Abstract
Background:
Obesity is associated with higher risk for coronary artery calcium (CAC), but the relationship between body mass index (BMI) and mortality is complex and frequently paradoxical.
Methods:
We analyzed BMI, CAC, and subsequent mortality using data from the CAC Consortium, a multi-centered cohort of individuals free of established CVD who underwent CAC testing. Mortality was assessed through linkage to the Social Security Death Index and cause of death from the National Death Index. Multivariable logistic regression was used to determine odds ratios (OR) for the association of clinically relevant BMI categories and prevalent CAC. Cox proportional hazards regression modeling was used to determine hazard ratios (HR) for CHD, CVD, and all-cause mortality according to categories of BMI and CAC.
Results:
Our sample included 36,509 individuals, mean age 54.1 [10.3] years, 34.4% female, median BMI 26.6 (IQR 24.1 – 30.1), 46.6% had zero CAC, and 10.5% had CAC ≥ 400. Compared to individuals with normal BMI, the multivariable-adjusted odds of CAC>0 were increased in those overweight (OR 1.13 [95% CI 1.1-1.2]) and obese (OR 1.5 [95% CI 1.4-1.6]). Over a median follow-up of 11.4 years there were 1,550 deaths (4.3%). Compared to normal BMI, obese individuals had a higher risk of CHD, CVD, and all-cause mortality while overweight individuals, despite a higher odds of CAC, showed no significant increase in mortality. In a sex-stratified analysis, the increase in CHD, CVD, and all-cause mortality in obese individuals appeared largely limited to men and there was a lower risk of all-cause mortality in overweight women (HR 0.79 [95% CI 0.63, 0.98]).
Conclusion:
In a large sample undergoing CAC scoring, obesity was associated with a higher risk of CAC and subsequent CHD, CVD, and all-cause mortality. However, overweight individuals did not have a higher risk of mortality despite a higher risk for CAC.
Keywords: Coronary Artery Calcium, Obesity, Cardiovascular Mortality
Introduction
Over one-third of US adults are defined as obese with a body mass index (BMI) ≥ 30.0 kg/m2, and approximately another 1/3 of US adults are overweight (25.0 - <30.0 kg/m2) (1). Obesity is associated with higher risk for multiple cardiovascular disease (CVD) risk factors as well as coronary artery calcium (CAC) (2,3). CAC, a direct measure of the coronary atherosclerotic burden, has been shown to be a strong predictor of coronary heart disease (CHD), CVD, and all-cause mortality (4,5). However, studies evaluating associations between overweight/obesity and CHD, CVD and all-cause mortality have frequently shown paradoxical results.
This obesity paradox has been demonstrated in the general population (6), as well as populations with CVD risk factors and established CVD (7–9). Additionally, some data suggest that the relationship between BMI and mortality may vary by sex (10). However, data supporting the obesity paradox have not been consistent, suggesting the paradox may be due to methodological issues and bias, with 2 large meta-analyses demonstrating a linear increase in risk for mortality in overweight and obese individuals compared to those with a normal BMI (11,12). The relationship between BMI and subclinical atherosclerosis as measured by CAC, and BMI and subsequent CHD, CVD, and all-cause mortality is complex and warrants further analysis. The CAC Consortium provides a novel opportunity to evaluate this relationship in a large cohort of asymptomatic individuals without established CVD at baseline with long term follow-up and cause-specific mortality.
Methods
Study design and population
The CAC Consortium is a multi-center retrospective cohort study of 66,636 patients intended to explore the association of CAC with long-term cause-specific mortality. Details of the rationale, design and baseline results of the CAC Consortium have been previously described (13). The four centers that contributed data to the CAC consortium have significant experience with CAC scoring and were spread across three states (California, Minnesota and Ohio) in the US. Patients were required to be at least 18 years old without cardiovascular symptoms or established CVD prior to CAC testing. Baseline data collected for the CAC consortium represent the years 1991 through 2010, with follow-up of the cohort through June 2014.
Consent for participation in research was obtained from all study participants at individual centers at the time of CAC scanning, while IRB approval for coordinating center activities including death ascertainment was obtained at the Johns Hopkins Hospital. The data that support the findings of this study are available from the corresponding author upon reasonable request. From the total cohort, we included 36,509 patients with data on BMI. Mean duration of follow-up for the cohort was 11.4 (IQR 9.3-12.6) years. In comparison to individuals missing data for BMI, individuals with data on BMI were slightly younger (mean age 54.1 [10.1] years versus 54.7 [10.9] years), more likely to be female (34.8% versus 30.9%), and had a lower prevalence of CAC (53.3% had CAC>0 versus 57.8%).
Measurement and definition of BMI
Height and weight were measured by staff at each scan site at the time of the CAC scan. Body Mass Index was calculated by weight in kg / height in m2. We categorized BMI according to guidelines for the management of overweight and obesity in adults (14): normal weight, BMI 21.0-24.9 kg/m2, overweight, BMI 25.0-29.9kg/m2, obese, BMI ≥ 30.0 kg/m2, and underweight, BMI < 21.0 kg/m2.
Definition of other CVD Risk Factors
Risk factor and laboratory data were obtained during the clinical visits that functioned as referrals for CAC testing or immediately prior to CAC testing with a separate in-person interview. Hypertension was noted from previous clinical diagnosis or anti-hypertensive treatment. Blood pressure measurements obtained at the CT scanning did not factor into the diagnosis of hypertension or lack thereof. Dyslipidemia was defined as a prior diagnosis of primary hyperlipidemia (low-density lipoprotein cholesterol (LDL-C) >160 mg/dL), prior diagnosis of dyslipidemia (elevated triglycerides >150 mg/dL and/or low high-density lipoprotein cholesterol <40 mg/dL in men and <50 mg/dL in women), or treatment with any lipid-lowering drug.
