The Association of Rate of Weight Gain During Early Adulthood With the Prevalence of Subclinical Coronary Artery Disease in Recently Diagnosed Type 2 Diabetes: The MAXWEL-CAD Study

Overweight and obesity are major risk factors for the development of type 2 diabetes (T2D) (1). Excess body fat increases insulin resistance, a condition characterized by increased insulin production and impaired glucose tolerance (2).

Even if it does not lead to obesity, weight gain by itself is associated with an increased risk of T2D and its complications (3,4). We recently showed a U-shaped association of BMI at T2D diagnosis but with higher cardiovascular disease and total mortality risks beyond a BMI of 30 kg/m² in white men and women living in Scotland (5). Although weight gain antedates the development of T2D by several years, quantitative investigation of the relationship between the amount and rate of weight gain and subclinical coronary artery disease (CAD) associated with T2D is lacking.

Weight gain may affect coronary vascular health by influencing various risk factors, such as insulin resistance, dyslipidemia, hypertension, the inflammatory process, and the prothrombotic state. Upper-airway tightness associated with weight gain can induce obstructive sleep apnea syndrome, leading to systemic insulin resistance (6).

Many studies have shown that overweight or obesity predisposes a person to heart diseases, such as heart failure and CAD (4,7). The pathophysiology of these entities and their link to obesity have been discussed previously (8).

Recent advances in computed tomography (CT) technology have allowed for a detailed evaluation of the coronary arteries, including the extent of stenosis and plaque composition (9,10). Noncalcified or mixed plaques are known to be more vulnerable to rupture than stabilized calcified plaques (11).

The current study examined the association between the rate of weight gain (Rate_{max_wt}) between the age of 20 years and the age at maximum lifetime weight with indicators of subclinical CAD at the time of T2D diagnosis. We hypothesized that rapid weight gain over and above the current BMI increases the risk of subclinical CAD in patients with newly diagnosed T2D.

We established a cohort in 2006 to investigate the effect of maximum body weight in lifetime on the development of T2D and its complications, the MAXWEL (12). We screened 5,321 individuals aged ≥30 years who visited the diabetes clinic for an initial evaluation at Seoul National University Bundang Hospital (SNUBH), Seongnam, Korea, from January 2007 to December 2009.

T2D was confirmed in 2,986 of these patients according to a glycosylated hemoglobin (A1C) ≥6.5% based on American Diabetes Association diagnostic criteria (13). Among this subset, we excluded 32 patients who had exertional chest pain with moderate to severe intensity that lasted 2–5 min over the sternum and that was relieved in 1–5 min by ceasing activities. Chest pain that radiated to either the shoulder or both arms was also included in this category. We also excluded patients with existing CAD or who received primary coronary intervention or coronary artery bypass graft according to their medical records or history taking if the medical record was not available. Finally, 2,044 participants without typical angina chest pain or existing CAD but with one or more risk factors were selected to enhance the power of the study. The risk factors were as follows: 1) BMI ≥25 kg/m² (overweight); 2) LDL cholesterol (LDL-C) level ≥130 mg/dL or taking statin medication; 3) triglyceride level ≥150 mg/dL, HDL cholesterol (HDL-C) level <40 mg/dL in men or <50 mg/dL in women, or taking lipid-lowering medications other than statins; 4) systolic blood pressure (SBP)/diastolic blood pressure (DBP) ≥140/90 mmHg or taking antihypertensive medication based on the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (14); 5) current smoker; and 6) a family history of CAD.

Patients with type 1 diabetes (measured by GAD antibody [n = 12]) or diabetes with a secondary cause (n = 9) were excluded. We also excluded patients with chronic diseases such as malignancy (n = 23), gastrointestinal disorder (n = 16), lung disease (n = 26), and eating disorder and/or depression (n = 14) or a history of taking medication for weight control for >3 months (n = 18).

