[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
Purchase Options:
[Skip to Content Landing]
Figure.
Mortality of Career National Football League (NFL) Players vs NFL Replacement Players Following Retirement From the NFL
Mortality of Career National Football League (NFL) Players vs NFL Replacement Players Following Retirement From the NFL

The figure plots Kaplan-Meier mortality curves for career NFL players vs NFL replacement players. The median duration of follow-up was 30 years (interquartile range, 27-33) in the career NFL player group and 31 years (interquartile range, 30-33) in the NFL replacement player group. The mortality curve for career NFL players was not significantly different from the mortality curve for NFL replacement players (log-rank test: χ21 = 3.26). The numbers at risk initially increase because players are not considered at risk until retirement (left truncation). The drop in the number at risk at the end of event time was due to differences in player birth cohort.

Table 1.  
NFL Long-term Mortality Study Populationa
NFL Long-term Mortality Study Populationa
Table 2.  
Mortality for Career NFL Players vs NFL Replacement Players
Mortality for Career NFL Players vs NFL Replacement Players
Table 3.  
Causes of Death in Career and Replacement Playersa
Causes of Death in Career and Replacement Playersa
1.
 Legal Issues Relating to Football Head Injuries (Part I & II): Hearings Before the Committee on the Judiciary House of Representatives. 111th Congress, 1st and 2nd sessions. Washington, DC: US Government Printing Office; 2010.
2.
C-SPAN. Concussion research and treatment. http://www.c-span.org/video/?406450-1/hearing-concussions. Published March 14, 2016. Accessed December 2, 2017.
3.
Horn  S, Gregory  P, Guskiewicz  KM.  Self-reported anabolic-androgenic steroids use and musculoskeletal injuries: findings from the center for the study of retired athletes health survey of retired NFL players.  Am J Phys Med Rehabil. 2009;88(3):192-200.PubMedGoogle ScholarCrossref
4.
Mez  J, Daneshvar  DH, Kiernan  PT,  et al.  Clinicopathological evaluation of chronic traumatic encephalopathy in players of American football.  JAMA. 2017;318(4):360-370.PubMedGoogle Scholar
5.
Omalu  BI, DeKosky  ST, Minster  RL, Kamboh  MI, Hamilton  RL, Wecht  CH.  Chronic traumatic encephalopathy in a National Football League player.  Neurosurgery. 2005;57(1):128-134.PubMedGoogle ScholarCrossref
6.
Holstein  J, Jones  R, Koonce  G.  Is There Life After Football? Surviving the NFL. New York: New York University Press; 2015.
7.
Chetty  R, Stepner  M, Abraham  S,  et al.  The association between income and life expectancy in the United States, 2001-2014.  JAMA. 2016;315(16):1750-1766.PubMedGoogle ScholarCrossref
8.
Weir  D, Jackson  J, Sonnega  A.  National Football League Player Care Foundation Study of Retired NFL Players. Ann Arbor: Institute for Social Research, University of Michigan; 2009.
9.
Baron  SL, Hein  MJ, Lehman  E, Gersic  CM.  Body mass index, playing position, race, and the cardiovascular mortality of retired professional football players.  Am J Cardiol. 2012;109(6):889-896.PubMedGoogle ScholarCrossref
10.
Baron  SL, Rinsky  R.  Health Hazard Evaluation Report: National Football League Players Mortality Study: Report No. HETA 88-085. Atlanta, GA: Centers for Disease Control and Prevention, National Institute for Occupational Safety and Health; 1994.
11.
Lehman  EJ, Hein  MJ, Gersic  CM.  Suicide mortality among retired National Football League players who played 5 or more seasons.  Am J Sports Med. 2016;44(10):2486-2491.PubMedGoogle ScholarCrossref
12.
Lehman  EJ, Hein  MJ, Baron  SL, Gersic  CM.  Neurodegenerative causes of death among retired National Football League players.  Neurology. 2012;79(19):1970-1974.PubMedGoogle ScholarCrossref
13.
Pincivero  DM, Bompa  TO.  A physiological review of American football.  Sports Med. 1997;23(4):247-260.PubMedGoogle ScholarCrossref
14.
Kluck  T.  Three-Week Professionals: Inside the 1987 NFL Players’ Strike. Lanham, MD: Rowman & Littlefield; 2015.
15.
Merrill  E. NFL replacements part of history. ESPN. http://espn.go.com/nfl/news/story?id=6642330. Published June 9, 2011. Accessed April 17, 2016.
16.
Pro Football Reference. Football stats and history. http://www.pro-football-reference.com. Accessed April 15, 2016.
17.
