Dyslipidaemias are disorders characterised by the presence of high total cholesterol (TC), high LDL-cholesterol, low HDL-cholesterol and high triglycerides (TAG)(Reference Faludi, Izar and Saraiva1). The increasing prevalence of dyslipidaemia in childhood(Reference Furtado, Almeida and Mascarenhas2–Reference Faria Neto, Bento and Baena4) is considered a major public health concern worldwide, especially due to its vital role in the pathogenesis of atherosclerosis and the development of CVD(Reference Faludi, Izar and Saraiva1,Reference Libby, Ridker and Maseri5) through their effect on inflammatory process(Reference Libby, Ridker and Maseri5,Reference Ross6) .
In this context, the diet is an important moderator of systemic inflammation(Reference Calder, Ahluwalia and Brouns7) and has been identified as a modifiable risk factor for managing dyslipidaemia and preventing atherosclerotic disease(Reference Faludi, Izar and Saraiva1,Reference Bamba8,Reference Jellinger, Smith and Mehta9) . Furthermore, studies in adults have shown the association of a pro-inflammatory diet with unfavourable lipoprotein profile(Reference Khan, Kwon and Shivappa10,Reference Mazidi, Shivappa and Wirth11) and subclinical atherosclerosis(Reference Bondonno, Lewis and Blekkenhorst12).
A validated tool for assessing the inflammatory potential of diet in paediatric populations(Reference Khan, Wirth and Ortaglia13), the Children’s Dietary Inflammatory Index (C-DIITM), has been directly associated with cardiometabolic risk and inflammation in children and adolescents(Reference Khan, Wirth and Ortaglia13–Reference Suhett, Hermsdorff and Ribeiro15). Moreover, we recently demonstrated the association between C-DII with anti- and pro-inflammatory adipokines in our sample from the Schoolchildren Health Assessment Survey(Reference Suhett, Hermsdorff and Ribeiro15).
In a study with adolescents, it was identified a direct association between C-DII with dyslipidaemia in overweight individuals (12–18 years), but not with isolated markers of lipid profile, reinforcing the need for further studies(Reference Sethna, Alanko and Wirth16). The effects of diet-related inflammation on lipid profile and atherosclerosis risk remain poorly explored at young ages(Reference Sethna, Alanko and Wirth16–Reference Sen, Rifas-Shiman and Shivappa18), and to date, no studies with children have evaluated these associations using the new C-DII. Therefore, we aimed to investigate the relationship of the C-DII with atherogenic risk in Brazilian schoolchildren. We hypothesise that the C-DII scores are directly associated with atherogenic risk.
Methods
Participants and study design
This cross-sectional representative study was carried out with participants from the Schoolchildren Health Assessment Survey (Pesquisa de Avaliação da Saúde do Escolar, in portuguese). Data collection was conducted in 2015 at the Universidade Federal de Viçosa by trained nutritionists.
The sample size calculation and the sampling process have been described previously(Reference Filgueiras, Suhett and Silva19,Reference Milagres, Rocha and Filgueiras20) . In brief, we recruited randomly 378 schoolchildren between the ages of 8 and 9 years from a total 1464 children in the same age group enrolled in one of the urban primary schools of Viçosa, Minas Gerais, Brazil.
Children were not included in the study when they used medications regularly or had a clinically diagnosed health condition that could interfere with nutritional status, body composition, lipid profile, blood pressure and/or glucose metabolism; whose parents did not sign informed consent form; or parents and guardian that, after three attempts, could not be reached.
Food intake and children’s dietary inflammatory index computation
To assess children’s food intake, we performed 24-h dietary recalls, completed by both mother/guardian and the child, across three non-consecutive days, including one weekend day. To enhance data reliability, the interviewers were previously trained. Moreover, we also used household utensils and a photograph album with food and beverage serving sizes to help the participants estimate the sizes of portions consumed(Reference Zabotto, Vianna and Gil21).
