New and Novel Rapid Diagnostics That are Impacting Infection Prevention and Antimicrobial Stewardship

Kaede V. Sullivan; Jennifer Dien Bard


Curr Opin Infect Dis. 2019;32(4):356-364. 

In This Article

Abstract and Introduction


Purpose of review: The current review summarizes advances in rapid diagnostic testing that impacts infection prevention and antimicrobial stewardship programs.

Recent findings: A variety of rapid diagnostic technologies to identify organisms in cultured blood are now available. When coupled with antimicrobial stewardship (ASP), these rapid technologies can optimize antimicrobial utilization and patient outcomes. Two rapid molecular panels that detect organisms related to pneumonia are available and may impact infection prevention surveillance definitions. Three molecular tests are available for the detection of meningitis and encephalitis pathogens. Still, the clinical impact of these broad, multiplexed panels need additional clarification. For Clostridioides difficile infections, ultrasensitive toxin A/B assays may provide enhanced sensitivity and specificity compared with enzyme immunoassay and molecular testing respectively. Finally, the adoption of Matrix-Assisted Laser Desorption Ionization-Time of Flight Mass Spectrometry (MALDI TOF MS) for rapid organism identification is growing. Recent US Food and Drug Administration-clearance of a MALDI TOF MS platform for identification of Nocardia, Mycobacteria, and molds may expedite antimicrobial decisions for infections that traditionally required days to weeks for an identification.

Summary: Tests with broad diagnostic scope and swift turnaround time are rapidly entering the market. Many impact infection prevention and ASP programs. Collaboration with the microbiology laboratory is crucial to ensure that new tests successfully optimize patient care.


In 2013, the Infectious Disease Society of America (IDSA) convened the IDSA Task Force, comprised of 18 representatives from academic medicine, government, and industry. The committee produced a document, entitled 'Better Tests, Better Care: Improved Diagnostics for Infectious Diseases,' that enumerated a list of unmet clinical needs in the realm of infectious diseases diagnostics. Among them were 'Rapid testing from clinical specimen (≤60 min)'; 'Rapid testing from clinical isolate (≤60 min)'; 'point of care or near-patient testing (≤60 min)'; and 'Simplicity (CLIA-waived)'.[1] Five years have now passed, and the diagnostic market of infectious diseases has grown with assays that aim to meet these needs. In this review, we discuss the recent advances in rapid diagnostic technologies (Table 1) that impact infection prevention and antimicrobial stewardship (ASP) programs.

Bloodstream Pathogen Diagnostics

There is currently a myriad of rapid diagnostic assays cleared by the US Food and Drug Administration (FDA) that identify bloodstream pathogens from cultured blood. Large, multiplexed, syndromic panels that employ molecular technologies for pathogen identification and detection of antimicrobial resistance determinants within 1–3 h are currently available and include the Verigene blood culture Gram-positive (BC-GP) and Gram-negative panels (Luminex Corporation, Austin, Texas, USA); the FilmArray Blood Culture Identification (BCID) panel (BioFire Diagnostics, Salt Lake City, Utah, USA); and the ePlex Blood Culture Identification Gram-positive (BCID-GP): Gram-negative (BCID-GN) and Fungal Pathogen (BCID-FP) assays (GenMark Diagnostics, Carlsbad, California, USA). These products and their coverage are summarized in Table 2. The Verigene and FilmArray assays detect common blood culture pathogens and have been widely adopted in the United States. Both systems detect bloodstream pathogens in positive blood cultures with sensitivity of more than 95% in monomicrobial bacteria compared with standard blood cultures.[2–4] However, both Verigene and FilmArray products may miss the detection of one or more organisms in polymicrobial bacteremia. Both systems may also have difficulty differentiating some organisms (Escherichia coli from Shigella spp.); Verigene BC-GP may erroneously report Streptococcus mitis group as Streptococcus pneumoniae.[2–4] More recently, the ePlex BCID-GP and BCID-FP assays were FDA-cleared in December 2018 and the BCID-GN in April 2019. With 20 bacterial targets, the former detects the largest number of Gram-positive species of any panel on the US market. It also detects four resistance markers (mecA, mecC, vanA, vanB) which are crucial for the detection of methicillin-resistant Staphylococcus and vancomycin-resistant enterococci. BCID-GP also includes 'pan' Gram-negative and 'pan' Candida targets, to detect Gram stain errors in which Gram-negative and/or yeast were not reported on the Gram stain. BCID-FP detects 11 Candida species, Cryptococcus neoformans, Cryptococcus gattii, Rhodotorula, and Fusarium representing the largest available fungal panel available for blood culture diagnosis. The ePlex BCID-GN panel identifies 20 aerobic and anaerobic Gram-negative pathogens.[5] It detects CTX-M (a common determinant of extended-spectrum beta-lactamase production), a variety of carbapenemases (KPC, NDM, VIM, IMP, OXA-23, and OXA-48), and has 'pan' Gram-positive and 'pan' Candida targets. Finally, the Sepsityper kit (Bruker Daltonics, Fremont, California, USA) provides the reagents required to process positive blood cultures for matrix-assisted laser desorption ionization time of flight mass spectrometry (MALDI–TOF MS).[6] This product is not FDA-cleared in the United States.

