Asthma is a chronic respiratory disease characterised by lower airway hyper-responsiveness and variable airflow limitation, reversible spontaneously or with treatment, that is most common among children( Reference O'Byrne 1 ). In recent decades, there has been a significant increase in asthma prevalence in different populations( Reference Anandan, Nurmatov and van Schayck 2 ). It is estimated that approximately 300 million people worldwide are asthmatic, with significant variations in prevalence between different countries and regions( Reference Cooper, Rodrigues and Cruz 3 ). The International Study of Asthma and Allergies in Childhood (ISAAC) found an average asthma prevalence of 24·3 % among urban Brazilian children( Reference Solé, Wandalsen and Camelo-Nunes 4 ). Recent studies in Salvador, Brazil have shown that a minor proportion of these cases was attributable to atopy, while the majority (80 %) were non-atopic( Reference Cooper, Rodrigues and Barreto 5 – Reference Barreto, Cunha and Fiaccone 7 ).
Asthma is a multifactorial illness; its causes are still not completely understood and there is no consensus on its aetiology. A large body of research emphasises the role of environmental and genetic factors in the onset of asthma( Reference von Mutius 8 ). Recently, there has been great interest in the relationship between psychosocial factors and asthma morbidity( Reference Priftis and Anthracopoulos 9 ). Food and nutrition insecurity (FNI) has been linked to psychological distress and maternal depressive symptoms, which can also have negative effects on the mother's children( Reference Murphy, Wehler and Pagano 10 , Reference Belsky, Moffitt and Arseneault 11 ). Associations between FNI and chronic diseases, including asthma, have been reported( Reference Hendrickson, O'Riordan and Arpilleda 12 , Reference Kirkpatrick, McIntyre and Potestio 13 ). A hypothesis suggests that family stress contributes to the occurrence of asthma( Reference Feitosa, Santos and do Carmo 14 , Reference do Carmo, Santos and Amorim 15 ). FNI is defined as limited or uncertain availability of, or the inability to acquire, adequate, safe and acceptable nutritious food due to limited financial resources. Approximately 30·2 % of Brazilian households are considered food insecure; this value rises to 46·1 % in the north-eastern region of the country( 16 ).
While there is compelling evidence for the associations between stress and asthma and between FNI and stress, the evidence for an association between asthma and FNI is limited. The present study aimed to evaluate the association between FNI and asthma in a population of children from a small municipality in north-eastern Brazil.
Materials and methods
Study/population/sample
A cross-sectional design was used to study children aged 6–12 years living in São Francisco do Conde, a municipality located in the metropolitan region of Salvador, north-east Brazil. This municipality has 33 183 inhabitants and a high urbanisation rate (80·2 %). The city council is the largest local employer. In Bahia state, out of 417 municipalities, São Francisco do Conde ranks third in the economic development index but has a far lower position regarding social (30th), education (139th) and health (178th) development indices( 17 ).
We used data from the Education Department of the Municipality of São Francisco do Conde for the year 2010 to estimate the sample size. From all 3734 registered students, there were 2649 from rural areas and 1085 from urban areas. These students were distributed across twenty-two schools in the county school system. To minimise travel costs and time for subject recruitment, only schools with 150 or more students were included in the sample. We performed a random selection of schools using clustered sampling. All students aged 6–12 years in each selected school were eligible to participate in the study. Considering that the prevalence of respiratory allergies, FNI, malnutrition and helminth infection varies from 15 % to 40 %, a sample size of 531 and 834 students from urban and rural areas, respectively, was determined taking into account 3 % error and 95 % confidence. We added 10 % to the total sample size to account for students declining to participate in the study, resulting in 1500 students.
Procedures
Data were collected between August and December 2010 by qualified and trained personnel. The principal of each selected school received a letter that included an explanation of the study's objectives and methodology and an invitation for his/her students to participate in the research study. The students’ parents attended informational meetings where they received extensive information on the project and were invited to enroll their children in the study. Participating parents signed a consent form and were invited to complete the study questionnaires.
Outcome variable
The prevalence of asthma and rhinitis symptoms was measured using a previously validated questionnaire from the ISAAC, phase III, translated into Portuguese( Reference Sole, Vanna and Yamada 18 ). For the present study, the presence of asthma was defined based on two questions: (i) ‘Has your child ever experienced asthma or wheezing in his/her lifetime?’ and (ii) ‘In the last 12 months, has your child experienced wheezing?’ The children who had a positive answer to both questions or to the second question only were considered to have asthma (wheezing) in the last 12 months.
Main independent variable
FNI was measured using the Brazilian Food Insecurity Scale (BFIS), which has been adapted and validated for the Brazilian non-institutionalised population( Reference Pérez-Escamilla, Segall-Corrêa and Kurdian Maranha 19 ). Questions about food insecurity were answered by the person responsible for feeding the family. The BFIS contains fifteen items that are used to measure food security among the population. Responses to these items are also used to assess different degrees of food insecurity experienced by families (mild, moderate and severe). We considered the presence or absence of residents under 18 years old in households when determining the level of food insecurity based on the scale score( Reference Marín-León, Segal-Corrêa and Panigassi 20 ). In households with children under 18 years old, the following cut-off points were used: 1–5 (mild FNI), 6–10 (moderate FNI) and 11–15 (severe FNI). A BFIS score of zero indicated food security. For the present study we adopted three dummies to characterise food and nutrition insecurity: dummy 1 was mild, dummy 2 was moderate, and dummy 3 was severe.
