Association of Obesity or Weight Change With Coronary Heart Disease Among Young Adults in South Korea

Key Points

Question  Does an association exist between body mass index or a change in body mass index and coronary heart disease among young adults in South Korea?

Findings  In this population-based longitudinal study of 2.6 million men and women aged 20 to 39 years, high body mass index and body mass index gain were significantly associated with elevated risk of coronary heart disease, whereas body mass index loss was associated with reduced risk of coronary heart disease.

Meaning  Obesity and weight gain were associated with increased risk of coronary heart disease among young adults in South Korea; prospective studies are needed to validate these findings.

Importance  Previous studies have shown a U- or J-shaped association of body mass index (BMI) or change in BMI with coronary heart disease (CHD) among middle-aged and elderly adults. However, whether a similar association exists among young adults is unclear.

Objective  To determine whether an association exists between BMI or BMI change with CHD among young adults.

Design, Setting, and Participants  This population-based longitudinal study used data obtained by the Korean National Health Insurance Service from 2002 to 2015. The study population comprised 2 611 450 men and women aged between 20 and 39 years who underwent 2 health examinations, the first between 2002 and 2003 and the second between 2004 and 2005.

Exposures  World Health Organization Western Pacific Region guideline BMI categories of underweight, normal weight, overweight, obese grade 1, and obese grade 2 derived during the first health examination and change in BMI calculated during the second health examination.

Main Outcomes and Measures  Body mass index (calculated as weight in kilograms divided by height in meters squared). Absolute risks (ARs), adjusted hazard ratios (aHRs), and 95% CIs for acute myocardial infarction or CHD during follow-up from 2006 to 2015.

Results  Data from 1 802 408 men with a mean (SD) age of 35.1 (4.8) years and 809 042 women with a mean (SD) age of 32.5 (6.3) years were included. The mean (SD) BMI was 23.2 (3.2) for the total population, 24.0 (3.0) for men, and 21.4 (2.9) for women. Compared with normal weight men, overweight (AR, 1.38%; aHR, 1.18 [95% CI, 1.14-1.22]), obese grade 1 (AR, 1.86%; aHR, 1.45 [95% CI, 1.41-1.50]), and obese grade 2 (AR, 2.69%; aHR, 1.97 [95% CI, 1.86-2.08]) men had an increased risk of CHD (P < .001 for trend). Similarly, compared with normal weight women, overweight (AR, 0.77%; aHR, 1.34 [95% CI, 1.24-1.46]), obese grade 1 (AR, 0.95%; aHR, 1.52 [95% CI, 1.39-1.66]), and obese grade 2 (AR, 1.01%; aHR, 1.64 [95% CI, 1.34-2.01]) women had an increased risk of CHD (P < .001 for trend). Compared with participants who maintained their weight at normal levels, those who became obese had elevated CHD risk among men (0.35% increase in AR; aHR, 1.35 [95% CI, 1.17-1.55]) and women (0.13% increase in AR; aHR, 1.31 [95% CI, 0.95-1.82]). Weight loss to normal levels among obese participants was associated with reduced CHD risk for men (0.58% decrease in AR; aHR, 0.77 [95% CI, 0.64-0.94]) and women (0.57% decrease in AR; aHR, 0.66 [95% CI, 0.45-0.98]).

Conclusions and Relevance  Obesity and weight gain were associated with elevated risk of CHD among young adults in this study. Studies that prospectively determine the association between weight change and CHD risk are needed to validate these findings.

Coronary heart disease (CHD) is the leading cause of death globally, with 8.1 million deaths in 2013 being due to CHD.¹ Furthermore, CHD-associated mortality is expected to increase by 100% among men and 80% among women from 1990 to 2020,² highlighting the importance of identifying and controlling risk factors for CHD. Despite previous implications of obesity being a risk factor for CHD, the association between obesity and CHD has been controversial. Recent studies have shown that being overweight or obese is associated with lower risk of cardiovascular disease (CVD) or CVD-associated death, whereas being underweight is associated with increased risk of CVD, a phenomenon called the obesity paradox.^3,4 Moreover, numerous previous studies investigating the association between weight change and CVD and CVD-related mortality have shown a U- or J-shaped association, in which weight gain or loss leads to increased risk of CVD-related death.^5,6

One plausible explanation for the obesity paradox may be that being slightly overweight or obese is associated with greater lean mass.⁷ Older adults who lose weight lose mostly lean mass,⁸ which could contribute to the increased risk of CVD events after weight loss. Because the body composition of fat and muscle mass is altered with age, the contributions of body mass index (BMI; calculated as weight in kilograms divided by height in meters squared) and change in BMI on CHD among middle-aged and elderly adults may differ from those among young adults. However, few studies have investigated the association between BMI or change in BMI with CHD among young adults.

