Armo-Annor et al. BMC Nutrition            (2021) 7:50 
https://doi.org/10.1186/s40795-021-00456-w
RESEARCH ARTICLE Open Access
Risk of anaemia among women engaged in
biomass-based fish smoking as their
primary livelihood in the central region of
Ghana: a comparative cross-sectional study
Daniel Armo-Annor1, Esi K. Colecraft1* , Seth Adu-Afarwuah1, Aaron Kobina Christian2 and Andrew D. Jones3
Abstract
Background: Fish smoking using biomass fuel is an important livelihood for women living in the coastal regions of
Ghana and may contribute to anaemia risk. We assessed whether women who smoke fish as their primary
livelihood are at increased risk of anaemia compared to women in other livelihoods in the Central Region of Ghana.
Methods: We conducted a comparative cross-sectional study of 330 randomly selected adult women (18–49 years)
whose primary livelihood was either fish smoking (FSL) involving the burning of biomass fuel (n = 175) or other
livelihoods (OL) not involving burning of firewood (n = 155). Data on participants’ recent diet were collected from a
single, quantitative 24-h dietary recall and qualitative 7-day food frequency questionnaire of animal-source food
(ASF) consumption. We further assessed participants’ haemoglobin concentration using the Urit 12 Hemocue
system. We compared total iron intakes, the proportion of dietary iron from animal and plant sources, mean
haemoglobin concentrations, and anaemia prevalence between FSL and OL women.
Results: Fish was the most frequently consumed ASF by both groups of women. Although OL women consumed
more diverse ASFs in the past week compared with the FSL women (3.4 ± 1.2 vs. 2.7 ± 1.3; p < 0.001), the
contribution of ASFs to total iron intake in the past day was greater for the FSL women (49.5% vs. 44.0%; p = 0.030).
Estimated total dietary iron intake in the past day was generally low (5.2 ± 4.7 mg) and did not differ by group. The
unadjusted prevalence of anaemia was 32 and 27.1% among the FSL and OL women, respectively (p = 0.33). After
covariates adjustment, the FSL women had statistically higher anaemia prevalence (36.4% vs. 20.5%; p = 0.032) and
80% greater risk of being anemic (RR: 1.8; 95% CI: 1.1, 3.0) than the OL women.
Conclusion: Women who use biomass fuel to smoke fish as their primary livelihood had an increased risk of
anaemia. Furthermore, the average 24-h dietary iron intake among both the FSL and OL women was below their
daily iron requirement. Interventions to enhance women’s dietary iron intake and reduce their livelihood related
biomass smoke exposure may be warranted in this population.
Keywords: Anaemia, Fish smoking, Livelihood, Biomass fuel, Ghana
* Correspondence: ekcolecraft@ug.edu.gh
1Department of Nutrition and Food Science, University of Ghana, P.O. Box LG
134, Accra, Ghana
Full list of author information is available at the end of the article
© The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,
which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if
changes were made. The images or other third party material in this article are included in the article's Creative Commons
licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons
licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain
permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the
data made available in this article, unless otherwise stated in a credit line to the data.
Armo-Annor et al. BMC Nutrition            (2021) 7:50 Page 2 of 11
Background Preserving fish using smoke (most commonly from
Anaemia prevalence among Ghanaian women of repro- burning wood) is widespread among approximately 185
ductive age (WRA) remains unacceptably high at 42% [1]. fishing villages along Ghana’s coast and inland fishing
Anaemia has multiple etiologies including iron deficiency communities where women make up about 70% of the
due primarily to suboptimal dietary iron intake, infections, workforce in the post-harvest fisheries value chain [16].
worm infestations, and inherited blood disorders [2, 3]. Indeed, fish smoking using firewood is the main liveli-
The relative contribution of the various causes to the total hood of the majority of Ghanaian women living in
burden of anaemia varies by region, population group and coastal communities. Depending on the species and de-
environmental factors [3]. According to the World Health sired dryness, fish may be smoked for 2 to 18 h using
Organization [4], about 50% of all anaemia is due to iron smoking ovens that emit considerable amounts of smoke
deficiency associated primarily with inadequate iron in- [17]. Thus, it is probable that being engaged in fish
take. However, in a systematic analysis of national surveys, smoking as a livelihood may expose women to chronic
the authors reported that countries with very high burden smoke inhalation during the fish smoking process which
of inflammation (determined by an index derived from in- may increase their risk of anaemia [18]. Furthermore,
fection, hygiene, and/or overnutrition indicators) had the women who smoke fish as their primary livelihood may
lowest prevalence of anaemia among WRA [5]. Further- consume less diverse animal-source foods (ASF) because
more, anaemia attributable to iron deficiency among of greater dependence on fish and so limit their intake
WRA was considerably lower (about 16%) for countries of comparatively richer sources of iron such as livestock
with a severe public health burden of anaemia (i.e. preva- meats. Kawarazuka and Bene [19] observed that due to
lence greater than 40%) [5]. variability in the iron content of different fish species,
Smoke from burning biomass fuel is another source of heavy dependence on fish at the expense of other ASFs
inflammation which has also been implicated as a pos- may influence dietary iron intakes and predispose
sible cause of anaemia in a few studies with children and women to iron deficiency anaemia.
