The SENDO Project is a dynamic, prospective cohort of Spanish children that began recruiting participants in 2015 (https://www.proyectosendo.es/). Its primary aim is to assess the effect of lifestyle and diet on the health of children and adolescents. Participants are invited to join the cohort by their pediatrician at their health care center or by the SENDO team of researchers through schools. Inclusion criteria are being a child 4 or 5 years old and residing in Spain. The sole exclusion criterion is the impossibility to access a device connected to the internet to fulfill the online questionnaires. Self-reported questionnaires are completed online by the child’s parents and collected at baseline and updated every year. The baseline questionnaire collects information related to lifestyle, diet, medical history, anthropometric and sociodemographic variables.

This cross-sectional study is based on baseline information of participants recruited between January 2015 and June 2023. A total of 1208 participants were recruited. We excluded 141 participants who reported extreme energy intake (< P1 or > P99) and 118 with extreme micronutrient intakes (≥ + 3 or ≤ -3 standard deviations (SD) from the mean). Lastly, 85 participants were additionally excluded for presenting incomplete questionnaires at baseline. Thus, the final sample included 864 participants.

The SENDO project adheres to the rules of the Declaration of Helsinki on the ethical principles for medical research in human beings. This study was approved by the Ethics Committee for Clinical Research of Navarra (P. 2016/122). An informed consent was obtained during the recruitment from all participants’ parents.

The usual dietary intake information was collected through a validated semi-quantitative 147-item food-frequency questionnaire (FFQ) [20] completed by the parents.

We calculated adherence to a provegetarian FP based on three scores: an overall provegetarian FP score [12] and two additional ones, that distinguish between healthful and unhealthful plant-based foods (healthful provegetarian and unhealthful provegetarian FP scores) [13]. The scoring criteria for all three scores are shown in Table 1 and has been described in previous studies [4, 12, 13, 15].

For the analysis of the overall provegetarian FP, we included 7 different plant food groups (fruits, vegetables, potatoes, nuts, legumes, grains, and olive oil) and 5 animal food groups (dairy, eggs, meat, fish and seafood, and animal fat), as originally proposed [13]. Food consumption (g/d) was adjusted for total energy intake by using the residual method. The energy-adjusted estimates were ranked into quintiles (for plant-based food groups) and reverse quintiles (for animal food groups). For plant-based food groups, a value of 1 was assigned to the first quintile, 2 to the second quintile, and so on, until 5 was assigned to the fifth quintile. For animal products quintiles were reversed (assigning a value of 5 to the first quintile, 4 to the second quintile, and so on, until a value of 1 was assigned to the fifth quintile). Final score ranged from 12 (lowest adherence) to 60 (highest adherence).

The healthful and unhealthful scores were constructed according to the criteria exposed by Satija et al. [4], by dividing plant-food groups according to their impact on health. The analysis included 17 items for both scores. Unlike the original index designed for adults, the item 'tea/coffee’ was not included in the score since our study focused on children. Food consumption was also adjusted for total energy intake (residual method) [21]. For the healthful provegetarian FP, we assigned quintile values for healthful plant food consumption and reverse quintiles values for animal food consumption and unhealthful plant food consumption. For the unhealthful provegetarian FP, we assigned quintile values for unhealthful plant food consumption and reverse quintiles values for animal food consumption and healthful plant food consumption. Final scores for both indexes ranged from 17 (lowest adherence) to 85 (highest adherence).

We also analyzed dietary intake through 16 different food groups: vegetables, fruits, legumes, grains, nuts, olive oil, potatoes, fruit juices, dairy, eggs, fish and seafood, meat, animal fat, sugar-sweetened beverages, pastries and sweets, beverages, and miscellaneous. Lastly, we separately assessed adherence to the Mediterranean diet by using the KIDMED score, which has been described previously [22].

The participants’ sociodemographic information was reported by their parents. The questionnaire on physical activity collected information on 14 types of activities, with 10 different possible answers, ranging from never to ≥ 11 h per week. Participants indicated the average time dedicated to each activity in the previous year. Screen time was assessed by averaging the daily hours spent using screens (TV, computer, or video games). Time spent on weekdays and weekends was assessed separately.

Parental attitudes towards their child’s dietary habits were assessed with 8 yes/no questions, scored positively for healthy attitudes and given no point for unhealthy ones. Parental knowledge on dietary recommendations for children were evaluated with questions on the recommended intake frequency of 18 food groups. Both scores have been described in previous articles [23, 24]

We assessed the intake of 20 micronutrients, including vitamins A, C, D, E, B1, B2, B3, B6, B12, folic acid, Calcium (Ca), Iodine (I), Iron (Fe), Phosphorous (P), Magnesium (Mg), Selenium (Se), Zinc (Zn), Chromium (Cr), Potassium (K) and Sodium (Na). Participants were considered to have an inadequate micronutrient intake when their intake was inferior to the estimated average requirement (EAR), (or to the adequate intake (AI) if the EAR was not available), as established by the Institute of Medicine [25].

Participants were divided into tertiles according to their provegetarian FP scores, with the highest tertile representing higher adherence and the lowest tertile representing lower adherence. Participants’ sociodemographic characteristics were presented according to their adherence to the overall provegetarian FP and information on nutrient and food group intake was presented according to adherence to all three provegetarian FPs. Numbers and percentages were used for categorical variables and means and SD were used for continuous variables. Linear trend tests across tertiles of each score were calculated by assigning the median of each tertile and treating this variable as continuous in regression models. We also assessed the prevalence of inadequate intake of each micronutrient in each tertile of the three patterns.

Multivariate analyses were used to assess the relationship between each provegetarian FPs and the risk of having ≥ 3 inadequate intakes, by using the EAR cut-point method. We fitted generalized estimating equations to account for intra-cluster correlation between siblings.

Analyses were progressively adjusted, in different models, for the following potential confounding factors: 1) age (continuous), sex, and total energy intake (continuous); 2) breastfeeding duration (none, < 6 months, 6–12 months, and > 12 months), number of siblings (1, 2, 3–4, 5 or more), parental knowledge about nutritional recommendations for children (low, medium or high) and parental attitudes towards child’s dietary habits (unhealthy, average or healthy); 3) physical activity (continuous) and screen time (continuous). The first tertile was used as category of reference.

For our analysis, we calculated 1) the number of micronutrients with inadequate intake (and 95% Confidence Interval (CI)) by tertiles of provegetarian scores and 2) the Odds Ratio (OR) and 95% CI for the inadequate intake of ≥ 3 micronutrients across tertiles of the three provegetarian FPs. We also estimated the adjusted proportions of children with inadequate intake of ≥ 3 micronutrients in each tertile of the three patterns.

Lastly, different sensitivity analysis were performed to assess the robustness of our findings: 1) we included the intake of supplements to the total intake of micronutrients calculated for each participant; and 2) we changed the outcome to an inadequate intake of ≥ 4 micronutrients.

Statistical analyses were carried out using Stata version 15.0 (Stata Corp., College Station, TX, USA). All p values are two-tailed and statistical significance was settled at the conventional cut-off point of p < 0.05.