Applied nutritional investigation

Parental tobacco use is associated with increased risk of child

malnutrition in Bangladesh
Cora M. Best, M.H.S.
a,c
, Kai Sun, Ph.D.
a,b
, Saskia de Pee, Ph.D.
c
,
Martin W. Bloem, M.D.
a,c
, Gudrun Stallkamp, M.Sc.
d
, and Richard D. Semba, M.D., M.P.H.
a,b,
*
a
Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA
b
Department of Ophthalmology, Johns Hopkins School of Medicine, Baltimore, Maryland, USA
c
World Food Program, Rome, Italy
d
Concern Worldwide, Dublin, Ireland
Abstract Objectives: We investigated the relation between parental tobacco use and malnutrition in children
5 y of age and compared expenditures on foods in households with and without tobacco use.
Methods: Tobacco use, child anthropometry, and other factors were examined in a stratied,
multistage cluster sample of 77 678 households from the Bangladesh Nutrition Surveillance Project
(20052006). Main outcome measurements were stunting, underweight, and wasting, and severe
stunting, severe underweight, and severe wasting. Secondary outcomes included the proportion of
household expenditures spent on food.
Results: The prevalence of parental tobacco use was 69.9%. Using the new World Health
Organization child growth standards, prevalences of stunting, underweight, and wasting were
46.0%, 37.6%, and 12.3%, respectively. After adjusting for potential confounders, parental tobacco
use was associated with an increased risk of stunting (odds ratio [OR] 1.17, 95% condence interval
[CI] 1.121.21, P 0.0001), underweight (OR 1.17, 95% CI 1.121.22, P 0.0001), and wasting
(OR 1.10, 95% CI 1.031.17, P 0.004), and severe stunting (OR 1.16, 95% CI 1.101.23, P
0.0001), severe underweight (OR 1.21, 95% CI 1.131.30, P 0.0001), and severe wasting (OR
1.14, 95% CI 0.981.32, P 0.09). Households with tobacco use spent proportionately less per
capita on food items and other necessities.
Conclusions: In Bangladesh parental tobacco use may exacerbate child malnutrition and divert
household funds away from food and other necessities. Further studies with a stronger analytic
approach are needed. These results suggest that tobacco control should be part of public health
strategies aimed at decreasing child malnutrition in developing countries. 2007 Elsevier Inc. All
rights reserved.
Keywords: Bangladesh; Malnutrition; Poverty; Smoking; Tobacco
Introduction
Cigarette smoking causes 5 million deaths worldwide
annually, and it is estimated that the annual death toll from
smoking will climb to 10 million deaths by 2030, with 7
million deaths in developing countries [1,2]. Cigarette
smoke damages the lower respiratory tract, increases oxi-
dative stress, and increases the risk of bronchitis, chronic
obstructive lung disease, cancer, and death [1]. As tobacco
control legislation in developed countries has exerted pres-
sure on tobacco companies, these companies have gradually
shifted their market from high-income to low-income coun-
tries, where many people are poorly informed about the
health risks of tobacco use and antismoking policy is rela-
tively weak [2]. Although much research has been focused
on the relation between smoking and adverse outcomes such
as cancer, respiratory disease, and cardiovascular disease,
the problem of smoking and its relation to malnutrition and
poverty has not been well characterized [2]. Tobacco use
may have adverse consequences for nutrition, health, and
This work was supported in part by a Lew R. Wasserman Merit Award
for Research to Prevent Blindness to Dr. Semba and by the Eye Foundation
of America.
* Corresponding author. Tel: 410-955-3572; fax: 410-955-0629.
E-mail address: rdsemba@jhmi.edu (R. D. Semba).
Nutrition 23 (2007) 731738
www.elsevier.com/locate/nut
0899-9007/07/$ see front matter 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.nut.2007.06.014
household budgets, especially among families living in pov-
erty in developing countries.
The amount of money spent on tobacco is especially
problematic in low-income countries [3,4]. Bangladesh is
one of the poorest countries in the world. Nearly half of the
population lives below the poverty line and consumes
2122 cal/d [2]. One investigation suggested that tobacco
expenditures exacerbate the effects of poverty and divert
household income away from food, clothing, housing,
health, and education [3]. It is estimated that the average
male cigarette smoker in Bangladesh spends more than
twice as much on cigarettes than the per capita expenditure
on clothing, housing, health, and education combined [3].
