Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
A cohort study of air-pollution and childhood obesity incidence
(USC Thesis Other)
A cohort study of air-pollution and childhood obesity incidence
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
1
A cohort study of air-pollution and childhood obesity incidence
by
Xiaoyi Zhou
A Thesis Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of
the Requirements for the Degree
MASTER OF SCIENCE
(Applied Biostatistics and Epidemiology)
August 2016
2
Acknowledgement
I would like to express my deep and sincere gratitude to my thesis chair Dr. Frank
Gilliland for his patience, encouragement, and enthusiastic supervision during this
process. He has been an excellent mentor to me. I would also like to thank my thesis
committee members Dr. Stanley Azen and Dr. Don Barkauskas for their helpful
suggestions and support.
3
Table of Contents
Abstract ...................................................................................................................... 4
Introduction ................................................................................................................ 5
Methods ..................................................................................................................... 7
Results ..................................................................................................................... 12
Discussion ................................................................................................................ 19
Conclusion ............................................................................................................... 23
Reference ................................................................................................................. 24
4
Abstract
Background: Air pollution, especially combustion products, could activate metabolic
disorders through inflammatory pathways, potentially leading to obesity. The effect of
air-pollution on BMI growth was shown by a previous study. Here we sought to
determine whether ambient air pollution and traffic air pollution are associated with
obesity incidence in children from age 8–18 years old.
Methods: 3,887 children aged 8-12 were selected from two prospective cohorts in the
Children’s Hospital Study (CHS) and followed for 9 years, with height and weight
measured annually. A child's weight status was determined using an age- and
sex-specific percentile for BMI based on the CDC’s chart. Average annual ambient
exposure was directly collected and measured at monitoring stations in 12 selected
communities. Dispersion models were used to estimate exposure to traffic-related air
pollution. Logistic regression was performed to analyze the association between
ambient-related and traffic-related air pollution and obesity prevalence at the study
entry; proportional hazard regression was performed for association with obesity
incidence over 8-18 age periods during 9-year follow up. Models were adjusted by
children baseline characteristics variables, with random effects of communities, sex,
and cohort.
Results: For obesity prevalence at age 8, ambient air pollutants including O 3
(OR=1.20, 95%CI: 1.18, 1.22), NO 2 (OR=1.22, 95%CI: 1.20, 1.23), and PM 2.5
(OR=1.21 95%CI: 1.13, 1.25), and total traffic NOX pollutants (OR=1.09, 95%CI:
0.98, 1.16) had significant influence on obesity at the study entry and continued
robust association after adjustments. After adjusted by all covariates, O 3 (HR=1.49,
95%CI: 1.08, 2.06), NO 2 level (HR=1.64, 95%CI: 1.06, 2.52), and NO level
(HR=1.43, 95%CI: 1.04, 1.85) were significantly associated with obesity incidence in
5
our cohorts. However, no significant association was found between traffic air
pollution and obesity incidence.
Conclusions: This result strengthens emerging evidence that exposure to ambient air
pollution contribute to development of childhood obesity. However, we still cannot
conclude that there is a significant association between childhood obesity incidence in
children aged 8–18 years and traffic air pollution based on the results of this study.
Keywords: Childhood obesity, Air pollution, Traffic, Ambient, California
Introduction
In both developed and developing countries, social economic development and
general changes in life style have brought about an epidemic of obesity
[1]
. When
considering this epidemic on a global scale, the rapidly increasing prevalence of
childhood overweight and obesity is of particular concern. In recent years, obesity
during childhood has emerged as one of today’s leading health threats in the United
States and most countries around the world. According to the National Health and
Nutrition Examination Survey (NHANES) data from 2011 to 2012, 16.9% children
and adolescents aged 2-19 years and 34.9% of adults aged 20 years or older were
obese
[2]
. 12.7 million children and adolescents have been affected for the past decade.
These increased rates of obesity during childhood increased the risks of obesity during
adulthood and short- and long-term risks of cancer, type 2 diabetes, hypertension, and
cardiovascular disease
[3]
.
A growing body of literature has shown that obesity magnifies the effects of air
pollution on physiologic responses in the lungs of experimental animals and humans.
Although the most commonly cited reasons for the growing obesity epidemic include
6
the rapidly widening availability of low-cost high-fat diets, certain food marketing
practices, and institutionally and culturally driven reductions in physical activity
[4]
,
the multifactorial causality for obesity including genetic, dietary, economic,
psychological, reproductive, pharmacological, and environmental factors has been
well established
[5]
. In addition, the relationship between air pollution and obesity has
caught scientists’ interest in recent years. Some research also indicates that air
pollution exposure, especially the combustion products from traffic pollution, may
initiate metabolic processes contributing to diabetes formation through the
inflammatory pathways
[6]
. For example, prenatal or postnatal environmental tobacco
smoke exposure appears to be associated with an increase in BMI in children
[7][8]
.
The metabolic mechanism was also supported by mice-based experiments
[9][10]
. An
experimental study reported that airborne fine particulate matter (diameter, <2.5
[PM 2.5]) exposure can induce metabolic abnormalities and obesity in mice fed a
normal diet
[9]
, suggesting that ambient air pollution may be associated with the
development of obesity. The health outcome of previous CHS papers that evaluated
the association of childhood obesity with growth in body mass index (BMI=kg/m
2
) in
children aged 5-11 years (Cohort E), found that traffic related pollutants level were
strongly associated with BMI growth during the study follow-up
[11]
. The effect size in
the adjusted model indicated about a 13.6% increase in annual BMI growth when
compared the lowest to the highest tenth percentile of air pollution exposure, resulting
in an increase of nearly 0.4 BMI units on attained BMI at age 10
[11]
.