Boublik  M, Schlegel  TF, Koonce  RC, Genuario  JW, Kinkartz  JD.  Quadriceps tendon injuries in National Football League players.  Am J Sports Med. 2013;41(8):1841-1846.PubMedGoogle ScholarCrossref
18.
Koning  R, Matheson  V, Nathan  A, Pantano  J.  The long-term game: an analysis of life expectancy of National Football League players.  Int J Financial Stud. 2014;2(1):168-178. doi:10.3390/ijfs2010168Google ScholarCrossref
19.
Parekh  SG, Wray  WH  III, Brimmo  O, Sennett  BJ, Wapner  KL.  Epidemiology and outcomes of Achilles tendon ruptures in the National Football League.  Foot Ankle Spec. 2009;2(6):283-286.PubMedGoogle ScholarCrossref
20.
Smith  RS, Efron  B, Mah  CD, Malhotra  A.  The impact of circadian misalignment on athletic performance in professional football players.  Sleep. 2013;36(12):1999-2001.PubMedGoogle ScholarCrossref
21.
Williams  T.  Long-term Mortality Effects of An NFL Career. Cambridge: Massachusetts Institute of Technology; 2012.
22.
Troan  J. NFL rosters. http://www.jt-sw.com/football/pro/rosters.nsf. Updated January 6, 2002. Accessed January 15, 2016.
23.
Oldest Living Pro Football Players. http://www.oldestlivingprofootball.com. Accessed June 10, 2016.
24.
National Football League. Players. http://www.nfl.com/players. Accessed June 10, 2016.
25.
National Center for Health Statistics; Centers for Disease Control and Prevention. National Death Index: user’s guide. https://www.cdc.gov/nchs/data/ndi/ndi_users_guide.pdf. Published 2013. Accessed January 12, 2018.
26.
Fillenbaum  GG, Burchett  BM, Blazer  DG.  Identifying a National Death Index match.  Am J Epidemiol. 2009;170(4):515-518.PubMedGoogle ScholarCrossref
27.
Freedman  LS.  Tables of the number of patients required in clinical trials using the logrank test.  Stat Med. 1982;1(2):121-129.PubMedGoogle ScholarCrossref
28.
Korn  EL, Graubard  BI, Midthune  D.  Time-to-event analysis of longitudinal follow-up of a survey: choice of the time-scale.  Am J Epidemiol. 1997;145(1):72-80.PubMedGoogle ScholarCrossref
29.
Pencina  MJ, Larson  MG, D’Agostino  RB.  Choice of time scale and its effect on significance of predictors in longitudinal studies.  Stat Med. 2007;26(6):1343-1359.PubMedGoogle ScholarCrossref
30.
Fogel  RW.  Economic growth, population theory, and physiology: the bearing of long-term processes on the making of economic policy.  Am Econ Rev. 1994;84(3):369-395.Google Scholar
31.
Grambsch  PM, Therneau  TM.  Proportional hazard tests and diagnostics based on weighted residuals.  Biometrika. 1994;81(3):515-526. doi:10.1093/biomet/81.3.515Google ScholarCrossref
32.
Dwyer-Lindgren  L, Bertozzi-Villa  A, Stubbs  RW,  et al.  US county-level trends in mortality rates for major causes of death, 1980-2014.  JAMA. 2016;316(22):2385-2401.PubMedGoogle ScholarCrossref
33.
Wang  H, Naghavi  M, Allen  C,  et al; GBD 2015 Mortality and Causes of Death Collaborators.  Global, regional, and national life expectancy, all-cause mortality, and cause-specific mortality for 249 causes of death, 1980-2015: a systematic analysis for the Global Burden of Disease Study 2015.  Lancet. 2016;388(10053):1459-1544.PubMedGoogle ScholarCrossref
34.
Case  A, Fertig  A, Paxson  C.  The lasting impact of childhood health and circumstance.  J Health Econ. 2005;24(2):365-389.PubMedGoogle ScholarCrossref
35.
Chetty  R, Hendren  N, Katz  LF.  The effects of exposure to better neighborhoods on children: new evidence from the Moving to Opportunity Experiment.  Am Econ Rev. 2016;106(4):855-902. doi:10.1257/aer.20150572Google ScholarCrossref
Original Investigation
February 27, 2018

Association Between Playing American Football in the National Football League and Long-term Mortality

Author Affiliations
  • 1Department of Medical Ethics and Health Policy, Perelman School of Medicine, University of Pennsylvania, Philadelphia
  • 2Leonard Davis Institute of Health Economics, University of Pennsylvania, Philadelphia
  • 3Department of Heart and Vascular (Cardiology), Clinical Cardiac Electrophysiology, The Everett Clinic, Everett, Washington
  • 4Providence Regional Medical Center, Everett, Washington
  • 5Department of Health Care Policy, Harvard Medical School, Boston, Massachusetts
  • 6Department of Medicine, Massachusetts General Hospital, Boston
  • 7National Bureau of Economic Research, Cambridge, Massachusetts
JAMA. 2018;319(8):800-806. doi:10.1001/jama.2018.0140
Key Points

Question  Is regular participation as a player in the National Football League (NFL) associated with long-term mortality risk?