In this large population-based longitudinal study, we used the Korean National Health Insurance Service (NHIS) database to assess whether an association exists between BMI or a change in BMI and the risk of developing acute myocardial infarction (AMI) or CHD among young adults.

The NHIS provides mandatory health insurance for all South Korean citizens, covering nearly all forms of health care, including health screening examinations for all employed and self-employed insured individuals aged 20 years or more as well as all dependents aged 40 years or older.⁹ The NHIS database includes data on sociodemographic characteristics, hospital admissions, outpatient department visits, pharmaceutical visits, and health screening examinations. Health screening examinations include information on health behavior obtained from a questionnaire, physical examinations, and blood tests. In total, 74.8% of those who were eligible participated in the health screening examinations in 2014.⁹ Several studies have used the NHIS database for epidemiologic studies, and its validity has been described in detail elsewhere.^9-11 The Seoul National University Hospital (Seoul, South Korea) institutional review board approved this study and waived the requirement for informed patient consent because data in the NHIS database are anonymized in adherence with strict confidentiality guidelines.

Data were obtained from the NHIS database between 2002 and 2015 for all individuals aged between 20 and 39 years in 2002. Among the 2 692 643 participants who underwent health examinations during the first (2002 and 2003) and second (2004 and 2005) health screening periods, 1721 individuals with missing BMI values were excluded. In addition, 90 individuals who died as well as 6168 individuals who had received a diagnosis of CHD before the index date of January 1, 2006, were excluded. The 73 214 individuals with missing covariate values were also excluded. The final study population consisted of 2 611 450 men and women.

The BMI was determined during each health examination period. When evaluating the association of BMI with AMI or CHD, BMI values from the second (2004 and 2005) health screening period were used. Participants were categorized by BMI using the World Health Organization Western Pacific Region guideline strata of underweight (<18.5), normal weight (18.5-22.9), overweight (23.0-24.9), obese grade 1 (25.0-29.9), or obese grade 2 (≥30.0).¹² The change in BMI was calculated by subtracting BMI values obtained during the second health examination from those of the first health examination, and the resulting change was categorized as normal (<23.0), overweight (23.0-24.9), or obese (≥25.0).

Hospital admission records and death certificates were used to identify AMI and CHD events. All diagnoses on hospital discharge or death are recorded in the NHIS database using International Classification of Diseases, 10th Revision (ICD-10) codes from the World Health Organization. Causes of death were determined by physicians at the time of death. Because hospital admission records must be submitted to the NHIS for hospitals to receive payment, the follow-up for CHD events is likely to be complete. Consistent with the American Heart Association, the ICD-10 codes for AMI of I21 and for CHD of I20 to I25 were used.¹³ We defined an event of CHD as 2 or more days of hospital admission or death with causes listed as ICD-10 codes for AMI or CHD. Multiple previous studies have used hospital admission records and death certificates to identify cardiovascular events using the NHIS data.^14,15

All participants were followed up starting January 1, 2006, and ending at a CHD event, date of death, or December 31, 2015, whichever came first. The following were considered covariates: age (continuous variable; years), household income (categorical variable; first, second, third, or fourth quartiles), physical activity (categorical variable; none, 1-2, 3-4, 5-6, or 7 times per week), alcohol consumption (categorical variable; none, 0-1, 1-2, 3-4, or 5 or more times per week), tobacco consumption (categorical variable; never, past, quit, light, moderate, or heavy), systolic blood pressure (continuous variable), fasting serum glucose level (continuous variable), total cholesterol (continuous variable), or Charlson comorbidity index (CCI; continuous variable).

