Impact of fine particulate matter and toxic gases on the health of school children in Dhaka, Bangladesh

Background. Air pollution exposure has a detrimental effect on children who spend more than 17% of their weekdays inside a school building. The purpose of this study is to look into the effects of particulate matter (PM) and toxic gases on health of the school children. Between April and November 2018, samples were collected in real time from ten different schools (both indoor and outdoor) over four hours on two consecutive days at each school. During the first two hours, when students were present in the classroom, measurements were conducted inside the classroom. After that the measurements were conducted outside the classroom but within the school premises - when students were playing on the playground or eating breakfast outside of classroom. Method. To evaluate the impact of air pollution, 250 students (on average 20 students from each school) aged from 9 to 12 were selected from ten schools. Automatic monitors (AEROCET 531S, USA) were employed to measure PM1.0, PM2.5, and PM10 concentrations. NO2, TVOC, and CO2 concentrations were measured using an AEROQUAL (500S, New Zealand), and the respiratory rate is measured by BSMI Peak Flow Meter (Made: BSMI, Origin: China). Monitors were placed at about 2.0 meters above the floor at breathing height and no student wore the sensors. The ANOVA test was conducted to see the statistical significance between air quality parameters and peak flow meter readings. Results. The mean ± standard deviation of PM1.0, PM2.5, and PM10 concentrations were 19.1 ± 3.6, 34.2 ± 10.1, and 131.3 ± 58.6 μgm−3, respectively. PM2.5 and PM10 concentrations exceeded WHO standards (15 and 45 μgm−3 of 24 h) by 2.3 and 2.9 times. The highest concentrations of toxic gases were found on school campuses where vehicle densities (measured manually) were high. The mean Hazard Quotient (HQ) for PM10 (2.5 ± 2.2 indoor; 3.6 ± 2.6 outdoor) and PM2.5 (1.8 ± 0.8 indoor; 1.9 ± 1.0 outdoor) among all participating students was >1 indicating an unacceptable risk for human health. Lung function associated with the PEF value has a negative correlation with PM1.0 and PM2.5 concentrations in most cases. Conclusions. The findings of this study are useful in gaining a general understanding of the school environment in Dhaka. It aimed to understand how children were personally exposed in school and to develop effective control strategies to mitigate negative effects.


Introduction
Long-term exposure to ambient air pollution has been associated with a range of unfavorable health effects, including mortality, cancer, and cardiovascular disease (CVD) [1]. PM with an aerodynamic diameter 10 μm (PM 10 ) can be deposited within the respiratory tract including those PM 2.5 (Particulate matter with an arrhythmias and convulsions), anesthetic (substantially reduced CNS activity), and fatal (severe acidosis and anoxia). The displacement of O 2 by CO 2 adds greatly to toxicity in high-level CO 2 exposure.
Children's lung development has been linked to prolonged exposure to PM 2.5 [11]. There is growing evidence that, in addition to age, gender, and height inequalities for PEF, there are variances in lung function across people of different races [37]. Pulmonary function, which is assessed in terms of lung volume changes can be useful in determining the effects of acute PM 2.5 exposure on the lungs [38]. Therefore, finding information on childhood lung function is crucial because the following evaluation and diagnosis of respiratory health. A 10 μgm −3 increase in acute PM 10 exposure was linked to a 0.19 l min −1 (95% CI: 0.30, 0.09) change in PEF, according to an assessment of 22 effect estimates from 15 studies [39]. According to Wu et al [40], in Beijing, where air pollution levels were high, healthy university students' nighttime PEF readings decreased with the increasing concentration of PM 2.5 . Furthermore, Rice et al [41] showed that short-term exposure to relatively low ambient PM 2.5 levels lowered the PEF values in adult men and women.
Bangladesh is the number one country in the World for the human death due to the environmental problem. It is also the topmost polluted country in the world for PM 2.5 concentrations for last five years consecutively (2018-2022). The country has been experiencing massive population growth as well as rapid industrial, commercial, infrastructure development, and also changing the basis of economy from agriculture to industry. As a result, the major components of the city environment are adversely affected resulting in continuous deterioration of air quality. The Dhaka city is also the top listed polluted capital in the World. The elderly and children have suffered the most because of the severe air pollution. Therefore, we have chosen the school children in Dhaka as our target subjects in this study.
However, we have focused on the overall school environment, collecting data on particulate matter, toxic gases, and the pulmonary function of the students. As far as we know, this is the first of its kind in Bangladesh. This study aims to characterize the exposure to a variety of particulate matter and toxic gases in school's environment in Dhaka city and their impact on children's health by comparing measured concentrations with relevant standards and suggesting ways to reduce the exposure of school children to unacceptable pollutants.

