High prevalence of vitamin D deficiency among school-age children in Tunisia: a first national cross-sectional study Authors: Radhouene Doggui, Jalila El Ati, Besma Mourou, Myriam El Ati-Hellal Published: 2026 ABSTRACT AbstractGlobally, vitamin D deficiency (VDD) represents a significant public health concern. Population-level data on vitamin D status in Tunisia are lacking. This study aimed to estimate the national prevalence of VDD among Tunisian school-aged children and identify associated factors. A national cross-sectional survey in 2023 randomly sampled 2670 children aged 6–12 years. Plasma 25-hydroxyvitamin D was measured. Dietary vitamin D (VD) intake and socio-demographics were assessed via questionnaires. Logistic regression identified factors associated to inadequate VD status. The prevalence of VDD (< 30 nmol/L) was 15.8%. Adjusted analysis showed that older children (10–12 years) had a 1.4-fold (P = 0.030) higher likelihood of VDD than younger children. Girls were twice as likely to be VD deficient as boys (adjusted relative prevalence ratio (aRPR) = 2.1; P = 0.001). Significant regional variations in VDD were observed (P < 0.0001). Low and moderate consumption of VD-rich foods was associated with significantly increased odds of severe deficiency (aRRR = 2.0, P = 0.026; aRRR = 1.6, P = 0.005, respectively) relative to high consumption. Finally, overweight status was associated with a 2.0-fold (P < 0.0001) increased odds of severe VDD compared to children with normal weight. VDD is prevalent and increases with age. Vitamin D supplementation strategies should be considered to optimize vitamin D status among children.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-54650-6.Keywords: 25-hydroxyvitamin D, Youths, Adolescents, Micronutrients, Determinants, Nutrition transitionSubject terms: Diagnostic markers, Risk factors, Nutrition, Public health Introduction Vitamin D, a fat-soluble prohormone, is crucial for regulating calcium and phosphorus homeostasis, thereby maintaining bone health by facilitating gastrointestinal calcium absorption and interaction with bone-resorbing cells1. Furthermore, vitamin D exerts pleiotropic effects on various extra-skeletal systems, and its deficiency has been associated with a range of health issues, including muscle pain and weakness, certain cancers, cardiovascular risk, autoimmune diseases, and infections2. Tunisia is a North African country located between approximately 30° North and 37° North latitude, within the Mediterranean basin. This geographic position is characterized by relatively high solar irradiance throughout the year. Under such conditions, cutaneous synthesis of vitamin D via exposure to ultraviolet B (UVB) radiation is expected to be sufficient during most months, particularly from spring to early autumn3,4. However, effective vitamin D production depends not only on ambient UV radiation but also on behavioral and environmental factors, including time spent outdoors, clothing practices, air pollution, and degree of urbanization. Seasonal variation in UVB exposure persists, with lower intensity during winter months, which may limit endogenous vitamin D synthesis despite the country’s generally sunny climate1. Despite the dual sources of vitamin D, i.e., cutaneous synthesis upon exposure to ultraviolet B radiation and dietary intake from foods such as fish, egg yolks, and fortified products, vitamin D deficiency remains a pervasive global public health concern5,6. A recent meta-analysis of 308 studies involving 7.9 million adults across 81 countries reported alarming global prevalence rates of vitamin D deficiency: 15.7% using a cut-off of < 30 nmol/L, 47.9% using a cut-off of < 50 nmol/L, and 76.6% using a cut-off of < 70 nmol/L7. Populations living in low-middle-income countries seem to be more vulnerable to vitamin D deficiency, 27% for serum 25(OH)D concentrations < 30 nmol/L, nearly two-fold higher than that observed in high-income countries7. Children, characterized by rapid growth and sharp physiological requirements, are particularly susceptible to vitamin D deficiency8. Indeed, studies are indicating that African children (2 days to 17 years)9 and those (< 5 years) living in the Eastern Mediterranean region are exposed vitamin D deficiency varying from 10 to 80% (based on different cut-off values)10. Acknowledging the widespread prevalence of vitamin D deficiency and its association with numerous adverse health outcomes, many national and international health authorities have underscored the critical need to enhance vitamin D intake11. The definition of vitamin D deficiency remains debated; however, thresholds below 25–30 nmol/L are widely recognized as indicative of increased risk of rickets and impaired musculoskeletal health in children. These cut-offs are supported by the Scientific Advisory Committee on Nutrition, which recommends maintaining serum 25(OH)D concentrations above this level for bone health12. However, there is no universal consensus regarding optimal serum 25-hydroxyvitamin D [25(OH)D] concentrations. While some expert groups, including the Institute of Medicine/National Academy of Medicine, define sufficiency primarily based on skeletal outcomes using thresholds around 50 nmol/L, other organizations such as the Endocrine Society recommend higher thresholds (≥ 70 nmol/L), particularly in clinical contexts and for potential extra-skeletal effects. This variability in definitions should be considered when comparing prevalence estimates across studies and populations. Like many countries from the