Evidence-based public health policy and practice

The effect of air pollution on pneumonia-related

emergency department visits in a region of
extensive sugar cane plantations: a 30-month
time-series study
Marcos Abdo Arbex,1,2 Luiz Alberto Amador Pereira,2,3 Regiane Carvalho-Oliveira,2
Paulo Hilrio do Nascimento Saldiva,2 Alfsio Lus Ferreira Braga2,3
1

Pulmonology Division,
Medical School of Araraquara
University Center, Araraquara,
Brazil
2
Environmental Epidemiology
Study Group, Department of
Pathology, Laboratory of
Experimental Air Pollution,
University of So Paulo, Faculty
of Medical Sciences, So
Paulo, Brazil
3
Environmental Exposure and
Risk Assessment Group,
Collective Health Post-Graduate
Program, Catholic University of
Santos, Santos, Brazil
Correspondence to
Dr Marcos Abdo Arbex, Rua
Voluntrios da Ptria, 2104,
Centro, Araraquara, So Paulo,
CEP 14801-320, Brazil;
arbexma@techs.com.br
Received 2 December 2013
Revised 20 March 2014
Accepted 24 March 2014
Published Online First
29 April 2014

ABSTRACT
Background In Brazil, many cities are surrounded by
sugar cane plantations, and when these plantations are
burnt prior to harvesting, millions of people are exposed
to the smoke from these res from May to November
every year.
Methods A daily time-series regression analysis was
conducted in a city located in the sugar cane plantation
region of So Paulo State, Brazil, between 1 February
2005 and 31 July 2007. The percentage increase in the
number of pneumonia-related emergency department
visits (PEDV) associated with a 10 mg/m3 increase in the
total suspended particles (TSP) concentration was
measured, including any effects that were delayed for up
to 6 days.
Results A total of 1505 PEDV (a median of two events
per day) were analysed. During the burning period, there
was an acute effect that began on the day of exposure
and remained for 2 days. An increase of 6% (95% CI
2.4 to 9.9) in PEDV was observed for the 2 days
following the TSP increase. This pattern and the size of
the effect were similar to those observed for the whole
period and also during the non-burning period.
Conclusions Increases in TSP concentrations were
found to be associated with increased PEDV in a region
affected by air pollution from sugar cane burning. This
nding reinforces the need for polices and efforts to ban
sugar cane burning prior to harvesting.

particularly in So Paulo State, which is responsible

for approximately 60% of Brazilian sugar cane production. Today, sugar cane plantations cover
approximately 7 million hectares of Brazilian land.5
In general, biomass fuels are considered to be
cleaner alternatives to petroleum-derived fuels, and
the air quality in the urban centres of Brazil has
improved following the conversion of most automobiles to an ethanol fuel source. Despite this
benet in urban centres, in Araraquara, a city
located in So Paulo State in the southeast of Brazil
(21470 4000 S, 48100 3200 W), 195 815 inhabitants in
2005, where sugar cane elds cover 50% of the
municipality area, the burning of the crops has
been associated with increases in the use of emergency room inhalation therapy and hospital admissions for asthma, hypertension.58 Also, a recent
study has shown the relationship between sugar
cane burning and global warming.9 However, it
remains unclear how sugar cane burning contributes to the exacerbations of respiratory infections.
The current study was designed to evaluate the
association between the daily variation uctuations
in TSP during sugar cane plantation burning and
non-burning periods and the incidence of
pneumonia-related emergency department visits
(PEDV) in the city of Araraquara.

METHODS
INTRODUCTION

To cite: Arbex MA,

Pereira LAA, CarvalhoOliveira R, et al. J Epidemiol
Community Health
2014;68:669674.

