A Prospective Household Observational Cohort Study of Influenza, Respiratory Syncytial Virus and Other Respiratory Pathogens Community Burden and Transmission Dynamics in South Africa

Background: Information on the community burden of influenza is essential to inform control, but is not routinely collected. For example, influenza transmission models, which are widely used to inform the efficacy and cost-effectiveness of vaccines, antivirals and non-pharmaceutical countermeasures, depend on valid epidemiological estimates of the community occurrence and transmission of disease. While community level studies such as the Flu Watch cohort study in the United Kingdom have been implemented in recent years in high-income countries, the community burden of influenza remains largely unknown in lower- and middle-income countries, especially in Africa where the burden of influenza may be higher. In addition, no data exist on the household transmission dynamics of influenza viruses. Such data could provide important insight to inform targeted interventions, for instance, how to maximize indirect effects of influenza vaccination to protect more vulnerable groups for which the available influenza vaccines are currently not licensed or the efficacy is low (e.g. children <2 years of age).

Human respiratory syncytial virus (RSV) is a major cause of childhood acute lower respiratory tract infection, especially among infants <3 months of age. While a RSV vaccine is currently not available, the efficacy of promising RSV candidate vaccines is being evaluated. Alternative strategies for RSV vaccination have been proposed, including vaccination of older age groups to indirectly protect vulnerable infants by reducing circulation of virus in the population or preventing chains of transmission to the infant. Nonetheless, to evaluate these effects knowledge of the RSV burden, transmission dynamics and source of infection for infants in the community is needed.

Pneumococcus is the leading bacterial cause of pneumonia. The pneumcoccal conjugate vaccine (PCV) was introduced into the expanded programme on immunisation in 2009. Data on carriage prevalence, dynamics of carriage and interaction with other respiratory pathogens will be helpful to guide interventions to target pneumococcal disease not prevented by vaccination. Both diphtheria and pertussis have been identified as emerging pathogens in South Africa in recent years, giving rise to outbreaks in 2014 and 2015 after years of low prevalence. Meningococcus is an epidemic-prone disease causing cyclical outbreaks in South Africa but little is known about the carriage prevalence in South Africa. Annual incidence of tuberculosis in Southern Africa is the highest globally, with Swaziland, Lesotho and South Africa have the highest rates of tuberculosis in the world. Moreover, in South Africa tuberculosis is the leading cause of death on death notification forms and, in HIV-infected individuals (both not yet receiving antiretrovirals and whilst taking antiretrovirals), is the leading serious opportunistic infection.

Data on the carriage prevalence and/or transmission dynamics of these pathogens are important to guide control measures, including targeted vaccination strategies. Prevention efforts to limit transmission in households are often recommended but data on household transmission and the impact of HIV infection are limited to date.

Objectives Primary

• To estimate the community burden of influenza and RSV, including: (i) the incidence of influenza and RSV infection in the community; (ii) the symptomatic fraction associated with influenza and RSV infection; (iii) the severity associated with symptomatic infections; and (iv) the fraction of individuals with symptomatic infection seeking medical care.
• To assess the transmission dynamics of influenza and RSV infections in the community, including: (i) the estimation of the household secondary infection risk (SIR), serial interval and length of shedding; (ii) the estimation of transmission of infection between age groups within the household and possibly the community; and (iii) the estimation of the effective reproductive number (Rt) and its variation over time in the community.

Secondary (selected)

