Mathematical Modeling to Assess the Drivers of the Recent Emergence of Typhoid Fever in Blantyre, Malawi

Typhoid fever, caused by infection with the human-restricted bacterial pathogen Salmonella enterica serovar Typhi, is a major cause of illness and mortality in regions of the world with limited access to improved water and sanitation [1]. Recent estimates have placed the global burden of typhoid fever at 11.9–26.9 million cases and 129 000–270 000 deaths per year [2–5]. Although the endemic burden of typhoid fever in South and Southeast Asia has long been recognized, less is known about the burden of disease in sub-Saharan Africa. In Malawi, as in much of sub-Saharan Africa, nontyphoidal Salmonella (NTS) serovars have been a much more common cause of bloodstream infections over the past 2 decades. Sequential epidemics of Salmonella enterica serovars Typhimurium and Enteritidis have been observed [6]. Salmonella Typhi represented just 2% (105/5061) of Salmonella isolates detected by sentinel surveillance at Queen Elizabeth Central Hospital (QECH) in Blantyre, the largest hospital in Malawi, between 1998 and 2004 [6]. Since 2011, however, there has been a substantial increase in the number of confirmed cases of Salmonella Typhi at QECH. The number of typhoid cases increased from 14 per year during 1998–2010 to 843 cases in 2013. This epidemic of typhoid fever has continued for at least 3 years [7] and coincides with numerous reports of ongoing epidemics of typhoid fever in settings across Africa [7–12]. These epidemics have closely followed the recent global emergence of the H58 haplotype of Salmonella Typhi [13]. The H58 lineage is highly clonal and differentiated from other haplotypes using a simple single nucleotide polymorphism–based typing scheme [14], and is associated with high levels of multidrug resistance [13, 15]. There are a variety of hypotheses that could explain the large increase in typhoid fever cases in Blantyre, including (1) an increase in population density in Blantyre beyond a critical threshold for transmission; (2) waning heterotypic immunity to Salmonella Enteritidis that is cross-protective against Salmonella Typhi; (3) an increase in the prevalence of multidrug resistant (MDR) strains, resulting in prolonged persistence; and (4) the emergence of the MDR H58 haplotype, which has become dominant in many places around the world and may be more transmissible than other strains [13]. The latter 2 hypotheses are tightly coupled, but describe slightly different mechanisms by which the H58 haplotype, and MDR strains more broadly, may have led to the outbreak. Mathematical models provide a platform for assessing the plausibility of these hypotheses to help direct future research efforts. We therefore adapted a mathematical model for the transmission dynamics of typhoid developed previously [16] by fitting to age-specific data on microbiologically confirmed cases of typhoid fever in Blantyre, Malawi. We modified the model to explore the 4 hypotheses outlined above, and examined whether each of them could explain the recent prolonged outbreak of typhoid fever cases in this setting.