This is in line with other large cohort studies which reported either a gradual increase or decrease in risk ratios for higher physical activity categories [12, 14]. In this study, physical activity was not significantly associated with fall risk. Three other cohort studies reported an increased fall risk in men [12] and a decreased fall risk in women [14] or in persons living in a residential care setting [13] in higher physical activity categories as compared with the lowest category. Perhaps lack of an association in our study is

due to an interaction with sex. However, the interaction term for physical activity by sex was not significant (p = 0.89). A second explanation may be that in our study, participants with high levels of physical activity were underrepresented causing an underestimation of the actual relationship. However, our sample is representative for the community-dwelling older population AZD7762 in the Netherlands. Third, these three studies and the current study differed in population (men [12] vs women [14] vs residential care setting [13]), physical activity measures (validated questionnaires [12] vs operational definitions

[14]), and outcome measures (4-month fall risk [12] vs proportion fallers [14]). It is likely that the contrasting findings are explained by differences in population and methodology. The association between physical activity and recurrent falling Bioactive Compound Library has been studied only once before. A study among persons (70+ years) living in a residential care setting showed that the risk of recurrent falling decreased at higher levels of physical Glutamate dehydrogenase activity [13]. Our findings in community-dwelling older persons are in line with this study: an increase of 100 units led to a 4% lower risk of recurrent falling. One hundred units equal 30 min per day of walking, 20 in of swimming, or 40 min

of billiards. Thus, if all older persons increase their physical activity level with 100 units, 4% may be prevented to become recurrent fallers. In addition, given the beneficial effects of physical activity on other health outcomes, it is important to observe that, other than expected in the literature, highly active persons do not have an increased risk of falling. Clinical trials are necessary to test whether increasing physical activity leads to a decrease in falls. Two recently published systematic reviews showed that multiple component exercise programs did reduce the fall risk in community-dwelling older persons [34, 35]. Increasing daily physical activity may be an important component of these exercise programs. It has been suggested that in this type of study adjustment should be made for baseline mobility [9]. Like physical performance and functional limitations, mobility is a measure of physical functioning. In the current study, physical functioning did not modify the relationship between physical activity and (recurrent) falling.