As in many countries, prevalence of tobacco use is highest
in the poorest of households, the same households that are
most likely to have malnourished children [2]. Efroymson et
al. [3] estimated that if a portion of the money spent on
tobacco by poor households in Bangladesh could be redi-
rected toward food purchases, this would provide enough
calories to prevent nearly 130 000 deaths from malnutrition
each year in children 5 y of age.
Previous estimates of the current prevalences of tobacco
use in Bangladesh are 48.3% in adult men and 20.9% in
adult women. The highest estimate is in men 3549 y of age
at 73% [3,5]. Since the 1980s, Bangladesh has had a grow-
ing negative trade balance in tobacco and tobacco products,
yet between 1993 and 1996 annual per-capita cigarette con-
sumption increased 33% [2]. Although it is difcult to
estimate the health costs associated with tobacco use in
Bangladesh, the Bangladesh Cancer Society estimates that
half of the annual cancer mortality of 75 000 people is due
to tobacco [2].
Although smoking appears to exacerbate poverty in de-
veloping countries, it is not well known whether smoking
contributes to malnutrition among children. We hypothe-
sized that among families in Bangladesh, 1) children are at
higher risk for malnutrition in households where a parent
uses tobacco and 2) household income spent on tobacco is
associated with lower expenditures on food. To examine
these hypotheses, we characterized tobacco use and child
malnutrition among families in Bangladesh.
Materials and methods
The study consisted of 77 678 households that partici-
pated in the Nutritional Surveillance Project (NSP) of Ban-
gladesh in 20052006. The NSP has been conducted by
Helen Keller International (HKI) and the Institute of Public
Health Nutrition of the Government of Bangladesh since
1989 [6]. The surveillance design is guided by the concep-
tual framework of the United Nations Childrens Fund on
the causes of malnutrition [7] and collects data on indicators
of health, nutrition, socioeconomic status, food production
and consumption, and health services [8]. The NSP used a
stratied multistage cluster sample of households in rural
areas and cities of the six major divisions of the Bangladesh:
Barisal, Chittagong, Dhaka, Khulni, Rajshahi, and Sylhet
[6]. Data were collected on a structured coded questionnaire
by two-person eld teams from partner non-governmental
organizations trained by the HKI. The questionnaire was
used to record data on children 059 mo of age, including
anthropometry, date of birth, and sex. The mother of the
child or other adult member of the household was asked to
provide information on the households composition, paren-
tal education, and household expenditures, along with other
socioeconomic, environmental, sanitation, and health indi-
cators.
The eld teams measured and recorded the weight of
each child 059 mo of age to a precision of 0.1 kg and
length/height to a precision of 0.1 cm. Birth dates were
estimated using a calendar of local and national events and
converted to the Gregorian calendar. Z scores of height for
age (stunting), weight for age (underweight), and weight for
height (wasting) were calculated using the new World
Health Organization (WHO) child growth standards [9].
The new WHO child growth standards have not yet been
widely applied, and for comparison purposes conventional
Z scores were also calculated using EpiInfo software (Cen-
ters for Disease Control and Prevention, Atlanta, GA,
USA), which uses the reference population of the U.S.
National Center for Health Statistics (NCHS). Children with
Z scores 2 SD for weight for height, weight for age, or
height for age were considered wasted, underweight, or
stunted, respectively, and Z scores 3 SD for weight for
height, weight for age, or height for age were considered
severely wasted, severely underweight, or severely stunted
[10].
In 20052006, the NSP included questions on paternal
and maternal tobacco use and weekly expenditures on to-
bacco products. In each household, data were gathered re-
garding expenditures the previous week on rice, wheat, dal,
eggs, sh, fruits, vegetables, milk, meat, poultry, sweet
biscuits, cooking oil, sugar, snacks, spices, and other foods.
Because few homes purchased signicant amounts of staple
foods (rice and wheat), the amount of rice produced and/or
received from relatives, friends, and aid programs was re-
ported in kilograms and assigned a monetary value deter-
mined by the daily market price of rice at the time of the
survey. This monetary value was included when calculating
the total monthly household expenditure per capita. The
previous monthly expenditure on housing, education, med-
ical care, agricultural inputs, livestock purchases, electric-
ity, fuel, loan payment, taxes, and other household items
was also recorded. Expenditure and price variables were
collected using the Bangladeshi taka.
