MVIIIHealth The MVIII app
← Back

Topics

Heart health

Everything in the library studied for it, medical and holistic together. Within each kind, what was shown to help comes first, strongest evidence first. What was tested and not shown to help stays on the list.

Strong several reviews agree · Moderate one review or consistent trials · Limited one or a few small studies · Untested no study in people, listed with what it rests on

Medication71Supplements27Practices16Daily life14Food and drink23Diets8Lab markers106

Medication

ACE inhibitorsStrong

Heart failure with reduced pumping function

Shown to lower deaths by about 20 to 23% and hospital readmissions for heart failure by about a third in people with heart failure or a weakened left ventricle.

The numbers and sources
Effect
Pooling placebo-controlled trials of at least 8 weeks in symptomatic heart failure, ACE inhibitors lowered total mortality (OR 0.77, 95% CI 0.67 to 0.88) and the combination of death or hospitalization for heart failure (OR 0.65, 95% CI 0.57 to 0.74). Most data came from enalapril, captopril, ramipril, quinapril and lisinopril [4]. In individual data from 5 long-term trials, 23.0% died on an ACE inhibitor and 26.8% on placebo (OR 0.80, 95% CI 0.74 to 0.87). Readmission for heart failure occurred in 13.7% against 18.9% (OR 0.67, 95% CI 0.61 to 0.74) and reinfarction in 8.9% against 11.0% [5].
Population
Randomized placebo-controlled trials in symptomatic heart failure [4]; 12,763 people with heart failure or left ventricular dysfunction in 5 trials, followed an average of 35 months [5].
Sources
[4], [5]
Caveats
Three of the 5 long-term trials enrolled people within a week of a heart attack [5]. Both overviews were published by 2000.
Antiplatelets: clopidogrel and low-dose aspirinStrongCommon

After a heart attack, stroke or with artery disease

Shown to cut serious vascular events by about a quarter in high-risk patients, about 36 fewer per 1,000 treated for two years after a heart attack.

The numbers and sources
How common
Common: between 1 in 100 and 1 in 10 people.
Effect
Among high-risk patients, antiplatelet therapy lowered serious vascular events by about one quarter, nonfatal heart attack by one third, nonfatal stroke by one quarter and vascular death by one sixth. The absolute reduction was 36 per 1,000 treated for two years after a heart attack, 36 per 1,000 after a stroke or transient ischemic attack, and 38 per 1,000 treated for one month during an acute heart attack. Aspirin was the most widely studied drug [1]. In secondary prevention trials with individual data, serious vascular events occurred at 6.7% a year on aspirin against 8.2% on control, total stroke at 2.08% against 2.54% and coronary events at 4.3% against 5.3% [2].
Population
High-risk patients in a collaborative meta-analysis of randomized antiplatelet trials [1]; secondary prevention trials with individual participant data [2].
Sources
[1], [2]
Caveats
Both analyses come from the same collaboration. The band above is the absolute reduction of 36 per 1,000 over two years [1].
Beta-blockersStrong

Heart failure with reduced pumping function

Shown to lower deaths by about a third in heart failure with reduced pumping function, with no difference found between individual beta-blockers.

The numbers and sources
Effect
A network meta-analysis of 21 trials found beta-blockers lowered mortality against placebo or standard treatment after a median 12 months (OR 0.69, 95% credible interval 0.56 to 0.80), with no clear difference between individual drugs [1]. In individual data from 11 trials, beta-blockers lowered all-cause and cardiovascular mortality in people in sinus rhythm across ejection fraction groups, except in the small group at 50% or above [2]. In the three largest trials, deaths were 11.8% against 17.3% on bisoprolol (HR 0.66, 95% CI 0.54 to 0.81) [3], 7.2% against 11.0% per patient-year on extended-release metoprolol (RR 0.66, 95% CI 0.53 to 0.81) [4], and 130 against 190 on carvedilol, a 35% lower risk (95% CI 19 to 48) [5].
Population
21 trials of six beta-blockers [1]; 14,262 people in sinus rhythm with a median ejection fraction of 27% [2]; 2,647 people with ejection fraction of 35% or less [3]; 3,991 people with ejection fraction of 40% or less [4]; 2,289 people with ejection fraction below 25% [5].
Sources
[1], [2], [3], [4], [5]
Caveats
All three large trials were stopped early for benefit. In people with heart failure and atrial fibrillation, beta-blockers did not improve prognosis [2].
Direct oral anticoagulants (DOACs)Strong

Stroke prevention in atrial fibrillation, against warfarin

Shown to cut stroke or systemic embolism by about 19% and death by about 10% against warfarin in adults with atrial fibrillation.

The numbers and sources
Effect
Across the four pivotal trials, DOACs lowered stroke or systemic embolism by 19% against warfarin (RR 0.81, 95% CI 0.73 to 0.91), mainly through fewer hemorrhagic strokes (RR 0.49, 95% CI 0.38 to 0.64), and lowered death from any cause (RR 0.90, 95% CI 0.85 to 0.95) [1]. In patient-level data from the same trials, stroke or systemic embolism occurred in 3.01% on a standard-dose DOAC against 3.69% on warfarin (HR 0.81, 95% CI 0.74 to 0.89) and death in 7.76% against 8.42% (HR 0.92, 95% CI 0.87 to 0.97) [2]. By trial, the yearly rate of stroke or systemic embolism was 1.11% on dabigatran 150 mg against 1.69% on warfarin [3], 1.7% on rivaroxaban against 2.2% [4], 1.27% on apixaban against 1.60% [5] and 1.18% on high-dose edoxaban against 1.50% [6].
Population
71,683 participants in four trials, 42,411 on a DOAC and 29,272 on warfarin [1][2]; 18,113 patients [3]; 14,264 patients [4]; 18,201 patients [5]; 21,105 patients [6].
Sources
[1], [2], [3], [4], [5], [6]
Caveats
Lower-dose regimens did not lower stroke or systemic embolism against warfarin (HR 1.06, 95% CI 0.95 to 1.19) [2] and were followed by more ischemic strokes (RR 1.28, 95% CI 1.02 to 1.60) [1]. Both meta-analyses pool the same four manufacturer-funded trials. The trials enrolled nonvalvular atrial fibrillation.
SGLT2 inhibitorsStrong

Heart failure hospitalization and cardiovascular death

Shown to cut cardiovascular death or heart failure hospitalization by about 23% in adults with heart failure, with or without diabetes.

The numbers and sources
Effect
Across five trials in heart failure, SGLT2 inhibitors lowered the composite of cardiovascular death or hospitalization for heart failure (HR 0.77, 95% CI 0.72 to 0.82), cardiovascular death (HR 0.87, 95% CI 0.79 to 0.95), first hospitalization for heart failure (HR 0.72, 95% CI 0.67 to 0.78) and death from any cause (HR 0.92, 95% CI 0.86 to 0.99). Effects were consistent across ejection fraction [1]. With dapagliflozin in reduced ejection fraction, worsening heart failure or cardiovascular death occurred in 16.3% against 21.2% on placebo over a median 18.2 months (HR 0.74, 95% CI 0.65 to 0.85) [2]. With empagliflozin in preserved ejection fraction, cardiovascular death or heart failure hospitalization occurred in 13.8% against 17.1% over a median 26.2 months (HR 0.79, 95% CI 0.69 to 0.90) [3]. Across 13 large placebo-controlled trials, the risk of cardiovascular death or heart failure hospitalization fell by 23% (RR 0.77, 95% CI 0.74 to 0.81), with similar effects in people with and without diabetes [4].
Population
21,947 participants in five heart failure trials [1]; 4,744 adults with heart failure and reduced ejection fraction [2]; 5,988 adults with heart failure and preserved ejection fraction [3]; 90,409 participants in 13 trials, 82.7% with diabetes [4].
Sources
[1], [2], [3], [4]
Caveats
The heart failure trials tested dapagliflozin, empagliflozin and sotagliflozin. Canagliflozin has no heart failure approval. In the preserved ejection fraction trial the effect was mainly on hospitalization [3]. Both single trials cited were funded by the manufacturers [2][3].
SGLT2 inhibitorsStrong

Heart attack, stroke and cardiovascular death in type 2 diabetes

Shown to cut major cardiovascular events by about 10% in adults with type 2 diabetes, with the benefit clearest in people with established cardiovascular disease.

The numbers and sources
Effect
Across three cardiovascular outcome trials, SGLT2 inhibitors lowered the composite of heart attack, stroke or cardiovascular death by 11% (HR 0.89, 95% CI 0.83 to 0.96). The benefit was seen in people with established cardiovascular disease (HR 0.86, 95% CI 0.80 to 0.93) and not in people without it (HR 1.00, 95% CI 0.87 to 1.16) [9]. Across six trials the estimate was similar (HR 0.90, 95% CI 0.85 to 0.95), and cardiovascular death fell (HR 0.85, 95% CI 0.78 to 0.93) with significant differences between trials [10]. With empagliflozin the composite occurred in 10.5% against 12.1% on placebo over a median 3.1 years (HR 0.86), cardiovascular death in 3.7% against 5.9% and death from any cause in 5.7% against 8.3%, with no significant difference in heart attack or stroke [11]. With dapagliflozin the composite occurred in 8.8% against 9.4% (HR 0.93, 95% CI 0.84 to 1.03) [12]. With canagliflozin it occurred at 26.9 against 31.5 per 1,000 patient-years (HR 0.86, 95% CI 0.75 to 0.97) [13].
Population
34,322 patients in three trials, 60.2% with established cardiovascular disease [9]; 46,969 patients in six trials, mean age 63.7 [10]; 7,020 adults with type 2 diabetes and cardiovascular disease [11]; 17,160 adults with type 2 diabetes [12]; adults with type 2 diabetes and raised cardiovascular risk, 65.6% with cardiovascular disease [13].
Sources
[9], [10], [11], [12], [13]
Caveats
The dapagliflozin trial did not find a lower rate of the composite [12]. The effect on cardiovascular death differed between drugs (I² 64.3%) [10]. One meta-analysis found the benefit only in people with established disease, and the other found no such difference [9][10].
StatinsStrong

Heart attack, stroke and death

Shown to cut major vascular events about 22% and deaths about 10% for each 1 mmol/L drop in LDL cholesterol, including in lower-risk people.

The numbers and sources
Effect
Per 1.0 mmol/L reduction in LDL cholesterol, major vascular events fell with a rate ratio of 0.78 (95% CI 0.76 to 0.80) and all-cause mortality with 0.90 (95% CI 0.87 to 0.93). More intensive regimens cut major vascular events a further 15% (95% CI 11 to 18) against less intensive ones [1]. In people with a 5-year risk under 10%, each 1 mmol/L reduction cut major vascular events by about 11 per 1,000 over 5 years. In people without vascular disease, all-cause mortality RR was 0.91 (95% CI 0.85 to 0.97) per 1 mmol/L [2].
Population
Individual data from 26 trials with about 170,000 participants, median follow-up about 5 years [1]; 27 trials with 174,149 participants, grouped by baseline risk [2].
Sources
[1], [2]
Caveats
Both analyses come from the same collaboration, and the same 5 intensity trials appear in both. Effects are expressed per unit of LDL lowered, not per drug or dose.
WarfarinStrongCommon

Stroke prevention in atrial fibrillation, against no treatment

Shown to cut stroke by about 64% against placebo or no treatment in adults with nonvalvular atrial fibrillation, with fewer deaths in one review.

The numbers and sources
How common
Common: between 1 in 100 and 1 in 10 people.
Effect
Across six trials, adjusted-dose warfarin lowered stroke by 64% (95% CI 49 to 74) against control [1]. In five trials of people with no earlier stroke, warfarin lowered all stroke (OR 0.39, 95% CI 0.26 to 0.59), disabling or fatal stroke (OR 0.47, 95% CI 0.28 to 0.80) and death (OR 0.69, 95% CI 0.50 to 0.94) over a mean 1.5 years. The reviewers estimated about 25 strokes prevented each year for every 1,000 people treated [2].
Population
2,900 participants in six trials, within a meta-analysis of 29 trials and 28,044 participants, mean age 71 [1]; 2,313 participants without earlier stroke in five trials, mean age 69 [2].
Sources
[1], [2]
Caveats
The band above is the estimate of 25 strokes prevented a year per 1,000 treated. Follow-up was short, and bleeding estimates from these trials were imprecise [2].
WarfarinStrong

Stroke prevention in atrial fibrillation, against aspirin

Shown to cut stroke by about 39% more than antiplatelet drugs in atrial fibrillation, and to halve stroke against aspirin in adults aged 75 or over.

The numbers and sources
Effect
Across 12 trials, adjusted-dose warfarin lowered stroke by 39% (95% CI 22 to 52) against antiplatelet therapy [1]. In people aged 75 or over, stroke, intracranial hemorrhage or systemic embolism occurred at 1.8% a year on warfarin against 3.8% on aspirin (RR 0.48, 95% CI 0.28 to 0.80). Extracranial hemorrhage occurred at 1.4% a year against 1.6% [3].
Population
12,963 participants in 12 trials [1]; 973 patients aged 75 or over, mean age 81.5, followed a mean 2.7 years [3].
Sources
[1], [3]
Caveats
Trial quality varied and was often incompletely reported [1].
ACE inhibitorsModerate

After a heart attack

Shown to lower deaths in the first 30 days after a heart attack, 7.1% against 7.6%, and more in long-term trials of people with a weakened left ventricle.

The numbers and sources
Effect
In individual data from 4 trials that started an ACE inhibitor within 36 hours of a heart attack and continued it 4 to 6 weeks, 30-day mortality was 7.1% on an ACE inhibitor and 7.6% in controls, a 7% proportional reduction (95% CI 2 to 11%). Nonfatal heart failure occurred in 14.6% against 15.2% [6]. In GISSI-3, the lisinopril trial among them, 6.4% on lisinopril and 7.2% of those not given it died by 6 weeks [3]. In 3 trials that enrolled people with heart failure or left ventricular dysfunction after a heart attack and treated them long term, 23.4% died on an ACE inhibitor and 29.1% on placebo (OR 0.74, 95% CI 0.66 to 0.83) [5].
Population
98,496 people in 4 trials of early short-term treatment [6]; 19,394 people in GISSI-3 [3]; 5,966 people in 3 long-term trials after a heart attack [5].
Sources
[6], [3], [5]
Caveats
GISSI-3 was unblinded [3]. Persistent low blood pressure was more frequent on an ACE inhibitor in the early trials, 17.6% against 9.3% [6].
ACE inhibitorsModerate

Heart attack, stroke and death in high-risk adults without heart failure

Shown to lower heart attack, stroke or cardiovascular death from 12.8% to 10.7% over about 4.5 years in adults with vascular disease and no heart failure.

The numbers and sources
Effect
In the HOPE trial, ramipril 10 mg a day for a mean of 5 years lowered the composite of heart attack, stroke or cardiovascular death from 17.8% on placebo to 14.0% (RR 0.78, 95% CI 0.70 to 0.86). Death from any cause was 10.4% against 12.2% and stroke 3.4% against 4.9% [7]. Combining HOPE with two later trials, ACE inhibitors lowered all-cause mortality (7.8% against 8.9%), nonfatal heart attack (5.3% against 6.4%) and stroke (2.2% against 2.8%). The composite of cardiovascular death, nonfatal heart attack or stroke occurred in 10.7% against 12.8% (OR 0.82, 95% CI 0.76 to 0.88) [8]. Across 20 trials in which at least two thirds of participants had hypertension, ACE inhibitors were associated with a 10% reduction in all-cause mortality (HR 0.90, 95% CI 0.84 to 0.97) [9].
Population
9,297 adults aged 55 or older with vascular disease, or diabetes plus one other risk factor, and no known heart failure [7]; 29,805 people in 3 trials followed a mean of about 4.5 years [8]; 158,998 people in 20 trials [9].
Sources
[7], [8], [9]
Caveats
The evidence in [7] is for ramipril. In ALLHAT, which compared drugs with each other, lisinopril did not differ from the diuretic chlorthalidone on fatal coronary disease or nonfatal heart attack (RR 0.99), and had higher 6-year rates of stroke (6.3% against 5.6%) and heart failure (8.7% against 7.7%) [10].
Antiplatelets: clopidogrel and low-dose aspirinModerate

Clopidogrel compared with aspirin

Shown to lower stroke, heart attack or vascular death slightly more than aspirin, 5.32% against 5.83% a year, in people with artery disease.

The numbers and sources
Effect
In people with a recent stroke, recent heart attack or peripheral artery disease, the yearly risk of ischemic stroke, heart attack or vascular death was 5.32% on clopidogrel against 5.83% on aspirin, a relative reduction of 8.7% (95% CI 0.3 to 16.5) [3]. A collaborative meta-analysis put the reduction in serious vascular events with clopidogrel against aspirin at 10% [1].
Population
19,185 patients followed a mean 1.91 years [3].
Sources
[3], [1]
Caveats
One large trial, with a confidence interval that nearly reached no difference [3].
Antiplatelets: clopidogrel and low-dose aspirinModerate

Clopidogrel added to aspirin during a heart attack or unstable angina

Shown to cut cardiovascular death, heart attack or stroke to 9.3% against 11.4% on aspirin alone in acute coronary syndromes, with more major bleeding.

The numbers and sources
Effect
In acute coronary syndromes without ST-segment elevation, cardiovascular death, nonfatal heart attack or stroke occurred in 9.3% on clopidogrel plus aspirin against 11.4% on aspirin alone (RR 0.80, 95% CI 0.72 to 0.90). Major bleeding occurred in 3.7% against 2.7% (RR 1.38) [4]. In acute heart attack, death, reinfarction or stroke occurred in 9.2% against 10.1% over about two weeks, a 9% reduction (95% CI 3 to 14), and death from any cause in 7.5% against 8.1% [5].
Population
12,562 patients treated for 3 to 12 months [4]; 45,852 patients with acute myocardial infarction [5].
Sources
[4], [5]
Caveats
In the second trial, fatal, transfused or cerebral bleeds did not differ (0.58% against 0.55%) over a short treatment period [5].
Antiplatelets: clopidogrel and low-dose aspirinModerate

After a coronary stent

Shown to cut stent clots to 0.4% against 1.4% and heart attacks to 2.1% against 4.1% when a thienopyridine was continued with aspirin beyond a year, with more bleeding.

The numbers and sources
Effect
After 12 months of aspirin plus clopidogrel or prasugrel following a drug-eluting stent, continuing the second drug for 18 more months lowered stent thrombosis (0.4% against 1.4%, HR 0.29, 95% CI 0.17 to 0.48), major cardiovascular and cerebrovascular events (4.3% against 5.9%) and heart attack (2.1% against 4.1%). Moderate or severe bleeding rose (2.5% against 1.6%), and death from any cause was 2.0% against 1.5% (HR 1.36, 95% CI 1.00 to 1.85) [6]. Across 10 trials, treatment beyond 12 months lowered heart attack (OR 0.53) and stent thrombosis (OR 0.33) and raised major bleeding (OR 1.62) and death from any cause (OR 1.30, 95% CI 1.02 to 1.66). Courses shorter than 12 months lowered major bleeding (OR 0.58) with no significant difference in ischemic outcomes [7].
Population
9,961 patients after a drug-eluting stent [6]; 32,287 patients in 10 trials [7].
Sources
[6], [7]
Caveats
The large trial tested clopidogrel or prasugrel, not clopidogrel alone, and both groups kept taking aspirin [6]. No trial here compared a stent with no antiplatelet drug.
Antiplatelets: clopidogrel and low-dose aspirinModerate

Clopidogrel with aspirin after a minor stroke or TIA

Shown to cut stroke in the next 90 days to 8.2% against 11.7% on aspirin alone after a minor stroke or TIA in one of two trials.

The numbers and sources
Effect
Started within 24 hours of a minor stroke or high-risk transient ischemic attack, clopidogrel plus aspirin lowered stroke at 90 days to 8.2% against 11.7% on aspirin alone (HR 0.68, 95% CI 0.57 to 0.81), with moderate or severe hemorrhage in 0.3% of each group [8]. In a second trial, major ischemic events occurred in 5.0% against 6.5% (HR 0.75, 95% CI 0.59 to 0.95), and major hemorrhage in 0.9% against 0.4% (HR 2.32, 95% CI 1.10 to 4.87) [9].
Population
5,170 patients in China [8]; 4,881 patients at 269 international sites [9].
Sources
[8], [9]
Caveats
Both trials tested a short course begun within a day of the event. The second was stopped early for both benefit and bleeding [9].
Antiplatelets: clopidogrel and low-dose aspirinModerate

Aspirin in the first weeks after a stroke or TIA

Shown to cut repeat ischemic stroke in the first 6 weeks by about 60%, with no further benefit seen after 12 weeks.

The numbers and sources
Effect
Across 12 trials, aspirin lowered the 6-week risk of recurrent ischemic stroke (84 of 8,452 against 175 of 7,326, HR 0.42, 95% CI 0.32 to 0.55) and of disabling or fatal ischemic stroke (HR 0.29, 95% CI 0.20 to 0.42). The effects did not depend on dose. After 12 weeks, stroke risk did not differ from control (OR 0.97, 95% CI 0.84 to 1.12) [10].
Population
15,778 participants in 12 trials of aspirin against control [10].
Sources
[10]
Caveats
A pooled time-course analysis of older trials.
Antiplatelets: clopidogrel and low-dose aspirinModerate

Stroke prevention in atrial fibrillation

Shown to cut stroke by about 22% against no treatment in atrial fibrillation, less than warfarin or apixaban did in direct comparisons.

The numbers and sources
Effect
Across eight trials, antiplatelet agents lowered stroke by 22% (95% CI 6 to 35) against control in nonvalvular atrial fibrillation. Adjusted-dose warfarin lowered stroke 39% more than antiplatelet therapy (95% CI 22 to 52) [17]. In a trial of people unsuited to warfarin, stroke or systemic embolism occurred at 3.7% a year on aspirin against 1.6% on apixaban (HR 0.45 for apixaban, 95% CI 0.32 to 0.62), with major bleeding at 1.2% against 1.4% a year [18].
Population
4,876 participants in eight antiplatelet trials and 12,963 in 12 trials against warfarin [17]; 5,599 patients [18].
Sources
[17], [18]
Caveats
Neither drug's label read here lists atrial fibrillation as an approved use.
ARBs (angiotensin receptor blockers)Moderate

Heart attack, stroke and death

Shown to lower stroke by about 10% and heart failure by 13% across 37 trials, with no reduction in heart attacks or deaths.

The numbers and sources
Effect
Across 37 randomized trials against placebo or other drugs, ARBs lowered stroke (RR 0.90, 95% CI 0.84 to 0.98), heart failure (RR 0.87, 95% CI 0.81 to 0.93) and new diabetes (RR 0.85, 95% CI 0.78 to 0.93). They did not change the risk of heart attack (RR 0.99, 95% CI 0.92 to 1.07), death or cardiovascular death [6]. In people who could not take ACE inhibitors, telmisartan did not significantly lower the composite of cardiovascular death, heart attack, stroke or heart failure hospitalization (15.7% against 17.0% on placebo, HR 0.92, 95% CI 0.81 to 1.05) [7]. Across hypertension trials, ARBs were not associated with lower all-cause mortality (HR 0.99, 95% CI 0.94 to 1.04) [8].
Population
147,020 participants in 37 trials with at least 1 year of follow-up [6]; 5,926 high-risk people intolerant to ACE inhibitors, median 56 months [7]; 158,998 people in 20 trials of renin-angiotensin drugs [8].
Sources
[6], [7], [8]
Caveats
The pooled trials mixed placebo and active comparators [6]. The evidence in [7] is for telmisartan.
ARBs (angiotensin receptor blockers)Moderate

Aortic growth in Marfan syndrome

Shown to about halve the yearly growth of the aortic root in Marfan syndrome against control. Losartan did not differ from atenolol.

The numbers and sources
Effect
In individual data from 4 trials, the annual increase in aortic root Z score was 0.07 on an ARB and 0.13 on control (difference -0.07, 95% CI -0.12 to -0.01) over a median 3 years [19]. In a head-to-head trial, the rate of change in aortic root Z score did not differ significantly between losartan and atenolol (-0.107 against -0.139 standard deviation units a year, P = 0.08) [20].
Population
676 people with Marfan syndrome and no prior aortic surgery in 4 trials of ARB against control, within a 7-trial analysis of 1,442 people [19]; 608 people aged 6 months to 25 years with an aortic root Z score above 3.0 [20].
Sources
[19], [20]
Caveats
The outcome is aortic size, not dissection or death. Not an FDA-approved use [10].
Calcium channel blockersModerate

Stroke and heart events against placebo

Shown to lower stroke from 3.4% to 1.9% and total cardiovascular events from 8.0% to 5.7% against placebo, with no significant change in deaths.

The numbers and sources
Effect
A Cochrane review found first-line calcium channel blockers lowered stroke (3.4% with control against 1.9%, RR 0.58, 95% CI 0.41 to 0.84) and total cardiovascular events (8.0% against 5.7%, RR 0.71, 95% CI 0.57 to 0.87). Coronary heart disease (RR 0.77, 95% CI 0.55 to 1.09) and mortality (RR 0.86, 95% CI 0.68 to 1.09) did not differ significantly [3]. In the Syst-Eur trial of nitrendipine-based treatment, stroke fell from 13.7 to 7.9 events per 1,000 patient-years (42% reduction) and all cardiovascular endpoints fell 31%, with no significant change in all-cause mortality [2].
Population
Adults with moderate to severe primary hypertension, mean age 56, within a review of 24 trials and 58,040 people across drug classes [3]; 4,695 adults over 60 with isolated systolic hypertension, median 2 years [2].
Sources
[3], [2]
Caveats
The Cochrane authors rated this evidence low quality [3]. The trial evidence is for a dihydropyridine [2].
Calcium channel blockersModerate

Angina and coronary artery disease

Shown to lower hospital stays for angina and artery procedures in coronary disease, with no change in deaths. The trials tested amlodipine and nifedipine.

The numbers and sources
Effect
In CAMELOT, a first cardiovascular event occurred in 16.6% on amlodipine and 23.1% on placebo over a mean 19 months (HR 0.691, 95% CI 0.540 to 0.884), mostly through fewer hospitalizations for angina and fewer revascularization procedures [8]. In ACTION, long-acting nifedipine did not change the main composite of death, heart attack, refractory angina, heart failure, stroke and peripheral revascularization (4.60 against 4.75 per 100 patient-years, HR 0.97, 95% CI 0.88 to 1.07) or deaths (HR 1.07, 95% CI 0.91 to 1.25). Death or any cardiovascular event or procedure was less frequent (9.32 against 10.50 per 100 patient-years, HR 0.89), mainly from less coronary angiography and intervention [9].
Population
1,318 people with coronary artery disease on angiography assigned to amlodipine 5 to 10 mg or placebo [8]; 7,665 people with treated stable angina, mean 4.9 years [9].
Sources
[8], [9]
Caveats
Both drugs are dihydropyridines. The diltiazem label also lists angina as an approved use, and its trial data were not sourced for this block [10].
Ezetimibe and fenofibrateModerate

Ezetimibe added to a statin: heart attacks and strokes

Shown to lower heart attacks, strokes and related events about 6% when added to a statin, 267 against 284 per 1,000, mostly in people with recent heart disease.

Compared with a statin alone or with placebo [3][4]

The numbers and sources
Effect
Adding ezetimibe to a statin lowered major cardiovascular events (RR 0.94, 95% CI 0.90 to 0.98), from 284 to 267 per 1,000, nonfatal heart attack (RR 0.88, 105 to 92 per 1,000) and nonfatal stroke (RR 0.83, 32 to 27 per 1,000) [3]. In the largest trial, the 7-year event rate was 32.7% on simvastatin with ezetimibe and 34.7% on simvastatin alone (HR 0.936, 95% CI 0.89 to 0.99), an absolute difference of 2.0 points [4].
Population
21,727 people in 10 trials, mostly with established heart disease [3]; 18,144 patients within 10 days of an acute coronary syndrome, median follow-up 6 years [4].
Sources
[3], [4]
Caveats
The pooled result is driven by one trial, which carried 41.5% to 98.4% of the weight [3]. That trial was funded by the maker [4]. Evidence for people without heart disease and for ezetimibe alone is limited [3].
SpironolactoneModerate

Heart failure with reduced pumping function

Shown to lower deaths in severe heart failure, 35% against 46% on placebo over 2 years, and heart failure hospital stays by 35%.

The numbers and sources
Effect
In the RALES trial, 35% died on spironolactone and 46% on placebo over a mean 24 months (RR 0.70, 95% CI 0.60 to 0.82), and hospitalization for worsening heart failure was 35% less frequent (RR 0.65, 95% CI 0.54 to 0.77) [1]. Individual data from four trials of this drug class found a lower risk of cardiovascular death or heart failure hospitalization (HR 0.77, 95% CI 0.72 to 0.83), with a larger effect in the two trials in reduced ejection fraction (HR 0.66, 95% CI 0.59 to 0.73), where all-cause death also fell (HR 0.73, 95% CI 0.65 to 0.83) [2].
Population
1,663 people with severe heart failure and ejection fraction of 35% or less, on an ACE inhibitor and a loop diuretic [1]; 13,846 people in four trials, one of spironolactone and one of eplerenone in reduced ejection fraction [2].
Sources
[1], [2]
Caveats
The trial was stopped early for benefit, at a mean dose of 26 mg a day [3]. The pooled analysis combines spironolactone with eplerenone and finerenone [2].
Thiazide diureticsModerate

Heart attack, stroke and death

Shown to lower stroke from 6.2% to 4.2%, coronary heart disease from 3.9% to 2.8% and deaths from 11.0% to 9.8% as first-line low-dose treatment.

The numbers and sources
Effect
A Cochrane review found first-line low-dose thiazides lowered mortality (11.0% with control against 9.8%, RR 0.89, 95% CI 0.82 to 0.97), total cardiovascular events (12.9% against 9.0%, RR 0.70), stroke (6.2% against 4.2%, RR 0.68, 95% CI 0.60 to 0.77) and coronary heart disease (3.9% against 2.8%, RR 0.72). High-dose thiazides lowered stroke (RR 0.47) and total cardiovascular events (RR 0.72) and did not lower mortality [4]. In the SHEP trial of chlorthalidone-based treatment, the 5-year incidence of stroke was 5.2 per 100 on treatment and 8.2 on placebo (RR 0.64), with major cardiovascular events also lower (RR 0.68) [5]. In ALLHAT, chlorthalidone did not differ from amlodipine or lisinopril on fatal coronary disease or nonfatal heart attack, and had a lower 6-year rate of heart failure than either (7.7% against 10.2% and 8.7%) [6].
Population
58,040 people in 24 trials across drug classes, mean age 56, followed 3 to 5 years [4]; 4,736 adults aged 60 and over with isolated systolic hypertension, mean age 72 [5]; 33,357 adults aged 55 or older in ALLHAT [6].
Sources
[4], [5], [6]
Caveats
The Cochrane authors rated the low-dose evidence high quality and the high-dose evidence low to moderate [4]. The two named trials used chlorthalidone.
TirzepatideModerate

Heart failure with preserved ejection fraction

Shown to cut cardiovascular death or worsening heart failure, 9.9% against 15.3% on placebo, in adults with obesity and heart failure with preserved ejection fraction.

The numbers and sources
Effect
Cardiovascular death or worsening heart failure occurred in 9.9% on tirzepatide and 15.3% on placebo (HR 0.62, 95% CI 0.41 to 0.95). Worsening heart failure events were 8.0% and 14.2% (HR 0.54, 95% CI 0.34 to 0.85). Cardiovascular deaths were 8 (2.2%) and 5 (1.4%) (HR 1.58, 95% CI 0.52 to 4.83). The KCCQ health status score rose 6.9 points more on tirzepatide (95% CI 3.3 to 10.6) [9]. Pooled with observational studies, tirzepatide was associated with fewer cardiovascular deaths or worsening heart failure events than standard therapy (HR 0.50, 95% CI 0.42 to 0.60). Heart failure hospitalization and all-cause mortality did not differ, and no outcome differed significantly against semaglutide [10].
Population
731 adults with heart failure, ejection fraction of 50% or more and BMI of 30 or more, median follow-up 104 weeks [9]; 47,710 patients with this type of heart failure and BMI of 30 or more in 5 studies [10].
Sources
[9], [10]
Caveats
SUMMIT had few deaths, so its death estimate is imprecise, and it was funded by Eli Lilly [9]. The meta-analysis mixed randomized and observational studies [10].
WarfarinModerate

Mechanical heart valves

Linked to about 1 major embolism per 100 patient-years with coumarin treatment in people with mechanical valves, against 4 without antithrombotic treatment.

The numbers and sources
Effect
Across 46 studies, the incidence of major embolism was 4 per 100 patient-years without antithrombotic treatment, 2.2 with antiplatelet treatment and 1 with coumarin treatment. Major bleeding on coumarins occurred at 1.4 per 100 patient-years [9]. In a randomized trial, stroke occurred in 5% on dabigatran and in none on warfarin [10]. The label gives a target INR of 2.5 or 3.0 by valve type and position [11].
Population
13,088 patients studied for 53,647 patient-years in 46 studies [9]; 252 patients with mechanical heart valves [10].
Sources
[9], [10], [11]
Caveats
The meta-analysis pooled mostly observational studies from 1970 to 1992, and embolism risk varied with valve type and position [9].
WarfarinModerate

Self-testing and self-managing the INR

Shown to cut clotting events by about 42% when people tested or managed their own INR, with no change in major bleeding.

The numbers and sources
Effect
Across 18 trials, self-monitoring or self-management lowered thromboembolic events (RR 0.58, 95% CI 0.45 to 0.75). Self-management lowered death from any cause (RR 0.55, 95% CI 0.36 to 0.84) and self-monitoring did not (RR 0.94, 95% CI 0.78 to 1.15). Major hemorrhage did not change (RR 0.95, 95% CI 0.80 to 1.12) [15].
Population
8,950 participants in 28 randomized trials [15].
Sources
[15]
Caveats
Evidence quality was rated low to moderate. Trial participants were people able and willing to test themselves.
AllopurinolLimited

Exercise time in stable angina

One crossover trial found high-dose allopurinol lengthened exercise time before ECG changes by 43 seconds against placebo in people with stable angina.

The numbers and sources
Effect
On allopurinol 600 mg a day for 6 weeks, median time to ST depression on a treadmill test rose from 232 to 298 seconds, against 249 seconds on placebo, a difference of 43 seconds (95% CI 31 to 58). Total exercise time was 58 seconds longer (95% CI 45 to 77) and time to chest pain 38 seconds longer (95% CI 17 to 55) [13].
Population
65 adults with coronary artery disease and stable angina, 60 analyzed in both periods [13].
Sources
[13]
Caveats
One small trial of 6 weeks. A later trial of 5,721 people with ischemic heart disease found no fewer cardiovascular events [11].
ARBs (angiotensin receptor blockers)Limited

Stroke in high blood pressure with an enlarged left ventricle

One trial found losartan lowered stroke by 25% against atenolol in adults with high blood pressure and an enlarged left ventricle, at similar blood pressure.

The numbers and sources
Effect
Over at least 4 years, the composite of death, heart attack or stroke occurred at 23.8 per 1,000 patient-years on losartan and 27.9 on atenolol (RR 0.87, 95% CI 0.77 to 0.98). Stroke occurred in 232 people on losartan and 309 on atenolol (RR 0.75, 95% CI 0.63 to 0.89). Heart attack did not differ (RR 1.07, 95% CI 0.88 to 1.31). Blood pressure fell by 30.2/16.6 mm Hg on losartan and 29.1/16.8 mm Hg on atenolol [9].
Population
9,193 adults aged 55 to 80 with high blood pressure and left ventricular hypertrophy on ECG [9].
Sources
[9]
Caveats
The comparison was atenolol, not placebo. The losartan label states there is evidence this benefit does not apply to Black patients [10].
Why limited
This rests on one trial, of 9,193 people. Strength counts independent studies that agree, so one trial is limited at any size.
Beta-blockersLimited

Heart rate control in atrial fibrillation

One small crossover trial found metoprolol and carvedilol lowered 24-hour heart rate in permanent atrial fibrillation, from 96 beats a minute to 82 and 84.

The numbers and sources
Effect
Mean 24-hour heart rate was 96 beats a minute untreated, 82 on metoprolol 100 mg a day and 84 on carvedilol 25 mg a day, against 75 on diltiazem 360 mg and 81 on verapamil 240 mg [14].
Population
60 adults with permanent atrial fibrillation, mean age 71, given each drug for 3 weeks in random order [14].
Sources
[14]
Caveats
One small investigator-blind trial of 3-week periods. In people with heart failure and atrial fibrillation, beta-blockers did not improve prognosis in pooled trial data [2].
Calcium channel blockersLimited

Heart rate control in atrial fibrillation

One small crossover trial found diltiazem and verapamil lowered 24-hour heart rate in permanent atrial fibrillation, from 96 beats a minute to 75 and 81.

The numbers and sources
Effect
Mean 24-hour heart rate was 96 beats a minute untreated, 75 on diltiazem 360 mg a day, 81 on verapamil 240 mg a day, 82 on metoprolol 100 mg a day and 84 on carvedilol 25 mg a day. All four drugs lowered heart rate against baseline [11].
Population
60 adults with permanent atrial fibrillation, mean age 71, given each drug for 3 weeks in random order [11].
Sources
[11]
Caveats
One small investigator-blind trial of 3-week periods. The evidence is for the non-dihydropyridines only.
Direct oral anticoagulants (DOACs)Limited

Stable coronary or peripheral artery disease, low-dose rivaroxaban with aspirin

One trial found cardiovascular death, stroke or heart attack in 4.1% on low-dose rivaroxaban plus aspirin against 5.4% on aspirin alone, with more major bleeding.

The numbers and sources
Effect
The primary outcome occurred in 4.1% on rivaroxaban 2.5 mg twice daily plus aspirin against 5.4% on aspirin alone (HR 0.76, 95% CI 0.66 to 0.86). Major bleeding occurred in 3.1% against 1.9% (HR 1.70, 95% CI 1.40 to 2.05), with no significant difference in intracranial or fatal bleeding. Deaths occurred in 3.4% against 4.1% (HR 0.82, 95% CI 0.71 to 0.96). Rivaroxaban 5 mg twice daily alone did not improve outcomes against aspirin and caused more major bleeding [13].
Population
27,395 participants with stable atherosclerotic vascular disease [13].
Sources
[13]
Caveats
The evidence is for rivaroxaban. The mortality result did not meet the trial's threshold for significance. Funded by the manufacturer [13].
Why limited
This rests on one trial, of 27,395 people. Strength counts independent studies that agree, so one trial is limited at any size.
Direct oral anticoagulants (DOACs)Limited

Atrial fibrillation when warfarin is unsuitable, against aspirin

One trial found stroke or systemic embolism at 1.6% a year on apixaban against 3.7% on aspirin, with no significant rise in major bleeding.

The numbers and sources
Effect
Stroke or systemic embolism occurred at 1.6% a year on apixaban and 3.7% a year on aspirin (HR 0.45, 95% CI 0.32 to 0.62). Major bleeding occurred at 1.4% a year against 1.2% (HR 1.13, 95% CI 0.74 to 1.75), with 11 and 13 intracranial bleeds. Deaths occurred at 3.5% a year against 4.4% (HR 0.79, 95% CI 0.62 to 1.02) [14].
Population
5,599 patients with atrial fibrillation for whom vitamin K antagonist therapy was unsuitable [14].
Sources
[14]
Caveats
The evidence is for apixaban. The trial was stopped early for benefit [14].
Why limited
This rests on one trial, of 5,599 people. Strength counts independent studies that agree, so one trial is limited at any size.
DulaglutideLimited

Heart attack, stroke and cardiovascular death

One large trial found major cardiovascular events in 12.0% on dulaglutide against 13.4% on placebo over 5.4 years in adults with type 2 diabetes.

