Heart Disease and Stroke Statistics—2022 Update: A Report From the American Heart Association

Abbreviations Table e165

About These Statistics e170

Cardiovascular Health e173

Smoking/Tobacco Use e194

Physical Activity and Sedentary Behavior e210

Nutrition e223

Overweight and Obesity e246

High Blood Cholesterol and Other Lipids e260

High Blood Pressure e274

Diabetes e290

Metabolic Syndrome e311

Adverse Pregnancy Outcomes e333

Kidney Disease e354

Sleep e366

Total Cardiovascular Diseases e374

Stroke (Cerebrovascular Diseases) e391

Brain Health e427

Congenital Cardiovascular Defects and Kawasaki Disease e445

Disorders of Heart Rhythm e462

Sudden Cardiac Arrest, Ventricular Arrhythmias, and Inherited Channelopathies e489

Subclinical Atherosclerosis e517

Coronary Heart Disease, Acute Coronary Syndrome, and Angina Pectoris e528

Cardiomyopathy and Heart Failure e547

Valvular Diseases e562

Venous Thromboembolism (Deep Vein Thrombosis and Pulmonary Embolism), Chronic Venous Insufficiency, Pulmonary Hypertension e580

Peripheral Artery Disease and Aortic Diseases e591

Quality of Care e611

Medical Procedures e625

Economic Cost of Cardiovascular Disease e630

At-a-Glance Summary Tables e633

Glossary e637

A report pooled NHANES (National Health and Nutrition Examination Survey) 2011 to 2016 data and individual-level data from 7 US community-based cohort studies and estimated that 70.0% of major CVD events in the United States were attributable to low and moderate CVH; 2.0 million major CVD events could potentially be prevented each year if all US adults attain high CVH; and even a partial improvement in CVH scores to the moderate level among all US adults with low overall CVH could lead to a reduction of 1.2 million major CVD events annually.

The large number of individuals in the United States who contracted severe illness because of coronavirus disease 2019 (COVID-19) resulted in a huge mortality toll. As of March 2021, the cumulative number of COVID-19 deaths in the United States was ≈545 000, which equates to ≈166 cases per 100 000 people, with higher rates of deaths occurring among US counties with metropolitan areas (≈185 deaths per 100 000), with a high percentage (>45.5%) of the population that is non-Hispanic (NH) Black (≈200 deaths per 100 000), with a high proportion (>37%) of the population that is Hispanic (≈219 deaths per 100 000), or with a high percentage (>17.3%) of the population that are living in poverty (≈211 deaths per 100 000 people).

Because of the high COVID-19 mortality rates, life expectancy in the United States for the year 2020 has been estimated to decline with disproportionate impacts on populations with high COVID-19 mortality rates. Provisional US life expectancy estimates for January to June 2020 indicate that between 2019 and the first half of 2020, life expectancy decreased from 74.7 to 72.0 years for NH Black individuals, from 81.8 to 79.9 years for Hispanic individuals, and from 78.8 to 78.0 years for NH White individuals.

The prevalence of cigarette use in the past 30 days among middle and high school students in the United States was 1.6% and 4.6%, respectively, in 2020.

Although there has been a consistent decline in adult and youth cigarette use in the United States in the past 2 decades, significant disparities persist. Substantially higher tobacco use prevalence rates are observed in American Indian/Alaska Native adults and youth and lesbian, gay, and bisexual adults.

Over the past 9 years, there has been a sharp increase in electronic cigarette use among adolescents, increasing from 1.5% to 19.6% between 2011 and 2020; electronic cigarettes are now the most commonly used tobacco product in this demographic.

According to nationwide self-reported PA (YRBSS [Youth Risk Behavior Surveillance System], 2019), the prevalence of high school students who engaged in ≥60 minutes of PA on at least 5 days of the week was 44.1% and was lower with each successive grade (from 9th [49.1%]–12th [40.0%] grades).

From nationwide self-reported PA (NHIS, 2018), the age-adjusted proportion who reported meeting the 2018 aerobic PA guidelines for Americans was 54.2%.

An umbrella review of 24 systematic reviews of adults ≥60 years of age concluded that those who are physically active are at a reduced risk of CVD mortality (25%–40% risk reduction), all-cause mortality (22%–35%), breast cancer (12%–17%), prostate cancer (9%–10%), and depression (17%–31%) while experiencing better quality of life, healthier aging trajectories, and improved cognitive functioning.

Data from the Nurses’ Health Study (1984–2014) and Health Professionals Follow-up Study showed that daily intake of 5 servings of fruit and vegetables (versus 2 servings/d) was associated with 13% lower total mortality, 12% lower CVD mortality, 10% lower cancer mortality, and 35% lower respiratory disease mortality.

NHANES data and meta-analyses of prospective cohort studies show that higher intakes of total fat, polyunsaturated fatty acids, and monounsaturated fatty acids are associated with lower total mortality. However, the evidence for saturated fatty acid intake as a risk or protective factor for total and CVD mortality remains controversial.

Meta-analytic evidence from randomized clinical trials does not support vitamin D supplementation for improving cardiometabolic health in children and adolescents between 4 and 19 years of age.

From NHANES data, the overall prevalence of obesity and severe obesity in youth 2 to 19 years of age increased from 13.9% to 19.3% and 2.6% to 6.1% between 1999 to 2000 and 2017 to 2018. Over the same period, the prevalence of obesity and severe obesity increased from 14.0% to 20.5% and from 3.7% to 6.9% for males and from 13.8% to 18.0% and from 3.6% to 5.2% for females.

From NHANES data, among adults, from 1999 to 2000 through 2017 to 2018, the prevalence of obesity among males increased from 27.5% to 43.0% and severe obesity increased from 3.1% to 6.9%. The prevalence of obesity among females increased from 33.4% to 41.9% and severe obesity from 6.2% to 11.5%.

Significant increases in the prevalence of obesity were seen between 1999 to 2000 through 2017 to 2018 in all age-race and ethnicity groups except for NH Black males, in whom the prevalence increased from 1999 through 2006.

In 2015 to 2018, unfavorable lipid measures of low-density lipoprotein cholesterol ≥130 mg/dL were present in 6.1% of male adolescents and 3.0% of female adolescents 12 to 19 years of age, triglycerides ≥130 mg/dL were present in 9.7% of male adolescents and 6.6% of female adolescents, and high-density lipoprotein cholesterol measures <40 mg/dL were present in 18.4% of male adolescents and 7.4% of female adolescents.

In 2015 to 2018, total cholesterol ≥200 mg/dL was present in 38.1% of adults, low-density lipoprotein cholesterol ≥130 mg/dL was present in 27.8% of adults, triglycerides ≥150 mg/dL were present in 21.1% of adults, high-density lipoprotein cholesterol <40 mg/dL was present in 17.2% of adults.

From 2009 to 2019, the death rate attributable to high BP increased 34.2%, and the actual number of deaths attributable to high BP rose 65.3%.

The 2019 age-adjusted death rate attributable primarily to high BP was 25.1 per 100 000 people. Age-adjusted death rates attributable to high BP (per 100 000 people) in 2019 were 25.7 for NH White males, 56.7 for NH Black males, 23.1 for Hispanic males, 17.4 for NH Asian/Pacific Islander males, 31.9 for NH American Indian/Alaska Native males, 20.6 for NH White females, 38.7 for NH Black females, 17.4 for Hispanic females, 14.5 for NH Asian/Pacific Islander females, and 22.4 for NH American Indian/Alaska Native females.

In an analysis of 18 262 adults ≥18 years of age with hypertension (defined as 140/90 mm Hg) in NHANES, the estimated age-adjusted proportion with controlled BP increased from 31.8% in 1999 to 2000 to 48.5% in 2007 to 2008, remained relatively stable at 53.8% in 2013 to 2014, but declined to 43.7% in 2017 to 2018.

In NHANES 2015 to 2018, an estimated 28.2 million adults (10.4%) had diagnosed diabetes, 9.8 million adults (3.8%) had undiagnosed diabetes, and 113.6 million adults (45.8%) had prediabetes.

In NHANES 2003 through 2016, among adults with diagnosed and undiagnosed diabetes, the proportion taking any medication increased from 58% in 2003 through 2004 to 67% in 2015 through 2016, with an increase in the use of metformin and insulin analogs and decrease in sulfonylureas, thiazolidinediones, and human insulin.

In NHANES 1988 through 2018, among adults with newly diagnosed type 2 diabetes, there was a significant increase in the proportion of individuals with hemoglobin A1c <7% (59.8% for 1998–1994 and 73.7% for 2009–2018) and decreases in mean hemoglobin A1c (7.0% and 6.7%), mean BP (130.1/77.5 and 126.0/72.1 mm Hg), and mean total cholesterol (219.4 and 182.4 mg/dL). The proportion with hemoglobin A1c <7.0%, BP <140/90 mm Hg, and total cholesterol <240 mg/dL improved from 31.6% to 56.2%.

In the HELENA study (Healthy Lifestyle in Europe by Nutrition in Adolescence) among 1037 European adolescents 12.5 to 17.5 years of age, those with mothers with low education showed a higher metabolic syndrome (MetS) risk score (β estimate, 0.54) compared with those with highly educated mothers. Adolescents who accumulated >3 disadvantages (defined as parents with low education, low family affluence, migrant origin, unemployed parents, or nontraditional families) had a higher MetS risk score compared with those who did not experience disadvantage (β estimate, 0.69).

In HCHS/SOL (Hispanic Community Health Study/Study of Latinos), socioeconomic status was inversely associated with prevalent MetS among Hispanic/Latino adults of diverse ancestry groups. Higher income and education and full-time employment status versus unemployed status were associated with a 4%, 3%, and 24% decreased odds of having MetS, respectively. The association with income was significant only among females and those with current health insurance.

In combined analysis from ARIC (Atherosclerosis Risk in Communities) and JHS (Jackson Heart Study), among 13 141 White and Black individuals with a mean follow-up of 18.6 years, risk of ischemic stroke increased consistently with MetS severity z score (hazard ratio [HR], 1.75) for those above the 75th percentile compared with those below the 25th percentile. Risk was highest for White females (HR, 2.63), although without significant interaction by sex and race.

Adverse pregnancy outcomes (including hypertensive disorders of pregnancy, gestational diabetes, preterm birth, and small for gestational age at birth) occur in 10% to 20% of pregnancies.

Among 2304 female-newborn dyads in the multinational HAPO study (Hyperglycemia and Adverse Pregnancy Outcome), lower CVH (based on 5 metrics: body mass index, BP, cholesterol, glucose, and smoking) at 28 weeks’ gestation was associated with a higher risk of preeclampsia; adjusted relative risks were 3.13, 5.34, and 9.30 for females with ≥1 intermediate, 1 poor, or ≥2 poor (versus all ideal) CVH metrics during pregnancy, respectively.

In analyses of Swedish national birth register data (>2 million–>4 million individuals), gestational age at birth was inversely associated with the risks for type 1 diabetes, type 2 diabetes, hypertension, and lipid disorders among individuals born preterm versus term.

Overall prevalence of chronic kidney disease (estimated glomerular filtration rate <60 mL·min⁻¹·1.73 m⁻²or albumin-to-creatinine ratio ≥30 mg/g) was 14.9% (2015–2018).

Age-, race-, and sex-adjusted prevalence of end-stage renal disease in the United States was 2242 per million people (in 2018) with highest rates among Black adults followed by American Indian/Alaska Native adults, Asian adults, and White adults.

Medicare spent $81 billion caring for people with chronic kidney disease and $49.2 billion on those with end-stage renal disease in 2018.

In data from the 2014 BRFSS (Behavioral Risk Factor Surveillance System), 11.8% of people reported a sleep duration ≤5 hours, 23.0% reported 6 hours, 29.5% reported 7 hours, 27.7% reported 8 hours, 4.4% reported 9 hours, and 3.6% reported ≥10 hours. Overall, 65.2% met the recommended sleep duration of ≥7 hours.

Analysis of the UK Biobank study (N=468 941) found that participants who reported short sleep (<7 h/d) or long sleep (>9 h/d) had an increased risk of incident HF compared with normal sleepers (7–9 h/d). In males, the adjusted HR was 1.24 for short sleep and 2.48 for long sleep. In females, the adjusted HR was 1.39 for short sleep and 1.99 for long sleep.

A meta-analysis of 15 prospective studies observed a significant association between the presence of obstructive sleep apnea and the risk of cerebrovascular disease (HR, 1.94).

In the Cardiovascular Lifetime Risk Pooling Project among 30 447 participants from 7 US cohort studies, among individuals ≥60 years of age with low CVH, the 35-year risk of CVD was highest in White males (65.5%), followed by White females (57.1%), Black females (51.9%), and Black males (48.4%). These estimated risks accounted for competing risks of death caused by non-CVD causes.

In a meta-analysis of 14 studies that focused on CVD among individuals diagnosed with COVID-19, preexisting CVD had a relative risk of 2.25 for death resulting from COVID-19.

In 2020, ≈19 million deaths were attributed to CVD globally, which amounted to an increase of 18.7% from 2010.

In the Greater Cincinnati Northern Kentucky Stroke Study, sex-specific ischemic stroke incidence rates declined significantly between 1993 to 1994 and 2015 for both males and females. In males, there was a decline from 282 to 211 per 100 000. In females, the decline was from 229 to 174 per 100 000. This trend was not observed for intracerebral hemorrhage or subarachnoid hemorrhage.

In the Northern Manhattan Study, among 3298 stroke-free participants followed up through 2019, Black and Hispanic females ≥70 years of age had higher risk of stroke compared with White females after controlling for age, sex, education, and insurance status (Black females/White females: HR, 1.76; Hispanic females/White females: HR, 1.77). This increased risk was not present among elderly Black or Hispanic males compared with White males.

A systematic analysis of data from the GBD study (Global Burden of Disease) showed that, in 2017, Alzheimer disease/Alzheimer disease and related dementia was the fourth most prevalent neurological disorder in the United States (2.9 million people). Among neurological disorders, Alzheimer disease/Alzheimer disease and related dementia was the leading cause of mortality in the United States (38 deaths per 100 000 population per year) ahead of stroke.

In 2017, Alzheimer disease/Alzheimer disease and related dementia had the fifth leading incidence rate of neurological disorders in the United States according to the GBD study data. The US age-standardized incidence rate of Alzheimer disease/Alzheimer disease and related dementia was 85 cases per 100 000 people).

In a meta-analysis of 12 randomized controlled trials (>92 000 participants; mean age, 69 years; 42% females), BP lowering with antihypertensive agents, compared with control, was associated with a lower risk of incident dementia or cognitive impairment (7.0% versus 7.5% of patients over a mean trial follow-up of 4.1 years; odds ratio [OR], 0.93; absolute risk reduction, 0.39%).

The 2017 all-age prevalence of congenital cardiovascular defects in the United States was estimated at 466 566 individuals, with 279 320 (60%) of these under the age of <20 years of age. The 2017 global prevalence of congenital cardiovascular defects was estimated at 157 per 100 000 people. with the highest prevalence estimates in countries with a low sustainable development index (238 per 100 000 people) and the lowest in those with a high-middle or high sustainable development index (112 and 135 per 100 000 people, respectively).

Congenital cardiovascular defects appear to be more common among infants born to mothers with low socioeconomic status. In Ontario, mothers who lived in the lowest-income neighborhoods had higher risk of having an infant with a congenital cardiovascular defect compared with mothers living in the highest-income neighborhoods (OR, 1.29). Furthermore, this discrepancy between low and high was also found across measures of neighborhood education (OR, 1.34) and employment rate (OR, 1.18).

Since May 2020, the Centers for Disease Control and Prevention has been tracking reports of multisystem inflammatory syndrome in children. As of June 28, 2021, 4196 cases and 37 attributable deaths (0.89%) have been reported. Median age of cases was 9 years; 62% of cases have occurred in children who are Hispanic or Latino (1246 cases) or Black (1175 cases); 99% tested positive for severe acute respiratory syndrome coronavirus 2 (reverse transcription–polymerase chain reaction, serology, or antigen test); and 60% of reported patients were male.

A systematic review and meta-analysis of 18 published studies reported short-term and long-term associations of air pollution with atrial fibrillation (AF). For 10-mg/m³increases in PM_2.5and PM₁₀concentrations, the OR of AF was 1.01 and 1.03, respectively. The corresponding ORs for long-term exposure were 1.07 for PM_2.5and 1.03 for PM₁₀. SO₂and NO₂were also associated with AF in the short term: ORs for 10-ppb increments were 1.05 and 1.03, respectively.

A multicenter, open-label, randomized trial evaluated a 2-week continuous electrocardiographic patch and an automated home BP machine with oscillometric AF screening capability for the detection of AF compared with usual care over a 6-month period in participants ≥75 years of age with hypertension. AF detection was 5.3% in the screening group compared with 0.5% in the control group (risk difference, 4.8%; number needed to screen, 21). By 6 months, anticoagulation was more frequently prescribed in the intervention group (4.1% versus 0.9%; risk difference, 3.2%).

AF has been associated with increased mortality in patients with COVID-19. A meta-analysis of studies published in 2020, including 23 studies and 108 745 patients with COVID-19, showed that AF was associated with increased mortality (pooled effect size, 1.14).

There was a 119% increase in out-of-hospital cardiac arrest during the pandemic compared with earlier control periods in a meta-analysis in 10 countries. For the patients with known outcomes (n=10 992), mortality was 85% compared with 62% for the control periods.

Coinciding with timing of the pandemic in the United States, CARES Registry (Cardiac Arrest Registry to Enhance Survival) data indicate increased delays to initiation of cardiopulmonary resuscitation for out-of-hospital cardiac arrest and reduced survival after out-of-hospital cardiac arrest. Accompanying these effects were reductions in the frequency of shockable rhythms, out-of-hospital cardiac arrest in public locations, and bystander automated external defibrillator use, whereas field termination of resuscitation efforts increased. There was no significant alteration in frequency of bystander cardiopulmonary resuscitation.

Survival to hospital discharge was 22.4% of 33 874 adult pulseless in-hospital cardiac arrests at 328 hospitals in Get With The Guidelines 2020 data. Among survivors, 79.5% had good functional status (Cerebral Performance Category 1 or 2) at hospital discharge.

In 3116 MESA (Multi-Ethnic Study of Atherosclerosis) participants (58±9 years of age, 63% females) who had no detectable coronary artery calcification (CAC) at baseline and were followed up over 10 years, CAC score >0, CAC score >10, and CAC score >100 were seen in 53%, 36%, and 8% of individuals at 10 years, respectively.

In a study with 12.3 years of mean follow-up, cancer-related mortality was 1.55-fold higher in individuals who had a CAC score ≥1000 at baseline compared with those who had a CAC score of 0 at baseline, after adjustment for age, sex, and risk factors.

In 9388 US and Finnish individuals with longitudinal measurement of CVD risk factors and carotid intima-media thickness, CVH declined from childhood to adulthood and was associated with thickening of the intima-media thickness.

In a European registry of high-volume percutaneous coronary intervention centers, the COVID-19 pandemic was associated with a significant increase in door-to-balloon and total ischemia times. Door-to-balloon time >30 minutes was 57.0% in the period of March to April 2020 compared with 52.9% in March to April 2019 (P=0.003), whereas total ischemia time >12 hours was 11.7% in the 2020 period compared with 9.1% in 2019 (P=0.001).

In a retrospective cohort study of Medicare fee-for-service patients (N=453 783) who were diagnosed with coronary artery disease, patients that received care at the most socioeconomically deprived practices had higher odds of being admitted for unstable angina (adjusted OR, 1.46) and higher 30-day mortality rates after acute myocardial infarction (adjusted OR, 1.31). After additional adjustment for patient-level area deprivation index, these associations were attenuated (unstable angina adjusted OR, 1.20; 30-day mortality after myocardial infarction adjusted OR, 1.31).

A pooled analysis of 21 randomized percutaneous coronary intervention trials including 32 877 patients (28% females) found that female sex was an independent risk factor for major adverse cardiovascular events (HR, 1.14) and ischemia-driven target lesion vascularization (HR, 1.23) but not of all-cause or cardiovascular mortality (HR, 0.91 and 0.97, respectively).

The lifetime risk of HF remains high, with variation across racial and ethnic groups ranging from 20% to 45% after 45 years of age.

Secular trends show that the incidence of HF with preserved ejection fraction is increasing and, in contrast, the incidence of HF with reduced ejection fraction is decreasing, whereas both HF subtypes have similar all-cause mortality rates.

Contemporary HF with reduced ejection fraction guideline-directed medical therapy is estimated to reduce the hazard of cardiovascular death or HF hospitalization by up to 62% compared with limited conventional therapy.

The number of elderly patients with calcific aortic stenosis is projected to more than double by 2050 in both the United States and Europe according to a simulation model in 7 decision analysis studies.

The pooled prevalence of all aortic stenosis in the elderly is 12.4%, and the prevalence of severe aortic stenosis is 3.4%. The annual volume of transcatheter aortic valve replacement (TAVR) has increased each year since 2011. After the US Food and Drug Administration approval of TAVR for low-risk patients in 2019, the TAVR volume exceeded all forms of surgical aortic valve replacement (n=72 991 versus n=57 626). From 2011 through 2018, extreme- and high-risk patients remained the largest cohort undergoing TAVR, but in 2019, the intermediate-risk cohort was the largest, and low-risk patients with a median 75 years of age increased to 8395, making up 11.5% of all patients undergoing TAVR.

