Use of traditional cooking fuels and the risk of young adult cataract in rural Bangladesh: a hospital-based case-control study
Use of traditional cooking fuels and the risk of young adult cataract in rural Bangladesh: a hospital-based case-control study
Joydhan Tanchangya 0
Alan F Geater 1
0 Impact Jibon Tari Floating Hospital, Impact Foundation Bangladesh , 7th Floor, House 23, Road 113/A, Gulshan-2, Dhaka-1212 , Bangladesh
1 Epidemiology Unit, Faculty of Medicine, Prince of Songkla University , Hat Yai, Songkhla, 90112 , Thailand
Background: This study aimed to investigate the independent relationship between the use of various traditional biomass cooking fuels and the occurrence of cataract in young adults in rural Bangladesh. Methods: A hospital-based age- and sex-matched case-control study incorporating two control groups was conducted. Cases were cataract patients aged 18 and 49 years diagnosed on the basis of any opacity of the crystalline lens or its capsule and visual acuity poorer than 6/18 on the Log Mar Visual Acuity Chart in either eye, or who had a pseudophakic lens as a result of cataract surgery within the previous 5 years. Non-eye-disease (NE) controls were selected from patients from ENT or Orthopaedics departments and non-cataract eye-disease (NC) controls from the Ophthalmology department. Data pertaining to history of exposure to various cooking fuels and to established risk factors for cataract were obtained by face-to-face interview and analyzed using conditional logistic regression. Results: Clean fuels were used by only 4% of subjects. A majority of males (64-80% depending on group) had never cooked, while the rest had used biomass cooking fuels, mainly wood/dry leaves, with only 6 having used rice straw and/or cow dung. All females of each group had used wood/dry leaves for cooking. Close to half had also used rice straw and/or cow dung. Among females, after controlling for family history of cataract and education and combining the two control groups, case status was shown to be significantly related to lifetime exposure to rice straw, fitted as a trend variable coded as never, median of all exposed, > median of all exposed (OR = 1.52, 95%CI 1.04-2.22), but not to lifetime exposure to wood/dry leaves. Case status among females showed an inverse association with ever use of cow dung as a cooking fuel (OR 0.43, 95%CI 0.22-0.81). Conclusions: In this population, where cooking is almost exclusively done using biomass fuels, cases of young adult cataract among females were more likely to have had an increased lifetime exposure to cooking with rice straw fuel and not to have cooked using cow dung fuel. There is a possibility that these apparent associations could have been the result of uncontrolled founding, for instance by wealth. The nature of the associations, therefore, needs to be further investigated.
Young adult cataract; risk factor; traditional cooking fuels; Bangladesh
A recent population-based survey in Bangladesh
estimated 650,000 people were blind aged 30 years and
older, around 80% due to cataract, and 130,000 new
cases developed annually [1,2]. The prevalence of
bilateral blindness was 2.9%, of severe visual impairment
1.2%-2.0% and of visual impairment 8.4% .
While the age-specific prevalence of blindness in
Bangladesh, as in other countries, was found to increase
with increasing age, there are some indications that the
incidence of young adult cataract may be higher than
expected compared with the age distribution of cataract
patients in other countries. Data from the Impact Jibon
Tari Floating Hospital in Bangladesh reveal that
cataract operations account for 90% of all eye surgery, and
that 14% of these patients are aged between 18 and 45
years. This contrasts with the 3.8% of cataract surgery
patients between the ages of 20 and 49 in USA . Even
though 66% of adults in the Bangladesh are below 45,
compared with only 47% in the USA , the large
discrepancy in proportions of young adults among cataract
surgery patients suggests a higher relative incidence of
cataract among young adults in Bangladesh, and thus
the possible existence of some specific risk factors for
early onset of cataract in this population.
Age-specific blindness in Bangladesh was also found to
be higher in women and the illiterate and disadvantaged
people . Most of those women were of low
socioeconomic status and commonly used wood, leaves, cow
dung and rice straw for household cooking and from a
young age. Exposure to cheaper cooking fuels has been
identified to be a risk factor in India and Nepal [6,7] and
is very common in rural Bangladesh. These exposures
might, therefore, be related to the high incidence of
young adult cataract in Bangladesh. Such exposures are
potentially avoidable and their identification as risk
factors for young adult cataract in the Bangladeshi setting
could indicate possible approaches for future
policymaking to prevent blindness.
Previous studies of risk factors for cataract have
considered cataract in the elderly. In our study, by contrast, we
focused on a younger age group to determine if exposures
to traditional biomass cooking fuels are actually risk
factors among this population in the rural Bangladesh setting.
