key: cord-0833028-ixsmbpwz authors: Belhadj, Moussa; Kazi Tani, Latifa Sarra; Dennouni Medjati, Nouria; Harek, Yahia; Dali Sahi, Majda; Sun, Qian; Heller, Raban; Behar, Ammaria; Charlet, Laurent; Schomburg, Lutz title: Se Status Prediction by Food Intake as Compared to Circulating Biomarkers in a West Algerian Population date: 2020-11-24 journal: Nutrients DOI: 10.3390/nu12123599 sha: 6e9d29ca5bdf1bad453c116ac91cc9ebf8af5a61 doc_id: 833028 cord_uid: ixsmbpwz Algeria is the largest country in Africa, located close to the Mediterranean coastal area, where nutrients consumption varies widely. Local data on selenium composition of foods are not available. We postulated a close correlation between selenium status predictions from food consumption analysis with a quantitative analysis of circulating biomarkers of selenium status. Population characteristics were recorded from 158 participants and dietary selenium intake was calculated by 24-h recall. The average total plasma selenium was 92.4 ± 18.5 µg/L and the mean of selenium intake was 62.7 µg/day. The selenoprotein P concentration was 5.5 ± 2.0 mg/L and glutathione peroxidase 3 activity was 247.3 ± 41.5 U/L. A direct comparison of the dietary-derived selenium status to the circulating selenium biomarkers showed no significant interrelation. Based on absolute intakes of meat, potato and eggs, a model was deduced that outperforms the intake composition-based prediction from all food components significantly (DeLong’s test, p = 0.029), yielding an area under the curve of 82%. Selenium status prediction from food intake remains a challenge. Imprecision of survey method or information on nutrient composition makes extrapolating selenium intake from food data providing incorrect insights into the nutritional status of a given population, and laboratory analyses are needed for reliable information. Selenium (Se) is an essential trace element for humans, mainly acquired through the daily diet [1] . The amount of Se in food items is variable, and it is hard to predict the Se content of a particular diet; its accumulation in plants depends on soil Se content and other soil parameters in a given area [2] . Dietary factors also determine the Se bioavailability; foods with high protein content (meat, fish, seafood) are characterised as better sources of Se, whereas high fat may impair bioavailability [3] . Most fruits and vegetables provide little Se because of their low content of protein and high content of water. The estimation of the Recommended Dietary Allowance (RDA) constitutes a considerable The amount and nature of each food item consumed during the last day was calculated in grams per day from a 24-h recall. Colour pictures of food samples with the weight indicated were provided to help participants make their choices as accurately as possible [29] . The quantities were converted into Se intakes using Ciqual (2017), a food nutritional composition table available online and free of charge [30] . Finally, all the answers were reviewed and completed if necessary. To estimate the contribution of different dietary sources to daily Se intake, the foods were divided into categories: fish and seafood, legumes (included lentils, beans, peas and chickpeas), eggs, meat, milk and dairy products, bread, fresh fruits, cereals (included pasta, rice, bread, couscous and all dishes made from flour or semolina), vegetables (included raw and cooked vegetables) and potatoes. Although potatoes and bread are included in vegetables and cereal, respectively, they were considered as separate food groups because of their high consumption rate. Circulating biomarkers of Se status were assessed in plasma, essentially as described [10, 31] . Blood samples were collected by venipuncture into 4 mL heparinised tubes, plasma and erythrocytes were separated by centrifugation at 1100× g (relative centrifugal force) using a Sigma 2-16P centrifuge for 15 min, and then they were frozen at −80 • C until the analyses were performed. Total Se concentration in plasma was determined by total reflection X-ray fluorescence (TXRF) analysis using a spiked gallium solution as standard and a benchtop TXRF analyser (S2 Picofox, Bruker nano GmbH, Berlin, Germany), and tested for accuracy by using a Seronorm serum standard (Sero AS, Billingstad, Norway) as described [32] . Plasma SELENOP was quantified by a validated commercial SELENOP-specific ELISA (selenOtest TM , selenOmed GmbH, Berlin, Germany) as previously described [33] . Enzymatic activity of plasma GPX3 was determined by a coupled enzymatic test, monitoring the consumption of NADPH at 340 nm [34] . Normal distribution of values was assessed by the Shapiro-Wilk test. Non-parametric test methods were assessed to investigate location shifts between groups (Mann-Whitney U test, Kruskal-Wallis test). Categorical variables were evaluated using Fisher's exact test. Relationship