Smoking was based on self-report of currently smoking or not currently smoking. Diabetes was defined as a prior diagnosis of diabetes or treatment with oral hypoglycemic drugs or insulin. Family history of CHD was primarily defined by a first degree relative with any history of CHD.
To account for partially missing risk factor data (12.5% of the cohort, with a majority [8.6%] missing only a single risk factor), previously validated and described multiple imputation process was conducted (13). Nearly identical mean and median atherosclerotic CVD risk score values and a robust correlation coefficient of 0.952 between the imputed and directly calculated scores were seen with the previous validation.
Computed tomography data
Standard protocol non-contrast cardiac-gated CT scans for CAC scoring were performed. Electron beam tomography was used in 93% of scans, while multi-detector CT (MDCT) was performed in 7% of scans. Previously no clinically significant difference in CAC scoring has been demonstrated between these scanning technologies (15). In this analysis, approximately 13% of patients were scanned with the Imatron C-100 scanner, 38% with the Imatron C-150, 38% with the C-300 and 3.5% with the e-speed scanner (GE-Imatron). The rest of the scans (7%) were performed on a 4-slice MDCT scanner (Somatom Volume Zoom, Siemens) and the General Electric LightSpeed VCT 64-slice platform (GE Healthcare). This has been previously described in the CAC consortium design and rationale (13). CAC was quantified in Agatston units in all participants and considered prevalent if CAC >0 and also categorized as CAC 0, CAC 1-99, 100-399, and CAC ≥ 400.
Follow-up and Death Ascertainment
Participants were followed for the primary outcome of death through linkage to the social security death index Death Master file using a previously validated algorithm (16) that is similar to the algorithm used by the National Death Index service. Death was determined if there was a match on social security number and one additional patient identifier. To prioritize specificity over sensitivity, a complete match on all other patient identifiers was required when the social security number was not available.
Death certificates were obtained from the National Death Index and organized according to common causes of death using ICD9 and ICD10 codes as previously described. Comparison of death rates between the CAC Consortium with the U.S. Census and the multi-ethnic study of atherosclerosis (MESA) has been previously published (13). Internal validation studies against known deaths identified via the electronic medical record revealed >90% specificity for identifying known deaths, with sensitivity between 72-90%.
Statistical Methods
First, we summarized baseline characteristics of the study population and the prevalence of CAC by BMI category, reporting mean +/− standard deviation for numerical variables and n (%) for categorical variables. The relationship between BMI and CAC was evaluated using restricted Cubic splines, with BMI (x-axis) and CAC (y-axis) as continuous variables. Cubic spline models were adjusted for age and stratified by sex. Similar spline models were created to assess the relationship of continuous BMI with CHD, CVD, and all-cause mortality.
We then determined odds ratios (OR) and 95% confidence intervals (CI) for according to BMI categories for CAC using multivariable logistic regression. Model 1 was adjusted for age and sex. Model 2 was adjusted for age, sex, scan site, smoking status, hypertension, hyperlipidemia, diabetes, and family history of CHD.
Cox proportional hazards regression modeling was used to determine hazard ratios to assess the relationship between clinically relevant categories of BMI and CHD, CVD, and total mortality. A formal test of proportional hazards, based on Schoenfeld residuals, was used to determine that the assumption was satisfied. Model 1 adjusted for age and sex. Model 2 adjusted for age, sex, scan site, smoking status, and family history of CHD. A third model adjusted for the variables in model 2 as well as CAC. Similar analyses were then performed stratified by sex. To test for effect modification by CAC on the relationship between BMI and CHD, CVD, and all-cause mortality, an age and sex adjusted interaction variable (BMI category x CAC) was tested in the multivariable adjusted models with CAC used as a categorical variable. All analyses were performed with Stata Software version 14.2 (College Station, TX).
RESULTS
Baseline demographics and prevalence of CAC
Table 1 shows the baseline distribution of demographics and risk factors for the total sample of 36,509 individuals as well as by BMI categories. The range of BMI for the sample was 18.5 kg/m2 to 74.3 kg/m2, with a median BMI of 26.6kg/m2 (Interquartile range 24.1kg/m2 – 30.1kg/m2). The sample studied was relatively young (mean age 54.1 [10.3] years), 34.4% female, and overall healthy without baseline cardiovascular disease by definition and with a low prevalence of smoking (9.7%) and diabetes (6.1%). The population did have a high prevalence of hyperlipidemia (57.7%) and family history of CHD (47.5%).
Table 1.
