Prevalence of anaemia in pregnancy and associated factors in northern Uganda: a cross-sectional study

Background: Anaemia in pregnancy is associated with poor maternal and foetal outcomes. Nonetheless, there is a paucity of recent literature on the predictors of anaemia during pregnancy in the context of northern Uganda, a region emerging out of decades of war. A study was undertaken to determine the prevalence and factors associated with anaemia among pregnant women in northern Uganda. Methods: In this cross-sectional study, 320 pregnant women seeking care at Lira Regional Referral Hospital were consecutively enrolled. Data were collected using a structured interviewer-administered questionnaire. Data collected included: demographic, obstetric, nutritional and dietary characteristics of study participants. Data analysis consisted of descriptive statistics, cross-tabulations and logistic regression with 95% confidence and a p-value of < 0.05 as significant using STATA version 14. Results: The mean age of the women was 25.3 ± 5.6 years while their mean gestational age was 25.4 ± 7.8 weeks. The overall prevalence of anaemia (Hb < 11 g/dl in the first and third trimesters and less than 10.5 g/dl in the second trimester) was 24.7%. Iron deficiency was prevalent in half of the women (50%) with anaemia. Factors independently associated with anaemia included taking antimalarial prophylaxis (AOR 0.44; 95% CI 0.19, 0.99) and consumption of legumes and cereals more than twice in the previous week (AOR 0.46; 95% CI 0.24, 0.89). Conclusion: One-quarter of pregnant women in this study population based in northern Uganda were anaemic. There is a need to strengthen interventions to control anaemia during pregnancy, particularly the intake of antimalarial prophylaxis and consumption of iron-rich locally available foods.


Introduction
Anaemia in pregnancy, defined as haemoglobin (Hb) concentration levels of less than 11 g/dl in the first and third trimesters and 10.5 g/dl in the second trimester, remains a public health problem. 1,2The global prevalence of anaemia during pregnancy is around 38%, which translates to 32 million anaemic pregnant women. 3The burden of anaemia during pregnancy is disproportionately higher in low-to middle-income countries with prevalence estimates of 24% to 56% compared with 22% in high-income countries. 3A recent systematic review estimates the burden of anaemia during pregnancy in Uganda to range from 23% to 37%. 4 According to Milman, 2 40% of global maternal deaths are attributable to maternal anaemia while 17% of low birthweight, preterm birth and perinatal mortality in low-and middleincome countries are attributed to anaemia during pregnancy. 1he heightened iron requirements that predispose pregnant women to anaemia are due to the expansion of maternal red cell mass, growth of the placenta and foetus development, which increases with increasing gestational age. 5,6The increased iron demands in pregnancy are exacerbated by food taboos, poor dietary habits, parasitic infestation, infections (especially malaria) and HIV/AIDS, and pre-existing anaemia before pregnancy. 7,8e mainstay strategy to control and treat anaemia during pregnancy has been supplementation with iron, coupled with routine screening for anaemia during antenatal care. 9ther broad strategies to control anaemia include prevention and control of malarial infection with intermittent preventive treatment (IPT) and use of insecticide-treated mosquito nets, treatment and control of helminth infestations, and promotion of diet diversification to increase dietary iron intake. 8,10[13] Hitherto, limited prospective studies have examined how effective these interventions have been in reducing anaemia during pregnancy in northern Uganda, partly because of the high costs associated with measuring anaemia, especially for big sample sizes. 14,15In addition, the contextual factors responsible for the preponderance of anaemia in the area remain widely unstudied.Current information on the burden and determinants of anaemia is vital in the design and evaluation of policies and strategies aimed at reducing the burden of anaemia in pregnancy in northern Uganda.Therefore, we aimed to assess the prevalence and factors associated with anaemia among pregnant women seeking antenatal care in a public referral hospital in northern Uganda.