Particulate air pollution is associated with an

increase in cardio respiratory morbimortality.13
Since the world oil crisis began 1970s, Brazil has
developed ethanol made from sugar cane to be
used as an alternative fuel.4 However, 60% of
Brazils sugar cane crops are manually harvested,
and for productivity and safety reasons, the entire
plantation is burned before harvesting, which
occurs every year from May to November.4 Sugar
cane burning produces large amounts of particles
and toxic gases that modify the environmental
characteristics of the areas where the sugar cane is
planted, harvested and industrialised. In contrast to
the sporadic nature and the short duration of forest
res, the cities surrounding the sugar cane plantations are exposed to the air pollution generated by
biomass burning continuously for at least 6 months
of every year.5 Currently, Brazil is the worlds
largest sugar cane producer, and this production is
concentrated mainly in the southeast region,

Arbex MA, et al. J Epidemiol Community Health 2014;68:669674. doi:10.1136/jech-2013-203709

This was an ecological time-series study using secondary health data. We obtained total daily records
of pertaining to emergency department visits for
pneumonia (International Code of Diseases, 10th
Revision J 1218) from one of the main hospitals
in Araraquara, So Paulo State, Brazil, from 1
February 2005 to 31 July 2007. The hospital
included in the study provides primary and specialised healthcare to the population that has one specic private health insurance plan, which covers
44% of Araraquara inhabitants. A PEDV was
dened as an emergency department visit in which
pneumonia was the primary diagnosis. We included
only records from the So Paulo Hospital for two
reasons: uniformity of clinical background and a
well-dened medical protocol for diagnosing and
treating pneumonia; clinical suspicion of the
disease needs to be conrmed by a chest X-ray.
After pneumonia diagnosis, the patients were hospitalised or referred to outpatient treatment,
depending on the severity of the disease.
669

Evidence-based public health policy and practice

Every year, based on the weather forecasts and crop characteristics, the sugar cane producers determine the harvesting period,
which usually occurs between the end of April and the end of
November. For the period under study (between 2005 and
2007), harvesting occurred from the nal week of April through
the end of November. During the 912 days in the current study,
there were 371 days without sugar cane burning and 541 days
with sugar cane burning.
We obtained the daily concentration of TSP (g/m3) using a
Handi-Vol sampler (Energtica, Brazil) with breglass lters and
a ow rate of 3 L/min. The equipment was placed in downtown
Araraquara in a location that was protected from rain. The
lters were weighed before and after every 24 h particle collection. The daily concentration of TSP in mg/m3 was estimated
using the following equation:
PC

DM
TV24h

In this equation, PC is the TSP concentration (g/m3) under

standard conditions of temperature and barometric pressure,
M is the mass difference of the lter weight (g) before and
after use, and TV24h is the total volume of sampled air in the 24
h period (m3). The local airport provided daily measurements
of temperature and humidity.
The daily number of PEDVs was used as the dependent variable in the generalised linear Poisson regression models,10 and
the TSP concentration was the independent variable. A natural
cubic spline11 was adopted to control for long-time trends.
Based on the seasonal pattern of occurrence of PEDVs and the
length of the time series, we used 14 degrees of freedom (df )
for the smoothing of the entire time trend in order to minimise
the sum of partial residual correlation coefcients and Akaike
Information Criterion (AIC).12 Sensitivity analysis using 1216
df in the smooth function did not change effect estimations.
Analysis of partial correlation coefcients has shown that there
was no serial correlation from the residuals and the use of autoregressive terms was not necessary.
Araraquara is a warm city, and a previous analysis had shown
that temperature is linearly correlated with respiratory diseases.8
However, we decided to explore the relation between PEDVs and
temperature using linear terms, lag zero to ve and
moving-averages and smooth functions. As this relation has been
shown to be linear and not delayed, we have decided to use linear
terms (lags 0 and 1) to control for the effects of temperature and
humidity on PEDVs. In addition, an indicator for the days of the
week was included to control for short-term trend and any unforeseen circumstances such as strikes at public institutions, natural disasters that could impede a person seeking medical attention.
The lag structures between air pollution and health have been
analysed using different approaches and time lags. This study
used the lag structure covering 06 days before the emergency
department admission using a third-degree polynomial
distributed-lag model,5 which imposes a certain amount of
restraint but allows for sufcient exibility in the estimation of a
biologically plausible lag structure to better control multicollinearity than unconstrained lag model. The SEs of the estimates
for each day were adjusted for overdispersion.
After dening the lag structure for TSP, we estimated the
2-day cumulative effect using the sum of effects at lag 0 and 1
obtained using the polynomial distributed lag model according
to the model proposed by Zanobetti et al.13 To assess potential
differences in the effects of TSP during the sugar cane burning
and non-burning periods, we built specic models for the
670

burning and non-burning periods. For these models, we

adopted 10 and 5 df for the smoothing time trend.
The effects of air pollutants are presented as percentage
increase or relative risk and 95% CIs for PEDVs due to a 10 g/
m3 increase in TSP. In addition, we have estimated the cumulative effect of the entire period for each IQR of TSP concentration increase. The statistical analyses were performed using SPSS
V.14 (SSPS, Inc.) and S-PLUS 4.5 (MathSoft, Inc.).

