Exercise Volume Versus Intensity and the Progression of Coronary Atherosclerosis in Middle-Aged and Older Athletes: Findings From the MARC-2 Study

Background: Physical activity and exercise training are associated with a lower risk for coronary events. However, cross-sectional studies in middle-aged and older male athletes revealed increased coronary artery calcification (CAC) and atherosclerotic plaques, which were related to the amount and intensity of lifelong exercise. We examined the longitudinal relationship between exercise training characteristics and coronary atherosclerosis. Methods: Middle-aged and older men from the MARC-1 (Measuring Athlete’s Risk of Cardiovascular Events 1) study were invited for follow-up in MARC-2 (Measuring Athlete’s Risk of Cardiovascular Events 2) study. The prevalence and severity of CAC and plaques were determined by coronary computed tomography angiography. The volume (metabolic equivalent of task [MET] hours/week) and intensity (moderate [3 to 6 MET hours/week]; vigorous [6 to 9 MET hours/week]; and very vigorous [≥9 MET hours/week]) of exercise training were quantified during follow-up. Linear and logistic regression analyses were performed to determine the association between exercise volume/intensity and markers of coronary atherosclerosis. Results: We included 289 (age, 54 [50 to 60] years [median (Q1 to Q3)]) of the original 318 MARC-1 participants with a follow-up of 6.3±0.5 years (mean±SD). Participants exercised for 41 (25 to 57) MET hours/week during follow-up, of which 0% (0 to 19%) was at moderate intensity, 44% (0 to 84%) was at vigorous intensity, and 34% (0 to 80%) was at very vigorous intensity. Prevalence of CAC and the median CAC score increased from 52% to 71% and 1 (0 to 32) to 31 (0 to 132), respectively. Exercise volume during follow-up was not associated with changes in CAC or plaque. Vigorous intensity exercise (per 10% increase) was associated with a lesser increase in CAC score (β, −0.05 [−0.09 to −0.01]; P=0.02), whereas very vigorous intensity exercise was associated with a greater increase in CAC score (β, 0.05 [0.01 to 0.09] per 10%; P=0.01). Very vigorous exercise was also associated with increased odds of dichotomized plaque progression (adjusted odds ratio [aOR], 1.09 [1.01 to 1.18] per 10%; aOR, 2.04 [0.93 to 4.15] for highest versus lowest very vigorous intensity tertiles, respectively), and specifically with increased calcified plaques (aOR, 1.07 [1.00 to 1.15] per 10%; aOR, 2.09 [1.09 to 4.00] for highest versus lowest tertile, respectively). Conclusions: Exercise intensity but not volume was associated with progression of coronary atherosclerosis during 6-year follow-up. It is intriguing that very vigorous intensity exercise was associated with greater CAC and calcified plaque progression, whereas vigorous intensity exercise was associated with less CAC progression.

P hysical activity and exercise training are associated with a significantly lower risk of cardiovascular disease (CVD) and CVD events, with an average 30% to 40% risk reduction for the most active individuals. 1,2 Computed tomography (CT) imaging permits the measurement of coronary artery calcification (CAC), which is an indicator of coronary atherosclerotic plaque burden and the risk of future cardiovascular events. [3][4][5] Coronary CT angiography (CCTA) allows more detailed characterization of atherosclerotic plaques, which is also predictive of future events. 4 Previous studies indicate that coronary plaque morphology provides prognostic information beyond CAC scores, with calcified plaque being less frequently associated with CVD events compared with mixed or noncalcified plaque. 4 Unexpectedly, emerging evidence suggests amateur athletes have increased coronary atherosclerosis compared with less active healthy controls. [6][7][8][9][10] Athletes more often had CAC scores ≥100 Agatston units, which was associated with their lifelong exercise volume and intensity of exercise training. 7 In the most active athletes, atherosclerotic plaque morphology appeared more often to be calcified and less mixed (ie, partially calcified). 6,8 However, previous studies were cross-sectional and could not determine whether exercise influenced the acceleration of coronary atherosclerosis and plaque morphology.
We used CAC scoring and CCTA to prospectively investigate the association between exercise volume and intensity and the progression of coronary atherosclerosis in a sizeable cohort of middle-aged and older male athletes. We hypothesized that higher volume and intensity exercise would be associated with the greatest increase in coronary atherosclerosis.

