This prospective, longitudinal study examined the effects of participation in team-based exercise training on cardiac structure and function. Competitive endurance athletes (EA, n = 40) and strength athletes (SA, n = 24) were studied with echocardiography at baseline and after 90 days of team training. Left ventricular (LV) mass increased by 11% in EA (116 ± 18 vs. 130 ± 19 g/m2; P < 0.001) and by 12% in SA (115 ± 14 vs. 132 ± 11 g/m2; P < 0.001; P value for the compared Δ = NS). EA experienced LV dilation (end-diastolic volume: 66.6 ± 10.0 vs. 74.7 ± 9.8 ml/m2, Δ = 8.0 ± 4.2 ml/m2; P < 0.001), enhanced diastolic function (lateral E′: 10.9 ± 0.8 vs. 12.4 ± 0.9 cm/s, P < 0.001), and biatrial enlargement, while SA experience LV hypertrophy (posterior wall: 4.5 ± 0.5 vs. 5.2 ± 0.5 mm/m2, P < 0.001) and diminished diastolic function (E′ basal lateral LV: 11.6 ± 1.3 vs. 10.2 ± 1.4 cm/s, P < 0.001). Further, EA experienced right ventricular (RV) dilation (end-diastolic area: 1,460 ± 220 vs. 1,650 ± 200 mm/m2, P < 0.001) coupled with enhanced systolic and diastolic function (E′ basal RV: 10.3 ± 1.5 vs. 11.4 ± 1.7 cm/s, P < 0.001), while SA had no change in RV parameters. We conclude that participation in 90 days of competitive athletics produces significant training-specific changes in cardiac structure and function. EA develop biventricular dilation with enhanced diastolic function, while SA develop isolated, concentric left ventricular hypertrophy with diminished diastolic relaxation.
- exercise physiology
- cardiac remodeling
- athlete's heart
an association between athletic participation and specific cardiac morphology has been well established. Increased left atrial size and left ventricular (LV) mass, wall thickness, and chamber size have been documented among trained athletes, and several recent reports have described right ventricular (RV) characteristics among such individuals (5, 26, 27, 30, 32, 37). Although copious data demonstrate a high prevalence of “abnormal” cardiac measurements among competitive athletes, such cross-sectional data are not sufficient to establish whether athletic training is causal in their development.
Several small prospective studies have reported LV changes in the context of exercise training (8, 9, 11, 21, 35, 36). At present, definitive longitudinal studies defining the LV structural and functional responses to sustained exercise training are lacking, and no such data exist regarding the RV. Further, the relationship between training discipline and cardiac remodeling has not been adequately characterized.
We sought to determine the impact of exercise training during a single season (90 days) of competitive athletics on cardiac structure and function. We hypothesized that significant structural and functional changes would occur and that the nature and magnitude of change would vary with training discipline. To address these hypotheses, we performed pre- and posttraining assessment of competitive university athletes.
University students participating in official competitive athletics affiliated with the Harvard University Department of Athletics participated in this study. University athletes, not elite competitors, were studied as we anticipated that they would enter the study period relatively detrained and would then exercise with a high enough intensity and consistency to maximize our ability to define and quantify training-induced changes. Written informed consent was obtained from all participants before involvement. The Harvard University institutional review board and the Partner's Human Research Committee approved the protocol before study initiation.
Individuals were considered eligible if they were ≥18 yr old and had been previously selected as members of an organized competitive team program. Seventy-five athletes were enrolled before the beginning of the 2006 fall semester training season. To assess for and to compare changes attributable to endurance vs. strength training, athletes from two distinct sporting disciplines were enrolled. The endurance athlete group (EA) consisted of long-distance male rowers (MR) and female rowers (FR), while the strength athlete group (SA) was comprised of American-style male football players (MF).
Height, weight, resting vital signs, medication use, and personal/family medical historical data were recorded at the time of enrollment. Baseline transthoracic echocardiography was performed as detailed below. Training volume during a prestudy period, defined as the 8 wk before baseline assessment, was collected. Prestudy period endurance activity was defined as running, cycling, swimming, rowing, or aerobic machine use at an effort sustainable for ≥20 min, while strength activity was defined as weight lifting, plyometric exercise, and sprint running drills.
