Pediatric Pulmonology 9:24-29 (1990)
Effects of Indoor Air Pollution on Lung Function of Primary School Children in Kuala-Lumpur B.H.O. Azizi, MRCP,’ and R.L. Henry, F R A C P ~ Summary. In a cross-sectionalstudy of 7-12 year-old primary school children in Kuala Lumpur city, lung function was assessed by spirometric and peak expiratory flow measurements. Spirometric and peak expiratoryflow measurements were successfully performed in 1,214 and 1,414 children, respectively. As expected, the main predictors of forced vital capacity (FVC), forced expiratory volume in one second (FEV,), forced expiratory flow between 25% and 75% of vital capacity (FEF2,-,), and peak expiratory flow rate (PEFR) were standing height, weight, age, and sex. In addition, lung function values of Chinese and Malays were generally higher than those of Indians. In multiple regression models which included host and environmentalfactors, asthma was associated with significant decreases in FEV,, FEF-, and PEFR. However, family history of chest illness, history of allergies, low paternal education, and hospitalization during the neonatal period were not independent predictors of lung function. Children sharing rooms with adult smokers had significantly lower levels of FEF_ ,., Exposures to wood or kerosene stoves were, but to mosquito repellents were not, associated with decreased lung function. Pediatr Pulmonol 1990; 9:24-29. Key words: Passive smoking; wood and kerosene stoves, mosquito repellents; family and personal medical history; ethnicity, paternal education; expiratory volume and flow rates; cross-sectional study, indoor air pollution.
INTRODUCTION
The effects of indoor environmental air pollution on lung function of children have been investigated in epidemiological studies of respiratory illness in developed countries. Most studies have been on the effects of passive smoking. Impaired lung function has been demonstrated in association with maternal and paternal smoking in several cross-sectional studies. Longitudinal studies have also shown impaired lung growth in association with parental However, others failed to show similar association^.^-^^ The differences in findings could be attributed to different methods of data collection and analysis; however, a recent reanalysis of results from two studies suggested that divergent results were not simply related to analytical methods but could be due to real differences in exposure or response under different climatic conditions. l 2 The effects of nitrogen dioxide from gas stoves have also been investigated. Two studies have shown small but statistically significant effects of exposure to gas stoves on lung function of but others with comparable sample sizes failed to demonstrate any a s s ~ c i a t i o n . ” ” ~Therefore, ”~ study results are not yet adequate to establish a causal relationship. l6 Similarly, the effects of kerosene and wood stoves have attracted the attention of several workers, I 6 , l 7 but in general more studies are needed to establish the clinical relevance of emissions from these sources. l 6 0 1990 Wiley-Liss, Inc.
Very few examinations on the effects of indoor pollution on lung function of children have been performed in developing countries. A cross-sectional study in China confirmed the adverse effect of passive smoking on children’s respiratory function, i’ but one of Papua New Guinea children failed to show any significant effect of wood burning on respiratory function. l9 We examined the relationships between exposure to indoor environmental factors, namely mosquito repellents, environmental tobacco smoke, and cooking stoves, and levels of lung function in Malaysian children. Host and familial factors were also considered.
From the Department of Paediatrics, Faculty of Medicine, National University of Malaysia, Jalan Raja Muda Abdul Aziz, Kuala Lumpur, Malaysia;’ The Discipline of Paediatrics, Faculty of Medicine, University of Newcastle, New South Wales, Australia.’ Received November 17, 1989; (revision) accepted for publication March 1, 1990. This study was supported by research grant UKM 13/87 awarded by the National University of Malaysia’s Research Fund and an additional grant from the Dean’s Office, Faculty of Medicine, the National University of Malaysia. Address correspondence and reprint requests to Dr. B.H.O. Azizi. Department of Paediatrics, Faculty of Medicine, National University of Malaysia, Jalan Raja Muda Abdul Aziz, 50300, Kuala Lumpur, Malaysia.
Indoor Air Pollution and Children’s Lung Function
MATERIALS AND METHODS
25
D. doctor-diagnosed asthma: a positive reply to “Has a doctor ever said that this child has asthma?” and E. chest illness: defined as a positive reply to “During the past 1 year had this child had any chest illness that has kept h i d h e r from hidher usual activities as much as 3 days?”