RESULTS
Table 1 presents the descriptive statistics of particles, weather
variables and PEDVs during the study period and stratied by
burning and non-burning periods. The mean TSP concentration
remained below the air quality standard according to WHO air
quality guidelines14 (80 mg/m3) during the sugar cane burning
period (541 days). This value was 2.5 times the TSP concentration found during the non-burning period (371 days)
(p<0.001). The PEDVs varied substantially between the two
periods, and they were 70% higher during the sugar cane
burning period. Also, the burning period was dryer and colder
than the non-burning period.
Figure 1 shows the daily records of the TSP concentrations
and pneumonia-related emergency room visits in Araraquara
from January 2005 to July 2007, stratied by burning and nonburning periods. During the burning periods, TSP daily concentrations surpassed the WHO air quality standard for the 24 h
average (120 mg/m3).14 In addition, it appears that the variations
in TSP did not account for all of the variability outcomes.
The TSP had a negative and statistically signicant Spearman
correlation ( p<0.01) in the burning and non-burning periods
with minimum temperatures of 0.23 and 0.25, respectively,
and humidities of 0.27 and 0.43, respectively. Conversely,
the TSP Spearman correlation with PEDVs was positive and
statistically signicant ( p<0.01) during both the burning and
the non-burning periods (burning=0.27; non-burning=0.21).

Table 1
study

Descriptive analyses of the main variables used in the

Periods and
descriptive
parameters

TSP
(g/m3)

Total (912 days)

Mean (SD)
55.6 (31.4)
Minimum
0.9
1st quartile
30.4
3rd quartile
76.5
Maximum
220.5
Non-burning (371 days)
Mean (SD)
30.8 (13.9)*
Minimum
0.9
1st quartile
19.9
3rd quartile
39.9
Maximum
79.5
Burning (541 days)
Mean (SD)
72.6 (28.7)
Minimum
9.3
1st quartile
53.6
3rd quartile
90.0
Maximum
220.5

Mean relative
humidity (%)

Minimum
temperature
(C)

Pneumonia
visits

61.9 (13.7)
22.0
53.7
70.8
91.5

17.2 (3.4)
5.0
15.0
20.0
27.0

1.65 (1.69)
0
0
2
13

67.9 (10.1)*
40.9
61.1
74.7
91.4

19.1 (1.8)*
10.0
18.0
20.0
27.0

1.2 (1.3)*
0
0
2
8

57.8 (14.3)
22.0
48.5
66.2
91.5

15.9 (3.6)
5.0
14.0
18.0
27.0

2.0 (1.8)
0
1
3
13

*Different from the non-burning period (p<0.001, MannWhitney test).

TSP, total suspended particles.

Arbex MA, et al. J Epidemiol Community Health 2014;68:669674. doi:10.1136/jech-2013-203709

Evidence-based public health policy and practice

Figure 1 The daily records of total suspended particle concentrations (mg/m3) and pneumonia-related emergency department visits for the entire
study period during both burning and non-burning periods.

During the burning period, humidity had a low and nonsignicant association with the outcome, whereas temperature
had a low and signicant correlation with the outcome (0.10;
p<0.5).
Figure 2 shows the lag structure and the cumulative effect of
increases in TSP on the pneumonia-related ED visits during the
entire study period.
There was an acute effect that began on the day of exposure
and continued for 2 days. The number of PEDVs increased

almost 6% in the 2 days following the TSP increase. For an IQR

of TSP concentration increase in the same period of time, an
increase of 29.4% (95% CI 14.1 to 50.9) on pneumonia-related
department room visits is expected. When the analysis was
stratied by the sugar cane harvesting periods (burning and nonburning), there were differences in the magnitude but not in the
pattern.
During the burning periods (gure 3), the lag structure and
the magnitude of the effect were quite similar to those observed

Figure 2 The percentage increases and 95% CIs for pneumoniarelated emergency department visits due to a 10 mg/m3 increase in the
daily total suspended particle concentrations during the entire study
period. The lag structure until 6 days after the exposure and the 2-day
cumulative effects are shown.