The study protocol complied with the principles enunci-
ated in the Helsinki Declaration [11]. The eld teams were
instructed to explain the purpose of the NSP and data
collection to each childs mother or caretaker and, if
present, the father and/or household head; data collection
proceeded only after written informed consent was ob-
732 C. M. Best et al. / Nutrition 23 (2007) 731738
tained. Participation was voluntary and all subjects were
free to withdraw at any stage of the interview. The NSP was
approved by the ethical review committee of the Bang-
ladesh Medical Research Council. The plan for secondary
data analysis was approved by the institutional review board
of the Johns Hopkins University School of Medicine.
In analyses where child malnutrition was the outcome
and there was more than one child in the household, the
youngest child in the household was used as the index of
child malnutrition for that particular household (i.e., house-
holds were not counted more than once). The status of
parental tobacco use was determined by whether the mother
Table 1
Characteristics of children and households where a parent uses tobacco or does not use tobacco among poor families in Bangladesh (n 77 678)
Characteristics n Tobacco use n No tobacco use P
Maternal age (y)
22 15 308 27.8 9316 39.2 0.0001
2326 14 325 26.0 7030 29.6
2730 10 844 19.7 4043 17.0
31 14 612 26.5 3379 14.2
Maternal education (y)
0 24 856 45.4 5723 24.2 0.0001
13 3550 6.5 1235 5.2
46 14 589 26.6 7030 29.7
79 9255 16.9 6645 28.1
10 2513 4.6 3038 12.8
Paternal education (y)
0 25 404 47.6 6259 27.7 0.0001
13 2893 5.4 953 4.2
46 11 629 21.8 5485 24.3
79 8328 15.6 4878 21.6
10 5121 9.6 4983 22.1
Child age (mo)
05 4941 9.8 2248 10.4 0.0001
611 5613 11.2 2596 12.0
1223 12 004 23.8 5422 25.0
2435 11 185 22.2 4748 21.9
3647 9487 18.8 3876 17.8
4859 7060 14.0 2800 12.9
Child malnutrition, conventional NCHS standard
Height-for-age Z score 2 21 388 39.3 7734 32.8 0.0001
Height-for-age Z score 3 6244 11.4 2011 8.5 0.0001
Weight-for-age Z score 2 25 343 46.0 9430 39.7 0.0001
Weight-for-age Z score 3 5797 10.5 1869 7.9 0.0001
Weight-for-height Z score 2 6418 11.8 2466 10.5 0.0001
Weight-for-height Z score 3 370 0.7 145 0.6 0.42
Child malnutrition, new WHO child growth standard*
Height-for-age Z score 2 26 418 48.1 9800 41.3 0.0001
Height-for-age Z score 3 9250 16.8 3097 13.0 0.0001
Weight-for-age Z score 2 21 783 39.5 7901 33.3 0.0001
Weight-for-age Z score 3 5961 10.8 1909 8.0 0.0001
Weight-for-height Z score 2 7043 12.8 2622 11.0 0.0001
Weight-for-height Z score 3 1166 2.1 402 1.7 0.001
No. of individuals eating from same cooking pot
24 21 900 39.7 10 253 43.1 0.0001
4 33 195 60.3 13 516 56.9
Geographic location
Rural 47 903 86.9 21 155 89.0 0.0001
Urban 7192 13.1 2614 11.0
Monthly household expenditure per capita (taka) 56 246 772 22 432 886 0.0001
Monthly household expenditure per capita (taka) in quintiles
Quintile 1 9890 17.9 3754 15.8 0.0001
Quintile 2 11 166 20.3 4045 17.0
Quintile 3 11 291 20.5 4449 18.7
Quintile 4 11 528 20.9 5086 21.4
Quintile 5 11 220 20.4 6435 27.1
Monthly household food expenditure per capita (taka) 56 246 358 22 432 405 0.0001
NCHS, U.S. National Center for Health Statistics; WHO, World Health Organization
* Stunting, underweight, and wasting dened as a Z score 2 SD for height for age, weight for age, and weight for height, respectively [9].
733 C. M. Best et al. / Nutrition 23 (2007) 731738
or the father in the household currently used tobacco (cig-
arette, bidi, hukah, or chewing tobacco). Child age was
categorized as 011, 1223, 2435, 3547, and 4859 mo.
Maternal and paternal education levels were categorized as
0, 13 y (rst half of primary), 46 y (second half of
primary), 79 y (junior high), and 10 y (at least high
school).