Compared with placebo [6]; other glucose-lowering drugs in a network meta-analysis [2]

The numbers and sources
Effect
Non-fatal heart attack, non-fatal stroke or cardiovascular death occurred in 12.0% on dulaglutide and 13.4% on placebo, 2.4 against 2.7 per 100 person-years (HR 0.88, 95% CI 0.79 to 0.99). Death from any cause did not differ (HR 0.90, 95% CI 0.80 to 1.01) [6]. In a network meta-analysis of people at increased cardiovascular risk on metformin-based treatment, the odds of stroke were lower with dulaglutide [2].
Population
9,901 adults aged 50 or older with type 2 diabetes and previous cardiovascular disease or risk factors, median HbA1c 7.2%, 46.3% women [6]; 21 trials in people at increased cardiovascular risk [2].
Sources
[6], [2]
Caveats
One trial, funded by the manufacturer. The absolute difference was 1.4 percentage points over a median 5.4 years [6].
Why limited
This rests on one trial, of 9,901 people. Strength counts independent studies that agree, so one trial is limited at any size.
Ezetimibe and fenofibrateLimited

Ezetimibe alone in adults 75 and older

One open-label trial found ezetimibe lowered heart events about a third against usual care in Japanese adults 75 or older without heart disease, with no change in deaths or strokes.

Compared with usual care with dietary counseling [5]

The numbers and sources
Effect
The composite of sudden cardiac death, heart attack, coronary revascularization or stroke was lower on ezetimibe 10 mg (HR 0.66, 95% CI 0.50 to 0.86), with fewer cardiac events (HR 0.60) and revascularizations (HR 0.38). Stroke, all-cause mortality and adverse events did not differ [5].
Population
3,796 adults aged 75 or older in Japan with raised LDL and no coronary disease, 3,411 analyzed, median follow-up 4.1 years [5].
Sources
[5]
Caveats
Open-label, stopped early, with follow-up problems and 10% excluded after randomization, which the authors say call for caution [5].
Why limited
This rests on one trial, of 3,796 people. Strength counts independent studies that agree, so one trial is limited at any size.
Ezetimibe and fenofibrateLimited

Ezetimibe with a moderate statin against a high-dose statin

One trial found a moderate statin with ezetimibe matched a high-intensity statin on heart events over 3 years, 9.1% against 9.9%, with fewer people stopping for intolerance.

Compared with high-intensity statin alone, rosuvastatin 20 mg [6]

The numbers and sources
Effect
The 3-year composite of cardiovascular death, major cardiovascular events or nonfatal stroke was 9.1% on rosuvastatin 10 mg with ezetimibe and 9.9% on rosuvastatin 20 mg, meeting the trial's noninferiority margin. LDL below 70 mg/dL was reached by 72% to 75% against 55% to 60%, and drug stopping or dose cuts for intolerance were 4.8% against 8.2% [6].
Population
3,780 patients with atherosclerotic heart disease at 26 centers in South Korea [6].
Sources
[6]
Caveats
Open-label, a noninferiority design, and funded by a manufacturer [6].
Why limited
This rests on one trial, of 3,780 people. Strength counts independent studies that agree, so one trial is limited at any size.
InsulinLimited

Heart disease in type 1 diabetes

One trial found 42% fewer cardiovascular events over a mean 17 years in people with type 1 diabetes first assigned to intensive insulin treatment.

The numbers and sources
Effect
Over a mean 17 years of follow-up, 46 cardiovascular events occurred in 31 patients first assigned to intensive treatment against 98 events in 52 patients assigned to conventional treatment. Intensive treatment lowered the risk of any cardiovascular event by 42% (95% CI 9 to 63) and of nonfatal heart attack, stroke or cardiovascular death by 57% (95% CI 12 to 79) [16].
Population
1,441 patients with type 1 diabetes treated for a mean 6.5 years, 93% of them then followed observationally [16].
Sources
[16]
Caveats
Events were few. After the randomized phase ended, follow-up was observational. A Cochrane review found too few heart attacks and strokes across trials to draw a conclusion [2].
Why limited
This rests on one trial, of 1,441 people. Strength counts independent studies that agree, so one trial is limited at any size.
LiraglutideLimited

Heart attack, stroke and cardiovascular death in type 2 diabetes

One trial found fewer major cardiovascular events, 13.0% against 14.9% on placebo, in adults with type 2 diabetes at high cardiovascular risk.

The numbers and sources
Effect
A first major cardiovascular event occurred in 13.0% on liraglutide and 14.9% on placebo (HR 0.87, 95% CI 0.78 to 0.97). Cardiovascular death was 4.7% against 6.0% (HR 0.78), and death from any cause 8.2% against 9.6% (HR 0.85). Nonfatal heart attack, nonfatal stroke and heart failure hospitalization were lower but not significantly [9].
Population
9,340 adults with type 2 diabetes at high cardiovascular risk, median follow-up 3.8 years [9].
Sources
[9]
Caveats
One trial, in people with type 2 diabetes. It does not describe people without diabetes. Funded by Novo Nordisk and the National Institutes of Health [9]. A re-analysis estimating the effect of staying on treatment found a 3.5-year risk difference of 1.1% (95% CI -0.4 to 2.6), against 1.5% (95% CI 0.1 to 2.9) by intention to treat [10].
Why limited
This rests on one trial, of 9,340 people. Strength counts independent studies that agree, so one trial is limited at any size.
LiraglutideLimited

Repeat stroke after a minor stroke in type 2 diabetes

One open-label trial found fewer repeat strokes within 90 days, 7.9% against 13.8%, in adults with type 2 diabetes after a minor stroke.

The numbers and sources
Effect
Within 90 days, stroke recurred in 7.9% on liraglutide added to standard care and 13.8% on standard care alone (HR 0.56, 95% CI 0.34 to 0.91). An excellent functional outcome was reached by 87.3% against 77.8% (OR 1.95). Symptomatic brain bleeding and death were low and similar between groups [11].
Population
636 adults with type 2 diabetes randomized within 24 hours of a minor ischemic stroke or high-risk transient ischemic attack, at 27 hospitals in China [11].
Sources
[11]
Caveats
One open-label trial over 90 days, which the authors describe as underpowered [11].
Loop diureticsLimited

Swelling and fluid in heart failure

Limited evidence: label trials found torsemide cut weight and swelling more than placebo in heart failure, and higher furosemide doses removed more fluid in acute heart failure.

Compared with placebo [1]; a lower furosemide dose [2]

The numbers and sources
Effect
In controlled trials in NYHA class II to IV heart failure, torsemide 10 to 20 mg a day produced significantly greater reductions in weight and swelling than placebo [1]. In acute heart failure, furosemide at 2.5 times the previous oral dose gave greater fluid removal than the previous dose itself, with a trend toward better symptom scores (P=0.06) and some transient worsening of kidney function [2].
Population
Patients with heart failure in label trials [1]; 308 patients hospitalized with acute decompensated heart failure [2].
Sources
[1], [2]
Caveats
The label gives no sizes for the placebo comparison [1]. The placebo-controlled trials behind the oldest uses are decades old, and no recent pooled placebo-controlled analysis was found.
SemaglutideLimited

Heart attack, stroke and cardiovascular death

One large trial found major cardiovascular events in 6.5% on semaglutide against 8.0% on placebo in adults with heart disease and overweight, without diabetes.

The numbers and sources
Effect
A major cardiovascular event occurred in 6.5% on semaglutide and 8.0% on placebo (HR 0.80, 95% CI 0.72 to 0.90) [5].
Population
17,604 adults aged 45 or older with existing cardiovascular disease and a BMI of 27 or more, without diabetes, followed a mean of 39.8 months [5].
Sources
[5]
Caveats
One trial, though large. It enrolled people with heart disease, so it does not describe people without it. Funded by Novo Nordisk [5].
Why limited
This rests on one trial, of 17,604 people. Strength counts independent studies that agree, so one trial is limited at any size.
WarfarinLimited

After a heart attack

One trial found death, repeat heart attack or stroke in 16.7% on warfarin against 20.0% on aspirin after a heart attack, with more major bleeding.

The numbers and sources
Effect
Death, nonfatal reinfarction or thromboembolic stroke occurred in 16.7% on warfarin against 20.0% on aspirin (rate ratio 0.81, 95% CI 0.69 to 0.95), and in 15.0% on warfarin plus aspirin (rate ratio 0.71, 95% CI 0.60 to 0.83). Major nonfatal bleeding occurred at 0.62% per treatment-year in both warfarin groups against 0.17% on aspirin [12].
Population
3,630 patients after acute myocardial infarction. Warfarin alone was dosed to an INR of 2.8 to 4.2 [12].
Sources
[12]
Caveats
One trial, published in 2002, at a higher INR target than the label gives for most uses.
Why limited
This rests on one trial, of 3,630 people. Strength counts independent studies that agree, so one trial is limited at any size.
Antiplatelets: clopidogrel and low-dose aspirinStrongUncommon

Aspirin for people without cardiovascular disease

Mixed results: aspirin cut first cardiovascular events by about 11%, 1 in 265 people, while raising major bleeding by 1 in 210.

The numbers and sources
How common
Uncommon: between 1 in 1,000 and 1 in 100 people.
Effect
Across 13 trials, aspirin lowered the composite cardiovascular outcome (HR 0.89, 95% credible interval 0.84 to 0.95, absolute reduction 0.38%, number needed to treat 265) and raised major bleeding (HR 1.43, 95% credible interval 1.30 to 1.56, absolute increase 0.47%, number needed to harm 210) [11]. An earlier individual-data analysis found serious vascular events at 0.51% a year on aspirin against 0.57% on control, with no difference in vascular death (0.19% against 0.19%) [2]. Of the three 2018 trials, one in adults aged 70 or over found no significant drop in cardiovascular disease (HR 0.95, 95% CI 0.83 to 1.08) [12], one in moderate-risk adults found events in 4.29% against 4.48% (HR 0.96, 95% CI 0.81 to 1.13) [13], and one in diabetes found serious vascular events in 8.5% against 9.6% (rate ratio 0.88, 95% CI 0.79 to 0.97) with major bleeding in 4.1% against 3.2% [14]. In 2022 the US Preventive Services Task Force concluded that the net benefit in adults aged 40 to 59 with a 10-year cardiovascular risk of 10% or more is small, and advised against starting low-dose aspirin for primary prevention at age 60 or older [15].
Population
164,225 participants in 13 trials, median age 62 [11]; primary prevention trials with individual participant data [2]; 19,114 adults aged 70 or over [12]; 12,546 moderate-risk adults [13]; 15,480 adults with diabetes [14].
Sources
[11], [2], [12], [13], [14], [15]
Caveats
The band above is the number needed to treat of 265. The trial in diabetes concluded that the absolute benefits were largely counterbalanced by bleeding [14]. Only one guideline body's statement was read for this page.
ARBs (angiotensin receptor blockers)Moderate

Heart failure

Mixed results: candesartan and valsartan trials found fewer heart failure hospital stays, and a Cochrane review found no significant drop in deaths or total hospital stays.

The numbers and sources
Effect
In people with reduced ejection fraction who could not take ACE inhibitors, cardiovascular death or heart failure admission occurred in 33% on candesartan and 40% on placebo (HR 0.77, 95% CI 0.67 to 0.89) [11]. Valsartan added to usual therapy did not change mortality and lowered a combined endpoint of death and illness (RR 0.87, 97.5% CI 0.77 to 0.97), mainly through fewer heart failure hospitalizations (13.8% against 18.2%) [12]. A Cochrane review of 22 studies in reduced ejection fraction found ARBs did not significantly lower total mortality (RR 0.87, 95% CI 0.76 to 1.00) or total hospitalizations (RR 0.94, 95% CI 0.88 to 1.01) against placebo. In 2 studies of people with ejection fraction above 40%, neither outcome differed from placebo [13].
Population
2,028 people with ejection fraction of 40% or less, median 33.7 months [11]; 5,010 people with heart failure [12]; 17,900 people with ejection fraction of 40% or less and 7,151 with ejection fraction above 40% [13].
Sources
[11], [12], [13]
Caveats
The trial evidence is for candesartan and valsartan. About 93% of people in the valsartan heart failure trials also took an ACE inhibitor [14].
Beta-blockersModerate

After a heart attack

Mixed results: older trials found 23% lower odds of death after a heart attack, and three recent trials in people with preserved heart function disagreed.

The numbers and sources
Effect
In a review of older trials, long-term beta-blockers lowered the odds of death by 23% (95% CI 15 to 31), with 42 people treated for 2 years to avoid one death. Short-term trials found a 4% reduction that was not significant [6]. In people with ejection fraction of 50% or more, death or a new heart attack occurred in 7.9% on a beta-blocker and 8.3% without (HR 0.96, 95% CI 0.79 to 1.16) [7]. In people with ejection fraction above 40%, death, reinfarction or heart failure hospitalization occurred at 22.5 against 21.7 events per 1,000 patient-years (HR 1.04, 95% CI 0.89 to 1.22) [8]. In a third trial in people with ejection fraction of 40% or more, death or a major cardiovascular event occurred in 14.2% against 16.3% (HR 0.85, 95% CI 0.75 to 0.98), with fewer new heart attacks (5.0% against 6.7%) [9].
Population
54,234 people in older randomized trials, mostly of propranolol, timolol and metoprolol [6]; 5,020 people, median 3.5 years [7]; 8,438 people, median 3.7 years [8]; 5,574 people, median 3.5 years [9].
Sources
[6], [7], [8], [9]
Caveats
The older review was published in 1999. The three recent trials were open-label.
Beta-blockersModerate

Heart attack and stroke in high blood pressure

Shown to lower stroke by about 20% against placebo in high blood pressure, with no drop in deaths and more strokes than on other drug classes.

The numbers and sources
Effect
In 13 trials, beta-blockers did not lower all-cause mortality against placebo (RR 0.99, 95% CI 0.88 to 1.11). Total cardiovascular disease fell (RR 0.88, 95% CI 0.79 to 0.97), reflecting fewer strokes (RR 0.80, 95% CI 0.66 to 0.96), with no difference in coronary heart disease (RR 0.93, 95% CI 0.81 to 1.07). Stroke was more frequent than on calcium channel blockers (RR 1.24, 95% CI 1.11 to 1.40) or renin-angiotensin drugs (RR 1.30, 95% CI 1.11 to 1.53) [12]. For atenolol, 4 placebo-controlled trials found no difference in all-cause mortality (RR 1.01, 95% CI 0.89 to 1.15) or heart attack (RR 0.99, 95% CI 0.83 to 1.19), and 5 trials against other drugs found higher mortality on atenolol (RR 1.13, 95% CI 1.02 to 1.25) [13].
Population
13 trials, with 23,613 people in placebo comparisons and 40,245 people on beta-blockers overall, three quarters of them on atenolol [12]; 6,825 people in 4 placebo-controlled atenolol trials and 17,671 in 5 comparative trials, mean 4.6 years [13].
Sources
[12], [13]
Caveats
Most of the evidence is for atenolol. The Cochrane authors rated most trials at high risk of bias and found no outcome trials of newer vasodilating beta-blockers [12].
Calcium channel blockersModerate

Compared with other blood pressure drugs

Mixed results: fewer strokes than beta-blockers or ACE inhibitors, more heart failure than diuretics or ACE inhibitors, and no difference in deaths from any cause.

The numbers and sources
Effect
In 23 trials of first-line treatment, all-cause mortality did not differ between calcium channel blockers and any other class. Against diuretics, heart failure was more frequent (RR 1.37, 95% CI 1.25 to 1.51). Against beta-blockers, major cardiovascular events (RR 0.84, 95% CI 0.77 to 0.92) and stroke (RR 0.77, 95% CI 0.67 to 0.88) were less frequent. Against ACE inhibitors, stroke was less frequent (RR 0.90, 95% CI 0.81 to 0.99) and heart failure more frequent (RR 1.16, 95% CI 1.06 to 1.28) [4]. In ALLHAT, amlodipine did not differ from chlorthalidone on fatal coronary disease or nonfatal heart attack (RR 0.98, 95% CI 0.90 to 1.07) and had a higher 6-year rate of heart failure (10.2% against 7.7%) [5]. An amlodipine-based regimen brought fewer strokes than an atenolol-based one (HR 0.77, 95% CI 0.66 to 0.89) [6]. Benazepril plus amlodipine brought fewer cardiovascular events than benazepril plus hydrochlorothiazide (9.6% against 11.8%, HR 0.80) [7].
Population
153,849 people with hypertension in 23 trials, 18 of dihydropyridines and 4 of non-dihydropyridines [4]; 33,357 adults aged 55 or older in ALLHAT, mean 4.9 years [5]; 19,257 adults with hypertension and three other risk factors [6]; 11,506 high-risk adults with hypertension [7].
Sources
[4], [5], [6], [7]
Caveats
Most of the evidence is for dihydropyridines, and the three named trials all used amlodipine. Certainty was rated low to moderate across comparisons [4].
Direct oral anticoagulants (DOACs)Moderate

Differences between the four drugs

Linked to less major and stomach bleeding with apixaban than with rivaroxaban or dabigatran in indirect comparisons and large cohorts. No randomized trial compared the drugs directly.

The numbers and sources
Effect
In a network meta-analysis of 23 trials, major bleeding was higher with dabigatran 150 mg twice daily (OR 1.33, 95% CI 1.09 to 1.62) and rivaroxaban 20 mg (OR 1.45, 95% CI 1.19 to 1.78) than with apixaban 5 mg twice daily. Stroke or systemic embolism was higher with rivaroxaban (OR 1.35) and edoxaban 60 mg (OR 1.33) than with dabigatran 150 mg [15]. In US Medicare patients, rivaroxaban was associated with more major ischemic or hemorrhagic events than apixaban (16.1 against 13.4 per 1,000 person-years, HR 1.18, 95% CI 1.12 to 1.24) [16]. In a multinational cohort, apixaban was associated with less gastrointestinal bleeding than dabigatran (HR 0.81), edoxaban (HR 0.77) or rivaroxaban (HR 0.72, 95% CI 0.66 to 0.79), with similar rates of ischemic stroke, intracranial bleeding and death [17].
Population
94,656 patients in 23 trials [15]; 581,451 Medicare beneficiaries aged 65 or older, mean age 77.0 [16]; 527,226 new DOAC users [17].
Sources
[15], [16], [17]
Caveats
The network meta-analysis compares drugs indirectly through warfarin, and the cohorts are observational, so differences between the people prescribed each drug may remain [17].
Ezetimibe and fenofibrateModerate

Fenofibrate and fibrates: heart attacks and strokes

Mixed results: pooled fibrate trials showed about 10% fewer major heart events and 13% fewer coronary events, but fenofibrate's main trial missed its coronary goal and deaths did not fall.

Compared with placebo [8][9]; placebo or usual care [10]

The numbers and sources
Effect
In people with type 2 diabetes, coronary events were 5.2% on fenofibrate and 5.9% on placebo (HR 0.89, 95% CI 0.75 to 1.05), not significant. Total cardiovascular events fell from 13.9% to 12.5% (HR 0.89, 95% CI 0.80 to 0.99), with 24% fewer nonfatal heart attacks and a nonsignificant rise in coronary deaths (HR 1.19) [8]. Across 18 fibrate trials, major cardiovascular events fell 10% (95% CI 0 to 18), coronary events 13%, and stroke, cardiovascular death and all-cause death did not change [9]. In primary prevention, fibrates lowered cardiovascular death, heart attack or stroke (RR 0.84, 95% CI 0.74 to 0.96), with absolute reductions under 1% [10]. The fenofibrate label states it was not shown to reduce coronary heart disease illness and death in type 2 diabetes [7].
Population
9,795 adults aged 50 to 75 with type 2 diabetes, 5 years [8]; 45,058 people in 18 trials [9]; 16,135 people in 6 primary prevention trials [10].
Sources
[8], [9], [10], [7]
Caveats
[9] and [10] pool fibrates as a class, including gemfibrozil and bezafibrate, not fenofibrate alone. More placebo patients started statins in [8], which the authors say may have masked a benefit.
Thiazide diureticsModerate

Chlorthalidone compared with hydrochlorothiazide

Mixed results: indirect comparisons favored chlorthalidone, and the one large head-to-head trial found no difference in heart events between the two.

The numbers and sources
Effect
In a pragmatic trial in adults 65 and older already on hydrochlorothiazide, a major cardiovascular event or non-cancer death occurred in 10.4% switched to chlorthalidone and 10.0% kept on hydrochlorothiazide over a median 2.4 years (HR 1.04, 95% CI 0.94 to 1.16) [7]. In a cohort of first-time users, the composite cardiovascular outcome did not differ (calibrated HR 1.00, 95% CI 0.85 to 1.17) [8]. Network meta-analyses that compared the drugs indirectly across separate trials found 21% fewer cardiovascular events on chlorthalidone (95% CI 12 to 28) [9], and 12% fewer on thiazide-like than thiazide-type diuretics after adjusting for blood pressure [10]. In 9 trials that compared them directly, chlorthalidone lowered systolic pressure by 3.26 mm Hg and diastolic by 2.41 mm Hg more [11].
Population
13,523 veterans with a mean age of 72 [7]; 730,225 first-time users in three US databases [8]; 9 trials with 50,946 people in the drug-adjusted analysis [9]; randomized trials of at least 1 year in adults with hypertension [10]; 9 head-to-head trials [11].
Sources
[7], [8], [9], [10], [11]
Caveats
In the head-to-head trial, 94.5% were taking hydrochlorothiazide 25 mg at entry and chlorthalidone was given at 12.5 or 25 mg [7]. The indirect comparisons do not randomize between the two drugs. For side effect differences, see the low potassium block below.
WarfarinModerate

Time in the therapeutic INR range

Linked to fewer major bleeds and clots when more time was spent in the target INR range, in retrospective studies but not clearly in randomized trials.

The numbers and sources
Effect
Across 47 studies in atrial fibrillation, time in therapeutic range ran from 29% to 75%, and was higher in randomized trials than in retrospective studies (64.9% against 56.4%). Time in range correlated negatively with major hemorrhage (r -0.59) and thromboembolic rates (r -0.59). In retrospective studies, a 6.9% improvement in time in range lowered major hemorrhage by 1 event per 100 patient-years (95% CI 0.29 to 1.71). The correlations were not significant within randomized trials [13]. In the edoxaban trial the median time in range on warfarin was 68.4% [14]. The label states that keeping the INR in range does not eliminate the risk of bleeding [11].
Population
47 studies from 38 articles of people with atrial fibrillation on vitamin K antagonists [13]; 21,105 patients [14].
Sources
[13], [14], [11]
Caveats
Correlations across studies, not a trial of raising time in range.
InsulinStrong

Heart attack, stroke and death in type 2 diabetes

Not shown to lower heart attack, stroke, cardiovascular death or death from any cause in type 2 diabetes, in one large trial and two meta-analyses.

The numbers and sources
Effect
Over a median 6.2 years, cardiovascular outcomes occurred at 2.94 per 100 person-years on insulin glargine and 2.85 on standard care (HR 1.02, 95% CI 0.94 to 1.11) [8]. Across 20 trials, insulin did not change death from any cause against other glucose-lowering drugs (RR 0.99, 95% CI 0.92 to 1.06) or against diet or placebo (RR 0.92, 95% CI 0.80 to 1.07), and did not change cardiovascular death (RR 0.99, 95% CI 0.90 to 1.09) [7]. A network meta-analysis of 301 trials found no significant differences between any drug class in cardiovascular or all-cause mortality [6]. Between two basal insulins, major cardiovascular events occurred in 8.5% on degludec and 9.3% on glargine (HR 0.91, 95% CI 0.78 to 1.06) [9].
Population
12,537 people with cardiovascular risk factors and impaired fasting glucose, impaired glucose tolerance or type 2 diabetes, mean age 63.5 [8]; 18,599 patients in 20 trials [7]; 301 trials [6]; 7,637 patients with type 2 diabetes [9].
Sources
[8], [7], [6], [9]
Caveats
The glargine trial was funded by the manufacturer and targeted a normal fasting glucose [8].
Loop diureticsStrong

Torsemide against furosemide in heart failure

Not shown to differ in deaths between torsemide and furosemide in heart failure, 26.1% against 26.2% in the largest trial and in two pooled analyses of trials.

Compared with furosemide [3][4][5][6]

The numbers and sources
Effect
After a hospital stay for heart failure, death occurred in 26.1% on torsemide and 26.2% on furosemide over a median 17.4 months (HR 1.02, 95% CI 0.89 to 1.18), and death or hospitalization at 12 months did not differ (HR 0.92, 95% CI 0.83 to 1.02) [3]. Pooled trials found no difference in deaths (OR 0.99, 95% CI 0.97 to 1.02) or heart failure hospitalization (OR 0.96) [4], and no difference in deaths (RR 0.98, 95% CI 0.87 to 1.10) or heart failure readmission (RR 0.85, 95% CI 0.52 to 1.38) [5]. A pooled analysis that added nonrandomized studies found fewer heart failure hospitalizations on torsemide (OR 0.73, 95% CI 0.54 to 0.99) and no difference in deaths [6].
Population
2,859 adults hospitalized with heart failure at 60 US hospitals, median age 65, 36.9% women [3]; 4,011 patients in 10 trials [4]; 12 trials [5]; 11,996 patients in 17 studies [6].
Sources
[3], [4], [5], [6]
Caveats
The large trial was open-label with crossover and loss to follow-up [3]. Hospitalization findings differed between analyses that did and did not include nonrandomized studies [4][6].
SulfonylureasStrong

Heart attack, stroke and death

Not shown to lower or raise heart attack, stroke or death against other diabetes drugs, in two large trials and a meta-analysis of 47 trials.

The numbers and sources
Effect
Over a median 6.3 years, the cardiovascular composite occurred in 12.0% on glimepiride and 11.8% on linagliptin (HR 0.98, 95.47% CI 0.84 to 1.14) [5]. Added to metformin, it occurred at 1.5 per 100 person-years on sulfonylureas and 1.5 on pioglitazone (HR 0.96, 95% CI 0.74 to 1.26) [6]. Across 47 trials of at least 52 weeks, sulfonylureas were not associated with death from any cause (OR 1.12, 95% CI 0.96 to 1.30), cardiovascular death (OR 1.12, 95% CI 0.87 to 1.42), heart attack (OR 0.92, 95% CI 0.76 to 1.12) or stroke (OR 1.16, 95% CI 0.81 to 1.66) [7]. A network meta-analysis of 301 trials found no significant differences between drug classes in cardiovascular or all-cause mortality [8].
Population
6,033 adults with type 2 diabetes and raised cardiovascular risk, mean age 64.0 [5]; 3,028 patients on metformin, 11% with a previous cardiovascular event [6]; 37,650 patients in 47 trials [7]; 301 trials [8].
Sources
[5], [6], [7], [8]
Caveats
One trial was stopped early for futility [6]. A Bayesian meta-analysis reached a different result. See the cardiovascular death block under risks.
WarfarinStrong

Compared with DOACs in atrial fibrillation

Not shown to match DOACs in atrial fibrillation: stroke or systemic embolism occurred in 3.69% on warfarin against 3.01%, with about twice the brain bleeds.

The numbers and sources
Effect
Across four trials, DOACs lowered stroke or systemic embolism by 19% against warfarin (RR 0.81, 95% CI 0.73 to 0.91) and death by 10% (RR 0.90, 95% CI 0.85 to 0.95) [4]. In patient-level data, stroke or systemic embolism occurred in 3.69% on warfarin against 3.01% on a standard-dose DOAC, death in 8.42% against 7.76% and intracranial bleeding in 1.40% against 0.63% [5]. In rheumatic heart disease with atrial fibrillation, the result went the other way, with fewer cardiovascular events or deaths on a vitamin K antagonist than on rivaroxaban [6].
Population
71,683 participants in four trials, 29,272 of them on warfarin [4][5]; 4,531 patients with rheumatic heart disease and atrial fibrillation [6].
Sources
[4], [5], [6]
Caveats
The four trials were funded by DOAC manufacturers. The gap in major bleeding was larger at trial centers where warfarin control was poorer [4].
ACE inhibitorsModerate

Compared with ARBs

Not shown to differ from ARBs in deaths or cardiovascular events in people with high blood pressure or vascular disease. Fewer people stopped ARBs over side effects.

The numbers and sources
Effect
In 9 head-to-head trials lasting at least a year, total mortality (RR 0.98, 95% CI 0.88 to 1.10), total cardiovascular events (RR 1.07, 95% CI 0.96 to 1.19) and cardiovascular mortality (RR 0.98, 95% CI 0.85 to 1.13) did not differ between ACE inhibitors and ARBs. Withdrawals over adverse effects were lower on ARBs (RR 0.83, 95% CI 0.74 to 0.93), an absolute difference of 1.8% over 4.1 years, mainly from dry cough on ACE inhibitors [14]. In ONTARGET, the composite of cardiovascular death, heart attack, stroke or heart failure hospitalization occurred in 16.5% on ramipril and 16.7% on telmisartan (RR 1.01, 95% CI 0.94 to 1.09) [15].
Population
11,007 people with primary hypertension in 9 trials [14]; 17,118 people with vascular disease or high-risk diabetes assigned to ramipril or telmisartan, median 56 months [15].
Sources
[14], [15]
Caveats
The Cochrane authors rated the evidence moderate for mortality and low for cardiovascular events, citing possible publication bias [14].
AllopurinolModerate

Compared with febuxostat

Not shown to differ from febuxostat in gout flares or major cardiovascular events. One trial found fewer deaths on allopurinol.

The numbers and sources
Effect
Across 3 trials, gout attacks over 8 to 24 weeks occurred in 21% on allopurinol up to 300 mg and 23% on febuxostat 80 mg (RR 0.89, 95% CI 0.71 to 1.1) [1]. At fixed doses, 21% on allopurinol 300 mg reached the urate target against 53% and 62% on febuxostat 80 and 120 mg, with flares over weeks 9 to 52 in 64%, 64% and 70% [2]. With doses raised to target, flares in the last phase were 36.5% on allopurinol and 43.5% on febuxostat [4]. In people with gout and cardiovascular disease, the main cardiovascular outcome occurred in 10.4% on allopurinol and 10.8% on febuxostat (HR 1.03), and death from any cause was more frequent on febuxostat (HR 1.22, 95% CI 1.01 to 1.47), as was cardiovascular death (HR 1.34, 95% CI 1.03 to 1.73) [6].
Population
1,136 participants in 3 trials [1]; 760 people with gout [2]; 940 people with gout [4]; 6,190 people with gout and cardiovascular disease followed a median of 32 months [6].
Sources
[1], [2], [4], [6]
Caveats
In the cardiovascular trial 56.6% stopped the study drug and 45.0% left follow-up [6].
ARBs (angiotensin receptor blockers)Moderate

Compared with ACE inhibitors

Not shown to differ from ACE inhibitors in deaths or cardiovascular events. Fewer people stopped ARBs over side effects, mainly cough.

The numbers and sources
Effect
In 9 head-to-head trials in hypertension, total mortality (RR 0.98, 95% CI 0.88 to 1.10) and total cardiovascular events (RR 1.07, 95% CI 0.96 to 1.19) did not differ, and withdrawals over adverse effects were lower on ARBs (RR 0.83, 95% CI 0.74 to 0.93) [15]. In ONTARGET, the main cardiovascular outcome occurred in 16.7% on telmisartan and 16.5% on ramipril (RR 1.01, 95% CI 0.94 to 1.09) [16]. After a heart attack with heart failure or left ventricular dysfunction, deaths did not differ between valsartan and captopril (979 against 958, HR 1.00, 97.5% CI 0.90 to 1.11) [17]. In heart failure trials, mortality did not differ between ARBs and ACE inhibitors (RR 1.05, 95% CI 0.91 to 1.22), and withdrawals over adverse effects were lower on ARBs (RR 0.63, 95% CI 0.52 to 0.76) [13].
Population
11,007 people in 9 hypertension trials [15]; 17,118 people assigned to telmisartan or ramipril [16]; 9,818 people assigned to valsartan or captopril after a heart attack, median 24.7 months [17]; heart failure trials within a 22-study review [13].
Sources
[15], [16], [17], [13]
Caveats
The evidence in [16] is for telmisartan and in [17] for valsartan.
Direct oral anticoagulants (DOACs)Moderate

Mechanical heart valves and rheumatic heart disease

Not shown to match warfarin with mechanical heart valves or rheumatic atrial fibrillation: two trials found more strokes, clots or deaths on a DOAC.

The numbers and sources
Effect
A trial of dabigatran in people with mechanical heart valves was stopped after 252 patients. Stroke occurred in 5% on dabigatran and none on warfarin, and major bleeding in 4% against 2% [18]. In rheumatic heart disease with atrial fibrillation, 560 patients on rivaroxaban and 446 on a vitamin K antagonist had a stroke, systemic embolism, heart attack or death, and survival time free of these was 76 days shorter on rivaroxaban (95% CI 31 to 121). Major bleeding did not differ [19]. The dabigatran label lists a mechanical prosthetic heart valve as a contraindication [20].
Population
252 patients with mechanical heart valves [18]; 4,531 patients with rheumatic heart disease and atrial fibrillation, mean age 50.5 [19].
Sources
[18], [19], [20]
Caveats
One trial each for dabigatran and rivaroxaban. The pivotal atrial fibrillation trials did not enroll these groups.
Ezetimibe and fenofibrateModerate

Ezetimibe: death from any cause

Not shown to lower deaths from any cause or from heart disease when ezetimibe was added to a statin or fenofibrate.

Compared with a statin or fenofibrate alone [3]

The numbers and sources
Effect
All-cause mortality did not differ (RR 0.98, 95% CI 0.91 to 1.05), nor did cardiovascular mortality (RR 1.00, 95% CI 0.89 to 1.12) [3].
Population
21,222 people in 8 trials for all-cause and 19,457 in 6 trials for cardiovascular mortality [3].
Sources
[3]
Caveats
Trials ran a median of up to 6 years.
Ezetimibe and fenofibrateModerate

Fenofibrate added to a statin

Not shown to lower heart attacks, strokes or heart deaths when fenofibrate was added to simvastatin in type 2 diabetes, 2.2% against 2.4% a year.

Compared with simvastatin with placebo [11]; statins alone [10]

The numbers and sources
Effect
The annual rate of nonfatal heart attack, nonfatal stroke or cardiovascular death was 2.2% on fenofibrate with simvastatin and 2.4% on simvastatin with placebo (HR 0.92, 95% CI 0.79 to 1.08), and no secondary outcome differed. Annual deaths were 1.5% and 1.6%. A subgroup with high triglycerides and low HDL suggested possible benefit, and results differed by sex [11]. In the Cochrane review, the two trials of fibrates on top of statins showed no benefit [10].
Population
5,518 patients with type 2 diabetes, mean follow-up 4.7 years [11]; 2 trials [10].
Sources
[11], [10]
Caveats
Subgroup findings were prespecified but not confirmed [11].
Loop diureticsModerate

Continuous drip against injections in acute heart failure

Not shown to differ in deaths, readmissions or kidney injury between a continuous intravenous drip and repeated injections of loop diuretics in acute heart failure.

Compared with bolus injection [7][2]

The numbers and sources
Effect
All-cause mortality (RR 1.53, 95% CI 0.81 to 2.90), readmission (RR 0.85, 95% CI 0.63 to 1.16) and acute kidney injury (RR 1.02, 95% CI 0.70 to 1.49) did not differ. Weight loss was 0.86 kg greater with a drip, but not in trials at lower risk of bias (MD 0.70 kg, 95% CI -0.06 to 1.46) [7]. In a double-blind trial, symptom scores and creatinine change at 72 hours did not differ between a furosemide drip and injections every 12 hours [2].
Population
681 adults in 7 trials [7]; 308 patients [2].
Sources
[7], [2]
Caveats
Low to very low certainty evidence [7].
SGLT2 inhibitorsModerate

After a heart attack

Not shown to lower heart failure hospitalization or death after a heart attack in two trials of about 10,500 people.

The numbers and sources
Effect
In people at raised risk of heart failure after a heart attack, a first heart failure hospitalization or death from any cause occurred in 8.2% on empagliflozin and 9.1% on placebo over a median 17.9 months (HR 0.90, 95% CI 0.76 to 1.06). Heart failure hospitalization alone occurred in 3.6% against 4.7% (HR 0.77, 95% CI 0.60 to 0.98) and death in 5.2% against 5.5% [21]. In people with a heart attack and no diabetes or chronic heart failure, cardiovascular death or heart failure hospitalization occurred in 2.5% on dapagliflozin and 2.6% on placebo (HR 0.95, 95% CI 0.64 to 1.40) over about a year [22].
Population
6,522 patients after acute myocardial infarction [21]; 4,017 patients after acute myocardial infarction with impaired heart function and no diabetes [22].
Sources
[21], [22]
Caveats
Not an approved use. The dapagliflozin trial changed its primary analysis during the trial because few events occurred. Its composite of ranked outcomes favored dapagliflozin (win ratio 1.34, 95% CI 1.20 to 1.50), driven by weight and new diabetes [22].
TirzepatideModerate

New heart failure in type 2 diabetes or obesity

Not shown to change the risk of new heart failure in people with type 2 diabetes or obesity.

The numbers and sources
Effect
Against placebo or other glucose-lowering drugs, tirzepatide did not change overall heart failure risk (RR 0.63, 95% CI 0.35 to 1.13), an absolute risk reduction of 0.17% [11]. A second meta-analysis found no association with heart failure hospitalization [12].
Population
13,378 participants in 11 RCTs [11]; 29,023 participants in 22 RCTs [12].
Sources
[11], [12]
Caveats
All 11 trials in the first review were funded by Eli Lilly, and its certainty was rated moderate [11].
TirzepatideModerate

Heart attack, stroke and cardiovascular death

Not shown to lower heart attack, stroke or cardiovascular death more than dulaglutide in type 2 diabetes with heart disease. Pooled trials found about 13% fewer events.

The numbers and sources
Effect
Against the GLP-1 drug dulaglutide, a major cardiovascular event occurred in 12.2% on tirzepatide and 13.1% on dulaglutide (HR 0.92, 95.3% CI 0.83 to 1.01). This met the trial's noninferiority margin but not superiority [13]. Across RCTs in type 2 diabetes or overweight and obesity, tirzepatide was associated with fewer major cardiovascular events (OR 0.87, 95% CI 0.79 to 0.94) and lower all-cause mortality (OR 0.84, 95% CI 0.75 to 0.93), with no association for the individual components of the composite [12]. A network meta-analysis estimated tirzepatide against placebo at HR 0.79 (95% CI 0.69 to 0.91) for major cardiovascular events in type 2 diabetes [14].
Population
13,165 adults with type 2 diabetes and atherosclerotic cardiovascular disease, mean age 64.1 [13]; 29,023 participants in 22 RCTs [12]; 11 trials in type 2 diabetes with cardiovascular disease or high risk [14].
Sources
[13], [12], [14]
Caveats
SURPASS-CVOT used an active comparator, not placebo, and was funded by Eli Lilly [13]. The network estimate is indirect, and SURPASS-CVOT was its only tirzepatide trial [14].
AllopurinolLimited

Heart attack, stroke and cardiovascular death

One trial found allopurinol did not lower heart attack, stroke or cardiovascular death in adults aged 60 or older with ischemic heart disease and no gout.

