Seroepidemiological investigations of Salmonella enterica serovar Typhi infection and the potential role of vaccination in the control of typhoid fever in Fiji

Abstract

Typhoid fever is a potentially-life threatening systemic disease caused by Salmonella Typhi, a human-restricted bacterium, spread through the faecal-oral route. Following a sustained rise in observed incidence in Fiji from 2004, in 2013, I undertook a nationally-representative cross-sectional serological survey of 1,531 participants to determine infection by age, assess putative risk factors, and quantify social contact patterns. These data were utilised in the development of a transmission dynamic model. The literature indicated that typhoid transmission models are relatively under-utilised, particularly in economic evaluation, with little to guide use of vaccination in place of or alongside water, sanitation and hygiene (WASH). The serosurvey found that iTaukei and non-iTaukei Fijians have similar risk of raised IgG antibodies to the Vi antigen expressed by S. Typhi. Seroprevalence increased with age, suggestive of endemic transmission or declining incidence. Unimproved sanitation may increase risk of seropositivity. Geospatial analysis suggested rainfall, proximity to major rivers and creeks, or flood-prone areas were risk factors for acquisition of anti-Vi IgG antibodies. Social mixing was assortative by ethnicity and age when assessed by mealtime contacts and highest in school-age children. Increasing number of age-adjusted contacts increases the odds ratio for being seropositive, though substantial uncertainties remain around the specificity and sensitivity of serological thresholds as indicators of past typhoid infection. An age- and ethnicity-structured transmission dynamic model fitted the serology and case surveillance data well when including a substantially reduced force of infection for high-dose infection being passed to non-iTaukei Fijians, and high generation of asymptomatic non-infectious cases per new infectious case. Surveillance reporting of infectious cases was estimated as one in five infectious adult cases and one in twelve infectious child cases. The fit to the data suggested endemic rather than declining transmission, and there was better fit with age-ethnic assortative mixing than with ethnically-assortative or homogeneous mixing. Vaccine scenarios suggested that of single dose routine programmes, school entry could be more effective than school leaver vaccination, reflecting age-contact transmission probabilities in the model. Modest reduction (10%) in per-case infectious transmission through effective WASH programmes offered substantial incidence reductions of around 25%, comparable to two-dose (school entry and exit) ViPS vaccination programmes. Potential benefits of conjugate vaccines were projected to be similar to more effective WASH programmes, with administration alongside other vaccines in the second year of life projected to offer approximately 50% incidence reduction, the most benefit of any single dose regimen; with the impact being greater if typhoid carrier daily infectious risk is lower than the daily infectiousness of acute typhoid fever cases.