Excess Mortality Attributable to Hospital-Acquired Antimicrobial-Resistant Infections: A 2-Year Prospective Surveillance Study in Northeast Thailand

Abstract

Background: Quantifying the excess mortality attributable to antimicrobial-resistant (AMR) bacterial infections is important for assessing the potential benefit of preventive interventions and for prioritization of resources. However, there are few data from low- and middle-income countries.

Methods: We conducted a 2-year prospective surveillance study to estimate the excess mortality attributable to AMR infections for all types of hospital-acquired infection (HAI), and included bacterial species that were both locally relevant and included in the World Health Organization priority list. Twenty-eight-day mortality was measured. Excess mortality and population attributable fraction (PAF) of mortality caused by AMR infections compared to antimicrobial-susceptible (AMS) infections, adjusted for predefined confounders, were calculated.

Results: We enrolled 2043 patients with HAIs. The crude 28-day mortality of patients with AMR and AMS infections was 35.5% (491/1385) and 23.1% (152/658), respectively. After adjusting for prespecified confounders, the estimated excess mortality attributable to AMR infections was 7.7 (95% confidence interval [CI], 2.2-13.2) percentage points. This suggests that 106 (95% CI, 30-182) deaths among 1385 patients with AMR infections might have been prevented if all of the AMR infections in this study were AMS infections. The overall PAF was 16.3% (95% CI, 1.2%-29.1%). Among the bacteria under evaluation, carbapenem-resistant Acinetobacter baumannii was responsible for the largest number of excess deaths. Among all types of infection, urinary tract infections were associated with the highest number of excess deaths, followed by lower respiratory tract infections and bloodstream infections.

Conclusions: Estimating and monitoring excess mortality attributable to AMR infections should be included in national action plans to prioritize targets of preventive interventions.

Clinical trials registration: NCT03411538.

Keywords: antimicrobial resistance; excess mortality; hospital-acquired infection; nosocomial infection.

Figure 1.

Study flow diagram. Infections with methicillin-resistant Staphylococcus aureus, ampicillin-resistant Enterococcus faecium and Enterococcus faecalis, third-generation cephalosporin-resistant Escherichia coli and Klebsiella pneumoniae, and carbapenem-resistant Pseudomonas aeruginosa and Acinetobacter baumannii are defined as antimicrobial-resistant infections. Infections caused by the bacteria that was susceptible to the corresponding antibiotic of interest are defined as antimicrobial-susceptible (AMS) infections. *Of the 658 patients with AMS infection, 11 were lost to follow-up. Among these 11 patients, 5 were discharged against medical advice with health condition not improving and were assumed to have died within 28 days of follow-up since the first date of culture-positive specimen collection. **Of the 1385 patients, 28 were lost to follow-up. Among these 28 patients, 23 were discharged against medical advice with health condition not improving, and were assumed to have died within 28 days of follow-up since the first date of culture-positive specimen collection. Abbreviations: AMR, antimicrobial-resistant; AMS, antimicrobial-susceptible; HAI, hospital-acquired infection.

Figure 2.

A level plot illustrating number of patients with hospital-acquired infections by each causative bacteria and type of infection. “>1 type of infection” is defined as infections that occurred with ≥2 types of infection including bloodstream infection (BSI), lower respiratory tract infection (LRTI), surgical site infection (SSI), urinary tract infection (UTI), and infections at other body sites (OTH). *Polymicrobial is defined as infections with >1 of the bacteria included in the evaluation.

Figure 3.

Expected mortality if all infections caused by each causative bacteria were antimicrobial susceptible (AMS; gray) or antimicrobial resistant (black). Excess mortality (percentage points; 95% confidence interval) is defined as the mortality in the study cohort that would be prevented if the infections of the specific type were AMS infections, adjusted for the predefined confounding factors. Polymicrobial is defined as infections with >1 of the bacteria in the evaluation. Abbreviations: 3GCREC, third-generation cephalosporin-resistant Escherichia coli; 3GCRKP, third-generation cephalosporin-resistant Klebsiella pneumoniae; 3GCSEC, third-generation cephalosporin-susceptible Escherichia coli; 3GCSKP, third-generation cephalosporin-susceptible Klebsiella pneumoniae; AMPREfc, ampicillin-resistant Enterococcus faecalis; AMPREfm, ampicillin-resistant Enterococcus faecium; AMPSEfc, ampicillin-susceptible Enterococcus faecalis; AMPSEfm, ampicillin-susceptible Enterococcus faecium; AMR, antimicrobial resistant; AMS, antimicrobial-susceptible; CRAB, carbapenem-resistant Acinetobacter baumannii; CRPA, carbapenem-resistant Pseudomonas aeruginosa; CSAB, carbapenem-susceptible Acinetobacter baumannii; CSPA, carbapenem-susceptible Pseudomonas aeruginosa; MRSA, methicillin-resistant Staphylococcus aureus; MSSA, methicillin-susceptible Staphylococcus aureus.

Figure 4.

Expected mortality if all infections for each type of infection were antimicrobial susceptible (AMS; gray) or antimicrobial resistant (black). Excess mortality (percentage points; 95% confidence interval) is defined as the mortality in the study cohort that would be prevented if the infections of the specific type were AMS infections, adjusted for the predefined confounding factors. Abbreviations: AMR, antimicrobial-resistant; AMS, antimicrobial-susceptible; BSI, bloodstream infection; LRTI, lower respiratory tract infection; OTH, other infections; SSI, surgical site infection; UTI, urinary tract infection; >1 type of infection, infections that occurred with ≥2 types including BSI, LRTI, SSI, UTI, and OTH.