Middle-Eastern and North African region, Tunisia is experiencing a nutrition transition marked by shifts in dietary patterns, increased sedentary behaviour, and rising obesity rates13,14. The situation may enhance the likelihood of suboptimal micronutrient intakes15. Existing data on vitamin D levels in Tunisia remain fragmented and primarily limited to small-scale, monocentric, or clinical studies focusing on specific subgroups such as neonates, women, or individuals with chronic diseases16–21. For example, a study conducted among 907 neonates reported a prevalence of approximately 57% (25(OH)D < 25 nmol/L)16. There has been no nationally representative assessment of vitamin D status among Tunisian school-age children (SAC). Such data are crucial for informing targeted policy interventions and designing effective nutrition programs. Therefore, the present study aims to: (i) estimate the prevalence of vitamin D deficiency in a nationally representative sample of Tunisian children aged 6 to 12 years and, (ii) identify associated sociodemographic, dietary, and lifestyle factors contributing to the risk of deficiency and insufficiency. Methods Study design and population A cross-sectional study was conducted among SAC between May and October 2023. The study targeted children aged 6–12 years because primary school enrollment in Tunisia is very high, reaching 99.2% during the 2021–2022 academic year22. Eligible participants were children enrolled in primary schools who provided assent and whose parents or legal guardians provided written informed consent. Children whose parents or legal guardians refused consent were excluded from the study. The minimum required sample size was calculated using the following formula: where n is the required sample size; t is the statistic corresponding to the confidence level (1.96 for a 95% confidence interval); p is the estimated prevalence of the outcome; d is the margin of error; and DEFF is the design effect. Since the prevalence of vitamin D deficiency among Tunisian SAC was unknown, a conservative prevalence estimate of 50% was used. With a margin of error of 5%, the minimum sample size for a simple random sample was estimated at 385 children. To account for the clustered school-based sampling design, the sample size was adjusted using a design effect of 3.5, resulting in a minimum required sample of approximately 1350 children. To further increase statistical power and compensate for potential missing data and unequal exposed/unexposed ratios, the sample size was doubled, yielding a final target sample of 2700 children. A proportional stratified cluster sampling design was used to ensure national representativeness. Stratification was based on the 24 Tunisian governorates, urban and rural settings, and public and private schools, according to the classification of the Ministry of Education. A two-stage cluster sampling procedure was then implemented23. In the first stage, 87 primary schools (Supplementary Fig. 1) were randomly selected from a comprehensive stratified list of schools, proportionally reflecting the national distribution of public and private schools (90.5% and 9.5%, respectively). In the second stage, 30 students per school were systematically selected from the 2022–2023 enrollment lists, with balanced representation by sex and age group (15 boys and 15 girls per school). Measurements Data were collected through face-to-face interviews with children and their parents or legal guardians at the schools. Information on children’s diet was obtained from their parents or legal guardians, as children under the age of 10 generally have difficulty reporting such information reliably. Vitamin D status Vitamin D status was assessed by measuring serum 25-hydroxyvitamin D (25(OH)D) concentrations, following the International Federation of Clinical Chemistry and Laboratory Medicine recommendations. Whole blood samples were collected via venipuncture into 4 ml EDTA tubes between 9 and 10 a.m. All samples were kept at 4–5 °C and were sent to the Research Laboratory SURVEN of the National Institute of Nutrition and Food Technology, Ministry of Health (Tunisia). The total serum 25(OH)D concentrations were quantified using the Mini VIDAS enzyme-linked fluorescent assay (ELFA) automated immunoassay system, employing the Enzyme-Linked Fluorescent Assay technique. The measuring range of the analytical method is 20.25 to 340 nmol/L. The inter-assay coefficients of variation for quality control samples ranged between 5 and 10% across all batches. However, participation in an external quality assessment scheme such as the Vitamin D External Quality Assessment Scheme was not conducted during the study period, and calibration against reference methods such as liquid chromatography–tandem mass spectrometry (LC–MS/MS) was not available. These aspects may limit comparability with studies using standardized reference methods. Vitamin D deficiency was defined as serum 25(OH)D < 30 nmol/L, a threshold associated with a significantly increased risk of osteomalacia and nutritional rickets25–28. Insufficient vitamin D status was defined as a concentration of 25(OH)D ≥ 30 to < 50 nmol/L, and optimal level as ≥ 50 nmol/L28. Results were also expressed using the 30 nmol/L and 50 nmol/L cut-off, as recommended by other organizations, such as the Institute of Medicine in the United States29. These thresholds were selected to ensure comparability with widely used public health definitions based on skeletal outcomes; however, alternative higher cut-offs have been proposed for clinical