Seven to 12 year-old primary school children in Kuala Lumpur city were the stubjects of this cross-sectional study. Children in the suburbs were excluded for logistic reasons and because of possible differences in ambient environments. Single-sex schools were excluded in order These definitions are similar to those used by Schento facilitate sampling. Seventy-five primary schools were eligible for inclusion. The schools were stratified accord- ker et a1.21 In the analysis of lung function, asthma, ing to their designated type, i.e. National, National Type defined as the presence of persistent wheeze and/or doc(Chinese), and National Type (Tamil). These designa- tor-diagnosed asthma, was included as an independent tions are based on the main medium of instruction, the variable. Indoor environmental variables of interest were expoNational schools being those using the Malaysian National Language (Bahasa Malaysia), while the others are sures to passive smoking, wood stoves, kerosene stoves, those that use Mandarin and Tamil respectively. National and mosquito repellents. These variables and host factors schools are attended by children of all ethnic groups and included in regression analyses are defined in Table 1. A short questionnaire covering children’s smoking have the largest enrollments, while Chinese and Tamil schools are attended mainly by the respective ethnic practices was administered privately during lung funcgroups and have smaller enrollments. In order to ensure tion testing sessions in the respective schools. Spirometry was performed by using a Vitalograph that each ethnic group was adequately represented two Model R spirometer which had been calibrated. Each schools were randomly selected from the first type and child was tested while standing and without a nose clip. one each from the remaining types. Further random seA practice blow was allowed after which forced expiralections of classes stratified for age groups were performed in one of the National schools and in the Chinese tory curves were recorded until two reproducible tracings school to arrive at the final sample of 2,109 children. (forced expiratory volume at one second (FEV,) and A modified version of the American Thoracic Soci- forced vital capacity (FVC) values repeatable to within ety’s ATS-DLD-78C respiratory questionnaire” was 0.2 liters) were achieved. A maximum of five attempts used. This questionnaire was translated into Bahasa Ma- were allowed. The largest values of FEV, and FVC were laysia, Mandarin, and Tamil. The English and translated recorded. Forced expiratory flow between 25% and 75% versions were distributed to parents through class teach- of vital capacity (FEF,,-,,) values were calculated from ers. Parents were allowed to answer the questionnaire in the best curves. All spirometric readings were converted the language of their choice. Minor changes were made to body temperature and pressure saturated (BTPS). Peak to several questions in the questionnaire on personal and expiratory flow rates (PEFR) were measured by using a demographic data to suit local situations. Additional standard Wright peak expiratory flow meter. After a questions on local indoor environmental factors were practice blow, the best of three blows were recorded. included. However, questions on respiratory symp- Tests of lung function were performed without knowltoms and illnesses were not altered from the original edge of the responses to the questionnaire. Standing heights in stockinged feet were measured in centimeters versions. The respiratory complaints listed by questionnaire and weights were measured fully clothed after removing heavy objects from pockets. were: The study was commenced in July 1987 and data colA. chronic cough: defined as a positive reply to lection was completed in October 1987. “Does he/she cough on most days (4 or more days per The means of lung function values for children exposed and unexposed to environmental factors were week) for as much as 3 months a year?” B. chronic phlegm: defined as a positive reply to compared by using the t-test. Multiple linear regressions “Does this child seem congested or bring up phlegm, were performed to examine the effects of environmental sputum or mucous from hidher chest on most days (4 or factors, host factors and asthma, after adjusting for more days per week) for as much as 3 months a year?” standing height, weight, age, sex, and race. The indeC. persistent wheeze: defined as positive responses to pendent variables were assessed simultaneously and a and b and/or c in the series of questions: stepwise procedures were not used. Regressions were “Does this child’s chest ever sound wheezy or performed by using both untransformed and log-transformed lung function data. Comparison of the untranswhistling: formed and transformed data showed some improvement a. when he/she has a cold? in the normality and homoscedasticity of the residuals. b. occasionally apart from colds? Regressions of untransformed and transformed data on c. most days or nights?”
26
Azizi and Henry TABLE 1-Environmental and Host Variables Used in Regression Analysis of Lung Function Data ~~
Variable Passive smoking Wood stove Kerosene stove Mosquito coil Fumigation mat Aerosol repellent Asthma Ethnicity Malay Indian Family history Allergy Neonatal illness Low paternal education Male sex
Definition
Positive responses, na (%)
Sharing a bedroom with adult smoker(s) Exposure to wood stove in contrast to electridgas stove Exposure to kerosene stove in contrast to electridgas stove Exposure to mosquito coil smoke in the bedroom for at least 3 nights a week Exposure to fumigation mat mosquito repellent in the bedroom for at least 3 nights a week Exposure to aerosol mosquito repellent in the bedroom for at least 3 nights a week Reported persistent wheeze and/or doctor-diagnosed asthma
209 (14.8) 80 (5.7) 401 (28.4) 415 (29.3)
Malay in contrast to Chinese Indian in contrast to Chinese History of chest illness in either parent Allergy to food, medicine, or pollen ever diagnosed History of hospitalization in neonatal period Father’s schooling only up to primary school Male sex
655 (59.6) 315 (41.5) 177 (12.5) 61 (4.3) 69 (4.9) 567 (42.5) 819 (57.9)
57 (4.0) 176 (12.4) 199 (14.1)
“of 1,414 children on whom either spirometry or PEFR was successfully performed.
dren who successfully performed lung function tests were compared with those who did not participate in lung function testing and those who did not perform the tests statisfactorily , with regard to respiratory symptoms and illnesses. No significant differences were found between the two groups. The mean values of lung function measurements were compared between exposed and unexposed groups with respect to indoor environmental factors (Table 2) In this crude analysis, passive smoking (defined as sharing a room with a smoker) and exposures to wood and kerosene stoves were associated with lower lung function, the differences being significant for the relationships between passive smoking and all lung functions and for the relationship between exposure to wood stove and PEFR. In the regression analysis, as expected, the main predictors of all measures of lung function were standing height, weight, age and sex ( P < 0.05). In addition, ethnicity was also a significant predictor of lung function RESULTS measures. The levels of FVC, FEV,, and FEF25--75were Questionnaires were returned by 1,621 children. higher in Chinese and Malays than Indians. Table 3 Eighty-three children were excluded because of grossly shows the effects of indoor environmental and host facincomplete questionnaires and a further 26 were ex- tors on lung function. Indoor environmental factors cluded because they belonged to minority ethnic groups. which affected lung function were passive smoking and All remaining 1,512 children were invited for lung func- exposure to kerosene and wood stoves. Passive smoking tion tests. Of these, 691 (45.5%) were Malays, 469 was associated with significantly lower FEF,,-,, , while (31 .O%) Chinese, and 352 (23.3%) Indians. Spirometry exposure to wood and kerosene stoves was associated was successfully performed by 1,214 and PEFR by with significantly lower FVC, FEV,, and PEFR. Lower FEF,,-,, values were also observed in those exposed to 1,414 children. In the subsequent analysis of lung function data 11 wood and kerosene stoves, but these differences were not children who admitted to smoking were excluded. Chil- statistically significant. However, exposure to mosquito
host and environmental factors yielded similar significant predictors. Results of regressions on the log-transformed data are presented in the form of percent predicted lung function values which were obtained by multiplying by 100 the exponentials of the coefficients of the predictor variables,”’3 thus allowing direct comparisons of lung function levels in exposed and unexposed individuals. This procedure was equivalent to calculating the lung function values from the prediction equations for exposed and unexposed individuals, keeping other variables constant, and then multiplying by 100 the ratio of the lung function values of the exposed to the unexposed. All analyses were performed by using the SAS Version 5 statistical package.,’ Least squares linear regression was used to analyze lung function data. P values of less than 0.05 were regarded as statistically significant.
