Author + information
- Received April 11, 2016
- Revision received July 7, 2016
- Accepted July 29, 2016
- Published online December 1, 2016.
- Jeffrey Lin, MDa,
- Francis Wang, MDb,
- Rory B. Weiner, MDa,b,
- James R. DeLuca, BAa,
- Meagan M. Wasfy, MDa,
- Brant Berkstresser, MS, ATCb,
- Gregory D. Lewis, MDa,
- Adolph M. Hutter Jr., MDa,
- Michael H. Picard, MDa and
- Aaron L. Baggish, MDa,b,∗ ()
- aCardiovascular Performance Program, Massachusetts General Hospital, Boston, Massachusetts
- bHarvard University Health Services, Cambridge, Massachusetts
- ↵∗Reprint requests and correspondence:
Dr. Aaron L. Baggish, Cardiovascular Performance Program, Massachusetts General Hospital, 55 Fruit Street, Yawkey Suite 5B, Boston, Massachusetts 02114.
Objectives This study sought to determine the relationships among American-style football (ASF) participation, acquired left ventricular (LV) hypertrophy, and LV systolic function as assessed using contemporary echocardiographic parameters.
Background Participation in ASF has been associated with development of hypertension and LV hypertrophy. To what degree these processes impact LV function is unknown.
Methods This was a prospective, longitudinal cohort study evaluating National Collegiate Athletic Association Division I football athletes stratified by field position (linemen: n = 30; vs. nonlinemen, n = 57) before and after a single competitive season, using transthoracic echocardiography. LV systolic function was measured using complementary parameters of global longitudinal strain (GLS) (using 2-dimensional speckle-tracking) and ejection fraction (EF) (2-dimensional biplane).
Results ASF participation was associated with field position-specific increases in systolic blood pressure (SBP) (a Δ SBP of 10 ± 8 mm Hg in linemen vs. a Δ SBP of 3 ± 7 mm Hg in nonlinemen; p < 0.001) and an overall increase in incident LV hypertrophy (pre-season = 8% vs. post-season = 25%, p < 0.05). Linemen who developed LV hypertrophy had concentric geometry (9 of 11 [82%]) with decreased GLS (Δ = −1.1%; p < 0.001), whereas nonlinemen demonstrated eccentric LV hypertrophy (8 of 10 [80%]) with increased GLS (Δ = +1.4%; p < 0.001). In contrast, LV ejection fraction in the total cohort, stratified by field position, was not significantly affected by ASF participation. Among the total cohort, lineman field position, post-season weight, SBP, average LV wall thickness, and relative wall thickness were all independent predictors of post-season GLS.
Conclusions ASF participation at a lineman field position may lead to a form of sport-related myocardial remodeling that is pathologic rather than adaptive. Future study will be required to determine if targeted efforts to control blood pressure, minimize weight gain, and to include an element of aerobic conditioning in this subset of athletes may attenuate this process and translate into tangible downstream health benefits.
Hypertension affects approximately 30% of adults in the United States and is associated with premature cardiovascular disease, including myocardial infarction, stroke, arrhythmias, and heart failure (1,2). Recently, it was demonstrated that hypertension in early life is a strong independent predictor of cardiovascular disease in later life and premature mortality (3,4). Additionally, elevated blood pressure in adolescent athletes predicts sustained hypertension 1 year later (5). As such, the identification and management of hypertension in at-risk younger people may improve long-term cardiovascular health.
American-style football (ASF) is a popular sport, with more than 5 million annual participants in the United States (6,7). Prior cross-sectional studies suggest an association between ASF participation and hypertension (8–11). Recent longitudinal work from our group demonstrated that participation in a single season of ASF is associated with increased systolic blood pressure (SBP) among linemen and that increases in blood pressure are associated with the development of concentric left ventricular (LV) hypertrophy (12). These findings are relevant in the context of prior data from the National Institute for Occupational Safety and Health which demonstrated risk of cardiovascular mortality among former professional linemen was increased compared to that in the general population, but provided no insights regarding causal mechanisms (13).
