Body mass index and cardiovascular outcomes in patients with acute coronary syndrome by diabetic status: Korean cohort study

Background: The “obesity paradox” has not been elucidated in the long-term outcomes of acute coronary syndrome (ACS). Therefore, we investigated the association between obesity and cardiovascular (CV) outcomes in ACS patients with and without diabetes. Methods: We identied 6,978 patients with ACS aged 40–79 years from the Korean National Health Insurance Service-Health Screening Cohort between 2003 and 2015. Baseline body mass index (BMI) was categorized as underweight (<18.5 kg/m 2 ), normal weight (18.5–22.9 kg/m 2 ), overweight (23.0– 24.9 kg/m 2 ), obese class I (25.0–29.9 kg/m 2 ), and obese class II ( ≥ 30.0 kg/m 2 ). The primary outcome was major adverse CV events (MACE)—CV death, myocardial infarction (MI), and stroke. The secondary outcomes were the individual components of MACE, hospitalization for heart failure (HHF), and all-cause death. Results: The study included 3,989 patients with and 2,989 without diabetes. Compared to normal-weight patients without diabetes (reference group), those with diabetes had a higher risk of MACE (hazard ratio [HR], 1.29; 95% condence interval [CI], 1.07–1.56). Obese patients without diabetes had a lower risk of MACE (HR, 0.78; 95% CI, 0.62–0.97) than those with diabetes (HR, 0.95; 95% CI 0.78–1.14). In patients without diabetes, obese BMI decreased the risk of HHF (HR, 0.62; 95% CI, 0.42–0.92) and stroke (HR, 0.61; 95% CI, 0.42–0.88), but not in those with diabetes. Conclusion: Among patients with ACS, obesity had rather protective effect on CV outcomes compared to normal weight, while this was not evident in patients with diabetes.


Background
Obesity is a well-established risk factor for cardiovascular (CV) disease that increases the risk of CV mortality [1,2]. However, several studies have reported that obese patients with high body mass index (BMI) have a better prognosis than patients with normal BMI [3]. However, the mechanism of this phenomenon, called "obesity paradox" is unclear.
Diabetes mellitus has been known to increase the risk of CV disease, and patients with diabetes experience worse CV outcomes than those without diabetes [4]. In a prospective cohort study, diabetes increased the risk of mortality by 140% in patients with previous myocardial infarction (MI) [5]. According to the Korean registry data, diabetes is an independent risk factor for 1-year mortality in patients undergoing percutaneous coronary intervention with MI [6].
There has been an ongoing debate regarding the relationship between obesity and CV outcomes according to the presence of diabetes. In a large-scale population study regarding diabetes mellitus, overweight or obese patients had a lower risk of major adverse CV events (MACE) and all-cause mortality than those with normal-weight patients [7]. However, there is limited evidence regarding the long-term prognosis in patients with established CV disease. Therefore, we investigated the association of obesity and diabetes mellitus with the clinical outcomes in patients treated with acute coronary syndrome (ACS).

Study population
This study used data from the nationwide administrative claims-based databases of the Korean National Health Insurance Service (NHIS), which covers > 98% of the entire Korean population. The NHIS-Health Screening Cohort is a dataset with a random sample of 10% of the population aged 40-79 years who completed a National Health Screening test in 2002 or 2003. The database provides information regarding the demographic characteristics, medical claims (including diagnostic and treatment codes), health surveys, physical examinations, and biochemical tests of 514,866 participants. The diagnostic codes are based on the International Classi cation of Diseases, 10 th revision (ICD-10) [8].
Patients with a previous history of ACS or cancer were excluded. Patients whose data were unavailable for analysis or not received the health screening test within 1 year were also excluded ( Figure 1). The follow-up period was de ned as the time from the date of the health screening test to the rst outcome event, date of death, or end of the study period (December 31, 2015), whichever came rst. The study protocol was reviewed and approved by the Institutional Review Board of Ajou University Hospital (AJIRB-MED-EXP-17-253). The requirement for informed consent was waived as the data used in the database were deidenti ed.
Diabetes mellitus was de ned using the following criteria: diagnosed with ICD-10 codes (E10-E14) more than twice, receiving glucose-lowering agents (ATC: A10B) for > 30 days, receiving insulin (ATC: A10A) as an outpatient, or fasting blood glucose level of ≥ 126 mg/dL. The primary outcome was MACE-CV death (ICD-10: I00-I99), MI (ICD-10: I21-I22), and stroke (ICD-10: I60-I64). The secondary outcomes were individual components of MACE, hospitalization for heart failure (HHF; ICD-10: I50), and all-cause death. All-cause and CV death were de ned by death status in the NHIS database, which was linked to the National Death Registry using unique resident registration numbers.