Indoor Air Pollution and Children’s Lung Function
27
TABLE 2-Mean Values of Lung Functions of Children Exposed (E) and Not Exposed (NE) to Indoor Environmental Factors Lung function
FEv 1 ( L)
FVC (L) Passive smoking Wood stove Kerosene stove Mosquito coil Fumigation mat Aerosol repellent
FEF25-75 (Lis)
PEFR (L/min)
E
NE
E
NE
E
NE
E
NE
1.42 1.43 1.47 1.52 I .58 1.55
1.54** 1.52 1.54 1.52 1.52 1.52
1.29 1.32 1.36 1.40 1.43 1.42
1.41*** 1.40 1.41 1.39 1.39 1.38
1.78 1.79 1.91 1.97 2.00 1.96
1.98*** 1.96 1.97 1.95 1.95 1.95
250.4 243.9 263.6 266.9 281.7 273.2
268.0*** 266.5’ 266.4 264.9 264.9 264.5
* P < 0.05. **P < 0.01. ***P < 0.001
TABLE &The Effects of Indoor Environmental and Host Factors on Lung Function % of predicted valuesa
FVC
98.1 Passive smoking 94.7* Wood stove 95.8*** Kerosene stove 99.9 Mosquito coil 98.9 Fumigation mat 101.5 Aerosol repellent 95.8** Malay 92.9*** Indian 99.3 Asthma 98.7 Family history 102.2 Allergy 99.3 Neonatal illness Low paternal education 100.5 0.78 Adiusted R-square
FEV,
FEF,,-,,
93.8** 98.1 93.1 95.3* 95.7*** 96.8 100.9 102.9 99.2 99.3 99.2 99.3 99.7 96.2** 93.1*** 93.0* 90.7*** 97.4* 99.6 98.2 101.9 102.7 102.9 106.3 98.8 102.3 0.42 0.77
PEFR
99.4 91.2*** 97.2* 100.0 104.7 104.7 101.3 97.1 94.3*** 98.9 101.o
101.7 99.1 0.60
aLog-transformed lung function data were regressed on terms for standing height, weight, age, sex, school of origin, and all the variables shown in the table simultaneously. The lung functions of children with specified factors are expressed as percentages of those of children without these factors. These values have been obtained by multiplying by 100 the exponentials of the regression coefficients of categorical independent variables (see text). *P < 0.05; **P < 0.01; ***P < 0.001; P values refer to factors which were significant in the regression equations.
coils and other mosquito repellents was not associated with lower lung function. Doctor-diagnosed asthma or persistent wheezing was significantly associated with lower FEV,, FEF,,,,, and PEFR independently of other factors. However, family history of chest illness, doctor-diagnosed allergy, hospitalization during the neonatal period (neonatal illness), and father’s low level of education were not associated with decreased lung function. No interactions were found between the significant environmental factors and age, sex, race, or other host and environmental factors.
DISCUSSION
This study examined the effects of host and environmental factors on lung function of urban Malaysian children. Ethnic diversity and the presence of several potentially harmful indoor environmental factors allowed a unique opportunity to look at some factors not previously examined, as well as those already studied, in different sociocultural and geographical situations. The possibility that climatic factors may influence exposure and outcome levels’’ has justified this study on children in a tropical environment. As expected, height, weight, age, and sex were the main determinants of lung function. In addition, significant ethnic differences were noted in spirometric lung function measures, with those of Indian children being generally lower than those of Malay and Chinese children. These differences were independent of all other factors. Ethnic differences have been reported in studies of children in developed countries and have been attributed to anthropometric variation^.^^ The associations of worse lung function with wheezing and asthma have been documented in many studies .3*24 The associations between asthma and lung function in this study indicated that responses to the questionnaire on wheezing and asthma were reasonably valid indicators of the presence or absence of obstructive respiratory conditions. Many previous studies did not include adjustments for asthma in their analyses of the effects of environmental factors on lung function. A recent review suggested that this could result in confounding bias and recommended that asthma should be included as an independent variable in regression models.’6 In our study, after adjusting for asthma and other host and environmental factors, several indoor environmental exposures were confirmed to have independent effects on lung function. The very small number of mothers who smoked did not allow an examination of the effects of maternal
28
Azizi and Henry
smoking. As a measure of exposure to passive smoking, severely affected. Also, sharing of bedrooms with chilsharing a bedroom with an adult smoker was used. Chil- dren is a fairly common practice particularly among dren who had to share rooms with at least one adult lower-income groups and an examination of this exposmoker showed significantly lower levels of FEF25-75. sure has cultural and practical relevance. In conclusion, this study confirmed previous findings This finding concurred with suggestions that the effects of adverse effects of parental smoking on lung function of passive smoking could be related to close physical of children. In addition, evidence was obtained that excontacts between smokers and children. The effect of posure to wood or kerosene stoves affected lung function maternal smoking on lung function, for instance, was adversely. more marked in female children who conceivably spent more time with their mothers than the male ~ h i l d r e n . ~ Several previous studies that showed significant effects of passive smoking on lung function reported asso- ACKNOWLEDGMENTS ciations with measures of flow such as MMER,1,2 The authors wish to thank Prof. Richard Heller, Prof. FEVo.,,3,4 and Vmax50%vc, Vmax7s%vc, and Annette Dobson, and Dr. Dianne O'Connell from the Vmax,,,v,.3 This study confirmed these findings of imCentre for Clinical Epidemiology and Biostatistics, Unipaired measures of expiratory flow rates in children exversity of Newcastle, Australia, for their encouragement posed to parental tobacco smoke. and advice, and the Ministry of Education of Malaysia Children exposed to wood or kerosene stoves had for allowing the study to be done on schoolchildren in lower lung function values than those exposed to gas or Kuala Lumpur. electric stoves. The lung function measures involved included FVC as well as FEV, and PEFR. Lower levels of FEF,, -7s were also noted although the differences were REFERENCES not statistically significant. Thus the effects of emissions from wood and kerosene stoves on lung function appear 1. Tager IB, Weiss ST, Rosner B, Speizer FE. Effect of parental cigarette smoking an pulmonary function of children. Am J Epto reduce both lung volume and expiratory flow. These idemiol. 