purposes. Dietary intake A semi-quantitative food frequency questionnaire (FFQ) assessed daily vitamin D intake. The FFQ was developed based on the methodology described by Morabia et al.30, ensuring the inclusion of food items and dishes that contribute to approximately 90% of vitamin D intake in the target population of children aged 6 to 11 years old. The validity of the questionnaire regarding the main food items and dishes contributing to daily vitamin D intakes was achieved using the dietary data from a previous representative survey conducted in 2019 among the same population of Tunisian SAC (detailed data not shown)31,32. Portion sizes were estimated based on the distribution of portions recorded in the 2019 survey. Three portion categories were defined for each food item: ‘reference’ (median portion size), ‘small’ (first quartile), and ‘large’ (third quartile). These categories were incorporated into the FFQ to improve the accuracy of dietary intake quantification. The frequency and quantity of vitamin D-rich food was assessed, such as fish and seafood, beef, lamb, chicken, turkey, liver, dairy products, butter, margarine, and eggs. Consumption frequency was evaluated over a one-month recall period. Nine predefined response categories were used to report dietary intake: ‘Never or less than once per month,’ ‘Once per month’, ‘2–3 times per month’, ‘Once per week’, ‘2–4 times per week’, ‘5–6 times per week’, ‘Once per day’, ‘2–3 times per day’, and ‘4 or more times per day’. The nutritional profile of the recipes was calculated by applying i) yield factors to the raw, edible ingredients to account for weight changes during cooking; and ii) retention factors to account for changes in nutrient content33. The analysis was based on the Tunisian food composition table34, supplemented with data from the United States Department of Agriculture table35, additional laboratory analyses, and the Food Processor software (version 8.3) to estimate the energy and nutrient content (both macro- and micronutrients) of the foods. Vitamin D-rich food consumption was categorized into terciles and classified as low, medium, and high. The daily recommended intake of vitamin D for all age groups and both gender of the study population is 15 µg/d6. Anthropometric measurements Standing height was recorded using a wall-mounted stadiometer (Person-check®, Kirchner and Wilhelm, Germany) displaying a precision of ± 0.1 cm, and weight was measured using a calibrated scale (Detecto, Webb City, MO, USA) with a precision of ± 0.1 kg. Height-for-age and BMI (Body mass index)-for-age z-scores were calculated using the World Health Organization reference for school-age children. Stunting was defined as height-for-age < − 2 z-scores, thinness as BMI-for-age < − 2 z-scores, and overweight as BMI-for-age ≥ + 1 z-score36. Sociodemographic and economic assessment Data on the parents’ level of education and professional occupation were collected. A household economic level proxy was computed by multiple correspondence analyses from the matrix of binary variables pertaining to type of housing, house ownership, household equipment (e.g. air conditioning), sanitation and appliances. The first component score was used as an index of economic level. This index displaying a gradient of household ‘wealth’, were categorized in ‘low’, ‘medium’ and ‘high’ levels using tertile values as cut-off values37,38. Statistical analysis Continuous variables were expressed as mean or median with standard error of mean (SEM) and as percentages for categorical variables. Linear regression was used to compare interval response variables across categorical groups. A multinomial logistic regression was used to compute the crude and adjusted association of vitamin D deficiency and vitamin D insufficiency. A sensitivity analysis was conducted using a binary logistic regression. Because few subjects had > 125 nmol/L, we included them in ≥ 50 nmol/L response category (considered children with optimal status). Bivariate multinomial logistic regression models (only for the primary analysis) were used to compute crude relative prevalence ratio (RPR) for the associations between each nutritional, demographic, and socioeconomic variable, and vitamin D status. The full adjusted multinomial logistic regression model included two sets of independent variables. The first set included child-specific characteristics namely, age, gender (reference: male), residence (rural), region (South West), vitamin D-rich food consumption (high consumption), BMI z-score (normal [− 2 to + 1[), and school type (private). The second set included parent-specific characteristics, namely education level (secondary education or higher), occupational status (unemployed), and household economic status (high economic status). A linear regression model with full adjustment variables was used to assess the association between the covariates and daily intake of vitamin D. The svy function in STATA version 16.0 was used to account for the complex survey design. A type I error rate of 0.05 was used for all statistical tests. Results Sample characteristics A total of 2670 children were surveyed, which corresponds to a response rate of 100%. The average age of participants was 9.4 (0.03) years, ranging from 6.0 to 12.8 years (Table 1). Seven out of 10 mothers had an uncompleted secondary schooling, while the prevalence was lower among fathers, with a proportion of 55.7%. Most of the mothers were not working (74.0%), while 93.2% of fathers were active. Based on the BMI-for-age z-score, the prevalence