METHODS Study Design and Population
The MARC-2 (Measuring Athletes' Risk of Cardiovascular Events 2) study, is a follow-up of the MARC-1 (Measuring Athletes' Risk of Cardiovascular Events 1) study. 11 In short, MARC-1 investigated the presence of subclinical coronary atherosclerosis in healthy, middle-aged, male amateur athletes using CT imaging. Asymptomatic middle-aged and older men (≥45 years of age) who had undergone sports medical evaluation without abnormalities were recruited between 2012 and 2014 (n=318). Men with abnormal sports medical examinations, known CVD, contrast allergy, and renal impairment were excluded. Surviving MARC-1 participants (n=314) were invited to participate in MARC-2. Data were collected between May 2019 and February 2020. For the present analysis, individuals who underwent a percutaneous coronary intervention during follow-up were excluded since CAC and plaque characteristics cannot reliably be assessed in that segment (n=2). Ethical approval was obtained from the Dutch Minister of Health, Welfare and Sport (no. 1456153-184955-PG). Participants provided written informed consent. The study was conducted according to the Declaration of Helsinki. The data that support the findings of this study are available from the corresponding author upon reasonable request.

Exercise Characteristics
Information regarding exercise characteristics was collected at MARC-1 and MARC-2 using a validated questionnaire. 12 Participants reported the type of sport, duration (in years) for each sport, frequency and duration of exercise sessions, and associated level of performance (recreational versus competitive). We assigned a metabolic equivalent of task (MET) for all reported sports using the Compendium of Physical Activities but used the reported level of performance (ie, performing the sport competitively at a regional level or higher) for tennis (5 versus 7.3) and swimming (8 versus 10) to choose the appropriate MET score from the Compendium since multiple options

Clinical Perspective
What Is New?
• Exercise intensity, but not exercise volume, was associated with coronary atherosclerosis progression in middle-aged and older male athletes during 6 years of follow-up. • The proportion of vigorous intensity exercise training was associated with less progression of coronary artery calcification, whereas the proportion of very vigorous intensity exercise was associated with greater progression of coronary artery calcification and atherosclerotic plaques.
What Are the Clinical Implications?  13 In addition, running, fitness, and cycling categories were assigned MET scores based on activity type, with MET scores of 11.8 and 7.0 allocated to running and jogging, respectively. Strength training was given a MET score of 3.5 and cardio fitness received a MET score of 5.5. Different types of cycling (eg, race, mountain bike, spinning) also received specific MET scores (Table S1). We calculated the exercise volume per sport by multiplying the MET score for the specific sport by the reported exercise volume (session duration × frequency/ week), months of practice per year, and total years of practice. For the current analyses, exercise volume (expressed in MET hours/week) during follow-up was used. Exercise volume is the sum of all sports activities between the baseline of MARC-1 and the follow-up assessment of MARC-2. We made tertiles of exercise volume (ie, low, medium, high) because the previously defined exercise groups (<1000, 1000 to 2000, >2000 MET min/week) did not fit the present study and would have produced skewed group distribution with 11% versus 26% versus 63% of individuals per subgroup. We also classified exercise intensity as light (<3 METs), moderate (3 to 6 METs), vigorous (6 to 9 METs), or very vigorous (≥9 METs) and calculated the percentage of exercise MET hours/week in these ranges as a proportion (%) of the total exercise volume. Since exercise in the light intensity range was negligible, with only 9 participants performing low volumes of light intensity exercise, we did not include light intensity in our analyses. Subsequently, we developed tertiles (ie, low, medium, high) based on the percentage of vigorous and very vigorous intensity exercise.