1979; 110:15-26. findings differ from findings in relation to passive smok2. Weiss ST, Tager IB, Speizer FE, Rosner B. Persistent wheeze: its ing where only measures of obstruction were affected. 1-4 relationship to respiratory illness cigarette smoking and level of It is possible that, in contrast to exposure to passive respiratory function in a population sample of children. Am Rev smoking, prolonged exposures to emissions from wood Respir Dis. 1980; 122:697-707. or kerosene stoves may result in restrictive as well as 3. Veda1 S , Schenker MB, Samet JB, Speizer FE.Risk factors for childhood respiratory disease. Analysis of pulmonary function. obstructive pulmonary abnormalities. The explanation Am Rev Respir Dis. 1984; 130:187-192. for this is not known but future studies should further 4. Hasselbladt V, Humble CG, Graham MG, Anderson HS. lndoor examine these findings. environmental determinants of lung function in children. Am Rev Mosquito coil smoke and other mosquito repellents Respir Dis. 1981; 123:479-485. 5. Burchfiel CM, Higgins MW, Keeler JB, Howatt WF, Butler WJ. were not associated with impaired lung function. Family Higgins IT. Passive smoking in childhood. Respiratory conditions history of chest illness and doctor-diagnosed allergy and pulmonary function in Tecumseh, Michigan. Am Rev Respir were not associated with lower levels of lung function. Dis. 1986; 133:966-973. This study did not confirm previously reported associa6. Yarnell JWG, St Leger AS. Respiratory illness, maternal smoktions between hospitalization in the neonatal period and ing habit and lung function in children. Br J Dis Chest. 1979; low paternal education with lowered lung f ~ n c t i o n . ~ 73:230-236. 7. Tager IB, Weiss ST, Munoz A, Rosner B. Speizer FE. LongituAlso, the study did not detect any interactions between dinal study of the effects of maternal smoking and pulmonary passive smoking, kerosene stoves, and wood stoves and function in children. N Eng J Med. 1983; 309:699-703. other host or environmental factors. We do not have an 8. Berkey CS, Ware JH, Dockery DW. Ferris BG. Jr. Indoor air explanation as to why interactions were not observed, for pollution and pulmonary function growth in preadolescent children. Am J Epidemiol. 1986; 123:250-260. instance, between asthma and passive smoking or be9. Schilling RSF, Letai AD, Hui SL, Beck GJ, Schoenberg JB, tween low paternal education and passive smoking. Bouhays A. Lung function, respiratory disease and smoking in This was a cross-sectional study without formal meafamilies. Am J Epidemiol. 1977; 106:274-283. sures of environmental air quality. In addition, we used 10. Dodge R. The effects of indoor pollution on Arizona children. strict criteria for exposure to passive smoking by considArch Environ Health. 1982; 37:151-155. ering only the sharing of a bedroom with an adult 11. Lebowitz MD, Arnet DB, Knudson R. The effect of passive smoking on pulmonary function in children. Environ Int. 1982; smoker. Children who did not share bedrooms with adult 8:371-373. smokers would be exposed to passive smoking as well, 12. Tager IB, Segal MR, Munoz A, Weiss ST, Speizer FE. The effect but as home ventilation is probably better in the tropics of maternal cigarette smoking on the pulmonary function of chil(where windows are often left open) than in temperate dren and adolescents. Analysis of data from two populations. Am _ . countries, these children would be expected to be less Rev Respir Dis. 1987; 136:1366-1370.
Indoor Air Pollution and Children’s Lung Function 13. Ware JH, Dockery DW, Spiro A 111, Speizer FE, Ferris BG, Jr. Passive smoking, gas cooking, and respiratory health of children living in six cities. Am Rev Respir Dis. 1984; 129:366-374. 14. Speizer FE, Fems B, Jr, Bishop YM, Spengler J. Respiratory disease rates and pulmonary function in children associated with NO2 exposure. Am Rev Respir Dis. 1980; 121:3-10. 15. Florey C du V, Melia RJ, Chinn S, Goldstein BD, Brooks AGF, John HH, Craighead IB, Webster X. The relation between respiratory illness in primary school children and the use of gas for cooking. 111. Nitrogen dioxide, respiratory illness and lung function. Int J Epidemiol. 1979; 8:347-353. 16. Samet JM, Marbury MC, Spengler JD. Health effects and sources of indoor air pollution. Part I. Am Rev Respir Dis. 1987; 136: 1486-1508. 17. Cooper KR, Alberti RR. Effect of kerosene heater emissions on indoor air quality and pulmonary functions. Am Rev Respir Dis. 1984; 129~624-631. 18. Yue C, Wanxian L. The effect of passive smoking on children’s
19.
20. 21. 22. 23. 24.
29
pulmonary function in Shanghai. Am J Public Health. 1986; 76: 515-518. Anderson HR. Respiratory abnormalities, smoking habits and ventilatory capacity in a highland community in Papua New Guinea. Prevalence and effects on mortality. Int J Epidemiol. 1979; 8:127-135. Fems BG, Jr. Epidemiology standardisation project. Am Rev Respir Dis [Suppl]. 1978; 118:7-53. Schenker MB, Samet JM, Speizer FE. Risk factors for childhood respiratory disease. The effects of host factors and home environment exposures. Am Rev Respir Dis. 1983; 128:1038-1043. SAS Institute Inc. SAS User’s Guide: Statistics, Version 5 Edition. Cary NC: SAS Institute, 1985. Rossiter CE, Weill H. Ethnic differences in lung function: evidence for proportional differences. Int J Epidemiol. 1974; 3 5 - 6 1 . Leeder SR, Corkhill RT, Wysock MJ, Holland WW, Colley JRT. Influence of personal and family factors on ventilatory function of children. Br J Prev SOCMed. 1976; 30:219-224.
