Trends in Cardiovascular Health Metrics in Obese Adults: National Health and Nutrition Examination Survey (NHANES), 1988–2014

National Health and Nutrition Examination Survey

NHANES is a serial cross‐sectional study involving nationally representative samples of the civilian noninstitutionalized US population. NHANES III was conducted during the period 1988–1994, and the continuous NHANES study has been conducted since 1999 with data released in 2‐year cycles. The survey is conducted by National Center for Health Statistics (NCHS) using a complex, stratified, multistage probability cluster design.17 The NCHS ethics review board approved all NHANES protocols before data collection, and all adult participants provided written informed consent.17 The participant response rate has consistently been ≈75% for interviews and ≈70% for examinations at mobile examination centers among adults (aged ≥20 years). This study was exempt from full board review by the institutional review boards at the University of Texas Medical Branch and the University of Alabama at Birmingham.

Race and ethnicity were self‐reported and classified as non‐Hispanic white, non‐Hispanic black, Mexican American, other Hispanic, and other. Standardized blood pressures were obtained by sphygmomanometer,18 and mean blood pressures from the last 3 measurements were used in this study. Blood lipids were measured enzymatically; hemoglobin A1c (HbA1c) was measured by high‐performance liquid chromatography. Detailed specimen collection and processing instructions are described in the NHANES Laboratory Procedures Manual.19 Obesity was defined according to body mass index (BMI; in kg/m²) status (BMI ≥30).

Cardiovascular Health Metrics

Cardiovascular health metrics included blood pressure, blood glucose, and lipid profile based on criteria modified from the American Heart Association (AHA)20 and the National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III)21 and the American College of Cardiology and AHA guidelines expert panel.22 Abnormal blood pressure status included systolic blood pressure ≥130 mm Hg, diastolic blood pressure ≥85 mm Hg, or use of antihypertensive medication; blood glucose status included HbA1c ≥5.7% or use of antidiabetic medication; and blood lipids status included total cholesterol ≥240 mg/dL, high‐density lipoprotein (HDL) cholesterol <40 mg/dL in men and <50 mg/dL in women, or use of lipid‐lowering medications. Participants with blood pressure health, blood glucose health, and blood lipid health were free from the above corresponding risk factors, according to these metrics. We also assessed trends in overall cardiovascular health status (presence of all 3 risk factors, presence of 1 or 2 risk factors, and freedom from all 3 risk factors). Ideal cardiovascular health was defined by the strict definition of the AHA: Blood pressure measured as untreated systolic <120 mm Hg and diastolic <80, blood glucose measured as untreated HbA1c <5.7%, and blood lipids measured as untreated total cholesterol <200 mg/dL and HDL ≥60 mg/dL.20

Statistical Analysis

Data from NHANES 1988–2014 were used to determine secular trends in the prevalence of each CVD risk factors and cardiovascular health status in the United States during the past 3 decades. All analyses using NHANES data took into account differential probabilities of selection and the complex sample design and nonresponse and noncoverage by using sample weights and SAS (SAS Institute) survey analysis procedures, following NHANES analytic and reporting guidelines.23, 24 Standard errors were calculated using Taylor series linearization. Age‐adjusted values25 were adjusted by the direct method to the year 2010 census population using the age groups 20 to 39, 40 to 59, and ≥60 years. Regression models were used to assess linear trends in the prevalence of healthy obese persons among obese adults during the period 1988–2014. Statistical analyses were carried out with SAS for Windows version 9.4 (SAS Institute). A 2‐sided P<0.05 was determined to be statistically significant.