between parameters was tested by Spearman's correlation analysis. As this was an exploratory post-hoc analysis, all p-values were to be interpreted descriptively, and no adjustment for multiple testing was adopted. Variable selection was performed via stepwise AIC selection [35, 36] . Differences between ROC curves were assessed by the DeLong's test for two correlated ROC curves [37] . All statistical tests used an α-level of 0.05. The results were considered as statistically significant when the p-value was less than 0.05, and differences are marked as follows: p < 0.05 (*), p < 0.01 (**) and p < 0.001 (***). All statistical calculations were performed with R version 4.0.2 [38] , applying the packages "tidyr" [39] , "dplyr" [40] , and "pROC" [41] . Figures were created by using the package "ggplot2" [42] . The characteristics of the subjects enrolled in this cross-sectional study were assessed by face-to-face interviews. One hundred fifty-eight subjects were recruited in total. The mean age was 49 (CI: [46] [47] [48] [49] [50] [51] years, and the majority were female (83.7%). On average, participants were overweight, with a mean BMI of 26.8 (CI: 26.9-28.5) kg/m 2 , and a fraction of subjects were hypertensive, current smokers and in a stable marriage. The majority of samples indicated that the subjects were non-deficient in Se status, with a small fraction only (7.6%) exhibiting a plasma Se concentration below the consented threshold of deficiency, i.e., below 70 µg/L (median (IQR): 59.9 (21.8, 69.4) versus 93.7 (70.8, 143.2) µg/L, p < 0.001) ( Table 1 ). The groups were very similar, and neither the anthropometric nor the sociodemographic parameters tested indicated a significant difference between the groups of subjects classified as Se-deficient or Se-replete (Table 1) . As the subjects were divided into two groups according to plasma Se concentrations, plasma SELENOP levels showed the expected difference between the groups. The 24 h food recall data were used to quantify absolute food intakes per food category. Data were then converted into Se intakes by using the food composition information from the ANSES French Food Composition Table Ciqual 2017 , and compared between the two groups of Se-deficient (plasma Se < 70 µg/L) and the Se-replete (plasma Se > 70 µg/L) subjects (Table 2 ). The diet-specific comparison of the groups with replete or marginal Se status revealed no particular food item that turned out to be significantly associated with the different Se status. Even the calculated total Se intake in the groups was not different, when comparing the subjects with measured Se deficiency (plasma Se < 70 µg/L) as compared to those with higher plasma Se status. The dietary food intakes were converted to daily Se intakes by the ANSES French Food Composition Table as highlighted above (Table 2 ). To test whether the results align with the expectation, i.e., providing an estimate on the resulting Se status, a direct comparison of the Se intake data with the measured Se status biomarkers was conducted. To this end, the subjects were divided according to their predicted Se intake, choosing the median Se intake as the threshold, i.e., whether daily intake was below or above 55 µg/day. The results indicate that the prediction of Se status based on the calculated daily Se intake and consumption pattern does not align with the Se status biomarkers measured, i.e., neither with total plasma Se nor with the protein SELENOP ( Figure 1 ). The threshold for Se deficiency is generally considered to be a total serum or plasma Se concentration of 70 µg/L. Using this boundary, the total study cohort was divided into Se-deficient or Se-replete subjects. To analyse the interrelation of the two major Se status biomarkers, i.e., total plasma Se and SELENOP, this boundary was chosen to test the correlation of both biomarkers in the Se-deficient and Se-replete groups, respectively ( Figure 2 ). The analysis indicates that there is a relatively tight and positive correlation between plasma Se and SELENOP concentrations, particularly in the Se-deficient subjects, with a weaker interaction in Se-replete subjects with plasma Se concentrations > 70 µg/L. The threshold for Se deficiency is generally considered to be a total serum or plasma Se concentration of 70 µg/L. Using this boundary, the total study cohort was divided into Se-deficient or Se-replete subjects. To analyse the interrelation of the two major Se status biomarkers, i.e., total plasma Se and SELENOP, this boundary was chosen to test the correlation of both biomarkers in the Se-deficient and Se-replete groups, respectively ( Figure 2 ). The analysis indicates that there is a relatively tight and positive correlation between plasma Se and SELENOP concentrations, particularly in the Se-deficient subjects, with a weaker interaction in Se-replete subjects with plasma Se concentrations > 70 