Abstract
Background
Our laboratory uses matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI) and the VITEK 2 system (DV2) directly from positive blood cultures (BC) for organism identification (ID) and antimicrobial susceptibility testing (AST). Our objective was to compare direct MALDI–DV2 with a commercial BC ID–AST platform, the Accelerate Pheno system (AXDX), in the ID–AST of clinical and seeded BC positive for gram-negative bacilli (GNB).
Methods
BC positive for GNB were collected over a 3-mo period and tested using AXDX and direct MALDI–DV2 and compared with conventional methods. A subset of sterile BC were seeded with multi-drug-resistant GNB.
Results
Twenty-nine clinical samples and 35 seeded samples were analyzed. Direct MALDI had a higher ID failure rate (31.0%) than AXDX (3.4%; p < 0.001). Time to ID–AST was 1.5–6.9 h, 5.8–16.5 h, and 21.6–33.0 h for AXDX, direct MALDI–DV2, and conventional methods, respectively (p < 0.001). For clinical samples, AXDX and DV2 had essential agreement (EA) or categorical agreement (CA) of more than 96%. For seeded samples, AXDX had EA, CA, VME, ME, and minor error (mE) of 93.2%, 89.0%, 2.2%, 0%, and 9.2%, respectively. AXDX had a large number of non-reports (6.1%) stemming from meropenem testing. DV2 had EA, CA, VME, ME, and mE of 97.5%, 94.7%, 1.3%, 0%, and 4.1%, respectively.
Conclusions
Direct MALDI–DV2 and AXDX both had high agreement for clinical samples, but direct MALDI–DV2 had higher agreement when challenged with MDR GNB.
Keywords: Accelerate pheno system, blood culture, direct MALDI, direct VITEK, susceptibility
Abstract
Historique
Le laboratoire des chercheurs fait appel à la spectrométrie de masse à temps de vol par désorption/ionisation laser assistée par matrice (MALDI) et au système VITEK 2 (DV2) directement à partir des hémocultures positives pour identifier les organismes et procéder aux antibiogrammes. Les chercheurs avaient l’objectif de comparer les MALDI–DV2 directs avec le système Accelerate Pheno (AXDX), une plateforme commerciale d’hémoculture, d’identification et d’antibiogramme, pour identifier les organismes et procéder aux antibiogrammes dans les hémocultures cliniques et les hémocultures ensemencées positives aux bacilles à Gram négatif (BGN).
Méthodologie
Les chercheurs ont colligé les hémocultures positives aux BGN sur une période de trois mois, ont procédé au dépistage à l’aide du système AXDX et des MALDI–DV2 directs et ont comparé les résultats aux méthodes classiques. Ils ont ensemencé un sous-groupe d’hémocultures stériles avec des BGN multirésistantes.
Résultats
Les chercheurs ont analysé 29 échantillons cliniques et 35 échantillons ensemencés. La MALDI directe présentait un taux d’échec d’identification plus élevé (31,0 %) que le système AXDX (3,4 %; p < 0,001). Il fallait de 1,5 à 6,9 heures, de 5,8 à 16,5 heures et de 21,6 à 33,0 heures pour identifier les organismes et procéder aux antibiogrammes à l’aide du système AXDX, des MALDI–DV2 directs et des méthodes classiques, respectivement (p < 0,001). Pour les échantillons cliniques, les systèmes AXDX et DV2 présentaient une concordance essentielle (CE) ou catégorique (CA) de plus de 96 %. Pour les échantillons ensemencés, le système AXDX présentait une CE, une CA, des erreurs très majeures (ETM), des erreurs majeures (EM) et des erreurs mineures (Em) de 93,2 %, 89,0 %, 2,2 %, 0 % et 9,2 %, respectivement. Le système AXDX était lié à un grand nombre d’absences de rapports (6,1 %) découlant des tests de méropénem. Le système DV2 présentait une CE, une CA, des ETM, des EM et des Em de 97,5 %, 94,7 %, 1,3 %, 0 % et 4,1 %, respectivement.
Conclusions
Les MALDI–DV2 directs et le système AXDX présentaient une concordance élevée pour les échantillons cliniques, mais les MALDI–DV2 directs étaient associés à une concordance plus élevée en cas de provocation par des BGN multirésistantes.
Mots-clés: hémoculture, MALDI directe, susceptibilité, système Accelerate Pheno, VITEK direct
Introduction
Time to antimicrobial identification (ID) and antibiotic susceptibility testing (AST) is a critical factor for patients with bloodstream infections (BSI) due to gram-negative enteric bacilli, especially in this era of ever-increasing antimicrobial resistance (1). Conventional microbiologic evaluation using phenotypic methods takes approximately 36–48 h from the time a blood culture flags positive to the reporting of final antimicrobial ID and AST results.
Different methods and commercial platforms are available to help reduce the turnaround time in ID and AST for positive blood cultures. For instance, using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS, abbreviated henceforth as MALDI) and the VITEK 2 system directly from positive blood culture isolates has decreased the organism ID and AST turnaround time in our laboratory to 18–24 h (ie, a full 1–1.5 d faster than the conventional pipeline). The Accelerate PhenoTest BC kit (AXDX; Accelerate Diagnostics Inc., Tucson, AZ), however, has been shown to reduce the turnaround time for final organism ID and AST from positive blood cultures to approximately 7 h, as long as the final organism is 1 of the 16 organisms detectable by AXDX (2).
For the 16 organisms detectable by AXDX, AXDX has a reported sensitivity of 97.4% and specificity of 99.3% for organism ID and 96.3% essential agreement (EA) and categorical agreement (CA) for AST compared with broth microdilution (3). Several studies have also found that AXDX was more rapid and more accurate for ID and AST of gram-negative BSI than MALDI or automated susceptibility testing from short-incubation cultures (4,5). In addition, AXDX can be run with minimum sample preparation, hands-on assay time, or training, making it feasible to run on nighttime shifts in our high-volume operations when there may be only one highly skilled technologist present.
The use of direct MALDI and VITEK 2 analyses of positive blood cultures to report a final ID and AST, respectively, has been shown to be non-inferior to conventional methods (6,7). This approach has also been documented to reduce mortality, decrease hospital length of stay, and decrease the time to prescription of optimal antibiotic therapy (8–11). Our laboratory recently implemented direct MALDI and VITEK 2 using a lysing procedure with an ammonium chloride solution for all positive blood cultures. For gram-negative bacilli, we demonstrated sensitivity (86%) and specificity (100%) for pathogen ID using direct MALDI and an EA (96.4%) and CA (99.0%) for direct VITEK 2 compared with our previous conventional protocol (12).
