Prevalence of Multidrug-Resistant Diarrheagenic Escherichia coli in Asia: A Systematic Review and Meta-Analysis

Abstract

Diarrhea is one of the leading causes of morbidity and mortality in developing countries. Diarrheagenic Escherichia coli (DEC) is an important bacterial agent for diarrhea in infants, children, and international travelers, and accounts for more than 30% of diarrheal cases in children less than 5 years old. However, the choices of antimicrobial agents are now being limited by the ineffectiveness of many first-line drugs, in relation to the emergence of antimicrobial-resistant E. coli strains. The aim of this systematic review and meta-analysis was to provide an updated prevalence of antimicrobial-resistant DEC in Asia. A comprehensive systematic search was conducted on three electronic databases (PubMed, ScienceDirect, and Scopus), where 40 eligible studies published between 2010 and 2022 were identified. Using meta-analysis of proportions and a random-effects model, the pooled prevalence of DEC in Asian diarrheal patients was 22.8% (95% CI: 16.5-29.2). The overall prevalence of multidrug-resistant (MDR) and extended-spectrum beta-lactamase (ESBL)-producing DEC strains was estimated to be 66.3% (95% CI: 58.9-73.7) and 48.6% (95% CI: 35.1-62.1), respectively. Considering antimicrobial drugs for DEC, the resistance prevalence was highest for the penicillin class of antibiotics, where 80.9% of the DEC isolates were resistant to amoxicillin and 73.5% were resistant to ampicillin. In contrast, resistance to carbapenems such as imipenem (0.1%), ertapenem (2.6%), and meropenem (7.9%) was the lowest. The relatively high prevalence estimation signifies that the multidrug-resistant DEC is a public health threat. Effective antibiotic treatment strategies, which may lead to better outcomes for the control of E. coli infections in Asia, are necessary.

Keywords: Asia; Escherichia coli; antibiotic resistance; diarrheagenic; epidemiology; meta-analysis; systematic review.

Figure 1

A PRISMA flow diagram of the study selection process. Three different electronic databases were utilized to search for eligible studies reporting antimicrobial-resistant diarrheagenic Escherichia coli (DEC) using defined search strategies. Records were combined and duplicates were removed using EndNote 20 software, followed by screening against predefined eligibility criteria before inclusion in the meta-analysis.

Figure 2

Forest and funnel plots representing the pooled Asian prevalence of DEC. The estimate of prevalence was calculated by pooling 36 selected studies using the random-effects model (top panel). Four studies (Bagus Wasito, E. 2017 [21], Hoang, P. H. 2017 [27], Margulieux, K. R. 2018 [18] and Pazhani, G. P. 2011 [44]) were excluded from the analysis due to the absence of data on the sample population. The distribution of effect estimates is shown by a funnel plot (bottom panel). Figures were generated using R software.

Figure 3

Pathotypes of DEC across Asian countries: (a) Proportions of different pathotypes of DEC. Enterotoxigenic E. coli constituted the highest proportion of DEC in Asia (n = 2178; 49%), followed by EPEC (n = 1374; 31%), EAEC (n = 593; 13%), STEC (n = 208; 5%), and EIEC (n = 73; 2%). (b) Forest plot representing the pooled Asian prevalence of enteropathogenic E. coli (EPEC). (c) Forest plot representing the pooled Asian prevalence of enteroaggregative E. coli (EAEC). (d) Forest plot representing the pooled Asian prevalence of enterotoxigenic E. coli (ETEC). (e) Forest plot representing the pooled Asian prevalence of enteroinvasive E. coli (EIEC).

Figure 4

Forest and funnel plots representing the pooled prevalence of multidrug-resistant DEC in Asia. The estimate of prevalence was calculated by pooling 26 selected studies using the random-effects model (top panel). The distribution of effect estimates is shown by a funnel plot (bottom panel). Figures were generated using R software.

Figure 5

Forest and funnel plots representing the pooled prevalence of extended-spectrum β-lactamase (ESBL)-producing DEC in Asia. The estimate of prevalence was calculated by pooling 16 selected studies using the random-effects model (top panel). The distribution of effect estimates is shown by a funnel plot (bottom panel). Figures were generated using R software.