Next Article in Journal
Nutritional Adequacy and Latent Tuberculosis Infection in End-Stage Renal Disease Patients
Next Article in Special Issue
Relation of Fruits and Vegetables with Major Cardiometabolic Risk Factors, Markers of Oxidation, and Inflammation
Previous Article in Journal
The Role of Early Pregnancy Maternal Selenium Levels on the Risk for Small-for-Gestational Age Newborns
Previous Article in Special Issue
Bioactive Compounds of Cooked Tomato Sauce Modulate Oxidative Stress and Arachidonic Acid Cascade Induced by Oxidized LDL in Macrophage Cultures
Erratum published on 10 December 2019, see Nutrients 2019, 11(12), 3021.
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:

Dietary Patterns and Cardiovascular Risk Factors in Spanish Adolescents: A Cross-Sectional Analysis of the SI! Program for Health Promotion in Secondary Schools

Patricia Bodega
Juan Miguel Fernández-Alvira
Gloria Santos-Beneit
Amaya de Cos-Gandoy
Rodrigo Fernández-Jiménez
Luis Alberto Moreno
Mercedes de Miguel
Vanesa Carral
Xavier Orrit
Isabel Carvajal
Carolina E. Storniolo
Anna Tresserra-Rimbau
Mónica Doménech
Ramón Estruch
Rosa María Lamuela-Raventós
6,7 and
Valentín Fuster
Foundation for Science, Health and Education (SHE), 08008 Barcelona, Spain
Centro Nacional de Investigaciones Cardiovasculares (CNIC), 28029 Madrid, Spain
The Zena and Michael A. Wiener Cardiovascular Institute, Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA
Centro de Investigación Biomédica En Red en enfermedades CardioVasculares (CIBERCV), 28029 Madrid, Spain
GENUD (Growth, Exercise, NUtrition and Development) Research Group, Faculty of Health Science, University of Zaragoza, 50009 Zaragoza, Spain
Consorcio CIBER, M.P. Fisiopatología de la Obesidad y Nutrición (CIBERObn), Instituto de Salud Carlos III (ISCIII), 28029 Madrid, Spain
Department of Nutrition, Food Science and Gastronomy, School of Pharmacy and Food Sciences, XaRTA, INSA, University of Barcelona, 08028 Barcelona, Spain
Universitat Rovira i Virgili, Departament de Bioquímica i Biotecnologia, Unitat de Nutrició Humana. Hospital Universitari Sant Joan de Reus, Institut d’Investigació Pere Virgili (IISPV), 43201 Reus, Spain
Department of Internal Medicine, Hospital Clínic, Institut d’Investigacions Biomèdiques August Pi I Sunyer (IDIBAPS), University of Barcelona, 08036 Barcelona, Spain
Authors to whom correspondence should be addressed.
Nutrients 2019, 11(10), 2297;
Submission received: 30 August 2019 / Revised: 23 September 2019 / Accepted: 24 September 2019 / Published: 26 September 2019
(This article belongs to the Special Issue Health Benefits of Vegetables and Fruits)


Previous studies on the association between dietary habits and cardiovascular risk factors (CVRF) in adolescents have generated conflicting results. The aim of this study was to describe dietary patterns (DP) in a large sample of Spanish adolescents and to assess their cross-sectional relationship with CVRF. In total, 1324 adolescents aged 12.5 ± 0.4 years (51.6% boys) from 24 secondary schools completed a self-reported food frequency questionnaire. DPs were derived by cluster analysis and principal component analysis (PCA). Anthropometric measurements, blood pressure, lipid profile, and glucose levels were assessed. Linear mixed models were applied to estimate the association between DPs and CVRF. Three DP-related clusters were obtained: Processed (29.2%); Traditional (39.1%); and Healthy (31.7%). Analogous patterns were obtained in the PCA. No overall differences in CVRF were observed between clusters except for z-BMI and and z-FMI values, total cholesterol, and non-HDL cholesterol, with the Processed cluster showing the lowest mean values. However, differences were small. In conclusion, the overall association between DPs, as assessed by two different methods, and most analyzed CVRF was weak and not clinically relevant in a large sample of adolescents. Prospective analysis may help to disentangle the direction of these associations.

1. Introduction

Cardiovascular (CV) diseases have a multifactorial origin, and modifiable behavioral risk factors such as insufficient physical activity, unhealthy diet, or smoking play an essential role in disease development and progression. Suboptimal diet is one of the leading risk factors for death and disability in the world [1,2].
Adult behavior has its roots in childhood [3]. Adolescence is a critical period of personal development when individuals acquire greater autonomy [4]. This period is characterized by biological, psychological, and social changes that expose the adolescent to unhealthy behaviors [5], and consequently the prevalence of unhealthy behaviors and risk factors in adolescents is alarmingly high [6]. The consumption of low nutritional quality diets is especially prevalent in groups with lower socioeconomic status [7,8]. Nevertheless, according to the results of the HELENA study, the proportion of European adolescents achieving the American Heart Association (AHA) ideal diet score is as low as 2% [9].
Previous reports have described the main dietary patterns (DP) in adolescent populations and their associations with several health outcomes [10,11,12]. However, results are inconsistent, with some studies showing positive associations between healthier DPs and better CV health profiles, while others have shown the opposite pattern. Information provided by these studies is also useful for improving current dietary guidelines [4,13]. The methods most frequently used to identify DPs are cluster analysis and principal component analysis (PCA). These methods differ in important respects: While cluster analysis classifies individuals into mutually exclusive groups or clusters with similar food intake [4], PCA uses the correlation between different food groups, identifying linear combinations of foods that are frequently consumed together [14]. These differences might limit comparability of results across studies using cluster analysis vs. those using PCA. To date, very few studies have analyzed DPs using both methods. Testing the comparability of cluster analysis and PCA in adolescent populations will thus help to clarify whether studies using these approaches are analogous.
The aim of this study was to define DPs using both cluster analysis and PCA and to assess their associations with CV risk factors (CVRF) in a large sample of Spanish adolescents enrolled in the SI! Program for Secondary School [15].

2. Materials and Methods

The SI! Program (Salud Integral—Comprehensive Health) ( is a school-based multicomponent educational intervention aimed at promoting CV health in childhood and adolescence [15,16,17,18,19]. The SI! Program for Secondary school enrolled 1326 adolescents and their families in 2017 from 24 public high schools in Madrid and Barcelona, Spain. The overall rationale and study design, as well as the protocol for data collection and measurements, have been previously described [15], and is registered at, number NCT03504059. Collected data at study baseline were used for the analyses. The study was approved by the ethical committees of the institutions involved in the project (University of Barcelona, Centro Nacional de Investigaciones Cardiovasculares (CNIC), and the SHE Foundation). Participants provided informed consent for all examinations.