Univariate and multivariate logistic regression models
were used to examine the relation between parental tobacco
use and the risk of stunting, underweight, and wasting. The
multivariate regression models included risk factors se-
lected on the basis of their known association with child
malnutrition in developing countries. Because information
on household income was not available, quintiles of total
monthly household expenditures per capita were included in
all multivariate models to serve as a proxy for socioeco-
nomic status. HKIs assigned sampling design weights were
used to adjust for population size, and all results are
weighted. The analyses were conducted using SAS Survey
(SAS Institute, Cary, NC, USA) [12]. P 0.05 was con-
sidered statistically signicant.
Results
The analysis included 77 678 households that were sur-
veyed between January 1, 2005 and January 31, 2006. The
overall prevalence of parental tobacco use was 69.9%. The
prevalences of paternal and maternal tobacco uses were
68.2% and 20.1%, respectively. Of all the households, the
father and mother used tobacco in 17.6%, only the father
used tobacco in 50.8%, only the mother used tobacco in
2.5%, and neither parent used tobacco in 29.1%.
The characteristics of households in which the father
and/or mother used tobacco were compared with house-
holds in which neither parent used tobacco (Table 1). In
households with tobacco use, the levels of paternal and
maternal education were lower and maternal age was older.
When using the new WHO child growth standards and the
conventional NCHS reference population, prevalences of
stunting, underweight, and wasting, and severe stunting and
severe underweight were signicantly higher in households
with tobacco use compared with no tobacco use. The prev-
alence of severe wasting was signicantly higher in house-
holds with tobacco use compared with no tobacco use using
the new WHO standards but not with the NCHS reference
population. There was a larger proportion of more than four
people eating from the same cooking pot in households with
parental tobacco use. The mean total monthly household
expenditure per capita and mean monthly household expen-
diture on food were lower in households with tobacco use
than in households with no tobacco use. The relation be-
tween stunting, underweight, wasting, and severe stunting,
severe underweight, and severe wasting and parental to-
bacco use was consistent and signicant in each quintile of
total monthly per capita household expenditure (Table 2).
In households where neither parent used tobacco, the
average proportion of household expenditures devoted to
food purchases was 58.0%. In households where at least one
parent used tobacco, this proportion was 56.6%. In a sub-
sample of 51 655 households where specic data on expen-
ditures on types of food were collected, families with no
tobacco use also devoted a greater proportion of the total
monthly household expenditure to animal foods, plant
foods, education, and medical care (Fig. 1). In contrast,
households with tobacco use spent a greater proportion of
money on clothing, housing, and other household items. In
households with no parental tobacco use, the median ex-
penditure was greater for animal food products (sh, eggs,
meat, poultry, and milk) and plant products (vegetables,
fruits, and dal), education and health care, and other house-
hold resources such as clothing, housing, electricity, fuel,
agricultural inputs, etc. As socioeconomic status increased,
the proportion of the total household expenditure devoted to
tobacco decreased. This proportion was 6.0% in the poorest
Table 2
Proportion of children 5 y of age with malnutrition in households by
parental tobacco use, stratied by quintile of monthly household
expenditure using new World Health Organization child growth standard
Tobacco
use (%)
No tobacco
use (%)
P
Quintile 1
Stunting 53.8 48.2 0.0001
Severe stunting 20.6 18.0 0.004
Underweight 45.8 41.6 0.0001
Severe underweight 14.5 11.4 0.0001
Wasting 16.1 13.4 0.0009
Severe wasting 2.9 2.5 0.22
Quintile 2
Stunting 51.4 45.0 0.0001
Severe stunting 18.7 15.4 0.0001
Underweight 42.5 36.5 0.0001
Severe underweight 11.8 9.9 0.005
Wasting 13.6 12.7 0.24
Severe wasting 2.3 1.5 0.015
Quintile 3
Stunting 49.3 44.9 0.0001
Severe stunting 17.6 14.0 0.0001
Underweight 40.6 35.9 0.0001
Severe underweight 11.5 8.8 0.0001
Wasting 13.2 11.2 0.005
Severe wasting 2.2 1.6 0.076
Quintile 4
Stunting 46.2 40.0 0.0001
Severe stunting 15.2 11.2 0.0001
Underweight 37.4 31.7 0.0001
Severe underweight 9.5 6.8 0.0001
Wasting 11.7 10.0 0.01
Severe wasting 1.8 1.6 0.44
Quintile 5
Stunting 40.2 33.4 0.0001
Severe stunting 12.4 9.4 0.0001
Underweight 32.1 25.7 0.0001
Severe underweight 7.3 5.4 0.0001
Wasting 9.9 9.3 0.28
Severe wasting 1.5 1.5 0.75
734 C. M. Best et al. / Nutrition 23 (2007) 731738
quintile (quintile 1) and decreased linearly to 1.8% in the
wealthiest quintile (quintile 5).