One hundred sixty-six subjects were excluded because they were unable to recall their maximum weight (Wt_max) or age at maximum weight (Age_{max_wt}). Another 36 with a high serum creatinine level (>1.5 mg/dL in men, >1.4 mg/dL in women) were excluded because of a contraindication to the radiocontrast dye. Finally, 1,724 participants gave written informed consent for CAD evaluation using cardiac CT angiography (CTA), which was performed within 3 months after T2D diagnosis. The protocol was reviewed and approved by the institutional review board of SNUBH (B-0909/083-008). This study was registered at ClinicalTrials.gov (NCT00816608).

Body weight at age 20 years (Wt_20y) was obtained in 54% of study participants from medical records, military service or college examination records, or personal records. Wt_max before T2D diagnosis and Age_{max_wt} were self-reported. Weight from pregnancy and to 1 year after delivery was disregarded. In 31.0% of subjects (n = 535) who were randomly selected from all participants, the recalled Wt_20y and Wt_max were validated by written documentation, including medical records, military service or college examination records, or personal recording. The absolute differences between the recalled Wt_20y and Wt_max and the written documentation were 1.2 ± 1.1 and 1.3 ± 1.2 kg, respectively. The actual differences between them were −0.2 ± 1.7 and −0.3 ± 1.8 kg, respectively. When agreement was defined as an absolute difference of ≤0.5 kg between the recalled Wt_20y and Wt_max and those obtained from written documentation, the agreement rates were high (r = 0.94 and r = 0.92, respectively; both P < 0.001) (Supplementary Fig. 1). We calculated the Rate_{max_wt} by dividing the weight change (in kilograms) after age 20 years by the time difference in years between age 20 years and Age_{max_wt}. The definitions of weight-related variables as well as the study design are shown in Fig. 1.

Interviews for lifestyle characteristics were conducted by designated physicians using a standardized survey. Smoking status was divided into three categories: nonsmokers, ex-smokers, and current smokers. Alcohol intake was assessed by frequency and quantity of beer, spirits, sake, and wine intake during the past 12 months. Alcohol intake in grams of alcohol per week fell into two categories: light to moderate (≤199.9 g/week) and heavy (≥200 g/week). Physical activity fell into two categories: no or irregular (two or fewer times per week) and regular (three or more times per week), with each period defined as exercising for at least 30 min. Family history of CAD at age <60 years included first-degree relatives.

At the time of T2D diagnosis, height and body weight were measured with the participants barefooted and wearing light clothing to the nearest 0.1 cm or 0.1 kg, respectively. SBP and DBP were measured twice with a 5-min interval after the participant had been sitting for 10 min. The mean value for each blood pressure measurement was used.

Fasting glucose level and other biochemical parameters were measured in a 12-h fasted state at the time of T2D diagnosis (Hitachi 747 chemistry analyzer; Hitachi, Tokyo, Japan). A1C level was measured in SNUBH, a National Glycohemoglobin Standardization Program Level II–certified laboratory.

Cardiac CTA was performed at the time of T2D diagnosis with a 64-slice CT scanner (Brilliance 64; Philips Medical Systems, Best, the Netherlands) and standard scanning protocol (15). The median time difference between T2D diagnosis and cardiac CTA was 34 days (range, 0‒92 days). All scans were analyzed independently in a blinded fashion by two experienced radiologists using a three-dimensional workstation (Brilliance; Philips Medical Systems). All coronary segments ≥1.5 mm in diameter were assessed. Image quality was evaluated on a per-segment basis, and noninterpretable segments were excluded from the analysis.

Coronary artery stenosis was assessed by tracing semiautomatically the contrast-enhanced portion of the coronary lumen at the maximal stenotic site and comparing this with the proximal and distal reference sites (16). Narrowing of ≥50% of the lumen was defined as significant stenosis according to the modified classification by the American Heart Association (AHA) (17).

Multivessel disease was defined as the presence of stenosis of ≥50% in two or more coronary vessels.

