ABSTRACT

OBJECTIVE

The objective of this study has been to analyze whether fine particulate matter (PM_2.5), as well as its synergistic effect with maximum temperature, humidity, and seasons, is associated with morbidity and mortality from cardiovascular diseases.

METHODS

This is an ecological study of time series. We have used as outcomes the daily death and hospitalization records of adults aged 45 years and over from 2009 to 2011 of the municipalities of Cuiabá and Várzea Grande, State of Mato Grosso, Brazil. We have used Poisson regression using generalized additive models, assuming a significance level of 5%. The model has been controlled for temporal trend, seasonality, average temperature, humidity, and season effects. Daily concentrations of PM_2.5 (particulate material with aerodynamic diameter less than 2.5 micrometers) have been obtained by converting the values of optical aerosol thickness. Maximum temperature, humidity, and seasons have been separately included in the model as dummy variables for the analysis of the synergistic effect of PM_2.5 with morbidity and mortality from cardiovascular disease. We have calculated the percentage increase of relative risk (%RR) of deaths and hospitalizations for the linear increase of 10 μg/m³ of PM_2.5.

RESULTS

Between 2009 and 2011, the increase in PM_2.5 was associated with a %RR 2.28 (95%CI 0.53–4.06) for hospitalizations on the same day of exposure and RR% 3.57 (95%CI 0.82–6.38) for deaths with a lag of three days. On hot days, %RR 4.90 (95%CI -0.61–9.38) was observed for deaths. No modification of the effect of PM_2.5 was observed for maximum temperature in relation to hospitalizations. On days with low humidity, %RR was 5.35 (95%CI -0.20–11.22) for deaths and 2.71 (95%CI -0.39–5.92) for hospitalizations. In the dry season, %RR was 2.35 (95%CI 0.59–4.15) for hospitalizations and 3.43 (95%CI 0.58–6.35) for deaths.

CONCLUSIONS

The PM_2.5 is associated with morbidity and mortality from cardiovascular diseases and its effects may be potentiated by heat and low humidity and during the dry season.

Particulate Matter, adverse effects; Air Pollutants; Cardiovascular Diseases, epidemiology; Risk Factors; Seasons; Ecological Studies; Time Series Studies

INTRODUCTION

There is evidence that exposure to fine particulate matter (PM_2.5) is associated with an increased risk of mortality, hospitalization, and exacerbation of cardiovascular diseases (CD). The magnitude of the reported effects, however, is related to the composition, amount, entryway, transport capacity, and deposition of PM in the organism. The PM_2.5 shows greater toxicity because of its multielementarity and because it can both reach the lower portions of the respiratory tract and directly contact the bloodstream⁵5. Franchini M, Mannucci PM. Air pollution and cardiovascular diseases. Thromb Res. 2012;129(3):230-4. https://doi.org/10.1016/j.thromres.2011.10.030
https://doi.org/10.1016/j.thromres.2011.... .

Geographic location and climatic seasonality may also influence the magnitude of adverse effects of PM on human health. Some studies have shown that climate can modify the association of PM_2.5 with morbidity and mortality. Although temperature is the most studied parameter⁸8. Kioumourtzoglou MA, Schwartz J, James P, Dominici F, Zanobetti A. PM2.5 and mortality in 207 US cities: modification by temperature and city characteristics. Epidemiology. 2016;27(2):221-7. https://doi.org/10.1097/EDE.0000000000000422
https://doi.org/10.1097/EDE.000000000000... ^,1010. Meng X, Zhang Y, Zhao Z, Duan X, Xu X, Kan H. Temperature modifies the acute effect of particulate air pollution on mortality in eight Chinese cities. Sci Total Environ. 2012;435-436:215-21. https://doi.org/10.1016/j.scitotenv.2012.07.008
https://doi.org/10.1016/j.scitotenv.2012... ^,1717. Ren C, Tong S. Temperature modifies the health effects of particulate matter in Brisbane, Australia. Int J Biometeorol. 2006;51(2):87-96. https://doi.org/10.1007/s00484-006-0054-7
https://doi.org/10.1007/s00484-006-0054-... ^,2323. Stafoggia M, Schwartz J, Forastiere F, Perucci CA; SISTI Group. Does temperature modify the association between air pollution and mortality? A multicity case-crossover analysis in Italy. Am J Epidemiol. 2008;167(12):1476-85. https://doi.org/10.1093/aje/kwn074
https://doi.org/10.1093/aje/kwn074... , some authors have observed that humidity, atmospheric pressure, and seasons can also act as modifiers of the effect of PM on health events¹⁴14. Qiu H, Yu IT, Wang X, Tian L, Tse LA, Wong TW. Cool and dry weather enhances the effects of air pollution on emergency IHD hospital admissions. Int J Cardiol. 2013;168(1):500-5. https://doi.org/10.1016/j.ijcard.2012.09.199
https://doi.org/10.1016/j.ijcard.2012.09... ^,1616. Ravljen M, Bilban M, Kajfež-Bogataj L, Hovelja T, Vavpotič D. Influence of daily individual meteorological parameters on the incidence of acute coronary syndrome. Int J Environ Res Public Health. 2014;11(11):11616-26. https://doi.org/10.3390/ijerph111111616
https://doi.org/10.3390/ijerph111111616... .