However, to our knowledge, there have been few previous statistical evaluations of
the effects of traffic-related and ambient air pollution on obesity incidence. One
possible reason is that a long follow-up cohort study with periodic measurements of
BMI and air-pollution data are difficult to perform. The Children’s Hospital Study
7
(CHS) also collected information on exposure to ambient particles and NRP
information during the follow up periods. This is an ideal opportunity to examine
the effects of air pollution exposures on the development of obesity by incidence
analysis.
We hypothesized that air pollution exposure is associated with increased risk of
developing obesity from age 8 to 18 in 12 Southern California communities from
CHS Cohort C and D. This paper will expand the earlier result of traffic as a risk
factor by examining the specific pathway of air pollution exposure from both ambient
and traffic air-pollution effects.
Methods
Study Design
A cohort of children attending elementary school (age 8–12 years) was enrolled
during the 1993-2001 and 1996-2004 respectively across 12 communities in Southern
California (N = 4,550). At baseline, a parent or guardian of each participating child
provided written informed consent and completed a written questionnaire that
supplied detailed information on family demographic characteristics (e.g. annual
family incomes, parental education), history of respiratory illness and associated risk
factors (e.g. second hands smoking, wheeze history, and asthma history), indoor
sources of exposure, physical activity patterns of the children, and household
characteristics. An update questionnaire was completed by each child in each year,
and anthropometric measures such as height and weight were obtained. Trained
technicians followed a standardized procedure to measure height and weight using a
calibrated medical scale at baseline and annually through the entire 9-year follow-up.
8
Measurements were recorded to the nearest 1 cm and 1 lb (0.45 kg), respectively.
These objective measures of height and weight allowed for accurate calculation of
BMI (kg/cm2). Homes of the children were geo-coded using the Tele Atlas geocoding
database to the corresponding road network. The research protocol was approved by
the Institutional Review Boards of the University of Southern California and the
University of California, Berkeley
[13]
. (More details on the built environment variable
compilation are described in
[13]
)
The analytical data set excluded 471 children who were already obese (BMI
percentile>=95%) at the baseline year, and was restricted to children who had two or
more measurements of height and weight (N = 2745).
Definition of obesity and overweight
BMI was calculated as weight (in kg) divided by height (in cm) squared (kg/m
2
), and
was categorized into age- and sex-specific percentiles based on the Centers for
Disease Control and Prevention (CDC) BMI growth charts: normal weight was
defined as a BMI below 85th percentile, overweight was defined as a BMI at or
greater than 85th percentile and below 95th percentile, and obesity was defined as a
BMI at or above 95th percentile.
Air-pollution Measurement
Ambient air pollution
Ambient air pollution data were acquired from a central monitor. Measurements were
made for the concentrations of several pollutants, including carbon monoxide (CO),
nitrogen dioxide (NO 2), nitrogen oxide (NO), elemental and organic carbon,
9
particulate matter ≤ 10 μm (PM10) and ≤ 2.5 μm (PM2.5) in aerodynamic diameter, and
ozone (O3), which are indicators of incremental increases due to primary emissions
from local vehicular traffic on top of background ambient levels. Measurements of
concentrations of these ambient pollutants in each 12 communities were obtained at
air pollution monitoring stations from 1992 to 2002. Twenty-four-hour pollutant
concentration averages were constructed for each above pollutants when at least 18
valid hourly values were available with not more than six successive hourly values
missing. For O3, the 1-h daytime maximum level, average level during 10am-6pm,
and average 24-hour ozone level were also included. In our analysis, we use ozone
average concentration during 10am-6pm to represent the ozone level.
Near-roadway air pollution exposure
Annual average exposure to local NRP at homes and schools was estimated from
CALINE4 dispersion models that incorporate distance to roadways, vehicle counts,
vehicle emission rates, wind speed and direction, and height of the mixing layer in
each community (Benson 1992). For our analysis, we characterized NRP exposure
using modeled NOx, including FCC2 (Highway NOx), FCC3 (Main road NOx),
FCC4 (), Freeway NOx (Local Freeway), non-freeway NOx (Local Non-freeway:
FCC2+FCC3+FCC4), and Total NOx (Caline4 est NOx from local traffic), to
represent the incremental contribution of local traffic to a more homogeneous
community background concentration of NOx that included both primary and
secondary pollution resulting from long range transport and regional atmospheric
photochemistry.
10
Other Covariates
Independent variables that were included as adjustments are listed in the Tables 1. A
final model was developed by including all additional confounders that individually
changed the effect of interest on the hazard ratio change in obesity incidence (slope)
by at least 10%. The categories for each of the covariates (with the reference category
underlined) are: sex (boy, girl), age (<9, 9-9.9, 10-10.9, 11-11.9, 12-12.9), ethnicity
(Non-Hispanic white, Hispanic white, Other), baseline overweight (No, Yes), annual
family incomes (Less than $15,000, $15,000 to $49,999, $50,000 or more), parental
education (Completed grade 12 or less, some college or technical school, more than
Completed 4 years of college), physical activity (mainly in the form of programmed
activities and team sports)(0, 1, 2 Weekly days of outdoor sports), insurance (No, Yes),
parental smoking (Never Smoking, past smoking, continual smoking, new smokers),
wheeze history (No, Yes), and asthma history (No, Yes). A missing indicator category
was included for missing values for all above covariates.
Statistical methods, time censoring
Obesity prevalence at the study entry among all study participates in CHS cohorts C
and D were first checked by conditional logistic regression. Community was treated
as random effect, and sex was adjusted as fixed effects. The main exposure variables
were set as standardized average pollutant concentrations during the year prior to the
baseline date.
The cox proportional hazard model then was performed to examine the effects of
air-pollution related risk factors on obesity incidence among children in CHS cohorts
C and D. New developed obesity is the event of interest. Children who were obese at
11
the study entry had been removed from cox regression. Time to obesity was accurate
to a day, and was calculated based on the date of children’s annual physical
examination. During a 9-year follow up, children will be censored once they have
developed obesity. In all our models, cohort was adjusted as a fixed effect, community
as a random effect, and sex as a strata factor. In the adjusted model, all the other
potential confounders in Table 1 were added and adjusted as fixed effects.