µg/L. In a subset of the samples (n = 98), we were able to analyse the GPX3 activity; the other samples had to be excluded for reasons of either insufficient residual volume or compromised sample quality. The results were correlated with the concentrations of Se (R = 0.16, p = 0.12) and SELENOP (R = 0.04, p = 0.69). There was no significant correlation between the estimated Se intake and the GPX3 activity in the set of samples analysed (R = 0.01, p = 0.95). In a subset of the samples (n = 98), we were able to analyse the GPX3 activity; the other samples had to be excluded for reasons of either insufficient residual volume or compromised sample quality. The results were correlated with the concentrations of Se (R = 0.16, p = 0.12) and SELENOP (R = 0.04, p = 0.69). There was no significant correlation between the estimated Se intake and the GPX3 activity in the set of samples analysed (R = 0.01, p = 0.95). Finally, the data were used to model Se status from the data on food intake (amount and food categories) in relation to the measured biomarkers of Se status. The analyses indicate that information on the food categories eggs, meat and potatoes provided the most reliable match and outperformed any other combination of variables when compared via stepwise AIC selection (Figure 3 ). Correlation analysis of plasma Se with SELENOP concentrations. All of the available plasma samples (n = 134) of the patients enrolled were analysed for total plasma Se and SELENOP concentrations. The samples were separated into two groups based on total plasma Se deficiency into Se-deficient (<70 µg/L, green) and Se-replete (>70 µg/L, red). The biomarkers showed a significant and positive linear correlation (Spearman, R = 0.79, p = 0.0036) in the Se-deficient samples, whereas Sereplete subjects revealed a non-significant, positive correlation (Spearman, R = 0.06, p = 0.47). In a subset of the samples (n = 98), we were able to analyse the GPX3 activity; the other samples had to be excluded for reasons of either insufficient residual volume or compromised sample quality. The results were correlated with the concentrations of Se (R = 0.16, p = 0.12) and SELENOP (R = 0.04, p = 0.69). There was no significant correlation between the estimated Se intake and the GPX3 activity in the set of samples analysed (R = 0.01, p = 0.95). Finally, the data were used to model Se status from the data on food intake (amount and food categories) in relation to the measured biomarkers of Se status. The analyses indicate that information on the food categories eggs, meat and potatoes provided the most reliable match and outperformed any other combination of variables when compared via stepwise AIC selection (Figure 3) . The essentiality of Se for human health is well established, and population-wide intake and status information is of high importance for the health care systems. However, the respective data are hard to obtain, and the best way to perform such analyses and how to predict Se status reliably has been intensively discussed [4, 43, 44] . In this study, we decided to compare nutritional Se intake prediction to laboratory analysis of Se status biomarkers in a North African population from Western Algeria. Our results indicate that the population on average consumes a wide variety of food items with some potentially Se-rich ingredients like sea food, meat, eggs and milk and leguminous plants. This impression is supported by the laboratory analyses of biomarkers of Se status including the most established parameters, i.e., total plasma Se, SELENOP and GPX3 [7, 12, 45] . Using the consented threshold for Se deficiency, i.e., serum or plasma Se concentrations below 70 µg/L, only a small fraction of less than 10% of subjects needed to be classified as insufficiently supplied with the essential trace element. However, there was no meaningful concordance when comparing the deduced Se status from the food intake patterns in combination with the food composition database with the measured biomarkers of Se status from the plasma samples. The most likely explanation for the observed mismatch between deduced values and measured concentrations lies in our assumption that using food composition data on Se contents of the different food categories would faithfully mirror the quality and Se content of the food items that have been consumed by the study participants. This assumption Nutrients 2020, 12, 3599 8 of 12 and strategy may yield accurate results for fat, carbohydrate or protein intakes, but unfortunately not for the trace element Se that presents itself again as difficult to grasp and predict, likely due to its complex geochemistry and uneven distribution [46] [47] [48] . Our laboratory analyses yielded average plasma Se and SELENOP concentrations in a range similar to what we determined in different European populations [31] . We did not observe a significant difference