Asthma and the environment
Urbanisation is associated with prevalence of
childhood asthma in diverse, small rural communities
in Ecuador
Alejandro Rodriguez,1,2 Maritza Vaca,1,2 Gisela Oviedo,1 Silvia Erazo,1 Martha E Chico,1
Carlos Teles,2 Mauricio L Barreto,2 Laura C Rodrigues,3 Philip J Cooper1,4,5,6
See Editorial, p 1025
< Additional materials are
published online only. To view
these files please visit the
journal online (http://thorax.bmj.
com).
1
Laboratorio de Investigaciones
FEPIS, Quinindé, Esmeraldas,
Ecuador
2
Instituto de Saúde Coletiva,
Universidade Federal da Bahia,
Salvador, Bahia, Brazil
3
Department of Epidemiology,
London School of Hygiene and
Tropical Medicine, London, UK
4
Molecular and Biochemical
Parasitology, Liverpool School of
Tropical Medicine, Liverpool, UK
5
Colegio de Ciencias de la
Salud, Universidad San
Fransisco de Quito
6
Centre for Infection, St
George’s University of London,
London, UK
Correspondence to
Professor Alejandro Rodrı́guez,
Centro de Investigaciones FEPIS
(Fundación Ecuatoriana para la
Investigación en Salud),
Quinindé, Esmeraldas, Ecuador;
rodriguez_alejo1@hotmail.com
Received 19 March 2011
Accepted 13 June 2011
Published Online First
8 August 2011
ABSTRACT
Background Studies conducted in transitional
communities from Africa and Asia have pointed to the
process of urbanisation as being responsible for the
increase in asthma prevalence in developing regions. In
Latin America, there are few published data available on
the potential impact of urbanisation on asthma
prevalence. The aim of the present study was to explore
how the process of urbanisation may explain differences
in asthma prevalence in transitional communities in
north-eastern Ecuador.
Methodology/principal findings An ecological study
was conducted in 59 communities in Esmeraldas
Province, Ecuador. Indicators of urbanisation were
grouped into three indices representing the processes
associated with urbanisation: socioeconomic, lifestyle
and urban infrastructure. Categorical principal
components analysis was used to generate scores for
each index and a fourth indexda summary urbanisation
indexdwas derived from the most representative
variables in each of the three indices. The authors
analysed the associations between community asthma
prevalence and the indices, as well as with each
indicator variable of every group. The overall prevalence
of asthma was 10.1% (range 0e31.4% between
communities). Three of the four indices presented
significant associations with community asthma
prevalence: socioeconomic (r¼0.295, p¼0.023), lifestyle
(r¼0.342, p¼0.008) and summary urbanisation index
(r¼0.355, p¼0.006). Variables reflecting better
socioeconomic status and a more urban lifestyle were
associated with greater asthma prevalence.
Conclusions These data provide evidence that the
prevalence of asthma increases with increasing levels of
urbanisation in transitional communities, and factors
associated with greater socioeconomic level and
changes towards a more urban lifestyle may be
particularly important.
INTRODUCTION
This paper is freely available
online under the BMJ Journals
unlocked scheme, see http://
thorax.bmj.com/site/about/
unlocked.xhtml
Differences have been observed in asthma prevalence between and within countries with different
levels of development, with greater asthma prevalence in developed countries and in urban areas of
developing countries.1 2 However, these differences
may have diminished over the past decade because
asthma prevalence may have reached a plateau in
some developed countries while continuing to rise
in urban and perhaps rural areas of developing
Thorax 2011;66:1043e1050. doi:10.1136/thoraxjnl-2011-200225
Key messages
What is the key question?
< This paper addresses the question of why the
prevalence of asthma is increasing in populations living in developing countries, particularly
in Latin America.
What is the bottom line?
< Although the causes of this increase are poorly
understood and are likely to be explained by
a complex set of causal factors, we believe that
the process of urbanisation may explain, at least
partly, these temporal trends. No studies to date
have investigated how the various processes
associated with urbanisation may affect asthma
prevalence.
Why read on?
< Our findings show how these processes extend
into rural communities undergoing the transition
from a traditional to a modern mode of life in
a developing country and show how urbanisation and specific factors involved in this process
are associated with asthma prevalence even in
the rural tropics.
< Our findings are, therefore, likely to be of general
relevance to those interested in the potential
role of urbanisation in explaining asthma
prevalence in the developing world.
countries.3 The causes of these variations are poorly
understood and may be explained by a complex set
of causal factors relating to changes in environment
and lifestyle.4 5
Evidence from epidemiological studies suggests
that the increase in asthma prevalence in developing countries may be explained by changes from
traditional/rural to modern/urban societies.6 7
Rural areas tend to have a low asthma prevalence
and this has been explained by the protection
against allergic diseases provided by factors associated with a traditional rural lifestyle.8 However,
numerous exposures related to urbanisation have
been identified as potential risk factors for asthma
(eg, reduction in the frequency of infections;
reduction in family size; increasing vaccine
coverage and use of antibiotics; increases in environmental pollution and household exposure to
1043
Thorax: first published as 10.1136/thoraxjnl-2011-200225 on 8 August 2011. Downloaded from http://thorax.bmj.com/ on June 19, 2020 by guest. Protected by copyright.