Individuals who quit were defined as those who reported having smoked tobacco during the first health examination but having quit during the second health examination, whereas past tobacco smokers were those who reported having quit in both the first and second health examinations. Light, moderate, and heavy smokers were those who reported having smoked 1 to 9, 10 to 19, or 20 or more cigarettes per day, respectively. Household income was derived from each patient’s insurance premium, and the CCI was calculated using the ICD-10 codes for major comorbidities between 2002 and 2005. The algorithm for the calculation of CCI using ICD-10 codes was adapted from another study.¹⁶

Absolute risk (AR) was calculated by determining the percentage of events per number of people based on BMI or BMI change. Cox proportional hazards regression analysis was conducted to obtain the adjusted hazard ratios (aHRs) and 95% CIs of AMI and CHD based on BMI and change in BMI. The risk of AMI and CHD were calculated after adjustments for all covariates. Restricted cubic splines¹⁷ of BMI or BMI change were used to graphically assess the association between BMI or a change in BMI with AMI or CHD. The lowest value of the Akaike information criteria was used to determine the number of knots, using a range of 3 to 7 knots. Stratified subgroup analyses were conducted by dividing the participants into subgroups for age, physical activity, smoking, and CCI. The aHRs of CHD for the covariates as well as per unit of BMI or change in BMI were determined. The aHRs of CHD were determined based on BMI after adjustments for metabolic mediators.

Statistical significance was defined as 2-sided P < .05. All analyses were conducted with SAS, version 9.4 (SAS Institute Inc) and Stata, version 14 (StataCorp).

In total, 30 372 CHD events were detected during 25 868 244 person-years of follow-up. The mean (SD) BMI was 23.2 (3.2) for the total population, 24.0 (3.0) for men, and 21.4 (2.9) for women (Table 1). The aHRs of the covariates for CHD are provided in eTable 1 in the Supplement. Older age, lower household income, and tobacco or alcohol consumption as well as higher blood pressure, fasting serum glucose level, total cholesterol level, and CCI were associated with increased risk of CHD.

Higher BMI was associated with increased risk of AMI and CHD among both men and women (Figure 1). A BMI gain was associated with greater risk of AMI and CHD among men and women, whereas a BMI loss was associated with decreased risk of AMI and CHD for both sexes (Figure 2). Compared with normal weight individuals, men who were in the overweight (AR, 1.38%; aHR, 1.18 [95% CI, 1.14-1.22]), obese grade 1 (AR, 1.86%; aHR, 1.45 [95% CI, 1.41-1.50]), or obese grade 2 (AR, 2.69%; aHR, 1.97 [95% CI, 1.86-2.08]) strata and women who were in the overweight (AR, 0.77%; aHR, 1.34 [95% CI, 1.24-1.46]), obese grade 1 (AR, 0.95%; aHR, 1.52 [95% CI, 1.39-1.66]), or obese grade 2 (AR, 1.01%; aHR, 1.64 [95% CI, 1.34-2.01]) strata had increased risk of CHD (P < .001 for trend for each sex) (Table 2). The risks of AMI and CHD were significantly elevated for every unit increase in BMI and change in BMI for both sexes (eTable 2 in the Supplement). The increased risks of CHD for overweight and obese participants were preserved when the study population was grouped by age, physical activity, smoking status, or CCI (eTable 3 in the Supplement). The association of high BMI with CHD appeared to be stronger among women with a history of smoking compared with that for men who had ever smoked.

Table 3 and eTable 4 in the Supplement indicate the association of change in the BMI strata with AMI or CHD. Weight gain among normal weight individuals to obese levels was associated with greater risk of CHD among men (0.35% increase in AR; aHR, 1.35 [95% CI, 1.17-1.55]) and women (0.13% increase in AR; aHR, 1.31 [95% CI, 0.95-1.82]). Compared with obese individuals who maintained their weight, those who reduced their weight to normal levels had reduced risk of CHD for men (0.58% decrease in AR; aHR, 0.77 [95% CI, 0.64-0.94]) and women (0.57% decrease in AR; aHR, 0.66 [95% CI, 0.45-0.98]). Compared with those who maintained normal weight, those who were overweight at baseline but became normal weight had elevated risk of CHD among both men (aHR, 1.13 [95% CI, 1.05-1.21]) and women (aHR, 1.24 [95% CI, 1.07-1.42]).