between men and women, which was in agreement with other independent studies on micronutrient status in Algeria [49, 50] , and also in agreement with other large population-wide studies in Europe [31] , the US [51] or in Se-deficient or Se-replete areas of China [52] . Moreover, we did not observe a higher Se status in married as compared to single subjects, in contrast to a recent study [53] . The challenge of predicting Se intake from food frequency data is not new, and other attempts have similarly struggled with poor congruence, e.g., a respective study conducted in Finland [54] . The major reason for the inconclusiveness lies most likely in the varying Se content of a given food item, as it depends mostly on the area where it was produced and the respective soil quality and Se content [55, 56] . In Algeria, most imported food groups are cereals (including wheat, meslin and corn) which cover more than 70% of its cereals needs. The cereals are grown in different regions of the world, mainly in America and Europe [57] . Similarly, milk, dairy products and legumes from different areas of the world contribute strongly to the Algerian nutrition [58] . The variation in the import origin of these products can be expected to have an impact on our analyses, as it causes strongly varying Se concentrations in the dietary items that formed the basis for our intake assessments and predictions [59] . On top of the variable international origin of the food items consumed in Algeria, local differences in Se content of the same nutrients are also known. Taking wheat as an example, a concentration range from as low as 21 µg/kg in Tiaret (western Algeria) to as high as 153 µg/kg in Khroub (eastern Algeria) has been reported in an Algerian study [60] . According to the Algerian Interprofessional office for Cereals, France is the main foreign supplier of cereals to Algeria [61] . French soil, as well as soils in other European countries, are rather poor sources of Se (with average Se contents as follows: France; 0.03 mg/kg, Finland; 0.08 mg/kg, Belgium; 0.11 mg/kg, Scotland; 0.17 mg/kg, Sweden; 0.30 mg/kg, and Norway; 0.63 mg/kg [62] ) and are considered to be Se deficient [63] . Soils in other areas of the world, e.g., in the United States of America, can be richer sources of Se, with concentrations of up to 0.95 mg/kg [64] . Our data therefore highlight the need for laboratory-based analyses of Se status in a representative sample of a given population, and the challenge when trying to deduce Se status from nutritional intake data. Moreover, the data agree with prior studies reporting a relatively moderate Se status in Western Algeria, with a small fraction of subjects only displaying an insufficient daily intake. On the one hand, the globalization of the food industry and the associated transport of food items across the world pose environmental problems and contribute to climate change, but on the other hand these transports distribute the micronutrients more evenly across the populations and also into regions at risk of low supply. This noteworthy development clearly hinders food frequency-based predictions and complicates nutritional intake analyses, but it also contributes to better health by preventing severe deficiencies in areas where certain micronutrients are sparse. The complex origin of dietary Se in the average Algerian food serves as a most instructive example for this notion. It appears impossible at present to correctly predict the average Se intake or resulting Se status of a given population from food intake information alone, at least as long as specific information on Se content of individual food items is not provided by the producers. Consequently, laboratory analyses of a representative sample of the population are needed to obtain the required information. To this end, different Se status biomarkers have been established and are available, and the results obtained usually agree reliably, especially in subjects with low Se status where insufficient intake causes low plasma Se levels and suppressed selenoprotein expression. Still, it would be helpful both for the health authorities and for the consumers alike to find specific information on the micronutrient contents on the commercial food items. This information should be provided at least for those nutrients that are A review of dietary selenium intake and selenium status in Europe and the Middle East Overview of Selenium Deficiency and Toxicity Worldwide: Affected Areas, Selenium-Related Health Issues, and Case Studies Current Knowledge on the Importance of Selenium in Food for Living Organisms: A Review Biomarkers of selenium status Regulation of Selenium Metabolism and Transport Insights for Setting of Nutrient Requirements, Gleaned by Comparison of Selenium Status Biomarkers in Turkeys and Chickens versus Rats, Mice, and Lambs The other view: The trace element