ORIGINAL ARTICLE
Asthma and the environment
METHODS
Study area and population
The study was conducted in small rural communities in the
districts of Eloy Alfaro and San Lorenzo in a humid tropical
region of Esmeraldas Province in north-eastern Ecuador. Located
in one of the poorest and remotest areas of the country, these
communities have been undergoing changes related to urbanisation over the last 20 years. The principal economic activity of
these communities is subsistence agriculture, hunting, fishing
and logging. Employment through commerce and the provision
of services is available in the larger communities. The educational
level of the population is generally very low and rates of illiteracy
among adults are high. Housing materials are generally wood and
bamboo for walls and corrugated iron for roofing, although the
use of cement blocks for walls is increasingly common. Drinking
water comes directly from rivers, wells or rainwater, but some
communities have piped but untreated water from streams.
There is no municipal sewage system in any of the communities
and disposal of faeces is by household or community pit latrines
or in the open. A number of communities are connected to the
national electrical grid but few have access to telephone services.
Transportation for most communities is by river, although small
roads are being built to interconnect some of these communities
with the provincial capital of Esmeraldas. Temporary migration
to the city of Esmeraldas and other urban centres in Ecuador is
very frequent among people of working age.14
Study design
An ecological analysis was conducted using data collected from
59 communities. The original study was designed to investigate
individual risk factors for allergy and asthma in Afro-Ecuadorian
children and has been described in detail.15 Communities with
a predominantly Afro-Ecuadorian population were included and
all children aged 7e15 years registered in a community census
were eligible. The mean number of children evaluated in the
communities was 73 children (range 15e331).
Data collection
Data were collected between May 2005 and March 2008.
Data on asthma symptoms and risk factors including relevant
lifestyle and socioeconomic factors were collected using a parentally administered questionnaire adapted from the International
Study of Asthma and Allergies in Childhood phase II.16 Data
on urban infrastructure characteristics were collected using
a checklist.
1044
Prevalence of asthma
Asthma was defined as a positive response to the question ‘Has
your child had wheeze in the chest in the last 12 months’.
Prevalence of asthma was defined for each community as the
number of children with wheeze in the past 12 months by the
number of children aged 7e15 evaluated in each community.
Measures of urbanisation
We selected variables into three different groups of indicators
(infrastructure, socioeconomic and lifestyle factors) to represent
some of the main characteristics of the process of urbanisation in
transitional communities (online supplement, table 1). The
indicators used for each of these three groups are shown in table 1.
Infrastructure indicators represent the ‘urban’ infrastructure of
each community based on the presence of general basic services
(online supplement, table 2). These variables were collected for
each community. Socioeconomic indicators represent the socioeconomic status of the households in the communities based on
indicators such as the presence of material possessions, access to
general services, parental education, household income. The third
group represents the acquisition of new lifestyle behaviours by
the study population based on personal habits and general
household characteristics. Data for the last two groups were
presented as averages or proportions per community using data
collected at the individual or household levels.
Statistical analysis
We calculated indices for each of the three groups and a summary
urbanisation index using Categorical Principal Component Analysis (CATPCA), a multivariate technique that summarises a group
of correlated variables (numeric, nominal or ordinal) in a single
component group or independent index.17 18 In CATPCA, the first
component explains the highest proportion of total variance in the
data set. The second component accounts for the majority of
variance not explained by the first component, and so on. The
variables are related to each component through their component
loadings that are measures of correlation ( 1 to 1) between each
variable and the components. A larger component loading for
a variable represents a greater contribution of that variable in the
component. The components produce a score for each community
that summarises the contribution of all variables in each component. This score works like an index providing a summary
measure, a range of z scores for each of the communities. Two
components were retained for each analysis and the scores
provided by these were used as indices. For the construction of the
summary urbanisation index, those variables with the highest
component loadings for each of the three group analyses were
included, also considering their theoretical relevance. Due to
sample size and to avoid ‘overloading’, the summary urbanisation
model kept the relationship of one variable to four observations.19
The urbanisation indices were interpreted such that higher values
of component scores for each observation indicated a higher
socioeconomic level, a more urban lifestyle and greater urban
infrastructure. Spearman’s correlations were calculated to explore
the relationships between the four indices and asthma prevalence.