The vast majority of diagnostics that are available for rapid pathogen identification in bloodstream infections detect pathogens from positive blood cultures and require incubation of inoculated blood culture broth to amplify the organism load. By contrast, the T2Dx system (T2 Biosystems, Lexington, Massachusetts, USA) detects pathogens directly from blood without incubation. T2Candida detects five Candida species, and T2Bacteria detects five bacterial species directly from whole blood within 3–6 h, circumventing the delays associated with blood culture incubation.[7,8] Weaknesses of these assays include the limited number of pathogens targeted and the absence of resistance markers. Standard blood cultures are still required to recover organism for antimicrobial susceptibility testing (AST) and isolation of organisms that are not detected by the T2 panels.

The only bloodstream panel that offers rapid identification and expedited AST is the PhenoTest BC (Accelerate Diagnostics, Tucson, Arizona, USA). This system detects 14 Gram-positive and Gram-negative species and 2 Candida species within 90 min using fluorescent in-situ hybridization. Minimum inhibitory concentrations are generated through morphokinetic cellular analysis of individual, immobilized bacterial cells in the presence or absence of antimicrobial agents.[9] This approach is unique; in that, other syndromic panels offer only limited genotypic antimicrobial susceptibility data.

Rapid identification and AST in bloodstream infections are most impactful when coupled with ASP intervention to guide adjustments in antimicrobial therapy. In 2016, IDSA and the Society for Healthcare Epidemiology jointly recommended the use of rapid identification systems in bloodstream infections and emphasized the importance of pairing it with ASP support.[10] Systematic reviews of rapid identification systems in bloodstream infections have suggested that implementation can lead to reductions in time to antimicrobial optimization, recurrence of bacteremia, length of stay (LOS), 30-day mortality, and hospital costs.[10,11] A randomized controlled trial of 617 adult and pediatric patients reinforced the importance of ASP involvement. Patients were randomized to one of three arms: standard blood cultures with no ASP intervention, FilmArray BCID and no ASP intervention, and FilmArray BCID with ASP intervention. Compared with the other two other arms, time to de-escalation of antimicrobials was significantly faster in the arm with FilmArray BCID coupled with ASP intervention.[12] Currently, there is no published data describing the impact of rapid identification and expedited AST in bloodstream infections on antimicrobial utilization and patient outcomes. Preliminary, unpublished data from an American study suggested possible reductions in time to optimal therapy, total antibiotic days, and LOS when positive blood cultures were tested using the PhenoTest BC.[13]


Two new tests that are designed to rapidly detect pathogens associated with pneumonia were FDA-cleared in 2018. If widely adopted and if clinical validated, these assays may require a re-examination of surveillance definitions for ventilator associated events. The Unyvero Lower Respiratory Tract Application (Curetis USA, San Diego, California, USA) was FDA-cleared in April, 2018 for endotracheal tube aspirates. Unyvero is a qualitative, multiplex PCR test that detects 19 bacterial species that are responsible for hospital-acquired pneumonia in addition to a variety of antibiotic resistance determinants. Curiously, Unyvero appears to be less sensitive than quantitative respiratory culture.[14] The reported limit of detection ranges from 104 to 106 CFU/ml for bacterial species, which is higher than that of a standard quantitative respiratory culture (typically 103 CFU/ml per isolate).[14] Because a positive result for a resistance marker cannot be linked to a bacterial isolate, a respiratory culture still needs to be performed to link organism identification with AST.[14]