In the following results, the new WHO child growth
standards were applied. The prevalence of stunting was
46.0%. In a univariate model (model 1) and a multivariate
model (model 2) adjusting for child age, child gender,
maternal age, maternal education level, total monthly
household expenditure per capita, and other factors, parental
tobacco use was associated with an increased risk of stunt-
ing (odds ratio [OR] 1.17, 95% condence interval [CI]
1.121.21, P 0.0001; Table 3). The prevalence of under-
weight was 37.6%. In a univariate model (model 1) and a
multivariate model (model 2) adjusting for child age, child
gender, maternal age, maternal education level, total
monthly household expenditure per capita, and other fac-
tors, parental tobacco use was associated with an increased
risk of a child being underweight (OR 1.17, 95% CI 1.12
1.22, P 0.0001). The prevalence of wasting was 12.3%.
In a univariate model (model 1), and a multivariate model
(model 2) adjusting for child age, child gender, maternal
age, maternal education level, total monthly household ex-
penditure per capita, and other factors, parental tobacco use
was associated with an increased risk of wasting (OR 1.10,
95% CI 1.031.17, P 0.004).
Using the conventional NCHS reference population, the
prevalences of stunting, underweight, and wasting were
37.4%, 44.1%, and 11.4%, respectively. When multivariate
models were analyzed, adjusting for the same variables as in
Table 3, paternal tobacco use was associated with an in-
creased risk of stunting (OR 1.17, 95% CI 1.121.23, P
0.0001), underweight (OR 1.15, 95% CI 1.111.20, P
0.0001), and wasting (OR 1.08, 95% CI 1.011.15, P
0.029).
Using the new WHO child growth standards, the preva-
lence of severe stunting was 15.7%. In a univariate model
(model 1) and a multivariate model (model 2) adjusting for
child age, child gender, maternal age, maternal education
level, total monthly household expenditure per capita, and
other factors, parental tobacco use was associated with an
increased risk of severe stunting (OR 1.16, 95% CI 1.10
1.23, P 0.0001; Table 4). The prevalence of severe
underweight was 10.0%. In a univariate model (model 1)
and a multivariate model (model 2) adjusting for child age,
child gender, maternal age, maternal education level, total
monthly household expenditure per capita, and other fac-
tors, parental tobacco use was associated with an increased
risk of a child being severely underweight (OR 1.21, 95%
CI 1.131.30, P 0.0001). The prevalence of severe wast-
ing was 2.0%. In a univariate model (model 1), parental
tobacco use was associated with an increased risk of severe
wasting (OR 1.26, 95% CI 1.101.44, P 0.001). In the
multivariate model (model 2) adjusting for the same factors
Fig. 1. Median monthly household expenditures per capita (taka) in Bangladeshi households with and without parental tobacco use.
735 C. M. Best et al. / Nutrition 23 (2007) 731738
as the previous multivariate models above, parental tobacco
use was marginally associated with severe wasting (OR
1.14, 95% CI 0.981.32, P 0.09).
Using the conventional NCHS reference population, the
prevalences of severe stunting, severe underweight, and
severe wasting were 10.5%, 9.7%, and 0.6%, respectively.
When multivariate models were run, adjusting for the same
variables as in Table 4, paternal tobacco use was associated
with an increased risk of severe stunting (OR 1.18, 95% CI
1.101.26, P 0.0001), severe underweight (OR 1.21,
95% CI 1.121.29, P 0.0001), and severe wasting (OR
1.02, 95% CI 0.791.31, P 0.89).
We also examined alternative multivariate logistic re-
gression models where weekly per-capita household expen-
diture on tobacco instead of paternal tobacco use was eval-
uated as a risk factor. Expenditures on tobacco were divided
into the highest half of expenditures on tobacco and lowest
half of expenditures on tobacco versus no expenditures on
tobacco (reference category). In multivariate models, ad-
justing for child age, child gender, maternal age, maternal
education level, total monthly household expenditure per
capita, and other factors, the highest half of tobacco expen-
ditures (OR 1.18, 95% CI 1.131.24, P 0.0001) and lower
half of tobacco expenditures (OR 1.18, 95% CI 1.121.23,
P 0.0001) were associated with stunting. In multivariate
models, adjusting for child age, child gender, maternal age,
maternal education level, total monthly household expendi-
ture per capita, and other factors, the highest half of tobacco
expenditures (OR 1.21, 95% CI 1.131.29, P 0.0001) and
lower half of tobacco expenditures (OR 1.14, 95% CI 1.07
1.22, P 0.0001) were associated with severe stunting. In
similar multivariate logistic regression models, tobacco ex-
penditures were also signicantly associated with under-
weight, severe underweight, and wasting but not with severe
wasting (data not shown).