2.1. Dietary Assessment

Dietary data were obtained using the Children’s Eating Habits Questionnaire (CEHQ-FFQ), which is a validated food frequency questionnaire (FFQ) [20,21]. The CEHQ-FFQ assesses the consumption of 43 food items during the preceding four weeks. Response options displayed from left to right were as follows: ‘Never/less than once a week’, ‘1–3 times a week’, ‘4–6 times a week’, ‘1 time per day’, ‘2 times per day’, ‘3 times per day’, ‘4 or more times per day’, and ‘I have no idea’ [21]. The CEHQ-FFQ does not provide information about total energy or food intake; instead, it focuses on the main food groups that have been shown to be related to childhood overweight and overall health [20,21]. Adolescents self-reported their frequencies of consumption using an online application under the guidance of trained nutritionists. For the analysis, a conversion factor ranging from 0 to 30 was used to transform questionnaire answers into weekly consumption frequencies [22].

2.2. Anthropometric and Clinical Measures

Body weight was measured with an OMRON BF511 electronic scale and height with a Seca 213 portable stadiometer, with adolescents wearing light clothes and no shoes. Body mass index (BMI) was calculated as body weight divided by height squared (kg/m2). Waist circumference (WC) was measured in triplicate with a Holtain tape to the nearest 0.1 cm. Total fat content was estimated by bioelectrical impedance analysis with a tetra polar OMRON BF511 device. Fat mass index (FMI) was calculated by dividing body fat mass (kg) by height squared (m2). Blood pressure (BP) was measured with an OMRON M6 monitor three times at 2–3 min intervals according to a standardized protocol [23], and the minimum reading was used for analysis. Age- and sex-adjusted BMI, BP, and WC percentiles and z-scores were calculated according to Center for Disease Control standards [24], the American Academy of Pediatrics [25], and NHANES III [26], respectively. Sex-specific FMI z-scores were calculated based on our sample. Maturation stages according to Tanner [27] were self-reported by participants with the help of pictograms. Blood glucose levels and lipid profiles (total cholesterol (TC), HDL-cholesterol, non-HDL-cholesterol, LDL-cholesterol, triglycerides (TG), and TC:HDL and LDL:HDL ratios) were analyzed using the CardioCheck Plus device in capillary blood sampled with a lancet [28]. All measurements were carried out by trained nutritionists and nurses.

2.3. Physical Activity

Physical activity levels were estimated by the triaxial Actigraph wGT3X-BT accelerometer, worn on the non-dominant wrist for 7 consecutive days. Records were considered valid if they provided data from a minimum of 4 days with at least 600 minutes of wear time. Chandler (2016) [29] cut-off points were applied for the calculation of time spent in different physical activity intensities. Time spent in moderate-to-vigorous physical activity was used as a covariate in the main models.

2.4. Family Information

Parents filled in a self-administered questionnaire on family socioeconomic status. Parental educational level was categorized according to the International Standard Classification of Education (ISCED) [30]: Low (families without studies or primary studies, 0 to 1 ISCED level), medium (secondary studies and professional training, 2 to 4 ISCED levels), and high (university studies, 5 to 6 ISCED level). For the analyses, the highest education level of either parent was taken into account. A migrant background was assumed if at least one of the parents was born outside Spain.

2.5. Statistical Methods

Continuous variables were expressed as mean and standard deviation or 95% confidence intervals, whereas categorical variables were expressed as frequency distributions and percentages.
DP clusters were identified by a combined methodology including hierarchical (Ward’s method) and non-hierarchical (k-means) approaches, as previously described [31]. Relative consumption frequency was calculated for each food item by dividing the consumption frequency of the specific food item by the sum of the consumption frequencies of all food items reported by each individual participant. Food items were standardized before clustering in order to avoid differences in variances that might affect the clustering results. Ward’s method based on squared Euclidean distances was applied, generating and comparing several cluster solutions from two to five clusters. With the resulting centroids, a K-means cluster analysis algorithm was applied with a predefined maximum 100 iterations to generate a fine-tuned clustering solution. The stability of the final solution was examined by randomly splitting the database into halves and repeating the same clustering procedure and comparing both solutions, until satisfactory results were observed (kappa degree of concordance = 0.91 and 0.88 for the first and second halves, respectively).
DPs were also derived by PCA using the same standardized relative consumption frequencies. The number of factors to be retained was mainly determined by the scree plot and the interpretability of the DPs [32]. Interpretability was improved by varimax (orthogonal) rotation. Food items with absolute factor loadings greater than 0.30 were considered important contributors to a DP. Labels were applied to clusters based on foods with higher consumption frequencies and to principal components based on foods with higher loadings within components. To compare the results of the two methods, mean principal component scores were calculated for each cluster.
Multilevel linear mixed-effect models were used to assess the association between DPs and CVRF (anthropometric measurements, BP, blood glucose, and lipid profile). Fixed effects were age (continuous variable), gender and fasting status (binary variables), z-BMI (continuous variable), moderate-to-vigorous physical activity (minutes, continuous variable), and migrant background and parental educational level (categorical variables). Schools and municipality (Madrid or Barcelona) were handled as random effects. In the models evaluating z-BMI, FMI, and WC as dependent variables, the Tanner stage (categorical variable, from I to V) was also included as a fixed effect. Statistical significance was set at a threshold of 0.05. Bonferroni correction was used to adjust for multiple pairwise comparisons. All analyses were performed using SPSS Statistics for Windows, version 19.0.