Potential confounding variables
The following potential confounding factors were considered in the study, where Ref. indicates the reference category: sex (male; female = Ref.), age (<10 years; ≥10 years = Ref.), mother's education level (≤4th grade; ≥5th grade = Ref.), place of residence (urban; rural = Ref.), per capita income (in Brazilian minimum salaries (BMS), <1/2 BMS; ≥1/2 BMS = Ref.), number of people living in the household (>4; ≤4 = Ref.), tobacco smoke exposure (yes; no = Ref.) and BMI. We used the WHO standard reference tables (2007)( Reference de Onis, Onyango and Borghi 21 ) to determine the percentile value of BMI for age and sex for each participant. To classify these BMI measurements, we used the WHO's criteria established in 2006( 22 ). For our analysis, overweight and obesity categories were aggregated. Therefore, individuals with BMI ≥ 85th percentile were considered to have excess weight.
Statistical analysis
For processing and construction of the database, we used Epi Info version 6·04. The data were entered in duplicate after reviewing the questionnaires and correcting for errors from the coding field.
The characteristics of the population were identified by descriptive analysis using categorised data for prevalence and means and standard deviations for continuous variables. The magnitude of the associations between asthma and different levels of food insecurity was expressed as odds ratios with 95 % confidence intervals. The statistical analysis used two-tailed tests and a significance level of 5 %.
The presence of confounding variables was explored using multivariable logistic regression models. We used the backward modelling strategy in the multivariable logistic regression analysis, considering as confounding variables those that caused a difference in the magnitude of the association of ≥10 % when excluded from the model( Reference Rothman and Greenland 23 ). However, gender and age were included and retained in the model independent of this criterion. The choice of variables for modelling was based on knowledge gathered from pre-existing literature.
The statistical analyses were performed using the statistical software package SPSS version 13·0.
Ethical issues
Ethical approval was provided by the Ethical Committee of the School of Nutrition, Federal University of Bahia, Brazil (registration number 27-09/CEPNUT). Written informed consent detailing all procedures to be carried out with the participants was signed by a parent or the legal guardian of each child.
Results
Of the total number of students initially selected (1500 students), 193 (12·9 %) did not participate in the study. These sample losses were due to refusal or the child moving to another city or transferring to another school. The final study sample consisted of 1307 students of both sexes, aged 6–12 years.
There was a slightly higher percentage of male students (53·6 %) and a higher percentage of children aged 6–9 years (51·3 %). Other descriptive characteristics of the study sample are shown in Table 1. It was found that 10·4 (95 % CI 8·7, 12·1) % of study participants had symptoms consistent with asthma (wheezing in the last 12 months). In addition, 64·5 % of participants were experiencing FNI, of which 42·5 % were mild, 15·9 % moderate and 6·1 % severe.
Frequency of asthma according to characteristics of the study population: elementary-school children (n 1307) aged 6–12 years, São Francisco do Conde, Bahia, north-eastern Brazil, 2010
FNS, food and nutrition security; FNI, food and nutrition insecurity.
*Different totals because of non-available data.
†Measured in Brazilian minimum salaries (BMS). Value in 2010 = R$ 510·00 (Brazilian reais; equivalent to $US 290·70).
‡Using the χ 2 test; other P values obtained using Mantel–Haenszel's linear association test.
The results of the multivariate unconditional logistic regression model adjusted for sex, age, tobacco smoke exposure, place of residence and anthropometric status showed that children with moderate FNI were 1·71 times more likely (95 % CI 1·01, 2·89) to have asthma (wheezing) compared with food-secure children. The odds increased to 2·51 (95 % CI 1·28, 4·93) under conditions of severe food insecurity (Table 2).
Univariate and multivariate associations between food and nutrition insecurity and asthma in the studied population: elementary-school children (n 1307) aged 6–12 years, São Francisco do Conde, Bahia, north-eastern Brazil, 2010
FNS, food and nutrition security; FNI, food and nutrition insecurity; Ref., reference category.
Statistically significant odds ratios and 95 % confidence intervals are denoted in bold font.
*Multivariate model adjusted for sex, age, tobacco smoke exposure, place of residence and anthropometric status.
†Measured in Brazilian minimum salaries (BMS). Value in 2010 = R$ 510·00 (Brazilian reais; equivalent to $US 290·70).
The goodness-of-fit test (Hosmer–Lemeshow test) indicated that the model showed satisfactory calibration (P > 0·05).
