Indoor Air 2005; 15: 6266 www.blackwellpublishing.com/ina Printed in Denmark.

All rights reserved

Copyright Blackwell Munksgaard 2004

INDOOR AIR doi:10.1111/j.1600-0668.2004.00324.x

Association of air-conditioning with respiratory symptoms in oce workers in tropical climate

Abstract To evaluate the association of heating, ventilation and air-conditioning systems (HVAC) and respiratory symptoms in a tropical city, self-administered questionnaires were given to 2000 individuals working in air-conditioned oce buildings and to 500 control workers in naturally ventilated buildings. Reported symptoms from the two populations were analyzed using chi-square tests, univariate and multiple logistic regressions models. Symptoms were the outcome variable and the odds ratios were adjusted by gender, age, accumulated work time, smoking habits and atopic background. There was a 79.8% response rate and there was a positive association of nasal symptoms (odds ratio, OR 1.59, 95% condence interval, CI 1.112.28), naso-ocular symptoms (OR 1.58, 95% CI 1.052.38), persistent cough (OR 3.04, 95% CI 2.004.63) sinusitis symptoms (OR 1.85, 95% CI 1.272.71) and building-related worsening of the symptoms (OR 4.92, 95% CI 2.938.27) with working in air-conditioned buildings. In conclusion, our study suggests that articial airconditioning is a matter of concern for respiratory symptoms in cities with hot and humid climate.

G. S. Graudenz1, C. H. Oliveira2, A. Tribess3, C. Mendes, Jr3, M. R. D. O. Latorre4, J. Kalil1

1 Division of Allergy and Clinical Immunology, Internal Medicine Department, School of Medicine, Millennium Institute, University of So Paulo, So Paulo, 2Pediatric Department, School of Medicine, University of Campinas, Campinas, 3Department of Mechanical Engineering, Polytechnic School and 4Department of Epidemiology, School of Public Health, University of So Paulo, So Paulo, Brazil

Key words: Air-conditioning; Air pollution; Indoor; Environmental exposure; Occupational health; Allergy. Gustavo Silveira Graudenz Rua Girassol 1317/21, So Paulo, SP 05433-002, Brazil Tel.: 55 11 30316642 Fax: 55 11 30321019 e-mail: gustavog@usp.br Received for review 23 March 2004. Accepted for publication 6 October 2004.

Practical Implications

This study suggests that indoor air-related respiratory symptoms are a matter of concern in places with hot and humid climate. The regression models were adjusted by confounders that could be used in further reanalysis of indoor air quality related symptoms and ventilation systems with expanded variety of climatic conditions.

Introduction

The life-style of the urban population has been rapidly changing along the last decades. The increased amount of time spent indoors, even in tropical countries, makes the indoor air quality (IAQ) a very important topic worldwide. Problems associated to indoor environment are one of the most common environmental health issues faced by clinicians (Ledford, 1994), but the factors associated with the perceived IAQ are not fully understood. They include temperature, humidity, air exchange rates, odors, air movement, ventilation, biological contaminants, volatile organic compounds, bacterial toxins, labor and psychosocial factors (Apter et al., 1994). One of the 62

most important factors aecting IAQ is the presence and the characteristics of the heating, ventilation and air-conditioning systems (HVAC) in the places studied (Mendell and Smith, 1990). In a recent reanalysis of prevalence studies, there was an increase in prevalence of indoor air-related symptoms between 30 and 200% in buildings with air-conditioning, with or without humidication, when compared with natural ventilation systems (Seppanen and Fisk, 2002). The authors stressed the need of data in tropical cities with hot and humid air most of the year. The geographic variations can lead to dierent IAQ perception (Spaul, 1994). The existing evidence of the association of HVAC systems and IAQ complains is to be determined in hot and humid air conditions.

