Antimicrobial resistance (AMR) is a major global threat. 1 Low- and middle-income countries (LMICs) have unique challenges that may contribute to AMR. 2 In many LMICs, antimicrobials can be purchased without prescriptions and are widely available. Reference Esimone, Nworu and Udeogaranya3–Reference Mena Lora, Rojas-Fermin, Bisono, Almonte and Bleasdale6 These factors may be driving an increase in antimicrobial use (AU) and AMR in LMICs over the past decade. Reference Semret and Haraoui7 The COVID-19 pandemic has further compounded this challenge, increasing AU in both inpatient and outpatient settings. 8–11
Developing antimicrobial stewardship programs (ASPs) in LMICs is critical to help optimize antimicrobial use and curb AMR. Understanding local susceptibility patterns is a key first step toward the development of treatment guidelines for ASPs. The lack of local susceptibility data to guide therapy could lead to spiraling empiricism, furthering antimicrobial pressure and worsening AMR. Data are lacking on local susceptibility in the Dominican Republic and the impact of the COVID-19 pandemic on AMR. We sought to describe resistance in the Dominican Republic and the impact of the COVID-19 pandemic.
Methods
We performed a retrospective review of antimicrobial susceptibility reports for Escherichia coli and Pseudomonas aeruginosa (PSAR) isolates from an outpatient clinical laboratory in the Dominican Republic. A report was developed for E. coli isolates from urine and PSAR isolates from bodily fluids processed between January 1, 2018, and December 31, 2021. Isolates processed between January 1, 2018, and December 31, 2019, were considered prepandemic isolates. Isolates processed between January 1, 2020, and December 31, 2021, were considered postpandemic isolates. These 2 groups reflect the annual antibiograms before and after COVID-19. The bodily fluid category represented samples from any abscess, body cavity, or anatomical site. The report included adults and children and was generated from deidentified susceptibility data extracted from all final culture results during the study period. The University of Illinois at Chicago Institutional Review Board approved this study.
Study site
Amadita Laboratories is a private commercial laboratory company that provides outpatient services nationwide in the Dominican Republic. There are 49 outpatient laboratory sites across the Dominican Republic, representing 13 urban areas and all geographic regions of the country (Fig. 1).
Laboratory workflow and microbiology testing
Samples are collected at the outpatient laboratory sites and are processed centrally in the city of Santo Domingo. The central laboratory is equipped with the VITEK MS, VITEK 2 XL, VITEK 2 Compact (bioMèrieux, Marcy-l'Étoile, France) and advanced expert systems (AESs) for initial analysis. The VITEK 2 XL and VITEK 2 Compact are used for the initial testing. VITEK 2 XL uses the AST-GN67, AST XN05, and AST XN08 susceptibility cards, and VITEK 2 Compat uses the AST-GN67, AST XN05, and AST XN08 cards. VITEK MS is used for matrix-assisted laser desorption/ionization time-of-flight mass spectroscopy (MALDI-TOF MS) analysis of organisms. AES software versions 8.0 and 9.2 were used for analysis and interpretation of the antibiogram. Additional confirmatory testing was performed using the Kirby-Bauer method via the MASTDISCS Combi AMP-C, ESBL ID MAST GROUP, and MASTDISCS Combi CARBA PLUS discs (Mast Group, Bootle, UK). Sensitivity discs for individual carbapenems and the E-test for meropenem were used to confirm resistance.
For E. coli, a minimum inhibitory concentration (MIC) of ≥4 to ceftriaxone was considered resistant and possible extended-spectrum β-lactamase (ESBL) whereas an MIC of ≥8 μg/mL to meropenem was considered resistant (carbapenem-resistant Enterobacterales or CRE). For PSAR, an MIC of ≥32 μg/mL to ceftazidime was considered resistant and possible ESBL while an MIC of ≥8 μg/mL for meropenem was considered resistant (carbapenem-resistant Pseudomonas auriginosa or CR-PSAR). Thresholds for resistance were based on Clinical and Laboratory Standards Institute (CLSI) M100.
Results
In total, 27,718 urine cultures with E. coli were reviewed from 2018 to 2021. On average, 6929.5 urine cultures were performed per year. For E. coli, the resistance rates to ceftriaxone were 25.21% in 2018, 26.05% in 2019, 24.84% in 2020 and 24.65% in 2021 (Table 1). The resistance rates to meropenem were 0.08% in 2018, 0.13% in 2019, 0.10% in 2020, and 0.30% in 2021 (Table 1). The average E. coli with ceftriaxone resistance rates were 25.63% and 24.75% before and after the COVID-19 pandemic, respectively (Fig. 2). The average rates of E. coli with CRE were 0.11% and 0.20% before and after the COVID-19 pandemic, respectively (Fig. 3). Between 2020 and 2021, a 200% increase in carbapenem-resistant E. coli occurred.