In the Brazilian Amazon and Cerrado, in particular, climatic phenomena intensify the health effects related to atmospheric pollution. In these regions, some ecological studies show that PM is associated with increased morbidity and mortality from respiratory⁶6. Ignotti E, Valente JG, Longo KM, Freitas SR, Hacon SS, Artaxo Netto P. Impact on human health of particulate matter emitted from burnings in the Brazilian Amazon region. Rev Saude Publica. 2010;44(1):121-30. https://doi.org/10.1590/S0034-89102010000100013
https://doi.org/10.1590/S0034-8910201000... ^,2222. Silva AMC, Mattos IE, Freitas SR, Longo KM, Hacon SS. Material particulado (PM2,5) de queima de biomassa e doenças respiratórias no sul da Amazônia brasileira. Rev Bras Epidemiol . 2010;13(2):337-51. https://doi.org/10.1590/S1415-790X2010000200015
https://doi.org/10.1590/S1415-790X201000... and cardiovascular diseases¹²12. Nunes KVR, Ignotti E, Hacon SS. Circulatory disease mortality rates in the elderly and exposure to PM2.5 generated by biomass burning in the Brazilian Amazon in 2005. Cad Saude Publica. 2013;29(3):589-98. https://doi.org/10.1590/S0102-311X2013000300016
https://doi.org/10.1590/S0102-311X201300... ^,1919. Rodrigues PCO, Santos ES, Ignotti E, Hacon SS. Space-time analysis to identify areas at risk of mortality from cardiovascular disease. BioMed Res Int. 2015;2015:ID 841645. https://doi.org/10.1155/2015/841645
https://doi.org/10.1155/2015/841645... in children and the elderly, especially during the dry season, in which accidental or intentional burning of pastures, sugarcane, and forest occurs. However, the modifying effects of climatic seasonality of the region on the morbidity and mortality associated with air pollution have not been investigated yet.

The characterization of the effects of PM_2.5, as well as their combined effect with climatic variables, can better explain the heterogeneity of the effects observed in the region and it can subsidize mitigation and control actions that are more effective and adequate to the needs of the local population in relation to air pollution. The objective of this study was to analyze whether fine particulate matter (PM_2.5), as well as its synergistic effect with maximum temperature, humidity, and seasons are associated with morbidity and mortality from cardiovascular diseases.

METHODS

This is an ecological study of time series of daily records of deaths and hospitalizations for cardiovascular diseases. We selected individuals aged 45 years and over who lived in the municipalities of Cuiabá and Várzea Grande and who were hospitalized and died from CD (Chapter IX of the Tenth Revision of the International Classification of Diseases [ICD-10], codes I00 to I99) between April 23, 2009 and December 31, 2011. The time series from 2009 to 2011, comprised of 983 days, was selected because of the availability of optical aerosol thickness data that enabled us to estimate PM_2.5.

The municipalities of Cuiabá and Várzea Grande, located in the State of Mato Grosso, Brazil, form a conurbation that groups a population of approximately 820,000 inhabitants, which is approximately 90% of the total population of the metropolitan region of the Cuiabá River Valley. The population aged 45 years and over represents approximately 60% of the total population of the cities^aaInstituto Brasileiro de Geografia e Estatística. IBGE Cidades: indicadores e dados populacionais. Rio de Janeiro (DF); IBGE; 2010 [cited 2013 May 6]. Available from: http://www.ibge.gov.br/cidadesat/xtras/perfil..

The climate of the region is limited to two remarkable seasons: the dry and rainy season. Between the months of May to October, the dry period, the prevalence of fires is historically high in the region. The most intense rainy season corresponds to the months of November to April. The cities are well known for their strong daily heat of 32°C, on average. The average maximum temperature reaches 41°C in the hottest months^bb Instituto Nacional de Meteorologia. Dados históricos. Brasília (DF): INMET; 2013 [cited 2013 Jun 15]. Available from: http://www.inmet.gov.br/portal/ php?codmun=510340& search=mato-grosso|cuiaba. Cardiovascular diseases appear as the first cause of general mortality between 2009 and 2011, with an average mortality rate of 135.35 deaths per 100,000 inhabitants, following the mortality profile of the State^ccMinistério da Saúde (BR), Departamento de Informática do SUS (DATASUS). Rede Interagencial de Informações para a Saúde - RIPSA. Indicadores e dados básicos – Brasil -2012 – IDB 2012. Brasília (DF); 2012 [cited 2014 May 13]. Available from: http://tabnet.datasus.gov.br/cgi/idb2012/matriz.htm . In the same period, diseases of the circulatory system were the fifth cause of hospitalization in the State of Mato Grosso and the third cause of hospitalization in Cuiabá and Várzea Grande, with an average hospitalization rate of 46.25 admissions per 10,000 inhabitants^ccMinistério da Saúde (BR), Departamento de Informática do SUS (DATASUS). Rede Interagencial de Informações para a Saúde - RIPSA. Indicadores e dados básicos – Brasil -2012 – IDB 2012. Brasília (DF); 2012 [cited 2014 May 13]. Available from: http://tabnet.datasus.gov.br/cgi/idb2012/matriz.htm .