Statistical analysis
All categorical variables are presented as frequencies and percentages. Normally distributed data were presented as mean ± standard deviation, whereas nonparametric data are presented as median and interquartile range by BMI. Cox proportional hazard regression analyses were performed to identify the association of BMI with the primary and secondary outcomes according to the presence of diabetes, calculating hazard ratio (HR) and 95% con dence interval (CI) and adjusting for the following potential confounders: sex, age, systolic blood pressure, fasting glucose level, total cholesterol level, drinking status, smoking status, physical exercise, household income, use of antihypertensive agents, use of statins, use of antiplatelet agents, previous history of heart failure, previous history of stroke, and index year. All analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).

Results
The study initially enrolled 26,597 patients who were diagnosed with ACS, and 6,978 participants ful lled the inclusion criteria ( Figure 1). The baseline characteristics according to the presence of diabetes mellitus are summarized in Table 1. Compared with patients without diabetes, those with diabetes (n = 3989, 57.1%) were older, were more likely to be females, and had higher BMI and systolic blood pressures. Patients with diabetes were less likely to be current smokers and had low alcohol consumption compared with those without diabetes; however, they had more frequent comorbidities and more frequently used concomitant medications. The baseline characteristics of patients with and without diabetes strati ed by BMI are summarized (Additional le 1: Supplementary Table 1). Obese patients were more likely to be younger, males, and physically active, and current smokers and more frequently used medications than normal-weighted patients.  The risk of the primary outcome across the BMI groups in post-ACS patients are described in Figure 2.  Figure 1). However, the obese group with diabetes had similar risks of events to the reference group (stroke: HR 1.11, 95% CI 0.84-1.47; HHF: HR 1.04, 95% CI 0.76-1.40, respectively). The risks of death (CV and allcause) and MI tended to be lower in the obese group with and without diabetes than in the reference group, although there was no statistical signi cance.
The subgroup analyses strati ed by age, sex, smoking, and clinical diagnosis are shown in Supplementary Figure 2 (Additional le 3). Obese female and elderly (≥ 65 years) patients without diabetes had a lower rate of MACE, while diabetes mellitus increased the risk of the primary outcome of the under-or normal-weight patients in male, younger (<65 years), and MI groups. In the smoking subgroup, there was no difference according to BMI and the presence of diabetes mellitus.