Also, in November 2018, the BioFire FilmArray Pneumonia Plus (BioFire) was FDA-cleared for sputum, endotracheal tube aspirate, bronchoalveolar lavage (BAL), and mini-BAL specimens. This is a semi-quantitative assay that detects 15 bacterial species that are responsible for hospital-acquired pneumonia. When a bacterial species is detected, the system assigns it to semi-quantitative 'bins' of 104, 105, 106, and more than 107 copies/ml. Biofire states in documentation submitted to the FDA that the semi-quantitative bins are not equivalent to CFU/ml.[15] At present, there is no published data correlating the bins to bacterial loads generated by quantitative respiratory cultures. This test also qualitatively detects Chlamydophila pneumoniae, Mycoplasma pneumoniae, Legionella pneumophila, and eight respiratory viruses. Like the Unyvero assay, standard respiratory culture is still required to link AST results to the identified organism. Of note, the manufacturers of both assays have stated that detection of bacterial nucleic acid may indicate colonization with normal respiratory flora and may not represent the causative agent of pneumonia.[14,15]

It is interesting to consider how these assays will be incorporated into surveillance definitions. The National Healthcare Safety Network, the body that sets surveillance definitions in the United States, has stated that pathogens grown in culture OR identified as a result of other laboratory testing (e.g., PCR) should be considered for Possible Ventilator-Associated Pneumonia status but they have not yet published guidance related to incorporating semi-quantitative molecular results.[16]

Mycobacterium Tuberculosis Complex

Mycobacterium tuberculosis complex has been considered an ideal candidate for rapid molecular diagnostics due to its propensity for severe disease, high transmissibility, extremely slow growth in culture, and its global distribution. Introduction of the rapid, automated, Xpert MTB/RIF assay (Cepheid) offered a variety of opportunities to improve patient care. From the infection prevention perspective, airborne precautions are typically discontinued if there are three consecutive, negative AFB (acid-fast bacilli) sputum smear results in specimens collected at 8–24-h intervals.[17] This typically requires approximately 2 days. In 2015, the FDA cleared Xpert MTB/RIF for testing of either one or two sputum specimens as an alternative to performing serial acid-fast stains on sputum when determining whether to continue airborne precautions in patients with suspected pulmonary tuberculosis (TB). This was based on data suggesting that a single Xpert MTB/RIF Assay result detects approximately 97% of patients who are smear–positive and culture-confirmed. Two serial Xpert MTB/RIF Assay results detected 100% of the same patients. Further, a single negative Xpert MTB/RIF Assay result was reported to have a negative predictive value of 99.7% for AFB smear-positive pulmonary TB and two negative Xpert MTB/RIF results had negative predictive value of 100%. If implemented, use of the Xpert MTB/RIF, therefore, has the potential to expedite discontinuation of airborne precautions earlier, using less labor.[18]

Central Nervous System Infections

Infections of the central nervous system (CNS), including meningitis and encephalitis, are potentially life-threatening conditions associated with significant morbidity, mortality, and economic burden. Although conventional microbiology testing methods (e.g., Gram stain and culture) suffer from low sensitivity and/or long turnaround time, new and emerging molecular tests have the potential to significantly improve these testing parameters.

There are currently only three FDA-cleared tests that use molecular techniques to detect pathogens directly from cerebrospinal fluid (CSF). The Xpert EV (Cepheid, Sunnyvale, California; to be discontinued in December 2019) detects enterovirus from CSF and the Simplexa HSV-1 & 2 Direct (Focus Diagnostics, Cypress, California, USA) detects herpes simplex viruses 1 and 2 (HSV-1, HSV-2). The most comprehensive test is the FilmArray Meningitis/Encephalitis panel (FA ME) (BioFire) which was FDA-cleared in October 2015. This panel detects 14 targets, including six bacterial targets (Haemophilus influenzae, Neisseria meningitidis, Listeria monocytogenes, S. pneumoniae, Streptococcus agalactiae, and E. coli K1), seven viral targets (HSV-1, HSV-2, varicella zoster virus, cytomegalovirus, human herpesvirus 6, enterovirus, and human parechovirus), and C. neoformans/gattii, within 1 h.