Findings  In this retrospective cohort study of 3812 NFL players who debuted between 1982 and 1992, there was no statistically significant difference in the risk of long-term all-cause mortality among career NFL players compared with NFL replacement players who participated in the NFL during a 3-game league-wide player strike in 1987 (adjusted hazard ratio, 1.38; 95% CI, 0.95-1.99).

Meaning  Career participation in NFL football, compared with participation as an NFL replacement player, was not associated with a statistically significant difference in the risk of all-cause mortality.

Abstract

Importance  Studies of the longevity of professional American football players have demonstrated lower mortality relative to the general population but they may have been susceptible to selection bias.

Objective  To examine the association between career participation in professional American football and mortality risk in retirement.

Design, Setting, and Participants  Retrospective cohort study involving 3812 retired US National Football League (NFL) players who debuted in the NFL between 1982 and 1992, including regular NFL players (n = 2933) and NFL “replacement players” (n = 879) who were temporarily hired to play during a 3-game league-wide player strike in 1987. Follow-up ended on December 31, 2016.

Exposures  NFL participation as a career player or as a replacement player.

Main Outcomes and Measures  The primary outcome was all-cause mortality by December 31, 2016. Cox proportional hazards models were estimated to compare the observed number of years from age 22 years until death (or censoring), adjusted for birth year, body mass index, height, and position played. Information on player death and cause of death was ascertained from a search of the National Death Index and web-based sources.

Results  Of the 3812 men included in this study (mean [SD] age at first NFL activity, 23.4 [1.5] years), there were 2933 career NFL players (median NFL tenure, 5 seasons [interquartile range {IQR}, 2-8]; median follow-up, 30 years [IQR, 27-33]) and 879 replacement players (median NFL tenure, 1 season [IQR, 1-1]; median follow-up, 31 years [IQR, 30-33]). At the end of follow-up, 144 NFL players (4.9%) and 37 replacement players (4.2%) were deceased (adjusted absolute risk difference, 1.0% [95% CI, −0.7% to 2.7%]; P = .25). The adjusted mortality hazard ratio for NFL players relative to replacements was 1.38 (95% CI, 0.95 to 1.99; P = .09). Among career NFL players, the most common causes of death were cardiometabolic disease (n = 51; 35.4%), transportation injuries (n = 20; 13.9%), unintentional injuries (n = 15; 10.4%), and neoplasms (n = 15; 10.4%). Among NFL replacement players, the leading causes of death were cardiometabolic diseases (n = 19; 51.4%), self-harm and interpersonal violence (n = 5; 13.5%), and neoplasms (n = 4; 10.8%).

Conclusions and Relevance  Among NFL football players who began their careers between 1982 and 1992, career participation in the NFL, compared with limited NFL exposure obtained primarily as an NFL replacement player during a league-wide strike, was not associated with a statistically significant difference in long-term all-cause mortality. Given the small number of events, analysis of longer periods of follow-up may be informative.

Introduction

Concerns around the long-term health of professional football players in the National Football League (NFL) have intensified since 2009.1,2 Career participation in professional football may be potentially harmful for several reasons, including exposure to harmful league-specific lifestyle factors and substance abuse (such as the use of anabolic steroids3) that may confer long-term harm, repeated exposure to head trauma that may result in adverse neurodegenerative changes (eg, chronic traumatic encephalopathy4,5), and difficulties in adjusting to life after retirement from the NFL.6 In contrast, other factors, such as exceptional access to training, health care, and greater income, may be protective of long-term health.7,8

Two studies of NFL pension-eligible retired players found significantly lower overall and cardiovascular mortality for career NFL players compared with age- and race/ethnicity-matched US general population controls.9,10 Similarly designed studies using these data have found no differences in mortality from suicide,11 although perhaps a higher risk of mortality from neurodegenerative diseases.12 A limitation of existing observational studies of mortality of NFL players is bias that may result from the comparison of participants who differ in underlying mortality risk. NFL players—and professional athletes more generally—are physiologically distinct from members of the general population, often in ways that suggest more robust health.13 A more valid comparison group would be individuals with similar athletic or physiologic attributes and lifestyles.

