POC Microbiology Implementation in Real Life: a European ExperienceMaurizio Sanguinetti

Institute of Microbiology

IRCCS‒Fondazione Policlinico Universitario “Agostino Gemelli”, Rome (Italy)

Educational WorkshopEW03: Antimicrobial susceptibility testing with EUCAST breakpoints and methods

Arranged with the European Committee on Antimicrobial Susceptibility Testing (EUCAST)

Convenors: Gunnar Kahlmeter (Växjö, SE)

Derek Brown (Peterborough, UK)

Faculty: Gunnar Kahlmeter (Växjö, SE)

Petra Apfalter (Linz, AT)

Derek Brown (Peterborough, UK)

Maiken C. Arendrup (Copenhagen, DK) Christian G. Giske (Stockholm, SE)

Rafael Cantón (Madrid, ES)

Point-of-care (POC) Microbiology – Ready or Not, Here it Comes

Faculty disclosure

• Invited speaker: Pfizer, Gilead, MSD, Astellas, Bruker Daltonics,

Becton Dickinson

• Consultant: Pfizer, Gilead, MSD, Thermo Fisher

• Research Grants: Pfizer, Gilead, MSD, BioMerieux, Becton

Dickinson, BioRad, Menarini

Objectives of the talk

• To describe setting up a POC microbiology laboratory in response to laboratory

restructuring in a tertiary care Hospital in Italy on the basis of the available

scientific literature

• To summarize lessons learned from this experience relevant to Canadian

microbiology laboratories

Rapid diagnostic tests: the John Bartlett’s game

changers in infectious diseases

• We concur with Dr. John Bartlett, retired infectious diseases physician, who

stated in 2010 that rapid diagnostic tests are ‘game changing’ in the

management of infectious diseases.

• However, in 2018, we concur with a modification of his statement as suggested

by Bauer and Goff, that is: “Rapid diagnostic tests and antimicrobial stewardship

economics will ‘change the game’ of managing patients with infectious diseases.”

Bartlett, Medscape, 2011Bauer and Goff, CID, 2015

PAST

• 1917: the first large-scale use of the immunoassay for the diagnosis of infectious disease. Dochez and Avery reported that pneumococcal polysaccharide can be detected by immunoassay of serum and urine from patients with lobar pneumonia. In a prescient comment, the authors suggested that antigen detection could enable a rapid diagnosis of infection (J Exp Med 26:477‒493)

• 1971: the enzyme-linked immunoassay (ELISA) accelerated the diagnosis of disease due to the high assay sensitivity. However, despite automation, the ELISA remains the dominant immunoassay platform technology in the non-POC central laboratory setting.

• Later, several companies developed technologies that led to the present lateral flow immunoassay (LFIA) platform.

Kozel and Burnham-Marusich, JCM, 2017

PRESENT

Most of the POC rapid diagnostics use the LFIA platform. As with the home pregnancy test (i.e. the at-home use of antigen testing), the rapid test for the diagnosis of streptococcal pharyngitis made popular the LFIA technology for diagnosis of infectious diseases. The tests here listed have received the Clinical Laboratory Improvement Amendments (CLIA) waivers that enable POC use.

Kozel and Burnham-Marusich, JCM, 2017

DIRECT DETECTION MOLECULAR METHODS (FROM BLOOD)

• One of few advances to address some of the most severe and costly infectious diseases

lies into improvements in diagnostic tests that reduce delays in getting test results. This

has the potential to save lives and curb healthcare costs.

• To date, emerging technologies enable nucleic acid-approaches at the POC, which is

defined as “the time and site of an encounter with the health care system” (thus, not

necessarily at the bedside).

• Some of these technologies have met the requirements for speed, low cost, and ease of

use for POC applications. By contrast to some antigen detection LFIAs, molecular

diagnostics can provide high sensitivity and a rapid diagnosis at the POC.