Baseline Characteristics of 36,509 patients with BMI measured enrolled in the CAC Consortium
| All (N=36509) | BMI* 18.5-21 (N=1982) | BMI 21-25 (N=9718) | BMI 25-30 (N=15475) | BMI >30 (N=9334) | |
|---|---|---|---|---|---|
| Age | 54.1 +/− 10.3 | 55.1 +/− 11.1 | 54.9 +/− 10.7 | 54 +/− 10.2 | 53.3 +/− 9.7 |
| Female | 12555, 34.4% | 1607, 81.1% | 4322, 44.5% | 3632, 23.5% | 2994, 32.1% |
| Race Category | |||||
| White | 30450, 92.6% | 1613, 90.5% | 8018, 92.6% | 12942, 93% | 7877, 92.4% |
| Asian | 732, 2.2% | 110, 6.2% | 283, 3.3% | 260, 1.9% | 79, 0.93% |
| Black | 607, 1.9% | 20, 1.1% | 131, 1.5% | 242, 1.7% | 214, 2.5% |
| Hispanic | 663, 2.0% | 138, 1.6% | 138, 1.6% | 299, 2.2% | 207, 2.4% |
| Hypertension | 11045, 30.3% | 416, 21% | 2370, 24.4% | 4655, 30.1% | 3604, 38.6% |
| Hyperlipidemia | 21066, 57.7% | 817, 41.2% | 5025, 51.7% | 9425, 60.9% | 5799, 62.1% |
| Current Smoker | 3542, 9.7% | 170, 8.6% | 889, 9.2% | 1480, 9.6% | 1003, 10.8% |
| Family History of CHD | 17331, 47.5% | 941, 47.5% | 4445, 45.7% | 7195, 46.5% | 4750, 50.9% |
| Diabetes | 2215, 6.1% | 96, 4.8% | 365, 3.8% | 773, 5.0% | 981, 10.5% |
| 10-Year ASCVD Risk Score | 7.0 +/− 8.5 | 5.6 +/− 9.1 | 6.7 +/− 8.8 | 7.3 +/− 8.4 | 7.1 +/− 8.3 |
| CAC Categories | |||||
| CAC 0 | 17004, 46.6% | 1246, 62.9% | 5062, 52.1% | 6823, 44.1% | 3873, 41.5% |
| CAC 1-99 | 10758, 29.5% | 428, 21.6% | 2581, 26.6% | 4717, 30.5% | 3032, 32.5% |
| CAC 100-399 | 4912, 13.5% | 189, 9.5% | 1195, 12.3% | 2175, 14.1% | 1353, 14.5% |
| CAC ≥400 | 3835, 10.5% | 119, 6.0% | 880, 9.1% | 1760, 11.4% | 1076. 11.5% |
BMI = Body Mass Index, CHD = Coronary Heart Disease, ASCVD = Atherosclerotic Coronary Vascular Disease, CAC = Coronary Artery Calcium
In general, the prevalence of CVD risk factors increased across categories of increasing BMI and the BMI ≥30 category had the highest prevalence of hypertension (38.6%), hyperlipidemia (62.1%), and diabetes mellitus (10.5%). Individuals with BMI >30 also had the highest percentage of smokers (10.8%) and family history of CHD (50.9%).
The association of BMI with CAC
The prevalence of CAC (CAC >0) for the total sample and across categories of BMI are also shown in Table 1. Of the total sample, 46.6% had zero CAC at baseline and 10.5% had CAC ≥ 400. The prevalence of zero CAC in the categories of underweight, normal weight, overweight, and obese was 62.9%, 52.1%, 44.1%, and 41.5% respectively. The prevalence of CAC ≥ 400 increased across the 4 BMI categories (6.0%, 9.1%, 11.4%, and 11.5%). Women had significantly less CAC at baseline compared to men, with 61.3% having zero CAC compared to 38.9% of men having zero CAC and 4.9% of women having CAC ≥ 400 compared to 13.5% of men. Age-adjusted cubic splines demonstrating the relationship of BMI and CAC according to sex are shown in Figure 1.
Figure 1. Age-adjusted cubic splines demonstrating the linear relationship of BMI and CAC according to sex in 36,509 participants from the CAC consortium.
BMI = Body Mass Index, CAC = Coronary Artery Calcium
The odds of prevalent CAC according to BMI categories are shown in Table 2 for the total sample and stratified by sex. After adjustment for age, sex, and traditional CVD risk factors, compared to individuals with normal BMI, both overweight (OR 1.1 [95% CI 1.06, 1.2]) and obese (OR 1.5 [95% CI 1.4, 1.6]) individuals were more likely to have CAC. When stratified by sex, the relationship between BMI and prevalent CAC appeared linear in men as underweight men were less likely to have CAC (OR 0.77 [95% CI 0.60, 0.99]) compared to normal weight men while overweight men (OR 1.2 [95% CI 1.1, 1.3]) and obese men (OR 1.5 [95% CI 1.4, 1.7]) showed a graded increase in the odds of prevalent CAC. In women, there was no significant increase in the odds of prevalent CAC in those with underweight or overweight though there was increased odds of CAC in obese women (OR 1.3 (95% CI 1.2, 1.5]).
Table 2.
Multivariable adjusted odds ratios for the prevalence of CAC according to BMI categories for the total sample and stratified by sex.
