Out of an initial screening of 126 children, 121 met the inclusion criteria. Of these, 114 completed the 8-week phase, with a drop-out rate of 5.8%, primarily due to discouragement, changes of address, social problems or scheduling conflicts. One hundred and nine participants completed the dietary intake (at baseline and 8 weeks), but two subjects were identified as outlier and were removed from the statistical analysis. The final sample was 107 children and adolescents: 81 subjects of intervention group and 26 of usual care group (Supplementary Figure S1).

Participants were randomly assigned to the usual care group or the intensive care group in a 1:3 allocation ratio using a computer-generated randomization sequence. The randomization sequence was generated by an independent researcher not involved in participant recruitment, enrollment, or outcome assessment, and allocation was concealed through a centralized assignment system. Randomization was performed after participant enrollment and completion of baseline assessments. The unequal allocation ratio was chosen to allow a greater number of participants to receive intensive intervention. Due to the nature of the intervention, blinding of participants and clinicians was not feasible. Randomization resulted in comparable baseline characteristics between the study groups, as shown in Table 1.

A multidisciplinary team comprising registered dieticians, pediatricians, physical activity experts and nurses carried out the 2 year-lifestyle intervention, which consisted of an 8-week intensive phase, followed by a 22-month follow-up period (20). This report presents data from the intensive phase. During this period, the usual care subjects received pediatric recommendations for a healthy diet based on the national guidelines of the Spanish Society of Community Nutrition (21) in one 30-min individual session with the dietitian, and 5 monitoring visits to evaluate anthropometric measurements with the research team. These recommendations focused on promoting a balanced and varied diet, healthy lifestyle habits, and adequate energy balance, with emphasis on regular physical activity and healthy culinary practices. In contrast, intervention subjects were advised to follow a moderately hypocaloric Mediterranean diet (22, 23). The dietary plan is detailed in Supplementary Table S1. Briefly, dietary intervention consists on a fixed full-day meal plan of five meals. Total daily energy was distributed among the day following the pattern: 20% on breakfast, 5%–10% on morning snack, 30%–35% on lunch, 10%–15% on afternoon snack and 20%–25% on dinner. The dietary plan was based in Mediterranean pattern; thus, it includes a high consumption of fruits, vegetables, whole grains, legumes, olive oil and minimally processed foods; a moderate consumption of dairy products, fish and poultry; and a low consumption of red meat. Intervention subjects received six 30-min individual sessions to both assess and address possible problems with dietary compliance along with anthropometric evaluation. In addition, the subjects and their parents (or legal guardians) attended a parallel group session in the third week. During the group sessions parents were told their role in the intervention and the obesity related comorbidities, while children were taught about different topics such as energy balance, portion sizes, groups of foods, the importance of the breakfast and physical activity. Both groups were encouraged to engage in an additional 200 min of physical activity per week at 60%–75% of their maximum heart rate. All participants were accompanied by their parents or legal guardians (20).

All measurements were taken at baseline and after the 8-week intervention by trained personnel using standard procedures and using calibrated equipment. Height was obtained with a Harpenden’s stadiometer of 1 mm precision (Seca 220, Vogel and Halke, Hamburg, Germany) and body weight using a digital scale (BC-418, TANITA, Tokyo, and Japan). Body Mass Index (BMI) was calculated from the ratio of weight to height squared (kg/m²) and converted after into standard deviation scores (SDs) for sex and age derived from Spanish reference data according to specific cutoff points for BMI (19). Waist and hip circumferences were measured with a non-stretchable measuring tape (Type SECA 200). Somatic maturity was evaluated according to Tanner stages by the pediatricians (24). Venous blood samples were collected after an overnight fast to measure glucose, insulin, and lipid profile using standard auto analyzer techniques. Insulin resistance and sensitivity were evaluated using Homeostasis Model Assessment of Insulin Resistance (HOMA-IR) and Quantitative Insulin Sensitivity Check Index (QUICKI), respectively. HOMA-IR was calculated as [fasting glucose (mg/dL) x fasting insulin (μU/mL)]/405 (25). QUICKI was calculated as 1/[log (fasting insulin (μU/mL)) + log (fasting glucose (mg/dL))] (26). Blood Pressure (BP) was measured three times using an electronic sphygmomanometer (OMRON M6, Hoofddorp, The Netherlands) on the right arm after the children had rested quietly for 15 min.