This study compares the performance of the AXDX platform with direct MALDI–VITEK 2 for the reporting of a final ID and AST on clinical and seeded blood culture samples positive for gram-negative enteric bacilli. The potential impact of more rapidly reporting final ID and AST using the AXDX system was estimated for patients with BSI due to gram-negative enteric organisms.
Methods
Calgary Laboratory Services (CLS) provides centralized laboratory services for the city of Calgary and surrounding areas, with a catchment population of more than 1.6 million individuals. Positive blood cultures identified as gram-negative bacilli on the basis of Gram stain were included if reported from 8:00 a.m. to 4:00 p.m. Monday through Friday in our laboratory. Only aerobic and facultatively anaerobic gram-negative bacilli were included in the analysis. Pellets from positive blood culture bottles, incubated using the BACT/ALERT VIRTUO system (bioMérieux, Laval, QC), were prepared using the protocol outlined in Prod’hom et al (13), with several modifications. First, 1.5 mL of positive blood culture medium was transferred into a microcentrifuge tube, and the medium sat for at least 30 seconds before 1.25 mL of medium was transferred into a 15 mL tube containing 11.25 mL of sterile water. The solution was then centrifuged at 1,000 × g for 10 minutes followed by removal of the supernatant and by the addition of 1 mL of an ammonium chloride solution (made up of 4 g of NH4Cl and 0.05 g of KHCO3 dissolved in 500 mL of distilled H2O). The pellet was subsequently re-suspended using a pipette, transferred into a microcentrifuge tube, and centrifuged at 200 × g for 10 min. The supernatant was again removed, and the pellet was re-suspended by adding 1 mL of sterile water and then vortexed. The solution was again centrifuged at 200 × g for 10 min, the supernatant discarded, and the pellet re-suspended in 200 μL of sterile water and then vortexed.
For direct MALDI analyses, MALDI plates were spotted with 1μl of pellet suspension and overlaid with 1 μl of 70% formic acid and allowed to dry. Matrix consisting of 28% acetonitrile was then applied and dried before mass spectrometry. Proteomics analyses using MALDI (VITEK MS) was then performed using standard procedures as outlined by the manufacturer. For direct susceptibility testing, a 0.5 McFarland standard was created from the final pellet solution, and AST using VITEK 2 was then performed using standard procedures outlined by the manufacturer.
Medium from each positive blood culture was also simultaneously inoculated onto sheep’s blood agar, chocolate agar, and MacConkey agar for 18–24 h to perform conventional ID methods using MALDI and VITEK 2 on the recovered isolate. Testing on MALDI and VITEK 2 were done using standard methods according to their respective manufacturer’s instructions. The Accelerate Pheno system (software version 1.3.1.15) using AXDX was inoculated and run according to the manufacturer’s instructions within 8 h of the time the blood culture flagged positive. Serial blood cultures from the same patient were excluded.
For the seeded samples, stored isolates were recovered from the CLS biorepository of multi-drug-resistant gram-negative enteric organisms, seeded into sterile blood culture bottles, and processed as outlined earlier. A total of 35 seeded samples of carbapenemase-resistant Escherichia coli, Klebsiella pneumoniae, K. oxytoca, or Enterobactae cloacae were included in the study. The samples included the following types of carbapenemase resistance: 5 NDMs, 4 VIMs, 4 IMPs, 9 OXAs, 6 KPC, 1 GES, 1 OXA/NDM, and 5 non-carbapenemase isolates with impermeability to carbapenems due to porin mutations. The enzymes were previously characterized with whole genome sequencing as part of several ongoing global molecular epidemiology studies using procedures previously described (14).
Time to ID and AST for direct and conventional MALDI–VITEK 2 was determined by comparing the time from Gram stain result to time of reporting of the final ID and AST results. Time to ID and AST for AXDX was based on the system run time but does not include other parameters such as time to instrument loading. At the time of the study, AXDX was not implemented in our laboratory and was not part of its normal workflow. Instead, samples for AXDX testing were prepared by author LC or author WS within several hours of Gram stain result. Preparation of sample and loading onto the AXDX instrument took, at maximum, 15 min to complete. Overall ID agreement for direct MALDI and AXDX was calculated by comparing the ID result with our reference standard (conventional MALDI).
AST results were interpreted as per the 2018 Clinical and Laboratory Standards Institute (CLSI) guidelines (15). EA, CA, very major error (VME), major error (ME), and minor error (mE) were calculated by comparing AST results with our reference standard (conventional VITEK 2 system). Amoxicillin–clavulanate, ampicillin, cefoxitin, and trimethoprim–sulfamethoxazole (TMP-SMX) were not compared because they were not available on the tested AXDX panel currently approved by the US Food and Drug Administration. The VITEK 2 system is unable to differentiate cefazolin intermediate from susceptible because of changes in recommended minimum inhibitory concentration (MIC) breakpoints for cefazolin (15). For the purposes of this study, a VITEK 2 cefazolin MIC reading of less than 4 was considered intermediate. Any major discrepancy (ME or VME) between direct VITEK 2 or AXDX compared with our reference standard (conventional VITEK 2 system) was resolved using broth microdilution on bacterial colonies that grew on blood agar using CLSI reference methods (15).
Isolates from conventional MALDI–VITEK 2 with cefoxitin MIC of 32 or higher were further assessed for the presence of ampC β-lactamases by the commercial D69C AmpC Detection Disc Set (Mast Group Ltd, Merseyside, UK). Isolates with ceftriaxone or ceftazidime MIC of 2 or higher were further assessed for the presence of extended-spectrum β-lactamases (ESBL) by the disk diffusion procedure described in the M100 CLSI guideline (15).
Patient data including demographics, antibiotic usage, changes in antibiotic management at time of AST results, and clinical outcomes were collected through electronic chart review.
Our sample size was 65 (30 clinical isolates, 35 seeded isolates) and was based on the number of AXDX kits supplied by Accelerated Diagnostics. Statistical analysis was conducted using Stata (version 14.1; StataCorp, College Station, TX). Categorical variables were compared using Pearson χ2 or Fisher exact test as appropriate. Continuous variables were compared using Wilcoxon signed rank tests.
This protocol was approved by the University of Calgary Conjoint Health Research Ethics Board (no. REB17-1705).
Results
Clinical samples
A total of 30 clinical samples were tested in the study. One isolate was excluded from ID agreement, time comparison, and AST comparison because it was identified as an anaerobe (Bacteroides fragilis), and another was excluded from AST comparison because it failed to provide an AST on AXDX, leaving 28 clinical samples.