PRESENT

Kozel and Burnham-Marusich, JCM, 2017

Strategic and practical issues for implementing molecular POC diagnostics

• Changes in clinical microbiology

• Different PoC scenarios

• Feasibility of different PoC platforms

• Impact on patient management

• Importance of guideline implementation

Lisby JG, ECCMID 2018

When the Laboratory is the POC: How do we transform IT • Determining how such POC tests would best incorporated into clinical diagnostic algorithms• Determining the clinical and economic benefits of a syndrome-based testing approach• On this basis we selected the following tests:

– T2 magnetic resonance– Automated microscopy (Accelerate Diagnostics)– β-D-glucan detection– Multiplex PCR (FilmArray, Unyvero, etc.)– LFIA tests (Legionella, S. pneumoniae, RSV)

• And these clinical syndromes– Sepsis (Target 1, implemented)– Gastroenteritis (Target 2, implemented)– Pneumonitis (Target 3, in progress)– Meningitis (Target 4, in progress)

Sanguinetti, unpublished

When the Laboratory is the POC: possible outcomes

• Better management of severe infections (e.g., sepsis) and reduction of the

associated mortality

• Incresed money reducing the length of stay and the use of

antibiotics/antifungals

• Better real-time assessment of the local microbial epidemiology of the

Hospital and to establish efficacious guidelines for the antimicrobial

(empirical) therapy

Target 1: Sepsis

• Objectives of the project

1. To speed up the “classical” blood culture workflow including POC or POC-like tests (MALDI-TOF mass spectrometry, FilmArray, Accelerate Pheno)

2. To implement rapid tests for the detection of microbial and fungal pathogens directly from blood

3. To include the β-D-glucan detection in an integrated diagnostic algorithm for the rapid de-escalation of antifungal therapy

Fiori et al, JCM, 2016

The GRES (Gestione Rapida Emocolture Sepsi) study

TotalN=743

Study period 1N=197

Study period 2N=233

Study period 3N=313

P-value

Mean time to start antibiotic therapy (SD), days 0.70 (1.4) 0.92 (1.7) 0.71 (1.4) 0.55 (1.1) 0.01

Percentage of effective therapy at the start of antibiotic therapy (SD)

92.6 (18.2) 93.9 (19.0) 91.4 (21.1) 92.6 (15.2) 0.37

Percentage of optimal therapy at the start of antibiotic therapy (SD)

62.2 (38.9) 59.0 (42.0) 51.9 (43.0) 71.9 (30.6) <0.001

Mean duration of antibiotic therapy (SD), days 19.3 (13.3) 21.9 (15.4) 19.3 (13.3) 17.7 (11.5) 0.002

Mean length of hospital stay (SD), days 26.5 (24.5) 29.7 (29.3) 26.8 (24.7) 24.2 (20.7) 0.04

Died at 30 days, n (%) 118 (15.9) 31 (15.7) 39 (16.7) 48 (15.3) 0.90

Study Period 1: the ID physician was called by a ward physician when patients had positive blood cultures. Study Period 2: the ID physician was called directly by the microbiologist immediately after a pathogen was isolated from blood culture. Study Period 3: the ID physician was called by the microbiologist immediately after a pathogen was identified from the blood culture, and all cases were re-evaluated at bedside 72 h after starting antibiotic therapy in order to optimize therapy.

Murri et al., Diag. Microbiol. Infect. Dis., 2018

Time Lapse Microscopy

Piperacillin-tazobactam Colistin

• This study describe results from a multicenter study evaluating the Accelerate Pheno system. A combination of fresh clinical and seeded blood cultures were tested, and results from the Accelerate Pheno system were compared to Vitek 2 results for identification (ID) and broth microdilution or disk diffusion for AST.

• Six common Gram-positive cocci and from the 4,142 AST results, the overall essential agreement (EA) and categorical agreement (CA) were 97.6% and 97.9%, respectively. Overall very major error (VME), major error (ME), and minor error (mE) rates were 1.0%, 0.7%, and 1.3%, respectively.