The study was designed as a hospital-based matched
case-control study with two types of control and
conducted from May to October 2009 in Impact Jibon Tari
Floating Hospital, Bangladesh. The hospital is located on
a river and moves to disadvantaged areas of the country
at 5- to 6-month intervals. It aims to provide health
services, both clinical and surgical, to address the problems
of disability in remote areas of the country. At the time
of the study, the hospital was located in Barisal district,
in the south-west of Bangladesh, a region criss-crossed
by many rivers.
Subjects were recruited from males and females aged
between 18-49 years who came to visit Impact Jibon
Tari Floating Hospital, Bangladesh, during the study
period. Owing to the relative paucity of healthcare
facilities above the primary level in rural areas, the potential
catchment area for patients is wide, covering much of
the district. Information on the location of the hospital
is disseminated through the local media.
Although over 90% of households in rural Bangladesh
have been reported to use traditional fuel of some kind
, exposure to any one type while actually doing the
cooking was expected to be much less. The required
sample size for each of the case, non-eye-disease control
and non-cataract eye-disease control groups, was
therefore based on having a power of 80% to detect an odds
ratio of at least 2 if between one third and one half of
controls were exposed and the correlation coefficient for
fuel use between cases and controls was 0.1 . The
sample size was calculated to be 153 per group.
Ophthalmologic patients aged between 18-49 years who
came to visit the eye specialist in Impact Jibon Tari
Floating Hospital and diagnosed as cataract by clinical
and slit-lamp examination results showing opacity of the
crystalline lens or its capsule and visual acuity poorer
than 6/18 on the Log Mar Visual Acuity Chart in either
eye, or having a pseudophakic lens as a result of cataract
surgery within the previous 5 years, were recruited as
cases in this study. Patients having congenital cataract,
severe mental disorder or cataract secondary to serious
eye disease, such as glaucoma, diabetes retinopathy or
severe injury, were excluded.
Non-eye-disease (NE) controls
Non-eye-disease control subjects were selected from
patients without any eye problem but attending the ear,
nose and throat or orthopedics departments in the same
hospital and 1:1 matched to cases on age (within same
5year age range) and sex. Inclusion criteria were: aged
between 18 and 49 years, having no history of prior
cataract/eye surgery and no eye problem within last 3 months.
Known cases of myopia were excluded. Potential subjects
were recruited from among out-patients fulfilling the
inclusion criteria on the same or following working day
as the case to which they were matched.
Non-cataract eye-disease (NC) controls
Non-cataract eye-disease controls were selected from
among patients attending in the same hospital with an
eye problem other than cataract and matched 1:1 with
cases on age (within same 5-year age range) and sex.
Inclusion criteria were: aged between 18 and 49 years,
having no history of prior cataract/eye surgery.
Recruitment of potential subjects was made from the first
patient subsequent to case attainment who fulfilled the
All eye examinations were carried out by an
ophthalmologist using same LogMar Visual Acuity Chart and
slit-lamp and in a similar manner for both cases and
controls. Matching was done to allow control of the
potential confounding effect arising from age and sex
(both documented risk factors for cataract and likely to
be associated with exposure to different types of cooking
fuel). The ideal control group would have comprised
subjects representative of the same catchment
population as the cases but not having cataract. However,
hospital-based controls were chosen for this study to
reduce the potential bias that could arise from a possible
lower willingness of community-based controls to
respond to the interviewer. A single group of controls
representative of non-case patients would have included
a large proportion of patients with other eye diseases as
the provision of health services for eye disease is a
major component of the health care provision of the
Jibon Tari Hospital. The two controls groups were
therefore recruited to avoid potential problems that
could arise from any similarity in exposure history
between cases and patients with other eye disease.
Data were collected using a structured questionnaire
through personal face-to-face interview following clinical
examination. The questionnaire covered general
sociodemographic characteristics, details of cooking history
and exposure to various cooking fuels, and other
previously documented risk factors for cataract.
Exposure to a cooking fuel in this study was defined as
use of the fuel by the person doing the cooking. History
of exposure was recorded on a matrix table comprising
fuel types in rows and age (years 11 to 49) in columns.
Recall of exposure earlier than age 11 was considered to
be unreliable. Information was also obtained regarding
the number of cooking sessions per day, hours spent
cooking per session and number of days when cooking
was done per week. These data were used to derive
parameters of frequency, intensity, duration and cumulative
lifetime exposure to each cooking fuel type.
Information on other documented or potential risk
factors for cataract, including a history of hypertension
and diabetes, was obtained by interview and not verified
independently. Family history of cataract referred to
parents, grandparents and siblings.
Data were entered using EpiData version 3.1  and
transferred into R version 2.10.0  for cleaning,
exploration and analysis. The distributions of variables
were explored and summarized within each outcome
group using mean and standard deviation or frequency.
Tabulation of independent variables was performed for
matched pairs of case vs non-eye-disease control and for
case vs eye-disease control. In addition, patterns of fuel
use among males and females were tabulated.