This study proposes such a comparison group. Specifically, in 1987, a 3-game league-wide NFL player strike necessitated the replacement of regular NFL players with substitute rosters. A retrospective cohort study was conducted to examine the risk of mortality during retirement among career NFL players compared with NFL “replacement players.”

Methods

The Partners Healthcare Human Research Committee approved the study protocol (2017P001511), which included use of the US National Death Index (NDI) in the manner that did not require participant consent.

Background and Research Strategy

This study’s primary objective was to examine the association between career exposure to the NFL and mortality. To obtain an appropriate comparison group for NFL athletes, the study used data from a period during which the NFL Players’ Association initiated a 24-day strike over collective bargaining rules around free agency.14 During this period, which occurred after the second week of the 1987 season, team owners enlisted replacement players to fill rosters and continue play during the fourth through sixth weeks of the season, after which the strike ended. NFL replacement players were typically former college football players, players who were released from their NFL team during the preseason, or former players from rival leagues.14,15 NFL replacement players in the 1987 season serve as an appropriate comparison group for career NFL players under the assumption that replacement players were experienced athletes fit enough to obtain an NFL roster position with the exit of another professional athlete, but relatively unexposed to an NFL career (ie, expected to play at most 3 games).

Study Population

Pro Football Reference (PFR), an online database covering the census of professional American football players since the 1920s, was used to obtain baseline characteristics of the study population.16 The PFR database, maintained by Sports Reference LLC and freely available online (http://www.pro-football-reference.com), draws from a wide variety of historic and current team and game statistics. It includes information on player demographics (date and location of birth and date of death); tenure (number of years and games played); position; and height and body mass index (BMI,calculated as weight in kilograms divided by height in meters squared), in addition to detailed information on professional football statistics. These data have been used in other studies.17-21

NFL replacement players were identified in the PFR database using team media guides and online compilations of team rosters from the 1987 season.16,22 Roster information specifically denoting NFL replacement players was crosschecked against game-by-game summaries from the fourth to sixth weeks of the 1987 season when the player strike was in force. We focused on career NFL players who debuted between 1982 and 1992 based on sample size considerations and to form a narrow window around the strike year and increase birth cohort overlap between NFL career and replacement players. Players who played primarily as place kickers and punters, as well as players who died prior to retirement from the NFL, were excluded from the cohort. Covariates, including birth year, BMI and height (recorded in the last year of participation in the NFL), and position, were drawn from the PFR database.

Outcome Ascertainment

The primary outcome was all-cause mortality as of December 31, 2016. All-cause mortality and dates and causes of death were ascertained and validated using 2 complementary strategies. In the first strategy, a research assistant blinded to the purpose of the study identified mortality events using a variety of sources, including football player databases (PFR, the Oldest Living Pro Football Players database,23 and NFL.com24), Google searches, genealogy databases (Ancestry.com and FamilySearch.org), and the Social Security Death Index. If no record of death was found, attempts were made to verify that the player was currently alive using additional Google searches as well as through professional social networks (eg, LinkedIn). In the second strategy, the NDI was used to search for matches based on player first and last name, exact dates of birth, sex, and state of birth (social security numbers were not available).25

Deaths that were initially found in web searches but not listed as exact matches in an NDI search were verified among probable matches in the NDI. Similarly, deaths found in the NDI but not in initial web searches were verified in a second, targeted web search. Potential deaths deemed sufficiently probable in the NDI were also subjected to targeted web searches. Full details of this approach are provided in the eAppendix in the Supplement. The combined use of web-based searches, the Social Security Death Index, and NDI has been shown to result in high sensitivity and specificity in other populations, such as elderly US individuals.26

Statistical Analysis

As the number of NFL replacement players was fixed (n = 879), the log-rank test–based method was used to calculate the number of career NFL players needed to detect a relative mortality hazard of 1.4 with 80% power and a 2-sided type I error of .05 (see the eAppendix and eFigure 1 in the Supplement for further details).27 The chosen hazard ratio reflects the hazard of death estimated in an analysis of NFL player mortality by length of tenure.21 The required sample size was 2850 career NFL players, which was met by focusing on career players debuting between 1982 and 1992.

