Letter to the EditorReply to Areta et al.: Time to withdraw and let the myth restB. Gualano, L. d. S. Gonçalves, V. d. S. Painelli, G. Yamaguchi, L. F. de Oliveira, B. Saunders, R. P. da Silva, E. Maciel, G. G. Artioli, and H. RoschelB. GualanoUniversity of Sao Paulo, São Paulo, Brazil, L. d. S. GonçalvesUniversity of Sao Paulo, São Paulo, Brazil, V. d. S. PainelliUniversity of Sao Paulo, São Paulo, Brazil, G. YamaguchiUniversity of Sao Paulo, São Paulo, Brazil, L. F. de OliveiraUniversity of Sao Paulo, São Paulo, Brazil, B. SaundersUniversity of Sao Paulo, São Paulo, Brazil, R. P. da SilvaUniversity of Sao Paulo, São Paulo, Brazil, E. MacielUniversity of Sao Paulo, São Paulo, Brazil, G. G. ArtioliUniversity of Sao Paulo, São Paulo, Brazil, and H. RoschelUniversity of Sao Paulo, São Paulo, BrazilPublished Online:22 Nov 2017https://doi.org/10.1152/japplphysiol.00567.2017MoreSectionsPDF (30 KB)Download PDF ToolsExport citationAdd to favoritesGet permissionsTrack citations ShareShare onFacebookTwitterLinkedInWeChat to the editor: It is undisputable that caffeine variation in coffee—which is not as dramatic in Brazil as reported elsewhere (2)—is a potential bias pertaining to our, and any, study assessing its habitual dietary intake. In fact, such variability exists in any natural food due to differences in environment, genetics, and food processing. Therefore, if one is to take the point made by Areta et al. (1), the entire literature involving assessment of natural foods should be disregarded, which is obviously not the case. The suggestion that “reported caffeine intakes” could be verified through direct measurement of caffeine is, at least, curious; considering that caffeine half-life is ~5 h, it is puzzling how this measure could reflect one’s habitual caffeine consumption.Notably, we carefully assessed caffeine intake following Food and Drug Administration recommendations for high-quality food composition assessment. Briefly, caffeine intake was determined using a questionnaire adapted (and not replicated, as incorrectly assumed) from a validated instrument; the questionnaire was completed with the help of a qualified nutritionist who extracted as much detail and specific information as possible. Exact brands of products were identified to determine exact caffeine content. Although Areta et al. claimed that our dietary assessment was flawed, no scientific evidence was provided to support this assumption.We used a 24-h caffeine withdrawal to isolate the acute effect of supplementation from diet. Areta et al. argue that this influenced our results. Nonetheless, meta-analytical data show that caffeine withdrawal does not influence the ergogenic effect of caffeine (3). Indeed, caffeine consumption has previously been shown to be equally effective in habitual consumers following no, 2 days, and 4 days of withdrawal. In fact, some of the authors themselves have confirmed that a withdrawal period from dietary caffeine does not enhance or reduce the effect of an acute dose of caffeine on exercise performance. Thus we feel justified in our use of a withdrawal period to ensure rigorous control of our protocol.Although we acknowledge that our rationale regarding increased adenosine upregulation has been shown only in experimental (animal) models, which we drew attention to within the text, if true as hypothesized, 24-h withdrawal will not have led to any significant reduction herein and therefore will unlikely have influenced results.The criticism of our sample size is unfounded: 1) the lack of correlation between habitual intake and performance improvements was based on our entire sample; and 2) even when broken down to habitual intake, each tertile showed a greater number of subjects than in previous studies addressing this topic. Certainly, the numerous analyses performed in this study mitigate the odds of type-2 error.Since Areta et al. doubt our conclusions due to supposed inaccuracies in caffeine measurement, we provide a requiem to their uncertainties by reanalyzing all data following removal of coffee as the “less stable” source of caffeine, thus considering only other sources (e.g., cola drinks, tea, supplements). Only two participants moved between tertiles, which remained significantly different regarding intake (P < 0.0001). Neither significant between-group differences in performance (P = 0.81) nor correlation between habitual intake and performance changes (r = 0.00023; P = 0.92) was shown. Finally, when participants were distributed into quintiles, with the lower quintile consuming no coffee at all and the higher quintile consuming 6–8 coffees/day, differences in performance were still absent. These sensitivity analyses reinforce our original conclusion that habitual dietary caffeine does not affect the response to caffeine supplementation (4), putting this myth to rest.GRANTSThis work was supported by CAPES-PROEX.AUTHOR CONTRIBUTIONSL.d.S.G., V.d.S.P., B.S., and B.G. drafted manuscript; L.d.S.G., V.d.S.P., B.S., G.G.A., H.R., and B.G. edited and revised manuscript; L.d.S.G., V.d.S.P., G.Y., L.F.d.O., B.S., R.P.d.S., E.M., G.G.A., H.R., and B.G. approved final version of manuscript.DISCLOSURESNo conflicts of interest, financial or otherwise, are declared by the authors.
Read full abstract