Exposure variables showing any indication of
differing within the matched pairs in either comparison (p <
0.2), in addition to selected parameters of cooking fuel
use, were included in initial sex-specific conditional
logistic regression models for matched pairs of case vs
non-eye-disease control and of case vs non-cataract
eye-disease control. These models allowed the
independent associations between use of each fuel type and
case status, controlling for confounding effects among
the fuels and of other covariates, to be revealed. The
models were refined by successive removal of variables
showing no statistically significant contribution to the
fit of either model, other than the selected cooking
fuel variables, which were retained in the models
irrespective of their statistical significance. Likelihood ratio
test p-values 0.05 were considered as indicating
Associations between case status and alternative,
ordinal, parameters of cooking fuel use were then explored
among these fuels to identify any dose-response
relationships, first using all subjects and subsequently in the
subset of female subjects. These ordinal exposure
variables comprised age at first exposure, frequency (times
per week during the years used), intensity (hours per
week during the years used), duration (years) and
lifetime exposure (lifetime hours of use). Each variable
except age at first exposure was cut into three levels:
never used, used for less than or equal to the median
among all users, and greater than the median for all
users. Age at first exposure was cut into never exposed,
older than or equal to the median age of all users, and
less than the median age of all users. These variables
were fitted first in categorical form and subsequently, if
appropriate, in trend-across-category form.
To evaluate the effect of matching, additional analyses
were performed using unconditional logistic regression
modelling including the same variables as above.
The study was approved by the Ethics Committee of
the Faculty of Medicine, Prince of Songkla University,
Thailand, and oral approval was granted after a
detailed presentation of the research proposal within
the management of Impact Foundation Bangladesh, the
authority of the study hospital, before conducting the
study. All potential participants in the study were
informed that the study was aimed at identifying
certain behaviours that might increase the risk of their
developing ailments common in rural Bangladesh.
Written informed consent was requested from all
potential respondents before their participation. For
participants who were willing but could not sign, a
finger print was taken after explaining the research
process. Only after the patient gave documented consent
was the interview conducted. Computerized data did
not indicate the identity of any patient.
None of the subjects selected for inclusion in the study
refused to participate. One subject who had been selected
as a NE control had insufficient time to be interviewed
and was replaced by the next eligible subject. A total of
459 subjects, including 153 cataract cases aged between
18-49 years with an equal number of each control group
matched on age and sex, were recruited and interviewed.
There were slightly more females than males. The mean
age of all participants was 41.8 (SD 6.3) years and 30
percent were aged 40 or less. In all groups, most of the
females were housewives, the commonest occupation
among males was farmer, and about 80 percent were
classified as having low socioeconomic status as
measured on the modified Kuppuswami scale using
education, occupation and income of the family head. Around
70 to 75 percent of subjects were classified as being
underweight. Cases had less commonly received
secondary or tertiary education than controls, particularly
noneye-disease controls, and more commonly reported a
history of current or past smoking and a greater
occupational exposure to sunlight. Cases also more frequently
reported a family history of cataract (Table 1). Very few
subjects reported a history of diagnosis of hypertension
(2%) or diabetes mellitus (1%).
Over half of the NE controls (56.2%) were diagnosed
as having various ear diseases followed by nasal, throat,
and orthopedic diseases. Among NC controls, refractive
error, corneal diseases and conjunctival diseases were
most common (Table 2).
Cooking history and exposure to various cooking fuels
among subjects in each group are shown in Table 3.
About two thirds of subjects in each group reported a
history of cooking, either regularly or occasionally. All
females in the study had cooked, almost all regularly,
whereas only 36%, 34% and 27% of males respectively
among cases, NE and NC controls had a history of
cooking, with a majority (70% to 85%) cooking only
occasionally. The commonest fuel used for cooking in all groups
was wood and/or dry leaves, followed respectively by cow
dung and rice straw. Further analysis of cooking fuel
exposure was confined to these three groups of cooking fuels.
Gas or kerosene was used by only a small number of
subjects in any group, comprising less than 4% of the total.
The proportion of cases using rice straw was higher than
that of either of the controls, whereas the proportion of
cases using cow dung was lower than that in each control
group. The lifetime duration of cooking activities was
somewhat higher among cases than either of the control
groups. Exposure parameters explored included age at first
exposure, frequency of exposure (times per week),
intensity of exposure (hours per week during the years used),
duration from first to most recent exposure in years
(irrespective of the frequency of intensity of exposure),
and total lifetime exposure (hours of actual exposure).