| Model 1* | Model 2† | |
|---|---|---|
| Total Sample (n=36,509) | ||
| BMI <21 | 0.92 (0.92, 1.03) | 0.99 (0.88, 1.1) |
| BMI 21-24.9 | 1.0 | 1.0 |
| BMI 25-29.9 | 1.2 (1.2, 1.3) | 1.1 (1.06, 1.2) |
| BMI ≥30 | 1.7 (1.6, 1.8) | 1.5 (1.4, 1.6) |
| Women (n=12,555) | ||
| BMI <21 | 0.93 (0.81, 1.1) | 1.00 (0.97, 1.2) |
| BMI 21-24.9 | 1.0 | 1.0 |
| BMI 25-29.9 | 1.2 (1.1, 1.3) | 1.1 (0.95, 1.2) |
| BMI ≥30 | 1.6 (1.4, 1.7) | 1.3 (1.2, 1.5) |
| Men (n=23,954) | ||
| BMI <21 | 0.79 (0.62, 1.01) | 0.77 (0.60, 0.99) |
| BMI 21-24.9 | 1.0 | 1.0 |
| BMI 25-29.9 | 1.3 (1.2, 1.4) | 1.2 (1.1, 1.3) |
| BMI ≥30 | 1.8 (1.7, 2.0) | 1.5 (1.4, 1.7) |
Model 1 is adjusted for age and sex
Model 2 is adjusted for age, sex, scan site, hypertension, hyperlipidemia, diabetes, smoking status, and family history
CAC = Coronary Artery Calcium, BMI = Body Mass Index
Association between BMI and CHD, CVD, and All-Cause Mortality
Overall, there were 1,550 deaths (4.3% of the sample) over a median 11.4 year follow up. The multivariable adjusted HR for CHD, CVD, and all-cause mortality according to categories of BMI are shown in Table 3. Compared to individuals with normal BMI there was a higher risk for CHD (HR 1.6 [95% CI 1.2, 2.3]), CVD (HR 1.5 [95% CI 1.1, 1.9), and all-cause mortality (HR 1.3 [95% CI 1.1, 1.4]) in obese individuals. However, there was no significant increase in risk for CHD (HR 0.97 [95% CI 0.70, 1.4]), CVD (HR 0.97 [95% CI 0.77, 1.2]), or all-cause mortality (HR 0.93 [95% CI 0.82, 1.1]) in overweight individuals. Additionally, there was an increase in the risk for all-cause mortality in underweight individuals (HR 1.4 [95% CI 1.1, 1.7]). The addition of CAC to the survival model resulted in modest attenuation of the hazard ratios for CHD, CVD, and all-cause mortality (Model 3, Table 3). There was not statistically significant evidence of effect modification by CAC on the relationship between BMI and mortality. Testing an age and sex adjusted interaction variable (BMI Category x CAC) showed no significant interaction for CHD mortality (p-value 0.09), CVD death (p-value 0.18), or all-cause mortality (p-value 0.80).
Table 3.
The adjusted hazard ratios for CHD, CVD, and all-cause mortality according to categories of BMI in 36,509 participants from the CAC consortium.
| Categories of BMI | Model 1 | Model 2 | Model 3 |
|---|---|---|---|
| CHD Mortality | |||
| BMI <21 | 1.2 (0.65, 2.4) | 1.2 (0.64, 2.4) | 1.3 (0.65, 2.4) |
| BMI 21-24.9 | 1.0 | 1.0 | 1.0 |
| BMI 25-29.9 | 0.92 (0.66, 1.3) | 0.97 (0.70, 1.4) | 0.93 (0.67, 1.3) |
| BMI ≥30 | 1.5 (1.04, 2.1) | 1.6 (1.2, 2.3) | 1.5 (1.04, 2.1) |
| CVD Mortality | |||
| BMI <21 | 1.1 (0.74, 1.8) | 1.2 (0.75, 1.8) | 1.2 (0.76, 1.8) |
| BMI 21-24.9 | 1.0 | 1.0 | 1.0 |
| BMI 25-29.9 | 0.95 (0.75, 1.2) | 0.97 (0.77, 1.2) | 0.94 (0.75, 1.2) |
| BMI ≥30 | 1.4 (1.1, 1.8) | 1.5 (1.1, 1.9) | 1.3 (1.04, 1.7) |
| All-Cause Mortality | |||
| BMI <21 | 1.3 ( 1.1, 1.7) | 1.4 ( 1.1, 1.7) | 1.4 ( 1.1, 1.7) |
| BMI 21-24.9 | 1.0 | 1.0 | 1.0 |
| BMI 25-29.9 | 0.92 (0.81, 1.04) | 0.93 (0.82, 1.1) | 0.92 (0.81, 1.0) |
| BMI ≥30 | 1.2 (1.1, 1.4) | 1.3 ( 1.1, 1.4) | 1.2 ( 1.1, 1.4) |
Model 1 is adjusted for age and sex
Model 2 is adjusted for age, sex, scan site, smoking status, and family history.
Model 3 is adjusted for the variables in model 2 plus CAC categories
BMI = Body Mass Index, CAC = Coronary Artery Calcium, CHD = Coronary Heart Disease, CVD = Cardiovascular Disease
When stratified by sex (Table 4), the increase in risk for CHD, CVD, and all-cause mortality in obese individuals appeared similar in women and men. However, overweight women had a significantly lower risk for all-cause mortality compared to normal weight women (HR 0.79 [95% CI 0.63, 0.98]) with no significant association in overweight men (HR 1.02 [0.87, 1.2]). The increase in mortality in underweight individuals appeared to be limited to men with a significant increase in CHD (HR 3.3 [95% CI 1.6, 6.8]), CVD (HR 2.4 [95% CI 1.3, 4.5]), and all-cause mortality (HR 2.1 [95% CI 1.5, 3.1]) in underweight men compared men with normal BMI. Cubic splines demonstrating the relationship between BMI and CHD, CVD, and all-cause mortality according to sex are shown in Figures 2–4. There was evidence of effect modification by sex on the relationship between BMI and mortality with an age adjusted interaction variable (BMI x sex) significant in the spline models for CHD mortality (p-value 0.004), CVD mortality (p-value 0.013), and all-cause mortality (p-value 0.004).
Table 4.
