Investigation of Indoor Air Quality in Houses of Macedonia

People who live in buildings are exposed to harmful effects of indoor air pollution for many years. Therefore, our research is aimed to investigate the indoor air quality in family houses. The measurements of indoor air temperature, relative humidity, total volatile organic compounds (TVOC), particulate matters (PM) and sound pressure level were carried out in 25 houses in several cities of the Republic of Macedonia. Mean values of indoor air temperature and relative humidity ranged from 18.9 °C to 25.6 °C and from 34.1% to 68.0%, respectively. With regard to TVOC, it can be stated that excessive occurrence was recorded. Mean values ranged from 50 μg/m3 to 2610 μg/m3. Recommended value (200 μg/m3) for human exposure to TVOC was exceeded in 32% of houses. Mean concentrations of PM2.5 (particular matter with diameter less than 2.5 µm) and PM10 (diameter less than 10 µm) are determined to be from 16.80 µg/m3 to 30.70 µg/m3 and from 38.30 µg/m3 to 74.60 µg/m3 individually. Mean values of sound pressure level ranged from 29.8 dB(A) to 50.6 dB(A). Dependence between characteristics of buildings (Year of construction, Year of renovation, Smoke and Heating system) and data from measurements (Temperature, Relative humidity, TVOC, PM2.5 and PM10) were analyzed using R software. Van der Waerden test shows dependence of Smoke on TVOC and PM2.5. Permutational multivariate analysis of variance shows the effect of interaction of Renovation and Smoke.


Introduction
Indoor environmental quality (IEQ) is an essential condition to establish a healthy housing environment [1] and is crucially linked to occupant' health and well-being [2]. Human beings have endeavored to create indoor environments in which they can feel comfortable. Human health is foremost when it comes to assessing the overall comfort of the environment. If the built environment is leading to sickness or negative impact on the occupant' health for any reason then it could lead to some design or technical flaw in the building system [3]. Building structures are linked with a range of health hazard, such as those attributable to extreme temperatures, indoor air pollution, noise, airborne infectious diseases or mold contamination [4]. Today, there is no secret that long-term as well as short-term exposure to PM 2.5 has been associated with increased respiratory and cardiovascular morbidity [5]. Study [6] showed that major contributor to total indoor volatile organic compounds in residences were households products, followed by combustion processes and environmental tobacco smoke, deodorizers and off-gassing of building materials. These chemicals can cause irritation of eyes contribution of the study is underlining the level of IAQ as well as finding dependence between building characteristics and measured indoor environmental parameters.

Selection of Objects
The research object is 25 homes situated in the Republic of Macedonia. Houses included in the IAQ study were selected from the area of south-western part of the Republic of Macedonia about 120 km from Skopje, in town Prilep. The details of the houses are shown in the Table 1. Selecting of activities which were considered in statistical analysis related with occurrence of indoor air pollutants. There are scientific evidence that smoking has impact to occurrence of particulate matters and VOCs (e.g., benzene, ethylbenzene, and styrene) [23][24][25][26]. Renovation (thermal insulation of buildings) and year of construction affects indoor air temperature and relative humidity as well as concentrations of pollutants [27]. Concentrations of particulate matters can be also influenced by type of heating system [28,29]. The following building characteristics were chosen for the analysis: type of building-single-family houses and apartments in multi-family buildings; age of the building (construction finished in years from 1960 to 2005); renovation, smoking, and heating system.

Measurements of Indoor Air Quality Factors
Indoor air temperature, relative humidity, sound pressure level, particulate matters and total volatile organic compounds were measured in the selected family houses in the period from December 2012 until March 2013. In this period, the outside air temperature was in the range of −10 • C to 10 • C and the external humidity from 30% to 79%.
Indoor air temperature and relative humidity were measured by a temperature and humidity meter (  The data from all the instruments was downloaded to a computer for further analysis. During the measurements all of the instruments were placed approximately in the middle of the living room in the height of 1.1 m above the floor. The measurement lasted for 1 hour and 30 min during normal operation of the building. Each measurement was repeated three times. Living rooms were selected as reference rooms because building users spend substantial part of day in these spaces. The doors and the windows were closed throughout the measurement. More data concerning the instruments used is shown in Table 2.