• To estimate the symptomatic fraction, the severity associated with symptomatic infections and the fraction of individuals with symptomatic infection seeking medical care among influenza or RSV positive cases by HIV serostatus and age.
• To estimate the SIR and length of shedding of influenza and RSV among HIV-infected and HIV-uninfected index cases and the rate of acquisition of influenza and RSV infection among HIV-infected and HIV-uninfected household members.
• To assess the role of asymptomatic infections in the household transmission of influenza and RSV.
• Determine the heterogeneity of influenza and RSV virus strains within household clusters and describe viral evolution within and between households as well as the association between virus strains and the duration of virus shedding and HIV status
• To describe and compare nasopharyngeal pneumococcal loads in healthy individuals by age and over time, and how the loads may be altered by respiratory viral infection, HIV-status and vaccination-status
• To identify the prevalence and duration of B. pertussis colonization within the community by age group, vaccine status and over time, the transmission dynamics within a household, and the proportion of individuals that develop symptomatic infection
• To determine the prevalence of N. meningitidis and C. diphtheria colonization within the community by age group at a single point in time each year
• To measure the annual incidence of tuberculosis infection in individuals living in study households and assess risk factors (including incident and prevalent HIV) for acquiring tuberculosis infection.
• To ascertain the impact of housing quality, fuel use and indoor air quality and selected measures of ventilation on tuberculosis infection and on household transmission of respiratory viruses and bacteria Methods: We will conduct a household-level community cohort study in a rural community in Mpumalanga Province where health and demographic surveillance (HDSS) is conducted (the Agincourt HDSS surveillance site) and selected urban areas in the Matlosana Muncipality in North West Province. Prospective, hospital-based surveillance for influenza- and RSV-associated severe acute respiratory illness has been conducted at these sites since 2010. The prospective cohort study will be conducted for 3 consecutive years, from 2016-2018, to include 3 consecutive RSV and influenza seasons. Approximately one hundred households (50 households per site, with an expected average number of household members of 5 to recruit a total of 500 enrolled individuals) will be randomly selected within the study population each year and consented to participate to the study. For each household member, serum will be collected quarterly (January-February, April-May, September-October and November-December) including before and after the RSV (February to May) and influenza (May to September) seasons and a final blood draw at the end of the year. The demographic characteristics of the household members will be collected on enrollment. Starting before the RSV and influenza seasons, upper respiratory tract samples will be collected twice-a-week by trained study nurses from all household members irrespective of the presence of respiratory symptoms throughout the year. During each visit the development of symptoms (using a severity score) and healthcare seeking behavior of symptomatic cases will be recorded. In addition, during winter at one household visit, an oropharyngeal swab will be collected from each individual for testing for meningococcus and diphtheria. Nasopharyngeal samples will be tested for the detection of RSV or influenza virus and pneumococcus and B. pertussis infection by real-time reverse-transcriptase polymerase chain reaction (rt-PCR). Blood samples will be tested for rise in RSV, influenza and pertussis antibody titers. From consenting patients, blood samples will be tested for HIV infection. For tuberculosis infection, we will do a tuberculin skin test (TST) at baseline and the end of the first year and annually thereafter in people are found to be TST negative at their previous test. HIV testing will be offered to all participants at baseline and then six monthly thereafter. All participants will have a sputum sample taken for culture at baseline and thereafter symptom screened for symptoms of tuberculosis monthly, and investigated with Xpert Mtuberculosis/Rif if symptoms of tuberculosis are present.

The sample size of approximately 1500 individuals over 3 consecutive seasons will allow the estimation of 20% risk of infection and a 10% risk of illness with 95% CI and 5% desired absolute precision in the community. Individuals in the households enrolled in the first and second year of the study will be have a single follow-up visit towards the end of the 2nd and 3rd year of the study (October). At this visit a blood specimen for testing for HIV infection, influenza and RSV serology will be collected and TST testing for individuals found to be TST negative at the previous test will be performed.

Impact: This study will improve understanding of the community burden of influenza, RSV, pertussis, tuberculosis and pneumococcal infection in South Africa. It will also provide data on the carriage prevalence of meningococcus and diphtheria. The data generated from this study will also provide important information on the transmission dynamics of influenza, RSV, pertussis and pneumococcus in the community allowing to better strategize interventions (including targeted vaccination strategies) and evaluate their potential impact. Moreover, there is an absence of prospective data on tuberculosis infection from high tuberculosis burden countries. The data generated will both inform modelling of transmission, sample size for prevention studies and surveillance assessing the impact of the National Tuberculosis Control Program.