However, these differentials were not statistically
Patterns of exposure to traditional fuels differed between
females and males (Table 4). Only males reported no
exposure to any of the three traditional fuel types, more
commonly in the NE and NC controls (79% and 70%
respectively) than in the cases (64%). All of the remaining
males had used wood/dry leaves as cooking fuel, but only
6 of these had ever used rice straw or cow dung. By
contrast, all females had used wood or dry leaves for cooking,
with around half of the cases and NE controls and over
half of NC controls using rice straw and/or cow dung in
addition. The most striking differences among the groups
was the higher percentage of cases (17.5%) than of either
control (6% and 9%) who used rice straw as the only
additional fuel, and the lower percentage of cases (16%) than
of controls (27% and 29%) who used cow dung as the sole
Comparison of socio-demographic variables among
cases and controls within each of the exposure patterns
revealed that only education differed significantly across
the groups, with no education being more common
among cases than controls in exposure patterns 0 and 4
(data not shown).
Because of the different combinations of cooking fuels
used by males and females, conditional logistic regression
modelling of ever use of the various cooking fuels was
performed separately for males and females. For the male
model, the 6 males who had cooked with rice straw and/
or cow dung were excluded. Traditional cooking fuel
types were fitted in binary form (ever used vs never used)
as appropriate for males and females. Other variables
initially included in the models were religion, educational
level, income of family, treatment or diagnosis of
hypertension, family history of cataract, sunlight exposure in
the workplace and smoking status, as the p-value for the
univariate association of each of these variables with case
status in comparison with at least one of the control
groups was < 0.2.
After refinement, the variables remaining in the model
for females, in addition to ever use of the traditional
cooking fuels, were education level of the subject and
family history of cataract. The same variables, with the
addition of smoking status, were retained in the model
for males (Table 5). Religion, family income, history of
hypertension, and workplace exposure to sunlight were
not significant in the multivariate setting.
Family history of cataract was strongly and
significantly associated with case status in both models within
each sex (ORs ranging from 2.14 to 3.59). Lower
education was associated with case status in the comparison
Current or past
* Likelihood ratio p-values obtained from univariate conditional logistic regression models.
Table 1 Distribution of socio-demographic characteristics among cases and controls (Continued)
with NE controls (compared to secondary education or
higher, ORs in males and females of 5.65 and 4.21 for
primary education, and 4.90 and 3.07 for less than
primary). There was also an association with current or
past smoking among males in comparison with NC
(OR = 3.40).
Overall, associations between use of traditional
cooking fuels and case status were more pronounced and
significant in the comparison with NE than with NC
controls. Nevertheless, as the odds ratios of each fuel in
both comparisons were in the same direction, the
associations with case status in comparison with the
combined control groups are also shown in Table 5. Among
females, case status was seen to be positively associated
with ever use of rice straw (OR = 1.95, 95%CI 1.03-3.69)
but negatively associated with ever use of cow dung
(OR = 0.45, 95%CI 0.24-0.84).
The level of exposure to each traditional fuel was
expected to be influenced by the level of use of other
traditional fuels, and to differ between males and
females. For example, lifetime exposure exclusively to
wood/dry leaves was lower among exposed males than
among females (median 3,650 hours, interquartile range,
IQR, 4,732-21,631 hours vs 30,285 hours, IQR 5,096
74,256 hours). Parameters reflecting the level of
exposure to each fuel were therefore fitted separately to
models containing family history of cataract, smoking
status and education and using the combined control
group. As the different exposure levels in males and
females should be reflected in these parameters of the
Table 2 Distribution of diagnosed diseases among
Ear diseases 86 (56.2) Refractive error
Nasal diseases 15 (9.8) Corneal diseases
Throat diseases 13 (8.5) Conjunctival diseases
Orthopedic diseases 11 (7.2) Lacrimal tract disease
Others 28 (18.3) Others
magnitude of exposure, initial models were constructed
using all subjects. The patterns of association were
similar for each parameter. The model for lifetime
exposures, controlling for family history of cataract,
education level and smoking status, using all subjects, is
shown in upper part of Table 6.
A significant dose-response relationship with case
status was revealed for lifetime exposure to rice straw as a
cooking fuel (OR = 1.53, 95%CI 1.06-2.24). However,
the odds ratios for lifetime exposure to wood/dry leaves,
while suggestive of a trend relationship, were not
statistically significant. By contrast, the odds ratios for case
status associated with use of cow dung indicated a
strong inverse association but with no evidence of a
trend relationship (ever vs never use OR = 0.42, 95%CI
In order to confirm these associations, a subsequent
model using the subset of females was constructed
(Table 6, lower part). This was done to avoid any effect
modification due to sex, which might have arisen from
the fact that only males were free of any exposure to
cooking with traditional fuels and all females had been
exposed to cooking with wood and/or dry leaves.
Associations between case status and lifetime exposure to
cooking with rice straw and with cow dung were almost
identical with those seen in the model using all subjects.
Unconditional binary regression models including the
same variables as above yielded essentially the same
relationships between lifetime exposures and case status
as in the conditional models.