The common household units for each food consumed were converted into g, mg or ml to evaluate daily energy (kcal) and nutrients intake. The food composition analysis was conducted in the software Dietpro® 5i, version 5.8 using the Brazilian Food Composition Table(Reference Milagres, Rocha and Filgueiras20) and the USDA Food Composition Database(22,23) .
We calculated the C-DII scores to determine the inflammatory potential of diet according to the recent method proposed for paediatric populations(Reference Khan, Wirth and Ortaglia13). This index characterises the child’s diet on a continuum scale from maximally anti- to pro-inflammatory. Briefly, the Pesquisa de Avaliação da Saúde do Escolar dietary data were related to a representative world database that provides a global estimate mean along with the standard deviation for each nutrient considered in the C-DII definition(Reference Khan, Wirth and Ortaglia13). First, we created a Z-score to express the individual exposure of the child relative to the ‘standard global mean’. Second, these Z-scores were converted to a centred proportion to avoid the effect of ‘right skewing’ (i.e. with values from 0 to 1), multiplied by 2 and then subtracting 1. As a result, we achieved a symmetrical distribution ranging from −1 (anti-inflammatory) to +1 (pro-inflammatory) and centred on 0 (null). Next, the scores were multiplied by the ‘overall food parameter-specific inflammatory effect score’. Finally, we summed each food parameter-specific C-DII scores to obtain the overall individual C-DII score(Reference Khan, Wirth and Ortaglia13). To minimise the energy intake difference among people, the energy-adjusted C-DII was used by converting all food to per 1000 kcal consumption.
For the C-DII calculation, we included sixteen of the twenty-five possible food parameters(Reference Khan, Wirth and Ortaglia13), as follows: carbohydrate, protein, total fat, SFA, MUFA, PUFA, fibre, Fe, Zn, Mg, vitamin C, vitamin B1, vitamin B2, vitamin B3, vitamin B6 and vitamin D. Higher C-DII scores represent a more pro-inflammatory diet.
Lipid profile and atherogenic risk
The children’s blood samples were collected at the Laboratory of Clinical Analysis of the Health Division of the Universidade Federal de Viçosa by venepuncture in the antecubital vein after 12 h of fasting. Bloods were packed in 1·5 ml microtubes and stored at −80°C until the lipid profile analysis was completed.
We evaluated serum TC, LDL-cholesterol, HDL-cholesterol, VLDL-cholesterol and TAG using an automated equipment (BioSystems 200 Mindray® model), according to the manufacturer recommendations (Bioclin® kits). Children with at least one of the following altered lipids were classified with dyslipidaemia: TC ≥ 170 mmol/l, HDL-cholesterol < 45 mmol/l, LDL-cholesterol ≥ 110 mmol/l and TAG ≥ 75 mmol/l(Reference Faludi, Izar and Saraiva1).
Moreover, we calculated atherogenic indexes(Reference Millán, Pintó and Muñoz24) such as the Castelli risk index I (TC/HDL-cholesterol), Castelli index risk II (LDL-cholesterol/HDL-cholesterol), lipoprotein combine index (LCI), atherogenic index of plasma and atherogenic coefficient (AC). LCI was calculated using the formula: (TC × TAG × LDL-cholesterol)/HDL-cholesterol. Atherogenic index of plasma is the base 10 log of the TAG/HDL-cholesterol. AC is the ratio of (TC − HDL-cholesterol)/HDL-cholesterol(Reference Millán, Pintó and Muñoz24–Reference Dobiášová and Frohlich26). Furthermore, we evaluated the accumulation of altered dyslipidaemia markers (high TC + high LDL-cholesterol + high TAG + low HDL-cholesterol), which contribute to the initiation and progression of endothelial dysfunction increasing the atherosclerosis risk(Reference Pires, Sena and Seiça27,Reference Kavey28) .
Sociodemographic characteristics and screen time
A semi-structured questionnaire was administer to parents and guardians to obtain sociodemographic characteristics and data on screen time. The collected data included: sex, age (years) and race (White, Brown and Black), household’s per capita income (US$) and screen time (h/d).