Sustained exposure to elevated blood pressure results in end-organ damage including hypertensive heart disease. Structurally, hypertensive heart disease is characterized by concentric LV hypertrophy, which is defined by an increase in LV mass and wall thickness without concomitant enlargement in LV cavity size (14). Functionally, patients with early stage hypertensive concentric LV hypertrophy typically demonstrate preserved LV ejection fraction (15–17) but often display subclinical systolic dysfunction as measured by global longitudinal strain (GLS), a well-validated echocardiographic parameter of myocardial systolic function with attendant prognostic implications (18,19).
We therefore undertook the present study with the following goals. First, we aimed to validate our prior observation that ASF participation is accompanied by an increase in blood pressure and LV hypertrophy. Second, we aimed to examine the functional correlates of ASF-induced LV hypertrophy in an attempt to delineate the functional implications of this form of “athlete’s heart.”
We used a prospective, longitudinal design to examine blood pressure, LV remodeling, and LV systolic function among competitive ASF participants at a National Collegiate Athletic Association Division 1 university. We enrolled varsity-level ASF athletes as part of the Harvard Athlete Initiative, an ongoing research program designed to address issues relevant to athletes’ health and exercise physiology. Clinical and anthropometric data related to blood pressure including age (years), height (cm), weight (kg), ethnicity (black vs. nonblack), and personal and family history of hypertension were obtained (20). Body surface area (BSA) was calculated using the Mosteller formula (21), and body mass index (BMI) was calculated using the following equation [weight (kg)/(height (m))2]. Resting blood pressure, heart rate, LV structural parameters, and indices of LV systolic function were assessed before and after a single season of ASF participation, as detailed below. Each participant provided written consent prior to enrollment, and all aspects of the study were approved by the Partners Human Research Committee.
ASF athletes were eligible to participate if they were ≥18 years of age and were recruited members of the Harvard University varsity football team. For this study, we enrolled first-year university athletes to ensure data capture of their initial season of college-level participation. The study period began at the time of enrollment and lasted for the entire ASF season (∼90 days). Field position for each participant was classified as either lineman or nonlineman. Linemen included players at tackle, guard, center, or defensive end positions, whereas nonlinemen included quarterbacks, running backs, wide receivers, tight ends, linebackers, cornerbacks, safeties, kickers, and punters (22). Each participant was tested for performance-enhancing drugs as dictated by National Collegiate Athletic Association standards. Subjects were excluded from the final analysis if they undertook training breaks of ≥3 days for any reason or if their echocardiographic images were unsuitable for GLS analysis.
Blood pressure measurement
A cardiologist (J.L., A.B., R.W., or M.W.) measured blood pressures at enrollment (pre-season) and during post-season assessment, using a manual sphygmomanometer and an appropriately sized arm cuff. Participants were supine for at least 10 min of quiet rest for this process. Reported values represent the average of triplicate measurements. Blood pressure was classified according to Joint National Commission 7th Report (JNC-7) guidelines into the following categories: normal, defined as SBP <120 mm Hg and SBP <80 mm Hg; pre-hypertension, defined as SBP of 120 to 139 mm Hg or diastolic blood pressure (DBP) of 80 to 89 mm Hg; and stage 1 or greater hypertension, defined as SBP ≥140 mm Hg or DBP ≥90 mm Hg (23). Joint National Commission 8th Report (JNC-8) guidelines were not used because they do not define blood pressure thresholds for pre-hypertension. The JNC-8 report does recommend pharmacologic therapy at a blood pressure of 140/90 mm Hg (24), which is consistent with the definition of hypertension used in this study.
Cardiac imaging was performed using a commercially available echocardiography system (Vivid-I, GE Healthcare, Milwaukee, Wisconsin) with a 1.9- to 3.8-mHz phased array transducer. Images were obtained from examinations performed by a sonographer credentialed in cardiac ultrasonography or a cardiologist trained in echocardiography. Participants were imaged at rest ≥12 h after the most recent training session at Harvard University during the pre-season and post-season. The post-season images were acquired within 5 days of the completion of the final game of the season. Two-dimensional imaging was performed from standard parasternal and apical transducer positions with frame rates confined to 60 to 100 Hz as determined on an individual basis for image optimization. All data were stored digitally, and 2 cardiologists (J.L., A.B) blinded to study time point performed all off-line analyses (EchoPac version 7.0 software, GE Healthcare).