Study design and setting
We conducted an analytical cross-sectional study between December 2017 and March 2018.The study was conducted in the antenatal clinic of Lira Regional Referral Hospital (LRRH).The hospital is one of the 14 regional referral hospitals in Uganda and is located in Lira City, which is proximately 340 kilometres by road to the north of Kampala, the capital city of Uganda.LRRH is found in the Lango sub-region, one of the four sub-regions in northern Uganda.The hospital has a bed capacity of approximately 350 and serves all 8 districts in the Lango sub-region.According to the clinic register, the monthly antenatal clinic attendance averages 640 pregnant women and it is run by at least two midwives each day.The facility was chosen because of the high patient load and the fact that it had a wide catchment area which would make our findings more generalisable.

Study participants
The study was conducted among pregnant women attending the antenatal clinic at LRRH.Confirmation of pregnancy was based on the antenatal card record of the woman.All pregnant women attending the antenatal clinic at the hospital were eligible to participate in the study irrespective of gestational age, gravidity or parity.Pregnant women with known haemoglobin disorders and those who had malarial infection at the time of data collection were excluded from the study.

Sample size and sampling procedure
The sample size was calculated using a single population proportion formula 16 with a z-score (z) of 1.96, a margin of error (e) of 0.05, and a prevalence (p) for the known population of 0.291. 17The calculated sample size was 349 plus a 10% allowance for non-response.Eligible participants were enrolled in the study using a consecutive sampling technique.The investigators enrolled the participants consecutively as the women came in for their ANC visits daily during the study period until the desired sample size was realised.This sampling technique was chosen because of time constraints given the big sample size.

Data-collection method, tool and procedures
We used a structured interviewer-administered questionnaire to collect data with the help of two trained research assistants.The questionnaire was developed from a literature review of previous studies. 8,10The questionnaire was reviewed, pre-tested and adjusted accordingly before data collection.The questionnaire was developed in English, translated to Lango, the local language, and then back-translated to English by two different language experts who speak English and Lango fluently.The questionnaire collected information from the participants on the following: demographic characteristics, obstetric history, nutritional status and dietary patterns, and anaemia prevention-related factors (iron supplementation, deworming, intake of antimalarial prophylaxis and intake of iron-rich foods).Data collection involved administration of the questionnaire and taking blood from the study participants for assessment of Hb and serum ferritin levels.The overall process of data collection lasted for 20-30 minutes.

Sample collection and processing
A venepuncture at the antecubital fossa was done to collect 5 ml of blood in each of two vacutainer tubes for Hb and serum ferritin analysis.The tubes were clearly labelled using a code that matched the participants' completed questionnaire.Samples were maintained at +4 0 C and transported to Mulago National Referral Hospital laboratory for analysis.Hb and serum ferritin concentrations were assessed using the Sysmex XS-1000i Automated Hematology Analyzer (Sysmex Corporation, Kobe, Japan) and Cobas 6000 automated analyser respectively (Roche Diagnostics, Indianapolis, IN, USA).Hb concentration was adjusted for altitude by subtracting 0.2 g/dl from the measured participant concentrations for defining and classifying anaemia based on the recommendation by the World Health Organization for study settings such as Lira City with altitudes between 1 000 and 1 500 metres above sea level. 18

Study variables and measures
The outcome variable was anaemia, defined as Hb less than 11 g/dl in the first and third trimesters and less than 10.5 g/dl in the second trimester. 18Anaemia was classified as: severe (Hb < 7.0 g/dl), moderate (Hb 7.0-9.9g/dl) or mild (Hb 10-10.9g/d). 18The independent variables included: age, marital status, level of education, occupation, level of income, gravidity, number of children, birth interval, gestational age, number of ANC visits, timing of the ANC visit, serum ferritin (sF) concentration levels, mid-upper-arm circumference (MUAC), frequency of eating meat and poultry, legumes and cereals, and vegetables and fruits, iron supplementation, intake of anti-malarial prophylaxis, intake of dewormers and food avoidance.
Iron deficiency (ID) as determined by sF concentration was classified as category I (sF <15 μg/l), which denotes ID that requires active intervention, category II (sF 15-29.9μg/l), which still signifies low iron stores, and category III (sF ≥30 μg/l), which indicates a normal iron-store level. 19Adherence to iron supplements was defined as a self-reported intake of iron pills on at least four days in the previous week.This was used as a proxy of how well the women were taking the total routine daily prescribed iron pills.Undernutrition status was defined as MUAC < 23.0 cm using an adult MUAC tape for pregnant women. 20The food frequency questionnaire 21 was adapted to measure the frequency of consumption of ironrich foods by participants.The frequency of intake of iron-rich foods was then defined as consumption of meat and poultry, legumes and cereals, vegetables and fruits at least twice or more than twice a week.