Discussion
This analysis reveals that in households in Bangladesh,
parental tobacco use is associated with an increased risk of
stunting, underweight, wasting, and severe malnutrition
(stunting and underweight) in children 059 mo of age.
Similar research on poor urban households in Indonesia also
detected an association between paternal smoking and an
increased risk of child malnutrition [13]. This study found
that paternal smoking in poor urban areas of Indonesia,
where the prevalence of smoking in adult males is 73.8%,
was associated with an increased risk of stunting and severe
wasting in children. In the present study, current parental
tobacco use increased the risk of child malnutrition, and
data were not available on long-term tobacco use and du-
Table 3
Logistic regression models for parental tobacco use and risk of moderate child malnutrition in households in Bangladesh
Characteristics Stunting Underweight Wasting
OR 95% CI P OR 95% CI P OR 95% CI P
Model 1
Parental tobacco use 1.32 1.271.37 0.0001 1.31 1.261.36 0.0001 1.18 1.121.25 0.0001
Model 2
Parental tobacco use 1.17 1.121.21 0.0001 1.17 1.121.22 0.0001 1.10 1.031.17 0.004
Male gender 1.09 1.051.13 0.0001 1.00 0.971.04 0.94 1.21 1.141.27 0.0001
Child age (mo)
05 0.40 0.370.43 0.0001 0.43 0.390.46 0.0001 0.75 0.660.85 0.0001
611 0.45 0.420.48 0.0001 0.53 0.490.57 0.0001 1.04 0.941.17 0.41
1223 1.18 1.111.26 0.0001 0.90 0.850.96 0.0008 1.54 1.401.69 0.0001
2435 1.29 1.221.38 0.0001 1.04 0.981.11 0.22 1.17 1.061.29 0.0013
3647 1.23 1.161.31 0.0001 0.98 0.911.04 0.44 0.98 0.881.08 0.66
4959 1.00 1.00 1.00
Maternal age (y) 0.93 0.980.99 0.0001 0.99 0.980.99 0.0001 1.00 0.991.00 0.33
Maternal education level (y)
0 2.67 2.462.91 0.0001 2.36 2.162.58 0.0001 1.65 1.451.89 0.0001
13 2.58 2.322.87 0.0001 2.22 1.992.47 0.0001 1.60 1.361.88 0.0001
46 2.08 1.912.26 0.0001 1.87 1.712.05 0.0001 1.44 1.261.65 0.0001
79 1.53 1.401.67 0.0001 1.40 1.281.54 0.0001 1.18 1.031.36 0.02
10 1.00 1.00 1.00
4 individuals eating from same cooking pot 1.02 0.981.07 0.27 0.98 0.941.02 0.37 0.96 0.901.02 0.14
Total monthly household expenditures per capita (taka)
Quintile 1 1.58 1.491.68 0.0001 1.64 1.541.74 0.0001 1.49 1.371.63 0.0001
Quintile 2 1.41 1.331.49 0.0001 1.41 1.331.50 0.0001 1.27 1.161.38 0.0001
Quintile 3 1.36 1.291.44 0.0001 1.35 1.271.43 0.0001 1.23 1.141.34 0.0001
Quintile 4 1.20 1.131.27 0.0001 1.20 1.131.27 0.0001 1.09 1.011.20 0.044
Quintile 5 1.00 1.00 1.00
Urban location 0.86 0.820.89 0.0001 0.88 0.830.92 0.0001 0.90 0.840.97 0.005
CI, condence interval; OR, odds ratio
736 C. M. Best et al. / Nutrition 23 (2007) 731738
ration of exposure. Studies have shown that cigarettes and
other forms of tobacco are addicting and that patterns of
tobacco use are regular and compulsive [14].
