Copyright 2001 American Medical Association. All Rights Reserved.
Applicable FARS/DFARS Restrictions Apply to Government Use.2001
In Reply: Drs Ellis and Elasy suggest that
inclusion of studies lasting less than 13 weeks would result in underestimation
of the effect of exercise on glycemic control. To address this concern, we
performed 2 analyses. First, among the 11 exercise vs nonexercise control
comparisons, we performed a subgroup analysis comparing the 6 studies lasting
8 to 12 weeks with the 5 studies lasting more than 12 weeks. The weighted
mean differences in HbA1c were almost identical in the 2 groups
of studies (−0.65% [P = .005] and −0.67%
[P = .004], respectively). Second, we performed a
meta-regression analysis regressing postintervention difference in HbA1c on study duration and found no significant association (r = 0.07, P = .84). Therefore, the inclusion of studies lasting just
8 to 12 weeks did not result in an underestimation of the effects of exercise
on HbA1c. Had we excluded these briefer studies, our results would
have been unchanged but statistical power would have been reduced.
Sigal RJ, Boulé NG, Kenny G. Exercise and Glycemic Control in Diabetes—Reply. JAMA. 2001;286(23):2941–2942. doi:10.1001/jama.286.23.2940
Customize your JAMA Network experience by selecting one or more topics from the list below.