The atherosclerotic coronary segment was identified for coronary artery segments with any plaque, which was defined as a structure of ≥1 mm² within and/or adjacent to the coronary vessel lumen that could be distinguished clearly from the lumen and surrounding epicardial fat. Every single interpretable coronary artery segment was evaluated for the presence of plaque and its characteristics. Each plaque was categorized into one of three types: 1) calcified plaque containing calcified tissue occupying ≥50% of the plaque area (density >130 Hounsfield units on native scans); 2) mixed plaque containing calcified tissue occupying <50% of the plaque area; and 3) noncalcified plaque, which had no calcified tissue (16).

The Agatston score was used as the coronary artery calcium score (CACS) (18). The CACS was categorized into two groups: <100 and ≥100.

All data are presented as the mean and SD or percentage and were analyzed with SPSS for Windows version 17.0 (IBM Corporation, Armonk, NY). The distributions of triglyceride level and Rate_{max_wt} were skewed (Kolmogorov-Smirnov Z = 1.216 and 1.325, respectively; both P < 0.05), and the values of triglyceride level and Rate_{max_wt} were normalized by logarithmic transformation for all analyses. The Student t and χ² tests were used to compare variables. Coronary artery stenosis, multivessel involvement, atherosclerotic coronary segments (any plaque), plaque type, and CACS were compared according to quartiles of Rate_{max_wt}.

To test independent associations of weight gain with various coronary lesions independent of Wt_20y, weight at T2D diagnosis, and other cardiovascular risk factors, we applied four separate multivariable logistic regression models to examine for associations with coronary artery stenosis (≥50%), multivessel involvement (two or more vessels), atherosclerotic coronary segments (any plaque), and high CACS (≥100). Additional multivariable logistic regression models using weight change instead of the Rate_{max_wt} were also conducted in association with these coronary artery lesions. Significance was defined as P < 0.05 for all analyses.

The baseline characteristics of the 1,724 participants are shown in Table 1. The percentage of men was 49.7%. The ranges of age and BMI at diagnosis of T2D were 30−68 years and 17.1−40.1 kg/m², respectively, with participants with more rapid weight gain being younger and heavier at diagnosis (Supplementary Fig. 2). Almost 20% had a family history of CAD. Among all participants, 50.8% were overweight as defined as BMI ≥25 kg/m²; 37.9% had an LDL-C level ≥130 mg/dL or were taking statin medication; 19.6% had a triglyceride level ≥150 mg/dL, HDL-C level <40 mg/dL in men, or <50 mg/dL in women or were taking lipid-lowering medication other than statins; and 24.5% had an SBP/DBP ≥140/90 mmHg or were taking antihypertensive medication at the time of T2D diagnosis.

The weight change was ∼12.9 kg; this value was calculated by subtracting the mean Wt_20y of 60.1 kg from the mean Wt_max of 73.0 kg (Table 1). The mean Age_{max_wt} was 41.3 years, and the mean change in time was 21.3 years. From these two variables, the Rate_{max_wt} was calculated as 0.59 ± 0.56 kg/year. There was no significant difference in the amount and rate of weight gain between men and women.

More participants in the higher quartiles of Rate_{max_wt} had significant coronary artery stenosis, particularly in the left main or left anterior descending artery and left circumflex artery, and more frequent multivessel involvement and atherosclerotic coronary segments than those in the lower quartiles of Rate_{max_wt} (Table 2). The log-transformed CACS did not differ among quartiles, but there was a trend for more participants in the higher quartiles of Rate_{max_wt} to have a CACS ≥100 compared with those in the lower quartiles.

In the multivariable linear regression for each phenotype of CAD (Table 3), high blood pressure or taking antihypertensive medication, more pack-years of smoking, family history of CAD, and high LDL-C level or taking statin medication were significantly associated with significant coronary artery stenosis, multivessel involvement, and atherosclerotic coronary segments. High blood pressure or taking antihypertensive medication, more pack-years of smoking, family history of CAD, and high LDL-C level or taking statin medication were significantly associated with a high CACS. A higher Rate_{max_wt} was significantly associated with significant coronary artery stenosis and atherosclerotic coronary segments. All variance-inflation factors among the independent variables were <1.29 in each regression model, suggesting that there was no significant collinearity among the covariates. Using weight change instead of the Rate_{max_wt} produced a significant, but attenuated, association with coronary artery stenosis and mixed/noncalcified plaque but not with multivessel involvement, atherosclerotic coronary segments (any plaque), or the CACS (Supplementary Tables 1 and 2).