This study aimed to test the hypothesis that exposure to
cooking with biomass fuels, such as wood or dry leaves,
cow dung and rice straw, is significantly associated with
the development of cataract among adults less than 50
years of age in rural Bangladesh. After adjusting for
family history of cataract, smoking status and level of
formal education, differences in exposure to the various
biomass fuels were found. Cooking with rice straw was
identified as being positively associated with young adult
* Likelihood ratio p-values obtained from univariate conditional logistic regression models.
Table 5 Conditional logistic models for ever use of cooking fuels
Cases vs non-eye-disease
Case vs non-cataract eye-disease
a-OR (95% CI)
a-OR (95% CI)
a-OR (95% CI) p- value*
CS = Cases, NE = Non-eye-disease controls; NC = Non-cataract-eye-disease controls.
* Likelihood ratio test.
# Six males who had been exposed to cooking with rice straw and/or cow dung are omitted from this model.
cataract, whereas cooking with cow dung was negatively
The positive association of cooking with rice straw
with case status was stronger in comparison with NE
than with NC controls. A possible, though unsupported,
explanation for this difference depending on the type of
controls employed is that the other eye diseases share
these risk factors with cataract patients, or that patients
with diseases included among the non-eye-disease
control group are less exposed to these particular cooking
fuels. Inter-comparison of the two types of control,
however, revealed no significant associations with the
use of these cooking fuels.
It is important to note that, in our study, the
association of young adult cataract with using rice straw for
cooking among females is relative to the use of other
biomass fuels. Because of the sex differences in the
patterns of exposure to cooking and the fuels used, the
comparator group for males was no exposure to
cooking. In contrast, other studies have examined the
associations between cataract and the use of cheaper,
biomass or solid fuels relative to the use of clean fuels.
The use of less expensive cooking fuels was more
common among patients with age-related cataract than
noncataract patients in India  and, compared with the
use of stoves burning clean fuels such as biogas,
liquefied petroleum gas or kerosene, the use of solid fuel in
unvented stoves was associated with cataract among
females of any age in the Nepal-India border area . In
both studies the associations remained significant after
adjustment for other risk factors, including low
A plausible mechanism underlying the association
between the use of certain biomass fuels and
development of cataract may be related to the constituents of
the large amounts of smoke produced from these fuels
Table 6 Conditional logistic regression models for lifetime exposure to cooking fuels
Number of patients
Case vs combined controls
crude-OR (95% CI)
a-OR (95% CI)
damaging the tissues of the eye following either systemic
absorption or even local diffusion through the cornea. It
has been suggested that such damage may be a result of
the endogenous generation of reactive oxygen species by
photodynamic action, similar to the purported
mechanism by which smoking tobacco may raise the risk of
Despite the relationship between risk of cataract and
use of cooking fuel being reported in a number of
studies, few have attempted to document the associations
for different types of biomass fuel. The component
materials have either not been specified or have been
specified but combined in the analysis and comparisons
made with the use of clean fuels. Thus, Mohan (1989)
and Mishra (1999) reported elevated risks of both
cataract and blindness among an Indian population with
exposure to the smoke of biomass cooking fuel,
specified as wood, crop residuals and/or cow dung, but
separate analyses of each of these materials were not
In view of these reports, the independent inverse
association between the use of cow dung as a cooking fuel
and case status in our study was unexpected. The
relationship held true in comparisons with each type of
control. While copious amounts of smoke are known to
emanate from burning cow dung, and indoor burning of
dung has been reported to produce higher PM10
concentrations than that of wood or straw , the opposite
directions of relationship in our study of rice straw and
cow dung might be related to the different complement
of smoke constituents. Although cow diet consists
largely of fresh grass, bacterial and enzymatic actions of
the bovine gastrointestinal tract result in considerable
transformations of the plant material.
Constituents of smoke from biomass fuels have been
reported to vary considerably with the type of stove
employed and with various other differences in the way
the fuel is prepared. Comparative information on the
constituents of smoke from different biomass fuels, or
from dung fuel separate from other biofuels, is scarcely
available in the scientific literature, despite several
studies of smoke constituents of biomass fuels combined
[17-21]. Mudway (2005), however, demonstrated that
particles derived from the burning of cow dung cake
burned in a traditional Indian cooking stove and
deposited in the human respiratory tract lining fluid had
considerable oxidative activity, which was mostly due to
their transitional metal content . If the postulated
mechanism whereby smoke from biomass fuels induces
cataract formation through the activity of reactive
oxygen species is true, then it is difficult to understand
why smoke from cow dung does not have a positive
association with young cataract, similar to that of rice
straw. Further comparative analyses are required to
identify differences in the smoke constituents and
elucidate possible differences in the mechanisms of action.