The adjusted hazard ratios for CHD, CVD, and all-cause mortality according to categories of BMI stratified by sex in 36,509 participants from the CAC consortium.
| Categories of BMI* | Model 1† | Model 2‡ |
|---|---|---|
| Women (n=12,555) | ||
| CHD Mortality | ||
| BMI <21 | 0.27 (0.06, 1.2) | 0.26 (0.06, 1.1) |
| BMI 21-24.9 | 1.0 | 1.0 |
| BMI 25-29.9 | 0.64 (0.32, 1.3) | 0.69 (0.34, 1.4) |
| BMI ≥30 | 1.4 (0.73, 2.7) | 1.6 (0.85, 3.2) |
| CVD Mortality | ||
| BMI <21 | 0.75 (0.42, 1.3) | 0.73 (0.41, 1.3) |
| BMI 21-24.9 | 1.0 | 1.0 |
| BMI 25-29.9 | 0.77 (0.51, 1.2) | 0.80 (0.53, 1.2) |
| BMI ≥30 | 1.3 (0.85, 1.9) | 1.4 (0.90, 2.1) |
| All-Cause Mortality | ||
| BMI <21 | 1.1 (0.84, 1.4) | 1.1 (0.82, 1.4) |
| BMI 21-24.9 | 1.0 | 1.0 |
| BMI 25-29.9 | 0.77 (0.62, 0.96) | 0.79 (0.63, 0.98) |
| BMI ≥30 | 1.2 (0.94, 1.5) | 1.2 (0.97, 1.5) |
| Men (n=23,954) | ||
| CHD Mortality | ||
| BMI <21 | 2.2 (1.2, 4.2) | 2.4 (1.3, 4.5) |
| BMI 21-24.9 | 1.0 | 1.0 |
| BMI 25-29.9 | 1.0 (0.78, 1.4) | 1.1 (0.80, 1.4) |
| BMI ≥30 | 1.5 (1.1, 2.0) | 1.5 (1.1, 2.1) |
| CVD Mortality | ||
| BMI <21 | 3.3 (1.6, 6.8) | 3.3 (1.6, 6.8) |
| BMI 21-24.9 | 1.0 | 1.0 |
| BMI 25-29.9 | 1.0 (0.70, 1.5) | 1.1 (0.74, 1.6) |
| BMI ≥30 | 1.5 (1.01 , 2.3) | 1.7 (1.1, 2.6) |
| All-Cause Mortality | ||
| BMI <21 | 1.9 (1.3, 2.8) | 2.1 (1.5, 3.1) |
| BMI 21-24.9 | 1.0 | 1.0 |
| BMI 25-29.9 | 1.01 (0.86, 1.2) | 1.02 (0.87, 1.2) |
| BMI ≥30 | 1.3 (1.1, 1.6) | 1.3 (1.1, 1.6) |
BMI = Body Mass Index, CHD = Coronary Heart Disease, CVD = Cardiovascular Disease
Model 1 is adjusted for age and sex
Model 2 is adjusted for age, sex, scan site, smoking status, and family history.
Figure 2. Age adjusted cubic splines of the relationship of BMI and CHD death according to sex in 36,509 participants from the CAC consortium.
A. Women
B. Men*
BMI = Body Mass Index, CHD = Coronary Heart Disease
The BMI x Sex interaction variable was statistically significant (p-value 0.004)
* Men with BMI <20 mg/kg2 excluded due to small numbers and very wide confidence intervals
Figure 4. Age adjusted cubic splines of the relationship of BMI and All Cause Mortality According to Sex in 36,509 participants from the CAC consortium.
A. Women
B. Men
BMI = Body Mass Index
The BMI x Sex interaction variable was statistically significant (p-value 0.004)
Discussion
In a large multicenter cohort of individuals undergoing CAC scoring for clinical indications, overweight and obese individuals were more likely to have prevalent CAC compared to those with normal BMI, including after adjustment for traditional risk factors. Obese individuals were found to be at higher risk for CHD, CVD, and all-cause mortality as were underweight individuals. However, overweight individuals were not found to be at higher risk for CHD, CVD, and all-cause mortality despite an increased risk for prevalent CAC with overweight women having a lower risk of all-cause mortality compared to women with a normal BMI. Our findings do not provide evidence for an obesity paradox in a population of healthy individuals undergoing CAC screening though they do suggest that the relationship between BMI and mortality may vary according to sex.
Comparison to prior research
Prior data from MESA demonstrated an increase in prevalent CAC in overweight (relative risk 1.05 [95% CI 1.00-1.10]) and obese individuals (relative risk 1.17 [95% CI 1.11-1.23]) compared to individuals with BMI <25.0 kg/m2 after adjustment for age, sex, and race (17). Our data showed a higher risk of CAC in overweight and obese individuals which may be related to the higher risk of the sample as they had clinical indications for CAC scoring as opposed to MESA being a community representative sample. Additionally, MESA excluded individuals with morbid obesity (BMI ≥ 40kg/m2) while the CAC consortium only excluded individuals who exceeded the weight limit of the CT scanner (usually 400 pounds).
In a meta-analysis of over 1.3 million patients with CHD (9), both overweight and obesity were associated with lower risk of all-cause mortality. However, in obese individuals the lower risk of mortality was no longer present after 5-years of follow-up, with a higher risk of death in individuals with more severe obesity (BMI >35kg/m2). This suggests the paradoxical relationship between obesity and mortality could possibly be due to earlier diagnosis of CHD in obese individuals. The obesity paradox was shown to vary according to sex in a prior sample of 3,811 individuals with systolic heart failure (10). The study found that overweight and obese men had higher adjusted mortality compared to normal weight men while overweight women had lower adjusted mortality compared to normal weight women (p-value for interaction <0.0001).