Effects of Indoor Air Pollution on Lung Function of Primary School Children in Kuala-Lumpur B.H.O. Azizi, MRCP,’ and R.L. Henry, F R A C P ~ Summary. In a cross-sectionalstudy of 7-12 year-old primary school children in Kuala Lumpur city, lung function was assessed by spirometric and peak expiratory flow measurements. Spirometric and peak expiratoryflow measurements were successfully performed in 1,214 and 1,414 children, respectively. As expected, the main predictors of forced vital capacity (FVC), forced expiratory volume in one second (FEV,), forced expiratory flow between 25% and 75% of vital capacity (FEF2,-,), and peak expiratory flow rate (PEFR) were standing height, weight, age, and sex. In addition, lung function values of Chinese and Malays were generally higher than those of Indians. In multiple regression models which included host and environmentalfactors, asthma was associated with significant decreases in FEV,, FEF-, and PEFR. However, family history of chest illness, history of allergies, low paternal education, and hospitalization during the neonatal period were not independent predictors of lung function. Children sharing rooms with adult smokers had significantly lower levels of FEF_ ,., Exposures to wood or kerosene stoves were, but to mosquito repellents were not, associated with decreased lung function. Pediatr Pulmonol 1990; 9:24-29. Key words: Passive smoking; wood and kerosene stoves, mosquito repellents; family and personal medical history; ethnicity, paternal education; expiratory volume and flow rates; cross-sectional study, indoor air pollution.
INTRODUCTION
The effects of indoor environmental air pollution on lung function of children have been investigated in epidemiological studies of respiratory illness in developed countries. Most studies have been on the effects of passive smoking. Impaired lung function has been demonstrated in association with maternal and paternal smoking in several cross-sectional studies. Longitudinal studies have also shown impaired lung growth in association with parental However, others failed to show similar association^.^-^^ The differences in findings could be attributed to different methods of data collection and analysis; however, a recent reanalysis of results from two studies suggested that divergent results were not simply related to analytical methods but could be due to real differences in exposure or response under different climatic conditions. l 2 The effects of nitrogen dioxide from gas stoves have also been investigated. Two studies have shown small but statistically significant effects of exposure to gas stoves on lung function of but others with comparable sample sizes failed to demonstrate any a s s ~ c i a t i o n . ” ” ~Therefore, ”~ study results are not yet adequate to establish a causal relationship. l6 Similarly, the effects of kerosene and wood stoves have attracted the attention of several workers, I 6 , l 7 but in general more studies are needed to establish the clinical relevance of emissions from these sources. l 6 0 1990 Wiley-Liss, Inc.
Very few examinations on the effects of indoor pollution on lung function of children have been performed in developing countries. A cross-sectional study in China confirmed the adverse effect of passive smoking on children’s respiratory function, i’ but one of Papua New Guinea children failed to show any significant effect of wood burning on respiratory function. l9 We examined the relationships between exposure to indoor environmental factors, namely mosquito repellents, environmental tobacco smoke, and cooking stoves, and levels of lung function in Malaysian children. Host and familial factors were also considered.
From the Department of Paediatrics, Faculty of Medicine, National University of Malaysia, Jalan Raja Muda Abdul Aziz, Kuala Lumpur, Malaysia;’ The Discipline of Paediatrics, Faculty of Medicine, University of Newcastle, New South Wales, Australia.’ Received November 17, 1989; (revision) accepted for publication March 1, 1990. This study was supported by research grant UKM 13/87 awarded by the National University of Malaysia’s Research Fund and an additional grant from the Dean’s Office, Faculty of Medicine, the National University of Malaysia. Address correspondence and reprint requests to Dr. B.H.O. Azizi. Department of Paediatrics, Faculty of Medicine, National University of Malaysia, Jalan Raja Muda Abdul Aziz, 50300, Kuala Lumpur, Malaysia.
Indoor Air Pollution and Children’s Lung Function
MATERIALS AND METHODS
25
D. doctor-diagnosed asthma: a positive reply to “Has a doctor ever said that this child has asthma?” and E. chest illness: defined as a positive reply to “During the past 1 year had this child had any chest illness that has kept h i d h e r from hidher usual activities as much as 3 days?”
Seven to 12 year-old primary school children in Kuala Lumpur city were the stubjects of this cross-sectional study. Children in the suburbs were excluded for logistic reasons and because of possible differences in ambient environments. Single-sex schools were excluded in order These definitions are similar to those used by Schento facilitate sampling. Seventy-five primary schools were eligible for inclusion. The schools were stratified accord- ker et a1.21 In the analysis of lung function, asthma, ing to their designated type, i.e. National, National Type defined as the presence of persistent wheeze and/or doc(Chinese), and National Type (Tamil). These designa- tor-diagnosed asthma, was included as an independent tions are based on the main medium of instruction, the variable. Indoor environmental variables of interest were expoNational schools being those using the Malaysian National Language (Bahasa Malaysia), while the others are sures to passive smoking, wood stoves, kerosene stoves, those that use Mandarin and Tamil respectively. National and mosquito repellents. These variables and host factors schools are attended by children of all ethnic groups and included in regression analyses are defined in Table 1. A short questionnaire covering children’s smoking have the largest enrollments, while Chinese and Tamil schools are attended mainly by the respective ethnic practices was administered privately during lung funcgroups and have smaller enrollments. In order to ensure tion testing sessions in the respective schools. Spirometry was performed by using a Vitalograph that each ethnic group was adequately represented two Model R spirometer which had been calibrated. Each schools were randomly selected from the first type and child was tested while standing and without a nose clip. one each from the remaining types. Further random seA practice blow was allowed after which forced expiralections of classes stratified for age groups were performed in one of the National schools and in the Chinese tory curves were recorded until two reproducible tracings school to arrive at the final sample of 2,109 children. (forced expiratory volume at one second (FEV,) and A modified version of the American Thoracic Soci- forced vital capacity (FVC) values repeatable to within ety’s ATS-DLD-78C