The study period began at the time of enrollment and lasted for 90 days. No effort was made to control training regimens during the study period, as the goal of this study was to examine the effects of participation in actual organized athletics. However, daily data were recorded on the duration and the type (endurance vs. strength) of training activities performed during the study period. EA performed rowing training aimed to optimize performance at a 5-km distance that consisted of long-duration open water and indoor ergometer sessions (1–3 h) at low stroke rates (20–24 stokes/min). Intermittent heart rate monitor use was encouraged to familiarize EA with an effort corresponding to a zone of 70–80% of maximum predicted heart rate. SA participated in team practice sessions consisting of plyometric exercises, sprint training, and tackling drills. In addition, all SA performed three to four weight training sessions per week under the supervision of a dedicated strength coach each consisting of three sets of 8–10 large-muscle group exercises (squats, dead-lift, bench press, etc.) with weights targeted to allow four to six repetitions per set. Training volumes during the prestudy and study period were characterized by total number of training hours per week and the hours per week dedicated to either strength or endurance activities. Identical data collection and echocardiography were repeated after the 90-day period of organized team training activity.
All potential subjects were questioned confidentially about anabolic steroid use and were excluded if a history of use was elicited. Individuals were excluded from the final data analysis if they undertook any breaks in training of ≥3 days during the study period.
Echocardiography was performed using a commercially available system (Vivid-I, GE Healthcare, Milwaukee, WI) with a 1.9- to 3.8-mHz phased-array transducer. Images were obtained after 20 min of quiet rest between 2 and 5 PM and were separated from the previous training session by ≥24 h. Two-dimensional, pulsed-Doppler, and color tissue-Doppler imaging from standard parasternal, apical, and subcostal positions were performed. The two-dimensional frame rate was 25–75/s, and the tissue Doppler frame rate was >100/s for all images. Echocardiography was performed by two trained sonographers, and each sonographer performed both baseline and poststudy imaging on the same athletes. All data were stored digitally, and poststudy off-line data analysis (EchoPac, version 6.5, GE Healthcare) was performed by one of two study cardiologists (A. L. Baggish, M. J. Wood) blinded to the study time point. Definitions of normalcy were adopted from the most recent American Society of Echocardiography guidelines (14).
LV ejection fraction, end-diastolic volume, and end-systolic volume were calculated using the modified Simpson's technique. RV fractional area change, a validated index of RV function, was calculated by outlining the endocardial borders of the RV in diastole and systole in the apical four-chamber view and calculating the difference between the two areas expressed as a percentage of end-diastolic RV area (1). LV mass was calculated using the area-length method. Relative wall thickness was defined as [interventricular septal thickness (mm) + posterior wall thickness (mm)]/LV internal end-diastolic diameter (mm). Left atrial volume was calculated using the biplane area-length method while right atrial major axis measurement was made in apical four-chamber view. Longitudinal tissue velocities were measured off-line from two-dimensional color-coded tissue Doppler images and reported as the average of three consecutive cardiac cycles. Resting heart rates were obtained from the final loop of each study. Cardiac output was determined by calculating the product of stroke volume and heart rate. Body surface area (BSA) was calculated using the Mosteller formula, and all measurements are presented both as raw data and adjusted for BSA when appropriate (20).
Measurements are presented as means ± SD. Comparison of baseline measurements for the three groups (MR, FR, and MF) was performed with ANOVA with Bonferroni correction for multiple comparisons. The paired t-test and the Wilcoxon matched-pair test were used to assess the significance of interval measurements within athlete groups. Correlation analysis was performed using the Spearman and Pearson's method as appropriate for data distribution. A P value of <0.05 was considered significant.
Baseline characteristics and training regimens.
