Study team
EXECUTIVE SUMMARY

Background
Influenza vaccine is recommended by the World Health Organization for pregnant women; however, vaccine uptake in most of Africa is negligible in the absence of national recommendations. Data on the programmatic impact and real-world effectiveness of a maternal immunization programme, particularly against severe influenza infection in infants in settings with a high prevalence of maternal HIV infection and in an African setting, will help to guide policy decisions related to implementation of maternal immunization programs in low- and middle-income countries. 

Objectives
Primary: To determine the effectiveness of antenatal maternal influenza vaccination against laboratory-confirmed influenza-associated hospitalised illness in infants <6 months of age 

Secondary: (a) To determine the effectiveness of antenatal maternal influenza vaccination against laboratory-confirmed influenza-associated hospitalised illness in infants <6 months of age stratified by maternal HIV status (b) To determine the effectiveness of antenatal maternal influenza vaccination against: (i) low birth-weight, (ii) preterm birth, (iii) small for gestational age (SGA) birth (iv) stillbirth; (c) To determine the effectiveness of influenza vaccination against laboratory-confirmed influenza associated hospitalised illness in pregnant and early postpartum (up to 42 days after delivery) women; (d) To assess occurrence and severity of adverse events following immunization with seasonal influenza vaccine in pregnant women, South Africa 

Tertiary: (a) Estimate the coverage achieved by a vaccination campaign of pregnant women in antenatal clinics, (b) Describe the spectrum of respiratory viral and bacterial infections amongst hospitalised children <6 months 

Methods
To address the primary and secondary outcomes we will conduct an unmatched test-negative case-control study in 5 sites linked to large hospitals in two provinces of South Africa (Chris Hani-Baragwanath Hospital, Rahima Moosa Mother & Child Hospital and Charlotte Maxeke Johannesburg Academic Hospital in Gauteng Province and in Cape Town the Red Cross Children’s Hospital, Tygerberg Hospital, Karl Bremer Hospital, Khayelitsha Hospital, Victoria Hospital and Mitchell’s Plain Hospital).  We will conduct a vaccination campaign in antenatal clinics targeting pregnant women before the influenza season and estimate coverage by dividing the number of vaccines delivered by the target population. Active surveillance will be conducted for hospitalised children <6 months with any non-surgical medical diagnosis in facilities serving the population targeted for the vaccination campaigns. Respiratory specimens from all hospitalised children during the influenza season will be tested for influenza and other respiratory viruses (e.g. respiratory syncytial virus) and bacteria (e.g. pertussis) by polymerase chain reaction. For the primary endpoint, cases will be hospitalised infants <6 months of age with laboratory-confirmed influenza-associated illness whose mothers were eligible to receive influenza vaccination and aged 0-6 months in the period of influenza circulation. Controls will be hospitalised infants as above, who test influenza negative. We will also conduct surveillance at maternity wards in a subset of surveillance hospitals for babies born to pregnant women, specifically to evaluate for low birth weight, preterm and SGA births. For these outcomes cases will be babies born with the endpoint of interest and controls will be babies born without the endpoint. Exposure will be the influenza vaccination status of the mother during the current season. Data will be collected using standardized case investigation forms and for the subset at maternity wards by collecting routine administrative data available at the wards. Maternal vaccination status will be ascertained through written documentation. Data on potential confounders will be collected. For the primary endpoint, we will calculate an adjusted odds ratio for maternal vaccination among cases vs. controls using logistic regression. Vaccine effectiveness = 1 - adjusted odd ratio for vaccination x 100%. Assuming 40% vaccine effectiveness and 50% coverage we will need to enrol 142 cases and 1838 controls (assuming a 95% confidence interval, 80%

Background
Burden of influenza in Africa
Seasonal and pandemic influenza-associated morbidity and mortality in the United States and other temperate countries are well-described 1-10. The influenza-associated disease burden in sub-Saharan Africa is not well established, particularly outside of the 2009 pandemic period. However, a high case fatality risk has been reported from several influenza outbreaks in Africa 11-15. The epidemiology of influenza may be different in sub-Saharan Africa due to less predictable seasonality and the high prevalence of possible risk factors for severe influenza including HIV, tuberculosis, and malnutrition 16-19. In South Africa, HIV-infected persons experience 4-8 times the age-adjusted incidence of severe influenza infection and 4 times the risk of influenza-associated death compared to HIV-uninfected persons 20. HIV-infected children hospitalised with influenza in South Africa were found to have more severe disease 21 and had an age-adjusted relative risk of death of 11.5 (95% confidence interval (CI) 9.6-12.6) compared to HIV-uninfected children 22. Following the 2009 pandemic, a global pooled analysis found immune compromise was associated with a pooled relative risk of death of 27.7, but lacked sufficient information about HIV status to determine if a specific association existed 23.  The same pooled analysis used data from 3 countries to assess tuberculosis as a risk factor for influenza-associated death.  