Anthropometry and body composition
The children’s anthropometric measurements were performed by a trained member of the research group. The children wore light clothes and were barefoot. Weight was measured using a digital electronic scale with a capacity of 150 kg and a sensitivity of 100 g (Tanita® Ironman Model BC 553; Tanita Corporation of America Inc.). The height was evaluated using a vertical stadiometer divided in cm and subdivided in mm (Alturexata®). We obtained height-for-age and BMI-for-age Z-scores using the WHO Anthro Plus software(Reference de Onis, Onyango and Borghi29).
We assessed the children’s body fat by dual-energy X-ray absorptiometry (Lunar Prodigy Advance; GE Medical Systems Lunar). The child was fasted during the exam wearing only light clothing and remained in supine position on the scanning bed until the end of the equipment evaluation.
Data analyses
Exposure
Children’s Dietary Inflammatory Index (C-DII) scores were expressed in quintiles.
Outcome
Lipid profile (TC, HDL-cholesterol, LDL-cholesterol, VLDL-cholesterol and TAG), atherogenic indexes (Castelli risk indexes I and II, LCI, atherogenic index of plasma and AC) and accumulation of altered dyslipidaemia markers (high TC + high LDL-cholesterol + high TAG + low HDL-cholesterol), were all fit as continuous variables.
Covariates
Child’s age, sex, race, screen time, the household’s per capita income and body fat (%).
Statistical analyses
The normality of the variables was assessed by the Kolmogorov–Smirnov test. Comparisons of mean and standard deviation between groups were verified by Student’s t test, and Pearson’s χ 2 test was used to evaluate the relationships between categorical variables. We evaluated the distribution of dietary intake using mean and standard deviation (sd) according to C-DII categories from multivariable linear regression. Then, we compared the distributions of outcomes (lipid profile, atherogenic indexes and accumulation of altered dyslipidaemia markers) by C-DII categories (exposure), using mean differences and 95 % CI from unadjusted and multivariable linear regression models adjusted by age, sex, race, screen time, per capita income and body fat (%). All the possible confounders included in the analysis were chosen after a literature review. The robust estimates of the variance were specified in all models, which are consistent with heteroscedasticity and non-normality(Reference White30).
Analyses were performed using the software Stata® version 14 (StataCorp LP) and Statistical Package for the Social Science® software version 21 (SPSS Inc.). The significance level was 0·05 for all hypothesis tests.
Results
In our sample, 52·12 % were girls and 68·52 % were Non-White children. The prevalence of dyslipidaemia was 70 %. The means ± sd of serum TC, HDL-cholesterol, LDL-cholesterol and TAG were 152·29 ± 26·40 mmol/l, 50·07 ± 9·99 mmol/l, 86·96 ± 23·38 mmol/l and 78·83 ± 35·41 mmol/l, respectively (data not shown). The children with dyslipidaemia had higher BMI-for-age (0·59 ± 1·45) and body fat (25·61 ± 10·44) compared with those with normal lipid profile (−0·01 ± 1·19 and 21·01 ± 8·40, respectively). The mean ± sd of C-DII score was 0·60 ± 0·94 and ranged from −2·16 (maximum anti-inflammatory diet) to 2·75 (maximum pro-inflammatory diet) (Table 1). Furthermore, children with hypercholesterolaemia and hypertriglyceridaemia reported consuming a more pro-inflammatory diet (Fig. 1).
Table 1 Characteristics of participants according to the presence or absence of dyslipidaemia, Viçosa, Minas Gerais, Brazil, 2015
C-DII, Children’s Dietary Inflammatory Index; sd, standard deviation.
* P < 0·05.
† Approximate exchange rates of real (R$) to dollar (US$) at the time of this study (US$1·00 = R$ 3·22).
‡ Pearson’s χ 2 test for categorical variables.
