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.ResultsSignificant 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].ConclusionCompared 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.

Background Coronavirus disease 2019 (COVID-19) elicits robust inflammatory reaction that may result in a declining albumin serum level. This meta-analysis aimed to evaluate the prognostic properties of hypoalbuminemia for poor prognosis and factors that may influence the relationship. Method A systematic literature search of PubMed was conducted from inception to April 22, 2021. The main exposure was albumin level below normal range–defined by the included studies. The outcome of interest was composite poor outcome that comprises of mortality, severity, and the requirement of mechanical ventilation or intensive care unit. Results There were 6200 patients from 19 studies. Meta-analysis showed that hypoalbuminemia was associated with composite poor outcome (OR 6.97 (95% CI 4.20–11.55), p<0.001; I 2 =91.3%, p<0.001). Meta-regression analysis showed that age (p=0.44), gender (p=0.76), HT (p = 0.97), DM (p = 0.40), CKD (p = 0.65), liver disease (p = 0.72), and malignancy (p = 0.84) did not affect the association. Subgroup analysis showed that hypoalbuminemia increased mortality (OR 6.26 (95% CI 3.26–12.04), p<0.001; I 2 =69.6%, p<0.01) and severity of the disease (OR 7.32 (95%CI 3.94-13.59), p<0.001; I 2 =92.5%, p<0.01). Pooled diagnostic analysis of hypoalbuminemia yielded a sensitivity of 0.63 (95% CI 0.52-0.72), specificity of 0.81 (95% CI 0.73-0.87), and AUC of 0.77. The probability of poor outcome was 70% in patients with hypoalbuminemia and 24% in patients with normal albumin level. Conclusion Hypoalbuminemia was associated with poor prognosis in COVID-19 patients.

Dietary and plasma carboxymethyl lysine (dCML, pCML) and plasma tumor necrosis factor-α (pTNF-α) may be associated with obesity in affluent society. However, evidence in women from low-middle income countries with predominantly traditional diets is lacking. We investigated the mediator effects of dCML, pCML and pTNF-α on body mass index (BMI) and waist circumference (WC) among Indonesian women. A cross-sectional study was conducted among 235 non-diabetic, non-anemic and non-smoking women aged 19–50 years from selected mountainous and coastal areas of West Sumatra and West Java. Dietary CML, pCML, pTNF-α were obtained from 2 × 24-h recalls, ultra-performance liquid chromatography-tandem mass spectrometry and enzyme-linked immunosorbent assay, respectively. Between-group differences were analyzed by the Chi-square or Mann-Whitney test and mediator effects by Structural Equation Modeling. The medians and interquartile-ranges of dCML, pCML and pTNF-α were 2.2 (1.7–3.0) mg/day, 22.2 (17.2–28.2) ng/mL and 0.68 (0.52–1.00) IU/mL, respectively, and significantly higher in the WC ≥ 80 cm than in the WC < 80 cm group, but not in BMI ≥ 25 kg/m2 compared to BMI < 25 kg/m2 group. Plasma CML and pTNF-α were positively and directly related to WC (β = 0.21 [95% CI: 0.09, 0.33] and β = 0.23 [95% CI: 0.11, 0.35], respectively). Dietary CML that correlated with dry-heat processing and cereals as the highest contributions was positively related to WC (β = 0.33 [95% CI: 0.12, 0.83]). Ethnicity, level of education, intake of fat, and intake of energy contributed to dCML, pCML and pTNF-α, and subsequently affected WC, while only ethnicity contributed to BMI through dCML, pCML and pTNF-α (β = 0.07 [95% CI: 0.01, 0.14]). A positive direct effect of dCML on pCML and of pCML and pTNF-α on WC was seen among Indonesian women. Dietary CML seems to have several paths that indirectly influence the increases in WC if compared to BMI. Thus, intake of CML-rich foods should be reduced, or the foods consumed in moderate amounts to avoid the risk of central obesity in this population.