Ultraprocessed Food Consumption and Cardiometabolic Risk Factors in Children

Key Points Question What is the association of consuming ultraprocessed foods (UPFs) with cardiometabolic risk factors in children? Findings In this cross-sectional study of 1426 children, higher consumption of UPFs was positively associated with body mass index, waist circumference, fat mass index, and fasting plasma glucose and negatively associated with high-density lipoprotein cholesterol concentrations. Meaning These findings highlight the need for public health initiatives to promote the replacement of UPFs with unprocessed or minimally processed foods.


Introduction
3][4] Emerging studies have shed light on the potential role of ultraprocessed foods (UPFs) in determining the risk of chronic diseases, [5][6][7] independent of their nutritional profiles. 8mmonly, UPFs represent a category of food products that undergo extensive industrial processing, often containing multiple ingredients, additives, and preservatives to make them not only convenient (ready to eat) but also palatable and appealing.This approach has been used to create the most widely used UPF classification, the NOVA Food Classification system. 9,10traprocessed foods are typically rich in saturated fats, sugars, sodium, and other substances (eg, additives) and lower in essential nutrients, all of which are associated with cardiometabolic health. 11,12e to their high availability and affordability and wide marketing to children, UPFs have become increasingly frequent in modern diets, particularly among children, adolescents, [13][14][15] and their families, and especially among individuals and families with low socioeconomic status and educational levels in which obesity is more prevalent. 16[20] Previous observational studies in adults have reported positive associations between UPF consumption and obesity, 21 type 2 diabetes, 22 CVD, 23 and all-cause mortality 24 ; however, the epidemiologic evidence in children remains limited and controversial. 25While the majority of studies have reported unfavorable associations with body mass index (BMI), others did not find this association, and few have focused on cardiometabolic risk factors. 25ven the public health burden of CVDs and the increasing availability of UPFs, having a better understanding of potential associations between UPF consumption and cardiometabolic risk factors in children is essential.Therefore, the aim of this study was to examine the associations between UPF consumption and cardiometabolic risk factors in a population of Spanish preschool children (aged 3-6 years).

Methodology Population and Study Design
This cross-sectional study was conducted using data from the Childhood Obesity Risk Assessment Longitudinal Study (CORALS), which followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.The ethics committee of each of recruitment center approved the study protocol, which was conducted following the standards of the Declaration of Helsinki. 26Parents or caregivers provided written informed consent.
CORALS is an ongoing prospective multicenter study conducted in 7 Spanish centers aiming to identify potential risk factors for childhood obesity over a 10-year follow-up period.A detailed description of the CORALS is published elsewhere. 27Between March 22, 2019, and June 30, 2022, eligible participants aged 3 to 6 years at enrollment were recruited from schools across 7 cities in Spain.To be enrolled in the study, parents or caregivers had to sign a consent form, attend the inclusion face-to-face visit and complete several questionnaires at home for data collection on leisure time physical activity, 3-day food consumption, and sociodemographics.The exclusion criteria included belonging to a family with difficulty collaborating due to low command of Spanish or unstable residence.

Dietary Intake of UPFs and NOVA Food Classification System
To estimate the dietary intake of UPFs, trained dietitians (B.P.-V., S.d.L.H.D., M.L.M.-B., K.A.P.-V., and R.V.-C.) used the validated, semiquantitative, 125-item food and beverage frequency COME-Kids questionnaire. 28Participants with energy intake below the first percentile or above the 99th percentile were excluded to minimize misreporting (details are provided in the eAppendix in Supplement 1).The NOVA Food Classification system was used to determine the consumption of food depending on its degree of processing 27 (details are provided in the eAppendix and eTable 1 in Supplement 1).

Outcomes Adiposity Measurements
Adiposity measurements and cardiometabolic risk factor assessments were conducted in health care centers.Weight and body fat mass were measured using a precision scale and an octopolar multifrequency bioelectrical impedance device (MC780MAS; Tanita).Height was measured using a portable seca 213 stadiometer according to standard procedures.Body mass index was calculated and categorized as underweight or normal weight or as overweight or obesity based on pediatric cutoffs. 29Waist circumference was determined using a flexible, nonextensible measuring tape.The fat mass index was estimated by dividing body fat mass in kilograms by height in meters squared. 30ist-to-height ratio was estimated by dividing waist circumference in centimeters by height in centimeters. 31

Cardiometabolic Risk Factors Assessment
Blood pressure was measured in the nondominant arm 3 times, with a 5-minute gap between each measurement, using an automatic oscillometer (M3 Intelligence HEM-75051-EV; OMRON Healthcare) equipped with a child-sized cuff.Eight-hour fasting blood samples were collected from participants, and serum total cholesterol, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol, triglycerides, plasma glucose, and insulin concentrations were measured using standard procedures.Homeostasis model assessment for insulin resistance (HOMA-IR) was calculated as fasting insulin (μIU/mL) × fasting glucose (mmol/L) / 22.5.

