ABSTRACT:

Objective:

To analyze the temporal trend of simultaneous occurrence of behavioral risk factors for chronic non-communicable diseases in Brazilian school adolescents.

Methods:

Ecological time-series study that analyzed data from the three editions of the National Survey of School Health with students in the 9th year of public and private schools. Cluster analysis was performed to identify the simultaneity of the following factors: irregular consumption of fresh or minimally processed foods, regular consumption of processed and ultra-processed foods, insufficient level of physical activity during leisure time, consumption of alcoholic beverages, use of cigarettes and illicit drugs. The cluster trend was tested using simple linear regression.

Results:

The sample was composed of 173,310 school adolescents. Sixty-four possible combinations were evaluated, resulting in the formation of six clusters. The grouping formed by all factors showed the highest ratios observed/expected prevalence in the last two years of the series. The highest ratios were observed for combinations in which alcohol and cigarettes were present. These substances integrated five of the six clusters. One of the groups (insufficient level of physical activity during leisure time + consumption of alcoholic beverages + use of cigarettes and illicit drugs) had an upward trend throughout the series.

Conclusion:

The trend analysis showed four out of six clusters remaining stable throughout the series, which shows that the factors are persistently present among school adolescents.

Keywords:
Noncommunicable diseases; Teenagers; Risk factors; Cluster analysis

INTRODUCTION

Adolescence is a growth phase marked by intense physical, psychological, and emotional changes¹1 World Health Organization. Launch: a lancet commission on adolescent health and wellbeing. Geneva: World Health Organization; 2016.. In this stage, certain acquired behaviors such as poor diet, sedentary lifestyle, alcohol consumption, tobacco use and illicit drug use constitute risk factors (RF) for the development of chronic non-communicable diseases (NCDs)²2 Freitas EAO, Martins MSAS, Espinosa MM. Experimentação do álcool e tabaco entre adolescentes da região Centro-Oeste/Brasil. Ciênc Saúde Colet 2019; 24(4): 1347-57. https://doi.org/10.1590/1413-81232018244.15582017
https://doi.org/10.1590/1413-81232018244... –⁴4 Rocha FL, Velasquez-Melendez G. Simultaneidade e agregamento de fatores de risco para doenças crônicas não transmissíveis em adolescentes brasileiros. Esc Anna Nery 2019; 23(3): e20180320. https://doi.org/10.1590/2177-9465-EAN-2018-0320
https://doi.org/10.1590/2177-9465-EAN-20... .

According to the World Health Organization (WHO), most deaths from NCDs are caused by a set of RFs that usually occur simultaneously⁵5 World Health Organization. Global status report in noncommunicable diseases 2014. Geneva: World Health Organization; 2014.. Studies show that multiple behaviors point to a higher risk of developing NCDs and mortality when compared to engaging in only one or no risky behavior⁶6 Meader N, King K, Wright K, Graham HM, Petticrew M, Power C, et al. Multiple risk behavior interventions: meta-analyses of RCTs. Am J Prev Med 2017; 53(1): e19-e30. https://doi.org/10.1016/j.amepre.2017.01.032
https://doi.org/10.1016/j.amepre.2017.01... .

In order to avoid this, the WHO proposed monitoring the health of adolescents through the creation of RF surveillance and health protection systems in school environment⁷7 World Health Organization. Political declaration of the high-level meeting of the general assembly on the prevention and control of non-communicable diseases. Genebra: World Health Organization; 2011.. From this perspective, the Surveillance System for Risk Factors for Chronic Non-Communicable Diseases was started in Brazil, under the responsibility of the Brazilian Institute of Geography and Statistics (IBGE) in partnership with the Ministries of Health and of Education. The system is based on data from the National Survey of School Health (PeNSE), which has been carried out since 2009 to obtain information to guide the knowledge and dimensioning of RFs and adolescent health protection⁸8 Instituto Brasileiro de Geografia e Estatística. Pesquisa nacional de saúde do escolar 2009. Rio de Janeiro: IBGE; 2009 [cited on Mar 9, 2021]. Disponível em: https://biblioteca.ibge.gov.br/visualizacao/livros/liv43063.pdf
https://biblioteca.ibge.gov.br/visualiza... –¹⁰10 Instituto Brasileiro de Geografia e Estatística. Pesquisa nacional de saúde do escolar: 2015. Rio de Janeiro: IBGE; 2016 [cited on Mar 9, 2021]. Disponível em: https://biblioteca.ibge.gov.br/visualizacao/livros/liv97870.pdf
https://biblioteca.ibge.gov.br/visualiza... .