Cardiac structural measurements were made in accordance with current guidelines (25). LV volumes and ejection fraction were measured and calculated using the modified Simpson biplane technique. LV mass was calculated using the area-length method, which was chosen because it accounts for LV morphology in both short- and long-axis dimensions, and LV hypertrophy was defined as LV mass index >102 g/m2 (25). Average LV wall thickness was calculated as follows: [interventricular septal thickness (mm) + posterior wall thickness (mm)/2]. Relative wall thickness was calculated as: [interventricular septal thickness (mm) + posterior wall thickness (mm)/LV end-diastolic diameter (mm)]. Concentric LV hypertrophy was defined as a relative wall thickness of >0.42 with an LV mass index of >102 g/m2; eccentric LV hypertrophy was defined as a relative wall thickness of ≤0.42 and an LV mass index of >102 g/m2 (25).
GLS measurements were made using commercially available speckle-tracking analysis software (EchoPac, version 7.0, GE Healthcare) as previously reported (26–28). Briefly, the highest quality digital 2-dimensional apical 4-chamber view was selected for analysis. The endocardium was traced, and a full-thickness myocardial region of interest was selected. The software then automatically partitioned the LV into 6 segments including apical (n = 2), septal (n = 2), and lateral (n = 2) territories and selected suitable speckles for tracking. The reliability of tracking was confirmed by the software’s reliability parameter (V = valid tracking; X = unacceptable tracking). When the software signaled poor tracking efficiency, the observer readjusted the endocardial trace and/or region of interest width until an acceptable tracking score was obtained. By convention, GLS values are presented as negative numbers with lower (i.e., more negative) values representing greater systolic shortening. Measurements were obtained from 3 consecutive cardiac cycles, and reported values represent a 3-cycle average. Participants in whom full 6-segment strain data over 3 consecutive cardiac cycles could not be obtained were excluded from final analyses.
The intraobserver and interobserver variability levels for LV mass and GLS were examined. Intraobserver variability was performed by a single investigator through blinded assessment of 10 randomly selected subjects on 2 separate occasions. Interobserver variability was assessed in a group of 10 randomly selected subjects by 2 investigators blinded to each other’s measurements and to study time point. Correlation coefficients for each measurement derived from simple linear regression were used to quantify variability with the following results: intraobserver LV mass: R2 = 0.946; intraobserver GLS: R2 = 0.968; interobserver LV mass: R2 = 0.921; and interobserver GLS: R2 = 0.972.
Categorical variables are presented as proportions and continuous variables as mean ± SD. Paired data were compared using paired Student t tests or McNemar exact test, whereas unpaired data were compared using Student independent sample t test or Fisher exact test as appropriate for data distribution. Correlation analyses were performed using the Spearman or Pearson technique as appropriate for data distribution. Univariate linear regression analyses were used to identify factors associated with post-season GLS. Parameters with a univariate p value of <0.10 were included in a forward stepwise multivariate linear regression model designed to identify factors independently associated with post-season GLS. All analyses were performed with SPSS version 22.0 software (IBM, Armonk, New York).
Among 190 eligible ASF participants enrolled from 2008 to 2014, 87 (all men, 18.8 ± 0.8 years of age) were included in this analysis. Subject attrition (n = 103) was due to intraseasonal injuries (25 of 103 athletes [24%]) and our strict a priori imaging quality exclusion criteria (78 of 103 athletes [76%]). Of note, there were no significant differences between field position (34% linemen and 66% nonlinemen), blood pressure trends, or LV structural parameters of those players who were excluded and those retained. The final cohort included 30 linemen (height: 191 ± 5 cm; weight: 111 ± 12 kg) and 57 nonlinemen (height: 185 ± 6 cm; weight: 90 ± 9 kg). Linemen had significantly higher BMI than nonlinemen (30.6 ± 3.5 kg/m2 vs. 26.4 ± 2.0 kg/m2; p < 0.001). Twenty-four athletes (27%) were of black ethnicity, and 22 athletes (25%) had a family history of hypertension. Weight (111 ± 12 to 114 ± 12 kg; p = 0.002), BMI (30.6 ± 3.5 to 31.2 ± 3.5 kg/m2; p = 0.007), and BSA (2.42 ± 0.13 to 2.44 ± 0.13 m2; p = 0.01) all increased significantly in the linemen group from baseline to post-season whereas no significant changes occurred among the nonlinemen (Table 1).