women. Anaemia associated with biomass smoke is be- The Central Region of Ghana, the site of this study, is
lieved to stem from altered haemoglobin metabolism among the regions of Ghana with the highest prevalence
and disruption of red bloods cells resulting from sys- of anaemia, and a large proportion of women in the area
temic inflammation and oxidative stress respectively, in- derive their livelihoods from fish smoking activities [1,
duced by pollutants in the smoke [6–9]. Kyu et al. [10] 17]. It is unclear the extent to which women fish
conducted a multilevel analysis of data from Demo- smokers consume fish and/or other ASFs and whether
graphic and Health Surveys for 29 countries and found these women have a greater risk of anaemia. The aims of
that country-level exposures to biofuel smoke were asso- this comparative cross-sectional study were to compare
ciated with up to a four-fold increased odds of anaemia among women whose primary livelihood is fish smoking
among children under five years of age [10]. Using data and those engaged in other livelihoods not involving the
from the 1998–1999 National Family Health Survey in burning of firewood: i) mean haemoglobin concentration
India, Mishra & Retherford [11] reported a significantly and prevalence of anaemia, and ii) animal-source food
higher risk of moderate-to-severe anaemia among intakes and their contribution to total iron intakes. We
preschool-aged children living in households where bio- hypothesized that women engaged in fish smoking as
fuels were used for cooking compared to those in house- their primary livelihood would have a greater risk of an-
holds using cleaner fuels [11]. However, Machisa et al. aemia possibly due to higher smoke exposure and/or
[12] did not find a significant association between house- consumption of less diverse ASFs.
hold use of biomass fuel and anaemia among preschool- Findings from this study will add to the body of evi-
aged children in Swaziland [12]. In India, pregnant dence on the association between biomass fuel use and
women in the Nagpur district whose main source of fuel anaemia among WRA and delineate whether there are
for cooking was biomass-based (i.e., wood, straw/shrubs/ unique considerations needed in efforts to address an-
grass, agricultural crop, or animal dung) had a higher aemia in WRA whose primary livelihoods expose them
adjusted relative risk of both mild and moderate-to- to chronic biomass smoke inhalation.
severe anaemia [13]. Similarly, Sukhsohale et al. [14]
found that among non-pregnant women from the same Methods
Indian district, those with the highest biomass smoke ex- Study design, setting and participants
posure index were significantly more likely to have an- This comparative cross-sectional study was carried out
aemia [14]. In contrast, blood haemoglobin in Biriwa, a fishing community of 7086 people in the
concentration did not differ between non-pregnant Gua- Mfantseman Municipality of the Central Region of
temalan women who cooked with biomass-based smoke Ghana [20]. Given its relatively large size, Biriwa was
ovens versus those who used a smokeless stove [15]. purposively selected from six fishing villages that were
Armo-Annor et al. BMC Nutrition            (2021) 7:50 Page 3 of 11
part of a larger pilot study conducted in the Central Re- questionnaire was pre-tested on women with similar
gion (the Invisible Fishers study) from May 2018 to Au- characteristics in a neighboring community before it was
gust 2019 examining the impacts of various fisheries administered to the actual study respondents. Face-to-
value chain and other behavior change interventions on face interviews were completed with participants in their
anaemia among women of reproductive age (WRA) [21]. preferred local language (Fante or Twi) or English at
The inclusion criteria for participation in the study their homes or workplaces. Data were recorded with An-
were, being an adult non-pregnant non-lactating WRA droid tablets by direct electronic data entry using the
(18 to 49 years), living in Biriwa and being willing to par- ODK.
ticipate in the study. A screening census was completed The research assistants obtained information on
to list all adult women in the community according to household characteristics, as well as personal social
their primary livelihood for the past two years; whether demographic characteristics, reproductive history,
fish smoking involving burning of firewood (FSL health, recent diet and use of firewood from the selected
women) or other livelihoods not involving burning of participant in each household (see Additional file 1).
firewood and living in a household where no one smokes Using the Urit12 HemoCue (URIT Medical Electronics
fish as a major economic activity (OL women). Due to Co., LTD, China) system, a lancet was used to prick the
concern about not achieving the estimated sample size, forefinger of each participant and a drop of blood was
no restrictions were placed on the number of women gently squeezed onto the sampling point of the system
who could be listed per household during the census. A to obtain a digital reading of the haemoglobin concen-
total of 355 eligible FSL and OL women in 311 house- tration in the sample. Anaemia was defined as having
holds were listed. Women participating in the Invisible haemoglobin concentration of less than 12 g/dl [22]
Fishers study in Biriwa (N = 10) were ineligible to par- based on one sample per participant. A one day 24-h re-
ticipate in this study. A sample size of 175 participants call method was used to record all foods and beverages
per group (i.e., FSL and OL) was estimated based on a (except water) consumed by the study participant in the
5% level of significance, 80% power, expected anaemia past 24 h [23]. Wooden food models and household
prevalence of 50% (using baseline anaemia prevalence measures were used to help participants estimate quan-
for the Central Region from the Invisible Fishers study) tities consumed. The frequency of consumption of dif-
and 35% (using prevalence of anaemia in the Upper ferent ASFs by the participant in the past week was
West Region of Ghana) where fish smoking is expected captured with an abbreviated food frequency question-
to be only minimally practiced [1] among the FSL naire listing seven categories of commonly consumed
women and OL women, respectively, and 3% contin- ASFs including fish and seafood, milk and milk products,
gency to cater for incomplete surveys. As 194 eligible livestock meats, eggs, poultry, organ meats and bush
FSL women were listed from the census, we randomly meats (see Additional file 1). This was a semi-
selected 175 to participate in the study using the RAND quantitative questionnaire which required participants to
function in excel. About 30% (n = 52) where from the specify the number of days in the past week they ate a
same households (two or three [two instances] women particular ASF without specifying the portion size. From
from the same household). The remaining 19 eligible the 24-h recall data, we determined each participant’s
women were put on a waiting list and replaced women dietary diversity score based on the 10 food groups used
who could not complete the study due to reasons such to compute the FAO’s Minimum Dietary Diversity for
as refusals, travel and relocation. The number of eligible Women indicator [24] and computed recent total iron
OL women listed (n = 161) was lower than the estimated intakes for participants. The iron content of foods con-
sample size so all of them were invited to participate in sumed in the past 24-h was determined using a food
the study. About 19% (n = 30) where from the same database (RIING food composition database, Nutrition
household (two or three [one instance] from the same Department, University of Ghana, unpublished). We es-
household). timated the bioavailability of the iron from the foods
consumed using a previously published method [25]. For
Data collection and measurement of variables each eating event, 40% of the iron content of meat, fish
Data collection took place from December 2018 to Feb- and poultry (MFP) consumed was considered as heme
ruary 2019 and was completed with semi-structured iron and available and the remaining 60% non-heme
questionnaires. Questionnaires were designed using the while 100% of iron from non-animal sources consumed
Kobotoolbox platform and loaded on Android tablets was considered non-heme. Bioavailability of non-heme
using the Open Data Kit (ODK) for data collection. iron was computed as 5, 10% or 15% of the total iron
Three research assistants with at least high school edu- content of the food source depending on the quantities
cation were recruited and trained by the primary re- of MFP and vitamin C consumed in the same eating
searcher to support the data collection activities. The event or meal [25]. The proportion of the total estimated
Armo-Annor et al. BMC Nutrition            (2021) 7:50 Page 4 of 11
bioavailable iron intake from all foods contributed by were associated with haemoglobin concentration or an-
ASFs consumed in the 24-h recall period was computed. aemia, respectively.
