Impact of a Rapid Blood Culture Diagnostic Test in a Children's Hospital Depends on Gram-Positive versus Gram-Negative Organism and Day versus Night Shift

Abstract

Rapid diagnostic tests (RDTs) for bloodstream infections (BSIs) decrease the time to organism identification and resistance detection. RDTs are associated with early deescalation of therapy for Gram-positive BSIs. However, it is less clear how RDTs influence antibiotic management for Gram-negative BSIs and whether RDT results are acted on during off-hours. We performed a single-center, retrospective review of children with BSI and Verigene (VG) testing at a children's hospital. Of the 301 positive cultures included in the study (196 Gram-positive, 44 Gram-negative, 32 polymicrobial, and 29 non-VG targets), the VG result had potential to impact antibiotic selection in 171 cases; among these, antibiotic changes occurred in 119 (70%) cases. For Gram-negative cultures, the Verigene result correlated with unnecessary antibiotic escalation and exposure to broader-spectrum antibiotics than needed. In contrast, for Gram-positive cultures, the VG results correlated with appropriate antibiotic selection. VG results permitted early deescalation for methicillin-susceptible Staphylococcus aureus (MSSA) (19/24 [79%]) and avoidance of antibiotics for skin contaminants (30/85 [35%]). Antibiotic changes occurred more quickly during the day than at night (4.6 versus 11.7 h, respectively; P < 0.05), and antibiotic escalations occurred more quickly than did deescalations (4.1 versus 10.1 h, P < 0.01). In a pediatric institution with a low prevalence of Gram-negative resistance, the VG RDT facilitated antibiotic optimization for Gram-positive BSIs but led to unnecessary escalation of antibiotics for Gram-negative BSIs. The time to action was slower for RDT results reported at night than during the day. Laboratories should consider these factors when implementing blood culture RDTs.

Keywords: antibiotic stewardship; bacteremia; pediatric; pediatric infectious diseases; rapid diagnostic tests.

FIG 1

Impact of Verigene result on antibiotic orders. (A) Definition of time periods for study purposes and overall antibiotic use during time periods. Period 1, time between culture collection and Gram stain result. Period 2, time between Gram stain and Verigene results. Period 3, time between Verigene result and final culture results. Period 4, time following final culture results and the completion of antibiotic therapy. Pie chart indicates the percentage (of all 301 cultures included in study) receiving 1, 2, 3+, or no antibiotics for each period. (B) Summary of antibiotic changes following Verigene results.

FIG 2

Antibiotic escalation following Verigene result. (A) Comparison between ESKAPE Gram-negative rods (GNR) (Enterobacter spp., Klebsiella pneumoniae, Acinetobacter spp., Pseudomonas aeruginosa) and all other organisms. Shown are the percentages of antibiotic changes that were escalations versus not escalation (includes deescalation and avoidance of antibiotics) in P3. P < 0.001 by chi-square analysis comparing the proportion of escalations between ESKAPE GNR and all other organisms. (B) Number of patients started on the specified antibiotics in P3, graphed by whether antibiotics were continued following the final culture results. Black bars indicate number of patients started on the antibiotic in P3 and continued on the antibiotic in P4. Gray bars indicate the number of patients started on the antibiotic in P3 but the antibiotic order was discontinued in P4. (C) Number of patients with the specified antibiotics discontinued in P3, graphed by whether antibiotics were restarted following the final culture results. Black bars indicate number of patients for whom the antibiotic was discontinued in P3 and not restarted in P4. Gray bars indicate the number of patients for whom the antibiotic was discontinued in P3 but restarted in P4. (B and C) Antibiotics are specified on the x axis labels. TZP, piperacillin-tazobactam.

FIG 3

Impact of Verigene on antibiotic selection for staphylococcal species. (A) Organisms (by final culture identification [ID]) for whom the proportion of cultures with possible antibiotic changes was significantly different from all cultures by chi-square analysis (CoNS and MRSA) or Fisher’s exact test (MSSA and MSSE). *, P < 0.05; ***, P < 0.001. All cultures, n = 301; CoNS, n = 31; MRSA, n = 26; MSSA, n = 27 (includes 1 polymicrobial culture with MSSA and CoNS); MSSE, n = 15. (B) Organisms (by final culture ID) for whom change was possible for at least 4 patients, graphed by the percentage of cases in which change was performed when possible. *, P < 0.05 by Fisher’s exact test between proportions of MSSA versus all other cultures for which change was performed. (C) Utilization of vancomycin, ampicillin, and cefazolin or nafcillin across period 1 (P1), period 2 (P2), period 3 (P3), and period 4 (P4). Bars indicate number of patients in each period receiving antibiotics, and shading specifies the antibiotic or antibiotic combination as depicted in the legend.

FIG 4

Time to antibiotic change following Verigene results. (A and B) Lengths of time between Verigene result reporting and antibiotic order discontinuation (for deescalations) or antibiotic administration (for escalations). (A) Deescalations and escalations. (B) Day shift and night shift. (A and B) Bars indicate median and 25 to 75% interquartile range, and error bars denote minimum to maximum values. Each filled circle represents an individual patient. *, P < 0.05; **, P < 0.01 by Mann-Whitney. (C) Times at which new antibiotics were administered (for escalations) or antibiotic orders were discontinued (deescalations) during period 3. Each symbol indicates an individual patient.