Covariates
Parents or caregivers were provided with a set of questionnaires to complete at home, gathering information on early life factors, maternal characteristics, and lifestyle patterns.To assess physical activity, the total time (in hours) of engagement in sports and physical activities per week was estimated using a previously validated questionnaire. 32An 18-item questionnaire for children was used to assess adherence to the Mediterranean diet, an indicator of diet quality. 27

Statistical Analysis
The current analysis was conducted using the CORALS database updated through January 20, 2023.
Analysis of descriptive baseline characteristics are reported as mean (SD) or median (IQR) for continuous variables and as numbers with percentages for categorical variables using one-way analysis of variance and χ 2 test, respectively.Consumption of UPFs (in grams per day) was adjusted for total energy intake using the residual regression method. 33Intake of UPFs in grams per day was calculated instead of energy percentage to account for foods with no energy content (eg, artificially sweetened beverages) and for nonnutritional concerns associated with food processing (eg, food additives).Participants were categorized by tertiles of energy-adjusted UPF consumption, ranging from tertile 1 for the lowest intake to tertile 3 for the highest intake.
Age-and sex-specific z scores of each outcome were estimated from standardized residuals conducted using linear regression models.Missing data of less than 5% for covariates were imputed to the mean and to the highest frequency category for quantitative and qualitative confounders, respectively. 34Multivariable linear regression models were fitted to assess the associations (β coefficient and 95% CI) between tertiles of energy-adjusted UPFs and z scores of cardiometabolic risk factors.The first tertile (lowest intake) was considered the reference.Models were adjusted for maternal education level (primary or lower, secondary or university), maternal BMI (underweight, normal overweight, obesity), physical activity (minutes per week), exclusive breastfeeding (yes or no), recruitment center size (<200, 200-400, >400 participants), and NOVA group 1, 2, or 3 (as detailed in eTable 1 in Supplement 1).To assess the linear trend, the median value of each tertile of UPF consumption was modeled as a continuous variable.The analysis was also conducted in a continuous form, with a 1-SD increment and using the same confounders.
Additionally, a simulation model was fitted to substitute 100 g of consumed UPFs with 100 g of unprocessed or minimally processed food to examine the association of healthier food consumption with the outcomes.The theoretical impact of substituting 1 food group for another was assessed by introducing both variables simultaneously as continuous variables into the model.Differences in the β coefficients, variances, and covariance were used to estimate the β coefficients and 95% CIs for the substitution association.Sensitivity analyses were conducted to assess associations according diet quality, maternal education, and socioprofessional level (details provided in the eAppendix in Supplement 1).
Data were analyzed using Stata, version 14 software (StataCorp LLC).All statistical tests were 2-sided, and P < .05 was deemed statistically significant.

Results
A total of 1426 participants (mean [SD] age, of 5.8 [1.1] years; 698 boys [49.0%] and 728 girls [51.0%]) were included in this study after excluding 54 participants lacking the food and beverage frequency questionnaire and 29 with missing data or implausible reported energy intake (eFigure 1 in Supplement 1).The characteristics of the study population across tertiles of energy-adjusted UPF consumption are shown in Table 1.Children in the third tertile (highest UPF consumption) had a higher BMI z score, waist-to-height ratio, fat mass index, systolic blood pressure, and overweight or obesity prevalence and lower HDL and LDL cholesterol.Mothers whose children were categorized in the highest tertile of energy-adjusted UPF consumption were younger, had a higher BMI, were more prone to be living with overweight or obesity, were less likely to have exclusively breastfed their children, and had lower educational achievement and employment rates.
General dietary characteristics of participants are shown in Table 2. Children in the top tertile were more likely to consume higher amounts of total energy, carbohydrates, yogurt, other dairy products, sugar and candy, and sugary beverages and lower amounts of protein, fat, monounsaturated and polyunsaturated fatty acids, fiber, milk, cheese, white meat, unprocessed red meat, eggs, fish, seafood, vegetables, fruits, nuts, legumes, whole and refined cereals, and oils and fat.
Children whose mothers had a low education level had higher z scores of waist circumference (β coefficient, 0.14; 95% CI, 0.05-0.