They found tuberculosis prevalence increased with level of disease severity from a median of 1.7% among hospitalised to a median of 2.6% of fatal cases 23. Surveillance data from South Africa have also revealed that tuberculosis coinfection is an independent risk factor for influenza-associated death amongst hospitalised patients with influenza [odds ratio (OR) 3.9 (95% CI 1.5-10.0)] 24. Few data are available about the relationship between malnutrition and influenza in children in sub-Saharan Africa.  Two studies from Argentina and Peru assessed malnutrition as a risk factor for death during the 2009 A(H1N1) pandemic period. Among hospitalised children in Argentina malnutrition was a risk factor for death among children with influenza A(H1N1)pdm09 infection [RR=3.07 (1.46-6.48)] 25. A study of 74 hospitalised children in Peru found that malnutrition was common but was not associated with severe disease [OR=1.7 (0.4-7.2)] 26.

Objectives
Primary: To determine the effectiveness of antenatal maternal influenza vaccination against laboratory-confirmed influenza-associated hospitalised illness in infants <6 months of age
Secondary: To determine the effectiveness of antenatal maternal influenza vaccination against laboratory-confirmed influenza-associated hospitalised illness in infants <6 months of age stratified by maternal HIV status  
To determine the effectiveness of antenatal maternal influenza vaccination against:
Low birth-weight (LBW)
Preterm birth
SGA birth 
Stillbirth
To determine the effectiveness of influenza vaccination against laboratory-confirmed influenza associated hospitalised illness in pregnant and early postpartum (up to 42 days after delivery) women. 
To assess occurrence and severity of adverse events following immunization with seasonal influenza vaccine in pregnant women, South Africa
Tertiary: Estimate the coverage achieved by a vaccination campaign of pregnant women in antenatal clinics
Describe the spectrum of respiratory viral and bacterial infections amongst hospitalised children <6 months  

Methods
Study design
We will conduct a vaccination campaign amongst pregnant women in several different areas of South Africa and monitor the effectiveness of the programme. We will then use an unmatched case-control study design to address the endpoints of:
1.    Laboratory-confirmed influenza-associated hospitalised illness in infants <6 months of age overall and by HIV status. 
2.    LBW
3.    Preterm birth 
4.    SGA birth
5.    Stillbirth
6.    Laboratory-confirmed influenza-associated hospitalised illness in pregnant and postpartum women

Study setting and population
The study will be conducted in Gauteng Province and Western Cape Province in South Africa. Both are urban areas with good access to hospital care for mothers and infants. We will conduct vaccination campaigns in Gauteng (City of Johannesburg including Soweto) and the Western Cape (City of Cape Town) amongst pregnant women. Babies born to women who were eligible to receive influenza vaccination during pregnancy (i.e. pregnant from April onwards but actual date will be determined based on actual campaign dates and resident in the area where vaccination was offered) will be the study population.

We will conduct surveillance for hospitalisations amongst infants <6 months of age. In City of Johannesburg we will conduct surveillance at Chris Hani-Baragwanath Academic Hospital (CHBAH), Rahima Moosa Mother and Child Hospital (RMMCH). In City of Cape Town we will conduct surveillance at Red Cross War Memorial Children’s Hospital (RCCH) and Mitchell’s Plain Hospital (MPH). We will conduct surveillance for adverse birth outcomes at CHBAH and RMMCH in City of Joburg and Groote Schuur, Mowbray Maternity, Mitchells Plain Hospital, and Mitchells Plain MOU in City of Cape Town.

For children admitted to RMMCH, RCCH and MPH, the majority will also be eligible to be enrolled into the National Institute for Communicable Diseases (NICD) pneumonia surveillance programme (Ethics number M140824 /REF 836_2014). Data collected for the study described in this protocol includes all data elements collected in the pneumonia surveillance programme. Patients enrolled into the study described in this protocol will be asked if they also consent to have their data used for pneumonia surveillance. If they consent to this, then data will be included in the pneumonia surveillance database.

We will conduct surveillance for LRTI hospitalisations among pregnant and postpartum women. In Gauteng we will conduct surveillance at CHBAH, RMMCH and Helen Joseph Hospital. In the Western Cape we will conduct surveillance at Mitchell’s Plain Hospital, Mowbray Maternity Hospital and Groote Schuur Hospital. For pregnant and post partum women admitted at RMMCH and Helen Joseph, these will be co-enrolled into the NICD pneumonia surveillance programme as described above. 
During the study active monitoring of vaccine adverse events will be conducted on a sample of the population targeted for vaccination (45 to 50 000 pregnant women).  

Published data report a low frequency of adverse events following immunization, between 0.5 to 1% of those vaccinated (1).  

Study Team