Physical Activity, sedentary behavior, and sleep duration were objectively measured during leisure time over 4 consecutive full days, 2 weekdays and 2 weekend days by triaxial accelerometry (Actigraph wGT3X-BT, Actigraph LLC, Penascola, FL, USA). Participants and parents were instructed to wear the accelerometer around the non-dominant waist all the time, including sleep time, and removing it just for water-related activities (bathing or showering). The accelerometer collected data in 60-s intervals (27), which were analyzed using ActiLife 6.0 software (Actigraph LLC, Penascola, FL, USA), and summarized as counts per min (CPM) using validated cutoff points to define time spent sedentary (<100 CPM) and moderate-to-vigorous physical activity (MVPA) (>2,296 CPM) (28). Sleep was assessed using the accelerometer Actigraph wGT3X-BT, and the data was analyzed using the Sadeh algorithm derived from fundamental research performed by Sadeh et al. (29). This algorithm was commonly used in younger adolescents (10 to 11 years old) (30).

Trained dieticians collected dietary intake data at baseline and after an 8-week period using a semi-quantitative 136-item Food-Frequency Questionnaire (FFQ) (23), previously validated in Spain (31, 32). Diet quality was evaluated through three dietary indices that provide different insights into various aspect of one’s dietary patterns, encompassing diversity of food intake (Diet Quality Index for Adolescents, DQI-A), lifestyle habits such as physical activity (Healthy Lifestyle Diet Index, HLD-I) and adherence to the Mediterranean diet (Mediterranean Diet Quality Index, KIDMED), as previously reported (20).

The DQI-A, which was previously validated and adapted for use in adolescents (33), is composed by the sum of three categories presented in percentages: quality, diversity, and equilibrium. Its total value ranges from −33% to 100% (33). The HLD-Index is composed of 10 items, eight of which refer to the frequency of consumption of fruit, vegetables, fish and seafood, sweets, regular soft drinks, grain, dairy products, meat, and meat products. The other two components indicate the level of physical activity through measuring the time spent on moderate to vigorous physical activity versus looking at screens (34). The HLD-Index employed in this study is a modified version of the original index due to a lack of information on screen hours, with a range from 0 and 36. Finally, the KIDMED index score evaluated the adequacy of Mediterranean dietary patterns in children and adolescents, with scores ranging from 0 to 12 (35).

The estimated daily food consumption was calculated by multiplying the portion size in grams by the consumption frequency for each item. To evaluate the consumption of UPFs, we classified all FFQ items according to NOVA classification system which divides food in four groups according to its nature, extent, and purpose of industrial food processing (36). UPFs are industrial formulations of food-derived substances (oils, fats, sugars, starch, protein isolates) that contain little or no whole food and often include flavorings, colorings, emulsifiers, and other cosmetic additives (e.g., carbonated drinks; sweet and savory packaged snacks; ice-cream, chocolate, and candies (confectionery); mass-produced packaged bread and buns; margarines and spreads; cookies, pastries, cakes, and cake mixes; breakfast “cereals,” “cereal” and “energy” bars; “energy drinks”; milk drinks, “fruit” yogurts, and “fruit” drinks; cocoa drinks; meat and chicken extracts and “instant” sauces; infant formulas, follow-on milks, and other baby products; “health” and “slimming” products such as powdered or “fortified” meal and dish substitutes; and many ready-to-heat products including prepared pies and pasta and pizza dishes; poultry and fish “nuggets” and “sticks,” sausages, burgers, hot dogs, and other reconstituted meat products; and powdered and packaged “instant” soups, noodles, and desserts). The sum of UPF items was estimated using the frequency of UPF consumption per person (servings/day).