The numbers and sources
Effect
Over a mean of 4.8 years, nonfatal heart attack, nonfatal stroke or cardiovascular death occurred in 11.0% on allopurinol up to 600 mg a day and 11.3% on usual care (HR 1.04, 95% CI 0.89 to 1.21). Death from any cause was 10.1% against 10.6% (HR 1.02, 95% CI 0.87 to 1.20) [11].
Population
5,721 UK adults aged 60 or older with ischemic heart disease and no history of gout, mean age 72 [11].
Sources
[11]
Caveats
Open label with no placebo, with outcomes assessed blind to treatment [11].
Why limited
This rests on one trial, of 5,721 people. Strength counts independent studies that agree, so one trial is limited at any size.
AllopurinolLimited

Heart failure

One trial found allopurinol did not improve clinical status, walking distance or quality of life over 24 weeks in heart failure with high uric acid.

The numbers and sources
Effect
At 24 weeks, clinical status had worsened in 45% on allopurinol and 46% on placebo and improved in 13% and 19%. Quality of life scores, 6-minute walk distance and ejection fraction did not differ. Uric acid was lower on allopurinol by 4.2 mg/dL at 12 weeks and 3.5 mg/dL at 24 weeks [12].
Population
253 people with symptomatic heart failure, an ejection fraction of 40% or below and uric acid of at least 9.5 mg/dL [12].
Sources
[12]
Caveats
One trial of 24 weeks, at a target dose of 600 mg a day.
Loop diureticsLimited

Stopping low-dose furosemide in stable heart failure

One trial found stopping low-dose furosemide in stable heart failure with no fluid retention did not worsen breathlessness over 90 days, and 75% stayed off it.

Compared with continuing furosemide, against placebo [8]

The numbers and sources
Effect
Breathlessness scores did not differ between withdrawal and continued furosemide (median AUC 1,875 against 1,541, P=0.94). 75.3% of the withdrawal group and 83.7% of the continuing group needed no extra furosemide (OR 1.69, 95% CI 0.82 to 3.49). Heart failure events were infrequent and similar [8].
Population
188 stable outpatients with heart failure, ejection fraction 45% or less, on furosemide 40 or 80 mg a day, 25% women [8].
Sources
[8]
Caveats
Small, 90 days, and limited to people with no signs of fluid retention.
Naltrexone-bupropionLimited

Heart attack, stroke and cardiovascular death

One trial stopped early: an early drop in major cardiovascular events did not hold at the later analysis, and cardiovascular safety remains uncertain.

The numbers and sources
Effect
At the 25% interim analysis, major cardiovascular events occurred in 1.3% on placebo and 0.8% on the combination (HR 0.59, 95% CI 0.39 to 0.90). At 50% of planned events the figures were 2.3% and 2.0% (HR 0.88, adjusted 99.7% CI 0.57 to 1.34). The trial was stopped early after the sponsor publicly released confidential interim data, and the authors state that cardiovascular safety remains uncertain [6].
Population
8,910 adults with overweight or obesity at raised cardiovascular risk, mean age 61, 85.2% with diabetes [6].
Sources
[6]
Caveats
The trial ended before it could test its planned noninferiority margin. The early interim result did not hold at the later analysis [6].
Why limited
This rests on one trial, of 8,910 people. Strength counts independent studies that agree, so one trial is limited at any size.
SpironolactoneLimited

Heart failure with preserved pumping function

One trial found no significant drop in cardiovascular death, cardiac arrest or heart failure hospital stay, 18.6% against 20.4%. Heart failure hospital stays alone fell 17%.

The numbers and sources
Effect
Over a mean 3.3 years, the composite of cardiovascular death, aborted cardiac arrest or heart failure hospitalization occurred in 18.6% on spironolactone and 20.4% on placebo (HR 0.89, 95% CI 0.77 to 1.04). Hospitalization for heart failure was less frequent (12.0% against 14.2%, HR 0.83, 95% CI 0.69 to 0.99). Total deaths and hospitalizations for any reason did not differ [4]. In the two trials of this drug class in mildly reduced or preserved ejection fraction, cardiovascular death or heart failure hospitalization was less frequent (HR 0.87, 95% CI 0.79 to 0.95) and cardiovascular death did not differ (HR 0.92, 95% CI 0.80 to 1.05) [2].
Population
3,445 people with symptomatic heart failure and ejection fraction of 45% or more [4]; the same trial pooled with a trial of finerenone [2].
Sources
[4], [2]
Caveats
Only [4] tested spironolactone in this group.
Why limited
This rests on one trial, of 3,445 people. Strength counts independent studies that agree, so one trial is limited at any size.
SpironolactoneLimited

After a heart attack

One trial found routine spironolactone after a heart attack did not lower cardiovascular death or heart failure against placebo over 3 years.

The numbers and sources
Effect
Cardiovascular death or new or worsening heart failure occurred at 1.7 events per 100 patient-years on spironolactone and 2.1 on placebo (HR 0.91, 95% CI 0.69 to 1.21). A first heart attack, stroke, heart failure event or cardiovascular death occurred in 7.9% against 8.3% (HR 0.96, 95% CI 0.81 to 1.13). Serious adverse events were reported in 7.2% against 6.8% [8].
Population
7,062 people with a heart attack treated by percutaneous coronary intervention, median 3 years [8].
Sources
[8]
Caveats
One trial, in people selected for a heart attack and not for heart failure.
Why limited
This rests on one trial, of 7,062 people. Strength counts independent studies that agree, so one trial is limited at any size.

Supplements

FolateStrong

Stroke and cardiovascular disease

Shown to lower stroke about 10% in adults, with no change in coronary heart disease. Effects were larger in people with low blood folate.

The numbers and sources
Effect
Across 30 RCTs with 82,334 participants, folic acid lowered stroke (RR 0.90, 95% CI 0.84 to 0.96) and overall cardiovascular disease (RR 0.96, 95% CI 0.92 to 0.99). Coronary heart disease did not change (RR 1.04, 95% CI 0.99 to 1.09) [3]. In the largest single trial, first stroke occurred in 2.7% on enalapril plus folic acid and 3.4% on enalapril alone (HR 0.79, 95% CI 0.68 to 0.93) [4]. A Cochrane review of B vitamin trials found the same 10% stroke reduction and no change in heart attack [5].
Population
Adults with and without cardiovascular disease [3][5]; 20,702 adults with high blood pressure in China, median 4.5 years [4].
Sources
[3], [4], [5]
Caveats
Stroke and overall cardiovascular effects were larger in people with lower blood folate at baseline and without existing cardiovascular disease [3]. The Chinese trial gave folic acid with enalapril and compared it with enalapril alone [4].
FolateStrong

Homocysteine

Shown to lower homocysteine in adults with high levels, kidney disease or type 2 diabetes. In kidney disease, cardiovascular events did not fall.

The numbers and sources
Effect
In 36 RCTs of adults with high homocysteine, folic acid alone lowered homocysteine (SMD -0.99, 95% CI -1.15 to -0.83, I² 90%), with the largest effects at 5 to 15 mg a day and after 6 months or more [12]. In 14 RCTs in chronic kidney disease, folate-based supplements lowered homocysteine by 8.09 µmol/L (95% CI -11.54 to -4.63, I² 98%), while cardiovascular events (RR 0.88, 95% CI 0.68 to 1.13), all-cause mortality and cardiovascular mortality did not change [13]. In 29 studies in type 2 diabetes, folic acid with B12 lowered homocysteine (SMD -2.77, 95% CI -3.23 to -2.30) [14].
Population
Adults with high homocysteine [12]; 8,625 adults with chronic kidney disease in 37 studies [13]; adults with type 2 diabetes [14].
Sources
[12], [13], [14]
Caveats
Heterogeneity was high in all three. The high-homocysteine review found substantial risk of bias and likely overestimation from small-study effects and publication bias, and no included trial reported adjudicated heart attacks, strokes or cardiovascular deaths [12]. The type 2 diabetes estimate is about three times the size of the other review's and combined folic acid with B12 [14]. Lower homocysteine did not translate into fewer cardiovascular events in kidney disease [13].
TMGStrong

Homocysteine

Shown to lower homocysteine about 1.2 to 1.3 μmol/L at 4 g a day or more in healthy adults. Heart disease was not measured.

The numbers and sources
Effect
Across 5 RCTs of at least 4 g a day for 6 to 24 weeks, betaine lowered plasma homocysteine by 1.23 μmol/L (95% CI -1.61 to -0.85) [1]. A later meta-analysis found a fall of 1.30 μmol/L (95% CI -1.61 to -0.98) [2].
Population
Healthy adults in 5 RCTs [1]; RCTs pooled up to February 2021 [2].
Sources
[1], [2]
Caveats
The two reviews overlap in trials. Neither measured heart disease or death.
Beta-sitosterolModerate

Blood pressure and other heart risk markers

Shown to lower systolic blood pressure by about 1.6 mmHg for plant sterols and stanols as a class. HDL, blood sugar and weight did not change.

The numbers and sources
Effect
Plant sterols and stanols lowered systolic blood pressure by 1.55 mmHg (95% CI 0.42 to 2.67), diastolic by 0.84 mmHg, apoB by 0.06 g/L and hs-CRP by 0.25 mg/L. HDL cholesterol, fasting glucose, HbA1c, body weight, waist and fat mass did not change [5].
Population
Adults of mixed health status in trials pooled across 14 meta-analyses. Blood pressure came from one of them [5].
Sources
[5]
Caveats
Class-level evidence, not beta-sitosterol alone. These are risk markers, not heart attacks or strokes. Certainty was moderate to high for most outcomes, downgraded for inconsistency, imprecision or publication bias [5].
CitrullineModerate

Blood flow and blood vessel function

Shown to raise flow-mediated dilation, a measure of artery widening, in middle-aged and older adults. Arterial stiffness did not change.

The numbers and sources
Effect
In 8 RCTs (176 participants), L-citrulline raised flow-mediated dilation, a measure of artery widening (mean difference 1.81, 95% CI 0.76 to 2.85), and did not change pulse wave velocity, a measure of arterial stiffness (-0.14, 95% CI -0.45 to 0.17). Watermelon changed neither [12]. In the postmenopausal review, 2 of 4 trials found flow-mediated dilation improved, and arterial stiffness findings were mixed [9].
Population
Middle-aged and older adults [12]; postmenopausal women [9].
Sources
[12], [9]
Caveats
Small trials. These are measures of vessel function, not heart attacks or strokes.
Coenzyme Q10Moderate

Heart failure

Shown to lower deaths and heart failure hospitalizations in people with heart failure on standard therapy. The largest trial was funded by a CoQ10 maker.

The numbers and sources
Effect
Across 33 RCTs, CoQ10 was associated with lower all-cause mortality (RR 0.64, 95% CI 0.48 to 0.85, moderate quality) and hospitalization for heart failure (RR 0.50), and 31.7 m more on the 6-minute walk test [6]. In Q-SYMBIO, 420 patients on 100 mg three times a day for 2 years had fewer major cardiovascular events (15% versus 26%, HR 0.50, 95% CI 0.32 to 0.80) and lower all-cause mortality (10% versus 18%). Short-term endpoints at 16 weeks did not change [7].
Population
Patients with heart failure on standard therapy [6][7].
Sources
[6], [7]
Caveats
Q-SYMBIO was funded by Kaneka Corp., a CoQ10 manufacturer, and the International Coenzyme Q10 Association [7]. Several pooled outcomes were rated low quality [6].
PotassiumModerate

Stroke, heart events and death

Shown to lower stroke about 14% and death about 12% with a potassium salt substitute, mostly from one trial in older rural Chinese adults at high stroke risk.

The numbers and sources
Effect
Stroke was lower on a salt substitute of 75% sodium chloride and 25% potassium chloride than on regular salt (29.14 against 33.65 per 1,000 person-years, rate ratio 0.86, 95% CI 0.77 to 0.96). Major cardiovascular events (rate ratio 0.87) and death (rate ratio 0.88, 95% CI 0.82 to 0.95) were also lower [6]. Pooled trials found lower total mortality (RR 0.89, 95% CI 0.85 to 0.94) and cardiovascular events (RR 0.89) [2]. In cohort studies, higher potassium intake went with a 24% lower stroke risk (RR 0.76, 95% CI 0.66 to 0.89) [1]. In a post hoc analysis of the same trial, total cancer did not differ (RR 0.94, 95% CI 0.81 to 1.08), and stomach cancer showed a non-significant trend lower (RR 0.77, 95% CI 0.54 to 1.08) [7].
Population
20,995 people in 600 rural Chinese villages with prior stroke, or aged 60 or older with high blood pressure, mean follow-up 4.74 years [6][7]; 5 trials reporting clinical outcomes [2]; 11 cohorts with 127,038 people [1].
Sources
[1], [2], [6], [7]
Caveats
One trial carries most of the clinical events, and it enrolled people at high stroke risk. Cochrane rated the effects as slight and the certainty moderate [3]. The trial was funded by Australia's National Health and Medical Research Council [6].
Fish oilLimited

Heart attack, stroke and cardiac death in people on hemodialysis

One trial found about 43% fewer serious cardiovascular events with 4 g a day in adults on hemodialysis, who have very high cardiovascular risk.

The numbers and sources
Effect
Over 3.5 years, serious cardiovascular events occurred at 0.31 per 1,000 patient-days on fish oil and 0.61 on a corn-oil placebo (HR 0.57, 95% CI 0.47 to 0.70). Cardiac death (HR 0.55), heart attack (HR 0.56), stroke (HR 0.37) and amputation for vascular disease (HR 0.57) were each lower. Adherence and adverse events did not differ meaningfully between groups [7].
Population
1,228 adults on maintenance hemodialysis at 26 sites in Canada and Australia, given 4 g of omega-3 fatty acids a day (1.6 g EPA and 0.8 g DHA) [7].
Sources
[7]
Caveats
One trial, in people on hemodialysis, who have a very high cardiovascular risk. It does not describe the general population [7].
Why limited
This rests on one trial, of 1,228 people. Strength counts independent studies that agree, so one trial is limited at any size.
Glucosamine and chondroitinLimited

Cardiovascular events

Limited evidence: habitual glucosamine use was linked to about 15% fewer cardiovascular events in UK adults. No trial has tested this.

The numbers and sources
Effect
Among 466,039 UK Biobank participants followed a median 7 years, habitual glucosamine use was associated with fewer cardiovascular events (HR 0.85, 95% CI 0.80 to 0.90) and cardiovascular deaths (HR 0.78) [6].
Population
UK adults without cardiovascular disease at baseline [6].
Sources
[6]
Caveats
Observational. No randomized trial has tested this.
NiacinLimitedStudied in men

Long-term mortality after a heart attack, before statins

One trial from before statins found 11% lower mortality at 15 years in men after a heart attack, appearing after treatment had stopped.

The numbers and sources
Effect
At a mean 15 years, about 9 years after the trial ended, mortality was 52.0% in the niacin group and 58.2% on placebo, 11% lower (P = 0.0004). During the trial, niacin lowered nonfatal repeat heart attacks and did not lower total mortality [4].
Population
8,341 men aged 30 to 64 with a prior heart attack, treated 1966 to 1975 [4].
Sources
[4]
Caveats
One trial from before statins. The mortality difference appeared after treatment stopped [4].
Why limited
This rests on one trial, of 8,341 people. Strength counts independent studies that agree, so one trial is limited at any size.
PotassiumLimited

Heart rhythm in people with a defibrillator

One open-label trial found fewer arrhythmia events, hospitalizations and deaths when potassium was raised with supplements or a drug in adults with a defibrillator.

The numbers and sources
Effect
In 1,200 adults with an implantable defibrillator and plasma potassium of 4.3 mmol/L or lower, a strategy of raising potassium to 4.5 to 5.0 mmol/L with potassium supplements, a mineralocorticoid receptor antagonist or both lowered the composite of sustained ventricular tachycardia or appropriate shocks, unplanned hospitalization for arrhythmia or heart failure, and death: 22.7% against 29.2% on standard care (HR 0.76, 95% CI 0.61 to 0.95) over a median 39.6 months [8]. Atrial fibrillation events were 8.0% against 12.2% (subdistribution HR 0.65, 95% CI 0.45 to 0.93) [9]. Among those on potassium supplements alone, 21% reached the primary endpoint, against 29% of controls [10].
Population
1,200 adults with cardiac disease and a defibrillator in Denmark, POTCAST trial [8][9][10].
Sources
[8], [9], [10]
Caveats
One open-label trial in a high-risk group. The strategy combined supplements and a prescription drug, so the effect of potassium supplements alone was not randomized [10]. Funded by the Independent Research Fund Denmark and others [8].
Why limited
This rests on one trial, of 1,200 people. Strength counts independent studies that agree, so one trial is limited at any size.
Fish oilModerate

Cardiovascular outcomes with purified EPA (prescription)

Mixed results: one review found fewer cardiovascular deaths with prescription purified EPA, and another found no change in cardiovascular death at high doses.

The numbers and sources
Effect
Across 16 RCTs (127,771 patients, median 3.7 years), cardiovascular death fell with purified EPA (HR 0.79, 95% CI 0.67 to 0.94) and less with mixed EPA and DHA (HR 0.92, 95% CI 0.84 to 1.00) [8]. Across 6 trials (42,738 adults) of 1.8 g a day or more of EPA-dominant omega-3, hospitalization for unstable angina fell (RR 0.75, 95% CI 0.66 to 0.87), while cardiovascular death, ischemic stroke and hs-CRP did not change significantly. Recurrent heart attack and revascularization fell only after excluding STRENGTH, the one mixed EPA and DHA outcome trial [9]. Across 10 imaging trials (860 patients) on statins, total coronary plaque volume changed by -10.0% with added EPA, -3.3% with added EPA and DHA and +1.9% on statin alone [10]. In 23 RCTs of omega-3 in general, plaque volume did not change [2].
Population
Adults in primary or secondary prevention trials [8]; adults with established cardiovascular disease or high risk [9]; patients with coronary artery disease [10][2].
Sources
[8], [9], [10], [2]
Caveats
Purified EPA (icosapent ethyl) is a prescription drug, not an over-the-counter fish oil. The two outcome reviews disagree on cardiovascular death [8][9]. Authors of [8] and [10] include a REDUCE-IT investigator with Amarin research funding and an Amarin speakers bureau member [8].
Vitamin B12Moderate

Stroke and heart attack

Mixed results: stroke was slightly lower, 4.3% against 5.1%, with B vitamins mostly combined with folic acid, while heart attack and death did not change.

The numbers and sources
Effect
Across 15 RCTs with 71,422 participants, homocysteine-lowering B vitamins (B6, folic acid or B12, alone or combined) did not change heart attack (RR 1.02, 95% CI 0.95 to 1.10) or death from any cause (RR 1.01, 95% CI 0.96 to 1.06). Stroke was lower: 4.3% against 5.1% (RR 0.90, 95% CI 0.82 to 0.99, 10 trials, 44,224 people) [8].
Population
Adults with and without existing cardiovascular disease, followed 1 to 7.3 years. People with end-stage kidney disease were excluded [8].
Sources
[8]
Caveats
Most trials combined B12 with folic acid, so the stroke result cannot be assigned to B12 alone. The authors rated the stroke evidence high quality [8].
Vitamin B6Moderate

Heart attack, stroke and death

Mixed results: stroke was about 10% lower in pooled trials of B6, mostly with folic acid and B12, while heart attack and death did not change.

The numbers and sources
Effect
Across 15 trials of vitamins B6, B9 or B12, heart attack (RR 1.02, 95% CI 0.95 to 1.10) and death from any cause (RR 1.01, 95% CI 0.96 to 1.06) did not change. Stroke was lower (RR 0.90, 95% CI 0.82 to 0.99) [8]. In HOPE-2, 50 mg of B6 with folic acid and B12 lowered homocysteine by 2.4 µmol/L. Major cardiovascular events did not change (RR 0.95, 95% CI 0.84 to 1.07) and stroke fell (RR 0.75, 95% CI 0.59 to 0.97) [9]. In NORVIT, 40 mg of B6 alone did not change recurrent events (RR 1.14, 95% CI 0.98 to 1.32) [10].
Population
71,422 participants in 15 trials [8]; 5,522 people aged 55 or older with vascular disease or diabetes, 5 years [9]; 3,749 people within 7 days of a heart attack, median 40 months [10].
Sources
[8], [9], [10]
Caveats
Most trials combined B6 with folic acid and B12, so B6 alone cannot be separated. The Cochrane authors rated the heart attack and mortality evidence high quality [8].
Vitamin K2Limited

Artery and valve calcification

Small trials found mixed results: vitamin K slowed artery calcification in some trials but not valve calcification in men. Calcification is a scan marker.

The numbers and sources
Effect
Across 13 controlled trials (2,162 people), vitamin K reduced vascular calcification by 9.1% (95% CI 0.5 to 17.7) against control, with no significant change in arterial stiffness [4]. Over 2 years, coronary calcium scores rose from a median 145 to 214 on placebo and from 135 to 184 on 360 µg MK-7 a day, a significant difference [5]. In men with aortic valve calcification, 720 µg MK-7 plus vitamin D for 2 years did not change valve calcium progression (difference 17 units, 95% CI -86 to 53) or aortic and coronary calcium [6]. In 398 people with severe coronary calcification, 720 µg vitamin K2 plus 25 µg vitamin D a day for 24 months slowed the rise in calcium score (196 against 248 units, difference -52, 95% CI -79 to -24). Total plaque volume rose similarly in both groups [7].
Population
13 controlled trials with 2,162 participants [4]; 180 Dutch patients with symptomatic coronary disease [5]; 365 men from the community averaging 71 years [6]; 398 adults with a calcium score of 400 or more, median age 71 [7].
Sources
[4], [5], [6], [7]
Caveats
Calcification is a scan marker, not heart attack or death. The meta-analysis notes few studies and substantial heterogeneity [4]. The valve trial was funded by the Novo Nordisk Foundation [6]. The two 720 µg trials gave vitamin D with vitamin K2 [6][7].
Fish oilStrong

Death and cardiovascular events

Not shown to change deaths, cardiovascular events or stroke in adults. Coronary events and heart attacks fell slightly, by about 8% to 9%.

The numbers and sources
Effect
Long-chain omega-3 made little or no difference to all-cause mortality (RR 0.97, 95% CI 0.93 to 1.01, 45 RCTs), cardiovascular events (RR 0.96, 95% CI 0.92 to 1.01, 43 RCTs) or stroke (RR 1.02, 95% CI 0.94 to 1.12). Coronary heart disease events fell slightly (RR 0.91, 95% CI 0.85 to 0.97, number needed to treat 167, low certainty) [1]. A meta-analysis of 13 large trials (127,477 people, mean 5.0 years) found lower myocardial infarction (RR 0.92, 95% CI 0.86 to 0.99), coronary death (RR 0.92, 95% CI 0.86 to 0.98) and total cardiovascular disease (RR 0.97, 95% CI 0.94 to 0.99) with REDUCE-IT excluded, and a linear dose-response for total cardiovascular disease [6].
Population
143,693 adults for mortality [1]; 127,477 adults in 13 cardiovascular outcome trials [6].
Sources
[1], [6]
Caveats
The two reviews agree on small effect sizes and differ in emphasis. Adding REDUCE-IT, a trial of high-dose purified EPA, strengthened the associations but introduced significant heterogeneity [6]. Effects in [1] did not differ by dose or trial duration.
B-complex vitaminsModerate

Heart attack, stroke and death

Not shown to change heart attack or death in 15 trials of B6, folic acid or B12. Stroke was about 10% lower.

The numbers and sources
Effect
Across 15 trials of vitamins B6, B9 or B12, heart attack (RR 1.02, 95% CI 0.95 to 1.10) and death from any cause (RR 1.01, 95% CI 0.96 to 1.06) did not change. Stroke was lower (RR 0.90, 95% CI 0.82 to 0.99) [1]. In HOPE-2, folic acid, B6 and B12 lowered homocysteine and did not change major cardiovascular events (RR 0.95, 95% CI 0.84 to 1.07) [2].
Population
71,422 participants in 15 trials [1]; 5,522 people aged 55 or older with vascular disease or diabetes, 5 years [2].
Sources
[1], [2]
Caveats
The Cochrane authors rated heart attack and mortality evidence high quality [1]. Trials tested folic acid, B6 and B12, not full B-complex products.
GinkgoModerate

Walking distance with leg artery disease

Not shown to give a clinically meaningful gain in pain-free walking distance in people with intermittent claudication.

The numbers and sources
Effect
In 11 placebo-controlled trials (477 participants), pain-free walking distance rose by an effect size of 3.57 kcal (95% CI -0.10 to 7.23), about 64.5 meters on a flat treadmill at 3.2 km/h. Publication bias was likely to have inflated this [8].
Population
People with intermittent claudication [8].
Sources
[8]
Caveats
The review authors found no evidence of a clinically significant benefit [8].
MultivitaminModerate

Cardiovascular disease

Not shown to lower heart attack, stroke or cardiovascular death in older US adults over 3.6 to 11 years.

The numbers and sources
Effect
In the Physicians' Health Study II, a daily multivitamin for a median 11.2 years did not change major cardiovascular events (HR 1.01, 95% CI 0.91 to 1.10), myocardial infarction (HR 0.93), stroke (HR 1.06) or cardiovascular death (HR 0.95) [1]. In COSMOS, the composite cardiovascular outcome did not differ (HR 0.98, 95% CI 0.86 to 1.12) over a median 3.6 years [2]. Among 8,905 COSMOS participants without hypertension, new self-reported hypertension occurred in 22.9% on the multivitamin and 23.6% on placebo (HR 0.98, 95% CI 0.90 to 1.06), and 2-year systolic pressure change did not differ [3].
Population
14,641 US male physicians aged 50 and older [1]; 21,442 US women aged 65 and older and men aged 60 and older [2][3].
Sources
[1], [2], [4], [3]
Caveats
The US Preventive Services Task Force evidence review covered these trials and found little or no benefit for cardiovascular disease from vitamin and mineral supplements [4].
NiacinModerate

Heart attack, stroke and death

Not shown to lower heart attacks, strokes or deaths in people with vascular disease, including those on statins.

The numbers and sources
Effect
In AIM-HIGH, the primary composite occurred in 16.4% on niacin and 16.2% on placebo (HR 1.02, 95% CI 0.87 to 1.21). The trial stopped at a mean 3 years for lack of efficacy [1]. In HPS2-THRIVE, major vascular events occurred in 13.2% and 13.7% (rate ratio 0.96, 95% CI 0.90 to 1.03) [2]. Across 23 trials, niacin did not change overall mortality (RR 1.05, 95% CI 0.97 to 1.12), heart attacks (RR 0.93, 95% CI 0.87 to 1.00) or strokes (RR 0.95, 95% CI 0.74 to 1.22) [3].
Population
3,414 people [1]; 25,673 adults [2]; 39,195 participants in 23 trials, median dose 2 g a day [3].
Sources
[1], [2], [3]
Caveats
AIM-HIGH was funded by the US National Heart, Lung, and Blood Institute and Abbott Laboratories [1]. HPS2-THRIVE was funded by Merck and others, and added laropiprant to reduce flushing [2]. The Cochrane authors rated the mortality evidence high quality [3].
SeleniumModerate

Heart attack, stroke and mortality

Not shown to lower death, heart attack, stroke or other cardiovascular events in people, with events mostly from two US trials.

The numbers and sources
Effect
Selenium did not change death from any cause (RR 0.97, 95% CI 0.88 to 1.08), cardiovascular death (RR 0.97, 95% CI 0.79 to 1.2) or all cardiovascular events (RR 1.03, 95% CI 0.95 to 1.11) [3]. A Cochrane review of antioxidants found no effect of selenium on mortality (RR 0.97, 95% CI 0.91 to 1.03, 17 trials) [4].
Population
12 trials with 19,715 people, mostly the US SELECT and NPC trials [3]; 17 low-bias selenium trials [4].
Sources
[3], [4]
Caveats
Clinical events came mostly from two US trials [3].
Vitamin AModerate

Heart disease

Not shown to lower heart attack, stroke or cardiovascular death with beta-carotene in male physicians. One trial in smokers found 26% more cardiovascular deaths, not significant.

The numbers and sources
Effect
Over 12 years, heart attack (468 against 489), stroke (367 against 382) and cardiovascular death (338 against 313) did not significantly differ between beta-carotene and placebo [4]. In CARET, cardiovascular death was 26% higher on beta-carotene with retinyl palmitate (RR 1.26, 95% CI 0.99 to 1.61) [5].
Population
22,071 US male physicians [4]; 18,314 smokers, former smokers and asbestos-exposed workers [5].
Sources
[4], [5]
Caveats
The CARET interval crosses 1. CARET combined beta-carotene with retinol, so the two cannot be separated [5].
Vitamin DModerate

Cancer, cardiovascular disease and death

Not shown to lower cancer, cardiovascular disease or death in adults. Reviews disagree on cancer death, which one found about 16% lower.

The numbers and sources
Effect
In VITAL, invasive cancer did not differ (HR 0.96, 95% CI 0.88 to 1.06), nor did major cardiovascular events (HR 0.97, 95% CI 0.85 to 1.12) or death from any cause (HR 0.99, 95% CI 0.87 to 1.12). Cancer death had HR 0.83 (95% CI 0.67 to 1.02) [18]. Across 52 trials (75,454 participants), all-cause mortality did not change (RR 0.98, 95% CI 0.95 to 1.02) and cancer death fell (RR 0.84, 95% CI 0.74 to 0.95) [19]. Across 9 RCTs (114,379 participants), cardiovascular events (RR 0.95, 95% CI 0.88 to 1.04), cardiovascular death (RR 1.04, 95% CI 0.87 to 1.24) and heart attack (RR 0.96, 95% CI 0.83 to 1.11) did not differ from placebo [20]. Across 5 RCTs (56,901 participants), melanoma incidence (RR 0.94, 95% CI 0.71 to 1.24) and recurrence (HR 1.23, 95% CI 0.29 to 5.18) did not differ [21]. In 455 people starting chemotherapy for metastatic colorectal cancer, 4,000 IU a day after a loading dose did not change progression-free survival against 400 IU a day (median 11.8 against 10.3 months, P 0.25) or overall survival (25.6 against 27.0 months) [22].
Population
25,871 US adults over a median 5.3 years [18]; adults in 52 RCTs [19]; adults in 9 placebo-controlled RCTs [20]; adults in 5 RCTs, with or without calcium [21]; US adults with untreated metastatic colorectal cancer, median follow-up 20 months [22].
Sources
[18], [19], [20], [21], [22]
Caveats
The sources disagree on cancer death. VITAL's interval crossed 1.0 [18] and the pooled analysis did not [19]. Melanoma was a secondary endpoint in some trials and events were few [21].
Vitamin EModerate

Heart attack, stroke and cardiovascular death

Not shown to lower heart attack, stroke or cardiovascular death in male physicians or in people with vascular disease or diabetes.

The numbers and sources
Effect
Over 8 years, 400 IU every other day did not change major cardiovascular events (HR 1.01, 95% CI 0.90 to 1.13) in male physicians [1]. Over a median 7 years, 400 IU a day did not change major cardiovascular events (RR 1.04, 95% CI 0.96 to 1.14) in people with vascular disease or diabetes [2]. Across 9 trials, total stroke did not change (RR 0.98, 95% CI 0.91 to 1.05) [3].
Population
14,641 US male physicians aged 50 or older [1]; 9,541 people aged 55 or older with vascular disease or diabetes [2]; 118,765 participants in 9 trials [3].
Sources
[1], [2], [3]
Caveats
The stroke total hides opposite effects by subtype (see Risks) [3].
TaurineLimited

Heart failure

Limited evidence: pooled heart failure studies found no significant change in ejection fraction or stroke volume.

The numbers and sources
Effect
Of 11 studies, only ejection fraction and stroke volume had enough data to pool, and neither association was significant in all-cause heart failure [8].
Population
Patients with heart failure and ejection fraction under 50%, 500 mg to 6 g a day for 2 to 48 weeks [8].
Sources
[8]
Caveats
One study was rated high quality. No study was specific to dilated cardiomyopathy.
Vitamin CLimitedStudied in men

Heart attack, stroke and cardiovascular death

One trial found no change in heart attack, stroke or cardiovascular death over 8 years in male physicians aged 50 or older.

The numbers and sources
Effect
Over 8 years, major cardiovascular events did not differ on 500 mg a day against placebo (HR 0.99, 95% CI 0.89 to 1.11). Heart attack, stroke and cardiovascular death each did not differ [10].
Population
14,641 US male physicians aged 50 or older in the Physicians' Health Study II [10].
Sources
[10]
Caveats
One trial, men only.
Why limited
This rests on one trial, of 14,641 people. Strength counts independent studies that agree, so one trial is limited at any size.
Vitamin K2LimitedStudied in men

Heart attack, stroke and cardiovascular death

One small trial found no difference in heart events or deaths, with far too few events to test an effect.

The numbers and sources
Effect
In the valve trial, cardiovascular events (10 against 10) and deaths (1 against 4) did not differ between MK-7 plus vitamin D and placebo [6].
Population
365 men with aortic valve calcification, 24 months [6].
Sources
[6]
Caveats
Clinical events were safety outcomes in a small trial, far too few to test an effect.

Practices

Cold water immersionModerate

Heart rate variability

Shown to raise heart rate variability for up to 15 minutes after cold exposure in healthy adults, alongside a slightly lower heart rate and slightly higher blood pressure.

The numbers and sources
Effect
Across 24 studies of cold water immersion and cryostimulation, heart rate variability indices rose after exposure, including RMSSD (SMD 0.61) and high-frequency power (SMD 0.46), persisting up to 15 minutes. Heart rate fell (SMD -0.16) and mean blood pressure rose (SMD 0.28) [13]. An umbrella review of post-exercise studies also reported better heart rate variability immediately after immersion [6].
Population
Healthy participants in 24 studies [13].
Sources
[13], [6]
Caveats
The pooled analysis mixes cold water immersion with whole-body and partial-body cryostimulation. Effects were measured within minutes, and the authors note they depend on individual characteristics and cooling technique [13].
Infrared saunaModerate

Heart function in heart failure

Shown to raise ejection fraction and lower BNP over 2 to 4 weeks in people with heart failure, in pooled trials of 60°C far-infrared sessions.

The numbers and sources
Effect
A meta-analysis of 7 controlled studies found that 15 minutes in a 60°C infrared sauna followed by 30 minutes of warm rest, 5 times a week for 2 to 4 weeks, was associated with lower B-type natriuretic peptide (BNP), a smaller cardiothoracic ratio and higher left ventricular ejection fraction. Left ventricular and left atrial diameters and blood pressure did not change [1]. A second meta-analysis reported ejection fraction 3.27 percentage points higher and BNP 116.66 pg/mL lower after 2 to 4 weeks of sauna treatment [2]. In 188 patients, ejection fraction rose from 31.6% to 34.6% and BNP fell from 542 to 394 pg/mL after 2 weeks of daily sessions, with no significant change in controls [3].
Population
People with chronic heart failure in 9 controlled studies, 7 pooled [1]. Participants in 16 articles [2]. 188 patients, 112 assigned to Waon therapy and 76 to control [3].
Sources
[1], [2], [3]
Caveats
The reviewers rated the strength of evidence from moderate to insufficient, and only short-term effects were measured [1]. Nearly all trials come from Japan, many from the group that developed Waon therapy. Patients could not be blinded to heat. The searches in [1] found no eligible Finnish sauna trial in heart failure.
Infrared saunaModerate

Walking distance and symptoms in heart failure

Shown to improve 6-minute walk distance and symptom class in people with heart failure, though the largest trial missed its main endpoint, the change in BNP.

The numbers and sources
Effect
In 149 hospitalized patients with advanced heart failure, 10 daily sessions improved NYHA class, 6-minute walk distance and cardiothoracic ratio only in the Waon group. The primary endpoint, the change in BNP, did not reach statistical significance [4]. Pooled sauna studies found 6-minute walk distance 48.11 m longer after 2 to 4 weeks [2]. In 41 patients, walk distance rose from 337 to 379 m after 3 weeks [5]. In 49 patients, physical and mental quality of life scores on the SF-36 improved after 3 weeks [6].
Population
76 Waon and 73 control patients at 19 Japanese institutes, mean age 66 [4]. Participants in 16 articles [2]. 41 patients [5] and 49 patients [6] with chronic heart failure, no control groups.
Sources
[4], [2], [5], [6]
Caveats
The largest trial did not meet its primary endpoint [4]. Walking and symptom measures depend on effort and expectation, and no trial was blinded. Two of the studies had no control group [5][6].
Slow breathing and breathworkModerate

Heart rate variability

Shown to raise vagally mediated heart rate variability during, right after and across weeks of slow breathing practice, in one meta-analysis of 223 studies.

Compared with control conditions the source abstract does not describe [12]; mindfulness meditation [4]

The numbers and sources
Effect
Vagally mediated heart rate variability rose with voluntary slow breathing during the session (172 studies), immediately after one session (16 studies) and after multi-session programs (49 studies) [12]. In the 1-month trial of 5 minutes a day, resting heart rate variability and resting heart rate measured by a wrist wearable did not change in any group [4].
Population
223 studies [12]. 114 adults invited into a remote trial [4].
Sources
[12], [4]
Caveats
The meta-analysis abstract gives no effect sizes [12]. Heart rate variability during slow breathing rises partly because breathing itself drives heart rhythm, so a rise in the moment is not a lasting change in resting tone.
GroundingLimited

Red blood cell clumping

One uncontrolled study of 10 adults found red blood cells carried more surface charge and clumped less on slides after 2 hours grounded. Blood viscosity itself was not measured.

Compared with blood samples taken before grounding, in the same people [8]

The numbers and sources
Effect
After 2 hours with conductive patches on the hands and feet wired to a rod in the earth, red blood cell zeta potential rose by an average of 2.70 and red cell aggregation on microscope slides fell significantly [8].
Population
10 healthy adults recruited by word of mouth [8].
Sources
[8]
Caveats
No control condition. Fingertip blood on a slide is not blood flow in the body, and no cardiovascular outcome was measured. Two of the authors disclosed contracts with and shares in the company sponsoring grounding research in a later paper [6].
Infrared saunaLimited

Cardiac events over 5 years in heart failure

One non-randomized study found cardiac events in 31.3% of heart failure patients who continued Waon therapy for 5 years, against 68.7% of matched controls.

The numbers and sources
Effect
At 60 months, cardiac events, defined as events due to heart failure or cardiac death, had occurred in 31.3% of the Waon group and 68.7% of the control group (P < 0.01). Eight patients died in the Waon group and 12 in the control group [7].
Population
129 patients in NYHA class III or IV, 64 treated daily for 5 days in hospital and at least twice a week after discharge, and 65 controls matched for age, sex and class [7].
Sources
[7]
Caveats
Not randomized, single center, and the group that kept attending sessions may differ from those who did not.
Infrared saunaLimited

Irregular heartbeats in heart failure

One small trial found fewer premature ventricular beats after 2 weeks of daily sessions in people with heart failure, 848 against 3,097 per 24 hours.

The numbers and sources
Effect
After 2 weeks, premature ventricular contractions numbered 848 ± 415 per 24 hours in the sauna group and 3,097 ± 1,033 in the untreated group (p < 0.01). Heart rate variability (SDNN) was higher, 142 against 112 ms, and BNP was lower [8].
Population
30 patients with NYHA class II or III heart failure and at least 200 premature ventricular contractions a day, 20 randomized to sauna and 10 to bed rest [8].
Sources
[8]
Caveats
One small trial from the group that developed the therapy.
Infrared saunaLimited

Blood vessel function

Small studies found better artery dilation after 2 to 3 weeks of daily sessions in men with coronary risk factors and in people with heart failure.