of overweight (including obesity) and obesity were 23.7% and 10.1%, respectively. Only 5.7% of children were thin. The prevalence of stunting was 2.6%. No significant differences in nutritional status were observed between boys and girls (data not shown). Table 1Socioeconomic and demographic characteristics and place of residence of 6–12 years’ Tunisian school children in 2023 (n = 2670).n%1 (95% CI2)Age (years) < 10159059.2 (57.1–61.3) ≥ 10108040.8 (38.7–42.9)GenderBoys130748.9 (47.7–50.2)Girls136351.1 (49.8–52.4)Living areaUrban187570.2 (59.4–79.2)Rural79529.8 (20.8–40.6)RegionGreater Tunis51921.4 (20.1–22.9)North East36114.0 (13.2–14.9)North West2719.2 (7.7–11.1)Centre East63716.2 (15.0–17.5)Centre West41915.0 (13.5–16.7)South East31118.1 (15.3–21.3)South West1526.0 (4.9–7.2)Education of fatherSecondary or more schooling75728.3 (26.8–29.8)Uncompleted secondary schooling90168.5 (66.9–70.0)Nor formal schooling or Primary schooling10123.3 (2.6–4.0)Education of motherSecondary or more schooling88237.0 (34.9–39.0)Uncompleted secondary schooling80555.7 (53.5–57.8)Nor formal schooling or Primary schooling9837.4 (6.4–8.5)Occupation of the fatherMiddle/upper executive63624.0 (20.9–25.7)Worker/employee185869.2 (65.8–72.4)Not working1766.8 (5.4–8.5)Occupation of the motherMiddle/upper executive36013.7 (11.2–26.6)Worker/employee35712.7 (10.6–15.3)Not working195373.6 (69.9–77.0)1 Weighted proportions accounting for sampling design. 2 95% confidence interval taking into account sampling design. Socioeconomic and demographic characteristics and place of residence of 6–12 years’ Tunisian school children in 2023 (n = 2670). 1 Weighted proportions accounting for sampling design. 2 95% confidence interval taking into account sampling design. Vitamin D status National The mean national serum concentration of 25(OH)D was 53.9 (1.5) nmol/L (Fig. 1), corresponding to a deficiency and insufficiency prevalences of 15.8% and 28.8%, respectively. The 25(OH)D values ranged from 19.8 to 154.3 nmol/L. Only 0.5% of children had a 25(OH)D concentration higher than 125 nmol/L, most of them had normal body corpulence (92%) and were from both Centre East or South East regions (83%). 25(OH)D serum concentrations among Tunisian children aged 6–12 years. P values: design-based linear regression for 25(OH)D (as figured by vertical reference lines) × factor variable. The red line refers to the median 25(OH)D concentration among all children. By gender and age Mean 25(OH)D concentration was significantly higher among boys (59.5 (1.6) nmol/L) compared to girls (48.5 (1.4) nmol/L; p < 0.0001). The prevalence of vitamin D deficiency was approximately twice as high in female participants (20.3%) compared to male participants (11.0%; p < 0.0001). Mean 25(OH)D concentration decreased with age (55.1 (1.6) nmol/L in 6–9 years vs. 52.1 (1.5) nmol/L in 10–12 years; p = 0.001), while only insufficiency was high among the 10–12 years (p = 0.013) (Table 2, Fig. 1). Table 2Association between vitamin D status and covariates among SAC (n = 2658) in Tunisia.25(OH)D of < 30 nmol/L25(OH)D of ≥ 30 to < 50 nmol/L25(OH)D ≥ 50 nmol/L(base response category)n%1(95% CI2)Crude analysisRPR3(95% CI2)Adjusted analysisaRPR3(95% CI2)%1(95% CI2)Crude analysisRPR3(95% CI2)Adjusted analysisaRPR3(95% CI2)%1 (95% CI2)Age (years)P4 = 0.26P4 = 0.013P4 = 0.0096—9158017.5 (13.1,21.9)1126.5 (22.7,30.7)1158.6 (52.4,64.6)10—12107816.9 (12.9,21.5)1.3 (1.0,1.7)1.4 (1.0,1.8)32.1 (27.7,36.8)1.4 (1.1,1.7)1.4 (1.1,1.7)51.1 (44.8,57.4)GenderP4 < 0.0001P4 < 0.0001P4 < 0.0001 Boys 130111.0 (7.9,15.3)1124.0 (20.2,28.3)1165.0 (58.9,70.7) Girls 135720.3 (15.8,25.7)2.6 (2.0,3.3)2.1 (1.3–3.3)33.3 (29.2,37.7)1.9 (1.6,2.4)2.4 (1.8,3.3)46.5 (40.1,52.9)Living areap4 = 0.22P4 = 0.95P4 = 0.35 Urban 187017.6 (13.1,23.2)0.6 (0.2,1.4)0.8 (0.3,2.0)28.8 (25.1,33.0)0.9 (0.5,1.5)0.8 (0.5,1.5)53.7 (46.6,60.6) Rural 78811.5 (5.9,21.3)1128.6 (21.2,37.3)1159.9 (48.7,70.2)Regionp4 = 0.003P4 = 0.42P4 = 0.0004 Greater Tunis 51035.9 (25.3,48.5)43.6 (7.7,247.8)68.0 (11.9,387.5)36.1 (30.2,42.5)5.2 (2.1,12.5)6.6 (2.7,16.1)28.1 (17.9,41.3) North East 3606.4 (3.2,12.4)3.2 (0.6,18.6)4.1 (0.7,23.3)26.6 (18.7,36.3)1.6 (0.7,3.8)1.6 (0.7,3.9)67.1 (56.1,76.5) North West 27010.4 (5.2,19.9)5.9 (0.9,36.8)7.7 (1.3,47.5)29.3 (18.8,42.6)2.0 (0.7,5.4)2.1 (0.8,5.7)60.3 (42.1,76.1) Centre East 63611.0 (5.3,21.4)6.3 (1.0,37.7)8.4 (1.4,50.3)29.1 (22.0,37.3)2.0 (0.8,4.6)2.2 (0.9,5.1)60.0 (47.8,71.1) Centre West 41911.2 (3.4,31.7)6.1 (0.8,48.5)8.7 (1.2,62.8)26.7 (18.2,37.3)1.7 (0.7,4.2)1.9 (0.8,4.6)62.2 (46.8,75.5) South East 31114.5 (6.3,30.0)8.2 (1.2,56.0)9.5 (1.4,65.3)26.0 (16.1,39.2)1.7 (0.6,5.0)1.8 (0.6,5.0)59.7 (39.6,76.9) South West 1522.3 (0.6,9.2)1119.4 (10.9,32.2)1178.3 (62.3,88.7)Frequency of vitamin D-rich food consumptionp4 < 0.0001P4 = 0.027P4 < 0.0001Low88421.7 (16.8,27.6)3.0 (2.2,3.9)2.0 (1.1,3.0)29.9 (25.6,34.7)1.6 (1.3,2.1)0.8 (0.5,1.2)48.5 (41.3,55.7)Middle88416.4 (12.0,22.0)2.1 (1.5,2.8)1.6 (1.2,2.5)31.7 (26.8,37.0)1.6 (1.3,2.0)1.1 (0.8,1.4)51.9 (45.5,58.4)High8909.9 (7.0,13.9)1124.9 (20.7,29.5)1165.3 (58.9,71.0)Body corpulencep4 < 0.0001P4 = 0.15P4 < 0.0001Thin (Z-score < -2)15611.9 (7.3,18.8)0.8 (0.5,1.3)0.9 (0.5,1.5)24.8 (18.6,32.3)0.8 (0.6,1.2)0.8 (0.6,1.2)63.4 (54.9,71.1)Normal (Z-score -2 et < + 2)188313.6 (10.4,17.7)1128.0 (24.2,32.2)1158.5 (52.3,64.3)overweight (Z-score ≥ + 1)61923.1 (16.8,30.9)2.2 (1.6,3.0)2.0 (1.5,2.7)32.0 (26.9,37.4)1.5 (1.2,1.9)1.4 (1.1,1.8)44.9 (38.3,51.9)Type of schoolp4 = 0.63P4 = 0.76P4 = 0.76Private25415.4 (11.5,20.3)1128.9 (25.1,32.9)1155.8 (49.4,62.0)Public240419.9 (7.0,44.9)1.4 (0.3,5.5)0.5 (0.1,2.0)27.5 (20.3,36.1)1.0 (0.6 – 1.8)0.6 (0.4,1.0)52.7 (33.8,70.8)Education of fatherp4 = 0.82P4 = 0.94P4 = 0.74 Secondary or more schooling 100316.2 (12.5,20.8)0.9 (0.7,1.2)0.7 (0.5,0.9)29.1 (24.7,33.9)0.9 (0.8,1.1)0.8 (0.7,1.0)54.8 (48.4,60.9) Primary schooling 90015.2 (11.0,20.6)1.0 (0.7,1.3)0.7 (0.5,1.0)28.5 (24.8,32.5)1.0 (0.8,1.2)0.8 (0.6,1.1)56.4 (49.9,62.6) No formal schooling 75515.9 (11.5,21.5)1128.6 (24.2,33.5)1155.6 (49.1,61.9)Education of motherp4 = 0.065P4 = 0.52P4 = 0.37 Secondary or more schooling 87513.3 (10.0,17.5)1.4 (1.1,1.8)1.5 (1.1,2.1)29.4 (24.9,34.3)1.0 (0.8,1.2)1.0 (0.8,1.3)57.4 (50.9,64.0) Primary schooling 80517.7 (13.1,23.4)1,3 (0.9,1.8)1.4 (1.0,2.0)27.3 (23.5,31.6)1.1 (0.9,1.3)1.1 (0.8,1.4)55.0 (49.0,61.0) No formal schooling 97816.4 (12.0,22.0)1129.4 (25.4,33.8)1154.3 (47.6,60.9)Occupation of the fatherp4 = 0.43P4 = 0.49P4 = 0.16 Middle/upper executive 35817.6 (12.7,23.9)1130.7 (26.7,35.0)1151.8 (54.2,58.3) Worker/employee 35515.3 (11.4,20.3)0.8 (0.5,1.2)1.1 (0.8,1.4)28.5 (24.1,33.2)0.9 (0.6,1.1)0.9 (0.6,1.2)56.3 (49.6,62.7) Not working 194513.5 (8.5,30.7)0.6 (0.4,1.2)0.9 (0.5,1.7)25.1 (17.7,34.3)0.7 (0.4,1.1)0.7 (0.4,1.2)55.5 (49.7,61.3)Occupation of the motherp4 = 0.048P4 = 0.13P4 = 0.037 Middle/upper executive 35522.0 (15.3,30.5)0.5 (0.3,0.9)0.6 (0.3,0.9)30.6 (25.4,36.4)1.0 (0.6,1.5)1.0 (0.7,1.4)47.5 (38.2,56.9) Worker/employee 35713.1 (8.7,20.0)0.6 (0.4,0.9)0.8 (0.5,1.1)33.5 (27.0,40.7)0.7 (0.5,1.0)0.9 (0.7,1.2)53.3 (45.1,61.3) Not working 194615.1 (11.2,20.0)1127.6 (23.8,31.8)1157.4 (51.3,63.4)Household economic levelp4 = 0.0008P4 = 0.77P4 = 0.005 High 88810.6 (7.1,15.6)1.7 (1.2,2.5)1.6 (1.2,2.2)27.8 (23.0,33.2)1.2 (0.9,1.5)1.1 (0.9,1.5)61.6 (54.4,68.4) Middle 88516.4 (12.2,21.8)2.3 (1.4,3.8)2.1 (1.4,3.0)28.9 (24.6,33.6)1.3 (0.9,1.8)1.2 (0.8,1.7)54.8 (48.0,61.4) Low 88520.2 (15.1,26.5)1129.6 (25.5,34.0)1150.3 (43.6,57.0)1 Weighted percentage (accounting for sampling design, including unequal probabilities of selection). 2 confidence interval. 3 Crude or adjusted relative prevalence ratio. 4 Chi-square test for the response variable x and its cofactors. The response variable was categorized as: (< 30 nmol/L, ≥ 30 and < 50 nmol/L, or ≥ 50 nmol/L (base response category)). Association between vitamin D status and covariates among SAC (n = 2658) in Tunisia. %1 (95% CI2) Crude analysis RPR3 (95% CI2) Adjusted analysis aRPR3 (95% CI2) %1 (95% CI2) Crude analysis RPR3 (95% CI2) Adjusted analysis aRPR3 (95% CI2) 1 Weighted percentage (accounting for sampling design, including unequal probabilities of selection). 2 confidence interval. 3 Crude or adjusted relative prevalence ratio. 4 Chi-square test for the response variable x and its cofactors. The response variable was categorized as: (< 30 nmol/L, ≥ 30 and < 50 nmol/L, or ≥ 50 nmol/L (base response category)). Place of residence No significant differences in the 25(OH)D concentration distribution were observed across living areas (p = 0.22). However, significant regional variation was noted, with mean 25(OH)D concentrations ranging from 40.3 (3.0) nmol/L in Greater Tunis to 66.3 (4.3) nmol/L in the South West. The prevalence of vitamin D deficiency was markedly higher among children residing in Greater Tunis (35.9%), exhibiting a 2.5- to 5.6-fold increase compared to other regions. Conversely, only 2.3% of children in the South West presented vitamin D deficiency. Vitamin D insufficiency was comparable across regions (p = 0.43) (Table 2, Fig. 1). Vitamin D-rich food consumption Children with a high vitamin D-rich food consumption exhibited a significantly higher mean 25(OH)D concentration (24.1 (0.7) nmol/L) compared to those with low consumption (19.6 (0.6) nmol/L; p < 0.0001). This corresponded to a two-fold lower prevalence of vitamin D deficiency in the high consumption (9.9%) versus the low consumption group (21.7%). Vitamin D insufficiency prevailed among those within the medium tercile of intake (31.7%, p = 0.027) (Table 2, Fig. 1). Overweight/obesity Mean 25(OH)D concentration varied significantly according to body corpulence status (p < 0.0001), ranging from 22.7 (0.8) nmol/L among thin children to 19.3 (0.7) nmol/L among those with overweight/obese. The prevalence of vitamin D deficiency was approximately twice as high among overweight/obese participants (23.1%) compared to thin participants (11.9%; p < 0.0001). No association between body corpulence and vitamin D insufficiency (p = 0.15) (Table 2, Fig. 1). Parents’ characteristics No significant associations were observed between most parental characteristics and child 25(OH)D concentration, except for maternal occupation. Participants whose mothers were not employed or held positions as employees/workers demonstrated a lower prevalence of vitamin D deficiency (15.1% and 13.1%, respectively) compared to those whose mothers were middle/upper executives (22.0%; p = 0.048). No association was found between parent characteristics and vitamin D insufficiency (p values between 0.13 and 0.94) (Table 2, Fig. 1). Household economic level The distribution of 25(OH)D concentrations varied significantly across household economic levels (p = 0.0008). The prevalence of vitamin D deficiency was approximately twice as high among children from households with high economic levels (20.2%) compared to those from low household economic level (10.6%). No association with vitamin D insufficiency was found (p = 0.77) (Table 2, Fig. 1). Dietary intake of vitamin D-rich food The national mean vitamin D-rich food intake was 6.6 (0.02) µg/d, with a median intake