Characteristics

The characteristics of participants from NHANES are provided in Table 1 for each of the sequential cross‐sectional surveys from 1988 to 2014. The study population over this time frame was composed of 18 626 obese adults. By way of illustration, the characteristics of obese participants from the most recent NHANES study, from 2011 to 2014, included a mean age of 47.2 years, mean systolic blood pressure of 124.4 mm Hg and diastolic blood pressure of 72.5 mm Hg, mean HbA1c of 5.9%, mean HDL cholesterol of 47.4 mg/dL, mean total cholesterol of 193.7 mg/dL, mean BMI of 36.0, and mean waist circumference of 114.8 cm. In examining trends in BMI and waist circumference in the analyses of participants with obesity, mean BMI was observed to significantly increase from 34.7 in 1988–1992 to 36.0 in 2011–2014 (Table 2), and waist circumference increased from 110.1 to 114.8 cm (Table 3).

Blood Pressure Health

Mean systolic blood pressure decreased from 126.1 mm Hg in 1988–1992 to 124.4 mm Hg in 2011–2014 (P for trend <0.001). The decrease was observed in all age groups, for both sexes, and in all racial and ethnic groups except for young adults (Table S1). Mean diastolic blood pressure also decreased from 76.6 to 72.5 mm Hg over the period 1988–2014 (Table S2). Prevalence of blood pressure health was stable during the 1988–2014 period and characterized just below 50% of the population (Table 4). In 2011–2014, 44.1% of men and 51.5% of women were below the threshold for the blood pressure risk factor. Prevalence of blood pressure health decreased as a function of age; >70% of young adults (aged 20–39 years) exhibited healthy blood pressure metrics, whereas only 20.4% of older adults (aged ≥60 years) did. Non‐Hispanic black participants had the lowest prevalence of blood pressure health, whereas Mexican American participants had the highest prevalence. For the prevalence of ideal blood pressure health (untreated systolic blood pressure <120 mm Hg and diastolic <80), we observed similar patterns, and the prevalence was stable, slightly above 30% (Table S3).

Blood Glucose Health

Mean HbA1c increased from 5.7% in 1988–1992 to 5.9% in 2011–2014 (P for trend <0.001). All age, sex, and racial and ethnic groups showed increases except non‐Hispanic black participants (Table S4). Prevalence of blood glucose health declined from 67.0% to 56.2% during the period 1988–2010 (P<0.001) but leveled off at ≈57% during 2011–2014 (Table 5). In 2011–2014, men and women had similar prevalence of ideal glucose health (57.4% versus 57.1%), and the prevalence declined by age (P<0.001); non‐Hispanic white participants had the highest prevalence, at 62.6%, whereas non‐Hispanic black participants had the lowest prevalence, at 44.2%.

An increase in the prevalence of diabetes mellitus (self‐reported or HbA1c ≥6.5%) contributed to declining rates of blood glucose health. Diabetes prevalence increased from 11.3% to 19.0% during the 1988–2014 period (Table S5). The prevalence increased with aging. A significant increase was observed in all age groups, for both sexes, and in all racial and ethnic groups except for young adults (aged 20–39 years).

Blood Lipid Health

Mean blood total cholesterol in obese adults decreased markedly by 20.8 mg/dL, from 214.5 mg/dL in the period 1988–1992 to 193.7 mg/dL in 2011–2014 (P<0.001). All subgroups experienced substantial decrements in mean total cholesterol during 1988–2014 (Table S6). In contrast, mean HDL cholesterol increased from 45.4 to 47.4 mg/dL during 1988–2014 (P<0.001 for trend) (Table S7). Prevalence of blood lipid health status changed little during the period 1988–2014 (Table 6). In 2011–2014, men had higher prevalence of lipids health than women (38.0% versus 31.7%), and the non‐Hispanic black group was the racial/ethnic group with the highest prevalence of lipids health (33.7%). In all constituencies, the prevalence of lipids health declined as a function of age; however, the prevalence of ideal blood cholesterol health (untreated total cholesterol <200 mg/dL and HDL ≥60 mg/dL) increased from 1988 to 2006 (3.3–5.2%) and then leveled off during 2007–2014 (Table S8).