The numbers and sources
Effect
In 25 men with at least one coronary risk factor, flow-mediated dilation rose from 4.0% to 5.8% after 2 weeks of daily sessions (p < 0.001) [9]. In 20 people with heart failure, flow-mediated dilation rose after 2 weeks while nothing changed in 10 untreated patients [10]. In uncontrolled heart failure studies it rose from 3.5% to 5.5% [5] and from 3.6% to 5.1% [6]. In 10 healthy women, a single 45-minute session did not change arterial stiffness compared with rest [11].
Population
25 men with risk factors, mean age 38 [9]. 30 people with heart failure [10]. 41 and 49 people with heart failure [5][6]. 10 healthy women [11].
Sources
[9], [10], [5], [6], [11]
Caveats
The risk factor study had no untreated comparison for the change over time [9]. All repeated-session studies are small and from Japan.
Infrared saunaLimited

Peripheral artery disease

One small trial found less leg pain and longer walking distance after 6 weeks of daily Waon therapy in people with peripheral artery disease.

The numbers and sources
Effect
Leg pain scores, ankle-brachial pressure index and 6-minute walking distance improved significantly after 6 weeks in the Waon group and not in the control group [15].
Population
21 people with peripheral artery disease on standard medication, 11 randomized to Waon therapy and 10 to control [15].
Sources
[15]
Caveats
Twenty-one people, and the abstract gives no sizes for the clinical changes.
SaunaLimited

Aerobic fitness when added to exercise

One trial found estimated VO2 max rose 2.7 mL/kg/min more when a 15-minute sauna followed each exercise session for 8 weeks, in 47 sedentary adults.

The numbers and sources
Effect
Estimated VO2 max rose 2.7 mL/kg/min more (95% CI 0.2 to 5.3) with exercise plus sauna than with exercise alone. Exercise alone raised it 6.2 mL/kg/min (95% CI 4.2 to 8.3) over the control group [3].
Population
47 adults aged 49 ± 9 with low physical activity and at least one cardiovascular risk factor, randomized to exercise plus sauna, exercise or control for 8 weeks [3].
Sources
[3]
Caveats
One trial of about 16 people per group. VO2 max was estimated, not measured by gas analysis. There was no sauna-only group.
SaunaLimited

Death from heart disease

Linked to fewer cardiovascular deaths in Finnish cohorts, with a hazard ratio of 0.37 for sudden cardiac death at 4 to 7 sessions a week against one.

The numbers and sources
Effect
In 2,315 Finnish men followed for a median 20.7 years, fatal cardiovascular disease occurred in 22.3% of men using a sauna once a week, 16.4% at 2 to 3 times and 12.0% at 4 to 7 times. The adjusted hazard ratio for sudden cardiac death was 0.78 (95% CI 0.57 to 1.07) at 2 to 3 sessions a week and 0.37 (95% CI 0.18 to 0.75) at 4 to 7 sessions. Sessions longer than 19 minutes carried a hazard ratio of 0.48 (95% CI 0.31 to 0.75) against sessions under 11 minutes [9]. In 1,688 men and women followed for 15.0 years, adjusted hazard ratios for cardiovascular death were 0.75 (95% CI 0.52 to 1.08) and 0.23 (95% CI 0.08 to 0.65) [10]. In 2,277 men over 26.1 years, 3 to 7 sessions a week against 2 or fewer carried a hazard ratio of 0.74 (95% CI 0.59 to 0.94) for cardiovascular death [11].
Population
Men aged 42 to 60 from eastern Finland in the Kuopio Ischemic Heart Disease study [9][11]. 1,688 participants with a mean age of 63, 51.4% women [10].
Sources
[9], [10], [11]
Caveats
Observational. All three analyses come from one Finnish cohort program and one research group, and sauna habits were self-reported at baseline. Only 201 men were in the 4 to 7 sessions group, with 10 sudden cardiac deaths [9]. The link may reflect other differences between frequent and infrequent sauna users.
SaunaLimited

Stroke

Linked to fewer strokes in Finnish men and women, with a hazard ratio of 0.38 at 4 to 7 sessions a week against one.

The numbers and sources
Effect
During a median 14.9 years, 155 strokes occurred. Compared with one session a week, 4 to 7 sessions carried a hazard ratio of 0.38 (95% CI 0.18 to 0.81) after adjustment for cardiovascular risk factors, physical activity and socioeconomic status [13].
Population
1,628 men and women aged 53 to 74 with no history of stroke, in the Kuopio Ischemic Heart Disease study [13].
Sources
[13]
Caveats
Observational, one cohort. Only 34 hemorrhagic strokes occurred, and the association there was described as modest [13].
SaunaModerate

Blood vessel function

Mixed results: pooled heat therapy trials found better artery dilation, while two 8-week sauna trials found no change in artery dilation or stiffness.

The numbers and sources
Effect
Heat therapy improved brachial artery flow-mediated dilation by 1.95 percentage points (95% CI 0.14 to 3.76, 5 studies) [1]. A sauna meta-analysis reported flow-mediated dilation 1.71 percentage points higher after 2 to 4 weeks [2]. In adults with stable coronary artery disease, 8 weeks of Finnish sauna did not change flow-mediated dilation (0.15% with sauna, 0.07% with control, interaction P = 0.909) or carotid-femoral pulse wave velocity (P = 0.816) [4]. In sedentary adults, pulse wave velocity did not differ between exercise plus sauna and exercise alone (-0.4 m/s, 95% CI -1.1 to 0.3) [3].
Population
Healthy and clinical populations [1][2]. 41 adults with coronary artery disease, 21 assigned to sauna at 79°C for 20 to 30 minutes, 4 times a week [4]. 47 sedentary adults [3].
Sources
[1], [2], [4], [3]
Caveats
Certainty was rated low for flow-mediated dilation [1]. In [4] the sauna group showed signs of heat acclimation, a lower resting core temperature and a higher sweat rate, without vascular change.
SaunaLimitedStudied in men

Blood clots in veins

Mixed results: Finnish men using a sauna 2 to 3 times a week had fewer venous clots than men using it once, and men at 4 or more did not.

The numbers and sources
Effect
Over 24.9 years, 146 of 2,242 men (6.5%) had a venous thromboembolism. Against one session a week or fewer, adjusted hazard ratios were 0.67 (95% CI 0.46 to 0.96) at 2 to 3 sessions and 0.92 (95% CI 0.51 to 1.68) at 4 or more [32].
Population
2,242 Finnish men aged 42 to 61 with no history of venous thromboembolism [32].
Sources
[32]
Caveats
Observational, one cohort. The most frequent users showed no association, so there is no dose pattern.
Infrared saunaLimitedStudied in women

Cardiovascular effects compared with exercise

One small trial found 45 minutes of infrared sauna raised body temperature but did not reproduce the breathing or cardiovascular responses of moderate cycling in 10 women.

The numbers and sources
Effect
Tympanic temperature was 1.05°C higher during infrared sauna than rest and 0.79°C higher than cycling. Respiratory rate was 6.66 breaths a minute higher during cycling than sauna. Blood pressure, arterial stiffness and heart rate variability did not differ between sauna, cycling and rest [11].
Population
10 healthy women, mean age 36, in a randomized crossover of rest, infrared sauna and indoor cycling [11].
Sources
[11]
Caveats
Ten women, single sessions.
SaunaLimited

Heart rate variability

One trial found no added change in heart rate variability when a 15-minute sauna followed exercise for 8 weeks, compared with exercise alone.

The numbers and sources
Effect
No heart rate variability index differed significantly between exercise plus sauna and exercise alone after 8 weeks, including RMSSD, SDNN and resting heart rate [31].
Population
38 sedentary adults aged 49 ± 9 with at least one cardiovascular risk factor, in three groups [31].
Sources
[31], [3]
Caveats
One small trial, drawn from the same study as the blood pressure and fitness findings in [3].

Daily life

Daily stepsStrong

Heart disease and stroke

Linked to less cardiovascular disease with more daily steps in pooled cohorts, a clear pattern in adults 60 and older and no clear pattern in younger adults.

Compared with people in the lowest quarter of steps [6]; 2,000 steps a day [3][4]; people in the lowest quarter, a median of 2,337 steps a day [2]

The numbers and sources
Effect
In adults aged 60 and older, cardiovascular disease risk against the lowest quarter was 0.80 in the second quarter, 0.62 in the third and 0.51 in the highest (95% CI 0.41 to 0.63). In adults under 60 the hazards were 0.79, 0.90 and 0.95, none significant [6]. Risk of new cardiovascular disease was lower from 2,735 steps a day (HR 0.89) and lowest near 7,126 steps (HR 0.49, 95% CI 0.45 to 0.55) [3]. Against 2,000 steps, 7,000 steps was linked to 25% fewer new cases (HR 0.75, 95% CI 0.67 to 0.85) and 47% fewer cardiovascular deaths (HR 0.53, 95% CI 0.37 to 0.77) [4]. Each extra 500 steps was linked to 7% lower cardiovascular mortality (HR 0.93, 95% CI 0.91 to 0.95) [2].
Population
20,152 adults in 8 cohorts, mean age 63.2, 1,523 events over a mean 6.2 years [6]. 111,309 people in 12 studies [3]. 6 studies for new disease and 3 for deaths [4]. 226,889 participants in 17 cohorts [2].
Sources
[6], [3], [4], [2]
Caveats
Observational. Early heart disease can lower activity before diagnosis, which is reverse causation. Heterogeneity was high for cardiovascular deaths (I² 78.2%), and certainty was rated low for that outcome [4].
Physical activityStrong

Heart disease and stroke

Linked to about 17% less cardiovascular disease and 23% to 29% fewer cardiovascular deaths at the guideline amount in pooled cohorts, with the largest gains moving from none to some.

Compared with people doing no non-occupational activity [2]; inactive people [3]; insufficiently active people, under 600 MET-minutes a week [4]

The numbers and sources
Effect
At 150 minutes a week of moderate activity, cardiovascular death had an RR of 0.71 (95% CI 0.66 to 0.77) [2]. Moving from inactive to that amount was linked to 23% fewer cardiovascular deaths (RR 0.77, 95% CI 0.71 to 0.84) and 17% less new cardiovascular disease (RR 0.83, 95% CI 0.77 to 0.89), after adjusting for body weight [3]. In the most active group, ischemic heart disease was 25% lower (RR 0.754) and ischemic stroke 26% lower (RR 0.736), with most of the gain at lower activity levels, up to 3,000 to 4,000 MET-minutes a week [4].
Population
28,884,209 person-years with 74,757 cardiovascular events [2]. 3,439,874 participants in 36 studies, followed an average 12.3 years [3]. 43 studies of heart disease and 26 of stroke [4].
Sources
[2], [3], [4]
Caveats
Observational. Early heart disease limits activity, which is reverse causation. The reviews draw on overlapping cohorts.
Sitting timeStrong

Heart disease

Linked to more cardiovascular disease and deaths from it with more sitting or TV time in pooled cohorts, with risk rising above about 6 hours of sitting a day.

Compared with people with less sedentary time, with physical activity held equal [1]; each hour less of sitting or TV a day [3]; each 2 hours less of TV a day [6]

The numbers and sources
Effect
Sedentary time was linked to cardiovascular death (HR 1.179, 95% CI 1.106 to 1.257) and new cardiovascular disease (HR 1.143, 95% CI 1.002 to 1.729) [1]. For cardiovascular death, each hour of sitting carried no added risk up to 6 hours (RR 1.01, 95% CI 0.99 to 1.02) and 4% more above 6 hours. Each hour of TV carried 8% more above 4 hours [3]. Each 2 hours of TV was linked to 15% more fatal or nonfatal cardiovascular disease (RR 1.15, 95% CI 1.06 to 1.23) [6].
Population
14 studies of cardiovascular disease and diabetes [1]. 34 studies [3]. 34,253 people with 1,052 cases in 4 cohorts [6].
Sources
[1], [3], [6]
Caveats
Observational. Heterogeneity in designs and in how sitting was measured was marked [1]. Confounding and reverse causation cannot be excluded.
Sleep durationStrong

Heart disease and stroke

Linked to more heart disease and stroke at both short and long sleep in pooled cohorts, with the lowest risk near 7 hours a night.

Compared with people sleeping 7 to 8 hours a night, the reference in most studies [1]; normal sleepers [3]; people with normal sleep duration [4]

The numbers and sources
Effect
Each hour under 7 was linked to 6% more total cardiovascular disease (RR 1.06, 95% CI 1.03 to 1.08) and 7% more coronary heart disease (RR 1.07, 95% CI 1.03 to 1.12). Each hour over 7 was linked to 12% more cardiovascular disease (RR 1.12) and 18% more stroke (RR 1.18, 95% CI 1.14 to 1.21) [1]. Short sleepers had higher risk of cardiovascular disease (RR 1.16, 95% CI 1.10 to 1.23), coronary heart disease (RR 1.26) and hypertension (RR 1.17) [3]. Long sleepers had higher risk of cardiovascular disease (RR 1.25), stroke (RR 1.46, 95% CI 1.26 to 1.69) and coronary heart disease (RR 1.24), and no higher risk of hypertension (RR 1.01, 95% CI 0.95 to 1.07) [4].
Population
3,582,016 participants in generally healthy populations, follow-up 2.3 to 34 years [1]. 5,172,710 participants in 153 studies [3]. 5,134,036 participants in 137 cohorts [4].
Sources
[1], [3], [4]
Caveats
Observational, with self-reported sleep. Confounding and reverse causation, where early heart disease or other illness changes sleep, cannot be ruled out. The three reviews draw on overlapping cohorts.
SmokingStrong

Heart disease and stroke after quitting

Linked to lower heart disease and stroke risk within 5 years of quitting, with the excess falling steadily with time but staying above never smokers for 10 to 15 years.

Compared with never smokers, with former and current smokers each set against them [7]; patients with heart disease who continued to smoke [8]; people who continued to smoke, and never smokers [9]; people who continued to smoke [10]

The numbers and sources
Effect
In adults 60 and older, cardiovascular death had a hazard ratio of 2.07 (95% CI 1.82 to 2.36) in current smokers and 1.37 (95% CI 1.25 to 1.49) in former smokers against never smokers. The excess fell continuously with time since quitting. Heart attack and stroke followed the same pattern [7]. In people with coronary heart disease, quitting was linked to 36% lower death from any cause than continuing to smoke (RR 0.64, 95% CI 0.58 to 0.71) [8]. In heavy smokers, quitting within 5 years was linked to lower cardiovascular disease (HR 0.61, 95% CI 0.49 to 0.76) than continuing. Against never smokers, the excess stopped being significant 10 to 15 years after quitting (HR 1.25, 95% CI 0.98 to 1.60) [9]. In adults without diabetes, recent quitters (HR 0.47) and long-term quitters (HR 0.46, 95% CI 0.34 to 0.63) had fewer cardiovascular events than smokers [10].
Population
503,905 adults 60 and older in 25 cohorts, with 37,952 cardiovascular deaths [7]. 20 cohorts of patients with coronary heart disease [8]. 8,770 Framingham participants followed a median 26.4 years, 2,371 of them heavy smokers [9]. 3,251 Framingham Offspring participants followed a mean 25 years [10].
Sources
[7], [8], [9], [10]
Caveats
Observational. Many cohorts did not fully control confounding or check smoking status, though restricting to higher-quality studies changed little [8]. In people with diabetes, the estimates were similar but not significant [10].
Social connectionStrong

Heart disease and stroke

Linked to about 29% more coronary heart disease and 32% more stroke with poor social relationships, and 34% more cardiovascular deaths with social isolation, in two meta-analyses.

Compared with people with better social relationships [4]; people who were not socially isolated [3]

The numbers and sources
Effect
Poor social relationships were linked to new coronary heart disease (RR 1.29, 95% CI 1.04 to 1.59) and stroke (RR 1.32, 95% CI 1.04 to 1.68), with no difference by gender [4]. Social isolation was linked to cardiovascular mortality (1.34, 95% CI 1.25 to 1.44), and to all-cause mortality in people who already had cardiovascular disease (1.28) [3].
Population
16 longitudinal datasets in high-income countries with 4,628 heart disease and 3,002 stroke events over 3 to 21 years [4]. 2,205,199 adults in 90 cohorts [3].
Sources
[4], [3]
Caveats
Observational. The lower confidence limits sit close to no difference [4]. Early heart disease can limit social life before diagnosis.
Time in natureStrong

Death and living near green space

Linked to fewer deaths in people living amid more greenery, in two meta-analyses, about 4% fewer per step of greenness within 500 meters of home.

Compared with people with lower greenspace exposure [2]; people with 0.1 lower NDVI greenness within 500 meters of home [12]

The numbers and sources
Effect
Greenspace exposure was linked to lower all-cause mortality (0.69, 95% CI 0.55 to 0.87) and cardiovascular mortality (0.84, 95% CI 0.76 to 0.93) [2]. Each 0.1 rise in satellite-measured greenness within 500 meters of home was linked to 4% lower mortality (HR 0.96, 95% CI 0.94 to 0.97, I² 95%) [12].
Population
143 studies [2]. 8,324,652 people in 9 cohorts from 7 countries [12].
Sources
[2], [12]
Caveats
Observational, and about where people live, not time they spend in nature. Income, air pollution and activity differ between greener and less green areas. Heterogeneity was very high [12], and the broader review notes poor study quality [2].
Pet ownershipModerate

Cardiovascular death

Linked to about 19% to 31% fewer cardiovascular deaths in pet or dog owners in two meta-analyses of cohorts. One review found no link for dogs alone.

Compared with people without a dog [1]; people without pets [2]; people without a dog, in single- and multiple-person households [3]

The numbers and sources
Effect
Dog ownership was linked to 31% lower cardiovascular death (RR 0.69, 95% CI 0.67 to 0.71) [1]. Pet owners had lower cardiovascular death after adjustment (HR 0.81, 95% CI 0.68 to 0.97), and cat owners too (HR 0.79, 95% CI 0.63 to 0.99), while dog ownership showed no significant link [2]. In a Swedish registry cohort, dog ownership was linked to lower cardiovascular death in single-person households (HR 0.64, 95% CI 0.59 to 0.70) and multiple-person households (HR 0.85, 95% CI 0.81 to 0.90) [3].
Population
3,837,005 participants in 10 studies, mean follow-up 10.1 years [1]. 26 studies [2]. 3,432,153 Swedish adults followed up to 12 years [3].
Sources
[1], [2], [3]
Caveats
Observational. People in better health are more able to keep and walk a dog, which is reverse causation. The first review did not adjust for confounders [1]. In the Swedish twin subsample, the lower risk was not replicated [3].
Strength trainingModerate

Heart disease

Linked to 17% less cardiovascular disease in one meta-analysis of cohorts, with two other reviews finding borderline lower cardiovascular death.

Compared with people doing no muscle-strengthening activity [1]; no exercise [2]; no resistance training [3]

The numbers and sources
Effect
Muscle-strengthening activity was linked to lower cardiovascular disease (RR 0.83, 95% CI 0.73 to 0.93), J-shaped with the largest drop, 18%, at 60 minutes a week [1]. Cardiovascular death had an HR of 0.83 (95% CI 0.67 to 1.03) in one review [2] and an RR of 0.81 (95% CI 0.66 to 1.00) in another [3].
Population
257,888 adults with 16,056 cases in 7 cohorts [1]. 11 studies [2]. 4 studies for cardiovascular death [3].
Sources
[1], [2], [3]
Caveats
Observational. Two of three estimates had intervals reaching no difference, and certainty was rated very low [1].
Volunteering and purposeModerate

Purpose in life and heart events

Linked to 17% fewer cardiovascular events in people with a stronger sense of purpose in life, in one meta-analysis.

Compared with people with a lower sense of purpose in life [10]

The numbers and sources
Effect
Higher purpose in life was linked to fewer cardiovascular events (adjusted RR 0.83, 95% CI 0.75 to 0.92), with similar results by country, questionnaire, age and baseline heart disease [10].
Population
136,265 participants in 10 prospective studies [10].
Sources
[10]
Caveats
Observational. One review with overlapping cohorts for death and heart events.
Shift workLimitedStudied in women

Heart disease after leaving night shift work

One cohort of women nurses linked more years since leaving rotating night shift work to lower coronary heart disease risk.

Compared with nurses with a history of shift work who had left it more recently [3]

The numbers and sources
Effect
Among women who had ever worked rotating night shifts, longer time since quitting shift work was linked to lower coronary heart disease risk in the younger cohort (P for trend under 0.001). In the older cohort, the link between years of shift work and heart disease was stronger in the first half of follow-up, which the authors read as waning risk after stopping [3].
Population
189,158 initially healthy women nurses in the Nurses' Health Studies, followed 24 years, with 10,822 coronary heart disease cases [3].
Sources
[3]
Caveats
Observational, in one occupation. Nurses who leave shift work may differ in health and income from those who stay. The abstract reports a trend, not a size [3].
SunlightLimitedStudied in women

Death from any cause

One cohort of Swedish women linked active sun habits to fewer deaths from heart disease and other non-cancer causes, and 0.6 to 2.1 years longer life.

Compared with women who avoided sun exposure [13]

The numbers and sources
Effect
Women with active sun exposure habits had lower risk of cardiovascular and non-cancer, non-cardiovascular death than women who avoided the sun. Life expectancy of avoiders was 0.6 to 2.1 years shorter than in the highest sun exposure group. Because they lived longer, cancer made up a larger share of deaths among sun-seeking women [13].
Population
29,518 Swedish women aged 25 to 64 at the start, followed 20 years [13].
Sources
[13]
Caveats
One observational cohort. Women who seek the sun may be healthier, more active and better off, and ill women may avoid going out, which is reverse causation.
VapingLimited

Blood vessel function after switching

One trial found better blood vessel function within 1 month in smokers who switched to e-cigarettes, with or without nicotine, than in those who kept smoking.

Compared with people who kept smoking, a nonrandomized preference group [6]

The numbers and sources
Effect
Within 1 month of switching, flow-mediated dilation improved by 1.49% (95% CI 0.93 to 2.04) against continued smoking, and arterial stiffness fell (pulse wave velocity -0.529 m/s, 95% CI -0.946 to -0.112). Women gained more than men. E-cigarettes with and without nicotine did not differ [6].
Population
Adults smoking at least 15 cigarettes a day for at least 2 years, free of heart disease, randomized to e-cigarettes with or without nicotine, with a smoking comparison group [6].
Sources
[6]
Caveats
The comparison with continued smoking was not randomized and was adjusted by propensity score. Follow-up was 1 month [6].
NappingStrong

Short naps and heart disease and death

Not linked to heart disease or death for naps under an hour in two pooled cohort analyses. One Swiss cohort linked napping once or twice weekly to fewer heart events.

Compared with people who did not nap [6][7][8]

The numbers and sources
Effect
Naps under 60 minutes a day were not linked to cardiovascular disease (p = 0.98) or death from any cause (p = 0.08). The dose-response curve for heart disease was J-shaped: risk dipped from 0 to 30 minutes, rose slightly to about 45 minutes and rose sharply after that [6]. Naps under an hour were not linked to death (HR 1.00, 95% CI 0.90 to 1.11) or cardiovascular disease (HR 1.03) [7]. In one cohort, napping 1 to 2 times a week was linked to fewer cardiovascular events (HR 0.52, 95% CI 0.28 to 0.95), and nap length was not linked to events [8].
Population
151,588 participants in 11 cohorts, mean follow-up 11 years [6]. 371,306 participants in 21 studies [7]. 3,462 Swiss adults without heart disease followed 5.3 years, 155 events [8].
Sources
[6], [7], [8]
Caveats
Observational, with naps self-reported. The dip near 30 minutes was not tested against a trial. The Swiss finding rests on few events with a wide interval [8].

Food and drink

CoffeeStrong

Heart disease and stroke

Linked to about 15% lower cardiovascular disease at 3 to 5 cups a day than no coffee, in pooled cohorts. Heavy drinking was not linked to higher risk.

Compared with people drinking no coffee [2]; the lowest intake group, median 0 cups a day [6]

The numbers and sources
Effect
At 3 to 4 cups a day against none, cardiovascular disease was lower (RR 0.85, 95% CI 0.80 to 0.90) and so was cardiovascular death (RR 0.81, 95% CI 0.72 to 0.90) [2]. Across 36 cohorts, cardiovascular disease risk was 0.89 (95% CI 0.84 to 0.94) at a median 1.5 cups a day, 0.85 (95% CI 0.80 to 0.90) at 3.5 cups and 0.95 (95% CI 0.87 to 1.03) at 5 cups, against the lowest intake. Links were nonlinear for both coronary heart disease and stroke [6]. Cardiovascular death was lowest at 3 cups a day, 21% (95% CI 16 to 26%) lower [4], and at 2.5 cups a day in a later update (RR 0.83, 95% CI 0.80 to 0.87) [3].
Population
1,279,804 participants and 36,352 cardiovascular cases in 36 cohorts [6]. Cohorts pooled in [2], [4] and [3].
Sources
[2], [6], [4], [3]
Caveats
Observational, with the same confounding and reverse causation limits as above. Genetic studies mostly did not support a causal effect [5].
Fish and seafoodStrong

Coronary heart disease

Linked to 9% lower coronary heart disease and 15% lower coronary death at the highest fish intake than the lowest, in pooled cohorts, with lean fish showing no link.

Compared with the lowest fish intake category [7][6]; 20 g a day less fish, in a dose-response model [7][5]

The numbers and sources
Effect
Across 22 cohorts, the highest fish intake was linked to lower coronary heart disease than the lowest (RR 0.91, 95% CI 0.84 to 0.97), and across 27 cohorts to lower coronary death (RR 0.85, 95% CI 0.77 to 0.94). Each extra 20 g a day was linked to 4% lower incidence and 4% lower death [7]. Each extra 20 g a day was also linked to lower cardiovascular death (RR 0.96, 95% CI 0.94 to 0.98) across 8 cohorts [5]. Fatty fish was linked to lower coronary heart disease (RR 0.92, 95% CI 0.86 to 0.97) and coronary death (RR 0.83, 95% CI 0.70 to 0.98). Lean fish was not, and neither type was significantly linked to overall cardiovascular disease [6].
Population
918,783 participants and 28,261 coronary cases, and 1,139,553 participants and 10,568 coronary deaths, in 40 cohorts [7]. 8 cohorts for cardiovascular death [5]. 20,531 coronary cases in the fatty fish cohorts and 21,636 in the lean fish cohorts [6].
Sources
[7], [5], [6]
Caveats
Observational, with moderate heterogeneity (I² 47.4% for incidence and 51.3% for death) [7]. Confounding by other habits and reverse causation are possible. Trials of advice to eat fish disagree (see the next block but one).
Fish and seafoodStrong

Stroke

Linked to 10% lower stroke at the highest fish intake than the lowest, and 6% lower for each 3 extra servings a week, in pooled cohorts.

Compared with the lowest fish intake category [8]; 3 fewer servings a week, in a dose-response model [9]

The numbers and sources
Effect
Across 33 cohorts, the highest against the lowest fish intake was linked to lower stroke (HR 0.90, 95% CI 0.85 to 0.96), with stronger links for hemorrhagic stroke (HR 0.88), women (HR 0.83) and Asia-Pacific cohorts (HR 0.87). Risk was 2% to 12% lower with intakes up to 100 to 700 g a week [8]. Across 15 cohorts, 3 extra servings a week were linked to 6% lower total stroke (RR 0.94, 95% CI 0.89 to 0.99), 0.90 (95% CI 0.84 to 0.97) for ischemic stroke and 0.90 (95% CI 0.76 to 1.06) for hemorrhagic stroke [9].
Population
33 prospective cohorts [8]. 383,838 participants and 9,360 strokes in 15 cohorts [9].
Sources
[8], [9]
Caveats
Observational. One review rated its evidence moderate by NutriGrade [8], and the other described the link as weak [9]. The two reviews differ on hemorrhagic stroke.
Green teaStrong

Heart attack and stroke

Linked to 19% lower heart attack and 36% lower stroke at 1 to 3 cups a day than under 1 cup, in pooled studies, mostly from Asia.

Compared with people drinking under 1 cup a day [3][2]

The numbers and sources
Effect
At 1 to 3 cups a day, heart attack (OR 0.81, 95% CI 0.67 to 0.98) and stroke (OR 0.64, 95% CI 0.47 to 0.86) were lower than at under 1 cup, and heart attack was lower at 4 or more cups (OR 0.68, 95% CI 0.56 to 0.84). Non-drinkers had higher cardiovascular disease (OR 1.19, 95% CI 1.09 to 1.29) than people drinking under 1 cup a day [3]. In Japanese cohorts, 5 or more cups a day was linked to lower heart disease and cerebrovascular death [2].
Population
259,267 people in 9 studies [3]. 313,381 Japanese adults [2].
Sources
[3], [2]
Caveats
Observational. The pooled analysis mixed observational studies and trials, and the abstract gives no study count by design [3].
NutsStrong

LDL cholesterol and ApoB

Shown to lower LDL cholesterol by about 4.8 mg/dL per daily ounce of tree nuts, and by 10.2 mg/dL at about 67 g a day, along with ApoB.

Compared with control arms without nuts [4]; control diets the abstract does not describe [5]

The numbers and sources
Effect
Per 1 oz (28.4 g) serving a day, tree nuts lowered total cholesterol by 4.7 mg/dL (95% CI 4.0 to 5.3), LDL cholesterol by 4.8 mg/dL (95% CI 4.2 to 5.5), ApoB by 3.7 mg/dL (95% CI 2.3 to 5.2) and triglycerides by 2.2 mg/dL (95% CI 0.5 to 3.8). HDL cholesterol, apolipoprotein A1 and blood pressure did not change significantly. Effects were stronger at 60 g a day or more and did not differ by nut type. ApoB fell more in people with type 2 diabetes (11.5 mg/dL) than in healthy people (2.5 mg/dL) [4]. Pooled individual data from 25 trials found a mean 67 g a day lowered total cholesterol by 10.9 mg/dL (5.1%) and LDL cholesterol by 10.2 mg/dL (7.4%). Triglycerides fell 20.6 mg/dL in people starting at 150 mg/dL or higher and not in others [5].
Population
2,582 adults without cardiovascular disease in 61 trials of 3 to 26 weeks [4]. 583 adults with normal or high cholesterol not on lipid-lowering drugs, in 25 trials in 7 countries [5].
Sources
[4], [5]
Caveats
Trials were short and many are shared between the two analyses. Effects were largest in people with higher starting LDL cholesterol and lower body mass index [5].
Olive oilStrong

Heart disease and stroke in cohorts

Linked to 16% less cardiovascular disease per extra 25 g a day in cohorts, mostly through less stroke. Coronary heart disease alone showed no clear link.

Compared with 25 g a day less, in a dose-response model [3][6]; the bottom third of intake [4]

The numbers and sources
Effect
Each extra 25 g a day was linked to less cardiovascular disease (RR 0.84, 95% CI 0.76 to 0.94) [3]. An earlier pooling found less stroke (RR 0.74, 95% CI 0.60 to 0.92) and fewer combined cardiovascular events (RR 0.82, 95% CI 0.70 to 0.96), and no significant link with coronary heart disease in cohorts (RR 0.96, 95% CI 0.78 to 1.18) [6]. The top third of intake was linked to fewer cardiovascular events and strokes than the bottom third [4].
Population
806,203 participants and 49,223 cardiovascular events in cohorts and one trial [3]. 101,460 participants for coronary heart disease and 38,673 for stroke [6]. 32 cohorts [4].
Sources
[3], [6], [4]
Caveats
Observational, with heterogeneity. Small-study effects were suspected [6].
Whole grainsStrong

Heart disease and stroke

Linked to 19% lower coronary heart disease and 22% lower cardiovascular disease per three daily servings of whole grains, in pooled cohorts. The stroke link did not reach significance.

Compared with 90 g a day less whole grain, in a dose-response model [3]; the lowest whole grain intake [4]

The numbers and sources
Effect
Per extra 90 g a day of whole grain, about three servings, the relative risk was 0.81 (95% CI 0.75 to 0.87) for coronary heart disease, 0.78 (95% CI 0.73 to 0.85) for cardiovascular disease and 0.88 (95% CI 0.75 to 1.03) for stroke [3]. The highest against the lowest whole grain intake was linked to lower cardiovascular death (RR 0.82, 95% CI 0.79 to 0.85, I² 0%), and each extra 16 g a day, about one serving, to 9% lower cardiovascular death [4]. A Lancet series of reviews observed similar findings for whole grains as for dietary fiber, where the highest intakes were linked to 15% to 30% lower death, coronary heart disease and stroke, and graded the whole grain evidence low to moderate in certainty [5].
Population
45 prospective studies [3]. 786,076 participants and 23,957 cardiovascular deaths in 14 cohorts [4]. 185 prospective studies and 58 trials [5].
Sources
[3], [4], [5]
Caveats
Observational. People who eat more whole grains differ in other habits such as smoking and exercise (confounding), and illness can change diet (reverse causation). The certainty was graded low to moderate [5].
Dark chocolateModerate

Blood vessel function and insulin resistance

Shown to improve blood vessel widening and lower insulin resistance in short trials of chocolate or cocoa, in one meta-analysis rated low to moderate quality.

Compared with control conditions the abstract does not describe [3]

The numbers and sources
Effect
Flow-mediated dilation rose 1.34% (95% CI 1.00 to 1.68) with regular intake and 3.19% after a single dose. HOMA-IR fell 0.67 (95% CI 0.36 to 0.98) through lower serum insulin. Changes in LDL cholesterol (-0.07 mmol/L, 95% CI -0.13 to 0.00) and HDL cholesterol were of borderline significance [3].
Population
1,297 participants in 42 trials of up to 18 weeks [3].
Sources
[3]
Caveats
The authors rated the evidence low to moderate quality and called for larger, longer, independently funded trials [3].
SaltModerate

Stroke, heart events and death, with salt substitutes

Shown to lower stroke by 14%, major heart events by 13% and death by 12% in one large trial of older Chinese adults, an effect that drives pooled results.

Compared with regular salt [10][7][8]

The numbers and sources
Effect
Over a mean 4.74 years, stroke was less frequent with a 75% sodium chloride, 25% potassium chloride substitute than with regular salt (29.14 against 33.65 events per 1,000 person-years, rate ratio 0.86, 95% CI 0.77 to 0.96), as were major cardiovascular events (rate ratio 0.87, 95% CI 0.80 to 0.94) and death (rate ratio 0.88, 95% CI 0.82 to 0.95) [10]. A Cochrane review rated substitutes as probably reducing non-fatal stroke, non-fatal acute coronary syndrome and cardiovascular death slightly, by 20, 150 and 180 fewer per 100,000 person-years [7]. A network meta-analysis estimated 7 to 17 fewer deaths per 1,000 people, and found that removing the large Chinese trial left no significant effect on death from any cause [8].
Population
20,995 people in 600 Chinese villages with past stroke, or aged 60 or older with high blood pressure, mean age 65.4, 49.5% women [10]. Adults in 3 trials with event data [7]. 34 trials [8].
Sources
[10], [7], [8]
Caveats
One open-label trial in rural China, where most salt is added at home, carries the pooled results, and no trial outside China reported deaths [8]. The reviewers describe very limited generalizability beyond Chinese populations who add much of their salt at home [8].
CoffeeLimited

Atrial fibrillation

One trial found less atrial fibrillation recurrence with about a cup a day than with abstaining from coffee, 47% against 64%, after cardioversion.

Compared with abstaining from coffee and caffeine [16]; people drinking one cup a day fewer [17]

The numbers and sources
Effect
In 200 coffee drinkers with persistent atrial fibrillation after electrical cardioversion, recurrence over 6 months was 47% with caffeinated coffee and 64% with abstinence (HR 0.61, 95% CI 0.42 to 0.89). Adverse events did not differ [16]. In 386,258 UK adults, each extra daily cup was linked to 3% lower risk of any arrhythmia (HR 0.97, 95% CI 0.96 to 0.98), and genetic analyses found no link between caffeine metabolism and arrhythmia [17].
Population
200 adults, mean age 69, 71% men, in an open-label randomized trial at 5 hospitals [16]. 386,258 UK Biobank participants followed a mean 4.5 years [17].
Sources
[16], [17]
Caveats
One open-label trial in people who already drank coffee and already had atrial fibrillation [16]. The cohort is observational.
Dark chocolateLimited

Heart disease and stroke, in cohorts

Linked to slightly lower heart disease and stroke in cohort meta-analyses, with the lowest risk near 45 g a week, rated low credibility.

Compared with people eating 20 g a week less [4]; people eating 10 g a day less [5]

The numbers and sources
Effect
Each 20 g a week more chocolate was linked to 1.8% lower cardiovascular disease (RR 0.982, 95% CI 0.972 to 0.992) and 4.4% lower stroke. Risk was lowest at about 45 g a week (RR 0.890), and the authors suggest higher intakes may lose the link [4]. Per 10 g a day, coronary heart disease (RR 0.96, 95% CI 0.93 to 0.99) and stroke (RR 0.90, 95% CI 0.82 to 0.98) were slightly lower, while heart failure and type 2 diabetes were not linked [5].
Population
405,304 participants with 35,093 cases in 23 studies [4]. 27 cohort analyses [5].
Sources
[4], [5]
Caveats
Observational. Most studies measured all chocolate, not dark chocolate. The credibility of the evidence was rated low to very low [5].
NutsLimited

Heart disease, in cohorts

Linked to 21% to 29% lower cardiovascular disease per daily serving in cohort meta-analyses. Stroke was not significantly linked.

Compared with people eating 28 g a day less [6]; people eating one serving a day less, and the lowest intake group [7]

The numbers and sources
Effect
Per 28 g a day more, nuts were linked to lower coronary heart disease (RR 0.71, 95% CI 0.63 to 0.80) and cardiovascular disease (RR 0.79, 95% CI 0.70 to 0.88), and not significantly to stroke (RR 0.93, 95% CI 0.83 to 1.05). Results were similar for tree nuts and peanuts [6]. Per serving a day, ischemic heart disease (RR 0.72) and cardiovascular disease (RR 0.71) were lower, and stroke did not differ between the highest and lowest intakes (RR 0.91, 95% CI 0.81 to 1.02) [7].
Population
20 prospective studies [6]. 18 prospective studies with 8,862 cardiovascular cases [7].
Sources
[6], [7]
Caveats
Observational. People who eat nuts may differ in diet, exercise and income, so confounding is possible, and people who are unwell may eat fewer nuts. Heterogeneity was moderate [6].
NutsLimited

Heart attack and stroke, in a diet trial with nuts

One trial found 28% fewer heart attacks, strokes or heart deaths with a Mediterranean diet plus mixed nuts than with advice to cut dietary fat.

Compared with a control diet with advice to reduce dietary fat [8]

The numbers and sources
Effect
Over a median 4.8 years, major cardiovascular events occurred in 3.4% of the Mediterranean diet with nuts group and 4.4% of the control group (HR 0.72, 95% CI 0.54 to 0.95) [8].
Population
7,447 Spanish adults aged 55 to 80 at high cardiovascular risk, 57% women [8].
Sources
[8]
Caveats
Nuts were one part of a whole diet change, so the trial does not isolate nuts. The original report was withdrawn after randomization problems, and these are reanalyzed estimates [8].
Why limited
This rests on one trial, of 7,447 people. Strength counts independent studies that agree, so one trial is limited at any size.
Olive oilLimited

Heart attack, stroke and cardiovascular death in a trial

One large trial found fewer major cardiovascular events on a Mediterranean diet with extra virgin olive oil than on low-fat advice, 3.8% against 4.4%.

