STUDY DESIGN: Longitudinal study with mortality follow-up. OBJECTIVE: Identify predictive factors for long-term mortality following tetraplegic spinal cord injury (TSCI). SETTING: The Tetrafigap survey is a multi-centre epidemiological survey on the long-term outcome of persons with TSCI, initiated in France in 1995 with the participation of 35 rehabilitation centres. METHODS: The mortality follow-up involves 1241 persons with TSCI who were admitted to one of the study rehabilitation units at the initial phase and who completed the initial self-administered questionnaire. There were 226 observed deaths (18.2%) during an 11-year period. Logistic regression methods, with estimates of odds ratios (ORs), incorporating clinical, functional and social participation data were used to determine the factors related to mortality. This was followed by multivariate analysis to determine the best predictive factors for long-term mortality. RESULTS: Risk of death increases significantly with age but not with the time elapsed since the accident. The risk of death is higher in men. Interestingly, clinical variables are not the best predictors of long-term mortality. Instead, the significant effect of poor social participation (being single, infrequent contact with friends) and functional limitations (full assistance required with dressing or eating) persists after adjustment for other variables. CONCLUSION: Once the medical situation becomes more stable, factors related to the long-term mortality of persons with TSCI are not exactly identical to those observed in the short acute-phase and during the first year after the accident. Social participation has a significant effect on mortality.