Baseline characteristics of career NFL players and NFL replacement players were compared using t tests and χ2 tests. Unadjusted and adjusted differences in absolute mortality risk were estimated across the 2 groups. Kaplan-Meier curves and Cox proportional hazard models were used to estimate the mortality hazard for career NFL players relative to NFL replacement players. Time was measured on an attained age time scale28 starting from the year the individual turned 22 years old—the time at which most NFL players graduated college and entered the league—to the end of the study (end line; December 31, 2016). The models accounted for left truncation, as individuals were at risk from the time of last football activity.29

Statistical significance of estimates from Kaplan-Meier curves was ascertained using the log-rank test. In Cox proportional hazards models, covariates that were adjusted for included birth year; BMI and player height (the latter of which was included given well-described associations with mortality in the literature30); and player position, which following prior studies, was grouped into 3 categories denoting different physiologic requirements (1: defensive back, quarterback, and wide receiver; 2: running back (including fullback and halfback), linebacker, and tight end; and 3: offensive linemen and defensive lineman).9 The proportional hazards assumption was formally tested by examining the relationship between Schoenfeld residuals and event time.31

Causes of death were descriptively examined among NFL career and replacement players, grouping International Classification of Diseases, Ninth Revision and International Statistical Classification of Diseases and Related Health Problems, Tenth Revision codes denoting the underlying cause of death into broad categories defined by the Global Burden of Disease Project.32,33 The percentage of deaths represented by each broad cause was calculated within each player group and presented descriptively.

Sensitivity Analyses

Two sensitivity analyses were conducted. First, the main hazard models were additionally adjusted for birth county–level socioeconomic measures (rates of unemployment, poverty, high school graduation, and nonwhite race/ethnicity percentage obtained from the 1970 US Census). This allowed assessment of the sensitivity of the findings to adjusting for a proxy for player race/ethnicity7 and differences in early childhood environments, which are key drivers of long-run socioeconomic status and health in population-based studies.34,35

Second, the models were reestimated after removing all NFL replacement players who had professional football experience (in the NFL or other professional leagues) prior to the strike year.

All statistical analyses were conducted using Stata version 15 (StataCorp). All 95% CIs and P values were based on 2-sided hypothesis tests, where P < .05 was considered statistically significant.

Results
Sample Characteristics

The study sample consisted of 2933 career NFL players and 879 NFL replacement players (Table 1). Four career NFL players were excluded from the sample because they died while actively employed on an NFL roster. By the study’s end point, the career NFL players’ mean age was 1.5 years younger than that of NFL replacement players (mean [SD] age, 52.7 [3.2] years vs 54.2 [1.9] years [difference, 1.5; 95% CI of difference, 1.25-1.70]; P < .001). The career NFL players’ mean age was 1 year younger than that of NFL replacement players at debut (mean [SD] age, 23.2 [1.2] years vs 24.2 [1.8] years [difference, 1.0; 95% CI of difference, 0.85-1.07]; P < .001) and 2.7 years older at retirement (mean [SD] age, 27.6 [3.9] years vs 24.9 [2.2] years [difference, 2.7; 95% CI of difference, 2.4-3.0]; P < .001). The 0.7-year difference between mean age at debut and retirement for NFL replacement players reflects that 155 (18%) played at least 1 game in seasons prior to the strike year and another 54 (6%) played in a subsequent season.

Mean BMI and mean height at the time of NFL participation were similar between career NFL players and NFL replacement players (mean [SD] BMI, 29.7 [3.7] vs 29.4 [3.3]; P = .03; mean [SD] height, 187.2 [6.8] cm vs 187.0 [6.4] cm; P = .62).

All-Cause Mortality

A total of 181 deaths occurred over a total time at risk of 95 441 person-years. A total of 144 deaths occurred among career NFL players (median follow-up, 30 years; interquartile range, 27-33) and 37 deaths occurred among NFL replacement players (median follow-up, 31 years; interquartile range, 30-33). The mortality risk by study end line was 4.9% for career NFL players and 4.2% for NFL replacement players, and the unadjusted absolute risk difference was 0.7% (95% CI, −0.9% to 2.2%; P = .39). Adjusting for birth year, BMI, height, and position group, the absolute risk difference was 1.0% (95% CI, −0.7 to 2.7; P = .25) (Table 2).

The Figure presents unadjusted time-to-death curves for career NFL players vs NFL replacement players. The mortality curve for career NFL players was not significantly different from the curve for NFL replacement players (log-rank test: χ21 = 3.26; P = .07). Table 2 presents the Cox proportional hazard estimates of differential mortality risk. Career NFL players had a mortality hazard of 1.39 (95% CI, 0.97 to 2.00; P = .08) relative to NFL replacement players. The estimate did not vary substantively with the inclusion of additional covariates: in the model with all covariates, the estimated mortality hazard was 1.38 (95% CI, 0.95 to 1.99; P = .09). Among statistically significant variables, BMI was positively associated with mortality risk (see eTable 1 in the Supplement for the full set of estimates). Formal tests did not reveal failure of the proportional hazard assumption (eFigure 2 in the Supplement).