Additionally, we calculated ASF diversity as the number
of different categories of ASF (out of a total of seven) Results
consumed by the participant in the past seven days. Background and household characteristics of women
Completed questionnaires were reviewed for complete- A total of 355 women were invited to participate in the
ness at the end of each day. Participants with missing or study and 330 (93%) of them completed the survey. Out
incomplete responses were contacted the following day of the 25 (7%) who did not participate, 12 (48%) refused
to complete the missing information. to participate (FSL women = 10, OL women = 2), 11
(44%) were traveling out of the study area at the time of
the survey (FSL women = 7, OL women = 4) and 2 (8%)
Data analysis FSL women could not be located on the scheduled day
Data were managed, cleaned and analyzed using the or after two additional attempts to interview them. Of
Statistical Software Package for Social Sciences (SPSS) the 330 women who completed the study, 175 (53%)
version 22.0 (Chicago, USA) and SAS for Windows Re- were FSL women and 155 (47%) were OL women. The
lease 9.4 (Cary, NC, USA). One duplicate record was OL women were largely engaged in occupations such as
identified and removed during data cleaning, otherwise hairdressing, dressmaking, petty trading, fish mongering,
all completed questionnaires were included in the ana- farming, and cleaning services as their primary liveli-
lysis. Bivariate analyses using Student’s T-tests for con- hoods, with < 1% employed in professional (nursing and
tinuous variables and Chi-Square tests for categorical teaching) occupations. The FSL women smoked fish on
variables were used to summarize differences in back- average 4.9 ± 1.3 times a week for a mean of 3.9 ± 1.7 h
ground and household characteristics of the FSL and OL per day. With the dominance of fish smoking in the
women. Additionally, we compared group differences in community 59% (n = 95) of the OL women indicated
mean ASF diversity (number of different ASFs con- that they occasionally assisted relatives or neighbours
sumed in the past seven days), total iron intakes, and with fish smoking for an average of less than twice per
mean percent contribution of ASFs to total iron intakes week. Only two OL women reported having the sickle
in the past 24-h. cell trait.
Blood haemoglobin concentrations were compared be- Background characteristics of the two groups of
tween the two groups of women using a general linear women in the study are summarized in Table 1. Com-
model and ANCOVA for unadjusted and adjusted com- pared to OL women, the FSL women were on average
parisons, respectively. The SAS PROC GLIMMIX pro- older (38 vs 29 years; p < 0.001), more likely to be mar-
cedure was used in both cases. Unadjusted and adjusted ried (82% vs. 57%; p < 0.001) and to have no formal edu-
means with their 95% CIs were calculated. Anaemia cation (88% vs. 50%; p < 0.001). The FSL women also
prevalence was compared using a simple logistic regres- had smoked fish as their primary livelihood for twice the
sion model for an unadjusted comparison, and multiple number of years that the OL women had spent in their
logistic regression model for an adjusted comparison. primary livelihoods. The OL women were less likely than
The SAS PROC GLIMMIX procedure was used in both the FSL women to have a supplementary income source
cases. A binary distribution and log-link function were and significantly more of them reported that they earned
specified in the SAS procedures so that relative risks be- less than 500 Ghana Cedis (about 100 US dollars) per
tween groups and their 95% CIs were calculated. Covari- month. There were no group differences in the propor-
ates for the ANCOVA and logistic regression models tion of women who had taken deworming medication or
were selected by correlating anaemia with each covariate an iron supplement in the past three or six months, re-
so that only those independent variables significantly as- spectively. The FSL women were more likely to report
sociated with the outcome at alpha = 0.2 [26] were se- usually sleeping under a mosquito net. There were no
lected for the multiple logistic regression model. This group differences in the use of firewood for cooking or
conservative level of alpha was chosen to minimize the in living with someone who smokes cigarettes in the
risk of type II error in variable selection [27]. In addition household.