Compared with a control diet of advice to reduce dietary fat [7]

The numbers and sources
Effect
Over a median 4.8 years, major cardiovascular events (heart attack, stroke or cardiovascular death) occurred in 3.8% of the extra virgin olive oil group and 4.4% of the control group (HR 0.69, 95% CI 0.53 to 0.91). Results were similar after leaving out 1,588 participants whose assignment departed from the protocol [7].
Population
7,447 Spanish adults aged 55 to 80 at high cardiovascular risk, 57% women, assigned to one of three diets [7].
Sources
[7]
Caveats
The olive oil was given with a whole Mediterranean diet and education sessions, so the trial does not isolate olive oil. The first report was withdrawn over randomization problems [7]. The mediterranean-diet page covers this trial's weight, diabetes and cognition results.
Why limited
This rests on one trial, of 7,447 people. Strength counts independent studies that agree, so one trial is limited at any size.
EggsStrong

Heart disease and stroke

Mixed results: most pooled cohorts linked up to one egg a day to no change in heart disease or stroke, while a 2022 analysis found 4% higher risk.

Compared with one fewer egg a day [5][6]; under one egg a week or never [7]; half an egg a day less [8]

The numbers and sources
Effect
Across 33 risk estimates, one more egg a day was not linked to cardiovascular disease (RR 0.98, 95% CI 0.93 to 1.03), coronary heart disease (0.96) or stroke (0.99). It was linked to lower risk in Asian cohorts (0.92, 95% CI 0.85 to 0.99) and not in US or European ones [5]. At 1 or more eggs a day against fewer than 1 a week, cardiovascular disease was not raised (HR 0.96, 95% CI 0.88 to 1.05) [7]. A 2022 update of 49 estimates linked each extra 50 g egg a day to higher risk (RR 1.04, 95% CI 1.00 to 1.08), 1.08 in US cohorts and no link in Asian ones [6]. In 6 pooled US cohorts, each extra half egg a day was linked to higher risk (HR 1.06, 95% CI 1.03 to 1.10), a link that disappeared after adjusting for dietary cholesterol [8].
Population
1,720,108 participants and 139,195 events [5]. 22 cohorts [7]. 3,601,401 participants and 255,479 events [6]. 29,615 US adults followed a median 17.5 years [8].
Sources
[5], [7], [6], [8]
Caveats
Observational. People eating more eggs in US cohorts had higher BMI, ate more red meat and were less often on statins [5]. Heterogeneity was high in the 2022 update (I² 80.1%) [6]. Results differ by region, which may reflect what eggs are eaten with.
AlcoholModerate

Heart attack and coronary heart disease

Mixed results: cohorts linked drinking to about 29% less coronary heart disease than non-drinking, but genetic studies found no protection and some found higher risk.

Compared with non-drinkers [2]; each 100 g a week less, among current drinkers [4]; genetically predicted lower intake [7][8][6]

The numbers and sources
Effect
Drinkers had less incident coronary heart disease (RR 0.71, 95% CI 0.66 to 0.77) and less coronary death (RR 0.75, 95% CI 0.68 to 0.81) than non-drinkers, lowest at 1 to 2 drinks a day [2]. Among current drinkers, each 100 g a week was linked to slightly lower heart attack risk (HR 0.94, 95% CI 0.91 to 0.97) [4]. In Chinese men, self-reported drinking of about 100 g a week showed a U-shaped link with heart attack, but genetically predicted intake was not linked to it (RR 0.96 per 280 g a week, 95% CI 0.78 to 1.18) [7]. A gene variant tied to drinking 17.2% less was linked to lower odds of coronary heart disease (OR 0.90, 95% CI 0.84 to 0.96) [8]. In UK adults, genetically predicted intake was linked to 1.4-fold higher coronary artery disease risk per standard deviation, with minimal rise at light intake [6].
Population
84 cohorts [2]. 599,912 current drinkers [4]. 512,715 Chinese adults, 161,498 genotyped [7]. 261,991 people of European descent in 56 studies [8]. 371,463 UK Biobank participants [6].
Sources
[2], [4], [7], [8], [6]
Caveats
Cohorts share the sick quitter and lifestyle problems above. Mendelian randomization uses genes that set drinking levels to sidestep those problems, and its results disagree with the cohorts. Genetic studies have their own limits, such as genes acting through other paths.
SaltModerate

Heart attack and stroke, from eating less salt

Mixed results: pooled trials found about 23% fewer cardiovascular events, which reviewers called underpowered and driven by one retirement home trial. A trial follow-up found 25% fewer.

Compared with control groups without salt reduction advice or substitution [11]; control groups without sodium reduction counseling [12]

The numbers and sources
Effect
A Cochrane review of 8 trials found weak evidence of fewer cardiovascular events with less salt (pooled RR 0.77, 95% CI 0.63 to 0.95, 6 trials, 5,912 people), driven by one trial in which retirement home kitchens cut salt. Results for cardiovascular death in people with hypertension were imprecise (RR 0.67, 95% CI 0.45 to 1.01) [11]. Ten to 15 years after two sodium reduction trials in adults with prehypertension, cardiovascular events were 25% less frequent in the former intervention groups (RR 0.75, 95% CI 0.57 to 0.99) [12].
Population
7,284 adults in 8 trials of 6 to 36 months [11]. 3,126 adults aged 30 to 54 with prehypertension, with event data from 2,415 [12].
Sources
[11], [12]
Caveats
The Cochrane authors state their review lacked the power to confirm clinically important effects [11]. The trial follow-up was observational after the trials ended, with 77% of participants reporting outcomes [12].
Fish and seafoodLimitedStudied in men

Death after a heart attack and in angina, in diet-advice trials

Mixed results: advice to eat fatty fish cut 2-year death 29% in men after a heart attack in one trial, while another found more cardiac death in men with angina.

Compared with men given no advice to eat fish [10][11][12]

The numbers and sources
Effect
In DART, men recovering from a heart attack who were advised to eat more fatty fish had 29% lower all-cause death over 2 years [10] (unadjusted HR 0.70, 95% CI 0.54 to 0.92), followed by higher risk over the next 3 years (HR 1.31, 95% CI 1.01 to 1.70), and no substantial long-term survival benefit [11]. In DART-2, men with angina advised to eat two portions of oily fish a week or take fish oil capsules had no fall in all-cause death, and higher cardiac death (adjusted HR 1.26, 95% CI 1.00 to 1.58) and sudden cardiac death (HR 1.54, 95% CI 1.06 to 2.23), largely in the group given capsules [12].
Population
2,033 men, mean age 56.5, recovering from a heart attack [10]. 3,114 men under 70 with angina, followed 3 to 9 years [12].
Sources
[10], [11], [12]
Caveats
Trials of dietary advice, so the fish actually eaten varied. DART-2 combined fish and capsule advice, and its authors call the excess cardiac death unexplained [12]. Trials of fish oil supplements are on the fish-oil page.
SaltLimited

Heart disease and stroke, in cohorts by sodium intake

Mixed results: in cohorts, higher measured sodium was linked to more cardiovascular events, while one large study found the link only where average intake exceeded 5 g a day.

Compared with people in the lowest quarter of sodium excretion [14]; communities with lower mean sodium intake [15]; people excreting 4.00 to 5.99 g of sodium a day [13]

The numbers and sources
Effect
In 10,709 adults with at least two 24-hour urine collections, the highest quarter of sodium excretion was linked to more cardiovascular events than the lowest (HR 1.60, 95% CI 1.19 to 2.14), and each extra 1,000 mg a day to 18% more (HR 1.18, 95% CI 1.08 to 1.29) [14]. Across 255 communities, sodium was linked to cardiovascular events and stroke only where mean intake was above 5 g a day, and in the lowest third of intake (mean 4.04 g a day) higher sodium was linked to fewer events [15]. In the same program, intakes below 3 g and above 7 g a day were both linked to more events than 4 to 6 g [13].
Population
10,709 adults in 6 cohorts, median follow-up 8.8 years [14]. 82,544 adults in 255 communities in 18 countries, median 8.1 years [15]. 101,945 adults in 17 countries [13].
Sources
[14], [15], [13]
Caveats
Observational. Studies using several 24-hour urine samples [14] found a straight-line link, while those using one morning sample [15][13] found a curve, and how sodium is measured is part of the disagreement. Confounding and reverse causation are possible.
NutsStrong

Inflammation and blood vessel function

Not shown to lower C-reactive protein in two meta-analyses of nut trials. Blood vessel widening, a measure of vessel function, improved in one.

Compared with control conditions the abstract does not describe [13]; control arms without nuts [4]

The numbers and sources
Effect
Across 25 trials, nuts made no significant difference to C-reactive protein (WMD -0.01 mg/L, 95% CI -0.06 to 0.03) or other inflammation markers. Flow-mediated dilation rose 0.79% (95% CI 0.35 to 1.23) across 8 trials [13]. Tree nuts had no significant effect on C-reactive protein in 61 trials [4].
Population
32 randomized trials in adults, 652 participants for vessel function [13]. 2,582 adults in 61 trials [4].
Sources
[13], [4]
Caveats
The authors note the C-reactive protein result may have been influenced by two studies [13].
DairyModerate

Heart disease, stroke and death from any cause

Not linked to heart disease, cardiovascular disease or death from any cause for total dairy or milk, in 29 pooled cohorts.

Compared with lower intakes, in dose-response models [11]

The numbers and sources
Effect
No associations were found for total dairy, high-fat or low-fat dairy, or milk with death from any cause, coronary heart disease or cardiovascular disease. Fermented dairy was linked to slightly lower death (RR 0.98, 95% CI 0.97 to 0.99, per 20 g a day) and cardiovascular disease (RR 0.98), and cheese to 2% lower cardiovascular disease per 10 g a day, but these links weakened when one large Swedish study was removed [11].
Population
938,465 participants with 93,158 deaths, 28,419 coronary cases and 25,416 cardiovascular cases in 29 cohorts [11].
Sources
[11]
Caveats
Observational. Heterogeneity was very high for the fermented dairy estimates (I² 94.4% and 87.5%) [11].
Dark chocolateLimited

Cardiovascular events, cocoa flavanol extract

One trial of a cocoa flavanol extract did not significantly lower total cardiovascular events in 21,442 older adults. Cardiovascular death, a secondary outcome, fell 27%.

Compared with placebo [6]

The numbers and sources
Effect
Over a median 3.6 years, total cardiovascular events occurred in 410 people taking cocoa extract and 456 taking placebo (HR 0.90, 95% CI 0.78 to 1.02). Cardiovascular death was lower (HR 0.73, 95% CI 0.54 to 0.98), and death from any cause did not differ significantly (HR 0.89, 95% CI 0.77 to 1.03). The authors reported no safety concerns [6].
Population
21,442 US women aged 65 and older and men aged 60 and older without major cardiovascular disease [6].
Sources
[6]
Caveats
This tested a capsule of cocoa extract with 500 mg of flavanols, not chocolate. Secondary outcomes do not carry the weight of the primary one.
Why limited
This rests on one trial, of 21,442 people. Strength counts independent studies that agree, so one trial is limited at any size.
Red meat and processed meatLimitedStudied in women

Heart disease, cancer and death on diets lower in red meat

One trial in 48,835 women found a diet lower in red meat made little or no difference to death, heart disease or cancer, at low certainty.

Compared with diets higher in red meat [3]

The numbers and sources
Effect
The trial found little or no effect of lower red meat intake on all-cause death (HR 0.99, 95% CI 0.95 to 1.03), cardiovascular death (HR 0.98), cardiovascular disease (HR 0.99), cancer death (HR 0.95, 95% CI 0.89 to 1.01), colorectal cancer (HR 1.04, 95% CI 0.90 to 1.20) or breast cancer (HR 0.97). The review found 12 eligible trials, most measuring only blood markers [3].
Population
48,835 women in one randomized trial, within a review of 12 trials lasting at least 6 months [3].
Sources
[3]
Caveats
The difference in red meat between the groups was small. The reviewers rated certainty low to very low [3].
Why limited
This rests on one trial, of 48,835 people. Strength counts independent studies that agree, so one trial is limited at any size.

Diets

DASH dietModerate

Heart disease, stroke and diabetes in cohorts

Linked to about 20% lower rates of cardiovascular disease, coronary heart disease, stroke and diabetes in cohort studies, at low certainty.

Compared with people who followed DASH less closely [7]

The numbers and sources
Effect
In 15 prospective cohorts (942,140 people), closer adherence to DASH was linked to lower cardiovascular disease (RR 0.80, 95% CI 0.76 to 0.85), coronary heart disease (RR 0.79), stroke (RR 0.81) and diabetes (RR 0.82, 95% CI 0.74 to 0.92) [7].
Population
Adults in prospective cohorts, observational [7].
Sources
[7]
Caveats
Observational. Certainty was low for cardiovascular disease incidence [7]. No trial of DASH has measured heart attacks or strokes.
Mediterranean dietModerate

Heart attack, stroke and cardiovascular death

Shown to lower heart attacks, strokes and cardiovascular deaths against minimal advice in adults at raised heart risk, with reviews disagreeing on how certain this is.

Compared with minimal intervention, such as a healthy diet brochure, and low-fat programmes, in a network meta-analysis [1]; a low-fat diet in the PREDIMED trial, and usual care in the Lyon Diet Heart Study [2]; a control diet of advice to reduce dietary fat, in Spanish adults [3]; a low-fat diet, in patients with coronary heart disease [4]; a prudent Western-type diet, in patients after a first heart attack [5]

The numbers and sources
Effect
In a network meta-analysis of 40 trials (35,548 people), Mediterranean programs lowered all-cause death (OR 0.72, 95% CI 0.56 to 0.92), cardiovascular death (OR 0.55, 95% CI 0.39 to 0.78), stroke (OR 0.65, 95% CI 0.46 to 0.93) and non-fatal heart attack (OR 0.48, 95% CI 0.36 to 0.65) against minimal intervention, at moderate certainty. Mediterranean and low-fat programs did not clearly differ for death or non-fatal heart attack [1]. A Cochrane review found low-quality evidence of little or no effect of the PREDIMED diets against a low-fat diet on cardiovascular death (HR 0.81, 95% CI 0.50 to 1.32) or total death (HR 1.0, 95% CI 0.81 to 1.24), and fewer strokes (HR 0.60, 95% CI 0.45 to 0.80, moderate quality) [2]. In PREDIMED, major cardiovascular events occurred in 3.8% on the olive oil diet, 3.4% on the nut diet and 4.4% on the low-fat control diet over a median 4.8 years (HR 0.69, 95% CI 0.53 to 0.91, and HR 0.72, 95% CI 0.54 to 0.95) [3]. In CORDIOPREV, events ran at 28.1 per 1,000 person-years on the Mediterranean diet against 37.7 on a low-fat diet over 7 years (adjusted HR 0.719 to 0.753) [4]. In the Lyon Diet Heart Study, cardiac death plus non-fatal heart attack totaled 14 events on a Mediterranean-type diet against 44 on a prudent Western-type diet over a mean 46 months [5].
Population
Adults at raised cardiovascular risk in 40 trials followed at least 9 months [1]; 30 RCTs with 12,461 people [2]; 7,447 Spanish adults aged 55 to 80 at high risk with no heart disease at entry [3]; 1,002 patients with coronary heart disease, 82.5% men [4]; 605 patients after a first heart attack [2][5].
Sources
[1], [2], [3], [4], [5]
Caveats
PREDIMED's first report was withdrawn after randomization problems at 2 of 11 sites. The reanalysis gave similar results after leaving out 1,588 affected participants [3]. The Cochrane review rated the death evidence low quality and the network meta-analysis rated it moderate [1][2]. In CORDIOPREV the effect appeared in men, and no difference was found in the 175 women [4]. CORDIOPREV was funded in part by olive oil foundations [4]. The Lyon trial added canola margarine and compared against usual care, and the Cochrane review rated it low quality [2].
Nordic dietModerate

Heart disease, diabetes and death in cohorts

Linked to lower death, heart attack, stroke and type 2 diabetes rates with higher adherence in cohort studies, at low certainty.

Compared with people with the lowest adherence to Nordic dietary patterns, in European cohorts [2]; people with lower adherence to the Nordic dietary pattern [9]; people in the lowest category of Nordic diet adherence [10]

The numbers and sources
Effect
In 13 cohorts (930,153 people), high against low adherence was linked to lower all-cause death (RR 0.78, 95% CI 0.69 to 0.87), cardiovascular death (RR 0.78), stroke (RR 0.88), heart attack (RR 0.80) and type 2 diabetes (RR 0.90, 95% CI 0.82 to 0.99) [10]. In 15 cohorts, total cardiovascular disease was 7% lower (RR 0.93, 95% CI 0.88 to 0.99, I² 88%) [2]. A 2025 meta-analysis of 32 cohorts found 12% lower type 2 diabetes and 8% lower cardiovascular disease [9].
Population
Adults in prospective cohorts, observational [2][9][10].
Sources
[2], [9], [10]
Caveats
Observational. Certainty ranged from very low to low [10], and was low for total cardiovascular disease [2].
Vegetarian dietModerate

Heart disease

Linked to about 22 to 28% lower coronary heart disease in cohort studies, with no significant link to total cardiovascular death.

Compared with omnivores, in cohort studies [9]; non-vegetarian dietary patterns, in cohorts [10]; meat eaters in the same UK cohort (EPIC-Oxford) [11]

The numbers and sources
Effect
In 10 cohorts, vegetarian diets were linked to lower incidence or death from ischemic heart disease (RR 0.75, 95% CI 0.68 to 0.82), and not to total cardiovascular disease, all-cause death or cancer death [9]. In 7 cohorts (197,737 participants), vegetarian diets were linked to lower coronary heart disease death (RR 0.78) and incidence (RR 0.72, 95% CI 0.61 to 0.85), and not to cardiovascular death (RR 0.92, 95% CI 0.84 to 1.02) [10]. In EPIC-Oxford over 18.1 years, vegetarians had 22% lower rates of ischemic heart disease than meat eaters (HR 0.78, 95% CI 0.70 to 0.87), about 10 fewer cases per 1,000 people over 10 years [11].
Population
Adults in prospective cohorts [9][10]; 48,188 UK adults [11]. All observational.
Sources
[9], [10], [11]
Caveats
Observational. Certainty was very low for all outcomes in [10]. The EPIC-Oxford link weakened after adjusting for cholesterol, blood pressure, diabetes and BMI (HR 0.90) [11]. Vegetarian groups in these cohorts include vegans.
Flexitarian dietLimited

Heart disease

Limited evidence: in a review of 4 trials in healthy adults, replacing part of animal protein with plant protein lowered LDL and total cholesterol. No study measured heart disease itself.

Compared with diets with a higher share of animal protein foods [9]

The numbers and sources
Effect
A 2026 systematic review found 4 dietary interventions, reported in 7 articles, that partially replaced animal protein foods with plant protein foods in healthy adults. LDL cholesterol, total cholesterol and body weight improved in those eating the higher share of plant protein. Vitamin B12 status fell on the diets with the most plant protein [9].
Population
Healthy adults with a mean or median age of 18 to 65 in randomized trials, 3 of 7 articles at low risk of bias [9].
Sources
[9]
Caveats
The abstract gives no pooled sizes of effect [9]. Cholesterol is a risk marker, not heart disease. No cohort or trial of flexitarian eating and heart attacks or heart deaths was located.
Low-fat dietModerate

Heart events against a Mediterranean diet

Shown to have more major heart events than a Mediterranean diet in two trials, one in people with heart disease and one in people at high heart risk.

Compared with diets with more fat

The numbers and sources
Effect
In people with coronary heart disease followed for 7 years, major cardiovascular events occurred in 111 of 500 on the low-fat diet and 87 of 502 on the Mediterranean diet, with adjusted hazard ratios of 0.719 to 0.753 for the Mediterranean diet. The difference appeared in men and not in the 175 women [11]. In people at high cardiovascular risk followed for a median of 4.8 years, events occurred in 4.4% of the reduced-fat control group against 3.8% and 3.4% on Mediterranean diets with olive oil or nuts (HR 0.69 and 0.72) [12].
Population
1,002 adults aged 20 to 75 with coronary heart disease in Spain, 82.5% men [11]; 7,447 adults aged 55 to 80 at high cardiovascular risk without cardiovascular disease, 57% women, in Spain [12].
Sources
[11], [12]
Caveats
Participants knew their diet [11]. In [12], the control group was given advice to reduce fat, and the results were republished after randomization problems were found at some sites. Both trials were in Spain.
Intermittent fastingLimited

Longevity and heart disease

Limited evidence: no pooled trial measured deaths, heart attacks or strokes, and cohort studies linked eating windows of 8 hours or less to more cardiovascular deaths.

Compared with no trial reported these outcomes [5]; US adults whose eating window was 12 to 14 hours [14]; US adults with other eating windows [15]

The numbers and sources
Effect
A Cochrane review of 18 pooled trials with 1,125 participants found no data on all-cause death, cardiovascular death, stroke, heart attack or heart failure [5]. In US survey cohorts, eating windows under 8 hours were linked to higher cardiovascular death against 12 to 14 hours [14], and mortality was lowest at about 11 to 12 hours a day [15]. These cohort results appear under Risks.
Population
Trials of 4 weeks to 6 months [5]; observational cohorts of 19,831 and 33,052 US adults [14][15].
Sources
[5], [14], [15]
Caveats
The cohorts measured habitual eating duration from two 24-hour diet recalls, not a chosen fasting plan [14][15].
Low-fat dietLimitedStudied in women

Heart attack and stroke against usual diet

One large trial in postmenopausal women found no change in coronary heart disease, stroke or total cardiovascular disease over 8.1 years against usual diet.

Compared with diets with more fat

The numbers and sources
Effect
Over a mean of 8.1 years, coronary heart disease (HR 0.97, 95% CI 0.90 to 1.06), stroke (HR 1.02, 95% CI 0.90 to 1.15) and total cardiovascular disease (HR 0.98, 95% CI 0.92 to 1.05) did not differ from the comparison group. By year 6, fat intake was 8.2% of energy lower in the low-fat group [9]. With follow-up through 2010, incidence still did not differ between groups [10].
Population
48,835 postmenopausal women aged 50 to 79 at 40 US centers [9][10].
Sources
[9], [10]
Caveats
Women only. The fat reduction reached was smaller than the 20% of energy target. Subgroup results differed by heart disease and blood pressure at entry (see Risks) [10].
Why limited
This rests on one trial, of 48,835 people. Strength counts independent studies that agree, so one trial is limited at any size.

Lab markers

25-hydroxyvitamin DStrong

Heart disease and stroke

Linked to more cardiovascular deaths at low levels in cohorts, while 21 pooled trials of vitamin D supplements found no fewer cardiovascular events.

The numbers and sources
Effect
Across cohort studies, the bottom third of baseline levels had a relative risk of death from cardiovascular disease of 1.35 against the top third (95% CI 1.13 to 1.61) [9]. In 21 randomized trials, vitamin D supplementation was not associated with fewer major adverse cardiovascular events (RR 1.00, 95% CI 0.95 to 1.06), heart attacks (RR 1.00), strokes (RR 1.06) or cardiovascular deaths (RR 0.98), and results were consistent across starting 25-hydroxyvitamin D levels [15]. In a trial of 2,000 IU a day, major cardiovascular events did not differ from placebo (HR 0.97, 95% CI 0.85 to 1.12) [11].
Population
73 cohort studies [9]. 83,291 participants in 21 trials, mean age 65.8, 74.4% women [15]. 25,871 US adults followed a median 5.3 years [11].
Sources
[9], [15], [11]
Caveats
The cohort finding is observational. Only 4 of the 21 trials had cardiovascular disease as a prespecified primary end point [15].
Apolipoprotein A1 (ApoA1)Strong

Coronary heart disease in cohort studies

Linked to more coronary heart disease at lower levels, about 62% more in the lowest than the highest third across 21 prospective studies.

The numbers and sources
Effect
In a meta-analysis of prospective studies in general populations, people in the bottom third of ApoA1 had a relative risk of coronary heart disease of 1.62 (95% CI 1.43 to 1.83) against those in the top third [8]. In a pooled analysis of individual records, ApoA1 and HDL cholesterol had very similar shape and size of association with coronary heart disease, and the authors concluded that either cholesterol levels or apolipoproteins can be measured for risk assessment [9]. In 27,673 initially healthy US women, the hazard ratio for cardiovascular disease in the top against the bottom fifth of ApoA1 was 0.63 (95% CI 0.52 to 0.77), against 0.52 (0.42 to 0.64) for HDL cholesterol [10]. Among patients on statins in 8 trials, each standard deviation higher ApoA1 went with a hazard ratio for major cardiovascular events of 0.79 (95% CI 0.72 to 0.82), including in those with LDL cholesterol below 50 mg/dL [11].
Population
6,333 coronary heart disease cases in 21 studies [8]. 91,307 participants with 4,499 cases in the 22 studies with apolipoproteins measured, within 302,430 people in 68 studies [9]. 27,673 women followed 11 years, 1,015 events [10]. 38,153 trial participants allocated to a statin, 5,387 with a major cardiovascular event [11].
Sources
[8], [9], [10], [11]
Caveats
Observational, including the analysis within statin trials. The literature-based meta-analysis found heterogeneity between studies that study characteristics only partly explained [8].
Apolipoprotein A1 (ApoA1)Strong

Heart attack and the ApoB to ApoA1 ratio

Linked to more heart attacks at a higher ApoB to ApoA1 ratio, about 59% higher odds per standard deviation in a 52-country case-control study.

The numbers and sources
Effect
In the INTERHEART study, the ApoB to ApoA1 ratio had an odds ratio for acute myocardial infarction of 1.59 per standard deviation (95% CI 1.53 to 1.64) and a population-attributable risk of 54%, against 32% for the ratio of total to HDL cholesterol [15]. In pooled prospective studies, the hazard ratio for coronary heart disease per standard deviation was 1.49 (95% CI 1.39 to 1.60) for the ApoB to ApoA1 ratio and 1.50 (1.38 to 1.62) for the ratio of non-HDL to HDL cholesterol [9]. A literature-based meta-analysis found a relative risk of 1.86 (95% CI 1.55 to 2.22) for the top against the bottom third of the ratio [8]. In a Swedish cohort, the ratio was strongly related to fatal myocardial infarction in men and women, and ApoA1 was noted to be protective [16].
Population
12,461 cases and 14,637 controls in 52 countries, with blood samples from 9,345 cases and 12,120 controls [15]. 91,307 participants in 22 prospective studies [9]. 3,730 coronary cases in 7 studies [8]. 175,553 people followed a mean of about 65 months, with 1,223 fatal myocardial infarctions [16].
Sources
[15], [9], [8], [16]
Caveats
Observational, and the largest study is a case-control study, which cannot show that the ratio was raised before the heart attack [15]. The ratio is a different quantity from ApoA1 and has no unit, so this page places no typed result against it. One Finnish study put its reference interval at 0.3 to 1.0 in men and 0.3 to 0.8 in women [5], and a European guideline describes such ratios as useful for risk estimation and not for diagnosis or as treatment targets [2].
Apolipoprotein B (ApoB)Strong

Heart attack and other cardiovascular events

Linked to more cardiovascular events at higher levels, about 43% more per standard deviation in pooled cohorts, a stronger link than for LDL cholesterol.

The numbers and sources
Effect
A meta-analysis of 12 reports found a standardized relative risk of fatal or nonfatal ischemic cardiovascular events of 1.43 for ApoB (95% CI 1.35 to 1.51), 1.34 for non-HDL cholesterol (1.24 to 1.44) and 1.25 for LDL cholesterol (1.18 to 1.33) [9]. In UK Biobank adults not taking lipid-lowering drugs, ApoB, non-HDL cholesterol and triglycerides were each associated with a first heart attack. When they were assessed together, only ApoB remained associated (adjusted HR 1.27 per standard deviation, 95% CI 1.15 to 1.40), and the same was found in statin-treated trial patients with atherosclerosis [10]. In 68 prospective studies, the hazard ratio for coronary heart disease per standard deviation was 1.49 (95% CI 1.39 to 1.60) for the ratio of ApoB to apolipoprotein A-I and 1.50 (1.38 to 1.62) for the ratio of non-HDL to HDL cholesterol [11].
Population
233,455 people with 22,950 events in 12 reports [9]. 389,529 UK Biobank participants followed a median 11.1 years, and 40,430 trial patients followed a median 2.5 years [10]. 302,430 people without vascular disease in 68 studies, a subset of whom had apolipoproteins measured [11].
Sources
[9], [10], [11]
Caveats
Observational. ApoB, LDL cholesterol and non-HDL cholesterol are highly correlated, which makes their separate contributions hard to tell apart [12]. The largest pooled analysis reported ApoB only as a ratio to apolipoprotein A-I [11].
Apolipoprotein B (ApoB)Strong

Cardiovascular events and death in trials that lowered ApoB

Shown in trials: each 10 mg/dL lowered went with about 7% fewer major cardiovascular events, except with drugs that act outside the LDL receptor.

The numbers and sources
Effect
Across 29 randomized trials, each 10 mg/dL decrease in ApoB went with a relative risk of 0.93 for major adverse cardiovascular events (95% CI 0.90 to 0.97), 0.93 for cardiovascular death (0.88 to 0.98) and 0.95 for death from any cause (0.92 to 0.99). For major events the relative risk was 0.88 (0.83 to 0.93) with statins and 0.96 (0.94 to 0.99) with non-statin drugs that clear ApoB through the LDL receptor. Drugs that lower ApoB by other routes (CETP inhibitors, fibrates, niacin, omega-3 fatty acids) showed no reduction (1.02, 0.81 to 1.30), and the reduction in death from any cause was limited to statins [14]. An earlier analysis of 25 trials found each 10 mg/dL decrease went with 9% less coronary heart disease, 6% less major cardiovascular disease and no decrease in stroke. Across all drug classes, the fall in ApoB did not consistently predict risk better than the fall in LDL or non-HDL cholesterol [15].
Population
332,912 patients in 29 trials of at least 1,000 people and 1 year [14]. 131,134 participants in 25 trials: 12 of statins, 4 of fibrates, 5 of niacin and 4 others [15].
Sources
[14], [15]
Caveats
Both analyses compare averages between trials, not individuals within them. None of the trials behind current treatment thresholds enrolled people by ApoB, and ApoB was often measured in only part of a trial population [2].
Cystatin CStrong

Heart attack, stroke and heart failure in cohort studies

Linked to more cardiovascular disease at higher levels, about 82% more for each doubling of cystatin C across 16 pooled cohorts.

The numbers and sources
Effect
Across 16 prospective cohorts, the relative risk of cardiovascular disease was 1.82 per doubling of cystatin C (95% CI 1.56 to 2.13) after adjustment for age, sex and traditional risk factors [16]. In elderly US adults, the highest fifth of cystatin C (1.29 mg/L or above) against the two lowest (0.99 mg/L or below) had hazard ratios of 2.27 (95% CI 1.73 to 2.97) for cardiovascular death, 1.48 (1.08 to 2.02) for heart attack and 1.47 (1.09 to 1.96) for stroke, while the highest fifth of creatinine was not associated with any of the three [11]. Among elderly adults without chronic kidney disease, each 0.18 mg/L higher went with hazard ratios of 1.28 (1.17 to 1.40) for heart failure, 1.22 (1.08 to 1.38) for stroke and 1.20 (1.06 to 1.36) for heart attack [10]. In 23 outpatient cohorts, a cystatin C-based kidney estimate at least 30% below the creatinine-based one went with hazard ratios of 1.35 (1.27 to 1.44) for atherosclerotic cardiovascular disease and 1.54 (1.40 to 1.68) for heart failure [14].
Population
76,481 participants in 16 cohorts, 37,126 with cystatin C measured [16]. 4,637 elderly adults followed from 1992 or 1993 to 2001 [11]. 4,663 elderly adults, 78% without chronic kidney disease, median follow-up 9.3 years [10]. 821,327 outpatients [14].
Sources
[16], [11], [10], [14]
Caveats
Observational. The genetic analysis in the next block was run to test whether this link is causal and did not support it [16].
Estimated GFR (eGFR)Strong

Death from any cause and from cardiovascular disease

Linked to more deaths at lower eGFR: against 95, risk was 18% higher at 60, 57% higher at 45 and about 3 times higher at 15.

The numbers and sources
Effect
In a collaborative meta-analysis of general population cohorts, risk of death was unrelated to eGFR between 75 and 105 mL/min/1.73 m2 and rose at lower values. Against an eGFR of 95, adjusted hazard ratios for death from any cause were 1.18 (95% CI 1.05 to 1.32) at 60, 1.57 (1.39 to 1.78) at 45 and 3.14 (2.39 to 4.13) at 15, with similar findings for cardiovascular death [6]. In a later individual-participant analysis of 114 cohorts, lower eGFR was significantly associated with higher all-cause and cardiovascular mortality, including in the mildest categories of chronic kidney disease [7].
Population
105,872 participants in 14 studies with urine albumin-to-creatinine measurements and 1,128,310 in 7 studies with dipstick measurements [6]. 27,503,140 individuals in 114 cohorts, mean age 54, mean follow-up 4.8 years [7].
Sources
[6], [7]
Caveats
Observational. Urine albumin predicted death independently of eGFR, so eGFR alone gives part of the picture [6]. KDIGO notes that creatinine-based eGFR has a J-shaped association with most outcomes, with risk also rising above 105 mL/min/1.73 m2, and that very low creatinine often reflects frailty or low muscle mass [3].
Fasting glucoseStrong

Heart disease and stroke

Linked to more coronary heart disease above 100 mg/dL in 102 pooled cohorts, 11% to 17% more below the diabetes range, with no link between 70 and 100 mg/dL.

The numbers and sources
Effect
In 102 prospective studies, fasting glucose was non-linearly related to vascular risk, with no significant association between 3.90 and 5.59 mmol/L (70 to 100 mg/dL). Against that range, hazard ratios for coronary heart disease were 1.07 (95% CI 0.97 to 1.18) below 3.90 mmol/L, 1.11 (1.04 to 1.18) at 5.60 to 6.09 mmol/L and 1.17 (1.08 to 1.26) at 6.10 to 6.99 mmol/L. Diabetes went with a hazard ratio of 2.00 (1.83 to 2.19). In people without known diabetes, adding fasting glucose to conventional risk factors did not significantly improve prediction of vascular disease [8]. In a meta-analysis of 53 cohorts, impaired fasting glucose went with a relative risk of cardiovascular disease of 1.13 by the American definition and 1.26 by the WHO definition, and of stroke of 1.06 and 1.17 [9].
Population
698,782 people without vascular disease in 102 prospective studies, with 52,765 vascular outcomes [8]. 1,611,339 people in 53 prospective cohort studies, median follow-up 9.5 years [9].
Sources
[8], [9]
Caveats
Observational. In the pooled cohorts, hazard ratios for diabetes did not change appreciably after further adjustment for lipid, inflammatory or kidney markers [8].
HDL cholesterol (HDL-C)Strong

Coronary heart disease in cohort studies

Linked to less coronary heart disease at higher levels, about 22% less for each 15 mg/dL higher across 68 pooled cohorts.

The numbers and sources
Effect
In 68 prospective studies, coronary heart disease rates per 1,000 person-years were 6.4 in the bottom third and 2.4 in the top third of HDL cholesterol. The adjusted hazard ratio was 0.78 (95% CI 0.74 to 0.82) per 15 mg/dL higher HDL cholesterol [8]. A separate pooled observational analysis found an odds ratio for heart attack of 0.62 (95% CI 0.58 to 0.66) per standard deviation higher [9].
Population
302,430 people without vascular disease in 68 prospective studies, mostly in Europe and North America [8].
Sources
[8], [9]
Caveats
Observational. Low HDL cholesterol travels with high triglycerides, obesity, smoking and type 2 diabetes, so an association does not show that HDL cholesterol itself is protective [2].
Hemoglobin A1c (HbA1c)Strong

Heart disease and stroke in people without diabetes

Linked to more coronary heart disease from 5.5% upward in adults without diabetes, though adding HbA1c to standard risk factors barely improved risk prediction.

The numbers and sources
Effect
In US adults without diabetes, hazard ratios for coronary heart disease against 5.0 to under 5.5% were 1.23 (95% CI 1.07 to 1.41) at 5.5 to under 6.0%, 1.78 (1.48 to 2.15) at 6.0 to under 6.5% and 1.95 (1.53 to 2.48) at 6.5% or greater, with similar results for stroke [8]. Across 73 prospective studies, the association with cardiovascular disease was approximately J-shaped. Adding HbA1c to conventional risk factors changed the C-index by 0.0018 (95% CI 0.0003 to 0.0033) and gave a net reclassification improvement of 0.42 (-0.63 to 1.48), which the authors called little incremental benefit [9].
Population
11,092 adults [8]. 294,998 participants without known diabetes or cardiovascular disease in 73 studies, median follow-up 9.9 years, 20,840 cardiovascular events [9].
Sources
[8], [9]
Caveats
Observational. The association weakened somewhat after adjustment for HDL cholesterol and C-reactive protein [9].
High-sensitivity C-reactive protein (hs-CRP)Strong

Coronary heart disease and stroke

Linked to more coronary heart disease and stroke at higher levels, about 37% more coronary disease per threefold rise after adjusting for standard risk factors.

The numbers and sources
Effect
In pooled individual records from 54 prospective studies, the risk ratio for coronary heart disease per one standard deviation higher log CRP (a threefold higher level) was 1.63 (95% CI 1.51 to 1.76) adjusted for age and sex, 1.37 (1.27 to 1.48) after conventional risk factors and 1.23 (1.07 to 1.42) after fibrinogen as well. For ischemic stroke the figures were 1.44, 1.27 and 1.32 [6]. In initially healthy US women followed for 30 years, the top fifth of hs-CRP had a hazard ratio for a first major cardiovascular event of 1.70 against the bottom fifth (95% CI 1.52 to 1.90), compared with 1.36 for LDL cholesterol [7]. In statin-treated patients from three trials, the top quarter of hs-CRP had hazard ratios of 1.31 for major cardiovascular events (95% CI 1.20 to 1.43) and 2.68 for cardiovascular death (2.22 to 3.23) against the bottom quarter [8].
Population
160,309 people without vascular disease in 54 studies, 27,769 outcomes [6]. 27,939 women, mean age 54.7, with 3,662 events [7]. 31,245 patients with or at high risk of atherosclerotic disease on statins [8].
Sources
[6], [7], [8]
Caveats
Observational. The pooled analysis found the association depends considerably on conventional risk factors and other markers of inflammation, and that CRP was tied about as strongly to deaths from several cancers and lung disease, which its authors say leaves its relevance unclear [6].
High-sensitivity C-reactive protein (hs-CRP)Strong

Cardiovascular events in trials that selected people by hs-CRP or lowered inflammation

Shown in trials: a drug that lowered hs-CRP without changing cholesterol cut cardiovascular events 15%, while one that did not lower it made no difference.

The numbers and sources
Effect
In people with a previous heart attack and hs-CRP of 2 mg/L or more, canakinumab lowered hs-CRP by 26 to 41 percentage points more than placebo without lowering lipids. At 150 mg it reduced the primary end point of heart attack, stroke or cardiovascular death (HR 0.85, 95% CI 0.74 to 0.98). Fatal infection was more frequent, and death from any cause did not differ (HR 0.94, 95% CI 0.83 to 1.06) [12]. Low-dose methotrexate did not lower C-reactive protein and did not reduce cardiovascular events (HR 0.96, 95% CI 0.79 to 1.16) [13]. In apparently healthy adults with hs-CRP of 2.0 mg/L or higher and LDL cholesterol below 130 mg/dL, rosuvastatin reduced the primary cardiovascular end point (HR 0.56, 95% CI 0.46 to 0.69) while lowering LDL cholesterol by 50% and hs-CRP by 37% [14].
Population
10,061 patients followed a median 3.7 years [12]. 4,786 patients with stable atherosclerosis and diabetes or metabolic syndrome followed a median 2.3 years [13]. 17,802 men and women followed a median 1.9 years [14].
Sources
[12], [13], [14]
Caveats
The statin trial lowered LDL cholesterol and hs-CRP together, and had no group with low hs-CRP, so it cannot show that hs-CRP identified who benefited [14]. Canakinumab acts on interleukin-1 beta, upstream of CRP, so the trial tests lowering inflammation, not lowering CRP itself. The canakinumab trial was funded by its maker [12].
HOMA-IRStrong

Heart disease and other cardiovascular events

Linked to more coronary heart disease in adults without diabetes, 64% more in the highest than the lowest HOMA-IR group across pooled cohorts.