Figure 3 The percentage increases and 95% CIs for pneumoniarelated emergency room visits due to a 10 mg/m3 increase in the daily
total suspended particle concentrations during the burning periods. The
lag structure until 6 days after the exposure and the 2-day cumulative
effect are depicted.

Arbex MA, et al. J Epidemiol Community Health 2014;68:669674. doi:10.1136/jech-2013-203709

671

Evidence-based public health policy and practice

for the entire period, including the estimate of the 2-day cumulative effect.
During the non-burning periods, despite the lack of signicance in the estimates (which can be explained by the small
numbers of days and events examined), the lag structure of
adverse TSP effects is similar to that observed in burning
periods (gure 4).

DISCUSSION
The results of this study show that air pollution by TSP during
the sugar cane burning periods is associated with increases in
pneumonia-related emergency room department visits in a city
located in the middle of the largest Brazilian sugar caneproducing region.
The time-series approach has been adopted in investigations
of environmental health, and it has provided more robust
results following improvements in modelling and statistical
tools. We adopted Poisson regression models, including covariates, confounders and effect modiers, that we had used in previous studies focusing on the same scenario: the exposure to air
pollutants generated from sugar cane burning. These models
were adjusted for the usual parameters required for time-series
analyses.11 12 Despite being widely used in many studies, the
ecological time-series approach encompasses limitations that
must be taken into account before making assumptions of
causality.
In this time series, the inclusion of both temperature and
long-term trend adjustment seems to provide a fair control of
seasonality. Data on inuenza incidence are missing in the city
and in most of the urban centres in Brazil since it is not mandatory to seek medical attention in Brazil for people with inuenza
and thus not used in the model to improve modelling of
seasonality.
We examined an outcome that encompassed all PEDVs. It
would be useful to stratify the patients by age to evaluate any
differences in susceptibility among the different age groups.
Nevertheless, this is an interesting study to be developed in
large cities where pneumonia rate is high. Several studies have
demonstrated associations between pneumonia-related hospital
admissions and particulates.1519 Although emergency department visits are a sensitive indicator for respiratory conditions,20
this variable has rarely been examined to explore the relationship between pneumonia and particulate air pollution. In

Figure 4 The percentage increases and 95% CIs for pneumoniarelated emergency department visits due to a 10 mg/m3 increase in the
daily total suspended particle concentrations during the non-burning
periods. The lag structure until six 6 after the exposure and the 2-day
cumulative effect are depicted.