Note. ESBL, extended-spectrum β-lactamase; CRE, carbapenem-resistant Enterobacterales.
In total, 2,111 bodily fluid cultures with PSAR were reviewed from 2018 to 2021. On average, 527.75 cultures were performed per year. For PSAR, the resistance rates to ceftazidime were 11.79% in 2018, 13.48% in 2019, 15.19% in 2020 and 15.13% in 2021 (Table 2). The resistance rates to meropenem were 2.12% in 2018, 2.53% in 2019, 2.33% in 2020, and 5.35% in 2021 (Table 2). The average rates of PSAR with ceftazidime resistance in bodily fluid were 12.64% and 15.16% before and after the COVID-19 pandemic, respectively (Fig. 2). The average rates of PSAR with carbapenem resistance were 2.33% and 3.84% before and after the COVID-19 pandemic, respectively (Fig. 3). Between 2020 and 2021, a 130% increase in CR-PSAR occurred.
Note. ESBL, extended-spectrum β-lactamase; PSAR, Pseudomonas aeruginosa; CR-PSAR carbapenem-resistant PSAR.
In conclusion, we detected high levels of resistance in E. coli isolated from urine samples and rising resistance to carbapenems in PSAR from bodily fluids. On average, resistance to ceftriaxone was present in 25.19% of E. coli isolates from urine each year. These resistance patterns suggest that ESBL is common in the Dominican Republic, at levels comparable to other countries in Latin American. Reference Gales, Castanheira, Jones and Sader12–Reference Guzmán-Blanco, Labarca, Villegas and Gotuzzo14 Carbapenem resistance was rare in both E. coli and PSAR, averaging 0.15% and 3.08% each year, respectively. However, an increase was seen after the COVID-19 pandemic, particularly in PSAR from bodily fluids. With bodily fluids as a source, this increase may be in more complex patients than those from urine samples and raises the concern of a similar effect in hospitalized patients.
In Latin America, the use of azithromycin for COVID-19 was widespread. 10,Reference Tapia15 Antibiotic use increased in hospitalized and ambulatory patients alike, despite efforts from professional societies to discourage their use. 8–10 Reports showing an increase in AMR in Latin America after the pandemic are emerging. Reference Thomas, Corso and Pasterán16 This finding mirrors a rise of AMR in the United States, where nosocomial infections by ESBL increased by 32%, CRE increased by 35% and AMR PSAR increased by 132% in the first year of the pandemic. 11
Excess antibiotic use is a significant problem in LMICs. Reference Muri-Gama, Figueras and Secoli4,Reference Shet, Sundaresan and Forsberg5 In the Dominican Republic, antimicrobials can be purchased without prescriptions and are available in pharmacies and stores. Reference Mena Lora, Rojas-Fermin, Bisono, Almonte and Bleasdale6 Aminopenicillins are commonly used and may contribute to high rates of ESBL. Studies have reported ceftriaxone resistance in >24% of E. coli at a pediatric hospital and 46% in hospitalized adults. Reference Luna, Lora, Roque, Franceschini, Perez and Cabán17,Reference Mena Lora, Rodriguez Abreu, Blanco, de Lara and Bleasdale18 Circulating ESBL genotypes in the Dominican Republic include blaCTX and blaTEM. Reference Roque, de Luna and Mena Lora19 Future ASP interventions must focus on both ambulatory and hospital settings. Treatment guidelines reflecting local susceptibilities reported herein can be an important patient safety and stewardship tool.
A strength of our study was the large number of samples and wide geographic representation of samples in the Dominican Republic. These samples may provide a valuable portrait of susceptibilities in these communities. The study also had several limitations. All cultures performed during the study period were included; thus, patients with recurrent infections may be overrepresented. The large number of isolates in the study may balance this weakness. Inferring resistance based on phenotypic resistance patterns rather than genotypes is the main limitation of our study. Ceftriaxone nonsusceptibility without ESBL genes can occur. MIC accuracy can be affected by bacterial inoculum, antimicrobial dilutions, and differing enzyme expression. Reference Villegas, Esparza and Reyes20 High inoculum can occur in severe infections and may have an impact on MICs. Ceftazidime resistance may indicate ESBL in PSAR. However, genes other than ESBL may also contribute to ceftazidime nonsusceptibility in PSAR. The lack of patient specific information, such as comorbidities or prior antibiotic use, is another weakness in our study. Despite these weaknesses, we have reported susceptibilities using the same methodology used in routine clinical practice and may provide valuable information for empiric antimicrobial selection and the development of local treatment guidelines.
Acknowledgments
The authors acknowledge the colossal efforts of healthcare workers and essential workers during the COVID-19 pandemic.
Financial support
No financial support was provided relevant to this article.
Conflicts of interest
All authors report no conflicts of interest relevant to this article.