Student’s t test for continuous variables and
Fig. 1 Children’s Dietary Inflammatory Index (C-DII) score according to the presence or absence of dyslipidaemia in schoolchildren from Viçosa, Minas Gerais, Brazil, 2015. Values are mean (95% CI). Student’s t test (P < 0·05*) adjusted for child’s age, sex, race, per capita income, screen time and body fat
Table 2 shows the association of food intake with C-DII after adjustment for potentially confounding factors. The C-DII was inversely associated with energy intake and the consumption of carbohydrate, protein, PUFA, fruits, vegetables, whole grains, rice and beans, and dairy products. Conversely, C-DII was directly associated with total fat, MUFA, saturated fat and unhealthy food groups intake as crackers and chips, processed meat, fast-food and fried snacks, industrial juices and soft drinks, and sugar.
Table 2 Distribution of nutrients and food groups intake according to Children’s Dietary Inflammatory Index (C-DII) in schoolchildren from Viçosa, Minas Gerais, Brazil, 2015
EI, energy intake; MUFA, monounsaturated fat; PUFA, polyunsaturated fat; sd, standard deviation.
* P < 0·05.
† From linear regression models with a variable representing ordinal C-DII categories introduced as continuous. Robust estimates of variance were specified in all models. Adjusted for child’s age, sex, race, per capita income, screen time and body fat.
The C-DII was directly associated with atherogenic indexes, regardless of child’s age, sex, race, per capita income, screen time and body fat (%). Compared with first, the fifth quintiles of C-DII were related to 0·17 (95 % CI −0·01, 0·35), 5·34 (95 % CI 0·43, 10·25) and 0·16 (95 % CI −0·02, 0·34) units higher TC/HDL-cholesterol ratio, LCI and AC, respectively. Every 1 sd of C-DII was associated with a 0·07 (95 % CI 0·01, 0·13), 1·94 (95 % CI 0·20, 3·67) and 0·06 (95 % CI 0·002, 0·12) units higher TC/HDL-c ratio, LCI and AC, respectively (Table 3). Additionally, a pro-inflammatory diet was associated with the accumulation of altered dyslipidaemia markers (high TC + high LDL-cholesterol + high TAG + low HDL-cholesterol). Children with the most pro-inflammatory diet (fifth quintile) had 0·32 (95 % CI 0·01, 0·63) unit higher accumulation of altered dyslipidaemia markers than children with the most anti-inflammatory diet (first quintile) (Fig. 2). Every 1 sd of C-DII was associated with a 0·12 (95 % CI 0·02, 0·22) unit higher accumulation of altered dyslipidaemia markers.
Table 3 Crude and adjusted analyses of the association of the Children’s Dietary Inflammatory Index (C-DII) with lipid profile and atherogenic indexes in schoolchildren from Viçosa, Minas Gerais, Brazil, 2015
95% IC, 95% confidence interval; sd, standard deviation; AC, atherogenic coefficient; AIP, atherogenic index of plasma; HDL-c, high-density lipoprotein cholesterol; LCI, lipoprotein combine index; LDL-c, low-density lipoprotein cholesterol; LDL-c/HDL-c, low-density lipoprotein cholesterol/high-density lipoprotein cholesterol ratio; TC, total cholesterol; TC/HDL-c, total cholesterol/high-density lipoprotein cholesterol ratio; TAG, triglycerides; VLDL-c, very-low-density lipoprotein cholesterol.
* P < 0·05.
† From linear regression models with each lipid marker as a continuous outcome and C-DII as predictor. Robust estimates of variance were specified in all models.
‡ From linear regression adjusted for child’s age, sex, race, per capita income, screen time and body fat.
§ From linear regression model with each lipid marker as continuous outcome and C-DII per 1 sd (continuous) as predictor.
‖ Test for linear trend when a variable representing ordinal C-DII categories were introduced as a continuous predictor in the linear regression.