of 7.0 (0.1) µg/d and an interquartile range of 5.4–7.4 µg/d (Supplementary Fig. 2). The percentage of vitamin D intake coverage was 43.7% (95% CI 43.3, 44.0) (min–max: 33.3%–83.7%). Vitamin D intake was significantly higher among boys (7.6 (0.02) µg/d) compared to girls (5.6 (0.02) µg/d; p < 0.0001) and among children aged 10–12 years (6.7 (0.04) µg/d) compared to those aged 6–9 years (6.5 (0.03) µg/d; p < 0.0001). Significant regional variations in vitamin D-rich food intake were observed (p = 0.039). Children living in the North East (coefficient = 0.29 µg/d; p = 0.009) and Centre East (coefficient = 0.23 µg/d; p = 0.026) regions had significantly higher vitamin D-rich food intakes than those in the South West region. Vitamin D-rich food intake was comparable in all other regions (p values between 0.13 and 0.56, detailed data no shown) compared to the South West region. Relationship between subject-level vitamin D status and living environment, socioeconomic, demographic factors, anthropometry and vitamin D-rich food consumption Table 2 presents the results of crude and adjusted multinomial logistic regression analyses for the association between vitamin D status (categorized into three levels) and covariates. The crude and adjusted models yielded consistent associations across all variables. In the adjusted analysis, children aged 10–12 years exhibited a 1.4-fold increased likelihood of vitamin D deficiency (aRPR = 1.4 (95% CI 1.0-1.8)). Similarly, girls were twice as likely to have vitamin D deficiency as boys (aRRR = 2.1 (95% CI 1.3-3.3)). Children residing in regions other than the North East demonstrated substantially higher odds of deficiency. The aRRR ranged from 7.7 (95% CI 1.3-47.5) for the North West region to 68.0 (95% CI 11.9-387.5) for Greater Tunis, compared to the South West region. Participants with low or moderate vitamin D-rich food consumption showed significantly elevated odds of vitamin D deficiency (aRRR = 2.0 (95% CI 1.1-3.0) and aRRR = 1.6 (95% CI 1.2-2.8), respectively, compared to those with high consumption. Overweight status was associated with a 2.0-fold (95% CI 1.5-2.7) increase in the odds of vitamin D deficiency compared to children with normal status. Parental education levels exhibited a complex association with vitamin D deficiency. Specifically, children with mothers holding a secondary or higher education level were more likely to have vitamin D deficiency (aRRR = 1.5 (95% CI 1.1-2.1)) vs. those without formal schooling), whereas children with fathers of similar educational attainment were less likely (aRRR = 0.7 (95% CI 0.5-0.9)). Children whose mothers held middle or upper executive professional positions demonstrated reduced odds of deficiency (aRRR = 0.6 (95% CI 0.3-0.9)) compared to those with mothers who did not work. Living in households with middle-to-high economic levels predicted vitamin D deficiency (aRRR ranging from 1.6 to 2.1, p values < 0.001). Finally, age, gender, and body corpulence status of SAC were significant predictors of insufficiency of vitamin D, with similar effect sizes observed for deficiency. In general, the sensitivity analysis showed relatively comparable results except few exceptions (Supplementary Tables 1 and 2). Discussion Interpretation of the key results This study presents the first nationally representative assessment of vitamin D status among Tunisian SAC (6–12 years), revealing a substantial burden vitamin D deficiency. The national mean serum 25(OH)D concentration was 53.9 nmol/L, and 15.8% of children were classified as deficient (< 30 nmol/L). This significant prevalence masks considerable sociodemographic, nutritional, and geographic disparities. Although previous studies in Tunisia provide useful contextual data, temporal comparisons of vitamin D status remain challenging due to differences in study populations, seasonality, and particularly the variability in 25(OH)D assay methods over time, which may limit the comparability of findings. Compared to other Mediterranean countries, the prevalence of deficiency among healthy Tunisian children was notably higher than reported in France (3% in 6–10 years)39, Italy (6% in 5–10 years)40, and Algeria (8.1% in 5–15 years)41, though it was comparable to estimates from Pakistan and Afghanistan10. A systematic review and meta-analysis42 reported that vitamin D deficiency (using comparable thresholds) affects approximately 18–34% of children in Africa, with substantial heterogeneity depending on region, lifestyle, and urbanization. The Tunisia prevalence is within the lower range reported across African settings, but remains noteworthy given Tunisia’s high solar irradiance, reinforcing the role of behavioral and environmental factors. A marked gender disparity was observed, with girls exhibiting significantly lower mean 25(OH)D concentrations and nearly double the prevalence of deficiency compared to boys. Adjusted logistic regression confirmed 1.4-fold higher odds of deficiency in girls. This gender disparity may reflect a combination of factors, including lower dietary intake of vitamin D-rich foods among girls, differences in sun exposure and outdoor activity, as well as potential biological differences in vitamin D metabolism43. Despite having higher intake coverage, older children were at greater risk of insufficiency or deficiency, possibly due to decreased sun exposure associated with sedentary behavior, increased indoor activities, and reduced outdoor time44,45. Physical activity may further contribute to vitamin D status through both indirect and direct