In 2018, there were an estimated ≈1 015 000 total venous thromboembolism cases in the United States.

In addition, 2019 data show that 37 571 deaths (any mention) resulted from pulmonary embolism and 27 574 deaths (any mention) resulted from pulmonary hypertension.

In the COVID-19 scenario, the incidence of venous thromboembolism was up to 31% in hospitalized patients. Among them, those who were admitted to the intensive care unit had a 2- to 3-fold greater risk of developing deep vein thrombosis or pulmonary embolism.

From 2011 to 2019, the global prevalence of peripheral artery disease was 5.56% with a higher prevalence in high- compared with low- to middle-income countries (7.37% versus 5.09%, respectively). In 2015, it was estimated that 236.62 million people ≥25 years of age were living with peripheral artery disease.

In an analysis of 393 017 patients who underwent lower extremity arterial revascularization, 50 750 (12.9%) had at least 1 subsequent hospitalization for major adverse limb events.

In a population-based screening study of 14 989 participants 60 to 74 years of age, male sex (OR, 1.9), hypertension (OR, 1.8), and family history (OR, 1.6) were associated with a heightened risk of ascending thoracic aortic aneurysm. Diabetes was associated with a lower risk (OR, 0.8).

Compared with 2019, a lower proportion of cases received bystander cardiopulmonary resuscitation in 2020, and use of automated external defibrillators was lower. There were also longer emergency medical services response times and lower survival to hospital discharge. Those are likely related to the COVID-19 pandemic.

In a Get With The Guidelines–HF study, inclusion in Medicare Advantage led to a higher proportion of discharge to home with no difference in mortality compared with fee-for-service programs.

In data from the PINNACLE Registry (Practice Innovation and Clinical Excellence), only about two-thirds of the individuals were treated with appropriate statin therapy as recommended in the American College of Cardiology/AHA guidelines. In addition, higher income was associated with higher likelihood of appropriate statin therapy.

As per the Society of Thoracic Surgeons/American College of Cardiology transcatheter valve therapy registry data, TAVR volumes continue to grow, with 13 723 TAVR procedures in 2011 to 2013 and 72 991 TAVR procedures in 2019. In 2019, 669 sites were performing TAVR. In 2019, TAVR volumes (n=72 991) exceeded the volumes for all forms of surgical aortic valve replacement (n=57 626).

In 2020, 3658 heart transplantations were performed in the United States, the most ever. The highest number of heart transplantations were performed in the states of California (496), Texas (302), Florida (288), and New York (250).

A global survey of 909 inpatient and outpatient centers performing cardiovascular diagnostic procedures in 108 countries compared procedural volumes for common cardiovascular diagnostic procedures between March 2019 and March 2020/April 2020. This survey indicated that cardiovascular diagnostic procedures decreased by 64% from March 2019 to April 2020.

The average annual direct and indirect cost of CVD in the United States was an estimated $378.0 billion in 2017 to 2018.

The estimated direct costs of CVD in the United States increased from $103.5 billion in 1996 to 1997 to $226.2 billion in 2017 to 2018.

By event type, hospital inpatient stays accounted for the highest direct cost ($99.6 billion) in 2017 to 2018 in the United States.

ACC NCDR’s Chest Pain–MI Registry (formerly the ACTION Registry)—quality information for AMI

ARIC—CHD and HF incidence rates

BRFSS—ongoing telephone health survey system

GBD—global disease prevalence, mortality, and healthy life expectancy

GCNKSS—stroke incidence rates and outcomes within a biracial population

GWTG—quality information for resuscitation, HF, and stroke

HCUP—hospital inpatient discharges and procedures

MEPS—data on specific health services that Americans use, how frequently they use them, the cost of these services, and how the costs are paid

NAMCS—physician office visits

NHAMCS—hospital outpatient and ED visits

NHANES—disease and risk factor prevalence and nutrition statistics

NHIS—disease and risk factor prevalence

NVSS—mortality for the United States

USRDS—kidney disease prevalence

WHO—mortality rates by country

YRBS—health-risk behaviors in youth and young adults

Multiple independent investigations (summaries of which are provided in this chapter) have confirmed the importance of having ideal levels of these components, along with the overall concept of CVH. Findings include strong inverse, stepwise associations in the United States of the number of CVH components at ideal levels with all-cause mortality, CVD mortality, IHD mortality, CVD, and HF; with subclinical measures of atherosclerosis such as carotid IMT, arterial stiffness, and CAC prevalence and progression; with physical functional impairment and frailty; with cognitive decline and depression; and with longevity.^3–8Similar relationships have also been seen in non-US populations.^3,4,9–22

A large Hispanic/Latino cohort study in the United States confirmed the associations between CVD and status of CVH components in this population and found that the levels of CVH components compared favorably with existing national estimates; however, some of the associations varied by sex and heritage.⁴

A study of Black people found that risk of incident HF was 61% lower among those with ≥4 ideal CVH components than among those with 0 to 2 ideal components.⁵

Ideal health behaviors and ideal health factors are each independently associated with lower CVD risk in a stepwise fashion; across any level of health behaviors, health factors are associated with incident CVD, and conversely, across any level of health factors, health behaviors are associated with incident CVD.²³

Analyses from the US Burden of Disease Collaborators demonstrated that poor levels of each of the 7 CVH components resulted in substantial mortality and morbidity in the United States in 2010. The leading risk factor related to overall disease burden was suboptimal diet, followed by tobacco smoking, high BMI, raised BP, high FPG, and physical inactivity.²⁴

A meta-analysis of 9 prospective cohort studies involving 12 878 participants reported that having the highest number of ideal CVH components was associated with a lower risk of all-cause mortality (RR, 0.55 [95% CI, 0.37–0.80]), cardiovascular mortality (RR, 0.25 [95% CI, 0.10–0.63]), CVD (RR, 0.20 [95% CI, 0.11–0.37]), and stroke (RR, 0.31 [95% CI, 0.25–0.38]) compared with having the lowest number of ideal components.²⁵

The adjusted PAFs for CVD mortality for individual components of CVH have been reported as follows²⁶:

Several studies have been published in which investigators have assigned individuals a CVH score ranging from 0 to 14 on the basis of the sum of points assigned to each component of CVH (poor=0, intermediate=1, ideal=2 points). With this approach, data from the REGARDS cohort were used to demonstrate an inverse stepwise association between a higher CVH score component and a lower incidence of stroke. On the basis of this score, every unit increase in CVH was associated with an 8% lower risk of incident stroke (HR, 0.92 [95% CI, 0.88–0.95]), with a similar effect size for White (HR, 0.91 [95% CI, 0.86–0.96]) and Black (HR, 0.93 [95% CI, 0.87–0.98]) participants.²⁷CVH score and components were also shown to predict MACEs (first occurrence of MI, stroke, acute ischemic syndrome, coronary revascularization, or death) over a median follow-up of 12 years in a biracial community-based population.²⁸

By combining the 7 CVH component scores and categorizing the total score to define overall CVH (low, 0–8 points; moderate, 9–11 points; high, 12–14 points), a report pooled NHANES 2011 to 2016 data and individual-level data from 7 US community-based cohort studies to estimate the age-, sex-, and race and ethnicity–adjusted PAF of major CVD events (nonfatal MI, stroke, HF, or CVD death) associated with CVH and found that 70.0% (95% CI, 56.5%–79.9%) of major CVD events in the United States were attributable to low and moderate CVH.²⁹According to the authors’ estimates, 2.0 (95% CI, 1.6–2.3) million major CVD events could potentially be prevented each year if all US adults attain high CVH, and even a partial improvement in CVH scores to the moderate level among all US adults with low overall CVH could lead to a reduction of 1.2 (95% CI, 1.0–1.4) million major CVD events annually.

A report from the Framingham Offspring Study showed increased risks of subsequent hypertension, diabetes, CKD, CVD, and mortality associated with having a shorter duration of ideal CVH in adulthood.³⁰

The Cardiovascular Lifetime Risk Pooling Project showed that adults with all optimal risk factor levels (similar to having ideal CVH factor levels of cholesterol, blood sugar, and BP, as well as not smoking) have substantially longer overall and CVD-free survival than those who have poor levels of ≥1 of these CVH factors. For example, at an index age of 45 years, males with optimal risk factor profiles lived on average 14 years longer free of all CVD events and 12 years longer overall than people with ≥2 risk factors.³¹

Better CVH as defined by the AHA is associated with lower incidence of HF,^3,5–7,22less subclinical vascular disease,^{8,15,17,33,34}better global cognitive performance and cognitive function,^16,35,36lower hazard of subsequent dementia,^37,38lower prevalence³⁹and incidence⁴⁰of depressive symptoms, lower loss of physical functional status,⁴¹longer leukocyte telomere length,⁴²less ESRD,⁴³less pneumonia, less chronic obstructive pulmonary disease,⁴⁴less VTE/PE,⁴⁵lower prevalence of aortic sclerosis and stenosis,⁴⁶lower risk of calcific aortic valve stenosis,⁴⁷better prognosis after MI,⁴⁸lower risk of AF,⁴⁹and lower odds of having elevated resting heart rate.⁵⁰Using the CVH scoring approach, the FHS demonstrated significantly lower odds of prevalent hepatic steatosis associated with more favorable CVH scores, and the decrease of liver fat associated with more favorable CVH scores was greater among people with a higher GRS for NAFLD.⁵¹In addition, a study based on NHANES data showed significantly decreased odds of ocular diseases (OR, 0.91 [95% CI, 0.87–0.95]), defined as age-related macular degeneration, any retinopathy, and cataract or glaucoma, and odds of diabetic retinopathy (OR, 0.71 [95% CI, 0.66–0.76]) associated with each unit increase in CVH among US adults.⁵²

In addition, a study among a sample of Hispanic/Latino people residing in the United States reported that greater positive psychological functioning (dispositional optimism) was associated with higher CVH scores as defined by the AHA.⁵³A study in college students found that both handgrip strength and muscle mass were positively associated with greater numbers of ideal CVH components,⁵⁴and a cross-sectional study found that greater cardiopulmonary fitness, upper-body flexibility, and lower-body muscular strength were associated with better CVH components in perimenopausal females.⁵⁵Furthermore, higher quality of life scores were associated with better CVH metrics,⁵⁶providing additional evidence to support the benefits of ideal CVH on general health and quality of life.

According to NHANES 1999 to 2006 data, several social risk factors (low family income, low education level, underrepresented racial groups, and single-living status) were related to lower likelihood of attaining better CVH as measured by Life’s Simple 7 scores.⁵⁷In addition, neighborhood factors and contextual relationships have been found to be related to health disparities in CVH, but more research is needed to better understand these complex relationships.⁵⁸A study focused on people with serious mental illness found that individuals of underrepresented races and ethnicities had significant lower CVH scores based on 5 of the Life’s Simple 7 components.⁵⁹

Having more ideal CVH components in middle age has been associated with lower non-CVD and CVD health care costs in later life.⁶⁰An investigation of 4906 participants in the Cooper Center Longitudinal Study reported that participants with ≥5 ideal CVH components exhibited 24.9% (95% CI, 11.7%–36.0%) lower median annual non-CVD costs and 74.5% (95% CI, 57.5%–84.7%) lower median CVD costs than those with ≤2 ideal CVH components.⁶⁰A report from a large, ethnically diverse insured population found that people with 6 or 7 and those with 3 to 5 of the CVH components in the ideal category had a $2021 and $940 lower annual mean health care expenditure, respectively, than those with 0 to 2 ideal health components.⁶¹

The national prevalence estimates for children (12–19 years of age) and adults (≥20 years of age) who meet ideal, intermediate, and poor levels of each of the 7 CVH components are displayed in Chart 2-1.⁶²The most current estimates at the time of publication were based on data from NHANES 2017 to 2018. NHANES 2017 to 2018 survey changed the PA assessments for children, so the PA status for children was updated according to data from respondents who were 18 to 19 years of age.

For most components of CVH, prevalence of ideal levels is higher in US children (12–19 years of age) than in US adults (≥20 years of age), except for the Healthy Diet Score, for which prevalence of ideal levels in children is lower than in adults. For PA, the contrast for adults versus children is not clear because the prevalence estimate for children was from a subgroup of children only.

Among US children (12–19 years of age; Chart 2-1), the unadjusted prevalence of ideal levels of CVH components currently varies from <1% for the Healthy Diet Score (ie, <1 in 100 US children meets at least 4 of the 5 dietary components) to >79% for smoking, BP, and diabetes components (95.7%, 89.1%, and 79.0% respectively; unpublished AHA tabulation).

Among US adults (Chart 2-1), the lowest prevalence of ideal levels for CVH components is <1% for the Healthy Diet Score in adults ≥20 years of age. The highest prevalence of ideal levels for a CVH component is for smoking (79.8% of adults report never having smoked or being a former smoker who has quit for >12 months). In 2017 to 2018, 52.4% of adults had ideal levels of TC (<200 mg/dL).

Age-standardized and age-specific prevalence estimates for ideal CVH and for ideal levels of individual CVH components for 2017 to 2018 are displayed in Table 2-2.

In 2017 to 2018, all individual components of CVH among adults were highest in the youngest age groups (20–39 years of age) and were lowest in the oldest age group (≥60 years of age), except smoking and the Healthy Diet Score, for which prevalence of ideal levels was highest in older adults. For the Healthy Diet Score, all age groups had a prevalence of ideal level <1% according to the 2017 to 2018 NHANES data.

Chart 2-2 displays the unadjusted prevalence estimates of ideal levels of CVH components for the population of US children (12–19 years of age) by race and ethnicity.

Chart 2-3 displays the adjusted prevalence estimates of ideal levels of CVH components for the population of US adults ≥20 years of age by race and ethnicity.

The trends in prevalence of meeting ideal criteria for the individual components of CVH from 1999 to 2000 to 2017 to 2018 (for diet, trends from 2003–2004 through 2017–2018) are shown in Chart 2-4 for children (12–19 years of age) and in Chart 2-5 for adults (≥20 years of age).

The leading risk factors for YLLs from 1990 to 2019 in the United States are presented in Table 2-3.

The leading causes of YLLs from 1990 to 2019 in the United States are presented in Table 2-4.

The leading causes of YLDs from 1990 to 2019 in the United States are presented in Table 2-6.

The leading global YLL risk factors from 1990 to 2019 are presented in Table 2-7.

The leading global YLL causes from 1990 to 2019 are presented in Table 2-8.

The leading global risk factors for YLDs from 1990 to 2019 are presented in Table 2-9.

The leading global causes of YLDs from 1990 to 2019 are presented in Table 2-10.

The large number of individuals in the United States who contracted severe illness attributable to COVID-19 resulted in a huge mortality toll, with disproportionate rates of deaths occurring among US counties with metropolitan areas and with higher proportions of the population who are NH Black and Hispanic people and in poverty.

As a result of the high COVID-19 mortality rates, life expectancy in the United States for 2020 has been estimated to decline with disproportionate impacts on populations with high COVID-19 mortality rates.

Provisional US life expectancy estimates for January to June 2020⁶⁴indicate that between 2019 and the first half of 2020, life expectancy (at birth) decreased from 74.7 to 72.0 years (−2.7 years) for NH Black individuals. Life expectancy decreased from 81.8 to 79.9 years (−1.9 years) for Hispanic individuals and decreased from 78.8 to 78.0 years (−0.8 year) for NH White individuals.

Renewed efforts to maintain and improve CVH will be foundational to successful reductions in mortality and disability in the United States and globally. Individuals with more favorable levels of CVH have significantly lower risk for several of the leading causes of death and YLD, including IHD,²³Alzheimer disease,⁶⁵stroke,^66,67CKD,⁶⁸diabetes,^69,70and breast cancer^71,72(Tables 2-4 and 2-6). In addition, 6 of the 10 leading US risk factors for YLL and 4 of the 10 leading risk factors for YLD in 2019 were components of CVH (Tables 2-3 and 2-5). Taken together, these data demonstrate the tremendous importance of continued efforts to improve CVH.

The expanding efforts of the AHA and American Stroke Association in areas of brain health are also well poised to drive toward improvement in several leading causes of death and disability that influence YLLs and YLDs, including stroke, Alzheimer disease, depression and anxiety disorders, and alcohol and substance use disorders.

Despite improvements observed in CVH and brain health over the past decade, further progress is needed to more fully realize these benefits for all Americans. Details are described in the AHA presidential advisory on brain health.⁷³

Renewal of efforts to improve CVH is a continuing challenge that requires collaboration throughout the global community in ways that aim targeted skills and resources at improving the top causes and risk factors for death and disability in countries. Such efforts are required in countries at all income levels with an emphasis on efforts to halt the continued worsening of the components of CVH (Tables 2-7 through 2-10).

Many challenges exist related to implementation of prevention and treatment programs in international settings; some challenges are unique to individual countries/cultures, whereas others are universal. Partnerships and collaborations with local, national, regional, and global partners are foundational to effectively addressing relevant national health priorities in ways that facilitate contextualization within individual countries and cultures.

Prevalence of cigarette use in the past 30 days for middle and high school students by sex and race and ethnicity in 2020 is shown in Chart 3-1.

In 2020⁷:

Of youth who smoked cigarettes in the past 30 days in 2019, 28.9% (95% CI, 23.1%–35.5%) of middle and high school students (corresponding to 330 000 users) reported smoking cigarettes on 20 to 30 days of the past 30 days.⁸

In 2020, tobacco use within the past month for middle and high school students varied by race and ethnicity: The prevalence of past 30-day cigarette use was 3.7% (95% CI, 2.8%–4.8%) in NH White youth compared with 2.5% (95% CI, 1.8%–3.5%) in NH Black youth and 3.6% (95% CI, 2.6%–4.9%) in Hispanic youth. For cigars, the respective percentages were 2.8% (95% CI, 2.1%–3.7%), 6.5% (95% CI, 5.2%–8.2%), and 4.0% (95% CI, 2.9%–5.4%).⁷

The percentage of high school (19.6% or 3 020 000 users) and middle school (4.7% or 550 000 users) students who used e-cigarettes in the past 30 days exceeded the proportion using cigarettes in 2020 (Chart 3-1).⁷

According to the NHIS 2019 data, among adults ≥18 years of age⁹:

According to data from BRFSS 2019, the state with the highest age-adjusted percentage of current cigarette smokers was West Virginia (25.4%). The states with the lowest age-adjusted percentage of current cigarette smokers were Utah (7.9%) and California (10.1%; Chart 3-2).¹⁰

In 2019, smoking prevalence was higher among adults ≥18 years of age who reported having a disability or activity limitation (21.1%) than among those reporting no disability or limitation (13.3%).⁹

Among individuals who reported cigarette use every day or some days, 34.5% reported having severe generalized anxiety disorder, 27.0% reported having moderate generalized anxiety disorder, and 21.5% reported having mild generalized anxiety disorder compared with 12.0% who reported having no/minimal generalized anxiety disorder.⁹

Among females who gave birth in 2017, 6.9% smoked cigarettes during pregnancy. Smoking prevalence during pregnancy was greatest for females 20 to 24 years of age (9.9%), followed by females 15 to 19 years of age (8.3%) and 25 to 29 years of age (7.9%).¹¹Rates were highest among NH American Indian or Alaska Native females (15%) and lowest in NH Asian females (1%). With respect to differences by education, cigarette smoking prevalence was highest among females who completed high school (12.2%), and lowest among females with a master’s degree and higher (0.3%).

E-cigarette prevalence in 2017 is shown in Chart 3-3. Comparing e-cigarette prevalence across the 50 states shows that the average age-adjusted prevalence was 5.3%. The lowest age-adjusted prevalence was observed in California (3.2%), and the highest prevalence was observed in Oklahoma (7.5%). The age-adjusted prevalence was 1.3% in Puerto Rico.¹⁰

According to the 2019 NSDUH, ≈1.60 million people ≥12 years of age had smoked cigarettes for the first time within the past 12 months compared with 1.83 million in 2018 (2019 NSDUH Table 4.2B).¹²Of new smokers in 2019, 541 000 were 12 to 17 years of age, 672 000 were 18 to 20 years of age, and 292 000 were 21 to 25 years of age; only 90 000 were ≥26 years of age when they first smoked cigarettes.