3. Results

A total of 1324 adolescents (of 1326 enrolled) completed the FFQs at baseline. Mean age of adolescents was 12.5 ± 0.4 years (51.6% boys), and participants were enrolled at 24 Secondary Schools in Barcelona and Madrid (Spain). Based on 43 food items and their relative consumption frequency, categorization into three clusters were considered the most interpretable and stable solution. The three identified clusters were labelled as follows: Processed (n = 386, 29.2%), Traditional (n = 518, 39.1%), and Healthy (n = 420, 31.7%). The mean z-score and standard deviation for all food items in the three clusters are presented in Table 1.
Adolescents in the Processed cluster more frequently consumed crisps, popcorn, sweetened drinks, hamburgers, hot dogs, ice cream, and sauces like ketchup and mayonnaise, whereas their consumption of unsweetened cereals, meat and non-fried fish, vegetables, fruit with no added sugar, and white bread was less frequent than in the other clusters. Compared with the other clusters, participants in the Traditional cluster had higher relative consumption frequencies of water, white bread, pasta and rice, sweetened milk and yogurt, cold cuts, and ready-to-cook meat products and lower scores for unsweetened cereals, boiled eggs, and vegetables. The highest scoring food items in the Healthy cluster were fruit with no added sugar, vegetables, unsweetened cereals, and unsweetened milk and yogurt, while this cluster had relatively low scores for fried potatoes, sauces, sweetened drinks and milk, and fast foods such as pizza and hamburgers.
PCA of the same variables also yielded three DPs, explaining 30.4% of the variation in the sample (Table 2). The Processed component was characterized by high loadings on high-sugar and high-fat foods such as chocolate and candy bars, pizza, crisps, popcorn, candies, and ice cream. The Traditional component featured high-factor loadings on fresh and fried fish and meat, cold cuts, pasta, rice, and sweetened milk. The Healthy component was characterized by high loadings on vegetables, fruit with no added sugar, and unsweetened cereals.
Comparison of the mean principal component values across the derived DP clusters is shown in Figure 1. The Processed component score was significantly higher in the Processed cluster than in the Healthy and Traditional clusters (p < 0.001), whereas the Traditional component score was significantly higher in the Traditional cluster than in the Healthy cluster (p < 0.001) but not the Processed cluster (p = 0.089). The Healthy component score was significantly higher in the Healthy cluster than in the Traditional and Processed clusters (p < 0.001).
The general characteristics of adolescents according to DP clustering are presented in Table 3. The proportion of girls in the Healthy cluster (55.7%) was significantly higher than in the other clusters. The Healthy cluster was also the most prevalent among participants whose parents had a high educational level; in contrast, the Processed cluster contained higher proportions of participants from families with a low or medium parental educational level (26.2% and 43.8%, respectively). The Processed cluster also contained the highest proportion of participants from a migrant background (37.8%).
The associations between cluster-analysis–derived DPs and CVRF are shown in Table 4. Mean z-BMI was significantly higher in the Traditional (mean 0.46, 95% CI: 0.35, 0.57) and Healthy (mean 0.44, 95% CI: 0.32, 0.57) clusters than in the Processed cluster (mean 0.11, 95% CI: −0.02, 0.24). Mean z-FMI was significantly higher in the Traditional cluster (mean 0.03, 95% CI: −0.03, 0.08) than in the Processed cluster (mean −0.06, 95% CI: −0.12, 0.00). TC was significantly higher in the Healthy cluster (mean 157.6 mg/dL, 95% CI: 150.7, 164.4) than in the Processed cluster (mean 148.6 mg/dL, 95% CI: 141.6, 155.5). This variation was reflected in the non-HDL cholesterol differences between the Healthy cluster (mean 95.1 mg/dL, 95% CI: 90.8, 99.5) and the Processed cluster (mean 89.0 mg/dL, 95% CI: 84.4, 93.6). No statistically significant differences were found between clusters for the remaining CVRF analyzed. Analysis of the association of PCA-derived DPs with CVRF showed similar results to those found by cluster analysis (Appendix A Table A1).

4. Discussion

Three DPs (Processed, Traditional, and Healthy) were derived by cluster analysis and PCA of the food consumption frequencies in a large sample of adolescents participating in the SI! Program for Secondary School in Spain. The similarity of the DPs derived by the two methods emphasizes the robustness of the results obtained. Because DPs are sensitive to the specific profile of the population studied, it is difficult to compare results between studies; however, similar DPs have been reported in other populations [14,31,33,34,35]. In seven-year-old children, Northstone et al. described a Processed DP characterized by the consumption of processed meat, coated poultry and fish products, chips, pizza, and sandwiches [33]. Likewise, Aranceta et al. defined a Snacky pattern (in a population aged from 2 to 24 years) characterized by high consumption of bakery products, sweets, salted snacks, and soft drinks [34]. Healthy patterns have been also frequently described. Pérez-Rodrigo et al. described a Mediterranean DP characterized by high consumption of vegetables, olive oil, fish, fruit, yogurt, and water in Spanish population aged 2 to 24 years [31]. The traditional patterns described in the literature reflect country-specific dietary habits that are especially difficult to compare between studies. For example, Smith et al. characterized a Traditional British cluster rich in full-fat milk, meat, and potatoes [14], while Lee et al. described a Korean Traditional pattern with high consumptions of grains, kimchi, fish and shellfish, beef, vegetables, seaweeds, oils, and oriental sauces [35].
Our analysis of associations between DPs and CVRF only found significant associations for z-BMI, z-FMI, TC, and non-HDL cholesterol, with z-BMI in particular being higher in the Traditional and Healthy clusters than in the Processed cluster. The literature on the association between DPs and CV risk and health factors shows conflicting results. Some studies have shown an association between high healthy DP scores and favorable biomarkers, a better CV profile, and a lower risk of being overweight [10,11,31,36,37]. However, other studies have shown opposite results [12,35,36,38,39]. A cross-sectional study in 12–16-year-old American adolescents found inverse associations between the prevalence of excess central body fat and the intakes of dairy products, whole grains, total fruit, and vegetables [38]. Truthmann et al. found that adolescents with a good quality diet (assessed by a dietary index) had a higher rate of obesity and higher BP than those with a poorer diet [36]. Likewise, data from the 2001 South Korea National Health and Nutrition Survey [35] showed that children and adolescents allocated to a Traditional diet group had higher rates of obesity than participants in the Westernized fast-food and Mixed-diet clusters. The same study found the highest levels of TG and the lowest levels of TC and HDL-cholesterol in participants allocated to the Traditional cluster, although these differences were not statistically significant. Surprisingly, the highest levels of HDL-cholesterol were found in participants belonging to the Westernized fast-food cluster [35]. Another study showed a similar pattern, with TC higher among participants with the Healthy DP than among those in the Unhealthy DP [12]. In another study, Golley et al. found no associations between dietary quality and serum lipid concentration [40].
The mechanisms underlying these associations are poorly understood. One major concern is participant under-reporting or over-reporting of intake, especially in relation to high-fat foods [39,41]. Other factors that might affect long-term body composition are the method of preparation (e.g., deep frying vs. stir frying) and serving style (e.g., with sauce or side dishes) [38,42,43]. Our study did not quantify total energy intake, and the results are based on frequency of food intake. Similar consumption frequencies thus might mask differences in portion size, preparation, and serving style that result in differing total calorie content. This might be especially relevant in the context of socioeconomic differences. In our study, adolescents from families reporting a relatively lower socioeconomic status were more frequently allocated to the Processed pattern but still could be consuming less total energy than required due to food insecurity, and thus confounding the associations with clinical markers. The participants in our study are also facing the onset of puberty, which may influence dietary habits, food intake reporting, and biomarker values. Adolescents are known to sometimes reduce their food intake or misreport consumption because of weight concerns [36]. Another possibility is that the lower BMI, FMI and TC in the Processed pattern might be a product of reverse causation, with participants with a high BMI, FMI or a worse cholesterol profile altering their food intake toward a healthier profile. Participants with a higher BMI/FMI might also be more susceptible to social desirability bias. Even if they reflect a genuine association, differences in body composition and biomarkers according to DPs are unlikely to be clinically relevant because all mean values remained within the normal range and differences were small. Future prospective analyses will provide insight into the long-term trends in body composition and CV health associated with DPs, clarifying the true direction of the associations.
Although the benefits of a healthy diet are widely known, fruit and vegetable intake is influenced by taste preferences, repeated exposure, social experience, and availability [44]. The value of derived DPs is that this approach considers diet as a whole rather than a collection of nutrients or components [14]. Cluster analysis classifies individuals into non-overlapping groups or clusters based on similarities between diets, while PCA exploits the correlation between different food groups by identifying linear combinations of foods that are frequently consumed together. Both approaches describe diet as a whole, but from slightly different perspectives, despite the aim of both methods being to explain maximum variation in food intake [4,14]. It is common practice to use just one of these approaches; however, we decided to use both methods to assess the consistency of the associations between the derived DPs and CVRF. The food items forming each DP were mostly the same, independently of the method used. The best agreement was found for the Healthy patterns, in which the 10 items showing the highest frequencies of consumption were the same in both approaches. Conversely, the lowest agreement was found for the Traditional patterns; they tended to correlate with the Processed patterns and showed weaker associations with clinical markers, making their interpretation more difficult.
A similar agreement between cluster and PCA analyses was found in two previous studies. In seven-year-old children from the ALSPAC cohort, Smith et al. derived three similar DPs by cluster analysis (Processed, Plant-based, and Traditional British) and PCA (Junk, Traditional, and Health-conscious) [14]. In another study on Irish adolescents aged 13–17 years from the Irish National Teens Food Survey, the DPs derived by cluster analysis were Healthy, Unhealthy, Rice/Pasta dishes, Sandwich, and Breakfast cereal and Main meal-type foods, whereas the PCA-derived DPs were Healthy foods, Traditional foods, Sandwich foods, and Unhealthy foods. Although the two methods identified different numbers of DPs, the authors concluded that the results were still comparable [45].
This study has several strengths. The SI! Program for Secondary School is a large multi-center study in which all nurses and nutritionists received prior training according to strictly standardized procedures for data collection and guiding participants in completing the questionnaire. Another strength is the use and comparison of two different methods to derive DPs, thus providing more robust evidence and consistency in the results found. Moreover, the study focuses on pre-adolescents, who receive less attention in lifestyle studies than younger children or older adolescents. Finally, follow-up of these adolescents will provide the opportunity to track changes in DPs and to assess their association with changes in CV health over time. Some study limitations should be acknowledged. The cross-sectional nature of the analyses allows the establishment of associations but not causation. Participation in the study was voluntary, and we therefore cannot exclude selection bias due to adolescents with worse CV health or unhealthier lifestyles being less willing to participate. Another limitation is the self-reported nature of the CEHQ-FFQ, which introduces the possibility of social desirability bias or difficulties in the accurate estimation of intake frequency for certain food groups.