ASF participation was associated with an increase in SBP in the total cohort (119 ± 10 to 124 ± 12 mm Hg; p < 0.001) but no significant change in DBP. Although both the linemen (pre-season: 122 ± 8 mm Hg vs. post-season: 132 ± 10 mm Hg; p < 0.001) and the nonlinemen (pre-season: 117 ± 11 mm Hg vs. post-season: 120 ± 13 mm Hg; p < 0.001) (Table 1) had increases in SBP, the magnitude of increase was significantly greater among linemen (linemen: Δ SPB = 10 ± 8 mm Hg vs. Δ SBP = 3 ± 7 mm Hg for nonlinemen; p < 0.001) (Figure 1A). At pre-season, the prevalence of pre-hypertension was similar between linemen (17 of 30 [57%]) and nonlinemen (29 of 57 [51%]; p = 0.31), and no athletes met criteria for overt hypertension. At post-season assessment, 90% of linemen met criteria for pre-hypertension (18 of 30 [60%]) or stage 1 hypertension (9 of 30 [30%]), whereas only 49% of linemen (28 of 57), a proportion similar to the pre-season, had blood pressure exceeding upper limits of normal (Figure 1B).
Left ventricular structure
LV mass index did not differ significantly at pre-season as a function of field position (linemen: 97 ± 15 g/m2 vs. 94 ± 12 g/m2 for nonlinemen; p = 0.10) (Table 2). Among the total cohort, ASF participation led to a significant increase in LV mass index (pre-season: 95 ± 13 g/m2 vs. 104 ± 15 g/m2 during post-season; p < 0.001), but LV mass index increased more among linemen (Δ in linemen: 13 ± 15 g/m2 vs. Δ of 7 ± 10 g/m2 in nonlinemen; p = 0.02). In the total cohort, the prevalence of LV hypertrophy increased from 8% (7 of 87) at pre-season to 24% (21 of 87; p < 0.001) at post-season. Among linemen with LV hypertrophy at post-season, 9 of 11 (82%) demonstrated concentric geometry, whereas, 8 of 10 nonlinemen (80%) with LV hypertrophy demonstrated eccentric geometry.
Left ventricular systolic function
LV ejection fraction was similar among linemen (59 ± 6%, range 49% to 69%) and nonlinemen (59 ± 6%, range 46% to 79%) at pre-season (p = 0.93) and was not significantly different at post-season. GLS was also similar among field position groups at pre-season (linemen = −21.2 ± 2.4% vs. −20.8 ± 2.7% in nonlinemen; p = 0.26). However, GLS was significantly changed at post-season in both of the position groups with changes occurring in opposite directions. Specifically, linemen demonstrated a significant decrement in GLS (Δ = 1.1%; p < 0.001), whereas nonlinemen experienced a significant increase (Δ = 1.4%; p < 0.001) (Figure 2). The observed GLS decrement among linemen correlated indirectly with changes in SBP (R = −0.47; p = 0.009) and LV mass index (R = −0.52; p = 0.003), whereas no such relationships were observed among nonlinemen (R = 0.01; p = 0.49; R = 0.09; p = 0.93, respectively). Univariate and multivariate predictors of post-season GLS are shown in Table 3. In a multivariate linear regression model, lineman position (β = 0.402; p < 0.001), post-season weight (β = 0.375; p < 0.001), SBP (β = 0.325; p < 0.001), average LV wall thickness (β = 0.234; p < 0.001), and LV relative wall thickness (β = 0.218; p = 0.003) were independent predictors of post-season GLS.
Key findings from this study are summarized as follows. First, our data confirm previously reported associations between ASF participation, increases in SBP, and development of concentric LV hypertrophy. These observations appear particularly relevant to ASF linemen. Second, we demonstrate that the development of concentric LV hypertrophy among linemen is characterized by concomitant reductions in LV longitudinal strain, a well-validated marker of subclinical LV dysfunction with attendant adverse prognosis. In aggregate, these findings advance our understanding of cardiac remodeling among athletes and challenge the pervasive notion that all sport-related cardiovascular changes are adaptive rather than pathologic.