• Eight species of Gram-negative rods were evaluated and from the 6,331 AST results, overall EA and CA were 95.4% and 94.3%, respectively. Overall VME, ME, and mE rates were 0.5%, 0.9%, and 4.8%, respectively.

• On the basis of these results the Accelerate Pheno system seems to be able to identify and provide phenotypic MIC and categorical AST results in a few hours directly from positive blood culture bottles and support accurate antimicrobial adjustment.

Economic and organizational impact of a rapid automated microbiological diagnostic testing: a preliminary study

• The aim of this study was to identify, quantify and evaluate materials use, times and workflows associated with Accelerate ID/AST System (AXDX) in order to build and populate a model to measuring the organizational and economic impact of this approach in a long term perspective.

• The analysis was carried out by comparing the results of AXDX with the conventional direct rapid methods performed on blood cultures (MALDI-TOF + VITEK 2 system).

• The preliminary analysis was performed on 24 positive blood culture samples and both procedures, conventional and AXDX were performed on each sample.

• The average time from Blood Draw to BC positivity was 10 h and 32 min. • 71% of samples resulted positive on night shift, leading to a median time of 6 hours of delay in the

workflow. • The median time to identification and for susceptibility testing was shorter for AXDX (7 h and 57 min � 7

min) than conventional methods (19 h and 26 min � 6 h and 51 min). • The hands-on labor time requested for AXDX was 10 min vs 40 min for standard tests.

Romano et al., ECCMID 2017

T2Candida Panel powered by T2 Magnetic Resonance (T2MR)

T2MR combines proven magnetic resonance with innovative nanotechnology toaccurately identify Candida pathogens in whole blood faster and easier than bloodculture-based diagnostics.T2Candida is cleared by the US FDA and EMA for the diagnosis of candidemia.

Pfaller et al., Future Microbiol, 2016

• T2Candida was devised to not amplify freely circulating, non-cell-associated DNA. Using primers for ribosomal DNA intervening transcribed spacer region 2, results are reported as positive or negative for C. albicans/C. tropicalis, C. glabrata/C. krusei, and C. parapsilosis

• >1500 control patients with Candida-negative blood cultures• 6 patients with Candida-positive blood cultures • 250 contrived blood samples spiked with C. albicans/C. tropicalis, C. glabrata/C. krusei, and C. parapsilosis

• T2Candida identified 98.1% of patients as noncandidemic, with a mean time to negative result of 4.2 � 0.9 hours. The overall sensitivity per patient (excluding invalid results) was found to be 91.1%, with a mean time to positive result of 4.4� 1.0 hours

• The limit of detection was 1–3 cfu/mL (1 cfu/mL for C. tropicalis and C. krusei, 2 cfu/mL for C. albicans and C. glabrata, and 3 cfu/mL for C. parapsilosis)

• “Assuming sensitivity of 90% and specificity of 98%, anticipated positive and negative predictivevalues (PPVs/NPVs) of T2Candida can be calculated”

• Thus, “T2Candida performance characteristics enable clinicians to assign clinical settings in whichT2Candida is most likely to be useful in guiding antifungal treatment decisions”. Indeed, “PPVs mayexceed a threshold that justifies antifungal treatment, while corresponding NPVs render activecandidemia extremely unlikely”.

Performance of T2Bacteria Panel for rapid identification of bacteria responsible of bloodstream infections

De Angelis et al., JAC, 2018

Positive results more than 4.5 times faster: Positive results for a bacterial infection with T2Bacteria were available in 5.5 hours, compared to a mean time of 25 hours for blood culture. Every hour of providing targeted therapy to patients faster can reduce patient mortality rates by 7.6%.Negative results 20 times faster: A negative result took 120 hours (5 days) for blood culture, compared to 6.1 hours with the T2Bacteria Panel. Excellent accuracy: The T2Bacteria Panel had 89.5% sensitivity and 98.4% specificity in comparison with clinical criterion of a true infection. The authors point out that blood culture-based diagnostics have a sensitivity as low as 65% for the first blood culture draw.Potential impact on therapy: The authors found that 66.7% of the clinically infected patients missed by blood culture and correctly identified by T2Bacteria were being inappropriately treated at the time of the T2Bacteria result.

