Evaluation of rapid polymerase chain reaction-based organism identification of gram-positive cocci for patients with a single positive blood culture

摘要

For patients with a single-positive blood culture growing gram-positive cocci, organism identification can provide supportive information for differentiating contamination from infection. We investigated the effect of a rapid blood culture identification panel (BCID) on vancomycin-prescribing patterns and patient outcomes for single positive blood culture (PBC) growing gram-positive cocci. Adult patients with single-positive blood culture growing gram-positive cocci with conventional organism identification (pre-BCID) were compared with organism identification by BCID (post-BCID). Antimicrobial Stewardship Program (ASP) review of PBC was performed in both study groups. Vancomycin prescribing patterns were studied. Secondary endpoints were the incidence of nephrotoxicity, length of stay (LOS), readmission rate, mortality, and hospital costs. A total of 188 patients (86 pre-BCID, 102 post-BCID) were included. Organism identification was known 21 h sooner in the post-BCID group (P < 0.001). Coagulase-negative staphylococci were the most commonly isolated organisms (73%). In patients where vancomycin was deemed unnecessary (n = 133), vancomycin use (51% pre-BCID vs 36% post-BCID; P = 0.09) and time from culture positivity to vancomycin discontinuation (1.5 vs. 1.7 days; P = 0.92) did not differ between groups. We found no differences in the development of nephrotoxicity, LOS, readmission, mortality, or hospital costs. Earlier identification of single positive blood culture growing gram-positive cocci did not significantly influence prescribing patterns of vancomycin. However, baseline antimicrobial stewardship review of single positive blood culture growing gram-positive cocci may have lessened the opportunity for detectable differences. Larger studies, accounting for the impact of ASP intervention, should be performed to determine the value of each individual component.