Cardiac CT
CT imaging was performed using the exact same 256-slice scanner (Philips Healthcare, The Netherlands) as in MARC-1, with electrocardiographic gating according to guidelines. 14 A noncontrast CT was acquired to calculate CAC scores (scan parameter: 120 kV, 210 mAs), which was followed by CCTA using the same protocol as in MARC-1. Participants experiencing signs of contrast reaction during baseline scan underwent only a noncontrast CT scan at follow-up (n=5). CT scans were processed on a workstation (IntelliSpace Portal; Philips Healthcare) by experienced technicians, and assessed by an experienced cardiovascular radiologist. Guidelines for interpretation and reporting of CCTA, including the 18-segment coronary artery model, were used to analyze plaque characteristics per segment. 15 The Agatston CAC score was used to quantify coronary artery calcification. 16 We quantified progression using delta CAC score (CAC score follow-up − CAC score baseline), which cannot be negative and any increase in CAC score indicates progression. CCTA was used to identify plaques not detected by noncontrast CT and to assess plaque characteristics. We classified plaques as calcified, noncalcified or mixed (ie, partially calcified) plaques. Study definitions regarding exercise and plaque characteristics are summarized in Table S2.

Statistical Analysis
All parameters were visually inspected for normality and checked for kurtosis and skewness. Continuous variables are reported as mean±SD when normally distributed or as median [interquartile range] when not. Categoric variables are presented as proportions. Participant characteristics at baseline and follow-up were compared using paired T-tests for normally distributed data and Wilcoxon signed rank tests for nonnormally distributed variables. The McNemar test was used for binary variables. Linear regression analyses were performed between exercise volume and intensity and change in CAC score and the number of plaques. CAC scores were transformed (Ln delta CAC score + 1) and analyses were adjusted for baseline confounders, CAC score or number of plaques at baseline, and time between CT scans. Binary logistic regression was used to calculate adjusted odds ratios (aORs) for the association between exercise volume and intensity and progression of plaque (dichotomized: increase/no change in number of plaques) and plaque types. All analyses were performed with exercise volume and intensity as continuous parameter and in tertiles. Exercise intensities were assessed as a proportion of MET hours/week and included separately, while adjusting for exercise volume. Additionally, we explored potential nonlinear associations using cubic spline regression analyses. We decided to adjust a priori for the following known cardiovascular risk factors at baseline: age, body mass index, systolic blood pressure, smoking (pack years), total cholesterol, family history of coronary heart disease, and use of antihypertensive, antidiabetic, and statin medications. Statistical significance was assumed at P<0.05. Since various analytic strategies are used to assess CAC progression, we performed multiple sensitivity analyses to include alternative methods to assess CAC progression. 17,18 For that purpose, we also calculated the delta CAC score (ie, CACat followup CACat baseline ) and the volume score using the squared root method, which entails removing density from the CAC score equation, using the squared root method. We also annualized the logarithmically transformed delta CAC score and squared root CAC and volume scores instead of including follow-up time as a covariate. Statistical analyses were performed using SPSS Statistics 25 (IBM Corp, USA), except for the cubic spline regression analyses which were conducted using R software version 4.0.3.

