Dr Charlie Simpson

Discrepancies in energy and macronutrient intakes between tests are apparent even when a solid prepackaged diet (Sdiet) is used to standardize dietary intake for preexperimental trials. It is unknown whether a liquid prepackaged diet (Ldiet) leads to improved adherence, resulting in lower variability in energy and macronutrient intakes. This study assesses the ability of athletes to replicate a diet when an Ldiet or Sdiet was used as a dietary standardization technique. In a cross-over design, 30 athletes were randomly assigned to either Sdiet or Ldiet. Each diet was consumed for two nonconsecutive days. Participants were instructed to consume all the meals provided and to return any leftovers. The coefficient of variation (CV) was calculated for each nutrient for the two methods and reported as the average CV. The Bland–Altman plots show that differences between Days 1 and 2 in energy and macronutrient intakes for both diets were close to zero, with the exception of some outliers. The %CV for Sdiet was higher than Ldiet (5% and 3% for energy, 5% and 3% for carbohydrate, 5% and 2% for protein, and 5% and 3% for fat, respectively). There was a strong positive correlation for energy and all macronutrients between Days 1 and 2 for both methods (r > .80; p

Participants are often asked to replicate their diet before each trial to control dietary intake. However, little is known about the reproducibility of the diet using different approaches. The aim was to assess the reproducibility of a diet when a controlled diet (Cdiet), food record (Frecord) and 24-h dietary recall (Drecall) were used as dietary standardisation techniques.

Subjects/Methods:

Thirty athletes completed six visits to the laboratory. On the first occasion, a Drecall was performed and subjects were asked to replicate exactly the same diet on the day before visit 2, when another Drecall was performed. The day before the third visit, subjects completed a Frecord, which was presented at visit 3 and assessed using Drecall to ensure comparability between methods. Subjects were asked to replicate this Frecord before visit 4, which was assessed using Drecall. Finally, subjects were provided with a Cdiet of known composition, which they consumed for 24 h before visits 5 and 6. For each method, the difference in energy and macronutrient intakes between both occasions was measured.

Results:

Despite finding no differences in mean energy and macronutrient intakes between visits for any technique, important within-subject differences were apparent. The range of percentage coefficient of variation for all variables was between 2.7 and 5.8% for Cdiet, 10.1 and 18.6% for Frecord and 7.1 and 11.7% for Drecall.

Conclusions:

This study has shown that Cdiet is the best approach to standardise dietary intake, especially when the expected effect of an intervention is small and an enhanced reliability is required.