The effects of exposure to ambient and traffic air pollution on childhood obesity
incidence were modeled as time-dependent variables and evaluated in several steps.
Traffic air-pollution exposures were estimated from CALINE4 dispersion models and
have records for each month. The mean of each of the traffic air-pollution exposures
was calculated for each study visit during the study period prior to that date, and then
standardized by the standard deviation of the corresponding cohort at the first year.
The baseline for air-pollution was set one year prior to the baseline date. The ambient
air-pollution exposures were measured in each year for each community, and then
were calculated and standardized with methods similar to those used for traffic
air-pollution. PM2.5 has 12 missing data in 12 communities in 1993, PM10 has 5
missing data in 5 communities in 1993, and NO has several random missing in each
community during study period. For these missing data, based on the sensitivity
analysis results, the baseline year was kept to be 1993 for both cohort C and D, so the
missing records were censored rather than being manipulated by the records of the
following year. Therefore, different pollutant exposures have different sample size. In
the combined C+D cohort, baseline age for two cohorts was set at age 9. Since more
than 50% of the information was not recorded for CO during the follow up periods,
CO was excluded in this analysis.
Based on the interaction test, since the pollutants effects on obesity were not differed
12
by cohort, therefore, results were presented in combined C and D cohorts to show the
final association. The cohort was adjusted as fixed effect and community were
adjusted as random effects in all models. The results are presented as hazard ratios
(HRs) with 95% CIs, per standard deviation increase. Statistical significance was
based on a probability of 0.05 for a two-sided hypothesis. Data were analyzed using
SAS (version 9.4) software.
Results
The mean age of children at study entry for cohort C and D was 9.7 years. Based on
Centers for Disease Control percentiles between the 85th and 95th percentile, rates of
overweight at the study entry were 25.7%. Obesity rates measured as BMI scores
equal to or greater than the 95th percentile were 13.1%.
Table 1 gives the descriptive statistics on children and the confounders selected
through the testing procedure described in the Methods above.
Based on the univariate cox regression analysis at the study entry, the number of boys
and girls were almost equal, but girls were less likely than boys to develop obesity
(HR=0.67, 95%CI: 0.58, 0.77). The majority of the cohort is White, Non-Hispanic
(56.7%), while a large proportion (34.3%) is of African and Asian American ethnicity.
Hispanic comprised about 8% of the cohort. African and Asian American children had
the highest risk (HR=1.29, 95%CI: 1.10, 1.51); Hispanic white children were 1.04
times higher than Non-Hispanic White children to become obese. Especially, for
children who were overweight at the study entry have a 24.10 (95%CI: 19.97, 29.08)
higher risk to develop obesity in the future than other children. 27.2% of the children
were exposed to second hand smoking from parents, and those exposed children were
13
more likely to be obese than unexposed children; 13% and 31% had asthma (HR=1.38,
95%CI: 1.14, 1.66) and wheeze (HR=1.28. 95%CI: 1.10, 1.48) respectively. Social
economics is usually regarded as a risk factor for obesity. In our data, children from
families with annual income range $15,000 - $49,999 (HR=1.03, 95%CI: 0.83, 1.28)
had a slightly lower chance than those from families with annual income less than
$15,000, while children from families whose annual income is more than $50,000 had
the obvious lower risk (HR=0.49, 95%CI: 0.39, 0.62). Low parental education was
found to be a risk factor of obesity, which is also correlated to family annual income.
The children with parents of the lowest education level have the highest obesity risk;
however, there is no linear trend for this association. However, the association
between annual family income and parental education level was not significant among
either baseline non-obese children or the whole cohort. Since family annual income
was highly correlated with parental education level, their relationship with obesity is
likely confounded by each other. Risk was lower among children who did outdoor
sports than those who didn’t. But there is no linear trend for the association. Children
who had outdoor sports 1 day per week (HR=0.61, 95%CI: 0.50, 0.73) had the lowest
risk, rather than more active children who did weekly outdoor sports more than one
day (HR=0.93, 95%CI: 0.78, 1.11), when compared to those didn’t do outdoor sports.
It is probably because the benefits of physical activities to prevent obesity were
counterbalanced by air-pollution. Additionally, obesity risk is significantly higher
among children have insurance in cohorts including baseline obese children (HR=0.88,
95%CI: 0.79, 0.98). (I don’t understand this previous sentence at all)
Table 1 Baseline Characteristics and Univariate association between the risk of
developing obesity during the follow-up among participants in cohorts C
14
(n=1,287) and D (n = 1,458) and exposures at Children’s Health Study
enrollment.