Discussion
The prevalence of asthma symptoms recorded among school children in São Francisco do Conde (10·4 %) is lower than the prevalence reported in other regions of Brazil( Reference Cooper, Rodrigues and Cruz 3 ). A possible explanation for this discrepancy is that most of the studies on asthma in Brazil have been carried out in large cities, whereas the present study was performed in a small town and its rural surroundings. However, this prevalence is comparable with those recorded in urban areas in Argentina, Chile and Mexico( Reference Cooper, Rodrigues and Cruz 3 ).
Using a multivariable logistic regression analysis adjusted for sex, age, tobacco smoke exposure, place of residence and anthropometric status, we observed a statistically significant, positive and markedly dose-dependent association between FNI and asthma (wheezing). The potential confounding variables used in the analysis have been identified in several previous studies as associated with both asthma and food insecurity in childhood( Reference Semba, Campbell and Sun 24 – Reference Baig-Ansari, Rahbar and Bhutta 26 ). The literature has consistently reported that FNI is associated with poor well-being and lower quality of life for families( Reference Belsky, Moffitt and Arseneault 11 , Reference McIntyre, Connor and Warren 27 – Reference Alaimo, Olson and Frongillo 29 ). However, the mechanisms for this relationship remain poorly understood. It is unclear whether the experience of food insecurity is associated with adverse health effects independent of the effects of poverty. Our study results are in agreement with a few previous investigations that evaluated the relationship between FNI and asthma( Reference Kirkpatrick, McIntyre and Potestio 13 ). The influence of food insecurity on asthma in children and adolescents is in accordance with the argument that food insecurity leads to various health problems, such as depression, anxiety and maternal stress. It is possible that when the mother's emotional condition is impaired the mother–child bond can be weakened, with consequent effects on the level of care and an increase in the risk for disease( Reference Priftis and Anthracopoulos 9 , Reference Schneider, Stein and Fritscher 30 – Reference Dean and Sharkey 33 ). We also consider that constant exposure to food insecurity is a source of chronic stress and can contribute to emotional and behavioural problems, such as anxiety, depression and conduct disorders in children, which can trigger or worsen asthma symptoms( Reference Murphy, Wehler and Pagano 10 , Reference Feitosa, Santos and do Carmo 14 ). Further research is warranted to better understand the pathways by which experiences of food insecurity may contribute to poor health outcomes, and specifically to asthma.
Limitations and strengths of the study
The current study has limitations inherent to cross-sectional studies. Specifically, because all of the variables are estimated at one time point in the cross-sectional design, temporality cannot be established, limiting the inference of a causal relationship between FNI and asthma. If FNI causes asthma, it must be mediated by a chain of other psychosocial related factors. An additional limitation is that there is no attempt to measure such factors in our study. The literature is controversial regarding whether psychosocial factors are confounding or mediator factors in the association between FNI and asthma( Reference Klinnert, Nelson and Price 34 – Reference Maes, Hadley and Tesfaye 37 ). We are in the group that argues in favour of psychosocial factors mediating the association between FNI and asthma. Then, the absence of this information precludes us to assess if psychosocial factors reduce or erase the association. Furthermore, we do not have dietary intake data available, making it difficult to ascertain if those with greater degrees of food insecurity have diets that are significantly different from those without food insecurity, especially for some groups of foods that have been associated with asthma prevalence. We emphasise the importance of conducting further modelling approaches, including some other predictor variables that potentially could measure factors that predispose children to both food insecurity and asthma and that were not included in our models.
The strengths of our study include its large and homogeneous study population; that it was carried out in a small town and its surrounding rural area; and the focus on food insecurity. It should be noted also that possible confounding effects related to socio-economic status were minimised in the present study.
Conclusions
The prevalence of asthma symptoms among the studied children reinforces the need to intervene on the factors that are associated with asthma vulnerability. Furthermore, the results point to severe food insecurity as a marker of vulnerability to asthma symptoms in this and maybe other similar populations. It could be inferred that public policies focused on decreasing food insecurity have a role in decreasing the burden of asthma during childhood and adolescence. The results also suggest the need for further studies to identify the mechanisms involved in the relationship between FNI and wheezing/asthma.
Acknowledgements
Sources of funding: This work was supported by the Fundação de Amparo à Pesquisa do Estado da Bahia (FAPESB; project nos. 7638/2009 and 7597/2009) and the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES; process no. 1254/10-3). Conflicts of interest: The authors declare no conflict of interest. Authors’ contributions: R.d.C.R.-S. participated in the study design, data collection, interpretation of results and writing of the manuscript. A.O.A.-M. participated in the interpretation of results and writing of the manuscript. S.J.B. participated in the field work and reviewed the manuscript. R.L.F. designed the work and reviewed the manuscript. S.M.d.S.C. participated in the interpretation of results and revision of the manuscript. M.L.B. reviewed the manuscript. E.d.J.P. reviewed the manuscript. L.A.d.S. participated in the interpretation of results and revision of the manuscript. L.C.R. participated in the interpretation of results and revision of the manuscript. N.M.A.-N. participated in the study design and revision of the manuscript.