We enrolled 75 athletes (MR = 20, FR = 20, MF = 35). Sixty-four individuals completed the training period and were included in the final analysis. All subject attrition (n = 11) was due to MF musculoskeletal injuries. Baseline characteristics are reported in Table 1. The mean age was 19.3 ± 0.9 yr with no significant differences among the three groups. Female athletes had lower height, weight, and BSA than male athletes. Male rowers had lower weight and BSA than male football players. Resting heart rate and diastolic blood pressure was significantly lower in both male and female rowers compared with football players, while systolic blood pressure did not vary among the groups. Two individuals (2/64, 3%) reported a history of hypertension, but neither individual was receiving medication for this entity. Noncardiovascular medication use was common (19/64, 30%) and included bronchodilators, oral contraceptives, and topical acne agents. A family history of hypertension was reported in 28% (18/64) of individuals.
Data characterizing the prestudy period and the study period are shown in Table 1. Aggregate weekly training hours during both periods were similar among the three groups. Prestudy and study period MR and FR training was comprised largely of endurance training, while MF spent the majority of time engaged in strength-training activities. Total training hours increased from prestudy to study period by a similar magnitude among the three groups. During the study period, MR and FR performed 1.5–2.5 daily hours of sustained rowing at effort targeted to be well below anaerobic threshold. In contrast SA participated in daily practice sessions consisting of repetitive burst activity in the form of short sprints, tackling drills, and weight lifting. Athletes in both groups engaged in organized sessions ≥5 days/wk through the entire study period.
Baseline echocardiographic parameters.
Baseline echocardiographic measurements are detailed in Table 2. All linear and volumetric measurements were significantly lower among FR compared with both MR and MF. After adjustment for BSA, MR and FR were well matched for all measurements with the exception of LV mass (FR = 103 ± 9 g/m2 vs. MR = 130 ± 14 g/m2, P < 0.001). In contrast, MF had smaller adjusted right and left atria, LV wall thicknesses, and LV volumes than rowers of both sexes. The majority of athletes in all groups had baseline LV mass in excess of proposed normal values (men < 102 g/m2; women < 88 g/m2); however, relative wall thickness was similar and within the normal range of limits in all athlete groups at baseline (FR = 0.39 ± 0.03, MR= 0.38 ± 0.03, MF = 0.38 ± 0.04). Aside from higher septal and lateral LV wall A1 velocities in MF, there were no significant differences in systolic or diastolic tissue velocities among the three groups at baseline.
Cardiovascular changes in EAs.
There were no significant differences among the mean changes of any parameters between MR and FR, and subsequent data labeled as “endurance athletes (EA)” include both sexes. Resting heart rate (60 ± 5 vs. 50 ± 6 beats/min, Δ = 10 ± 11 beats/min; P = 0.01) and diastolic blood pressure (56 ± 5 vs. 52 ± 3 mmHg, Δ = 4 ± 3 mmHg; P = 0.004) were reduced at the completion of the study period, while height, weight, and systolic blood pressure were unchanged.
EA experienced significant increases in left atrial volume (28.9 ± 5.7 vs. 31.3 ± 6.2 ml/m2, Δ = 2.4 ± 1.5 ml/m2; P < 0.001) and right atrial major dimension (24.9 ± 1.8 vs. 26.4 ± 1.9 mm/m2, Δ = 1.5 ± 0.7 mm/m2; P < 0.001). Measurements of LV structure, including internal end-diastolic diameter (24.9 ± 1.7 vs. 26.3 ± 1.8 mm/m2, Δ = 1.3 ± 0.8 mm/m2; P < 0.001), end-diastolic volume (67 ± 10 vs. 75 ± 10 ml/m2, Δ = 8 ± 4 ml/m2; P < 0.001), end-systolic volume (25 ± 4 vs. 30 ± 4 ml/m2, Δ = 5 ± 3 ml/m2; P < 0.001), and mass (116 ± 14 vs. 130 ± 16 g/m2, Δ = 14 ± 9 g/m2; P < 0.001), increased significantly. While LV ejection fraction [62 ± 5% vs. 59 ± 5%, Δ = 3 ± 5%, P = not significant (NS)] was not significantly changed, resting cardiac output was significantly decreased following training (2,490 ± 590 vs. 2,240 ± 690 l−1·min−1·m−2, P = 0.005). There were nonsignificant trends toward increased LV posterior wall thickness (5.0 ± 0.4 vs. 5.2 ± 0.4 mm/m2, Δ = 0.2 ± 0.3 mm/m2; P = NS) and increased interventricular septal thickness (4.7 ± 0.5 vs. 5.0 ± 0.5 mm/m2, Δ = 0.3 ± 0.4 mm/m2; P = NS); however, relative wall thickness was unchanged (0.39 ± 0.02 vs. 0.40 ± 0.02, Δ = 0.02 ± 0.02; P = NS). There was a significant increase in peak systolic and both early and late peak diastolic tissue velocities measured at both the lateral wall and the interventricular septum (Table 3).