This type of strategy focused on the development of healthy behaviors and the opportunity to prevent diseases and injuries at an early age is an indispensable method for health promotion, even though it is not being fully explored¹1 World Health Organization. Launch: a lancet commission on adolescent health and wellbeing. Geneva: World Health Organization; 2016.,¹¹11 Oliveira-Campos M, Oliveira MM, Silva SU, Santos MAS, Barufaldi LA, Oliveira PPV, et al. Fatores de risco e proteção para as doenças crônicas não-transmissíveis em adolescentes nas capitais brasileiras. Rev Bras Epidemiol 2018; 21 (suppl 1): E180002. https://doi.org/10.1590/1980-549720180002.supl.1
https://doi.org/10.1590/1980-54972018000... .

Therefore, this study aimed to analyze food consumption, physical activity, consumption of alcoholic beverages, use of cigarettes and illicit drugs among students of the 9th grade of elementary school (former 8th grade) in public and private schools of Brazilian capitals and the Federal District (DF), based on data collected from PeNSE, to answer the following question: is there a temporal trend of RF clusters for CNCDs among Brazilian school adolescents?

In Brazil, most epidemiological studies that evaluated RFs among adolescents sought to analyze their occurrence in isolation¹¹11 Oliveira-Campos M, Oliveira MM, Silva SU, Santos MAS, Barufaldi LA, Oliveira PPV, et al. Fatores de risco e proteção para as doenças crônicas não-transmissíveis em adolescentes nas capitais brasileiras. Rev Bras Epidemiol 2018; 21 (suppl 1): E180002. https://doi.org/10.1590/1980-549720180002.supl.1
https://doi.org/10.1590/1980-54972018000... –¹⁴14 Malta DC, Stopa SR, Santos MAS, Andrade SSCA, Oliveira MM, Prado RR, et al. Fatores de risco e proteção de doenças e agravos não transmissíveis em adolescentes segundo raça/cor: Pesquisa Nacional de Saúde do Escolar. Rev Bras Epidemiol 2017; 20(2): 247-59. https://doi.org/10.1590/1980-5497201700020006
https://doi.org/10.1590/1980-54972017000... . Although it is relevant to identify and implement actions aimed at the prevalence of RF in isolation, it is also important to consider their simultaneity, since behavior is multi-determined, and an individual assessment of factors may not reflect the true risk of individuals¹⁵15 Hu FB. Dietary pattern analysis: a new direction in nutritional epidemiology. Curr Opin Lipidol 2002; 13(1): 3-9. https://doi.org/10.1097/00041433-200202000-00002
https://doi.org/10.1097/00041433-2002020... . To date, few studies have analyzed simultaneous RFs in adolescents, and some have focused on evaluating data from only one edition of PeNSE⁴4 Rocha FL, Velasquez-Melendez G. Simultaneidade e agregamento de fatores de risco para doenças crônicas não transmissíveis em adolescentes brasileiros. Esc Anna Nery 2019; 23(3): e20180320. https://doi.org/10.1590/2177-9465-EAN-2018-0320
https://doi.org/10.1590/2177-9465-EAN-20... ,¹⁵15 Hu FB. Dietary pattern analysis: a new direction in nutritional epidemiology. Curr Opin Lipidol 2002; 13(1): 3-9. https://doi.org/10.1097/00041433-200202000-00002
https://doi.org/10.1097/00041433-2002020... ,¹⁶16 Ricardo CZ, Azeredo CM, Rezende LFM, Levy RB. Co-occurrence and clustering of the four major non-communicable disease risk factors in Brazilian adolescents: analysis of a national school-based survey. PLoS One 2019; 14(7): e0219370. https://doi.org/10.1371/journal.pone.0219370
https://doi.org/10.1371/journal.pone.021... . Thus, this work fills a gap when it assesses the simultaneous occurrence of main behavioral RFs for CNCDs among adolescents, using data from the three editions of PeNSE.

Thus, this paper’s purpose was to analyze the temporal trend of simultaneous behavioral RFs for CNCDs in Brazilian school adolescents with data from PeNSE 2009, 2012 and 2015.

METHODS

This is a time series ecological study which verified the main RFs for the development of CNCDs in Brazilian school adolescents, using secondary data from the 2009, 2012 and 2015 editions of PeNSE, available on the IBGE website (http://www.ibge.gov.br).

Only data from 9th grade students that contained information common to the three editions were included, the adolescents being selected from public and private schools of the 26 Brazilian capitals and the Federal District. This allowed a temporal comparison between editions. In this way, the responses of 173,310 school adolescents were analyzed, 60,973 from PeNSE 2009, 61,145 from PeNSE 2012 and 51,192 from PeNSE 2015.

Irregular consumption of in natura or minimally processed foods, regular consumption of processed and ultra-processed foods, insufficient level of leisure-time physical activity, consumption of alcoholic beverages, cigarette use and illicit drug use were considered RFs.

