Spirometry Alteration Due to Exposure to Atmospheric Pollutants in Rural Homes in Bogotá, Colombia

This study was aimed to determine the association between the concentrations of particulate matter less than 2.5 microns (PM 2.5 ) , carbon monoxide (CO) and black carbon (BC) within the home and alterations in respiratory function in a population living in rural dwellings in the districts of Usme and Sumapaz in Bogotá, Colombia. Cross-sectional study that recruited people from the rural area of Bogotá. In total, 68 participants were recruited by non-random sampling. Indoor air monitoring of PM 2.5 , BC and CO were carried out. Additionally, a pulmonary function evaluation was performed using spirometry. Pulmonary function variables were included in a multiple linear regression by successive steps, adjusted in terms of the main variables described as modiers of the values in spirometry, these are: age, height, sex and history of tobacco use. Assumptions of multicollinearity and randomization of variances in the residuals were evaluated. Negative associations were found between spirometry parameters and the concentration of indoor air pollutants. For FEV1 / FVC and theoretical FEV1 there was a statistically signicant association with the concentration of carbon monoxide (p = 0.003 and 0.019). Environmental concentrations were higher in homes where biomass is used for cooking these differences were only statistically signicant for BC and CO (p = 0.008 and 0.03 respectively). Carbon monoxide concentrations in rural homes caused alterations in respiratory function.


Introduction
In the world, approximately 2.8 billion people use solid fuels for cooking (wood, animal manure, agricultural waste, vegetal charcoal and charcoal), mainly among rural and low-income populations (Bonjour, et al., 2013). The combustion of this type of solid energy sources leads to high levels of indoor air pollution (IAP), which depends on the system used to cook the food. . Health effects of indoor air pollution include increased blood pressure, dyspnea, childhood pneumonia, lung cancer, low birth weight, and cardiovascular disease (Pratiti et al., 2020).
In this way, IAP problems are recognized as important risk factors for human health. Indoor air pollution has been associated with 3.5 million deaths and it causes 4.5% of the lost years of life due to disability in the world during 2010 (Lim, et al., 2012). A systematic review with meta-analysis shows that there were positive associations between the use of solid fuels and Chronic Obstructive Pulmonary Disease (COPD) (OR = 2.80, 95% CI 1.85 to 4.0) and chronic bronchitis (OR = 2.32, 95% CI 1.92 to 2.80), also show that exposure to wood smoke, while doing housework, presents a greater risk of developing COPD and chronic bronchitis than other fuels. Therefore, efforts are recommended to reduce the use of this type of solid fuel (Kurmi et al., 2010). In this sense, IAP is a global public health threat that requires efforts from research and policy formulation (Alvis-Guzmán & De la Hoz-Restrepo, 2008).
Regarding Colombia, the national survey of quality of life carried out by the National Administrative Department of Statistics (DANE) in 2018, shows that 21.7% of the population uses rewood as fuel for cooking (DANE, 2018). In a study conducted on a total of 2.1 million inhabitants living in extreme poverty in Colombia, the use of solid fuels for cooking food was 20 times higher in the rural population and almost three times higher in indigenous ethnic groups. Furthermore, the authors report an increase in the prevalence of cardiorespiratory limitations, as well as vision limitation (Soto-Moreno & Ballester-Díez, 2013). According to the National Planning Department (DNP), morbidity related to indoor air pollution was estimated at more than one million cases per year and more than two thousand deaths from this cause nationwide (National Planning Department, 2018).
Carbon monoxide (CO), black carbon (BC), and particulate matter (PM 2.5 ) have been used as indicators of the components of indoor air that affect people's health. Particulate matter and carbon monoxide have been used in various studies to determine exposure to indoor and outdoor air pollution and have been associated with alterations in respiratory function measured by spirometry or respiratory symptoms (Steinvil et al., 2008;Steinvil et al., 2009;Pope et al., 2015). In a cohort study that evaluated mean environmental concentrations of various pollutants, including black carbon, it was signi cantly associated with increases in the proportion of emphysema during the 10-year follow-up (Wang et al., 2019). Likewise, the effects of black carbon on health problems can be seven or eight times greater than those related to PM 2.5 . In any case, the concentrations of these two pollutants directly impact the life expectancy of the exposed population (Janssen et al., 2011).
AIP can affect all population groups, as well as those vulnerable due to their health or age, such as people suffering from heart, respiratory, cognitive, physical or other conditions which might make them stay most of the time at home, thus increasing their exhibition (Bruce, Perez-Padilla, & Albalak, 2000). However, in Colombia, current scienti c studies are mostly focused on outdoor air quality, without recognizing the possible sources of indoor pollutant emissions, or monitoring the concentrations inside the house to which those who remain for longer time are exposed, such as women, children, and older adults. For these reasons, the objective of this study is to determine the association between concentrations of PM 2.5 , CO and intradomiciliary BC and alterations in respiratory function in a population of inhabitants of rural dwellings in the districts of Usme and Sumapaz in Bogotá.