672

addition, in contrast to the large amount of information relating

fossil fuel-generated particulate matter to adverse effects on
human health, there are only a limited number of studies that
have directly evaluated effects of air pollution resulting from
outdoor biomass burning.2 3 However, biomass combustion is
becoming increasingly important as a source of ambient air pollution. There has been a worldwide increase in severe vegetation
res, which are associated with climate change and shifts in
population settlement patterns.21 A large number of people are
exposed to air pollution from both forest and agricultural res.
As we have demonstrated in previous studies conducted in the
same city, exposure to particles generated from sugar cane
burning is deleterious to the cardiorespiratory system.5 6 8 In
the present study, we analysed 30 months of data, which
allowed us to investigate two different air pollution exposure
patterns in a single community: periods of air pollution from
sugar cane burning (burning) and periods with a predominance
of fossil fuel-generated pollutants (non-burning). We know of
no previous studies that have examined the effects of outdoor
biomass burning on an equally long period.
It was observed that when comparing both periods, daily
means of PEDVs and TSP concentrations were higher during
the burning period than the non-burning period (2.5 and
1.7-fold, respectively); this effect was acute and continued for
2 days. However, the pneumonia data presented two new situations that had not arisen in our previous studies. First, we found
a non-signicant effect during the non-burning period, which
could be explained by the small number of PEDVs during this
period. Second, the effect was greater than that observed during
the burning period, which may have been due to differences in
both the composition and size of the particles during these two
periods.
Monitoring of particles could be a matter for discussion and
considered a limitation in the study, and two topics must be discussed: TSP measurements instead of inhalable or ne particles
and a single monitoring site.
Previous studies5 8 have demonstrated that PM10 or even
PM2,5 measurements are rarely and intermittently performed by
the So Paulo State Environmental Agency (CETESB) in the
region where this study was conducted. Time-series analyses
require regular data on exposure and outcomes over a long
period to obtain accurate results. We performed our own measurements of the TSP concentrations during the study period to
assure that data were available for both the burning and the
non-burning periods. Moreover, previous studies have demonstrated that the majority of particles emitted from biomass
burning are ultrane (smaller than 0.1 m), with only a small
fraction in the large size range.22 Approximately 90% of the
total particulate emissions in agricultural burning emissions
consist of particulate matter with a diameter smaller than
10 mm (PM10).22 Ward et al23 found that the total particulate
matter emitted from forest res with aming combustion
contains 8095% ne particles (PM2.5), and the total particulate
matter emitted from smouldering combustion contains up to
90% ne particles. A previous study has shown that sugar cane
burning is the largest source of ne and ultrane particles in
the sugar cane region during burning periods.24 Thus,
during burning periods, daily variations in TSP can be assumed
to be a proxy for the daily variations in ne and ultrane
particles.5 8 Despite the impossibility to measure ne and inhalable particles directly in the study, WHO has suggested that if
ne or inhalable particles cannot be measured, TSP should be
measured.22

Arbex MA, et al. J Epidemiol Community Health 2014;68:669674. doi:10.1136/jech-2013-203709

Evidence-based public health policy and practice

At the end of the 1990s, we carried out a study using two
samplers in order to measure TSP located downtown and in the
rural area of Araraquara6 The Pearson correlation coefcient
was 0.92 between the measurements from the two sites. With
this it is possible and plausible to assume that only one monitoring station can provide a true estimation of the inhabitants
exposure to air pollutants.
In the study, the monitor was located in the downtown of
Araraquara where the trafc volume is the same along the year.
Trucks used to transport the sugar cane from the elds to the sugar
cane mills do not pass through the roads in the city since the sugar
cane mills are located along the highway. Furthermore, the distance between the monitoring site and the highway is 5 km.
Studies have demonstrated that the effect of the coarse fraction (PM10) on the respiratory system may be equally or more
severe than that attributed to ne and ultrane particles,18 25 26
Brunekreef and Forsberg have suggested that in environments
with predominantly ne particles, the effects will be more
evident than those attributed to the coarse fraction, at least in
part because of the difference in particle number and larger
active surface areas.27
The airway epithelium is a physical barrier between the mucociliary and immune systems, which acts as the rst defence
against respiratory pathogens. The airway epithelium initiates
multiple innate and adaptive immune responses for the efcient
targeting of bacterial and viral pathogens. These defence
mechanisms usually result in rapid pathogen clearance.28 In addition, alveolar macrophages (AMs) possess a potent phagocytic
property that plays a key role in the defence against respiratory
infections.29
Exposure to environmental agents can affect the respiratory
epithelium and can lead to an inammatory process30 that compromises aerodynamic ltration, mucociliary clearance, particle
transport, detoxication by alveolar macrophages and immunological defences against viruses and bacteria. Particles damage
the airway epithelium31 and affect the alveolar macrophage
integrity.26 Consequently, particulate matter may increase the
susceptibility to infection and/or may lead to more severe infections. Experimental studies have shown that animals exposed to
particulates by inhalation had decreased intrapulmonary killing
of bacteria, which further demonstrates that particulate inhalation affects the bactericidal activity and/or clearance mechanisms of the respiratory tract.31
Our study examined two different environmental scenarios:
the presence and absence of sugar cane burning. Experimental
studies have evaluated the adverse effects of particulates on the
respiratory system by comparing emissions during outdoor
biomass burning and non-biomass burning periods. Wegesser
et al have shown that during a 2008 forest re episode in
California, the wildre PM, especially the coarse PM fraction,
contained chemical components that are toxic to the lungs, particularly to alveolar macrophages. In addition, they demonstrated in mouse bioassays that the coarse PM fraction was more
toxic to the lungs than an equal dose of PM collected from
ambient air without vegetation res from the same region
during a comparable season.32 Mazzoli-Rocha et al have demonstrated that a single low dose of ambient particles produced by
automobiles and sugar cane burning induced signicant alterations in the pulmonary mechanics and lung histology of mice
and that the biomass particles were at least as toxic as those produced by automobiles.33

