Concern exists about cardiovascular disease (CVD) in professional football players. We examined whether playing position and size influence CVD mortality in 3,439 National Football League players with ≥5 pension-credited playing seasons from 1959 to 1988. Standardized mortality ratios (SMRs) compared player mortality through 2007 to the United States population of men stratified by age, race, and calendar year. Cox proportional hazards models evaluated associations of playing-time body mass index (BMI), race, and position with CVD mortality. Overall player mortality was significantly decreased (SMR 0.53, 95% confidence interval [CI] 0.48 to 0.59) as was mortality from cancer (SMR 0.58, 95% CI 0.46 to 0.72), and CVD (SMR 0.68, 95% CI 0.56 to 0.81). CVD mortality was increased for defensive linemen (SMR 1.42, 95% CI 1.02 to 1.92) but not for offensive linemen (SMR 0.70, 95% CI 0.45 to 1.05). Defensive linemen’s cardiomyopathy mortality was also increased (SMR 5.34, 95% CI 2.30 to 10.5). Internal analyses found that CVD mortality was increased for players of nonwhite race (hazard ratio 1.69, 95% CI 1.13 to 2.51). After adjusting for age, race, and calendar year, CVD mortality was increased for those with a playing-time BMI ≥30 kg/m 2 (hazard ratio 2.02, 95% CI 1.06 to 3.85) and for defensive linemen compared to offensive linemen (hazard ratio 2.07, 95% CI 1.24 to 3.46). In conclusion, National Football League players from the 1959 through 1988 seasons had decreased overall mortality but those with a playing-time BMI ≥30 kg/m 2 had 2 times the risk of CVD mortality compared to other players and African-American players and defensive linemen had higher CVD mortality compared to other players even after adjusting for playing-time BMI.
In 1990 the National Football League (NFL) Players Association requested that the National Institute for Occupational Safety and Health (NIOSH) investigate the rate and causes of death of NFL players because of concerns about player longevity and excess cardiovascular disease (CVD) mortality. In 1994 the NIOSH reported that players from the 1959 through 1988 seasons experienced decreased overall mortality, but CVD mortality was higher in linemen compared to nonlinemen and in players with an increased playing-time body mass index (BMI) compared to other players. Subsequent studies of retired and active professional players have documented higher rates of CVD risk factors—including hypertension, increased left atrial size, and metabolic syndrome—in relation to player position and size. The 1994 report on NFL mortality was based on a small number of cardiovascular deaths (38) owing to the relatively short vital status follow-up period. We report on 16 additional years of follow-up, expanding our understanding of players’ mortality as they age beyond their status as elite athletes.
Methods
A cohort of 3,439 NFL players was constructed from a 1990 NFL pension fund database and included all pension-vested players (≥5 pension-credited playing seasons) from the 1959 through 1988 seasons. The pension fund database includes each player’s name, Social Security number, date of birth, and year of each credited season. Player race was collected because of known racial differences in United States mortality. Race was determined based on players’ self-identification using standard racial categories on a response form accompanying a letter introducing the study sent to players’ last known address. For nonrespondents (48%) race was assigned using available pictures in yearly media guides. Position played and height and weight during the last season played were assigned using data reported annually by the NFL teams and compiled in a commercial publication. BMI (weight in kilograms divided by height in meters squared) was categorized using standard cutpoints: normal (18.5 to <25 kg/m 2 ), overweight (25 to <30 kg/m 2 ), and obese (≥30 kg/m 2 ). Player position was categorized based on physiologic studies of college and professional players demonstrating differences in size, body composition, strength, and endurance: category I (defensive back, punter, kicker, quarterback, and wide receiver), category II (fullback, halfback, linebacker, running back, and tight end), and category III (linemen). Because of our interest in linemen, we performed additional analyses on the 2 major subgroups, offensive and defensive linemen.
Vital status was ascertained from multiple sources: NFL pension fund death records, Social Security Administration Death Master File, and Internal Revenue Service records. Players were matched to the National Death Index from 1979, when the National Death Index began, through the study end date, December 31, 2007. The National Death Index provided underlying causes of death coded to the revision of the International Classification of Diseases in effect at the time of death. When death information was not provided by the National Death Index, hardcopy death certificates were obtained from state vital statistics offices and were coded by a certified nosologist.
NFL player mortality was compared to the general United States population using the NIOSH life-table analysis system. Because all cohort members had a minimum of 5 credited seasons, a risk begin date was determined as the approximate date at the end of the fifth credited season (assigned as February 1). Each cohort member accumulated person-years at risk for each year of life from the risk begin date until the date of death or the study end date, whichever came first. To calculate the expected number of deaths, person-years at risk were stratified into 5-year intervals by age and calendar time and then multiplied by the appropriate United States race- and cause-specific mortality rates for men in 119 cause-of-death categories. For each cause of death the expected numbers of deaths were summed across the strata and the ratio of observed to expected number of deaths was expressed as the standardized mortality ratio (SMR). Ninety-five percent confidence intervals (CIs) were obtained using exact methods when the observed number of deaths was ≤10 or approximate methods when the observed number of deaths was >10. SMRs were also calculated by position category and for the offensive and defensive linemen subgroups.
Internal analyses compared mortality across player position categories. Because stratified SMRs are generally not comparable, directly standardized rate ratios and 95% CIs for each position category relative to category I were obtained using the life-table analysis system. We also used Cox proportional hazards regression models to evaluate the association of position category and BMI with CVD mortality ( International Classification of Diseases, Tenth Revision codes I00 through I78). Attained age was used as the time scale; consequently, age was controlled for in the models. BMI was treated as a categorical variable using standard cutpoints but additional models treating BMI as a continuous variable were evaluated. In addition to position category and BMI, other covariates included race (white/Hispanic, African-American/all other races), era of play (play ended before vs in 1980 [median] or later), and time since last played (<10 vs ≥10 years). We also controlled for the calendar year of follow up (by decades) because recent studies have found that age-adjusted CVD mortality has been decreasing over time. Because BMI and position category are strongly associated, in our models we first examined the independent effect of BMI on CVD mortality. Additional models examined the added effect of position category on CVD mortality while adjusting for the role of BMI.