Materials and method 2.1. Study area and sampling sites
In total, samples were obtained from ten schools in Dhaka. Kamlapur High School (S1), New Model Bohumukhi High School (S2), University Laboratory High School (S3), Khilgaon Girls School and College (S4), Motijheel Model High School (S5), Maniknagar Model High School (S6), Maniknagar Model High School (S7), Nawabkatra Government High School (S8), Nababpur Government High School (S9) and Badda High School (S10) were the 10 schools. The schools were selected from a variety of sites in Dhaka city (figure 1). These were selected on the basis of overcrowded areas, residential areas associated with low vehicle traffic, dense traffic area, and less populated areas surrounded by plants and lakes. Each school occupies the entire structure, and each floor is solely dedicated to schoolwork. From April to November 2018, all of the schools were subjected to twoday indoor and outdoor air quality tests, as well as PEF measurements.  All of the schools opened their doors between 7:00 and 8:00 a.m., and in the meanwhile, basic maintenance was completed. The school buildings in all cases are naturally ventilated via openable windows and doors, and they have white painted concrete walls, chalkboards, and dusters, as well as concrete cement flooring. The number of students in each classroom was between 35 and 40. The arrangement of classrooms differs from school to school. In essence, most classrooms were located on the 1st floor, but others also comprised the 2nd, 3rd, and 4th floors. Table 1 contains information about each sampling site as well as site-specific parameters.

Sampling periods
A total of ten schools were selected for monitoring of PM 1.0 , PM 2.5 , PM 10 , NO 2 , CO 2, and TVOC with the aim to cover the maximum areas in Dhaka city for acquiring the concentration gradients in the ambient air pollutants. As the study is based on school children, students of class 5 to class 7 were selected, aged between 9-12 years. Each classroom has approximately the same number of students. On average, 20 to 25 students were selected from each school to collect the peak flow through a peak flow meter. Particulate matter and toxic gases monitoring was conducted from April 2018 to October 2018, as each school is very active during this period and also this time period gives an idea of the average weather forecast for the whole year (March-May: Pre-monsoon; June-September: Monsoon; October-November: Post-monsoon). As part of the study activities, sampling was performed inside and outside of the selected classrooms. The morning recess generally takes place between 10:00 am and 10:30 am. Nearly all the schools had a small kitchen where most of the time tea and coffee were prepared for the teachers and a canteen in a corner of the schoolyard where snacks were sold to the students. In addition, the classroom and playground were cleaned using the same schedule (e.g., early morning and late afternoon on weekdays). Children between the ages of 9 and 12 spend their time outdoors on average 1 h of their 5 h school day for recess and other playground activities.