For the analysis of food consumption, the Food Guide for the Brazilian Population¹⁷17 Brasil. Ministério da Saúde. Secretaria de Atenção à Saúde. Departamento de Atenção Básica. Guia alimentar para a população brasileira. Brasília: Ministério da Saúde; 2014. Disponível em: https://bvsms.saude.gov.br/bvs/publicacoes/guia_alimentar_populacao_brasileira_2ed.pdf
https://bvsms.saude.gov.br/bvs/publicaco... , which classifies foods as in natura or minimally processed, processed and ultra-processed, was taken into account. Thus, two major food groups were considered:

1. In natura or minimally processed food;

2. Processed and ultra-processed food.

To identify RFs, we chose the concepts of “regular consumption” (≥5 times in the previous week), which was validated through a 24-hour recall among adolescents¹⁸18 Tavares LF, Castro IRR, Levy RB, Cardoso LO, Passos MD, Brito FSB. Validade relativa de indicadores de práticas alimentares da Pesquisa Nacional de Saúde do Escolar entre adolescentes do Rio de Janeiro, Brasil. Cad Saúde Pública 2014; 30(5): 1029-41. https://doi.org/10.1590/0102-311X00000413
https://doi.org/10.1590/0102-311X0000041... , and “irregular consumption” (<5 times in the previous week)¹⁹19 Azeredo CM, Rezende LFM, Mallinson PAC, Ricardo CZ, Kinra S, Levy RB, et al. Progress and setbacks in socioeconomic inequalities in adolescent health-related behaviours in Brazil: results from three cross-sectional surveys 2009-2015. BMJ Open 2019; 9(3): e025338. https://doi.org/10.1136/bmjopen-2018-025338
https://doi.org/10.1136/bmjopen-2018-025... . Thus, consumption of in natura or minimally processed foods was considered irregular when it had happened less than five times in the seven days prior to the survey; and consumption of processed and ultra-processed foods was considered regular when consumed five or more times in the seven days prior to the survey.

Regarding physical activity, activities performed during leisure time were considered, not counting physical education classes (sports, dance, gymnastics, weight training). Adolescents were asked about the frequency and duration of these activities in the seven days prior to the survey. Based on this, the variable “insufficient level of physical activity during leisure time” was generated, considering adolescents who did not practice at least 300 minutes of physical activity per week²⁰20 World Health Organization. Child and adolescent health and development: progress report 2009: highlights. Geneva: World Health Organization; 2010..

According to the WHO²¹21 World Health Organization. Global status report on alcohol and health 2014. Genebra: World Health Organization; 2014., adolescents are more vulnerable to damages caused by the consumption of alcohol, tobacco and illicit drugs, and therefore these substances should not be marketed to this group. Thus, any level of consumption of these substances in the last 30 days prior to the survey was considered as RF¹⁶16 Ricardo CZ, Azeredo CM, Rezende LFM, Levy RB. Co-occurrence and clustering of the four major non-communicable disease risk factors in Brazilian adolescents: analysis of a national school-based survey. PLoS One 2019; 14(7): e0219370. https://doi.org/10.1371/journal.pone.0219370
https://doi.org/10.1371/journal.pone.021... .

For the sociodemographic analysis, the following variables were considered: sex (male; female); skin color or race (white; black; yellow; brown; indigenous); age (13 years or younger; 14 to 15 years; 16 years or older); region of residence (North; Northeast; South; Southeast; Midwest); and school network (public; private).

The data used for the development of this study were downloaded directly from the IBGE website by the researchers, in June and July 2021; then, they were organized in Microsoft Excel 2010 spreadsheets for further statistical analysis.

The distribution of sociodemographic characteristics and RFs for CNCDs was presented as percentages and respective 95% confidence intervals (95%CI) for each year of the series. The RFs were coded as dichotomous variables (1=presence and 0=absence). The prevalence of multiple RFs was calculated by the sum of indicators presented.

The presence of clusters was studied by comparing observed (OP) and expected (EP) prevalence. The EP for each combination was calculated by multiplying the probabilities of each RF, based on its distribution in the studied population. Clustering was defined when OP/EP ratio was greater than 1.20²²22 Singh A, Rouxel P, Watt RG, Tsakos G. Social inequalities in clustering of oral health related behaviors in a national sample of British adults. Prev Med 2013; 57(2): 102-6. https://doi.org/10.1016/j.ypmed.2013.04.018
https://doi.org/10.1016/j.ypmed.2013.04.... . Confidence intervals (CI) for the OP/EP ratios were obtained using the Newton method and assuming the Poisson distribution²³23 StataCorp. Stata 14 Base Reference Manual. Confidence intervals for means, proportions, and variances. Texas: Stata Press Publication; 2015.; combinations in which the 95%CI did not contain a null value were considered clusters. The RFs were assumed to have occurred independently.

