Dr Vicky EnneUniversity College London

v.enne@ucl.ac.ukTwitter: @Beat_AMR_Bugs

@HAP_DiagnosticsWebsite: www.ucl.ac.uk/inhale-project

The INHALE Project: What have we learnt about the molecular

diagnostics of HAP/VAP?

Sample arrives in lab

Microscopy/Gram Stain

(a few min)

Culture(16h-48h)

Identify growth: MALDI-TOF/Biochemical testing

(a few min – 24h)

Antimicrobial susceptibility testing

(24-48h)

Further testing / reference laboratory

(up to 2 weeks)

The Need for Rapid Diagnostics

Potential Benefits of Rapid Diagnostics

• Improved treatment outcome for patient

• Improved infection control and outbreak monitoring

• Reduction in empirical antimicrobial prescriptions• Preservation of broad spectrum antimicrobials

• Reduction in duration of treatment

• Reduction in cost of treatment

• Overall reduction antimicrobial consumption

• Potential reduction in levels of resistance

Potential Problems Associated with Molecular Diagnostics for AMR

• Inability to distinguish colonisation/infection• Host of resistance gene may not be clear

• Discrepancy between genotype and phenotype1. Presence of unknown resistance mechanisms

leading to treatment failure

2. Genotypic markers of resistance not resulting in clinically significant resistance. Patient may be denied therapeutic drug needlessly

• Rapid molecular methods should be supplemented with culture-based techniques

What types of test are available & what do they look for?

• Single target – simple Yes/No answer

• Gene (DNA/RNA)

• Protein

• Whole bacteria

• Something else

• Multiple targets

• One or more of targets as above

• Combine pathogen & resistance gene detection

• “Syndromic” testing

• A complete answer to particular question

• Host biomarkers

• Alone or in combination with pathogen detection

Types of Technology

• PCR Based

• Real-time/qPCR

• Droplet PCR

• Isothermal amplification

• Microarray

• MALDI-TOF MS + PCR

• PNA-FISH

• Rapid phenotypic testing

• Mass spectroscopy

• Whole genome/next generation sequencing

• Transcriptomics

What Factors Contribute to the Adoption of Molecular Diagnostics?

In the Clinical Laboratory

• Can it be integrated into the clinical pathway/laboratory workflow?• How much does it cost – who

pays?• Is it easy to use, how much

training is required?• How much space is needed?• How long does it take?

• Is the machine reliable• How often does it break down? • Is the customer support

adequate?

At the Bedside

• Are the results accurate• Does the test do “what it says on

the tin”?

• Does is look for the important pathogens/resistance genes?

• Do doctors trust the results & know what to do with them?

• Does using the test actually make a difference:• To levels of antibiotic use?

• To patient outcomes?

INHALE Programme Outline

Y1 Y2 Y3 Y4 Y5 Y6

WP1 – Diagnostic evaluation

WP2 – ICU Epidemiology

WP3 – Preparation/analysis

WP3 RCT

WP4 – Behaviour Study

WP5 – Health Economic Analysis

WP6 – Management and Dissemination

Project Participants• UCLH/UCL/UEA (Lead organisations)

• 15 Critical Care Units in England representing diverse case-

mix

WP1 – Clinical Laboratory Evaluation

• Evaluate three molecular tests for HAP/VAP

• Curetis Unyvero (PCR + microarray)

• Biomerieux Filmarray LRTI panel (nested PCR)

• Oxford Nanopore MinION (NGS sequencing)

• 700 surplus specimens from pneumonia patients

• Compare results to those from routine culture

• Supplement imperfect gold standard with 16s metagenomics / comprehensive culture ID / Checkpoints microarray for resistance genes

The Curetis UnyVero Pneumonia Test

• ‘Sample-in, answer-out’; cartridge based system

• Automated sample prep, multiplex PCR and microarray detection of product in 5-6h

• Pneumonia cartridge first product to be launched (39 targets)

bioMerieux Biofire FilmArray Pneumonia Panel

• FDA cleared & CE-IVD Marked Pouch based nested PCR

• 1h turnaround, 2 min hands on

• Semi-quantitative detection of pneumonia pathogens & AMR genes

FilmArray Pneumonia Panel Targets

Antibiotic Resistance GenesMethicilin Resistance

mecA/mecC and MREJ

ESBL

CTX-M

Carbapenemases

KPC

NDM

Oxa48-like

VIM

IMP

BacteriaSemi - Quantitative

Acinetobacter calcoaceticus-baumannii complex

Serratia marcescens

Proteus spp.