The number of new smokers 12 to 17 years of age in 2019 (541 000) decreased from 2018 (571 000). The number of new smokers 18 to 25 years of age in 2019 (964 000) also decreased from 2018 (1.14 million) (2019 NSDUH Table 4.2B).¹²

According to data from the PATH Study between 2013 and 2016, in youth 12 to 15 years of age, use of an e-cigarette was independently associated with new ever use of combustible cigarettes (OR, 4.09 [95% CI, 2.97–5.63]) and past 30-day use (OR, 2.75 [95% CI, 1.60–4.73]) at 2 years of follow-up. For youth who tried another non–e-cigarette tobacco product, a similar strength of association for cigarette use at 2 years was observed.¹³

Per NSDUH data for individuals 12 to 17 years of age, overall, the lifetime use of tobacco products declined from 13.4% to 12.8% between 2018 and 2019, with lifetime cigarette use declining from 9.6% to 9.0% during the same time period (2019 NSDUH Tables 2.1B and 2.2B).¹²

According to NSDUH data, the lifetime use of tobacco products in individuals ≥18 years of age did not decline significantly between 2018 (66.3%) and 2019 (65.8%). Lifetime cigarette use declined in a similar interval from 60.3% to 59.5% (2019 NSDUH Tables 2.1B). Similar to the patterns in youth, lifetime risk of tobacco products varied by demographic factors (2019 NSDUH Table 2.8B)¹²:

In 2019, the lifetime use of smokeless tobacco for adults ≥18 years of age was 16.6% (2019 NSDUH Table 2.4B).¹²

On the basis of weighted NHIS data (2019), the current smoking status among males 18 to 24 years of age declined from 28.0% in 2005 to 15.3% in 2019; for females 18 to 24 years of age, smoking declined from 20.7% to 12.7% over the same time period.⁹

According to data from the BRFSS, the prevalence of e-cigarette use increased from 4.3% to 4.5% between 2016 and 2019 in US adults. Increases in e-cigarette use over this period were significant for middle-aged adults, females, and former smokers.¹⁷

A 2010 report of the US Surgeon General on how tobacco causes disease summarized an extensive body of literature on smoking and CVD and the mechanisms through which smoking is thought to cause CVD.¹⁸There is a sharp increase in CVD risk with low levels of exposure to cigarette smoke, including secondhand smoke, and a less rapid further increase in risk as the number of cigarettes per day increases. Similar health risks for CHD events were reported in a systematic review of regular cigar smoking.¹⁹

Smoking is an independent risk factor for CHD and appears to have a multiplicative effect with the other major risk factors for CHD: high serum levels of lipids, untreated hypertension, and diabetes.¹⁸

Cigarette smoking and other traditional CHD risk factors might have a synergistic interaction in HIV-positive individuals.²⁰

Among the US Black population, cigarette use is associated with elevated measures of subclinical PAD in a dose-dependent manner. Current smokers had an increased adjusted odds of ABI <1 (OR, 2.2 [95% CI, 1.5–3.3]).²¹

A meta-analysis of 75 cohort studies (≈2.4 million individuals) demonstrated a 25% greater risk for CHD in female smokers than in male smokers (RR, 1.25 [95% CI, 1.12–1.39]).²²

Cigarette smoking is a risk factor for both ischemic stroke and SAH in adjusted analyses and has a synergistic effect on other stroke risk factors such as oral contraceptive use.²³

A meta-analysis comparing pooled data of ≈3.8 million smokers and nonsmokers found a similar risk of stroke associated with current smoking in females and males.²⁴

Current smokers have a 2 to 4 times increased risk of stroke compared with nonsmokers or those who have quit for >10 years.^23,25Among JHS participants without a history of stroke (N=4410), risk of stroke was higher among current smokers compared with individuals who never smoked (HR, 2.48; 95% CI, 1.60–3.83).²⁶

A meta-analysis of 26 studies reported that compared with never smoking, current smoking (RR, 1.75 [95% CI, 1.54–1.99]) and former smoking (RR, 1.16 [95% CI, 1.08–1.24]) were associated with increased risk of HF.²⁷In MESA, compared with never smoking, current smoking was associated with an adjusted doubling in incident HF (HR, 2.05 [95% CI, 1.36–3.09]). The increased risk was similar for HFpEF (HR, 2.51) and HFrEF (HR, 2.58).²⁸

Short-term exposure to hookah smoking is associated with a significant increase in BP and heart rate and changes in cardiac function and blood flow, similar to those associated with cigarette smoking.²⁹The short-term vascular impairment associated with hookah smoking is masked by the high levels of carbon monoxide–—a vasodilator molecule—released from the charcoal briquettes used to heat the flavored tobacco product.³⁰In a recent meta-analysis of 42 studies, compared with nonsmokers, hookah smokers had significantly lower HDL-C and higher LDL-C, triglycerides, and fasting glucose.³¹The long-term effects of hookah smoking remain unclear.

Current use of smokeless tobacco was associated with an adjusted 1.27-fold increased risk of CVD events compared with never using. The CVD rate was 11.3 per 1000 person-years in never users and 21.4 in current users of smokeless tobacco.³²

The long-term CVD risks associated with e-cigarette use are not known because of a lack of longitudinal data.^33,34However, e-cigarette use has been linked to elevated levels of preclinical biomarkers associated with cardiovascular injury such as markers for sympathetic activation, oxidative stress, inflammation, thrombosis, and vascular dysfunction.³⁵In addition, daily and some-day use of e-cigarettes may be associated with MI and CHD.^36,37

Dual use of e-cigarettes and combustible cigarettes was associated with significantly higher odds of CVD (OR, 1.36 [95% CI, 1.18–1.56]) compared with exclusive combustible cigarette use.³⁷The association of dual use (relative to exclusive cigarette use) with CVD was 1.57 (95% CI, 1.18–2.07) for daily e-cigarette users and 1.31 (95% CI, 1.13–1.53) for occasional e-cigarette users.

In a pooled analysis of data collected from 10 randomized trials (N=2564), smokers had a higher risk of death or HF hospitalization (HR, 1.49 [95% CI, 1.09–2.02]), as well as reinfarction (HR, 1.97 [95% CI, 1.17–3.33) after primary PCI in STEMI.³⁸

Genetic factors contribute to smoking behavior; in analyses of up to 346 813 participants, common and rare variants in dozens of loci have been found to be associated with smoking initiation, number of cigarettes smoked per day, and smoking cessation.^39,40

Genetics might also modify adverse CVH outcomes among smokers, with variation in ADAMTS7 associated with loss of cardioprotection in smokers.⁴¹

Mendelian randomization analysis has linked genetic liability to smoking to ASCVD, including increased risk of PAD (OR, 2.13 [95% CI, 1.78–2.56]; P=3.6×10⁻¹⁶), CAD (OR, 1.48 [95% CI, 1.25–1.75]; P=4.4×10⁻⁶), and stroke (OR, 1.40 [95% CI, 1.02–1.92]; P=0.04).⁴²

Such legislation is likely to reduce the rates of smoking during adolescence—a time during which the majority of smokers start smoking—by limiting access because most people who buy cigarettes for adolescents are <21 years of age.

In addition, in several towns where Tobacco 21 laws were enacted before federal legislation, reductions of up to 47% in smoking prevalence among high school students have been reported.⁴⁷Furthermore, the National Academy of Medicine estimates that the nationwide Tobacco 21 law could result in 249 000 fewer premature deaths, 45 000 fewer lung cancer deaths, and 4.2 million fewer life-years lost among Americans born between 2010 and 2019.⁴⁷

Before the federal minimum age of sale increase, 19 states (Hawaii, California, New Jersey, Oregon, Maine, Massachusetts, Illinois, Virginia, Delaware, Arkansas, Texas, Vermont, Connecticut, Maryland, Ohio, New York, Washington, Pennsylvania, and Utah), Washington, DC, and at least 470 localities (including New York City, NY; Chicago, IL; San Antonio, TX; Boston, MA; Cleveland, OH; and both Kansas Cities [Kansas and Missouri]) passed legislation setting the minimum age for the purchase of tobacco to 21 years.⁴⁸

According to NHIS 2017 data, 61.7% of adult ever-smokers had stopped smoking; the quit rate has increased 6 percentage points since 2012 (55.1%).⁴⁹

Data from clinical settings suggest wide variation in counseling practices related to smoking cessation. In a study based on national registry data, only 1 in 3 smokers who visited a cardiology practice received smoking cessation assistance.⁵²

According to cross-sectional MEPS data from 2006 to 2015, receiving advice to quit increased over time from 60.2% in 2006 to 2007 to 64.9% in 2014 to 2015. In addition, in 2014 to 2015, use of prescription smoking cessation medicine was significantly lower among NH Black (OR, 0.51 [95% CI, 0.38–0.69]), NH Asian (OR, 0.31 [95% CI, 0.10–0.93]), and Hispanic (OR, 0.53 [95% CI, 0.36–0.78]) individuals compared with White individuals. Use of prescription smoking cessation medicine was also significantly lower among those without health insurance (OR, 0.58 [95% CI, 0.41–0.83]) and higher among females (OR, 1.28 [95% CI, 1.10–1.52]).⁵³In 2014 to 2015, receipt of doctor’s advice to quit among US adult smokers was significantly lower in NH Black (59.7% [95% CI, 56.1%–63.1%]) and Hispanic (57.9% [95% CI, 53.5%–62.2%]) individuals compared with NH White individuals (66.6% [95% CI, 64.1%–69.1%]).

Smoking cessation reduces the risk of cardiovascular morbidity and mortality for smokers with and without CHD.

Among 726 smokers included in the Wisconsin Smokers Health Study, smoking cessation was associated with less progression of carotid plaque but not IMT.⁵⁸

Cessation medications (including sustained-release bupropion, varenicline, nicotine gum, lozenge, nasal spray, and patch) are effective for helping smokers quit.^59,60

EVITA was an RCT that examined the efficacy of varenicline versus placebo for smoking cessation among smokers who were hospitalized for ACS. At 24 weeks, rates of smoking abstinence and reduction were significantly higher among patients randomized to varenicline. The abstinence rates at 24 weeks were higher in the varenicline (47.3%) than the placebo (32.5%) group (P=0.012; number needed to treat, 6.8). Continuous abstinence rates and reduction rates (≥50% of daily cigarette consumption) were also higher in the varenicline group.⁶¹

The EAGLES trial⁶²demonstrated the efficacy and safety of 12 weeks of varenicline, bupropion, or nicotine patch in motivated-to-quit patients who smoked with major depressive disorder, bipolar disorder, anxiety disorders, posttraumatic stress disorder, obsessive-compulsive disorder, social phobia, psychotic disorders including schizophrenia and schizoaffective disorders, and borderline personality disorder. Of note, these participants were all clinically stable from a psychiatric perspective and were believed not to be at high risk for self-injury.⁶²

Extended use of a nicotine patch (24 compared with 8 weeks) has been demonstrated to be safe and efficacious in randomized clinical trials.⁶³

An RCT demonstrated the effectiveness of individual- and group-oriented financial incentives for tobacco abstinence through at least 12 months of follow-up.⁶⁴

In addition to medications, smoke-free policies, increases in tobacco prices, cessation advice from health care professionals, and quit lines and other counseling have contributed to smoking cessation.^51,65

Mass media antismoking campaigns such as the CDC’s Tips campaign (Tips From Former Smokers) have been shown to reduce smoking-attributable morbidity and mortality and are cost-effective. Investigators estimated that the Tips campaign cost about $48 million, saved ≈179 099 QALYs, and prevented ≈17 000 premature deaths in the United States.⁶⁶

Despite states having collected $25.6 billion in 2012 from the 1998 Tobacco Master Settlement Agreement and tobacco taxes, <2% of those funds are spent on tobacco prevention and cessation programs.⁶⁷

A randomized trial of e-cigarettes and behavioral support versus nicotine-replacement therapy and behavioral support in adults attending the UK National Health Service stop-smoking services found that 1-year cigarette abstinence rates were 18% in the e-cigarette group compared with 9.9% in the nicotine-replacement therapy group (RR, 1.83 [95% CI, 1.30–2.58]; P<0.001). However, among participants abstinent at 1 year, in the nicotine-replacement therapy group, only 9% were still using nicotine-replacement therapy, whereas 80% of those in the e-cigarette group were still using e-cigarettes.⁶⁸

In a meta-analysis of 55 observational studies and 9 RCTs, e-cigarettes were not associated with increased smoking cessation, but e-cigarette provision was associated with increased smoking cessation.⁶⁹

According to the 2020 Surgeon General’s report on smoking cessation, >480 000 Americans die as a result of cigarette smoking and >41 000 die of secondhand smoke exposure each year, ≈1 in 5 deaths annually.

Of risk factors evaluated by the US Burden of Disease Collaborators, tobacco use was the second leading risk factor for death in the United States and the leading cause of DALYs, accounting for 11% of DALYs, in 2016.⁷⁰Overall mortality among US smokers is 3 times higher than that for never-smokers.⁵⁴

On average, on the basis of 2016 data, male smokers die 12 years earlier than male never-smokers, and female smokers die 11 years earlier than female never-smokers.^16,71

Increased CVD mortality risks persist for older (≥60 years of age) smokers as well. A meta-analysis of 25 studies comparing CVD risks in 503 905 cohort participants ≥60 years of age reported an HR for cardiovascular mortality of 2.07 (95% CI, 1.82–2.36) compared with never-smokers and 1.37 (95% CI, 1.25–1.49) compared with former smokers.⁷²

In a sample of Native American individuals (SHS), among whom the prevalence of tobacco use is highest in the United States, the PAR for total mortality was 18.4% for males and 10.9% for females.⁷³

Since the first report on the dangers of smoking was issued by the US Surgeon General in 1964, tobacco control efforts have contributed to a reduction of 8 million premature smoking-attributable deaths.⁷⁴

If current smoking trends continue, 5.6 million US children will die of smoking prematurely during adulthood.¹⁸

Electronic nicotine delivery systems are battery-operated devices that deliver nicotine, flavors, and other chemicals to the user in an aerosol without any combustion. Although e-cigarettes—the most common form of electronic nicotine delivery systems—were introduced into the United States only around 2007, there are currently >450 e-cigarette brands and vaping products on the market, and sales in the United States were projected to be $2 billion in 2014. Juul came on the market in 2015 and has rapidly become the most popular vaping product sold in the United States. The popularity of the Juul likely relates to several factors, including its slim and modern design, appealing flavors, and intensity of nicotine delivery, which approximates the experience of combustible cigarettes.⁷⁵Besides e-cigarettes and Juul, e-hookahs (ie, e-waterpipes) are a new category of vaping devices recently patented by Philip Morris in 2019.^76,77Unlike e-cigarettes and Juul, e-hookahs are used through traditional water pipes, allowing the flavored aerosol to pass through the water-filled bowl before being inhaled.⁷⁸The popularity of e-hookahs is driven in part by unsubstantiated claims that the presence of water “filters out toxins,” rendering e-hookahs as healthier tobacco alternatives.^79,80

E-cigarette use has become prevalent among never-smokers. In 2016, an estimated 1.9 million tobacco users exclusively used e-cigarettes in the United States. Of these exclusive e-cigarette users, 60% were <25 years of age.⁸¹

Current e-cigarette user prevalence for 2017 in the United States is shown in Chart 3-3.

According to the NYTS, in 2020, e-cigarettes were the most commonly used tobacco products in youth: In the past 30 days, 4.7% (550 000) of middle school and 19.6% (3.0 million) of high school students endorsed use (Chart 3-1).⁷An exponential increase in current e-cigarette use in high school students was observed between 2011 (1.5%) and 2020 (19.6%).^7,82A significant increase in current e-cigarette use also was observed for middle school students, for whom the corresponding values were 0.6% and 4.7% in the 2 periods.^2,7Among high school students, rates of use were slightly higher among males (20.4%) than females (18.7%) and most pronounced among NH White students (23.2%). In middle school students, rates of use were approximately equal between males (4.5%) and females (4.8%) and in Hispanic students (7.1%).⁷

According to the NYTS, current exclusive e-cigarette use among US youth who have never used combustibles, including cigarettes, increased exponentially from 2014 to 2019.⁸³Among high school students, current exclusive e-cigarette use increased from 1.4% (95% CI, 1.0%–2.1%) in 2014 to 9.2% (95% CI, 8.2%–10.2%) in 2019 and from 0.9% (95% CI, 0.6%–1.3%) in 2014 to 4.5% (95% CI, 3.7%–5.2%) in 2019 among middle school students.

Frequent use of e-cigarettes among high school students who were current e-cigarette users increased from 27.7% in 2018 to 34.2% in 2019. In middle school students, the percentage frequently using e-cigarettes among current users increased from 16.2% in 2018 to 18.0% in 2019.^2,8

Current use of e-cigarettes among high school students declined from 27.5% in 2019 to 19.6% in 2020.⁷In middle school students, current e-cigarette use declined from 10.5% in 2019 to 4.7% in 2020.

In 2016, 20.5 million US middle and high school students (80%) were exposed to e-cigarette advertising.⁸⁴

In 2019, the prevalence of current e-cigarette use in adults, defined as use every day or on some days, was 4.5% according to data from the NHIS. The prevalence of current e-cigarette use was highest in individuals 18 to 24 years of age (9.3%) and among those reporting severe generalized anxiety disorder (10.1%).⁹

According to data from BRFSS 2016 to 2018, current use of e-cigarettes in adults ≥18 years of age was higher in sexual and gender minority individuals.^85,86Data from 2017 and 2018 data sets show that the prevalence of current e-cigarette use among sexual and gender minority adults was 13.0% (95% CI, 12.0%–14.2%) versus 4.8% (95% CI, 4.6%–4.9%) among heterosexuals.⁸⁵In 2016, with respect to sexual orientation, 9.0% of bisexual and 7.0% of lesbian/gay individuals were current e-cigarette users compared with 4.6% of heterosexual people. Individuals who were transgender (8.7%) were current e-cigarette users at a higher rate than cisgender individuals (4.7%). Across US states, the highest prevalence of current e-cigarette use was observed in Oklahoma (7.0%) and the lowest in South Dakota (3.1%).⁸⁶

Limited data exist on the prevalence of other electronic nicotine delivery devices besides e-cigarettes. According to nationally representative data from the PATH study, in 2014 to 2015, 7.7% of youth 12 to 17 years of age reported ever e-hookah use.⁸⁷Among adults >18 years of age, 4.6% reported ever e-hookah use, and 26.8% of them reported current use.

E-cigarettes contain lower levels of most tobacco-related toxic constituents compared with traditional cigarettes,⁸⁸including volatile organic compounds.^89,90However, nicotine levels have been found to be consistent across long-term cigarette and long-term e-cigarette users.^35,91

E-cigarette use has a significant cross-sectional association with a less favorable perception of physical and mental health and with depression.^92,93

According to the BRFSS 2016 and 2017, e-cigarettes are associated with a 39% increased odds of self-reported asthma (OR, 1.39 [95% CI, 1.15–1.68]) and self-reported chronic obstructive pulmonary disease (OR, 1.75 [95% CI, 1.25–2.45]) among never users of combustible cigarette.^94,95There is a dose-response relationship such that higher frequency of e-cigarette use was associated with more asthma or chronic obstructive pulmonary disease.

An outbreak of e-cigarette or vaping product use–associated lung injury peaked in September 2019 after increasing rapidly between June and August 2019. Surveillance data and product testing indicate that tetrahydrocannabinol-containing e-cigarettes or vaping products are linked to most e-cigarette or vaping product use–associated lung injury cases. In particular, vitamin E acetate, an additive in some tetrahydrocannabinol-containing e-cigarettes or vaping, has been identified as the primary source of risk, although exposure to other e-cigarette– or vaping-related toxicants may also play a role. As of February 18, 2020, a total of 2807 hospitalized e-cigarette or vaping product use–associated lung injury cases or deaths have occurred in the United States.⁹⁶

Effective August 8, 2016, the FDA’s Deeming Rule prohibited sale of e-cigarettes to individuals <18 years of age.⁹⁷

In January 2020, the FDA issued a policy prioritizing enforcement against the development and distribution of certain unauthorized flavored e-cigarette products such as fruit and mint flavors (ie, any flavors other than tobacco and menthol).⁹⁸

According to data from the BRFSS 2016 and 2017, e-cigarette use among adults is associated with state-level regulations and policies regarding e-cigarettes: OR of 0.90 (95% CI, 0.83–0.98) for laws prohibiting e-cigarette use in indoor areas; OR of 0.90 (95% CI, 0.85–0.95) for laws requiring retailers to purchase a license to sell e-cigarettes; OR of 1.04 (95% CI, 0.99–1.09) for laws prohibiting self-service displays of e-cigarettes; OR of 0.86 (95% CI, 0.74–0.99) for laws prohibiting sales of tobacco products, including e-cigarettes, to people <21 years of age; and OR of 0.89 (95% CI, 0.83–0.96) for laws applying taxes to e-cigarettes.⁹⁹

Data from the US Surgeon General on the consequences of secondhand smoke indicate the following:

A meta-analysis of 23 prospective and 17 case-control studies of cardiovascular risks associated with secondhand smoke exposure demonstrated 18%, 23%, 23%, and 29% increased risks for total mortality, total CVD, CHD, and stroke, respectively, in those exposed to secondhand smoke.¹⁰¹

A meta-analysis of 24 studies demonstrated that secondhand smoke can increase risks for preterm birth by 20%.¹⁰²

A study using the Framingham Offspring cohort found that there was an 18% increase in AF among offspring for every 1–cigarette pack per day increase in parental smoking. In addition, offspring with parents who smoked had 1.34 (95% CI, 1.17–1.54) times the odds of smoking compared with offspring with nonsmoking parents.¹⁰³

As of September 30, 2020, 15 states (California, Colorado, Delaware, Hawaii, Massachusetts, Minnesota, New Jersey, New Mexico, New York, North Dakota, Oregon, Rhode Island, South Dakota, Utah, and Vermont), the District of Columbia, and Puerto Rico have passed comprehensive smoke-free indoor air laws that include e-cigarettes. These laws prohibit smoking and the use of e-cigarettes in indoor areas of private worksites, restaurants, and bars.^48,104

Pooled data from 17 studies in North America, Europe, and Australia suggest that smoke-free legislation can reduce the incidence of acute coronary events by 10% (RR, 0.90 [95% CI, 0.86–0.94]).¹⁰⁵

The percentage of the US nonsmoking population with serum cotinine ≥0.05 ng/mL (which indicates exposure to secondhand smoke) declined from 52.5% in 1999 to 2000 to 24.7% in 2017 to 2018, with declines occurring for both children and adults. During 2017 to 2018, the percentage of nonsmokers with detectable serum cotinine was 38.2% for those 3 to 11 years of age, 33.2% for those 12 to 19 years of age, and 21.2% for those ≥20 years of age. The percentage was higher for NH Black individuals (48.0%) than for NH White individuals (22.0%) and Mexican American individuals (16.6%). People living below the poverty level (44.7%) had higher rates of secondhand smoke exposure than their counterparts (21.3% of those living above the poverty level; NHANES).^106,107

In the United States, cigarette smoking was associated with 8.7% of annual aggregated health care spending from 2006 to 2010, which represented roughly $170 billion per year, 60% of which was paid by public programs (eg, Medicare and Medicaid).¹⁰⁸

According to the CDC and Federal Trade Commission, the tobacco industry spends about $9.06 billion on cigarette and smokeless tobacco advertising annually, equivalent to $25 million per day.¹⁰⁹In 2018, total US e-cigarette advertising expenditures (including print, radio, television, internet, and outdoors) were estimated to be $110 million, which increased remarkably from $48 million in 2017.¹¹⁰

In 2018, 216.9 billion cigarettes were sold by major manufacturers in the United States, which represents a 5.3% decrease (12.2 billion units) from 2017.¹¹¹

Cigarette prices in the United States increased steeply between the early 1970s and 2018, in large part because of excise taxes on tobacco products. The increase in cigarette prices appeared to be larger than general inflation: Per pack in 1970, the average cost was $0.38 and tax was $0.18, whereas in 2018, the average cost was $6.90 and average tax was $2.82.¹¹²

From 2012 through 2016, e-cigarette sales significantly increased while national e-cigarette prices significantly decreased. Together, these trends highlight the rapidly changing landscape of the US e-cigarette marketplace.¹¹²

Despite the morbidity and mortality resulting from tobacco use, Dieleman et al¹¹³estimated that tobacco interventions were among the bottom third of health care expenditures of the 154 health conditions they analyzed. They estimated that in 2019 the United States spent $1.9 billion (95% CI, $1.5–$2.3 billion) on tobacco interventions, the majority (75.6%) on individuals 20 to 64 years of age. Almost half of the funding (48.5%) for the intervention came from public insurance.