Statistical Analyses
Characteristics of buildings (Year of construction, Year of renovation, Smoke and Heating system) and data from measurements (Temperature, Relative humidity, TVOC, PM 2.5 and PM 10 ) were used for statistical analysis using R software (R Foundation for Statistical Computing, Vienna, Austria, version 3.2.5) [30]. Since Normality property is violated we focused on nonparametric methods. Permutational multivariate analysis of variance (PERMANOVA, [31]) was used for multivariate analysis to determine the impact of continuous and categorical factors on measurements. To test null hypothesis ("The centroids of the groups, as defined in the space of the chosen resemblance measure, are equivalent for all groups."), a pseudo-F-statistic, modelled on the classical F-statistic used in classical Analysis of variance (ANOVA), is constructed directly from the dissimilarity values in the matrix [32]. Important fact is that PERMANOVA is quite robust to correlations and heterogeneous variances. The PERMANOVA used the "adonis" procedure in the vegan package [33]. Ordinations were plotted with non-metric multidimensional scaling (NMDS) [34]; using the default settings of vegan' "metaMDS" procedure and "ordiellipse", which adds ellipses enclosing all points in the group (ellipsoid hulls) or ellipses of standard deviation, standard error or confidence areas. This allows us to visualize the level of similarity of individual cases of a dataset. It uses adequate dissimilarity measures, whereas standard distance on real line was chosen. Consequently, Van der Waerden normal scores test in the PMCMR package was used for univariate cases [35]. The advantage of the Van Der Waerden test is that it provides the high efficiency of the standard ANOVA analysis when the normality assumptions are in fact satisfied, but it also provides the robustness of the Kruskal-Wallis test when the normality assumptions are not satisfied.

Results and Discussion
The following table (Table 3)

Temperature and Relative Humidity
The mean indoor air temperature ranges from 19.3 • C to 25.6 • C in single family houses and from 18.9 • C to 25.1 • C in apartments. Standard deviation (S.D.) ranges from 0.4 to 1.0. Similar values were found in study [18]. The mean indoor temperature set of 157 single-family houses and 148 apartments was found to be 21.4 • C and 22.5 • C, respectively. This study also points out that the values decrease in the single-family houses than in apartments which may have been caused by building characteristics (e.g., less exposed facades, sharing internal walls in case of apartments) but also by the occupant' behavior related to the selection of the heating set-point (e.g., the elderly living mainly in apartments prefer slightly higher temperatures). Such conclusions can be deductive for our study. Because it is not possible to say whether higher or lower values of temperature were achieved in renovated or non-renovated houses.
In our study the mean relative humidity ranges from 36.0% to 64.0% in single family houses and from 34.1% to 68.0% in apartments. Standard deviation ranges from 0.3 to 1.4. According to study [18] the mean relative humidity was higher in the single-family houses than in the apartments (34% vs. 31%). These values correspond with the required range of 30%-70%.

Total Volatile Organic Compounds
In Table 3, there are TVOC concentrations measured in the selected houses. As can be seen, the mean values of TVOC concentrations ranged from 50 µg/m 3 to 2610 µg/m 3 . The lowest mean level (50 µg/m 3 ) was measured in house (House 6) with a maximum mean relative humidity. The recommended value (200 µg/m 3 ) [36] for human exposure to TVOC was exceeded in 32% of all houses. However in 73% of houses with allowed smoking were observed with very high concentrations ranged from 206 to 2610 µg/m 3 . Study [37] ascertains that in the pre-occupancy stage, the median TVOC levels were low in two houses (less than 150 µg/m 3 ) and high in the other houses (between 500 and more that 3000 µg/m 3 ). In study [38] the indoor total VOC levels were fairly low (1283 µg/m 3 ) compared to other studies (210-6000 µg/m 3 ) [39]. As noted above the 157 single-family houses and 148 apartments were monitored from the occurrence of TVOC concentrations too [18]. The mean TVOC concentration in single family houses was higher (306 µg/m 3 ) than in apartments (174 µg/m 3 ). This study takes a note that a previous study [40] conducted more than ten years ago found higher concentrations of 388 µg/m 3 in single-family houses and 317 µg/m 3 in apartments. Very high levels of TVOC concentrations in family houses were recorded in our study.