Baseline non-obese children
(n=2,745)
The whole cohort including
baseline obese children (n=3,154)
Frequency (%) HR (95%CI) p Frequency (%) HR (95%CI) p
Baseline Characteristics
Sex
Boys
1352 (49.25%)
1.00
1574 (49.90%)
1.00
Girls 1393 (50.75%) 0.67 (0.58, 0.77) <0.0001 1580 (50.10%) 0.73 (0.67, 0.79) <0.0001
Age
(8,9] 109 (4.0%) 1.00 124 (3.93%) 1.00
(9,10] 1951 (71.1%) 0.60 (0.45, 0.81) 0.0007 2255 (71.50%) 0.89 (0.73, 1.08) 0.24
(10,12] 685 (25.0%) 0.63 (0.46, 0.86) 0.0043 775 (24.57%) 0.81 (0.66, 1.00) 0.05
Race/ Ethnicity
Non-Hispanic Whites 1557 (56.7%) 1.00 1730 (54.85%) 1
Hispanic Whites 226 (8.2%) 1.04 (0.79, 1.36) 0.0053 258 (8.18%) 1.28 (1.11, 1.48) 0.0009
Others 941 (34.3%) 1.29 (1.10, 1.51) 0.0019 1141 (36.18%) 1.45 (1.33, 1.59) <0.0001
Missing 21 (0.8%) 25 (0.79%)
Individual and Household
Characteristics
Overweight
No 2281 (83.1%) 1.00 2281 (72.32%) 1
Yes
464 (16.9%)
24.10(19.97,
29.08) 0.0004
873 (27.68%) 47.01(39.47,
56.00) <0.0001
Physical Activity
Weekly days of outdoor
sports 0 1274 (46.41%) 1.00
1466 (46.48%)
1
1 851 (31.0%) 0.61 (0.50, 0.73) <0.0001 972 (30.82%) 0.83 (0.75, 0.91) 0.0001
>1 620 (22.6%) 0.93 (0.78, 1.11) 0.0054 716 (22.70%) 1.03 (0.93, 1.14) 0.59
Child had health
insurance
No 382 (13.9%) 1.00 458 (14.52%) 1
Yes 2293 (83.5%) 0.96 (0.78, 1.17) 0.0004 2614 (82.88%) 0.88 (0.79, 0.98) <0.0001
Missing 70 (2.6%) 82 (2.60%)
Wheeze
No 1746 (63.6%) 1.00 1985 (62.94%) 1
Yes 854 (31.1%) 1.28 (1.10, 1.48) 0.0013 1001 (31.74%) 1.33 (1.22, 1.45) <0.0001
15
Missing 145 (5.3%) 168 (5.33%)
Asthma
No 2319 (84.5%) 1.00 2643 (83.80%) 1
Yes 365 (13.3%) 1.38 (1.14, 1.66) 0.0003 436 (13.82%) 1.37 (1.23, 1.52) <0.0001
Missing 61 (2.2%) 75 (2.38%)
Annual Family income
Less than $15,000 377 (13.7%) 1.00 472 (14.97%) 1
$15,000 to $49,999 948 (34.5%) 1.03 (0.83, 1.28) 0.78 1099 (34.84%) 0.80 (0.72, 0.90) 0.8
$50,000 or more 1020 (37.2%) 0.49 (0.39, 0.62) <0.0001 1118 (35.45%) 0.50 (0.44, 0.57) 0.5
Missing 400 (14.6%) 465 (14.74%)
Parental education
Completed grade 12
or less 851 (31.0%) 1.00
996 (31.58%)
1
Some college or
technical school 1173 (42.7%) 0.90 (0.83, 1.18) 0.9
1338 (42.42%)
0.94 (0.86, 1.04)
0.23
More than
Completed 4 years of
college 618 (22.5%) 0.94 (0.81, 1.22) 0.94
691 (21.91%)
0.81 (0.72, 0.92)
0.0006
Missing 103 (3.75%) 129 (4.09%)
Paternal smoking
Never Smoking 1854 (67.5%) 1.00 2110 (66.90%) 1
Past Smoking 318 (11.6%) 0.80 (0.61, 1.03) 0.09 358 (11.35%) 0.97 (0.85, 1.11) 0.07
Continual Smoking 298 (10.9%) 1.22 (0.97, 1.54) 0.08 347 (11.00%) 1.25 (1.10, 1.41) 0.04
New Smoker 130 (4.74%) 1.78 (1.36, 2.33) <0.0001 162 (5.14%) 1.62 (1.39, 1.89) <0.0001
Missing 145 (5.28%) 177 (5.61%)
Based on logistic regression in Table 2, the prevalence of obesity at the study entry
among children 8 year-old was positively associated with O3 (OR=1.20, 95%CI: 1.18,
1.22), NO2(OR=1.22, 95%CI: 1.20, 1.23), and PM2.5 (OR=1.21 95%CI: 1.13, 1.25).
And those associations were robust after adjustment by covariates in Table1. However,
we have not found the association in other ambient pollutants, probably because of the
missing data. Freeway NOX (OR=1.05, 95%CI: 0.99, 1.13), Non-freeway NOX
(OR=1.02, 95%CI: 1.00, 1.03), and Total NOX (OR=1.10, 95%CI: 1.09, 1.19). The
association with Total NOX (OR=1.09, 95%CI: 0.98, 1.16) was also robust even after
16
adjustment by the covariates in Table1.
Table 2 The association between obesity prevalence and air pollution
Unadjusted Adjusted
OR (95%CI) P OR (95%CI) P
O
3
2
1.20 (1.18, 1.22) 0.001 1.23 (1.18, 1.32) 0.003
NO
2
1.22 (1.20, 1.23) 0.03 1.24 (1.22, 1.26) 0.02
NO 1.19 (1.16, 1.23) 0.12 1.20 (1.16, 1.22) 0.09
PM10 1.08 (1.05, 1.12) 0.10 1.11 (1.05, 1.13) 0.07
PM2.5 1.21 (1.13, 1.25) 0.03 1.22 (1.18, 1.25) 0.05
FCC2 1.00 (0.99, 1.01) 0.14 1.12 (1.10. 1.13) 0.15
FCC3 1.04 (1.00, 1.01) 0.17 1.11 (1.05, 1.13) 0.13
FCC4 1.05 (0.99, 1.12) 0.12 1.13 (1.09,1.15) 0.11
Freeway NO X 1.05 (0.99, 1.13) 0.04 1.10 (0.98, 1.18) 0.06
Non-Freeway NO
X
1.02 (1.00, 1.03) 0.05 1.15 (0.99, 1.17) 0.07
Total NO
X
1.10 (1.09, 1.19) 0.01 1.09 (0.98, 1.16) 0.04
Table 3 presents the unadjusted and adjusted hazard ratio over the follow up among
children who were not obese at the study entry in the two combined cohorts. Based on
the interaction test for ambient air pollution, the interaction between ambient
pollutants and the cohort only exists in an eight-hour ozone level and PM10. But this
interaction was not found in the other pollutants, which means that the association
between other pollutants and obesity were not different in two cohorts except for
ozone and PM10 (Table 5). Additionally, the association with ozone levels becomes
more significant in the combined cohort than in each separate cohort. This is probably
because the increased sample size enhanced the power. Therefore, the ambient air
pollution results were presented with the two cohorts combined.