The prevalence of mixed and noncalcified plaques was higher with higher Rate_{max_wt}, whereas there was no relationship between the Rate_{max_wt} and the presence of calcified plaques (Supplementary Fig. 3). We used a multivariable logistic regression model to investigate further the independent risk of the Rate_{max_wt} on plaque type. After adjusting for the same variables used in the model for CAD, high SBP or taking antihypertensive medication, heavy smoking, family history of CAD, and high LDL-C level or taking statin medication were all significantly associated with any kind of plaque (Table 4). However, the highest Rate_{max_wt} quartile was significantly associated with the presence of mixed and noncalcified plaques but not with calcified plaques.

In this study of patients with newly diagnosed T2D and at least one risk factor for CAD, rapid weight gain from young adulthood was significantly associated with significant stenosis, multivessel involvement, and the presence of atheromatous plaques in the coronary arteries independent of current BMI. High blood pressure, heavy smoking history, and higher A1C and LDL-C levels were also independent predictors of the presence of CAD in these patients, lending external validity to the findings.

Few studies have investigated the relationship between weight gain during early childhood or adolescence and cardiovascular risk factors in later life (19,20). The Bogalusa Heart Study confirmed that obesity directly effects coronary atherogenesis in youth (21) and that this process is accelerated in the presence of combined metabolic abnormalities (22). In population-based long-term follow-up studies in subjects of mainly white European extraction, being overweight in childhood was associated with an increased risk of CAD (23,24). Excess weight gain was associated with more extensive atherosclerosis, insulin resistance, dyslipidemia, and high blood pressure in the DCCT-EDIC (Diabetes Control and Complications Trial-Epidemiology of Diabetes Interventions and Complications) follow-up study (25). A study of Japanese people reported that rapid weight gain during early childhood was associated with a poor lipid profile and high blood pressure in adolescence (26). In a previous study of 115,818 American women, weight gain after 18 years of age was a strong predictor of CAD risk among those within a normal BMI range (27). In the ARIC (Atherosclerosis Risk in Communities) study, weight gain since age 25 years was associated with elevated risk of CAD and stroke after controlling for baseline BMI and other covariates (28). These studies focused mainly on obesity or the amount of weight gain during a certain period rather than on weight dynamics. In the current study, the rate of or actual weight gain showed an association with cardiac CTA findings. However, the Rate_{max_wt} was associated with more types of subclinical atherosclerosis, and the degree of its association seemed to be higher than actual weight gain (Supplementary Table 1 and 2). These findings suggest that the Rate_{max_wt} may be more important in the development of atherosclerosis than actual weight gain, a testable hypothesis that future studies can address. Thus, we believe that the current study is possibly the first to show the influence of rapid weight gain on the risk of subclinical CAD at the time of T2D diagnosis, although conceptually, more rapid weight gain is broadly linked to earlier development of obesity.

Of note, Tirosh et al. (29) reported that the risk of CAD in midlife is associated with an elevated BMI in adolescence and young adulthood, whereas the risk of T2D is associated mainly with increased BMI close to the time of diagnosis. Recently, Reis et al. (30) showed that longer periods of obesity over the level of current BMI were more strongly associated with coronary artery calcification. These findings, together with the current results in an Asian population, a group at greater cardiometabolic risk for a given BMI, indicate that the processes of atherosclerosis development depend not only on the level of obesity at the time of measurement but also on the rate of weight change in younger years and linked to this, the overall duration of obesity.

In this study, 1.2% (n = 21) of participants had a BMI <18.5 kg/m² at age 20. There was no significant alteration in the association of rapid weight gain with coronary artery stenosis after excluding these underweight participants.