Association of HVAC and respiratory symptoms in a tropical city The use of air-conditioning systems is widespread in Brazil and the standard equipments lack heating or humidication systems even in inland dryer regions. Sao Paulo is the largest city in South America with around 10 million inhabitants. High temperatures (15 35C) and high relative humidity (6094%) occur throughout the year. A typical day has a temperature of 32C and relative air humidity of 70%, which corresponds to an absolute air humidity of 23 g of water per kg of dry air. The role of respiratory allergy such as asthma and allergic rhinitis, is still a matter of debate because it is consistently involved as a risk factor to develop IAQ-related symptoms in epidemiological studies (Bjornsson et al., 1998; Brooks, 1994; Ooi and Goh, 1996). However, no signicant exposure of aeroallergens to explain IAQ-related symptoms is usually found (Graudenz et al., 2002a; Menzies et al., 1998). Most previous epidemiological studies have not used standardized questionnaires for the diagnosis of allergic symptoms. Therefore, the use of specic methods for addressing the host susceptibility could add more information for this group of atopic individuals, specially in Brazil with high prevalence of allergic diseases (Sole et al., 2001; Vanna et al., 2001). The aim of this study was to evaluate the association of respiratory symptoms of full-time workers from buildings with air-conditioning without humidication systems in comparison with naturally ventilated buildings in a city with hot and humid climate. not working full-time (8 h a day or more) under the HVAC systems were excluded. The control group was composed of 500 workers, mostly clerical workers and trainees, staying full-time in three public buildings with natural ventilation via leakage paths and intentional openings, with no thermal control or presence of sealed windows. They were located in the same city of the air-conditioned buildings and evaluated at the same season of the year. Stas allocated in sections equipped with any kind of air-conditioning or mechanical air exhaust system were excluded. Control buildings ranged from 30 to 45 years old, had partial carpeted coverings and were also dependent of articial illumination. Control individuals who were exposed to any kind of HVAC or airconditioning systems, for more than 36 h a week were arbitrarily excluded. In all buildings, oces were equipped with laser printers, fax machines and video terminal displays. No print shops, internal loading docks, stored chemicals or gas-red heating systems were present in any building. There were no reports of major building renovation or water incursion in the locations studied. None of the buildings studied were previously known as sick buildings. After the authorization of the managers and workers of the companies settled in the buildings, and the approval of the Ethics Committee of the Sao Paulo University, a self-administered questionnaire regarding allergy background, smoking status, respiratory symptoms, previous medical diagnosis of asthma and rhinitis and work relation with respiratory symptoms was given by a medical doctor and explained to the managers of each studied section of the buildings. The questionnaire was a combination of two previously standardized and validated for Portuguese language questionnaires; the ATS-DLD-78,(Esteves et al., 1999a,b; Ferris, 1978) and the ISAAC (Asher et al., 1995; Esteves et al., 1999a,b). The main questions on asthma symptoms were: Did you have wheezing or whistling in your chest in the last 12 months when you did not have u? Have you ever had attacks of breathlessness with wheezing? The core questions on nasal symptoms were: In the past 6 months have you had a problem with sneezing, a runny or a blocked nose when you did not have u? In the past 6 months has this nose problem been accompanied by itchywatery eyes? The question about sinusitis was: In the past 12 months did you have purulent nasal discharge accompanied by frontal or maxillary headache? And where do the above symptoms more frequently occur?
Statistical analysis

Material and methods Office buildings

A total of 2000 full-time oce workers from eight dierent oce buildings, located in the downtown area of Sao Paulo were studied. All oces were sealed, with HVAC systems without humidication devices or opening windows, fully carpeted, and with articial uorescent illumination. None of the selected places were at the ground level, varying from the 3rd to the 35th oor and with 1535 years of commercial use. All HVAC systems had periodical maintenance, operational and organizational standard procedures to assure ltering, exchange air rates, total particulate matter, dioxide carbon levels, temperature and humidity under national safety limits (Sanitaria, Agencia Nacional de Vigilancia Sanitaria, 2003). Previous evaluation of indoor biological and allergenic contents from these air-conditioned buildings showed no contaminations (Graudenz et al., 2002b). The workers did not know the characteristics of their air-conditioning system, or the specic objective of the study that was conducted during an evaluation of the non-smoking policy in commercial oces of Sao Paulo. Individuals