RV end-diastolic annular diameter (22.1 ± 2.3 vs. 23.6 ± 2.5 mm/m2, Δ = 1.4 ± 0.8; P < 0.001) and end-diastolic area (1,460 ± 220 vs. 1,650 ± 200 mm2/m2, Δ = 190 ± 150 mm2/m2; P < 0.001) increased. This was accompanied by a reduction in RV end-systolic area (1,000 ± 150 vs. 840 ± 100 mm2/m2, Δ = 170 ± 120 mm2/m2; P < 0.001) and a resultant rise in fractional area change (34 ± 9% vs. 46 ± 5%, Δ = 12 ± 8%, P < 0.001). RV peak systolic, peak early-diastolic, and peak late-diastolic tissue velocities increased significantly (Table 3).
Cardiovascular changes in SAs.
Systolic blood pressure (115 ± 11 vs. 122 ± 5 mmHg, P = 0.01) was higher after the training period, while diastolic blood pressure, resting heart rate, height, and weight were unchanged.
There were no significant changes in left atrial volume or right atrial major dimension. LV posterior wall thickness (4.5 ± 0.5 vs. 5.2 ± 0.5 mm/m2, Δ = 0.7 ± 0.4 mm/m2; P < 0.001) and the interventricular septal thickness (4.2 ± 0.5 vs. 5.0 ± 0.5 mm/m2, Δ = 0.7 ± 0.3 mm/m2; P < 0.001) increased significantly. LV mass (115 ± 13 vs. 132 ± 11 g/m2, Δ = 17 ± 9 g/m2; P < 0.001) and relative wall thickness (0.38 ± 0.03 vs. 0.44 ± 0.04, Δ = 0.06 ± 0.02; P < 0.001) also increased during the study period. LV linear measurements, volume, ejection fraction, and cardiac output were unchanged. However, early and late LV diastolic tissue velocities were all significantly reduced. Right ventricular parameters, including end-diastolic annular diameter, end-diastolic area, end-systolic area, fractional area change, and tissue Doppler velocities, were unchanged.
Differential LV remodeling: correlations between size, mass, and function.
We sought to further characterize the LV mass increase that was observed in both endurance and strength athlete groups. Correlations between change in LV mass and change in LV end-diastolic volume, aggregate LV wall thickness (defined as interventricular septal thickness change plus posterior wall thickness change), and LV early diastolic tissue velocity were performed. The results are displayed in Figs. 1, 2, and 3. With EA there was a significant positive correlation between the increase in LV mass and the increase in LV end-diastolic volume (r2 = 0.82, Fig. 1A). No such correlation existed in SA (r2 = 0.02, Fig. 1B). In contrast, change in LV mass was highly although negatively correlated with change in aggregate LV wall thickness in SA (r2 = 0.67) but not in EA (r2 = 0.00, Fig. 2, A and B). Change in lateral LV wall early-diastolic tissue velocity correlated with change in LV mass in both groups (EA r2 = 0.90; SA r2 = 0.79). However, this correlation was positive in EA and negative in SA (Fig. 3, A and B).
We prospectively studied EA and SA participating in organized team training. This study provides longitudinal data demonstrating significant training-specific changes in cardiac structure and function during a 3-mo period of exercise training. In summary, we observed a significant increase in LV mass in both EA and SA. Although this increase in LV mass was of similar magnitude in both groups, it was associated with distinct training-specific structural and function profiles. EA, performing daily sustained aerobic exercise, experienced LV dilation, enhanced LV diastolic function, biatrial enlargement, and RV dilation with increased systolic and diastolic function. In contrast, SA, performing repetitive short burst-type power exercises, experienced concentric LV hypertrophy, a reduction in LV diastolic function, and demonstrated no changes in atrial dimensions or RV parameters.