Consideration, however, must also be given to the
possibility that the apparent protective effect against the
development of young adult cataract of using cow dung
as a cooking fuel could be due to uncontrolled
confounding. Exposure to cow dung as cooking fuel is more
common among middle class families in rural areas in
Bangladesh. Cows are usually used for cultivation and
dairy products, so more frequently kept by land and
farm owners, whereas poor families can rarely afford to
buy or keep cattle. Use of cow dung as a fuel thus may
be acting as a proxy for higher socioeconomic status,
which itself has been identified in some previous studies
to be associated with a lower prevalence of cataract (of
any type) [1,6,23-25]. Nevertheless, adjusting our models
for family income level or for the composite
socioeconomic status indicator based on the Kuppuswami scale
had no discernable effect on the relationship between
case status and use of cow dung as a cooking fuel, so
that confounding, if it is to be invoked as the
explanation for the relationship, must involve an as yet
It is of interest, however, that an Indian study of the
relationship between fuel use and ocular morbidity in
which separate independent associations between
different types of cooking fuel and cataract were examined
reported a significantly increased risk for wood but not
for cattle dung or for gas, kerosene or coal . On the
other hand, eye irritation was significantly associated
with the use of coal and cattle dung but not the other
Other variables related to case status in our study
family history of cataract, a history of cigarette smoking,
and low educational attainment - have each been
recognized as risk factors for cataract in other studies [27-30].
The relationship with low educational achievement may
be explained by the generally poorer nutritional status
of less educated people. Poor nutritional status  as
well as experience of dehydrational crises , have
been identified as independent risk factors for cataract.
Unlike the findings of some previous studies ,
working in sunlight was not identified as being associated
with case status.
A limitation of this study stems from difficulties in
recalling lifetime use of various cooking fuels, although
recall was stimulated during the interview by referring
to significant life events of each patient. However, it is
unlikely that recall misinformation was differential as all
patients, both cases and controls, were visiting the
hospital for treatment of some ailment, and the specific
hypothesis under study was not known to the subjects.
Such random errors that may have occurred would
therefore tend to reduce the observed strength of
association between exposures and outcome. Interviewer
knowledge of subject status and the hypothesis under
study, which could theoretically introduce bias and an
overestimation of associations, is unlikely to have
introduced significant distortion of the data as the interviews
were carried out strictly according to the structured
Variables of exposure to cooking fuels in this study
were confined to those subjects who did the cooking. As
other family members could also be exposed, albeit
probably to a lower extent, the associations might have
been underestimated. Indeed it has been shown that the
PM10 concentrations in living rooms were only slightly
lower than, and closely followed, those in the kitchen
throughout the day, in poor households in Bangladesh
. Unfortunately, in our study, it is not known
whether those subjects who were not exposed to
cooking generally remained in the house while cooking was
done or were at work away from home.
Both the case-control design of the study and the fact
that the relationship between use of rice straw as
cooking fuels and development of young adult cataract was
not consistently significant in comparisons with the two
controls, preclude our drawing firm conclusions
regarding a causal relationship between the cooking with rice
straw and the risk of developing young adult cataract.
The inverse relationship between the use of cow dung
and case status, however, was much more consistent.
Nevertheless, a plausible explanation for the association
The study did not classify the cases with respect to
type of cataract. Unless all types share the same risk
factors, any heterogeneity of cataract types would have the
effect of diluting the true relationships with exposure.
Finally, since this study was conducted in a charitable
non-government organization hospital, catering for
disadvantaged villagers in remote parts of the country,
the range of socio-economic status among subjects was
not very wide. Such restricted variability may have
prevented the identification of certain risk factors that may
be seen in studies with a wider variety of patient
backgrounds. On the other hand, the location of the hospital
at the time of the study in a poorly accessible rural
district with only meagre permanent healthcare facilities
makes it unlikely that the cases and controls were
drawn from different catchment populations.
The strength of this study lies in its separation of
different types of traditional cooking fuel, which allowed
the identification of contrasting directions of association
among these fuel types.
Our results may add to the understanding gained from
earlier studies regarding the relationship between the
use of cheaper biomass cooking fuels and the risk of
cataract. Our study provides evidence of differing
histories of cooking with several types of biomass fuel
between patients developing young adult cataract and
control subjects in a population whose cooking fuel
exposure is almost exclusively to biomass fuels.
Interestingly, the study revealed evidence of lower exposure to
cooking with cow dung among cataract patients than
among control subjects. In view of the lack of plausible
biological explanation for a protective effect of cooking
with cow dung and the possibility that the relationship
may have resulted from inadequately controlled
confounding, further studies of the relationship between
exposure to cow-dung smoke and cataract in other
settings, as well as comparative analysis of the constituents
of the smoke from rice straw and that from cow dung,
are needed. Various means of reducing indoor air
pollution derived from the use of biomass cooking fuels have
been described, such as changing to the use of
alternative fuels such as biogas or liquefied petroleum gas,
adopting stoves that can reduce the free emission of
smoke from biomass fuels, and ensuring adequate
ventilation of cooking areas of the house [16,34]. Should the
use of rice straw as a cooking fuel be confirmed as a
risk factor, then adoption of these pollution-reducing
measures may benefit the health of Bangladeshi villagers
by reducing the incidence of young adult cataract.