5. Conclusions

Cluster analysis and PCA identified three consistent DPs (Processed, Traditional, and Healthy) in a large sample of Spanish adolescents. Overall, no significant associations were found between DPs and CVRF, with the exception of z-BMI, z-FMI, TC, and non-HDL cholesterol. Given the limitations of the cross-sectional design (including the potential for reverse causation in the association between DPs and CVRF), disentangling the direction of the associations between DPs and lipid profile and body composition in adolescents will require future prospective analyses.

Author Contributions

V.F. conceived the overall study. J.M.F.-A., G.S.-B., R.F.-J., M.d.M., C.E.S., A.T.-R., M.D., R.E., and R.M.L.-R. coordinated data collection and measurements. I.C., V.C., and X.O. coordinated the implementation of the educational program. P.B., J.M.F.-A., and A.d.C.-G. conducted statistical analyses. R.F.-J. and L.A.M. provided scientific support over the course of this work. P.B. drafted the first version of the manuscript along with J.M.F.-A. and G.S.-B. All authors revised the manuscript critically for intellectual content and approved the published version.


This research was funded by the SHE Foundation, the “la Caixa” Foundation (LCF/PR/CE16//10700001 and LCF/PR/MS19/12220001), and the Fundació la Marató de TV3 (369/C/2016). R.F-J. received funding from the European Union Horizon 2020 Research and Innovation Programme under Marie Skłodowska-Curie grant No 707642. Support was also provided by the Ministerio de Ciencia, Innovación y Universidades (AGL2016–75329-R) and Generalitat de Catalunya. The CNIC is supported by the Instituto de Salud Carlos III (ISCIII), the Ministerio de Ciencia, Innovación y Universidades (MCNU) and the Pro CNIC Foundation, and is a Severo Ochoa Center of Excellence (SEV-2015-0505). A.T.-R. is supported by the Ministerio de Ciencia, Innovación y Universidades through a Juan de la Cierva-Formación postdoctoral fellowship (FJCI-2016-28694).


The authors thank the SHE Foundation (intellectual owner of the SI! Program), especially Carla Rodríguez and Domènec Haro, and all the staff who helped with data collection. We also especially thank all participating adolescents and their families, teachers, and schools. Simon Bartlett (CNIC) provided English editing.

Conflicts of Interest

The authors declare no conflict of interest.

Appendix A

Table A1. Cardiovascular risk factors of participants according to tertiles of principal component analysis (PCA).
Table A1. Cardiovascular risk factors of participants according to tertiles of principal component analysis (PCA).
T1T2T3p For TrendT1T2T3p For TrendT1T2T3p For Trend
Mean *
(95% CI)
Mean *
(95% CI)
Mean *
Mean *
(95% CI)
Mean *
(95% CI)
Mean *
(95% CI)
Mean *
(95% CI)
Mean *
(95% CI)
Mean *
(95% CI)
z-BMI0.50 a
(0.38, 0.62)
0.41 b
(0.30, 0.53)
0.16 a,b
(0.04, 0.27)
(0.26, 0.49)
(0.26, 0.49)
(0.20, 0.43)
(0.20, 0.43)
(0.27, 0.51)
(0.25, 0.48)
(0.20, 0.38)
(0.17, 0.34)
(0.17, 0.35)
(0.18, 0.35)
(0.17, 0.34)
(0.19, 0.37)
(0.21, 0.38)
(0.14, 0.32)
(0.19, 0.37)
z-Systolic BP−0.16
(−0.33, 0.02)
(−0.26, 0.09)
(−0.26, 0.08)
(−0.28, 0.07)
(−0.26, 0.08)
(−0.3, 0.04)
(−0.31, 0.04)
(−0.26, 0.08)
(−0.27, 0.07)
z-Diastolic BP−0.04
(−0.21, 0.13)
(−0.23, 0.11)
(−0.23, 0.11)
(−0.20, 0.14)
(−0.26, 0.08)
(−0.22, 0.12)
(−0.22, 0.12)
(−0.24, 0.11)
(−0.22, 0.12)
(−0.04, 0.07)
(−0.04, 0.07)
(−0.09, 0.02)
(−0.07, 0.04)
(−0.05, 0.06)
(−0.06, 0.05)
(−0.03, 0.08)
(−0.08, 0.03)
(−0.07, 0.04)
Glucose (mg/dL)102.9
(100.2, 105.7)
(100.0, 105.5)
(101.4, 106.9)
(101.0, 106.5)
(99.9, 105.4)
(100.8, 106.2)
(99.5, 105.0)
(100.0, 105.6)
(102.1, 107.6)
Total cholesterol (mg/dL)155.3
(148.6, 162.1)
(1476.0, 160.3)
(144.0., 157.4)
(146.6, 160.0)
(146.8, 160.2)
(146.2, 159.5)
(146.8, 160.2)
(145.8, 159.2)
(146.8, 160.2)
HDL-cholesterol (mg/dL)62.7
(60.1, 65.3)
(59.9, 65.1)
(59.8, 65.1)
(59.7, 64.9)
(60.0, 65.2)
(60.1, 65.3)
(59.9, 65.2)
(59.5, 64.7)
(60.3, 65.6)
LDL-cholesterol (mg/dL)80.0
(75.6, 84.4)
(72.8, 81.5)
(71.5, 80.4)
(72.7, 81.6)
(73.6, 82.3)
(73.6, 82.3)
(74.0, 82.8)
(73.2, 82.0)
(72.7, 81.5)
Non-HDL cholesterol (mg/dL)93.7
(89.4, 98.0)
(88.0, 96.5)
(86.0, 94.7)
(88.2, 96.8)
(87.9, 96.5)
(87.4, 95.9)
(88.2, 96.9)
(87.6, 96.2)
(87.6, 96.1)
Triglycerides (mg/dL)76.8
(72.4, 81.1)
(76.4, 85.0)
(74.5, 83.1)
(75.1, 83.7)
(75.0, 83.5)
(73.5, 81.9)
(74.6, 83.2)
(75.0, 83.5)
(73.8, 82.3)
TC:HDL ratio2.55
(2.47, 2.63)
(2.44, 2.60)
(2.46, 2.63)
(2.50, 2.66)
(2.46, 2.63)
(2.42, 2.58)
(2.46, 2.63)
(2.47, 2.63)
(2.45, 2.61)
LDL:HDL ratio1.34
(1.27, 1.42)
(1.23, 1.38)
(1.25, 1.40)
(1.28, 1.43)
(1.25, 1.40)
(1.22, 1.36)
(1.25, 1.40)
(1.26, 1.41)
(1.24, 1.38)
BMI: body mass index (kg/m2); WC: waist circumference; BP: blood pressure; FMI: fat mass index; TC: total cholesterol. * Estimated marginal means from multilevel linear mixed-effects models. Variables in the models were age, gender, fasting status, parental education level, migrant background, z-BMI, tanner stage and moderate and vigorous physical activity and included a random school and a random municipality effect to account for the clustered study design. a Significant differences between lowest and highest tertile after Bonferroni correction; b Significant differences between medium and highest tertile after Bonferroni correction.


