Particulate matter is a complex mixture of solid particles and liquid droplets found in the air, which comes from various natural and anthropogenic sources. This form of pollutant can have different sizes and can be composed of many types of materials and chemicals [1] [2] . Numerous epidemiological studies have found that the expo- sure to PM, especially the fine fraction (PM_2.5) has adverse effects on human health, especially for vulnerable populations [1] [3] [4] . Children are more vulnerable than adults to the effects of exposure to polluted air, due to their stage of physical growth, immature immune system, and developing respiratory organs with a more suscep- tible and reactive respiratory epithelium [3] [5] . Extensive evidence has associated exposure to PM_2.5 from ve- hicular traffic and fuel burning with impaired pulmonary function and increased respiratory complaints on children [3] [4] [6] -[11] . However, few studies have reported on the effects of PM from the soil and dust of mine tailings [12] [13] .

Chile is one of the largest copper producers worldwide, and therefore copper is one of the country’s major sources of economic income of the country [14] . Most mines are in the central and northern Chile, distributed along the Cordillera de Los Andes. El Salvador is an open-pit mine, located at 2600 m above sea level (26˚15' South Lat. S.; 69˚ West Long). The chemical composition of this mineral corresponds to cuprous primary por- phyry mineralization, which one is characterized by alkali feldspar-biotite-anhydrite-chalcopyrite and bornite- chalcopyrite-pyrite mineral assemblages [15] . As a result of the mining operations, between 1939 and 1975, more than 150 × 10⁶ Mg of tailing were discharged continuously into the Rio Salado without any treatment, being deposited in the bay of Chañaral. This modified the coastline, expanding the area of the beach significantly and causing a heavy siltation and pollution of the bay, directly affecting more than 20 km of coastline and covering about 12 km² [16] [17] (Figure 1). This resulted in pollution of tailings sands rich in Cu, Fe, As, Zn, Cn, Pb, As, Hg, Mo and other heavy metals [18] [19] . Coastal winds carry the particulate material contamination over the town of Chañaral. Neary and Garcia-Chevesich report a high incidence of cancers and skin, respiratory, and eye diseases that would be associated with exposure to particulate matter as a result of prevailing coastal winds [17] .

The objective of the present study was to evaluate the relationship between lung function and exposure to en- vironmental PM_2.5, among a panel of schoolchildren living near a beach highly contaminated with mine tailings.

We invited 158 children (and their parents) to participate, via meetings to provide information about the study’s purpose and participation requirements. We were able to recruit 119 children, 9 of whom (7.5%) abandoned the study, of which 6 did so before completing the questionnaire, and 3 before starting the measurements. Therefore, 110 children were followed. Figure 3 shows the location of the homes of the participating schoolschildren. The anthropometric and sociodemographic data for the 110 participants are shown in Table 1. The majority of children were male (58.18%), and average age at recruitment was 11.2 years (SD = 2.7). The anthropometric variables showed similar distributions for both sexes, with no significant differences. For both the mother and the father of the children, the most common education level category was 9 to 12 years (58.18% and 48.18%, respectively), followed by the category 8 or fewer years of education. Smoking prevalence of at the time of the questionnaire was similar for the father and mother of the child. Asthma and rhinitis prevalence was 9.1%, and 10.9%, respectively.

The spirometry values for the children are presented in Table 2. Spirometry values increased progressively with age’s groups; there were significant differences for all spirometry values between age groups, as well as between sexes. There were no significant differences between groups according to asthma and rhinitis diagnosis, education level or smoking status of parents.

Table 3 shows the average levels by minute as well as the 24-hour average for the PM and meteorological variables during the study period. The PM_2.5 levels by minute showed a range of 0.01 to 172.5 µg/m³. The meteorological variables showed narrower ranges of variability. There were no extreme temperatures recorded during the study period; furthermore, the relative humidity and barometric pressure were relatively stable and showed no relationship with PM variation. Figure 4 displays a time series for the 24-hour average (daily) PM_2.5 concentrations and wind velocities as well as the 25^th and 75^th percentiles (P25, P75) of the daily measurements for these variables throughout the study period, along with Chilean norms for 24-hour average PM_2.5 concentration. As shown, the variability of the wind velocity was higher during the first 3 months of the study, corresponding with the higher PM levels recorded for the same period. The 24-hour average of PM_2.5 exceeded the Chilean norm of 50 μg/m³ only one occasion.