The numbers and sources
Effect
In a meta-analysis of adults without diabetes, the relative risk of coronary heart disease for high against low values was 1.64 (95% CI 1.35 to 2.00) for HOMA-IR, 1.52 (1.31 to 1.76) for fasting glucose and 1.12 (0.92 to 1.37) for fasting insulin. Per standard deviation higher it was 1.46 (1.26 to 1.69) for HOMA-IR, 1.21 (1.13 to 1.30) for glucose and 1.04 (0.96 to 1.12) for insulin [11]. A later meta-analysis of cohorts of healthy adults and adults with metabolic syndrome components found that a higher HOMA-IR went with more non-fatal major cardiovascular events (HR 1.46, 95% CI 1.08 to 1.97) and found no association with fatal ones [12].
Population
516,325 participants in 65 studies of glucose, insulin or HOMA-IR [11]. 215,878 participants in 38 cohort studies [12].
Sources
[11], [12]
Caveats
Observational. Each study compared its own high and low groups, measured with its own insulin assay, so the pooled groups share no common cut point.
LDL cholesterol (LDL-C)Strong

Heart attack, stroke and other artery disease

Linked to more atherosclerotic cardiovascular disease across over 200 cohort, genetic and trial studies, with risk rising with the level and the years of exposure.

The numbers and sources
Effect
A European Atherosclerosis Society consensus panel reviewed separate meta-analyses of over 200 prospective cohort studies, Mendelian randomization studies and randomized trials. It reported a consistent, dose-dependent, log-linear association between LDL cholesterol and atherosclerotic cardiovascular disease that grows with duration of exposure, and concluded that LDL causes the disease [9]. In a pooled analysis of prospective studies, directly measured LDL cholesterol had a hazard ratio for coronary heart disease of 1.38 per 33 mg/dL higher (95% CI 1.09 to 1.73) [10]. In Mendelian randomization analyses, each 1 mmol/L (38.7 mg/dL) lower lifelong LDL cholesterol went with a 54.5% lower risk of coronary heart disease (95% CI 48.8 to 59.5), about 3 times the reduction per unit seen with a statin started later in life [11].
Population
More than 2 million participants with over 150,000 cardiovascular events [9]. 302,430 people without vascular disease in 68 prospective studies, a subset of whom had LDL cholesterol measured directly [10]. 312,321 participants in genetic studies of 9 variants in 6 genes [11].
Sources
[9], [10], [11]
Caveats
The cohort findings are observational. The consensus states that the benefit of lowering holds when the fall in LDL cholesterol is matched by a fall in LDL particle number and the treatment has no offsetting harms [9].
LDL cholesterol (LDL-C)Strong

Cardiovascular events in trials that lowered LDL cholesterol

Shown in trials: each 1 mmol/L (39 mg/dL) lowered cut major vascular events about 22% with statins, and similarly with other drugs acting through the LDL receptor.

The numbers and sources
Effect
In 26 statin trials, each 1.0 mmol/L reduction in LDL cholesterol lowered major vascular events with a rate ratio of 0.78 (95% CI 0.76 to 0.80), including in people whose LDL cholesterol was already below 2 mmol/L [12]. Across 49 trials of 9 kinds of LDL-lowering treatment, the relative risk per 1 mmol/L (38.7 mg/dL) was 0.77 for statins (95% CI 0.71 to 0.84) and 0.75 for diet, bile acid sequestrants, ileal bypass and ezetimibe (95% CI 0.66 to 0.86). CETP inhibitors produced less risk reduction than their LDL lowering predicted (observed 1.01 against an expected 0.90) [13]. In populations starting at a median of 70 mg/dL or less, the relative risk per 1 mmol/L was 0.79 (95% CI 0.71 to 0.87) for statins and non-statins combined [14].
Population
About 170,000 participants in 26 trials, median follow-up about 5 years [12]. 312,175 participants in 49 trials [13]. A low-LDL subgroup of the statin trials plus 50,627 patients in 3 trials of non-statin drugs added to statins [14].
Sources
[12], [13], [14]
Caveats
Effects are expressed per unit of LDL cholesterol lowered, over about 5 years of treatment. The same fall in LDL cholesterol did not bring the same benefit with every drug class [13].
Lipoprotein(a)Strong

Heart attack and stroke in cohort studies

Linked to more heart attacks and strokes at higher levels, about 11% more cardiovascular disease for each 50 nmol/L higher in 460,506 UK adults.

The numbers and sources
Effect
In 36 prospective studies, the risk ratio for coronary heart disease was 1.13 (95% CI 1.09 to 1.18) per 3.5-fold higher usual lipoprotein(a) after adjustment for lipids and other risk factors, and 1.10 (1.02 to 1.18) for ischemic stroke. It was 1.01 (0.98 to 1.05) for death from nonvascular causes [10]. In the UK Biobank, the relationship with atherosclerotic cardiovascular disease appeared linear, with a hazard ratio of 1.11 (95% CI 1.10 to 1.12) per 50 nmol/L. Median levels were 19, 31, 75 and 16 nmol/L in white, South Asian, black and Chinese participants, and the risk per 50 nmol/L was similar across groups. A level of 150 nmol/L or above was found in 12.2% of people without cardiovascular disease and went with a hazard ratio of 1.50 (95% CI 1.44 to 1.56) [11].
Population
126,634 participants in 36 prospective studies, 1.3 million person-years [10]. 460,506 middle-aged UK Biobank participants, median follow-up 11.2 years, 22,401 events [11].
Sources
[10], [11]
Caveats
Observational. The pooled cohorts used many different assays [10]. A European panel states that risk rises continuously, with no level at which it begins [4].
Lipoprotein(a)Strong

Coronary heart disease in genetic studies

Shown in genetic studies to track coronary heart disease: variants that raise lipoprotein(a) raised risk, about 22% more heart attacks for each doubling.

The numbers and sources
Effect
In Danish studies, genetically elevated lipoprotein(a) went with a hazard ratio for heart attack of 1.22 (95% CI 1.09 to 1.37) per doubling, against 1.08 (1.03 to 1.12) for measured levels [12]. Two variants at the LPA gene had odds ratios for coronary disease of 1.70 and 1.92. Carrying one variant gave an odds ratio of 1.51 (95% CI 1.38 to 1.66) and two or more 2.57 (1.80 to 3.67), and the association disappeared after adjustment for the lipoprotein(a) level [13]. In a Mendelian randomization analysis, each 10 mg/dL lower genetically predicted lipoprotein(a) went with 5.8% lower coronary heart disease risk (OR 0.942, 95% CI 0.933 to 0.951). A difference of 101.5 mg/dL (95% CI 71.0 to 137.0) carried the same risk difference as 38.67 mg/dL of LDL cholesterol [14].
Population
8,637, 29,388 and 2,461 white Danish participants in three studies [12]. 3,145 cases and 3,352 controls, replicated in 4,846 cases and 4,594 controls [13]. 20,793 coronary heart disease cases and 27,540 controls with individual data, plus summary data on 62,240 cases and 127,299 controls, all of European ancestry [14].
Sources
[12], [13], [14]
Caveats
Genetic studies estimate the effect of lifelong differences, which may not match what lowering later in life does. Participants were of European ancestry [12][14].
Magnesium levelStrong

Heart and artery disease

Linked to less cardiovascular disease at higher levels, 30% lower risk per 0.2 mmol/L (0.49 mg/dL) across pooled cohorts, with genetic evidence pointing the same way for coronary disease.

The numbers and sources
Effect
In a meta-analysis of prospective studies, each 0.2 mmol/L (0.49 mg/dL) higher circulating magnesium was associated with a 30% lower risk of cardiovascular disease (RR 0.70, 95% CI 0.56 to 0.88), with trends toward lower ischemic heart disease (RR 0.83, 95% CI 0.75 to 1.05) and fatal ischemic heart disease (RR 0.61, 95% CI 0.37 to 1.00) [7]. A second meta-analysis found a relative risk of coronary heart disease of 0.86 (95% CI 0.74 to 0.996) comparing the highest with the lowest category, with no significant linear trend [8]. In a Mendelian randomization study, each 0.1 mmol/L higher genetically predicted serum magnesium had an odds ratio for coronary artery disease of 0.88 (95% CI 0.78 to 0.99) [9].
Population
313,041 people in 16 prospective studies, with 11,995 cardiovascular disease events [7]. 38,808 people in 5 cohorts, average follow-up 10.5 years [8]. 60,801 coronary artery disease cases and 123,504 non-cases in a genetic consortium analysis of 48 studies [9].
Sources
[7], [8], [9]
Caveats
The cohort findings are observational, and the second meta-analysis reports that its results were sensitive to the exclusion of individual studies [8]. The authors of each study call for trials of magnesium to test whether raising the level lowers risk [7][9].
Non-HDL cholesterol (non-HDL-C)Strong

Heart attack, stroke and other cardiovascular disease in cohort studies

Linked to more cardiovascular disease at higher levels, about 50% more coronary heart disease for each 43 mg/dL higher across 68 pooled cohorts.

The numbers and sources
Effect
In 68 prospective studies, coronary heart disease rates per 1,000 person-years were 2.3 in the bottom third and 6.7 in the top third of non-HDL cholesterol. The adjusted hazard ratio per 43 mg/dL higher was 1.50 (95% CI 1.39 to 1.61) for coronary heart disease and 1.12 (1.04 to 1.20) for ischemic stroke [8]. In a consortium of 38 cohorts, 30-year cardiovascular event rates rose from 7.7% at non-HDL cholesterol below 2.6 mmol/L (about 100 mg/dL) to 33.7% at 5.7 mmol/L (about 220 mg/dL) or above in women, and from 12.8% to 43.6% in men [9]. A meta-analysis of 9 prospective studies found a relative risk of coronary heart disease of 1.79 (95% CI 1.68 to 1.91) for the highest against the lowest category [10]. Across 12 reports, the standardized relative risk of ischemic cardiovascular events was 1.34 (95% CI 1.24 to 1.44) for non-HDL cholesterol, between that of LDL cholesterol (1.25, 1.18 to 1.33) and ApoB (1.43, 1.35 to 1.51) [11].
Population
302,430 people without vascular disease in 68 prospective studies [8]. 398,846 people in 38 cohorts, median follow-up 13.5 years, 54,542 cardiovascular events [9]. 448,732 people in 9 prospective studies [10]. 233,455 people with 22,950 events in 12 reports [11].
Sources
[8], [9], [10], [11]
Caveats
Observational. In the pooled cohorts the hazard ratios were at least as strong in participants who did not fast as in those who did [8]. Each study in the category comparison defined its own highest and lowest groups [10].
Non-HDL cholesterol (non-HDL-C)Strong

Cardiovascular events in trials that lowered non-HDL cholesterol

Shown in trials: each 1 mmol/L (39 mg/dL) of non-HDL cholesterol lowered went with about 21% fewer major vascular events across 49 trials.

The numbers and sources
Effect
In a meta-regression of 24 trials of non-statin drugs and 25 statin trials, the risk ratio for major vascular events was 0.79 (95% CI 0.76 to 0.82) per 1 mmol/L reduction in non-HDL cholesterol [12]. In an earlier analysis of 30 trials, each 1% decrease in non-HDL cholesterol with a statin went with a 4.5-year relative risk of coronary heart disease of 0.99 (95% Bayesian interval 0.98 to 1.00), and the fibrate trials did not differ from the statin trials [13]. Across 7 placebo-controlled statin trials, coronary risk fell 20.0% (95% CI 15.2 to 24.7) per standard deviation decrease in non-HDL cholesterol, 20.1% for LDL cholesterol and 24.4% for ApoB [14]. Among patients on statins in 8 trials, each standard deviation higher on-treatment non-HDL cholesterol went with a hazard ratio of 1.16 (95% CI 1.12 to 1.19) for major cardiovascular events, against 1.13 (1.10 to 1.17) for LDL cholesterol, a significant difference [15].
Population
374,358 patients in 49 trials, with 46,180 major cardiovascular events [12]. 14 statin trials (100,827 people), 7 fibrate trials (21,647), 6 niacin trials (4,445) and 3 others [13]. 7 placebo-controlled statin trials [14]. 62,154 patients in 8 statin trials, of whom 38,153 were allocated to a statin [15].
Sources
[12], [13], [14], [15]
Caveats
The meta-regressions compare trials with each other, not people within a trial, and cover drugs that lower non-HDL cholesterol in different ways [12][13]. The niacin trials did not follow the pattern of the statin trials, with inconsistent results in five small trials of niacin combined with other drugs [13].
Total cholesterolStrong

Death from coronary heart disease and total cholesterol

Linked to more coronary heart disease deaths at higher levels: each 1 mmol/L (39 mg/dL) lower total cholesterol went with a third fewer at ages 50 to 69.

The numbers and sources
Effect
In a meta-analysis of 61 prospective studies, each 1 mmol/L lower usual total cholesterol went with lower death from ischemic heart disease in both sexes: hazard ratio 0.44 (95% CI 0.42 to 0.48) at ages 40 to 49, 0.66 (0.65 to 0.68) at ages 50 to 69 and 0.83 (0.81 to 0.85) at ages 70 to 89, with no apparent threshold. The ratio of total to HDL cholesterol predicted these deaths more than twice as informatively as total cholesterol alone [5].
Population
Almost 900,000 adults without previous disease in 61 prospective studies, mostly in western Europe or North America, with more than 55,000 vascular deaths [5].
Sources
[5]
Caveats
Observational. The proportional difference in risk was smaller at older ages and at higher blood pressure [5].
Total cholesterolStrong

Cardiovascular disease and non-HDL cholesterol

Linked to more cardiovascular disease at higher non-HDL cholesterol, about 50% more coronary heart disease for each 43 mg/dL higher across 68 pooled cohorts.

The numbers and sources
Effect
In 68 prospective studies, coronary heart disease rates per 1,000 person-years were 2.3 in the bottom third and 6.7 in the top third of non-HDL cholesterol, and the adjusted hazard ratio was 1.50 (95% CI 1.39 to 1.61) per 43 mg/dL higher [6]. In a consortium of 38 cohorts, 30-year cardiovascular event rates rose from 7.7% at non-HDL cholesterol below 2.6 mmol/L (about 100 mg/dL) to 33.7% at 5.7 mmol/L (about 220 mg/dL) or above in women, and from 12.8% to 43.6% in men. Adjusted hazard ratios for the highest against the lowest category were 1.9 (95% CI 1.6 to 2.2) in women and 2.3 (2.0 to 2.5) in men [7].
Population
302,430 people without vascular disease in 68 prospective studies [6]. 398,846 people in 38 cohorts, median follow-up 13.5 years, 54,542 cardiovascular events [7].
Sources
[6], [7]
Caveats
Observational. In the pooled cohorts the hazard ratios were at least as strong in participants who did not fast as in those who did [6].
Total cholesterolStrong

Cardiovascular events in trials that lowered non-HDL cholesterol

Shown in trials: each 1 mmol/L (39 mg/dL) of non-HDL cholesterol lowered went with about 21% fewer major vascular events across 49 trials.

The numbers and sources
Effect
In a meta-regression of 24 trials of non-statin drugs and 25 statin trials, the risk ratio for major vascular events was 0.79 (95% CI 0.76 to 0.82) per 1 mmol/L reduction in non-HDL cholesterol [8]. Among patients on statins in 8 trials, each standard deviation higher on-treatment non-HDL cholesterol went with a hazard ratio of 1.16 (95% CI 1.12 to 1.19) for major cardiovascular events, against 1.13 (1.10 to 1.17) for LDL cholesterol, a significant difference [9]. Statin trials found that each 1.0 mmol/L reduction in LDL cholesterol lowered major vascular events with a rate ratio of 0.78 (95% CI 0.76 to 0.80) [10].
Population
374,358 patients in 49 trials, with 46,180 major cardiovascular events [8]. 62,154 patients in 8 statin trials, of whom 38,153 were allocated to a statin [9]. About 170,000 participants in 26 statin trials [10].
Sources
[8], [9], [10]
Caveats
The meta-regression compares trials with each other, not people within a trial, and covers drugs that lower non-HDL cholesterol in different ways [8].
Transferrin saturationStrong

Heart failure hospital stays in trials of intravenous iron

Shown in trials: intravenous iron cut heart failure hospital stays in iron-deficient patients, possibly more when saturation was below 20%, with no clear effect on deaths.

The numbers and sources
Effect
In a meta-analysis of 10 trials, intravenous iron lowered the combined rate of repeat heart failure hospital stays and cardiovascular death (rate ratio 0.75, 95% CI 0.61 to 0.93). Effects on cardiovascular death (odds ratio 0.86, 0.70 to 1.05) and death from any cause (0.93, 0.78 to 1.12) were inconclusive. Patients with a transferrin saturation below 20% had an odds ratio of 0.67 (0.49 to 0.92) for the main outcome, against 0.99 (0.74 to 1.30) at 20% or above (P for difference 0.07) [13]. In pooled patient-level data from 3 trials of one iron product, the rate ratio for cardiovascular hospital stays and cardiovascular death was 0.86 (0.75 to 0.98), with no apparent effect on death. That analysis reported a lower risk of cardiovascular death in the subgroup with saturation below 15% and possibly unfavorable effects at 24% or above [14].
Population
3,373 people with heart failure and iron deficiency in 10 randomized trials [13]. 4,501 people in 3 placebo-controlled trials of ferric carboxymaltose with at least 52 weeks of follow-up [14].
Sources
[13], [14]
Caveats
The findings by saturation come from subgroup analyses, and the test for a difference between subgroups did not reach significance [13]. The trials enrolled people with heart failure who met a definition of iron deficiency, so they say nothing about iron given to other people with a low saturation.
Urine albumin to creatinine ratioStrong

Death from any cause and cardiovascular death

Linked to more deaths at higher ratios with no threshold, 63% more at about 30 mg/g than at about 5 mg/g across 105,872 people.

The numbers and sources
Effect
In a meta-analysis of general population cohorts, the ratio was related to death on a continuous scale without a threshold. Against a ratio of 0.6 mg/mmol (about 5 mg/g), the adjusted hazard ratio for death from any cause was 1.20 (95% CI 1.15 to 1.26) at 1.1 mg/mmol (10 mg/g), 1.63 (1.50 to 1.77) at 3.4 mg/mmol (about 30 mg/g) and 2.22 (1.97 to 2.51) at 33.9 mg/mmol (about 300 mg/g). Findings were similar for cardiovascular death, and the ratio and estimated GFR each added risk without interaction [5]. In a later meta-analysis of 114 cohorts, higher ratios were associated with higher rates of each of 10 outcomes studied: kidney failure, death from any cause, cardiovascular death, acute kidney injury, hospitalization, coronary heart disease, stroke, heart failure, atrial fibrillation and peripheral artery disease [8].
Population
105,872 participants in 14 general population cohorts with ratio measurements [5]. 9,067,753 people with albuminuria data in 114 cohorts, with a median ratio of 11 mg/g in the main population [8].
Sources
[5], [8]
Caveats
Observational, with the ratio measured once at baseline. The first meta-analysis printed its ratios in mg/mmol, and the mg/g figures other than 10 are converted here with the factor on this page [5].
White blood cell count (WBC)Strong

Coronary heart disease

Linked to more coronary heart disease at higher counts, about 40% more in the top third than the bottom third across pooled prospective studies.

The numbers and sources
Effect
In a meta-analysis of the 7 largest prospective studies, people in the top third of white cell count had a risk ratio for coronary heart disease of 1.4 (95% CI 1.3 to 1.5) against the bottom third, a difference of 8.4 against 5.6 thousand/µL [8]. A later meta-analysis by cell type found the neutrophil count had the strongest association (risk ratio 1.33, 95% CI 1.17 to 1.50, top against bottom third), with substantial differences between studies. In its own US cohort the figure for neutrophils was 1.09 (0.93 to 1.29) [9].
Population
5,337 coronary heart disease cases in 7 long-term prospective studies [8]. 1,764 cases in 5 prospective studies, including 4,625 US adults followed 18 years [9].
Sources
[8], [9]
Caveats
Observational. The authors of the first analysis state that the mechanisms are not clear and call for further study of whether the association is causal [8].
Absolute neutrophil count (ANC)Moderate

Cardiovascular disease

Linked to more heart failure, heart attack and peripheral arterial disease at higher counts within the usual range, in two large United Kingdom cohorts.

The numbers and sources
Effect
In linked English health records, a count of 6,000 to 7,000 cells/µL against 2,000 to 3,000 had adjusted hazard ratios of 2.04 (95% CI 1.82 to 2.29) for heart failure, 1.95 (1.72 to 2.21) for peripheral arterial disease, 1.78 (1.51 to 2.10) for unheralded coronary death, 1.58 (1.42 to 1.76) for nonfatal heart attack and 1.36 (1.17 to 1.57) for ischemic stroke. There was no association with stable angina or bleeding in the brain. The associations were linear, with higher risk already at 3,000 to 4,000 [12]. In UK Biobank, the highest tenth of neutrophil count against the fifth had hazard ratios for cardiovascular death of 1.59 (1.22 to 2.08) in men and 2.15 (1.38 to 3.35) in women. In an analysis that left out people with existing illness and the first two years, each 1,000 cells/µL higher count had a hazard ratio of 1.07 (1.04 to 1.11) in both sexes combined [13].
Population
775,231 adults aged 30 or older with no history of cardiovascular disease, 55,004 cases over a median 3.8 years [12]. 478,259 UK Biobank participants followed a median 7.0 years [13].
Sources
[12], [13]
Caveats
Observational. Smoking, infection and long-term illness raise the count and also raise cardiovascular risk, and adjustment may not remove this. The UK Biobank authors call for trials of treatments that lower neutrophils to investigate causality [13].
AlbuminModerate

Heart disease and stroke

Linked to more cardiovascular disease at lower albumin, about 85% more for each 1 g/dL lower in a meta-analysis of 14 studies.

The numbers and sources
Effect
In a Danish general population cohort, the hazard ratio for each 10 g/L (1 g/dL) lower albumin was 1.17 (95% CI 1.08 to 1.28) for ischemic heart disease, 1.25 (1.09 to 1.43) for heart attack, 1.37 (1.21 to 1.54) for any stroke and 1.46 (1.28 to 1.68) for ischemic stroke. In the accompanying meta-analysis, the risk ratio for a cardiovascular event per 10 g/L lower albumin was 1.85 (95% CI 1.39 to 2.47), with the Danish cohort carrying 57% of the weight [10].
Population
100,520 people without prior cardiovascular disease followed for a median of 8.5 years, and 150,652 individuals in 14 studies [10].
Sources
[10]
Caveats
Observational. The estimates weakened after adjustment for C-reactive protein, and the authors write that the association is partly explained by albumin falling during inflammation [10].
Alkaline phosphatase (ALP)Moderate

Heart disease and stroke

Linked to slightly more cardiovascular disease, 8% more per standard deviation of log alkaline phosphatase in a meta-analysis of 29 cohorts.

The numbers and sources
Effect
The pooled, fully adjusted relative risk of cardiovascular disease was 1.08 (95% CI 1.03 to 1.14) per standard deviation of log baseline alkaline phosphatase, with a log-linear relation [11]. A second meta-analysis found a linear association with cardiovascular deaths (RR 1.02, 95% CI 1.01 to 1.04) [9].
Population
29 cohorts with over 1.23 million participants and 20,406 cardiovascular outcomes, which also covered other liver enzymes [11]. 24 studies with 147,634 participants [9].
Sources
[11], [9]
Caveats
Observational. The authors of the second meta-analysis describe the relationship between alkaline phosphatase and cardiovascular disease as still ambiguous [9].
Apolipoprotein B (ApoB)Moderate

Coronary heart disease in genetic studies

Shown to track coronary heart disease in genetic studies, with about 23% lower odds for each 10 mg/dL lower lifelong ApoB.

The numbers and sources
Effect
In Mendelian randomization analyses, gene variants that lower triglycerides (in LPL) and gene variants that lower LDL cholesterol (in LDLR) were associated with the same lower odds of coronary heart disease per 10 mg/dL lower ApoB: OR 0.771 (95% CI 0.741 to 0.802) and 0.773 (0.747 to 0.801). After adjusting for ApoB, the associations of triglycerides and of LDL cholesterol with coronary heart disease became null [13]. A second multivariable analysis found that only ApoB kept a robust effect (OR 1.92 per standard deviation higher, 95% CI 1.31 to 2.81), while the LDL cholesterol estimate reversed (OR 0.85, 95% CI 0.57 to 1.27) [12].
Population
654,783 participants in 63 cohort or case-control studies, including 91,129 with coronary heart disease [13]. Genetic data from up to 441,016 UK Biobank participants, applied to 60,801 coronary cases and 123,504 controls [12].
Sources
[13], [12]
Caveats
Gene variants act from birth, so these estimates describe lifelong differences, not a change made later in life. The authors of one analysis note that lipid traits are highly correlated, which limits conclusions about any single one [12].
Blood urea nitrogen (BUN)Moderate

Death in heart failure and after a heart attack

Linked to more deaths in people hospitalized with heart failure, about 2.3 times the risk in the top quarter of BUN in one cohort.

The numbers and sources
Effect
In 541 patients admitted for decompensated heart failure, the risk of death rose with each quartile of BUN, with an adjusted relative risk of 2.3 (95% CI 1.3 to 4.1) in the top against the bottom quartile [6]. In a US registry of heart failure admissions, a BUN of 43 mg/dL or above on admission was the single strongest predictor of death in hospital among 39 variables examined [4]. In older Medicare patients, adjusted 1-year mortality rose 3% per 5 mg/dL of BUN after a heart attack and 7% per 5 mg/dL in heart failure, with risk climbing steadily above 17 mg/dL and 16 mg/dL [7].
Population
541 patients, mean age 63, followed for a mean of 343 days [6]. 33,046 hospitalizations in the derivation cohort and 32,229 in the validation cohort, mean age 72.5 [4]. 44,437 patients hospitalized for heart attack and 56,652 for heart failure, aged 65 or older [7].
Sources
[6], [4], [7]
Caveats
Observational, in people already in hospital with heart disease. The authors of the first study suggest that a raised BUN in this setting reflects reduced blood flow to the kidneys from the heart failure itself [6]. These cut points were derived to predict death in hospital patients and are not reference ranges.
Calcium levelModerate

Death and cardiovascular disease

Linked to slightly more deaths and cardiovascular disease at higher calcium, 13% more deaths per standard deviation, in pooled studies of people without a calcium disorder.

The numbers and sources
Effect
In a systematic review of studies in people selected neither for kidney disease nor for abnormal calcium, 5 of 11 studies reported a significant positive association between serum calcium and death, and 8 of 13 between serum calcium and cardiovascular disease. Pooled across 8 studies each, the hazard ratio per standard deviation of serum calcium was 1.13 (95% CI 1.09 to 1.18) for death and 1.08 (1.04 to 1.13) for cardiovascular disease. With hazard ratios adjusted for cardiovascular risk factors, the pooled figure for cardiovascular disease was 1.04 (1.01 to 1.08) [5].
Population
Longitudinal studies with at least 1 year of follow-up, located in a search to December 2014 [5].
Sources
[5]
Caveats
Observational. The authors state that these data cannot determine causality [5]. The abstract gives no total head count.
Calcium levelModerate

Coronary artery disease in a genetic study

Shown in a genetic study to go with more coronary artery disease, 25% higher odds for each 0.5 mg/dL of genetically higher calcium.

The numbers and sources
Effect
Using 6 gene variants related to serum calcium, the odds ratio per 0.5 mg/dL of genetically predicted higher calcium, about 1 standard deviation, was 1.25 (95% CI 1.08 to 1.45) for coronary artery disease and 1.24 (1.05 to 1.46) for heart attack [6].
Population
184,305 individuals, 60,801 with coronary artery disease and 123,504 without [6].
Sources
[6]
Caveats
Mendelian randomization. The 6 variants explained about 0.8% of the variation in serum calcium. The estimate describes lifelong genetic exposure, and the authors state that whether it carries over to shorter exposures is unknown [6].
Copper levelModerate

Stroke, heart attack and cardiovascular death at higher levels

Linked to more strokes and cardiovascular deaths at higher serum copper in two meta-analyses of cohorts, while a genetic study pointed the other way for heart disease.

The numbers and sources
Effect
In a meta-analysis of 17 cohort studies, higher circulating copper went with more stroke (OR 1.52, 95% CI 1.30 to 1.78), coronary artery disease death (RR 2.77, 1.82 to 4.19) and cardiovascular death (RR 1.79, 1.52 to 2.11), and not with acute myocardial infarction (RR 2.01, 0.63 to 6.47). Per 20 µg/dL higher copper, the odds ratio for stroke was 1.23 (1.14 to 1.33) and the hazard ratio for cardiovascular death was 1.28 (1.07 to 1.53) [7]. A second meta-analysis, comparing the highest with the lowest category, found pooled odds ratios of 1.49 (1.22 to 1.82) for stroke, 1.31 (1.17 to 1.46) for myocardial infarction and 1.60 (1.39 to 1.86) for cardiovascular death [8]. A Mendelian randomization study found the reverse direction: genetically predicted copper was associated with less ischemic heart disease (OR 0.94, 95% CI 0.90 to 0.98) [9].
Population
47,813 people in 17 cohort studies [7]. 41,322 participants in 16 observational studies [8]. Up to 76,014 cases of ischemic heart disease and 264,785 controls [9].
Sources
[7], [8], [9]
Caveats
Observational, and the two meta-analyses share studies and disagree on myocardial infarction. Serum copper rises with inflammation because ceruloplasmin is an acute phase protein [3][4], so a higher level may mark inflammation already present. The genetic study used variants tied to copper in red blood cells, not serum copper, and its authors call for further investigation [9]. No trial that lowered serum copper and counted cardiovascular events was located.
Estimated GFR (eGFR)Moderate

Heart disease, stroke, heart failure and hospital admission

Linked to more coronary heart disease, stroke, heart failure, atrial fibrillation and hospital admissions at lower eGFR in an analysis of 114 cohorts.

The numbers and sources
Effect
In an individual-participant meta-analysis, lower eGFR was significantly associated with higher risk of each of 10 outcomes studied, among them coronary heart disease, stroke, heart failure, atrial fibrillation, peripheral artery disease and any hospitalization. Among people with a urine albumin-to-creatinine ratio below 10 mg/g, an eGFR of 45 to 59 against 90 to 104 mL/min/1.73 m2 went with more hospitalizations (adjusted HR 1.3, 95% CI 1.2 to 1.3), 161 against 79 events per 1,000 person-years [7].
Population
27,503,140 individuals from 114 global cohorts with creatinine-based eGFR, and 720,736 from 20 cohorts with eGFR based on creatinine and cystatin C, from 1980 to 2021 [7].
Sources
[7]
Caveats
Observational and retrospective. The analysis used the 2021 CKD-EPI equations. Confounding by age and other illness may remain after adjustment [7].
Fasting insulin and HOMA-IRModerate

High blood pressure and stroke

Linked to more new high blood pressure, 63% more in the highest against the lowest fasting insulin group across 10 cohorts, and not linked to stroke.

The numbers and sources
Effect
Comparing the highest with the lowest group of fasting insulin, the pooled relative risk was 1.63 (95% CI 1.35 to 1.97) for hypertension and 1.18 (0.87 to 1.60) for stroke. Each 50 pmol/L higher fasting insulin went with a relative risk of 1.25 (1.14 to 1.36) for hypertension and 0.999 (0.99 to 1.01) for stroke [11].
Population
36,617 people and 4,491 cases in 10 studies of hypertension, and 27,887 people and 1,550 cases in 7 studies of stroke [11].
Sources
[11]
Caveats
Observational, from one meta-analysis, so confounding by body weight and other factors cannot be ruled out.
Free T4 (free thyroxine)Moderate

Atrial fibrillation

Linked to more atrial fibrillation at higher free T4 inside the reference range: 45% higher risk in the top quarter than the bottom quarter in pooled cohorts.

The numbers and sources
Effect
Among people with a TSH of 0.45 to 4.49 mIU/L in 11 pooled cohorts, the hazard ratio for new atrial fibrillation was 1.45 (95% CI 1.26 to 1.66) for the highest quartile of free T4 against the lowest, with a significant trend across quartiles. TSH itself was not associated [7]. In older US adults whose thyroid tests were all in the reference range, higher free T4 was associated with atrial fibrillation (P below 0.001), and the top quartile had 11.0 more cases per 1,000 person-years than the bottom quartile [8].
Population
30,085 adults in 11 prospective cohorts, 2,574 new cases of atrial fibrillation [7]. 2,843 older adults in the Cardiovascular Health Study not taking thyroid medication [8].
Sources
[7], [8]
Caveats
Observational. Quartiles were set within each cohort, so the finding is not tied to one ng/dL value [7].
Free T4 (free thyroxine)Moderate

Cardiovascular disease and death

Linked in a J shape: risk of cardiovascular disease and death was lowest at the 20th to 40th percentiles of free T4, and the top fifth had 34% more deaths.

The numbers and sources
Effect
In pooled data on 134,346 adults, risk was lowest at the 20th to 40th percentiles of free T4, medians of 13.5 to 14.8 pmol/L. Against that band, the top fifth had hazard ratios of 1.20 (95% CI 1.11 to 1.31) for cardiovascular events or death combined, 1.34 (1.20 to 1.49) for death from any cause, 1.57 (1.31 to 1.89) for cardiovascular death and 1.22 (1.11 to 1.33) for cardiovascular events [9]. In older US adults with thyroid tests in the reference range, the top quartile of free T4 had more coronary heart disease, heart failure and deaths than the bottom quartile, by 8.0, 7.8 and 14.3 per 1,000 person-years [8].
Population
134,346 adults in 26 cohorts, median age 59 [9]. 2,843 older adults [8].
Sources
[9], [8]
Caveats
Observational, so illness that shifts free T4 before an event cannot be ruled out. The pooled figures are adjusted for age and sex, and the percentile medians are printed in pmol/L [9].
FructosamineModerate

Heart attack, stroke and death from any cause

Linked to more heart attacks and deaths at high levels in US and Swedish cohorts, and to slightly more deaths at the lowest levels.

The numbers and sources
Effect
In a US cohort followed for two decades, elevated fructosamine and glycated albumin were each significantly associated with coronary heart disease, ischemic stroke, heart failure and death after adjustment for cardiovascular risk factors. The associations were similar in size to those for HbA1c and tended to be J-shaped, with raised risk at the lowest levels as well [7]. In a Swedish cohort, the hazard ratio in people above 2.70 mmol/L on that cohort's scale, against its reference range, was 2.88 (95% CI 2.70 to 3.07) for heart attack and 2.31 (2.21 to 2.41) for death from any cause, and the associations weakened but remained after adjustment for glucose [8]. Among Swedish adults without diabetes, the lowest tenth of fructosamine had a hazard ratio for death of 1.20 against the middle eight tenths, falling to 1.11 (95% CI 1.07 to 1.15) after adjustment for six other blood markers [9].
Population
11,104 adults with and without diabetes, with 1,096 new cases of coronary heart disease, 605 ischemic strokes, 1,432 cases of heart failure and 2,860 deaths [7]. 338,443 people followed 19 years on average, with 21,526 heart attacks and 73,458 deaths [8]. 215,011 people without diabetes, 41,388 deaths over 25 years [9].
Sources
[7], [8], [9]
Caveats
Observational. The Swedish cohort reported fructosamine in mmol/L on a scale about ten times this page's, so its cut points are not placed here. Its authors found that smoking and chronic inflammation at least partly explained the higher mortality at low fructosamine [9]. No source here tested whether lowering fructosamine changes these outcomes, and a laboratory guideline states that treatment goals have not been established [2].
Gamma-glutamyl transferase (GGT)Moderate

Heart disease and stroke

Linked to more cardiovascular disease, 23% higher risk per standard deviation of log GGT in a meta-analysis of 29 cohorts.

The numbers and sources
Effect
In a meta-analysis of 29 prospective cohorts, the fully adjusted relative risk of cardiovascular disease was 1.23 (95% CI 1.16 to 1.29) per standard deviation of log baseline GGT, in a log-linear pattern [9]. In US adults, GGT above 51 U/L in men or 33 U/L in women was not significantly associated with cardiovascular death (HR 1.3, 95% CI 0.80 to 2.0) [4].
Population
Over 1.23 million participants and 20,406 cardiovascular outcomes in 29 cohorts, which also covered other liver enzymes [9]. 14,950 US adults [4].
Sources
[9], [4]
Caveats
Observational. The cohorts cannot separate GGT from alcohol intake, body weight and other factors that move with it, beyond the adjustments each study made.
HematocritModerate

Death from any cause

Linked to more deaths at both low and high hematocrit in an Iranian cohort of 49,983 adults and in the Framingham cohort.

The numbers and sources
Effect
In 49,983 Iranian adults followed for a mean of 5 years, the relation between hematocrit and death was U-shaped in both sexes. In women, against 35% to 39%, adjusted hazard ratios for death were 2.12 (95% CI 1.26 to 3.58) at 15% to 29%, 1.58 (1.29 to 1.93) at 45% to 49% and 2.05 (1.17 to 3.60) at 50% to 54%. In men, against 40% to 44%, they were 2.12 (1.30 to 3.45) at 15% to 29%, 1.64 (1.22 to 2.21) at 30% to 34% and 1.29 (1.04 to 1.61) at 50% to 54%. Deaths from cardiovascular disease followed the same pattern [7]. In the Framingham cohort over 34 years, people in the highest fifth of hematocrit had more deaths from any cause and more cardiovascular disease, and women with lower values also had more cardiovascular events, which the authors describe as a J-shaped or U-shaped relation [8].
Population
49,983 adults aged 40 to 75 in the Golestan Cohort Study in northeast Iran, 2,262 deaths [7]. 5,209 men and women in the Framingham cohort [8].
Sources
[7], [8]
Caveats
Observational. In the Iranian cohort, leaving out the first 2 years of follow-up weakened the association at low hematocrit in both sexes and removed the association at high hematocrit in men, which points to illness before death as part of the explanation. That cohort measured hematocrit by spinning blood in a tube, not by analyzer [7]. The Framingham abstract gives no hazard ratios [8].
LDL particle number (LDL-P)Moderate

Heart attack, stroke and other cardiovascular disease in cohort studies

Linked to more cardiovascular disease at higher levels, about 2.5 times the rate in the top against the bottom fifth of 27,673 US women.