RESULTS
A total of 291 men (age, 54 [50 to 60] years at MARC-1 baseline and 60 [56 to 66] years at follow-up) were included in MARC-2 (92% of original cohort [ Figure S1]). Characteristics of MARC-1 participants that did (n=291) and did not (n=27) participate in MARC-2 were comparable (Table S3). Two participants (1%) received coronary stents before MARC-2 and were excluded from the current analyses. Five participants (2%) underwent noncontrast CT scans only in MARC-2 because of mild contrast reactions during their MARC-1 CCTAs. Two participants (1%) had (false) regression of CAC (from 1 to 0 and from 12.2 to 3.6), which was caused by a sampling error related to partial volume effect (calcium just below 130 Hounsfield units at follow-up) and unreliable measurement related to movement artefact. These individuals were not included in the delta CAC score analyses. Therefore, 287 men were included in the MARC-2 CAC analyses and 284 in the plaque analyses. Average follow-up between CT-scans was 6.3±0.5 years. Average blood pressure and use of antihypertensives and statins increased during follow-up, whereas total cholesterol remained similar (  (Figure 1). Exercise volume during followup was greater than the average lifelong exercise volume that was assessed upon enrollment in MARC-1 (Table  S4). Hence, exercise duration, number of exercise sessions, and MET hours per week were greater at MARC-2 measurement compared with MARC-1 (Table S4). Characteristics of exercise intensity remained similar over time with the largest portion being vigorous intensity exercise followed by very vigorous intensity exercise. Characteristics of the exercise volume and vigorous and very vigorous intensity tertile groups are summarized in Tables S5, S6 and S7. CAC was present in 151 (52%) men at baseline and increased to 205 (71%) men at follow-up. Median CAC scores increased from 1 [0 to 32] to 31 [0 to 132] (Table 2; Figure 2). CAC score categories increased over time ( Table 2): CAC scores ≥100 increased from 15% to 31%, CAC scores ≥400 increased from 6% to 13%, and CAC scores ≥1000 increased from 1% to 6% at follow-up (Figure 2). At follow-up, 26% had CAC scores in the 75th MESA (Multi-Ethnic Study of Atherosclerosis) percentile or higher, 19 compared with 22% at baseline (P=0.11). Plaque prevalence increased from 64% at baseline to 83% at follow-up (Table 2; Figure 2). A total of 216 (75%) participants showed an increased number of coronary atherosclerotic plaques and had a median change in the number of plaques of 3 [1 to 5], which consisted of 1 [0 to 3] calcified plaque, 1 [0 to 2] mixed plaque, and 0 [0 to 1] noncalcified plaques. A total of 150 (52%) participants showed an increase in the number of calcified plaques: 145 (50%) showed increased mixed plaques and 105 (36%) showed increased noncalcified plaques.
We found no association between exercise volumes and CAC and plaque progression using multivariable adjusted linear ( Figure 3) and logistic regression analyses ( Figure 4 Figure 4). Very vigorous intensity exercise was not associated with the absolute increase in number of plaques ( Figure 3), but was associated with plaque progression using logistic regression analyses (aOR,    Figure 4). This appeared to be specifically driven by an increase in calcified plaque ( Figure 4B). The absolute values are reported in Tables S8 and S9. The positive and negative value of the adjusted betas for very vigorous and vigorous intensity exercise, respectively, cannot be interpreted in isolation since these values are part of a multivariable model with 12 to 14 variables depending on the specific analysis. Since all participants either remained at CAC score 0 or had an increase in their CAC score, the increase in CAC score was lower in participants performing mainly vigorous exercise (indicated by negative beta) and greater in those performing mainly very vigorous exercise (indicated by positive beta).
Sensitivity analyses using the squared root method and annualized scores to assess CAC progression were calculated and reported in Table S10. These outcomes showed similar results as our main analyses, supporting the robustness of our findings.
Finally, cubic spline regression analyses and nonlinearity tests indicated no nonlinear associations for most analyses (P>0.05; data not shown), except for the association between MET hours/week and calcified plaque progression (P=0.048), and between both vigorous (P=0.01) and very vigorous intensity (P=0.03) and mixed plaque progression which could also be observed in the categorical data analyses (Figure 4).

DISCUSSION
We investigated progression of coronary atherosclerosis in 289 middle-aged and older male athletes using CAC scoring and CCTA and assessed its association with exercise characteristics during 6.3 years follow-up. Exercise intensity, but not volume, was associated with progression of coronary atherosclerosis. Vigorous exercise was associated with less CAC progression, whereas very vigorous exercise was associated with greater progression of CAC and (mainly calcified) plaque.