As the results showed, O 3 (HR=1.58, 95%CI: 1.16, 2.17) increased the childhood
obesity incidence, and the association was robust after adjustment (HR=1.49, 95%CI:
1.08, 2.06). Especially, after adjusting all covariates and the random effect of regional
17
difference, the Ozone level 10am-6pm (HR=1.49, 95%CI: 1.08, 2.06), NO2 level
(HR=1.64, 95%CI: 1.06, 2.52), and NO level (HR=1.43, 95%CI: 1.04, 1.85) were
significantly associated with childhood obesity incidence in our cohorts. PM2.5
showed a positive but not significant association (p>0.1).
Table 3 Regression association between ambient-related air pollution and the
risk of developing obesity among children from age 8 to 18 in cohort C and D.
Unadjusted
1
Adjusted
2
HR (95%CI) p HR (95%CI) p
O 3
3
(n=17887) 1.58 (1.16, 2.17) 0.004 1.49 (1.08, 2.06) 0.015
NO
2
(n=17887) 1.51 (0.98, 2.33) 0.06 1.64 (1.06, 2.52) 0.03
NO (n=14353) 1.24 (0.81, 1,43) 0.06 1.43 (1.04, 1.85) 0.04
PM 10 (n=17347) 0.76 (0.56, 1.05) 0.09 0.76 (0.56, 1.05) 0.10
PM
2.5
(n=16548) 1.58 (0.93, 2.70) 0.09 1.56 (0.90, 2.72) 0.11
*Results were showed as hazard ratio per standard deviation increase in corresponding risk
factors
*Both unadjusted and adjusted analysis were strata by sex and cohort, and randomized by
community
*Since different pollutants have different missing data, so the sample size for each analysis
were showed in table
1 Both unadjusted and adjusted analysis were strata by sex, randomized by cohort, and add
community as fixed effect.
2 Adjusted by baseline age, ethnicity, baseline overweight, annual family incomes, parental
education, physical activity, insurance, second hands smoking, wheeze, and asthma as fixed
effects.
3 O 3 level here was represented by average ozone level during 10am to 6pm.
4 Since different pollutants have different missing data, so the sample size for each analysis
differs: O 3 (n=17887), NO 2 (n=17887), NO (n=14353), PM 10 (n=17347), PM 2.5 (n=16548).
Results for the traffic air pollution exposures are shown in Table 4. Since the test of
interaction shows there is no interaction between traffic pollutants and cohort (Table
5), which means the effects of traffic air pollution were not different by cohort;
therefore, the analysis results of two combined cohorts were presented in this paper.
In combined cohort C+D, there was no evidence of an association between childhood
18
obesity incidence and near-roadway air pollution exposures including FCC2
(Highway air pollution), FCC3 (Main-road air pollution), FCC4 (Local air pollution),
Freeway NOX, Non-Freeway NOX, and Total NOX, which were not found to have a
significant association with obesity incidence. The proportional hazard ratios were all
less than 1 and not significant for both crude and adjusted hazard ratios. After
adjustment by all covariates, only the Freeway NOX (HR=1.02; 95%CI: 0.89, 1.18)
has positive effects on childhood obesity incidence among all near-roadway air
pollution exposures.
As a sensitivity analysis, the model was stratified by both sex and baseline
over-weight, which would be expected as another potential predictor on developing
obesity. The results were consistent with the results of the model, stratified only by
sex, with increased risks of obesity incidence associated with increased exposed level
of traffic air-pollution. This association was not as statistically significant as the
previous one.
Table 4 Incidence analysis of childhood obesity incidence and traffic-related
air-pollution
Unadjusted Adjusted**
HR (95%CI) P HR (95%CI) p
FCC2 0.95 (0.84, 1.09) 0.47 0.96 (0.81, 1.14) 0.64
FCC3 1.05 (0.93, 1.18) 0.41 0.96 (0.84, 1.09) 0.51
FCC4 0.86 (0.73, 1.00) 0.15 0.72 (0.60, 0.87) 0.16
Freeway NO X 0.97 (0.84, 1.11) 0.62 1.02 (0.89, 1.18) 0.75
Non-Freeway NO X 0.99 (0.87, 1.13) 0.91 0.89 (0.76, 1.03) 0.12
Total NO X 0.97 (0.84, 1.11) 0.65 0.97 (0.83, 1.13) 0.70
*Results were showed as hazard ratio per standard deviation increase in corresponding risk
factors.
*Both unadjussted and adjusted analysis were strata by sex, randomized by cohort, and add
community as fixed effect.
**Adjusted by baseline age, ethnicity, baseline overweight, annual family incomes, parental
education, physical activity, insurance, second hands smoking, wheeze, and asthma.
19
Table 5 Interaction between cohort and air pollution exposures on the association
with childhood obesity
p-value p-value
FCC2 0.68 O3 10am-6pm 0.05
FCC3 0.51 NO2 24hr avg 0.42
FCC4 0.53 NO 24hr avg 0.66
Freeway NO
X
0.38 PM10 24hr avg 0.05
Non-Freeway NO X
(FCC2+FCC3+FCC4)
0.70 PM2.5 FRM mass 0.29
Local Traffic (Total NO
X
) 0.41
Discussion
We hypothesized that ambient-related air pollution and traffic-related air pollution
would positively associate with prevalence and incidence of childhood obesity. In
these two cohorts of children from 12 communities across Southern California,
ambient air pollutants including O3 and NO 2, and PM 2.5, and total traffic NOX
pollutants had significant influence on obesity prevalence at the study entry and
continued robust association after adjustments. For obesity incidence during the
follow up, traffic air pollution did not show a significant effect on obesity incidence,
while O3, NO2, and NO increased obesity occurrence.