Performance of the T2Bacteria Panel in clinical blood samples according to Clinical Infection used as the gold standard assessed by detection channel

De Angelis et al., JAC, 2018

Prospective, randomized, controlled single-center unblinded study, performed in a mixed ICU. A total of 110 patients were randomly assigned to a strategy in which empirical antifungal treatment duration was determined by (1,3)-β-d-glucan, mannan, and anti-mannan serum assays, performed on day 0 and day 4; or to a routine care strategy, based on international guidelines, which recommend 14 days of treatment.

The primary endpoint of this study was the percentage of patients receiving early discontinuation of empirical antifungal treatment, defined as a discontinuation, for reasons other than death, strictly before day 7 after empirical antifungal treatment initiation.

No significant difference was found in the percentage of patients with subsequent proven invasive Candida infection, mechanical ventilation-free days, length of ICU stay, cost, and ICU mortality between the two study groups.

ICU: intensive care unit.Rouze A, et al. ICM, 2017

• Objectives:– To determine the effects of a strategy that uses (1,3)-β-D-glucan (BDG) results for antifungal treatment of ICU patients at

high risk of invasive candidiasis.• Methods:

– Patients were included in the analysis if they exhibited sepsis at the time of BDG testing, and they met Candida-score components (i.e., severe sepsis, total parenteral nutrition, surgery, or multifocal Candida colonization) to reach a ≥3 value.

• Results:– 198 patients were studied– Of 63 BDG-positive patients, 47 with candidemia and 16 with probable Candida infection, all received antifungal therapy– Of 135 BDG-negative patients, 110 (55.5%) did not receive antifungal therapy, whereas 25 (12.6%) were initially treated.

In 14 of these 25 patients antifungals were discontinued as negative BDG results were notified. Candidemia was subsequently diagnosed only in one patient who did not receive prior antifungal therapy

– The median antifungal therapy duration in candidemic patients differed from that in non-candidemic patients (14 days [IQR, 6–18 d] vs 4 days [IQR, 3–7 d]; p <0.001)

– Thus, unnecessary antifungal therapy was avoided in ~73% of potentially treatable patients and it was shortened in another ~20%

– A rough cost analysis showed also that adopting a BDG-based therapeutic approach might lead to a not negligible cost reduction by saving approximately €3500 per patient.

• Positive predictive value (PPV) and negative predictive value (NPV)

for candidemia of (1,3)-β-D-glucan (BDG) and procalcitonin (PCT)

considered both separately and in combination (BDG+PCT)

• The reported PPV for candidemia (also readable as NPV for

bacteremia) was obtained when both markers were concordant in

indicating candidaemia (BDG ≥80 pg/ml and PCT <2 ng/ml), while

the reported NPV for candidemia (also readable as PPV for

bacteremia) was obtained when both markers when concordant in

indicating bacteremia (BDG <80 pg/ml and PCT ≥2 ng/ml)

Giacobbe DR, et al. Crit Care, 2017

Issues encountered during the project implementation

• Necessity to train the technicians for the use of innovative procedures

• Implementation of specific algorithms for the validation of the new tests

• Necessity to have the Microbiology POC-Lab open 24 h/day-7 days/week

• Compliance of the clinicians in receiving the results also in the night

Target 2: Gastroenteritis (C. difficile infection)

• Objectives of the project

1. To implement a diagnostic procedure including a CLIA test for GDH and Toxins A and B, and a confirmatory molecular test (BDMax platform)

2. To setup a clinical-diagnostic algorithm at the Hospital level to manage the C. difficile infection

3. To include the Fecal Microbiota Transplantation (FMT) in this algorithm

Sample Multistep Algorithm for Rapid Diagnosis of Clostridium difficile Infection