mechanisms46. Children who engage in regular outdoor physical activity are more likely to be exposed to sunlight, thereby enhancing cutaneous vitamin D synthesis. In addition, emerging evidence suggests that skeletal muscle may serve as a functional reservoir for vitamin D metabolites, contributing to their conservation and regulation in the body. Therefore, lower levels of physical activity, often associated with sedentary behavior and overweight, may negatively impact vitamin D status independently of dietary intake47. The higher prevalence of deficiency observed among older and overweight children in our study may partly reflect these mechanisms. Notably, sun exposure remains a stronger predictor of vitamin D status than dietary intake alone48. Given Tunisia’s high solar potential, the persistence of vitamin D deficiency observed in this study highlights the importance of effective, rather than theoretical, sun exposure. Although UVB radiation is adequate for dermal synthesis during most of the year, actual exposure may be substantially reduced by modern lifestyle factors. Urbanization, increased time spent indoors, limited access to safe outdoor spaces, clothing habits, and atmospheric pollution can all significantly attenuate UVB penetration and skin exposure49. The markedly higher prevalence of deficiency in Greater Tunis compared to southern regions supports this hypothesis, suggesting that environmental and behavioral determinants of sun exposure play a central role. Significant regional disparities were also evident. Children in Greater Tunis had the highest prevalence of deficiency (35.9%), with adjusted odds more than 30 times higher than children in the South West region. As vitamin D intake was comparable across regions, these differences are likely due to lifestyle and environmental factors. Greater Tunis, as the most urbanized region, is characterized by indoor living, higher levels of air pollution, and limited access to outdoor spaces, all of which may reduce sun exposure, a key determinant of endogenous vitamin D synthesis50. A strong inverse association was also observed between overweight/obesity and vitamin D status. Overweight children had twice the risk of deficiency, a finding consistent with previous studies51,52. Several meta-analyses reported a 24–52% increased risk of vitamin D deficiency among overweight or obese individuals51,53,54. While the exact mechanisms remain unclear, hypotheses include the sequestration of vitamin D in adipose tissue, dilution effects due to larger body volume, and decreased bioavailability55,56. There may also be reciprocal causality, as the widespread expression of vitamin D receptors in adipose tissue could enhance vitamin D clearance in individuals with greater fat mass. In the context of the ongoing nutrition transition and rising obesity rates, it is plausible that diets high in ultra-processed or “junk” foods (e.g., sugar-sweetened beverages, fast foods) contribute to poorer vitamin D status57,58. However, bi-directional Mendelian randomization analyses suggest a causal link from obesity to vitamin D deficiency, rather than the reverse59. Vitamin D-rich food consumption emerged as a significant protective factor. Children in the lowest tercile of intake had twice the prevalence of deficiency compared to those in the highest. The national average daily intake was 7 µg/day, associated with a mean serum 25(OH)D level of 54 nmol/L. This yields a ratio of intake to 25(OH)D concentration of 0.13, which is lower than the 0.3 ratio proposed by the U.S. Institute of Medicine—assuming 15 µg/day achieves a serum level of 50 nmol/L6. Given that vitamin D is fat-soluble, serum 25(OH)D reflects cumulative intake over several months, and the cross-sectional design of this study limits precise estimation of the intake–status relationship. Interestingly, an inverse association was also observed between household economic level and vitamin D status, with higher deficiency rates among children from wealthier households. While higher socioeconomic status is typically associated with healthier lifestyle behaviors60,61, vitamin D intake did not vary significantly across economic strata. Thus, other behavioral factors -such as reduced sun exposure due to less outdoor time- may explain this finding. Parental characteristics and household economic level were found to be stronger predictors of vitamin D deficiency than of insufficiency, underscoring the multifactorial nature of vitamin D status. In contrast, a similar study in India using multinomial logistic regression did not distinguish between deficiency and insufficiency62, potentially due to different model specifications and confounding variables. Evidence supporting extra-skeletal benefits of higher 25(OH)D concentrations remain heterogeneous, particularly in pediatric populations, and causal relationships have not been firmly established. In addition, data linking vitamin D status to functional health outcomes in children from North Africa are scarce. Most available studies in Tunisia have focused on specific subgroups or clinical populations rather than nationally representative samples. This highlights the need for longitudinal and intervention studies to better define optimal vitamin D thresholds in this context63. Strengths and limitations This study has several notable strengths. It is the first to assess vitamin D status among Tunisian SAC using a nationally representative sample, ensuring generalizability of the findings. The high response rate and use of rigorous, standardized measurement