The study was conducted at Pumwani Maternity Hospital which is a referral maternity hospital located on the East of Nairobi City (country’s capital city). The hospital is a pioneer in the provision of maternity care in sub-Saharan Africa. It has 354 obstetric beds, 144 baby cots and 2 theatres. Daily normal deliveries are 50 - 100, and Caesarean Sections are 10 - 15. To date the Hospital remains the largest maternity hospital in the country and sub-Saharan Africa. It is equivalent to a Provincial Hospital in status and is reported to be third busiest maternity hospital in African.

A hospital based cross-sectional study was conducted from 8^th June to 18^th August, 2015. The study population included in this study was all 2^nd and 3^rd trimester pregnant women attending ANC at Pumwani Maternity Hospital.

The sample size was determined by using single population proportion formula (Fischer et al., 1998; n = Z²pq/d²) by considering 95% CI, 55.1% proportion of anaemia during pregnancy from the study carried out by the Ministry of Health, Kenya in 2013 and absolute precision of 0.05 which gives a sample size of 380. However, since the population during the study period was below 10,000, sample size adjustment was done using the following

formula, resulted in 258 subjects.

where:

nf = The desired sample size (when the study population is less than 10,000);

N = Total population (around 800 antenatal mothers, in the 2^nd and 3^rd trimesters in 10 weeks period);

n = The desired sample size (when the study population is more than 10,000) which is 380.

A review of ANC attendance register showed that about 16 pregnant women in the second and third trimester attend ANC at Pumwani Maternity Hospital daily. Accordingly the study was designed to be performed for 10 weeks during which period an estimated 800 women in second and third trimester were expected to visit for antenatal care. Systematic random sampling method was used to select study participants. The estimated number of 800 pregnant women (2^nd and 3^rd trimester) in 10 weeks period was divided by the minimum adjusted sample size (258) to give the sampling interval of 3. The first pregnant woman included in the sample was chosen randomly by blindly picking one of three pieces of paper named for the first three visitors. After that, every third pregnant woman who came for antenatal care was recruited in the study until the desired sample size was attained. To avoid double participant recruitment, the research assistants confirmed whether the participants had previously visited the ANC at the hospital and participated in the study during the data collection period.

A pre-tested semi-structured questionnaire was used to collect data by trained research assistants. Data regarding participants’ demographic and socio-economic characteristics, obstetric and medical history, ANC visits and iron and folic supplementation, dietary habits and anthropometric measurements were obtained. Haemoglobin concentration was measured using portable HEMOCUE B-Hb photometer. One drop of capillary blood via finger prick was used to estimate haemoglobin level.

Data were analysed using SPSS Version 20. Descriptive statistics such as frequency, percentage, mean and standard deviation were used to describe selected variables. Pearson’s chi-square test and odds ratio (OR) with corresponding 95% confidence intervals (CI) computed to find association between independent and dependent variable (anaemia). A P-value of <0.05 was considered statistically significant. Variables having a P-value <0.05 in the bivariate analysis were subjected into a multivariate analysis to determine factors independently predicting anaemia during pregnancy.

Ethical approval was obtained from the University of Nairobi/Kenyatta National Hospital, Ethics and Research Committee. Permission to conduct the study was also obtained from the hospital administration. Written informed consent was obtained from each study participant. Confidentiality was maintained at all levels of the study. Pregnant women who found to be anaemic were advised by the research assistants to seek proper treatment and follow-up.

Table 1 shows the socio-demographic and economic characteristics of the pregnant women. The mean age of

the women was 26.4 years (SD ± 4.7). Out of the 258 participants, 104 (40.3%) and 102 (39.5%) were in the age group of 18 - 24 years and 25 - 30 years respectively. The remaining 52 (20.2%) were 31 years and above. All of the participants were from Nairobi County. Majority of the women were married (84.9%), employed (67.4%) and Christian followers (90.3%). Regarding educational level of the participants 54 (21%), 121 (46.9%) and 78 (30.3%) attended primary, secondary and tertiary level of education respectively. The highest percentage 127 (49.2%) of the participants had monthly income between 10,000 - 30,000 Ksh. About a quarter 68 (26.4%) of them had monthly income of less than 10,000 Ksh.