IDSA recommends the use of PCR to detect HSV encephalitis and enterovirus meningitis.[19,20] Expedited diagnosis of enterovirus meningitis by PCR has been reported to shorten LOS and decrease duration of antimicrobial therapy.[21] In addition, prompt 'rule-out' of CNS infection involving HSV1/2 using the Simplexa HSV-1 & 2 Direct has been reported to reduce duration of acyclovir administration in pediatric patients.[22] By contrast, experiences with the more comprehensive FA ME panel have been mixed. In a prospective, multicenter study involving 1560 CSF specimens and 136 positive FA ME tests, sensitivity and specificity were calculated using culture for bacterial pathogens and PCR/sequencing for all other pathogens. Sensitivity was 100% for nine of the 14 analytes: E. coli K1, H. influenzae, S. pneumoniae, cytomegalovirus, herpes simplex virus 1 and 2, human parechovirus, varicella zoster virus, and C. neoformans/gattii; 95.7% for enterovirus; and 85.7% for human herpes virus 6. Specificity was 99.2% for all targets. The study raised concerns about false-positive results due to contamination of the testing system with respiratory flora (e.g., S. pneumoniae, H. influenzae).[23] Although other recent studies reported satisfactory concordance between FA ME and traditional culture results and clinical presentation, this study underscored the importance of interpreting FA ME results in conjunction with other available information including clinical presentation, other CSF findings (cell count, glucose, protein), and radiographic imaging.[24–27] With respect to patient outcomes, a recent study reported that implementation of FA ME was not associated with reductions in duration of antimicrobial use or LOS in patients who tested negative on the FA ME assay. However, the sample size was small, and implementation of FA ME was not coupled with AS intervention.[27]

Clostridioides Difficile Infections

Clostridioides difficile toxins (toxin A and toxin B) are sine qua non for C. difficile infection (CDI), but historically, detection has been challenging. The traditional cell cytotoxicity neutralization assay (CCNA) is highly sensitive and specific, but slow and labor intensive. Toxin A/B enzyme immunoassays (EIA) require less labor but they lack sensitivity. Highly sensitive nucleic acid amplification tests (NAATs) gradually replaced EIA over the last decade but their specificity has come under fire due to published data suggesting that individuals who test NAAT-positive but toxin EIA-negative may be clinically indistinguishable from asymptomatic controls.[28]

Two 'ultrasensitive' C. difficile toxin assays are being marketed as a possible solution. Although neither assay is commercially available in the United States at this time, both claim higher sensitivity than toxin EIA and improved specificity compared with NAAT. The Singulex Clarity C. difficile toxins A/B assay (Singulex, Alameda, California, USA) measures toxin A and toxin B in stool specimens using the Singulex Clarity system. This is a rapid (<1 h), automated test that uses single-molecule counting technology.[29] Using the cell CCNA as the reference standard and a collection of 103 stool samples collected from patients with suspected CDI, Sandlund et al.[29] analyzed area under receiver operating characteristic curves and reported that at the cutoff value of 16.7 pg/ml, sensitivity and specificity of the Singulex Clarity C. difficile toxins A/B assay were optimized at 96.3% [95% confidence interval (CI), 81.0–99.9%] and 96.1% (95% CI, 88.9–99.2%), respectively, compared with CCNA. Quanterix Inc (Lexington, Massachusetts, USA) has also developed an ultrasensitive toxin assay, but uses 'Simoa' technology, which is based on capture, labeling, and detection of single protein molecules on paramagnetic beads in arrays of femtoliter-sized wells.[30] The limit of detection for both assays has been reported to be approximately 1 pg toxin/ml stool, suggesting that they are about 1000-fold more sensitive than commercially available EIAs.[29,30] Pollack et al. used the Simoa assay to compare toxin A and toxin B loads in asymptomatic individuals and patients with diarrhea. They reported that toxin loads were similar in the two groups and concluded that, if used alone, toxin concentrations generated by this assay may not be able to discriminate between symptomatic patients with CDI and colonized patients who have diarrhea from a different cause. They suggested that ultrasensitive toxin testing may need to be coupled with blood and stool inflammatory markers and, possibly, microbiome analysis to clearly define predictors of CDI versus colonization.[30]

Mass Spectrometry for Organism Identification

Over the last decade, MALDI–TOF MS has emerged as the fastest and most cost-efficient method of organism identification available on the market. There are two FDA-cleared MALDI–TOF MS systems that are commercially available in the United States: the Bruker Biotyper CA system (Bruker, Billerica, Massachusetts, USA) and VITEK MS (bioMérieux, Durham, North Carolina, USA). Both systems identify Gram-positive and Gram-negative organisms, anaerobes, and various yeast species within minutes. Implementation of MALDI–TOF MS has been reported to reduce time to optimal antimicrobial therapy and hospital costs in patients with bloodstream infections.[31] In February 2019, Vitek MS also received FDA-clearance for identification of Nocardia, a variety of Mycobacteria species and filamentous fungi. Identification of these organisms has traditionally involved labor intensive and time-consuming procedures that require specialty skills in phenotypic and molecular testing. To date, the use of MALDI–TOF MS for identification of these organisms has been largely confined to academic and/or reference laboratories due to a lack of access to high-quality spectral databases for these organisms, and the specialized cell lysis and protein extraction procedures that are required when preparing isolates with mycolic acid (in the case of Nocardia and Mycobacteria spp.) or chitin (in fungi) in the cell wall for MALDI–TOF MS. Widespread adoption of accurate and reliable MALDI–TOF MS procedures for these organisms' identification may allow ASP teams to expedite optimization of antibiotics and antifungals in infections that have traditionally required a prolonged wait for the identification information.