protocols (including a validated immunoassay method (VIDAS ELFA) further strengthen the reliability of the data. The use of a 30 nmol/L threshold to define deficiency represents a key strength of this study, as it aligns with public health recommendations for preventing rickets and ensuring musculoskeletal health in children. Nonetheless, several limitations should be acknowledged. First, the cross-sectional design precludes causal inference. Second, the absence of data on sun exposure and physical activity limits the ability to fully explain variability in vitamin D status. Third, although internal quality control procedures were applied, the lack of participation in an external quality assessment scheme such as the Vitamin D External Quality Assessment Scheme and the absence of calibration against reference methods (e.g., LC–MS/MS) may affect comparability with other studies. Fourth, liquid chromatography-tandem mass spectrometry (LC–MS/MS) is considered the reference method for the quantification of serum 25-hydroxyvitamin D due to its high analytical specificity and its capacity to distinguish between vitamin D2 and D3 metabolites64,65. In comparison, the VIDAS ELFA assay is more commonly used in large-scale studies because of its operational simplicity and high throughput. However, this assay may be affected by inter-assay variability, cross-reactivity with other vitamin D metabolites, and calibration differences, which can lead to systematic overestimation or underestimation of 25(OH)D concentrations relative to LC–MS/MS66. Despite these limitations, the VIDAS ELFA assay remain widely accepted for population-based assessments when standardized protocols and quality controls are applied67. In addition, key determinants of vitamin D status such as individual sun exposure (duration, timing, and skin surface exposed) and physical activity levels were not assessed, limiting the ability to fully explain the observed variability despite favorable environmental UV conditions. Conclusion This first nationally representative study of children aged 6–12 years in Tunisia reveals a high prevalence of vitamin D deficiency, particularly in urbanized settings such as Greater Tunis. Key determinants include sex, age, overweight status, inadequate dietary intake, and socioeconomic conditions. In the context of an ongoing nutrition transition and a rising burden of childhood overweight and obesity, the prevalence of vitamin D deficiency is likely to increase further. These findings highlight a well-documented but increasingly widespread paradox: the persistence of vitamin D deficiency in sun-rich environments. They underscore the need for comprehensive and context-specific public health responses, including food fortification strategies, promotion of healthy dietary behaviors, and urban planning initiatives that encourage outdoor physical activity. Beyond Tunisia, these results have broader relevance. They may inform public health policies in countries with similar climatic conditions and in populations experiencing rapid urbanization, lifestyle changes, and shifts toward more sedentary behaviors, where the risk of micronutrient deficiencies coexists with emerging non communicable disease burdens. Supplementary Information Supplementary Information 1. Supplementary Information 2. Supplementary Information 3. Author contributions JEA performed the conceptualization and the methodology of the study. RD, MEAH, BM and JEA analyzed and interpreted the data. RD wrote the original draft preparation. All authors reviewed, edited and agreed to the published version of the manuscript. Funding This study was carried out with the financial support of the European Union as part of the ‘Support for Sustainable Development in the Agricultural and Artisanal Fishing Sectors in Tunisia (ADAPT)’ programme. Data availability The data that support the findings of this study are available from Tunisian Ministry of Health but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors (JEA, email: jalila.elati@rns.tn) upon reasonable request and with permission of Tunisian Ministry of Health. Declarations Competing interests The authors declare no competing interests. Ethical approval All applicable institutional and governmental regulations concerning the ethical use of human volunteers were respected during this study. The study was conducted in accordance with the Declaration of Helsinki and the survey protocol was approved by the Ethics Committee on Human Research of the National Institute of Nutrition and Food Technology (Approval N° 22/2022) and the National Council of Statistics (Approval N° 03/2023). The National Personal Data Protection Agency issued three favourable opinions regarding the transfer of data abroad (Approval N° 22/02–7064), health data (Approval N° 22/02-7063), and personal data (Approval N° 22/01-1539). Additionally, the Family Judge provided a favourable opinion for the study (Approval N° 15152, dated 14 December 2022). The child’s parent/s or authorised legal guardian provides informed consent for their child’s participation. All collected information was handled with the strictest confidentiality. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Information 1. Supplementary Information 2. Supplementary Information 3. Data Availability Statement The data that support the findings of this study are available from Tunisian Ministry of Health but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors (JEA, email: jalila.elati@rns.tn) upon reasonable request and with permission of Tunisian Ministry of Health.