Assuming that during the burning periods there are primarily

ne and ultrane particles in the sugar cane regions, our results
are similar to those presented by Peel et al,20 in which PEDVs
in the Atlanta metropolitan area were analysed. Both studies
demonstrated similar magnitudes of acute effects from PM on
the days that the exposure occurred.
In terms of hospitalisation for pneumonia, an outcome that
represents patients with more severe diseases, Host et al18
found a 2.5% increased risk of lower respiratory tract infections
after short-term PM2.5 exposure in France; Dominici et al17
found a 1.4% increase in hospitalisation for lower respiratory
tract infections in elderly patients; Zanobetti and Schwartz34
found a 6.5% increase in hospitalisations for pneumonia among
the elderly; and Belleudi et al35 found a 2.8% increase in hospitalisations for lower respiratory tract infections at lag 2 in
Rome. A recent review showed that each 10 g/m3 increase in
long-term ambient PM2.5 concentrations is associated with an
approximately 12% increased risk of acute lower respiratory
infections36
Several studies have demonstrated positive associations
between acute respiratory infection and indoor biomass
burning.2 3 Episodic forest burning is also associated with
adverse respiratory health effects3739 despite limited number of
studies that have evaluated pneumonia as an outcome.
Uncontrolled res and prescribed burning occur on a regular
basis in many parts of the world. Despite forest res being a globally important source of PM, the public health effects are difcult to evaluate because forest res are sporadic, short-lived
event that rarely occur in areas with a high population density.
The comparison between our study and the previous investigations of biomass burning is not trivial: in both indoor wood
smoke and forest burning, the particulate matter reaches
extremely high levels, as opposed to sugar cane burning, which
causes a less intense but more prolonged increase.
Biofuels such as ethanol are renewable sources of alternative
energy. In Brazil, ethanol has been used since the 1970s, and its
usage in automobiles has surpassed that of fossil fuels. The main
concern regarding alcohol production lies in the rudimentary
technique of harvesting sugar cane by burning the crop. The
burning of sugar cane plantations before manual harvesting is
performed to protect the sugar cane workers to the risks of venomous animals such as snakes and scorpions and to facilitate the
sugar cane harvesting by removing dry sugar cane straws and
dead growth. The mechanisation of sugar cane harvesting
(burning is not needed for mechanised harvesting) can contribute to a remarkable decrease in air pollution.4 40
In So Paulo State, almost 23 million people live in cities
close or are surrounded by sugarcane plantations where the
crops are usually burnt before harvesting, thus exposing them to
the pollutants emitted for 6 months every year. Similar exposure
also takes place in other sugar cane-producing regions of Brazil,
Central and North America and Southeast Asia.41 42 Hence, the
magnitude of this situation is not only a Brazilian concern, but
of a global one as well.
In summary, increases in TSP concentrations were found to
be associated with an increased PEDV visits in a populated
region, which remains under the inuence of the particles generated from sugar cane burning for a long period of time every
year. Despite the importance of ethanol production to Brazils
economy, this nding reinforces the need for polices and efforts
to ban sugar cane burning prior to harvesting.

Arbex MA, et al. J Epidemiol Community Health 2014;68:669674. doi:10.1136/jech-2013-203709

673

Evidence-based public health policy and practice

14

What is already known on this subject

15

Air pollution generated from the burning of both biomass

and fossil fuels is deleterious to the cardiorespiratory system.
Pneumonia is one of the largest causes of mortality and
morbidity.
The adverse effects of air pollutants on pneumonia-related
emergency department visits have been underestimated.

16

17

18

19

What this study adds

20

This is the longest study analysing the effects of exposure to

particles generated from outdoor biomass burning, and it
compares periods with and without sugar cane burning.
The number of emergency department visits is a sensitive
and important indicator of the effect of air pollutants on
pneumonia that is highly prevalent worldwide and is
associated with high lethality. For pneumonia, the effect is
acute and occurs predominantly on the day of exposure to
the pollutants.
This study suggests that during periods with higher
concentrations of combustion particles the effect of daily
TSP exposure is higher.