Data management and analysis
Every questionnaire was checked for completeness at the end of each interview.Data were entered in SPSS version 25 (IBM Corp, Armonk, NY, USA) and exported to STATA version 14 (StataCorp, College Station, TX, USA) for analysis.Data were scanned for out-of-range and missing values before commencing data analysis.Categorical variables were summarised as proportions and continuous variables as means with their standard deviations as appropriate.We used logistic regression to test for association between the dependent and independent variables at a 95% level of confidence.Based on the scientific literature and biological plausibility, we included the following factors in our multivariate logistic model (Table 3): maternal age, 22 gravity, 23 parity, 23,24 gestational age, 23,24 number of ANC visits, 25 intake of antimalarial 26 and iron supplementation, 11,27,28 deworming 29 and MUAC. 23All the variables in the model were assessed for collinearity, which was considered present if the variables had a variance inflation factor (VIF) of > 10.In situations of collinearity, we retained the variable with greater biological plausibility.

Ethical considerations
The research protocol was reviewed and approved by the Makerere University School of Health Sciences Research and Ethics Committee (SHSREC REF: 2017-061).Administrative clearance was also obtained from the office of the District Health Officer, Lira District, and Lira Regional Referral Hospital administration.Written informed consent was obtained from participants.Pregnant women who were below 18 years of age were regarded as emancipated minors.Privacy and confidentiality were maintained throughout the study by conducting the interviews in a private place, the use of codes instead of participants' names, and password protection of entered data.Pregnant women found to have low Hb and serum ferritin concentrations were contacted and advised to seek treatment from LRRH or any other nearby health facility.

Results
Out of the 349 participants approached, 320 accepted to take part in the study, giving a response rate of 91.7%.

Demographic and obstetric characteristics of study participants
The mean age of the study participants was 25.3 ± 5.6 years of age while the mean gestational age of the study participants was 25.4 ± 7.8 weeks.The majority of the study participants (51.2%) were young adults and had completed a primary level of education (52%) (Table 1).Most of the participants were multigravida (72.5%) and in either the second or third trimester (92.5%) (Table 1).

Prevalence of anaemia among study participants
The prevalence of anaemia in pregnancy (defined as Hb < 11 g/dl in the first and third trimester and 10.5 g/dl in the

Nutritional and dietary characteristics of study participants
The mean MUAC measurement of the study participants was 28.1 ± 3.6 cm based on the adult MUAC tape for pregnant women.Nearly half of the participants (45%) had iron deficiency (sF < 30 μg/l).The majority of the participants (53.4%) were not on the iron supplementation programme while more than three-quarters of the eligible participants (78.1%) had never taken deworming medicine.Most of the participants consumed meat and poultry (52.2%), legumes and cereals (83.4%), and vegetables and fruits (94.1%) more than twice a week (Table 2).

Factors associated with anaemia among pregnant women in northern Uganda
Pregnant women who were taking antimalarial prophylaxis were 56% less likely to be anaemic compared with pregnant women who were not taking antimalarial prophylaxis (AOR 0.44; 95% CI 0.19, 0.99).Pregnant women who consumed legumes and cereals more than twice in the previous week were 54% less likely to have anaemia compared with those who consumed legumes and cereals less than twice in the previous week (AOR 0.46; 95% CI 0.24, 0.89), as indicated in Table 3.