The ID agreement and number of failures are provided in Table 1. The number of failures was high for the direct MALDI (31.0%) compared with AXDX (3.4%). AST failures were similar between AXDX and direct VITEK (p = 0.32). Compared with conventional MALDI–VITEK 2, direct MALDI–VITEK 2 was more rapid for both ID (mean 5.8 versus 21.6 h, p < 0.001) and AST (mean 16.5 versus 33.0 h, p < 0.001). AXDX obtained faster ID (mean 1.5 h) and AST (mean 6.9 h) results than direct MALDI–VITEK 2 (p < 0.001).
Table 1: ID agreement and number of failures for AXDX and DM compared with conventional MALDI
Note: p < 0.001 for all comparisons
| Sample type and method | No. (%), 95% CI | |
|---|---|---|
| Agreement | Non-report | |
| Clinical (n = 29) | ||
| AXDX | 28 (100) | 1 (3.4), −3.6 to 10.5 |
| DM | 20 (100) | 9 (31.0), 13.1 to 48.9 |
| Seeded (n = 35) | ||
| AXDX | 31 (93.9), 79.8 to 99.3 | 2 (5.7), −2.4 to 13.8 |
| DM | 19 (90.5), 76.8 to 104.2 | 14 (40.0), 22.9 to 57.1 |
AXDX = Accelerate Pheno™ system; DM = Direct matrix-assisted laser desorption/ionization time-of-flight mass spectrometry; ID = Identification
For the clinical samples, AXDX had overall EA, CA, and mE of 97.0%, 96.3%, and 3.7%, respectively. Direct VITEK 2 had overall EA, CA, and mE of 99.3%, 99.3%, and 0.7%, respectively. When comparing individual antibiotics within the AXDX and direct VITEK 2 systems, each antibiotic had an EA and CA of more than 90% with the exception of tobramycin for direct VITEK 2 (three minor errors, corresponding to a CA of 89%). Moreover, both systems had no VME or ME. Further details are provided in Table 2.
Table 2: AXDX and direct VITEK 2 system AST results for clinical samples and seeded samples
| No. (%)* | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| System and antibiotic | EA | CA | VME | ME | mE | NR | S | I | R |
| Clinical samples (n = 28**) | |||||||||
| AXDX | |||||||||
| Amikacin | 27 (96.4) | 28 (100) | 0 | 0 | 0 | 0 | 28 | 0 | 0 |
| Cefazolin† | 23 (100) | 21 (91.3) | 0 | 0 | 2 (8.7) | 0 | 0 | 19 | 4 |
| Cefepime | 26 (92.9) | 26 (92.9) | 0 | 0 | 2 (7.1) | 0 | 27 | 1 | 0 |
| Ceftazidime | 26 (92.9) | 26 (92.9) | 0 | 0 | 2 (7.1) | 0 | 27 | 0 | 1 |
| Ceftriaxone | 27 (100) | 27 (100) | 0 | 0 | 0 | 1 | 24 | 0 | 3 |
| Ciprofloxacin | 26 (92.9) | 27 (96.4) | 0 | 0 | 1 (3.6) | 0 | 24 | 0 | 4 |
| Ertapenem | 27 (100) | 27 (100) | 0 | 0 | 0 | 1 | 27 | 0 | 0 |
| Gentamicin | 28 (100) | 28 (100) | 0 | 0 | 0 | 0 | 25 | 0 | 3 |
| Meropenem | 27 (96.4) | 28 (100) | 0 | 0 | 0 | 0 | 28 | 0 | 0 |
| Piperacillin or tazobactam | 27 (96.4) | 27 (96.4) | 0 | 0 | 1 (3.6) | 0 | 27 | 0 | 1 |
| Tobramycin | 28 (100) | 25 (89.3) | 0 | 0 | 3 (10.7) | 0 | 24 | 3 | 1 |
| All antibiotics | 292 (97.0), 95% CI 94.3% to 98.5% | 290 (96.3), 95% CI 93.5% to 98.0% | 0 | 0 | 11 (3.7), 95% CI 2.0% to 6.6% | 2 (0.7) | 261 (86.7) | 23 (7.6) | 17 (5.6) |
| Direct VITEK 2 system | |||||||||
| Amikacin | 28 (100) | 28 (100) | 0 | 0 | 0 | 0 | 28 | 0 | 0 |
| Cefazolin | 28 (100) | 28 (100) | 0 | 0 | 0 | 0 | 0‡ | 22 | 6 |
| Cefepime | 27 (96.4) | 27 (96.4) | 0 | 0 | 1 (3.6) | 0 | 27 | 0 | 1 |
| Ceftazidime | 28 (100) | 28 (100) | 0 | 0 | 0 | 0 | 27 | 0 | 1 |
| Ceftriaxone | 27 (100) | 27 (100) | 0 | 0 | 0 | 1 | 24 | 0 | 3 |
| Ciprofloxacin | 27 (96.4) | 27 (96.4) | 0 | 0 | 1 (3.6) | 0 | 23 | 1 | 4 |
| Ertapenem | 27 (100) | 27 (100) | 0 | 0 | 0 | 1 | 27 | 0 | 0 |
| Gentamicin | 28 (100) | 28 (100) | 0 | 0 | 0 | 0 | 25 | 0 | 3 |
| Meropenem | 28 (100) | 28 (100) | 0 | 0 | 0 | 0 | 28 | 0 | 0 |
| Piperacillin or tazobactam | 28 (100) | 28 (100) | 0 | 0 | 0 | 0 | 27 | 0 | 1 |
| Tobramycin | 28 (100) | 28 (100) | 0 | 0 | 0 | 0 | 24 | 3 | 1 |
| All antibiotics | 304 (99.3), 95% CI 97.5% to 99.98% | 304 (99.3), 95% CI 97.5% to 99.98% | 0 | 0 | 2 (0.7), 95% CI 0.02% to 2.7% | 2 (0.7) | 282 (92.2) | 4 (1.3) | 20 (6.5) |
| Seeded samples (n = 29**) | |||||||||
| AXDX | |||||||||
| Amikacin | 25 (86.2) | 25 (86.2) | 2 (28.6) | 0 | 2 (6.9) | 0 | 24 | 1 | 4 |
| Cefazolin† | 20 (100) | 20 (100) | 0 | 0 | 0 | 1 | 0 | 0 | 20 |
| Cefepime | 24 (82.8) | 22 (75.9) | 1 (5.6) | 0 | 6 (20.7) | 0 | 7 | 3 | 19 |
| Ceftazidime | 28 (96.6) | 26 (89.7) | 0 | 0 | 3 (10.3) | 0 | 2 | 2 | 25 |
| Ceftriaxone | 27 (96.4) | 26 (92.9) | 0 | 0 | 2 (7.1) | 1 | 3 | 1 | 24 |
| Ciprofloxacin | 29 (100) | 27 (93.1) | 0 | 0 | 2 (6.9) | 0 | 3 | 0 | 26 |
| Ertapenem | 24 (88.9) | 23 (85.2) | 2 (8.0) | 0 | 2 (7.4) | 2 | 5 | 1 | 21 |
| Gentamicin | 27 (93.1) | 26 (89.7) | 0 | 0 | 3 (10.3) | 0 | 15 | 2 | 12 |
| Meropenem | 13 (92.9) | 14 (100) | 0 | 0 | 0 | 15 | 1 | 0 | 13 |
| Piperacillin or tazobactam | 27 (93.1) | 25 (86.2) | 0 | 0 | 4 (13.8) | 0 | 4 | 0 | 25 |