Measurement and instrumentation
The measurements were taken at each site from 07:30 a.m. to 12:00 p.m during the school day on two consecutive days. This time period was selected because the student movement was more in the morning when classes began and in the middle of the day when schools were going to end. The mass concentration of particles (PM 1.0 , PM 2.5, and PM 10 ) was monitored by AEROCET 531S (Met One Instrument, Washington, USA). The AEROCET 531S counts and measures particles in six distinct size ranges then convert count data to mass measurements (μgm −3 ) with a ±10% accuracy using a proprietary algorithm. It collected data on a minute-byminute basis. The device was compared to a typical filter-based device called SIBATA (model: 090860-504, Saitama, Japan), and the results were within 10% of each other. The AEROCET was placed in the classroom facing the blackboard at the top of the bench. For outdoor sampling, the device was continuously moved from one place to another to cover the whole playground area. Each site's ambient temperature and relative humidity were assessed simultaneously with particulate matter measurements. For the measurement of toxic gas, samplers were placed inside the classroom, next to the classroom on the balcony, playground, and canteen near the cooking stoves. Air gaseous pollutants CO 2 , NO 2 , and TVOC were continuously monitored every 15 min from 7:30 a.m. to 12:00 p.m. using the AEROQUAL 500 SERIES (AEROQUAL Ltd Auckland, New Zealand). We then calculate an hourly average of them (the first two hours in the classroom and the remainder of the time in the playground). Each of the contaminants has its own set of sensors. For CO 2 measurement, AEROQUAL employs a nondispersive infrared approach with a precision of 10 ppm plus 5%. NO 2 concentrations are detected with a precision of 0.02 ppm plus 10% using gas-sensitive electrochemical sensors. It employs a photoionization detector for TVOC and has a precision of 0.2 ppm plus 10%. Both of the monitors were placed at about 2.0 meters above the floor at breathing height and no student wore the sensors. The activity of the lungs of the students was measured two times each day. 1st measurement was taken inside the classroom and the 2nd one was taken at the playground after the recess time. At first, all the students were lined up. One of our research students measured the PEF of each student individually using a Peak Expiratory Flow Meter and noted any physical issues they had. We collected the PEF value from each student four times: before they entered their class, during the class gap, before recess, and after they returned from the playground. They attempted multiple PEF values each time to get the most accurate result possible. The integrated scale measures of PEF meter are liters per minute from 100 to 850. It consists of three peak flow zones with a color code green, yellow and red. If the peak flow rate is between 80% and 100%, it shows that the asthma is under control and in the steady green zone. When the peak flow rate ranges from 50% to 80%, it indicates a severe asthma attack that mimics the caution yellow zone. If the peak flow rate is less than 50%, it is a sign of a medical emergency with a hazardous red zone. However, ANOVA test was conducted to see the statistical significance between air quality parameters and peak flow meter values.

Relationship between indoor and outdoor particles (I/O)
The I/O ratio is widely used to represent the relationship between indoor and outdoor particles [42].
The indoor and outdoor particle concentrations are represented by C in and C out , respectively.

Risk assessment for health
Human health risk assessment refers to the process of determining the likelihood of adverse health effects from contaminated environment exposure. The chronic daily intake (CDI) was computed using equation (2), and the exposure was assessed by measuring how much PM 10 and PM 2.5 impacted human health [43].
Where, C is the contaminant concentration (μgm −3 ), IR is the average inhalation rate (m 3 /day), ET is the exposure time (hours/ day), ED is the exposure duration for scenario (years), EF is the exposure frequency, BW is the body weight (kg), and AT is the average time for a lifetime (days). The hazard quotient (HQ) was calculated by using equation (3).
Where, Rfd is the inhalation reference dose (μg/kg/day) sought from WHO recommended values. To calculate the senses of control, the non-hazard level is HQ < 1, and the hazard level is HQ > 1 [43].

Temporal trends of particulate matter concentrations
The study took place over six months (April 2018 to October 2018) in indoor and outdoor schools in Dhaka city, Bangladesh.