A 1994 report from the National Institute for Occupational Safety and Health evaluated the medical status of 6,848 retired NFL players (13). While the study documented a 46% decreased rate of death among former NFL players compared to a general population of age-matched men, it showed a 52% increase in cardiovascular mortality among linemen. Subsequent studies have confirmed a similar increase in heart disease mortality in active and former ASF linemen as well as a higher risk of metabolic syndrome and subclinical carotid atherosclerosis (11,29,30). Although elevated body mass has been proposed as a key explanatory factor, it is unlikely this parameter explains the complete mechanistic story (10,31). Data from the current study introduce a more comprehensive mechanistic explanation in the form of acquired, early life hypertension with resultant target organ damage in the form of concentric LV hypertrophy with subclinical LV systolic dysfunction.
Conventional assessment of LV systolic function relies on measurement of LV ejection fraction, a relatively crude volumetric technique that is exquisitely load-dependent, highly reliant on image quality, and based on key geometric assumptions that may not be valid in the setting of significant cardiac remodeling (32). Speckle-tracking echocardiography minimizes these limitations and more accurately captures subtle yet clinically relevant changes in cardiac function. Global longitudinal strain, the systolic function parameter used in this study, has emerged as a highly reproducible metric with powerful prognostic implications in numerous clinical populations. Among patients with hypertensive heart disease and preserved ejection fraction, mean GLS is uniformly decreased and inversely correlated with SBP, LV mass, and the presence of concentric remodeling or hypertrophy (33–35).
Previous studies of GLS among athlete cohorts almost exclusively confined to athletes in endurance sports have demonstrated preservation or increase in contractile function, highlighting the generally adaptive nature of exercise-induced cardiac remodeling. In a study evaluating GLS in rowers, age-matched hypertensive subjects not receiving therapy, and normotensive sedentary controls, the hypertensive group had significantly lower GLS (17.5 ± 2.8%) than that of rowers (−22.2 ± 2.7%) and controls (21.1 ± 2.0%) (36). In addition, prior work by our group documented increases in GLS after 90 days of endurance training among competitive rowers (26). The present study builds on that work in 2 principal ways. First, the finding of LV hypertrophy with GLS augmentation among nonlineman ASF athletes, a group exposed to a mixture of isometric and isotonic physiology (i.e., skills training and aerobic conditioning), suggests an adaptive form of remodeling similar to that seen in other athlete cohorts. Second, in stark comparison, LV hypertrophy with relative impairment of GLS among ASF lineman, a group exposed to predominantly isometric physiology (i.e., blocking and tackling drills), suggests maladaptive remodeling. To our knowledge, this study is the first to document a relative impairment of GLS among athletes and to provide plausible physiologic and anatomic correlates in the forms of acquired hypertension and concentric LV hypertrophy, respectively.
There are discrete clinical and scientific implications of the data presented. First, our data confirm the notion that ASF linemen are at substantial risk for the development of hypertension. Clinicians who care for ASF athletes should be encouraged to screen for hypertension and to implement lifestyle and pharmacologic therapies as suggested by current guidelines. Additionally, this study clearly documents position-specific changes in blood pressure and cardiac structure and function in response to ASF participation. This finding reinforces the notion that not all forms of athletic participation result in similar cardiovascular remodeling and, for the first time, identifies an athletic population that appears to remodel with maladaptive attributes. The cardiovascular changes seen in linemen, namely, concentric LVH with reduced GLS, should be viewed as a unique entity from traditional “athlete’s heart” with pathologic rather than adaptive characteristics.