The records on mortality and hospitalization for CD used as outcomes were obtained, respectively, from the Mortality Information System (SIM/SUS) and the Hospital Information System (SIH/SUS) databases. Information of the daily averages of temperature and humidity came from the National Institute of Meteorology (INMET).

Aerosol Optical Depth (AOD) data were obtained from the Cuiabá-Miranda station (Latitude: -15.43; Longitude: -56.01) available on the Aerosol Robotic Network (AERONET) website. Estimates of PM_2.5 were generated by converting the AOD values (500 nanometers), using a method developed and validated for the Brazilian Amazon and Cerrado^ddPaixão MMA. Propriedades ópticas de aerossóis naturais e de queimadas da Amazônia [dissertation]. São Paulo: Instituto de Física da Universidade de São Paulo; 2011 [cited 2016 Jul 22]. Available from: http://www.teses.usp.br/teses/disponiveis/43/43134/tde-30092011-150552/ . The calculation basis of this method allows us to obtain values of concentrations of PM_2.5 very close to the data measured by air monitoring stations²2. Baars H, Ansmann A, Althausen D, Engelmann R, Heese B, Müller D, et al. Aerosol profiling with lidar in the Amazon Basin during the wet and dry season. J Geophys Res Atmospheres. 2012;117(D21):1-16. https://doi.org/10.1029/2012JD018338
https://doi.org/10.1029/2012JD018338... and it has been used to supply the lack of atmospheric pollution information in some areas¹1. Andrade Filho VS, Artaxo P, Hacon S, Carmo CN, Cirino G. Aerosols from biomass burning and respiratory diseases in children, Manaus, Northern Brazil. Rev Saude Publica. 2013;47(2):239-47. https://doi.org/10.1590/S0034-8910.2013047004011
https://doi.org/10.1590/S0034-8910.20130... ^,1919. Rodrigues PCO, Santos ES, Ignotti E, Hacon SS. Space-time analysis to identify areas at risk of mortality from cardiovascular disease. BioMed Res Int. 2015;2015:ID 841645. https://doi.org/10.1155/2015/841645
https://doi.org/10.1155/2015/841645... . Although the measurements obtained represent average values for the entire atmospheric column with a spatial resolution of up to 10 km, the validation^ddPaixão MMA. Propriedades ópticas de aerossóis naturais e de queimadas da Amazônia [dissertation]. São Paulo: Instituto de Física da Universidade de São Paulo; 2011 [cited 2016 Jul 22]. Available from: http://www.teses.usp.br/teses/disponiveis/43/43134/tde-30092011-150552/ showed that the conversion is adequate for the deforestation arc and the estimated PM_2.5 obtained can be used as an alternative for studies on the impact of PM on human health²2. Baars H, Ansmann A, Althausen D, Engelmann R, Heese B, Müller D, et al. Aerosol profiling with lidar in the Amazon Basin during the wet and dry season. J Geophys Res Atmospheres. 2012;117(D21):1-16. https://doi.org/10.1029/2012JD018338
https://doi.org/10.1029/2012JD018338... .

A time series regression was performed, and we constructed explanatory models to count the hospitalizations and deaths from CD over time. The time series method has been widely applied in ecological epidemiological studies, mainly in the evaluation of the acute effect, because it presents better performance in the analysis of the linear effects of air pollution on mortality from CD¹⁰10. Meng X, Zhang Y, Zhao Z, Duan X, Xu X, Kan H. Temperature modifies the acute effect of particulate air pollution on mortality in eight Chinese cities. Sci Total Environ. 2012;435-436:215-21. https://doi.org/10.1016/j.scitotenv.2012.07.008
https://doi.org/10.1016/j.scitotenv.2012... ^,1313. Pinheiro SLLA, Saldiva PHN, Schwartz J, Zanobetti A. Isolated and synergistic effects of PM10 and average temperature on cardiovascular and respiratory mortality. Rev Saude Publica. 2014;48(6):881-8. https://doi.org/10.1590/S0034-8910.2014048005218
https://doi.org/10.1590/S0034-8910.20140... .

The generalized additive model (GAM) class was used, with Poisson regression, assuming a significance level of 5%. The temporal trend and the seasonality of the series were controlled introducing the following variables: days of the week and a natural cubic spline of the time with nine degrees of freedom. For the meteorological variables, we used the natural cubic spline of the average temperature and relative humidity, with three and two degrees of freedom, respectively. In the modeling process, we sought to reduce the Akaike information criterion (AIC) and improve partial autocorrelation (PACF).

