi.nlm.nih.gov/16810415/
47. https://doi.org/10.1093/eurheartj/ehv641
48. https://doi.org/10.1001/jamacardio.2016.4700
49. https://doi.org/10.1016/j.jacl.2022.09.001
50. https://www.nejm.org/doi/full/10.1056/NEJMc2031173
51. https://www.accessdata.fda.gov/scripts/cder/daf/
52. https://www.nejm.org/doi/full/10.1056/NEJMoa2514428
53. https://www.jacc.org/doi/10.1016/j.jacc.2012.09.017
54. https://www.accessdata.fda.gov/drugsatfda_docs/appletter/2026/220848Orig1s000ltr.pdf
55. https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=100ec543-fbd0-44fc-b740-db9cdff39145
56. https://jamanetwork.com/journals/jama/article-abstract/2841258
57. https://www.healthline.com/health-news/fda-approves-new-cholesterol-pill-lipfendra
59. https://www.ncbi.nlm.nih.gov/books/NBK611296/