Several limitations of this study set the stage for future work. First, we used a longitudinal, repeated measures study design in an effort to establish temporal and causal relationships between ASF participation, SBP increase, and myocardial remodeling. This approach, however, does not permit definitive determination of underlying mechanisms. Future studies examining and controlling for factors including inflammation, peripheral vascular dysfunction, distinct cellular pathway activation, and diffuse myocardial fibrosis represent logical next steps. Second, we acknowledge that possible confounding factors known to impact blood pressure including nonsteroidal anti-inflammatory drug use, psychosocial stress level, and dietary intake were not standardized. Whether these factors impacted our data and to what degree efforts to control these factors will benefit ASF athletes remain uncertain. Third, our study duration was relatively brief, as many athletes accrue numerous years of ASF exposure. Carefully designed longer term studies that capture blood pressure and cardiac parameters during more lengthy ASF careers and in the post-ASF years are warranted. Finally, we acknowledge the potential impact of subject attrition due to the unavoidably high rate of participant exclusion from the final analysis (103 of 190 athletes [54%]) due to injury and technically inadequate echocardiographic images for strain analysis. Although there were no differences in baseline parameters, blood pressure trends, or prevalent LV hypertrophy among those excluded, we cannot eliminate the possibility of some attrition bias.
Data from the present study advance our understanding of the cardiovascular response to participation in ASF. ASF participants at the lineman field positions appear to be at risk for incident hypertension and secondary maladaptive cardiac remodeling. The development of concentric LV hypertrophy with relative impairment of GLS, a phenotype unequivocally associated with adverse prognoses in other clinical populations, is concerning in this population of young, otherwise healthy athletes. These findings underscore the need for future work designed to confirm the scope and long-term clinical implications of this problem which may affect millions of young people.
COMPETENCY IN MEDICAL KNOWLEDGE: The cardiovascular care of ASF athletes requires specific consideration of the unique cardiovascular demands during training and competition in this sport. Understanding the effects of football participation on cardiovascular remodeling using echocardiography represents an important step in providing more specialized long-term care for this group of athletes.
TRANSLATIONAL OUTLOOK: Our findings demonstrate that ASF athletes, especially those at a lineman position, are at increased risk for early life hypertension and associated secondary cardiac remodeling including a decrement in left ventricular systolic function as measured by global longitudinal strain. This decrease in function appears to be mediated partly by acquired resting hypertension and weight gain, highlighting the importance of the modifiable factors in mitigating the risk of later life heart disease in football linemen.
The authors thank the Harvard University coaching staff, athletic department administration, and student athletes for their continued support of the Harvard Athlete Initiative.
This work was supported by U.S. National Institutes of Health/National Heart, Lung, and Blood Institute research grant HL RO1 125869 (to Dr. Baggish) and American Heart Association grant FTF220328 (to Dr. Baggish). The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- American-style football
- global longitudinal strain
- left ventricular
- Received April 11, 2016.
- Revision received July 7, 2016.
- Accepted July 29, 2016.
- American College of Cardiology Foundation
- Gray L.,
- Lee I.M.,
- Sesso H.D.,
- Batty G.D.
- Yano Y.,
- Stamler J.,
- Garside D.B.,
- et al.
- ↵U.S. Census. Participation in Selected Sports Activities. 2006. Available at: http://www.allcountries.org/uscensus/435_participation_in_selected_sports_activities.html. Accessed September 12, 2015.
- Irick E.
- Karpinos A.R.,
- Roumie C.L.,
- Nian H.,
- Diamond A.B.,
- Rothman R.L.
- Weiner R.B.,
- Wang F.,
- Isaacs S.K.,
- et al.
- Baron S.,
- Rinsky R.
- Radulescu D.,
- Stoicescu L.,
- Buzdugan E.,
- Donca V.
- Drazner M.H.
- Mentz R.J.,
- Khouri M.G.
- Shah A.M.,
- Solomon S.D.
- Lang R.M.,
- Badano L.P.,
- Mor-Avi V.,
- et al.
- Baggish A.L.,
- Yared K.,
- Wang F.,
- et al.
- Wasfy M.,
- Weiner R.B.,
- Wang F.,
- et al.
- Weiner R.B.,
- Hutter A.M. Jr..,
- Wang F.,
- et al.
- McCarthy C.G.,
- Webb R.C.
- Biering-Sorensen T.,
- Solomon S.D.
- Gonçalves S.,
- Cortez-Dias N.,
- Nunes A.,
- et al.
- Kraigher-Krainer E.,
- Shah A.M.,
- Gupta D.K.,
- et al.
- Narayanan A.,
- Aurigemma G.P.,
- Chinali M.,
- Hill J.C.,
- Meyer T.E.,
- Tighe D.A.