We chose to use the GAM model because we do not need to define a priori the relationships and structures between the health indicator and the explanatory variables. In addition, we used the ARES2 library available in the R application as a tool to define these relationships. The Poisson regression model, with logarithmic link function, was chosen because the mortality data obtained by counting have a Poisson probability distribution⁷7. Junger W, Leon AP. Ares: a library for time series analysis in air pollution and health effects studies using R. Epidemiology. 2009;20(6):S217. https://doi.org/10.1097/01.ede.0000362727.55583.4a
https://doi.org/10.1097/01.ede.000036272... ^,1313. Pinheiro SLLA, Saldiva PHN, Schwartz J, Zanobetti A. Isolated and synergistic effects of PM10 and average temperature on cardiovascular and respiratory mortality. Rev Saude Publica. 2014;48(6):881-8. https://doi.org/10.1590/S0034-8910.2014048005218
https://doi.org/10.1590/S0034-8910.20140... .

The percentage increase of relative risk (%RR) of deaths and hospitalizations for CD was calculated. We investigated associations of current day exposures with single lag and the cumulative effect of up to ten days using a polynomial distributed lag model. The 10-day lag period was selected to allow a more accurate estimate of the effects of PM_2.5 on mortality and hospitalizations for CD, as these effects may not occur immediately or on the same day as the exposure⁵5. Franchini M, Mannucci PM. Air pollution and cardiovascular diseases. Thromb Res. 2012;129(3):230-4. https://doi.org/10.1016/j.thromres.2011.10.030
https://doi.org/10.1016/j.thromres.2011.... ^,1818. Roberts S. Interactions between particulate air pollution and temperature in air pollution mortality time series studies. Environ Res. 2004;96(3):328-37. https://doi.org/10.1016/j.envres.2004.01.015
https://doi.org/10.1016/j.envres.2004.01... ^,2323. Stafoggia M, Schwartz J, Forastiere F, Perucci CA; SISTI Group. Does temperature modify the association between air pollution and mortality? A multicity case-crossover analysis in Italy. Am J Epidemiol. 2008;167(12):1476-85. https://doi.org/10.1093/aje/kwn074
https://doi.org/10.1093/aje/kwn074... . The %RR corresponds to the linear increase of 10 μg/m³in the levels of PM_2.5.

The synergistic effect between PM_2.5 and maximum temperature, humidity, and season was tested by stratifying the final models built for hospitalizations and deaths. The stratification method consists in the insertion of a dummy variable in the final models, observing the new characteristics of the same model with stratification⁷7. Junger W, Leon AP. Ares: a library for time series analysis in air pollution and health effects studies using R. Epidemiology. 2009;20(6):S217. https://doi.org/10.1097/01.ede.0000362727.55583.4a
https://doi.org/10.1097/01.ede.000036272... . We chose to use stratification with this method to use fewer parameters and to offer a simple and quantitative comparison of the estimated effects of pollutants in the different strata in relation to other methods to detect the modifying effect¹⁰10. Meng X, Zhang Y, Zhao Z, Duan X, Xu X, Kan H. Temperature modifies the acute effect of particulate air pollution on mortality in eight Chinese cities. Sci Total Environ. 2012;435-436:215-21. https://doi.org/10.1016/j.scitotenv.2012.07.008
https://doi.org/10.1016/j.scitotenv.2012... ^,1313. Pinheiro SLLA, Saldiva PHN, Schwartz J, Zanobetti A. Isolated and synergistic effects of PM10 and average temperature on cardiovascular and respiratory mortality. Rev Saude Publica. 2014;48(6):881-8. https://doi.org/10.1590/S0034-8910.2014048005218
https://doi.org/10.1590/S0034-8910.20140... . Three dummy variables were included in the model: (i) 90th percentile of the maximum temperature (37.9°C), (ii) 10th percentile of relative humidity (54.5%), and (iii) dry (May – October) and rainy season (January to March, November and December). The statistical significance of the differences between the estimates of the effect among the strata established by the dummy variables was determined by calculating the 95% confidence interval (95% CI) and the p-value⁷7. Junger W, Leon AP. Ares: a library for time series analysis in air pollution and health effects studies using R. Epidemiology. 2009;20(6):S217. https://doi.org/10.1097/01.ede.0000362727.55583.4a
https://doi.org/10.1097/01.ede.000036272... .

The analyses were performed in the R 3.0.2 application using the Ares2 library⁷7. Junger W, Leon AP. Ares: a library for time series analysis in air pollution and health effects studies using R. Epidemiology. 2009;20(6):S217. https://doi.org/10.1097/01.ede.0000362727.55583.4a
https://doi.org/10.1097/01.ede.000036272... . This study has been approved by the Ethics Committee of the Escola Nacional de Saúde Pública (CAAE 18634613.0.0000.5240).

RESULTS

During the study period, 8,610 hospitalizations and 3,024 deaths were recorded. Particulate material and relative humidity showed a heterogeneous distribution during the year, presenting standard deviations (SD) of 15.66 μg/m³ and 11.35%, respectively. The maximum values of the average and maximum temperatures are almost three times higher than their respective minimum values (Table 1).

Average temperature and maximum temperature showed a direct correlation of approximately 0.25 with PM_2.5. Humidity presented an inverse correlation with all variables, being strongly associated with maximum temperature (-0.600). Deaths and hospitalizations for CD presented a linear correlation below 0.100 with the meteorological and pollution variables (Table 2).