Presence of All 3 CVD Risk Factors

Table 7 shows the trends in prevalence of the presence of all 3 risk factors among obese adults. Prevalence increased by 37% from 16.4% to 22.4% during the period 1988–2014. An increase was noted in all ages, sexes, and racial/ethnic subgroups. The increase occurred in parallel with a decline in the prevalence of healthy blood glucose, which is the predominant explanation accounting for the rise in the prevalence of presence of all 3 risk factors.

Presence of 1 to 2 CVD Risk Factors

Table 8 shows the trends in the prevalence of presence of 1 to 2 risk factors among obese adults. The prevalence decreased significantly from 69.6% to 62.4% during the 1988–2014 period. Only young adults (aged 20–39 years) and non‐Hispanic black participants did not see a decrease trend in prevalence over this time frame.

Freedom From CVD Risk Factors

As can be seen in Table 9, the prevalence of participants free from all 3 CVD risk factors remained stable at ≈15% among the general adult obese population and in all sex and racial/ethnic subgroups during the period 1988–2014. The prevalence remained highest in participants aged 20 to 39 years and declined progressively in the subgroups aged 40 to 59 and ≥60 years. Very few adult obese participants met the criteria for ideal cardiovascular health, and the prevalence of ideal cardiovascular health remained stable at ≈2% during the 1988–2014 period (Table 10).

We summarized the trends in cardiovascular health status in obese adults during NHANES 1988–2014 in Figure. Prevalence of freedom from all 3 CVD risk factors remained stable at ≈15% during 1988–2014, whereas prevalence of the presence of all 3 risk factors was greatly increased during this period. In addition, prevalence of ideal cardiovascular health remained very low at <2% over the period 1988–2014.

Cardiometabolic Health in People With Obesity

Obese participants who are free from cardiometabolic disease risk factors are termed “metabolically healthy obese.”9, 26 The existence of the metabolically healthy obese condition has been controversial27, 28, 29, 30; however, we have recently added some clarification to this issue by pointing out that conclusions rest heavily on the definition of metabolically healthy obese.31 We4, 5, 10 and others12, 13, 14 have shown that 1 or 2 metabolic syndrome traits confer increased risk of diabetes mellitus and CVD outcomes and that such persons are relatively insulin resistant. The problem with many reports minimizing the importance of the metabolically healthy obese condition is the inclusion of participants with 1 or 2 risk factors among metabolically healthy participants. When “metabolically healthy” is defined as the absence of all metabolic syndrome risk factors, there is minimal or no contribution of BMI to CVD events or mortality, and cumulative rates of diabetes mellitus remain quite low regardless of BMI. Obese participants with metabolic syndrome consistently have much greater risks for diabetes mellitus and CVD than their counterparts assigned to the metabolically healthy subgroup.12, 13, 14 We have also defined a subgroup of patients with suboptimal metabolic health who have 1 or 2 risk factors but who do not meet criteria for metabolic syndrome, and those persons have intermediate risks of incident diabetes mellitus and CVD outcomes.10, 31 BMI interacts with metabolic status to significantly augment diabetes risks such that progressive increments in BMI augment the risk of diabetes mellitus; however, the impact of BMI on rates of diabetes mellitus is greater among those with metabolic syndrome (ie, metabolically unhealthy obese), less so in those with suboptimal metabolic health, and least in metabolically healthy obese persons. In any event, cardiometabolic risk factors confer much higher risk of diabetes mellitus, CVD, stroke, and mortality than BMI per se.