The population characteristics according to exposure to air-conditioning systems were analyzed using chisquare tests. Univariate and multiple logistic models were estimated in order to analyze the variables 63

Graudenz et al. associated with nasal and eye symptoms, recurrent wheezing and building-related worsening of the symptoms. The comparison of the crude symptoms prevalence and the symptoms-adjusted odds ratios in natural ventilation and HVAC groups assessed the role of airconditioning as determinant of the symptoms. The symptoms were the dependent variables and the independent variables were gender, age, accumulated working time in the building, smoking habits, passive smoking, history of familiar allergy, previous medical diagnosis of asthma and rhinitis and the air-conditioning system groups. The variables that remained at the nal model were either signicant (P < 0.05) or were confounding.
Results

Upper airways symptoms such as nasal, naso-ocular and sinusitis had a signicant association with working under the HVAC system in the univariate logistic regression. Breathlessness episodes, persistent cough and building relation of the symptoms also had signicant association with air-conditioning (Table 3). Upper airways symptoms such as nasal, naso-ocular and sinusitis maintained positive signicant associations with working under the HVAC systems, whereas lower airways as chest wheezing or breathlessness episodes showed no signicant association with airconditioning in the multivariate model. Persistent cough and building relation of the symptoms had signicant association with the HVAC system after the adjustment (Table 4).
Discussion

The response rate was 79.8%, totalizing 1994 responders of 2500. Among the responders, 1885 subjects identied themselves and 139 decided to be anonymous. The characteristics of the population studied are presented in Table 1. The group of individuals working under HVAC systems had a higher prevalence of nasal, naso-ocular, sinusitis symptoms, persistent cough and building relation of the symptoms, when compared with individuals under natural ventilation (P < 0.05) (Table 2).
Table 1 Population characteristics according to their ventilation system Variable Response rate Gender (male) Age (years; s.e.m.) Accumulated work time (years; s.e.m.) Smoking (active) Passive smoking At home At work Allergy Familial allergy Previous diagnosis of asthma Previous diagnosis of allergic rhinitis HVAC group 77.1 45.8 29; 0.29 3.00; 0.18 15.4 16.2 7.4 14.1 10.9 32.1 Control group 90.6 85.9 20; 0.09 3.00; 0.11 4.8 14.1 10.4 16.2 12.8 31.8

Values are presented as percentage unless indicated. HVAC, heating, ventilation and air-conditioning, control group, natural ventilation; s.e.m., standard error of mean.

Table 2 Symptoms distribution according to ventilation system Natural ventilation (n 453) n (%) 271 120 114 95 69 82 41 (60) (38.3) (25.2) (21.1) (15.3) (18.1) (9.1) HVAC (n 1541) n (%) 1023 563 538 515 237 370 534 (66.4) (46.9) (35.0) (33.7) (15.4) (24.1) (34.7)

Symptom Nasal symptoms Naso-ocular symptoms Sinusitis symptoms Persistent cough Chest wheezing Breathlessness episodes Building-related symptoms

