Abstract
Aim: The potential renal acid load (PRAL) has been described in relation to different health outcomes. Outcomes over time and conclusions made are often based on baseline dietary intake values. However, to study reliable long-term associations, parameters calculated based on dietary intake data, such as PRAL, must be stable over time. Therefore, the aim of the present study was to assess the stability of PRAL and its components over a 10-year time period. Methods: PRAL values of three-day dietary intake data from 197 women and 373 men on two assessment moments (2002–2004 and 2012–2014) were calculated. Pearson correlation and intra-class correlation coefficients were used for assessing the stability of PRAL and its components. Level of agreement between the two assessment moments was estimated after splitting up subjects in quintiles of PRAL, calculating kappa values and changes of quintiles over time. Results: No significant differences in mean PRAL over time were found. Stability of PRAL and its components was low. Poor agreement between the first and second assessment was shown by low kappa values and change of most of the subjects to an adjacent and non-adjacent quintile after 10 years. Conclusions: Based on nutrition assessments carried out using three-day dietary records, stability of PRAL over a 10-year time period could not be confirmed, even though no significant difference between mean PRAL and its components over time was found. Therefore, interpretation of longitudinal outcomes based on PRAL and its component calculated at baseline should be interpreted with caution.
| Original language | English |
|---|---|
| Pages (from-to) | 139-143 |
| Number of pages | 5 |
| Journal | Nutrition & Dietetics |
| Volume | 77 |
| Issue number | 1 |
| DOIs | |
| Publication status | Published - Feb 2020 |
Bibliographical note
Funding Information:Data collection was possible through funding by the Policy Research Centre Sport. The Policy Research Centre Sport is funded by the Flemish government. The present study used longitudinal data from the Flemish Policy Research Centre Sport, Physical Activity, and Health (data collected in 2002–2004 and 2012–2014). A random original sample of 1569 healthy volunteers between 18 and 75 years, sufficiently representative of the Flemish population in terms of geographic distribution, age, gender and educational level, was selected in 2002. All the participants of the present study received information about the tests and measures and signed an informed consent prior to participation. The study was approved by the ethical and medical committee of the KU Leuven. Sampling, dropout analysis, characteristics of the participants and dietary assessment analyses have been described in depth elsewhere. In brief, three-day dietary records, including two weekdays and one weekend day, from 570 healthy participants (373 men (65.4%) and 197 women (34.6%)) were available. Amounts of foods and drinks consumed were weighed if possible. If not, weighed estimates were made using standard household measures. Becel Nutrition software (Unilever Co.; Rotterdam, The Netherlands) was used to analyse dietary records. The calculated macro- and micronutrients were used to calculate PRAL for each participant on both measure points using the following formula : PRALmEq/day=0.49×proteing/day+0.037*phosphorusmg/day–0.021*potassiummg/day–0.026*magnesiummg/day–0.013*calciummg/day. Statistical analysis was performed using SPSS 24.0. To check the distribution of PRAL values, a Kolmogorov–Smirnov test was performed. In case of outliers for the different components affecting PRAL, a plausibility check with total kcal intake was performed. To check if mean protein intake and micronutrients used for calculating PRAL were in line with national recommendations a one-sample t-test was applied. Paired samples t-tests were performed to compare PRAL value and the different PRAL components between 2002–2004 and 2012–2014. Stability between the two assessments was computed using Pearson correlation and intra-class correlation coefficient (ICC). Stability is considered with a correlation of r ≥ 0.50, following Bloom's stability definition. To estimate the measure of agreement between the two test times, PRAL was categorised into quintiles using the 20–40–60–80 percentiles. Changes of PRAL were estimated using the Cohen's Kappa measure of agreement. Frequencies of participants in the same quintile, in an adjacent quintile and non-adjacent quintile for 2002–2004 versus 2012–2014, were calculated. Level of significance was set at p < 0.05. Findings of the study were reported in accordance with the STROBE-nut guidelines.
Publisher Copyright:
© 2018 Dietitians Association of Australia