The proportional hazards assumption for BMI was tested using a time-dependent interaction term between age (<55 vs ≥55 years) and BMI category. To examine the impact of possible racial misclassification we conducted additional stratified analyses comparing findings for players with self-reported versus observationally assigned race. All regression modeling used the SAS 9.2 procedure PHREG (SAS Institute, Cary, North Carolina). A 2-sided p value <0.05 was considered statistically significant. The study protocol was approved by the NIOSH institutional review board.
Results
The initial cohort included 3,732 NFL players but 292 players with unknown race and 1 “player” who was actually a trainer were excluded. By the end of follow-up in 2007, the final cohort of 3,439 players contributed 104,776 person-years at risk and 334 deaths. On average the cohort was followed for 26.8 ± 8.7 years (mean ± SD) after retirement from the NFL. For players still alive, the median age at the study end date was 57 years; 60% of the players were white (including 15 Hispanics) and 39% were African-American ( Table 1 ). Players were evenly distributed among the 3 position categories and 10% of the cohort had a BMI <25 kg/m 2 and 34% had a BMI ≥30 kg/m 2 . BMI varied across position categories with <1% of category I, 18% of category II, and 82% of category III players having a BMI ≥30 kg/m 2 . Demographically, the players were similar across the position categories except a larger proportion of category III players were of white race. Within category III offensive and defensive linemen were similar except offensive linemen were more likely to be of white race and to have a BMI ≥30 kg/m 2 .
| Variable | Overall (n = 3,439) | Position Category ⁎ | |||
|---|---|---|---|---|---|
| I | II | III—Defensive | III—Offensive | ||
| (n = 1,180) | (n = 1,097) | (n = 498) | (n = 664) | ||
| Race | |||||
| White | 2,070 (60%) | 631 (53%) | 606 (55%) | 291 (58%) | 542 (82%) |
| African-American | 1,355 (39%) | 548 (46%) | 486 (44%) | 203 (41%) | 118 (18%) |
| Other | 14 (<1%) | 1 (<1%) | 5 (<1%) | 4 (1%) | 4 (1%) |
| Vital status as of December 31, 2007 | |||||
| Alive | 3,105 (90%) | 1,095 (93%) | 998 (91%) | 422 (85%) | 590 (89%) |
| Dead | 334 (10%) | 85 (7%) | 99 (9%) | 76 (15%) | 74 (11%) |
| First credited season | |||||
| Median (range) | 1973 (1950–1984) | 1973 (1950–1984) | 1974 (1950–1984) | 1972 (1950–1984) | 1972 (1950–1984) |
| Before 1980 | 2,685 (78%) | 912 (77%) | 847 (77%) | 398 (80%) | 528 (80%) |
| 1980 and after | 754 (22%) | 268 (23%) | 250 (23%) | 100 (20%) | 136 (20%) |
| Age (years) at first credited season † | |||||
| Median (range) | 22 (19–34) | 22 (19–34) | 22 (21–28) | 23 (21–28) | 22 (21–28) |
| ≤21 | 789 (23%) | 248 (21%) | 258 (24%) | 112 (22%) | 171 (26%) |
| 22 | 1,614 (47%) | 572 (48%) | 532 (48%) | 214 (43%) | 296 (45%) |
| 23 | 701 (20%) | 245 (21%) | 215 (20%) | 106 (21%) | 135 (20%) |
| ≥24 | 335 (10%) | 115 (10%) | 92 (8%) | 66 (13%) | 62 (9%) |
| Number of credited seasons ‡ | |||||
| Median (range) | 8 (5–25) | 8 (5–25) | 7 (5–17) | 8 (5–20) | 8 (5–17) |
| Age (years) at death | |||||
| Median (range) | 54 (27–81) | 54 (27–80) | 53 (28–78) | 53 (31–81) | 55 (29–76) |
| Age (years) at date last observed—nondeceased | |||||
| Median (range) | 57 (45–88) | 57 (45–88) | 56 (45–82) | 58 (45–84) | 57 (45–83) |
| <50 | 633 (20%) | 218 (20%) | 212 (21%) | 84 (20%) | 119 (20%) |
| 50–54 | 738 (24%) | 267 (24%) | 263 (26%) | 82 (19%) | 126 (21%) |
| 55–59 | 565 (18%) | 197 (18%) | 162 (16%) | 90 (21%) | 116 (20%) |
| 60–69 | 890 (29%) | 309 (28%) | 282 (28%) | 120 (28%) | 179 (30%) |
| ≥70 | 279 (9%) | 104 (9%) | 79 (8%) | 46 (11%) | 50 (8%) |
| Body mass index (kg/m 2 ) | |||||
| Median (range) | 29 (22–39) | 26 (22–34) | 29 (24–34) | 31 (24–37) | 31 (25–39) |
| 18.5–<25 | 342 (10%) | 337 (29%) | 4 (<1%) | 1 (<1%) | 0 (0%) |
| 25–<30 | 1,944 (57%) | 841 (71%) | 897 (82%) | 114 (23%) | 92 (14%) |
| 30–<35 | 1,125 (33%) | 2 (<1%) | 196 (18%) | 372 (75%) | 555 (84%) |
| ≥35 | 28 (1%) | 0 (0%) | 0 (0%) | 11 (2%) | 17 (3%) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