P-value (v2) 0.011 0.007 <0.001 <0.001 0.959 0.008 <0.001

This work suggests that sealed buildings with airconditioning are associated with a higher prevalence of work-related upper respiratory symptoms than buildings with natural ventilation, in spite of being a tropical country with high relative and absolute humidity rates all around the year. All self-administered questionnaires can lead to bias. Workers experiencing more symptoms and perception of disease are more likely to complete the questionnaire. Dierences in response rates were attributed to dierent work organizations and collaborative prole, however a selection bias cannot be ruled out, in spite of good percentage of overall responders (79.8%). There are some potential information biases such as dierent job organizations factors (public to private practices), job satisfaction, quantity of work and job-related stress and other unknown job-related factors that could inuence the outcomes (Crawford and Bolas, 1996), in spite of the evidence that these factors could not account for the dierences between the buildings (Skov et al., 1989). The participants were informed about the general objective of studying the environmental health, launched together with the non-smoking policy in large public and oce buildings. The blinding of the study objective and the blinding of the division of the groups probably minimized the information bias and other selection biases seems unlikely as the survey was not based on previous sick-buildings symptoms reports. The adequacy of the control population is always questionable, because there are virtually no oce buildings without any kind of air-conditioning or mechanical ventilation. Dierences between both groups such as age, gender, accumulated work time, smoking habits and atopic background were adjusted in the multivariate regression. The reported previous diagnosis of asthma of 10.912.8% and allergic rhinitis from 31.8 to 32.1%, are in accordance with previous prevalence studies of allergic diseases in Brazil (Sole et al., 2001; Vanna et al., 2001). The overall respiratory

64

Association of HVAC and respiratory symptoms in a tropical city

Table 3 Univariate analysis of risk factors associated with respiratory symptoms Outcome Nasal symptoms Naso-ocular symptoms Sinusitis Persistent cough Chest wheezing Breathlessness Building related worsening Working under HVAC 1.33 1.42 1.56 1.90 1.00 1.42 5.38 (1.071.65) (1.101.83) (1.241.97) (1.482.44) (0.751.33) (1.091.85) (3.807.63) Atopic background 9.63 (7.1512.98) 2.80 (2.343.59)a 2.83 (2.323.45)a 2.93 (2.233.85)b 8.24 (6.1511.05)b 17.26 (12.5523.74)b 0.82 (0.661.02)a
a

Female gender 1.57 1.42 1.47 1.43 0.99 1.72 1.59 (1.291.91) (1.101.83) (1.211.79) (1.171.75) (0.771.29) (1.382.14) (1.271.99)

Active smoking habits 0.61 1.22 1.99 0.89 1.60 0.86 0.80 (0.470.79) (0.891.69) (1.452.73) (0.661.19) (1.152.22) (0.621.19) (0.571.13)

Second-hand smoke 1.86 1.29 1.15 1.03 1.28 1.43 2.03 (1.312.64) (0.941.79) (0.891.49) (0.791.34) (0.861.91) (1.021.99) (1.432.89)

Values are presented as OR (95% CI). Analysis of odds ratio for univariate logistic regression. Atopic background: aprevious diagnosis of allergic rhinitis, bprevious diagnosis of asthma.

Table 4 Working with HVAC systems as a risk factor for respiratory symptoms as compared with natural ventilation OR crude 1.32 1.42 1.60 1.90 1.01 1.43 5.33 95% CI crude OR 1.071.64 1.101.83 1.262.03 1.482.43 0.751.35 1.101.87 3.807.47 OR adjustedb 1.59 1.58 1.85 3.04 1.33 1.36 4.92 95% CI adjusted ORb 1.112.28 1.052.38 1.272.71 2.004.63 0.842.11 0.852.17 2.938.27

Outcome Nasal symptoms Naso-ocular symptoms Sinusitis Persistent cough Chest wheezing Breathlessness Building-related worsening
a b

H & La 0.57 0.55 0.94 0.41 0.59 0.00 0.20

Hosmer and Lemeshow goodness of fit. Adjusted by age, gender, atopic background, smoking habits and accumulated work time.

