Low bone mass, functional impairment, low muscle strength, and postural instability are predictors of the risk of fracture in an elderly person. The purpose of this study was to investigate the functional impact of an unvarying long-term exercise program to be carried out at home. The exercises had been shown to delay bone loss in an elderly population. At the Department of Physical Medicine and Rehabilitation, University of Vienna, postmenopausal women who had been stratified into exercise or control groups 5 to 10 yr ago were called in for a follow-up examination. Frequency of training, habits, and pain causing disability in activities of daily living were recorded. Walking velocity, muscle strength, and postural stability were measured. Functional assessment, blood analysis, and x-rays of the spine were performed additionally. One hundred twenty-four women aged 68.3 +/- 6.8 yr (mean +/- SD) underwent a follow-up investigation at the outpatient clinic. After 7.7 +/- 1.1 yr the compliance of the training group was still 36%. Self-chosen gait velocity was slightly higher in the regular exercisers than in the controls. No intergroup differences were found for pain induced disability, muscle strength, body sway, and fracture rate. The pain disability index was significantly associated with corrected self-chosen gait velocity. The results suggest that an unvarying home-based exercise program may support general agility but does not yield enough force to improve muscle strength and postural stability in healthy, nondisabled, postmenopausal women who start exercising at the age of 60 yr. Further studies are needed to define more appropriate exercise programs for a comprehensive improvement of functional outcome in a population at high risk for osteoporosis.