Copyright 2015 American Medical Association. All rights reserved.

adverse effects, a disulfiram-like reaction when ingested with alco-hol, and peripheral neuropathy with prolonged therapy.70

Treatment by Disease SeverityTable 3 lists definitions of CDI severity, definitions for recurrent dis-ease, and factors associated with recurrence.15,16,20 Figure 3 pro-vides a possible approach for CDI treatment according to disease se-verity. However, the approach in Figure 3 has not been validated.71-75

Treating Mild to Moderate CDIFor mild to moderate CDI, oral metronidazole remains the preferredtherapy, in part because of its low cost.15,16,63 The standard dose is500 mg orally, 3 times daily for 10 to 14 days. For patients unable totake oral medications, metronidazole can be administered intrave-nously at the same dose, although metronidazole is not recom-mendedasmonotherapywhenadministeredintravenously.15,16 Basedon a recent study64 that showed a lower clinical success rate for met-ronidazole vs vancomycin, it may be reasonable to consider vanco-mycin for mild to moderate CDI.

Treating Severe or Complicated CDIVancomycin is the preferred therapy for severe or complicatedCDI.15,16,63 Taking vancomycin 125 mg orally, 4 times daily for 10 to14 days, is noninferior to higher doses in the absence of compli-cated infection.22 However, expert opinion often favors higher dosesin severe or complicated disease.15,16

Vancomycin may also be administered rectally in the setting ofileus, as an adjunctive therapy, although evidence is limited to casereports.15,76,77 Rectally administered vancomycin is not typically usedalone because rectally administered vancomycin may not reach theentire affected area.78 Intravenous metronidazole achieves detect-able levels throughout the colon79 and may be an adjunctive therapyfor ileus or severe/complicated CDI, typically with vancomycin whenadministered orally, rectally, or by both methods. However, there areno randomized trials supporting this practice.15,16 Treatment fail-ures have occurred in patients with ileus administered intravenousmetronidazole monotherapy.56,77

Promptsurgicalevaluationshouldbeobtainedinpatientswithcom-plicated CDI. Early intervention can reduce mortality.80,81 Subtotal ortotal colectomy with end ileostomy is often performed when surgeryisrequired,althoughtherearenewercolon-preservingtechniques.80,81

Treating Recurrent CDIRecurrent CDI is more common in older patients and in those withconcomitant antibiotic use, presence of comorbidities, concomi-tant use of proton pump inhibitors, and worse initial diseaseseverity.11,16 Inadequate antibody response after an episode of CDIis associated with increased recurrence rates.82,83

Figure 2. Sample Multistep Algorithm for Rapid Diagnosis of Clostridium difficile Infection

Overview of Diagnostic Steps

Perform 2 initial rapid diagnostic stool tests in a symptomatic patient

Perform a third test on discordant samples

Check result of the third test

Discordant resultsConcordant results

Final result

Detailed Diagnostic Steps

1. Perform EIA for GDH in stool sample2. Perform EIA for Toxin A and B in stool sample

Perform PCR for tcdB Gene

Testing consistent with C difficile infection

Testing not consistent with C difficile infection

GDH +/Toxin +a GDH +/Toxin -a GDH -/Toxin +a

tcdB Gene + tcdB Gene -

GDH -/Toxin -a

Abbreviations: EIA, enzymeimmunoassay; GDH, glutamatedehydrogenase; PCR, polymerasechain reaction.a Toxin-positive (+) indicates

presence of either Toxin A or Toxin Bwhen a combined test is performed,as in this example. Toxin-negative(−) indicates that neither Toxin Anor B is present. Adapted underCreative Commons License.116

Table 3. CDI Classification Based on Disease Severity

CDI DiseaseCategory

Clinical and LaboratorySigns Associated Risk Factors

Mild to moderate Diarrhea withoutsystemic signs ofinfection, white bloodcell count <15 000 cells/mL, and serumcreatinine <1.5 timesbaseline15