to covariates selected through correlation analysis (i.e.,
marital status, fever in the past two weeks, fish smoking Recent diet, animal-source food diversity and iron intakes
livelihood, dewormed in past 3 months, ever been preg- among women
nant and age), other covariates including number of days Starchy staples (including grains, white roots and tubers,
spent smoking fish, fuel for cooking, ASF diversity, and and plantains) were consumed by almost all the women
access to a toilet facility were selected for the final in the past day and at least 89% of women consumed
ANCOVA and multiple logistic regression models if they meat, poultry, and fish in the past day (Fig. 1). Less than
Armo-Annor et al. BMC Nutrition            (2021) 7:50 Page 5 of 11
Table 1 Background characteristics of women in fish smoking livelihoods (FSL) or other livelihoods (OL) in the study
Characteristic FSL women OL women aP-value
(n = 175) (n = 155)
Age, years 38 ± 8b 29 ± 8 < 0.001
Married 143 (82)c 89 (57) < 0.001
No Formal education 87 (50) 18 (12) < 0.001
Years in primary occupation 12 ± 7 6 ± 5 < 0.001
Has supplementary income source 78 (45) 45 (29) < 0.001
Average monthly income < GH₵ 500d 103 (59) 130 (84) < 0.001
Has been pregnant before 170 (97) 107 (69) < 0.001
Parity 5 ± 3 2 ± 2 < 0.001
Dewormed in past 3 months 47c (27) 47 (30) 0.50
Used iron supplements in past 6 months 25 (14) 27 (17) 0.44
Usually sleeps under mosquito net 106 (61) 72 (47) < 0.001
Had fever in past 2 weeks 47 (27) 29 (19) 0.08
Household size 6 ± 3 5 ± 3 < 0.001
Household has improved toilet facility 44 (25) 36 (41) < 0.001
Use firewood as main fuel for cooking 28 (16) 17 (11) 0.18
Cigarette smoker in household 9 (5) 5 (3) 0.39
aSignificance associated with independent T-test for continuous variables, and chi-squared statistic for categorical variables bMean ± SD; cn (%); dexchange rate at
time of the study I USD = 5GHC
50% of women in both groups consumed foods from the however, significantly higher for the FSL women (49.5%
other eight food groupings. Consumption prevalence for vs. 44.0%; p = 0.030). Differences in the types of ASFs
the different food groups was similar in the two groups consumed by the women in the past week are depicted
of women except for nuts and seeds where consumption in Fig. 2. Fish and seafood were the predominant ASF
prevalence was significantly higher among FSL women consumed by both groups of women in the past week.
and significantly more OL women consumed foods from Less than 50% of FSL women consumed other types of
the dairy and pulses food groups. In the past day, the ASF in the past week, whereas at least 60% of the OL
total iron intake for both groups of women from both women consumed foods made with dairy, livestock
plant and animal sources was about 5 mg (Table 2). meats and eggs in the past week in addition to consum-
Mean percent contribution of iron from ASFs was, ing foods from the fish and seafood group. Consumption
Fig. 1 Food groups consumed by women in the study in the past 24 h
Armo-Annor et al. BMC Nutrition            (2021) 7:50 Page 6 of 11
Table 2 Quantities of Animal Source Foods (ASF), total iron intake and contribution of ASF to total iron intakes in the past day
Variable aFSL women bOL women Mann-Whitney U cP-value
(n = 175) (n = 155)
Fish and shell fishes (g) 90.8d 78.0 10,077.5 0.68
Livestock meat (g) 11.0 19.0 404.0 0.15
Milk and milk products (g) 30.0 32.0 376.0 0.67
Eggs (g) 124.4 100.0 172.0 0.15
Poultry (g) 78.0 78.0 110.5 0.88
Total iron intake from all sources (mg) 5.3 ± 5.2e 5.0 ± 4.2 – 0.57
Mean percent contribution of iron from ASF 49.5 ± 22.2 44.0 ± 23.9 – 0.030
aFSL = Fish Smoking Livelihood bOL = Other Livelihood cP-value associated with Mann-Whitney U test or independent T-test, dMedian, eMean ± SD
of poultry and organ meats was also higher among OL 0.33). After controlling for covariates, livelihood type
women. Organ meats and bush meats (game) were only was associated with haemoglobin concentration and the
minimally consumed. Mean ASF diversity was signifi- prevalence of anaemia. That is, the FSL women had sig-
cantly higher among the OL women (3.4 ± 1.2 vs. 2.7 ± nificantly lower mean ± SD haemoglobin concentration
1.3; p < 0.001). However, with the exception of fish and (12.2 ± 0.2 g/dl vs. 12.8 ± 0.2 g/dl; p = 0.018) (Table 4),
seafood where the mean frequency of consumption in and higher prevalence of anaemia (36.4% vs. 20.5%; p =
the last week was at least 12 times, the mean frequency 0.032) than the OL women. The risk of anaemia was
of consuming the other ASF groups averaged about two 80% greater in the FSL women compared with the OL
or fewer times in the past week (Table 3). women (Table 5).
Unadjusted and adjusted differences in haemoglobin Discussion
concentration and anaemia prevalence between FSL The study examined whether being engaged in fish
women and OL women smoking as a primary livelihood is associated with a
In unadjusted analyses, mean ± SD haemoglobin concen- higher risk of having anaemia among Ghanaian women
tration did not differ between the two groups of women compared to being engaged in a primary livelihood that
(FSL, 12.3 ± 1.9 g/dl; OL, 12.6 ± 1.8 g/dl; p = 0.17). About does not involve burning of biomass smoke. Regardless
one-third of all the study women had mild (13.6%), of livelihood type, anaemia prevalence among women in
moderate (13%) or severe (3%) anaemia (Fig. 3) and the study was high although lower than the national
there was no group difference in the prevalence of any average for rural communities (43%) and for the Central
anaemia among the women (FSL, 32%; OL, 27%; p = Region of Ghana (48%) [1]. The results show that
Fig. 2 Types of Animal Source Foods (ASF) consumed in the past 7-days by women in the study
Armo-Annor et al. BMC Nutrition            (2021) 7:50 Page 7 of 11
Table 3 Differences in mean ± SD frequency of consuming different Animal Source Foods (ASF) in the past week
ASF group FSLa women (n = 175) OLb women (n = 155) cp-value
Fish and seafood 14.2 ± 3.8 12.4 ± 4.7 < 0.001
Milk and milk products 1.2 ± 2.2 1.9 ± 2.1 < 0.001
Livestock meats 1.1 ± 1.5 1.3 ± 1.6 0.09
Eggs 0.9 ± 1.4 1.2 ± 1.4 0.09
Poultry 0.5 ± 0.9 0.9 ± 1.5 < 0.001
Organ meats < 0.1 ± 0.2 0.1 ± 0.3 0.09
Bush meats < 0.1 ± 0.1 0.1 ± 0.2 0.36
aFish Smoking Livelihood bOther Livelihoods cP-value associated independent T-test
women who burn biomass fuel to smoke fish as their deficiency anaemia, but unlike iron-deficiency, does not
primary livelihood had lower blood haemoglobin con- depress iron stores [30].