The GBD 2020 study produces comprehensive and comparable estimates of disease burden for 370 reported causes and 88 risk factors for 204 countries and territories from 1990 to 2020. Oceania, East and Central Asia, and Central and Eastern Europe had the highest age-standardized mortality rates attributable to tobacco (Chart 3-5).

Tobacco caused 8.09 (95% UI, 3.18–12.76) million deaths in 2020, with 6.27 (95% UI, 2.24–9.88) million among males and 1.82 (95% UI, 0.83–2.95) million among females (Table 3-1).¹¹⁴

GBD investigators estimated that in 2019 tobacco was the second leading risk of mortality (high SBP was number 1), and tobacco ranked third in DALYs globally.¹¹⁵

In 2015, there were a total of 933.1 million (95% UI, 831.3–1054.3 million) smokers globally, of whom 82.3% were male. The annualized rate of change in smoking prevalence between 1990 to 2015 was −1.7% in females and −1.3% in males.¹¹⁶

Worldwide, ≈80% of tobacco users live in low- and middle-income countries.¹¹⁷

The WHO estimated that the economic cost of smoking-attributable diseases accounted for US $422 billion in 2012, which represented ≈5.7% of global health expenditures.¹¹⁸The total economic costs, including both health expenditures and lost productivity, amounted to approximately US $1436 billion, which was roughly equal to 1.8% of the world’s annual gross domestic product. The WHO further estimated that 40% of the expenditures were in developing countries.

To help combat the global problem of tobacco exposure, in 2003, the WHO adopted the Framework Convention on Tobacco Control treaty. From this emerged a set of evidence-based policies with the goal of reducing the demand for tobacco, entitled MPOWER. MPOWER policies outline the following strategies for nations to reduce tobacco use: (1) monitor tobacco use and prevention policies; (2) protect individuals from tobacco smoke; (3) offer to help with tobacco cessation; (4) warn about tobacco-related dangers; (5) enforce bans on tobacco advertising; (6) raise taxes on tobacco; and (7) reduce the sale of cigarettes. More than half of all nations have implemented at least 1 MPOWER policy.^86,119In 2018, population cost coverage (either partial or full) for quit interventions increased to 78% in middle-income countries and to 97% in high-income countries; 5 billion people are now covered by at least 1 MPOWER measure. However, only 23 countries offered comprehensive cessation support in the same year.¹²⁰

The CDC examined data from 28 countries from the 2008 to 2016 Global Adult Tobacco Survey and reported that the median prevalence of tobacco smoking was 22.5% with wide heterogeneity (3.9% in Nigeria to 38.2% in Greece). Among current smokers, quit attempts over the prior 12 months also varied with a median of 42.5% (ranging from 14.4% in China to 59.6% in Senegal). Knowledge that smoking causes heart attacks (median, 83.6%; range, 38.7% in China to 95.5% in Turkey) and stroke (median 73.6%; range, 27.2% in China to 89.2% in Romania) varied widely across countries.¹²¹

Using parental report, from 2018 to 2019, the nationwide prevalence of youth who were active for ≥60 minutes every day of the week was higher for youth 6 to 11 years of age (28.3%) compared with youth 12 to 17 years of age (16.5%; Chart 4-2).¹³

Using nationwide self-reported PA (YRBSS, 2019)⁴:

With the use of accelerometry (NHANES, 2003–2006),¹⁴youth 6 to 19 years of age had a median of 53 min/d of moderate to vigorous PA.

With regard to measured cardiorespiratory fitness (NHANES, 2012),¹⁵for adolescents 12 to 15 years of age, boys at each age were more likely to have adequate levels of cardiorespiratory fitness than girls.

With regard to self-reported muscle-strengthening activities (YRBSS, 2019),⁴the proportion of high school students who participated in muscle-strengthening activities (such as push-ups, sit-ups, or weight lifting) on ≥3 d/wk was 49.5% nationwide and was lower in 12th grade (45.9%) compared with 9th grade (52.4%). More high school boys (59.0%) than girls (39.7%) reported having participated in muscle-strengthening activities on ≥3 d/wk.

From a nonrepresentative sample of US parents of youth 5 to 13 years of age, there is an indication that PA declined from before COVID-19 to early COVID-19 in 2020.¹⁶The longer-term impacts of the pandemic on PA and sedentary behavior patterns are not known.

Only 25.9% of students attended physical education classes in school daily (28.9% of boys and 22.8% of girls; YRBSS, 2019).⁴

Daily physical education class participation was lower with successively higher grades from the 9th grade (34.7%) through the 12th grade (19.7%; YRBSS, 2019).⁴

Just more than half (57.4%) of high school students played on at least 1 school or community sports team in the previous year (54.6% of girls and 60.2% of boys); this number was lower in 12th grade (49.8%) compared with 9th grade (61.9%; YRBSS, 2019).⁴

Research suggests that screen time (watching television or using a computer) is associated with less PA among children.¹⁷In addition, television viewing is associated with poor nutritional choices, overeating, and weight gain (Chapter 5, Nutrition).

Nationwide, 46.1% of high school students used a computer, tablet, or smartphone for activities other than school work (eg, video games, texting, social media) for ≥3 h/d on an average school day (YRBSS, 2019; Chart 4-3).⁴The prevalence differed by race and ethnicity and was high among both boys (47.5%) and girls (44.6%; YRBSS, 2019).⁴

Among high school students, the prevalence of watching television ≥3 h/d was 19.8% (YRBSS, 2019; Chart 4-4).⁴The prevalence varied by race and ethnicity and was higher among boys than girls. (31.6%).⁴

According to NHIS (2018), for self-reported leisure-time aerobic PA:

According to NHANES (2003–2006), adults from urban areas reported more transportation activity, but adults from rural areas reported spending more time in household PA and total PA than individuals from urban areas.²⁰

According to NHIS (2015), the prevalence of any walking for transportation in the United States varied by geographic location, ranging from 17.8% for adults living in the East South Central region to 43.5% for adults living in New England.²¹

From NHIS (2018) data, 25.4% of adults did not engage in leisure-time PA (no sessions of leisure-time PA of ≥10 minutes in duration).²²Trends in physical inactivity over time (1998–2018) are shown in Chart 4-7.

According to accelerometer-assessed PA (NHANES, 2005–2006),²³US adults were estimated to participate in 45.1 min/wk (SE, 4.6 min/wk) of moderate PA and 18.6 min/wk (SE, 6.6 min/wk) of vigorous PA. Levels of moderate and vigorous PA were lower in older adults (60–69 years of age; moderate, 32.7 min/wk [SE, 3.6 min/wk]; vigorous, 1.4 min/wk [SE, 0.7 min/wk]) compared with adults in younger age groups (eg, 40–49 years of age; moderate, 54.1 min/wk [SE, 12.8 min/wk]; vigorous, 24.9 min/wk [SE, 16.6 min/wk]).

Accelerometer data (NHANES, 2003–2006) also revealed that rural-dwelling adults were generally more active than urban-dwelling adults (mean, 325 bout min/d versus 314 bout min/d).²⁰Self-reported data from the same sample indicated higher total (438 min/wk versus 371 min/wk) and household PA (202 min/wk versus 124 min/wk), similar leisure PA (207 min/wk versus 206 min/wk), and lower transportation PA (30 min/wk versus 41 min/wk) among rural- compared with urban-dwelling adults.

In a nonrepresentative sample of adults from 14 countries, a cross-sectional study indicated that self-reported PA declined from before to after COVID-19 restrictions in 2020.²⁴The decline was greater for occupational activity compared with leisure activity, for more compared with less active adults, and for younger compared with older adults.

Activity tracker companies also documented declines in PA among their users during the COVID-19 pandemic. Comparing the week of March 22, 2020, with the same week in 2019 showed that Fitbit-measured steps declined worldwide (eg, declined 24% Argentina, 4% Australia, 15% Brazil, 14% Canada, 16% China, 13% Mexico, 14% Norway, 7% South Africa, 38% Spain, 9% United Kingdom, 12% United States), with the greatest decline occurring in Europe.²⁵Users of Garmin activity trackers also documented a decline in average daily steps during the month of March 2020 both globally and for the United States, as well as a shift to indoor fitness-oriented activities.²⁶The total number of steps decreased by 7.3% from 2019 to 2020 for Garmin users.²⁷It is important to note that those who own and wear activity trackers are not representative of the general population.^28,29

According to NHANES (2015–2016), 25.7% reported sitting >8 h/d; the time spent sitting was successively higher with older age.³⁰

A Nielsen report indicated that in January 2020 US adults spent on average 12 hours 21 minutes connected to media (eg, television, radio, smartphone, tablet, internet on computer), higher than in January 2018 (11 hours 6 minutes) and January 2019 (11 hours 27 minutes).³¹These habits affect time available for PA and contribute to sedentary behavior.

Among high school students nationwide, the prevalence of being physically active for ≥60 minutes for at least 5 d/wk decreased from 49.5% in 2011 to 44.1% in 2019.³²Similarly, the prevalence of being physically active for ≥60 minutes on all 7 days in a week decreased from 28.7% in 2011 to 23.2% in 2019.³²

Nationwide, the prevalence of high school students who reported attending physical education classes at least once per week (on an average week while in school) did not change substantively between 1991 (48.9%) and 2019 (52.2%).³²However, the prevalence of attending physical education classes on all 5 days of the week decreased from 1991 (41.6%) to 2019 (25.9%).

The prevalence of high school students playing ≥1 team sports in the past year did not substantively change between 1999 (55.1%) and 2019 (57.4%).³²

Among high school students nationwide, the prevalence of playing video or computer games or using a computer ≥3 hours/d increased from 22.1% in 2003 to 46.1% in 2019.³²However, watching television for ≥3 h/d decreased from 42.8% in 1999 to 19.8% in 2019.

The prevalence of physical inactivity among adults ≥18 years of age, overall and by sex, decreased from 1998 to 2018 (Chart 4-7).

The age-adjusted percentage of US adults who reported meeting both the muscle-strengthening and aerobic guidelines increased from 18.2% in 2008 to 24.0% in 2018.³³The percentage of US adults who reported meeting the aerobic guidelines increased from 43.5% in 2008 to 54.2% in 2018.³³

Sitting and watching television or videos at least 2 h/d remained high over time for adults ≥20 years of age (64.7% in 2003–2004 to 65.1% in 2015–2016).³⁵

The proportion of adults ≥25 years of age who met the 2018 guidelines for aerobic PA was higher with successively higher educational attainment category (Chart 4-8). This pattern was similar for meeting recommendations for both aerobic and strengthening activities.

In 26 high- and 34 middle-income countries between 2001 and 2016, the levels of insufficient PA were greater when there were greater income inequalities (defined as the difference between those with the highest and lowest incomes).³⁶

Genetic factors have been shown to contribute to the propensity to exercise; however, more work is needed to identify genetic factors that contribute to PA.^37,38

Genome-wide association analysis in >377 000 individuals identified multiple variants associated with habitual PA, including CADM2 and APOE.³⁷

A GWAS of 91 105 individuals with device-measured PA identified 14 significant loci.³⁹

Multiethnic analysis of >20 000 individuals identified several loci associated with leisure-time PA in individuals of European and African ancestry.⁴⁰Specifically, 4 previous loci (GABRG3, CYP19A1, PAPSS2 and CASR) were replicated. Among African Americans, 2 variants were identified (rs116550874 and rs3792874) and among European Americans, 1 variant was identified (rs28524846) as being associated with leisure-time PA.

Genetic variants have been identified, but few have been replicated by other studies.⁴¹

Community-level interventions are effective in promoting PA.⁴³Communities can encourage walking with street design that includes sidewalks, improved street lighting, and landscaping design that reduces traffic speed to improve pedestrian safety.⁴⁴Nationwide, in 2017, the most prominent barriers to bicycling included heavy traffic and lack of separated paths or trails.⁴⁵In a qualitative study across 10 US cities, other barriers to bicycling were identified.⁴⁶

Park prescriptions, which prescribe PA in local parks, may increase park use, time spent in parks, and recreational PA.⁴⁷

The COVID-19 pandemic affected walking and bicycling for transportation and leisure through environmental and policy changes designed to limit or accommodate shifting users.⁴⁸The short- and long-term impacts of the environmental and policy changes on representative patterns of walking and bicycling are not yet known.

Schools can provide opportunities for PA through physical education, recess, before- and after-school activity programs, and PA breaks, as well as offering by a place for PA for the community.⁴⁹

Worksites can offer access to onsite exercise facilities or employer-subsidized offsite exercise facilities to encourage PA among employees.

Worksite interventions for sedentary occupations such as providing “activity-permissive” workstations and email contacts that promote breaks have reported increased occupational light activity, and the more adherent individuals observed improvements in cardiometabolic outcomes.^50,51

In an analysis from NHIS, among 67 762 adults with >20 years of follow-up, 8.7% of all-cause mortality was attributed to a PA level of <150 min/wk of moderate-intensity PA.⁵²

A meta-analysis of 23 studies revealed an association between participating in more transportation-related PA and lower all-cause mortality, CVD, and diabetes.⁵³

In the UK Biobank of 263 540 participants, commuting by bicycle was associated with a lower risk of CVD mortality and all-cause mortality (HR, 0.48 and 0.59, respectively). Commuting by walking was associated with a lower risk of CVD mortality (HR, 0.64) but not all-cause mortality.⁵⁴Data on participants in NHANES enrolled from 1999 to 2006 indicated that participation in moderate to vigorous walking, bicycling, or running was most beneficial for reducing all-cause and CVD mortality.⁵⁵

A meta-analysis including 193 696 adults reported that high occupational PA was associated with a greater risk of all-cause mortality in males (HR, 1.18 [95% CI, 1.05–1.34]) compared with low occupational PA.⁵⁶However, a lower risk of all-cause mortality was observed among females with high occupational PA (HR, 0.90 [95% CI, 0.80–1.01]) compared with those with low occupational PA. There are several limitations to the literature that demonstrate these seemingly paradoxical results and likely other confounding factors such as fitness, SES, preexisting CVD, type of occupation, and other domains of PA that may modify this relationship.⁵⁷

A harmonized meta-analysis that included >1 million participants across 16 studies compared the risk associated with sitting time and television viewing in physically active and inactive study participants. For inactive individuals (defined as the lowest quartile of PA), those sitting >8 h/d had a higher all-cause mortality risk than those sitting <4 h/d (HR, 1.27 [95% CI, 1.22–1.32]). For active individuals (top quartile for PA), sitting time was not associated with all-cause mortality (HR, 1.04 [95% CI, 0.98–1.10]), but active people who watched television ≥5 h/d did have higher mortality risk (HR, 1.15 [95% CI, 1.05–1.27]).⁵⁸

An umbrella review of 24 systematic reviews of older adults concluded that those who are physically active are at a reduced risk of CVD mortality (25%–40% risk reduction), all-cause mortality (22%–35%), breast cancer (12%–17%), prostate cancer (9%–10%), and depression (17%–31%) while experiencing better quality of life, healthier aging trajectories, and improved cognitive functioning.⁵⁹Another review indicated that sedentary behavior, specifically transportation-related sitting time, was associated with a lower risk of CVD and less favorable cardiovascular risk factors, whereas less consistent associations were found when the exposure focused on occupational sitting.⁶⁰

With the use of an isotemporal substitution approach in a subsample of the CPS-II, among participants with the lowest level of PA, replacing 30 min/d of sitting with light-intensity PA or moderate- to vigorous-intensity PA was associated with 14% (HR, 0.86 [95% CI, 0.81–0.89]) or 45% (HR, 0.55 [95% CI, 0.47–0.62]) lower mortality, respectively. For the individuals with the highest PA levels, substitution was not associated with differences in mortality risk.⁶¹

In a review of 15 cohort studies, adults in the highest category of total, light, and moderate to vigorous PA had 67%, 40%, and 56% lower risk for mortality compared with adults in the lowest categories, respectively.⁶²

Among individuals 70 years of age who wore an accelerometer for 1 week, both light PA and moderate PA were associated with a lower risk and sedentary behavior was associated with an increased risk of all-cause mortality, stroke, and MI.⁶³

Among participants 40 to 79 years of age in the population-based European Prospective Investigation Into Cancer and Nutrition–Norfolk Study, higher levels of accelerometer-assessed total and moderate to vigorous PA were associated with a lower incident CVD risk; models indicated an initial steep decrease in the HR followed by a flattening of the curve.⁶⁴

Among females ≥63 years of age who wore an accelerometer for 1 week, those who spent more time standing (quartile 4 versus 1 HR, 0.63 [95% CI, 0.49–0.81]) and more time standing with ambulation (quartile 4 versus 1 HR, 0.50 [95% CI, 0.35–0.71]) had a lower risk of all-cause mortality.⁶⁵

In a harmonization meta-analysis of 8 prospective studies of adults measured with accelerometry, over a median of 5.8 years of follow-up, the highest 3 quartiles of light (HR, 0.38–0.60 across quartiles) and moderate to vigorous (HR, 0.52–0.64 across quartiles) PA compared with the lowest quartile (least active) were associated with a lower risk of all-cause mortality.⁶⁶Time in sedentary behavior was associated with a higher risk of all-cause mortality (HR, 1.28–2.63 across quartiles) compared with the lowest quartile (least sedentary). In a follow-up analysis of 9 prospective studies, 30 to 40 min/d of moderate to vigorous PA attenuated the adverse association between sedentary behavior and mortality.⁶⁷

Step counting is recommended as an effective method for translating PA guidelines and monitoring PA levels because of its simplicity and the increase in step-counting devices.^10,68Results from a systematic review revealed that for every 1000 steps taken at baseline, risk reductions ranged from 6% to 36% for all-cause mortality and 5% to 21% for CVD.⁶⁹More evidence is needed to set target volumes of PA based on steps per day and to determine the role of cadence (steps per minute, a proxy for intensity of ambulation) in these relationships.^10,68

Among a Swedish cohort of 266 109 adults 18 to 74 years of age, risk of CVD morbidity and all-cause mortality decreased 2.6% and 2.3% per 1–mL·min⁻¹·kg⁻¹increase, respectively, in cardiorespiratory fitness estimated from a submaximal bicycle test.⁷⁰The risk reduction with higher cardiorespiratory fitness was observed for both males and females across ages.

