From global to local; invasive and antimicrobial-resistant Salmonella

Background: Invasive nontyphoidal Salmonella (iNTS) is a leading cause of bacteraemia globally and of childhood mortality in sub-Saharan Africa. However, little is known about the iNTS disease burden, temporal trends, and risk factors including co-infections with HIV and malaria. Antimicrobial resistant (AMR) infections due to iNTS serovars are becoming increasingly common, making empirical treatment problematic and potentially resulting in worse health outcomes. Despite the importance of AMR in NTS, the sources are poorly understood. My PhD objectives were to: 1) estimate the global burden of iNTS; 2) compare health outcomes from AMR and susceptible NTS infections; 3) identify sources of AMR NTS in humans; 4) describe the epidemiology of iNTS infections in Queensland, Australia; and 5) identify risk factors associated with iNTS disease. Methods: I used data from a variety of sources to investigate the health impacts of NTS from global to local settings. These data sources included: 1) published studies and grey literature identified through systematic reviews; 2) The Vietnam Initiative on Zoonotic Infections (VIZIONs); a community-based study of zoonotic disease transmission; 3) The study on bacterial etiology and antimicrobial susceptibility profile of bloodstream infections at the Hospital for Tropical Diseases in Ho Chi Minh City, Vietnam; 4) The study on aetiology of diarrhoea in provincial hospitals in southern Vietnam; 5) The Vietnam Bacterial Resistance (VIBRE) study; and 6) Salmonella data from the Notifiable Conditions Register held at Queensland Health. For specific projects I developed univariable and multivariable regression models including Bayesian meta-regression. Results: In 2017, there were 534.6 thousand iNTS cases resulting in 77.5 thousand deaths and accounting for 4.26 million disability-adjusted life years (DALYs) globally. Of these, 8.2% of cases and 24.3% of deaths were attributable to co-infection with HIV. The highest incidence rates occurred in sub-Saharan Africa and in children <5 years of age. The mean all-age case fatality for iNTS was 14.5%, with higher estimates among children and the elderly, HIV-infected, and in areas of low socio-economic development. From a systematic review, multidrug-resistant (MDR) NTS infections were associated with higher risk for bacteraemia, hospitalisation, and mortality. I identified that the source of isolates in Vietnam was a significant predictor of AMR, with NTS isolated from human blood being more resistant to antimicrobials. In addition, human NTS diarrhoeal isolates were almost 3 times more likely to be MDR when compared with isolates from asymptomatic humans. Source attribution models of 242 different sequence types and AMR (ST-AMR) profiles across NTS isolates from humans and animals identified an unsampled source a principal contributor for AMR in humans, with chickens and pigs being the second most important contributors. Lastly, I identified a 6.5% annual increase in the number of iNTS cases from 2007-2016 in the sub-tropical Australian state of Queensland. The highest notification rates occurred in summer months, among males, infants, and the elderly, and in remote areas. Overall, serovar Virchow was the most prevalent iNTS serovar; however, as in Vietnam, serovar Choleraesuis was the most invasive. Conclusions: I demonstrated the significant global burden of iNTS disease, estimated the impact of HIV infections, identified temporal trends, and health gap metrics including DALYs. I provided evidence that AMR NTS infections result in worse health outcomes including higher risk for iNTS. Although antimicrobial use in food animals plays a role in the emergence of AMR in humans, it does not explain its full extent suggesting the role of alternative sources. Lastly, I estimated the incidence of iNTS disease in Queensland and described risk factors that may provide hypotheses for future prevention efforts.