LE8 is an updated metric for assessing CVH, encompassing 4 health behaviors: diet (updated), physical activity, nicotine exposure (updated), sleep health (new); and health factors like BMI, blood lipids (updated), blood glucose (updated), and BP; each metric range from 0 to 100 points (8). Given the absence of nicotine exposure data, we included 7 metrics (physical activity, sleep health, BMI, blood lipids, blood glucose, and BP) to construct LE8 as defined by the AHA. This approach is validated by a study using data from three large nationwide prospective cohorts (NHS, NHSII, and HPFS) and a nationally representative survey (NHANES). The study demonstrated that CVH-related factors routinely measured in many research and clinical settings can accurately estimate individuals’ overall CVH, even when not all eight LE8 metrics are available (37).

The sample size calculation indicated that 13 subjects were needed in usual care and 39 in the intervention group, based on a 5% error rate, 90% power, a 1:3 ratio, and a mean difference of 0.50 (SD 0.47) units in BMI-SDS after the intervention (38). Variables were described using mean ± SD; all statistical analyzes were two-tailed and a p-value < 0.05 was considered as statistically significant. The change in each variable was calculated as the difference between post- and pre- intervention values for each subject. Student’s unpaired or paired t-test were used for comparison between groups or within group changes, respectively. Differences in fasting glucose and systolic BP (SBP) between groups, which showed baseline differences, were analyzed using ANCOVA adjusted for baseline values. Differences in LE8 and UPF consumption were evaluated using ANCOVA adjusted for changes in total energy intake and somatic maturity (Tanner stage), given the wide age range of participants and the weight-loss intervention design; this adjustment allowed isolation of the associations of interest independently of overall caloric reduction. Associations between CVH and lifestyle or clinical variables were assessed using multiple linear regression adjusted for the same covariates. Stata/MP 14.1 for Mac (version 14.1, College Station, TX: StataCorp LP, USA) was used as statistical software.

Linear regression models were fitted using bootstrapped (10,000 samples) mediation techniques contained in the PROCESS R macro (version 4.3), using Model 4. The analyzes aimed to determine whether the effect of the weight loss program (independent variable) on the change in LE8 (dependent variable) was mediated by the change in UFPs consumption (mediator). The mediation model included the following paths: (a) the effect of the weight loss program on the change in UPFs consumption (path a); (b) the effect of the change in UPFs consumption on the change in LE8 score, controlling for the weight loss program (path b); (c) the total effect of the weight loss program on the change in LE8 score (path c); (c’): the total effect of the weight loss program on the change in LE8 score after accounting for the mediator (path c’). The indirect effect (ab) was estimated using percentile bootstrap 95% confidence intervals. Given the use of modified versions of the HLD-I and LE8 scores, all analyzes involving these indices were conducted with an exploratory aim. No formal psychometric or measurement-model validation procedures were performed, as the study was not designed for instrument validation. Accordingly, the statistical analyzes focused on examining associations and mediation effects with clinical and anthropometric outcomes, rather than on evaluating the measurement properties of the indices.

A total of 121 children with abdominal obesity (11.25 years old; 62% girls) were randomized into two groups (usual care or intervention) for the IGENOI lifestyle intervention. Of these, 114 participants completed the 8-week intensive intervention. This sub-study consisted of 107 children (11.3 years old; 63% girls) with abdominal obesity (Supplementary Figure S1).

Baseline anthropometric, clinical and lifestyle parameters were not differ statistically between the intervention (n = 81) and usual care (n = 26) groups (Table 1). Both groups exhibited similar health behaviors, health factor scores based on LE8, overall LE8 score, CVH status and UPFs consumption at baseline (Table 2), except for significantly lower glucose levels (−3.18 mg/dL, p = 0.034) and a higher SBP (5.71 mm HG, p = 0.047) in the intervention group compared to the usual care group.