Our finding that the prevalence of metabolically healthy obese participants who are free from all CVD risk factors has remained relatively stable at ≈15% during the past 3 decades in the United States agrees with a large collaborative study in Europe encompassing 163 517 participants (the BioSHaRE‐EU Healthy Obese Project). This European study demonstrated that the frequency of metabolically healthy obesity was 7% to 28% in women and 2% to 19% in men in various European countries.32 Although these persons have lower risks of total and CVD mortality and type 2 diabetes mellitus, the proportion of metabolically healthy persons among all obese adults is relatively low. We also observed that the prevalence of ideal cardiovascular health among obese people was extremely low over the past 3 decades, remaining consistently <2%, and that very few adult obese participants met the criteria for ideal blood cholesterol health (<5%). These results are consistent with previously reported very low prevalence of meeting cardiovascular health among the general adult population in the United States.33 Consistent with a previous report on the prevalence of CVD risk factors in US adults from 1960 to 2000,16 our results indicated that blood glucose health continued to deteriorate in obese people during the period 1988–2014. Further studies are need to unravel the underlying causes of this troubling phenomena. The improvements in mean blood pressure and blood cholesterol may reflect the healthy trends in the US population, including decreases in the proportion of dietary saturated fat; increases in fruits and vegetables and the proportion of polyunsaturated and monounsaturated fats in the diet34, 35, 36, 37; decreased smoking38, 39, 40; and increased awareness, identification, and pharmacological treatment of high cholesterol and blood pressure levels.25, 41

Obese participants with metabolic syndrome are at greater risk of diabetes and CVD outcomes.12, 13, 14 In our study, the prevalence of obese people with all 3 CVD risk factors kept increasing over the past 3 decades in the United States. The increasing trend of obese people with all 3 CVD risk factors, commensurate with a decline in those with 1 to 2 risk factors, suggested an overall deterioration in health among people with obesity. In recent years, almost 1 in 4 obese adults had all 3 CVD risk factors. During the period 1988–2010, adults in the United States experienced a significant decrease in diet quality and an increase in the prevalence of sedentary lifestyles.33 Those changes may partially explain our findings regarding the increasing trend in prevalence of presence of all 3 CVD risk factors in obese adults.

Implications

Although some data point to a leveling off of prevalence rates of obesity in recent years, our data indicate that overall metabolic health is declining among people with obesity in the United States. Furthermore, deteriorating metabolic health can be attributed primarily to worsening glycemic health because mean values of HbA1c have been increasing, whereas mean blood pressure and lipid metrics have been improving. These patterns of worsening metabolic health constitute an increase in risk of type 2 diabetes mellitus and underlie increasing prevalence rates for diabetes mellitus. Without intensive interventions, obese people do not lose sufficient body weight to become lean, regardless of their age31; however, moderate weight loss of 10% of body weight is sufficient to dramatically reduce progression to diabetes among high‐risk persons.42, 43 The deteriorated blood glucose health among obese adults in the United States calls for lifestyle interventions (diet and exercise) on a national scale. Community‐based public health intervention programs may help increase physical activity and diet quality to alleviate the problem.44, 45 Obese adults at high risk of diabetes and CVD may require more intense approaches to achieve weight loss using lifestyle interventions alone or in combination with weight loss medications.42, 43, 46, 47 A complication‐centric approach may be adopted to guide the selection of weight loss therapy options (eg, more aggressive treatment in those obese patients with complications or at higher risk of complications to optimize outcomes, the benefit–risk ratio of the intervention, and the cost‐effectiveness of care).48, 49 In the context of the current data, those obese adults who are metabolically unhealthy or perhaps those with suboptimal metabolic health represent patients who will benefit most from intensive obesity management. As suggested by Yang et al,33 coordinated efforts aligning CVD prevention and control activities across the public and private sectors in the United States are needed reduce the burden of CVD among the obese population.50

Strengths and Limitations

The main strengths of this study involve the use of sequential surveys over time, consisting of nationally representative data from NHANES. These data from the past 3 decades enabled us to assess the secular trends of prevalence of individual CVD risk factors and overall CVD health status among obese people in a nationally representative sample. Limitations of this study include the sample size, which may not be large enough to permit extensive subgroup analyses of each age, sex, and racial or ethnic group. In addition, we were able to assess only the trends in major racial and ethnic groups (eg, non‐Hispanic black, non‐Hispanic white, and Mexican American); we could not further differentiate other racial and ethnic groups because of the sample size limit. Furthermore, because the duration of physical activity (in minutes) changed substantially since the 2007–2008 cycle, we did not assess physical activity among obese adults during the 1988–2014 period.