Reliability of hunger-related assessments during 24-hour fasts and their relationship to body composition and subsequent energy compensation

Many diets employ regular periods of fasting that extend beyond a typical overnight fast (i.e. intermittent fasting [IF]). Evaluation of acute fasting responses provides information concerning the potential theoretical rationale for IF. The purpose of the present investigation was to assess the test-retest reliability of hunger-related variables during 24-hour fasts and the relationship between these variables and body composition, as well as subsequent energy intake (EI) after fasting. Eleven participants (6 F, 5 M) completed two 24-hour fasts after being provided a 3-day standardized weight-maintenance diet. From 16 to 24 h of fasting, participants were directly observed and provided hourly assessments of hunger, desire to eat (DTE), prospective food consumption (PFC), fullness and energy. After the fast, participants were allowed ad libitum food consumption, and compensation was calculated as EI relative to weight-maintenance energy needs. Test-retest reliability for hunger-related assessments at particular durations of fasting was evaluated using intraclass correlation coefficients (ICC), changes in dependent variables were evaluated using ANOVA with repeated measures, and relationships between variables were explored using bivariate correlations. At 16 h of fasting, the ICCs for all hunger-related assessments were statistically significant (r = 0.67-0.91; p ≤ 0.05). However, as the fast progressed, reliability varied substantially. When averaged across the nine measurements, the ICCs were: 0.81 (fullness), 0.74 (PFC), 0.67 (energy), 0.44 (DTE) and 0.36 (hunger). Body fat percentage was significantly correlated with changes in PFC (r = 0.62, p = 0.04), hunger (r = 0.66, p = 0.03), DTE (r = 0.71, p = 0.02), and fullness (r = −0.63, p = 0.04), but not energy (r = −0.16, p = 0.64). Average EI compensation was only 60% of weight-maintenance needs, but substantial variability was observed (7 to 110% compensation). Compensation was significantly correlated with changes in PFC (r = 0.72, p = 0.01), hunger (r = 0.63, p = 0.04) and DTE (r = 0.60, p = 0.05), but not fullness (r = 0.58, p = 0.06) or energy (r = 0.35, p = 0.29). Compensation and body fat percentage were not correlated (r = 0.03, p = 0.94). The percent of energy intake from fat and protein increased after the fast (29.9 to 37.3% and 13.8 to 16.8%; p < 0.05), while the percent of energy intake from carbohydrate decreased (56.4 to 46.0%; p = 0.02). These results may have implications for IF programs. It is possible that the implementation of multiple “test fasts,” in which subjective variables and subsequent energy intake are evaluated, could be used to identify candidates who may be more likely to benefit from an IF program.