Fig. 2 Association between the Children’s Dietary Inflammatory Index (C-DII) and the accumulation of altered dyslipidaemia markers (high TC + high TAG + high LDL-cholesterol + low HDL-cholesterol) in schoolchildren from Viçosa, Minas Gerais, Brazil, 2015. Q, quintiles; Q1: reference. †From linear regression models with the number of altered dyslipidaemia markers as continuous outcome and a variable representing ordinal C-DII categories introduced as continuous predictor. Robust estimates of variance were specified in all the models. Adjusted for child’s age, sex, race, per capita income, screen time and body fat. (P < 0·05*)
Discussion
Our cross-sectional study supports the hypothesis that a pro-inflammatory diet is associated with a higher atherogenic risk in schoolchildren.
The findings showed a direct association between C-DII and atherogenic indexes, but not individual parameters. Previous studies with paediatric population(Reference Sethna, Alanko and Wirth16,Reference Sen, Rifas-Shiman and Shivappa18) also found no significant associations between the pro-inflammatory diet and isolated traditional markers of lipid profile, corroborating with our results. However, the atherogenic indexes are good indicators of CVD risk because these consider the balance between atherogenic and protective lipoproteins, thus, increasing their predictive capacity compared with isolated traditional markers(Reference Millán, Pintó and Muñoz24,Reference Criqui and Golomb31) . In this context, the assessment of atherogenic indexes, also obtained at young ages, should be encouraged in paediatric clinical practice as interesting complementary methods of screening and monitoring cardiometabolic risk(Reference Castro, Hermsdorff and Milagres32,Reference Sapunar, Aguilar-Farías and Navarro33) .
Current studies with adults have shown the association between a pro-inflammatory diet and an increased risk of dyslipidaemia(Reference Khan, Kwon and Shivappa10), elevated TAG/HDL-cholesterol ratio and apoB(Reference Mazidi, Shivappa and Wirth11), as well as the development of CVD and increased mortality(Reference Bondonno, Lewis and Blekkenhorst12,Reference Puddu, Shivappa and Menotti34–Reference Shivappa, Steck and Hussey36) . Though results do exist in adults, the investigations regarding the impacts of dietary inflammatory potential on children’s lipid profiles are limited(Reference Rahbarinejad, Asghari and Yuzbashian17,Reference Sen, Rifas-Shiman and Shivappa18) ; to date, no studies have been found evaluating the relationship between C-DII and atherogenic indexes. Taken together, more research conducted in children is needed to clarify this issue.
We also observed a positive association of the C-DII with the accumulation of lipid abnormalities characterising a dyslipidaemia pattern. This result suggests a potential role of diet in the early occurrence of CVD through inflammatory modulation mechanisms. The adherence to an anti-inflammatory diet with a higher intake of nutrients and antioxidant compounds sources as fruits and vegetables may improve lipid profile(Reference Giannini, Diesse and D’adamo37,Reference Cadario, Prodam and Pasqualicchio38) and reduce inflammation(Reference Sureda, Bibiloni and Julibert39,Reference Arouca, Michels and Moreno40) and atherosclerosis risk(Reference Shivappa, Steck and Hussey36). In contrast, a poor-quality diet that includes excessive intake of pro-inflammatory food parameters, such as sugar and dietary fats (e.g. ultraprocessed foods), increases lipogenesis(Reference Rupérez, Mesana and Moreno41,Reference Siri-Tarino and Krauss42) , activates the NF-kB pathway(Reference Calder, Albers and Antoine43,Reference Galland44) and may subsequently alter lipid profile in children(Reference Leffa, Hoffman and Rauber45,Reference Rauber, Campagnolo and Hoffman46) . Moreover, emerging evidence demonstrate that the pro-inflammatory diet is directly associated with the secretion of pro-inflammatory cytokines and expression of adhesion molecules(Reference Khan, Wirth and Ortaglia13,Reference Almeida-de-Souza, Santos and Barros47,Reference Shivappa, Hebert and Marcos48) which contributes to vascular tissue inflammation, endothelial dysfunction(Reference Libby, Ridker and Maseri5,Reference Ross6) and carotid intima-media thickening(Reference Sethna, Alanko and Wirth16). Atherosclerosis is a chronic inflammatory disease(Reference Libby, Ridker and Maseri5,Reference Ross6) that often originates in childhood(Reference Holman, McGill and Strong49) and is related to future cardiovascular events(Reference Willeit, Tschiderer and Allara50). Therefore, a diet rich in anti-inflammatory properties might benefit children’s health and growth, preventing the early onset of the atherosclerotic process and subsequent heart diseases in adulthood.