respiratory questionnaire” was 0.2 liters) were achieved. A maximum of five attempts used. This questionnaire was translated into Bahasa Ma- were allowed. The largest values of FEV, and FVC were laysia, Mandarin, and Tamil. The English and translated recorded. Forced expiratory flow between 25% and 75% versions were distributed to parents through class teach- of vital capacity (FEF,,-,,) values were calculated from ers. Parents were allowed to answer the questionnaire in the best curves. All spirometric readings were converted the language of their choice. Minor changes were made to body temperature and pressure saturated (BTPS). Peak to several questions in the questionnaire on personal and expiratory flow rates (PEFR) were measured by using a demographic data to suit local situations. Additional standard Wright peak expiratory flow meter. After a questions on local indoor environmental factors were practice blow, the best of three blows were recorded. included. However, questions on respiratory symp- Tests of lung function were performed without knowltoms and illnesses were not altered from the original edge of the responses to the questionnaire. Standing heights in stockinged feet were measured in centimeters versions. The respiratory complaints listed by questionnaire and weights were measured fully clothed after removing heavy objects from pockets. were: The study was commenced in July 1987 and data colA. chronic cough: defined as a positive reply to lection was completed in October 1987. “Does he/she cough on most days (4 or more days per The means of lung function values for children exposed and unexposed to environmental factors were week) for as much as 3 months a year?” B. chronic phlegm: defined as a positive reply to compared by using the t-test. Multiple linear regressions “Does this child seem congested or bring up phlegm, were performed to examine the effects of environmental sputum or mucous from hidher chest on most days (4 or factors, host factors and asthma, after adjusting for more days per week) for as much as 3 months a year?” standing height, weight, age, sex, and race. The indeC. persistent wheeze: defined as positive responses to pendent variables were assessed simultaneously and a and b and/or c in the series of questions: stepwise procedures were not used. Regressions were “Does this child’s chest ever sound wheezy or performed by using both untransformed and log-transformed lung function data. Comparison of the untranswhistling: formed and transformed data showed some improvement a. when he/she has a cold? in the normality and homoscedasticity of the residuals. b. occasionally apart from colds? Regressions of untransformed and transformed data on c. most days or nights?”
26
Azizi and Henry TABLE 1-Environmental and Host Variables Used in Regression Analysis of Lung Function Data ~~
Variable Passive smoking Wood stove Kerosene stove Mosquito coil Fumigation mat Aerosol repellent Asthma Ethnicity Malay Indian Family history Allergy Neonatal illness Low paternal education Male sex
Definition
Positive responses, na (%)
Sharing a bedroom with adult smoker(s) Exposure to wood stove in contrast to electridgas stove Exposure to kerosene stove in contrast to electridgas stove Exposure to mosquito coil smoke in the bedroom for at least 3 nights a week Exposure to fumigation mat mosquito repellent in the bedroom for at least 3 nights a week Exposure to aerosol mosquito repellent in the bedroom for at least 3 nights a week Reported persistent wheeze and/or doctor-diagnosed asthma
209 (14.8) 80 (5.7) 401 (28.4) 415 (29.3)
Malay in contrast to Chinese Indian in contrast to Chinese History of chest illness in either parent Allergy to food, medicine, or pollen ever diagnosed History of hospitalization in neonatal period Father’s schooling only up to primary school Male sex
655 (59.6) 315 (41.5) 177 (12.5) 61 (4.3) 69 (4.9) 567 (42.5) 819 (57.9)
57 (4.0) 176 (12.4) 199 (14.1)
“of 1,414 children on whom either spirometry or PEFR was successfully performed.
dren who successfully performed lung function tests were compared with those who did not participate in lung function testing and those who did not perform the tests statisfactorily , with regard to respiratory symptoms and illnesses. No significant differences were found between the two groups. The mean values of lung function measurements were compared between exposed and unexposed groups with respect to indoor environmental factors (Table 2) In this crude analysis, passive smoking (defined as sharing a room with a smoker) and exposures to wood and kerosene stoves were associated with lower lung function, the differences being significant for the relationships between passive smoking and all lung functions and for the relationship between exposure to wood stove and PEFR. In the regression analysis, as expected, the main predictors of all measures of lung function were standing height, weight, age and sex ( P < 0.05). In addition, ethnicity was also a significant predictor of lung function RESULTS measures. The levels of FVC, FEV,, and FEF25--75were Questionnaires were returned by 1,621 children. higher in Chinese and Malays than Indians. Table 3 Eighty-three children were excluded because of grossly shows the effects of indoor environmental and host facincomplete questionnaires and a further 26 were ex- tors on lung function. Indoor environmental factors cluded because they belonged to minority ethnic groups. which affected lung function were passive smoking and All remaining 1,512 children were invited for lung func- exposure to kerosene and wood stoves. Passive smoking tion tests. Of these, 691 (45.5%) were Malays, 469 was associated with significantly lower FEF,,-,, , while (31 .O%) Chinese, and 352 (23.3%) Indians. Spirometry exposure to wood and kerosene stoves was associated was successfully performed by 1,214 and PEFR by with significantly lower FVC, FEV,, and PEFR. Lower FEF,,-,, values were also observed in those exposed to 1,414 children. In the subsequent analysis of lung function data 11 wood and kerosene stoves, but these differences were not children who admitted to smoking were excluded. Chil- statistically significant. However, exposure to mosquito
host and environmental factors yielded similar significant predictors. Results of regressions on the log-transformed data are presented in the form of percent predicted lung function values which were obtained by multiplying by 100 the exponentials of the coefficients of the predictor variables,”’3 thus allowing direct comparisons of lung function levels in exposed and unexposed individuals. This procedure was equivalent to calculating the lung function values from the prediction equations for exposed and unexposed individuals, keeping other variables constant, and then multiplying by 100 the ratio of the lung function values of the exposed to the unexposed. All analyses were performed by using the SAS Version 5 statistical package.,’ Least squares linear regression was used to analyze lung function data. P values of less than 0.05 were regarded as statistically significant.