| Tobramycin | 28 (96.6) | 26 (89.7) | 0 | 0 | 3 (10.3) | 0 | 7 | 7 | 15 |
| All antibiotics | 272 (93.2), 95% CI 89.6% to 95.6% | 260 (89.0), 95% CI 84.9% to 92.2% | 5 (2.2), 95% CI 0.8% to 5.0% | 0 0% | 27 (9.2), 95% CI 6.4% to 13.1% | 19 (6.1), 95% CI 3.9% to 9.4% | 71 (24.3) | 17 (5.8) | 204 (69.9) |
| Direct VITEK 2 system | |||||||||
| Amikacin | 28 (96.6) | 27 (93.1) | 1 (14.3) | 0 | 1 (3.4) | 0 | 22 | 2 | 5 |
| Cefazolin | 29 (100) | 29 (100) | 0 | 0 | 0 | 0 | 0‡ | 0 | 29 |
| Cefepime | 28 (96.6) | 27 (93.1) | 0 | 0 | 2 (6.9) | 0 | 12 | 3 | 14 |
| Ceftazidime | 29 (100) | 29 (100) | 0 | 0 | 0 | 0 | 3 | 0 | 26 |
| Ceftriaxone | 28 (96.6) | 28 (96.6) | 0 | 0 | 1 (3.4) | 0 | 2 | 0 | 27 |
| Ciprofloxacin | 29 (100) | 29 (100) | 0 | 0 | 0 | 0 | 3 | 2 | 24 |
| Ertapenem | 28 (96.6) | 28 (96.6) | 1 (4.0) | 0 | 0 | 0 | 3 | 1 | 25 |
| Gentamicin | 29 (100) | 27 (93.1) | 0 | 0 | 2 (6.9) | 0 | 15 | 3 | 11 |
| Meropenem | 25 (89.3) | 23 (82.1) | 1 (4.5) | 0 | 4 (14.3) | 1 | 6 | 3 | 19 |
| Piperacillin or tazobactam | 28 (96.6) | 27 (93.1) | 0 | 0 | 2 (6.9) | 0 | 3 | 2 | 24 |
| Tobramycin | 29 (100) | 28 (96.6) | 0 | 0 | 1 (3.4) | 0 | 7 | 7 | 15 |
| All antibiotics | 310 (97.5), 95% CI 95.0% to 98.8% | 302 (94.7), 95% CI 91.6% to 96.7% | 3 (1.3), 95% CI 0.3% to 3.9% | 0 (0) | 13 (4.1), 95% CI 2.4% to 7.3% | 1 (0.3) | 76 (23.9) | 23 (7.2) | 219 (68.9) |
Note: MICs were truncated to overlapping reportable range before analysis
*Unless otherwise specified
**For clinical samples, 1 isolate was excluded from AST comparison because it failed to provide an AST on AXDX. For seeded samples, 6 isolates were excluded from AST comparison (2 failed to identify on AXDX, 2 failed to provide AST with AXDX, and 2 failed to provide AST with direct VITEK 2)
† AXDX only reports cefazolin for E. coli and Klebsiella spp
‡VITEK 2 system unable to differentiate cefazolin I from S. For the purposes of this study, a VITEK 2 system cefazolin MIC reading of <4 was considered intermediate AST = antimicrobial susceptibility testing; AXDX = Accelerate PhenoTest™ BC kit; CA = Categorical agreement; EA = Essential agreement; I = Intermediate; mE = Minor error; ME = Major error; MIC = Minimum inhibitory concentration; NR = Non-report; R = Resistant; S = Susceptible; VME = Very major error
Details of patient characteristics from clinical samples are provided in Table 3. The mean age was 68 y, and 54% were male. Eleven percent of patients had nosocomial bacteremia, with urosepsis being the main diagnosis (50%). Shorter time to ID and AST may have had a major impact in 4 of 28 (14%) clinical cases in which the organism isolated was resistant to the empiric antibiotics used. Empiric antibiotics were escalated on the basis of organism ID (2 of 4 cases; 50%) and AST (2 of 4 cases; 50%). Once AST was known, 57% of patients had their antibiotics narrowed.
Table 3:
Clinical samples—baseline characteristics and changes to antibiotics made after AST results (N = 28)
| Characteristic | % of patients* |
|---|---|
| Age, y, mean (SD), range | 68 (3), 24–100 |
| Gender | |
| Female | 46 |
| Male | 54 |
| Diagnosis | |
| Urosepsis | 50 |
| Fever NYD | 21 |
| Liver or biliary sepsis | 11 |
| Other | 18 |
| Nosocomial bacteremia | 11 |
| Death ≤3 mo | 11 |
| Organism identified | |
| E. coli | 61 |
| K. pneumoniae | 14 |
| K. oxytoca | 7 |
| P. mirabilis | 7 |
| P. aeruginosa | 4 |
| E. cloacae | 4 |
| C. freundii | 4 |
| Empiric antibiotics used | |
| Ceftriaxone | 62 |
| Piperacillin or tazobactam | 31 |
| Ertapenem | 4 |
| Meropenem | 4 |
| Organism resistant to empiric antibiotics | 14 |
| Antibiotics narrowed after AST results | 57 |
Note: Nosocomial bacteremia: positive blood cultures (for gram-negative bacilli) that were drawn >48 h after hospital admission. Clinical data were unknown for 1 patient.
* Unless otherwise specified
AST = Antimicrobial susceptibility testing; NYD = Not yet determined
Seeded samples
A total of 35 seeded isolates were tested in the study. Two isolates were excluded from ID comparison because they failed to ID on AXDX and therefore could not provide ID and AST results (E. coli NDM-6, E. coli OXA-48). Six isolates were excluded from AST comparison, including the two that were excluded from ID comparison, 2 that failed to provide AST with AXDX (K. pneumoniae OXA-48, K. oxytoca OXA-48), and 2 that failed to provide AST with direct VITEK 2 (K. pneumoniae OXA-232, K. pneumoniae NDM-1).
ID agreement was similar between direct MALDI and AXDX (p = 0.65) but overall had lower agreement than that found for the clinical samples (Table 1). Seeded samples again showed a high number of non-reports (NR) (40.0%) when using direct MALDI compared with AXDX (5.7%; p < 0.001).