After identifying the RF clusters in each year of the survey, the temporal trend was tested across the series. As in a previous study¹³13 Malta DC, Andreazzi MAR, Oliveira-Campos M, Andrade SSCA, Sá NNB, Moura L, et al. Tendência dos fatores de risco e proteção de doenças crônicas não transmissíveis em adolescentes, Pesquisa Nacional de Saúde do Escolar (PeNSE 2009 e 2012). Rev Bras Epidemiol 2014; 17(1):77-91. https://doi.org/10.1590/1809-4503201400050007
https://doi.org/10.1590/1809-45032014000... , a simple linear regression model was used, whose dependent variable (indicator) was the value of the OP/EP ratio of each combination tested, and whose independent variable was time (year of research).

All analyses were performed in the STATA program, version 14.1 (Stata Corporation, College Station, Texas)²³23 StataCorp. Stata 14 Base Reference Manual. Confidence intervals for means, proportions, and variances. Texas: Stata Press Publication; 2015. and took into account the weighting for the complex sample structure as established by PeNSE, so that it could represent the Brazilian population of adolescents residing in the 26 capitals and in the Federal District.

The three editions of PeNSE were approved by the National Research Ethics Committees (CONEP), opinions 11,537/2009, 16,805/2012 and 1,006,467/2015. For this study, the recommendation of Resolution 510/2016 by the National Health Council on information in the public domain was followed”²⁴24 Brasil. Ministério da Saúde. Conselho Nacional de Saúde. Resolução n° 510, de 7 de abril de 2016. O Plenário do Conselho Nacional de Saúde em sua Quinquagésima Nona Reunião Extraordinária, realizada nos dias 06 e 07 de abril de 2016, no uso de suas competências regimentais e atribuições conferidas pela Lei n o 8.080, de 19 de setembro de 1990, pela Lei n o 8.142, de 28 de dezembro de 1990, pelo Decreto n o 5.839, de 11 de julho de 2006. [Internet]. [cited on Dec 11, 2020]. Disponível em: https://www.in.gov.br/materia//asset_publisher/Kujrw0TZC2Mb/content/id/22917581
https://www.in.gov.br/materia//asset_pub... .

RESULTS

The sample consisted of 173,310 school adolescents residing in the 26 Brazilian capitals and in the Federal District in 2009, 2012 and 2015. Table 1 describes their sociodemographic characteristics and the RFs for CNCD.

In the entire series, the highest prevalence was observed for adolescents aged between 14 and 15 years old, living with both parents, studying in public schools and residing in the Southeast Region. As for gender, the highest percentages in the three years of the series seem to be among female adolescents, which shows significant differences only in the first two years. With regard to skin color, the percentage of school adolescents who reported being brown had a small increase throughout the whole series, being significantly higher in 2015 (Table 1).

In the three editions of the series, insufficient level of physical activity during leisure time was the most frequent RF among students (present in more than 84% of them), followed by the regular consumption of processed and ultra-processed foods. Data suggest, during the years of research, a decrease in the regular consumption of processed and ultra-processed foods and consumption of alcoholic beverages. On the other hand, the irregular consumption of fresh or minimally processed foods and the use of illicit drugs seem to have increased (Table 1).

Table 2 describes the OP and EP of the 64 possible combinations evaluated, among which six had OP/EP above 1.20, which corresponded to the grouping of RF. Table 1 (supplementary material) describes the factors that make up the six risk behaviors for NCDs in Brazilian school adolescents.

Cluster 1 resulted in OP/EP of 26.00 in 2009, 38.00 in 2012 and 45.00 in 2015, but remained stable for the period (p>0.05). Considering that the higher the OP/EP ratio, the greater the number of times the clustering can be observed, if these factors are independent, cluster 1 will still be the one with the strongest relationships and/or with the greatest possibility of occurring in the last two years of the series (Table 2).

The highest OP/EP ratios were found for combinations in which alcohol and cigarettes were present; these substances were part of five of the six identified clusters, indicating an association between factors (Table 2).

The trend analysis showed that clusters 1, 2, 3 and 6 remained stable throughout the series, while cluster 4 decreased by 0.04% per year (p<0.017) and cluster 5 increased by 1.5% (p<0.041) (Table 2).

As for the clusters with the highest OP in the entire series, 6 stood out, in which the irregular consumption of fresh or minimally processed foods, along with insufficient level of physical activity during leisure time, had a prevalence corresponding to 5% in 2009, 6.78% in 2012 and 6.7% in 2015. It is also worth mentioning that the OPs for the occurrence of no RF were 2.55% in 2009, 2.96% in 2012 and 3.74% in 2015 (Table 2).

As for the combinations, the one formed by the regular consumption of processed and ultra-processed foods and insufficient level of physical activity during leisure time had the highest OP of all three editions of the survey (34.4% in 2009; 29.67% in 2012; and 28.95% in 2015). Despite the high frequency of these combinations, the OP/EP ratio was close to 1, with no cluster formation (Table 2).