Exercise Volume
We found no association between exercise volume during follow-up and progression of coronary atherosclerosis. Our previous study, MARC-1, was cross-sectional and examined the effect of 36 (27 to 42) years of exercise on coronary atherosclerosis and found that the most active athletes had higher prevalence of CAC and atherosclerotic plaques. 6 In contrast, MARC-2 assessed the additional effects of exercise on coronary atherosclerosis after only 6.3±0.5 years follow-up, which may be insufficient to find an effect of exercise volume on CAC or plaque progression. Indeed, a study in 61 recreational athletes (74% male) found no significant difference in exercise volume between individuals with or without progression of CAC (68 [ 20 In contrast, a Korean population study categorized 25 485 young (42±6 years) Korean men and women as inactive, moderately active, and health-enhancing physically active. Moderately active was defined as ≥3 days of vigorous intensity activity for ≥20 min/day, ≥5 days of moderate intensity activity for ≥30 min/day, or ≥5 days of activities attaining ≥600 MET min/week. Health-enhancing physical activity was defined as ≥3 days of vigorous intensity activity attaining ≥1500 MET min/week or 7 days of activities attaining ≥3000 MET min/week. The latter group had a higher estimated 5-year progression of CAC during 3.0 (2.0 to 4.2) years follow-up. 21 It  Data is presented as n (%) or median (interquartile range). Since plaque characteristics for 2 different groups (n=184 and n=239) are described, no P values for within-person changes could be calculated. CAC indicates coronary artery calcification; and MESA, Multi-Ethnic Study of Atherosclerosis. *n=289 at baseline and follow-up. †n=184 at baseline; n=239 at follow-up.
is also possible that exercise volume is involved in the initiation of coronary atherosclerosis, but does not contribute to progression. The Heinz Nixdorf recall study found that CAC score itself predicts CAC progression whereas other risk factors explain little of the variance in CAC progression. 22 Overall, the results from MARC-1 and MARC-2 suggest that exercise volume plays a role in CAC prevalence, but not in CAC progression during 6-years of follow-up.

Exercise Intensity
Previous studies investigating coronary atherosclerosis in athletes are limited by their cross-sectional design and do not address the separate roles of exercise volume versus intensity because they either compare athletes with nonathletes 8,9 or do not examine intensity and volume separately. 10,20,21 We found that vigorous exercise lowered CAC progression, whereas very vigorous exercise accelerated it. Very vigorous exercise was also associated with increased progression of (calcified) plaque. These findings align with our cross-sectional MARC-1 observation that specific exercise intensities accelerate the development of calcified plaque (seen with CAC scoring and CCTA). 6 The lower progression in the more vigorous exercise group or the increased progression in the more very vigorous intensity exercise group could not be explained by the fact that the very vigorous exercise group was older or had a higher prevalence of risk factors (Tables S5 to S7), for which we also adjusted in our analyses. Thus, our longitudinal MARC-2 observations suggest that very vigorous exercise is related to calcified plaque formation, which increases the probability that certain mechanisms contribute to the increased coronary atherosclerosis in athletes. Higher intensity exercise produces higher catecholamine levels. 23 Catecholamines are relatively stable at lower intensity exercise, but increase exponentially with higher intensity exercise. 23,24 Catecholamines increase heart rate and blood pressure, both of which increase coronary mechanical stress. Increased heart rate itself has been demonstrated to accelerate atherosclerosis probably by increasing the frequency of turbulent blood flow. 25 Catecholamines also produce long-lasting proinflammatory changes in monocytes. 26 Consequently, increases in blood pressure, heart rate, mechanical stress, disrupted laminar flow and inflammation become more likely potential mechanisms of increased coronary atherosclerosis in athletes. 7

Exercise Type
We cannot separate the roles of exercise intensity and type of sport since most of our athletes cycled or ran, activities that were classified as vigorous or very vigorous, respectively (Table S1). Many athletes (71%), however, performed more than one type of sport, further obscuring the role of sport type. We previously observed that cyclists had less coronary atherosclerosis than runners, but did tend to have more calcified plaques, which we hypothesized could be attributable to differences in exercise intensity or differences in skeletal stress, bone turnover, and calcification between the 2 activities. 27 Future studies should examine differences in atherosclerosis based on exercise intensity within the separate communities of running and cycling.