The chemical composition of the environmental PM₁₀ in Chañaral during the study period was analyzed as part of another study, and therefore we will not go into detail here regarding the procedures and analyses performed to determine these values. Briefly, the average concentrations of metals and metalloids found in the PM₁₀ in Chañaral were, in descending order: Cl > Si > S > Ca > Al > Fe > K > Cu > Mg > Ti > Zn; comparatively, these average levels are higher than those reported in other studies carried out in the central and northern zones of Chile [26] [27] . On the other hand, to PM_2.5 the descending order were S > Na > Cl > Ca > Si > Fe > Al > Mg > K > Cu > Ti > and Zn.

Table 4 shows the regression coefficients and 95% confidence intervals for the average associations between PM_2.5 concentration and lung function in GEE models. We fitted this model including only 105 schoolchildren because 5 subjects abandoned the study during the first weeks of follow-up. There were significant negative associations between PM_2.5 levels and the lung function variables analyzed. The regression coefficients represent the average decrease in lung function values for a 1-unit increase in PM_2.5 concentration. In the unadjusted mo- del, lag12 max PM_2.5 concentration was negatively associated with decreased FEV₁ (β −0.75 ml, 95% CI −1.4,

SD: standard deviation.

−0.03) i.e. for every 1 unit increase in the maximum concentration of 12 hour PM_2.5 decreases the FEV₁ 0.75 ml with a confidence interval between −1.4 and −0.03 ml, which it does not include the value of invalidity (0) allows us to conclude the negative association is significant and not due to chance; lag4 and lag12 avg PM_2.5 were also negatively associated with FVC (β −2.42 ml, 95% CI −4.7, −0.1; and β −5.07 ml, 95% CI −8.9, −1.1, respectively), as were lag4, lag12, and lag24 max PM_2.5 levels (β −1.74 ml, 95% CI −2.7, −0.8; β −1.90 ml, 95% CI −2.8, −1.01; and β −2.01 ml, 95% CI −2.9, −1.03, respectively). In the unadjusted analysis, the only flow value showing a significant negative association with PM was PEF, which was negatively associated with lag24

^*Number of schoolchildren; ^**Number of spirometry tests. SD: standard deviation, FEV₁ (ml): forced expiratory volume during the first second (milliliter), FVC (ml): forced vital capacity (milliliters), PEF (ml/sec): peak expiratory flow (milliliters/seconds), FEF_{25 - 75} (ml/sec): forced expiratory flow 25 - 75 (milliliters/seconds).

^*By minute; ^**By day. SD: standard deviation. P25: 25^th percentile, P50: 50^th percentile, P75: 75^th percentile, min - max: minimum - maximum, PM: particulate matter, µg/m³: micrograms/cubic meters, ˚C: degrees Celsius, m/s: meters/second, mmHg: millimeters of mercury.

^*GEE model adjusted for age, sex, weight, wind speed, ambient temperature; Significant values in bold. FEV₁ (ml): forced expiratory volume in one second (ml), FVC (ml): forced vital capacity (milliliters), PEF (mL/sec): peak expiratory flow (milliliters/second), FEF_{25 - 75} (mL/sec): forced expiratory flow 25 - 75 (milliliters/second). Coef_unadj: unadjusted coefficient, Coef_adj: adjusted coefficient, PM: particulate matter, µg/m³: microgram/cubic meter, P75: 75^th percentile.

max PM_2.5 concentration (β −2.93 ml/sec, 95% CI −5.7, −0.1). After adjusting for age, sex, weight, environmental temperature, and wind speed, only the negative associations between lag12 max and lag24 max PM_2.5 levels and FVC remained significant. Non-significant relationship with lung function was observed when we fitted a model considering the distance of households of schoolchildren to tailings deposit (as a continuous variable and according to categories of strata), the diagnosis of asthma and rhinitis as well as parental smoking were considered (data not shown).

In order to study the short-term effect of environmental exposure to PM on lung function values, we found a negative association between environmental exposure to PM and spirometry values, in an urban population of schoolchildren living near a beach contaminated with mine tailings. This decrease in lung function was especially marked for FVC impairment associated with fine particulate matter exposure (PM_2.5).

In other studies with similar panel designs, the PM value most commonly reported has been the median of 24-h average concentrations during the period studied. For PM_2.5, the median 24-h concentration in this study (12.5 µg/m³) moderately exceeded the value reported by Trenga et al. [11] at 11.2 µg/m³ for a residential area in Seattle, United States and markedly exceeded the value reported by Dales et al. at 6.5 µg/m³ for an area affected by heavy truck traffic in Windsor, Canada [8] ; however, a study performed by Moshammer et al. in a zone exposed to industrial pollution and vehicular traffic in Linz, Austria [6] reported a median value of 15.79 µg/m³, exceeding the value reported in the present study.