List of abbreviations used
NE: Non-eye-disease control; NC: Non-cataract eye-disease control; WD/DL:
Wood and/or dry leaves; RS: Rice straw; CD: Cow dung; OR: Odds ratio; a-OR:
Adjusted odds ratio; CI: Confidence interval.
This study is part of the research study of the first author to fulfil the
requirements of the degree of M.Sc. in Epidemiology, Prince of Songkla
University, Thailand. Fredskorpset, Norway, is gratefully acknowledged for
providing partial funding for the study through the Asia Health Research
Network. The authors would like to acknowledge all the doctors of Impact
Jibon Tari Floating Hospital, Bangladesh, who helped to examine and
diagnose the patients, all staff for facilitating data collection, and all
participants in the study. Special thanks go to Dr Pradip Sen Gupta for
advice on proposal development. The authors also wish to thank the
management of the Impact Foundation Bangladesh, the authority of Impact
Jibon Tari Floating Hospital, for permission to conduct the study in that
JT designed the study, carried out the entire data collection, conducted data
analysis and interpretation of the data. AG supervised the entire study and
was involved equally during design, data analysis and interpretation of the
data. Both authors read and approved the final manuscript
The authors declare that they have no competing interests.
1. Dineen BP , Bourne RR , Ali SM , Huq DM , Johnson GJ : Prevalence and causes of blindness and visual impairment in Bangladeshi adults: results of the National Blindness and Low Vision Survey of Bangladesh. Br J Ophthalmol 2003 , 87 : 820 - 828 .
2. Bourne RR , Dineen BP , Ali SM , Huq DM , Johnson GJ : Outcomes of cataract surgery in Bangladesh: results from a population based nationwide survey . Br J Ophthalmol 2003 , 87 : 813 - 819 .
3. Wadud Z , Kuper H , Polack S , Lindfield R , Akm MR , Choudhury KA , Lindfield T , Limburg H , Foster A : Rapid assessment of avoidable blindness and needs assessment of cataract surgical services in Satkhira District , Bangladesh. Br J Ophthalmol 2006 , 90 : 1225 - 1229 .
4. Erie JC , Baratz KH , Hodge DO , Schleck CD , Burke JP : Incidence of cataract surgery from 1980 through 2004: 25-year population-based study . J Cataract Refract Surg 2007 , 33 : 1273 - 1277 .
5. U.S. Census Bureau, International Data Base. [http://www.census.gov/ipc/ www/idb/index.php], Accessed 19 August , 2010 .
6. Ughade SN , Zodpey SP , Khanolkar VA : Risk factors for cataract: a case control study . Indian J Ophthalmol 1998 , 46 : 221 - 227 .
7. Pokhrel AK , Smith KR , Khalakdina A , Deuja A , Bates MN : Case-control study of indoor cooking smoke exposure and cataract in Nepal and India . Int J Epidemiol 2005 , 34 : 702 - 708 .
8. Local Government Engineering Department: Reducing greenhouse gas emissions by promoting bioenergy technologies for heat applications . Report no EP/RAS/106/GEF Country report. Bioenergy study - Bangladesh 2006 [http://www.lged-rein.org/archive_file/Bioenergy% 20Study -Bangladesh.pdf], Accessed 05 November , 2010 .
9. Dupont WD : Power calculations for matched case-control studies . Biometrics 1988 , 44 : 1157 - 1168 .
10. Lauritsen JM , Bruus M : A comprehensive tool for validated entry and documentation of data . Odense Denmark., The EpiData Association; 2003 , ( EpiData (version 3.1)) .
11. R Development Core Team ( 2009 ): R: A language and environment for Statistical computing Vienna , Austria: R Foundation for Statistical Computing, Vienna, Austria; 2010 .
12. Shalini VK , Luthra M , Srinivas L , Rao SH , Basti S , Reddy M , Balasubramanian D : Oxidative damage to the eye lens caused by cigarette smoke and fuel smoke condensates . Indian J Biochem Biophys 1994 , 31 : 261 - 266 .
13. Wegener A , Kaegler M , Stinn W : Frequency and nature of spontaneous age-related eye lesions observed in a 2-year inhalation toxicity study in rats . Ophthalmic Res 2002 , 34 : 281 - 287 .
14. Mohan M , Sperduto RD , Angra SK , Milton RC , Mathur RL , Underwood BA , Jaffery N , Pandya CB , Chhabra VK , Vajpayee RB , et al: India- US case-control study of age-related cataracts . India-US Case-Control Study Group. Arch Ophthalmol 1989 , 107 : 670 - 676 .