Previous cross-sectional studies have documented a high prevalence of LV dilation among endurance-trained athletes (19, 25, 34). Several small-scale longitudinal studies suggest that such dilation may be a direct consequence of exercise training, as interval increases in LV major dimension have been documented among recreational male joggers, elite distance runners, rowers, swimmers, and cyclists (9, 11, 15, 35, 36). Before this report, longitudinal data integrating LV structure and function in EA have been limited. duManoir et al. (8) recently reported significant increases in resting LV cavity dimensions, stroke area, and mass after 10 wk of 3 times weekly rowing sessions. As noted by the authors, athletes in this study performed combined strength and endurance training that differs from the purer endurance-type rowing training to which athletes in our study were subjected. One prior study documented changes in transmitral and pulmonary vein Doppler flow patterns consistent with enhanced LV relaxation among elite rowers after 3 mo of training (21). Although this conclusion is consistent with ours, it is based solely on highly vascular-volume dependent indexes of LV diastolic function that have been largely replaced by the tissue Doppler myocardial velocity interrogation techniques used in the present study.
EA in the present study experienced significant increases in LV mass and end-diastolic volume and a nonsignificant trend toward increased wall thickness. The increase in LV mass correlated tightly with the increase in LV end-diastolic volume but not with changes in wall thickness. Although LV stroke volume was increased following training, resting cardiac output, due to a marked reduction in resting heart rate, was significantly reduced. This finding deserves further study. In addition, EA experienced significant increases in LV diastolic tissue velocities, which were positively correlated with the magnitude of LV mass increase. In aggregate, our data demonstrate direct evidence that endurance training produces LV dilation with accompanying augmentation of LV diastolic function.
No previous longitudinal data describing the impact of strength training on LV parameters have been reported, and cross-sectional data have been conflicting (6, 7, 10, 12, 13, 29, 31). In this study, SA experienced LV mass increase, an increase in LV wall thickness, but no change in LV chamber dimensions, volume, or resting cardiac output. The change in LV mass was positively correlated with the change in LV wall thickness but was unrelated to LV chamber volume. In addition, SA experienced a significant reduction in LV diastolic tissue velocities, which were highly although negatively correlated with rising LV mass. These data suggest that the LV undergoes significant concentric hypertrophy with resultant reduction of diastolic function when exposed to a sustained period of strength training.
The effects of single exercise events on RV function have been reported, and several cross-sectional studies of RV parameters among trained athletes have been described (22, 23). Scharhag and colleagues (32) performed MRI on 21 male endurance athletes and 21 controls. They detected higher RV mass, end-diastolic volume, and stroke volume among athletes compared with controls but similar RV ejection fractions and RV-to-LV measurement ratios between the two groups. An echocardiographic assessment of RV function comparing short- and long-distance swimmers reported higher early-diastolic RV tissue velocities among swimmers engaging in long-distance training than in those performing sprint work (4). We present novel longitudinal data describing the RV response to athletic training. EA experienced RV dilation and a significant accentuation of function as measured by fractional area change, peak systolic tissue velocity, and peak early- and late-diastolic tissue velocity. In contrast, SA exhibited no significant changes in RV dimensions or systolic function and a trend toward reduction in RV diastolic function.