symptoms prevalence in air-conditioned buildings is similar to other studies in Brazil (Costa and Brickus, 2000; Graudenz et al., 2002a; Santos et al., 1992), conrming indoor air related problems as an important issue in Brazil. Atopic related (allergic) conditions such as previous diagnosis of asthma or allergic rhinitis, were consistently associated with IAQ complains as demonstrated in other studies (Bjornsson et al., 1998; Brooks, 1994; Ooi and Goh, 1996), possibly because of a lower threshold of the respiratory mucosal immune system to dierent stimuli, as most studies on allergen exposure assessment and symptoms in oce buildings show low allergen levels and no correlation of symptoms with measured allergens (Graudenz et al., 2002b; Menzies and Bourbeau, 1997). The association of nasal symptoms of itching, pruritus and nasal discharge with HVAC systems, when compared with natural ventilation (OR 1.59, 95% CI 1.112.28) are similar to those reported by Jaakkola and Miettinen, 1995, and to the study of reanalysis of epidemiologic studies by Mendell and Smith (1990). The ocular involvement found in our study have also similar OR to these two previous studies. The positive association of persistent cough to HVAC without humidication systems was signicant (OR 3.04, 95% CI 2.004.63). Coughing is a nonspecic symptom, which could reect irritation or

inammation either from upper or lower airways, but its association to building occupancy is an important marker of poor IAQ (Allermann et al., 2003; Wan and Li, 1999; WHO, 1983). Of great interest is the association of sinusitis symptoms and this type of airconditioning system. Sinusitis can be a result of dierent conditions including viral, bacterial, fungal, toxic and allergic conditions that aect the sinus system. Sinusitis can be aggravated by concomitant rhinitis commonly reported in indoor air surveys, but its diagnosis frequently needs an imaging complementary exam. Lower respiratory symptoms such as wheezing, shortness of breath and tight chest symptoms can be related to IAQ problems (Burge et al., 1985; Jaakkola and Jaakkola, 1999; Mendell et al., 1996). Symptoms of wheezing and shortness of breath were not associated with air-conditioning in our population sample. This is in accordance with another study with similar design (Robertson et al., 1985), suggesting that in absence of conditions such as hypersensivity pneumonitis, or humidier fever, upper airways could be more sensitive to air-conditioning-related conditions than lower airways in this context (Graudenz et al., 2002b). These results should be carefully interpreted as lower respiratory symptoms are rare symptoms and the non-statistically signicance may be misguiding, because of large CIs. Prospective studies and detailed investigation on cases may add additional information on this specic subject. More research in dierent climatic and geographic conditions are needed to conrm IAQ issues as a global matter of public health.
Acknowledgements

This work was partially supported by grants obtained from the following Brazilian institutions: FAPESP (Sao Paulo State Foundation for Research Support) and CNPq (National Research Council). We thank Dr Paulo Hilario Saldiva for insightful discussions. Dr Latorre thanks the CNPq for the research fellowship. 65