15. Mishra VK , Retherford RD , Smith KR : Biomass cooking fuels and prevalence of blindness in India . J Environ Med 1999 , 1 : 189 - 99 .
16. Dasgupta S , Huq M , Khaliquzzaman M , Pandey K , Wheeler D : Indoor air quality for poor families: new evidence from Bangladesh . Indoor Air 2006 , 16 : 426 - 44 .
17. Fu PQ , Kawamura K , Pavuluri CM , Swaminathan T : Interactive comment on Molecular characterization of urban organic aerosol in tropical India: contributions of biomass/biofuel burning, plastic burning, and fossil fuel combustion . Atmos Chem Phys Discuss 2009 , 9 : 21669 - 716 .
18. Hasan M , Salam A , Alam AMS : Identification and characterization to trace metals in black sold materials deposited from biomass burning at the cooking stoves in Bangladesh . Biomass & Bioenergy 2009 , 33 : 1376 - 80 .
19. Bhargava A , Khanna RN , Bhargava SK , Kumar S : Exposure risk to carcinogenic PAHs in indoor-air during biomass combustion whilst cooking in rural India . Atmospheric Environment 2004 , 38 : 4761 - 4767 .
20. Viau C , Hakizimana G , Bouchard M : Indoor exposure to polycyclic aromatic hydrocarbons and carbon monoxide in traditional houses in Burundi . Int Arch Occup Environ Health 2000 , 73 : 331 - 338 .
21. Venkataraman C , Rao GU : Emission factors of carbon monoxide and sizeresolved aerosols from biofuel combustion . Environ Sci Technol 2001 , 35 : 2100 - 2107 .
22. Mudway IS , Duggan ST , Venkataraman C , Habib G , Kelly FJ , Grigg J : Combustion of dried animal dung as biofuel results in the generation of highly redox active fine particulates . Part Fibre Toxicol 2005 , 2 : 6 .
23. Krishnaiah S , Vilas K , Shamanna BR , Rao GN , Thomas R , Balasubramanian D : Smoking and its association with cataract: results of the Andhra Pradesh eye disease study from India . Invest Ophthalmol Vis Sci 2005 , 46 : 58 - 65 .
24. Hodge WG , Whitcher JP , Satariano W : Risk factors for age-related cataracts . Epidemiol Rev 1995 , 17 : 336 - 346 .
25. Leske MC , Wu SY , Connell AM , Hyman L , Schachat AP : Lens opacities, demographic factors and nutritional supplements in the Barbados Eye Study . Int J Epidemiol 1997 , 26 : 1314 - 1322 .
26. Saha A , Kulkarni PK , Shah A , Patel M , Saiyed HN : Ocular morbidity and fuel use: an experience from India . Occup Environ Med 2005 , 62 : 66 - 69 .
27. Congdon N , Broman KW , Lai H , Munoz B , Bowie H , Gilber D , Wojciechowski R , Alston C , West SK : Nuclear cataract shows significant familial aggregation in an older population after adjustment for possible shared environmental factors . Invest Ophthalmol Vis Sci 2004 , 45 : 2182 - 2186 .
28. Christen WG , Manson JE , Seddon JM , Glynn RJ , Buring JE , Rosner B , Hennekens CH : A prospective study of cigarette smoking and risk of cataract in men . JAMA 1992 , 268 : 989 - 993 .
29. Tan JS , Wang JJ , Younan C , Cumming RG , Rochtchina E , Mitchell P : Smoking and the long-term incidence of cataract: the Blue Mountains Eye Study . Ophthalmic Epidemiol 2008 , 15 : 155 - 161 .
30. Athanasiov PA , Casson RJ , Sullivan T , Newland HS , Shein WK , Muecke JS , Selva D , Aung T : Cataract in rural Myanmar: prevalence and risk factors from the Meiktila Eye Study . Br J Ophthalmol 2008 , 92 : 1169 - 1174 .
31. Jacques PF , Hartz SC , Chylack LT , McGandy RB , Sadowski JA : Nutritional status in persons with and without senile cataract: blood vitamin and mineral levels . Am J Clin Nutr 1988 , 48 : 152 - 158 .
32. Zodpey SP , Ughade SN , Khanolkar VA , Shrikhande SN : Dehydrational crisis from severe diarrhoea and risk of age-related cataract . J Indian Med Assoc 1999 , 97 : 13 - 15 , 24 .
33. Collman GW , Shore DL , Shy CM , Checkoway H , Luria AS : Sunlight and other risk factors for cataracts: an epidemiologic study . Am J Public Health 1988 , 78 : 1459 - 1462 .
34. Bilkis AB , Samir KP , Hossain MD , Swapan K , Hopke PK : Indoor air pollution from particulate matter emissions in different households in rural areas of Bangladesh . Building and Environment 2008 , 44 : 898 - 903 .