In the modeling by the simple lag method, we observed a cyclical effect of reduction and increase every three days, both for hospitalizations and for deaths associated with the linear increase of PM2.5. However, the effect was more acute for hospitalizations. The hospitalizations for CD presented %RR 2.28 (95%CI 0.53–4.06) for the same day of exposure (Lag0) and %RR 2.00 (95%CI 0.31–3.71) for the sigle lag of one day (Lag1). Mortality presented %RR 3.57 (95%CI 0.82–6.38) for the sigle lag of three days (Lag3) (Figure 1).

The accumulated %RR for the period of ten days was 1.81 (95%CI 0.03–3.61) for deaths and 2.64 (95%CI 1.60–3.69) for hospitalizations related to the increase in PM_2.5. The distributed lag method reinforced the acute effect pattern for hospitalizations. There was an increase in %RR of mortality from the second day of exposure and decrease in %RR for hospitalizations throughout the days (Figure 2).

In the dry season, approximately 50% of the observations of maximum temperature were between 34.8°C and 39.10°C, while approximately 50% of the relative humidity data were between 66% and 51%. The rainy season had approximately 50% of the observations between 33°C and 36.4°C, while approximately 50% of the relative humidity data were between 77% and 65%. The average PM_2.5 was 18.71 μg/m³ (SD = 19.59 μg/m³) and 14.88 μg/m³ (SD = 7.18 μg/m³) in the dry and rainy seasons, respectively.

The results of the analyses on the modifying effect show that temperature, humidity, and season can present synergistic effects of both deaths and hospitalizations for CD related to PM_2.5. The %RR of mortality from CD related to PM_2.5 was 4.90 (95%CI -0.61–9.38) on days whose maximum temperature was above 37.9°C. The %RR of mortality and hospitalizations for CD related to PM_2.5 were 5.35 (95%CI -0.20–11.22) and 2.71 (95%CI -0.39–5.92), respectively, on days with relative humidity below 54.5%. During the dry season, we observed %RR of 3.43 (95%CI 0.58–6.35) for mortality and %RR 2.35 (95%CI 0.59–4.15) for hospitalizations. We observed no modification of the effect of PM_2.5 by the maximum temperature related to hospitalizations (Table 3).

DISCUSSION

The results of this study reinforce findings from the international⁵5. Franchini M, Mannucci PM. Air pollution and cardiovascular diseases. Thromb Res. 2012;129(3):230-4. https://doi.org/10.1016/j.thromres.2011.10.030
https://doi.org/10.1016/j.thromres.2011.... and Brazilian¹¹11. Nardocci AC, Freitas CU, Ponce de Leon ACM, Monteiro AC, Junger WL, Gouveia NC. Poluição do ar e doenças respiratórias e cardiovasculares: estudo de séries temporais em Cubatão, São Paulo, Brasil. Cad Saude Publica 2013;29(9):1867-76. https://doi.org/10.1590/0102-311X00150012
https://doi.org/10.1590/0102-311X0015001... ^,2020. Romieu I, Gouveia N, Cifuentes LA, Leon AP, Junger W, Vera J, et al. Multicity study of air pollution and mortality in Latin America (the ESCALA study). Res Rep Health Eff Inst. 2012;(171):5-86 literature by pointing out the relationship between PM_2.5 and the increased risk of hospitalization and mortality from CD in adults aged 45 years and over. The adverse effects of PM on hospitalizations for CD coincided with the day of the increase of pollutant (Lag0) and the risk decreased on subsequent days, while the effects of PM on mortality from CD increased after the increase of pollutant, culminating in mortality mainly after the third day of exposure (lag3). This relationship corroborates the literature, which indicates that cardiovascular diseases are related to more outcomes that are nonspecific and with many-associated causes⁵5. Franchini M, Mannucci PM. Air pollution and cardiovascular diseases. Thromb Res. 2012;129(3):230-4. https://doi.org/10.1016/j.thromres.2011.10.030
https://doi.org/10.1016/j.thromres.2011.... . Thus, although both outcomes are related to more severe cases, hospitalizations are acute, whereas mortality may be associated with the complication of the diseases²⁰20. Romieu I, Gouveia N, Cifuentes LA, Leon AP, Junger W, Vera J, et al. Multicity study of air pollution and mortality in Latin America (the ESCALA study). Res Rep Health Eff Inst. 2012;(171):5-86.