Discussion
This study showed that obesity and diabetes mellitus affected the long-term outcomes of patients with ACS. While diabetes mellitus deteriorated the clinical outcomes, obesity ameliorated the results; the "obesity paradox" was obvious in post-ACS patients regardless of the presence of diabetes. Obesity decreased the risk of MACE driven primarily by stroke and reduced the rate of HHF exclusively in patients without diabetes mellitus. These were more evident in female and elderly patients without diabetes.
However, patients with diabetes did not show such bene t. All-cause and CV mortality seemed to be lowered in overweight and obese patients with and without diabetes.
As the prevalence of diabetes mellitus has increased worldwide, the prevalence also correspondingly increased in Korea, accounting for approximately 5.1 million people with diabetes in 2016 [10]. Patients with diabetes not only have a higher risk for developing coronary heart disease, but their prognosis worsens after acute MI [11,12]. In a previous cohort study, patients with diabetes had a lower socioeconomic status than those without diabetes, which would also affect the adverse CV results [5]. The present study also showed similar ndings.
Obesity increases the risk of CV disease via dysregulated metabolism sharing the common mechanism with diabetes (e.g., insulin resistance and in ammation) [13]. BMI increases the prevalence of diabetes mellitus up to 5 times depending on the race [14,15]. However, in contrast to the biological effect of obesity, overweight or obese patients show a better prognosis than normal-weight patients, which is called the "obesity paradox" [13]. In a systematic review for the general population, all-cause mortality was the lowest in the overweight BMI group [16]. In a meta-analysis study, obese patients with coronary artery disease had no increased risk for all-cause and CV death, which was most evident in BMI 25.0-29.9 kg/m 2 [17]. A study including 12 million Koreans also reported the inverse relationship between BMI and all-cause mortality. The optimal BMI for lowering all-cause death was higher in men than in women and was increased with aging, which corresponded to the overweight or obese BMI categories [18]. Our study showed that overweight or obese patients did not have an increased risk of all-cause and CV mortality compared with normal-weight patients, regardless of the presence of diabetes. In the overall population, the risk of all-cause and CV mortality mortalities was signi cantly lower in the overweight and obese groups than in the reference group (Additional le 1: Supplementary Table 2). As observed in the present study, obese patients had a favorable outcome of MACE with lower event rates of stroke and HHF compared with the normal-weight group. Stroke was more prevalent in obese patients, but the "obesity paradox" has remained. In the REACH study, the higher BMI groups had a comparable risk of stroke in patients with established CV disease than the reference group [19]. A study of 5,202 patients with a previous history of CV disease reported a consistent nding that obesity was protective for clinical outcomes including stroke and HHF. The risks of stroke and HHF increased by 10% and 5%, respectively, with the decreased weight [20].
The association of BMI and long-term outcomes of patients with and without diabetes mellitus after CV events has not been recognized. In a study of 19,579 patients with diabetes and established CV disease, the increasing BMI enhanced the clinical outcomes during a 2-year follow-up [19]. The PROactive study revealed that the lowest mortality was seen in patients with a BMI of 30-35 kg/m 2 . Patients with diabetes and CV comorbidity had the risk of all-cause and CV death, which increased by 13% and 7% for each 1% of weight loss [22]. The MONICA/KORA study with 4,504 patients after MI demonstrated the relationship between obesity and CV outcomes regarding the presence of diabetes during 6 years. Higher BMI decreased the risk of long-term survival in patients without diabetes, although diabetes attenuated the protective role of obesity [23].
The present study adds to the accumulating evidence that "obesity paradox" is obvious in patients with CV disease. It is a novel nding on the relationship between BMI and diabetes mellitus, that diabetes deteriorates the prognosis of ACS patients for the long-term duration in the Korean population while obesity was protective.
The cardiometabolic consequences of obesity have a more relevant impact on CV outcomes than obesity.
A study on the population with CV disease demonstrated that the cardiometabolic dysfunction increased the risk of CV morbidity and all-cause mortality. Even in comparison with normal-weight patients with cardiometabolic dysfunction, overweight and obese patients had a comparable prognosis of CV disease [24]. A propensity-matched study of 7,788 patients with heart failure also reported a consistent result: the paramount difference in mortality between obese patients with and without diabetes [25]. The impaired cardiometabolic function, such as diabetes mellitus, might offset the protective role of obesity.
Most epidemiologic studies regarding "obesity paradox" used BMI as the obesity parameter. BMI is a simple and feasible measure, but it has some pitfalls. The HORIZONS-AMI trial including patients with STsegment elevation MI showed that obese patients have more comorbidities, vulnerable features of the coronary lesions, and premature onset of MI. However, BMI did not predict the acute and long-term mortality [26]. The ethnic difference in BMI should also be considered. The pooled analysis for post-ACS patients reported that Asians had similar BMI between people with and without diabetes (24.7 vs.
24.2 kg/m 2 ), whereas the Caucasians with diabetes had higher BMI than those without diabetes (29.3 vs. 27.2 kg/m 2 ) [27]. The effect of obesity and diabetes for CV outcomes might be distinguishable between the ethnicities. In contrast to BMI, central obesity (e.g., waist-to-hip ratio) was more related to the clinical outcomes, and body fat distribution or epicardial fat were considered reliable measures for predicting the prognosis in CV disease [23]. A further study with these parameters is needed to clarify the interrelationship of diabetes mellitus and obesity in patients with CV disease.
Our ndings showed that the aim of treatment in post-ACS patients with and without diabetes should be to prevent or control cardiometabolic complications and not merely weight reduction. Meanwhile, lowering the weight should not be underemphasized as intentional weight loss is still associated with an improvement of comorbidities and long-term prognosis in CV disease [19].
This study has some limitations. The data obtained from the claims data could not exclude the possibility of a diagnosis code inaccuracy, disease misclassi cation, or the bias of the observed difference-in-difference estimate toward the null. Additionally, this study was not able to consider the duration of diabetes, the quality or grade of diabetic treatment, and the information on the treatment of overweight or obesity. To identify the impact of obesity and diabetes on the long-term results of ACS patients, a well-designed prospective study is needed.

Conclusion
Obesity had a protective effect on MACE over normal weight in patients diagnosed with ACS, exclusively for stroke and HHF, but the presence of diabetes mellitus negated this effect. Especially for obese patients, the clinical practice might focus on improving the cardiometabolic pro le rather than just losing the body weight for a better long-term prognosis.

Consent for publication
Not applicable. The requirement for informed consent was waived as the data in this database were deidenti ed.

Availability of data and materials
Data cannot be shared publicly as the access of National Health Insurance Service (NHIS) data is available only at the NHIS center, Wonju, Republic of Korea. The contact information for the NHIS center of Republic of Korea is as follows: +82-33-736-2431-3 (Tel) and https://nhiss.nhis.or.kr (website).