Some strengths of this study should be pointed out. First, this is a representative study with the sample composed exclusively of children, an important group in which to investigate food consumption because eating habits are created in childhood. Second, we evaluated the association of C-DII and atherogenic risk using traditional lipoproteins as well as non-traditional atherogenic indexes. Furthermore, body composition was analysed with dual-energy X-ray absorptiometry, a gold-standard method for assessing body composition, and the models were adjusted by body fat (%). We highlight that, according to our knowledge, this is the first epidemiological study to evaluate the relationship of C-DII with atherogenic risk in children. Our study provides novel information to public health practice and a better understanding of the role of the inflammatory dietary factors in the early occurrence of dyslipidaemia and underlying atherosclerosis.
This study has some limitations. First, the application of 24-h dietary recalls may present memory bias; however, the interviewers were previously trained, and the 24-h dietary recalls were administered on three non-consecutive days, using food a photo album and standard measurement tools; and all are recognisably methodological strategies to guarantee better accuracy. Second, we calculated the C-DII score with sixteen of the total twenty-five food parameters included in the C-DII definition. Nevertheless, our C-DII scores were similar to previous paediatric studies(Reference Sen, Rifas-Shiman and Shivappa18,Reference Navarro, Shivappa and Hébert51) . Finally, the cross-sectional design makes it impossible to establish causal relationships. Thus, further longitudinal studies with paediatric population are required to confirm our findings and determine the long-term effects of dietary-related inflammation on lipid profile and cardiovascular risk.
We conclude that pro-inflammatory diet, as indicated by higher C-DII scores, is associated with higher atherogenic risk in Brazilian schoolchildren. These results reinforce the importance of effective nutritional policies to promote healthy eating habits, improve the children’s lipid profile and prevent later CVD.
Acknowledgements
Acknowledgements: We thank BioClin (Belo Horizonte, MG, Brazil) for the lipid profile kits granted, the Pesquisa de Avaliação da Saúde do Escolar participants and parents. Financial support: Our study was supported by Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG – CDS-APQ-02979-16), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq – 478910/2013-4 and 407547/2012-6) and the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES – code 001). H.H.M.H. has a Research Fellowship from CNPq (1D-level). The CNPq, FAPEMIG and CAPES had no role in the design, analysis or writing of this article. Conflict of interest: The authors declare no conflicts of interest. However, Dr James R. Hébert discloses that he owns controlling interest in Connecting Health Innovations LLC (CHI), a company that has licensed the right to his invention of the dietary inflammatory index (DII®) from the University of South Carolina in order to develop computer and smart phone applications for patient counseling and dietary intervention in clinical settings. Dr Nitin Shivappa is an employee of CHI. Authorship: L.G.S.: Conceptualisation; data analysis; investigation; writing – original draft; writing – review & editing. S.A.V.R.: Writing – review & editing, final approval of the version to be published. H.H.M.H.: Writing – review & editing, final approval of the version to be published. M.A.S.: Writing – review & editing, final approval of the version to be published. N.S.: Writing – review & editing, final approval of the version to be published. J.R.H.: Writing – review & editing, final approval of the version to be published. J.F.de.N.: Project administration, writing – review & editing, final approval of the version to be published. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving research participants were approved by the Human Research Ethics Committee of the Universidade Federal de Viçosa (reference number 663.171/2014). Written informed consent was obtained from all parents.