Clinical Relevance
CAC and atherosclerotic plaques are strongly associated with future risk of cardiovascular events in the general and patient populations. 4,28-30 Although we and others have found an increased prevalence and severity of coronary atherosclerosis in athletes, which was associated with their lifelong exercise volume, athletes have a superior life expectancy compared with the general popu-lation. 31,32 This may partially be explained by beneficial coronary adaptations to exercise and lower risk plaque morphology. 7 The effect of very vigorous intensity exercise may mimic the effect of statins on CAC and plaque composition. Statins increase plaque calcification in some studies, 33 but decrease atheroma volume and cardiovascular risk. 34 Nevertheless, individuals in the MESA study with CAC scores >1000 had risk for a major adverse cardiovascular event of 3.4 per 100 person-years, similar to that of a secondary prevention population, 35 and 6% of our population had a CAC score >1000 on their MARC-2 evaluation. CAC scores that increase from <100 to >100 are associated with increased CVD mortality in athletes, 10 even though the absolute risk of similar CAC scores is lower with increasing physical activity 10 and cardiorespiratory fitness. 36

Limitations
The strengths of our study are its longitudinal design and its assessment of CAC scoring and CCTA in a large cohort of athletes. Its limitations are that participants were aware of their MARC-1 results and could have altered their lifestyle, exercise and medication regimens. Others previously showed that CAC scanning impacts risk factor control and that higher baseline CAC scores are associated with better subsequent risk factor control, 37 which may have diluted the effect of exercise. MARC-1 participants with high CAC scores and/or significant stenoses were advised to seek medical follow-up. The prevalence of statin use was increased in MARC-2, and since statins may increase CAC this could have affected the MARC-2 results. This is, however, unlikely because sensitivity analyses in MARC-2 excluding statin users did not alter the results (data not shown). Both MARC-1 and this follow-up study only included men, 99% of whom identified as White. We only included men in MARC-1 because of their higher probability of coronary atherosclerosis and risk for exercise-related cardiac arrest 19,38 ; therefore, our results can be applied only to middle-aged and older White men.

Conclusions
Exercise intensity, but not volume, is associated with progression of coronary atherosclerosis in middle-aged and older male athletes during 6 years of follow-up. Progression occurred both in those exercising vigorously and those exercising very vigorously, however vigorous exer-cise was associated with less CAC progression, whereas very vigorous exercise was associated with greater CAC progression. This acceleration in CAC may reflect increased plaque calcification. Future studies assessing the CVD risk associated with coronary atherosclerosis in athletes are required to establish the clinical implications of these findings.

Figure 4. Association between exercise volume and intensity and plaque progression.
Association with any plaque (A and B), calcified plaque (C and D), mixed plaque (E and F), and noncalcified plaque (G and H). Plaque progression was dichotomized and defined as yes or no. Adjusted odds ratios from multivariable logistic regression analyses are reported with 95% CIs, with odds ratios <1 indicating less progression. The models are adjusted for the following baseline confounders: age, body mass index, systolic blood pressure, use of antihypertensive medication, pack-years smoked, total cholesterol, family history of coronary heart disease, use of statin, diabetes, time between computed tomography scans, and baseline number of plaques. The exercise intensity analyses are additionally adjusted for exercise volume during follow-up. MET indicates metabolic equivalent of task.
JA Wiley Publishing, J&J, Abbvie, Abbott, Medtronic, Serapta and Cryolife. B.K.V. reports serving as a regular speaker for Philips Healthcare. All relationships are modest. The other authors report no conflicts.