Particulate Matters
Box plot of PM 2.5 concentrations in selected houses is shown in Figure 1. Mean concentrations of PM 2.5 ranged from 16.80 µg/m 3 to 30.70 µg/m 3 . Similar results to our observation were ascertained in study [37], in which the mass concentrations of PM 2.5 were always below 30 µg/m 3 and ranged from 6 to 28 µg/m 3 . Escobedo et al (2014) [41] found that the houses using natural gas for cooking had average 24 h indoor concentration of 8.0 µg/m 3 for PM 2.5 . Households using electricity had corresponding value of 4.7 µg/m 3 . This study confirms that only the homes with an indoor cigarette-smoking event during the two weeks prior to the survey had the highest concentration of 28 µg/m 3 . The average indoor PM 2.5 concentration was determined to be 7.0 µg/m 3 [41]. A comprehensive study [42] confirms that indoor pollutants such as cigarette smoking and cooking are a major source of indoor PM 2.5 concentration and their impact is much greater than that infiltrated from outside. The following results are gained: (a) smoking led to an increase in the level of indoor concentration to as much as 1280 µg/m 3 , which took several hours to settle down; (b) indoor concentration in a room subjected to smoking was about 0.6 times higher than that of a control room; (c) cooking activities contributed to the PM 2.5 concentration in the kitchen, to a level of 3000 µg/m 3 within a short period of time; and (d) human activities such as walking, dressing and sweeping contributed to an increase of indoor concentration by about 33%.  [36] for human exposure to TVOC was exceeded in 32% of all houses. However in 73% of houses with allowed smoking were observed with very high concentrations ranged from 206 to 2610 μg/m 3 . Study [37] ascertains that in the pre-occupancy stage, the median TVOC levels were low in two houses (less than 150 μg/m 3 ) and high in the other houses (between 500 and more that 3000 μg/m 3 ). In study [38] the indoor total VOC levels were fairly low (1283 μg/m 3 ) compared to other studies (210-6000 μg/m 3 ) [39]. As noted above the 157 single-family houses and 148 apartments were monitored from the occurrence of TVOC concentrations too [18]. The mean TVOC concentration in single family houses was higher (306 μg/m 3 ) than in apartments (174 μg/m 3 ). This study takes a note that a previous study [40] conducted more than ten years ago found higher concentrations of 388 μg/m 3 in single-family houses and 317 μg/m 3 in apartments. Very high levels of TVOC concentrations in family houses were recorded in our study.

Particulate Matters
Box plot of PM2.5 concentrations in selected houses is shown in Figure 1. Mean concentrations of PM2.5 ranged from 16.80 µ g/m 3 to 30.70 µ g/m 3 . Similar results to our observation were ascertained in study [37], in which the mass concentrations of PM2.5 were always below 30 μg/m 3 and ranged from 6 to 28 μg/m 3 . Escobedo et al (2014) [41] found that the houses using natural gas for cooking had average 24 h indoor concentration of 8.0 μg/m 3 for PM2.5. Households using electricity had corresponding value of 4.7 μg/m 3 . This study confirms that only the homes with an indoor cigarettesmoking event during the two weeks prior to the survey had the highest concentration of 28 μg/m 3 . The average indoor PM2.5 concentration was determined to be 7.0 μg/m 3 [41]. A comprehensive study [42] confirms that indoor pollutants such as cigarette smoking and cooking are a major source of indoor PM2.5 concentration and their impact is much greater than that infiltrated from outside. The following results are gained: (a) smoking led to an increase in the level of indoor concentration to as much as 1280 μg/m 3 , which took several hours to settle down; (b) indoor concentration in a room subjected to smoking was about 0.6 times higher than that of a control room; (c) cooking activities contributed to the PM2.5 concentration in the kitchen, to a level of 3000 μg/m 3 within a short period of time; and (d) human activities such as walking, dressing and sweeping contributed to an increase of indoor concentration by about 33%.   [17]. This study points out that Summer/Autumn indoor PM10 (22.6 ± 9.0 μg/m 3 ) was significantly higher than Winter/Spring indoor PM10 (18.3 ± 6.6 μg/m 3 ). This study also suggests that heating was significantly   [17]. This study points out that Summer/Autumn indoor PM 10 (22.6 ± 9.0 µg/m 3 ) was significantly higher than Winter/Spring indoor PM 10 (18.3 ± 6.6 µg/m 3 ). This study also suggests that heating was significantly correlated with PM 10 which may be due to the use of wood heaters in some homes during the Winter/Spring period.    Study [37] shows that monitored bedrooms were generally the quietest (less than 30-33 dB(A)), except in two houses (up to 48 and 36 dB(A), respectively). Mean noise level in kitchens and drawing rooms was 53.58 and 55.67 dB(A)respectively in rural and urban houses in India [45]. Study of Ryu and Jeon [46] showed that noise sensitivity influenced the annoyance level caused by both indoor and outdoor noise. Bivariate analysis in study of Hammersen et al. (2016) revealed associations between high levels of noise annoyance and impaired mental health for all noise sources except air traffic [47].     Study [37] shows that monitored bedrooms were generally the quietest (less than 30-33 dB(A)), except in two houses (up to 48 and 36 dB(A), respectively). Mean noise level in kitchens and drawing rooms was 53.58 and 55.67 dB(A)respectively in rural and urban houses in India [45]. Study of Ryu and Jeon [46] showed that noise sensitivity influenced the annoyance level caused by both indoor and outdoor noise. Bivariate analysis in study of Hammersen et al. (2016) revealed associations between high levels of noise annoyance and impaired mental health for all noise sources except air traffic [47]. Study [37] shows that monitored bedrooms were generally the quietest (less than 30-33 dB(A)), except in two houses (up to 48 and 36 dB(A), respectively). Mean noise level in kitchens and drawing rooms was 53.58 and 55.67 dB(A)respectively in rural and urban houses in India [45]. Study of Ryu and Jeon [46] showed that noise sensitivity influenced the annoyance level caused by both indoor and outdoor noise. Bivariate analysis in study of Hammersen et al. (2016) revealed associations between high levels of noise annoyance and impaired mental health for all noise sources except air traffic [47].

