Purpose: To compare total fluid intake (TFI), defined as the sum of water and all other fluid types, assessed with a 24-h dietary (food and fluid) recall with mean TFI assessed with a 7-day fluid-specific record among adolescents and adults.
Methods: This repeated cross-sectional study compared TFI as assessed by two fluid assessment instruments using a crossover approach. 290 adolescents (17.3 ± 0.8 years, 50 % boys) and 289 adults (43 ± 9.3 years, 50 % men) from Indonesia completed the study.
Results: Significant correlations were observed between fluid intake assessed with the 24-h recall and the 7-day fluid record (r = 0.333; p < 0.001). The Bland-Altman method, however, showed an underestimation (bias) of mean TFI by a 24-h recall when compared with the 7-day fluid record [mean difference (95 % CI) -382 mL (-299, -465); p < 0.001]. The mean difference also increased with increasing TFI: Mean difference for the lowest and highest quartiles of TFI was 139 versus -1265 mL/day. The 7-day fluid record recorded two (95 % CI -1.9, -2.4; p < 0.0001) extra drinking acts compared with the 24-h recall, whereas the mean volume per drinking act was significantly higher with the 24-h recall [mean difference (95 % CI) 39 mL (31, 47); p < 0.001].
Conclusion: Compared with a 7-day fluid record, a 24-h dietary recall significantly underestimated TFI. Subjects recalled two less drinking acts, while estimating the volume consumed per drinking act to be larger. Since the adequate intakes for total water intake are based on median intakes observed in national surveys that most frequently used the 24-h recall method, they may potentially be underestimated.