Reply to M Bes-Rastrollo and MA Martinez-Gonzalez

摘要

We appreciate the interest that Bes-Rastrollo and Martinez-Gonzalez have shown in our recent article in the Journal (1), and we welcome the opportunity to share our response. They mention 3 articles that were not included in our analysis. We did not mention these articles because they were published after our article was submitted for publication. Sichieri et al (2) found no statistically significant reduction in body mass index (P = 0.33) between the treatment and control groups, but they did find a statistically significant reduction in body mass index for the subgroup of females who were overweight at baseline. Libuda et al (3) also found no statistically significant association between soda consumption and measures of body mass index or body fat in most of their analyses. The only reported statistically significant association was for girls, and that relation was not significant at the 0.05 confidence level (0.055 SD score/MJ increase in regular soft drink consumption; P = 0.08) These results are entirely consistent with the findings from our meta-analysis. The third study did report a statistically significant association between the odds of being overweight and between-meal consumption of sugar-sweetened beverages (4); however, the study design was different from those included in our meta-analysis in that it examined only a subset of sugar-sweetened beverage consumption. On the basis of our sensitivity tests, we strongly doubt that the inclusion of these studies would make any substantive changes in our conclusions.