The numbers and sources
Effect
In initially healthy US women, the adjusted hazard ratio for cardiovascular disease in the top against the bottom fifth was 2.51 (95% CI 1.91 to 3.30) for LDL particle number, 2.57 (1.98 to 3.33) for ApoB, 2.52 (1.95 to 3.25) for non-HDL cholesterol and 1.74 (1.40 to 2.16) for LDL cholesterol. Adding LDL particle number to a model with the total to HDL cholesterol ratio and other risk factors reclassified essentially no one, and the authors called its prediction comparable but not superior to standard lipids or apolipoproteins [4]. In a multi-ethnic US cohort, the hazard ratio per standard deviation was 1.32 (95% CI 1.19 to 1.47) for LDL particle number and 1.20 (1.08 to 1.34) for LDL cholesterol [5]. In a later analysis of that cohort, the top against the bottom quarter had a hazard ratio for coronary heart disease of 1.77 (95% CI 1.21 to 2.58), which weakened after adjustment for the standard lipid panel [8]. A laboratory medicine working group reviewed 25 studies with 85 outcomes in which both markers were measured, and found LDL particle number and ApoB each significantly associated with the outcome in 50 comparisons and neither in 17 [3].
Population
27,673 women followed 11 years, 1,015 events [4]. Participants in a cohort of 6,814 adults aged 45 to 84, 319 events over 5.5 years [5]. 4,679 participants of the same cohort [8]. 25 clinical studies [3].
Sources
[4], [5], [8], [3]
Caveats
Observational. None of the studies here assigned treatment by LDL particle number, so they do not show what happens to events when it is lowered. The multi-ethnic cohort analysis was written with employees of the company that ran the test [5].
LDL particle number (LDL-P)Moderate

Heart disease when LDL particle number and LDL cholesterol disagree

Linked to coronary events when the two disagree: women with low LDL cholesterol and high LDL particle number had about 2.3 times the risk of women low on both.

The numbers and sources
Effect
Among 27,533 healthy women split at the medians, 24.3% had LDL particle number and LDL cholesterol on opposite sides. Among women with LDL cholesterol below the median of 121 mg/dL, those with LDL particle number at or above its median had an age-adjusted hazard ratio for coronary events of 2.32 (95% CI 1.88 to 2.85) against women low on both. Among women with LDL cholesterol at or above the median, those with a low LDL particle number had a hazard ratio of 0.42 (0.33 to 0.53). The same pattern held for ApoB (2.48 and 0.34) and non-HDL cholesterol (2.92 and 0.40) [9]. In a multi-ethnic cohort, among people whose two values differed by 12 percentile points or more, LDL particle number was associated with cardiovascular events (HR 1.45 per standard deviation, 95% CI 1.19 to 1.78) and LDL cholesterol was not (1.07, 0.88 to 1.30) [5]. In the Framingham Offspring cohort, people below the 25th percentile of LDL particle number had 59 events per 1,000 person-years, against 81 for those equally low on LDL cholesterol [10].
Population
27,533 women, median follow-up 17.2 years, 1,070 coronary events [9]. Participants in a cohort of 6,814 adults, 319 events [5]. 3,066 middle-aged white adults followed a median 14.8 years, 531 events [10].
Sources
[9], [5], [10]
Caveats
Observational. High and low were defined by each cohort's own medians or percentiles, not by a fixed cut point. After adjustment for HDL cholesterol, triglycerides and other factors, risk in the women's cohort remained under or overestimated by about 20% to 50% [9].
Lipoprotein(a)Moderate

Aortic valve stenosis

Linked to more aortic valve stenosis in cohort and genetic studies, about 3 times more at levels above 90 mg/dL in 77,680 Danish adults.

The numbers and sources
Effect
Against levels below 5 mg/dL, adjusted hazard ratios for aortic valve stenosis were 1.6 (95% CI 1.1 to 2.4) at 20 to 64 mg/dL, 2.0 (1.2 to 3.4) at 65 to 90 mg/dL and 2.9 (1.8 to 4.9) above 90 mg/dL. Genetic analysis gave a relative risk of 1.6 (95% CI 1.2 to 2.1) for a 10-fold higher lipoprotein(a) [15]. In a genome-wide study, one LPA variant was associated with aortic valve calcification (odds ratio per allele 2.05), with incident aortic stenosis (HR per allele 1.68, 95% CI 1.32 to 2.15) and with aortic valve replacement (HR 1.54, 95% CI 1.05 to 2.27) [16].
Population
77,680 Danish adults followed for up to 20 years, 454 diagnosed with aortic valve stenosis [15]. 6,942 people with CT scans in the discovery cohorts, with replication in further cohorts [16].
Sources
[15], [16]
Caveats
The cohort reported mass units (mg/dL), which do not convert to nmol/L by a fixed factor. The cohort analysis is observational, and the genetic analyses estimate the effect of lifelong differences [15][16].
NT-proBNPModerate

Heart failure, heart disease and stroke in people without cardiovascular disease

Linked to about twice the rate of heart disease, stroke or heart failure in the top third against the bottom third among 95,617 people without cardiovascular disease.

The numbers and sources
Effect
In an individual-participant meta-analysis, people in the top third of NT-proBNP had a risk ratio of 1.76 (95% CI 1.56 to 1.98) for coronary heart disease or stroke and 2.00 (1.77 to 2.26) for coronary heart disease, stroke or heart failure against the bottom third, after adjustment for conventional risk factors. Adding NT-proBNP to those risk factors raised the C-index by 0.019 for the combined outcome. The authors describe it as a strong predictor of first-onset heart failure [6].
Population
95,617 participants without a history of cardiovascular disease in 40 prospective studies, with 5,500 coronary heart disease, 4,002 stroke and 2,212 heart failure outcomes [6].
Sources
[6]
Caveats
Observational. The comparison is between thirds of the distribution in people without cardiovascular disease, not between the bands of the guideline ranges on this page [6].
NT-proBNPModerate

Death from any cause and from cardiovascular disease

Linked to about 2.4 times the deaths from any cause in the highest against the lowest category across 11 general population studies.

The numbers and sources
Effect
In a meta-analysis of prospective studies in the general population, the highest category of NT-proBNP against the lowest had a hazard ratio of 2.44 (95% CI 2.11 to 2.83) for death from any cause, 3.77 (2.85 to 5.00) for cardiovascular death and 2.35 (1.45 to 3.82) for death from coronary heart disease [7]. Among patients with acute heart failure in a pooled emergency department study, a level above 5,180 pg/mL at presentation went with death within 76 days (OR 5.2, 95% CI 2.2 to 8.1) [5]. In a meta-analysis by kidney function, the unadjusted risk ratio for death with a raised NT-proBNP was 3.01 (2.53 to 3.58) in patients with preserved kidney function and 3.25 (2.45 to 4.30) in patients with reduced function [8].
Population
25,715 people in 11 studies [7]. 720 patients with acute heart failure among 1,256 with acute breathlessness [5]. 32,203 patients in 30 studies [8].
Sources
[7], [5], [8]
Caveats
Observational. The studies defined the highest and lowest categories by different cut points [7].
Omega-3 indexModerate

Death from coronary heart disease

Linked to fewer coronary deaths at higher levels: an estimated 30% lower risk at an index of 8% than at 4% across 10 pooled cohorts.

The numbers and sources
Effect
In an analysis of 10 cohort studies, the mean omega-3 index was 6.1% (SD 2.1%) and each standard deviation higher went with a hazard ratio for fatal coronary heart disease of 0.85 (95% CI 0.80 to 0.91). The median index was 4.2% in the lowest fifth and 8.3% in the highest, from which the authors estimated about 30% lower risk at 8% than at 4% [3]. In the Framingham Offspring cohort, people in the highest fifth of the index (above 6.8%) had a 39% lower risk of a first cardiovascular event than those in the lowest fifth (below 4.2%) [6].
Population
10 cohort studies drawn from a published 19-cohort meta-analysis [3]. 2,500 participants, mean age 66, 54% women, 245 cardiovascular events [6].
Sources
[3], [6]
Caveats
Observational, so confounding cannot be ruled out. Both papers were led by a developer of the index. The Framingham authors list a relatively short follow-up and a single measurement as limits [6].
Parathyroid hormone (PTH)Moderate

Heart attack, stroke and cardiovascular death

Linked to more cardiovascular events at higher PTH: 45% more in people with the highest levels than in those with the lowest across 10 cohorts.

The numbers and sources
Effect
Comparing the highest with the lowest PTH concentrations, the pooled hazard ratio was 1.45 (95% CI 1.24 to 1.71) for all cardiovascular events, 1.50 (1.18 to 1.91) for fatal events and 1.48 (1.14 to 1.92) for non-fatal events [13].
Population
12 prospective studies in the meta-analysis, 10 of them of total cardiovascular events, lasting 2 to 14 years, in mainly white populations with a mean age of 55 to 75 [13].
Sources
[13]
Caveats
Observational, with moderate differences between studies. The comparison is of each study's highest against its lowest PTH group, not of fixed levels. PTH also runs higher with low vitamin D and reduced kidney function (see the blocks above), so it may mark those and not cause events. In a trial that gave a drug for raised PTH, events did not fall significantly (see the next block).
Potassium levelModerate

Heart rhythm problems and cardiovascular death

Linked to more rhythm problems at low potassium and more cardiovascular deaths at high potassium in a meta-analysis of 24 observational studies.

The numbers and sources
Effect
In the older general population, low potassium was associated with more supraventricular arrhythmias (RR 1.62, 95% CI 1.02 to 2.55) and high potassium with more cardiovascular deaths (RR 1.38, 1.14 to 1.66). In patients with a heart attack, high potassium was associated with more ventricular arrhythmias (RR 2.33, 1.60 to 3.38). In people with high blood pressure, a composite cardiovascular outcome was 2.6 times as frequent with hypokalemia and 1.7 times as frequent with hyperkalemia [10].
Population
310,825 participants in 24 observational studies [10].
Sources
[10]
Caveats
Observational, and the studies defined low and high potassium by different cut points. The authors call for controlled trials [10].
Red cell distribution width (RDW)Moderate

Heart attack and other artery disease

Linked to more coronary disease at higher RDW, both in volunteers free of disease and in people who had already had a heart attack.

The numbers and sources
Effect
In UK Biobank volunteers free of coronary disease at the start, an RDW of 15% or above went with more new coronary artery disease than an RDW below 12.5% (subhazard ratio 1.67, 95% CI 1.40 to 1.99), and with more heart failure, peripheral vascular disease, atrial fibrillation and stroke [7]. In people with a prior heart attack, each 1 point higher RDW had a hazard ratio for death of 1.14 (1.05 to 1.24). The top quarter had a hazard ratio of 1.78 (1.28 to 2.47) against the bottom quarter, and higher RDW also went with more coronary death or nonfatal heart attack, new heart failure and stroke [8]. Among older adults in community studies, the hazard ratio for cardiovascular death was 1.15 per 1 point (1.12 to 1.25) [6].
Population
240,477 UK volunteers aged 40 to 70 [7]. 4,111 participants with a prior heart attack and no heart failure in a pravastatin trial, median follow-up 59.7 months [8]. 11,827 older adults in 7 studies [6].
Sources
[7], [8], [6]
Caveats
Observational. The analysis in people with a prior heart attack was planned after the trial ended [8]. None of the sources shows that RDW itself changes risk.
Red cell distribution width (RDW)Moderate

Outcomes in people with heart failure

Linked to more deaths in people with heart failure, about 10% more per 1 point of RDW across 9 pooled studies.

The numbers and sources
Effect
Among 36 routine laboratory values in a heart failure trial program, higher RDW showed the strongest association with illness and death (adjusted hazard ratio 1.17 per standard deviation). In a second cohort it was 1.29 per standard deviation for death, second only to age [9]. A meta-analysis found that each 1 point higher RDW at baseline had a hazard ratio for death of 1.10 (95% CI 1.07 to 1.13), pooling 9 of its 17 studies [10].
Population
2,679 people with chronic heart failure in a trial program and 2,140 in a hospital database [9]. 18,288 people with heart failure in 17 cohort studies [10].
Sources
[9], [10]
Caveats
Observational. The meta-analysis found evidence of publication bias among the pooled studies [10].
Thyroid stimulating hormone (TSH)Moderate

Coronary heart disease

Linked to more coronary heart disease at TSH of 10 mIU/L or above, with no link at 4.5 to 6.9 mIU/L or inside the reference range, in pooled cohorts.

The numbers and sources
Effect
In pooled data on 55,287 adults, the hazard ratio for coronary heart disease events against a TSH of 0.50 to 4.49 mIU/L was 1.00 (95% CI 0.86 to 1.18) at 4.5 to 6.9 mIU/L, 1.17 (0.96 to 1.43) at 7.0 to 9.9 mIU/L and 1.89 (1.28 to 2.80) at 10 to 19.9 mIU/L. For coronary deaths the figures were 1.09 (0.91 to 1.30), 1.42 (1.03 to 1.95) and 1.58 (1.10 to 2.27) [16]. Among 55,412 people with TSH inside 0.45 to 4.49 mIU/L, each 1 mIU/L higher TSH had a hazard ratio of 0.97 (0.90 to 1.04) for coronary death and 1.00 (0.97 to 1.03) for a first coronary event [11]. In people with TSH below 0.45 mIU/L and normal free T4, the hazard ratio for coronary death was 1.29 (1.02 to 1.62) [17].
Population
55,287 adults in 11 prospective cohorts, with events data from 7 [16]. 55,412 adults in 14 cohorts with no known thyroid or cardiovascular disease [11]. 52,674 adults in 10 cohorts [17].
Sources
[16], [11], [17]
Caveats
Observational. All three analyses come from one research collaboration and share cohorts. The high TSH groups had normal thyroxine, so these figures do not describe overt thyroid disease.
Thyroid stimulating hormone (TSH)Moderate

Atrial fibrillation

Linked to more atrial fibrillation at TSH below 0.45 mIU/L, with no link inside the reference range or at mildly raised TSH, in pooled cohorts.

The numbers and sources
Effect
In people with TSH below 0.45 mIU/L and normal free T4, the hazard ratio for new atrial fibrillation was 1.68 (95% CI 1.16 to 2.43), and risk was higher below 0.10 mIU/L than at 0.10 to 0.44 mIU/L [17]. In 30,085 people with TSH of 0.45 to 19.9 mIU/L, baseline TSH was not significantly associated with new atrial fibrillation, either inside the reference range or in subclinical hypothyroidism [19].
Population
8,711 adults in 5 cohorts with atrial fibrillation data, 785 cases [17]. 30,085 adults in 11 cohorts, 2,574 cases [19].
Sources
[17], [19]
Caveats
Observational. In the second analysis, higher free T4 inside the reference range was tied to atrial fibrillation while TSH was not [19].
TriglyceridesModerate

Coronary heart disease in genetic studies

Shown in genetic studies to track coronary heart disease, an effect that followed the number of ApoB-carrying particles more than triglycerides themselves.

The numbers and sources
Effect
Each copy of a variant in the APOA5 gene went with 16.0% higher triglycerides (95% CI 12.9 to 18.7) and an odds ratio for coronary heart disease of 1.18 (95% CI 1.11 to 1.26), in line with the hazard ratio of 1.10 seen for the same triglyceride difference in cohorts [13]. Across 185 common variants, the strength of a variant's effect on triglycerides correlated with its effect on coronary artery disease after accounting for LDL and HDL cholesterol [14]. In a Mendelian randomization analysis, triglyceride-lowering LPL variants and LDL-lowering LDLR variants went with the same lower risk per 10 mg/dL lower ApoB (odds ratios 0.771 and 0.773). After adjustment for ApoB, the association of triglycerides with coronary heart disease was null (OR 1.014, 95% CI 0.965 to 1.065) [15].
Population
20,842 coronary heart disease patients and 35,206 controls [13]. Variants mapped in genome-wide studies of plasma lipids [14]. 654,783 participants, including 91,129 coronary heart disease cases, in 63 studies [15].
Sources
[13], [14], [15]
Caveats
Variants that change triglycerides also change other lipoproteins. The APOA5 variant went with lower HDL cholesterol and higher ApoB [13]. The authors of the largest analysis concluded that the benefit of lowering triglycerides may be proportional to the change in ApoB [15].
Apolipoprotein B (ApoB)Limited

Death and heart attack when ApoB and LDL cholesterol disagree

One cohort of statin-treated adults linked high ApoB with low LDL cholesterol to 21% more deaths and 49% more heart attacks than when both were low.

The numbers and sources
Effect
Compared with both values below the median, ApoB above the median with LDL cholesterol below it had hazard ratios of 1.21 for death from any cause (95% CI 1.07 to 1.36) and 1.49 for heart attack (95% CI 1.15 to 1.92). High LDL cholesterol with low ApoB was not associated with a higher risk of either outcome [16].
Population
13,015 statin-treated adults in the Copenhagen General Population Study, median follow-up 8 years [16].
Sources
[16]
Caveats
Observational, in one Danish cohort. High and low were defined by the cohort's own medians, which the abstract does not report [16].
C-peptideLimited

Death in adults without diabetes

Limited evidence: two analyses of one US survey linked the highest quarter of fasting C-peptide to about 70% to 80% more deaths in adults without diabetes.

The numbers and sources
Effect
In 5,153 adults aged 40 to 74 without diabetes, the highest quarter of fasting C-peptide, 0.984 nmol/L (about 2.97 ng/mL) or above, had an adjusted hazard ratio of 1.72 (95% CI 1.34 to 2.21) for death from any cause and 1.60 (1.07 to 2.39) for cardiovascular death against the lowest quarter, 0.418 nmol/L (about 1.26 ng/mL) or below. C-peptide predicted these deaths better than fasting glucose, insulin or indices built from them [7]. In 5,902 adults aged 40 and older from the same survey, the highest quarter had a hazard ratio of 1.80 (1.33 to 2.43) for death from any cause and 3.20 (2.07 to 4.93) for cardiovascular death against the lowest [8].
Population
5,153 adults without diabetes in the third US National Health and Nutrition Examination Survey of 1988 to 1994, followed for a median of 14.4 years with 513 deaths [7]. 5,902 adults with no history of diabetes in the same survey, mean C-peptide 0.78 nmol/L [8].
Sources
[7], [8]
Caveats
Observational, and both analyses draw on the same survey, so they are not independent. The authors read a high C-peptide in a person without diabetes as a sign of insulin resistance. Neither source tested whether lowering it changes the risk of death [7][8].
Erythrocyte sedimentation rate (ESR)LimitedStudied in men

Death from coronary heart disease

Limited evidence: one Norwegian cohort of 2,014 men linked a higher ESR to more deaths from coronary heart disease over 23 years.

The numbers and sources
Effect
After adjustment for age, hemoglobin, smoking, cholesterol and systolic blood pressure, ESR was a strong short-term and long-term predictor of death from coronary heart disease, most of all in men who had developed angina or had a positive exercise test at a second survey 7 years later. More deaths from other causes and more nonfatal heart attacks were seen only in the upper range of ESR [11].
Population
2,014 apparently healthy men aged 40 to 60 examined in 1972 to 1975 and followed for 23 years [11].
Sources
[11]
Caveats
One observational cohort of men. The abstract gives no hazard ratios and no ESR levels [11].
Free T3 (free triiodothyronine)Limited

Death in people with heart disease

One cohort found more deaths within a year among heart patients with low free T3: 14.4% against 3% with normal free T3.

The numbers and sources
Effect
Among 573 consecutive patients with heart disease, 173 had a free T3 below 3.1 pmol/L. Over 1 year, deaths from any cause were 14.4% in that group and 3% in the group at 3.1 pmol/L or above, and cardiac deaths were 7.5% and 1.5%. In a Cox model free T3 was the strongest predictor of death from any cause (hazard ratio 3.582) [6].
Population
573 consecutive cardiac patients at one center, followed for 1 year [6].
Sources
[6]
Caveats
One observational cohort of hospital patients. A US guideline states that T3 levels are low in severe illness without thyroid disease, because less T4 is converted to T3 [2], so low free T3 may mark the severity of illness.
Free T4 (free thyroxine)Limited

Sudden cardiac death

One cohort found more sudden cardiac death at higher free T4, with a hazard ratio of 2.28 per 1 ng/dL, including in people with normal TSH.

The numbers and sources
Effect
In the Rotterdam Study, higher free T4 was associated with sudden cardiac death even in the normal range of thyroid function, with a hazard ratio of 2.28 per 1 ng/dL (95% CI 1.31 to 3.97). The absolute 10-year risk rose from 1% to 4% with increasing free T4 among people with TSH of 0.4 to 4.0 mIU/L [10].
Population
10,318 adults aged 45 and older, 261 sudden cardiac deaths over a median 9.1 years [10].
Sources
[10]
Caveats
One observational cohort with 261 events [10].
HematocritLimited

Cardiovascular death and clots in a trial that held hematocrit below 45% in polycythemia vera

One trial found a hematocrit target below 45% in polycythemia vera went with fewer cardiovascular deaths and major clots, 2.7% against 9.8% at a 45% to 50% target.

The numbers and sources
Effect
Adults with JAK2-positive polycythemia vera were assigned to treatment aimed at a hematocrit below 45% or at 45% to 50%. After a median of 31 months, death from cardiovascular causes or a major clot had occurred in 5 of 182 patients in the lower target group (2.7%) and 18 of 183 in the higher target group (9.8%), a hazard ratio in the higher target group of 3.91 (95% CI 1.45 to 10.53). The rate of adverse events did not differ significantly between groups [10].
Population
365 adults with polycythemia vera in Italy, treated with phlebotomy, hydroxyurea or both [10].
Sources
[10]
Caveats
One trial of 365 people with a bone marrow disorder that raises red cell production, so it does not show what a hematocrit of 45% to 50% means in people without that disorder [10].
HematocritLimitedStudied in men

Heart events and clots at a hematocrit of 52% or above in men on testosterone therapy

Limited evidence: one database study linked a hematocrit of 52% or above on testosterone therapy to more heart events and clots in the first year, 5.15% against 3.87%.

The numbers and sources
Effect
Among men with low testosterone who received testosterone therapy, 5,842 who reached a hematocrit of 52% or above were matched to 5,842 who did not. A major cardiovascular event or venous clot in the first year occurred in 301 (5.15%) against 226 (3.87%), an odds ratio of 1.35 (95% CI 1.13 to 1.61). Among men whose hematocrit stayed under 52%, the odds were similar with and without testosterone therapy (odds ratio 1.07, 0.98 to 1.18). The authors report that national guidelines stop or change therapy at thresholds between 50% and 55%, and that the basis for those thresholds is not cited in the guidelines [12].
Population
Men over 18 with total testosterone below 350 ng/dL in a US multi-institution records database, 86% white [12].
Sources
[12]
Caveats
Retrospective, from medical records. Men who reached 52% started with a higher hematocrit, a mean of 47.4%, and the groups could not be matched on it, so the authors cannot say whether the risk comes from reaching 52% or from the higher starting value. The guideline thresholds are quoted from this study, not from the guidelines themselves [12].
Magnesium levelLimited

Atrial fibrillation

Limited evidence: one US cohort linked serum magnesium of 1.77 mg/dL or below to about 50% more atrial fibrillation over up to 20 years.

The numbers and sources
Effect
The incidence of atrial fibrillation was 9.4 per 1,000 person-years (95% CI 6.7 to 11.9) in the lowest quartile of serum magnesium (1.77 mg/dL or below) and 6.3 per 1,000 person-years (95% CI 4.1 to 8.4) in the highest (1.99 mg/dL or above). After adjustment, the lowest quartile had a hazard ratio of 1.52 (95% CI 1.00 to 2.31) compared with the upper quartiles. Results were similar after excluding people on diuretics [11].
Population
3,530 Framingham Offspring Study participants free of atrial fibrillation and cardiovascular disease, mean age 44, 228 cases over up to 20 years [11].
Sources
[11]
Caveats
One observational cohort, and the confidence interval reaches 1.00 [11].
NT-proBNPLimited

Treatment offered because the level was raised, in people with diabetes

One trial of 300 people with diabetes and NT-proBNP above 125 pg/mL found intensified heart drug treatment cut cardiac hospital stays or deaths by 65%.

The numbers and sources
Effect
People with type 2 diabetes, an NT-proBNP above 125 pg/mL and no cardiac disease were assigned to usual diabetes care or to additional treatment at a cardiac clinic, where renin-angiotensin system blockers and beta-blockers were raised to the highest tolerated doses. After 2 years, hospitalization or death due to cardiac disease was less frequent in the intensified group (HR 0.351, 95% CI 0.127 to 0.975) [10].
Population
300 people with type 2 diabetes, mean age 67.5, with a median NT-proBNP of 265.5 pg/mL [10].
Sources
[10]
Caveats
One trial of 300 people with a wide confidence interval. It tested treating people selected by the marker. It did not compare them with people whose level was not raised, and it does not show what lowering the level itself does [10].
Platelet countLimited

Heart attack, stroke and other artery disease

Limited evidence: one Danish cohort linked a count of 301 to 450 thousand/µL to 32% more cardiovascular disease than a count of 201 to 250.

The numbers and sources
Effect
The adjusted hazard ratio for a later heart attack, ischemic stroke or peripheral vascular disease was 1.32 (95% CI 1.03 to 1.69) at 301 to 450 thousand/µL against 201 to 250. A count of 100 to 200 was not associated with later bleeding or venous clots [7]. In an Italian cohort, counts outside the limits for age and sex were not related to cardiovascular death [8].
Population
21,252 adults aged 20 and older in a Danish general population study, linked to national registers [7]. 21,563 Italian adults [8].
Sources
[7], [8]
Caveats
One cohort, observational, adjusted for age, sex, smoking, hemoglobin, white cell count, C-reactive protein and other illness. It does not show that the platelet count itself changes risk [7].
Potassium levelLimited

Raising potassium to the upper part of the range

One trial found raising potassium to 4.5 to 5.0 mEq/L cut a combined count of arrhythmia events, hospital stays and deaths by 24% in people with implanted defibrillators.

The numbers and sources
Effect
People with a baseline plasma potassium of 4.3 mmol/L or lower were assigned to a regimen aimed at 4.5 to 5.0 mmol/L or to standard care. Over a median 39.6 months, sustained ventricular tachycardia or appropriate defibrillator therapy, unplanned hospitalization for arrhythmia or heart failure, or death occurred in 22.7% of the higher potassium group and 29.2% of the standard care group (HR 0.76, 95% CI 0.61 to 0.95). Hospitalization for hyperkalemia or hypokalemia was similar in the two groups [11].
Population
1,200 people in Denmark with an implantable cardioverter-defibrillator, at high risk of ventricular arrhythmias [11].
Sources
[11]
Caveats
One open-label trial in people with an implanted defibrillator, so it does not show what a higher potassium does in people without heart disease. The regimen used potassium supplementation, a mineralocorticoid receptor antagonist or both, with dietary guidance [11].
Why limited
This rests on one trial, of 1,200 people. Strength counts independent studies that agree, so one trial is limited at any size.
Red cell folateLimited

Cardiovascular death in adults with hypertension

Limited evidence: two US survey analyses linked the highest red cell folate levels to more cardiovascular deaths in adults with hypertension.

The numbers and sources
Effect
In adults with hypertension surveyed from 1991 to 1994, the high folate quartile had a hazard ratio for cardiovascular death of 1.44 (95% CI 1.31 to 1.58) against the intermediate quartiles and 2.13 (1.77 to 2.57) for death from heart attack. The low quartile did not differ (1.09, 0.94 to 1.27) [9]. In adults with hypertension surveyed from 1999 to 2014, the highest quartile against the second had hazard ratios of 1.68 (1.30 to 2.16) for cardiovascular death and 1.30 (1.16 to 1.46) for death from any cause, and the lowest quartile was not associated with either. The associations were non-linear, with inflection points at 819.7 and 760.1 ng/mL [10].
Population
2,986 US adults with hypertension, 579 cardiovascular deaths over an average 11.7 years [9]. 13,986 US adults with hypertension, 548 cardiovascular deaths over a median 7.0 years [10].
Sources
[9], [10]
Caveats
Observational, and limited to people with hypertension. Supplement use and illness may differ between people with high and low levels. No trial cited here tested the outcome of lowering a high red cell folate [9][10].
Serum creatinineLimitedStudied in men

Stroke and death at higher levels

Limited evidence: one British cohort of 7,690 middle-aged men linked creatinine above 1.31 mg/dL to 60% more strokes over about 15 years.

The numbers and sources
Effect
Median creatinine was 98 µmol/L (1.11 mg/dL). Above 116 µmol/L (1.31 mg/dL), the top tenth, the relative risk of stroke was 1.6 (95% CI 1.1 to 2.1) against all other men after adjustment for cardiovascular risk factors. For major ischemic heart disease, the relative risk at 130 µmol/L (1.47 mg/dL) or above was 1.2 (0.8 to 1.7) after adjustment. Death from any cause and from cardiovascular disease was significantly raised only above the 97.5th percentile, and there was no association with death from cancer [6].
Population
7,690 men aged 40 to 59 from 24 British towns, followed for an average of 14.75 years, with 287 strokes, 967 major ischemic heart disease events and 1,259 deaths [6].
Sources
[6]
Caveats
One observational cohort of middle-aged men in one country [6]. Studies that convert creatinine to an estimated filtration rate are covered on the eGFR page.
Serum ironLimited

Coronary artery disease in a genetic study

Limited evidence: one genetic study found higher serum iron went with slightly less coronary artery disease, an odds ratio of 0.94 per standard deviation.

The numbers and sources
Effect
In a Mendelian randomization study using three gene variants tied to iron status, the odds ratio for coronary artery disease was 0.94 per standard deviation higher serum iron (95% CI 0.88 to 1.00, P = 0.039). Results for transferrin saturation (0.95, 0.91 to 0.99) and ferritin (0.85, 0.73 to 0.98) pointed the same way [7].
Population
Gene associations with iron status from 48,972 people, and with coronary artery disease from 60,801 cases and 123,504 controls combined with a second set of 63,746 cases and 130,681 controls [7].
Sources
[7]
Caveats
A genetic estimate of lifelong difference in iron status, resting on three variants, with a confidence interval that reaches 1.00. It does not show what a single high or low reading means, or what changing serum iron does [7].
Serum ironLimited

Stroke in a genetic study

Limited evidence: one genetic study found higher serum iron went with slightly more stroke, an odds ratio of 1.07 per standard deviation.

The numbers and sources
Effect
In a Mendelian randomization study, the odds ratio for stroke was 1.07 per standard deviation higher genetically determined serum iron (95% CI 1.01 to 1.14), 1.06 for transferrin saturation (1.01 to 1.11) and 1.18 for ferritin (1.02 to 1.36). Among types of ischemic stroke, the finding was driven by cardioembolic stroke [8].
Population
Gene associations with iron status from 48,972 people, and with stroke from the MEGASTROKE consortium [8].
Sources
[8]
Caveats
The same gene variants pointed toward less coronary artery disease in a separate study, so the genetic evidence runs in opposite directions for two vascular outcomes [7][8].
Urine albumin to creatinine ratioLimited

Cardiovascular events at ratios below 30 mg/g

One cohort of 1,568 adults without diabetes or high blood pressure linked ratios above the group's median, far below 30 mg/g, to nearly three times the cardiovascular events.

The numbers and sources
Effect
Participants with a ratio at or above the sex-specific median (3.9 mg/g for men, 7.5 mg/g for women) had a hazard ratio of 2.92 (95% CI 1.57 to 5.44) for a first cardiovascular event against those below the median, and 1.75 (0.95 to 3.22) for death. The finding held when people with a ratio of 30 mg/g or more were excluded [9].
Population
1,568 adults in the Framingham Offspring Study, mean age 55, free of cardiovascular disease, hypertension and diabetes, followed for a median of 6 years, with 54 first cardiovascular events [9].
Sources
[9]
Caveats
One cohort with 54 events and wide intervals. The larger pooled analyses above also found risk rising below 30 mg/g [5].
Fasting glucoseStrong

Diabetes and heart disease in trials that treated raised glucose before diabetes

Mixed results: in trials of people with raised glucose, lifestyle change cut new diabetes 58%, while neither it nor insulin aimed at normal fasting glucose cut cardiovascular events.

The numbers and sources
Effect
In people with elevated fasting and post-load glucose, a lifestyle program reduced the incidence of diabetes by 58% (95% CI 48 to 66) and metformin by 31% (95% CI 17 to 43) against placebo over 2.8 years [12]. Over a median 21 years of follow-up of the same participants, neither metformin (HR 1.03, 95% CI 0.78 to 1.37) nor the lifestyle program (HR 1.14, 95% CI 0.87 to 1.50) reduced major cardiovascular events [13]. In a trial that gave insulin glargine to reach a fasting glucose of 95 mg/dL or below, cardiovascular outcomes did not differ from standard care over 6.2 years (HR 1.02, 95% CI 0.94 to 1.11). New diabetes was less frequent (30% against 35%), severe hypoglycemia was more frequent (1.00 against 0.31 per 100 person-years) and weight rose [14].
Population
3,234 people without diabetes, mean age 51, mean body mass index 34.0 [12][13]. 12,537 people, mean age 63.5, with cardiovascular risk factors plus impaired fasting glucose, impaired glucose tolerance or type 2 diabetes [14].
Sources
[12], [13], [14]
Caveats
A marker that tracks risk is not always one whose lowering changes it. In the long follow-up, all groups were offered a lifestyle program and many used statins and blood pressure drugs, which the authors say may have diluted any effect [13]. The insulin trial included people who already had type 2 diabetes [14].
Fasting insulin and HOMA-IRStrong

Coronary heart disease

Mixed results: two meta-analyses found no clear link between fasting insulin and coronary heart disease and one found 50% more, while HOMA-IR was linked in the one that pooled it.

The numbers and sources
Effect
In 14 Western prospective studies, the odds ratio for coronary heart disease in the top against the bottom third of fasting insulin was 1.12 (95% CI 0.98 to 1.28), and the authors concluded the association is likely to be more modest than previously suspected [9]. In a meta-analysis of adults without diabetes, the relative risk of coronary heart disease for high against low values was 1.12 (95% CI 0.92 to 1.37) for fasting insulin, 1.52 (1.31 to 1.76) for fasting glucose and 1.64 (1.35 to 2.00) for HOMA-IR. Per standard deviation it was 1.04 (0.96 to 1.12) for insulin and 1.46 (1.26 to 1.69) for HOMA-IR [10]. A third meta-analysis found a relative risk of 1.50 (95% CI 1.28 to 1.77) for the highest against the lowest fasting insulin group, and 1.16 (1.10 to 1.22) per 50 pmol/L [11].
Population
2,649 coronary heart disease cases in 14 studies of fasting insulin [9]. 516,325 participants in 65 studies of glucose, insulin or HOMA-IR [10]. 22,379 people and 1,986 cases in 9 studies [11].
Sources
[9], [10], [11]
Caveats
Observational. Each study compared its own high and low groups, measured with its own insulin assay, so the pooled groups share no common cut point.
Hemoglobin A1c (HbA1c)Strong

Heart attack and death in trials that lowered HbA1c in type 2 diabetes

Mixed results: pooled trials of intensive glucose lowering found 9% fewer major cardiovascular events and no fewer deaths, and one trial aiming below 6.0% found more deaths.

The numbers and sources
Effect
A prospectively planned meta-analysis found that more intensive glucose control reduced major cardiovascular events by 9% (HR 0.91, 95% CI 0.84 to 0.99) and heart attack by 15% (HR 0.85, 95% CI 0.76 to 0.94). Death from any cause (HR 1.04, 95% CI 0.90 to 1.20) and cardiovascular death (HR 1.10, 95% CI 0.84 to 1.42) were not reduced, and major hypoglycemia rose (HR 2.48, 95% CI 1.91 to 3.21) [15]. In one of the pooled trials, a strategy targeting HbA1c below 6.0% reached a median of 6.4% against 7.5% with standard therapy. Deaths were higher in the intensive group (257 against 203, HR 1.22, 95% CI 1.01 to 1.46), and the intensive strategy was stopped after a mean 3.5 years [16].
Population
27,049 participants with type 2 diabetes and 2,370 major vascular events over 4.4 years [15]. 10,251 patients, mean age 62, with cardiovascular disease or risk factors and a median HbA1c of 8.1% [16].
Sources
[15], [16]
Caveats
A marker that tracks risk is not always one whose lowering reduces it. These trials used the drugs and insulin regimens of their time in older, high-risk people with long-standing diabetes, and do not describe people without diabetes.
High-sensitivity C-reactive protein (hs-CRP)Strong

Risk prediction beyond standard risk factors

Mixed results: adding CRP to standard risk factors improved prediction slightly in 52 pooled cohorts, and the US Preventive Services Task Force judged the evidence insufficient.

The numbers and sources
Effect
Across 52 prospective studies, adding CRP to a model of conventional risk factors raised the C-index by 0.0039 and gave a net reclassification improvement of 1.52%. The authors estimated that measuring CRP in people at intermediate risk could prevent one additional cardiovascular event over 10 years for every 400 to 500 people screened [15]. The US Preventive Services Task Force found that adding hs-CRP to existing risk models gives small improvements in discrimination and reclassification whose clinical meaning is largely unknown, and concluded the evidence is insufficient to assess the balance of benefits and harms [16].
Population
246,669 participants without cardiovascular disease in 52 studies [15]. Asymptomatic adults [16].
Sources
[15], [16]
Caveats
The estimate of events prevented assumes statin treatment decisions under older US guidelines [15].
Omega-3 indexStrong

Heart attack and coronary death in trials of omega-3 supplements

Mixed results: a pooled analysis of 10 supplement trials found no significant reduction in coronary events, while a later one of 13 trials found reductions of 5% to 8%.

The numbers and sources
Effect
In a meta-analysis of 10 trials, allocation to omega-3 supplements had no significant association with coronary heart disease death (RR 0.93, 99% CI 0.83 to 1.03), nonfatal heart attack (RR 0.97, 99% CI 0.87 to 1.08), any coronary heart disease event (RR 0.96, 95% CI 0.90 to 1.01) or major vascular events (RR 0.97, 95% CI 0.93 to 1.01) [8]. A later meta-analysis of 13 trials found, in its analysis that left out the REDUCE-IT trial of icosapent ethyl, lower risks of heart attack (RR 0.92, 95% CI 0.86 to 0.99), coronary heart disease death (RR 0.92, 95% CI 0.86 to 0.98) and total coronary heart disease (RR 0.95, 95% CI 0.91 to 0.99) [9].
Population
77,917 people at high cardiovascular risk in 10 trials lasting a mean 4.4 years, EPA doses of 226 to 1,800 mg a day [8]. 127,477 participants in 13 trials, mean treatment 5.0 years [9].
Sources
[8], [9]
Caveats
These analyses compared people by the dose they were assigned, not by the omega-3 index they reached. One developer of the index argues that events track blood levels more than dose and that uptake varies from person to person, which he says makes the effects hard to show in trials of this design [1]. No trial reported here treated people to an index target.
Total cholesterolStrong

Stroke

Mixed results: higher total cholesterol went with fewer bleeding strokes in pooled cohorts, and higher non-HDL cholesterol with slightly more ischemic strokes.

The numbers and sources
Effect
In a meta-analysis of 23 prospective studies, the relative risk of hemorrhagic stroke was 0.69 (95% CI 0.59 to 0.81) for high against low total cholesterol and 0.85 (0.80 to 0.91) per 1 mmol/L higher [11]. In 61 prospective studies, total cholesterol showed no independent positive association with death from stroke, and at ages 70 to 89 and at systolic blood pressure above about 145 mm Hg it was negatively related to death from hemorrhagic and total stroke [5]. In 68 pooled cohorts, the hazard ratio for ischemic stroke was 1.12 (95% CI 1.04 to 1.20) per 43 mg/dL higher non-HDL cholesterol [6].
Population
1,430,141 participants with 7,960 hemorrhagic strokes [11]. Almost 900,000 adults with about 12,000 stroke deaths [5]. 302,430 people with 2,534 ischemic strokes [6].
Sources
[11], [5], [6]
Caveats
Observational. The authors of the largest analysis call the absence of a positive link with stroke death unexplained, and note that statin trials reduce total stroke rates [5].
TriglyceridesStrong

Coronary heart disease in cohort studies

Mixed results: pooled cohorts linked higher triglycerides to more coronary heart disease, but in the largest analysis the link disappeared after adjustment for other risk factors.