  1. GBD 2017 Diet Collaborators. Health effects of dietary risks in 195 countries, 1990–2017: A systematic analysis for the Global Burden of Disease Study 2017. Lancet 2019, 393, 1958–1972. [Google Scholar] [CrossRef]
  2. Mozaffarian, D. Dietary and Policy Priorities for Cardiovascular Disease, Diabetes, and Obesity: A Comprehensive Review. Circulation 2016, 133, 187–225. [Google Scholar] [CrossRef]
  3. Fernandez-Jimenez, R.; Al-Kazaz, M.; Jaslow, R.; Carvajal, I.; Fuster, V. Children Present a Window of Opportunity for Promoting Health: JACC Review Topic of the Week. J. Am. Coll. Cardiol. 2018, 72, 3310–3319. [Google Scholar] [CrossRef]
  4. Ambrosini, G.L. Childhood dietary patterns and later obesity: A review of the evidence. Proc. Nutr. Soc. 2014, 73, 137–146. [Google Scholar] [CrossRef]
  5. Aparicio, E.; Canals, J.; Voltas, N.; Valenzano, A.; Arija, V. Emotional Symptoms and Dietary Patterns in Early Adolescence: A School-Based Follow-up Study. J. Nutr. Educ. Behav. 2017, 49, 405–414. [Google Scholar] [CrossRef]
  6. Caleyachetty, R.; Echouffo-Tcheugui, J.B.; Tait, C.A.; Schilsky, S.; Forrester, T.; Kengne, A.P. Prevalence of behavioural risk factors for cardiovascular disease in adolescents in low-income and middle-income countries: An individual participant data meta-analysis. Lancet Diabetes Endocrinol. 2015, 3, 535–544. [Google Scholar] [CrossRef]
  7. Watts, A.W.; Mason, S.M.; Loth, K.; Larson, N.; Neumark-Sztainer, D. Socioeconomic differences in overweight and weight-related behaviors across adolescence and young adulthood: 10-year longitudinal findings from Project EAT. Prev. Med. 2016, 87, 194–199. [Google Scholar] [CrossRef] [Green Version]
  8. De Winter, A.F.; Visser, L.; Verhulst, F.C.; Vollebergh, W.A.M.; Reijneveld, S.A. Longitudinal patterns and predictors of multiple health risk behaviors among adolescents: The TRAILS study. Prev. Med. 2016, 84, 76–82. [Google Scholar] [CrossRef]
  9. Henriksson, P.; Henriksson, H.; Gracia-Marco, L.; Labayen, I.; Ortega, F.B.; Huybrechts, I.; Espana-Romero, V.; Manios, Y.; Widhalm, K.; Dallongeville, J.; et al. Prevalence of ideal cardiovascular health in European adolescents: The HELENA study. Int. J. Cardiol. 2017, 240, 428–432. [Google Scholar] [CrossRef]
  10. Cunha, C.M.; Costa, P.R.F.; de Oliveira, L.P.M.; Queiroz, V.A.O.; Pitangueira, J.C.D.; Oliveira, A.M. Dietary patterns and cardiometabolic risk factors among adolescents: Systematic review and meta-analysis. Br. J. Nutr. 2018, 119, 859–879. [Google Scholar] [CrossRef]
  11. Rocha, N.P.; Milagres, L.C.; Longo, G.Z.; Ribeiro, A.Q.; Novaes, J.F. Association between dietary pattern and cardiometabolic risk in children and adolescents: A systematic review. J. Pediatr. 2017, 93, 214–222. [Google Scholar] [CrossRef]
  12. Shang, X.; Li, Y.; Liu, A.; Zhang, Q.; Hu, X.; Du, S.; Ma, J.; Xu, G.; Li, Y.; Guo, H.; et al. Dietary pattern and its association with the prevalence of obesity and related cardiometabolic risk factors among Chinese children. PLoS ONE 2012, 7, e43183. [Google Scholar] [CrossRef]
  13. Ambrosini, G.L.; Emmett, P.M.; Northstone, K.; Howe, L.D.; Tilling, K.; Jebb, S.A. Identification of a dietary pattern prospectively associated with increased adiposity during childhood and adolescence. Int. J. Obes. 2012, 36, 1299–1305. [Google Scholar] [CrossRef] [Green Version]
  14. Smith, A.D.; Emmett, P.M.; Newby, P.K.; Northstone, K. A comparison of dietary patterns derived by cluster and principal components analysis in a UK cohort of children. Eur. J. Clin. Nutr. 2011, 65, 1102–1109. [Google Scholar] [CrossRef] [Green Version]
  15. Fernandez-Jimenez, R.; Santos-Beneit, G.; Tresserra-Rimbau, A.; Bodega, P.; de Miguel, M.; de Cos-Gandoy, A.; Rodriguez, C.; Carral, V.; Orrit, X.; Haro, D.; et al. Rationale and design of the school-based SI! Program to face obesity and promote health among Spanish adolescents: A cluster-randomized controlled trial. Am. Heart J. 2019, 215, 27–40. [Google Scholar] [CrossRef]
  16. Penalvo, J.L.; Santos-Beneit, G.; Sotos-Prieto, M.; Bodega, P.; Oliva, B.; Orrit, X.; Rodriguez, C.; Fernandez-Alvira, J.M.; Redondo, J.; Vedanthan, R.; et al. The SI! Program for Cardiovascular Health Promotion in Early Childhood: A Cluster-Randomized Trial. J. Am. Coll. Cardiol. 2015, 66, 1525–1534. [Google Scholar] [CrossRef]
  17. Santos-Beneit, G.; Bodega, P.; de Miguel, M.; Rodriguez, C.; Carral, V.; Orrit, X.; Haro, D.; Carvajal, I.; de Cos-Gandoy, A.; Penalvo, J.L.; et al. Rationale and design of the SI! Program for health promotion in elementary students aged 6 to 11 years: A cluster randomized trial. Am. Heart J. 2019, 210, 9–17. [Google Scholar] [CrossRef]
  18. Fernandez-Jimenez, R.; Jaslow, R.; Bansilal, S.; Santana, M.; Diaz-Munoz, R.; Latina, J.; Soto, A.V.; Vedanthan, R.; Al-Kazaz, M.; Giannarelli, C.; et al. Child Health Promotion in Underserved Communities: The FAMILIA Trial. J. Am. Coll. Cardiol. 2019, 73, 2011–2021. [Google Scholar] [CrossRef]
  19. Cespedes, J.; Briceno, G.; Farkouh, M.E.; Vedanthan, R.; Baxter, J.; Leal, M.; Boffetta, P.; Hunn, M.; Dennis, R.; Fuster, V. Promotion of cardiovascular health in preschool children: 36-month cohort follow-up. Am. J. Med. 2013, 126, 1122–1126. [Google Scholar] [CrossRef]
  20. Bel-Serrat, S.; Mouratidou, T.; Pala, V.; Huybrechts, I.; Bornhorst, C.; Fernandez-Alvira, J.M.; Hadjigeorgiou, C.; Eiben, G.; Hebestreit, A.; Lissner, L.; et al. Relative validity of the Children’s Eating Habits Questionnaire-food frequency section among young European children: The IDEFICS Study. Public Health Nutr. 2014, 17, 266–276. [Google Scholar] [CrossRef]
  21. Lanfer, A.; Hebestreit, A.; Ahrens, W.; Krogh, V.; Sieri, S.; Lissner, L.; Eiben, G.; Siani, A.; Huybrechts, I.; Loit, H.M.; et al. Reproducibility of food consumption frequencies derived from the Children’s Eating Habits Questionnaire used in the IDEFICS study. Int. J. Obes. 2011, 35, S61–S68. [Google Scholar] [CrossRef] [PubMed]
  22. Fernandez-Alvira, J.M.; Bornhorst, C.; Bammann, K.; Gwozdz, W.; Krogh, V.; Hebestreit, A.; Barba, G.; Reisch, L.; Eiben, G.; Iglesia, I.; et al. Prospective associations between socio-economic status and dietary patterns in European children: The Identification and Prevention of Dietary- and Lifestyle-induced Health Effects in Children and Infants (IDEFICS) Study. Br. J. Nutr. 2015, 113, 517–525. [Google Scholar] [CrossRef] [PubMed]
  23. Santos-Beneit, G.; Sotos-Prieto, M.; Pocock, S.; Redondo, J.; Fuster, V.; Penalvo, J.L. Association between anthropometry and high blood pressure in a representative sample of preschoolers in madrid. Rev. Española Cardiol. 2015, 68, 477–484. [Google Scholar] [CrossRef]
