Exercise Type in Dieting Obese Older Adults.

THE AUTHORS REPLY:

The current protocol was informed by our previous study¹ that included a weight loss alone group as compared with non-weight-loss groups. Direct effects that were due to weight loss were established in that study. Differential effects were established in the present trial by the head-to-head comparison of aerobic exercise with resistance exercise during matched weight loss. Aerobic exercise resulted in greater improvements in cardiorespiratory fitness, and resistance exercise in strength; the combination exercise group had the greatest improvement in physical function. An important point of our present trial is that weight loss alone was harmful with respect to lean mass and bone mineral density but that the addition of resistance exercise or the combination of resistance and aerobic exercise attenuated the loss. Our ongoing study aims to determine whether an increase in bone strength compensates for bone loss that remains after resistance exercise or combined exercise is added to weight loss. The existence of an “obesity paradox” in older adults is unclear.² Whether a Mediterranean or DASH diet would be as effective as a multifactorial intervention of weight loss and exercise in reversing frailty in obese older adults seems doubtful.

The differential outcomes due to the two exercise modes, aerobic and resistance, required that the combined exercise ensured sufficient physiological responses from both exercise interventions. Therefore, the combined exercise group did the full aerobic program plus the full resistance program to ensure an adequate dose of each type of training. Moreover, to test the interference effect,³ it was essential to balance the aerobic and resistance training between groups – reducing the training volume in the combination group would have made it impossible to differentiate a less optimal response due to the lower volume vs. interference effect. Accordingly, the combined exercise improved functional status the most, indicating no interference effect. We agree that a confidence interval of difference includes more information than P-values in simple situations. Although, like P-values confidence intervals alone do not establish clinical importance, one needs some clinical idea of minimally important differences.⁴ The 95% confidence interval for the difference in the change in PPT scores between the combination group and either exercise group was 0.5 to 2.7 (in each comparison); the information needed to compute confidence intervals between groups is provided in Table 2 of our article.