Graudenz et al.
References
Allermann, L., Meyer, H.W. et al. (2003) Inflammatory potential of dust from schools and building related symptoms, Occup. Environ. Med., 60, E5(15). Apter, A., Bracker, A. et al. (1994) Epidemiology of the sick building syndrome, J. Allergy Clin. Immunol., 94, 277288. Asher, M.I., Keil, U. et al. (1995) International Study of Asthma and Allergies in Childhood (ISAAC): rationale and methods, Eur. Respir. J., 8, 483491. Bjornsson, E., Janson, C. et al. (1998) Symptoms related to the sick building syndrome in a general population sample: associations with atopy, bronchial hyperresponsiveness and anxiety, Int. J. Tuberc. Lung. Dis., 2, 10231028. Brooks, S.M. (1994) Host susceptibility to indoor air pollution, J. Allergy Clin. Immunol., 94, 344351. Burge, P.S., Finnegan, M. et al. (1985) Occupational asthma in a factory with a contaminated humidifier, Thorax, 40, 248254. Costa, M.F. and Brickus, L.S. (2000) Effect of ventilation systems on prevalence of symptoms associated with sick buildings in Brazilian commercial establishments (in process citation), Arch. Environ. Health, 55, 279283. Crawford, J.O. and Bolas, S.M. (1996) Sick building syndrome, work factors and occupational stress, Scand. J. Work Environ. Health, 22, 243250. Esteves, A.R., Sole, D. et al. (1999a) Adaptation and validity of the ATS-DLD-78-C questionnaire for asthma diagnosis in children under 13 years of age, Braz. Pediatr. News, 1, 18. Esteves, P.C., Tripia, S.G. et al. (1999b) Validacao do questionario ISAAC para rinite alergica perene e sazonal em Curitiba, Rev. Bras. Alerg. Imunopatol., 22, 106113. Ferris, B.G. (1978) Epidemiology standardization project (American Thoracic Society), Am. Rev. Respir. Dis., 118, 1120. Graudenz, G.S., Kalil, J. et al. (2002a) Transporte passivo de alergenos de cao e gato em edif cios de escritorios, Rev. Bras. Alerg. Imunopatol., 25, 99106. Graudenz, G.S., Kalil, J. et al. (2002b) Upper respiratory symptoms associated with aging of the ventilation system in artificially ventilated offices in Sao Paulo, Brazil, Chest, 122, 729735. Jaakkola, M.S. and Jaakkola, J.J. (1999) Office equipment and supplies: a modern occupational health concern? Am. J. Epidemiol., 150, 12231228. Jaakkola, J.J. and Miettinen, P. (1995) Type of ventilation system in office buildings and sick building syndrome, Am. J. Epidemiol., 141, 755765. Ledford, D.K. (1994) Indoor allergens, J. Allergy Clin. Immunol., 94, 327334. Mendell, M.J. and Smith, A.H. (1990) Consistent pattern of elevated symptoms in air-conditioned office buildings: a reanalysis of epidemiologic studies (see comments), Am. J. Public Health, 80, 11931199. Mendell, M.J., Fisk, W.J. et al. (1996) Elevated symptom prevalence associated with ventilation type in office buildings, Epidemiology, 7, 583589. Menzies, D. and Bourbeau, J. (1997) Building-related illnesses, N. Engl. J. Med., 337, 15241531. Menzies, D., Comtois, P. et al. (1998) Aeroallergens and work-related respiratory symptoms among office workers, J. Allergy Clin. Immunol., 101, 3844. Ooi, P.L. and Goh, K.T. (1996) Sick-building syndrome in a tropical city (letter). Lancet, 347, 841842. Robertson, A.S., Burge, P.S. et al. (1985) Comparison of health problems related to work and environmental measurements in two office buildings with different ventilation systems, Br. Med. J. (Clin. Res. Ed.), 291, 373376. Sanitaria, Agencia Nacional de Vigilancia Sanitaria (2003) Norma Regulatoria 9 de 16 de janeiro de2003. Diario da Uniao, Brasilia, Brasil, Ministerio da Saude. Santos, U.P., Rumel, D. et al. (1992) Sick building syndrome in bank employees, Rev. Saude Publica, 26, 400404. Seppanen, O. and Fisk, W.J. (2002) Association of ventilation system type with SBS symptoms in office workers, Indoor Air, 12, 98112. Skov, P., Valbjorn, O. et al. (1989) Influence of personal characteristics, job-related factors and psychosocial factors on the sick building syndrome. Danish Indoor Climate Study Group, Scand. J. Work Environ. Health, 15, 286295. Sole, D., Yamada, E. et al. (2001) International Study of Asthma and Allergies in Childhood (ISAAC): prevalence of asthma and asthma-related symptoms among Brazilian schoolchildren, J. Invest. Allergol. Clin. Immunol., 11, 123128. Spaul, W.A. (1994) Building-related factors to consider in indoor air quality evaluations, J. Allergy Clin. Immunol., 94, 385 389. Vanna, A.T., Yamada, E. et al. (2001) International Study of Asthma and Allergies in Childhood: validation of the rhinitis symptom questionnaire and prevalence of rhinitis in schoolchildren in Sao Paulo, Brazil, Pediatr. Allergy Immunol., 12, 95101. Wan, G.H. and Li, C.S. (1999) Dampness and airway inflammation and systemic symptoms in office building workers, Arch. Environ. Health, 54, 5863. WHO (1983) Indoor air pollutants: exposure and health eects, World Health Organization Reports, Copenhagen, Denmark.

66