Due to the controversy of suggesting that central site measurements may not be representative of individual or residential community exposure, Trenga et al. studied the differences between concentrations measured in the central site and those measured immediately outside personal residences. The author found a strong correlation between the two values (r = 0.77) [11] . This finding is very important, as most studies on the health effects of air pollution rely on central site measurements, including the present study.

The average metal and metalloid concentrations found in the PM₁₀ in Chañaral were generally greater than those reported by other studies in different cities and mining zones in northern Chile and are consistent with the metal measured for this population in different studies [17] [18] [26] [28] . Several studies in animal models suggest that the bioavailable metal transition is one the primary determinants of the acute inflammatory response for both the combustion source and ambient PM samples [2] [29] -[31] . Genotoxic and epigenotoxic effects on human bronchial epithelial cells have been reported due to variable concentrations of transition metals and organic compounds [32] [33] .

Pulmonary development, immune function, and respiratory response to various air pollutants are interrelated via complex multifactorial processes [34] , possibly explaining the high degree of variability for lung function values reported.

The results of the association analysis for PM and lung function are largely consistent with the literature. However, the type of measure reported varies by study. Some authors have reported changes in lung function for each 10 µg/m³ of change in PM or change in interquartile range (IQR) of PM, while other authors have focused on variations in spirometry values as compared to predicted values. Furthermore, the time lags used vary among studies, making it difficult to compare results. PM_2.5 was negatively associated with FEF_{25 - 75} in the study by Trenga et al., carried out on children with asthma, and with FEV₁ and PEF in the study by Moshammer et al., carried out on healthy children in Linz, Austria. Dales et al. only studied FEV₁ values in asthmatic children, finding a negative association with 12-h lag PM_2.5 concentration. In our study, we found negative unadjusted associations between FEV₁ and 12-h lag maximum PM_2.5 concentration and between PEF and 24-h lag maximum PM_2.5 concentration, which is consistent with the findings reported by Moshammer et al. in healthy children. The same authors also found a marked association between average and maximum 4 and 12-h lag PM_2.5 levels with FVC as well as 24-h lag maximum levels with FVC.

Most of the significant associations found with PM_2.5 levels were for FVC. This measure is the maximum capacity of air expelled during a forced expiration and represents a concrete indicator of pulmonary capacity. Decreased FVC indicates a restrictive ventilatory defect. One of the causes described in the literature for this type of limitation is inhalation of organic and inorganic dust. However, the findings in this study represent acute variation in FVC rather than a progressive decrease in function over time. To confirm the long-term effects of exposure, it would be necessary to extend the study to follow the children for several more years. The sample of schoolchildren studied was mostly healthy, and the associations between PM concentration and lung function did not vary according to asthma or rhinitis diagnosis; that is, exposure to increased PM_2.5 levels impairs respiratory function in the short-term regardless of asthma or rhinitis diagnosis.

While changes in lung function as a result of chronic exposure become evident at more advanced ages, the fact that we found associations between short-term PM_2.5 exposure and lung function indicates that schoolchildren in Chañaral are currently affected by the exposure. Further chronic effects may emerge during adulthood. Moreover, given that the PM_2.5 studied contained metallic particles, long-term exposure may have other silent and cumulative effects not only on the respiratory system but also on other organs due to bio-accumulation of heavy metals.

These findings underscore the need for further studies in communities exposed to air pollution from various sources in order to uncover other acute or chronic effects of exposure to pollution from mine tailings.

The present study is the first report of the respiratory health effects of exposure to PM from mine tailings among the inhabitants of Chañaral. Increased PM_2.5 levels associated with toxic metals affect the respiratory function of schoolchildren living in the city and the variation in FVC suggests that these children may be vulnerable to effects of long-term exposure. Measures to control or decrease exposure in this population are needed, and we hope that the evidence reported here will contribute to such efforts.

The authors would like to thank the schoolchildren who took part in the study as well as their parents; authorities in the departments of education, environment, and community health for the Municipality of Chañaral; the staff who carried out the fieldwork; Dr. Benigno Linares for his assistance in the questionnaire’s design; and the Chilean National Fund for Research and Development in Health (FONIS N˚ SA11|2244).

The authors declare that they have no conflicts of interest.