The contribution of metabolic mediators to the risk of BMI on CHD are given in eTable 5 in the Supplement. Before adjustments for the metabolic mediators, the aHRs (95% CI) for individuals in the obese grade 2 stratum were 2.51 (2.38-2.65) for men and 1.95 (1.59-2.37) for women. After adjustments for all 3 metabolic mediators, the aHRs (95% CI) for individuals in the obese grade 2 stratum were 1.97 (1.86-2.08) for men and 1.64 (1.34-2.01) for women, corresponding to an excess risk of obesity to CHD of 35.8% (34.5%-37.7%) for men and 32.6% (26.3%-42.4%) for women.

In this large-scale, population-based, longitudinal study, we showed that high BMI was associated with increased risk of AMI and CHD among young adults. Furthermore, BMI gain was associated with increased risk, whereas BMI loss was associated with reduced risk of AMI and CHD. To our knowledge, this is the first study to show that losing weight and maintaining normal BMI is associated with decreased risk of CHD among young adults.

The results from previous prospective studies with similar age groups are in accordance with the increased risk of CHD with high BMI observed in our study. For example, in a study investigating the association of BMI with CVD-related mortality, Selmer and Tverdal¹⁸ found that the risk of dying of CHD increased for every 5-unit BMI increase among persons aged between 30 and 59 years (HR, 1.30; 95% CI, 1.21-1.39). Similarly, Seidell and colleagues¹⁹ showed that the risk of dying of CHD increased among overweight and obese individuals aged between 30 and 54 years. In a large cohort study investigating weight gain from early to middle adulthood among US women in the Nurses’ Health Study and US men in the Health Professionals Follow-Up Study cohorts, Zheng and colleagues²⁰ showed that weight gain from age 18 or 21 to 55 years was associated with elevated risk of CVD. Another previous study²¹ investigating weight change and CHD using the Honolulu Heart Program cohort revealed that gaining 2.5 kg or more was associated with increased risk of CHD.

Overweight and obesity increases the risk of CHD by elevating the risk of other cardiovascular risk factors, such as hypertension, dyslipidemia, and type 2 diabetes.^22,23 Obesity is associated with higher sympathetic nervous system activity, leptin concentrations, and angiotensin-aldosterone activity, which may lead to greater salt retention and higher blood pressure.²⁴ Obesity may also contribute to the development of type 2 diabetes and dyslipidemia by elevating C-reactive protein levels, thereby promoting a systemic inflammatory state.²⁵ We showed in the present study that blood pressure, total cholesterol levels, and fasting serum glucose levels may contribute to approximately one-third of the excess risk of CHD associated with obesity among young adults. These results are consistent with those of the BMI Mediated Effects study,²⁶ which found that blood pressure, cholesterol levels, and blood glucose levels accounted for 46% (95% CI, 42%-50%) of the excess risk of CHD with elevated BMI.

Unlike our results, previous studies have indicated a significantly increased risk of CHD among underweight individuals (HR, 1.70; 95% CI, 1.42-2.05) compared with those with normal BMI.^27,28 In addition, other studies^21,29 have failed to show that weight loss is associated with reduced risk of CHD, and a study investigating weight loss strategies and the risk of CHD using the Nurses’ Health Study cohort³⁰ found that compared with no intervention, for individuals with BMIs greater than 25 and without chronic diseases, a 5% loss in BMI was associated with increased risk of CHD.

This discrepancy regarding the association between being underweight or losing weight and CHD may be due to the younger age of our study population compared with that in previous studies. Because the prevalence of sarcopenia, the reduction of lean mass, increases with age,³¹ young adults tend to have greater lean mass and lower fat mass than their middle-aged and elderly counterparts with similar BMI. Several previous studies using computed tomography,³² dual-energy x-ray absorptiometry,³³ or other measures of body composition, such as calf muscle area,³¹ have shown that young adults tend to have a higher muscle to fat mass ratio than older adults. Greater muscle mass is associated with better exercise capacity and cardiorespiratory fitness, which may in turn lead to decreased risk of CHD.^34,35 Because increased levels of leptin, a hormone released by adipocytes, have been associated with increased risk of CVD,³⁶ the greater proportion of fat mass among elderly adults who lose weight may elevate the risk of CHD.²²