Study Area
The research was carried out in the rural territories of Usme and Sumapaz (Bogotá, Colombia) located between 2,600 and 3,100 meters above sea level, it is located 31 km from the urban area of Bogotá, with a temperature that ranges between 4 and 19 °C (District Planning Secretariat, 2020). All the monitored houses were built in brick and cement, and because it is a rural area, the houses were scattered, that is, they were not close to each other, forming a neighborhood. It is important to highlight that, due to the temperature conditions, the use of stoves represents a heating alternative for rural homes, which in the face of spaces with limited ventilation contributes to the accumulation of pollutants.

Population
The research was developed from a cross-sectional study that recruited inhabitants of rural houses. The inclusion criteria included people who lived in rural homes in the towns of Usme and Sumapaz, it was established that they had lived for at least one year in the rural region. For the evaluation of pulmonary function, people who had respiratory symptoms such as coughing, expectoration or dyspnea, health conditions that prevented the performance of spirometry, as well as people who were unable to perform the maneuver for the examination were excluded.
Due to the geographic dispersion of the region and the limitation of resources of the study initially, it was established to calculate a sample size of 50 individuals with a con dence of 95% and a prevalence of COPD in rural areas of 15% and an error of 10%; however, the study resources allowed obtaining a total of 68 participants with an estimate of the standard error of 8%.
Regarding the measurement of air quality, in some homes the evaluation equipment was manipulated or disconnected from the electrical sources by the participants, which caused loss in the evaluations of the environmental variables for 19 people. Environmental and lung function assessments were carried out over a 4-month period between November 2018 and February 2019.

Air pollutant monitoring
In the indoor air quality monitoring, levels of particulate matter of less than 2.5 microns, as well as black carbon and carbon monoxide were measured. In this study, the measurement site was a room or living room that was always adjoining the kitchen. The measurement was carried out for at least 48 continuous hours for PM 2.5 and CO, and 12 hours for black carbon in each of the houses.
Particulate matter: the Dusttrak II 8530 aerosol pollutant monitor was used to monitor particulate matter. This instrument uses optical dispersion to deliver real-time concentrations of PM 2.5 . The detection range is 0.001 to 400 mg / m3 with a resolution greater than 0.001 mg/m 3 or 1% of the reading. Carbon monoxide: Carbon monoxide was determined with the Q-trak Model 7575 Indoor Air Quality Monitor; It is a multi-function electrochemical sensor, which can measure CO, carbon dioxide (CO2), temperature and relative humidity (RH). It uses an electrochemical sensor to measure the CO concentration range from 0 to 500 ppm with an accuracy of ± 3% of reading or 3 ppm (whichever is greater), a non-dispersive infrared (NDIR) sensor to determine the CO2 concentration between 0 and 5000 pm with an accuracy of ± 3% of reading or ± 50 ppm (whichever is greater), a thermistor to measure the temperature between 0 and 60 °C with an accuracy of ± 0.5 °C and a thin-lm capacitive element to measure relative humidity between 5 and 95% with an accuracy of ± 3% (TSI, 2018).  Table 1 summarizes the information on the monitoring characteristics and the correction factors used for the PM 2.5 and BC readings.
Additionally, to obtain information that would allow an effective results analysis to be carried out, a question format was used, which is established based on what was done by Rumchev (2017) and considers both the characteristics of the dwelling and sociodemographic data additional to the one which had been already collected for the project (Rumchev, K. et al. (2017). Each home receives a format to keep track of the stove they use in order to contrast with the data recorded by the monitoring equipment. as well as to record emission activities, such as external events or use of the stove to produce products for commercial purposes.

Lung function assessment
After 48 hours of environmental monitoring, an evaluation of lung function was carried out. To do it, in each house, a table was located in a closed room where the spirometer, disposable mouthpieces, antibacterial lters were placed. Likewise, a computer to register the evaluation information was set next to it. To guarantee reproducibility and avoid biases in the collection of information, the spirometries were performed by a physiotherapist trained by an expert from the National University of Colombia. The training included handling, installation and performance tests of the equipment, time management to carry out the test, and voice commands for the motivation of the participants and uni cation of the test approval criteria. 5 piloting spirometries were performed in the Human Body Movement Department of the National University.
In order to carry out the spirometry, the professional explained the test to each participant. Then, there was a mock test which was corrected and provided feedback. Spirometry was performed following the protocol established by Miller et al (2005), who are part of the guidelines of the American Thorax Association and the European Respiratory Society. The equipment used was a SpiroBank G brand spirometer and Winspro software; the data was recorded in digital medium. The test was considered adequate after performing and recording 3 maneuvers which had met the acceptability criteria (correct start, stable plateau, plotting the curves without artifacts, slow and asymptotic termination and adequate duration), and repeatability criteria. The result of the maneuver that presented the least variability was selected (Miller et al., 2005).
In the same room, the physiotherapist evaluated the height with a height rod on a at surface. Using a previously calibrated scale, body weight was determined, and oxygen saturation and heart rate were measured with a pulse oximeter. (2) obstructive defect, indicated by a decrease in the FEV1/FVC ratio (less than 70%), a FEV1 <80% and an FVC ≥LLN; (3) combined defect, indicated by FVC <LLN, FEV1 <LLN and the normal FEV1/FVC ratio, increased or decreased, depending on the pattern that predominates the most (Pellegrino, et al., 2005; Romero de Ávila Cabezón, and others, 2013).

Statistical Analysis
Statistical analysis of the data was performed for each of the pollutants under study, so that it would be possible to show the averages of the evaluations during the observed periods. In the evaluation variables of lung function, a description was made by means and standard deviation. For continuous variables, tests were performed to determine the normality of the distribution (Lilliefors, 1967). To make comparisons between the concentrations of pollutants and fuels used in cooking, non-parametric tests (Mann-Whitney test) were performed, with a signi cance p value of 0.05. To establish an association between the lung function variables and the mean concentrations of the pollutants, a bivariate correlation matrix was made, which used the Sperman test for contrast, with signi cance values of 0.05.
Pulmonary function variables were included in a multiple linear regression by successive steps, adjusted for the main variables described as modi ers of the spirometry values, these are age, height, sex and history of tobacco use, multicollinearity assumptions were evaluated, randomization of variances in the residuals and the absence of in uential individuals by extreme values that modify the model (Cook's distance). The data were analyzed in the SPSS software.
For the development of the project, the informed consent document was registered by a person of legal age living in each dwelling and was endorsed by the ethics committee of the EAN University.

Population and emission sources
The population sample is mainly female with 66.2%. They presented a low prevalence of tobacco consumption with only 10.3%, with an average age of 42 years. The 68 people selected lived in 33 independent houses and in 64% of the homes there was at least one person with a disability. Regarding the stoves found in homes that consume wood as fuel, 98% were replaces, and 2% were a traditional type of stove ( Figure 2). In the case of those that use gas as fuel, 100% of the households used a stove with a cylinder valve. The main sources of PM 2.5 , BC and CO emissions in the studied households are the use of solid fuel as rewood for cooking food and external burning of waste. When they occur, there are peaks in concentrations as evidenced in were obtained in homes where they cook with gas and the highest where they use rewood. As expected, carbon monoxide levels were not recorded in homes where gas is used; the detection limit of the equipment is 0 to 500 ppm. For households where rewood was used, the 24-hour average of carbon monoxide concentrations was between 0.34 ppm and 1.31 ppm.
Black carbon concentrations were always correlated with those of PM 2.5 , the evolution of pollutant concentrations during the measurement period. As an example, Figure 3 shows a home where they used rewood and Figure 4 shows the concentrations where they used gas. The 12h average concentrations of black carbon measured were between 0.64 μg/m 3 and 51.2 μg/m 3 . The concentrations were compared between the types of fuel used for cooking food inside homes, nding higher concentrations of all pollutants in homes where cooking with biomass, these differences were only statistically signi cant for BC and CO (p= 0.008 and 0.03, respectively).

Pulmonary function
The study evaluated the interaction between the indoor air quality pollutants in homes and the respiratory variables (Table 3). Pulmonary function analyses revealed that the population had some type of spirometry alteration in 10.3% of the sample (n = 7). 7.4% of the people had a restriction pattern (n = 5) and 1.5% had an obstructive pattern (n = 1) and a mixed pattern (n = 1) ( Table 2). Negative correlations were found between spirometry parameters and the concentration of indoor air pollutants. For FEV1/FVC, negative associations were found with all the pollutants evaluated, but there was only a statistically signi cant association with the concentration of carbon monoxide (p = 0.003). For the theoretical FEV1 there was also a negative association with monoxide (p = 0.019) ( Table 4).
A negative association was found between carbon monoxide concentrations and FEV1/FVC, theoretical FEV1 and PEF values, in the adjusted and unadjusted measurements ( Table 4). The successive steps model excluded low in ation values of the variance. For CO, however, when evaluating the adjusted residual values with respect to the predictor values we found a random pattern in the graph, when measuring the cook distance, we found only one case with extreme mean in the model, but it was the dwelling that used the most rewood for cooking food. The model was not statistically signi cant for the vital capacity FVC, FET.  Table 5 compares the CO concentrations measured in this study and the ambient concentrations reported in the Bogotá Air Quality Monitoring Network. It is evident that the 8-hour average concentrations of CO in homes where they use rewood can reach concentrations higher than those of the station closest to the monitoring site (Tunal Station) and is also higher than the highest average registered in the monitoring network (Puente Aranda) (District Secretary for the Environment, 2019). This reveals the high concentrations within households in this rural area and the importance of researching possible negative effects on people's health and strategies to improve air quality.

Discussion
Other investigations against the association between CO and respiratory involvement in a rural area of Ghana in pregnant women (n = 840) reported a positive association between CO exposure by 1 ppm increase and a composite score of respiratory symptoms such as current cough (that was > 5 days), wheezing and/or dyspnea (OR: 1.2, p = 0.03) and CO was also positively associated with wheezing (OR: 1.3, p = 0.05), as well as with phlegm (OR: 1.2, p = 0.08) and visits to the clinic for respiratory infection were reported within the 4 weeks prior to the study (OR: 1.2, p = 0.09) (Van Vliet et al, 2019).
Likewise, other studies have found prevalence regarding the use of biomass and the presence of symptoms such as headache and/or recent respiratory symptoms (cough, wheezing, shortness of breath or runny nose) with the use of wood stoves and/or coal with or without a closed combustion chamber and chimney, taking into account that groups of people who used stoves without any protection had marginally higher prevalence. Biomarkers for indoor air pollution exposure were also positively associated with red, itchy and wet eyes and eye pain when cooking (Li, Z. et al, 2016). Furthermore, Li, S et al., On the other hand, in the crude and adjusted measurements, the CO altered several spirometry indicators in asymptomatic patients. In studies evaluating exposure to monoxide, no statistically signi cant associations were found. However, in a population of non-smoking women exposed to wood smoke, CO concentrations in exhaled breath were negatively associated with lower FEV1 (

Conclusions
Although these ndings allow us to collect evidence of the impact of gases from the use of solid fuels on health, we must recognize that this research has numerous limitations, rstly, associated with its cross-sectional design that limits the monitoring of the population, as well as the identi cation of cumulative impacts to exposure. The geographic context of the Colombian rurality makes populations live in conditions of great dispersion and of di cult access. For this reason, individuals were selected by means of convenience. Additionally, the losses of environmental evaluation further limited the number of evaluations, thus limiting the external validity of the results, which must be considered when extrapolating these data to other populations.
Likewise, this study found associations between IAP and spirometry alterations. It is very important for the health of the rural population of Colombia to carry out studies that allow examining the changes in the incidence of chronic lung diseases and if it is possible to carry out interventions such as health policy to reduce exposure to IAP due to the use of solid fuel, and it will require larger sample size and/or duration of follow-up.