As opposed to patients with clinical CHD or heart failure, our sample included asymptomatic individuals without established CVD. Therefore, comparisons to primary prevention populations may be more relevant. In a systematic review and meta-analysis including over 2.8 million individuals from prospective studies of general populations, Flegal et al (6) found that overweight individuals were at lower risk for all-cause mortality (HR 0.95 [95% CI 0.91-0.96]) while obese individuals had a higher risk for all-cause mortality (HR1.18 [95% CI 1.12-1.15]), which was largely driven by an increase in those with more severe obesity (BMI > 35kg/m2). We found a similar increase in risk for all-cause mortality with obesity and additionally demonstrated an increased risk for CHD and CVD death in obese individuals. Additionally, we found no statistically significant evidence of effect modification by CAC the relationship between BMI and mortality, and only partial mediation of effect when CAC was placed in the BMI-mortality model, suggesting the increase risk of death in obese individuals may be additionally mediated by non-atherosclerotic mechanisms. In contrast to our study, Flegal et al did not find evidence that the relationship between BMI and mortality varied according to sex.
The rationale for the paradoxical relationship between BMI and mortality
There are multiple potential mechanisms for the paradoxical relationship between BMI and mortality (18). Overweight and obese individuals may have greater metabolic reserve and more favorable neuroendocrine profiles (19,20). Additionally, adipose tissue has been shown to capable of neutralizing substances with potential adverse metabolic effects (21,22). Systematically, obese patients may seek medical treatment earlier and have a greater likelihood to be placed on optimal medical management for longer durations (23–26).
Methodological issues to consider when evaluating the relationship between BMI and mortality
The paradoxical relationship between BMI and mortality may be heavily influenced by methodological issues and bias. Some individuals with lower BMI at the time of a CVD diagnosis may represent individuals with more advanced disease and unintentional weight loss prior to diagnosis. Tobacco use is also known to be a strong confounder of the relationship between BMI and mortality and inclusion of smoking in a regression model may not adequately control for its impact (27). Additionally, some analyses have introduced bias by comparing overweight and obese individuals to those with BMI < 25kg/m2, which then groups underweight individuals with normal weight individuals and, as seen in our analysis, underweight individuals represent a unique population at elevated risk for mortality. Cardiorespiratory fitness is also known to be a significant confounder of the relationship between BMI and mortality (28,29). After accounting for potential weight loss due to underlying illness and robust control for tobacco use, large studies in both men and women have found no evidence of an obesity paradox in the general population, alternatively showing a positive linear relationship between BMI and mortality in women and men (30, 31).
Strengths and Limitations
Our analysis benefited from data from multiple experienced centers geographically dispersed throughout the US with a large sample size of both men and women with long term follow-up and data on CAC and cause-specific mortality. Our sample also had a low rate of tobacco use and exclusion of smokers did not meaningfully change our results.
Potential limitations include that BMI can potentially be an inaccurate measure of body fat in some individuals (32) though the obesity paradox has been demonstrated using percent body fat and waist circumference as well (33,34). Knowledge of CAC results may have influenced downstream treatment recommendations, including more preventive therapies in overweight and obese individuals who were more likely to have CAC. There were individuals in our analysis with missing data (12.5%) but the majority were missing only 1 risk factor and there was excellent correlation between mean and median ASCVD risk scores calculated directly compared to scores calculated using imputed data.
Additionally, menopause is known to alter CVD risk in women but menopausal status was not available in our database and therefore could not be adjusted for in the multivariable models. The majority of our sample was white and we were unable to examine associations specifically among non-white race/ethnicities. Other limitations include self-reporting of risk factors and the inadequacies of vital status ascertainment in the United States, with a potential for up to 10% underestimation of mortality (35), though this would be expected to be non-differential across BMI and CAC groups.
Conclusions
In a large sample of men and women without established CVD but clinical indications for CAC scoring we found an increased risk for CAC in overweight and obese individuals. Subsequently, we found an increased risk for CHD, CVD, and all-cause mortality in obese men and women. Our findings support healthy lifestyle behaviors aimed at maintaining a healthy body weight and avoiding obesity in men and women.
Figure 3. Age adjusted cubic splines of the relationship of BMI and CVD Death according to sex in 36,509 participants from the CAC consortium.
A. Women
B. Men*
BMI = Body Mass Index, CVD = Cardiovascular Disease
The BMI x Sex interaction variable was statistically significant (p-value 0.013)
* Men with BMI <20 mg/kg2 excluded due to small numbers and very wide confidence intervals
Clinical Perspective.
The relationship between BMI and mortality is complex. An analysis of a large database of over 36,000 individuals undergoing coronary artery calcium (CAC) scoring found that, compared to those with normal BMI, overweight and obese individuals had a higher risk of baseline CAC, with a similar linear relationship in both men and women. For obese individuals, there was a subsequent increase in risk for death from coronary heart disease, cardiovascular disease, and all-cause mortality. However, overweight individuals, despite a higher likelihood of CAC, did not have a higher risk of morality with actually a lower risk of all-cause mortality in overweight women compared to women with normal BMI. The relationship between BMI and mortality did vary according to sex, with a statistically significant interaction between BMI and sex. There did not appear to be an interaction between BMI and baseline CAC. Overall, these findings are consistent with prior data showing an increased risk of CAC in those with higher BMI but with a novel finding that the relationship between BMI and mortality in a generally healthy, primary prevention population may vary according to sex. Further analyses to explore mechanisms of the relationship between BMI and mortality as well as the potential interaction according to sex are warranted.
