Accuracy and reliability of direct disc diffusion antibiotic susceptibility test from flagged-positive of blood culture

Abstract

Introduction: Antibiotic susceptibility tests (AST) done on blood cultures are critical for the treatment of patients suspected to be suffering from bloodstream infection. The objective of this study was to evaluate the accuracy and reliability of disc diffusion AST conducted directly (direct AST) from flagged-positive blood cultures, especially for Gram-positive cocci bacteria.

Methodology: This study compared direct AST with conventional AST (broth micro-dilution). The antibiotics studied were piperacillin/tazobactam, gentamicin, ceftazidime, erythromycin, and penicillin. Accuracy was determined by calculating very major, major, and minor errors. The reliability was determined by categorical agreement and weighted Kappa index.

Results: Gram-positive cocci bacteria were grown in pairs of blood culture bottles and tested with the two methods of AST. No very major errors were detected among the five types of antibiotics. Major errors of 2.56% and minor errors of 4.93% were found when testing gentamicin. The major and minor errors when testing erythromycin were 2.85% and 1.23%, respectively. Perfect agreements (categorical agreement: 100%; weighted Kappa index: 1) of the two AST methods were observed with piperacillin/tazobactam, ceftazidime, and penicillin. Almost perfect agreement was found with gentamicin and erythromycin. Categorical agreement results when testing antibiotics gentamicin and erythromycin were 93.83% and 97.53%, respectively. In addition, the weighted-Kappa index when testing these two antibiotics were 0.92 and 0.96, respectively.

Conclusions: The accuracy and reliability of the direct AST was within acceptable limits.