Objective: To determine the feasibility and benefit of implementing the World Health Organization-recommended Global Antimicrobial Resistance Surveillance System (GLASS) for the surveillance of sputum specimens collected from patients at Siriraj Hospital.

Materials and Methods: All sputum specimens sent for culture between December 2016 and June 2017 at Siriraj Hospital were retrieved from the microbiology laboratory. A locally-developed web application program was used to transfer the sputum culture results into GLASS and to enter the clinical data of the patients with positive sputum cultures. The relevant clinical data of each patient with a positive sputum culture were collected, including the nature of the reported organisms, acquisition of infection, type and severity of infection, concordance of antibiotic treatment, outcomes of treatment, hospitalization cost, and in-hospital mortality. These data were extracted from the medical records and hospital database.

Results: Three hundred eighty-one patients with positive cultures for 1,050 bacterial isolates from 2,367 sputum specimens collected during the present study period were included. The most common isolated bacteria were A. baumannii, followed by P. aeruginosa, S. maltophilia, K. pneumoniae, and S. aureus. Among the 1,050 bacterial isolates, the rate of true infection was 58%, with P. aeruginosa and A. baumannii, the common causes of pneumonia. The rate of colonization was 42% with A. baumannii and S. maltophilia, the most common colonized bacteria. Most of the bacteria isolated from the sputum specimens were from patients who had hospital-associated infections (HAI, 70.1%), which the most common causative bacteria were A. baumannii (38.1%) and P. aeruginosa (34%); whereas methicillin-susceptible S. aureus (36.1%) and K. pneumoniae (29.1%) were observed in patients with community-associated infection (CAI). Among the patients with HAI, 60% had ventilator-associated pneumonia (VAP), particularly late-onset VAP (51.8%), and hospital-acquired pneumonia (HAP, 34%). The patients with HAI significantly had higher rates of sepsis (p<0.001) and of receiving non-concordant empirical antibiotics (p<0.001), more unfavorable outcomes (p<0.001), a longer length of hospital stay (p<0.001), and higher hospitalization costs (p<0.001) and higher in-hospital mortality (p<0.001) than those with CAI. For the antibiotic susceptibility profiles, K. pneumoniae colonization isolates were more resistant to antibiotics than the isolates causing true infections. Overall mortality of the patients with infections was 42.8%. In-hospital mortality of the patients with HAI, HAP, and late-onset VAP caused by antimicrobial resistant (AMR) bacteria was significantly higher than those with such infections caused by non-AMR bacteria.

Conclusion: GLASS provides more applicable and more reliable data for the AMR surveillance of sputum specimens than conventional laboratorybased surveillance in terms of the nature of the reported organisms, acquisition of infection, type and severity of infection, antibiotic susceptibility of isolated bacteria, concordance of antibiotic treatment, and the burden of respiratory tract infections. The information on HAP and VAP is useful for developing local clinical guidelines for choosing appropriate empirical antibiotics. However, GLASS has limitations in the AMR surveillance of sputum specimens from patients with CAP and it requires more time and resources than laboratory-based surveillance.

Keywords: Global antimicrobial resistance surveillance system, GLASS, Sputum specimens

Received 22 Jan 2020 | Revised 5 Feb 2020 | Accepted 7 Feb 2020