Acknowledgments
Sources of funding
Dr. Blaha has received support from NIH award L30 HL110027 for previous CAC project
Disclosures
Dr. Blaha has received grants from the NIH, FDA, AHA, Aetna Foundation, and Amgen Foundation.
Dr. Budoff has received a grant from General Electric.
References
- 1.Virani SS, Alonso A, Benjamin EJ, Bittencourt MS, Callaway CW, Carson AP, Chamberlain AM, Chang AR, Cheng S, Delling FN, et al. Heart Disease and Stroke Statistics-2020 Update: A Report From the American Heart Association. Circulation. 2020;141:e139–e596. [DOI] [PubMed] [Google Scholar]
- 2.Nguyen NT, Magno CP, Lane KT, Hinojosa MW, Lane JS. Association of hypertension, diabetes, dyslipidemia, and metabolic syndrome with obesity: findings from the National Health and Nutrition Examination Survey, 1999 to 2004. J Am Coll Surg. 2008;207:928–34. [DOI] [PubMed] [Google Scholar]
- 3.Chang Y, Kim BK, Yun KE, Cho J, Zhang Y, Rampal S, Zhao D, Jung HS, Choi Y, Ahn J et al. Metabolically-healthy obesity and coronary artery calcification. J Am Coll Cardiol. 2014;63:2679–2686. [DOI] [PubMed] [Google Scholar]
- 4.Detrano R, Guerci AD, Carr JJ, Bild DE, Burke G, Folsom AR, Liu K, Shea S, Szklo M, Bluemke DA et al. Coronary Calcium as a Predictor of Coronary Events in Four Racial or Ethnic Groups. N Engl J Med. 2008;358:1336–1345. [DOI] [PubMed] [Google Scholar]
- 5.Shaw LJ, Giambrone AE, Blaha MJ, Knapper JT, Berman DS, Bellam N, Quyyumi A, Budoff MJ, Callister TQ, Min JK. Long-Term Prognosis After Coronary Artery Calcification Testing in Asymptomatic Patients: A Cohort Study. Ann Intern Med. 2015;163:14–21. [DOI] [PubMed] [Google Scholar]
- 6.Flegal KM, Kit BK, Orpana H, Graubard BI. Association of All-Cause Mortality With Overweight and Obesity Using Standard Body Mass Index Categories: A Systematic Review and Meta-analysis. JAMA. 2013;309:71–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Uretsky S, Messerli FH, Bangalore S, Champion A, Cooper-DeHoff RM, Zhou Q, Pepine CJ. Obesity paradox in patients with hypertension and coronary artery disease. Am J Med. 2007;120:863–870. [DOI] [PubMed] [Google Scholar]
- 8.Romero-Corral A, Montori VM, Somers VK, Korinek J, Thomas RJ, Allison TG, Mookadam F, Lopez-Jimenez F. Association of body weight with total mortality and with cardiovascular events in coronary artery disease: a systematic review of cohort studies. Lancet. 2006;368:666–678. [DOI] [PubMed] [Google Scholar]
- 9.Wang ZJ, Zhou YJ, Galper BZ, Gao F, Yeh RW, Mauri L. Association of body mass index with mortality and cardiovascular events for patients with coronary artery disease: a systematic review and meta-analysis. Heart. 2015;101:1631–1638. [DOI] [PubMed] [Google Scholar]
- 10.Vest AR, Wu Y, Hachamovitch R, Young JB, Cho L. The heart failure overweight/obesity survival paradox: the missing sex link. JACC Heart Fail. 2015;3:917–926. [DOI] [PubMed] [Google Scholar]
- 11.Aune D, Sen A, Prasad M, Norat T, Janszky I, Tonstad S, Romundstad P, Vatten LJ. BMI and all cause mortality: systematic review and non-linear dose-response meta-analysis of 230 cohort studies with 3.74 million deaths among 30.3 million participants. BMJ. 2016;353:i2156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Prospective Studies Collaboration. Body-mass index and cause-specific mortality in 900 000 adults: collaborative analyses of 57 prospective studies. Lancet. 2009;373:1083–1096. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Blaha MJ, Whelton SP, Al Rifai M, Dardari ZA, Shaw LJ, Al-Mallah MH, Matsushita K, Rumberger JA, Berman DS, Budoff MJ et al. Rationale and design of the coronary artery calcium consortium: A multicenter cohort study. J Cardiovasc Comput Tomogr. 2017;11:54–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Jensen MD, Ryan DH, Apovian CM, Ard JD, Comuzzie AG, Donato KA, Hu FB, Hubbard VS, Jakicic JM, Kushner RF, et al. 2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and The Obesity Society. Circulation. 2014;129:S102–138. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Mao SS, Pal RS, McKay CR, Gao YG, Gopal A, Ahmadi N, Child J, Carson S, Takasu J, Sarlak B et al. Comparison of coronary artery calcium scores between electron beam computed tomography and 64-multidetector computed tomographic scanner. J Comput Assist Tomogr. 2009;33:175–178. [DOI] [PubMed] [Google Scholar]