Indoor and outdoor sources contribution to PM
Indoor/outdoor ratios (I/O) have been used for a long time to assess the difference between indoor and outdoor concentrations as an indicator of indoor sources [44]. Figure 3(a) shows the average I/O ratio of PM varies from 0.1 to 1.8. It observed that there were in total 4 schools (S3, S4, S5, and S9) that gave I/O ratio less than 1 for PM 10 , whereas there are in total 3 schools (S5, S6, S9) which showed an I/O ratio less than 1 for PM 1.0 and PM 2.5 . Due to the location of these schools, their outdoor air is more polluted in comparison with the indoor air. The I/O ratios of PM 10 were highest for S2 and S10 with the values of 1.8 and 1.3 whereas Badda high school (S10) had the highest I/O ratio of PM 2.5 with the range of 1.3-2.6. Figures 3(a), (b), (c) pointed a weak relationship between indoor and outdoor PM concentrations. The strongest correlation between PM 1.0 , PM 2.5, and PM 10 was reported in PM 1.0 (R 2 = 0.372), indicating that outdoor concentrations can only explain 37% of the variation in indoor concentrations. The lowest correlation found in PM 2.5 (R 2 = 0.007) and PM 10 (R 2 = 0.021) indicate that the indoor and outdoor PM concentration was quite independent of each other. Because of their smaller size and lower mass, PM 1.0 particles can float upwards or disperse further, but they can also fit through unsealed gaps in windows and doors, among other things. As a result, most of them can enter the classroom from the school playground causing an impact on indoor particle concentration. For PM 10 , the greater source of PM 10 in indoor environment is caused due to the number of students, poor ventilation system, dust in shoes, particle from chalk and other activities in the classroom.
3.3. Temporal trends of NO 2 , CO 2, and TVOC concentrations NO 2 , CO 2, and TVOC levels were measured on an hourly basis in selected schools in Dhaka city ( figure 4). All the gases are collected for four hours including the first two hours in the indoor environment and the remaining hours in the outdoor environment. The NO 2 concentration ranged from 0.061 to 0.102 ppm. It had been seen that the concentration of NO 2 was quite high in the outdoor environment (9:30 to 11:30 am) of S6 with the value of 0.122 ppm, because of the presence of the cooking stove beside the playground. The lowest concentration was observed at S4 with the average value of 0.061 ppm, which is placed in a residential area related to a very low amount of traffic.
The drop in TVOC concentration table B1 (Appendix-B) during the morning hours at all sites could be attributed to the opening of floor doors and windows for various causes. Because of the temporary increase in ventilation, the TVOC concentration inside the floor may have been diluted. The TVOC concentration was significantly higher in S4 (1439.250 ppm) than in other sites, which could be attributed to the painting of a new building adjacent to the playground during the sampling period. The following table also showed that the indoor concentration of TVOC was very high at S7 with the average value 1994 ppm. This school was placed inside a busy lane. Many tea stalls also welding shops were situated around the school area. The classroom of the school was placed on the 1st floor where TVOC concentration was measured. Thus, it was easily affected by the adjacent tea stalls and welding shop attributed to the high value of TVOC.
The measurements revealed that CO 2 concentrations increased in the early morning and decreased before lunch. The highest concentrations are caused by insufficient ventilation (closed windows) in the morning (1135.0 ppm). Following the start of class, the windows were opened, and CO 2 concentrations began to fall until early or late afternoon when they settled near the lowest amount (550.0 ppm).

The relationship between fine particulate matter and peak flow rate
The baseline survey targeted around 250 students (115 boys and 135 girls) of class 5 to class 7 around 9-12 years old of 10 schools in Dhaka City (table 2). Children were randomly selected from each classroom to understand the average health difficulties of the students while avoiding a group of students with a particular illness. All students in these classes were given a structured questionnaire (age, gender, health issues) and told to fill it out with their parents and submit it to their respective class teachers. All of the students were instructed on how to apply Peak Flow Meter. During sampling, their health conditions are also collected. It was observed that most of the students suffered from cough and respiratory problems. The high concentration of particulate matter in every school is mainly responsible for this issue. The status of other health problems was given in table C1 (Appendix-C). From the ages of 9 to 12 years, there was no significant difference in body height and weight between boys and girls. The selection of the sampling location ensures that it completely encompasses the entire city of Dhaka. Figure 5 shows the link between particulate matter and lung function of students. According to the associations between indoor-outdoor four-hour averages of PM 1.0 and PM 2.5 concentrations and lung function measures among school occupants, it was observed that in most schools, the average value of the Peak Expiratory Flow Meter is negatively associated with increasing exposure. The average PEF rate for boys and girls in each school is listed in table E.1 (Appendix-E). Also using linear regression figure D1 (Appendix-D), we found that the association between PEF and PM 1.0 , PM 2.5 data is very weak. The link between particulate matter and peak expiratory flow rate is not statistically significant, according to the results of the ANOVA test, which showed that the p-values for PM 1.0 (0.53) and PM 2.5 (0.19) are more than 0.05.

Hazard quotient assessment and associated health risks
The Hazard Quotient (HQ) calculated for school children at different locations in Dhaka city, Bangladesh was shown in figure 6. The results of 250 participants studying in schools were collected to further determination of health risks. The ages of subjects enrolled were between 9 to 12 years including (60.5%) for girls whereas the rest of the students are boys. Details of average body weight are obtained from [45]. The Exposure frequency days yearly were estimated as the active period that students have to attend school. Data used in this calculation are tabulated in table S4. The mean HQ for PM 10 (2.5 ± 2.2 for indoor and 3.6 ± 2.6 for outdoor) and PM 2.5 (1.8 ± 0.8 for indoor and 1.9 ± 1.0 for outdoor) among all participants was >1, indicating a risk to human health that is unacceptable.

Discussion
We collected the particulate matter data from ten schools in Dhaka city and the observations indicated that indoor and outdoor PM 10 concentrations in majority of the schools were above the WHO 2021 recommended standard value of 45 μgm −3 over a 24 h period for PM 10 . With the exception of S6, the indoor PM 2.5 concentrations of all schools did not meet the WHO 24 h PM 2.5 (15 μgm −3 ) recommended value, WHO 2021. The outdoor concentrations of PM 1.0 , PM 2.5, and PM 10 were low at S10, which is adjacent to a lake, with limited traffic. In addition, the highest concentration of PM 1.0 was observed at S7 which is in an area where local tea stalls are placed  beside the school playground. In addition, Mugda bishow road, one of the busiest places, had the greatest indoor concentration of PM 2.5 and PM 10 . The highest outdoor concentration of PM 10 is at Motijheel (S5) which is located in a residential area, but there was a large amount of dust in the school playground, and also the area of the playground is bounded by buildings rather than an open space. S1 and S9 exhibited a higher outdoor PM 10 concentration due to the high number of vehicles. The average outdoor concentration in most schools was higher than the average indoor concentration. These values are also higher than those reported in school-based research in Texas [46], where the outdoor concentration was 13.4 μgm −3 , and in Sweden [47] where the outdoor concentration was 9.7 μgm 3 . The I/O ratios of PM were determined to quantify the impact of outdoor air and indoor sources on indoor air quality. The I/O ratio can vary significantly depending on factors such as location, building design, and inhabitant activities [48]. The I/O ratio varies from site to site depending on many influencing factors such as indoor source, ventilation pattern, different household activities, penetration factor, particle deposition rate, and outdoor concentration. The I/O ratios of PM 10 were highest in S2 and S10, which used chalk duster on the blackboard to instruct their students in the classroom, while the rest of the schools used black marker on the white board to instruct their students in the classroom. The majority of them have more students in the classroom and a poor ventilation system, both of which affect the concentration of particles inside. In contrast, the I/O ratio of PM 1.0 , PM 2.5, and PM 10 are quite low for S5, S6, and S9. Because interior PM levels were more likely to be influenced by indoor sources such as the presence of students and the intensity of their indoor activities, it was discovered that the number of students at these schools was relatively low in comparison to the classroom size. S10 had the highest I/O ratio of PM 2.5 which also indicate that the outdoor PM 2.5 concentration of the school is quite low in comparison with the indoor PM 2.5 concentrations. Because the school was built beside a lake, there was less activities or emission of particulate matter [49] that helped to improve the outdoor environment. Ventilation and infiltration have a role in the transfer of contaminants from the outdoors to the inside environment. Air contaminants from the outdoors can enter the inside environment in a variety of ways, diluting or accumulating depending on the ventilation condition.
Both outdoor and indoor factors may have an impact on the increasing NO 2 concentration. Previous research has found that NO 2 concentrations in schools in urban, industrial, and rural areas of Central and Southern Spain were as follows: In rural areas, 0.004 ppm for kindergarten and 0.005 ppm for primary classrooms; in urban areas, 0.021 ppm for kindergarten and 0.015 ppm for primary classrooms; and in industrial areas, 0.013 ppm for kindergarten and 0.014 ppm for primary classrooms. The median NO 2 values in the outdoor environment, on the other hand, were 0.001 ppm, 0.007 ppm, and 0.005 ppm, respectively, for rural, urban, and industrial areas [50]. In schools of Dhaka city, the gas stoves at the canteen, chalk dust, and road traffic are the main source of NO 2 . Concentrations of NO 2 are varied due to the daily traffic pattern. Traffic densities were measured manually. Previous studied also showed that traffic density has very high correlation with atmospheric particulate pollution [51]. It had been seen that the concentration of NO 2 was quite high in the outdoor environment (9:30 to 11:30 am) of S6 because of the presence of the cooking stove beside the playground. The lowest concentration was observed at S4 which is placed in a residential area related to a very low amount of traffic. It is well known that when paint solvents are applied freshly, they emit VOCs into the air [34]. The majority of urethane coating and paint solvents are bonded to urethane formaldehyde, causing formaldehyde to be released into the atmosphere and causing pollution [52]. . HQs associated with PM 2.5 and PM 10 at the ten schools for 9-12 age students.
Because of their smaller size, particulate materials (PM 1.0 and PM 2.5 ) have a greater surface area-to-mass ratio of the tissue in which they can move, as well as a high probability of contact, oxidative stress, and inflammation, which occurs in extrapulmonary organs [53]. PEF levels are used to assess how effectively the lungs are functioning and responding to treatment. Furthermore, several students experienced cold-related health issues, implying that there is an effect of air pollution on their breathing. The child specialist would be preferred to provide any type of treatment based on the child's health condition, while we only recommend ways to improve the school environment. Notably, the current findings revealed that inhalation exposure to various outdoor air pollutants, as well as indoor air pollutants, poses considerable danger. The weak linear regression value of the association between PEF and PM 1.0 , and PM 2.5 data means that the lung function is quite good when the PM exposure rate is low. It also found that the average PEF value is low and indicates that lung function decreases when the particulate concentration is high. Chen et al [54] found that higher exposure to PM led to reduced lung function. The level of physical activity during exposure and the individuals' prior exposure to traffic-related air pollution both affect the relationships between different pollutant exposures and respiratory measurements [55]. The literature contains a variety of reference values that vary by demographic, ethnic group, age, sex, height, and weight of the patient. For each age and sex, we compared the mean PEF values of Bangladeshi school children to those of Turkish (N-2,791), British (N-339), Chinese (N-3,196), Irish (N-2,828) and Greek (N-522) children. In most cases, the PEF values of Bangladeshi school children for both sexes according to age were lower than those of Chinese, Irish, Turkish, and British children [37]. Controlling PM sources, increasing ventilation, and employing air cleaners are all examples of ways to improve school classrooms and playgrounds [56]. In this study, the hazard quotient (HQ) is calculated to evaluate the potential for children (non-cancer health hazards) to occur from exposure to a contaminant with available noncancer health guidelines [57]. HQs less than 1 indicate a non-cancer hazard should not be an issue. When an HQ is greater than 1, retain those contaminants and conduct an in-depth toxicological effects analysis. In other words, an HQ above 1 means there is an exceedance of the non-cancer health guideline.

Conclusion
Air quality and its relationship with the health of the school children were investigated in the highly populated and populated megacity Dhaka, Bangladesh from April to November 2018. Relatively high particulate matter concentration was observed at the school premises. A poor correlation was obtained between indoor and outdoor particulate matter. Indoor and outdoor NO 2 , TVOC, and CO 2 concentrations at ten different schools revealed that the concentrations were raised in schools near the road including an overcrowded and high-traffic density region. The HQ for PM 10 and PM 2.5 had a moderate health risk of 9 to 12 years aged school children in Dhaka. A negative correlation between the concentrations of PM 1.0 , and PM 2.5 with the peak flow meter reading were found among the children in ten different schools in Dhaka. The very high indoor PM levels and increased inhalation of fine particulate matter may be clogging the airways and lowering lung performance. The study has several limitations (e.g., number of schools, students, and sampling time/period) as it was a pilot study. In future, longer sampling periods with more students form many schools at different seasons (e.g., winter, monsoon) will be needed with primary health data of the students. According to our study, we believe that schools in Dhaka city need to learn more about air pollution. The school grounds need to have a healthy atmosphere because students spend most of the time there. However, the policy makers need to take immediate actions to improve the air quality situations in the school premises in Dhaka, Bangladesh by changing the design of the school building, classroom materials, and emission control from vehicles with proper traffic management during children pick and drop off. It might also be possible to continuously monitor the school environment by setting up a real-time particulate matter detector for 24 h.

Acknowledgments
Authors acknowledge the support from all the school authorities for allowing us to conduct the sampling in their premises. Authors also acknowledge the financial support for the instruments used in this study from the Ministry of Education, The Government Republic of Bangladesh (Project no.: PS 14138).

Data availability statement
The data that support the findings of this study are available upon reasonable request from the authors.  Figure D1. Linear regression of PEF-PM estimated from ten schools during this study.