  24. Fryar, C.D.; Gu, Q.; Ogden, C.L. Anthropometric reference data for children and adults: United States, 2007–2010. Vital Health Stat. 2012, 11, 1–48. [Google Scholar]
  25. Flynn, J.T.; Kaelber, D.C.; Baker-Smith, C.M.; Blowey, D.; Carroll, A.E.; Daniels, S.R.; de Ferranti, S.D.; Dionne, J.M.; Falkner, B.; Flinn, S.K.; et al. Clinical Practice Guideline for Screening and Management of High Blood Pressure in Children and Adolescents. Pediatrics 2017, 140. [Google Scholar] [CrossRef] [PubMed]
  26. Sharma, A.K.; Metzger, D.L.; Daymont, C.; Hadjiyannakis, S.; Rodd, C.J. LMS tables for waist-circumference and waist-height ratio Z-scores in children aged 5–19 y in NHANES III: Association with cardio-metabolic risks. Pediatr. Res. 2015, 78, 723–729. [Google Scholar] [CrossRef] [PubMed]
  27. Tanner, J.M.; Whitehouse, R.H. Clinical longitudinal standards for height, weight, height velocity, weight velocity, and stages of puberty. Arch. Dis. Child. 1976, 51, 170–179. [Google Scholar] [CrossRef] [PubMed]
  28. Whitehead, S.J.; Ford, C.; Gama, R. A combined laboratory and field evaluation of the Cholestech LDX and CardioChek PA point-of-care testing lipid and glucose analysers. Ann. Clin. Biochem. 2014, 51, 54–67. [Google Scholar] [CrossRef] [PubMed]
  29. Chandler, J.L.; Brazendale, K.; Beets, M.W.; Mealing, B.A. Classification of physical activity intensities using a wrist-worn accelerometer in 8-12-year-old children. Pediatr. Obes. 2016, 11, 120–127. [Google Scholar] [CrossRef] [PubMed]
  30. UNESCO. International Standard Classification of Education, ISCED 2011; UNESCO Institute for Statistics: Montreal, QC, Canada, 2012. [Google Scholar]
  31. Perez-Rodrigo, C.; Gil, A.; Gonzalez-Gross, M.; Ortega, R.M.; Serra-Majem, L.; Varela-Moreiras, G.; Aranceta-Bartrina, J. Clustering of Dietary Patterns, Lifestyles, and Overweight among Spanish Children and Adolescents in the ANIBES Study. Nutrients 2016, 8, 11. [Google Scholar] [CrossRef]
  32. Northstone, K.; Emmett, P. Multivariate analysis of diet in children at four and seven years of age and associations with socio-demographic characteristics. Eur. J. Clin. Nutr. 2005, 59, 751–760. [Google Scholar] [CrossRef] [Green Version]
  33. Northstone, K.; Smith, A.D.; Cribb, V.L.; Emmett, P.M. Dietary patterns in UK adolescents obtained from a dual-source FFQ and their associations with socio-economic position, nutrient intake and modes of eating. Public Health Nutr. 2014, 17, 1476–1485. [Google Scholar] [CrossRef] [PubMed]
  34. Aranceta, J.; Perez-Rodrigo, C.; Ribas, L.; Serra-Majem, L. Sociodemographic and lifestyle determinants of food patterns in Spanish children and adolescents: The enKid study. Eur. J. Clin. Nutr. 2003, 57, S40–S44. [Google Scholar] [CrossRef] [PubMed]
  35. Lee, J.W.; Hwang, J.; Cho, H.S. Dietary patterns of children and adolescents analyzed from 2001 Korea National Health and Nutrition Survey. Nutr. Res. Pract. 2007, 1, 84–88. [Google Scholar] [CrossRef] [PubMed]
  36. Truthmann, J.; Richter, A.; Thiele, S.; Drescher, L.; Roosen, J.; Mensink, G.B. Associations of dietary indices with biomarkers of dietary exposure and cardiovascular status among adolescents in Germany. Nutr. Metab. 2012, 9, 92. [Google Scholar] [CrossRef] [PubMed]
  37. Leermakers, E.T.M.; van den Hooven, E.H.; Franco, O.H.; Jaddoe, V.W.V.; Moll, H.A.; Kiefte-de Jong, J.C.; Voortman, T. A priori and a posteriori derived dietary patterns in infancy and cardiometabolic health in childhood: The role of body composition. Clin. Nutr. 2018, 37, 1589–1595. [Google Scholar] [CrossRef] [PubMed]
  38. Bradlee, M.L.; Singer, M.R.; Qureshi, M.M.; Moore, L.L. Food group intake and central obesity among children and adolescents in the Third National Health and Nutrition Examination Survey (NHANES III). Public Health Nutr. 2010, 13, 797–805. [Google Scholar] [CrossRef] [PubMed]
  39. Smith, A.D.; Emmett, P.M.; Newby, P.K.; Northstone, K. Dietary patterns and changes in body composition in children between 9 and 11 years. Food Nutr. Res. 2014, 58. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  40. Golley, R.K.; McNaughton, S.A.; Hendrie, G.A. A dietary guideline adherence score is positively associated with dietary biomarkers but not lipid profile in healthy children. J. Nutr. 2015, 145, 128–133. [Google Scholar] [CrossRef] [PubMed]
  41. Togo, P.; Osler, M.; Sorensen, T.I.; Heitmann, B.L. Food intake patterns and body mass index in observational studies. Int. J. Obes. Relat. Metab. Disord. 2001, 25, 1741–1751. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  42. Newby, P.K. Are dietary intakes and eating behaviors related to childhood obesity? A comprehensive review of the evidence. J. Law Med. Ethics 2007, 35, 35–60. [Google Scholar] [CrossRef]
  43. Livingstone, M.B.; Robson, P.J. Measurement of dietary intake in children. Proc. Nutr. Soc. 2000, 59, 279–293. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  44. Blanchette, L.; Brug, J. Determinants of fruit and vegetable consumption among 6–12-year-old children and effective interventions to increase consumption. J. Hum. Nutr. Diet. 2005, 18, 431–443. [Google Scholar] [CrossRef] [PubMed]
  45. Hearty, A.P.; Gibney, M.J. Dietary patterns in Irish adolescents: A comparison of cluster and principal component analyses. Public Health Nutr. 2013, 16, 848–857. [Google Scholar] [CrossRef] [PubMed]
Figure 1. Mean standardized principal component scores for each cluster.
Figure 1. Mean standardized principal component scores for each cluster.
Nutrients 11 02297 g001
Table 1. Z-Scores for relative consumption frequencies characterizing each cluster.
Table 1. Z-Scores for relative consumption frequencies characterizing each cluster.
Food ItemsProcessed
(n = 386)
(n = 518)
(n = 420)
Cooked vegetables, potatoes, beans−0.280.75−0.230.550.541.36
Fried potatoes, potato croquettes0.421.23−0.080.78−0.300.87
Raw vegetables−0.320.67−0.220.730.561.27
Fresh fruit with no added sugar−0.490.57−0.120.860.601.16
Fresh fruit with added sugar0.291.15−0.150.82−0.081.00
Fruit juice0.331.24−0.120.86−0.150.84
Sweetened drinks0.531.43−0.150.68−0.300.58
Diet drinks0.381.38−0.180.55−0.130.92
Breakfast cereals, muesli (sweetened)0.050.990.001.00−0.051.02
Porridge, oatmeal, cereals, muesli (unsweetened)−0.090.78−0.320.460.481.40
Plain unsweetened milk−0.270.67−0.190.820.481.26
Sweetened milk−0.050.780.351.18−0.390.77
Plain unsweetened yogurt or kefir−0.050.88−0.210.820.301.21