- 16.Al-Mallah MH, Keteyian SJ, Brawner CA, Whelton S, Blaha MJ. Rationale and design of the henry ford exercise testing project (the FIT project). Clin Cardiol. 2014;37:456–461. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Burke GL, Bertoni AG, Shea S, Tracy R, Watson KE, Blumenthal RS, Chung H, Carnethon MR. The impact of obesity on cardiovascular disease risk factors and subclinical vascular disease: the Multi-Ethnic Study of Atherosclerosis. Arch Intern Med. 2008;168:928–935. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lavie CJ, De Schutter A, Parto P, Jahangir E, Kokkinos P, Ortega FB, Arena R, Milani RV. Obesity and prevalence of cardiovascular diseases and prognosis—the obesity paradox updated. Prog cardiovasc dis. 2016:58;537–547. [DOI] [PubMed] [Google Scholar]
- 19.Oreopoulos A, Padwal R, Kalantar-Zadeh K, Fonarow GC, Norris CM, McAlister FA. Body mass index and mortality in heart failure: a meta-analysis. Am Heart J. 2008;156:13–22. [DOI] [PubMed] [Google Scholar]
- 20.Kalantar-Zadeh K, Block G, Horwich T, Fonarow GC. Reverse epidemiology of conventional cardiovascular risk factors in patients with chronic heart failure. J Am Coll Cardiol 2004;43:1439–44. [DOI] [PubMed] [Google Scholar]
- 21.Rauchhaus M, Coats AJS, Anker SD. The endotoxin-lipoprotein hypothesis. Lancet 2000;356:930–933. [DOI] [PubMed] [Google Scholar]
- 22.Mohamed-Ali V, Goodrick S, Bulmer K, Holly JM, Yudkin JS, Coppack SW. Production of soluble tumor necrosis factor receptors by human subcutaneous adipose tissue in vivo. Am J Physiol. 1999;277:E971–5. [DOI] [PubMed] [Google Scholar]
- 23.Oreopoulos A, McAlister FA, Kalantar-Zadeh K, Padwal R, Ezekowitz JA, Sharma AM, Kovesdy CP, Fonarow GC, Norris CM. The relationship between body mass index, treatment, and mortality in patients with established coronary artery disease: a report from APPROACH. Eur Heart J. 2009;30:2584–2592. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Chang VW, Asch DA, Werner RM. Quality of care among obese patients. JAMA. 2010;303:1274–1281. [DOI] [PubMed] [Google Scholar]
- 25.Schenkeveld L, Magro M, Oemrawsingh RM, Lenzen M, de Jaegere P, van Geuns RJ, Serruys PW, van Domburg RT. The influence of optimal medical treatment on the “obesity paradox,” body mass index and long term mortality in patients treated with percutaneous coronary intervention: a prospective cohort study. BMJ Open. 2012;2:e000535. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Steinberg BA, Cannon CP, Hernandez AF, Pan W, Peterson ED, Fonarow GC. Medical therapies and invasive treatments for coronary artery disease by body mass: the “obesity paradox” in the Get With The Guidelines database. Am J Cardiol. 2007;100:1331–1335. [DOI] [PubMed] [Google Scholar]
- 27.Manson JE, Stampfer MJ, Hennekens CH, Willett WC. Body Weight and longevity. A reassessment. JAMA. 1987;257:353–358. [PubMed] [Google Scholar]
- 28.Barry VW, Baruth M, Beets MW, Durstine JL, Liu J, Blair SN. Fitness vs. fatness on all-cause mortality: a meta-analysis. Prog Cardiovasc Dis. 2014;56:382–390. [DOI] [PubMed] [Google Scholar]
- 29.McAuley PA, Keteyian SJ, Brawner CA, Dardari ZA, Al Rifai M, Ehrman JK, Al-Mallah MH, Whelton SP, Blaha MJ. Exercise Capacity and the Obesity Paradox in Heart Failure: The FIT (Henry Ford Exercise Testing) Project. Mayo Clin Proc. 2018;93:701–708. [DOI] [PubMed] [Google Scholar]
- 30.Gelber RP, Kurth T, Manson JE, Buring JE, Gaziano JM. Body mass index and mortality in men: evaluating the shape of the association. Int J Obes (Lond). 2007;31:1240–1247. [DOI] [PubMed] [Google Scholar]
- 31.Manson JE, Willett WC, Stampfer MJ, Colditz GA, Hunter DJ, Hankinson SE, Hennekens CH, Speizer FE. Body weight and mortality among women. N Engl J Med. 1995;333:677–685. [DOI] [PubMed] [Google Scholar]
- 32.De Schutter A, Lavie CJ, Arce K, Menendez SG, Milani RV. Correlation and discrepancies between obesity by body mass index and body fat in patients with coronary heart disease. J Cardiopulm Rehabil Prev 2013;33:77–83. [DOI] [PubMed] [Google Scholar]
- 33.Lavie CJ, De Schutter A, Patel DA, Romero-Corral A, Artham SM, Milani RV. Body composition and survival in stable coronary heart disease: impact of lean mass index and body fat in the “obesity paradox”. J Am Coll Cardiol. 2012;60: 1374–1380. [DOI] [PubMed] [Google Scholar]
- 34.McAuley PA, Artero EG, Sui X, Lee DC, Church TS, Lavie CJ, Myers JN, España-Romero V, Blair SN. The obesity paradox, cardiorespiratory fitness, and coronary heart disease. Mayo Clin Proc. 2012;87:443–451. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.da Graca B, Giovanni F, Nicewander D. Consequences for healthcare quality and research of the exclusion of records from the Death Master File. Circ Cardiovasc Qual Outcomes. 2013;6:124–128. [DOI] [PubMed] [Google Scholar]