The relationship between appetite scores and subsequent energy intake: An analysis based on 23 randomized controlled studies

Abstract

Background/Objectives: Several studies have assessed relationships of self-reported appetite (eating motivations, mainly by Visual Analogue Scales, VAS) with subsequent energy intake (EI), though usually in small data sets with limited power and variable designs. The objectives were therefore to better quantify the relationships of self-reports (incorporating subject characteristics) to subsequent EI, and to estimate the quantitative differences in VAS corresponding to consistent, significant differences in EI. Subjects/Methods: Data were derived from an opportunity sample of 23 randomized controlled studies involving 549 subjects, testing the effects of various food ingredients in meal replacers or 100–150 ml mini-drinks. In all studies, scores on several VAS were recorded for 30 min to 5 h post-meal, when EI was assessed by ad libitum meal consumption. The relationships between pre-meal VAS scores and EI were examined using correlation, linear models (including subject characteristics) and a cross-validation procedure. Results: VAS correlations with subsequent EI were statistically significant, but of low magnitude, up to r = 0.26. Hunger, age, gender, body weight and estimated basal metabolic rate explained 25% of the total variance in EI. Without hunger the prediction of EI was modestly but significantly lower (19%, P < 0.001). A change of ≥15–25 mm on a 100 mm VAS was the minimum effect consistently corresponding to a significant change in subsequent EI, depending on the starting VAS level. Conclusions: Eating motivations add in a small but consistently significant way to other known predictors of acute EI. Differences of about 15 mm on a 100 mm VAS appear to be the minimum effect expected to result in consistent, significant differences in subsequent EI.