Indoor Air Pollution and Children’s Lung Function
27
TABLE 2-Mean Values of Lung Functions of Children Exposed (E) and Not Exposed (NE) to Indoor Environmental Factors Lung function
FEv 1 ( L)
FVC (L) Passive smoking Wood stove Kerosene stove Mosquito coil Fumigation mat Aerosol repellent
FEF25-75 (Lis)
PEFR (L/min)
E
NE
E
NE
E
NE
E
NE
1.42 1.43 1.47 1.52 I .58 1.55
1.54** 1.52 1.54 1.52 1.52 1.52
1.29 1.32 1.36 1.40 1.43 1.42
1.41*** 1.40 1.41 1.39 1.39 1.38
1.78 1.79 1.91 1.97 2.00 1.96
1.98*** 1.96 1.97 1.95 1.95 1.95
250.4 243.9 263.6 266.9 281.7 273.2
268.0*** 266.5’ 266.4 264.9 264.9 264.5
* P < 0.05. **P < 0.01. ***P < 0.001
TABLE &The Effects of Indoor Environmental and Host Factors on Lung Function % of predicted valuesa
FVC
98.1 Passive smoking 94.7* Wood stove 95.8*** Kerosene stove 99.9 Mosquito coil 98.9 Fumigation mat 101.5 Aerosol repellent 95.8** Malay 92.9*** Indian 99.3 Asthma 98.7 Family history 102.2 Allergy 99.3 Neonatal illness Low paternal education 100.5 0.78 Adiusted R-square
FEV,
FEF,,-,,
93.8** 98.1 93.1 95.3* 95.7*** 96.8 100.9 102.9 99.2 99.3 99.2 99.3 99.7 96.2** 93.1*** 93.0* 90.7*** 97.4* 99.6 98.2 101.9 102.7 102.9 106.3 98.8 102.3 0.42 0.77
PEFR
99.4 91.2*** 97.2* 100.0 104.7 104.7 101.3 97.1 94.3*** 98.9 101.o
101.7 99.1 0.60
aLog-transformed lung function data were regressed on terms for standing height, weight, age, sex, school of origin, and all the variables shown in the table simultaneously. The lung functions of children with specified factors are expressed as percentages of those of children without these factors. These values have been obtained by multiplying by 100 the exponentials of the regression coefficients of categorical independent variables (see text). *P < 0.05; **P < 0.01; ***P < 0.001; P values refer to factors which were significant in the regression equations.
coils and other mosquito repellents was not associated with lower lung function. Doctor-diagnosed asthma or persistent wheezing was significantly associated with lower FEV,, FEF,,,,, and PEFR independently of other factors. However, family history of chest illness, doctor-diagnosed allergy, hospitalization during the neonatal period (neonatal illness), and father’s low level of education were not associated with decreased lung function. No interactions were found between the significant environmental factors and age, sex, race, or other host and environmental factors.
DISCUSSION
This study examined the effects of host and environmental factors on lung function of urban Malaysian children. Ethnic diversity and the presence of several potentially harmful indoor environmental factors allowed a unique opportunity to look at some factors not previously examined, as well as those already studied, in different sociocultural and geographical situations. The possibility that climatic factors may influence exposure and outcome levels’’ has justified this study on children in a tropical environment. As expected, height, weight, age, and sex were the main determinants of lung function. In addition, significant ethnic differences were noted in spirometric lung function measures, with those of Indian children being generally lower than those of Malay and Chinese children. These differences were independent of all other factors. Ethnic differences have been reported in studies of children in developed countries and have been attributed to anthropometric variation^.^^ The associations of worse lung function with wheezing and asthma have been documented in many studies .3*24 The associations between asthma and lung function in this study indicated that responses to the questionnaire on wheezing and asthma were reasonably valid indicators of the presence or absence of obstructive respiratory conditions. Many previous studies did not include adjustments for asthma in their analyses of the effects of environmental factors on lung function. A recent review suggested that this could result in confounding bias and recommended that asthma should be included as an independent variable in regression models.’6 In our study, after adjusting for asthma and other host and environmental factors, several indoor environmental exposures were confirmed to have independent effects on lung function. The very small number of mothers who smoked did not allow an examination of the effects of maternal
28
Azizi and Henry
smoking. As a measure of exposure to passive smoking, severely affected. Also, sharing of bedrooms with chilsharing a bedroom with an adult smoker was used. Chil- dren is a fairly common practice particularly among dren who had to share rooms with at least one adult lower-income groups and an examination of this exposmoker showed significantly lower levels of FEF25-75. sure has cultural and practical relevance. In conclusion, this study confirmed previous findings This finding concurred with suggestions that the effects of adverse effects of parental smoking on lung function of passive smoking could be related to close physical of children. In addition, evidence was obtained that excontacts between smokers and children. The effect of posure to wood or kerosene stoves affected lung function maternal smoking on lung function, for instance, was adversely. more marked in female children who conceivably spent more time with their mothers than the male ~ h i l d r e n . ~ Several previous studies that showed significant effects of passive smoking on lung function reported asso- ACKNOWLEDGMENTS ciations with measures of flow such as MMER,1,2 The authors wish to thank Prof. Richard Heller, Prof. FEVo.,,3,4 and Vmax50%vc, Vmax7s%vc, and Annette Dobson, and Dr. Dianne O'Connell from the Vmax,,,v,.3 This study confirmed these findings of imCentre for Clinical Epidemiology and Biostatistics, Unipaired measures of expiratory flow rates in children exversity of Newcastle, Australia, for their encouragement posed to parental tobacco smoke. and advice, and the Ministry of Education of Malaysia Children exposed to wood or kerosene stoves had for allowing the study to be done on schoolchildren in lower lung function values than those exposed to gas or Kuala Lumpur. electric stoves. The lung function measures involved included FVC as well as FEV, and PEFR. Lower levels of FEF,, -7s were also noted although the differences were REFERENCES not statistically significant. Thus the effects of emissions from wood and kerosene stoves on lung function appear 1. Tager IB, Weiss ST, Rosner B, Speizer FE. Effect of parental cigarette smoking an pulmonary function of children. Am J Epto reduce both lung volume and expiratory flow. These idemiol. 