The numbers and sources
Effect
In a meta-analysis of 29 Western prospective studies, people in the top third of usual triglyceride values had an adjusted odds ratio for coronary heart disease of 1.72 (95% CI 1.56 to 1.90) against the bottom third. The authors noted that adjustment for established risk factors substantially weakened the association [10]. In 68 prospective studies, coronary heart disease rates per 1,000 person-years were 2.6 in the bottom third and 6.2 in the top third of triglycerides. After adjustment for several conventional factors, the hazard ratio per standard deviation higher triglycerides was 0.99 (95% CI 0.94 to 1.05), and the authors concluded that vascular risk can be assessed without regard to triglyceride [11]. In a Danish cohort followed for a mean 26 years, multifactorially adjusted hazard ratios for heart attack at non-fasting triglycerides of 5 mmol/L (442.5 mg/dL) or more, against below 1 mmol/L, were 5.4 in women and 2.4 in men [12].
Population
262,525 participants with 10,158 coronary heart disease cases in 29 studies [10]. 302,430 people without vascular disease in 68 prospective studies [11]. 7,587 women and 6,394 men aged 20 to 93 in Copenhagen [12].
Sources
[10], [11], [12]
Caveats
Observational. Triglycerides travel with low HDL cholesterol, obesity and diabetes, so how much of the link belongs to triglycerides themselves depends on what an analysis adjusts for [10][11].
TriglyceridesStrong

Cardiovascular events in trials that lowered triglycerides

Mixed results: fibrates cut cardiovascular events 10% in pooled trials and icosapent ethyl 25% in one, while pemafibrate lowered triglycerides 26% without fewer events.

The numbers and sources
Effect
Across 18 trials, fibrates reduced major cardiovascular events by 10% (95% CI 0 to 18) and coronary events by 13% (95% CI 7 to 19), with no effect on death from any cause [16]. Icosapent ethyl, 4 g a day, reduced a composite of cardiovascular events against placebo (17.2% against 22.0%, HR 0.75, 95% CI 0.68 to 0.83) in statin-treated patients with triglycerides of 135 to 499 mg/dL [17]. A different omega-3 formulation, 4 g a day against corn oil, made no difference (HR 0.99, 95% CI 0.90 to 1.09) and the trial was stopped early [18]. Pemafibrate lowered triglycerides by 26.2% against placebo and did not reduce cardiovascular events (HR 1.03, 95% CI 0.91 to 1.15), while ApoB rose 4.8% [19]. In a meta-regression of 49 trials, each 1 mmol/L reduction in triglycerides went with a risk ratio of 0.84 (95% CI 0.75 to 0.94) for major vascular events after accounting for LDL cholesterol. With the icosapent ethyl trial removed, the risk ratio was 0.91 (95% CI 0.81 to 1.006) [20].
Population
45,058 participants in 18 fibrate trials [16]. 8,179 patients followed a median 4.9 years [17]. 13,078 statin-treated patients at high cardiovascular risk [18]. 10,497 patients with type 2 diabetes, triglycerides of 200 to 499 mg/dL and low HDL cholesterol, followed a median 3.4 years [19]. 374,358 patients in 49 trials [20].
Sources
[16], [17], [18], [19], [20]
Caveats
A marker that tracks risk is not always one whose lowering changes it. The meta-regression's authors concluded that the benefit of high-dose eicosapentaenoic acid appeared to exceed its lipid-lowering effect [20]. The icosapent ethyl and pemafibrate trials were funded by the drugs' makers [17][19].
Bilirubin (total)Moderate

Heart disease and stroke

Mixed results: pooled cohorts linked higher bilirubin to 7% less cardiovascular disease per standard deviation, while genetic studies found no effect on heart disease.

The numbers and sources
Effect
In a meta-analysis of 12 population cohorts, the adjusted relative risk of cardiovascular disease was 0.93 (95% CI 0.90 to 0.97) per standard deviation higher total bilirubin, 0.95 (0.92 to 0.99) for coronary heart disease and 0.93 (0.88 to 0.98) for stroke. Adding bilirubin to a risk model built on established risk factors did not improve prediction [6]. In three Danish studies, people with the UGT1A1 genotype that raised bilirubin by 95% had no lower odds of ischemic heart disease (OR 1.03, 95% CI 0.96 to 1.11) or heart attack (1.01, 0.92 to 1.12). A meta-analysis of genetic studies gave an odds ratio of 1.01 (0.88 to 1.16) for genotypes with about twice the bilirubin [7].
Population
12 prospective studies with 9,378 cardiovascular events, including 7,222 participants in the PREVEND cohort [6]. 67,068 genotyped Danish adults, and 14,711 cases with 60,324 controls in the genetic meta-analysis [7].
Sources
[6], [7]
Caveats
The cohort finding is observational. The authors of the genetic studies conclude that their data suggest bilirubin is not causally associated with the risk of ischemic heart disease [7].
CortisolModerate

Cardiovascular disease

Mixed results: higher morning cortisol went with 18% more cardiovascular disease per standard deviation in pooled cohorts, while genetic studies found no clear effect on heart disease or stroke.

The numbers and sources
Effect
In two nested case-control studies, each standard deviation higher morning plasma cortisol had an odds ratio for new cardiovascular disease of 1.28 (95% CI 1.06 to 1.54), and in a meta-analysis with earlier prospective studies it was 1.18 (1.06 to 1.31). The same paper's genetic analysis gave 1.06 (0.98 to 1.15) [5]. A second genetic study found genetically predicted cortisol unrelated to ischemic heart disease (0.98, 0.93 to 1.03) or ischemic stroke (0.99, 0.91 to 1.08) [6]. A third found each standard deviation of genetically higher cortisol associated with more atrial fibrillation (1.20, 1.06 to 1.35), an association that disappeared after adjusting for genetically predicted blood pressure and waist size (0.99, 0.72 to 1.38) [7].
Population
696 people with new cardiovascular disease and 6,680 controls in the pooled cohorts, and 122,737 coronary heart disease cases with 547,261 controls in the genetic analysis [5]. Genetic variants from a consortium of 12,597 people, applied to large outcome studies [6][7].
Sources
[5], [6], [7]
Caveats
The cohort finding is observational, and the genetic estimates disagree with one another. The genetic studies rest on a small number of variants [6][7].
EstradiolModerateStudied in women

Fractures, stroke and breast cancer in trials of estrogen therapy after menopause

Mixed results: in one large trial estrogen alone cut fractures by 27% and raised strokes by 33% in postmenopausal women, with little or no difference in breast cancer.

The numbers and sources
Effect
A Cochrane review drew its main results from the Women's Health Initiative. Among 10,739 women who had had a hysterectomy, oral conjugated equine estrogen for an average of 7 years, against placebo, gave relative risks of 0.73 for all clinical fractures (95% CI 0.65 to 0.80), 1.33 for stroke (1.06 to 1.67), 1.78 for gallbladder disease requiring surgery (1.42 to 2.24), 0.94 for coronary events (0.78 to 1.13), 1.32 for venous clots (1.00 to 1.74) and 0.79 for breast cancer (0.61 to 1.01). Among 16,608 women with a uterus, estrogen with a progestin for 5.6 years gave 0.78 for fractures (0.71 to 0.86), 1.39 for stroke (1.09 to 2.09), 2.03 for venous clots (1.55 to 6.64) and 1.27 for breast cancer (1.03 to 1.56) [10].
Population
24 randomized trials with 45,660 women, most of them postmenopausal American women with a mean age over 60 [10].
Sources
[10], [7]
Caveats
A marker that tracks an outcome is not always one whose correction changes it. Higher estradiol went with more breast cancer in cohorts [7], and estrogen given alone did not raise it in this trial. The trials set a dose and did not target a blood level. The review's authors caution that the results rest on one study of oral therapy, and that only about 30% of the women were aged 50 to 59 [10].
Folate levelModerate

Stroke in trials of folic acid

Mixed results: folic acid cut first stroke 21% in Chinese adults with hypertension, while a pooled analysis of 8 other trials found no significant effect.

The numbers and sources
Effect
In a Chinese trial, enalapril with folic acid 0.8 mg lowered first stroke compared with enalapril alone (2.7% against 3.4%, HR 0.79, 95% CI 0.68 to 0.93) over a median 4.5 years. Heart attack (HR 1.04, 95% CI 0.60 to 1.82) and death from any cause (HR 0.94, 95% CI 0.81 to 1.10) did not differ [9]. In a meta-analysis of 8 trials of folic acid in people at raised cardiovascular risk, the rate ratio for stroke was 0.96 (95% CI 0.87 to 1.06) and for major vascular events 1.01 (95% CI 0.97 to 1.05) [10]. A Cochrane review of B vitamin trials found a small reduction in stroke against placebo (RR 0.90, 95% CI 0.82 to 0.99, 10 trials) and no difference in heart attack (RR 1.02, 95% CI 0.95 to 1.10) [11].
Population
20,702 adults with hypertension and no history of stroke or heart attack [9]. 37,485 participants in 8 trials, median follow-up 5 years [10]. 71,422 participants in 15 trials [11].
Sources
[9], [10], [11]
Caveats
These trials assigned folic acid, alone or with other B vitamins, and report outcomes by assigned treatment, not by the folate level reached. The Cochrane authors note that the stroke evidence for folic acid with enalapril came from one very large trial [11].
Free T3 (free triiodothyronine)Moderate

Heart function and death in trials that gave T3

Mixed results: trials giving T3 after heart surgery or in heart failure with low T3 improved measures of heart function, and deaths after surgery did not differ.

The numbers and sources
Effect
In a systematic review of 14 randomized trials of T3 after surgery, 13 of them cardiac surgery, intravenous T3 increased cardiac index after coronary bypass. Mortality was not significantly altered by high-dose intravenous T3 and could not be assessed for the other regimens [8]. In a meta-analysis of randomized trials in heart failure with low T3, thyroid hormone treatment raised free T3 and improved left ventricular ejection fraction (weighted mean difference 5.61, 95% CI 4.38 to 6.85) and cardiac output (0.65, 0.42 to 0.89). The abstract reports no data on deaths [9].
Population
14 randomized trials of postoperative T3 in adults [8]. Randomized trials in heart failure with low T3 found in English and Chinese databases to November 2021 [9].
Sources
[8], [9]
Caveats
A marker that tracks an outcome is not always one whose correction changes it. The surgical trials were small and short, with T3 given for 6 to 24 hours in most. The heart failure analysis had high heterogeneity, with I-squared of 63% to 95% for its main outcomes [8][9].
HDL cholesterol (HDL-C)Moderate

Stroke

Mixed results: 68 pooled cohorts found no clear link with ischemic stroke, and a meta-analysis tied higher HDL cholesterol to slightly more bleeding in the brain.

The numbers and sources
Effect
The adjusted hazard ratio for ischemic stroke was 0.93 (95% CI 0.84 to 1.02) per 15 mg/dL higher HDL cholesterol [8]. In a meta-analysis of prospective studies, the relative risk of hemorrhagic stroke was 0.98 (95% CI 0.80 to 1.19) for high against low HDL cholesterol and 1.11 (0.99 to 1.25) per 1 mmol/L higher. For intracerebral hemorrhage the pooled relative risk was 1.17 (95% CI 1.02 to 1.35) [19].
Population
302,430 people in 68 prospective studies, with 2,534 ischemic strokes [8]. 1,430,141 participants in 23 prospective studies, with 7,960 hemorrhagic strokes [19].
Sources
[8], [19]
Caveats
Observational. Most intervals include no association [8][19].
HomocysteineModerate

Heart disease and stroke in cohorts and genetic studies

Mixed results: cohorts linked 25% lower homocysteine to 11% less heart disease and 19% less stroke, while large genetic datasets found little or no effect on heart disease.

The numbers and sources
Effect
In a meta-analysis of individual data, a 25% lower usual homocysteine level (about 3 µmol/L) was associated in the prospective studies with an 11% lower risk of ischemic heart disease (OR 0.89, 95% CI 0.83 to 0.96) and a 19% lower risk of stroke (OR 0.81, 95% CI 0.69 to 0.95), after adjustment for known risk factors. Associations were stronger in retrospective studies, where blood was drawn after disease onset [5]. In 19 unpublished genetic datasets, people with the MTHFR variant that raises homocysteine by about 20% had no excess of coronary heart disease (OR 1.02, 95% CI 0.98 to 1.07). In 86 published studies of the same variant the odds ratio was 1.15 (95% CI 1.09 to 1.21), a discrepancy the authors attribute to publication bias or methodological problems [6].
Population
30 prospective or retrospective studies with 5,073 ischemic heart disease events and 1,113 stroke events [5]. 48,175 coronary heart disease cases and 67,961 controls in the unpublished datasets, and 28,617 cases and 41,857 controls in the published ones [6].
Sources
[5], [6]
Caveats
The cohort authors conclude that raised homocysteine is at most a modest independent predictor in healthy populations [5]. The genetic datasets did not measure homocysteine itself and tested heart disease, not stroke [6].
HomocysteineModerate

Stroke in trials that lowered homocysteine

Mixed results: a Cochrane review found 10% fewer strokes with homocysteine-lowering B vitamins, while a pooled analysis of 8 trials found no significant effect.

The numbers and sources
Effect
A Cochrane review found a small reduction in stroke with homocysteine-lowering B vitamins against placebo (4.3% against 5.1%, RR 0.90, 95% CI 0.82 to 0.99, 10 trials) [8]. An individual-data meta-analysis of 8 trials found a rate ratio for stroke of 0.96 (95% CI 0.87 to 1.06) [7]. In a Chinese trial in adults with hypertension, adding folic acid 0.8 mg to enalapril lowered first stroke (HR 0.79, 95% CI 0.68 to 0.93) over a median 4.5 years, with no difference in heart attack or death from any cause [9].
Population
44,224 participants in the 10 trials with stroke data [8]. 37,485 participants in 8 trials [7]. 20,702 adults with hypertension and no history of stroke or heart attack [9].
Sources
[8], [7], [9]
Caveats
The Cochrane authors calculate that about 143 people would need treatment with folic acid plus enalapril for 5.4 years to prevent 1 stroke, and note that this evidence came from one very large trial [8].
LDL cholesterol (LDL-C)Moderate

Bleeding stroke at low levels

Mixed results: one cohort linked LDL cholesterol below 70 mg/dL to more bleeding strokes, while pooled trials lowering it to very low levels found no increase.

The numbers and sources
Effect
In a cohort with repeated LDL cholesterol measurements, people averaging below 70 mg/dL had more intracerebral hemorrhage than those at 70 to 99 mg/dL: adjusted HR 1.65 (95% CI 1.32 to 2.05) at 50 to 69 mg/dL and 2.69 (2.03 to 3.57) below 50 mg/dL. Risk at 100 mg/dL or above was similar to that at 70 to 99 [16]. In a meta-analysis of trials in populations starting at a median of 70 mg/dL or less and reaching a median as low as 21 mg/dL, LDL lowering was not associated with an increased risk of hemorrhagic stroke [14]. Statin trials found no significant effect on deaths due to stroke (RR 0.96, 95% CI 0.84 to 1.09) [12].
Population
96,043 adults, mean age 51, free of stroke, heart attack and cancer at baseline, 753 hemorrhages over 9 years [16]. 50,627 patients in 3 trials of non-statin drugs plus a statin trial subgroup [14].
Sources
[16], [14], [12]
Caveats
The cohort is observational and cannot show that low LDL cholesterol caused the hemorrhages. The trials ran for a few years, so they say little about decades at very low levels.
Lipoprotein(a)Moderate

Cardiovascular events in trials that lowered lipoprotein(a)

Mixed results: a trial designed to test lowering took lipoprotein(a) from 236 to 64 nmol/L with no fewer cardiovascular events, while PCSK9 inhibitor trial analyses tied lowering to fewer.

The numbers and sources
Effect
In a placebo-controlled trial of monthly pelacarsen in patients with cardiovascular disease and lipoprotein(a) of at least 70 mg/dL, the median baseline level was 236 nmol/L. At 12 months it was 64.4 nmol/L on pelacarsen and 233.4 nmol/L on placebo. Over a median 4.7 years, a major adverse cardiovascular event occurred in 12.3% and 12.1% (HR 1.02, 95% CI 0.90 to 1.15). Among patients starting at 90 mg/dL or above the hazard ratio was 1.00 (0.87 to 1.15) [17]. Further analyses by treatment exposure, by the level reached at 3 months and by subgroup also found no association with fewer events [18]. Adding niacin to a statin lowered lipoprotein(a) 21% and did not reduce cardiovascular events [19]. Evolocumab lowered lipoprotein(a) by a median 26.9%, and reduced coronary events by 23% (HR 0.77, 95% CI 0.67 to 0.88) in patients with baseline levels above the median against 7% (HR 0.93, 0.80 to 1.08) at or below it, an interaction with a P value of 0.07 [20]. With alirocumab, each 1 mg/dL reduction in lipoprotein(a) went with a hazard ratio of 0.994 (95% CI 0.990 to 0.999) for major cardiovascular events, independent of the fall in LDL cholesterol [21].
Population
8,350 patients with cardiovascular disease randomized and 8,322 analyzed, mean age 59.7, 27.0% women [17]. Patients with cardiovascular disease and low HDL cholesterol in a niacin trial [19]. 25,096 patients with cardiovascular disease, median follow-up 2.2 years [20]. 18,924 patients after an acute coronary syndrome, median follow-up 2.8 years [21].
Sources
[17], [18], [19], [20], [21]
Caveats
A marker that tracks risk is not always one whose lowering changes it. The pelacarsen trial enrolled people who already had cardiovascular disease, was funded by Novartis and was published in November 2026 [17]. The PCSK9 inhibitor findings are secondary analyses of drugs that mainly lower LDL cholesterol [20][21]. The sources for this page include no outcome results for other drugs built to lower lipoprotein(a).
Omega-3 indexModerate

Atrial fibrillation

Mixed results: higher blood EPA plus DHA went with slightly less atrial fibrillation in pooled cohorts, while trials of omega-3 supplements found 25% more.

The numbers and sources
Effect
Across 17 cohorts, higher levels of EPA plus DHA in blood or fat tissue had a hazard ratio for new atrial fibrillation of 0.93 (95% CI 0.87 to 0.99), comparing the 90th with the 10th percentile. EPA alone showed no association (HR 1.00, 95% CI 0.95 to 1.05) [10]. In a meta-analysis of 7 trials with cardiovascular outcomes, marine omega-3 supplements were associated with more atrial fibrillation (HR 1.25, 95% CI 1.07 to 1.46), with a hazard ratio of 1.49 (95% CI 1.04 to 2.15) in trials testing more than 1 g a day and 1.12 (95% CI 1.03 to 1.22) at 1 g a day or less [11].
Population
54,799 participants in 17 cohorts, 7,720 cases over a median 13.3 years [10]. 81,210 patients in 7 trials, mean age 65, weighted average follow-up 4.9 years [11].
Sources
[10], [11]
Caveats
The cohorts measured levels in blood or fat tissue under habitual intake and are observational. The trials tested supplements, with more atrial fibrillation at higher doses [10][11].
Uric acid (serum urate)Moderate

Coronary heart disease, in cohorts and in a trial that lowered urate

Mixed results: cohorts link high uric acid to slightly more coronary heart disease, while a genetic study and an allopurinol trial in 5,721 people found no effect.

The numbers and sources
Effect
In a meta-analysis of 26 prospective studies, hyperuricemia went with an adjusted risk ratio of 1.09 (95% CI 1.03 to 1.16) for coronary heart disease and 1.16 (1.01 to 1.30) for coronary death. Each 1 mg/dL higher went with a risk ratio for coronary death of 1.12 (1.05 to 1.19). The association was not significant in men [16]. In a Mendelian randomization study, genetically higher serum urate was not associated with coronary heart disease, ischemic stroke or heart failure [11]. In a randomized trial in people aged 60 or older with ischemic heart disease and no gout, allopurinol titrated up to 600 mg daily did not change the rate of heart attack, stroke or cardiovascular death against usual care (hazard ratio 1.04, 95% CI 0.89 to 1.21) over a mean of 4.8 years [17].
Population
402,997 adults in 26 studies [16]. 54,501 coronary heart disease cases and 68,275 controls [11]. 5,721 participants, mean age 72, 75.5% men [17].
Sources
[16], [11], [17]
Caveats
A marker that tracks risk is not always one whose lowering changes it. The trial's abstract does not report how far serum urate fell, and it enrolled older people who already had heart disease [17].
DHEA sulfate (DHEA-S)Limited

Heart failure and coronary heart disease

One cohort found low DHEA-S went with 30% to 42% more heart failure hospitalizations and was not tied to coronary heart disease.

The numbers and sources
Effect
DHEA-S below the 15th percentile for sex (55.4 µg/dL in men, 27.4 µg/dL in women) had hazard ratios for heart failure hospitalization of 1.30 in men (95% CI 1.07 to 1.58) and 1.42 in women (1.13 to 1.79). In men a greater percentage fall in DHEA-S over 15 years also went with more heart failure hospitalization (1.94, 1.11 to 3.39). Neither a low level nor the change was associated with new coronary heart disease [5].
Population
8,143 adults in the Atherosclerosis Risk in Communities study, mean age 63, without cardiovascular disease at the start. 3,706 had DHEA-S measured twice, 15 years apart [5].
Sources
[5]
Caveats
One observational cohort. The cut points are percentiles of that study's own participants [5].
Free testosteroneLimitedStudied in men

Outcomes in trials that raised testosterone

Limited evidence: the large testosterone treatment trials enrolled men by total testosterone, so they do not show what correcting a low free testosterone alone does.

The numbers and sources
Effect
A trial of 790 men aged 65 or older enrolled on a total testosterone below 275 ng/dL and found improved sexual function with no benefit for vitality or walking distance [15]. A trial of 5,246 men enrolled on two fasting total testosterone levels below 300 ng/dL and found no difference in major cardiac events (hazard ratio 0.96, 95% CI 0.78 to 1.17) [16]. The American Urological Association states that total and free testosterone are not interchangeable and that no clear data point to consistent thresholds between the two below which symptoms appear and above which treatment benefits occur [3].
Population
790 men treated for 1 year [15]. 5,246 men aged 45 to 80, mean follow-up 33.0 months [16].
Sources
[15], [16], [3]
Caveats
A marker that tracks an outcome is not always one whose correction changes it. Neither trial reports results by free testosterone level in its abstract.
Total iron binding capacity (TIBC)Limited

Coronary artery disease and stroke in genetic studies of transferrin

Limited evidence: two genetic studies tied higher transferrin, the protein behind TIBC, to slightly more coronary artery disease and slightly less stroke.

The numbers and sources
Effect
In Mendelian randomization studies using three gene variants tied to iron status, the odds ratio per standard deviation higher transferrin was 1.08 for coronary artery disease (95% CI 1.01 to 1.16) and 0.92 for stroke (0.86 to 0.99). The authors read a higher transferrin as a sign of lower iron status. The same studies found higher serum iron associated with less coronary artery disease and more stroke [8][9].
Population
Gene associations with iron status from 48,972 people. Coronary artery disease data from 60,801 cases and 123,504 controls combined with a second set of 63,746 cases and 130,681 controls. Stroke data from the MEGASTROKE consortium [8][9].
Sources
[8], [9]
Caveats
The studies measured transferrin, not TIBC. They estimate a lifelong genetic difference in iron status from three variants and do not show what a single reading means. The two outcomes point in opposite directions [8][9].
HDL cholesterol (HDL-C)Strong

Heart attack and death in trials of drugs that raised HDL cholesterol

Not shown to cut heart attacks or deaths when raised by drugs: pooled trials of niacin, fibrates and CETP inhibitors in 117,411 patients found no reduction on statins.

The numbers and sources
Effect
In a meta-analysis of 39 trials, no significant effect on death from any cause was seen for niacin (OR 1.03, 95% CI 0.92 to 1.15), fibrates (0.98, 0.89 to 1.08) or CETP inhibitors (1.16, 0.93 to 1.44). Niacin reduced non-fatal heart attack in trials run before statins (OR 0.69, 95% CI 0.56 to 0.85) and not in trials where statins were taken (0.96, 0.85 to 1.09) [10]. A meta-regression of 108 trials found no association between the change in HDL cholesterol a treatment produced and coronary events or deaths, once the change in LDL cholesterol was accounted for [11]. Among CETP inhibitors, torcetrapib raised HDL cholesterol 72.1% and increased cardiovascular events (HR 1.25, 95% CI 1.09 to 1.44) and death from any cause (HR 1.58, 95% CI 1.14 to 2.19) [12]. Dalcetrapib raised it 31 to 40% with no effect on events (HR 1.04, 95% CI 0.93 to 1.16) [13]. Evacetrapib raised it 133.2% and lowered LDL cholesterol 31.1% with no effect on events (HR 1.01, 95% CI 0.91 to 1.11) [14]. Anacetrapib raised it by 43 mg/dL, lowered non-HDL cholesterol by 17 mg/dL and reduced major coronary events (rate ratio 0.91, 95% CI 0.85 to 0.97) [15]. Adding niacin to a statin raised HDL cholesterol from 35 to 42 mg/dL in one trial (HR 1.02, 95% CI 0.87 to 1.21) and by 6 mg/dL in another (rate ratio 0.96, 95% CI 0.90 to 1.03), without fewer events in either [16][17].
Population
117,411 patients in 39 trials [10]. 299,310 participants in 108 trials [11]. 15,067, 15,871, 12,092 and 30,449 patients at high cardiovascular risk in the four CETP inhibitor trials [12][13][14][15]. 3,414 and 25,673 patients with vascular disease in the two niacin trials [16][17].
Sources
[10], [11], [12], [13], [14], [15], [16], [17]
Caveats
A marker that tracks risk is not always one whose raising changes it. Torcetrapib also raised blood pressure, an effect its investigators considered off-target [12]. The one trial with fewer events also lowered non-HDL cholesterol [15]. Niacin with laropiprant increased serious adverse events [17].
HomocysteineStrong

Heart attack and death in trials that lowered homocysteine

Not shown to fall when homocysteine is lowered: B vitamins cut homocysteine about 25%, and heart attacks and deaths did not differ from placebo across pooled trials.

The numbers and sources
Effect
In a meta-analysis of individual data from 8 placebo-controlled trials, folic acid lowered homocysteine by an average of 25% and had no significant effect on major vascular events (rate ratio 1.01, 95% CI 0.97 to 1.05) or major coronary events (1.03, 95% CI 0.97 to 1.10), overall or in any subgroup studied. Nor did it change overall cancer incidence (1.05, 95% CI 0.98 to 1.13) or death from any cause (1.02, 95% CI 0.97 to 1.08) [7]. A Cochrane review of 15 trials found no difference from placebo in heart attack (RR 1.02, 95% CI 0.95 to 1.10, 12 trials) or death from any cause (RR 1.01, 95% CI 0.96 to 1.06, 11 trials), and judged that further trials were unlikely to change the finding [8].
Population
37,485 people at increased cardiovascular risk in 8 trials, median follow-up 5 years, 9,326 major vascular events [7]. 71,422 participants in 15 trials with follow-up of 1 to 7.3 years [8].
Sources
[7], [8]
Caveats
The pooled trials ran for a median of 5 years in people at increased risk of cardiovascular disease, so they say little about longer treatment or lower-risk people [7].
Alanine aminotransferase (ALT)Moderate

Heart disease and stroke

Not linked to cardiovascular disease overall in a meta-analysis of 29 cohorts, with a slightly lower rate of coronary heart disease and a slightly higher rate of stroke.

The numbers and sources
Effect
A meta-analysis of 29 prospective cohorts found no evidence of an association of ALT with cardiovascular disease. In stratified analyses ALT was somewhat inversely associated with coronary heart disease (relative risk 0.95, 95% CI 0.90 to 1.00) and positively associated with stroke (1.01, 1.00 to 1.02) [15]. In US adults, ALT above 30 U/L in men or 19 U/L in women was not associated with cardiovascular death (HR 0.90, 95% CI 0.56 to 1.4) [11].
Population
Over 1.23 million participants and 20,406 cardiovascular outcomes in 29 cohorts, which also covered other liver enzymes [15]. 14,950 US adults [11].
Sources
[15], [11]
Caveats
Observational. The authors asked for more study of the opposite directions seen for coronary heart disease and stroke [15].
Apolipoprotein A1 (ApoA1)Moderate

Coronary heart disease in genetic studies

Not shown to protect against coronary disease in genetic studies: gene variants that raise ApoA1 did not go with less disease once ApoB was accounted for.

The numbers and sources
Effect
In a Mendelian randomization analysis using a variant in the APOA1 gene region, each standard deviation higher ApoA1 had an odds ratio for coronary artery disease of 1.13 (95% CI 0.98 to 1.30), while the observational hazard ratio in the same paper was 0.81 (0.75 to 0.88) [12]. In a second analysis, the odds ratio per standard deviation higher ApoA1 was 0.83 (95% CI 0.77 to 0.89) when ApoA1 was analyzed alone and 0.94 (0.76 to 1.17) after adjustment for ApoB, which itself kept an odds ratio of 1.68 (1.54 to 1.84) [13].
Population
Genetic data on 20,370 Finnish participants combined with an earlier genome-wide study, applied to 122,733 coronary cases, and 918 incident cases among 11,535 people for the observational estimate [12]. Genetic data from up to 441,016 UK Biobank participants, applied to 60,801 coronary cases and 123,504 controls [13].
Sources
[12], [13]
Caveats
Gene variants act from birth, so these estimates describe lifelong differences, not a change made later in life. The authors of one analysis note that its single-trait estimates weakened under methods more robust to variants that act through several traits [13].
Aspartate aminotransferase (AST)Moderate

Heart disease and stroke

Not linked to cardiovascular disease in a meta-analysis of 29 general population cohorts.

The numbers and sources
Effect
A meta-analysis of 29 prospective cohorts found no evidence of an association of AST with cardiovascular disease, defined as coronary heart disease, stroke or both. Two other enzymes in the same analysis, GGT and alkaline phosphatase, were associated with it [7].
Population
Over 1.23 million participants and 20,406 cardiovascular outcomes across the 29 cohorts, which also covered other liver enzymes [7].
Sources
[7]
Caveats
Observational. The abstract gives no pooled estimate for AST [7].
Cystatin CModerate

Cardiovascular disease in genetic studies

Not shown to cause cardiovascular disease in a genetic study: people with a variant that lowers cystatin C had the same risk.

The numbers and sources
Effect
In a Mendelian randomization analysis using a common variant in the cystatin C gene, each copy of which lowered serum cystatin C by 6.13%, the causal relative risk of cardiovascular disease was 1.00 per doubling of cystatin C (95% CI 0.82 to 1.22). This differed significantly from the observational estimate of 1.82. No causal effect was detected for coronary heart disease, ischemic stroke or heart failure taken singly [16].
Population
Genetic data from 43 studies with 252,216 participants and 63,292 cardiovascular events [16].
Sources
[16]
Caveats
One analysis using a single variant that explained 2.8% of the variation in cystatin C. It tests cystatin C itself as a cause and does not test kidney function, which cystatin C is measured to reflect [16].
HDL cholesterol (HDL-C)Moderate

Heart attack in genetic studies

Not shown to protect against heart attack in genetic studies: gene variants that raise HDL cholesterol alone did not lower risk.

The numbers and sources
Effect
Carriers of a variant in the endothelial lipase gene had HDL cholesterol 0.14 mmol/L higher, with other risk factors similar. That difference predicted a 13% lower risk of heart attack (OR 0.87), and the observed odds ratio was 0.99 (95% CI 0.88 to 1.11). A one standard deviation rise in HDL cholesterol from a 14-variant genetic score was also not associated with heart attack (OR 0.93, 95% CI 0.68 to 1.26). The same method applied to LDL cholesterol gave an odds ratio of 2.13 (95% CI 1.69 to 2.69) per standard deviation, in line with the observational estimate [9].
Population
20,913 heart attack cases and 95,407 controls in 20 studies for the single variant, and up to 12,482 cases and 41,331 controls for the genetic score [9].
Sources
[9]
Caveats
One Mendelian randomization study. Its authors concluded that some genetic mechanisms that raise HDL cholesterol do not seem to lower risk, which does not rule out a role for other properties of HDL particles [9].
High-sensitivity C-reactive protein (hs-CRP)Moderate

Coronary heart disease in genetic studies

Not shown to cause coronary heart disease: gene variants that raise CRP did not raise risk in 194,418 people.

The numbers and sources
Effect
In a Mendelian randomization meta-analysis, variants in the CRP gene each changed CRP concentration by up to 30% per allele and were unrelated to other risk factors. The risk ratio for coronary heart disease was 1.00 (95% CI 0.90 to 1.13) per standard deviation higher genetically raised log CRP, against 1.33 (1.23 to 1.43) for the same difference in measured CRP in prospective studies (P = 0.001 for the difference). The authors concluded that CRP concentration itself is unlikely to be even a modest causal factor in coronary heart disease [9].
Population
194,418 participants in 47 studies in 15 countries, including 46,557 with coronary heart disease [9].
Sources
[9]
Caveats
The finding concerns CRP itself. It does not test whether the inflammation that CRP reflects is a cause [9].
Apolipoprotein A1 (ApoA1)Limited

Heart attack, stroke and cardiovascular death in a trial that infused ApoA1

One trial found four weekly infusions of human ApoA1 after a heart attack did not significantly lower heart attack, stroke or cardiovascular death against placebo.

The numbers and sources
Effect
In the AEGIS-II trial, the composite of heart attack, stroke or cardiovascular death occurred in 4.8% of patients given ApoA1 (CSL112) and 5.2% given placebo at 90 days (HR 0.93, 95% CI 0.81 to 1.05, P = 0.24). At 180 days it was 6.9% against 7.6% (HR 0.91, 0.81 to 1.01), and at 365 days 9.8% against 10.5% (HR 0.93, 0.85 to 1.02). Adverse events were similar, with more hypersensitivity events reported with ApoA1 [14]. In a meta-analysis of people on statins, a rise in ApoA1 after starting treatment went with fewer major cardiovascular events (HR 0.93 per standard deviation, 95% CI 0.90 to 0.97), while a rise in HDL cholesterol did not (0.98, 0.94 to 1.01) [11].
Population
18,219 patients with acute myocardial infarction, multivessel coronary disease and other risk factors, given four weekly infusions of 6 g or placebo [14]. 38,153 trial participants allocated to a statin [11].
Sources
[14], [11]
Caveats
The trial tested a short course of infused ApoA1 soon after a heart attack and was funded by the maker of the product. It did not test a lasting rise in a person's own ApoA1 [14]. The statin finding compares people within trials and is observational [11].
Why limited
This rests on one trial, of 18,219 people. Strength counts independent studies that agree, so one trial is limited at any size.
HematocritLimited

Death and heart attack in a dialysis trial that raised hematocrit to 42%

One trial found raising hematocrit to 42% with epoetin did not cut deaths or heart attacks in hemodialysis patients with heart disease, and the trial was halted.

The numbers and sources
Effect
Patients were assigned to epoetin doses aimed at a hematocrit of 42% or of 30%. After 29 months there were 183 deaths and 19 first nonfatal heart attacks in the 42% group and 150 deaths and 14 nonfatal heart attacks in the 30% group, a risk ratio of 1.3 (95% CI 0.9 to 1.9). The difference did not reach the preset stopping boundary, and the study was halted. Within each group, death rates fell as the hematocrit reached rose [11].
Population
1,233 people on hemodialysis with heart failure or ischemic heart disease, treated for a median of 14 months [11].
Sources
[11]
Caveats
One trial in people on dialysis with heart disease, in whom hematocrit was raised with a drug. The higher target group also had a decline in the adequacy of dialysis and more intravenous iron, so the trial cannot separate hematocrit from the treatment used to raise it [11].
Why limited
This rests on one trial, of 1,233 people. Strength counts independent studies that agree, so one trial is limited at any size.
NT-proBNPLimited

Treatment steered by the level in heart failure

One trial found that adjusting heart failure drugs toward an NT-proBNP below 1,000 pg/mL did not reduce hospital stays or cardiovascular deaths against usual care.

The numbers and sources
Effect
Patients with heart failure and reduced ejection fraction were assigned to therapy titrated toward an NT-proBNP below 1,000 pg/mL or to usual care. Over a median 15 months, a first heart failure hospitalization or cardiovascular death occurred in 37% of each group (adjusted HR 0.98, 95% CI 0.79 to 1.22). Cardiovascular death occurred in 12% and 13%. The falls in NT-proBNP achieved did not differ significantly between the groups, and the trial was stopped early for futility [9].
Population
894 patients at 45 sites in the United States and Canada, median age 63, with an ejection fraction of 40% or below and a recent heart failure event [9].
Sources
[9]
Caveats
The two groups reached similar NT-proBNP levels, so the trial compared two strategies of care more than two levels of the marker. Its abstract notes that earlier, smaller trials of guided therapy had inconsistent results [9].
Parathyroid hormone (PTH)Limited

Death and cardiovascular events in a trial of cinacalcet in dialysis patients

One trial found cinacalcet, a drug for raised PTH, did not significantly reduce death or cardiovascular events in dialysis patients: 48.2% against 49.2% on placebo.

The numbers and sources
Effect
In patients on hemodialysis with moderate to severe secondary hyperparathyroidism, the composite of death, heart attack, hospitalization for unstable angina, heart failure or a peripheral vascular event occurred in 938 of 1,948 (48.2%) on cinacalcet and 952 of 1,935 (49.2%) on placebo, a relative hazard of 0.93 (95% CI 0.85 to 1.02, P = 0.11). Low blood calcium and gastrointestinal adverse events were significantly more frequent on cinacalcet [14].
Population
3,883 patients with a median intact PTH of 693 pg/mL, followed for up to 64 months [14].
Sources
[14]
Caveats
A marker that tracks an outcome is not always one whose correction changes it. This is the unadjusted intention-to-treat result, and the trial was funded by the drug's maker. It studied people with kidney failure and very high PTH, so it says nothing about results near the reference range [14].
Why limited
This rests on one trial, of 3,883 people. Strength counts independent studies that agree, so one trial is limited at any size.
Total testosteroneLimitedStudied in men

Heart attack and stroke in trials that raised testosterone

One trial found raising testosterone in men with levels below 300 ng/dL did not change heart attack, stroke or cardiovascular death over 33 months.

The numbers and sources
Effect
In 5,246 men aged 45 to 80 with symptoms, two fasting testosterone levels below 300 ng/dL and existing or high risk of cardiovascular disease, a first major cardiac event occurred in 7.0% on testosterone gel and 7.3% on placebo (hazard ratio 0.96, 95% CI 0.78 to 1.17), which met the trial's test for noninferiority. Atrial fibrillation, acute kidney injury and pulmonary embolism were more frequent in the testosterone group [22]. In a 1-year imaging substudy of an earlier trial, noncalcified coronary plaque volume grew more with testosterone than with placebo (estimated difference 41 cubic millimeters, 95% CI 14 to 67), with no major cardiovascular events in either group [23].
Population
5,246 men, mean treatment 21.7 months and mean follow-up 33.0 months [22]. 138 men aged 65 or older who completed coronary imaging [23].
Sources
[22], [23]
Caveats
The trial was funded by a maker of testosterone gel and others, and treatment lasted under 2 years on average [22]. In cohorts, very low testosterone went with more cardiovascular deaths [16], which this trial did not reduce.
Why limited
This rests on one trial, of 5,246 people. Strength counts independent studies that agree, so one trial is limited at any size.