Sweet yogurt, fermented milk beverages0.020.860.151.18−0.210.84
Fresh or frozen fish, not fried−0.280.97−0.080.830.361.11
Fried fish, fish fingers0.170.950.021.00−0.181.02
Cold cuts, preserved, ready to cook meat products−0.220.830.311.14−0.180.85
Fresh meat, not fried−0.200.980.010.910.181.09
Fried meat0.−0.310.80
Fried or scrambled eggs0.−0.240.87
Boiled or poached eggs0.221.23−0.230.730.090.99
Mayonnaise, mayonnaise-based products0.411.28−0.060.95−0.300.53
Meat replacement products, plant milk−0.060.52−0.190.350.291.62
Jam, honey0.091.06−0.200.720.161.19
Chocolate or nut-based spreads0.421.20−0.010.93−0.370.69
Butter, margarine on bread0.271.19−0.160.86−0.050.93
Reduced-fat products on bread0.241.21−0.180.770.000.99
White bread, white roll, white crispbread−0.310.660.431.20−0.240.78
Wholemeal bread, dark roll, dark crispbread−0.080.85−0.150.850.261.22
Pasta, noodles, rice0.020.990.021.05−0.050.95
Cereals, milled0.261.47−0.220.420.030.90
Pizza as main dish0.401.23−0.040.86−0.320.78
Hamburgers, hot dogs, kebabs, wraps, falafel0.511.29−0.100.80−0.340.69
Nuts, seeds, dried fruits0.221.31−0.200.730.040.91
Crisps, maize (corn) crisps, popcorn0.541.34−0.050.74−0.440.61
Savory pastries, fritters0.441.22−0.150.80−0.220.85
Chocolate, candy bars0.381.28−0.080.92−0.250.64
Candies, loose candies, marshmallows0.431.30−0.100.83−0.280.70
Biscuits, packaged cakes, pastries, puddings0.140.910.131.19−0.290.74
Ice cream, milk, or fruit-based bars0.471.33−0.170.79−0.220.69
Values are expressed as mean with standard deviation. Bold entries are the highest mean value within a row.
Table 2. Principal components and factor loadings.
Table 2. Principal components and factor loadings.
Food ItemsProcessedTraditionalHealthy
Cooked vegetables, potatoes, beans0.070.140.61
Fried potatoes, potato croquettes0.570.140.11
Raw vegetables0.030.100.57
Fresh fruit with no added sugar−0.03−0.040.58
Fresh fruit with added sugar0.440.240.20
Fruit juice0.380.210.14
Sweetened drinks0.520.140.02
Diet drinks0.400.110.17
Breakfast cereals, muesli (sweetened)0.320.220.17
Porridge, oatmeal, cereals, muesli (unsweetened)0.26−0.040.55
Plain unsweetened milk0.07−0.080.50
Sweetened milk0.250.43−0.03
Plain unsweetened yogurt or kefir0.100.100.43
Sweet yogurt, fermented milk beverages0.290.350.07
Fresh or frozen fish, not fried−
Fried fish, fish fingers0.200.500.19
Cold cuts, preserved, ready to cook meat products0.020.530.04
Fresh meat, not fried−0.030.520.32
Fried meat0.150.690.11
Fried or scrambled eggs0.380.520.16
Boiled or poached eggs0.420.100.27
Mayonnaise, mayonnaise-based products0.360.52−0.05
Meat replacement products, plant milk0.030.000.31
Jam, honey0.350.250.21
Chocolate or nut-based spreads0.630.36−0.04
Butter, margarine on bread0.280.410.00
Reduced-fat products on bread0.420.330.16
White bread, white roll, white crispbread0.100.32−0.05
Wholemeal bread, dark roll, dark crispbread0.100.260.34
Pasta, noodles, rice0.230.490.15
Cereals, milled0.310.190.30
Pizza as main dish0.630.110.03
Hamburgers, hot dogs, kebabs, wraps, falafel0.490.350.06
Nuts, seeds, dried fruits0.370.330.24
Crisps, maize (corn) crisps, popcorn0.630.34−0.08
Savouy pastries, fritters0.570.230.06
Chocolate, candy bars0.650.16−0.02
Candies, loose candies, marshmallows0.60−0.040.01
Biscuits, packaged cakes, pastries, puddings0.530.090.01
Ice cream, milk-, or fruit-based bars0.610.190.11
Bold entries are factor loadings >0.30, which are considered important contributors to the patterns.
Table 3. Adolescent and family characteristics in the total population and stratified by dietary pattern cluster.
Table 3. Adolescent and family characteristics in the total population and stratified by dietary pattern cluster.
Age, years, mean (SD)12.54 (0.45)12.66 (0.55)12.52 (0.40)12.46 (0.37)<0.001
Gender, n (%) <0.001
Boys683 (51.59%)230 (59.59%)267 (51.54%)186 (44.29%)
Girls641 (48.41%)156 (40.41%)251 (48.46%)234 (55.71%)
Municipality, n (%) 0.302
Barcelona901 (68.05%)265 (68.65%)362 (69.88%)274 (65.24%)
Madrid423 (31.95%)121 (31.35%)156 (30.12%)146 (34.76%)
Parental education level, n (%) <0.001
Low214 (16.16%)101 (26.17%)68 (13.13%)45 (10.71%)
Medium529 (39.95%)169 (43.78%)222 (42.86%)138 (32.86%)
High518 (39.12%)91 (23.58%)203 (39.19%)224 (53.33%)
Unknown63 (4.76%)25 (6.48%)25 (4.83%)13 (3.10%)
Parental origin, n (%) <0.001
Spanish850 (64.20%)216 (55.96%)371 (71.62%)263 (62.62%)
Migrant background397 (29.98%)146 (37.82%)113 (21.81%)138 (32.86%)
Unknown77 (5.82%)24 (6.22%)34 (6.56%)19 (4.52%)
MVPA, min/day, mean (SD)76.63 (22.75)76.47 (23.94)76.56 (23.45)76.55 (23.43)0.996
MVPA: Moderate-to-vigorous physical activity. Values are expressed as mean (standard deviation) for continuous variables and as frequencies (percentages) for categorical variables. p-values from X2 and ANOVA test.
Table 4. Cardiovascular risk factors according to dietary pattern clustering.
Table 4. Cardiovascular risk factors according to dietary pattern clustering.
nProcessed ClusterTraditional ClusterHealthy Clusterp For Trend
Mean *95% CIMean *95% CIMean *95% CI
z-BMI11090.11 a,b−0.02, 0.240.46 a0.35, 0.570.44 b0.32, 0.57<0.001
z-WC11090.270.17, 0.360.260.17, 0.340.280.19, 0.370.719
z-Systolic BP1106−0.08−0.26, 0.10−0.12−0.29, 0.05−0.13−0.31, 0.050.760
z-Diastolic BP1106−0.07−0.24, 0.11−0.04−0.21, 0.13−0.06−0.24, 0.110.813
z-FMI1109−0.06−0.12, 0.000.03−0.03, 0.080.01−0.05, 0.060.048
Glucose (mg/dL)1109104.2101.4, 107.1102.499.8, 105.1103.5100.7, 106.30.274
TC (mg/dL)1109148.6 b141.6, 155.5153.1146.4, 159.8157.6 b150.7, 164.40.003
HDL-cholesterol (mg/dL)110761.358.5, 64.062. 960.3, 65.563.360.6, 66.00.212
LDL-cholesterol (mg/dL)85575.971.2, 80.777.373.0., 81.679.675.1, 84.10.279
Non-HDL cholesterol (mg/dL)103889.0 b84.4, 93.691.987.7, 96.095.1 b90.8, 99.50.027
Triglycerides (mg/dL)110879.074.4, 83.778.374.3, 82.379.174.7, 83.50.942
TC:HDL ratio9562.562.47, 2.652.502.43, 2.582.572.49, 2.650.318
LDL:HDL ratio7821.331.25, 1.411.291.22, 1.361.351.28, 1.430.386
BMI: Body mass index (kg/m2); WC: Waist circumference; BP: Blood pressure; FMI: Fat mass index; TC: Total cholesterol. * Estimated marginal means from multilevel linear mixed-effects models. Variables in the models were age, gender, fasting status, parental education level, migrant background, z-BMI, tanner stage, and moderate and vigorous physical activity and included a random school and a random municipality effect to account for the clustered study design. a Significant differences between Processed and Traditional clusters after Bonferroni correction; b Significant differences between Processed and Healthy clusters after Bonferroni correction.