The new WHO child growth standards were applied in
this study, and it is notable that the prevalence of stunting,
severe stunting, wasting, and severe wasting were higher
when compared with the prevalence when the conventional
NCHS reference population was used. The new WHO child
growth standards are considered to be more accurate in
dening child malnutrition and are now being more widely
applied in epidemiologic studies.
The large proportion of household nances devoted to
food expenditures indicates that poverty is widespread in
this sample. In all socioeconomic status quintiles food ex-
penditures account for 50% of total expenditures on av-
erage. Poorer households spent proportionally more money
on tobacco products than wealthier households, suggesting
tobacco addiction imposes the greatest strain on the house-
hold budgets of the poorest sector of society. These are the
same households in which children are most likely to have
malnutrition. Households with no tobacco use were able to
devote a larger proportion of money to purchase micronu-
trient-rich animal and plant foods such as eggs, sh, meat,
milk, green leafy vegetables, and fruits. Thus, efforts to
decrease tobacco use, which would lead to a decrease in
expenditures on tobacco and medical care due to tobacco
use in the household, could increase disposable income and
contribute to poverty reduction and nutritional interven-
tions.
Tobacco control legislation in Bangladesh had been
weak until the recent passing of the Tobacco Control Act in
2005. This legislation bans all forms of direct and indirect
advertising for tobacco products, excluding point-of-pur-
chase advertising [15]. The sale of tobacco products in
vending machines has been outlawed, and pack warnings
now must convey more specic health messages and cover
30% of the package. The Tobacco Control Act also bans
all forms of sponsorship to promote tobacco and smoking in
public and government facilities. Tobacco activists in Ban-
gladesh are concerned that the portion of the legislation that
allows for the creation of smoking areas within smoke-free
public places lacks clarity and that, without greater descrip-
tion, smoke-free public places will not be sufciently free of
tobacco smoke [15]. British American Tobacco remains one
of the nations most protable industries, but the govern-
ment has agreed to develop programs to discourage the
entrance of new tobacco companies into the country. Cur-
rently, there is no research available that reviews the com-
pliance and success of this legislation. But this act signies
a major success for tobacco control in Bangladesh and for
the Framework Convention on Tobacco Control to which it
is signatory.
Table 4
Logistic regression models for parental tobacco use and risk of severe child malnutrition in households in Bangladesh
Characteristics Severe stunting Severe underweight Severe wasting
OR 95% CI P OR 95% CI P OR 95% CI P
Model 1
Parental tobacco use 1.35 1.281.42 0.0001 1.39 1.301.48 0.0001 1.26 1.101.44 0.0001
Model 2
Parental tobacco use 1.16 1.101.23 0.0001 1.21 1.131.30 0.0001 1.14 0.981.32 0.09
Male gender 1.16 1.111.22 0.0001 1.05 0.991.12 0.08 1.50 1.321.71 0.0001
Child age (mo)
05 0.62 0.550.70 0.0001 0.84 0.730.97 0.016 3.18 2.354.30 0.0001
611 0.62 0.550.70 0.0001 0.96 0.851.10 0.58 2.65 1.963.59 0.0001
1223 1.64 1.501.78 0.0001 1.54 1.391.71 0.0001 3.04 2.314.00 0.0001
2435 1.62 1.491.77 0.0001 1.56 1.401.73 0.0001 1.95 1.462.59 0.0001
3647 1.43 1.311.56 0.0001 1.24 1.111.38 0.0002 1.31 0.961.78 0.09
4959 1.00 1.00 1.00
Maternal age (y) 0.99 0.980.99 0.0001 0.99 0.990.99 0.01 1.00 0.991.01 0.79
Maternal education level (y)
0 3.05 2.643.52 0.0001 2.97 2.483.55 0.0001 1.66 1.212.26 0.0015
13 2.54 2.163.00 0.0001 2.64 2.153.24 0.0001 1.07 0.721.59 0.76
46 2.24 1.942.59 0.0001 2.28 1.902.84 0.0001 1.27 0.931.74 0.14
79 1.46 1.261.70 0.0001 1.54 1.271.86 0.0001 0.97 0.701.35 0.86
10 1.00 1.00 1.00
4 individuals eating from same cooking pot 1.07 1.011.12 0.022 1.07 1.011.14 0.039 0.98 0.851.13 0.77
Total monthly household expenditure per capita (taka)
Quintile 1 1.61 1.491.75 0.0001 1.80 1.631.99 0.0001 1.49 1.331.83 0.0001
Quintile 2 1.41 1.301.53 0.0001 1.46 1.321.61 0.0001 1.16 0.941.42 0.17
Quintile 3 1.34 1.241.46 0.0001 1.44 1.301.59 0.0001 1.19 0.971.46 0.09
Quintile 4 1.16 1.071.26 0.0004 1.23 1.111.37 0.0001 1.03 0.831.27 0.82
Quintile 5 1.00 1.00 1.00
Urban location 0.99 0.931.05 0.70 0.96 0.891.04 0.32 0.92 0.781.09 0.33
CI, condence interval; OR, odds ratio
737 C. M. Best et al. / Nutrition 23 (2007) 731738
The high prevalence of adult tobacco use in Bangladesh
indicates that the nations public health efforts in tobacco
control must be of sustained intensity. A decrease in tobacco
consumption in Bangladesh could improve immediate
health outcomes, such as the incidence of cancer, cardio-
vascular, and respiratory diseases, and intermediate health
outcomes that are mediated by poverty, such as child mal-
nutrition. Compliance with the recent Tobacco Control Act
is essential for a reduction in the negative health outcomes
caused by tobacco addiction and exposure to tobacco
smoke.