In a study of 36 956 Brazilian adolescents, higher self-reported moderate to vigorous PA levels (≥600 min/wk compared with 0 min/wk; adjusted proportional OR, 0.80 [95% CI, 0.6–0.95]) and lower amounts of screen time (≥6 h/d compared with ≤2 h/d; OR, 1.23 [95% CI, 1.10–1.37]) were associated with lower cardiometabolic risk.⁷¹

Among the NHANES 2003 to 2006 cohort of youths 6 to 17 years of age assigned to 4 latent classes with the use of accelerometry-assessed PA, those in the highest latent class PA had lower SBP (−4.1 mm Hg [95% CI, −7.7 to −0.6]), lower glucose levels (−4.3 mg/dL [95% CI, −7.8 to −0.7]), and lower insulin levels (−6.8 μU/mL [95% CI, −8.7 to −5.0]) than youths in the lowest latent class PA group.⁷²

An umbrella review of 21 systematic reviews found that greater amounts and higher intensities of PA and limiting sedentary behavior were associated with improved health outcomes (eg, cardiometabolic health, cardiorespiratory fitness, adiposity, and cognition) among youth 5 to 17 years of age.⁷³However, the evidence base available was insufficient to fully describe the dose-response relationship or whether the association varied by type or domain of PA or sedentary behavior.

A meta-analysis of 37 RCTs of walking interventions in apparently healthy adults indicated favorable effects on cardiovascular risk factors, including body fat, BMI, SBP, DBP, fasting glucose, and maximal cardiorespiratory fitness.⁷⁴

Multisession behavioral counseling can improve PA among those with elevated lipid levels or BP and reduce LDL, BP, adiposity, and cardiovascular events.⁷⁵The US Preventive Services Task Force recommends “offering or referring adults with CVD risk factors to behavioral counseling interventions to promote a healthy diet and PA” (Grade B).⁷⁶

In a meta-analysis of 11 studies investigating the role of exercise among individuals with MetS, aerobic exercise significantly improved DBP (−1.6 mm Hg; P=0.01), WC (−3.4 cm; P=0.01), fasting glucose (−0.15 mmol/L; P=0.03), and HDL-C (0.05 mmol/L; P=0.02).⁷⁷

In a dose-response meta-analysis of 29 studies with 330 222 participants that evaluated the association between PA levels and risk of hypertension, each 10–MET h/wk higher level of leisure-time PA was associated with a 6% lower risk of hypertension (RR, 0.94 [95% CI, 0.92–0.96]).⁷⁸

In an umbrella review of 17 meta-analyses and 1 systematic review, there was a strong inverse dose-response relationship between PA and incident hypertension, and PA reduced the risk of CVD progression among hypertensive adults.⁷⁹

A systematic review reported favorable dose-response relationships between daily step counts and both type 2 diabetes (25% reduction in 5-year dysglycemia incidence per 2000–step/d increase) and MetS (29% reduction in 6-year metabolic score per 2000–step/d increase).⁶⁸

In a prospective cohort study of 130 843 participants from 17 countries, compared with low levels of self-reported PA (<150 min/wk of moderate-intensity PA), moderate-intensity PA (150–750 min/wk) and high-intensity PA (>750 min/wk) were associated with a graded lower risk of major cardiovascular events (HR for high versus low, 0.75 [95% CI, 0.69–0.82]; moderate versus low, 0.86 [95% CI, 0.78–0.93]; high versus moderate, 0.88 [95% CI, 0.82–0.94]) over an average 6.9 years of follow-up.⁸⁰

In the 2-year LIFE study of older adults (mean age, 78.9 years), higher levels of accelerometer-assessed PA and daily steps were associated with lower risk of adverse cardiovascular events.⁸¹

A systematic review reported a favorable dose-response relationship between daily step counts and cardiovascular events (defined as cardiovascular death, nonfatal MI, or nonfatal stroke; 8% yearly rate reduction per 2000–step/d increase).⁶⁸

In the WHI, every 1–h/d increase in accelerometer-assessed light-intensity PA was associated with a lower risk of CHD (HR, 0.86 [95% CI, 0.73–1.00]) and lower CVD (HR, 0.92 [95% CI, 0.85–0.99]).⁸²

The Rotterdam Study evaluated the contribution of specific PA types to CVD-free life expectancy. Higher levels of cycling were associated with a greater CVD-free life span in males (3.1 years) and females (2.4 years). Furthermore, high levels of domestic work in females (2.4 years) and high levels of gardening in males (2 years) were also associated with an increased CVD-free life span.⁸³

With an average of 27 years of follow-up, estimates from 13 534 ARIC participants indicated that those who engaged in past-year leisure-time PA at least at median levels had a longer life expectancy free of nonfatal CHD (1.5–1.6 years), stroke (1.8 years), and HF (1.6–1.7 years) compared with those who did not engage in leisure-time PA.⁸⁴In addition, those watching less television had longer life expectancy free of CHD, stroke, and HF of close to 1 year.

According to data from the NHANES-III survey, adults with poor PA (OR, 1.30 [95% CI, 1.10–1.54]) and intermediate PA (OR, 1.19 [95% CI, 1.02–1.38]) had an increased odds of subclinical myocardial injury (based on the ECG) compared with those with ideal PA.⁸⁵

A meta-analysis summarizing 10 studies found that the pooled fully adjusted risk of venous thromboembolism was 0.87 (95% CI, 0.79–0.95) when the most physically active group was compared with the least physically active group.⁸⁶

In a dose-response meta-analysis of 9 prospective cohort studies (N=720 425), higher levels of sedentary behavior were associated with greater risk of CVD in a nonlinear relationship (HR for highest versus lowest sedentary behavior, 1.14 [95% CI, 1.09–1.19]).⁸⁷

In a meta-analysis of 12 prospective cohort studies (N=370 460), there was an inverse dose-dependent association between self-reported PA and risk of HF. PA levels at the guideline-recommended minimum (500 MET min/wk) were associated with 10% lower risk of HF. PA at 2 and 4 times the guideline-recommended levels was associated with 19% and 35% lower risk of HF, respectively.⁸⁸

In 2020, the WHO began a review that concluded that services and programs are needed to increase PA and limit sedentary behavior among adults living with chronic conditions, including diabetes and hypertension.⁸⁹

In a prospective cohort study of 15 486 participants with stable CAD from 39 countries, higher levels of PA were associated with a lower risk of mortality such that doubling the exercise volume was associated with a 10% lower risk of all-cause mortality.⁹⁰

Among 1746 patients with CAD followed up for 2 years, those who remained inactive or became inactive had a 4.9- and 2.4-fold higher risk of cardiac death, respectively, than patients who remained at least irregularly active during the follow-up period.⁹¹

In a prospective cohort study of 3307 individuals with CHD, participants who maintained high PA levels over longitudinal follow-up had a lower risk of mortality than those who were inactive over time (HR, 0.64 [95% CI, 0.50–0.83]).⁹²

A study of females in the WHI observational study after MI demonstrated that compared with those who maintained low PA levels, participants with improvement in PA levels (HR, 0.54 [95% CI, 0.36–0.86]) or with sustained high PA levels (HR, 0.52 [95% CI, 0.36–0.73]) had lower risks of mortality.⁹³

Among males after an MI, those who maintained high PA had a 39% lower risk of all-cause mortality, and those who walked for at least 30 min/d had a 29% lower risk of all-cause mortality.⁹⁴

Exercise and resistance training are recommended for adults after stroke.⁹⁵In a review pooling 499 patients with stroke, exercise programs adhering to these guidelines indicated improved walking speed and endurance, but no differences for PA or other mobility outcomes, compared with usual care.⁹⁶An RCT found that higher doses of walking during inpatient rehabilitation 1 to 4 weeks after stroke provided greater walking endurance and gait speed and improved quality of life compared with usual care physical therapy.⁹⁷

Among 2370 individuals with CVD who responded to the Taiwan NHIS, achieving more total PA, leisure-time PA, and domestic and work-related PA was associated with lower mortality at the 7-year follow-up.⁹⁸

The economic consequences of physical inactivity are substantial. A global analysis of 142 countries (93.2% of the world’s population) concluded that physical inactivity cost health care systems $53.8 billion in 2013, including $9.7 billion paid by individual households.⁹⁹

Increasing population levels of PA could increase productivity, particularly through presenteeism, and lead to substantial economic gains.¹⁰⁰

Prevalence of physical inactivity in 2016 was reported to be 27.5% (95% CI, 25.0%–32.2%) of the population globally. These rates have not changed substantially since 2001, at which time prevalence of physical inactivity was 28.5% (95% CI, 23.9%–33.9%). Critically, it appears that the number of females reporting insufficient PA is 8% higher than the number of males globally.¹⁰¹

The GBD 2020 study produces comprehensive and comparable estimates of disease burden for 370 reported causes and 88 risk factors for 204 countries and territories from 1990 to 2020.

The adjusted PAF for achieving <150 minutes of moderate to vigorous PA per week was 8.0% for all-cause and 4.6% for major CVD in a study of 17 low-, middle-, and high-income countries in 130 843 participants without preexisting CVD.⁸⁰

Consumption of whole grains was low with sex and racial variations and ranged from 0.6 (Mexican American males) to 0.9 (NH White males) servings per day. For each of these groups, <10% of adults met guidelines of ≥3 servings per day.

Whole fruit consumption similarly showed a sex and racial difference and ranged from 1.1 (NH Black males) to 1.7 (Mexican American females) servings per day. For each of those groups except Mexican American females, <10% of adults met guidelines of ≥2 cups/d. When 100% fruit juices were included, the number of servings increased, and the proportions of adults consuming ≥2 cups/d increased.

Nonstarchy vegetable consumption ranged from 1.5 (NH Black males) to 2.3 (NH White females) servings per day. The proportion of adults meeting guidelines of ≥2.5 cups/d was <10%.

Consumption of fish and shellfish ranged from 1.0 (NH White individuals) to 1.9 (NH Black females) servings per week. The proportions of adults meeting guidelines of ≥2 servings per week were ≈18% of NH White adults, ≈28% of NH Black adults, and ≈19% of Mexican American adults.

Weekly consumption of nuts and seeds was ≈6 servings among NH White adults, ≈3 servings among NH Black adults, and ≈ 4 servings among Mexican American adults. Approximately 1 in 3 White adults, 1 in 5 NH Black adults, and 1 in 4 Mexican American adults met guidelines of ≥4 servings per week.

Consumption of processed meats was lowest among Mexican American females (1.0 servings per week) and highest among NH White males (≈2.5 servings per week). Between 59% (NH White males) and 87% (Mexican American females) of adults consumed ≤2 servings per week.

Consumption of SSBs was lowest among NH White females (6.4 servings per week) and highest among NH Black individuals and Mexican American males (≈10 servings per week). The proportions of adults meeting guidelines of <36 oz/wk were ≈61% for NH White adults, 48% for Mexican American adults, and 41% for NH Black adults.

Consumption of sweets and bakery desserts ranged from 4.4 servings per week among Mexican American females to 3.3 servings per week among NH Black males. The majority of NH White, NH Black, and Mexican American adults consumed <2.5 servings per week.

The proportion of total energy intake from added sugars ranged from 11.8% for NH White males to 20.4% for NH Black females. Between 16.6% of NH Black females and 38.3% of Mexican American males consumed ≤6.5% of total energy intake from added sugars.

Consumption of EPA and DHA ranged from 0.079 to 0.124 g/d in each sex and racial or ethnic subgroup. Fewer than 9% of US adults met the guideline of ≥0.250 g/d.

Two-fifths to one-third of adults consumed <10% of total calories from saturated fat, and approximately one-half to two-thirds consumed <300 mg dietary cholesterol per day.

The ratio of (PUFAs+monounsaturated fatty acids)/SFAs ranged from 1.8 in NH White males and Mexican American males to 2.6 in NH Black females. The proportion with a ratio ≥2.5 ranged from 40.6% in NH Black females to 11.2% in NH White males.

Only ≈5% of NH White adults, ≈4% of Black adults, and ≈15% of Mexican American adults consumed ≥28 g dietary fiber per day.

Fewer than 10% of adults consumed <2.3 g sodium per day. Estimated mean sodium intake by 24-hour urinary excretion was 4205 mg/d for males and 3039 mg/d for females in 2013 to 2014. Estimates of sodium intake by race, sex, and source are shown in Charts 5-3 and 5-4. Sodium added to food outside the home accounts for more than two-thirds of total sodium intake in the United States (Chart 5-4).³Top sources of sodium intake vary by race and ethnicity, with the largest contributor being yeast breads for NH White adults, sandwiches for NH Black adults, burritos and tacos for Hispanic adults, and soups for NH Asian adults.⁴

Whole grain consumption was low with an estimated average intake of 0.95 serving per day (95% CI, 0.88–1.03) among US youth 2 to 19 years of age. Youth with higher parental education had higher intake.

Whole fruit consumption was low with an estimated average intake of 0.68 serving per day (95% CI, 0.58–0.77). The consumption pattern decreased with age. NH Asian youth and those of other races, including multiracial youth, had the highest intake of whole fruit, followed by NH White youth, other Hispanic youth, Mexican American youth, and NH Black youth. The average intake of 100% fruit juice was 0.46 serving per day (95% CI, 0.39–0.53). The consumption pattern also decreased with age. NH White youth had the lowest intake of fruit juice, followed by NH Asian youth and other races, including multiracial youth, Mexican American youth, other Hispanic youth, and NH Black youth.

Nonstarchy vegetable consumption was low with an estimated average intake of 0.57 serving per day (95% CI, 0.53–0.62). The consumption pattern increased with age.

Consumption of fish and shellfish was low with an estimated average intake of 0.06 serving per day (95% CI, 0.04–0.07). The consumption pattern increased with age. Hispanic youth had the highest intake of fish and shellfish, followed by NH Asian youth and other races, including multiracial youth, NH Black youth, Mexican American youth, and NH White youth.

Consumption of nuts and seeds was low with an estimated average intake of 0.40 serving per day (95% CI, 0.33–0.47). NH White youth had the highest intake of nuts and seeds, followed by NH Asian youth and other races, including multiracial youth, other Hispanic youth, NH Black youth, and Mexican American youth. The consumption pattern of nuts and seeds increased with attainment of parental education and parental income.

Consumption of unprocessed red meats was 0.31 serving per day (95% CI, 0.27–0.34) on average with higher intake among youth with attainment of parental education less than high school and high school graduate, and lower among youth with parental education of some college or above and college graduate or above.

Consumption of processed meats was 0.27 serving per day (95% CI, 0.24–0.29) on average with higher intake among males and lower intake among females. NH White youth had the highest intake of processed meat, followed by NH Black youth, Mexican American youth, NH Asian youth, and those of other races, including multiracial youth and other Hispanic youth.

Consumption of SSBs was 1.0 serving per day (95% CI, 0.89–1.11) on average among US youth. The consumption pattern of SSBs increased with age. NH Black youth had the highest intake of SSBs, followed by Mexican American youth, NH White youth, other Hispanic youth, NH Asian youth, and those of other races, including multiracial youth.

Consumption of sweets and bakery desserts contributed to an average of 6.07% of calories (95% CI, 5.55%–6.60%) among US youth, with no significant heterogeneity across age, sex, race and ethnicity, parental education, and household income.

Consumption of EPA and DHA was low with an estimated average intake of 0.04 g/d (95% CI, 0.03–0.05). The consumption pattern of EPA and DHA increased with age. NH Asian youth and those of other races, including multiracial youth, had the highest intake of EPA and DHA, followed by other Hispanic youth, Mexican American youth, NH White youth, and NH Black youth.

Consumption of SFAs was ≈12.1% of calories (95% CI, 11.8%–12.4%) among US youth. Consumption of dietary cholesterol was 254 mg/d (95% CI, 244–264) with NH White youth having the lowest intake (238 mg/d [95% CI, 226–250]) and Mexican American youth having the highest intake (292 [95% CI, 275–309]).

Consumption of dietary fiber was 15.6 g/d (95% CI, 15.1–16.0) on average among US youth, with no significant heterogeneity across age, sex, race and ethnicity, parental education, and household income.

Consumption of sodium was 3.33 g/d (95% CI, 3.28–3.37) on average among US youth. The consumption pattern increased with age. NH Asian youth and those of other races, including multiracial youth, had the highest intake of sodium, followed by NH Black youth, Mexican American youth, and NH White youth.

Societal and environmental factors independently associated with diet quality, adiposity, or weight gain include education, income, race and ethnicity, and (at least cross-sectionally) neighborhood availability of supermarkets.^12,13

Other local food-environment characteristics such as availability of grocery stores (ie, smaller stores than supermarkets), convenience stores, and fast food restaurants are not consistently associated with diet quality or adiposity and could be linked to social determinants of health for CVH.^14,15

Disparities may be driven in part by an overabundance of unhealthy food options. In a study of neighborhood-level data from 4 US cities (Birmingham, AL; Chicago, IL; Minneapolis, MN; and Oakland, CA), past neighborhood-level income was inversely associated with current density of convenience stores.¹⁶The percentage of the White population was inversely associated with density of fast food restaurants in low-income neighborhoods and with density of smaller grocery stores across all income levels.

In a study using NHANES and Nielsen Homescan data to examine disparities in calories from store-bought consumer packaged goods over time, calories from store-bought beverages decreased between 2003 to 2006 and 2009 to 2012. However, the decline in calories from consumer packaged goods was slower for NH Black people, Mexican American people, and lowest-income households.¹⁷

Genetic factors may contribute to food preferences and modulate the association between dietary components and adverse CVH outcomes.^18–20However, there is a paucity of gene-diet interaction studies with independent replication to support personalizing dietary recommendations according to genotype.

In a randomized trial of 609 overweight-obese, nondiabetic participants that compared the effects of healthy low-fat and healthy low-carbohydrate weight loss diets, neither genotype pattern (3 SNP multilocus genotype responsiveness pattern) nor insulin secretion (30 minutes after glucose challenge) modified the effects of diet on weight loss.²¹

The interactions between a GRS composed of 97 BMI-associated variants and 3 diet-quality scores were examined in a pooled analysis of 30 904 participants from the Nurses’ Health Study, the HPFS, and the Women’s Genome Health Study. Higher diet quality was found to attenuate the association between GRS and BMI (P for interaction terms <0.005 for AHEI-2010 score, Alternative Mediterranean Diet score, and DASH diet score).²²A 10-unit increase in the GRS was associated with a 0.84-unit (95% CI, 0.72–0.96) increase in BMI for those in the highest tertile of AHEI score compared with a 1.14-unit (95% CI, 0.99–1.29) increase in BMI in those in the lowest tertile of AHEI score.