selenium as a micronutrient in thyroid disease, diabetes, and beyond. Hormones Selenoproteins: Molecular pathways and physiological roles Hierarchical regulation of selenoprotein expression and sex-specific effects of selenium The choice of biomarkers determines the selenium status in young German vegans and vegetarians Optimization of selenoprotein P and other plasma selenium biomarkers for the assessment of the selenium nutritional requirement: A placebo-controlled, double-blind study of selenomethionine supplementation in selenium-deficient Chinese subjects Establishing optimal selenium status: Results of a randomized, double-blind, placebocontrolled trial German Nutrition, S. Revised reference values for selenium intake Systematic study of the selenium fractionation in human plasma from a cancer prevention trial using HPLC hyphenated to ICP-MS and ESI-MS/MS Dietary Reference Intakes for Vitamin C, Vitamin E, Selenium, and Carotenoids Selenium in diet, blood, and toenails in relation to human health in a seleniferous area Endemic selenium intoxication of humans in China Selenium toxicity from a misformulated dietary supplement, adverse health effects, and the temporal response in the nail biologic monitor Selenium and selenocysteine: Roles in cancer, health, and development Selenoprotein-P Deficiency Predicts Cardiovascular Disease and Death Selenium deficiency and viral infection Selenoprotein P status correlates to cancer-specific mortality in renal cancer patients systemic immune response syndrome, sepsis, and outcome in critically ill patients Bogner-Flatz, V. Selenium and Selenoprotein P Deficiency Correlates With Complications and Adverse Outcome After Major Trauma Association between regional selenium status and reported outcome of COVID-19 cases in China Selenium Deficiency Is Associated with Mortality Risk from COVID-19 Dietary assessment methods for intakes of iron, calcium, selenium, zinc and iodine Food-chain selenium and human health: Emphasis on intake Voici Quelques Photos Pour Aider vos Patients à Estimer les Quantités ANSES French Food Composition Table Ciqual Selenium status is associated with colorectal cancer risk in the European prospective investigation of cancer and nutrition cohort Determinants of selenium status in healthy adults Sex-specific and inter-individual differences in biomarkers of selenium status identified by a calibrated ELISA for selenoprotein P Assays of glutathione peroxidase Variable selection with stepwise and best subset approaches Likelihood of a model and information criteria Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing Easily Install and Load the 'Tidyverse Dplyr: A Grammar of Data Manipulation pROC: An open-source package for R and S+ to analyze and compare ROC curves Elegant Graphics for Data Analysis Assessment of requirements for selenium and adequacy of selenium status: A review Selenoprotein P-expression, functions, and roles in mammals EPA's selenium revision heads into murky waters The labour pains of biochemical selenology: The history of selenoprotein biosynthesis Selenium cycling across soil-plant-atmosphere interfaces: A critical review Whole blood selenium levels in healthy adults from the west of Algeria Blood micronutrients in Algeria, relationships with sex and age Serum Selenium Levels in the US Population: Third National Health and Nutrition Examination Survey Low Population Selenium Status Is Associated with Increased Prevalence of Thyroid Disease Relationship between serum selenium, sociodemographic variables, other trace elements and lipid profile in an adult Spanish population Reproducibility and validity of a food frequency questionnaire in a case-control study on breast cancer Selenium in global food systems Selenium accumulation in plant foods Rapport de Synthèse sur L'agriculture en Algérie Premier Trimestre de L'année. 2020 Biogeochemistry of selenium and its impact on food chain quality and human health Selenium content in wheat and estimation of the selenium daily intake in different regions of Algeria Algeria-Company Information Selenium Distribution and Translocation in Rice (Oryza sativa L.) under Different Naturally Seleniferous Soils Associations of Spatial Disparities of Alzheimer's Disease Mortality Rates with Soil Selenium and Sulfur Concentrations and Four Common Risk Factors in the United States Global virtual water trade and the hydrogical cycle: Patterns, drivers, and socio-environmental impacts Selenium deficiency risk predicted to increase under future climate change We thank Vartitér Seher for excellent technical assistance. We acknowledge financial support by the Open Access Publication Fund of Charité-Universitätsmedizin Berlin.Conflicts of Interest: L.S. holds shares in selenOmed GmbH, a company involved in Se status assessment and supplementation. The other authors declare no competing interest with respect to this study.