Statistical Analysis
Further advantage of Permutational multivariate analysis of variance is that factor variables can be continuous but categorical as well. It reveals that we cannot conclude factors: Temperature, Relative humidity, Heating system and Year of construction. They have any effect on dependent variable TVOC, PM 2.5 , PM 10 . Table 4 shows the opposite case. Evidently the factor Smoke influences TVOC, PM 2.5 , and PM 10 in the case of multivariate dependence. Nevertheless, Van der Waerden test reveals that we cannot conclude statistical dependence of Smoke on PM 10 (similar situation is in other cases), only on TVOC and PM 2.5 (Table 5).
Significance levels α are indicated with stars (asterisks). If the p-values are less than or equal to the significance level, then the outcome is said to be statistically significant (* significant at p < 0.05 = α).  For differences of TVOC and PM 2.5 we can see it on box plots (Figure 4). In the case of Renovation factor the situation is more complicated. There is no evidence of influence concerning only itself. But PERMANOVA shows us the effect of interaction of Renovation and Smoke. This means hidden dependence. In the Figure 5 of NMDS are obviously two particular groups of houses differentiated by the factor Smoke. Evidently there is an intersection, which confirms previous ideas (interaction). Confidence areas (ellipses) show that we cannot include specific house in exactly one of the groups for 100%, but there is quite high probability to be in the group Smoke/Y.

Statistical Analysis
Further advantage of Permutational multivariate analysis of variance is that factor variables can be continuous but categorical as well. It reveals that we cannot conclude factors: Temperature, Relative humidity, Heating system and Year of construction. They have any effect on dependent variable TVOC, PM2.5, PM10. Table 4 shows the opposite case. Evidently the factor Smoke influences TVOC, PM2.5, and PM10 in the case of multivariate dependence. Nevertheless, Van der Waerden test reveals that we cannot conclude statistical dependence of Smoke on PM10 (similar situation is in other cases), only on TVOC and PM2.5 (Table 5).
Significance levels α are indicated with stars (asterisks). If the p-values are less than or equal to the significance level, then the outcome is said to be statistically significant (* significant at p < 0.05 = α).  For differences of TVOC and PM2.5 we can see it on box plots (Figure 4). In the case of Renovation factor the situation is more complicated. There is no evidence of influence concerning only itself. But PERMANOVA shows us the effect of interaction of Renovation and Smoke. This means hidden dependence. In the Figure 5 of NMDS are obviously two particular groups of houses differentiated by the factor Smoke. Evidently there is an intersection, which confirms previous ideas (interaction). Confidence areas (ellipses) show that we cannot include specific house in exactly one of the groups for 100%, but there is quite high probability to be in the group Smoke/Y.
A. Dependence between smoke and TVOC B. Dependence between smoke and PM2.5