1979; 110:15-26. findings differ from findings in relation to passive smok2. Weiss ST, Tager IB, Speizer FE, Rosner B. Persistent wheeze: its ing where only measures of obstruction were affected. 1-4 relationship to respiratory illness cigarette smoking and level of It is possible that, in contrast to exposure to passive respiratory function in a population sample of children. Am Rev smoking, prolonged exposures to emissions from wood Respir Dis. 1980; 122:697-707. or kerosene stoves may result in restrictive as well as 3. Veda1 S , Schenker MB, Samet JB, Speizer FE.Risk factors for childhood respiratory disease. Analysis of pulmonary function. obstructive pulmonary abnormalities. The explanation Am Rev Respir Dis. 1984; 130:187-192. for this is not known but future studies should further 4. Hasselbladt V, Humble CG, Graham MG, Anderson HS. lndoor examine these findings. environmental determinants of lung function in children. Am Rev Mosquito coil smoke and other mosquito repellents Respir Dis. 1981; 123:479-485. 5. Burchfiel CM, Higgins MW, Keeler JB, Howatt WF, Butler WJ. were not associated with impaired lung function. Family Higgins IT. Passive smoking in childhood. Respiratory conditions history of chest illness and doctor-diagnosed allergy and pulmonary function in Tecumseh, Michigan. Am Rev Respir were not associated with lower levels of lung function. Dis. 1986; 133:966-973. This study did not confirm previously reported associa6. Yarnell JWG, St Leger AS. Respiratory illness, maternal smoktions between hospitalization in the neonatal period and ing habit and lung function in children. Br J Dis Chest. 1979; low paternal education with lowered lung f ~ n c t i o n . ~ 73:230-236. 7. Tager IB, Weiss ST, Munoz A, Rosner B. Speizer FE. LongituAlso, the study did not detect any interactions between dinal study of the effects of maternal smoking and pulmonary passive smoking, kerosene stoves, and wood stoves and function in children. N Eng J Med. 1983; 309:699-703. other host or environmental factors. We do not have an 8. Berkey CS, Ware JH, Dockery DW. Ferris BG. Jr. Indoor air explanation as to why interactions were not observed, for pollution and pulmonary function growth in preadolescent children. Am J Epidemiol. 1986; 123:250-260. instance, between asthma and passive smoking or be9. Schilling RSF, Letai AD, Hui SL, Beck GJ, Schoenberg JB, tween low paternal education and passive smoking. Bouhays A. Lung function, respiratory disease and smoking in This was a cross-sectional study without formal meafamilies. Am J Epidemiol. 1977; 106:274-283. sures of environmental air quality. In addition, we used 10. Dodge R. The effects of indoor pollution on Arizona children. strict criteria for exposure to passive smoking by considArch Environ Health. 1982; 37:151-155. ering only the sharing of a bedroom with an adult 11. Lebowitz MD, Arnet DB, Knudson R. The effect of passive smoking on pulmonary function in children. Environ Int. 1982; smoker. Children who did not share bedrooms with adult 8:371-373. smokers would be exposed to passive smoking as well, 12. Tager IB, Segal MR, Munoz A, Weiss ST, Speizer FE. The effect but as home ventilation is probably better in the tropics of maternal cigarette smoking on the pulmonary function of chil(where windows are often left open) than in temperate dren and adolescents. Analysis of data from two populations. Am _ . countries, these children would be expected to be less Rev Respir Dis. 1987; 136:1366-1370.
Indoor Air Pollution and Children’s Lung Function 13. Ware JH, Dockery DW, Spiro A 111, Speizer FE, Ferris BG, Jr. Passive smoking, gas cooking, and respiratory health of children living in six cities. Am Rev Respir Dis. 1984; 129:366-374. 14. Speizer FE, Fems B, Jr, Bishop YM, Spengler J. Respiratory disease rates and pulmonary function in children associated with NO2 exposure. Am Rev Respir Dis. 1980; 121:3-10. 15. Florey C du V, Melia RJ, Chinn S, Goldstein BD, Brooks AGF, John HH, Craighead IB, Webster X. The relation between respiratory illness in primary school children and the use of gas for cooking. 111. Nitrogen dioxide, respiratory illness and lung function. Int J Epidemiol. 1979; 8:347-353. 16. Samet JM, Marbury MC, Spengler JD. Health effects and sources of indoor air pollution. Part I. Am Rev Respir Dis. 1987; 136: 1486-1508. 17. Cooper KR, Alberti RR. Effect of kerosene heater emissions on indoor air quality and pulmonary functions. Am Rev Respir Dis. 1984; 129~624-631. 18. Yue C, Wanxian L. The effect of passive smoking on children’s
19.
20. 21. 22. 23. 24.
29
pulmonary function in Shanghai. Am J Public Health. 1986; 76: 515-518. Anderson HR. Respiratory abnormalities, smoking habits and ventilatory capacity in a highland community in Papua New Guinea. Prevalence and effects on mortality. Int J Epidemiol. 1979; 8:127-135. Fems BG, Jr. Epidemiology standardisation project. Am Rev Respir Dis [Suppl]. 1978; 118:7-53. Schenker MB, Samet JM, Speizer FE. Risk factors for childhood respiratory disease. The effects of host factors and home environment exposures. Am Rev Respir Dis. 1983; 128:1038-1043. SAS Institute Inc. SAS User’s Guide: Statistics, Version 5 Edition. Cary NC: SAS Institute, 1985. Rossiter CE, Weill H. Ethnic differences in lung function: evidence for proportional differences. Int J Epidemiol. 1974; 3 5 - 6 1 . Leeder SR, Corkhill RT, Wysock MJ, Holland WW, Colley JRT. Influence of personal and family factors on ventilatory function of children. Br J Prev SOCMed. 1976; 30:219-224.