DISCUSSION

Studies have addressed the issue of RF for NCDs in adolescents in order to monitor the situation of risk and vulnerability⁴4 Rocha FL, Velasquez-Melendez G. Simultaneidade e agregamento de fatores de risco para doenças crônicas não transmissíveis em adolescentes brasileiros. Esc Anna Nery 2019; 23(3): e20180320. https://doi.org/10.1590/2177-9465-EAN-2018-0320
https://doi.org/10.1590/2177-9465-EAN-20... ,¹¹11 Oliveira-Campos M, Oliveira MM, Silva SU, Santos MAS, Barufaldi LA, Oliveira PPV, et al. Fatores de risco e proteção para as doenças crônicas não-transmissíveis em adolescentes nas capitais brasileiras. Rev Bras Epidemiol 2018; 21 (suppl 1): E180002. https://doi.org/10.1590/1980-549720180002.supl.1
https://doi.org/10.1590/1980-54972018000... –¹⁶16 Ricardo CZ, Azeredo CM, Rezende LFM, Levy RB. Co-occurrence and clustering of the four major non-communicable disease risk factors in Brazilian adolescents: analysis of a national school-based survey. PLoS One 2019; 14(7): e0219370. https://doi.org/10.1371/journal.pone.0219370
https://doi.org/10.1371/journal.pone.021... ,²⁵25 Caleyachetty R, Echouffo-Tcheugui JB, Tait CA, Schilsky S, Forrester T, Kengne AP. 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(7): 535-44. https://doi.org/10.1016/S2213-8587(15)00076-5
https://doi.org/10.1016/S2213-8587(15)00... . Thus, this study emphasized the identification of frequency, trend and prevalence—isolated and simultaneous—, of behavioral RFs and sociodemographic characteristics for the development of CNCDs in Brazilian school adolescents, using data from PeNSE.

In the three years of the series, adolescents aged between 14 and 15 years were the majority of the sample and the highest percentages were found among females. A study carried out with data from the Global School-based Student Health Survey (GSHS) between 2007 and 2016 analyzed 304,779 adolescents aged 11 to 17 years from 89 countries, also reporting a higher prevalence of female adolescents (52.2%)²⁶26 Uddin R, Lee EY, Khan SR, Tremblay MS, Khan A. Clustering of lifestyle risk factors for non-communicable diseases in 304,779 adolescents from 89 countries: a global perspective. Prev Med 2020; 131: 105955. https://doi.org/10.1016/j.ypmed.2019.105955
https://doi.org/10.1016/j.ypmed.2019.105... . A study carried out in Liberia based on GSHS data also observed that more than half of those evaluated were women (52.5%) aged 15 years (85.3%)²⁷27 Atorkey P, Asante KO. Clustering of multiple health risk factors among a sample of adolescents in Liberia: a latent class analysis. J Public Health (Berl.) 2021. https://doi.org/10.1007/s10389-020-01465-y
https://doi.org/10.1007/s10389-020-01465... .

The isolated analysis of RFs showed that insufficient level of physical activity during leisure time and regular consumption of processed and ultra-processed foods were the most common ones in all three editions. Although our study had a possibility of drop in regular consumption of processed and ultra-processed foods, the prevalence of this indicator across the years was still high and worrying, especially considering that irregular consumption of in natura or minimally processed foods seems to be increasing.

Our results show that factors such as insufficient level of physical activity and inadequate diet are the main conditions associated with CNCDs in adolescents¹⁶16 Ricardo CZ, Azeredo CM, Rezende LFM, Levy RB. Co-occurrence and clustering of the four major non-communicable disease risk factors in Brazilian adolescents: analysis of a national school-based survey. PLoS One 2019; 14(7): e0219370. https://doi.org/10.1371/journal.pone.0219370
https://doi.org/10.1371/journal.pone.021... ,²⁵25 Caleyachetty R, Echouffo-Tcheugui JB, Tait CA, Schilsky S, Forrester T, Kengne AP. 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(7): 535-44. https://doi.org/10.1016/S2213-8587(15)00076-5
https://doi.org/10.1016/S2213-8587(15)00... , and this is no different when evaluating studies carried out in other regions of the world. A survey conducted with data from the National Health and Nutrition Examination Survey (NHANES) from 2011 to 2016, which evaluated 7,714 school children and adolescents in the United States, showed that insufficient physical activity levels (53.4%) and consumption of food without nutritional quality (60.0%) were among the most prevalent factors in all editions of the survey²⁸28 Mayne SL, Virudachalam S, Fiks AG. Clustering of unhealthy behaviors in a nationally representative sample of U.S. children and adolescents. Prev Med 2020; 130: 105892. https://doi.org/10.1016/j.ypmed.2019.105892
https://doi.org/10.1016/j.ypmed.2019.105... .

Also, a study carried out with data from GSHS in 2015 and 2016 in Nepal, with 5,795 children aged between 13 and 17 years, also obtained the highest prevalence for insufficient intake of fruits and vegetables (95.33%), followed by insufficient physical activity (84.77%)²⁹29 Dhungana RR, Bista B, Pandey AR, Courten M. Prevalence, clustering and sociodemographic distributions of non-communicable disease risk factors in Nepalese adolescents: secondary analysis of a nationwide school survey. BMJ Open 2019; 9(5): e028263. https://doi.org/10.1136/bmjopen-2018-028263
https://doi.org/10.1136/bmjopen-2018-028... . Another survey based on cross-sectional data from the National Longitudinal Survey of Children and Youth (NLSCY) found that physical inactivity was the most prevalent RF among adolescents (62%)³⁰30 Alamian A, Paradis G. Clustering of chronic disease behavioral risk factors in Canadian children and adolescents. Prev Med 2009; 48(5): 493-9. https://doi.org/10.1016/j.ypmed.2009.02.015
https://doi.org/10.1016/j.ypmed.2009.02.... .

Regarding the cluster analysis, of the 64 possible combinations evaluated in this study, six corresponded to the grouping of RFs. The comparison with other studies⁴4 Rocha FL, Velasquez-Melendez G. Simultaneidade e agregamento de fatores de risco para doenças crônicas não transmissíveis em adolescentes brasileiros. Esc Anna Nery 2019; 23(3): e20180320. https://doi.org/10.1590/2177-9465-EAN-2018-0320
https://doi.org/10.1590/2177-9465-EAN-20... ,¹⁶16 Ricardo CZ, Azeredo CM, Rezende LFM, Levy RB. Co-occurrence and clustering of the four major non-communicable disease risk factors in Brazilian adolescents: analysis of a national school-based survey. PLoS One 2019; 14(7): e0219370. https://doi.org/10.1371/journal.pone.0219370
https://doi.org/10.1371/journal.pone.021... ,²⁶26 Uddin R, Lee EY, Khan SR, Tremblay MS, Khan A. Clustering of lifestyle risk factors for non-communicable diseases in 304,779 adolescents from 89 countries: a global perspective. Prev Med 2020; 131: 105955. https://doi.org/10.1016/j.ypmed.2019.105955
https://doi.org/10.1016/j.ypmed.2019.105... ,³⁰30 Alamian A, Paradis G. Clustering of chronic disease behavioral risk factors in Canadian children and adolescents. Prev Med 2009; 48(5): 493-9. https://doi.org/10.1016/j.ypmed.2009.02.015
https://doi.org/10.1016/j.ypmed.2009.02.... –³²32 Poortinga W. The prevalence and clustering of four major lifestyle risk factors in an English adult population. Prev Med 2007; 44(2): 124-8. https://doi.org/10.1016/j.ypmed.2006.10.006
https://doi.org/10.1016/j.ypmed.2006.10.... is still restricted by the fact that our study used a larger number of factors and/or different RFs to identify the clusters. In addition, different methods were adopted in previous studies for the simultaneous assessment and classification of factors.

For example, a cross-sectional investigation carried out in Sousse, Tunisia, with adolescents aged 11 to 16 years, performed cluster analysis with the variables tobacco use, physical inactivity, unhealthy diet and overweight and formed four clusters³³33 El Ghardallou M, Maatoug J, Harrabi I, Fredj SB, Jihene S, Dendana E, et al. Socio-demographic association of non communicable diseases’ risk factors in a representative population of school children: a cross-sectional study in Sousse (Tunisia). Int J Adolesc Med Health 2016; 29(5): /j/ijamh.2017.29.issue-5/ijamh-2015-0109/ijamh-2015-0109.xml. https://doi.org/10.1515/ijamh-2015-0109
https://doi.org/10.1515/ijamh-2015-0109... . Another study that used NHANES data from 2011 to 2016 and evaluated excessive screen time, low-quality diet, low physical activity level, fast food consumption and smoking identified the formation of 20 unhealthy clusters²⁸28 Mayne SL, Virudachalam S, Fiks AG. Clustering of unhealthy behaviors in a nationally representative sample of U.S. children and adolescents. Prev Med 2020; 130: 105892. https://doi.org/10.1016/j.ypmed.2019.105892
https://doi.org/10.1016/j.ypmed.2019.105... . Although both studies used cluster analysis, it is not possible to make direct comparisons, given the different formations of clusters with different RFs.

When analyzing clusters, we found that the highest OP/EP ratio was found when the six RFs were present simultaneously, that is, the chance of them to occur in conjunction is greater than in isolation. A cross-sectional, multicenter study carried out with 73,624 Brazilian adolescents showed that the OP/EP ratio was always higher in participants who had several RFs³⁴34 Oliveira G, Silva TLN, Silva IB, Coutinho ESF, Bloch KV, Oliveira ERA. Agregação dos fatores de risco cardiovascular: álcool, fumo, excesso de peso e sono de curta duração em adolescentes do estudo ERICA. Cad Saúde Pública 2019; 35(12): e00223318. https://doi.org/10.1590/0102-311X00223318
https://doi.org/10.1590/0102-311X0022331... .

In this research, the group formed by two RFs (low consumption of fresh or minimally processed foods and insufficient level of leisure-time physical activity) presented the highest OP between groups. A similar study carried out in 89 countries also found that the most prevalent grouping was composed of low consumption of fruits and vegetables and physical inactivity²⁶26 Uddin R, Lee EY, Khan SR, Tremblay MS, Khan A. Clustering of lifestyle risk factors for non-communicable diseases in 304,779 adolescents from 89 countries: a global perspective. Prev Med 2020; 131: 105955. https://doi.org/10.1016/j.ypmed.2019.105955
https://doi.org/10.1016/j.ypmed.2019.105... .

The combination consisting of regular consumption of processed and ultra-processed foods and insufficient level of physical activity during leisure time had the highest OP in all three editions; however, despite the high frequency, these factors did not group together.

A global disease investigation analyzed the effects of individual dietary factors on NCD mortality in 195 countries, quantifying the overall impact of poor dietary habits on mortality from these diseases. According to results, one in five deaths could be prevented by diet improvements alone. In addition, researchers have indicated that an inadequate diet can cause more deaths than other RFs. Thus, it is possible to assess the level of impact of dietary factors on the development of CNCD³⁵35 GBD 2017 Diet Collaborators. Health effects of dietary risks in 195 countriens, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet 2019; 393(10184): 1958-72. https://doi.org/10.1016/S0140-6736(19)30041-8
https://doi.org/10.1016/S0140-6736(19)30... .

There is evidence that consuming fresh and minimally processed foods is protective against cardiovascular diseases, cancer, among others⁵5 World Health Organization. Global status report in noncommunicable diseases 2014. Geneva: World Health Organization; 2014.,¹⁷17 Brasil. Ministério da Saúde. Secretaria de Atenção à Saúde. Departamento de Atenção Básica. Guia alimentar para a população brasileira. Brasília: Ministério da Saúde; 2014. Disponível em: https://bvsms.saude.gov.br/bvs/publicacoes/guia_alimentar_populacao_brasileira_2ed.pdf
https://bvsms.saude.gov.br/bvs/publicaco... ,³⁶36 World Health Organization. Diet, nutrition and the precention of chronic diseases. Geneva: World Health Organization; 2003. Disponível em: http://apps.who.int/iris/bitstream/handle/10665/42665/WHO_TRS_916.pdf?sequence=1
http://apps.who.int/iris/bitstream/handl... . However, in this research, the consumption of these foods remains insufficient among adolescents, a result that does not differ from studies such as the Health Behavior in School-Aged Children (HSBC). Carried out in Europe, it showed that only 36% of adolescents aged 13 years eat fruit—and this percentage is even lower in adolescents aged 15 years (31%)¹¹11 Oliveira-Campos M, Oliveira MM, Silva SU, Santos MAS, Barufaldi LA, Oliveira PPV, et al. Fatores de risco e proteção para as doenças crônicas não-transmissíveis em adolescentes nas capitais brasileiras. Rev Bras Epidemiol 2018; 21 (suppl 1): E180002. https://doi.org/10.1590/1980-549720180002.supl.1
https://doi.org/10.1590/1980-54972018000... .

Regarding the OP of not having any of the RFs evaluated in this study, our results showed low prevalence throughout the series. These results are still lower than those found in a cross-sectional study that assessed the prevalence and grouping of five RFs in Canadian adolescents and reported 10% of them without any of the RFs³⁰30 Alamian A, Paradis G. Clustering of chronic disease behavioral risk factors in Canadian children and adolescents. Prev Med 2009; 48(5): 493-9. https://doi.org/10.1016/j.ypmed.2009.02.015
https://doi.org/10.1016/j.ypmed.2009.02.... . Likewise, a study conducted in southwestern Nigeria observed that 4.6% of adolescents did not present any of the RFs studied³¹31 Odunaiya NA, Grimmer K, Louw QA. High prevalence and clustering of modifiable CVD risk factors among rural adolescents in southwest Nigeria: implication for grass root prevention. BMC Public Health 2015; 15: 661. https://doi.org/10.1186/s12889-015-2028-3
https://doi.org/10.1186/s12889-015-2028-... . Although there is a methodological difference between studies, if we compare adolescents from some regions of the world, Brazilians evaluated in the three years of PeNSE may be less likely to have no RFs for CNCD.

