OBJECTIVE: To assess the association between Injury Severity Score (ISS) and subsequent risk of early retirement. DESIGN: Observational cohort study with follow-up based on prospectively collected data. Hospital-based data were linked to national register data on pension reception and vital status. SETTING: Level-one urban trauma centre. PARTICIPANTS: Patients aged 18-64 years entering the trauma centre in Copenhagen during 1999-2007 who were alive after three days were followed until early retirement, death or emigration. MAIN OUTCOME MEASURES: Primary outcome was early retirement, defined as receiving disability pension (unintentional) or voluntary early retirement pension (intentional) before the regular age of retirement (65 years). Relative risk of early retirement according to ISS (low, ISS 1-15 vs. high, ISS 16-75) was assessed using Cox proportional hazards regression, adjusted for age and gender. RESULTS: Of all 6687 patients admitted to the trauma centre, a total of 1722 trauma patients were included and followed for a median of 6.2 years (interquartile range (IQR) 3.7-9.1). Of these, 1305 (75.8%) were males, median age was 35.0 years (IQR 25.4-46.5), and median ISS was 16 (IQR 9-25). Three hundred and twenty-two patients retired during follow-up. Patients with high ISS, compared to patients with low ISS, had an increased risk of early retirement, adjusted hazard ratio 2.60 (95% confidence interval (CI) 2.05-3.30; p<0.001). Relative increase in retirement risk was 1.04 (95% CI 1.03-1.05) per ISS point and 1.03 (95% CI 1.03-1.04) per year older. Gender was not found to be a significant risk factor (p=0.69). Five-year absolute risks of early retirement were 9.9% (95% CI 7.8-12.0%) for the low ISS group and 24.6% (95% CI 21.6-27.5%) for the high ISS group. CONCLUSIONS: The risk of early retirement is 2.6 times higher in severely injured patients (ISS 16-75) than the risk in low to moderately injured patients (ISS 1-15) and they have a high absolute 5-year risk as well. Early, targeted interventions to assist with return to work might be able to reduce this risk.