Antimicrobial drug resistance in Salmonella Typhi and Paratyphi isolates worldwide, 1990 to 2017: A systematic review of the literature

Abstract

Salmonella Typhi and Paratyphi are WHO priority pathogens that cause serious bloodstream infections. Antimicrobial resistance presents a significant public health burden as it negatively impacts our ability to treat and control enteric fever.

The aim of this project is to map the spatial distribution of drug resistant S. Typhi and Paratyphi and to incorporate the impact of AMR into the Global Burden of Disease (GBD) study estimates.

Methods & Materials

We conducted a systematic review of the global published literature between 1990 and 2017 to analyse the prevalence of drug resistance in S. Typhi and Paratyphi blood culture isolates.

Antimicrobial susceptibility data to all tested drugs were extracted. We calculated the median prevalence (percentage) of resistance according to serotype, GBD super-region and five-year time-period.

Results

Data from 518 studies (1990-2015) have been analysed to date; heterogeneity was high (I2 > 80%) within most subgroups.

Between 1990-1994 and 2010-2015 the median prevalence (MP) of multidrug resistant (MDR; resistance to chloramphenicol, co-trimoxazole and ampicillin) Typhi decreased in South Asia and North Africa/Middle-East from 42% [IQR 23-66] to 6% [0-19] and 43% [34-48] to 7% [0-33], respectively; remained constant in Southeast Asia, East Asia and Oceania; and increased in sub-Saharan Africa from 6% [3-8] to 33% [4-83]. The MP of MDR Paratyphi was less than 11% for all time-periods in South Asia.

The MP of nalidixic acid resistant (NAR; an indicator of fluoroquinolone resistance) Typhi increased from < 5% in 1990-1994 to > 80% in 2010-2015 in South Asia and Southeast Asia but remained low in sub-Saharan Africa (2% [1-3] in 1990-1994; 7% [4-28] in 2010-2015). The MP of NAR Paratyphi was > 81% for all time-periods in South Asia.

Conclusion

Our results indicate that Salmonella Typhi strains in South Asia and North Africa/Middle East are regaining sensitivity to first-line drugs; fluoroquinolone resistance is now highly prevalent in South and Southeast Asia.

High heterogeneity indicates further spatial-temporal variation exists alongside multiple biases. We intend to adjust for biases and develop a spatial-temporal model to produce accurate temporal national and subnational estimates of resistance. These results will be incorporated into the GBD study, providing a vital source of information.