A particular strength of this study is that it is a large cohort study which was followed
up for 9 years with two cohorts. A total of 4,550 children were recruited, and 2,745
were left in this analysis. The information on yearly BMI was collected by CHS for
each participants, and the exposure to ambient particles and NRP information during
the follow up periods were also collected and calculated, provide an opportunity for
incidence analysis separate from previous studies.
20
This is one of the first studies to examine the association between childhood obesity
incidence and air pollution. While certain factors concurred with obesity research
conducted in the cohort E of Children’s Hospital Study
[11]
and a study of ambient air
pollution and the childhood obesity prevalence in The Seven Northeastern Cities
Study
[22]
, the majority of the findings were different from the findings in CHS cohort
E.
Consistent with the study of ambient air pollution and obesity, we found there is a
significant positive association between childhood obesity prevalence and ambient air
pollutants, including O3 (OR=1.20, 95%CI: 1.18, 1.22), NO2 (OR=1.22, 95%CI: 1.20,
1.23), and PM 2.5 (OR=1.21 95%CI: 1.13, 1.25); and also, there is a positive
association between NO, PM10 and obesity incidence but with a p value larger than
0.1. The Seven Northeastern Cities Study found after adjusting for confounding
factors, an increased prevalence of obesity was also associated with an increase in O 3
(OR=1.14, 95% CI: 1.04-1.24), PM 10 (OR=1.19, 95% CI: 1.11-1.26), and NO 2
(OR=1.13, 95% CI: 1.04-1.22). Additionally, our study also explored the relationship
between obesity and ambient air pollutions. It is showed O3 (HR=1.49, 95%CI: 1.08,
2.06), NO2 level (HR=1.64, 95%CI: 1.06, 2.52), and NO level (HR=1.43, 95%CI:
1.04, 1.85) were significantly associated with obesity incidence in our cohorts.
Compared with the study which examined BMI growth and traffic air pollution in
cohort E
[11]
, the findings here are different from the previous ones. In cohort E,
Non-freeway NOX levels were significantly and positively associated with BMI at age
10, while the freeway-related exposures were not associated with BMI growth. The
association between BMI and non-freeway NOX was reduced but remained significant
[11]
. But in our study, traffic air pollution didn’t show a significant association with the
incidence of childhood obesity. Only one negative and significant association showed
21
up in FCC4 (OR=0.69, 95%CI: 0.54, 0.89) (Local street traffic). And the association
is robust after adjustment (OR=0.60, 95%CI: 0.47, 0.76). For all the other traffic
pollutants, we do not find associations with obesity incidence in these two cohorts.
This difference in findings from the two cohorts may have resulted from mobility
differences by age and outcome variable. In the present study, most of the children
were 8-18 years old during the follow up, while most children in cohort E were less
than 10 years old for most of the follow up. These children of different ages were
different in many aspects, such as physical activity, food intake, body development,
metabolic mechanism, etc. On the outcome data, we used obesity incidence, which
has not been used in any previous published paper; instead, the previous paper use
BMI growth for statistical analysis. Also, several other studies have demonstrated
significant correlations between BMI growth or obesity prevalence and traffic air
pollution, and the significant association between obesity prevalence and ambient air
pollution.
The confounding variables in this study were first selected based on previous studies
and common sense. Then they were evaluated with the confounder test. Those with
more than a 10% change in beta were kept. But there may still be a potential for
confounding by other factors, such as smoking history. Since most children in this
study were followed up until 18, some of those teenagers might have been smoking
during these periods. However, we only considered second-hand smoking in this
study but the self-smoking history is a more important risk factor.
Socioeconomic status in the home and neighborhood has been controlled in the model
22
as a potential confounder, because children living in areas with high air pollution
concentrations could come from families of lower socioeconomic status (as indicated
by family income and parental highest education level as two fixed effects in the
present study)
[14]
. After adjusted by all covariates, in both traffic air-pollution and
ambient air-pollution, the higher the family income level, the lower the probability of
obesity in their children, which is the same as we expected based on previous
knowledge about association between lower socioeconomic position and higher
traffic-related pollution exposures in California
[14]
. However, the parents’ highest
education level shows flipped effects on obesity with an increased hazards ratio in
higher education level group. The logistic regression performed at study entry shows
children in higher education level families have significantly lower chances for
obesity. The possible reason for the flipped effects might be due to the significant
correlation between family income level and parental educational level (p<0.0001),
which leads to the interaction in the cox regression model when adjusted for all
covariates.
Inactive children (outdoor sports less than one day per week) have the highest rates of
developing obesity. This is in accord with previous reports that overweight and obese
children typically participate in less frequent and shorter periods of physical activity
than normal weight children
[23].
Another possible explanation is that parents tried to
restrict air pollution exposure to their children, since the children who lack of physical
activity might be from families residing in heavily polluted areas. Besides, the lowest
risk of obesity is among the children with moderate activities (one day of outdoor
sports weekly), rather than more active children who did outdoor sports more than one
day per week. It is possible that highly exposed to outdoor activities induced the
metabolic mechanism that leading to the obesity.
23
In addition to outcome limitation, other limitations of this study are also considered.
Another limitation of this study is the lack of information on food intake. No diet
information was controlled in the study, but dietary factors are directly and strongly
related to obesity thus should be fully evaluated. Moreover, prenatal air pollution
exposure can also be a significant risk factor for later childhood obesity. In utero
exposure to combustion products such as cigarette smoking and near-roadway
pollution (NRP), contribute to increased BMI and childhood obesity in children
[15] [16]
.
Large epidemiological surveys have also indicated that maternal prenatal smoking is
associated with increased occurrence of overweight and obesity among children and
early adolescents
[17-19]
. There are reported results indicating that active maternal
smoking during pregnancy is a key factor predicting overweight in childhood
[17]
.
Effect modifications between main air pollution exposures and covariates mentioned
in the paper were checked. No significant interaction was found. However, further
exploration still needs to be done by checking the effect modification between
air-pollution and other genetic and/or environmental exposure variables, including
diet, Native American genetics admixture among Hispanic Whites, and genetic risk
score. Because diet information is only restricted for later years of the study, we can’t
evaluate the diet in these two cohorts. Genetic risk score is still under consideration in
constructing the present study, so we can look at the genetic admixture for effect
modification in future research.
Conclusion
Although this paper provides no evidence that traffic-related air pollution is
24
significantly associated with the development of obesity in children, we do find a
strong influence on childhood obesity by ambient air pollutants including O3, NO2, NO,
and PM2.5. It’s also meaningful since it points to a clearer direction for future work.
Future research is necessary to refine the hypothesis, i.e. assessing the role of air
pollution in the development of obesity in children. We could continue the
longitudinal study by exploring the mixed effects among ambient air pollutants and
traffic air pollutants. Moreover, we could also continue exploring the cross-sectional
associations between traffic air pollution and BMIZ and/or obesity status at baseline
visit, air pollution association with BMI growth curve during the study follow-up, and
exploring the interaction between air pollution with genetic admixture in the
cross-sectional and longitudinal associations of air pollution exposures with obesity
outcomes.
The role of air pollution in the development of obesity in children is important for
guiding possible interventions. If a robust association were found exist, we can sure to
take measures controlling traffic which would then reduce obesity formation in
children.
Reference
[1] Caballero B. The global epidemic of obesity: an overview. Epidemiol Rev 2007;
29: 1–5.
[2] Ogden, Cynthia L., Margaret D. Carroll, Brian K. Kit, and Katherine M. Flegal.
"Prevalence of childhood and adult obesity in the United States, 2011-2012." Jama
311, no. 8 (2014): 806-814.
[3] Dietz WH, Robinson T. Clinical practice. Overweight children and adolescents. N
25
Engl J Med. 2005; 352:2100–2109.
[4] McAllister EJ, Dhurandhar NV , Keith SW, et al. Ten putative contributors to the
obesity epidemic. Crit Rev Food Sci Nutr 2009; 49:868–913.
[5] Keith SW, Redden DT, Katzmarzyk PT, et al. Putative contributors to the secular
increase in obesity: exploring the roads less traveled. Int J Obes (Lond) 2006; 30:
1585–1594
[6] Brook RD, Jerrett M, Brook JR, Bard RL, Finkelstein MM: The relationship
between diabetes mellitus and traffic-related air pollution. J Occup Env Med 2008,
50:32–38.
[7] Ino T, Shibuya T, Saito K, Inaba Y . Relationship between body mass index of
offspring and maternal smoking during pregnancy. Int J Obes (Lond) 2012; 36:
554–558.
[8] Bolton JL, Smith SH, Huff NC, et al. Prenatal air pollution exposure induces
neuroinflammation and predisposes offspring to weight gain in adulthood in a
sex-specific manner. FASEB J 2012; 26:4743–4754
[9] Xu, Xiaohua, Zubin Yavar, Matt Verdin, Zhekang Ying, Georgeta Mihai, Thomas
Kampfrath, Aixia Wang et al. "Effect of early particulate air pollution exposure on
obesity in mice role of p47phox." Arteriosclerosis, thrombosis, and vascular biology
30, no. 12 (2010): 2518-2527.
[10] Sun Q, Yue P, Deiuliis JA, Lumeng CN, Kampfrath T, Mikolaj MB, Cai Y ,
Ostrowski MC, Lu B, Parthasarathy S, Brook RD, Moffatt-Bruce SD, Chen LC,
Rajagopalan S: Ambient air pollution exaggerates adipose inflammation and insulin
resistance in a mouse model of diet-induced obesity. Circulation 2009, 119:538–546.
[11] Jerrett, Michael, Rob McConnell, Jennifer Wolch, Roger Chang, Claudia Lam,
26
Genevieve Dunton, Frank Gilliland, Fred Lurmann, Talat Islam, and Kiros Berhane.
"Traffic-related air pollution and obesity formation in children: a longitudinal,
multilevel analysis." Environ Health 13, no. 1 (2014): 49.
[12] McConnell, Rob, Ernest Shen, Frank D. Gilliland, Michael Jerrett, Jennifer
Wolch, Chih-Chieh Chang, Frederick Lurmann, and Kiros Berhane. "A longitudinal
cohort study of body mass index and childhood exposure to secondhand tobacco
smoke and air pollution: the Southern California Children’s Health Study."
Environmental health perspectives 123, no. 4 (2015): 360.
[13] Jerrett M, McConnell R, Chang CCR, Wolch J, Reynolds K, Lurmann F,
Gilliland F, Berhane K: Automobile traffic around the home and attained body mass
index: a longitudinal cohort study of children aged 10–18 years. Prev Med 2010,
50(Suppl 1):S50–S58.
[14] Krebs NF, Jacobson MS; American Academy of Pediatrics Committee on
Nutrition. Prevention of pediatric overweight and obesity. Pediatrics 2003;
112:424–430.
[15] Ino, Toshihiro. "Maternal smoking during pregnancy and offspring obesity:
Meta ‐analysis." Pediatrics International 52, no. 1 (2010): 94-99.
[16] Rundle A, Hoepner L, Hassoun A, Oberfield S, Freyer G, Holmes D, et al. 2012.
Association of childhood obesity with maternal exposure to ambient air polycyclic
aromatic hydrocarbons during pregnancy. Am J Epidemiol 175(11):1163–1172.
[17] Al Mamun A, Lawlor DA, Alati R, O’Callaghan MJ, Williams GM, Najman JM.
Does maternal smoking during pregnancy have a direct effect on future offspring
obesity? Evidence from a prospective birth cohort study. Am J Epidemiol 2006;
164:317–325.
27
[18] Mendez MA, Torrent M, Ferrer C, Ribas-Fit o N, Sunyer J. Maternal smoking
very early in pregnancy is related to child overweight at age 5-7 y. Am J Clin Nutr
2008;87:1906–19013.
[19] Raum E, Ku ¨pper-Nybelen J, Lamerz A, Hebebrand J, Herpertz-Dahlmann B,
Brenner H. Tobacco smoke exposure before, during, and after pregnancy and risk of
overweight at age 6. Obesity (Silver Spring) 2011;19:2411–2417.
[20] Gauderman WJ, V ora H, McConnell R, Berhane K, Gilliland F, Thomas D,
Lurmann F, Avol E, Kunzli N, Jerrett M, Peters J. Effect of exposure to traffic on lung
development from 10 to 18 years of age: a cohort study. Lancet. 2007; 369:571–577.
[21] McConnell R, Berhane K, Yao L, Jerrett M, Lurmann F, Gilliland F, Kunzli N,
Gauderman J, Avol E, Thomas D, Peters J. Traffic, susceptibility, and childhood
asthma. Environ Health Perspect. 2006; 114:766–772
[22] Dong, Guang ‐Hui, Zhengmin Min Qian, Miao ‐Miao Liu, Da Wang, Wan ‐Hui Ren,
Louise H. Flick, John Fu et al. "Ambient air pollution and the prevalence of obesity in
chinese children: The seven northeastern cities study." Obesity 22, no. 3 (2014):
795-800.
[23] De Bourdeaudhuij I, Lefevre J, Deforche B, Wijndaele K, Matton L, Philippaerts
R. Physical activity and psychosocial correlates in normal weight and overweight 11 to
19 year olds. Obes Res 2005; 13:1097–1105.
Abstract (if available)
Abstract
Background: Air pollution, especially combustion products, could activate metabolic disorders through inflammatory pathways, potentially leading to obesity. The effect of air-pollution on BMI growth was shown by a previous study. Here we sought to determine whether ambient air pollution and traffic air pollution are associated with obesity incidence in children from age 8–18 years old. ❧ Methods: 3,887 children aged 8-12 were selected from two prospective cohorts in the Children’s Hospital Study (CHS) and followed for 9 years, with height and weight measured annually. A child's weight status was determined using an age- and sex-specific percentile for BMI based on the CDC’s chart. Average annual ambient exposure was directly collected and measured at monitoring stations in 12 selected communities. Dispersion models were used to estimate exposure to traffic-related air pollution. Logistic regression was performed to analyze the association between ambient-related and traffic-related air pollution and obesity prevalence at the study entry
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Native American ancestry among Hispanic Whites is associated with higher risk of childhood obesity: a longitudinal analysis of Children’s Health Study data
PDF
Air pollution and childhood obesity
PDF
Street connectivity and childhood obesity: a longitudinal, multilevel analysis
PDF
Assessment of the mortality burden associated with ambient air pollution in rural and urban areas of India
PDF
Obesity paradox in acute heart failure decompensation
PDF
Associations between ambient air pollution and hypertensive disorders of pregnancy
PDF
Association of traffic-related air pollution and lens opacities in the Los Angeles Latino Eye Study
PDF
Disparities in exposure to traffic-related pollution sources by self-identified and ancestral Hispanic descent in participants of the USC Children’s Health Study
PDF
Air pollution, smoking, and multigenerational DNA methylation Signatures: a study of two southern California cohorts
PDF
Air pollution and breast cancer survival in California teachers: using address histories and individual-level data
PDF
Association of maternal and environmental factors with infant feeding behaviors in a birth cohort study
PDF
Long-term effects on vertebral body height growth of dose sculpting intensity modulated radiation therapy for children with neuroblastoma
PDF
A pilot study of a global approach to assessing air pollution exposure in port communities: passive air monitoring of nitrogen dioxide concentrations
PDF
Associations between isoflavone soy protein (ISP) supplementation and breast cancer in postmenopausal women in the Women’s Isoflavone Soy Health (WISH) clinical trial
PDF
Predictive factors of breast cancer survival: a population-based study
PDF
Examining exposure to extreme heat and air pollution and its effects on all-cause, cardiovascular, and respiratory mortality in California: effect modification by the social deprivation index
PDF
A novel risk-based treatment strategy evaluated in pediatric head and neck non-rhabdomyosarcoma soft tissue sarcomas (NRSTS) patients: a survival analysis from the Children's Oncology Group study...
PDF
Psychometric study of an English version of Perceived Stress Scale in minority adolescents
PDF
The associations between ultra-processed food consumption and type 2 diabetes and obesity among young adults
PDF
Spatial analysis of PM₂.₅ air pollution in association with hospital admissions in California
Asset Metadata
Creator
Zhou, Xiaoyi
(author)
Core Title
A cohort study of air-pollution and childhood obesity incidence
School
Keck School of Medicine
Degree
Master of Science
Degree Program
Applied Biostatistics and Epidemiology
Publication Date
07/08/2016
Defense Date
04/08/2016
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
Air pollution,ambient,California,childhood obesity,OAI-PMH Harvest,Traffic
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Gilliland, Frank (
committee chair
), Azen, Stanley (
committee member
), Barkauskas, Donald (
committee member
)
Creator Email
xiaoyizh@usc.edu,zhouxysherry@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c40-266377
Unique identifier
UC11281550
Identifier
etd-ZhouXiaoyi-4527.pdf (filename),usctheses-c40-266377 (legacy record id)
Legacy Identifier
etd-ZhouXiaoyi-4527.pdf
Dmrecord
266377
Document Type
Thesis
Format
application/pdf (imt)
Rights
Zhou, Xiaoyi
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
ambient
childhood obesity