Causes of Death

Table 3 presents the distribution of deaths among career NFL players and replacement players by cause of death categories (eTable 2 in the Supplement lists detailed causes of death recorded in the NDI). A known cause of death was identified for 176 of 181 players (97%), including 35 of 37 NFL replacement players (95%) and 141 of 144 career NFL players (98%).

Among career NFL players (144 deaths), the most common causes of death were cardiometabolic disease (n = 51; 35.4%), transportation injuries (n = 20; 13.9%), unintentional injuries (n = 15; 10.4%), and neoplasms (n = 15; 10.4%). Among NFL replacement players (37 deaths), the leading causes of death were cardiometabolic diseases (n = 19; 51.4%), self-harm and interpersonal violence (n = 5; 13.5%), and neoplasms (n = 4; 10.8%). Ten of the 15 career NFL player deaths (and 1 of the 2 replacement player deaths) from unintentional injuries were due to accidental poisonings (from substances, medications, and/or biological substances, most of which were unspecified). There were 7 deaths (4.9%) from neurological causes in the career NFL player group and none in the NFL replacement player group. All 7 of these deaths were due to amyotrophic lateral sclerosis.

Sensitivity Analyses

Findings were similar after adjusting for county-of-birth race/ethnicity distributions and socioeconomic characteristics (adjusted absolute risk difference, 0.9% [95% CI, −0.8% to 2.6%], P = .29; adjusted hazard ratio, 1.36 [95% CI, 0.94 to 1.96], P = .11; eTable 3 in the Supplement) and in models excluding NFL replacement players with professional experience prior to 1987 (adjusted absolute risk difference, 0.8% [95% CI, −1.0% to 2.6%], P = .41; adjusted hazard ratio, 1.33 [95% CI, 0.90 to 1.96], P = .15; eTable 4 in the Supplement).

Discussion

In this study of NFL players who debuted between 1982 and 1992, career participation in the NFL, compared with limited NFL exposure obtained primarily as a replacement player during a league-wide strike, was not associated with a statistically significant difference in the risk of long-term all-cause mortality.

In this study, the use of NFL replacement players provided a comparison group that is more likely closely matched in terms of unobserved characteristics that affect long-term health than age- and race/ethnicity-matched general population controls, the main comparison group used in prior studies of the association between NFL participation and mortality to date.9,10,12 Other research has examined differences in mortality by position played or length of tenure, both of which may be associated with mortality.9,18,21 However, these approaches cannot address the question of whether playing professional football is associated with an increased risk of mortality relative to not playing. The use of NFL replacement players as a control group in the current study allows a more direct comparison.

While the present study could not precisely determine an association between career NFL participation and mortality, the 95% CIs are not consistent with a large inverse association with mortality such as found in research that used general population controls.9,10,12 Moreover, the estimates in this study may represent lower bounds on the association between career NFL participation and mortality as it is possible that there remained systematic differences between the player groups—such as better physical health, skill, and earnings, particularly among NFL players with longer tenures—that are positively correlated with long-term health.

Limitations

This study has several limitations. First, the residual differences between career NFL players and NFL replacement players that may bias the results could not be addressed. Additional research that collects data on medical comorbidities, lifestyle factors, and earnings across the 2 groups could further address this bias. A complementary approach might focus on groups of NFL career and replacement players who are more similar in their duration of NFL exposure. For example, comparing career NFL players with shorter tenures vs NFL replacement players may avoid healthy worker bias from players who are able to persist in the league because of better unobserved health. Another consideration for researchers is how to model exposure for NFL replacement players who participated in the NFL outside of the strike year or played in other professional leagues in the United States and Canada. In the current analyses, excluding these players did not substantively change the findings.

Second, the estimates were based on a small number of events: 181 deaths, of which only 37 occurred among NFL replacement players. Consequently, the present analysis could be underpowered to detect meaningful associations. Additional analyses with longer-term follow-up may be informative.

Third, the data were drawn from online databases. While these have been used in other analyses,17-21 it is possible that some information was misreported, particularly for NFL replacement players. A consequence of the use of web-based sources was that reliable information on player race/ethnicity was not available. However, in sensitivity analysis, the use of county-of-birth measures of race/ethnicity, which follows from other research in which reliable information on individual race/ethnicity was not available,7 did not appreciably change the point estimates.

Fourth, the data used in this study cannot be used to identify mechanisms underlying career participation in professional football and health. In this regard, studies that account for differences in medical comorbidities or lifestyles may provide useful insights.