In the seeded samples, AXDX had overall EA, CA, VME, ME, mE, and NR of 93.2%, 89.0%, 2.2%, 0%, 9.2%, and 6.1%, respectively (Table 2). AXDX had multiple antibiotics that did not meet acceptable performance parameters. These included amikacin with 2 VME (28.6%), ertapenem with 2 VME (8.0%), cefepime with 1 VME (5.6%) and 6 mE (20.7%), and piperacillin–tazobactam with 4 mE (13.8%). Moreover, more than 50% of meropenem AST for AXDX gave a non-report. Direct VITEK 2 had overall EA, CA, VME, ME, mE, and NR of 97.5%, 94.7%, 1.3%, 0%, 4.1%, and 0.3%, respectively. Each antibiotic assessed using direct VITEK 2 had one or less VME and had EA and CA of more than 90% with the exception of meropenem (1 VME and 4 mE, corresponding to a CA of 82.1%).
Combined clinical and seeded samples, for all antimicrobials, by organism
AST results for the clinical and seeded samples for all antimicrobials by organism are displayed in Table A.2 in the Appendix (N = 57). AXDX had EA, CA, VME, mE, and NR of 95.1%, 92.7%, 2.3%, 6.4%, and 3.5%, respectively. Direct VITEK 2 had EA, CA, VME, mE, and NR of 98.2%, 97.0%, 1.3%, 2.4%, and 0.5%, respectively. There were statistically significant differences among EA (p = 0.003), CA (p < 0.001), mE (p = 0.001), and NR (p < 0.001) for the AXDX compared with direct VITEK 2. No major differences in the results were seen between organisms.
Discussion
Time to blood culture isolate ID and susceptibility testing are critical factors in patient care. For gram-negative bacteremia, determining the ID of the gram-negative bacteria can lead to improved patient management, such as broadening antibiotic therapy to a carbapenem when organisms known to be resistant to cephalosporins are identified (eg, organisms with inducible chromosomal AmpC beta-lactamases) (16).
In our clinical isolates, organisms with inducible chromosomal AmpC beta-lactamases were identified in 7% of isolates. As a result of increases in antibiotic resistance, particularly among gram-negative bacteremia harbouring ESBL, organism ID alone is not sufficient to gain reassurance that the organism is sensitive to first-line empiric antimicrobial therapies such as third-generation cephalosporins (17–19). We detected a 3 in 29 (10%) incidence of ESBL-producing organisms among our clinical isolates, 2 of which were resistant to the empiric antibiotic therapy the patient was given (ceftriaxone). An ESBL prevalence of 10% is lower than that seen in our centre, which has an approximately 13% prevalence of ESBL-producing E. coli among community patients and emergency room patients with E. coli bacteremia (20). Therefore, the clinical impact of rapid ID and AST results may be even higher. In addition, 57% of patients had their therapy narrowed shortly after the results of AST, suggesting that narrowing of therapy can occur sooner with more rapid AST results. Narrowing of antibiotic therapy not only reduces antimicrobial resistance but can also improve patient outcomes such as a reduction in mortality (21).
In our study, time to ID and AST for direct MALDI–VITEK 2 was, on average, 5.8 h and 16.5 h after Gram stain result, respectively. This corresponds to a reduction of 15.8 h for direct MALDI (73% reduction) and 16.5 h (50% reduction) for direct VITEK 2 in comparison with conventional methods. AXDX was the fastest, reducing time to ID to 1.5 h (93% reduction) and time to AST to 6.9 h (79% reduction) compared with conventional methods. Compared with direct MALDI–VITEK 2, AXDX reduced time to ID and time to AST by 74% and 58%, respectively. Moreover, AXDX had the most convenient workflow compared with direct MALDI–VITEK 2 because the processing of the blood culture pellet for MALDI–VITEK 2 is labour intensive (22). Time to ID for our direct MALDI and conventional MALDI was longer than expected as a result of wait times in MALDI processing. Our laboratory has since brought in another machine to reduce this turnaround time.
The main limitation with AXDX is its cost. AXDX costs more than $200 per test, excluding the upfront cost of the instrument and software (23). In comparison, in our institution direct MALDI and direct VITEK 2 cost $14 and $15 per test, respectively, when including reagent cost, labour costs, and the costs of each instrument. AXDX also generates large amounts of plastic waste, and each system module can process only one blood culture isolate approximately every 7 h (maximum eight modules per system). Because of these limitations, AXDX is difficult to consider implementing in our high-volume laboratory. It may have a niche in other settings, such as in lower volume laboratories, laboratories without access to MALDI–VITEK 2, or at times when personnel are limited (eg, overnight).
In the overall results, direct VITEK 2 had higher EA and CA and lower mE and NR (all ps < 0.05) than AXDX. When broken down, AST results were similar between AXDX and direct VITEK 2 in the clinical samples, but direct VITEK 2 had higher agreement than AXDX when challenged with multi-drug-resistant organisms. For clinical samples, both AXDX and direct VITEK 2 achieved more than 96% overall EA and CA with no VME or ME and only low numbers of mE (<4%). In the seeded samples, AXDX had a higher number of overall VMEs (especially for amikacin and ertapenem) than direct VITEK 2. Interestingly, Schneider et al found higher rates of VME using direct VITEK 2 (3.2%), as opposed to AXDX (1.3%), when compared with conventional VITEK (22). This may be related to differences in their pellet-making procedure from positive blood cultures.
AXDX also had a high number of non-reportable results that occurred with meropenem (>50% of the time). A study testing AXDX with carbapenemase-producing organisms also found major issues with meropenem AST results, with VMEs occurring in 25% of meropenem tested (5). These results are problematic because meropenem is a commonly used carbapenem and using more conventional methods, such as Kirby-Bauer disk diffusion, to clarify meropenem sensitivities will take away AXDX’s speed advantage. However, AXDX plans to address this issue in a future software release (already released in Europe).
Strengths of our study included direct MALDI–VITEK 2 being an established protocol in our laboratory; therefore, the times to ID and AST for direct MALDI–VITEK 2 were obtained through retrospective data from normal laboratory workflow. Another strength of our study was the ability to challenge AXDX and direct MALDI–VITEK 2 with a wide variety of rare carbapenem-resistant isolates that were collected worldwide.
Several study limitations are worth mentioning. Our sample size was small and limited by the number of kits that were provided by Accelerate Diagnostics. Moreover, our reference for AST was not a true gold standard (VITEK 2 from bacterial colonies grown on blood agar). Broth microdilution was done only for discrepant results between AXDX and direct VITEK 2 for ME and VMEs. By using VITEK 2 as our reference standard, the results of direct VITEK 2 may appear better than they actually are.
Many different methods are available to reduce time to ID and AST for positive blood cultures, all with differing advantages and disadvantages. Short-incubation MALDI and AST (using an automated AST platform such as VITEK 2), direct MALDI and AST, AXDX, peptide nucleic acid fluorescence in situ hybridization, polymerase chain reaction–based assays, and the T2Bacteria panel (T2 Biosystems, Lexington, MA) are the most notable (24,25). The T2Bacteria panel has the advantage of not requiring blood culture incubation before ID, but it does not provide susceptibility information. The main limitation with most of these panels is their high cost. Given that many major laboratories already have a MALDI, a VITEK 2, or both, short-incubation MALDI–VITEK 2 and direct MALDI–VITEK 2 are attractive options. The main limitation identified with using direct MALDI is its high failure rate, which, in this study, was as high as 40%. However, our lab has since improved the blood culture pellet extraction process to improve failure rates down to approximately (20%) (unpublished data), which is more in keeping with other published in-house extraction methods for direct MALDI (26–29). Also, commercial kits such as the MALDI Sepsityper® (Bruker Daltonics, Billerica, MA) are now available for purifying and extracting bacteria from positive blood cultures for use in performing direct MALDI. However, one head-to-head study of an in-house extraction method versus the MALDI Sepsityper did not find any statistically significant differences in direct MALDI failure rates (26).
Short-incubation MALDI is a similar alternative to direct MALDI but has many of the same challenges associated with it. The failure rate for ID using short-incubation MALDI is comparable with that of direct MALDI at approximately 20% (30,31). The downside to short-incubation MALDI and VITEK 2 compared with direct MALDI is the number of hours required to incubate before undergoing MALDI and VITEK 2. As such, studies examining short-incubation MALDI–VITEK 2 had increased time to ID and AST compared with our direct method (32,33). However, the extraction process required for direct MALDI–VITEK 2 has its own limitations, such as the hands-on personnel time required (our extraction process takes approximately 45 min to complete).
Conclusions
AXDX and MALDI–VITEK 2 directly from blood cultures positive for gram-negative bacilli are an effective means to obtain ID and AST results compared with conventional methods. Faster results may improve patient care by speeding the time to antibiotic de-escalation and reducing the time patients are on less effective empiric antibiotic therapy resulting from bacterial drug resistance. Although direct MALDI had high failure rates for ID (≤40%), direct VITEK 2 had higher agreement, and higher reportability for meropenem, than AXDX in the AST for multi-drug-resistant organisms. Direct MALDI–VITEK 2 and AXDX each have their own advantages and disadvantages with respect to costs, turnaround time, workflow, performance, and environmental impact, and all these factors must be taken into account when deciding on the best diagnostic test for blood culture ID and AST for a laboratory.
Appendix
Table A.1:
AST results for clinical and seeded samples using conventional MALDI/VITEK 2 system
| Antibiotic | Susceptible | Intermediate | Resistant | N/A |
|---|---|---|---|---|
| Clinical samples (n = 28) | ||||
| Amikacin | 28 | 0 | 0 | 0 |
| Cefazolin | 19 | 2 | 7 | 0 |
| Cefepime | 27 | 1 | 0 | 0 |
| Ceftazidime | 27 | 0 | 1 | 0 |
| Ceftriaxone | 24 | 0 | 3 | 1 |
| Ciprofloxacin | 24 | 0 | 4 | 0 |
| Ertapenem | 27 | 0 | 0 | 1 |
| Gentamicin | 25 | 0 | 3 | 0 |
| Meropenem | 28 | 0 | 0 | 0 |
| Piperacillin or tazobactam | 27 | 0 | 1 | 0 |
| Tobramycin | 24 | 3 | 1 | 0 |
| Total | 280 | 6 | 20 | 2 |
| Seeded samples (n = 29) | ||||
| Amikacin | 21 | 1 | 7 | 0 |
| Cefazolin | 0 | 0 | 29 | 0 |
| Cefepime | 8 | 3 | 18 | 0 |
| Ceftazidime | 3 | 1 | 25 | 0 |
| Ceftriaxone | 2 | 1 | 26 | 0 |
| Ciprofloxacin | 3 | 2 | 24 | 0 |
| Ertapenem | 3 | 1 | 25 | 0 |
| Gentamicin | 13 | 5 | 11 | 0 |
| Meropenem | 4 | 3 | 22 | 0 |
| Piperacillin or tazobactam | 2 | 4 | 23 | 0 |
| Tobramycin | 7 | 6 | 16 | 0 |
| Total | 66 | 27 | 226 | 0 |
AST = Antimicrobial susceptibility testing; MALDI = Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry
Table A.2:
AXDX and direct VITEK 2 system AST results for clinical samples and seeded samples, for all antimicrobials, by organism (N = 57)
| No. (%), 95% CI | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Organism | n | EA | CA | VME | ME | mE | NR | S | I | R |
| AXDX * | ||||||||||
| E. coli | 23 | 245 (97.6) | 241 (96.0) | 0 | 0 | 10 (4.0) | 2 | 182 | 25 | 44 |
| Klebsiella spp | 21 | 213 (95.5) | 210 (94.2) | 2 (1.6) | 0 | 11 (4.9) | 8 | 74 | 13 | 136 |
| Enterobacter spp | 8 | 65 (90.3) | 58 (80.6) | 2 (4.3) | 0 | 12 (16.7) | 8 | 27 | 9 | 36 |
| Citrobacter spp | 2 | 16 (84.2) | 16 (84.2) | 1 (11.1) | 0 | 2 (10.5) | 1 | 13 | 0 | 6 |
| P. mirabilis | 2 | 20 (100) | 20 (100) | 0 | 0 | 0 | 0 | 20 | 0 | 0 |
| P. aeruginosa | 1 | 5 (62.5) | 5 (62.5) | 0 | 0 | 3 (37.5) | 2 | 8 | 0 | 0 |
| All organisms | 564 (95.1), 93.1% to 96.7% | 550 (92.7), 90.4% to 94.7% | 5 (2.3) 0.07% to 5.2% | 0 (0%) | 38 (6.4), 4.6% to 8.7% | 21 (3.5), 2.2% to 5.4% | 324 (54.6) | 47 (7.9) | 222 (37.4) | |
| Direct VITEK 2 system † | ||||||||||
| E. coli | 23 | 248 (98.0) | 247 (97.6) | 0 | 0 | 6 (2.4) | 0 | 195 | 11 | 47 |
| Klebsiella spp | 21 | 230 (99.6) | 224 (97.0) | 1 (0.8) | 0 | 6 (2.6) | 0 | 87 | 9 | 135 |
| Enterobacter spp | 8 | 84 (96.6) | 82 (94.3) | 2 (4.3) | 0 | 3 (3.4) | 1 | 33 | 7 | 47 |
| Citrobacter spp | 2 | 21 (95.5) | 22 (100) | 0 | 0 | 0 | 0 | 13 | 0 | 9 |
| P. mirabilis | 2 | 22 (100) | 22 (100) | 0 | 0 | 0 | 0 | 22 | 0 | 0 |
| P. aeruginosa | 1 | 9 (100) | 9 (100) | 0 | 0 | 0 | 2 | 8 | 0 | 1 |
| All organisms (%) [95% CI] | 614 (98.2), 96.9% to 99.1% | 606 (97.0), 95.3% to 98.2% | 3 (1.3), 0.3% to 3.6% | 0 (0) | 15 (2.4), 1.3% to 3.9% | 3 (0.5), 0.1% to 1.4% | 358 (57.4) | 27 (4.3) | 239 (38.3) | |
Note: MICs were truncated to overlapping reportable range prior to analysis. For clinical samples, 1 isolate was excluded from AST comparison because it failed to provide an AST on AXDX. For seeded samples, 6 isolates were excluded from AST comparison (2 failed to identify on AXDX, 2 failed to provide AST with AXDX, and 2 failed to provide AST with direct VITEK 2)
* AXDX only reports cefazolin for E. coli and Klebsiella spp
† VITEK 2 system unable to differentiate cefazolin I from S. For the purposes of this study, a VITEK 2 system cefazolin minimum inhibitory concentration (MIC) reading of <4 was considered intermediate.
AXDX = Accelerate PhenoTest™ BC kit; CA = Categorical agreement; EA = Essential agreement; I = Intermediate; mE = Minor error; ME = Major error; NR = Non-report; R = Resistant; S = Susceptible; VME = Very major error
Ethics Approval:
The study protocol was approved by an ethics committee and the ethics certificate information is available from the authors upon request.
Disclosures:
The use of the Accelerate Pheno™ system and its kits was provided by Accelerate Diagnostics, Inc. The authors acted independently with respect to study design, data collection, data analysis, and the preparation and contents of this publication.
Peer Review:
This manuscript has been peer reviewed.
Animal Studies:
N/A.
References
- 1.Carroll KC, Pfaller MA, Landry ML, et al. Manual of clinical microbiology. 12th ed. Washington, DC: ASM Press; 2019. [Google Scholar]
- 2.Charnot-Katsikas A, Tesic V, Love N, et al. Use of the Accelerate Pheno System for identification and antimicrobial susceptibility testing of pathogens in positive blood cultures and impact on time to results and workflow. J Clin Microbiol. 2017;56(1):e01166–17. 10.1128/JCM.01166-17. Medline: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Pancholi P, Carroll KC, Buchan BW, et al. Multicenter evaluation of the Accelerate PhenoTest BC Kit for rapid identification and phenotypic antimicrobial susceptibility testing using morphokinetic cellular analysis. J Clin Microbiol. 2018;56(4):e01329–17. 10.1128/JCM.01329-17. Medline: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Matic N, Willey B, Gascon B, et al. Clinical impact of rapid identification and phenotypic antimicrobial susceptibility testing by Accelerate Pheno™ System for gram-negative bloodstream infections. Open Forum Infect Dis. 2017;4(1):S594–5. 10.1093/ofid/ofx163.1562. [DOI] [Google Scholar]
- 5.Willey BM, Gascon B, Lee S, et al. Prospective evaluation of Accelerate Pheno™ System (AXDX) version 1.1 for reducing turn around time in identification/antimicrobial susceptibility testing of gram-negative bacilli from positive blood culture using AXDX PhenoTest™ BC Kits. Open Forum Infect Dis. 2017;4(1):S594–5. 10.1093/ofid/ofx163.1561. [DOI] [Google Scholar]
- 6.Scott JS, Sterling Sa, To H, Seals SR, Jones AE. Diagnostic performance of matrix-assisted laser desorption ionisation time-of-flight mass spectrometry in blood bacterial infections: a systematic review and meta-analysis. Infect Dis. 2016;48(7):530–6. 10.3109/23744235.2016.1165350. Medline: [DOI] [PubMed] [Google Scholar]
- 7.Machen A, Drake T, Wang YF. Same day identification and full panel antimicrobial susceptibility testing of bacteria from positive blood culture bottles made possible by a combined lysis-filtration method with MALDI-TOF VITEK mass spectrometry and the VITEK2 system. PLoS ONE. 2014;9(2):e87870. 10.1371/journal.pone.0087870. Medline: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Lockwood AM, Perez KK, Musick WL, et al. Integrating rapid diagnostics and antimicrobial stewardship in two community hospitals improved process measures and antibiotic adjustment time. Infect Control Hosp Epidemiol. 2016;37(4):425–432. 10.1017/ice.2015.313. Medline: [DOI] [PubMed] [Google Scholar]
- 9.Perez KK, Olsen RJ, Musick WL, et al. Integrating rapid pathogen identification and antimicrobial stewardship significantly decreases hospital costs. Arch Pathol Lab Med. 2013;137(9):1247–1254. 10.5858/arpa.2012-0651-OA. Medline: [DOI] [PubMed] [Google Scholar]
- 10.Perez KK, Olsen RJ, Musick WL, et al. Integrating rapid diagnostics and antimicrobial stewardship improves outcomes in patients with antibiotic-resistant gram-negative bacteremia. J Infect. 2014;69(3):216–225. 10.1016/j.jinf.2014.05.005. Medline: [DOI] [PubMed] [Google Scholar]
- 11.Verroken A, Defourny L, le Polain de Waroux O, et al. Clinical impact of MALDI-TOF MS identification and rapid susceptibility testing on adequate antimicrobial treatment in sepsis with positive blood cultures. PLoS One. 2016;11(5):e0156299. 10.1371/journal.pone.0156299. Medline: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Gregson D, Church D, Chan WW, Willscroft H, Campbell L, Pitout J. An evaluation of Vitek2TM direct susceptibility testing (DST) of gram-negative bacteria (GNB) from positive blood cultures (+BC). Open Forum Infect Dis. 2017;4(Supplement 1):S594–5. 10.1093/ofid/ofx163.1560 [DOI] [Google Scholar]
- 13.Prod’hom G, Bizzini A, Durussel C, Bille J, Greub G. Matrix-assisted laserdesorption ionization-time of flight mass spectrometry for direct bacterial identification from positive blood culture pellets. J Clin Microbiol. 2010;48(4):1481–3. 10.1128/JCM.01780-09. Medline: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Peirano G, Matsumura Y, Adams MD, et al. Genomic epidemiology of global carbapenemase-producing Enterobacter spp., 2008–2014. Emerg Infect Dis. 2018;24:1010–9. 10.3201/eid2406.171648. Medline: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing. 28th ed. CLSI supplement M100. Wayne, PA: Clinical and Laboratory Standards Institute; 2018. [Google Scholar]
- 16.Clerc O, Prod’hom G, Vogne C, Bizzini A, Calandra T, Greub G. Impact of matrix-assisted laser desorption ionization time-of-flight mass spectrometry on the clinical management of patients with gram-negative bacteremia: a prospective observational study. Clin Infect Dis. 2013;56(8):1101–7. 10.1093/cid/cis1204. Medline: [DOI] [PubMed] [Google Scholar]
- 17.McDanel J, Schweizer M, Crabb V, et al. Incidence of extended-spectrum β-lactamase (ESBL)-producing Escherichia coli and Klebsiella infections in the United States: A systematic literature review. Infect Control Hosp Epidemiol. 2017;38(10):1209–15. 10.1017/ice.2017.156. Medline: [DOI] [PubMed] [Google Scholar]
- 18.Denisuik AJ, Garbutt LA, Golden AR, et al. Antimicrobial-resistant pathogens in Canadian ICUs: results of the CANWARD 2007 to 2016 study. J Antimicrob Chemother. 2019;74(3):645–53. 10.1093/jac/dky477. Medline: [DOI] [PubMed] [Google Scholar]
- 19.Laupland KB, Pitout JD. Extended-spectrum beta- lactamase-producing Enterobacteriaceae: an emerging public-health concern. Lancet Infect Dis. 2008;8(3): 159–66. 10.1016/S1473-3099(08)70041-0. [DOI] [PubMed] [Google Scholar]
- 20.Alberta Precision Laboratories. Antibiotic % susceptibility patterns: blood cultures, emergency and community, January–December 2018 [Internet]. Calgary: Alberta Precision Laboratories; 2017. [cited June 23, 2020]. Available from: https://www.albertahealthservices.ca/assets/wf/lab/wf-lab-antibiogram-cal-blood-cultures-emerg-and-community-2018.pdf. [Google Scholar]
- 21.Schuts EC, Hulscher MEJ, Mouton JW, et al. Current evidence on hospital antimicrobial stewardship objectives: a systematic review and meta-analysis. Lancet Infect Dis. 2016;16(7):847–56. 10.1016/S1473-3099(16)00065-7. [DOI] [PubMed] [Google Scholar]
- 22.Schneider JG, Wood JB, Smith NW, et al. Direct antimicrobial susceptibility testing of positive blood cultures: a comparison of the Accelerate PhenoTM and VITEK® 2 systems. Diagn Microbiol Infect Dis. 2019;95(3):114841. 10.1016/j.diagmicrobio.2019.05.013. Medline: [DOI] [PubMed] [Google Scholar]
- 23.Accelerate Diagnostics. Accelerate Diagnostics reports Q4 and full-year financial results for 2017 [Internet]. Tucson, AZ: Globe Newswire; 2018. [cited February 28, 2020]. Available from: https://www.globenewswire.com/news-release/2018/02/15/1349156/0/en/Accelerate-Diagnostics-reports-Q4-and-Full-Year-Financial-Results-for-2017.html. [Google Scholar]
- 24.Wolk DM, Johnson K. Rapid diagnostics for blood cultures: supporting decisions for antimicrobial therapy and value-based care. J Appl Lab Med. 2019;3(4):686–97. 10.1373/jalm.2018.028159. Medline: [DOI] [PubMed] [Google Scholar]
- 25.Florio W, Morici P, Ghelardi E, Barnini S, Lupetti A. Recent advances in the microbiological diagnosis of bloodstream infections. Crit Rev Microbiol. 2018;44(3):351–70. 10.1080/1040841X.2017.1407745. Medline: [DOI] [PubMed] [Google Scholar]
- 26.Martiny D, Dediste A, Vandenberg O. Comparison of an in-house method and the commercial Sepsityper™ kit for bacterial identification directly from positive blood culture broths by matrix-assisted laser desorption-ionisation time-of-flight mass spectrometry. Eur J Clin Microbiol Infect Dis. 2012;31(9):2269–81. 10.1007/s10096-012-1566-1. Medline: [DOI] [PubMed] [Google Scholar]
- 27.Thomin J, Ghislain AG, Foubert F, Corvec S. Assessment of four protocols for rapid bacterial identification from positive blood culture pellets by matrix-assisted laser desorption ionization-time of flight mass spectrometry (Vitek® MS). J Microbiol Methods 2015:54–56. 10.1016/j.mimet.2015.05.010. Medline: [DOI] [PubMed] [Google Scholar]
- 28.Barnini S, Bucculeri V, Morici P, Ghelardi E, Florio W, Lupetti A. A new rapid method for direct antimicrobial susceptibility testing of bacteria from positive blood cultures. BMC Microbiol. 2016;16:185. 10.1186/s12866-016-0805-5. Medline: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Jo SJ, Park KG, Han K, Park DJ, Park YJ. Direct identification and antimicrobial susceptibility testing of bacteria from positive blood culture bottles by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry and the Vitek 2 system. Ann Lab Med. 2016;36(2):117–23. 10.3343/alm.2016.36.2.117. Medline: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Zabbe JB, Zanardo L, Mégraud F, Bessède E MALDI- TOF mass spectrometry for early identification of bacteria grown in blood culture bottles. J Microbiol Methods. 2015;115:45–46. 10.1016/j.mimet.2015.04.009. Medline: [DOI] [PubMed] [Google Scholar]
- 31.Verroken A, Defourny L, Lechgar L, Magnette A, Delmée M, Glupczynski Y. Reducing time to identification of positive blood cultures with MALDI-TOF MS analysis after a 5-h subculture. Eur J Clin Microbiol Infect Dis. 2015;34:405–13. 10.1007/s10096-014-2242-4. Medline: [DOI] [PubMed] [Google Scholar]
- 32.Delport JA, Strikwerda A, Armstrong A, Schaus D, John M. MALDI-ToF short incubation identification from blood cultures is associated with reduced length of hospitalization and a decrease in bacteremia associated mortality. Eur J Clin Microbiol Infect Dis. 2017;36:1181–6. 10.1007/s10096-017-2906-y. Medline: [DOI] [PubMed] [Google Scholar]
- 33.Mauri C, Principe L, Bracco S, et al. Identification by mass spectrometry and automated susceptibility testing from positive bottles: a simple, rapid, and standardized approach to reduce the turnaround time in the management of blood cultures. BMC Infect Dis. 2017;17:749. 10.1186/s12879-017-2851-5. Medline: [DOI] [PMC free article] [PubMed] [Google Scholar]