Time trend of asthma in children and adolescents in Brazil , 1998-2008

MÉTODOS: Foram analisados os dados de prevalência de asma da Pesquisa Nacional por Amostra de Domicílios, dos anos de 1998, 2003 e 2008. A amostra foi constituída por 141.402, 144.443 e 134.032 indivíduos em 1998, 2003 e 2008, respectivamente, e a análise foi ajustada pelo desenho amostral. As tendências de asma foram descritas por sexo, regiões do Brasil e local de residência, em crianças (zero a nove anos) e adolescentes (dez a 19 anos).


INTRODUCTION
Asthma affects approximately 300 million people worldwide, 12 with negative consequences for individuals and society.These consequences, such as days and costs resulting from hospitalizations, have an economic and social impact in Brazil, a where asthma is an important public health problem. 8A Brazilian study assessed death certifi cates and included asthma as a cause of death, revealing a mortality rate of 2.29 per 100,000 inhabitants, standardized for the Brazilian population in 2000. 18Asthma affects all ages, especially children and adolescents. 6Although several studies on asthma prevalence have been performed worldwide, there is no consensus on a single accurate measure to defi ne this disease, a which is usually assessed with a medical diagnosis or the presence of its symptoms.prevalence of asthma in childhood, with a reduction in mortality rates in the same period. 1A study conducted by Pearce et al, 15 with data from the International Study of Asthma and Allergy in Childhood (ISAAC), showed a reduction in asthma in English-speaking countries and Western Europe.In contrast, there was an increase in the asthma prevalence throughout the years in countries where the prevalence was considered to be low.In Latin America, the prevalence of asthma symptoms has increased. 15Nonetheless, in South America, a systematic review pointed to inconsistencies in the decreasing or increasing pattern of asthma prevalence in studies that use the ISAAC methodology. 3e Pesquisa Nacional por Amostra de Domicílios (PNAD -National Household Sample Survey), performed by the Instituto Brasileiro de Geografi a e Estatística (IBGE -Brazilian Institute of Geography and Statistics), regularly investigates health aspects of the population.This survey is conducted with a representative sample of states, Brazilian regions and the country.A total of three studies on the health of the Brazilian population have been conducted, including the question of the medical diagnosis of asthma, which enables the establishment of a time trend of this problem in Brazil, thus contributing to the monitoring of the medical diagnosis of asthma throughout time and the implementation of public policies.
The present study aimed to analyze the trends of asthma prevalence in children and adolescents between 1998 and 2008, in Brazil.

METHODS
The study was conducted with data derived from the 1998, 2003 and 2008 PNAD health supplements.
The PNAD sampling plan follows three stages (cities, census tracts and households), being representative for the country, regions, states and certain metropolitan areas. 21 The following closed-ended question was used to defi ne asthma prevalences in each year: "Has any doctor or health professional said that (…) has/have asthma or bronchitis?".
The analyses were performed for the country and also stratifi ed per age groups, Brazilian regions and area of residence.Age groups were dichotomized into participants aged between zero and nine years and participants aged between ten and 19 years.Asthma prevalences were also analyzed according to regions (North, Northeast, Southeast, South and Central-West).Finally, the area of residence was dichotomized into "urban" (city or village in an urban area, city or village in a non-urban area and isolated urban area) and "rural" (rural cluster with an urban extension, isolated rural cluster and settlement, isolated rural cluster -nucleus, isolated rural cluster -other types of cluster, and rural area except for rural cluster).
The databases were obtained from the IBGE.Data were analyzed in the STATA statistical package, version 11.0, using the svyset command to define sample weights and clusters and the svy command for all analyses performed, considering the sampling process complexity.Asthma prevalences were obtained for the country and per region, in addition to annual percentage variations.Moreover, this information was stratifi ed per area of residence and sex.
The mean annual percentage variation of the prevalences observed was calculated with the values found in 1998 and 2008.The total percentage variation was calculated, with the mean annual variation being estimated as a tenth root: where p 1998 and p 2008 are the prevalences calculated for 1998 and 2008, respectively.The Wald test for linear trend was used to observe the time trend.The sample characteristics in each year of the PNAD are shown in Table 1.In all years studied, there was a slight predominance of males among children and adolescents.The most frequent area of residence was the urban area in all years studied.There was an increase of approximately 5.0 percentage points in children living in the urban area in 2008, when compared to 1998.Among adolescents, this increase totaled 2.5 percentage points in the same period.In addition, there was a reduction in the proportion of children and adolescents living in the Northeast, Southeast and South regions, whereas the North and Central-West regions had an increase in the proportion of residents aged between zero and 19 years (Table 1).The mean per capita household income increased in the period studied, totaling R$ 359.30 in 2008 (SD = 580.90) in the zero-to-nine-year age group, and R$ 407.20 in the ten-to-19-year age group (SD = 606.70).