The complex physiology that underlies the training modalities studied in this report remains an area of active research, and as such, mechanistic explanations for the observed changes remain speculative. Sustained elevation in cardiac output underlies endurance training and must be handled by both the right and left sides of the heart. It is plausible that this repetitive volume challenge accounts for the biatrial enlargement and the biventricular structural and functional changes we observed in EA. In contrast, strength training-induced change appeared to be confined to the LV. Several authors have demonstrated that strength-training activities produce significant increases in systemic blood pressure with attendant alterations in LV systolic function (17, 18). In addition, aortic root remodeling and arterial compliance reduction have been demonstrated among power athletes (2, 24). Finally, recent animal study data demonstrate that transient elevation in LV afterload leads to LV diastolic dysfunction in otherwise healthy hearts (16). In aggregate, these studies provide strong evidence that repetitive power activity has the capacity to alter structure and function of the heart and proximal arteries. We propose that daily sessions of strength training, accompanied both by transient increases in LV afterload and by small but significant increases in resting systolic blood pressure, are responsible for the isolated concentric LV hypertrophy and reduced diastolic LV function observed in SA. The relative contributions to these changes of the dramatic but transient intraexercise systemic hypertension and the subtle yet persistent resting blood pressure elevation remain uncertain. Further characterization of exercise hemodynamics and precise characterization of the underlying cellular and molecular mechanisms are warranted.
There are several important implications of our findings. First, we present conclusive evidence that vigorous exercise training produces significant cardiac remodeling. Our data strongly refute the notion that individuals with “athlete's heart” morphology are simply more likely to rise to the competitive ranks of sport and that their training endeavors are at least in part directly responsible for their cardiac morphology. Second, we have demonstrated that the type of exercise training (endurance vs. strength challenge) determines the specific characteristics of the cardiac remodeling. This finding may ultimately prove relevant to the many individuals with intrinsic heart disease who engage in exercise for recreation or rehabilitation and should be considered during the design of future therapeutic exercise trials. Further study regarding the impact of specific exercise modalities on health outcomes in such populations is warranted. Finally, it is noteworthy that we detected no individuals with cardiac measurements similar to those observed in pathological cardiomyopathies and among previously studied cohorts of elite athletes (3, 25, 28, 33). This likely reflects our decision to study highly competitive although not elite athletes. We chose to study university athletes as we anticipated that they would enter the study period relatively detrained and would then exercise with a high enough intensity and consistency to maximize our ability to define and quantify training-induced changes. It is possible that the changes we observed would continue to evolve during further sustained high-intensity training, eventually resulting in the overtly abnormal measurements sometimes encountered among elite competitors. Further study will be required to establish with certainty a precise dose-response relationship and the implications of training-induced cardiac remodeling with respect to long-term health outcomes.
This study was designed to determine the effects of real-world athletic participation on cardiac parameters, and thus we made no effort to control the training regimens to which they were subjected. This element of our study design introduces several potential limitations. First, athletes entered the organized study period having performed some amount of individualized training during the 8 wk prior and, although age matched, may have differed in the number of previous years of formal athletic training. The nature of this prestudy training variability and differences in lifelong training exposure may explain some differences in baseline measurements and may have reduced the amount of potential change observable during the actual study period. Second, while the vast majority of training performed by EA involved sustained aerobic conditioning characterized by high heart rate/high cardiac output while SA were largely exposed to brief, repetitive burst activity resulting in LV pressure challenge, athletes in both groups were exposed to small amounts of crossover training stimuli, possibly reducing the magnitude and sport specificity of observed change. However, both of these limitations are simultaneously strengths as they allow characterization of the cardiac response to genuine athletic participation. Next, although all enrolled individuals were questioned about anabolic steroids and excluded if a history of use was reported, we did not perform serum assessment for relevant agents and cannot exclude that this factored into our observations. In addition, body composition has recently been shown to be an important consideration in the assessment of cardiac morphology among athletes. Although we did not observe any significant training-induced changes in body mass, we cannot exclude the possibility that baseline differences among groups or training-induced changes in the ratio of adipose to lean muscle mass factored into our observations. Finally, there are currently no female football players available for study. As such, we cannot comment on the female response to such strength training.
We employed a prospective and longitudinal study design to determine if participation in organized team athletics produces changes in cardiac parameters. We present data demonstrating training-specific cardiac changes over the course of a 3-mo season of competitive athletics. EA experienced biatrial enlargement and biventricular dilation with an accompanying enhancement in biventricular function. In contrast, SA demonstrated concentric LV hypertrophy with diminished LV diastolic function and experienced no changes in atrial dimensions or RV structure and function. The longitudinal data from this study provide convincing evidence that athletic training has a causal role in the development of training-specific profiles of cardiac structure and function.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society