Greater frequencies of comorbidities and chronic diseases observed among middle-aged and elderly adults, in whom weight loss may act as a surrogate marker for worsening health, compared with younger adults might also have contributed to the increased risk of CHD among underweight individuals or to the lack of CHD risk-reducing benefit with weight loss in previous studies. Despite our study population consisting of young adults, a population generally associated with low rates of chronic diseases, we attempted to account for the possibility of serious illnesses contributing to weight loss. We conducted a sensitivity analysis by excluding patients who had received a diagnosis of CHD within the first 4 years of follow-up as well as by analyzing subgroups of individuals who had never smoked tobacco and had no chronic conditions (CCI of 0). Although we do not present a figure of this analysis, our results continued to indicate a reduced risk of CHD with weight loss.

There are several limitations to consider when interpreting the results of our study. First, owing to the observational nature of our study and because the reasons for the weight changes were unknown, we could not establish a cause-effect relationship between obesity or weight change and CHD. Thus, future prospective studies comparing intentional and unintentional weight loss are needed. Although we attempted to account for this by adjusting for and conducting subgroup analyses for physical activity according to the number of times per week (as determined in a self-reported questionnaire), this measure may be insufficiently sensitive to accurately determine the contribution of physical activity to the association between obesity and CHD. Therefore, studies investigating the association of BMI with CHD using a more detailed measure of physical activity, for example, with metabolic equivalent values, are needed.

Second, because waist circumference was not available, obesity was defined solely by BMI, which may not adequately reflect certain aspects of body composition, such as fat distribution, particularly because young adults have a higher proportion of muscle mass for the same BMI than do middle-aged and elderly adults.^31-33 Thus, future studies are needed that use other measures of adiposity, such as waist circumference, waist to hip ratio, and waist to height ratio. Third, the present study population comprised young adults who had health insurance through their employers or were self-insured, a group that may have certain sociodemographic tendencies, such as high income and healthy lifestyle behaviors. Although we attempted to take this into account by adjusting for household income, physical activity, and tobacco and alcohol consumption, future studies with a broader study population are needed to validate the findings of our study. Finally, the results on whether becoming normal weight among overweight or obese individuals eliminates the risk associated with having been overweight or obese are mixed. Although the reasons for this are unclear, the duration of obesity may play an important role, as longer duration obesity may lead to longer exposure to the CHD risk-increasing contributions of obesity. Therefore, future studies investigating the association between duration of obesity and CHD are also needed.

Despite these limitations, our study has a number of strengths. To our knowledge, this is the first study to investigate the association of BMI and change in BMI with CHD among young adults. Unlike the results from previous studies with study populations of middle-aged and elderly adults, our results show that weight reduction was associated with reduced risk of CHD among young adults. Furthermore, our large study population enhances the generalizability of our results. We were also able to adjust for key metabolic mediators, including blood glucose levels, blood pressure, and total cholesterol levels. Finally, we attempted to take into account the possibility of reverse causality by extensive subgroup and sensitivity analyses, which enhanced the reliability of our findings.

Obesity and weight gain were associated with greater CHD risk among young adults. The loss of weight for obese individuals was associated with reduced risk of CHD. Future prospective studies that investigate the association of intentional and unintentional weight change with CHD are needed to clarify the interpretations of our findings.

Accepted for Publication: April 11, 2018.

Corresponding Author: Sang Min Park, MD, PhD, MPH, Department of Biomedical Sciences, College of Medicine, Seoul National University, 101 Daehak-ro, Jongno-gu, Seoul 03080, South Korea (smpark.snuh@gmail.com).

Published Online: June 18, 2018. doi:10.1001/jamainternmed.2018.2310

Author Contributions: Dr S. M. Park had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Choi, Son, Yun, S. M. Park.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Choi, S. M. Park.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Choi, K. Kim.

Administrative, technical, or material support: S. M. Kim, S. Y. Park, Y.-Y. Kim.

Supervision: S. M. Park.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by a grant (20170322652-00) from the Ministry of Health and Welfare of South Korea.

Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.

Additional Information: This study used NHIS data (NHIS-2017-1-143) from the Korean NHIS.