The sample of children and adolescents
Figure 1 shows the asthma prevalences according to the area of residence and sex in children aged between zero and nine years, per year of PNAD.Regardless of the area of residence, there was an increase in prevalence in the period studied.In the urban area, the prevalences varied from 8.6% in 1998 to 8.7% in 2003 and 9.1% in 2008.In the rural area, they varied between 4.9% and 5.9%, from 1998 to 2008.The prevalence of asthma was higher in male children, in both rural and urban areas.
Table 2 shows the prevalence of asthma and annual variation in children aged between zero and nine years, in general and per Brazilian region.The prevalence of asthma in children varied from 7.7% to 8.5% in 1998 and 2008, respectively.With the exception of the Central-West region, where there was a decline that varied between 0.7% (when the female sex was evaluated) and 3.0% (in the rural area), all Brazilian regions and the country as a whole showed an increase in asthma prevalence, although not statistically significant in some cases.In Brazil, the annual increase was lower in the urban area (0.6%, p=0.006) than in the rural area (1.8%, p=0.002).In the rural area, annual increases in asthma prevalence were 3.5% (p<0.001) in the Northeast region and 2.2% (p=0.035) in the Southeast region.
The trend in asthma prevalence in adolescents according to area of residence and sex is shown in Figure 2.This prevalence also increased in the period studied, although the differences between sexes were not as relevant as they were in children.The exception The annual variations and prevalences for adolescents are shown in Table 3.The variation in asthma prevalence was between 4.4% and 5.5%.Although with lower prevalences, adolescents had higher percentage variations than children.The percentage variation in the country points to an annual increase in asthma prevalence of 2.2% (p<0.001) in the ten-to-19-year age group.This increase corresponded to 1.9% (p<0.001) in the urban area and 3.5% (p<0.001) in the rural area.Although the North region showed an annual decrease of 0.5% in this prevalence in the ten-to-19-year age group, this same region pointed to an increase of 0.9% in the urban area, both of which were not statistically signifi cant.Once again, the annual increase in asthma prevalence in the rural area was emphasized, totaling 4.8% (p<0.001) in the Northeast region, and 4.3% (p=0.005),4.4% (p=0.010) and 2.1% (p=0.268) in the Southeast, South and Central-West regions, respectively.

DISCUSSION
The results of the present study indicate a general increase in the diagnosis of asthma in children and adolescents in different Brazilian regions, in both sexes, and in the rural and urban areas, between 1998 and 2008.The greatest increases were observed in males and in those living in the rural area.
Despite the national scope, with representativeness for states and regions, and methodological rigor, analyses based on PNAD data can have certain limitations.One of them is the possible underestimation of asthma prevalence, as it depends on the diagnosis by a medical doctor or health professional.If the asthma assessment criteria had taken its symptoms into account, prevalences could have been higher in all the years studied.Another point to be considered as a limitation is that the outcome was assessed in all years in a retrospective way, which could also have led to an underestimation of magnitude.One of the main diffi culties in the comparison of results is the way the disease is assessed, which could vary from the presence of symptoms such as wheezing in the last 12 months; a medical diagnosis; to tests being performed, such as the bronchoprovocation test with methacoline conducted in a hospital environment.As an example, asthma prevalences are usually higher in studies that use the presence of symptoms to defi ne the outcome, 22 which can be partly explained by the greater access to health services.
The recent literature on time trends in asthma shows the large use of the comparison between stages I and III of ISAAC, 2,5,11,15,16,20 which observed this prevalence in the six-and-seven-year age group, in the 13-and-14-year age group, or in both.However, the majority of these studies were performed in Europe and Asia; only three of these were conducted in Latin America, of which one was in Chile 11 and two in Brazil. 16,20The ISAAC investigates the prevalences of wheezing in the last 12 months and self-reported asthma in life as outcomes of this disease.Nonetheless, ISAAC studies make a comparison between two periods and, consequently, characterize its evolution rather than its time trend, as the latter requires at least three points in time.
The present study showed the increase in prevalence of asthma in children and adolescents during a period of ten years (from 1998 to 2008).This result is similar to that of other studies, although these assessed the outcome differently.In Greece, Anthracopoulos et al 4 found an increase from 8.0% in 1991 to 12.4% in 2003 by assessing the medical diagnosis of children aged between eight and ten years.In the United States, data from the Centers for Disease Control and Prevention pointed to an increase of 4.6% from 1980 to 1996 and a stability in the prevalence of asthma in individuals aged between zero and 17 years, between 1997 and 2007. 1 By analyzing information from all countries that participated in the ISAAC, Pearce et al 15 observed an annual increase in asthma prevalence in adolescents that varied between 0.07% in Africa and 0.93% in Oceania.Only in the Indian subcontinent was there an annual decrease of 0.01%.In children aged between six and seven years, the annual increase varied from 0.12% on the Asian side of the Pacifi c Ocean to 0.74% in North America.In Africa, Eastern Mediterranean and Latin America, there was an annual decrease in prevalence of 0.01%, 0.05% and 0.15%, respectively.
There were disagreements among the fi ndings when studies conducted in South America were compared.
In Chile, 11 a study derived from the ISAAC project pointed to a reduction in prevalence from 12.5% in 1994 to 10.7% in 2002 in the six-to-seven-year age group, although not being statistically signifi cant.In contrast, among adolescents aged between 13 and 14 years, there was an increase from 10.2% to 14.9% in the same period.In Brazil, 20 also as part of the ISAAC, the prevalence of asthma remained stable in individuals aged between 13 and 14 years (from 14.9% in 1994 to 14.7% in 2002).However, it should be emphasized that studies using the ISAAC methodology provide a selfreported prevalence of asthma rather than the medical diagnosis of this disease, as in the PNAD.The prevalence of asthma increased at least ten times in recent decades, especially in developed countries, but also in developing countries such as Brazil and other Latin American countries. 8Some explanations can be attributed to the increase in primary health care services and the number of family health teams in all Brazilian regions.As an example, the expansion of family health teams between 1999 and 2004 was greater in the Northeast region than in the South region. 9Considering the fact that the outcome was assessed with the question about medical diagnosis, it is assumed that the increase in access to and use of health services could partly explain the increase in asthma prevalences, especially in the rural area.
Another factor that could explain the increase in asthma in rural populations is their greater exposure to pesticides.Although this is only a hypothesis in the theoretical sphere and based on scientifi c speculation, researchers consider it to be an important aspect to be approached in future Brazilian studies.
The highest prevalences of asthma in males coincide with what has been described in previous publications.
A study conducted with students in 1984, 1994, 1999 and 2004 by Osman et al 14 found higher prevalences and time trends of asthma in males.In children, the highest prevalences of asthma in males, when compared to females, can be explained by physiological disadvantages. 23Boys at this age tend to have higher IgE concentrations than girls, 19 in addition to lower respiratory fl ow rates. 10,17thma management conditions in Brazil are below what is ideal.Considering the fact that asthma is a disease sensitive to primary health care management, hospitalizations are viewed as inevitable. 13A study on the trend of medical diagnosis of asthma in the United States 7 pointed out that, in addition to medication, good asthma self-management training is also required.Thus, it becomes important to improve primary health care coverage in Brazil, including public policies aimed at asthma control and management.
Time trends of asthma show different patterns according to geographical areas.In the United States, between 1980 and 2007, there was a stabilization in the a Global Strategy for Asthma Management and Prevention: updated 2009.Cape Town: University of Cape Town Lung Institute; 2009.

Table 1 .
Description of the sample of children and adolescents according to demographic and socioeconomic characteristics.Brazil, 1998-2008.Asthma prevalence in children aged between zero and nine years per place area of residence and year, according to sex.Brazil, 1998-2008.
a Adjusted for cluster effect and sample weight.Figure 1.

Table 2 .
Asthma prevalences and annual variation in children aged from zero to nine years, according to sex, area of residence and Brazilian region.Brazil,1998-2008.In the North region, in 1998 and 2003, the PNAD only considered the state of Tocantins for the rural area. a

Table 3 .
Asthma prevalences and annual variation in adolescents aged from ten to 19 years, according to sex, area of residence and Brazilian region.Brazil, 1998-2008.In the North region, in 1998 and 2003, the PNAD only considered the state of Tocantins for the rural area. a