Prevalence of hepatitis B virus infection and associated risk factors among pregnant women attending antenatal clinic in Mulago Hospital, Uganda: a cross-sectional study

AIM: To determine the prevalence and factors associated with hepatitis B virus infection among pregnant women attending antenatal clinic in Mulago Hospital. DESIGN: Cross-sectional observational study. SETTING: Mulago National Referral Hospital, Uganda, antenatal clinic. PARTICIPANTS: We randomly selected 340 pregnant women attending their first antenatal visit at Mulago Hospital antenatal clinic. PRIMARY OUTCOME: Hepatitis B surface antigen positivity. RESULTS: We recruited 340 participants, with a mean age of 27±5.7 years, and a median gravidity of 3. The prevalence of hepatitis B virus infection among pregnant women attending the antenatal clinic in Mulago Hospital, in our study, was 2.9% (95% CI 1.58% to 5.40%, n=10). Factors positively associated with hepatitis B virus infection were: marital status (adjusted OR (aOR)=11.37, p=0.002), having a hepatitis B positive family member (aOR=49.52, p<0.001) and having had a blood or body fluid splash to mucous membranes from a hepatitis B positive patient (aOR=61.69, p=0.015). Other factors such as age, socioeconomic status, number of sexual partners, HIV serostatus, piercing of ears and history of blood transfusion were not significantly associated with hepatitis B virus infection in this study. CONCLUSION: The prevalence of hepatitis B virus infection among pregnant women attending antenatal clinic in Mulago Hospital was of intermediate endemicity. We found that marital status, having a hepatitis B positive family member at home and having had a blood or body fluid splash to mucous membranes from a hepatitis B positive patient were independently associated with hepatitis B infection. Factors such as age, HIV status, history of blood transfusion, piercing of ears and social status were not associated with hepatitis B status in this study. This was a cross-sectional study, conducted at Mulago Hospital antenatal clinic. Mulago Hospital was founded in 1917 in Kampala City, the capital of Uganda, and is the Ugandan national referral hospital. It is a public general and teaching hospital, with a bed capacity of over 1500 beds. Mulago Hospital is located on Mulago hill, approximately 5 km northeast of Kampala’s central business district. Approximately 2500 women attend for antenatal care at Mulago Hospital every month. We purposively selected Mulago Hospital because it serves people from all ethnic groups and regions of Uganda. It also serves a high number of refugees, such as those from Rwanda, South Sudan and Congo. Kawempe General Hospital, in Kawempe division, temporarily houses the Mulago Hospital antenatal clinic, as the main Mulago Hospital is undergoing major renovation. The clinic continues to offer preventive and curative services, such as health education, immunisation, prenatal maternal and fetal monitoring. Between 200 and 250 pregnant women visit the clinic on every Tuesday, Wednesday and Thursday, of whom about 100 each day are attending their first antenatal visit. We enrolled only mothers attending their first visit into the study, as the flow of first visit attendees in the antenatal care (ANC) clinic was convenient for recruitment, unlike their repeat visit counterparts. We enrolled women after their health education session, as they awaited examination. The study population was pregnant women attending antenatal clinic, between December 2018 and February 2019, in Mulago Hospital, Kawempe site. All pregnant women who were attending their first antenatal visit at Mulago Hospital, who were willing to participate in the study. Women below 18 years were eligible for recruitment as emancipated minors. We excluded women who could not understand either English or Luganda, which were the two languages used for the consenting process and administration of the questionnaires. The dependent variable was hepatitis B infection among pregnant women, assessed by HBsAg positivity. The independent variables were factors associated with positive hepatitis B status. These included the sociodemographic factors such as age, marital status, level of education, socioeconomic status and sexual behaviour. We also included obstetric factors such as gravidity and parity and other factors such as history of blood transfusion, history of caring for a hepatitis B positive patient, immunisation status, ear piercings, tattoos and intravenous drug use in our model. We estimated the sample size for prevalence using the Wiegand18 Leslie formula, 1965.19 Using a prevalence from Northern Uganda of 11.8%,5 the required sample size to determine prevalence would be 160 women. We based the sample size estimates for factors associated with hepatitis B infection in pregnancy on the hypothesis that the prevalence of infection would be higher among women ≤20 years than among those >20 years, as found by Bayo et al.5 We compared the proportions in subpopulations (Fleiss,20 Statistical Methods for Rates and Proportions, formulas pages 3.18 and 3.19) using an OpenEpi calculator (accessed at http://www.openepi.com/SampleSize/SSCC.htm). Assuming power of 80%, type 1 error of 5% and frequency of hepatitis B infection of 20% among women ≤20 years and 8.7% among women >20 years,5 we calculated a total sample size required for risk factor modelling of 340. Three hundred and forty women were included in the study. We used a random probability sampling method. During the day of data collection, we used the register book for the antenatal clinic attendance and the booking cards for the first visit attendees to identify eligible