centrations and a higher burden of anaemia than those In previous studies linking anaemia with smoke expos-
whose primary livelihoods did not involve burning of ure, the source of smoke exposure was household use of
biofuels. Our findings corroborate those from several biomass fuel for cooking indoors. However, in the
previous epidemiological studies that have reported an present study the prevalence of using biomass fuel for
association between use of biomass fuel and increased household cooking was similar for the two groups of
likelihood of anaemia among preschool-aged children women and was not significantly associated with an-
and women [10, 11, 13, 14, 28]. The biological mecha- aemia prevalence. In the study community, fish smoking
nisms by which exposure to biomass fuel lead to lower is largely done outdoors although some of the women
haemoglobin concentrations and higher anaemia risk are smoked fish in semi-enclosed structures. Our results
believed to involve cytokine-mediated inflammation in suggest that exposure to outdoor smoke from fish smok-
response to pollutants in biomass smoke such as carbon ing among the FSL women might have had a similar im-
monoxide, transition metals, and particulate matter less pact on haemoglobin concentration and anaemia
than 10 μm (PM10) and less than 2.5 μm (PM2.5) in prevalence as did exposure to indoor biomass fuel in
diameter [13, 29]. Anaemia caused by systemic inflam- those previous studies. In fact outdoor exposure has also
mation impairs iron homeostasis and red blood cell syn- been shown to be associated with anaemia in children.
thesis leading to low serum iron levels similar to iron- For example, Morales-Ancajima et al. [31], observed that
Fig. 3 Prevalence of anaemia among women in fish smoking livelihoods and other livelihoods in the study
Armo-Annor et al. BMC Nutrition            (2021) 7:50 Page 8 of 11
Table 4 Unadjusted and Adjusted differences in mean haemoglobin concentration between women in the study
FSL womena (n = 175) OL womenb Difference in means (95% CI)c dP-value
(n = 155)
Haemoglobin concentration (g/dL)
Unadjusted 12.3 ± 1.9 12.6 ± 1.8 −0.3 (−0.7, 0.1) 0.17
Adjustede 12.2 ± 0.2 12.8 ± 0.2 −0.6 (−1.2, − 0.1) 0.018
a Women engaged in Fish smoking livelihood. b Women engaged in other livelihoods unrelated to fish smoking c95% CI = 95% Confidence Interval dP-values are
based on general linear regression model (unadjusted) and ANCOVA (adjusted) for haemoglobin concentration eAdjusted for marital status, toilet facility, fever,
fuel used for cooking, cigarette smoker in household, exposed to cigarette smoke outside the home at least once in past 12months, number of days spent
smoking fish in a week, and Animal Source Food (ASF) diversity in past 7-days
Peruvian children living in areas of Lima with higher for covariates, a fish smoking livelihood was associated
outdoor concentrations of the biomass smoke pollutant with a lower haemoglobin concentration and anaemia
particulate matter (PM2.5) had lower mean haemoglobin prevalence. Number of days of fish smoking may not ne-
concentrations and higher anaemia prevalence than cessarily reflect intensity of smoke exposure experienced
those living in areas with lower concentrations of the by women as, depending on fish species and the type
pollutant [31]. Honda et al. [32] reported that PM2.5 was and stage of fish smoking, the number of actual hours
responsible for a 0.81 g/dL decrease in average haemo- spent near the smoking stove may differ widely [17].
globin among old American adults [32]. Furthermore, there may have been limited variability in
Spending more days smoking fish was not associated the women’s fish smoking enterprises thus limiting dif-
with a higher risk of anaemia, though after controlling ferences in the types of fish being smoked and in
Table 5 Unadjusted and adjusted odds ratios for having anaemia among women in the study
Variable (referent) Unadjusted Adjusted
Relative Risk (95% Cl)b aP-value Relative Risk (95% Cl)b aP-value
Livelihood type
FSL (OL)c 1.2 (0.8, 1.7) 0.331 1.8 (1.1, 3.0) 0.032
Age
< 35 years (≥35 years) 1.3 (0.9, 2.0) 0.17
Marital status
Married (Single) 0.9 (0.6, 1.3) 0.55
Ever been pregnant
Yes (No) 0.8 (0.4, 1.3) 0.33
Dewormed in past 3mos
No (Yes) 0.8 (0.6, 1.2) 0.30
Fever in past 2 weeks
Yes (No) 1.2 (0.8, 1.7) 0.47
Household with toilet facility
No (Yes) 0.7 (0.5, 1.1) 0.09
Cooking fuel
Biomass fuel (other fuel) 1.2 (0.8, 2.0) 0.38
number of days spent smoking fish
More days (Fewer days) 0.7 (0.5, 1.2) 0.20
ASF diversity
Low (High) 1.4 (1.0, 2.1) 0.08
% of iron from ASF
Low (High) 1.2 (0.9, 1.7) 0.22
a P-value is based on logistic regression: Dependent variable = Anaemia (Present/Not Present)
b 95% CI = 95% Confidence Interval cFish smoking Livelihood (Other Livelihoods)
Armo-Annor et al. BMC Nutrition            (2021) 7:50 Page 9 of 11
smoking method. Alternatively, the lack of association inflammation from infections is reportedly common
between number of fish smoking days and anaemia may among developing country populations [40, 41]. Efforts
mean that the differences observed by livelihood type to enhance the women’s dietary diversity and consump-
may not be due entirely to smoke exposure from smok- tion of iron-rich foods are also warranted in the
ing fish. Given the multiplicity of anaemia causes even community.