Discussion
To our knowledge, this study is the first to assess the associations between UPF consumption and various cardiometabolic risk factors in young children.In this large cross-sectional study, UPF consumption was positively associated with z scores of BMI, waist circumference, fat mass index, and fasting plasma glucose concentration and inversely associated with HDL cholesterol concentration.
6][37] Consumption of UPFs at age 4 years was associated with increased BMI z scores at age 10 years in the Generation XXI cohort, while no association was found at age 7 years. 35Another study showed that high UPF consumption in children aged 7 to 13 years was associated with increased BMI growth trajectories. 36Similarly, lower UPF intake in Spanish children aged 4 to 7 years was associated with lower BMI z scores at age 7 years, though this association became nonsignificant after adjusting for maternal factors. 37global study by Neri et al 38 revealed that increased UPF consumption was associated with higher dietary energy density and intake of free sugars, alongside decreased total fiber intake, potentially contributing to childhood obesity.Additionally, findings from the Avon Longitudinal Study of Parents and Children showed that high UPF consumption was associated with unfavorable fat mass index trajectories from age 7 to 24 years. 36Similarly, in a Brazilian cohort, UPF consumption during preschool years was associated with increases in waist circumference from preschool to school age. 39Other studies found no significant association between UPF consumption and HDL cholesterol and fasting plasma glucose concentrations. 37,40Therefore, to our knowledge, our study is the first in children to find significant associations with the aforementioned risk factors and is in line with other studies assessing adult populations. 41r results provide new insight into the association between UPF consumption and health and the importance of recognizing that early dietary habits in childhood might have future implications on cardiometabolic health.While the magnitude of the associations reported in our study may be considered of limited clinical relevance, it is important to note that our study consisted of young children.Therefore, if such minimal differences can reveal a significant association, they may serve as an early warning of future cardiometabolic conditions.
3][44] In addition, our results support the findings of other European studies that have shown that children of mothers with lower education or with lower socioeconomic status are more likely to consume UPFs.These findings suggest that educational and socioeconomic factors may contribute to the purchase of low-cost and unhealthy foods, such as UPFs, increasing the risk of health disorders. 37,45,46veral possible mechanisms could explain our results.First, UPFs contain higher amounts of sodium, energy, fat, and sugar and lower amounts of fiber, which are well recognized as contributors to cardiometabolic risk factors. 47[50] Moreover, excessive consumption of energy, saturated fat, and sugar contributes to weight gain and a higher risk of obesity, which is an important risk factor in CVD. 51Furthermore, our study showed that children who consumed high amounts of UPFs tended to have lower intakes of fruits and vegetables, which, along with a healthy dietary pattern, are known to be beneficial for cardiometabolic health. 52st of the associations were maintained in our study after further adjusting the models to Mediterranean diet adherence, suggesting that other intrinsic UPF factors may play an important role in determining these associations (eg, additives).4][55][56][57][58] Food processing generates contaminants such as acrylamide and acrolein, which have been linked, respectively, to increased odds and risk of cardiovascular disease. 59,60traprocessed foods may contain chemicals such as bisphenols and perfluoroalkyl substances that have been associated with a higher risk of cardiometabolic outcomes in children. 61,62e NOVA Food Classification system has sparked debate among researchers due to disagreements over UPF definitions, bias concerns, and the system's contribution to dietary guidelines. 63,64The NOVA system itself has some limitations, as it does not consider that certain minimal processing could improve the final product (eg, fermentation in milk) and adopts a vague definition of what is considered a cosmetic additive, which has led to considering carotenoids as an additive that increases the potential harmfulness of a product. 65Despite these limitations, NOVA categories have consistently shown associations with cardiometabolic health in adults.

Strengths and Limitations
This study has several strengths.Most importantly, the study was conducted in a large sample size from 7 different geographic areas of Spain.[23][24] Our study also has several limitations.First, because the study is observational, we cannot draw conclusions on cause and effect.Second, our study involved preschool children from Spain, which means that the generalization of our findings to different populations is not appropriate.Third, some grade of misclassification could be present in our study since UPF consumption was estimated from a food and beverage frequency questionnaire that was not specifically developed to assess this type of food, which could result in either an overestimation or underestimation of consumption within various NOVA categories.Additionally, imprecise estimations could also arise from the use of a food and beverage frequency questionnaire, which may be influenced by social desirability bias.Finally, we cannot dismiss that associations may be due to residual confounding or that undetected cardiometabolic disorders in our study population may exist due to age.

Table 1 .
General Characteristics of Study Participants Across Tertiles of Energy-Adjusted Ultraprocessed Food Consumption a P values for comparisons were tested by 1-way analysis of variance or χ 2 test, as appropriate, according to tertiles of energy-adjusted ultraprocessed food consumption.

Table 2 .
Baseline Dietary Characteristics of Participants Across Tertiles of Energy-Adjusted Ultraprocessed Food Consumption in the DietSimilar positive associations among fat mass index, BMI, and plasma glucose were observed, irrespective of the animal or vegetable origin of the UPFs consumed.Substitution of 100 g of UPFs with 100 g of unprocessed or minimally processed foods was associated with a decrease in z scores a P values for comparisons were tested by one-way analysis of variance or χ 2 test, as appropriate, according to energy-adjusted ultraprocessed food consumption.

Table 3 .
Association Between Energy-Adjusted Ultraprocessed Food Consumption in Tertiles and 1-SD Increments and Cardiometabolic Risk Factor z Scores Abbreviations: BMI, body mass index; HDL, high-density lipoprotein; HOMA-IR, homeostasis model assessment for insulin resistance; LDL, low-density lipoprotein.