Hypothesis: age-associated changes in gait represent, in part, a tendency towards parkinsonism.

In 144 healthy, active, asymptomatic volunteers (aged 30-88 years), age explained 33% of the variance in free walking speed. This could not be ascribed simply to differences in physique at maturity, since height had a more important effect on speed (p less than 0.0001) than did leg length (p less than 0.01). Multiple linear regression was used to fit models for speed, by selection from four sets of variables: common physical, body sway, functional anatomy of spine and lower limbs, and psychometric. The best model (which included age) explained little more of the variance (41%) than did age alone. Moreover, incorporation of age into the alternative model, built in its absence, still contributed significantly (p less than 0.0001) to the variance explained. The effect of age on gait appeared to be complex, with an increment in double support time disproportionate to the reduction in speed. We formulate by exclusion, the hypothesis that age-associated changes in gait represent, in part, a tendency towards parkinsonism.