UNLABELLED: ABSTRACT: BACKGROUND: Longitudinal studies on physical activity patterns around retirement age are scarce and provide divergent findings. Little is known about changes in sedentary behaviour in this context. Our aim was to investigate relationships between retirement and 3-year changes in leisure-time physical activity (LTPA) patterns and sedentary behaviour in middle-aged French adults. METHODS: Past-year LTPA and sedentary behaviour (watching television) were assessed in 1998 and 2001 using the Modifiable Activity Questionnaire on participants in the SU.VI.MAX (Supplementation with Antioxidants and Minerals) study. A total of 698 men and 691 women aged 45-64 were included in this analysis. Comparisons were made between subjects who had retired between 1998 and 2001 and those who continued to work, using the Chi-square test, Student t-test, Wilcoxon rank test or covariance analysis where appropriate. RESULTS: 20.1% of men and 15.6% of women retired during follow-up. The baseline LTPA level was similar between subjects who retired during follow-up and those who continued to work. Mean LTPA increased by about 2 h/week in men and women who had retired, whereas no change was observed in employed persons. The positive change in LTPA following retirement was mainly related to an increase in activities of moderate intensity, such as walking. Retirement did not modify the ranking of the most frequently performed LTPAs, but the number of participants and the duration increased through retirement. In men, the increase in time spent watching TV was more than twice as high in retirees as in workers (+40.5 vs. +15.0 min/day, P < 0.0001). The same tendency was observed among women, but was borderline non-significant (+33.5 vs. +19.9 min/day, P = 0.05). In women, retirees who increased their walking duration by 2 h/week or more also decreased time spent watching TV by 11.5 min/day. CONCLUSIONS: Retirement was associated with both an increase in LTPAs and in time spent watching TV, suggesting that retirement is an important period not only for promoting physical activity, but also for limiting sedentary behaviour.