In a study of ≈9000 women from the WHI, a GRS for LDL-C, composed of 1760 LDL-associated variants, explained 3.7% (95% CI, 0.09%–11.9%) of the variance in 1-year LDL-C changes in a dietary fat intervention arm but was not associated with changes in the control arm.²³

Nationally representative data from 37 233 US adults were analyzed to examine the association between low-carbohydrate and low-fat diets and mortality. Neither low-carbohydrate nor low-fat diets were associated with total mortality; however, diet quality and sources of macronutrients appeared to play a role in that healthy low-carbohydrate (HR, 0.91 [95% CI, 0.87–0.95]; P<0.001) and low-fat (HR, 0.89 [95% CI, 0.85–0.93]; P<0.001) diets were associated with lower mortality and unhealthy low-carbohydrate (HR, 1.07 [95% CI, 1.02–1.11]; P=0.01) and low-fat (HR, 1.06 [95% CI, 1.01–1.12]; P=0.04) diets were linked to higher mortality.²⁴

Essential to any healthy diet, higher intakes of fruit and vegetables are associated with lower mortality. Specifically, data from 66 719 females from the Nurses’ Health Study (1984–2014) and 42 016 males from the HPFS (1986–2014) showed that daily intake of 5 servings of fruit and vegetables (versus 2 servings per day) was associated with lower total mortality (HR, 0.87 [95% CI, 0.85–0.90]), CVD mortality (HR, 0.88 [95% CI, 0.83–0.94]), cancer mortality (HR, 0.90 [95% CI, 0.86–0.95]), and respiratory disease mortality (HR, 0.65 [95% CI, 0.59–0.72]).²⁵

NHANES III (1988–1994) data from 3733 overweight/obese (BMI ≥25 kg/m²) adults (20–90 years of age) were analyzed to assess the relationship between the DII score and mortality. Results show that the DII scores of metabolically unhealthy obese/overweight individuals were associated with increased mortality risk (HR_{tertile 3 versus tertile 1}, 1.44 [95% CI, 1.11–1.86]; P_trend=0.008; HR_{1SD increase}, 1.08 [95% CI, 0.99–1.18]) and, more specifically, CVD-related mortality (HR_{T3 versus T1}, 3.29 [95% CI, 2.01–5.37]; P_trend< 0.001; HR_{1SD increase}, 1.40 [95% CI, 1.18–1.66]). These associations were not observed among MHO adults, and no cancer mortality risk was observed for either metabolically unhealthy obese/overweight or MHO individuals. The SUN (N=18 566) and PREDIMED (N=6790) Spanish cohort studies similarly analyzed the DII score in relation to mortality. Significant associations were found in differences between the highest and lowest quartiles of the DII score and mortality in both SUN (HR, 1.85 [95% CI, 1.15–2.98]; P_trend=0.004)²⁶and PREDIMED (HR, 1.42 [95% CI, 1.00–2.02]; P_trend=0.009). A subsequent meta-analysis of 12 studies examined the association between the DII score and mortality and found the DII score to be significantly associated with a 23% increase in mortality (95% CI, 16%–32%) in the highest versus lowest quartiles of the DII score.^26,27

NHANES 1999 to 2010 data from 20 256 US adults (mean, 47.5 years of age) were analyzed to evaluate the relationship between dietary uricemia score and dietary atherogenic score (which were derived in regression models on 37 micronutrients and macronutrients predicting levels of serum uric acid and apolipoprotein B, respectively) and all-cause and cause-specific mortality. Individuals in the highest dietary uricemia score quartile were at greater risk for all-cause (HR, 1.17 [95% CI, 1.07–2.30]), cancer (HR, 1.06 [95% CI, 1.01–1.14]), and CVD (HR, 1.36 [95% CI, 1.21–1.59]) mortality. Similar patterns were noted in the dietary atherogenic score, with those in the highest quartiles (versus those in the lowest) experiencing increased risk for all-cause (25%), cancer (11%), and CVD (40%) mortality.²⁸

A number of studies examined the relationship between sugar intake and all- and cause-specific mortality. A 6-year cohort study of 13 440 US adults (mean, 63.6 years of age) found that higher consumption (each additional 12-oz serving per day) of sugary beverages (HR, 1.11 [95% CI, 1.03–1.19]) and 100% fruit juices (HR, 1.24 [95% CI, 1.09–1.42]) was associated with higher all-cause (but not CHD-specific) mortality.²⁹In 2 Swedish studies (MDCS; n=24 272 and NSHDS; n=24 475), higher sugar consumption (>20% energy intake) was linked to higher mortality risk (HR, 1.30 [95% CI, 1.12–1.51]), and low sugar consumption (<5% energy intake) was also associated with higher mortality risk (HR, 1.23 [95% CI, 1.11–1.35]) in the MDCS study.³⁰

A systematic review of 18 cohort studies (N=251 497) examined the relationship between glycemic index and glycemic load with risk of all-cause mortality and CVD and found no associations between glycemic index or glycemic load and CVD or all-cause mortality. However, a positive association was found with all-cause mortality among females with the highest (versus lowest) glycemic index (RR, 1.17 [95% CI, 1.02–1.35]).³¹Using data from 137 851 participants between 35 and 70 years of age living in high-, middle-, and low-income countries across 5 continents with a median follow-up of 9.5 years, the international PURE study reported that a high glycemic index was associated with an increased risk of a major cardiovascular event or death among participants with (HR, 1.51 [95% CI, 1.25–1.82]) and without (HR, 1.21 [95% CI, 1.11–1.34) preexisting CVD at baseline.³²

In an assessment of the relationship between dairy intake and mortality, data from 3 large prospective cohort studies with 217 755 US adults showed a dose-response relationship in which 2 daily servings of dairy were associated with the lowest CVD mortality and higher intake was linked to higher mortality, especially cancer mortality. Compared with other subtypes of dairy (eg, skim/low-fat milk, cheese, yogurt, ice cream/sherbet), whole milk (and additional 0.5 serving per day) was associated with higher risks of cancer mortality (HR, 1.11 [95% CI, 1.06–1.17]), CVD mortality (HR, 1.09 [95% CI, 1.03–1.15]), and total mortality (HR, 1.11 [95% CI, 1.09–1.14]). A similar large cohort study of 45 009 Italian participants found no dose-response relationship between dairy (eg, milk, cheese, yogurt, butter) consumption and mortality, and no differences were present between full-fat and reduced-fat milk. However, there was a significant reduction of 25% in risk of all-cause mortality among those consuming 160 to 200 g/d (HR, 0.75 [95% CI, 0.61–0.91]) milk versus nonconsumers. Another European study examined the relationship between dietary protein and protein sources and mortality among 2641 Finnish males. Higher meat intake (HR, 1.23 [95% CI, 1.04–1.47]) and higher ratio of animal to plant protein (HR, 1.23 [95% CI, 1.02–1.49]) were associated with higher mortality. This relationship was more pronounced among those with a history of CVD, cancer, and type 2 diabetes.^33–35In addition, several meta-analyses of prospective cohort studies have consistently reported that higher plant protein intake is inversely associated with total and CVD mortality, lending support for dietary recommendations to replace foods high in animal protein with plant protein sources.^36–38

The association between nut and peanut butter consumption and mortality has also been assessed. In a large prospective cohort study of 566 398 US adults (50–71 years of age at baseline) with a median follow-up of 15.5 years, nut consumption was inversely related to mortality (HR, 0.78 [95% CI, 0.76–0.81]; P≤0.001) and was associated with reductions in cancer, CVD, and infectious, respiratory, and liver and renal disease mortality (but not Alzheimer- or diabetes-related mortality). No significant relationships were found between peanut butter and cause-specific or all-cause mortality (HR, 1.00 [95% CI, 0.98–1.04]; P=0.001).³⁹

Moderate egg consumption and all-cause and cause-specific⁴⁰mortality were investigated in a large cohort of 40 621 adults (29–69 years of age) in the EPIC-Spain prospective cohort study across 18 years. Mean egg consumption was 22 g/d (SD, 15.8 g/d) in females and 30.9 g/d (SD, 23.1 g/d) in males, and no association was found between the highest and lowest quartiles of egg consumption and all-cause mortality (HR, 1.01 [95% CI, 0.91–1.11]; P=0.96) or cancer and CVD mortality. However, egg consumption appears to be linked to deaths resulting from other causes (HR, 0.76 [95% CI, 0.63–0.93]; P=0.003), specifically nervous system–related deaths (HR, 0.59 [95% CI, 0.35–1.00]; P=0.036).⁴⁰

The association between dietary choline and overall- and cause-specific mortality was examined in a large, nationally representative study of 20 325 US adults (mean, 47.4 years of age). Higher choline consumption was found to be associated with worse lipid profiles, poorer glycemic control, and lower CRP levels (all comparisons P<0.001). Those with highest compared with lowest consumption had increased risk of total (RR, 1.23 [95% CI, 1.09–1.38]), stroke (RR, 1.30 [95% CI, 1.02–1.66]), and CVD (RR, 1.33 [95% CI, 1.19–1.48]) mortality (all comparisons P<0.001).⁴¹A subsequently performed meta-analysis confirmed these results and found choline to be linked to higher mortality risk (RR, 1.12 [95% CI, 1.08–1.17]; I²=2.9) and CVD mortality risk (RR, 1.28 [95% CI, 1.17–1.39]; I²=9.6).⁴¹

The observational findings for benefits of the Mediterranean diet have been confirmed in a large primary prevention trial in Spain among patients with CVD risk factors.⁴²The PREDIMED trial demonstrated an ≈30% relative reduction in the risk of stroke, MI, and death attributable to cardiovascular causes in those patients randomized to unrestricted-calorie Mediterranean-style diets supplemented with extra virgin olive oil or mixed nuts,⁴²without changes in body weight.⁴³In a subgroup analysis of 3541 patients without diabetes in the PREDIMED trial, HRs for incident diabetes were 0.60 (95% CI, 0.43–0.85) for the Mediterranean diet with olive oil group and 0.82 (95% CI, 0.61–1.10) for the Mediterranean diet with nuts group compared with the control group.

In a randomized crossover trial of 118 overweight omnivores at low-moderate CVD risk, a reduced-calorie lacto-ovo-vegetarian diet was compared with a reduced-calorie Mediterranean diet by providing face-to-face, individual counseling sessions. Both diets were equally successful in reducing body weight and fat mass. LDL-C, uric acid, and vitamin B₁₂were lower during the vegetarian diet, whereas triglycerides were lower during the Mediterranean diet, without substantial differences between oxidative stress markers and inflammatory cytokines.⁴⁴

In a systematic review and meta-analysis of 29 observational studies, the RR for the highest versus the lowest category of the Mediterranean diet was 0.81 (95% CI, 0.74–0.88) for CVD, 0.70 (95% CI, 0.62–0.80) for CHD/AMI, 0.73 (95% CI, 0.59–0.91) for unspecified stroke (ischemic/hemorrhagic), 0.82 (95% CI, 0.73–0.92) for ischemic stroke, and 1.01 (95% CI, 0.74–1.37) for hemorrhagic stroke.⁴⁵

In a meta-analysis of 20 prospective cohort studies, the RR for each 4-point increment of the Mediterranean diet score was 0.84 (95% CI, 0.81–0.88) for unspecified stroke, 0.86 (95% CI, 0.81–0.91) for ischemic stroke, and 0.83 (95% CI, 0.74–0.93) for hemorrhagic stroke.⁴⁶

In another systematic review, a meta-analysis of 3 RCTs showed a beneficial effect of the Mediterranean diet on total CVD incidence (RR, 0.62 [95% CI, 0.50–0.78]) and total MI incidence (RR, 0.65 [95% CI, 0.49–0.88]).⁴⁷

Another meta-analysis of 38 prospective cohort studies showed that the RR for the highest versus the lowest categories of Mediterranean diet adherence was 0.79 (95% CI, 0.77–0.82) for total CVD mortality, 0.73 (95% CI, 0.62–0.86) for CHD incidence, 0.83 (95% CI, 0.75–0.92) for CHD mortality, 0.80 (95% CI, 0.71–0.90) for stroke incidence, 0.87 (95% CI, 0.80–0.96) for stroke mortality, and 0.73 (95% CI, 0.61–0.88) for MI incidence.⁴⁷

Compared with a usual Western diet, a DASH-type dietary pattern with low sodium reduced SBP by 5.3, 7.5, 9.7, and 20.8 mm Hg in adults with baseline SBP <130, 130 to 139, 140 to 149, and ≥150 mm Hg, respectively.⁴⁸In an umbrella review of systematic reviews, a meta-analysis of 33 controlled trials showed that the DASH diet was associated with decreased SBP (mean difference, −5.2 mm Hg [95% CI, −7.0 to −3.4]), DBP (−2.60 mm Hg [95% CI, −3.50 to −1.70]), TC (−0.20 mmol/L [95% CI, −0.31 to −0.10]), LDL-C (−0.10 mmol/L [95% CI, −0.20 to −0.01]), HbA1c (−0.53% [95% CI, −0.62 to −0.43]), fasting blood insulin (−0.15 μU/mL [95% CI, −0.22 to −0.08]), and body weight (−1.42 kg [95% CI, −2.03 to −0.82]).⁴⁹A meta-analysis of 15 prospective cohort studies showed that the DASH diet was associated with decreased incident CVD (RR, 0.80 [95% CI, 0.76–0.85]), CHD (0.79 [95% CI, 0.71–0.88]), stroke (0.81 [95% CI, 0.72–0.92]), and diabetes (0.82 [95% CI, 0.74–0.92]).⁴⁹In another systematic review and meta-analysis of 7 prospective cohort studies, the RR for each 4-point increment of DASH diet score was 0.95 (95% CI, 0.94–0.97) for CAD.⁵⁰

Compared with a higher-carbohydrate DASH diet, a DASH-type diet with higher protein lowered BP by 1.4 mm Hg, LDL-C by 3.3 mg/dL, and triglycerides by 16 mg/dL but also lowered HDL-C by 1.3 mg/dL. Compared with a higher-carbohydrate DASH diet, a DASH-type diet with higher unsaturated fat lowered BP by 1.3 mm Hg, increased HDL-C by 1.1 mg/dL, and lowered triglycerides by 10 mg/dL.⁵¹The DASH-type diet higher in unsaturated fat also improved glucose-insulin homeostasis compared with the higher-carbohydrate DASH diet.

A secondary analysis of the AHS-2 among NH White participants showed that vegetarian dietary patterns (vegan, lacto-ovo vegetarian, and pescatarian) at baseline were associated with lower prevalence of hypertension at 1 to 3 years of follow-up compared with the nonvegetarian patterns: PR was 0.46 (95% CI, 0.25–0.83) for vegans, 0.57 (95% CI, 0.45–0.73) for lacto-ovo-vegetarians, and 0.62 (95% CI, 0.42–0.91) for pescatarian. This association remained after adjustment for BMI among the lacto-ovo-vegetarians.⁵²

In a systematic review and meta-analysis of 9 prospective cohort studies, higher adherence to a plant-based dietary pattern was significantly associated with lower risk of type 2 diabetes (RR, 0.77 [95% CI, 0.71–0.84]).⁵³

In an RCT of 48 835 postmenopausal females, a low-fat dietary pattern (lower fat and higher carbohydrates, vegetables, and fruit) intervention led to significant reductions in breast cancer followed by death (HR, 0.84 [95% CI, 0.74–0.96]) and in diabetes requiring insulin (HR, 0.87 [95% CI, 0.77–0.98]) over a median follow-up of 19.6 years compared with usual diet.⁵⁴

In a prospective cohort study of 105 159 adults followed up for a median of 5.2 years, for a 10% increment in the percentage of ultraprocessed foods in the diet, the HR was 1.12 (95% CI, 1.05–1.20) for overall CVD, 1.13 (95% CI, 1.02–1.24) for CHD, and 1.11 (95% CI, 1.01–1.21) for cerebrovascular disease.⁵⁵

An umbrella review of 16 meta-analyses of 116 primary prospective cohort studies with 4.8 million participants reported moderate-quality evidence for the inverse association of healthy dietary patterns with the risk of type 2 diabetes (RR, 0.81 [95% CI, 0.76–0.86]) and for a positive association between unhealthy dietary patterns and the risk of type 2 diabetes (RR, 1.44 [95% CI, 1.33–1.56]) and MetS (RR, 1.29 [95% CI, 1.09–1.52]).⁵⁶

A meta-analysis of 7 RCTs with 425 participants for an average duration of 8.6 weeks found that compared with breakfast consumption, breakfast skipping led to modest weight loss (WMD, −0.54 kg [95% CI, −1.05 to −0.03]) but a modest increase in LDL-C (WMD, 9.24 mg/dL [95% CI, 2.18−16.30]).⁵⁷Another meta-analysis of 23 RCTs with 1397 participants reported that fasting and energy-restricting diets resulted in significant reductions in SBP (WMD, −1.88 mm Hg [95% CI, −2.50 to −1.25]) and DBP (WMD, −1.32 mm Hg [95% CI, −1.81 to −0.84]), and the SBP-lowering effects were stronger with fasting (WMD, −3.26 mm Hg) than energy restriction (WMD, −1.09 mm Hg).⁵⁸

In meta-analyses of RCTs comparing higher and lower fiber intake, higher fiber intake lowered body weight (−0.37 kg [95% CI, −0.63 to −0.11]), TC (−0.15 mmol/L [95% CI, −0.22 to −0.07]), and SBP (−1.27 mm Hg [95% CI, −2.50 to −0.04]) and tended to lower HbA1c (−0.54% [95% CI, −1.28% to 0.20%]).⁵⁹In similar meta-analyses of RCTs for whole grains and glycemic index, higher whole grain intake significantly reduced only body weight (−0.62 kg [95% CI, −1.19 to −0.05]), whereas no consistent health effects were found for glycemic index. In meta-analyses of observational studies, higher total dietary fiber intake was associated with a lower risk of incident CHD (RR, 0.76 [95% CI, 0.69–0.83]), CHD mortality (RR, 0.69 [95% CI, 0.60–0.81]), and incident stroke (RR, 0.78 [95% CI, 0.69–0.88]).⁵⁹Higher whole grain intake was associated with a lower risk of incident CHD (RR, 0.80 [95% CI, 0.70–0.91]), CHD mortality (RR, 0.66 [95% CI, 0.56–0.77]), and stroke death (RR, 0.74 [95% CI, 0.58–0.94]). In a meta-analysis of 40 prospective cohort studies in the United States, Asia, and Europe, total dietary fiber (HR, 0.92 [95% CI, 0.88–0.96)] and cereal fiber (HR, 0.83 [95% CI, 0.77–0.90]) were shown to be associated with decreased risk of developing type 2 diabetes among adults with overweight or obesity in US-based studies. The same meta-analysis also reported increased risks of type 2 diabetes with higher glycemic index or glycemic load in US and Asian studies.⁶⁰

In a randomized trial of 609 participants without diabetes with a BMI of 28 to 40 kg/m²that compared the effects of healthy low-fat and healthy low-carbohydrate weight loss diets, weight loss at 12 months did not differ between groups.²¹A meta-analysis of 12 randomized studies confirmed the benefit of consuming low-carbohydrate healthy diets for multiple CVD risk factors, including reductions in body weight, triglycerides, LDL-C, SBP, and DBP, as well as increases in HDL-C, although the effects are modest in general and the sustainability is uncertain.⁶¹

A study of NHANES 1999 to 2010 data from 24 144 participants comparing those in the fourth versus first quartiles of consumption of dietary fats by type found an inverse association between total fat (HR, 0.90 [95% CI, 0.82–0.99]) and PUFA (0.81 [95% CI, 0.78–0.84]) but an increased association between SFA (1.08 [95% CI, 1.04–1.11]), and all-cause mortality. In the same study, a meta-analysis of 29 prospective cohorts (N=1 164 029) was also conducted and corroborated the findings for the inverse association between total fat and PUFA and all-cause mortality. In addition, the meta-analysis showed an inverse association between monounsaturated fatty acid (HR, 0.94 [95% CI, 0.89–0.99) intake and all-cause mortality and between monounsaturated fatty acid (0.80 [95% CI, 0.67–0.96]) and PUFA (0.84 [95% CI, 0.80–0.90]) intake and stroke mortality. A positive association between SFA (HR, 1.10 [95% CI, 1.01–1.21]) intake and CHD mortality was observed.⁶²However, another meta-analysis reported a protective association between dietary SFA intake and risk for stroke (RR, 0.87 [95% CI, 0.78–0.96]), and there was a linear relation in that every 10–g/d increase in SFA intake was associated with a 6% lower RR of stroke (RR, 0.94 [95% CI, 0.89–0.98]).⁶³A recent review underscores the controversy surrounding SFA intake as a risk or protective factor for CVD and total mortality and recommends against arbitrary population-wide upper limits on SFA intake without regard to the types of SFA, the food sources, the overall micronutrient distributions, and the health outcomes of interest.⁶⁴Gut microbiota is associated with the risk of obesity, type 2 diabetes, and many other cardiometabolic diseases. In a 6-month randomized controlled feeding trial of 217 healthy young adults with BMI <28 kg/m², the high-fat diet (fat, 40% energy) had overall unfavorable effects on gut microbiota: increased Alistipes (P=0.04) and Bacteroides (P<0.001) and decreased Faecalibacterium (P=0.04). The low-fat diet (fat, 20% energy) appeared to have beneficial effects on gut microbiota: increased α-diversity assessed by the Shannon index (P=0.03) and increased abundance of Blautia (P=0.007) and Faecalibacterium (P=0.04).⁶⁵

In the WHI RCT (N=48 835), reduction of total fat consumption from 37.8% energy (baseline) to 24.3% energy (at 1 year) and 28.8% energy (at 6 years) had no effect on incidence of CHD (RR, 0.98 [95% CI, 0.88–1.09]), stroke (RR, 1.02 [95% CI, 0.90–1.15]), or total CVD (RR, 0.98 [95% CI, 0.92–1.05]) over a mean follow-up of 8.1 years.⁶⁶In a matched case-control study of 2428 postmenopausal females nested in the WHI Observational Study, higher plasma phospholipid long-chain SFAs (OR, 1.18 [95% CI, 1.09–1.28]) and lower PUFA n-3 (OR, 0.93 [95% CI, 0.88–0.99]) were associated with increased CHD risk. Replacing 1 mol% PUFA n-6 or trans fatty acid with an equivalent amount of PUFA n-3 was associated with 10% lower CHD risk (OR, 0.90 [95% CI, 0.84–0.96]).⁶⁷

In a study using NHANES 2007 to 2014 data (N=18 434 participants), ORs for newly diagnosed hypertension comparing the highest and lowest tertiles were 0.60 (95% CI, 0.50–0.73) for dietary n-3 fatty acids, 0.52 (95% CI, 0.43–0.62) for dietary n-6 fatty acids, and 0.95 (95% CI, 0.79–1.14) for n-6:n-3 ratio.⁶⁸

In a prospective study of 3042 CVD-free adults followed up for a mean of 8.4 years, exclusive olive oil use was inversely associated with the risk of developing CVD (RR, 0.07 [95% CI, 0.01–0.66]) compared with no olive oil consumption.⁶⁹In the same study, adults with ≥50 mg/dL lipoprotein(a) had 2 times higher CVD risk than those with <50 mg/dL lipoprotein(a) (HR, 2.18 [95% CI, 1.11–4.28]), driven mainly by the lipoprotein(a) effect in males.⁷⁰

In a systematic review and dose-response meta-analysis of 123 prospective studies, the risk of CHD, stroke, and HF was inversely associated with consumption of whole grain, vegetables and fruits, nuts, and fish.⁷¹In contrast, the risk of these conditions was positively associated with consumption of egg, red meat, processed meat, and SSBs.