Antibiotic use, previoushospitalization, longerduration of hospitalization,use of proton pumpinhibitors, receipt ofchemotherapy, chronickidney disease, and presenceof a feeding tube10-14

Severe Systemic signs ofinfection, and/or whiteblood cell count≥15 000 cells/mL,or serum creatinine≥1.5 times thepremorbid level15

Advanced age, infection withBI/NAP1/027 strain114,115

Severe,complicated

Systemic signs ofinfection includinghypotension, ileus,or megacolon15

See above,a plus recentsurgery, history ofinflammatory bowel disease,and intravenousimmunoglobulin treatment43

Recurrent Recurrence within 8weeks of successfullycompleting treatmentfor CDI16,20

Patient age ≥65 y,concomitant antibiotic use,presence of significantcomorbidities, concomitantuse of proton pumpinhibitors, and increasedinitial disease severity16

Abbreviation: CDI, Clostridium difficile infection.a Also includes associated risk factors for mild to moderate and severe CDI

disease categories.

Diagnosis and Treatment of C difficile in Adults Review Clinical Review & Education

jama.com (Reprinted) JAMA January 27, 2015 Volume 313, Number 4 403

Copyright 2015 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by maurizio sanguinetti on 10/27/2015

Bagdasarian et al., JAMA, 2015

Adopted procedure for the managementof C. difficile infection at the “A. Gemelli” Hospital

• Principal steps of the procedure

1. Identification of the patient with suspected C. difficile infection

2. Microbiological real-time evaluation of the patient

3. Management of the patient on the basis of microbiological results (including the possible institution of FMT)

“A. Gemelli” Hospital, Official Procedures

Outcomes of the implementation of C. difficile procedure at the “A. Gemelli” Hospital

0

5

10

15

20

25

30

2008-2012 (beforeprocedure)

2012-2016 (afterprocedure)

Length of stay (days)

0

5

10

15

20

25

2008-2012 (beforeprocedure)

2012-2016 (afterprocedure)

Mortality (%)

De Belvis et al., in preparation

• 130 patients treated (mean age, 71 years; range, 29‒94 years) • Mean number of recurrences of Clostridium difficile infection (rCDI) = 3 (range = 2‒7)• Mean Charlson Comorbidity Index score = 3• Inpatients/Outpatients = 90/40

Fecal Microbiota Transplantation at the “A. Gemelli” Hospital in Rome (Italy)Overall Preliminary Data (years 2013‒2017)

• Overall 192 infusion procedures (71 patients received one infusion, 59 multiple infusions)• Fresh material/Frozen material= 126/66

FMT procedures/yearo June‒December 2013 = 9 (7 pts)o 2014 = 24 (15 pts)o 2015 = 36 (25 pts)o January‒July 2016 = 32 (18 pts)o July 2016‒July 2017 = 91 (65 pts)

RESULTS:o Severe colitis (pseudomembranous colitis) in 51 patientso Follow-up range = 1‒59 monthso Resolution of rCDI in 127/130 treated patients (98%)o No patients experienced further recurrences after FMTo No more surgery for CDI in our hospital since 2014

Cammarota G, et al., Ann. Internal Med., 2015Cammarota et al., Aliment Pharmacol Ther, in revision

Conclusions

• More than ever, it is emphasized the primary role of a clinical microbiology laboratory, that

is to assist clinicians in the diagnosis and treatment of infectious diseases, and to support

infection control specialists in their tasks.

• In many instances, however, the delay between the collection of the specimen and the

results of the microbial culture makes the latter unhelpful for the clinician.

• Therefore, advances in diagnostic methods (such as POC or POC-like tests) and/or

procedures might satisfy the ‘need for speed’, particularly when clinicians, infection control

specialists, and microbiologists confront with a septic patient or, in general, with a patient

with a systemic infection, or deal with an emerging epidemic.