This case-control study examines the association of vigorous and mild physical activity with fall-related fractures in a community-dwelling population age 65 and older in South Florida. Vigorous physical activity was defined as exercising, doing heavy housecleaning, or other hard labor three or more times per week in the month prior to the index date; mild physical activity was defined as the number of hours per day subjects reported spending on their feet. A case was any subject who sustained a fall-related fracture (ICD-9CM-800 through ICD-9CM-829) over a 21-month period (n = 471). Controls were at 10% random sample selected from the Health Care Financing Administration Medicare files (n = 712). The presence of any limitation in activities of daily living (ADL) significantly modified the effect of vigorous physical activity. Physically active subjects with no limitations (ADL = 0) were less likely to sustain a fall-related fracture than were inactive subjects with an adjusted odds ratio (aOR) of 0.6, (0.5-0.8 95% CI), and active subjects with any limitation (ADL > or = 1) had an aOR of 3.2 (1.1-9.8 95% CI). Limiting this analysis to 159 hip fracture cases produced similar results. Mild physical activity was not associated with fracture. These results suggest that vigorous physical activity is associated with a lower fracture risk among elderly persons who have no limitation in ADL and with a higher risk among those with any limitations.