With regard to the highest OP/EP ratios found in this study, the correlation between alcohol and cigarettes was present in five of the six clusters identified. Similarly, in a cross-sectional study carried out with 4,724 young Canadians, the use of these products was present in three of four clusters identified, representing the strongest combination of the factors, although they are also less prevalent³⁰30 Alamian A, Paradis G. Clustering of chronic disease behavioral risk factors in Canadian children and adolescents. Prev Med 2009; 48(5): 493-9. https://doi.org/10.1016/j.ypmed.2009.02.015
https://doi.org/10.1016/j.ypmed.2009.02.... . Likewise, Rocha and Velasquez-Melendez⁴4 Rocha FL, Velasquez-Melendez G. Simultaneidade e agregamento de fatores de risco para doenças crônicas não transmissíveis em adolescentes brasileiros. Esc Anna Nery 2019; 23(3): e20180320. https://doi.org/10.1590/2177-9465-EAN-2018-0320
https://doi.org/10.1590/2177-9465-EAN-20... reported the group formed by low consumption of fruits and vegetables, smoking and alcohol as the most prevalent.

The positive and simultaneous relationship between alcohol and tobacco is well known, since the former tends to encourage the use of the latter³⁷37 Sanchez A, Norman GJ, Sallis JF, Calfas KJ, Cella J, Patrick K. Patterns and correlats of physical activy and nutrition behaviors in adolescentes. Am J Prev Med 2007; 32(2): 124-30. https://doi.org/10.1016/j.amepre.2006.10.012
https://doi.org/10.1016/j.amepre.2006.10... . This may explain the OP and EP relationships of these variables verified in this study.

The trend analysis showed that four of the six clusters remained stable throughout the series, which shows that these factors are persistently present among adolescents. However, the simultaneous occurrence of regular consumption of processed and ultra-processed foods, insufficient level of leisure-time physical activity, consumption of alcoholic beverages and cigarette use has decreased significantly throughout the series. On the other hand, the cluster formed by insufficient level of physical activity during leisure time, consumption of alcoholic beverages, use of cigarettes and illicit drugs increased, showing that, although the chances of this grouping occurring among students are already high, it grew over the years.

To date, no study has evaluated the trend of RF clusters among adolescents, which highlights the innovation of our results. As already mentioned, it is important to consider the simultaneity of RFs among adolescents, and the importance of health actions aimed not only at the control of RFs seen individually, but considering the individual as a whole, understanding that these factors behave in a multideterminant way¹⁵15 Hu FB. Dietary pattern analysis: a new direction in nutritional epidemiology. Curr Opin Lipidol 2002; 13(1): 3-9. https://doi.org/10.1097/00041433-200202000-00002
https://doi.org/10.1097/00041433-2002020... .

Some limitations were identified in this study because, as it is a cross-sectional study, it prevents inferences on causal directions of reported associations. In addition, when conducting research on risk behaviors, there may be a systematic bias of social desirability, since participants may provide answers that underestimate the prevalence⁴4 Rocha FL, Velasquez-Melendez G. Simultaneidade e agregamento de fatores de risco para doenças crônicas não transmissíveis em adolescentes brasileiros. Esc Anna Nery 2019; 23(3): e20180320. https://doi.org/10.1590/2177-9465-EAN-2018-0320
https://doi.org/10.1590/2177-9465-EAN-20... .

The cluster method may have limitations, since, when identifying patterns, it is possible that the same individual adheres to more than one³⁸38 Cardoso LO, Alves LC, Castro IRR, Leite IC, Machado CJ. Uso do método Grade of Membership na identificação de perfil de consumo e comportamento alimentar de adolescentes do Rio de Janeiro, Brasil. Cad Saúde Pública 2011; 27(2): 335-46. https://doi.org/10.1590/S0102-311X2011000200015
https://doi.org/10.1590/S0102-311X201100... . It is also worth noting that the temporal trend shows the changes of a population group over a period, but not of individuals³⁹39 Campos AQ, Rech SR. Método para pesquisa de tendências: uma revisão do modelo Futuro do Presente. ModaPalavra E-periódico 2016; 9(17): 27-47. https://doi.org/10.5965/1982615x09172016027
https://doi.org/10.5965/1982615x09172016... .

Then, health actions originating from macropolitics, represented by official national health and education bodies, should induce health promotion and the prevention of CNCDs among adolescents, mainly aiming to face the various factors that compromise the full development of this group. It is important that continuous evaluations of these actions are carried out to correct whatever is not achieved. It is also up to school managers, together with students’ families and the community, to identify contexts that expose adolescents to health risk behaviors, as well as engage in social actions to monitor what has been implemented.

REFERENCES

Publication Dates

History