Share and Cite

MDPI and ACS Style

Bodega, P.; Fernández-Alvira, J.M.; Santos-Beneit, G.; de Cos-Gandoy, A.; Fernández-Jiménez, R.; Moreno, L.A.; de Miguel, M.; Carral, V.; Orrit, X.; Carvajal, I.; et al. Dietary Patterns and Cardiovascular Risk Factors in Spanish Adolescents: A Cross-Sectional Analysis of the SI! Program for Health Promotion in Secondary Schools. Nutrients 2019, 11, 2297.

AMA Style

Bodega P, Fernández-Alvira JM, Santos-Beneit G, de Cos-Gandoy A, Fernández-Jiménez R, Moreno LA, de Miguel M, Carral V, Orrit X, Carvajal I, et al. Dietary Patterns and Cardiovascular Risk Factors in Spanish Adolescents: A Cross-Sectional Analysis of the SI! Program for Health Promotion in Secondary Schools. Nutrients. 2019; 11(10):2297.

Chicago/Turabian Style

Bodega, Patricia, Juan Miguel Fernández-Alvira, Gloria Santos-Beneit, Amaya de Cos-Gandoy, Rodrigo Fernández-Jiménez, Luis Alberto Moreno, Mercedes de Miguel, Vanesa Carral, Xavier Orrit, Isabel Carvajal, and et al. 2019. "Dietary Patterns and Cardiovascular Risk Factors in Spanish Adolescents: A Cross-Sectional Analysis of the SI! Program for Health Promotion in Secondary Schools" Nutrients 11, no. 10: 2297.

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop