High risk of respiratory diseases in children in the fire period in Western Amazon

ABSTRACT OBJECTIVE To analyze the toxicological risk of exposure to ozone (O3) and fine particulate matter (PM2.5) among schoolchildren.. METHODS Toxicological risk assessment was used to evaluate the risk of exposure to O3 and PM2.5 from biomass burning among schoolchildren aged six to 14 years, residents of Rio Branco, Acre, Southern Amazon, Brazil. We used Monte Carlo simulation to estimate the potential intake dose of both pollutants. RESULTS During the slash-and-burn periods, O3 and PM2.5 concentrations reached 119.4 µg/m3 and 51.1 µg/m3, respectively. The schoolchildren incorporated medium potential doses regarding exposure to O3 (2.83 μg/kg.day, 95%CI 2.72–2.94). For exposure to PM2.5, we did not find toxicological risk (0.93 μg/kg.day, 95%CI 0.86–0.99). The toxicological risk for exposure to O3 was greater than 1 for all children (QR = 2.75; 95%CI 2.64–2.86). CONCLUSIONS Schoolchildren were exposed to high doses of O3 during the dry season of the region. This posed a toxicological risk, especially to those who had previous diseases.


INTRODUCTION
Ozone (O 3 ) and fine particulate matter (PM 2.5 ) are the pollutants with the greatest impact on public health, even at low concentrations a . Annually, approximately 0.7 million deaths by respiratory disease and 3.5 million deaths by cardiopulmonary disease worldwide are attributed to exposure to O 3 and PM 2.5 , respectively, originating from anthropogenic activities 1 .
Since O 3 reaches the lower airways of children, its oxidizing and cytotoxic properties decrease their pulmonary function 7 . Several studies have also showed that exposure to PM 2.5 is an important risk factor for health, especially for cardiopulmonary diseases 15 , even when PM 2.5 derived from biomass burning [8][9][10]13,14 .
In the Amazon region of Brazil, high peaks of atmospheric pollution occur during the dry season. Intense slash-and-burn has been observed in the last few years in Rio Branco, AC, exposing the local population to high levels of atmospheric pollution b .
Monitoring pollutants at soil level is crucial to observe the effects of exposure on human health, especially in children and older adults. The number of monitoring networks in Brazil is growing, but there is no monitoring network in the Amazon region to continuously oversee the main pollutants, even though this region has gained international attention due to its significant amount of pollutants.
The aim of this study was to analyze the toxicological risk of exposure to O 3 and PM 2.5 among schoolchildren.

Study Design
This study is a risk assessment in which we estimated the potential intake dose and the toxicological risk of the pollutants O 3 and PM 2.5 for children aged six to 14 years. This study assessed the risk of exposure to O 3 and PM 2.5 located in an area of biomass burning activities in the Brazilian Amazon. We conducted this study in Rio Branco (the largest city in Acre state, with 336,038 inhabitants c ) between August and October, 2009, during the dry season b . The United States Environmental Protection Agency (EPA) d and the Agency for Toxic Substances and Disease Register e methodologies were used to assess toxicological risk, adapted to estimate the potential intake dose of O 3 and PM 2.5 pollutants.

Study Area and Population
According to education officials in Rio Branco, the public school assessed had similar demographic features as the local population. The school is in the same area as Horto Florestal (approximately 870 meters away), where PM 2.5 and O 3 were hourly measured. The advantage of this location is less traffic in its surroundings when compared with the downtown. It is also in the opposite direction to the industrial area of the city, with mainly brick factories, which prevents interference from pollutants from other sources.
A continuous air quality monitoring station was established and supervised by the atmospheric pollution study group of Instituto de Física of Universidade de São Paulo. Missing data were not attributed for days when monitoring failed. The PM 2.5 concentrations were estimated based on real time measurements of the PM 10 (combination of coarse and fine particulate matter) mass applied to the daily ratio of PM 2.5(FCS) /PM 10(FCS) . Hourly, PM 10 levels were measured by a Tapered Element Oscillating Monitor (TEOM), and PM 2.5 concentrations were obtained by a Fine and Coarse Particulate Matter Sampler (FCS), collected by inertial impaction in 47 mm polycarbonate filters with 4 µm diameter pores. Daily averages were estimated based on the PM 2.5 concentrations that were measured every day in hourly intervals, from 12 a.m. to 11 p.m. The lognormal distribution fits the model best.
Among the 250 children randomly selected from the sample, 237 (95.0%) agreed to participate in the study.

Study Variables
The variables sex, age, and asthma were provided in an individual survey with the children's parents or guardians. The survey was conducted by duly qualified research assistants. Eight questions specifically addressed asthma symptoms, which were related to wheezing, shortness of breath, and coughing, according to the method of the International Study of Asthma and Allergies in Childhood g .
The children's weight and height were obtained in a single measurement at the beginning of the study. Project researchers used a mechanical anthropometric scale with a ruler.
The potential O 3 and PM 2.5 intake dose was estimated for all schoolchildren. The participants were separated into groups stratified by age, sex, presence of asthma, and body mass index (BMI). The average of the eight hours with the greatest O 3 concentration and the average of daily PM 2.5 concentrations were compared among the groups. The equation to estimate the daily potential intake dose and the toxicological risk of O 3 and PM 2.5 followed the general EPA equation 11  We assumed a retention factor of FR = 1, which represents the highest exposure and the highest potential impact on subjects' health.

FA = absorption factor:
We assumed an absorption factor of FA = 1, which represents the highest exposure and the highest potential impact on subjects' health.  16 . Therefore, we assumed the occurrence of constant exposure. According to the EPA 16 , the measurement of an individual's exposure to O 3 is normally conducted throughout the exposure period.
The schoolchildren' s exposure time to PM 2.5 ranged from two to eight hours. This corresponds to the period when children are outdoors, according to recommendations in the Highlights of the Child-Specific Exposure Factors Handbook. h We did not select this exposure time according to its daily variation because the concentration of PM 2.5 , in contrast to O 3 , can vary throughout the day and is independent of ultraviolet radiation i . Therefore, we assumed the exposure to this pollutant was uniform for each 24-hour period. The average exposure time was 122 days, which corresponds to the longest dry period in the region being studied in 2009.
We assumed a constant distribution for the variables average time, duration of exposure, frequency of exposure, and exposure time for O 3 .

RQ I RfD
In which: Risk quotients are classified as follows: RQ ≤ 1: unlikely risk, even in population groups that are sensitive to adverse health effects; RQ > 1: there is a risk of non-carcinogenic adverse effects on human health. I = potential intake dose (μg/kg.day); RfD: reference dose for each pollutant; We estimated each pollutant' s RfD in this study in μg/kg.day units to compare them with the potential intake dose estimated in the exposure assessment. To achieve this, we applied the RfD in the potential intake dose equation above, with average inhalation rates and body weights of all children and environmental variables (PM 2.5 and O 3 ) of the location being studied 2,d .
According to Collins et al. 3 and McDonnell et al. 12 , the estimated RfD for O 3 was obtained assuming the lowest-observed-adverse-effect level (LOAEL) that matches the lowest pollutant dose that may cause observed side effects on human health, including sensitive groups, over a given time of exposure. Studies have found a relationship in which healthy adults and children exposed to 0.12 ppm of O 3 experienced reduced pulmonary function for a one-hour exposure. Expanding the data to the intraspecies uncertainty factor, which was 10, from no-observed-adverse-effect level (NOAEL) to LOAEL, which was 10 In contrast, to obtain RfD for PM 2.5 , we used NOAEL, which corresponds to the maximum dose without any noticeable adverse effects on human health, corresponding to 5.8 μg/m 3i . For PM 2.5 exposures above 5.8 µg/m 3 , we observed an estimated risk of mortality caused by respiratory diseases.

Statistical Analysis
Monte Carlo simulations were used to estimate the potential intake dose in the different subgroups of children for both pollutants being studied. Probabilistic models were used to assess dose by the general equation of the potential dose. The probability distributions for each input model variable were defined after a descriptive analysis and by the adhesion Kolmogorov-Smirnov test results. The input model variables and the assumed probability distributions are presented in Table 1. We estimated average O 3 and PM 2.5 doses according to individual characteristics of schoolchildren, by 1,000 simulations for each category under analysis. In the group of schoolchildren, differences between averages of O 3 and PM 2.5 doses for each category under study were compared using t student and ANOVA tests when appropriate, at a significance level of 5% (95%CI). Model entry variables with the most influence in estimating the dose were identified by Spearman correlation coefficients. Application R 2.13 was used in simulations and statistical analyses.

Ethical Aspects
This study was approved by the Ethics Committee of the National School of Public Health (CEP/ESNP/FIOCRUZ -Protocol 25/07 -on March 7, 2007). The children's parents or guardians signed an informed consent form.

RESULTS
The highest O 3 concentrations were recorded in December with two peaks over 100 µg/m 3 , which exceeds the air quality standard levels prescribed by the WHO. It did not rain on those days, and relative humidity was 76.0% and 80.0% (Figure 1).  Figure 2).
The lognormal probability distribution was used to simulate the concentration, inhalation rate, and body weight of schoolchildren with the results of the adhesion Kolmogorov-Smirnov test placed in the best fit for the data. The uniform probability distribution was assumed for exposure time (ET) while the exposure frequency (EF), duration (ED), and average time (AT) were maintained constant in the model (Table 1).  The potential average dose of O 3 was higher than that of the PM 2.5 dose. The doses differed depending on age. Schoolchildren aged six to eight years inhaled a higher potential average dose than those aged nine to 14 years for exposure both to O 3 and PM 2.5 .The comparison between sexes showed statistically significant differences only for exposure to O 3 (p = 0.008).
The differences between children with and without asthma were significant for exposures to O 3 and PM 2.5 . Among normal-weight schoolchildren, we estimated an average potential dose for O 3 and PM 2.5 exposure. Both exposure doses significantly differed between normal-weight and overweight schoolchildren ( Table 2).
Based on the estimated reference RfD dose of 1.03 µg/kg.day of O 3 and 1.14 µg/kg.day of PM 2.5 , we estimated toxicological risks by the ratio between average potential doses and RfD.
Regarding O 3 exposure, 95,0% of schoolchildren exposed to this pollutant had risk quotients above 1, which means a toxicological risk of exposure to this pollutant. For PM 2.5 , we did not find any toxicological risk for children arising from exposure to this pollutant ( Figure 3).
The variables O 3 and PM 2.5 concentration were the ones most strongly correlated with the potential intake dose (r = 0.38 and r = 0.68, respectively). The variable weight was negatively related to the average potential dose, for both O 3 (r = -0.29) and PM 2.5 (r = -0.12).

DISCUSSION
We verified that schoolchildren aged six to 14 years experienced toxicological risks for O 3 from biomass burning in 2009, in the "arch of deforestation", located in Rio Branco.   We did not find health risks for children exposed to PM 2.5 . However, during the study the concentrations of this pollutant surpassed the levels prescribed by the EPA and WHO. The highest daily average concentration of PM 2.5 , measured on September 14, 2009, was 46.0% higher than the air quality standard prescribed by the EPA, which is 35 μg/m 3i .
Our results were similar to a study conducted in Mexico, which also showed that toxicological risk to the chemical components of PM 2.5 was 1.81 for children aged between 6-12 years, but no risk was observed when each chemical component was individually analyzed 5 . However, although any toxicological risk for PM 2.5 was observed, exposed individuals may experience non-observable health effects caused by exposure to particulate matter. Several international epidemiological studies showed harmful effects associated with even low concentrations of PM 2.5 . The doses of exposure were lower than those estimated in Rio Branco during the 2009 dry season 15 . Potential health effects depend on the multi-element composition of particulate matter and its aerodynamic characteristics, its capacity for reaction with other elements or compounds, persistence in the environment, transportation capacity across long distances, exposure time, local climate conditions, and human susceptibility, with several possible impacts on human health 3,13,15 .
Oliveira et al. 14  Even though both studies used the same methodology, the reference concentration for particles released from diesel combustion applied by Oliveira et al. 14 was lower than the PM 2.5 NOAEL applied in this study (5.0 μg/m 3 and 5.8 µg/m 3 , respectively). However, even using the same reference concentration as Oliveira et al. 14 , the toxicological risk in our study would not be > 1 for PM 2.5 . Furthermore, the average PM 2.5 concentrations were 2.5 times higher in Tangará da Serra compared with Rio Branco, which could explain the different findings.
The pollutant concentration is a major factor in determining the toxicological risk, since the risk is strongly related with the potential average dose inhaled by schoolchildren exposed to O 3 and PM 2.5 . Therefore, the variable pollutant concentration had the greatest influence in the sensitivity analysis over potential intake doses in both studies.
It is currently understood that, according to the EPA 12 , the use of NOAEL for PM 2.5 is more appropriate because it is specific PM 2.5 . Although there is no research similar to ours addressing children' s exposure to O 3 , a study showed the association between the breathable dose of an individual exposed to O 3 and changes in pulmonary function for different levels and exposure duration 12 .
In Rio Branco, O 3 reached maximum levels of 119.4 µg/m 3 , which coincided with the scarce rainfall in the period. Rainfall can increase O 3 levels because it transfers NO 2 , an important O 3 precursor, closer to the surface, increasing NO 2 levels and consequently O 3 formation reactions 6 . Another factor that favors the formation of O 3 in Rio Branco is the extension of its forests: approximately 87.0% of its territory still has exuberant forests. Ozone is typically formed when precursors from combustion emissions, such as NO x , reach an area with abundant volatile organic compounds (VOC) and solar radiation. The VOC in the Brazilian Amazon are abundantly available in forest areas where vegetation is the greatest natural source 6 .
Even if Rio Branco does not have many slash-and-burns like other regions of the Amazon, its population may be subject to a large amount of O 3 precursor pollutants from other states such as Rondonia and Mato Grosso 4 .
Schoolchildren aged six to eight years incorporated the highest average potential doses of O 3 and consequently experienced highest toxicological risk, with 20.0% higher risk of effects on health when compared with 9-11 and 12-14 age groups. Because of their physiological growth and pulmonary development, children are vulnerable to environmental pollutants a . In this study, 19.0% of children were classified as asthmatic, according to the International Study of Asthma and Allergies in Childhood (ISAAC) score g . Asthmatic schoolchildren inhaled high average potential doses for O 3 exposure. There is evidence in the literature that asthmatic children are more vulnerable to adverse effects caused by exposure to O 3 , following the hypothesis that inhaling high doses of O 3 could lead to airway hyperactivity and inflammation, and that this would make individuals with asthma more likely to experience pulmonary obstructions 11 . In a cohort study, the incidence of new asthma diagnoses increased among children living in regions with high O 3 concentrations 11 .
The toxicological risk for exposure to O 3 in schoolchildren evidenced in our study indicates that air quality standards prescribed by the EPA and WHO do not protect human health from exposure to this pollutant. The O 3 LOAEL used in the present study corresponds to the lowest dose of the pollutant that can cause an adverse effect on human health, including vulnerable subgroups, during a certain exposure time. It is eight times lower than the level established as the air standard quality for O 3 in Brazil by Conselho Nacional do Meio Ambiente (CONAMA -National Council for the Environment) j , which is 160 µg/m 3 . This is the maximum tolerable concentration of O 3 during an average one-hour period. The CONAMA is responsible for setting air quality standards for pollutants in Brazil. Its latest update in environmental legislation occurred in 1990, which we consider out of date.
Limitations of this study include insufficient coverage of the population exposed to slashand-burns by air quality monitoring networks across longer periods, which would allow for evaluating a trend of exposure to the main pollutants released by burns in the region. Another limitation is the quality of healthcare data, their standardization, and accessibility. Lack of agreement in environmental agencies on the reference concentration for O 3 is also associated with the lack of continuous air quality monitoring networks. Children' s inhalation rate was obtained from an international study, since there are no similar studies in Brazil providing measurement parameters for individuals' daily inhalation rate according to age group, sex, and BMI. Finally, it was also difficult to acquire accurate PM 2.5 measurements, which were obtained from the daily ratio between PM 2.5(AFG) /PM 10(AFG) applied to real time PM 10 (TEOM) mass measurements.
We conclude that schoolchildren residing in Rio Branco were exposed to high doses of O 3 during the dry season of the region, and this poses toxicological risk. Schoolchildren aged six to eight years incorporated the highest average potential doses of O 3 and consequently experienced the highest toxicological risk.