potential participants. We invited every 10th eligible woman to take part in the study. If the 10th woman declined participation in the study, then the next eligible woman was invited. There was an average daily attendance of ~100 eligible potential participants on 3 days per week. Data were collected over a 3-month period (December 2018 and February 2019). The antenatal clinic midwives, who were the research assistants, approached potential participants and explained the study purpose to them. They then screened them for eligibility and obtained informed consent for participation in the study. We used interviewer-administered semi-structured questionnaires to get sociodemographic information including age, gestational age, gravidity, parity, education and occupation, and other factors that could influence infection with hepatitis B virus such as injection drug use, tattoos, family history of hepatitis B infection, sexual history, among others. Obstetric factors like gravidity and parity, previous history of taking care of a hepatitis B positive patient, immunisation against hepatitis B were all assessed. The laboratory technician collected 4 mL of whole blood from the antecubital fossa under aseptic technique, into BD Clot Activator Tube vacutainers (labelled with the participant’s study number), using WHO (2010) guidelines on blood collection.21 The technician then took the blood to the Kawempe Hospital laboratory, centrifuged and transferred an aliquot of serum, about 2 mL, into cryovials. We performed immunochromatographic assay using HBsAg rapid test strip, FaStep, for HBsAg testing following the manufacturer’s instructions. This rapid test kit has a relative sensitivity of >99.9%, a relative specificity of 99.9% and an accuracy of >99.9%. We did quality control on this test kit using known positive and negative samples from MBN Clinical Laboratories, an ISO 9001 certified laboratory, before it was used to test any participant’s samples. We performed the HBsAg rapid tests using the following procedure: the laboratory technician labelled a rapid test strip with the patient identifiers, and placed it on a flat surface. He pipetted three drops of serum from the cryovial and applied it to the sample pad of the test strip. We read the result after 15 min. A positive result was one where two-coloured bands appeared on the membrane, one in the control region and another in the test region. A negative result was one where only one band appeared in the control region, and none in the test region. An invalid result was one where a band failed to appear in the control region. We produced results and gave a copy to the patient after being counselled by the principal investigator and the nurse counsellor. We sent all HBsAg positive samples to MBN Clinical Laboratories, Kampala, an ISO 9001 accredited laboratory, for confirmatory HBsAg testing, HBeAg testing and liver function tests. We took 5% of all HBsAg negative samples to MBN, for external quality assurance. We packed and transported all samples at 2–8°C, on the same day of collection. We referred hepatitis B positive mothers to the hepatitis B clinic in Mulago Hospital (Kiruddu) for further care, where the physicians performed their viral load assays and further assessments for treatment. We planned to vaccinate their babies against hepatitis B within 24 hours of birth. The study endeavoured to meet the cost of the HepB-BD vaccine for all study participants who were hepatitis B positive. The principal investigator double-entered the raw study data into Epidata V.3.1 and imported into STATA V.14.1. Numerical results are presented as frequencies or percentages. To assess factors associated with hepatitis B infection in pregnancy, we performed bivariate analysis and computed crude ORs and 95% CIs. Student’s t-test was used to assess statistical significance between groups for continuous variables, while the Fischer’s exact test was used for categorical variables. To assess the independent association of these risk factors, we performed multivariate analysis. All independent variables with a p value of <0.2 at bivariate analysis and those with biological significance were included in multivariate logistic regression. Stepwise regression method was used to determine the model of best fit. We considered all associations with p value <0.05 as statistically significant. We explained information regarding the study to the subjects and got written consent for participation in the study in English/Luganda. We interviewed participants and collected samples, following local guidelines regarding research participant privacy. Only study numbers, with unique participant identifiers, were used to label the questionnaires and the samples. We kept filled questionnaires separate from signed consent forms and locked away securely all study material, with only access to study team. Participants were free to drop out at any point of the study, with an assurance that if they did so this would not affect their ongoing antenatal care. All researchers in the study were fully immunised against hepatitis B. They wore protective gear such as gloves, masks and eye glasses during laboratory procedures. These measures protected both mothers and health workers from acquiring hepatitis B infection from each other during the study. While there were patients involved in this study as study participants, we did not involve them at the stage of the study design. However, the intended outcome in favour of the patient was that we made available hepatitis B birth dose vaccination for all exposed babies. We circulated a summary of the findings using SMS messaging to all study participants.