In a dose-response meta-analysis of prospective cohort studies in adults, each 250–mL/d increase in SSB and ASB intake was associated with an increased risk in obesity (RR, 1.12 [95% CI, 1.05–1.19] for SSB; 1.21 [95% CI, 1.09–1.35] for ASB), type 2 diabetes (1.19 [95% CI, 1.13–1.25] for SSB; 1.15 [95% CI, 1.05–1.26] for ASB), hypertension (1.10 [95% CI, 1.06–1.14] for SSB; 1.08 [95% CI, 1.06–1.10] for ASB), and total mortality (1.04 [95% CI, 1.01–1.07] for SSB; 1.06, [95% CI, 1.02–1.10] for ASB).⁷²A network meta-analysis of isocaloric substitution interventions in 38 RCTs involving 1383 participants suggested beneficial effects of replacing sucrose and fructose with starch for LDL-C and replacing fructose with glucose for insulin resistance and uric acid; however, the evidence was judged to be of low to moderate certainty and warrants replication.⁷³In a prospective study of 512 891 adults in China (only 18% consumed fresh fruit daily), individuals who ate fresh fruit daily had 40% lower risk of CVD death (RR, 0.60 [95% CI, 0.54–0.67]), 34% lower risk of incident CHD (RR, 0.66 [95% CI, 0.58–0.75]), 25% lower risk of ischemic stroke (RR, 0.75 [95% CI, 0.72–0.79]), and 36% lower risk of hemorrhagic stroke (RR, 0.64 [95% CI, 0.56–0.74]).⁷⁴

In a meta-analysis of 45 prospective studies, whole grain intake was associated with a lower risk of CHD (HR, 0.81 [95% CI, 0.75–0.87]) and CVD (HR, 0.78 [95% CI, 0.73–0.85]) but was not significantly associated with stroke (HR, 0.88 [95% CI, 0.75–1.03]).⁷⁵In another meta-analysis of 8 cohort or case-control studies, whole grain or cereal fiber intake was inversely associated with type 2 diabetes (RR, 0.68 [95% CI, 0.64–0.73]).⁷⁶

In a meta-analysis of 14 prospective cohort studies, every 20–g/d higher intake of fish was associated with 4% reduced risk of CVD mortality (RR, 0.96 [95% CI, 0.94–0.98]).⁷⁷The association was stronger in Asian cohorts than Western cohorts. Another meta-analysis reported similar results on the beneficial association of higher fish intake with CHD incidence (RR, 0.91 [95% CI, 0.84–0.97]) and mortality (0.85 [95% CI, 0.77–0.94]).⁷⁸In the REGARDS study, individuals who consumed ≥2 servings of fried fish per week had a greater risk of CVD over 5.1 years of follow-up than those who consumed <1 serving per month (HR, 1.63 [95% CI, 1.11–2.40]).⁷⁹

In a meta-analysis of prospective cohort and case-control studies from multiple countries, consumption of unprocessed red meat was not significantly associated with incidence of CHD. In contrast, each 50-g serving per day of processed meats was associated with a higher incidence of CHD (RR, 1.42 [95% CI, 1.07–1.89]).⁸⁰In an RCT (N=113 healthy adults), LDL-C and apolipoprotein B were significantly higher with red and white meat than with nonmeat consumption for 4 weeks, regardless of SFA content. Regardless of protein source, high SFA content (≈14% total energy) significantly increased LDL-C, apolipoprotein B, and large LDL particles compared with low SFA content (≈7% total energy).⁸¹

In a study of 169 310 female nurses and 41 526 male health professionals, consumption of 1 serving of nuts ≥5 times per week was associated with lower risk of CVD (HR, 0.86 [95% CI, 0.79–0.93]) and CHD (HR, 0.80 [95% CI, 0.72–0.89]) compared with never or almost never consuming nuts. Results were largely consistent for peanuts, tree nuts, and walnuts.⁸²In a meta-analysis of 61 trials (N=2582), tree nut consumption lowered TC by 4.7 mg/dL, LDL-C by 4.8 mg/dL, apolipoprotein B by 3.7 mg/dL, and triglycerides by 2.2 mg/dL. No heterogeneity by nut type was observed.⁸³In another meta-analysis of 5 prospective observational studies, consumption of legumes (beans) was associated with lower incidence of CHD (RR per 4 weekly 100-g servings, 0.86 [95% CI, 0.78–0.94]).⁸⁴

An umbrella review of 41 meta-analyses with 45 unique health outcomes concluded that milk consumption was more beneficial than harmful; for example, in dose-response analyses, an increment of 200 mL (≈1 cup) milk intake per day was associated with a lower risk of common cardiometabolic disease, such as CVD, stroke, hypertension, type 2 diabetes, MetS, and obesity.⁸⁵A meta-analysis of 10 cohort studies also showed that fermented dairy foods intake was associated with reduced CVD risk (OR, 0.83 [95% CI 0.76–0.91]), in particular cheese (0.87 [95% CI, 0.80–0.94]) and yogurt (0.78 [95% CI, 0.67–0.89]).⁸⁶

In a crossover RCT (n=25 normocholesterolemic and 27 moderately hypercholesterolemic participants), 8-week consumption of moderate amounts of a soluble green/roasted (35:65) coffee blend significantly reduced TC, LDL-C, very–low-density lipoprotein cholesterol, triglycerides, SBP, DBP, heart rate, and body weight among participants with moderate hypercholesterolemia. The beneficial influence on SBP, DBP, heart rate, and body weight was also observed in healthy participants.⁸⁷

In a cross-sectional study of 12 285 adults, for males, consumption of >30 g alcohol per day was significantly associated with a higher risk of MetS (OR, 1.73 [95% CI, 1.25–2.39]), HBP (OR, 2.76 [95% CI, 1.64–4.65]), elevated blood glucose (OR, 1.70 [95% CI, 1.24–2.32]), and abdominal obesity (OR, 1.77 [95% CI, 1.07–2.92]) compared with nondrinking.⁸⁸In males, drinkers at all levels had a lower risk of coronary disease than nondrinkers, whereas alcohol consumption was not associated with the risk of hypertension or stroke.⁸⁹In females, consumption of 10.1 to 15.0 g alcohol per day was associated only with a higher risk of elevated blood glucose (OR, 1.65 [95% CI, 1.14–2.38]) compared with nondrinking.⁸⁸Compared with nondrinkers, consumption of 0.1 to 10.0 g alcohol per day was associated with a lower risk of coronary disease and stroke and consumption of 0.1 to 15.0 g/d was associated with a lower risk of hypertension in females.⁸⁹

In a meta-regression analysis of 133 RCTs, a 100–mmol/d (2300–mg/d) reduction in sodium was associated with a 7.7–mm Hg (95% CI, −10.4 to −5.0) lower SBP and a 3.0–mm Hg (95% CI, −4.6 to −1.4) lower DBP among people with >131/78 mm Hg SBP/DBP. The association was weak in people with ≤131/78 mm Hg SBP/DBP: A 100–mmol/d reduction in sodium was associated with a 1.46–mm Hg (95% CI, −2.7 to −0.20) lower SBP and a 0.07–mm Hg (95% CI, −1.5 to 1.4) lower DBP.⁹⁰The effects of sodium reduction on BP appear to be stronger in individuals who are older, hypertensive, and Black.^91,92

In a systematic review and nonlinear dose-response meta-analysis of 14 prospective cohort studies and 1 case-control study, a 1–g/d increment in sodium intake was associated with a 6% increase in stroke risk (RR, 1.06 [95% CI, 1.02–1.10]), and a 1-unit increment in dietary sodium–to–potassium ratio (millimoles per millimole) was associated with a 22% increase in stroke risk (RR, 1.22 [95% CI, 1.04–1.41]).⁹³

Nearly all observational studies demonstrate an association between higher estimated sodium intakes (eg, >4000 mg/d) and a higher risk of CVD events, in particular stroke.^94–98Some studies have also observed higher CVD risk at estimated low intakes (eg, <3000 g/d), which suggests a potential J-shaped relationship with risk. An AHA science advisory suggested that variation in methodology might account for inconsistencies in the relationship between sodium and CVD in observational studies. Increased risk at low sodium intake in some observational studies could be related to reverse causation (illness causing low intake) or imprecise estimation of sodium intake through a single dietary recall or a single urine excretion.⁹⁸

In a meta-analysis of 133 RCTs with 12 197 participants, interventions with reduced sodium versus usual sodium resulted in a mean reduction of 130 mmol (95% CI, 115–145) in 24-hour urinary sodium, 4.26 mm Hg (95% CI, 3.62–4.89) in SBP, and 2.07 mm Hg (95% CI, 1.67–2.48) in DBP. The results also showed a dose-response relationship between each 50–mmol reduction in 24-hour sodium excretion and a 1.10–mm Hg (95% CI, 0.66–1.54) reduction in SBP and a 0.33–mm Hg (95% CI, 0.04–0.63 mm Hg) reduction in DBP. BP-lowering effects of sodium reductions were stronger in older people, populations that are not White, and those with higher baseline SBP levels.⁹⁹

In a secondary analysis of the PREMIER trial, changes in phosphorus intake were not significantly associated with changes in BP. Phosphorus type (plant, animal, or added) significantly modified this association, with only added phosphorus associated with increases in SBP (mean coefficient, 1.24 mm Hg/100 mg [95% CI, 0.36–2.12]) and DBP (0.83 mm Hg/100 mg [95% CI, 0.22–1.44]). An increase in urinary phosphorus excretion was significantly associated with an increase in DBP (0.14 mm Hg/100 mg [95% CI, 0.01–0.28]).¹⁰⁰

In a systematic review and meta-analysis of 18 prospective cohort studies, the highest magnesium intake category was associated with an 11% decrease in total stroke risk (RR, 0.89 [95% CI, 0.83–0.94]) and a 12% decrease in ischemic stroke risk (RR, 0.88 [95% CI, 0.81–0.95]) compared with the lowest magnesium intake category. After further adjustment for calcium intake, the inverse association remained for total stroke (RR, 0.89 [95% CI, 0.80–0.99]).¹⁰¹

In an RCT of 15 480 adults with diabetes and no history of ASCVD, 1 g n-3 fatty acids had no effect on first serious vascular event (RR, 0.97 [95% CI, 0.87–1.08]) or a composite outcome of first serious vascular event or revascularization (RR, 1.00 [95% CI, 0.91–1.09]) or mortality (RR, 0.95 [95% CI, 0.86–1.05]) compared with placebo (1 g olive oil).¹⁰²

A 2017 AHA science advisory summarized available evidence and suggested fish oil supplementation only for secondary prevention of CHD and SCD (Class IIa recommendation) and for secondary prevention of outcomes in patients with HF (Class IIa recommendation).¹⁰³

A meta-analysis of 77 917 participants in 10 RCTs with ≥500 participants treated for ≥1 year found that fish oil supplementation (EPA dose range, 226–1800 mg/d; DHA dose range, 0–1700 mg/d) had no significant effect on CHD death (RR, 0.94 [95% CI, 0.81–1.03]), nonfatal MI (RR, 0.97 [95% CI, 0.87–1.08]), or any CHD events (RR, 0.97 [95% CI, 0.93–1.01]).¹⁰⁴However, an updated meta-analysis of 124 477 participants (that included additional data from 3 large RCTs) found that marine omega-3 supplementation significantly lowered the risk of MI (RR, 0.92 [95% CI, 0.86–0.99]; P=0.020), CHD death (RR, 0.92 [95% CI, 0.86–0.98]; P=0.014), total CHD (RR, 0.95 [95% CI, 0.91–0.99]; P=0.008), CVD death (RR, 0.93 [95% CI, 0.88–0.99]; P=0.013), and total CVD (RR, 0.97 [95% CI, 0.94–0.99]; P=0.015). In addition, significant linear dose-response risk reductions were found for total CVD and major vascular events.¹⁰⁵

An observational study of 197 761 US veterans assessed omega-3 fatty acid supplement use and fish intake years on ischemic stroke over 3.2 years (2.2–4.3 years) and incident nonfatal CAD over 3.6 (2.4–4.7 years). It was found that omega-3 fatty acid supplement use was independently associated with a decreased risk of ischemic stroke (HR, 0.88 [95% CI, 0.81–0.95]) but not with nonfatal CAD. Fish intake was not independently associated with either outcome.¹⁰⁶

Results from a meta-analysis of 62 RCTs with 3772 participants showed that flaxseed supplementation improved TC (WMD, −5.389 mg/dL [95% CI, −9.483 to −1.295 mg/dL]), triglycerides (−9.422 mg/dL [95% CI, −15.514 to −3.330 mg/dL]), and LDL-C (−4.206 mg/dL [95% CI, −7.260 to −1.151 mg/dL]) concentrations.¹⁰⁷

In an RCT of 25 871 adults (males ≥50 years of age and females ≥55 years of age), the effects of daily supplementation of 2000 IU vitamin D and 1 g marine n-3 fatty acids on the prevention of cancer and CVD were examined.¹⁰⁸Vitamin D had no effect on major cardiovascular events (HR, 0.97 [95% CI, 0.85–1.12]), cancer (HR, 0.96 [95% CI, 0.88–1.06]), or any secondary outcomes. Marine n-3 fatty acid supplementation had no effect on major cardiovascular events (HR, 0.92 [95% CI, 0.80–1.06]), invasive cancer (HR, 1.03 [95% CI, 0.93–1.13]), or any secondary outcomes.

A secondary RCT data analysis study conducted across 3 years with 161 patients with advanced HF assessed the effects of daily vitamin D supplementation of 4000 IU on lipid parameters (TC, HDL-C, LDL-C, TC/HDL-C ratio, LDL-C/HDL-C ratio, and triglycerides) and vascular calcification parameters (fetuin-A and nonphosphorylated undercarboxylated matrix Gla protein). Long-term vitamin D supplementation did not improve lipid profiles and did not affect vascular calcification markers in these patients. In addition, no sex-specific vitamin D effects were found.¹⁰⁹A similar study, a post hoc analysis of the EVITA trial, assessing daily vitamin D₃supplementation of 4000 IU, also found no improvement in cardiac function among patients with advanced HF. However, subgroup analyses among those ≥50 years of age indicated improvements of 2.73% in LVEF (95% CI, 0.14%–5.31%) at the 12-month follow-up and 2.60% (95% CI, −2.47% to 7.67%) improvement at the 36-month follow-up.¹¹⁰

A Cochrane review of 1 RCT with 1355 females (with previous preeclampsia) from various hospital sites in Argentina, South Africa, and Zimbabwe who began calcium supplementation before conception (500 mg daily until 20 weeks’ gestation) found that calcium made little to no difference in developing serious health problems during pregnancy, including preeclampsia¹¹¹(RR, 0.80 [95% CI, 0.61–1.06]; P=0.121; low-quality evidence), severe maternal morbidity and mortality (RR, 0.93 [95% CI, 0.68–1.26]; low-quality evidence), pregnancy loss or stillbirth at any age (RR, 0.83 [95% CI, 0.61–1.14]; low-quality evidence), or a cesarean section (RR, 1.11 [95% CI, 0.96–1.28]; low-quality evidence). Calcium was found to slightly reduce the risk of a composite outcome of preeclampsia or pregnancy loss or stillbirth at any aage (RR, 0.82 [95% CI, 0.66–1.00]; low-quality evidence). Results should be interpreted with caution, particularly because ≈25% of the sample was lost to follow-up.¹¹²

The VITAL-HF, an ancillary study of the VITAL RCT, examined whether vitamin D₃(2000 IU/d) or marine omega-3 fatty acids (n-3; 1 g/d, including EPA 460 mg+ DHA 380 mg) were associated with first HF-related hospitalization or recurrent hospitalization for HF among 25 871 adults with HF between 2011 and 2017. No significant relationships were found between either vitamin D or n-3 fatty acid supplementation and first HF hospitalization. However, marine n-3 supplementation (326 events) significantly reduced recurrent HF hospitalization compared with placebo (379 events; HR, 0.86 [95% CI, 0.74–0.998]; P=0.048).¹¹³

A secondary analysis of the WHI examining the efficacy of calcium and vitamin D supplementation on AF prevention found that calcium and vitamin D had no reduction in incidence of AF compared with placebo (HR, 1.02 [95% CI, 0.92–1.13]). Although a relationship between baseline CVD risk factors and vitamin D deficiency was present, no significant association was found between baseline 25-hydroxyvitamin D serum levels and incident AF (HR, 0.92 in lowest versus highest subgroup [95% CI, 0.66–1.28]). Similarly, using data from the WHI RCT, another study examined whether calcium and vitamin D supplementation (1000 mg elemental calcium carbonate and 400 IU vitamin D₃/d) moderated the effects of premenopausal hormone therapy on CVD events among 27 347 females. Females reporting prior hysterectomy (n=16 608) were randomized to the conjugated equine estrogens (0.625 mg/d)+medroxyprogesterone (2.5 mg/d) trial, and those without prior hysterectomy (n=10 739) were randomized to the conjugated equine estrogen trial (0.625 mg/d). In the conjugated equine estrogen trial, receiving calcium and vitamin D was associated with lowered stroke risk (HR, 0.49 [95% CI, 0.25–0.97]). In both trials, in females with a low intake of vitamin D, a significant synergist effect of calcium and vitamin D and hormone therapy on LDL-C was observed (P=0.03).¹¹⁴

A meta-analysis of 14 RCTs with 1088 participants 4 to 19 years of age concluded that the evidence does not support vitamin D supplementation for improving cardiometabolic health in children and adolescents.¹¹⁵Another review article similarly reported that vitamin D supplementation had no beneficial effects on SBP and DBP in children and adolescents.¹¹⁶

Meta-analyses of RCTs examining the effects of multivitamins, vitamin D, calcium, vitamin C, B-complex, antioxidants, and vitamin B₃(niacin) have demonstrated no salutary cardiovascular benefits.¹¹⁷

An umbrella review of 10 systematic reviews and meta-analyses examined the relationship between vitamin C supplementation and CVD biomarkers (ie, cardiovascular arterial stiffness, BP, lipid profile, endothelial function, and glycemic control) and found weak evidence for salutary effects from vitamin C supplementation on CVD biomarkers. However, subgroup analyses revealed that specific groups of participants (ie, those who were older or with higher BMI, elevated CVD risk, and lower intake of vitamin C) may benefit from vitamin C supplementation.¹¹⁸

A 2-sample mendelian randomization study including 7781 individuals of European descent examined the relationship between vitamin E and risk of CAD and found higher vitamin E to be associated with a higher risk of CAD and MI. Specifically, each 1–mg/L increase in vitamin E was significantly associated with CAD (OR, 1.05 [95% CI, 1.03–1.06]), MI (OR, 1.04 [95% CI 1.03–1.05]); elevated TC (SD, 0.043 [95% CI, 0.038–0.04]), LDL-C (SD, 0.021 [95% CI, 0.016–0.027]), and triglycerides (SD, 0.026 [95% CI, 0.021–0.031]); and lower levels of HDL-C (SD, −0.019 [95% CI, −0.024 to −0.014]).¹¹⁹

Meta-analyses of folic acid RCTs suggested reductions in stroke risk (RR, 0.80 [95% CI, 0.69–0.93]) and CVD (RR, 0.83 [95% CI, 0.73–0.93]), although the benefit was driven mainly by the China Stroke Primary Prevention Trial, a large RCT of 20 702 adults with hypertension and no history of stroke or MI.¹²⁰

A meta-analysis of price comparisons of healthy versus unhealthy diet patterns found that the healthiest diet patterns cost, on average, ≈$1.50 more per person per day to consume.¹²³

In a 1-year (2013–2014) RCT of 30 after-school programs in South Carolina, site leaders in the intervention group received assistance in establishing snack budgets and menus and identifying low-cost outlets to purchase snacks that met healthy eating standards. The intervention was successful in increasing the number of days that fruits and vegetables were served (3.9 d/wk versus 0.7 d/wk) and decreasing the number of days that SSBs (0.1 d/wk versus 1.8 d/wk) and sugary foods (0.3 d/wk versus 2.7 d/wk) were served.¹²⁴Cost in the intervention group was minimized by identifying low-cost grocery outlets or large bulk warehouse stores; cost increased by $0.02 per snack in the intervention group compared with a $0.01 per snack decrease in the control group.