within individuals, one risk factor such as smoke expos-
ure is unlikely to explain the total burden of anaemia in
an individual or population [33]. In addition to biomass Limitations
smoke exposure being a potential cause of anaemia, an- Our study has some limitations. First the two groups of
aemia may be caused by various factors associated with women were from the same community where fish
nutritional deficiencies, infection and infestations and smoking is the dominant livelihood for women so that
genetic haemoglobin disorders [34]. Hence, being en- we did not have a true ‘unexposed’ group because: 1)
gaged in fish smoking as a primary livelihood may pre- more than 50% of women classified as being in primary
dispose women to other anaemia risk factors that were livelihoods unrelated to fish smoking occasionally
not measured in this study. The use of biomass fuel has assisted with smoking activities of families and friends;
been linked to certain infectious diseases such as tuber- and, 2) fish smoking was generally an outdoor activity
culosis that are also associated with increased risk of an- indicating community-wide exposure to smoke. This
aemia [13, 35]. may have underestimated the effect observed. Further,
Fish was the predominant ASF in the diets of both smoke exposure was not directly measured and so we
groups of women. This was expected as fish is the cannot be truly certain whether the association observed
cheapest animal protein source across Ghana and con- was due to smoke exposure per se or other unmeasured
tributes to more than 50% of total animal protein intake characteristics that the women had. Iron intake was
[36, 37]. The two groups of women had similar 24-h based on one 24-h recall which may not necessarily re-
total iron intakes from all sources and the contribution flect usual intakes among the participants.
of iron from ASF was higher for the FSL women who
had higher mean frequency of fish consumption in the
Conclusion
past week. While the prevalence of consuming other
The adverse effects of biomass smoke exposure on mor-
ASF besides fish was higher for OL women, mean con-
bidity and mortality associated with respiratory and car-
sumption frequency for iron-rich ASF such as livestock
diovascular conditions and cancers in developing
meats, organ meats, and poultry was low and not signifi-
country populations has been long recognized, however
cantly different between the two groups. This suggests
the association with anaemia has received less attention.
that fish was the main contributor to total iron intakes
Our study suggests that women who smoke fish as their
for both groups of women. This is plausible as the fish
primary livelihood may have additional risks of anaemia
commonly available in the study community were the
besides those posed by other well recognized contribu-
smaller fish species such as anchovy, sardinella, and her-
tors to anaemia such as poor diet, and infections. This
rings that tend to be richer sources of iron than the lar-
added risk may be especially important for reducing
ger species because they are eaten whole with bones,
population-level anaemia risk given how ubiquitous fish
head and viscera which have the highest concentration
smoking activities are across Ghana and throughout
of iron and other micronutrients [19]. Cooking methods
many communities in West Africa as described previ-
may also affect the iron content of animal protein
ously. The study also showed that the women’s iron in-
sources. For example, Pourkhalili et al. [38] reported
take in the past 24-h was well below the required iron
higher iron retention in grilled lamb meat compared to
intake with probability the women had underlying iron
when the meat is boiled [38]. Total iron intake for both
deficiency. Therefore, efforts to reduce smoke emissions
groups of women in the past day was just around 5mg
from fish smoking ovens and to enhance dietary iron in-
which falls substantially short of the recommended daily
take may be warranted in efforts to address different
intake of 19.6 for adult women assuming dietary iron
causes of anaemia in our study community and beyond
bioavailability of 15%. At even lower dietary bioavailabil-
given the potential for impact at scale. Further research
ity, levels expected for predominantly plant-based diets,
is needed to understand the mechanisms of the associ-
the recommended daily intakes are even higher (62 mg,
ation and to determine causality.
31 mg, and 25.8 mg for 5, 10 and 12% bioavailability, re-
spectively) [39]. In addition to smoke exposure, nutri-
tional iron inadequacy is probably an important Abbreviations
ASF: Animal Source Food; FAO: Food and Agriculture Organization; FSL: Fish
contributor to anaemia in the study population. Coexist- Smoking Livelihood; OL: Other Livelihood PM:Particulate Matter; WHO: World
ence of iron-deficiency anaemia and anaemia of Health Organization; WRA: Women of Reproductive Age
Armo-Annor et al. BMC Nutrition            (2021) 7:50 Page 10 of 11
Supplementary Information anemia burden from 1990 to 2010. Blood. 2014;123(5):615–25. https://doi.
The online version contains supplementary material available at https://doi. org/10.1182/blood-2013-06-508325.
org/10.1186/s40795-021-00456-w. 4. WHO. The global prevalence of anaemia in 2011. WHO Rep [Internet]. 2015;
48 [cited 2020 Feb 26]. Available from: http://apps.who.int/iris/bitstream/1
Additional file 1. Questionnaire on the risk of anaemia among women 0665/177094/1/9789241564960_eng.pdf?ua=1.
engaged in biomass-based fish smoking as their primary livelihood in the 5. Petry N, Olofin I, Hurrell RF, Boy E, Wirth JP, Moursi M, et al. The proportion
Central Region of Ghana. The word file contains the questionnaire used of anemia associated with iron deficiency in low, medium, and high human
to gather information on the personal social demographic characteristics, development index countries: a systematic analysis of national surveys.
reproductive history, health and use of firewood of the study participants Nutrients. 2016;8(11):1–17.
and another section showing the modified food frequency questionnaire. 6. Bissinger R, Bhuyan M, Qadri SM, Lang F. Oxidative stress , eryptosis and
anemia : a pivotal mechanistic nexus in systemic diseases. FEBS J 2019;
286(5):826–854, https://doi.org/10.1111/febs.14606.
Acknowledgements 7. Naeher LP, Brauer M, Lipsett M, Zelikoff JT, Simpson CD, Koenig JQ, et al.