Fifth, the results may be specific to NFL players who debuted between 1982 and 1992. It is possible that cohorts who debuted at other times faced different mortality risks.

Conclusions

Among NFL football players who began their careers between 1982 and 1992, career participation in the NFL, compared with limited NFL exposure obtained primarily as an NFL replacement player during a league-wide strike, was not associated with a statistically significant difference in long-term all-cause mortality. Given the small number of events, analysis of longer periods of follow-up may be informative.

Back to top
Article Information

Corresponding Author: Atheendar S. Venkataramani, MD, PhD, Department of Medical Ethics and Health Policy, Perelman School of Medicine, University of Pennsylvania, Blockley Hall, 1102, 423 Guardian Dr, Philadelphia, PA 19104 (atheenv@pennmedicine.upenn.edu).

Accepted for Publication: January 9, 2018.

Published Online: February 1, 2018. doi:10.1001/jama.2018.0140

Author Contributions: Dr Venkataramani had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: All authors.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Venkataramani.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Venkataramani, Jena.

Administrative, technical, or material support: Gandhavadi.

Supervision: Gandhavadi, Jena.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report salary support from the National Institutes of Health outside the submitted work (Dr Venkataramani: Mentored Career Development Award, grant K23MH106362; Dr Jena: Early Independence Award, grant 1DP5OD017897). Dr Jena reported receiving consulting fees unrelated to this work from Pfizer Inc, Hill Rom Services Inc, Bristol-Myers Squibb, Novartis Pharmaceuticals, Vertex Pharmaceuticals, and Precision Health Economics, a company providing consulting services to the life sciences industry. No other disclosures were reported.