Changes in scores measured by LE8 are summarized in Table 3. The intervention and usual care groups showed a significant increase in three (p < 0.001) and two points (p = 0.001) in the KIDMED score, respectively. Both groups showed significant decreases in BMI, HDL, non-HDL and glucose levels, with the mean score for blood lipids in the LE8 significantly improved (+11.88 points, p < 0.001) only in the intervention group. Additionally, the intervention group increased moderate to vigorous physical activity (+5.58 min/day, p = 0.028), and the mean score for BMI in the LE8 (10.25 points, p < 0.001) with a significant difference between groups (compared to the usual care, +7.53 points, p = 0.021). Furthermore, a significant decrease in BP was only observed in the intervention group, with a significant difference between groups of −7.39 mmHg for SBP (p = 0.032), −5.18 mmHg for diastolic BP (DBP, p = 0.029) and in 18.57 points for BP of LE8 (p = 0.002) compared to the usual care group.

In Figure 1, we show the effect of the 8-week intervention on the key endpoints. The impact of the lifestyle intervention on CVH, as measured by LE8, is shown in Figure 1A. A significant increase in +5.94 points of LE8 (p < 0.001) was observed in the intervention group, but not in the usual care group (−2.47 points, p = 0.248), with a significant difference between groups of +8.41 points (p < 0.001).

Consumption of UPF after the 8-week intervention in children with abdominal obesity is shown in Figure 1B. The intervention group showed a significant decrease about 2.74 portions of UPF per day (p < 0.001), while the usual care group decreased by 2.15 portions/day (p < 0.001). The difference between groups statistically significant (−0.59 portions per day, p = 0.032).

Changes in other anthropometric, clinical and lifestyle parameters not included in the LE8 are summarized in Table 1. Participants of both groups showed a significant decrease in anthropometric measurements (BMI-SDS, WC, HC, and fat mass percentage), lipids metabolism (TAG and total cholesterol) and total energy, along with an increase in DQI-A after the intervention. The intervention group also showed a significant decrease in LDL levels, insulin levels and HOMA index, while increased QUICKI index and HLD-I. Compared to the usual care group, participants in the intervention group showed a greater reduction in fat mass (−1.04%, p = 0.033) and a greater increase in HLD-I (2.65 points, p = 0.002) and DQI-A (5.33%, p = 0.020).

Multiple linear regressions were performed to assess the association between changes in LE8 and changes in anthropometric, clinical, and lifestyle measures (Table 4). We found a negative fitted association between changes in LE8 and changes in BMI-SDS (β −6.832, p = 0.039), HC (β −0.805, p = 0.010), total cholesterol (β −0.194, p = 0.004), LDL (β −0.211, p = 0.015), non-HDL (β −0.221, p = 0.015), insulin (β −0.370, p = 0.048), SBP (β −0.366, p < 0.001) and DBP (β −0.104, p = 0.026); and a positive association between changes in LE8 and changes in QUICKI index (β 145.393, p = 0.020), DQI-A (β 0.353, p = 0.006), HLD-I (β 1.573, p < 0.001) and KIDMED (β 2.859, p < 0.001).

Finally, a multivariable-adjusted model was used to evaluate the possible association between changes in UPF consumption and changes in LE8 (Figure 2). A significant negative association between changes in UPF consumption and changes in LE8 was found (p = 0.024) after adjusting for change in energy intake and somatic maturity. Notably, we observed that each decrease of one portion/day in UPF consumption was independently associated with an increase of 1.5 points in the LE8.

Regarding our mediation analysis, we found that the weight loss program has a significant positive total effect on LE8 score (β = 7.776; p < 0.001; path c). The program also significantly reduced UFP consumption (β = −0.933; p = 0.027; path a). When including both the weight loss program and the change in UPF consumption in the regression model, the effect of UPF consumption on LE8 was not statistically significant (β = −0.983; p = 0.076; path b). The weight loss program continued to have a significant direct effect on LE8 (β = 6.859; p = 0.004; path c’), although this effect was attenuated by the reduction in UPF consumption. The indirect effect through UPF reduction was significant (β = 0.917; 95% percentile bootstrap CI 0.011 to 2.319), indicating that approximately 11.8% of the total effect of the weight loss program on LE8 was mediated by changes in UPF consumption (Figure 3).