A significant association was found in this study between the Rate_{max_wt} and the presence of mixed or noncalcified plaques. There is no clear explanation for this finding, but conceivably, a rapid increase in weight could be more damaging to the vascular system than a slow increase, given the shorter period of time available to adapt to weight increase (31). Rapid weight and adiposity gains lead to alterations in gene expression of growth factors and cytokines, such as transforming growth factor-β, that are known to play important roles in vascular integrity (32) and plaque stability (33). Of course, it is possible that other, not easily measureable characteristics are also associated with rapid weight gain and plaque promotion. Regardless of the mechanism, the current findings further strengthen the concept that CAD development depends more on the dynamics of weight change (i.e., its trajectory over life) than the current level of adiposity. This important point is not well appreciated but has clinical and public health ramifications.

A scientific statement from the AHA (34) and the expert consensus document from the American College of Cardiology and AHA (35) concluded that cardiac CTA may be helpful in screening for CADs in asymptomatic individuals at an intermediate risk for CAD. In the current study, we used cardiac CTA to screen for subclinical CAD in patients with newly diagnosed T2D who had at least one cardiovascular risk factor. Advanced CT techniques, such as 64-slice cardiac CTA, can measure the degree of stenosis with high diagnostic accuracy and can provide details about plaque composition, including whether the plaque is noncalcified or mixed and rupture prone (11). In studies that investigated the prognostic value of cardiac CTA, coronary artery stenosis and great plaque burden were stronger predictors of future cardiovascular events than were routine clinical predictors, exercise stress test results, or the CACS in patients without known CAD (36,37). Similarly, in studies of outpatients at low to intermediate risk, coronary artery stenosis severity, the presence of noncalcified and/or mixed plaque, and multivessel involvement were superior to traditional risk factors, including the CACS, for predicting CAD (11,38).

The MAXWEL cohort has several strengths. First, weight information at age 20 years was obtained accurately from official written documents in 54% participants. Second, identification of T2D was based on laboratory results and not on self-report, and we carefully excluded patients with likely type 1 diabetes. Third, only patients with newly diagnosed T2D were included, which enabled us to assess their glycemic level and the status of diabetic macrovascular complications within 3 months after diagnosis. Finally, studies in people of Asian ethnicity that comprehensively assessed the coronary arteries with multislice CT are limited; hence, the present data are novel and important.

There are also some limitations of this study. First, identification of Wt_max and Age_{max_wt} was based on the participants’ self-report. Although there was high agreement rate between the recalled Wt_max and the written documentation, Age_{max_wt} could possibly be prone to recall error, and thus, the rate of weight change is likewise potentially influenced. Second, we excluded participants who had lost weight between the age of 20 years and the time of T2D diagnosis (36 patients [2.1%]), but this small number was unlikely to have biased the results. Third, other factors that can contribute to body weight, such as smoking habit, alcohol consumption, physical activity, and dietary pattern, in a time-dependent manner as well as consumption of diet pills and psychological stress were not considered. Fourth, we did not assess weight fluctuation, which may affect low-grade systemic inflammation (39) or metabolic abnormalities (40). However, other studies reported that weight fluctuation was not related to cardiovascular morbidity or mortality (41,42). Finally, a possibility of unknown duration of diabetes exists, which may have influenced the study results.

In conclusion, a greater Rate_{max_wt} from young adulthood was associated with early development of CAD, including the presence of noncalcified or mixed plaques in patients with newly diagnosed T2D independent of weight/BMI at the time of cardiovascular phenotyping. The current findings imply that the rate of evolution of atherosclerosis may be affected by weight gain from young adulthood, a modifiable and avoidable risk factor (43). It is true that small changes early on can be leveraged to great benefits later in life. Taken together with other emerging data, the results suggest that population-based approaches for weight control during adolescence, including behavioral factors, should be highlighted to reduce the risk of macrovascular complications in general and in people at elevated risk for diabetes.

Funding. This study was supported by a National Research Foundation grant funded by the Korean government (2006-2005410) and Seoul National University Bundang Hospital.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. S.L. contributed to researching the data and writing the manuscript. S.H.C., K.M.K., S.I.C., E.J.C., M.J.K., K.S.P., H.C.J., and N.S. contributed to the discussion and reviewed and edited the manuscript. S.L. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.