Abstract
OBJECTIVES: Career duration is often used as a metric of neurotrauma exposure in studies of elite athletes. However, as a proxy metric, career length may not accurately represent causal factors, and associations with health outcomes may be susceptible to selection effects. To date, relationships between professional American-style football (ASF) career length and postcareer health remain incompletely characterised. METHODS: We conducted a survey-based cross-sectional cohort study of former professional ASF players. Flexible regression methods measured associations between self-reported career duration and four self-reported health conditions: pain, arthritis, mood and cognitive symptoms. We also measured associations between career duration and four self-reported ASF exposures: prior concussion signs and symptoms (CSS), performance enhancing drugs, intracareer surgeries and average snaps per game. Models were adjusted for age and race. RESULTS: Among 4189 former players (52±14 years of age, 39% black, 34% lineman position), the average career length was 6.7±3.9 professional seasons (range=1-20+). We observed inverted U-shaped relationships between career duration and outcomes (all p<0.001), indicating that adverse health effects were more common among men with intermediate career durations than those with shorter or longer careers. Similar findings were observed for play-related exposures (eg, CSS and snaps). CONCLUSIONS: Relationships between ASF career duration and subsequent health status are non-linear. Attenuation of the associations among longer career players may reflect selection effects and suggest career length may serve as a poor proxy for true causal factors. Findings highlight the need for cautious use of career duration as a proxy exposure metric in studies of former athletes.