Developments in Detection of Multidrug Resistant Organisms

mecC methicillin-resistant Staphylococcus aureus. Methicillin-resistant Staphylococcus aureus (MRSA) mediated by mecC (as opposed to mecA) was first described in 2011 in both livestock and humans in the United Kingdom and Denmark. The prevalence in humans is thought to be less than 1% but epidemiological data in the United States is lacking.[32] Genetically, mecC is only 70% identical to mecA. As a result, PCR assays that are designed to detect mecA are unable to detect mecC MRSA. In addition, immunoassays that are designed to detect PBP2a in MRSA (the protein product resulting from translation of mecA) may not detect mecC product. However, from the infection prevention perspective, the impact of mecC on patient screening may be minimal. The Xpert MRSA NxG (Cepheid)'s PCR assay has been designed to detect both mecA and mecC. For institutions that use culture methods for MRSA screening, using a collection of 111 MRSA isolates with mecC, Kriegeskorte et al.[33] reported that 100% of the isolates grew on a variety of widely used MRSA screening media.

Carbapenem-resistant Organisms

Carbapenem-resistant Enterobacteriaceae (CRE) are a significant clinical problem world-wide. Although carbapenem resistance is frequently mediated by AmpC-producing or extended spectrum beta-lactamase-producing organisms with additional chromosomally mediated porin mutations, reports of carbapenemase producing CRE (CP-CRE) that harbor blaKPC and, increasingly, blaNDM have increased over time in the United States.[34] Some reports have suggested a recent decline in incidence of CP-CRE; however, clinical teams continue to face challenging therapeutic dilemmas in multidrug resistant CP-CRE.[35]

There are a variety of phenotypic methods for the detection of carbapenemase production in CRE. These include the Carba NP test and the modified carbapenem inactivation (mCIM) method. mCIM has a sensitivity and specificity of 98 and 99%, respectively, but requires overnight incubation.[35] The Carba NP procedure provides a result in about 2 h but has a lower sensitivity (84%), and the test requires laborious reagent preparation.[35] In the United States, the RAPIDEC Carba NP (bioMérieux) was FDA cleared in April 2017 for use on Enterobacteriaceae and Pseudomonas aeruginosa isolates. This kit is based on the Carba NP test and has a sensitivity and specificity of 98 and 99%.[35] It provides preprepared reagents, reducing labor requirements. This assay detects carbapenemase activity in isolates growing in pure culture within 2 h. Until recently, the availability of ceftazidime-avibactam and meropenem-vaborbactam (which both have excellent activity against KPC-mediated CRE) and the rarity of resistance to these agents have tempered the perceived need for carbapenemase testing and characterization of CP-CRE.[36] Emergence of endemic cases involving Enterobacteriaceae harboring blaNDM and other metallo-β-lactamases (which are resistant to ceftazidime-avibactam and meropenem-vaborbactam) in the United States has begun to change this picture.[36] At this time, rapid differentiation of KPC and metallo-β-lactamase production requires molecular testing. The Xpert Carba-R (Cepheid) was FDA-cleared in March 2016 and detects blaKPC, blaNDM, blaVIM, blaIMP, and blaOXA-48 from pure colonies of Enterobacteriaceae, Acinetobacter baumannii complex, and P. aeruginosa. In June of the same year, the Carba-R was FDA-cleared for rectal and peri-rectal screening swabs. The Xpert system is fully automated, uses real-time PCR technology, and generates results in about 2 h. Although very sensitive (100%) and specific (98%), the Xpert, on the contrary, is costly and novel carbapenemases with mutant primer and probe sequences may not be detected.[37] A less costly method may be in development: a rapid lateral flow immunoassay has been described in the literature that detects KPC, NDM, VIM, IMP, and OXA-48.[38] The Carba5 assay is still in development, but commercialization of a rapid, affordable product that performs with satisfactory sensitivity and specificity for KPC and NDM in the United States may provide hospitals with a viable alternative.