Impact of urban atmospheric environment on hospital admissions in the elderly Impacto do ambiente atmosférico urbano nas internações hospitalares de idosos

OBJECTIVE: To analyze the impact of intra-urban atmospheric conditions on circulatory and respiratory diseases in elder adults. METHODS: Cross-sectional study based on data from 33,212 hospital admissions in adults over 60 years in the city of São Paulo, southeastern Brazil, from 2003 to 2007. The association between atmospheric variables from Congonhas airport and bioclimatic index, Physiological Equivalent Temperature, was analyzed according to the district’s socioenvironmental profi le. Descriptive statistical analysis and regression models were used. RESULTS: There was an increase in hospital admissions due to circulatory diseases as average and lowest temperatures decreased. The likelihood of being admitted to the hospital increased by 12% with 1oC decrease in the bioclimatic index and with 1oC increase in the highest temperatures in the group with lower socioenvironmental conditions. The risk of admission due to respiratory diseases increased with inadequate air quality in districts with higher socioenvironmental conditions. CONCLUSIONS: The associations between morbidity and climate variables and the comfort index varied in different groups and diseases. Lower and higher temperatures increased the risk of hospital admission in the elderly. Districts with lower socioenvironmental conditions showed greater adverse health impacts. DESCRIPTORS: Aged. Hospitalization. Cardiovascular Diseases. Respiratory Tract Diseases. Climate Effects. Air Pollution, adverse effects. 695 Rev Saúde Pública 2012;46(4):694-701 Global climate changes pose public health risks. However, there is little evidence based on data to build models to predict risks in the context of climate changes, particularly in tropical cities. Studies on thermal comfort in urban areas apply complex indexes for urban and tourism planning as well as for the assessment of health effects of climate in developed countries. The Physiological Equivalent Temperature (PET) is a complex index that refl ects the interactions between humans and the thermal environment.8,11,10 Knowing local weather phenomena is essential for understanding health impacts on different social groups living in environments with dynamic climatic attributes. In developing countries population growth, urbanization and increasing consumption go together. It is thus important to understand health-related effects of climate oscillations in the cities to foresee potential effects of future climate changes as trends indicate that cities in developing countries will increasingly contribute to global climate changes.a According to Campbell-Lendrum & Corvalán,5 “decision makers RESUMO OBJETIVO: Analisar o impacto das condições climáticas intra-urbanas nas doenças circulatórias e respiratórias em idosos. MÉTODOS: Estudo transversal baseado em 33.212 internações hospitalares de idosos na cidade de São Paulo, SP, de 2003 a 2007. As variáveis atmosféricas coletadas no aeroporto de Congonhas e o índice bioclimático Physiological Equivalent Temperature foram analisados segundo o perfi l socioeconômico da vizinhança. Análise estatística descritiva e modelos de regressão foram usados. RESULTADOS: Houve aumento nas internações por doenças circulatórias quando houve diminuição nas temperaturas médias e mínimas. A probabilidade de internações foi 12% maior com diminuição de 1oC no índice bioclimático e com aumento de 1oC nas temperaturas máximas no grupo com pior perfi l socioeconômico. Houve risco aumentado de internações respiratórias durante má qualidade do ar em distritos com melhor perfi l socioeconômico. CONCLUSÕES: Associações entre morbidade, variáveis climáticas e índice de conforto não apresentaram padrão de comportamento entre os diferentes grupos e doenças. Desconforto para o frio e extremo de calor representaram maior risco para internações de idosos. Distritos com piores condições sociais e ambientais apresentaram maiores impactos à saúde. DESCRIPTORES: Idoso. Hospitalização. Doenças Cardiovasculares. Doenças Respiratórias. Efeitos do Clima. Poluição do Ar, efeitos adversos.

Global climate changes pose public health risks.However, there is little evidence based on data to build models to predict risks in the context of climate changes, particularly in tropical cities.
Studies on thermal comfort in urban areas apply complex indexes for urban and tourism planning as well as for the assessment of health effects of climate in developed countries.The Physiological Equivalent Temperature (PET) is a complex index that refl ects the interactions between humans and the thermal environment. 8,11,10Knowing local weather phenomena is essential for understanding health impacts on different social groups living in environments with dynamic climatic attributes.
In developing countries population growth, urbanization and increasing consumption go together.It is thus important to understand health-related effects of climate oscillations in the cities to foresee potential effects of future climate changes as trends indicate that cities in developing countries will increasingly contribute to global climate changes.a According to Campbell-Lendrum & Corvalán, 5 "decision makers would therefore benefi t from assessments that can assist them to select development policies that can bring synergies or optimize trade-offs between protecting the local and global environment while also bringing health gains." The present study aimed to analyze the impact of intra--urban atmospheric conditions on cardiovascular and respiratory diseases in elder adults.We selected a local scale that is designed to refl ect data from meteorological standard stations (approximately 150 km 2 ).An area surrounding the meteorological station was defi ned, covering 14 districts.Our sample included a diversity of urban spaces to understand the association between intra-urban climate and health inequalities (Figure).

Cross
The impact of atmospheric variables on morbidity in areas with different socioenvironmental conditions was assessed based on the Environmental Atlas of São Paulo.This index was calculated based on the following indicators: population density; population growth rate; median age; median family income; average education level; infant mortality rate; housing stock (m²/inhabitant); percentage of slum dwellers; mortality from external causes; vegetation and urban climate.Regression analysis was performed using SPSS version 18.The 14 districts were combined for a larger sample of hospital data and categorized as higher, intermediate and lower socioenvironmental conditions.
The regression model was built as follows: information of daily admissions was collected in a database; the standardized incidence of daily hospital admissions was calculated based on annual population projections for the age group studied.

Hospital admissions per day
Rate =_________________ × 10,000 inhabitants elderly population living in the southeastern region for that year The median admission rate was calculated and dichotomized into low (patient admissions below the median) and high (patient admissions above the median).
The climate variables selected where those with statistical signifi cance in the logistic regression model.Associations with p≤ 0.05 were included in the logistic regression adjusted model.
Air pollution was categorized as adequate (adequate air quality according to CETESB) or inadequate (barely adequate and inadequate air quality).

RESULTS
Atmospheric variables from Congonhas Airport monitoring stations showed an impact of the urban surrounding on the climate.The results showed an association with environmental atmospheric variables in different ways.

Congonhas Airport
No signifi cant association between hospital admissions due to circulatory diseases in the elderly and meteorological variables was found in the districts with higher socioenvironmental conditions.There was a signifi cant association (p<0.00)with air pollution: the relative risk of high admission rate was 1.36 times greater in days with inadequate air quality (Table 2).
There was a statistically significant association between hospital admissions and relative humidity and air pollution in districts with intermediate socioenvironmental conditions.A high admission rate was 1.6% more likely with every one-unit decrease in relative humidity (p<0.05) and the relative risk of a high rate was 1.43 times higher in days with inadequate air quality.
There were signifi cant associations (p<0.05) with the PET variables and maximum temperatures in districts with lower socioenvironmental conditions.The PET index showed that the relative risk of high admission rates was 12% higher with every 1°C decrease in the comfort index, and about 1.12 higher with every 1ºC increase in maximum temperatures.Air pollution showed no statistically signifi cant association for the group of districts with lower socio-environmental conditions.
Cold days with wide thermal fl uctuations were associated with hospital admissions due to circulatory diseases in the entire southeastern region of São Paulo, regardless of socioenvironmental conditions.
The regression model showed maximum air temperatures and relative humidity as explanatory variables in the group of districts with higher socioenvironmental conditions (Table 3).Relative humidity remained signifi cant (p=0.05) after adjustment.The relative risk of high admission rate was 1.011 with one-unit decrease in relative humidity.There was no association of respiratory disease admissions with atmospheric environmental variables and air pollution in the group of districts with lower socioenvironmental conditions.

DISCUSSION
The intra-urban subdivisions showed complex relationships.There was a strong association between daily hospital admissions due to circulatory diseases and some atmospheric variablesmaximum temperature and PET index -in the districts with lower socioenvironmental conditions.There was also an increased risk of excess admissions in days of cold discomfort as well as days of high maximum temperatures in these districts.In areas with lower socioenvironmental indicators people are more vulnerable to atmospheric conditions in the two extremes: cold and hot weather.No statistically signifi cant association was found between atmospheric variables and hospital admissions due to circulatory conditions in districts with higher socioenvironmental conditions.Relative humidity was associated with a higher risk of hospital admission in districts with intermediate conditions, but with weak statistical significance (p=0.04).Other studies conducted in São Paulo also reported a relationship between colder days and an increase in circulatory diseases. 7,20e etiology of circulatory system diseases is complex and some risk factors include hypertension, smoking, dyslipidemia, diabetes, and obesity.Stressful psychosocial factors as work-related stress, sadness, depression, anxiety, and strain have been associated with greater risk of myocardial infarction. 18In our study, climate parameters constitute a risk factor, which is important to understand in view of global climate changes.
Poor urbanization can aggravate the climate effects on health.The association between socioeconomic factors and environment makes the elderly a vulnerable population.
The analysis of respiratory disease admissions showed a better association with minimum temperature and wind speed in districts with intermediate conditions.
Confounding factors for respiratory diseases in adults include work, smoking, and nutritional status among others.The causal relationship of atmospheric variables is more complex and less evident.
The variable hospital admissions dichotomized into low and high rates solved the problem of days with zero admissions in the regression model, allowing the analysis of groups of districts with different profi les and to understand health inequalities within a large city.
Studies using other methods pointed to decrease and/ or fl uctuations in air temperature, decrease in humidity, and increase in temperature range as aggravating factors of respiratory diseases, but in young children. 1,4,14,16search studies in cities in developing countries 12 including São Paulo showed that low temperatures and high temperature range were associated with great seasonal fl uctuations of mortality, with highest death rates occurring over relatively colder periods.
The population here studied is sensitive to impacts of climate, but this impact was different for disease groups and different districts, and was dependent on socioenvironmental conditions.Risk factors in subtropical cities may differ from those in higher latitudes.
Comfort indicators are increasingly used as exposure parameters to poor health conditions as, compared to comfortable conditions, cold and hot climate stress aggravate more health problems. 6,9,13,19,22ermal discomfort was associated with higher risk of developing circulatory diseases in people ove r 60 years in São Paulo.Hot weather is a factor of great concern in developed cities and some studies indicate high death risk among vulnerable elderly during hot days, especially from cardiopulmonary problems. 2,15study 3 pointed out the health effects of heat in people living in cities of developing countries.An increase in mortality risk per unit of temperature increase in those over 65 years old was found in São Paulo.
São Paulo is characterized by wide fl uctuations in daily and interday changes in temperature, with frequent changes from a very hot to a very cold day.Daily temperature ranges may exceed 20ºC, especially in the fall and spring, and may have a health impact in the elderly. 17study in São Paulo showed more extreme temperatures and more pronounced temperature fl uctuations in slum area than in urban environments with better structure. 21r pollution also increased circulatory and respiratory diseases in the entire population in the southeastern region of the city.However, there was a higher relative risk of high admission rates with inadequate air quality in districts with higher socioenvironmental conditions.This fi nding may be explained by the location of the monitoring station in those districts more exposed to heavy traffi c.
A limitation of the study is that the sample was drawn from SUS hospital data, and refers to a subset of the population that utilizes public hospitals.Non-availability of systematic and reliable information on admissions in private hospitals and with private health insurance limited access to a wider universe.However, as most elderly population cannot afford private health insurance, the study sample may refl ect most hospital admissions in those over 60.Hospitalizations refl ect an acute impact of a health condition, but climate effects on health may have a diversity of effects, from annoyance to weakness, subclinical and clinical damage and even increased mortality.A more complete database is required to assess all impacts of climate on people's health and more accurate.It is necessary to analyze events that do not necessarily lead to hospitalization.More comprehensive information on diseases must be produced and systematized in order to increase knowledge of atmospheric effects with refi ned models and to encourage the search for solutions to minimize health impacts.
In tropical urban areas, in addition to investigating the characteristics particular to the climate and social diversity, it is important to understand the relationship between social and environmental inequalities and health-disease.Studies of urban bioclimatology are complex and an interdisciplinary approach is required to build up knowledge on the impact of climate on human health.The present study found associations between morbidity, climate variables, and comfort index, but diverse and specifi c effects of weather on health.The group of districts studied with lower socioenvironmental conditions showed greater vulnerability to cardiovascular diseases associated to the negative effects of the atmospheric environment.These results may provide input to develop policies for mitigating climate-related risks in cities in the developing world as most studies on health effects of climate have been carried out in cities of developed countries.

Figure .
Figure.Location of Congonhas airport weather station in the southeastern region of the city.São Paulo, Southeastern Brazil.Source: Google Earth 2010.

Table 1 .
Outdoor thermal comfort assessed using the Physiological Equivalent Temperature index.City of São Paulo, southeastern Brazil.

Table 2 .
Atmospheric variables and hospital admissions due to circulatory diseases in the elderly, controlled for air pollution, in the southern/southeastern region of the city.São Paulo, southeastern Brazil, 2003 to 2007.

Table 3 .
Atmospheric variables and hospital admissions due to respiratory diseases in the elderly, controlled for air polluti on, in the southern/southeastern region of the city.São Paulo, southeastern Brazil, 2003 to 2007.Minimum temperature and wind speed remained statistically signifi cant (p<0.05) in the adjusted model for the group of districts with intermediate socioenvironmental conditions.The relative risk of high admission rate was 1.034 with every 1˚C decrease in minimum temperature and elderly were 1.142 more likely to be admitted with decrease in wind speed.The variables daily temperature range and air pollution lost statistical signifi cance in the adjusted model.
PET: Physiological Equivalent Temperature; RH: relative humidity Poor housing conditions such as poor construction, ventilation, and thermal insulation may increase health risks.The literature on this subject is scarce, but future studies may help understand health effects and the potential for adaptation to global climate changes.Research at different geographical scales, from local to microclimate conditions, is needed.The atmospheric environment should be taken into account in environmental health programs.Interventions including urban planning and controlling factors that affect climate (trees, plazas, streets) can reduce emissions of pollutants, among others.At the dwelling level, improvement of thermal insulation may help protect dwellers.Education interventions including awareness campaigns of adverse effects of climate changes and orientation can help protect people against cold/heat and air pollution.