Estimating the subnational prevalence of antimicrobial resistant Salmonella enterica serovars Typhi and Paratyphi A infections in 75 endemic countries, 1990-2019: a modelling study

Abstract

Background: Enteric fever, a systemic infection caused by Salmonella enterica serovars Typhi and Paratyphi A, remains a major cause of morbidity and mortality in low-income and middle-income countries. Enteric fever is preventable through the provision of clean water and adequate sanitation and can be successfully treated with antibiotics. However, high levels of antimicrobial resistance (AMR) compromise the effectiveness of treatment. We provide estimates of the prevalence of AMR S Typhi and S Paratyphi A in 75 endemic countries, including 30 locations without data.

Methods: We used a Bayesian spatiotemporal modelling framework to estimate the percentage of multidrug resistance (MDR), fluoroquinolone non-susceptibility (FQNS), and third-generation cephalosporin resistance in S Typhi and S Paratyphi A infections for 1403 administrative level one districts in 75 endemic countries from 1990 to 2019. We incorporated data from a comprehensive systematic review, public health surveillance networks, and large multicountry studies on enteric fever. Estimates of the prevalence of AMR and the number of AMR infections (based on enteric fever incidence estimates by the Global Burden of Diseases study) were produced at the country, super-region, and total endemic area level for each year of the study.

Findings: We collated data from 601 sources, comprising 184 225 isolates of S Typhi and S Paratyphi A, covering 45 countries over 30 years. We identified a decline of MDR S Typhi in south Asia and southeast Asia, whereas in sub-Saharan Africa, the overall prevalence increased from 6·0% (95% uncertainty interval 4·3-8·0) in 1990 to 72·7% (67·7-77·3) in 2019. Starting from low levels in 1990, the prevalence of FQNS S Typhi increased rapidly, reaching 95·2% (91·4-97·7) in south Asia in 2019. This corresponded to 2·5 million (1·5-3·8) MDR S Typhi infections and 7·4 million (4·7-11·3) FQNS S Typhi infections in endemic countries in 2019. The prevalence of third-generation cephalosporin-resistant S Typhi remained low across the whole endemic area over the study period, except for Pakistan where prevalence of third-generation cephalosporin resistance in S Typhi reached 61·0% (58·0-63·8) in 2019. For S Paratyphi A, we estimated low prevalence of MDR and third-generation cephalosporin resistance in all endemic countries, but a drastic increase of FQNS, which reached 95·0% (93·7-96·1; 3·5 million [2·2-5·6] infections) in 2019.

Interpretation: This study provides a comprehensive and detailed analysis of the prevalence of MDR, FQNS, and third-generation cephalosporin resistance in S Typhi and S Paratyphi A infections in endemic countries, spanning the last 30 years. Our analysis highlights the increasing levels of AMR in this preventable infection and serves as a resource to guide urgently needed public health interventions, such as improvements in water, sanitation, and hygiene and typhoid fever vaccination campaigns.

Funding: Fleming Fund, UK Department of Health and Social Care; Wellcome Trust; and Bill and Melinda Gates Foundation.

Figure 1

Model estimates of the prevalence of multidrug resistance in S Typhi isolates in endemic countries for 1990, 2000, 2010, and 2019 at the administrative division level one resolution Results represent the mean of 1000 draws of the stacked ensemble plus the spatial temporal conditional autoregressive model. Estimates are presented for 75 endemic countries and were not produced for countries deemed non-endemic (shown in grey). Multidrug resistance was defined as concurrent resistance to ampicillin, chloramphenicol, and trimethoprim-sulfamethoxazole.

Figure 2

Model estimates of the number of MDR and non-MDR Salmonella enterica Typhi infections, and the prevalence of MDR S Typhi isolates by GBD super-region and year Stacked bar plots show the number of S Typhi infections susceptible to first-line antibiotics (non-MDR; orange) and the number of MDR S Typhi infections (blue) for each year, 1990–2019. The prevalence of MDR S Typhi infections is shown by the solid red line, with 95% uncertainty intervals shown by the red dashed line. Estimates are shown for modelled disease-endemic countries within each GBD super-region. MDR=multidrug resistance. GBD=Global Burden of Diseases.

Figure 3

Model estimates of the prevalence of fluoroquinolone non-susceptibility in Salmonella enterica Typhi in all endemic countries for 1990, 2000, 2010, and 2019 at the administrative division level one resolution Results represent the mean of 1000 draws of the stacked ensemble plus the spatial temporal conditional autoregressive model. Estimates are presented for 75 endemic countries and were not produced for countries deemed non-endemic (shown in grey). Fluoroquinolone non-susceptibility was defined as ciprofloxacin minimum inhibitory concentration of ≥0·125 μg/mL or nalidixic acid resistance.

Figure 4

Model estimates of the number of FQNS and fluoroquinolone susceptible Salmonella enterica Typhi infections, and the prevalence of FQNS S Typhi isolates by GBD super-region and year Stacked bar plots show the number of S Typhi infections susceptible to fluoroquinolones (non-FQNS; orange) and the number of FQNS S Typhi infections (blue) for each year, 1990–2019. The prevalence of FQNS S Typhi infections is shown by the solid purple line, with uncertainty intervals shown by the purple dashed line. Estimates are shown for modelled endemic countries within each GBD super-region. FQNS=fluoroquinolone non-susceptibility. GBD=Global Burden of Diseases.

Figure 5

Model estimates of the prevalence of multidrug resistance in S Paratyphi A in all endemic countries for 1990, 2000, 2010, and 2019 at the administrative division level one resolution. Results represent the mean of 1000 draws of the stacked ensemble plus STCAR model. Estimates are presented for 18 endemic countries and were not produced for countries deemed non-endemic (shown in grey). Multidrug resistance was defined as concurrent resistance to ampicillin, chloramphenicol, and trimethoprim–sulfamethoxazole.