Heat had a synergistic effect with PM_2.5 on mortality from CD in this study. This synergism between temperature and mortality from CD is still controversial in the scientific literature; however, some authors have observed that adverse effects of particulate matter may be more apparent in warmer seasons⁸8. Kioumourtzoglou MA, Schwartz J, James P, Dominici F, Zanobetti A. PM2.5 and mortality in 207 US cities: modification by temperature and city characteristics. Epidemiology. 2016;27(2):221-7. https://doi.org/10.1097/EDE.0000000000000422
https://doi.org/10.1097/EDE.000000000000... ^,2323. Stafoggia M, Schwartz J, Forastiere F, Perucci CA; SISTI Group. Does temperature modify the association between air pollution and mortality? A multicity case-crossover analysis in Italy. Am J Epidemiol. 2008;167(12):1476-85. https://doi.org/10.1093/aje/kwn074
https://doi.org/10.1093/aje/kwn074... and at higher temperatures¹⁰10. Meng X, Zhang Y, Zhao Z, Duan X, Xu X, Kan H. Temperature modifies the acute effect of particulate air pollution on mortality in eight Chinese cities. Sci Total Environ. 2012;435-436:215-21. https://doi.org/10.1016/j.scitotenv.2012.07.008
https://doi.org/10.1016/j.scitotenv.2012... ^,1717. Ren C, Tong S. Temperature modifies the health effects of particulate matter in Brisbane, Australia. Int J Biometeorol. 2006;51(2):87-96. https://doi.org/10.1007/s00484-006-0054-7
https://doi.org/10.1007/s00484-006-0054-... . Ren and Tong¹⁷17. Ren C, Tong S. Temperature modifies the health effects of particulate matter in Brisbane, Australia. Int J Biometeorol. 2006;51(2):87-96. https://doi.org/10.1007/s00484-006-0054-7
https://doi.org/10.1007/s00484-006-0054-... have observed %RR of 2.32 (95%CI -0.26–4.97) for hospitalizations for CD and 6.95 (95%CI 0.95–13.33) for mortality from CD related to PM₁₀ on days with temperature above 27°C in Brisbane, Australia, between 1996 and 2001. Meng et al.¹⁰10. Meng X, Zhang Y, Zhao Z, Duan X, Xu X, Kan H. Temperature modifies the acute effect of particulate air pollution on mortality in eight Chinese cities. Sci Total Environ. 2012;435-436:215-21. https://doi.org/10.1016/j.scitotenv.2012.07.008
https://doi.org/10.1016/j.scitotenv.2012... have observed %RR of 1.57% (95%CI 0.69–2.46) for cardiovascular mortality related to PM₁₀ on days with temperature above 30°C (> 95th percentile) for eight cities in China between 2001 and 2008, whose RR% was 3.19% (95%CI 2.55–3.84) for the city of Wuhan. In the United States, an ecological study carried out for 207 cities has observed an increase in mortality from CD related to PM_2.5 in cities with higher average temperatures⁸8. Kioumourtzoglou MA, Schwartz J, James P, Dominici F, Zanobetti A. PM2.5 and mortality in 207 US cities: modification by temperature and city characteristics. Epidemiology. 2016;27(2):221-7. https://doi.org/10.1097/EDE.0000000000000422
https://doi.org/10.1097/EDE.000000000000... .

Modification of the effects of PM_2.5 by high temperatures on human health may be related to the direct or indirect response of the organism to heat stress¹⁸18. Roberts S. Interactions between particulate air pollution and temperature in air pollution mortality time series studies. Environ Res. 2004;96(3):328-37. https://doi.org/10.1016/j.envres.2004.01.015
https://doi.org/10.1016/j.envres.2004.01... . Body thermoregulation is directly linked to the circulatory regulation of an individual. In this way, hot days can overload the body temperature-regulating system and change the physiological response to toxic agents, increasing individual vulnerability to the effects of PM⁴4. Cheng X, Su H. Effects of climatic temperature stress on cardiovascular diseases. Eur J Intern Med. 2010;21(3):164-7. https://doi.org/10.1016/j.ejim.2010.03.001
https://doi.org/10.1016/j.ejim.2010.03.0... ^,1717. Ren C, Tong S. Temperature modifies the health effects of particulate matter in Brisbane, Australia. Int J Biometeorol. 2006;51(2):87-96. https://doi.org/10.1007/s00484-006-0054-7
https://doi.org/10.1007/s00484-006-0054-... . On warmer days, a greater exposure to air pollution may occur because of the propensity to keep windows open or to spend more time outdoors²³23. Stafoggia M, Schwartz J, Forastiere F, Perucci CA; SISTI Group. Does temperature modify the association between air pollution and mortality? A multicity case-crossover analysis in Italy. Am J Epidemiol. 2008;167(12):1476-85. https://doi.org/10.1093/aje/kwn074
https://doi.org/10.1093/aje/kwn074... . Some studies suggest that high temperatures can also be considered an indirect measure of the composition of urban atmospheric particles⁸8. Kioumourtzoglou MA, Schwartz J, James P, Dominici F, Zanobetti A. PM2.5 and mortality in 207 US cities: modification by temperature and city characteristics. Epidemiology. 2016;27(2):221-7. https://doi.org/10.1097/EDE.0000000000000422
https://doi.org/10.1097/EDE.000000000000... , since they are strongly associated with organic and elemental sulfate and carbon²⁴24. Tai APK, Mickley LJ, Jacob DJ. Correlations between fine particulate matter (PM2.5) and meteorological variables in the United States: implications for the sensitivity of PM2.5 to climate change. Atmos Environ. 2010;44(32):3976-84. https://doi.org/10.1016/j.atmosenv.2010.06.060
https://doi.org/10.1016/j.atmosenv.2010.... , ozone³3. Bell ML, McDermott A, Zeger SL, Samet JM, Dominici F. Ozone and short-term mortality in 95 US urban communities, 1987-2000. JAMA. 2004;292(19):2372-8. https://doi.org/10.1001/jama.292.19.2372
https://doi.org/10.1001/jama.292.19.2372... , and the concentration of semi-volatile particles²¹21. Seinfeld JH, Pandis SN. Atmospheric chemistry and physics: from air pollution to climate change. Hoboken (NJ): Wiley; 2006..

Our results showed evidence that the action of PM_2.5 on hospitalizations and mortality from CD can be exacerbated on days of low relative humidity (below 54.5%). The effects of humidity related to the prevalence of hospitalizations and mortality, as well as their impact on the effects of pollution, are little reported in the literature. Qiu et al.¹⁴14. Qiu H, Yu IT, Wang X, Tian L, Tse LA, Wong TW. Cool and dry weather enhances the effects of air pollution on emergency IHD hospital admissions. Int J Cardiol. 2013;168(1):500-5. https://doi.org/10.1016/j.ijcard.2012.09.199
https://doi.org/10.1016/j.ijcard.2012.09... , in China, have observed an increase of 1.67 (95%CI 1.26–2.08) in emergency hospitalizations for ischemic heart diseases related to PM₁₀ on days when relative humidity was below 80%. Ravljen et al.¹⁶16. Ravljen M, Bilban M, Kajfež-Bogataj L, Hovelja T, Vavpotič D. Influence of daily individual meteorological parameters on the incidence of acute coronary syndrome. Int J Environ Res Public Health. 2014;11(11):11616-26. https://doi.org/10.3390/ijerph111111616
https://doi.org/10.3390/ijerph111111616... , in Slovenia, have identified that the increase in 1% in the average daily humidity decreases the incidence of acute coronary syndrome by approximately 3 ‰ (95%CI -6.00–-1.00). In Cuiabá and Várzea Grande, Brazil, Rodrigues et al.¹⁹19. Rodrigues PCO, Santos ES, Ignotti E, Hacon SS. Space-time analysis to identify areas at risk of mortality from cardiovascular disease. BioMed Res Int. 2015;2015:ID 841645. https://doi.org/10.1155/2015/841645
https://doi.org/10.1155/2015/841645... have observed that high-risk temporal clusters for mortality from CD had higher daily averages of PM_2.5 and lower daily averages of relative humidity. However, the comparison between these results is difficult because of the great methodological differences.

Seasonality is described by several authors as a modifier of the effects of PM associated with hospitalizations and mortality from CD⁸8. Kioumourtzoglou MA, Schwartz J, James P, Dominici F, Zanobetti A. PM2.5 and mortality in 207 US cities: modification by temperature and city characteristics. Epidemiology. 2016;27(2):221-7. https://doi.org/10.1097/EDE.0000000000000422
https://doi.org/10.1097/EDE.000000000000... ^,2323. Stafoggia M, Schwartz J, Forastiere F, Perucci CA; SISTI Group. Does temperature modify the association between air pollution and mortality? A multicity case-crossover analysis in Italy. Am J Epidemiol. 2008;167(12):1476-85. https://doi.org/10.1093/aje/kwn074
https://doi.org/10.1093/aje/kwn074... ^,2424. Tai APK, Mickley LJ, Jacob DJ. Correlations between fine particulate matter (PM2.5) and meteorological variables in the United States: implications for the sensitivity of PM2.5 to climate change. Atmos Environ. 2010;44(32):3976-84. https://doi.org/10.1016/j.atmosenv.2010.06.060
https://doi.org/10.1016/j.atmosenv.2010.... . Some studies point to seasons as indirect measures of temperature, such as Kioumourtzoglou et al.⁸8. Kioumourtzoglou MA, Schwartz J, James P, Dominici F, Zanobetti A. PM2.5 and mortality in 207 US cities: modification by temperature and city characteristics. Epidemiology. 2016;27(2):221-7. https://doi.org/10.1097/EDE.0000000000000422
https://doi.org/10.1097/EDE.000000000000... , in the USA, and Staffogia et al.²³23. Stafoggia M, Schwartz J, Forastiere F, Perucci CA; SISTI Group. Does temperature modify the association between air pollution and mortality? A multicity case-crossover analysis in Italy. Am J Epidemiol. 2008;167(12):1476-85. https://doi.org/10.1093/aje/kwn074
https://doi.org/10.1093/aje/kwn074... , in Italy, who have found an increase in mortality from CD related to PM_2.5 and PM₁₀, respectively, in the summer. Atmospheric chemical speciation studies suggest that the prevalence of more toxic components may occur at higher concentrations during different times of the year, according to the climatic and pollutant emission characteristics of each site, with rainfall being the most effective mechanism for the deposition of pollutants¹⁵15. Qiu H, Yu IT, Wang X, Tian L, Tse LA, Wong TW. Differential effects of fine and coarse particles on daily emergency cardiovascular hospitalizations in Hong Kong. Atmos Environ. 2013;64:296-302. https://doi.org/10.1016/j.atmosenv.2012.09.060
https://doi.org/10.1016/j.atmosenv.2012.... ^,2424. Tai APK, Mickley LJ, Jacob DJ. Correlations between fine particulate matter (PM2.5) and meteorological variables in the United States: implications for the sensitivity of PM2.5 to climate change. Atmos Environ. 2010;44(32):3976-84. https://doi.org/10.1016/j.atmosenv.2010.06.060
https://doi.org/10.1016/j.atmosenv.2010.... ^,2525. Villalobos AM, Barraza F, Jorquera H, Schauer JJ. Chemical speciation and source apportionment of fine particulate matter in Santiago, Chile, 2013. Sci Total Environ. 2015;512-513:133-42. https://doi.org/10.1016/j.scitotenv.2015.01.006
https://doi.org/10.1016/j.scitotenv.2015... . The findings of this study show that the dry season may increase both hospitalizations and deaths from CD related to PM_2.5, which is compatible with both hypotheses, firstly because the dry season also had higher temperatures than the rainy season in our database, and secondly because the dry season is also the period with the highest prevalence of fires in the municipalities and in the North and Midwest regions of Brazil.

It should be taken into account that other sources of air pollution may exist, such as vehicular traffic, burning of urban waste, or the transport of air pollutants from biomass burning and industrial emissions from neighboring cities. Although they also had an important influence on morbidity and mortality from CD, these sources were not included in our analyses, as no data were available on these variables. The local climatic condition must also be considered. Cuiabá and Várzea Grande are located in a depression with extensive Chapadões (mountain ranges) at their edge and present great periodicity of atmospheric stability (clean sky and low wind speed), which can hinder the dispersion of pollutants and increase the incidence of heat islands and thermal inversion episodes. In addition, the dry season may characterize an indirect measure of the effects of low humidity¹⁹19. Rodrigues PCO, Santos ES, Ignotti E, Hacon SS. Space-time analysis to identify areas at risk of mortality from cardiovascular disease. BioMed Res Int. 2015;2015:ID 841645. https://doi.org/10.1155/2015/841645
https://doi.org/10.1155/2015/841645... ^,2222. Silva AMC, Mattos IE, Freitas SR, Longo KM, Hacon SS. Material particulado (PM2,5) de queima de biomassa e doenças respiratórias no sul da Amazônia brasileira. Rev Bras Epidemiol . 2010;13(2):337-51. https://doi.org/10.1590/S1415-790X2010000200015
https://doi.org/10.1590/S1415-790X201000... , also observed in this study.

As limitations of the study, we can highlight the filling bias inherent to the use of the database on hospitalizations and mortality of the Brazilian Unified Health System (SUS) that may not accurately portray the distribution of the variables studied. However, studies have indicated the reliability of the data contained in health information systems⁹9. Lemos C, Chaves LDP, Azevedo ALCS. Sistemas de informação hospitalar no âmbito do SUS: revisão integrativa de pesquisas. Rev Eletron Enferm [Internet]. 2010 [cited 2015 Dec 17];12(1):177-85. Available from: http://www.fen.ufg.br/revista/v12/n1/v12n1a22.htm
http://www.fen.ufg.br/revista/v12/n1/v12... . Another important point is related to the use of estimated values for PM, obtained by remote sensing techniques, which may underestimate or not correspond to the real values of personal exposure; however, its use is feasible and advantageous both for this study and for remote regions that do not have atmospheric monitoring stations¹1. Andrade Filho VS, Artaxo P, Hacon S, Carmo CN, Cirino G. Aerosols from biomass burning and respiratory diseases in children, Manaus, Northern Brazil. Rev Saude Publica. 2013;47(2):239-47. https://doi.org/10.1590/S0034-8910.2013047004011
https://doi.org/10.1590/S0034-8910.20130... ^,22. Baars H, Ansmann A, Althausen D, Engelmann R, Heese B, Müller D, et al. Aerosol profiling with lidar in the Amazon Basin during the wet and dry season. J Geophys Res Atmospheres. 2012;117(D21):1-16. https://doi.org/10.1029/2012JD018338
https://doi.org/10.1029/2012JD018338... . It is important to mention that the ecological character of our analyses does not allow us to identify any causal effect; that is, the association observed between the aggregates does not mean that the same association occurs at the individual level.

We concluded that PM_2.5 has an association with hospitalizations and mortality from CD in medium-sized municipalities. Low relative humidity and dry season may increase both hospitalizations and deaths from CD related to PM_2.5. Heat is related to the increased risk of mortality from CD associated with PM_2.5.

This is the first study that addresses the interactions of climatic factors with pollution in the Cerrado region. The use of estimated pollution data, obtained by remote sensing techniques, provides results compatible with those found in other ecological studies, indicating that this tool is effective in obtaining the concentrations of PM. The results of this research can be used for an evaluation on the improvement of the quality of the data on mortality from CD and air pollution to base future studies and, mainly, help in the planning of strategies to mitigate the impact of air pollution in human health.

REFERENCES

Publication Dates

History