The obstetric history of the participants is presented in Table 2. The mean gestational age was 28.2 weeks (SD ± 8.1). The majority 165 (64.0%) of the participants were in the third trimester of pregnancy and 93 (36.0%) were in the second trimester of pregnancy. One hundred and fifty one (58.5%) of the women were multigravida while 107 (41.5%) were primigravida. Of the multigravida women 121 (80.1%) had an inter-pregnancy interval of two years or more. The remaining 30 (19.9%) had birth interval of less than two years.

Table 3 shows ANC visits and IFAS supplementation during the current pregnancy. Majority (92.2%) of the participants started attending ANC after 12 weeks of gestation. Most 217 (84.1%) of the women attended ANC

for the first time between 12 and 24 weeks of their gestational age. Almost half 89 (50.9%) of the women started taking IFAS at 19 - 24 weeks of gestation. Sixty seven (38.3%) of them started taking IFAS at 12 - 18 weeks of their gestation. Only 13 (7.4%) of the women had started taking IFAS before 12 weeks of their gestation.

Majority 213 (82.6%) of them reported they did not experience any health problem. Very small number 8 (3.1%) and 23 (8.9%) of the women reported they had diarrhoea and febrile illnesses for the last one month preceding the study respectively. Majority of the women were sero-negative while 20 (7.8%) were sero-positive (Table 4).

An assessment of weight, height and MUAC was done for all the pregnant women. Majority169 (65.5%) of the women had above 24.9 body mass index (BMI) and 88 (34.1%) of them had normal BMI (18.5 - 24.9). The mean MUAC for the women was 25.1 cm (SD ± 3.5). About three quarter 187 (72.5%) of the women had MUAC of 23 cm or above and about a quarter 71 (27.5%) of the women had MUAC of less than 23 cm (Table 5).

The majority 241 (93.4%) of the women said they drink beverages (tea, cocoa or coffee) and 66.0% of them drink these beverages in less than 20 minutes before/after meals. More than half 143 (55.4%) of the women ate meals three times per day. One third 86 (33.3%) of the women avoided certain foods such as red meat 28 (32.6%) and green vegetables 26 (30.2%) due to vomiting 37 (43.0%) and heart burn 19.8%. A few mothers 51 (19.8%) craved for non-food substances and the common type of craving was stone 44 (86.3%). All of the women reported they never smoke cigarettes and almost all (98.4%) of the women said never taken alcohol (Table 6).

The cut-off point of Hb concentration for diagnosis of anaemia during pregnancy was taken at 11 g/dl (McLean et al., 2008). The overall prevalence of anaemia (Hb < 11 g/dl) was 57.0% with 95% CI of 51.0% to 63.0%.The mean Hb concentration was 10.32 gram/deciliter (g/dl) with a SD of ±1.72 g/dl. In terms of severity, mild anaemia was 26.5%, moderate anaemia was 70.7% and severe anaemia was 2.7%.

In the bivariate analysis, participants’ age, marital status, employment status, iron and folic acid supplementa-

tion during the current pregnancy, febrile illness, HIV status and nutritional status (MUAC) were significantly associated with anaemia among the pregnant women. However, following multivariate analysis maternal age, employment status, nutritional status (MUAC) and iron and folic acid supplementation during the current pregnancy remained significantly and independently associated with anaemia.

Pregnant women aged 31 years and above were about 3 fold more likely to be anaemic [AOR = 2.71; 95% CI = 1.25 - 5.88; P = 0.012] compared to those aged 18 - 24 years. Government/private employed pregnant women had 2.9 times [AOR = 2.94; 95% CI = 1.47 - 5.88; P = 0.002] and those who were self-employed had 1.9 times [AOR = 1.91; 95% CI = 1.03 - 3.53; P = 0.039] more likely to be anaemic compared to those who were housewives. Pregnant women who did not take iron supplementation during the current pregnancy were 2 times more likely to develop anaemia [AOR = 2.04; 95% CI = 1.14 - 3.64; P = 0.016] than those who took the supplements. Pregnant women with MUAC of less than 23 cm were about 2.5 fold more likely to be anaemic [AOR = 2.52; 95% CI = 1.36 - 4.67; P = 0.003] compared to those who had MUAC of 23 cm and above (Table 7).

Abbreviations: COR = Crude Odds Ratio, AOR = Adjusted Odds Ratio, CI = Confidence Interval, ^*P < 0.05, ^**P < 0.01, ^***P < 0.001.