Discussion
In this study, we assessed the prevalence and of anaemia and associated factors among pregnant women seeking antenatal care in Lira City, northern Uganda.We found that one in four pregnant women seeking antenatal care was anaemic, with the taking of antimalarial prophylaxis and consumption of legumes and cereals as the predictors.These findings suggest gaps in the implementation of interventions to control and treat anaemia among pregnant women and the role of diet in preventing anaemia during pregnancy.
The prevalence of anaemia in our study was comparable to the 30% reported by a systematic review conducted among Prevalence of anaemia in pregnancy and associated factors in northern Uganda pregnant women in Uganda 4 and another study done in Ethiopia. 30However, our prevalence was lower than the 53% and 59% reported among pregnant women in Sudan 31 and India, 32 but higher than the 5% and 16% reported among pregnant Iranian women in the first and third trimesters respectively. 33The differences in the reported prevalence across studies could be attributed to the differences in the cut-off for Hb level, methods of measuring maternal Hb across the studies and gestational age of study participants.For example, the Indian study 32 defined anaemia as Hb < 11 g/dl across trimesters while we set our cut-off for anaemia at 11 g/ dl for the first and third trimesters and Hb < 10.5 g/dl for the second trimester.Additionally, most of our study participants (48.7%) were in the third trimester, a trimester when iron stores are transferred to the foetus and maternal serum ferritin concentrations fall. 34The result of our study indicates that anaemia in pregnancy is still a moderate public health problem in our setting, based on the World Health Organization's classification of the public health significance of anaemia. 18egnant women who consumed legumes and cereals more than twice in the previous week had 54% lower odds of having anaemia in pregnancy compared with those who consumed legumes and cereals less than twice in the previous week.Cereals and legumes such as beans, millet (kalo), and sorghum (abir) are good sources of non-heme iron 35 although less bioavailable at 1% to 10% than the hemeiron sources at 20% to 30%. 36Fortunately, these kinds of foodstuffs are part of the staple foods in northern Uganda.The association of anaemia and consumption of legumes and cereals in this study is consistent with a finding from rural Ethiopia 37 but contrary to findings from rural parts of Ghana 25 and urban areas of Ethiopia. 38Discrepancies in results across studies could be attributed to variations in demographic and dietary characteristics of study participants.
The results of this study serve to reinforce the need to emphasise the consumption of locally available iron-rich foods, which are usually affordable and readily available for most pregnant women.
The odds of anaemia among pregnant women taking antimalarial prophylaxis were 56% less than the odds of anaemia among pregnant women who were not taking antimalarial prophylaxis.Malaria is endemic in Uganda and the relationship between malarial infestation and anaemia is well documented in a previous study. 39Malarial infestation causes rupture of red blood cells, which lowers the number of red blood cells thus predisposing the individual to anaemia; thus the intake of antimalarial prophylaxis ameliorates this effect.Our finding is similar to that reported in other parts of Uganda, Malawi and Nigeria. 26,40,41Therefore, all pregnant women should be advised to take the full dose of antimalarial intermittent presumptive therapy (IPT) to prevent anaemia in pregnancy.
This study had some limitations.One of these is that we did not correct ferritin for inflammation or verify the serum ferritin results by doing a transferrin saturation test.Serum ferritin is an acute-phase protein that is affected by infections.Recall bias might have affected responses on food consumption frequency.Finally, the results of this study may not be generalisable to women in the community who do not attend antenatal clinics, as this was a hospital-based study.Therefore, the findings of this study should be interpreted within these limitations.

Conclusions
Anaemia in pregnancy is a moderate public health problem in our study population from northern Uganda and is associated with the taking of antimalarial prophylaxis and consumption of legumes and cereals less than twice in the previous week.
The results of this study give insight into the impact of interventions to control anaemia during pregnancy and the targeted areas of emphasis during maternal nutrition education.We recommend a community-based study to expand data on the burden of anaemia in pregnancy in northern Uganda.More studies are needed to unravel the optimal strategies to improve the compliance of pregnant women with interventions to control anaemia during pregnancy.

Table 2 :
Nutritional and dietary characteristics of study participants (N = 320) n = 296: study participants in the second and third trimester; MUAC: mid-upper arm circumference

Table 1 :
Demographic Ugx: Uganda shillings; n = 232: study participants who were multigravida; n = 168: study participants coming for at least the second ANC visit.

Table 3 :
Multivariate analysis of factors associated with anaemia among pregnant women in Lira City, northern Uganda *Statistically significant variables at p < 0.05; 1: reference group; COR: crude odds ratio; AOR: adjusted odds ratio.