A study evaluated the health care costs associated with following the Healthy US-Style eating pattern (measured by the HEI) and the Healthy Mediterranean-Style eating pattern (measured by the Mediterranean diet score) and found that a 20% increase in compliance with the HEI was estimated to result in annual cost savings of $31.5 billion (range, $23.9–$38.9 billion). Half of the cost savings were attributed to the reduction in costs associated with CVD, whereas the other half were attributed to cancer and type 2 diabetes cost reductions. Similarly, a 20% increase in conformance with the Mediterranean diet score resulted in annual cost savings of $16.7 billion (range, $6.7–$25.4 billion). The biggest contributors to these costs savings were HD ($5.4 billion), type 2 diabetes ($4.6 billion), AD ($2.6 billion), stroke ($1.0 billion), and, to a lesser degree, site-specific cancer (125

Based on combined data from NHANES (2013–2016) and a community-based randomized trial of cash and subsidized CSA intervention, a microsimulation model was developed to assess the cost-effectiveness of improving dietary quality (as measured by the HEI) on CVD and type 2 diabetes in US adults with low income. The implementation of the model in the short term (10-year time horizon) and long term (life-course time horizon) demonstrated that both a cash transfer ($300) and subsidized CSA ($300/y subsidy) lowered total discounted DALYs accumulated over the life course attributable to CVD and diabetes complications from 24 797 per 10 000 people (95% CI, 24 584–25 001) at baseline to 23 463 per 10 000 (95% CI, 23 241–23 666) under the cash intervention and 22 304 per 10 000 (95% CI, 22 084–22 510) under the CSA intervention. Both interventions demonstrated incremental cost-effectiveness ratios of 126

A global cost-effectiveness analysis modeled the cost-effectiveness of a so-called soft regulation national policy to reduce sodium intake in countries around the world using the UK experience (government-supported industry agreements, government monitoring of industry compliance, public health campaign).¹²⁷Model estimates were based on sodium intake, BP, and CVD data from 183 countries. Country-specific cost data were used to estimate the cost-effectiveness ratio, defined as purchasing power parity–adjusted international dollars (equivalent to country-specific purchasing power of US $1) per DALY saved over 10 years. Globally, the estimated average cost-effectiveness ratio was $204 (international dollars) per DALY (95% CI, 149–322) saved. The estimated cost-effectiveness ratio was highly favorable in high-, middle-, and low-income countries. A US study examined the cost-effectiveness of implementing voluntary sodium target reformulation among people ever working in the food system and those in the processed food industry and found benefits in both. Achieving FDA reformulations across 10 years could lead to 20-year health gains in those who had ever worked in the food system of 180 000 QALYs (95% UI, 150 000–209 000) and health care–related savings of $5.2 billion (95% UI, 3.5–8.3 billion) with an incremental cost-effectiveness ratio of $62 000 (95% UI, 1000–171 000) per each QALY gained. Those working in the processed food industry could see similar improvements of 32 000 gained QALYs (95% UI, 27 000–37 000), health cost savings of $1 billion (95% UI, 0.7–1.6 billion), and an incremental cost-effectiveness ratio of $486 000 (95% UI, 148 000–1 094 000) for each QALY gained. The long-term reformulation would cost the industry $16.6 billion (95% UI, 12–31 billion). This highlights that potential health benefits and cost savings are greater than the costs associated with sodium reformulation.¹²⁸

A policy review of worldwide consumption of SSBs found that SSB consumption has increased significantly, which is problematic given the mounting evidence illustrating the association between high SSB daily intake and heightened risk of obesity and CVD. This review also presents evidence in support of an SSB tax because of its effectiveness in lowering SSB consumption in several countries to date.¹²⁹In the United States, a validated microsimulation model (CVD PREDICT) was used to assess cost-effectiveness, CVD reductions, and QALYs gained as a result of imposing a penny-per-ounce tax on SSBs. Cost savings were identified for the US government ($106.56 billion) and private sector ($15.60 billion). A 100% price pass-through led to reductions of 4494 (2.06%) lifetime MI events (95% UI, 2640–6599) and 1540 (1.42%) total IHD deaths (95% UI, 995–2118) versus no tax and to a gain of 0.020 lifetime QALYs. The lifetime cost to the beverage industry is $0.92 billion (or $49.72 billion if electing to absorb half the proposed SSB tax).¹³⁰Similar evidence was found in the Philippines, where a 13%/L SSB tax was associated with fewer deaths resulting from diabetes (−5913), IHD (−10 339), and stroke (−7950) across 20 years and averting 13 890 cases of catastrophic expenditure. In addition, health care savings of $627 million and annual revenue increases of $813 million were projected over 20 years.¹³¹

The GBD 2020 study produces comprehensive and comparable estimates of disease burden for 370 reported causes and 88 risk factors for 204 countries and territories from 1990 to 2020. The age-standardized mortality rate attributable to dietary risks was highest in Central Asia (Chart 5-6).

An updated report from the GBD 2019 Study estimated the impact of 15 dietary risk factors on mortality and DALYs worldwide using a comparative risk assessment approach.¹³⁹In 2019, an estimated 7.9 million deaths (95% UI, 6.5–9.8 million; 14% of all deaths) and 188 million DALYs (95% UI, 156–225 million; 7% of all DALYs) were attributable to dietary risks. The leading dietary risk factors were high sodium intake (1.9 million [95% UI, 0.5–4.2 million] deaths), low whole grain intake (1.8 million [95% UI, 0.9–2.3 million] deaths), and low legume intake (1.1 million [95% UI, 0.3–1.8 million] deaths). Countries with low-middle Socio-Demographic Index and middle Socio-Demographic Index had the highest age-standardized rates of diet-related deaths (119 [95% UI, 96–147] and 116 [95% UI, 92–147] deaths per 100 000 population), whereas countries with high Socio-Demographic Index had the lowest age-standardized rates of diet-related deaths (56 [95% UI, 47–69] deaths per 100 000 population). Age-standardized diet-related death rates decreased between 1990 and 2019 from 154 (95% UI, 128–186) to 101 (95% UI, 82–124) deaths per 100 000 population, although the proportion of deaths attributable to dietary risks was largely stable.

For adults, the NHLBI weight categories are as follows: overweight (BMI, 25.0–29.9 kg/m²) and obese class I (BMI, 30.0–35.0 kg/m²), class II (BMI, 35.0–39.9 kg/m²), and class III (BMI ≥40.0 kg/m²). BMI cutoffs often misclassify obesity in those with muscle mass on the upper and lower tails of the distribution. BMI categories also vary in prognostic value by race and ethnicity; they appear to overestimate risk in Black people and underestimate risk in Asian people.⁴For this reason, lower BMI cutoffs have been recommended to identify increased health risks for Asian and South Asian populations.⁵

For youth, sex-specific BMI-for-age 2000 CDC growth charts for the United States are used,⁶and overweight is defined as 85th to <95th percentile and obesity as ≥95th percentile. A 2013 AHA scientific statement recommended a definition of severe obesity for children ≥2 years of age and adolescents of BMI ≥120% of the 95th percentile for age and sex or an absolute BMI ≥35 kg/m², whichever is lower.⁷NHANES typically uses a definition of severe obesity for children ≥2 years of age and adolescents of BMI ≥120% of the 95th percentile for age and sex.⁸

Current obesity guidelines define WC ≥40 in (102 cm) for males and ≥35 in (88 cm) for females as being associated with increased cardiovascular risk⁹; however, different cutoffs have been recommended for various racial and ethnic groups, for example, ≥90 cm for Asian males and ≥80 cm for Asian females^4,10and >97 cm for Hispanic/Latino women.¹¹WC measurement is recommended for those with BMI of 25 to 34.9 kg/m²to provide additional information on CVD risk.¹²

According to 2015 to 2018 data from NHANES, the overall prevalence of obesity (≥95th percentile) among youth 2 to 19 years of age was 19.0% (Table 6-1). A similar prevalence was found with the use of NHANES data from 2017 to 2018, with higher prevalence in older age groups (Chart 6-1).^13,14

According to 2015 to 2018 data from NHANES, prevalence of obesity was lower for NH Asian boys and girls than youth in other racial and ethnic groups (Table 6-1).¹⁴Similar prevalences were found with the use of NHANES data from 2017 to 2018 (Chart 6-2).¹³

Prevalence of childhood obesity varies by SES.

According to NHANES 1999 to 2014, prevalence of obesity among adolescents 12 to 19 years of age was 21.6% (95% CI, 18.5%–24.7%) in the South region, 20.8% (95% CI, 17.6%–24.0%) in the Midwest region, 18.2% (95% CI, 13.1%–23.4%) in the Northeast region, and 15.8% (95% CI, 12.6%–19.1%) in the West region.¹⁶

According to self-reported height and weight data from the YRBSS 2019, 15.5% of US high school students had obesity and 16.1% were overweight. Obesity was more common in males (18.9%) than females (11.9%) and in Black students (21.1%) and Hispanic students (19.2%) than in White students (13.1%).¹⁷

According to NHANES 2015 to 2018, among US adults ≥20 years of age, the age-adjusted prevalence of obesity was 39.9% in males and 41.1% in females (Table 6-1). The prevalence of severe obesity (BMI ≥40 kg/m²) was 6.2% in males and 10.5% in females.

In both males and females according to NHANES 2015 to 2018, the prevalence of obesity was lowest in NH Asian adults. Among males, the prevalence of obesity was highest among Hispanic males. Among females, the prevalence of obesity was highest among NH Black and Hispanic females (Table 6-1).

According to NHANES 2017 to 2018, the age-adjusted prevalence of obesity was 44.8% among middle-aged (40–59 years of age) adults, 42.8% among older (≥60 years of age) adults, and 40.0% among younger (20–39 years of age) adults. No significant differences by age groups or between males and females were observed (Chart 6-3).¹⁸

Among females, according to 2001 to 2014 NHANES, obesity prevalence was inversely associated with income and educational attainment among females. For example, females with a household income ≤130% of the federal poverty level had a prevalence of obesity of 45.2%, those with household income of 130% to 350% of the federal poverty level had a prevalence of 42.9%, and those with household income >350% of the federal poverty level had a prevalence of 29.7%. Among males, the relationship is not as clear. Males with a household income ≤130% of the federal poverty level had a prevalence of obesity of 31.5%; those with household income of 130% to 350% of the federal poverty level had a prevalence of 38.5%; and those with household income >350% of the federal poverty level had a prevalence of 32.6%.¹⁹

In NHANES 2013 through 2016, the age-adjusted prevalence of obesity and severe obesity was generally higher among individuals living in areas with higher levels of urbanization. For example, females living in nonmetropolitan statistical areas had a prevalence of obesity of 47.2% compared with 38.1% among females living in large metropolitan statistical areas.²⁰

Self-reported BMI weight and height data are available through BRFSS.^21,22

According to NHANES data, overall prevalence of obesity and severe obesity in youth 2 to 19 years of age increased from 13.9% to 19.3% and 2.6% to 6.1% between 1999 to 2000 and 2017 to 2018. Over the same period, prevalence of obesity and severe obesity increased from 14.0% to 20.5% and from 3.7% to 6.9% for males and from 13.8% to 18.0% and from 3.6% to 5.2% for females.¹³

Among children 2 to 5 years of age, prevalence of obesity was 10.3% in 1999 to 2000 and 13.4% in 2017 to 2018, 9.5% and 14.7% for males, and 11.2% and 12.2% for females.¹³Among children 6 to 11 years of age, the prevalence of obesity was 15.1% in 1999 to 2000 and 20.3% in 2017 to 2018, 15.8% and 21.3% for males, and 14.3% and 19.2% for females. Among adolescents 12 to 19 years of age, the prevalence of obesity was 14.8% in 1999 to 2000 and 21.2% in 2017 to 2018, 14.8% and 22.5% for males, and 14.8% and 19.9% for females.

The change in the prevalence of obesity between 1999 and 2018 was not significant for youth <6 years of age but was for adolescents.⁸

From 1999 through 2000 to 2017 through 2018, the prevalence of obesity for US children 2 to 19 years of age increased from 11.0% to 16.1% for NH White children, from 18.8% to 24.2% for NH Black children, and from 20.2% to 26.9% for Mexican American children.¹³For NH Asian children, data have been available since 2011 to 2012, and prevalence of obesity remained stable for NH Asian children from 8.6% in 2011 to 2012 to 8.7% in 2017 to 2018.

According to the YRBSS, among US high school students, prevalence of obesity increased from 10.6% in 1999 to 15.5% in 2019.¹⁷

From NHANES data, from 1999 to 2000 through 2017 to 2018, the age-standardized prevalence of obesity and severe obesity (BMI ≥40 kg/m²) increased significantly from 30.5% to 42.4% and from 4.7% to 9.2%, respectively (Chart 6-8).¹⁸

From NHANES data, from 1999 to 2000 through 2017 to 2018, prevalence of obesity among males increased from 27.5% (95% CI, 24.3%–30.8%) to 43.0% (95% CI, 37.6–48.6%), and severe obesity increased from 3.1% (95% CI, 1.9%–4.7%) to 6.9% (95% CI, 5.1%–9.1%). Prevalence of obesity among females increased from 33.4% (95% CI, 29.8%–37.1%) to 41.9% (95% CI, 37.8%–46.1%) and severe obesity from 6.2% (95% CI, 5.0%–7.7%) to 11.5% (95% CI, 8.9%–14.5%).⁸

Significant increases in the prevalence of obesity were seen between 1999 to 2000 and 2017 to 2018 in all age-race and ethnicity groups except for NH Black males, in whom the prevalence increased from 1999 through 2006 (Chart 6-9).⁸

Comparing NHANES 1999 and 2016 shows an increase in mean weight, WC, and BMI in adults. No changes in height were seen in most demographic subgroups, and height decreased in some subgroups.²³

Overweight and obesity have considerable genetic components, with heritability estimates ranging from ≈30% to 75%.^24,25Estimates suggest that as much as 21% of variation in BMI can be explained by genetic variation in commonly occurring SNPs.²⁶This suggests a role for DNA methylation variants in explaining the genetic contributions to obesity.²⁷

Monogenic or mendelian causes of obesity include variants with strong effects in genes that control appetite and energy balance (eg, LEP, MC4R, POMC) and obesity that occurs in the context of genetic syndromes (eg, Prader-Willi syndrome).²⁸

GWASs in diverse populations have implicated multiple loci for obesity, defined mostly by BMI, WC, or waist-hip ratio. The FTO locus is the most well-established obesity locus, first reported in 2007^29,30and replicated in many studies with diverse populations and age groups since then.^31–35The mechanisms underlying the association remain incompletely elucidated but could be related to mitochondrial thermogenesis⁵or food intake.³⁶

Other GWASs have reported numerous additional loci,³⁷with >300 putative loci, most of which explain only a small proportion of the variance in obesity, have not been mechanistically defined, and have unclear clinical significance.

A GWAS of BMI in >330 000 individuals identified 97 loci, accounting for ≈2.7% of BMI variation, with genes related to synaptic function, glutamate signaling, insulin secretion, energy metabolism, lipid biology, and adipogenesis.²⁶

A meta-analysis of GWASs of childhood BMI in >46 000 children from 33 studies identified 15 genetic loci associated with childhood BMI; although most of these are loci found from adult BMI GWASs, 3 novel loci were identified, suggesting that the genetics of BMI are common in children and adults. Of note, a risk score combining all 15 loci explained only 2% of the variance in childhood BMI.³⁸

Variants associated with lean mass also have been reported.^39,40Fine mapping of loci, including efforts focused on GWASs in African ancestry, in addition to mechanistic studies, is required to define functionality of obesity-associated loci.⁴¹

Aggregating individual genetic variants associated with BMI into a GRS comprising 2.1 million common variants demonstrates the potential clinical utility of GRS over individual variants. In a study of 300 000 individuals, a BMI GRS was associated with a 13-kg gradient in weight and a 25-fold gradient in risk of severe obesity across GRS deciles.⁴²However, genetics are not deterministic for obesity; in fact, 17% of individuals in the top decile of the BMI GRS had a normal BMI.

It is important to note that a high GRS was associated with increased risk of 6 cardiometabolic diseases (28% increased risk of CAD, 72% increased risk of diabetes, 38% increased risk of hypertension, 34% increased risk of congestive HF, 23% increased risk of ischemic stroke, and 41% increased risk of VTE).⁴²

A mendelian randomization study has shown that a high BMI GRS is associated with shorter life span in the UK Biobank (HR of per 1-SD BMI GRS for increase in mortality, 1.07 [95% CI, 1.05–1.09]).⁴³

A large GWAS of obesity in >240 000 individuals of predominantly European ancestry revealed an interaction with smoking, which highlights the need to consider gene-environment interactions in genetic studies of obesity.⁴⁴Furthermore, a study of gene-environment interactions in the UK Biobank study found that gene-environment interactions increased the proportion of BMI variance explained by a GRS from 5.2% to 7.1%.⁴⁵

Rare variants have also been found to be associated with nonsyndromic obesity; in a study of 2737 individuals with severe obesity, rare variants in 3 novel genes (PHIP, DGKI, ZMYM4) were identified.⁴⁶

Genetic variants also are associated with weight loss response to dietary intervention.⁴⁷

Epigenetic modifications such as DNA methylation have both genetic and environmental contributors and may contribute to risk of and adverse consequences of obesity. An epigenome-wide association study in 479 people demonstrated that increased methylation at the HIF3A locus in circulating white blood cells and in adipose tissue was associated with increased BMI.⁴⁸

Beyond genetics, other molecular technologies have identified BMI and obesity biomarkers that have elucidated novel biology. For example, metabolomic profiling has uncovered that branched chain amino acids and related catabolic byproducts are dysregulated in patients with obesity.⁴⁹Branched chain amino acid biomarkers are also associated with response to weight loss interventions⁵⁰and cardiometabolic diseases.⁵¹

The microbiome has also been shown to be associated with BMI, with several microbial taxa associated with BMI.⁵²

In a 2016 meta-analysis based on studies conducted from 1958 to 2010, 70% of adults with obesity did not have obesity in childhood or adolescence.⁵³

The CDC Prevention Status Reports highlight the status of public health policies and practices to address public health problems, including obesity, by state. Reports rate the extent to which the state has implemented the policies or practices identified from systemic reviews, national strategies or action plans, or expert bodies.⁵⁴Obesity reduction policies and programs implemented by country are also available online.⁵⁵

The randomized Look AHEAD trial showed that among adults with type 2 diabetes who had overweight or obesity, an intensive lifestyle intervention produced a greater percentage of weight loss at 4 years than diabetes support education.^56,57After 8 years of intervention, the percentage of weight loss ≥5% and ≥10% was greater in the intensive lifestyle intervention group than in the diabetes support education group (50.3% and 26.9% for the intensive lifestyle group versus 35.7% and 17.2% for the diabetes support education group).⁵⁷

A comprehensive review and meta-analysis of 34 RCTs suggested that dietary weight loss interventions reduce all-cause mortality (RR, 0.82 [95% CI, 0.71–0.95]), but the benefit on lowering cardiovascular mortality was less clear.⁵⁸

A systematic review conducted for the US Preventive Services Task Force in 2018 found that behavior-based weight loss interventions with or without weight loss medications led to increased weight loss compared with usual care.⁵⁹These interventions also decreased the risk of incident diabetes.

Benefits reported for bariatric surgery include substantial weight loss; remission of diabetes, hypertension, and dyslipidemia; reduced incidence of mortality; reduction in microvascular disease; and fewer CVD events.^60,61

A meta-analysis of 3.74 million deaths among 30.3 million participants found that overweight and obesity were associated with higher risk of all-cause mortality, with the lowest mortality observed at BMI of 22 to 23 kg/m²among healthy never smokers.⁷¹

In 10 large population cohorts in the United States, individual-level data from adults 20 to 79 years of age with 3.2 million person-years of follow-up (1964–2015) demonstrated that obesity was associated with a shorter total longevity and increased cardiovascular morbidity and mortality.⁷²

According to data from the National Adult Cardiac Surgery registry from 2002 to 2013, there was lower mortality in individuals with overweight and class I and II obesity (OR, 0.79 [95% CI, 0.76–0.83], 0.81 [95% CI, 0.76–0.86], and 0.83 [95% CI, 0.74–0.94], respectively) relative to normal-weight individuals, as well as greater mortality risk in those who were underweight (OR, 1.51 [95% CI, 1.41–1.62]), with these results persisting after adjustment for residual confounding and reverse causation.⁷³

Fluctuation of weight is associated with cardiovascular events and death. In 9509 participants of the Treating to New Targets trial, those in the quintile of highest body weight fluctuation had the highest rates of cardiovascular events, MI, stroke, and death (85% higher, 117% higher, 136% higher, and 124% higher, respectively, compared with those in the lowest quintile of body weight fluctuation).⁷⁴

A systematic review and meta-analysis of 15 prospective cohort studies with 200 777 participants showed that children and adolescents who had obesity were ≈5 times more likely to have obesity in adulthood than those who did not have obesity. Approximately 55% of children with obesity will remain with obesity in adolescence; 80% of adolescents with obesity will remain with obesity in their adulthood; and 70% of these adolescents will remain with obesity at >30 years of age.⁵³

Children and adolescents who are overweight and have obesity are at increased risk for future adverse health effects⁷⁵such as increased prevalence of traditional cardiovascular risk factors, including hypertension, hyperlipidemia, and diabetes.^76,77Among 8579 youths in NHANES, higher BMI was associated with higher SBP and DBP, lower HDL-C, and higher triglyceride and HbA1c levels.⁷⁸

A systematic review and meta-analysis of 37 studies showed that high childhood BMI was associated with an increased incidence of adult diabetes (OR, 1.70 [95% CI, 1.30–2.22]) and CHD (OR, 1.20 [95% CI, 1.10–1.31]) but not stroke; however, the accuracy with which childhood BMI predicted any adult morbidity was low. Only 31% of future diabetes and 22% of future hypertension and CHD occurred in those who as youth ≥12 years of age had been classified as having overweight or obesity.⁷⁷

A study examining longitudinal data from 2.3 million adolescents (16–19 years of age) demonstrated increased cardiovascular mortality in adulthood among youth with obesity compared with youth with BMI in the 5th to 24th percentile, with an HR of 4.9 (95% CI, 3.9–6.1) for death attributable to CHD, 2.6 (95% CI, 1.7–4.1) for death attributable to stroke, 2.1 (95% CI, 1.5–2.9) for sudden death, and 3.5 (95% CI, 2.9–4.1) for death attributable to total cardiovascular causes, after adjustment for sex, age, birth year, sociodemographic characteristics, and height.⁷⁹