We appreciate the contributions of Claudia Ewa, Griffith Ammah, Princess Woodsmoke health effects: a review. Inhal Toxicol. 2007;19(1):67–106.
Abakah and Eric Arthur to the completion of this study and publication. We https://doi.org/10.1080/08958370600985875.
are also grateful to all the women who took part in the study. 8. Padhy P, Padhi B. Effects of biomass combustion smoke on hemotological
and antioxidant provile among children (8-13 years) in India. Inhal Toxicol.
Authors’ contributions 2009;21(8):705–11. https://doi.org/10.1080/08958370802448961.
DAA, EKC, SAA and AKC designed the research questionnaire. DAA 9. Weiss G, Goodnough LT. Anemia of chronic disease. N Engl J Med. 2005;
conducted the data collection. EKC, SAA, DAA and AKC wrote the methods 352(10):1011–23. https://doi.org/10.1056/NEJMra041809.
section. DAA, SAA and AKC performed statistical analysis. EKC and DAA 10. Kyu HH, Georgiades K, Boyle MH. Biofuel smoke and child Anemia in 29
discussed the results. EKC, DAA, SAA, AKC and ADJ critically reviewed and developing countries: a multilevel analysis. Ann Epidemiol. 2010;20(11):811–
edited the manuscript. All authors read and approved the final manuscript. 7. https://doi.org/10.1016/j.annepidem.2010.07.096.
11. Mishra V, Retherford R. Does biofuel smoke contribute to anemia and
Funding stunting in early childhood. Int J Epidemiol. 2006;36(1):1–33.
This research was supported with a grant from the Bill & Melinda Gates 12. Machisa M, Wichmann J, Nyasulu PS. Biomass fuel use for household
Foundation (OPP1182940). The funders did not participate in the design of cooking in Swaziland: is there an association with anaemia and stunting in
the study, data collection, analysis nor interpretation of data in the writing of children aged 6-36 months? Trans R Soc Trop Med Hyg. 2013;107(9):535–44.
this manuscript. All the information and opinions expressed in the https://doi.org/10.1093/trstmh/trt055.
manuscripts are based solely on the authors’ research and interpretation of 13. Page CM, Patel A, Hibberd PL. Does smoke from biomass fuel contribute to
the research. Anemia in pregnant women in Nagpur, India? A cross-sectional study. PLoS
One. 2015;10(5):1–13.
14. Sukhsohale N, Narlawar U, Phatak M. Indoor air pollution from biomass
Availability of data and materials
combustion and its adverse health effects in Central India: an exposure-
The datasets used or analysed during the current study are available from
response study. Indian J Community Med Off Publ Indian Assoc Prev Soc
the corresponding author on reasonable request.
Med. 2013;38(3):162–7.
15. Neufeld LM, Haas JD, Ruel MT, Grajeda R, Naeher LP. Smoky indoor cooking
Declarations fires are associated with elevated hemoglobin concentration in iron-
deficient women. Pan Am J Public Heal [Internet]. 2004;15(2):110–8.
Ethics approval and consent to participate Available from: http://www.scielosp.org/pdf/rpsp/v15n2/20820.pdf. https://
Ethical approval for the study was obtained from the Ethics Committee for doi.org/10.1590/S1020-49892004000200006.
the College of Basic and Applied Science (ECBAS) of the University of Ghana 16. FAO. Fishery and aquaculture country profiles (Ghana). [Internet]. FAO
(approval reference: ECBAS 003/18–19). The study was explained to all Fisheries and Aquaculture Department. 2016 [cited 2019 Jun 26]. Available
eligible participants at the time of recruitment and each participant signed from: http://www.fao.org/fishery/facp/NGA/en.
or if illiterate, thumb-printed an informed consent document before being 17. Sakyi M, Jia C, Ampofo-Yeboah A, Aglago A. Fish smoking in Ghana: a
enrolled in the study. review. J Fish Sci. 2019;13(3):13–24.
18. Dadzie EK, Ephraim RKD, Afrifa J, Quaicoe R, Bediako BS, Duku BI, et al.
Consent for publication Persistent exposure to wood smoke is associated with variations in
Not applicable. biochemical and hematological indices among regular wood burners in the
Cape Coast metropolis, Ghana. Sci African. 2019;4:e00100. https://doi.org/1
Competing interests 0.1016/j.sciaf.2019.e00100.
The authors declare that they have no competing interests. 19. Kawarazuka N, Bene C. The potential role of small fish species in improving
micronutrient deficiencies in developing countries : building evidence.
Author details Public Health Nutr. 2011;14(11):1927–38. https://doi.org/10.1017/S13
1Department of Nutrition and Food Science, University of Ghana, P.O. Box LG 68980011000814.
134, Accra, Ghana. 2Regional Institute for Population Studies, University of 20. Ghana Statistical Service. 2010 Population and housing census. Ghana Stat Serv
Ghana, P.O. Box LG 96, Accra, Ghana. 3School of Public Health, University of [Internet]. 2012;1–117 [cited 2018 Sep 12]. Available from: http://www.statsgha
Michigan, Ann Arbor, MI 48109-2029, USA. na.gov.gh/docfiles/2010phc/Census2010_Summary_report_of_final_results.pdf.
21. Invisible Fishers. The invisible fishers - full text view - ClinicalTrials.gov
Received: 26 August 2020 Accepted: 7 July 2021 [Internet]. 2018 [cited 2020 Jun 12]. Available from: https://clinicaltrials.gov/
ct2/show/study/NCT03498755.