We also analysed the associations between each constituent
variable for each of the infrastructure, socioeconomic and lifestyle
indicator groups and community asthma prevalence using bivariate and multivariate linear regression, weighted for the inverse of
the variance of the mean of community sizes. For multivariate
analyses, we constructed a model for each of the indicator groups
introducing all variables of each group separately through backwards stepwise regression. The final models selected were those
that explained the most variation in asthma prevalence between
Thorax 2011;66:1043e1050. doi:10.1136/thoraxjnl-2011-200225
Thorax: first published as 10.1136/thoraxjnl-2011-200225 on 8 August 2011. Downloaded from http://thorax.bmj.com/ on June 19, 2020 by guest. Protected by copyright.
allergens; changes in diet, lifestyle and socioeconomic factors).5 8
There is evidence that the prevalence of allergic diseases is
increasing with increasing levels of urbanisation.9 10 Nevertheless, such studies have not explored in detail which elements of
the process of urbanisation are responsible, and there are no
published studies evaluating the effects of urbanisation on
asthma prevalence in Latin America.11 12
Urbanisation is defined as a gradual process of transformation
where rural areas and their populations lose their rural characteristics and gradually become urban.13 This process occurs
through improvements in urban infrastructure, public services,
changes in social structures and lifestyle and population growth,
among other factors. The present study developed indices of
urbanisation to measure some of the factors that mould this
process in transitional communities, and used these indices to
explore how elements of the process of urbanisation may
explain differences in asthma prevalence between small rural
communities in a tropical region of Ecuador.
Asthma and the environment
Infrastructure, socioeconomic and lifestyle indicators: definitions and descriptive characteristics
Index
Indicators
Definition
Categories
N (%)
Infrastructure
(General characteristics
of the communities)
Administrative grade
Measures the level of development based on the political/
administrative division of the communities
Spatial organisation
Used as a proxy variable for population density. This indicator
identifies the concentration of houses
Transport access
Identifies the type of access used to arrive at communities
Electrical grid
Identifies the presence of a connection to the electrical grid
Piped water system
Identifies the presence of a piped water system (untreated
water only)
Telephone system
Identifies the presence of access to the national telephone
network
Health centre
Identifies the presence of a health centre
Pharmacy
Identifies the presence a pharmacy
Educational institution
Identifies the presence of secondary schools
Shops
Estimates the commercial infrastructure through number
of shops
Towns
Parish
Dispersed
Blocks
Neighbourhoods
River
Road
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
0e1
2e5
6e15
>15
42
17
33
20
6
40
19
41
18
12
47
17
42
17
42
19
40
13
46
16
22
14
7
Index
Indicators
Definition
Minimum (%) Mean (%) Maximum (%)
Estimates the percentage of households in which the father has any
0
secondary education
Mother’s education
Estimates the percentage of households in which the mother has any
0
secondary education
Household income
Estimates the percentage of households with an income equal or
0
above US$150/month
Access to electricity
Estimates the percentage of households with electricity
0
(electrical grid or generator)
Material goods
Estimates the percentage of households with (all of) refrigerator,
0
TV and stereo system
Cement house
Estimates the percentage of households with cement walls
0
Gas for cooking
Estimates the percentage of households that use propane gas for cooking
0
Motor vehicles
Estimates the percentage of households with motor vehicles (boat or car)
0
Uncrowded household Estimates the percentage of households that are not crowded (#3 persons
15.8
by bedroom)
Consumption
Estimates the percentage of the study children who consume hamburgers
0
Lifestyle (% study
at least once a month
children or household of hamburgers
by community)
Consumption of
Estimates the percentage of the study children who consume fizzy drinks daily
0
fizzy drinks
No physical exercise Estimates the percentage of the study population that do no physical activity
0
daily (proxy of sedentarism)
Television in house
Estimates the percentage of houses that have a television
0
TV viewing
Estimates the percentage of the study children who watch television >1 h daily
0
No farming activities Estimates the percentage of households that do not work in agricultural activities 0
Cat in house
Estimates the percentage of households that have a cat living inside the house
0
Dog in house
Estimates the percentage of households that have a dog living inside the house
4.2
Migration
Estimates the percentage of the study children who has lived for at least
3.9
3 months outside the study area in the past
Parasite infection rate Percentage of population with intestinal helminth Ascaris lumbricoides,
14
Trichuris trichiura, Strongyloides stercoralis
Socioeconomics
(% households by
community)
Father’s education
communities and were those with the smallest mean square error
and the highest value of adjusted R2.20 For bivariate analyses,
variables with p<0.10 were considered relevant and statistically
significant associations in the multivariate models were inferred
by p<0.05. Calculations were done with and without outliers
and extreme observations of the asthma prevalence; given that
results were very similar, we present results with all observations.
All the analyses were carried out with SPSS V.15.
Thorax 2011;66:1043e1050. doi:10.1136/thoraxjnl-2011-200225
(71.2)
(28.8)
(55.9)
(33.9)
(10.2)
(67.8)
(32.2)
(69.5)
(30.5)
(20.3)
(79.7)
(28.8)
(71.2)
(28.8)
(71.2)
(67.8)
(32.2)
(22.0)
(78.0)
(27.1)
(37.3)
(23.7)
(11.9)
17.8
40.7
19.4
47.8
16.5
53.5
68.3
100
14.4
42.4
12.3
63.3
10.1
58.8
48.2
100
48.2
89.7
21.1
58.3
17.4
88.0
25.6
65.6
55.5
65.9
11.8
34
52.6
29.8
96.6
100
68
81.5
96
76
76.3
100
RESULTS
Based on local updated censuses, we sampled approximately
95% of the eligible population or 4183 children. The overall
prevalence of asthma was 10.1% (range 0e31.4%). The indicator
variables considered for infrastructure, socioeconomic and lifestyle indices are provided in table 1 together with frequencies
and mean values where appropriate, showing significant
heterogeneity between communities.
1045
Thorax: first published as 10.1136/thoraxjnl-2011-200225 on 8 August 2011. Downloaded from http://thorax.bmj.com/ on June 19, 2020 by guest. Protected by copyright.
Table 1
Asthma and the environment
Components loading by group
Components loading for
summary urbanisation index
Component 1
Groups
Indicators
Component 1
Component 2
Infrastructure
Administrative grade
Spatial organisation
Transport access
Electrical grid
Piped water system
Telephone system
Health centre
Pharmacy
Educational institution
Shops
% of variance
Father’s education
Mother’s education
Household income
Access to electricity
Material goods
Cement house
Gas for cooking
Motor vehicles
Uncrowded household
% of variance
Consumption of hamburgers
Consumption of fizzy drinks
No physical exercise
Television in house
TV viewing
No farming activities
Cat in house
Dog in house
Migration
Parasite infection rate
% of variance
0.611
0.753
0.741
0.449
0.746
0.762
0.875
0.681
0.785
0.877
54.4
0.658
0.768
0.821
0.787
0.854
0.705
0.834
0.456
0.255
50.1
0.837
0.749
0.672
0.849
0.889
0.663
0.646
0.595
0.793
0.036
50.7
0.337
0.420
0.188
0.764
0.016
0.252
0.021
0.138
0.337
0.015
11.1
0.436
0.268
0.184
0.187
0.005
0.143
0.304
0.692
0.662
15.1
0.269
0.080
0.522
0.259
0.061
0.542
0.382
0.398
0.193
0.732
15.9
Socioeconomic
Lifestyle
CATPCA results are provided in table 2. Most indicator variables showed high loadings (>0.6) for the first components with
Cronbach’s a >0.85. The total variance explained by the first
component for infrastructure, socioeconomic and lifestyle
indices was 54.4%, 50.1% and 50.7%, respectively. Five indicators
were selected from each group to be included in the summary
urbanisation index (table 2): those with the highest component
loadings for each group with a single exception ‘No agricultural
activity’, which was included because of its theoretical importance in distinguishing rural from urban populations.13 The first
component of the summary analysis accounted for 54.0% of
total variation, and 13/15 variables had factor loadings >0.6.
The associations between the four indices and asthma prevalence are illustrated in figure 1. Significant correlations (p<0.05)
with asthma prevalence were observed for three of the four
indices: socioeconomic (r¼0.295), lifestyle (r¼0.342) and
summary urbanisation index (r¼0.355). There was a weak
association between the urban infrastructure index and asthma
prevalence (table 3). The summary urbanisation index showed
high correlations with the other three indices (>0.840) that were
highly correlated (table 3).
The results of bivariate analyses between indicator variables
for each index and asthma prevalence are shown in table 4.
Bivariate analyses for infrastructure indicators showed that
electrical grid (b¼2.226), healthcare centre (b¼2.076) and
community pharmacy (b¼2.296) were associated with asthma
prevalence (p<0.10). In the multivariate analysis for this group,
1046
Component 2
0.530
0.664
0.696
0.678
0.348
0.507
0.511
0.746
0.684
0.461
0.736
0.790
0.736
0.816
0.220
0.095
0.486
0.172
0.808
0.363
0.763
0.654
0.262
0.312
0.873
0.820
0.757
0.344
0.419
0.038
54.0
16.0
3/10 variables were kept in the final model (p<0.05), explaining
17.7% of variability in asthma prevalence across communities
(adjusted R2¼0.177; p¼0.008): presence of access roads
(b¼ 6.069), presence of electrical grid (b¼5.117) and number of
shops (>15/shops: b¼5.065). Thus, asthma prevalence decreased
6.1% in communities with roads compared with river access;
asthma prevalence increased 5.1% for communities with an
electric grid connection compared with those without; and
asthma prevalence increased 5.1% for those communities with
more than 15 shops compared with those with 0e1 shops.
Bivariate analyses for socioeconomic indicators showed that
asthma prevalence was associated with household income
(b¼0.072), having access to electricity (b¼0.029), number of
household material goods (b¼0.104), use of gas for cooking
(b¼0.046) and having a motor vehicle (b¼0.164). In the multivariate analysis, 5/9 indicators remained in the final model,
explaining 17.4% of variability in asthma prevalence across
communities (adjusted R2¼0.174; p¼0.009), of which three
were statistically associated: having access to electricity
(b¼0.045), having a cement house (b¼-0.127) and having
a motor vehicle (b¼0.168). Thus, in the multivariate model,
asthma prevalence increased 4.5% if all households had access to
electricity, increased 16.8% if all households had motor vehicles,
but decreased 12.7% if all households were cement-built. Bivariate analyses for lifestyle indicators showed associations with:
consumption of hamburgers (b¼0.104) and fizzy drinks
(b¼0.121), watching TV daily (b¼0.048), no agricultural
Thorax 2011;66:1043e1050. doi:10.1136/thoraxjnl-2011-200225
Thorax: first published as 10.1136/thoraxjnl-2011-200225 on 8 August 2011. Downloaded from http://thorax.bmj.com/ on June 19, 2020 by guest. Protected by copyright.
Table 2 Component loadings by infrastructure, socioeconomic and lifestyle indices
Asthma and the environment
activity (b¼0.065) and having a household cat (b¼0.091). In the
multivariate analysis, 5/10 lifestyle indicators remained in the
final model, explaining 19.8% of variability in asthma prevalence
(adjusted R2¼0.198; p¼0.005), of which consumption of fizzy
drinks was statistically significant (b¼0.132).
DISCUSSION
In the present study, we conducted an ecological analysis to
understand better if urbanisation (and specific factors associated
with this process) may explain the variation in asthma prevalence between rural transitional communities. Clearly, urbanisation is a highly complex process and is strongly associated
with other aspects of modernisation, especially in developing
countries. We chose to measure urbanisation by exploring variables representative of infrastructure, socioeconomic and lifestyle indicators that together constitute much of what is
understood by urbanisation. A population does not have to live
in an urban environment to experience many of the processes
associated with urbanisation and the effects of individual factors
that constitute urbanisation on asthma prevalence may be more
clearly measured in transitional societies where there is likely to
be sufficient heterogeneity in exposure to such factors.
Cross-sectional studies conducted in developing countries
have shown a higher prevalence of asthma reported in urban
Table 3
compared with rural populations, although the magnitude of
the difference is variable: South Africa (urban¼33.0%, periurban¼34.4%; rural¼17.0%), Zimbabwe (urban rich¼5.8%;
urban poor¼3.1%; rural¼0.1%), Ethiopia (urban¼3.6%,
rural¼1.6%), Ghana (urban rich¼4.7%, urban poor¼2.2%,
rural¼1.4%), Kenya (urban¼22.9%, rural¼13.2%) and Saudi
Arabia (urban¼14.9%, rural¼5.4%).21e26 Similarly, crosssectional studies conducted at different times have provided
evidence for temporal increases in the prevalence of asthma in
both urban and rural populations,27 28 although the urbanerural
difference in asthma prevalence may be narrowing.10 29 Previous
ecological studies have compared asthma prevalence between
cities20 within a country or between countries represented by
one or more cities.2 Inter-country comparisons such as the
International Study of Asthma and Allergies in Childhood2 are
unlikely to provide useful insights into the effects of urbanisation on asthma prevalence given that most study centres were in
cities and different countries are likely to have different experiences of the individual urbanisation processes. Intra-country
comparisons may be more useful: an ecological comparison
between several cities in Brazil found that asthma prevalence
was greater in cities with more poverty and inequality.20
The present study, which compared small rural communities,
had significant heterogeneity in the levels of exposure to the
Associations between community asthma prevalence and urbanisation indices
Asthma prevalence
Infrastructure
Indices
r
p
r
p
r
Socioeconomic
p
r
Lifestyle
p
Infrastructure
Socioeconomic
Lifestyle
Summary urbanisation
0.173
0.295
0.342
0.355
0.190
0.023
0.008
0.006
1
0.757
0.671
0.840
e
<0.001
<0.001
<0.001
e
1
0.819
0.937
e
e
<0.001
<0.001
e
e
1
0.895
e
e
e
<0.001
Results shown are Spearman’s rank correlation coefficients.
Thorax 2011;66:1043e1050. doi:10.1136/thoraxjnl-2011-200225
1047
Thorax: first published as 10.1136/thoraxjnl-2011-200225 on 8 August 2011. Downloaded from http://thorax.bmj.com/ on June 19, 2020 by guest. Protected by copyright.
Figure 1 Scatter plots of the
relationships between community
asthma prevalence (measured by the
proportion of children with wheezing in
the last 12 months) and z scores for the
first components of infrastructure (A),
socioeconomic (B), lifestyle (C) and
summary urbanisation (D) indices. The
regression line is shown for each
relationship. Red squares represent
outliers and extreme observations
identified on residual analysis in
bivariate linear regression.
Asthma and the environment
Associations between community asthma prevalence and individual indicators of infrastructure, socioeconomic and lifestyle indices
Bivariate analyses
Groups
Indicators
Infrastructure
Administrative grade (parish)
Spatial organisation
(a) Disperse (reference)
(b) Blocks
(c) Neighbourhood
Transport access (road)
Electrical grid (Yes)
Piped-water system (Yes)
Telephone system (Yes)
Health centre (Yes)
Pharmacy (Yes)
Educational institution (Yes)
Shops
(a) 0e1 (reference)
(b) 2e5
(c) 6e15
(d) >15
Father’s education
Mother’s education
Household income
Access to electricity
Material goods
Cement house
Gas for cooking
Motor vehicles
Uncrowded household
Hamburger consumption
Fizzy drink consumption
No physical exercise
Television in house
TV viewing
No farming activities
Cat in house
Dog in house
Migration
Parasite infection rate
Socioeconomic
Lifestyle
b
Multivariate analyses
p
0.017
0.989
0.147
0.615
1.377
2.226
0.455
0.939
2.076
2.296
0.449
e
0.919
0.712
0.263
0.142
0.721
0.443
0.089
0.058
0.715
0.407
1.603
1.994
0.072
0.023
0.072
0.029
0.104
0.009
0.046
0.164
0.053
0.104
0.121
0.064
0.034
0.048
0.065
0.091
0.050
0.061
0.009
e
0.866
0.484
0.390
0.19
0.62
0.077
0.077
0.05
0.865
0.019
0.012
0.235
0.015
0.004
0.12
0.111
0.046
0.05
0.025
0.191
0.22
0.777
e
e
b
CI (95%)
p
6.059
5.117
9.313 to 2.805
1.702 to 8.532
<0.001
0.004
0.706
3.974
5.065
3.664 to 5.077
0.433 to 8.381
0.168 to 9.961
0.747
0.076
0.043
0.074
0.099
0.045
0.193 to 0.046
0.11 to 0.208
0.005 to 0.084
0.221
0.078
0.028
e
e
0.127
0.247 to
0.008
0.037
0.168
0.043 to 0.294
0.009
0.076
0.132
0.069
0.034
0.025 to 0.176
0.041 to 0.223
0.012 to 0.149
0.088 to 0.020
0.137
0.005
0.093
0.214
0.049
0.035 to 0.133
0.248
Results shown are for bivariate and multivariate linear regression analyses.
urbanisation indicators between communities and was able to
measure the potential effects of these indicators on asthma
prevalence. The study was able to demonstrate that lifestyle
indicators and socioeconomic indicators had stronger overall
effects on asthma prevalence than infrastructure indicators
(figure 1), indicating that a higher asthma prevalence was
present in communities with a higher socioeconomic level and
a more urbanised lifestyle. Indeed, asthma prevalence increased
with increasing household income, access to electricity, material
goods, gas for cooking and possession of motor vehicles. Our
findings are consistent with the findings of other studies
conducted in transitional communities that have shown an
association between better socioeconomic status and asthma
prevalence (ie, urban rich vs urban poor).22 24 Rural residence has
been repeatedly shown to be protective against allergic diseases
and this effect has been attributed to farming exposures.8
Consistent with these, communities with a higher percentage of
households not involved in agricultural activities had a higher
prevalence of asthma. Rates of parasite infection with geohelminth infections were not associated with asthma prevalence in
our study population in agreement with the findings of previous
studies.11 30 The increased consumption of fast foods such as
hamburgers and soft drinks (that represent new food habits)
1048
was associated with greater asthma prevalence and is consistent
with individual-level studies that have shown changes in diet of
populations undergoing transition is associated with an increase
in asthma.31 32 Factors associated with a sedentary lifestyle such
as limited physical activity and watching TV for more than 1 h
daily were associated with greater asthma prevalence, and are
consistent with the findings of previous studies.33 34 Somewhat
paradoxically, ownership of motorised vehicles was associated
with an increased asthma prevalence but road access with
a decreased asthma prevalence. Road access does not necessarily
mean a higher level of urbanisation than river access because
most roads are dirt tracks with limited public transportation and
accessible communities have few privately-owned vehicles.
Communities with river access actually had higher levels
of ownership of motor vehicles (motorised canoes), and such
ownership could be associated, for example, with greater
levels of environmental pollution with exhaust particulates,
a potential mechanism mediating an increase in asthma.
The lifestyle index that showed the strongest association with
the prevalence of asthma was also strongly associated with the
other urbanisation indices. These relationships suggest that the
growing influence of urban lifestyle on rural societies may
explain at least partly the increase in asthma prevalence in rural
Thorax 2011;66:1043e1050. doi:10.1136/thoraxjnl-2011-200225
Thorax: first published as 10.1136/thoraxjnl-2011-200225 on 8 August 2011. Downloaded from http://thorax.bmj.com/ on June 19, 2020 by guest. Protected by copyright.
Table 4
Asthma and the environment
Acknowledgements The Ecuadorian Elimination Programme for Onchocerciasis
(Dr Eduardo Gomez, Dra. Raquel Lovato, Lcda. Margarita Padilla, Lcda Anabel Ponce,
Lcda Ing Sandra Barrera, Magdalena Cortez) and CECOMET (Dr Gregorio Montalvo
and Lcda Monica Marquez) are thanked for support in visiting communities and
Thorax 2011;66:1043e1050. doi:10.1136/thoraxjnl-2011-200225
providing community censuses. The health promoters, school teachers, parents and
children are thanked for their enthusiastic cooperation. The study forms part of the
SCAALA (Social Changes, Asthma, and Allergies in Latin America) programme of
research.
Funding Wellcome Trust, UK, HCPC Latin American Centres of Excellence Programme
(ref 072405/Z/03/Z). The funders had no role in study design, data collection and
analysis, decision to publish or preparation of the manuscript.
Competing interests None.
Ethics approval The study protocol was approved by the ethics committee of the
Hospital Pedro Vicente Maldonado, Ecuador.
Contributors Study design: PJC, AR, MC, LCR, MLB. Data collection; AR, MV, GO,
SE. Data analysis: AR,CT. Draft manuscript: AR, PJC. Manuscript review: AR, PJC, LR,
MLB.
Provenance and peer review Not commissioned; externally peer reviewed.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
Woolcock AJ, Peat JK. Evidence for the increase in asthma worldwide. Ciba Found
Symp 1997;206:122e34.
Anon. Worldwide variation in prevalence of symptoms of asthma, allergic
rhinoconjunctivitid, and atopic eczema: ISSAC. The international Study of Asthma and
Allergies in Childhood ISAAC. Lancet 1998;351:1225e32.
Pearce N, Ait-Khaled N, Beasley R, et al. Worldwide trends in the prevalence of
asthma symptoms: phase three of the International Study of Asthma and Allergies in
Childhood (ISAAC). Thorax 2007;62:758e66.
Beasley R, Crane J, Lai CK, et al. Prevalence and etiology of asthma. J Allergy Clin
Immunol 2000;105:S466e72.
Von Hertzen L, Haahtela T. Asthma and atopydthe price of affluence? Allergy
2004;59:124e37.
Weinberg EG. Urbanization and childhood asthma: an African perspective. Allergy
Clin Inmunol 2000;105:224e31.
Singh M. The burden of asthma in children. An Asian perspective. Paediatr Respir
Rev 2005;6:14e19.
Von Hertzen L, Haahtela T. Disconnection of man and the soil: reason for the
asthma and atopic epidemic? Allergy 2006;117:334e44.
Nicolaou N, Siddique N, Custovic A. Allergic disease in urban and rural populations:
increasing prevalence with increasing urbanization. Allergy 2005;60:1359e60.
Viinanen A, Munhbayarlah S, Narantsetseg L, et al. Prevalence of asthma, allergic
rhinocongunctivitis and allergic sensitisation in Mongolia. Allergy 2005;60:1370e77.
Cooper PJ, Rodrigues LC, Cruz AA, et al. Asthma in Latin America: a public health
challenge and research opportunity. Allergy 2009;64:5e17.
Fischer GB, Camargos PAM, Mocelin HT. The burden of asthma in children: a Latin
American perspective. Pediatr Respir Rev 2005;6:8e13.
Germani G. El proceso de urbanización en los paı́ses avanzados y en los paı́ses en
desarrollo. In: Germani G, ed. Sociologı́a de la Modernización. Buenos Aires: PAIDOS
Press, 1971:146e76.
Caicedo CM. Documento de Diagnóstico del Área de Salud # 7 Borbón. Cantón Eloy
Alfaro provincia de Esmeraldas, 2004.
Cooper PJ, Chico ME, Rodriguez A, et al. Risk factors for asthma and allergy
associated with urban migration: background and methodology of a cross-sectional
study in Afro-Ecuadorian school children in Northeastern Ecuador (EsmeraldasSCAALA Study). BMC Pulm Med 2006;6:24e31.
Weiland SK, Bjorksten B, Brunekreef B, et al. Phase II of the International Study of
Asthma and Allergies in Childhood (ISAAC II): rationale and methods. Eur Respir J
2004;24:406e12.
Maroco J. Análise de Componentes Principais. In: Maroco J, ed. Análise
EstatisticadCom utilização do SPSS. Lisboa: Silabo Press, 2003:231e57.
Meulman J. Optimal Scaling Methods for Multivariate Categorical Data Analysis.
Leiden: Leiden University, 2000.
Hair JF, Anderson RE, Tathan RL, et al. Análise factorial. In Análise Multivariada de
dados. Porto Alegre: Bookman Press, 2005:88e130.
Cunha SS, Pujades MR, Barreto ML, et al. Ecological study of socio-economic
indicators and prevalence of asthma in schoolchildren in urban Brazil. BMC Public
Health 2007;7:205.
Steinman HA, Donson H, Kawalski M, et al. Bronchial hyper-responsiveness and
atopy in urban, peri-urban and rural South African Children. Pediatr Allergy Immunol
2003;14:383e93.
Keeley DJ, Neill P, Gallivan S. Comparison of the prevalence of reversible airways
obstruction in rural and urban Zimbabwean children. Thorax 1991;46:549e53.
Yemaneberhan H, Bekele Z, Venn A, et al. Prevalence of wheeze and asthma and
relation to atopy in urban and rural Ethiopia. Lancet 1997;350:85e90.
Addo Yobo EO, Custovic A, Taggart SC, et al. Exercise induced bronchospasm in
Ghana: differences in prevalence between urban and rural schoolchildren. Thorax
1997;52:161e5.
Ng’ang’a LW, Odhiambo JA, Mungai MW, et al. Prevalence of exercise induced
bronchospasm in Kenyan school children: an urban-rural comparison. Thorax
1998;53:919e26.
Hijazi N, Abalkhail B, Seaton A. Asthma and respiratory symptoms in urban and rural
Saudi Arabia. Eur Respir 1998;12:41e4.
1049
Thorax: first published as 10.1136/thoraxjnl-2011-200225 on 8 August 2011. Downloaded from http://thorax.bmj.com/ on June 19, 2020 by guest. Protected by copyright.
populations and the declining gap between urban and rural
asthma prevalence. This is illustrated in the present study by the
high rates of television viewing, consumption of soft drinks and
presence of material goods in some of the communities. The
weak associations between the infrastructure index (and most
indicators that constituted this index) with asthma prevalence
might be related to the fact that in communities undergoing
transition in developing countries some elements of the urbanisation process develop less faster than others changes (eg, social
and consumption changes).13 For example, a high proportion of
communities lacked basic services but had a high percentage of
television ownership. The implementation of basic infrastructure and basic services tends to lag behind other urbanisation
processes because of a lack of economic resources from highly
centralised political establishments that are reluctant to invest
in populations that do not represent their primary political
constituencies. The disparities in the levels of urbanisation in
developing countries could help explain the variation of asthma
prevalence between and within countries of Latin America.
Our study findings from an ecological analysis are subject to
some limitations. First, the associations observed at the
community level may not reflect associations at the individual
level (ecological fallacy). Despite the fact that we collected
detailed information on urbanisation indicators, other data that
are likely to be important (eg, environmental sampling for air
pollution) were not available. Smoking was not included in our
analysis because in the study area, a widespread traditional
lifestyle habit is to smoke a traditional cigarette, ‘cachimba’, and
thus smoking in the context of this study, given that we did not
collect data to distinguish smoking of traditional versus modern
cigarettes, is unlikely to be a useful explanatory variable for
urbanisation. We defined asthma as having a wheeze in the
previous 12 months. This is a widely used definition and has the
advantage of simplicity, but of course not all wheezing is
asthma. However, we believe that this definition is likely to be
less subject to misclassification compared with other definitions
(such as doctor diagnosis of asthma) used in rural populations
with limited access to healthcare. Our analysis is limited in
degree to which we can attribute urbanisation as being a causal
factor for asthma prevalencedmany of the variables included in
our analysis are markers of lifestyle and while being associated
with asthma prevalence may not necessarily be causal.
In conclusion, we have explored the effects of factors associated with urbanisation on the community prevalence of asthma
in rural transitional communities in a developing country.
Clearly, urbanisation is a complex process that comprises an
array of changes that modify the living environment and human
behaviours. The consequence of such changes that penetrate
inexorably over time into rural environments may be to deprive
urbanising populations of childhood exposures that are important for the development of normal airway function. Our data
provide evidence that the process of urbanisation is associated
with a greater prevalence of asthma in rural communities in
Ecuador. The high level of socioeconomic inequality and the
rapid but variable rate of urbanisation occurring in Latin
America may explain temporal trends and differences in asthma
prevalence in this region.35 To explore these findings better,
future studies should pay attention to the social mechanisms
and processes that modify the rural way of life towards an urban
lifestyle.
Asthma and the environment
28.
29.
30.
31.
1050
Van Niekerk CH, Weinberg EG, Shore SC, et al. Prevalence of asthma:
a comparative study of urban and rural Xhosa children. Clin Allergy
1979;9:319e24.
Addo-Yobo EO, Woodcock A, Allotey A, et al. Exercise-induced bronchospasm and
atopy in Ghana: two surveys ten years apart. PLoS Med 2007;4:e70.
Calvert J, Burney PGJ. Increase in prevalence of exercise induced bronchospasm in
a rural and urban population of African school children. Curr Allergy Clin Immunol
2003;16:117.
Moncayo A, Vaca M, Oviedo G, et al. Risk factors for atopic and non atopic asthma
in rural area of Ecuador. Thorax 2010;65:409e16.
Hijazi N, Abalkhail B, Seaton A. Diet and childhood asthma in a society in transition:
a study in urban and rural Saudi Arabia. Thorax 2000;55:775e9.
32.
33.
34.
35.
Hooper R, Calvert J, Thompson RL, et al. Urban/rural differences in diet and atopy in
South Africa. Allergy 2008;63:425e31.
Jones SE, Merkle SL, Fulton JE, et al. Relationship between asthma, overweight,
and physical activity among U.S. high school students. J Community Health
2006;31:469e78.
Firrincieli V, Keller A, Ehrensberger R, et al. Decreased physical activity among
Head Start children with a history of wheezing: use of an accelerometer to measure
activity. Pediatr Pulmonol 2005;40:57e63.
Asher MI, Montefort S, Björkstén B, et al; The ISSAC Phase three Study Group.
Worldwide time trends in the prevalence of symptoms of asthma, allergic
rhinoconjunctivitis, and eczema in childhood: ISAAC Phases One and Three repeat
multicountry cross-sectional surveys. Lancet 2006;368:733e43.
Thorax 2011;66:1043e1050. doi:10.1136/thoraxjnl-2011-200225
Thorax: first published as 10.1136/thoraxjnl-2011-200225 on 8 August 2011. Downloaded from http://thorax.bmj.com/ on June 19, 2020 by guest. Protected by copyright.
27.