References
1.
 Legal Issues Relating to Football Head Injuries (Part I & II): Hearings Before the Committee on the Judiciary House of Representatives. 111th Congress, 1st and 2nd sessions. Washington, DC: US Government Printing Office; 2010.
2.
C-SPAN. Concussion research and treatment. http://www.c-span.org/video/?406450-1/hearing-concussions. Published March 14, 2016. Accessed December 2, 2017.
3.
Horn  S, Gregory  P, Guskiewicz  KM.  Self-reported anabolic-androgenic steroids use and musculoskeletal injuries: findings from the center for the study of retired athletes health survey of retired NFL players.  Am J Phys Med Rehabil. 2009;88(3):192-200.PubMedGoogle ScholarCrossref
4.
Mez  J, Daneshvar  DH, Kiernan  PT,  et al.  Clinicopathological evaluation of chronic traumatic encephalopathy in players of American football.  JAMA. 2017;318(4):360-370.PubMedGoogle Scholar
5.
Omalu  BI, DeKosky  ST, Minster  RL, Kamboh  MI, Hamilton  RL, Wecht  CH.  Chronic traumatic encephalopathy in a National Football League player.  Neurosurgery. 2005;57(1):128-134.PubMedGoogle ScholarCrossref
6.
Holstein  J, Jones  R, Koonce  G.  Is There Life After Football? Surviving the NFL. New York: New York University Press; 2015.
7.
Chetty  R, Stepner  M, Abraham  S,  et al.  The association between income and life expectancy in the United States, 2001-2014.  JAMA. 2016;315(16):1750-1766.PubMedGoogle ScholarCrossref
8.
Weir  D, Jackson  J, Sonnega  A.  National Football League Player Care Foundation Study of Retired NFL Players. Ann Arbor: Institute for Social Research, University of Michigan; 2009.
9.
Baron  SL, Hein  MJ, Lehman  E, Gersic  CM.  Body mass index, playing position, race, and the cardiovascular mortality of retired professional football players.  Am J Cardiol. 2012;109(6):889-896.PubMedGoogle ScholarCrossref
10.
Baron  SL, Rinsky  R.  Health Hazard Evaluation Report: National Football League Players Mortality Study: Report No. HETA 88-085. Atlanta, GA: Centers for Disease Control and Prevention, National Institute for Occupational Safety and Health; 1994.
11.
Lehman  EJ, Hein  MJ, Gersic  CM.  Suicide mortality among retired National Football League players who played 5 or more seasons.  Am J Sports Med. 2016;44(10):2486-2491.PubMedGoogle ScholarCrossref
12.
Lehman  EJ, Hein  MJ, Baron  SL, Gersic  CM.  Neurodegenerative causes of death among retired National Football League players.  Neurology. 2012;79(19):1970-1974.PubMedGoogle ScholarCrossref
13.
Pincivero  DM, Bompa  TO.  A physiological review of American football.  Sports Med. 1997;23(4):247-260.PubMedGoogle ScholarCrossref
14.
Kluck  T.  Three-Week Professionals: Inside the 1987 NFL Players’ Strike. Lanham, MD: Rowman & Littlefield; 2015.
15.
Merrill  E. NFL replacements part of history. ESPN. http://espn.go.com/nfl/news/story?id=6642330. Published June 9, 2011. Accessed April 17, 2016.
16.
Pro Football Reference. Football stats and history. http://www.pro-football-reference.com. Accessed April 15, 2016.
17.
Boublik  M, Schlegel  TF, Koonce  RC, Genuario  JW, Kinkartz  JD.  Quadriceps tendon injuries in National Football League players.  Am J Sports Med. 2013;41(8):1841-1846.PubMedGoogle ScholarCrossref
18.
Koning  R, Matheson  V, Nathan  A, Pantano  J.  The long-term game: an analysis of life expectancy of National Football League players.  Int J Financial Stud. 2014;2(1):168-178. doi:10.3390/ijfs2010168Google ScholarCrossref
19.
Parekh  SG, Wray  WH  III, Brimmo  O, Sennett  BJ, Wapner  KL.  Epidemiology and outcomes of Achilles tendon ruptures in the National Football League.  Foot Ankle Spec. 2009;2(6):283-286.PubMedGoogle ScholarCrossref
20.
Smith  RS, Efron  B, Mah  CD, Malhotra  A.  The impact of circadian misalignment on athletic performance in professional football players.  Sleep. 2013;36(12):1999-2001.PubMedGoogle ScholarCrossref
21.
Williams  T.  Long-term Mortality Effects of An NFL Career. Cambridge: Massachusetts Institute of Technology; 2012.
22.
Troan  J. NFL rosters. http://www.jt-sw.com/football/pro/rosters.nsf. Updated January 6, 2002. Accessed January 15, 2016.
23.
Oldest Living Pro Football Players. http://www.oldestlivingprofootball.com. Accessed June 10, 2016.
24.
National Football League. Players. http://www.nfl.com/players. Accessed June 10, 2016.
25.
National Center for Health Statistics; Centers for Disease Control and Prevention. National Death Index: user’s guide. https://www.cdc.gov/nchs/data/ndi/ndi_users_guide.pdf. Published 2013. Accessed January 12, 2018.
26.
Fillenbaum  GG, Burchett  BM, Blazer  DG.  Identifying a National Death Index match.  Am J Epidemiol. 2009;170(4):515-518.PubMedGoogle ScholarCrossref
27.
Freedman  LS.  Tables of the number of patients required in clinical trials using the logrank test.  Stat Med. 1982;1(2):121-129.PubMedGoogle ScholarCrossref
28.
Korn  EL, Graubard  BI, Midthune  D.  Time-to-event analysis of longitudinal follow-up of a survey: choice of the time-scale.  Am J Epidemiol. 1997;145(1):72-80.PubMedGoogle ScholarCrossref
29.
Pencina  MJ, Larson  MG, D’Agostino  RB.  Choice of time scale and its effect on significance of predictors in longitudinal studies.  Stat Med. 2007;26(6):1343-1359.PubMedGoogle ScholarCrossref
30.
Fogel  RW.  Economic growth, population theory, and physiology: the bearing of long-term processes on the making of economic policy.  Am Econ Rev. 1994;84(3):369-395.Google Scholar
31.
Grambsch  PM, Therneau  TM.  Proportional hazard tests and diagnostics based on weighted residuals.  Biometrika. 1994;81(3):515-526. doi:10.1093/biomet/81.3.515Google ScholarCrossref
32.
Dwyer-Lindgren  L, Bertozzi-Villa  A, Stubbs  RW,  et al.  US county-level trends in mortality rates for major causes of death, 1980-2014.  JAMA. 2016;316(22):2385-2401.PubMedGoogle ScholarCrossref
33.
Wang  H, Naghavi  M, Allen  C,  et al; GBD 2015 Mortality and Causes of Death Collaborators.  Global, regional, and national life expectancy, all-cause mortality, and cause-specific mortality for 249 causes of death, 1980-2015: a systematic analysis for the Global Burden of Disease Study 2015.  Lancet. 2016;388(10053):1459-1544.PubMedGoogle ScholarCrossref
34.
Case  A, Fertig  A, Paxson  C.  The lasting impact of childhood health and circumstance.  J Health Econ. 2005;24(2):365-389.PubMedGoogle ScholarCrossref
35.
Chetty  R, Hendren  N, Katz  LF.  The effects of exposure to better neighborhoods on children: new evidence from the Moving to Opportunity Experiment.  Am Econ Rev. 2016;106(4):855-902. doi:10.1257/aer.20150572Google ScholarCrossref
×