Comparison of appetite assessment methods in older adults from the APPETITE study.

Abstract

Poor appetite is an important health concern in older adults. Numerous methods exist for appetite assessment, without a consensual gold standard. This study aims to compare Simplified Nutritional Appetite Questionnaire (SNAQ) and two single-item appetite questions with appetite ratings from visual analogue scales (VAS) and energy intake (EI). In 126 community-dwelling adults aged ≥65 years from the APPETITE trial, appetite was assessed using SNAQ, SNAQ 1^st item, Centre for Epidemiologic Studies-Depression scale (CES-D) 2^nd item and using VAS appetite ratings (fasting, 0, 30, 60, 120, 180 min post-breakfast, post-lunch) and EI (ad libitum lunch) in a test meal setting. Spearman correlations were calculated between SNAQ, single-items, VAS-fasting, and EI. Differences in VAS-fasting and EI between normal and poor appetite groups (based on SNAQ, single-items) were examined using Mann-Whitney-U-test. Repeated measures Generalized Linear Models were used to compare all VAS ratings across the test morning and post-breakfast response ratings with VAS-fasting as a covariate between appetite groups. SNAQ score was correlated with VAS-fasting (r=0.26, p<0.001) and EI (r=0.22, p<0.05). VAS-fasting was lower in the SNAQ-based poor appetite group (p=0.01). Time/group interaction effects (SNAQ, η_p²=0.02; CES-D-item, η_p²=0.02) for all VAS ratings, and group (SNAQ-item, η_p²=0.04) and interaction effects (CES-D-item, η_p²=0.03) for post-breakfast ratings were observed (p<0.05, respectively). SNAQ identified differences in VAS-fasting, possibly reflecting processes related to the drive to eat, while the two single-items identified appetite differences in response to a standardised breakfast. Different methods appear to capture different aspects of appetite, which should be considered when choosing an assessment method.