Conclusions
Parental use of tobacco increases the risk of child mal-
nutrition among households in Bangladesh and diverts
money from necessities such as animal and plant foods,
education, and health care. Tobacco control should be con-
sidered as part of integrated public health strategies aimed at
decreasing child malnutrition in developing countries.
Acknowledgments
The authors acknowledge the contribution of the survey
participants and the entire NSP team.
References
[1] Centers for Disease Control and Prevention. Use of cigarettes and
other tobacco products among students aged 1315 yearsworld-
wide, 19992005. MMWR 2006;55:556.
[2] de Beyrer J, Brigden LW, editors. Tobacco control policy: strategies,
successes, and setbacks. Washington, DC: The World Bank and
Research for International Tobacco Control; 2003.
[3] Efroymson D, Ahmed S, Townsend J, Alam SM, Dey AR, Saha R, et
al. Hungry for tobacco: an analysis of the economic impact of tobacco
consumption on the poor in Bangladesh. Tob Control 2001;10:2127.
[4] Jenkins CNH, Dai PX, Ngoc DH, Kinh HV, Hoang TT, Bales S, et al.
Tobacco use in Vietnam: prevalence, predictors, and the role of the
transnational tobacco corporations. JAMA 1997;227:172631.
[5] Esson KM, Leeder S. The millennium development goals and tobacco
control: an opportunity for partnership. Geneva: World Health Orga-
nization; 2004.
[6] Bloem MW, Moench-Pfanner R, Panagides D, editors. Health &
nutritional surveillance for development. Singapore: Helen Keller
Worldwide; 2003.
[7] de Pee S, Bloem MW. Assessing and communicating the impact of
nutrition and health programs. In: Semba RD, Bloem MW, editors.
Nutrition and health in developing countries. Totowa, NJ: Humana
Press; 2001, p. 483506.
[8] Mason JB, Habicht JP, Tabatabai H, Valverde V. Nutritional surveil-
lance. Geneva: World Health Organization; 1984.
[9] World Health Organization. WHO child growth standards, SAS mac-
ros. Available at: http://www.who.int/childgrowth/standards/en.
[10] de Onis M. Child growth and development. In: Semba RD, Bloem
MW, editors. Nutrition and health in developing countries. Totowa,
NJ: Humana Press; 2001, p. 7191.
[11] World Medical Association. World Medical Association Declaration
of Helsinki: ethical principles for medical research involving human
subjects. Bull World Health Organ 2001;79:3734.
[12] SAS onlinedoc 9.1.3. Available at: http://support.sas.com/onlinedoc/
913/docMainpage.jsp. Accessed February 14, 2007.
[13] Semba RD, Kalm LM, de Pee S, Ricks MO, Sari M, Bloem MW.
Paternal smoking is associated with increased risk of child malnutri-
tion among poor urban families in Indonesia. Pub Heath Nutr 2007;
10:715.
[14] US Department of Health and Human Services. The health conse-
quences of nicotine addiction: a report of the surgeon general. Rock-
ville, MD: CDC Center for Health Promotion and Education; 1988.
[15] Bangladesh Anti-Tobacco Alliance. Overview of the tobacco control
act 2005. BATA. Available at: http://bata.globallink.org/lawpass.htm.
Accessed November 2006.
738 C. M. Best et al. / Nutrition 23 (2007) 731738