Conclusions
This study has discussed the indoor air quality in 25 homes in Prilep, Republic of Macedonia during the winter period. Results were compared with other surveys performed over the world. Because of the limited set of houses, the findings cannot be generalized. Our results indicated that indoor relative humidity in homes meets the requirement of 30%-70%. The mean values of sound pressure level were high in most of the houses. The recommended value for TVOC (200 μg/m 3 ) was exceeded in 32% of houses. Concentrations of PM10 were high in 64% of the investigated houses. This study shows that family house users are highly exposed to excessive noise, high concentrations of total volatile organic compounds as well excessive occurrence of particulate matters in the indoor air. The results of the present study are comparable to the results of other studies mentioned above. (Section 3). Some differences are related to different boundary conditions and the chosen methodology for measuring the environmental parameters.
By statistical analysis correlations between smoke and TVOC; smoke and PM2.5 as well as hidden dependence between renovation and smoke were found.
The study confirmed the high concentrations of the environmental parameters. General knowledge of indoor air quality in family houses is often very low and the occupants do not know that exposure to these pollutants has an impact on their health and comfort. Therefore, the indoor air quality needs to be investigated and people need to be informed of the possible health consequences. Long-term measurements of indoor environmental parameters need to be performed in order for these parameters to be generalized. This will be the objective of our future research work.
Acknowledgments: This study was supported by the Grant Agency of the Slovak Republic for the support of project No. 1/0307/16.

Author Contributions:
Silvia Vilcekova had the original idea and design of the study; Zoran Apostoloski carried out measurements, Ludmila Meciarova and Jozef Kiselak analyzed data and conducted the statistical analyses. Silvia Vilcekova and Eva Kridlova Burdova interpreted the results, prepared the text and provides critical version of the manuscript, which was revised by all authors. All authors read and approve the final manuscript.

Conflicts of Interest:
The authors declare no conflict of interest.

Abbreviations
The following abbreviations are used in this manuscript:

IEQ
Indoor environmental quality OEQ Overall Environmental Satisfaction IAQ Indoor Air Quality PM Particulate Matter

Conclusions
This study has discussed the indoor air quality in 25 homes in Prilep, Republic of Macedonia during the winter period. Results were compared with other surveys performed over the world. Because of the limited set of houses, the findings cannot be generalized. Our results indicated that indoor relative humidity in homes meets the requirement of 30%-70%. The mean values of sound pressure level were high in most of the houses. The recommended value for TVOC (200 µg/m 3 ) was exceeded in 32% of houses. Concentrations of PM 10 were high in 64% of the investigated houses. This study shows that family house users are highly exposed to excessive noise, high concentrations of total volatile organic compounds as well excessive occurrence of particulate matters in the indoor air. The results of the present study are comparable to the results of other studies mentioned above. (Section 3). Some differences are related to different boundary conditions and the chosen methodology for measuring the environmental parameters.
By statistical analysis correlations between smoke and TVOC; smoke and PM 2.5 as well as hidden dependence between renovation and smoke were found.
The study confirmed the high concentrations of the environmental parameters. General knowledge of indoor air quality in family houses is often very low and the occupants do not know that exposure to these pollutants has an impact on their health and comfort. Therefore, the indoor air quality needs to be investigated and people need to be informed of the possible health consequences. Long-term measurements of indoor environmental parameters need to be performed in order for these parameters to be generalized. This will be the objective of our future research work.
Acknowledgments: This study was supported by the Grant Agency of the Slovak Republic for the support of project No. 1/0307/16.

Author Contributions:
Silvia Vilcekova had the original idea and design of the study; Zoran Apostoloski carried out measurements, Ludmila Meciarova and Jozef Kiselak analyzed data and conducted the statistical analyses. Silvia Vilcekova and Eva Kridlova Burdova interpreted the results, prepared the text and provides critical version of the manuscript, which was revised by all authors. All authors read and approve the final manuscript.

Conflicts of Interest:
The authors declare no conflict of interest.

Abbreviations
The following abbreviations are used in this manuscript: