Repeated 24 Hour Dietary Recalls for the Assessment of Energy and Added Sugar Intakes: Methodological Considerations

Abstract

Non-communicable diseases (NCDs) pose a growing global health challenge, with Lebanon experiencing a high NCD-related mortality rate of 83%. The pivotal role of diet, such as excessive intakes of energy, fats, salt, and added sugar (AS), has been established by several studies as a major contributor to this public health crisis. It is therefore imperative to have proper ways of assessing dietary intakes at the population level. The limited availability of food composition databases (such as for AS) and the challenges associated with certain dietary assessment methods further complicate dietary assessment strategies.

This study uses four repeated 24-hour dietary recalls (24 HRs), with the aim of assessing dietary energy intakes (EI) in a sample of Lebanese adults, comparing EI estimates to total energy expenditure (TEE) based on indirect calorimetry and physical activity (PA) assessment, and determining AS intake. Another objective is to evaluate the consistency of the four 24 HRs in terms of energy, macronutrients, and AS intakes, and to compare energy estimates generated by four vs. two 24 HRs. This will allow us to evaluate whether four 24 HRs are needed, or if two non-consecutive 24 HRs would be sufficient to assess the participants’ dietary intakes.

Conducted as an observational study on a convenience sample of 77 participants from AUB, the research spanned over a 4-week timeframe for each participant. Assessments included four repeated 24 HRs (including one weekend day), completion of forms and questionnaires, PA assessment using the IPAQ, as well as BMR and anthropometric measurements. Analysis of the four 24 HRs was done using the Nutritionist Pro (NutriPro) software (version 7.1.0, 2019, Nutritionist Pro, Axxya Systems, USA) to estimate energy and macronutrients’ intakes, while AS intake was determined using the 10-step approach published by Louie et al (given that comprehensive information on AS content is not available in the food composition database). Subsequent data analysis was conducted using SPSS.

Results showed that average EI was 2260.91 ± 875.073 kcal/day for the total population, 2690.026 ± 831.94 kcal/day for males, and 1670.88 ± 521.404 kcal/day for females. As for TEE, it was estimated at 2560.26 ± 676.61 kcal/day for the total population, 2926.73 ± 1100.98 kcal/day for males, and 2198.4 ± 427.26 kcal/day for females. AS sugar intake was estimated at 43.9 ± 27.32 g/day for the total population, 49.066 ± 29.093 g/day for males, and 36.8 ± 23.28 g/day for females, representing 7.83% of EI in the total population, 7.14% in men and 8.79% in women. Data showed that when comparing the four different 24HRs, there was no significant difference in EI, AS or macronutrient intakes between the days. Similarly, no significant difference was observed when comparing two vs. four 24 HRs in terms of energy, macronutrients and AS intakes. Although there was a significant moderate correlation between EI and TEE (spearman correlation = 0.37), there was a significant mean difference between EI and TEE in women (-527.52 ± 641.52 kcal/day), individuals who have completed high school level education nor higher (-402.2 ± 1094.83 kcal/day), and individuals with a higher BMI (-707.94 ± 1208.78 kcal/day). Multiple linear regression was conducted to examine the predictors of the mean difference (EI-TEE). Results showed that only BMI remained significantly associated with the difference between EI and TEE [beta (95% CI): -869.18 (-1476.35; -262.02)] after the adjustment for other confounders, with 71.9% of people with high BMI having EI<TEE.

These study findings showed a consistency in the reporting of dietary recalls across four different days distributed along a month period, and suggest that two repeated recalls may be sufficient for the assessment of dietary intake in Lebanese adults. This was observed for total EI, macronutrient intakes as well as AS intake. Only BMI was identified as a significant predictor of the mean difference between EI and TEE in the study sample - a finding that has been reported by several previous studies conducted in other parts of the world. Larger and longer studies are needed to further confirm the results of this study.