front |1 |2 |3 |4 |5 |6 |7 |8 |9 |10 |11 |12 |13 |14 |15 |16 |17 |18 |19 |20 |21 |22 |23 |24 |review

Table 3 contains the estimates for the simple relation of physical activity to baseline body weight and to weight change over the 12-year follow-up for the entire cohort and for survivors and non-survivors. The regression estimates represent either the difference in body weight at baseline (total activity score (TAS)) or the change in rate of weight loss (TAS*time) per unit increase in baseline total activity score. Among the combined cohort, active older people initially weighed more at baseline (1.29 lbs for each unit increase in activity score) than their less active counterparts. The interaction term of activity and time suggests that each unit increase in baseline activity score attenuated weight loss by 0.16 lbs per year among the entire cohort. This relation is explained primarily by the association of activity to the rate of weight loss in those who did not survive the follow-up (beta=0.20 (0.09, 0.31)) and not in the survivors, confirming the effect modification of this relation by survival status, as well as possible confounding and/or effect modification by disability and chronic disease.