2 Physical activity as a risk or protective factor | 33 recommended to include a larger study population with more events. However, when conducting sensitivity analyses in a larger study population, results were the same compared to the primary analysis. Last, as in all longitudinal studies, there is the risk of subject attrition, which can lead to a motivated and healthy study sample, and an overestimation of physical activity and an underestimation of frailty, falls and fractures. To minimize a healthy survivor effect, we took all available data into account in the GEE analyses and not only the cases having complete data. The results of our study have implications for clinical practice and public health. In this study, we found that more physical activity does not decrease, but does also not increase fall and fracture risk. Since physical activity has major health benefits, such as more muscle and bone strength, but also a reduced risk of, for example, cardiovascular disease and diabetes, the advice remains to encourage older adults to be physically active (121). Because frailty appeared related to falls or fractures in our population, frail older adults should be monitored. Because of the ageing population, the number of older adults is increasing, of which a growing proportion will be frail. As frailty does not modify the interaction between physical activity and both falls and fractures, also in this group of frail older adults the advice remains to stay physically active. Thus, frailty is important for fall and fractures risk, but not specifically in the context of physical activity. Further research should investigate the extent and intensity to which physical activity is safe for frail older adults taken into account their higher risk of falls. Besides, a broader approach is needed to prevent falls and fractures than only looking at physical activity, when acting on all aspects of frailty. CONCLUSION Longer durations of physical activity did not decrease or increase the risk of falls or fractures in our sample of community-dwelling adults of 75 years or older. However, frail older adults in our study did have an increased fall and fracture risk compared to the non-frail participants. ACKNOWLEDGEMENTS We are grateful to all participants of the LASA study. We are grateful to the fieldwork team and all researchers at LASA for their ongoing commitment to the study.
RkJQdWJsaXNoZXIy MjY0ODMw