22. WHO. Worldwide prevalence of anaemia, WHO Vitamin and Mineral
References Nutrition Information System, 1993–2005. Public Heal Nutr. 2009;12(4):444–
1. GDHS. Ghana demographic health survey [Internet]. [GSS, GHS, AND ICF 54 Available from: http://whqlibdoc.who.int/publications/2008/9789241
INTERNATIONAL]. 2014 [cited 2019 Apr 8]. Available from: https://dhsprogra 596657_eng.pdf.
m.com/pubs/pdf/FR307/FR307.pdf. 23. FAO. Dietary assessment: a resource guide to method selection and
2. CDC. Hemoglobinopathies current practices for screening, confirmation and application in low resource settings [Internet]. 2018 [cited 2020 Jul 9]. p.
follow up [Internet]. 2015 [cited 2019 Jun 7]. Available from: https://www.cdc. 152. Available from: http://www.fao.org/3/i9940en/I9940EN.pdf.
gov/ncbddd/sicklecell/documents/nbs_hemoglobinopathy-testing_122015.pdf. 24. FAO. Minimum dietary diversity for women a guide to measurement
3. Kassebaum NJ, Jasrasaria R, Naghavi M, Wulf SK, Johns N, Lozano R, et al. [Internet]. 2016 [cited 2019 Jun 26]. Available from: http://www.fao.org/3/a-
RED CELLS , IRON , AND ERYTHROPOIESIS A systematic analysis of global i5486e.pdf.
Armo-Annor et al. BMC Nutrition            (2021) 7:50 Page 11 of 11
25. Tseng M, Chakraborty H, Robinson DT, Mendez M, Kohlmeier L. Adjustment
of iron intake for dietary enhancers and inhibitors in population studies :
bioavailable iron in rural and urban residing Russian women and children. J
Nutr. 1997;127(8):1456–68.
26. De Freitas JS, Costa PS, Costa LR, Naghettini AV. Evaluation of clinical and
laboratory variables associated with anemia in pediatric patients on
hemodialysis. J Pediatr (Rio J). 2015;91(1):87–92. https://doi.org/10.1016/j.
jped.2014.05.009.
27. Mickey RM, Greenland S. The impact of confounder selection criteria on
effect estimation. Am J Epidemiol. 1989;129(1):125–37. https://doi.org/10.1
093/oxfordjournals.aje.a115101.
28. Accinelli RA, Leon-Abarca JA. Solid fuel use is associated with anemia in
children. Environ Res. 2017;158(January):431–5 Available from: http://www.
ncbi.nlm.nih.gov/pubmed/28689034.
29. Bishwajit G, Yaya S, Tang S, Hossain A, Fan Y, Akter M, et al. Association of
Living Arrangement Conditions and Socioeconomic Differentials with
Anemia status among women in rural Bangladesh. Biomed Res Int. 2016;
2016:1–6. https://doi.org/10.1155/2016/4571686.
30. Dutta A, Ray M, Banerjee A. Systemic inflammatory changes and increased
oxidative stress in rural Indian women cooking with biomass fuels. Toxicol
Appl Pharmacol. 2012;261(3):255–62. https://doi.org/10.1016/j.taap.2012.04.
004.
31. Morales-ancajima VC, Tapia V, Vu BN, Liu Y, Alarc DE, Gonzales GF. Increased
outdoor PM 2 . 5 concentration is associated with moderate / severe
anemia in children aged 6–59 months in Lima, Peru. Journal environmental
and public health. 2019;2019.
32. Honda T, Pun V, Manjourides J, Helen S. Anemia prevalence and
hemoglobin levels are associated with long-term exposure to air pollution
in an older population. Environ Int. 2017;101(2):125–32 Available from:
http://www.embase.com/search/results?subaction=viewrecord&from=
export&id=L614272295%0Ahttp://dx.doi.org/10.1016/j.envint.2017.01.017.
33. Chaparro CM, Suchdev PS. Anemia epidemiology, pathophysiology, and
etiology in low- and middle-income countries. Ann N Y Acad Sci. 2019;
1450(1):15–31. https://doi.org/10.1111/nyas.14092.
34. Balarajan Y, Ramakrishnan U, Özaltin E, Shankar AH, Subramanian SV.
Anaemia in low-income and middle-income countries. Lancet. 2011;
378(9809):2123–35. https://doi.org/10.1016/S0140-6736(10)62304-5.
35. Gil-Santana L, Cruz LAB, Arriaga MB, Miranda PFC, Fukutani KF, Silveira-
Mattos PS, et al. Tuberculosis-associated anemia is linked to a distinct
inflammatory profile that persists after initiation of antitubercular therapy.
Sci Rep. 2019;9(1):1–8.
36. FAO. The state of world fisheries and aquaculture. Meeting the sustainable
development goals. [Internet]. Rome; 2018 [cited 2019 May 28]. p. 210.
Available from: http://www.fao.org/3/i9540en/I9540EN.pdf.
37. GAIN. Ghana fish and seafood report. Global Agricultural Information
Network [internet]. 2019;1–15 [cited 2020 May 28]. Available from: http://a
griexchange.apeda.gov.in/MarketReport/Reports/Fish_and_Seafood_
Report_Accra_Ghana_3-8-2019.pdf.
38. Pourkhalili A, Mirlohi M, Rahimi E. Heme iron content in lamb meat is
differentially altered upon boiling, grilling, or frying as assessed by four
distinct analytical methods. Sci World J. 2013;2013:1–5. https://doi.org/10.11
55/2013/374030.
39. WHO. Vitamin and mineral requirements in human nutrition second edition.
[Internet]. 2nd edition. World Health Organization, 2004 [cited 2020 May 28]
p 342. Aailable from: https://apps.who.int/iris/rest/bitstreams/50670/retrieve.
40. Ganz T. Anemia of inflammation. N Engl J Med. 2019;381(12):1148–57.
https://doi.org/10.1056/NEJMra1804281.
41. Shaw JG, Friedman JF. Iron deficiency Anemia : focus on infectious diseases
in lesser developed countries. Hindawi Publ Corp. 2011;2011:1–6.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.