21

22
23

24
25

26

27

Contributors All authors designed the study, acquired the data, wrote the
manuscript and approved the nal version. MAA and ALFB also analysed the data.

28

Competing interests None.

29

Provenance and peer review Not commissioned; externally peer reviewed.

30

REFERENCES
1
2
3
4

5
6

10
11
12

13

674

Brunekreef B, Holgate ST. Air pollution and health. Lancet 2002;360:123342.

Naeher LP, Brauer M, Lipsett M, et al. Woodsmoke health effects: A review. Inhal
Toxicol 2007;106:67106.
Torres-Duque C, Maldonado D, Perez-Padilla R, et al. Biomass fuels and respiratory
diseases. A review of the evidence. Proc Am Thor Soc 2008;5:57790.
Uriarte MC, Yackulic CB, Cooper T, et al. Expansion of sugar cane production in
So Paulo, Brazil: implication of re occurrence and respiratory health. Agriculture
Ecosystems Environ 2009;132:4856.
Arbex MA, Saldiva PH, Pereira LAA, et al. Impact of outdoor biomass burning on
hypertension hospital admissions. J Epidemiol Community Health 2010;64:5739.
Arbex MA, Bohm GM, Saldiva PHN, et al. Assessment of the effects of sugar cane
burning on plantation burning on daily counts of inhalation therapy. J. Air & Waste
Manage Assoc 2000;50:17459.
Canado JED, Saldiva PHN, Pereira LAA, et al. The impact of sugar cane burning
emissions on the respiratory system of children and elderly. Environ Health Perspect
2006;114:7259.
Arbex MA, Martins LC, Oliveira RC, et al. Outdoor biomass burning and asthma
hospital admissions in a sugar cane plantation area in Brazil. J Epidemiol
Community Health 2007;61:395400.
Galdos M, Calvalett O, Seabra JEA. et al. Trends in global warming and human
health impacts related to Brazilian sugar cane ethanol production considering black
carbon emissions. Energy Applied 2013;104:57682.
McCullagh P, Nelder JA. Generalized Linear Models. 2nd edn. London: Chapman
and Hall, 1989.
Green PJ, Silverman BW. Non parametric regression and generalized linear models.
A Roughness Penalty Approach. London: Chapman and Hall, 1994.
Akaike H. Information theory and an extension of the maximum likelihood principal.
In: Petrov BN, Csaki F, eds. 2nd International Symposium on Information Theory.
Budapest: Akademiai Kiado, 1973:26781.
Zanobetti A, Wand MP, Schwartz J, et al. Generalized additive distributed lags
models: quantifying mortality displacement. Biostatistics 2000;1:27992.

31
32
33
34
35

36

37
38

39

40

41

42

World Health Organization. Guidelines for Air Quality. Geneva: WHO, 1999.
http://www.who.int/environmentalinformation/air/guidelines/aqguides3.pdf (acessed
1 Oct 2013).
Laumbach RJ, Kipen HM. Respiratory health effects of air pollution: Update on
biomass smoke and trafc pollution. J Allergy Clin Immunol 2012;129:311.
Lin M, Stieb DM, Chen Y. Coarse particulate matter and hospitalizations for
respiratory infections in children younger than 15 years in Toronto: A case-crossover
analysis. Pediatrics 2005;116:e23540.
Dominici F, Peng RD, Bell ML, et al. Fine particulate air pollution and hospital
admissions for cardiovascular and respiratory diseases. JAMA 2006;294:
112734.
Host S, Larrieu S, Pascal L, et al. Short term associations between ne and coarse
particles and hospital admissions for cardiorespiratory diseases in six French cities.
Occup Environ Med 2008;65:54451.
Neupane B, Jerrett M, Burnett RT, et al. Long-term exposure to ambient air
pollution and risk of hospitalization with community-acquired pneumonia in older
adults. Am J Respir Crit Care Med 2010;181:4753.
Peel JL, Tolbert PE, Klein M, et al. Ambient air pollution and respiratory emergency
department visits. Epidemiology 2005;16:16474.
Johnston FH, Bailie RS, Pilotto LS, et al. Ambient biomass smoke and
cardio-respiratory hospital admissions in Darwin, Australia. BMC Public Health
2007;7:240.
Schwela D, Soldemmer J, Morawska L, et al. Health guidelines for vegetation re
events. Geneva, Switzerland: World Health Organization, 1999.
Ward D. Review of smoke components in United States Department of Agriculture
(USDA): Health hazards of smoke. Recommendations of the Concensus Conference.
Missoula April 1997.
Lara LL, Artaxo P, Martinelli LA, et al. Properties of aerosols from sugar-cane
burning emissions on Southeastern Brazil. Atmos Environ 2005;39:462737.
Becker S, Fenton MJ, Soukup JM. Involvement of microbial components and toll-like
receptors 2 and 4 in cytokine responses to air pollution particles. Am J Respir Cell
Mol Biol 2002;27:61118.
Becker S, Soukup JM, Sioutas C, et al. Response of human alveolar macrophages to
ultrane, ne, and coarse urban air pollution particles. Exp Lung Res
2003;29:2944.
Brunekreef B, Forsberg B. Epidemiological evidence of effects of coarse airborne
particles on health. Eur Respir J 2005;26:30918.
Vareille M, Kieninger E, Edwards MR, et al. The airway epithelium: soldier in the
ght against respiratory viruses. Clin Microbiol Rev 2011;24:21029.
Knapp S, Leemans JC, Florquin S, et al. Alveolar macrophages have a protective
anti-inammatory during murine pneumococcal pneumonia. Am J Respir Crit Care
Med 2003;167:1719.
Berstein JA, Alexis N, Barnes C, et al. Heath effects of air pollution. J Allergy Clin
Immunol 2004;114:111622.
Gilmour MI, Daniels M, McCrillis RC, et al. Air pollutant-enhanced respiratory
disease in experimental animals. Environ Health Perpect 2001;109:61922.
Wegesser TC, Pinkerton KE, Last JA. California wildres of 2008: Coarse and ne
particulate matter toxicity. Environ Health Perspect 2009;117:8937.
Mazzoli-Rocha F, Magalhes CB, Malm O, et al. Comparative respiratory toxicity of
particles produced by trafc and sugar cane burning. Environ Res 2008;108:3541.
Zanobetti A, Schwartz J. Air pollution and emergency admissions in Boston, MA.
J Epidemiol Community Health 2006;60:8905.
Belleudi V, Faustini A, Stafoggia M,, et al Impact of ne and ultrane particles on
emergency hospital admissions for cardiac and respiratory diseases. Epidemiology
2010;21:41423.
Mehta S, Shin H, Burnett R, et al. Ambient particulate air pollution and acute lower
respiratory infections: a systematic review and implications for estimating the global
burden of disease. Air Qual Atmos Health 2013;6:6983.
Knzli N, Avol E, Wu J, et al. Health effects of the 2003 Southern California wildre
on children. Am J Respir Crit Care Med 2006;174:12218.
Morgan G, Sheppeard V, Khalaj B, et al. Effects of bushre smoke on daily
mortality and hospital admissions in Sydney, Australia. Epidemiology
2010;21:4754.
Henderson SB, Brauer M, MacNab Y, et al. Three measures of forest re smoke
exposure ante their associations with respiratory and cardiovascular health outcomes
in a population-based cohort. Environ Health Pespect 2011;119:126671.
Gotto DM, Lana M, Obuti CA, et al. Effects of biomass burning on nasal
mucociliary clearance and mucus properties after sugar cane harvesting. Environ Res
2011;111:6649.
Frana DA, Longo KM, Neto TGM, et al. Pre-harvest sugar cane burning:
Determination of emission factors through laboratory measurements. Atmosphere
2012;3:16480.
Souza KF, Carvalho LRF, Allen AG, et al. Diurnal and nocturnal measurements of
PAH nitro-PAH, and oxy-PAH compounds in atmospheric particulate matter of a
sugar cane burning region. Atmos Environ 2014;83:193201.

Arbex MA, et al. J Epidemiol Community Health 2014;68:669674. doi:10.1136/jech-2013-203709