Klebsiella pneumoniae group

Enterobacter aerogenes

Enterobacter cloacae

Escherichia coli

Haemophilus influenzae

Moraxella catarrhalis

Pseudomonas aeruginosa

Staphylococcus aureus

Streptococcus pneumoniae

Klebsiella oxytoca

Streptococcus pyogenes

Streptococcus agalactiae

VirusesQualitative

Influenza A

Influenza B

Adenovirus

Coronavirus

Parainfluenza virus

Respiratory Syncytial virus

Human Rhinovirus/Enterovirus

Human Metapneumovirus

Middle East Respiratory Syndrome Coronavirus (MERS-CoV)

Atypical BacteriaQualitative

Legionella pneumophila

Mycoplasma pneumoniae

Chlamydia pneumoniae

Oxford Nanopore MinION NGS Diagnostics

• Rapid, low-cost NGS sequencing based on nanopore technology

• Capability to obtain full organism & resistance gene profile – not limited to selected targets

• First trial in the world evaluating suitability of technology for primary diagnosis direct from clinical samples

• Cost approx. $120/sample

INHALE Laboratory Work Flow for MinIONProcessing: Total time to Result = 7h

Deplete human DNA with saponin

based method (2h)

Extract pathogen DNA using robot

& magnetic particles (1h)

Manually wash DNA with

magnetic beads (30 min)

Fragment DNA & Set up Barcoding

PCR (15 min)

PCR to tag DNA with barcodes (3h)

Pool and wash

samples (30 min)

Library Prep

(15 min)

Load onto flow cell & start sequencing (first results in

30 min)

INHALE sputum sample

arrives in research lab

Routine Microbiology finds Organisms Commonly Associated with HAP/VAP

0

10

20

30

40

50

60

70

80

90

100

Staphylococcus…

Pseudomonas…

Esch

eric

hia

co

li

Haem

ophilu

s…

Klebsiella…

Sten

otrophomo…

E. c

loa

cea

e g

rou

p

Kle

bsi

ella

oxy

toca

Oth

er

Pro

teu

s m

ira

bili

s

Serratia…

Streptococcus…

Co

lifo

rm

Acinetobacter…

Enterobacter…

Moraxella

…

Pseudomonas…

Citrobacter…

Pro

teu

s sp

p.

Streptococcus…

Ente

rob

act

er s

pp

.

Morganella…

Nu

mb

er

of

de

tect

ion

s Median time to result from sample collection: 70.1 hMedian transport time to lab: 6.1h

% Culture Positivity Varies Significantly According to Site

0 10 20 30 40 50 60 70 80 90 100

O

N

M

L

K

J

I

H

G

F

E

D

C

B

A

Ho

spit

al

PCR Positive (%) Culture Positive (%)

Example Results: UCLH Sample 111 (ETT)

Method Organism Resistance Phenotype Resistance genotype

Routine microbiology

Escherichia coli Penicillins, amoxy-clav, aztreonam, 2/3/4 gen cephalosporins, co-trimoxazole, trimethoprim

NA

Unyvero Escherichia coli (+++) Penicillins, aztreonam, 2/3/4 gen cephalosporins, sulphonamides (predicted)

blaTEM-1

sul1blaCTX-M

GyrA wt

FilmArray Escherichia coli (>107) Penicillins, aztreonam, 2/3/4 gen cephalosporins (predicted)

blaCTX-M

MinION Escherichia coli (90.5% of reads)

Penicillins, aztreonam, 2/3/4 generation cephalosporins, co-trimoxazole trimethoprim (predicted)

blaTEM-1

sul1dfrA7blaCTX-M

sul2GyrA wt

Example Results: RFH Sample 59 (SPU)

Method Organism Resistance Phenotype Resistance genotype

Routine microbiology

P. aeruginosa Susceptible NA

Curetis Unyvero E. coli (+++)K. pneumoniae (+)P. aeruginosa (+++)S. maltophila (+++)

Penicillins, sulfonamides,methicillin (S. aureusonly) (predicted)

mecAblaSHV

sul1blaTEM

Biofire Filmarray E. cloaceae complex (10^4)E. coli (>10^7)K. pneumoniae (10^6)P. aeruginosa (>10^7)

NA None detected

Bayesian Latent Class Analysis demonstrates Culture is Less Sensitive than PCR

Routine Culture FilmArray Unyvero

Top 5 Bacteria Sen Spec Sen Spec Sen Spec

P. aeruginosa 64.8 99.7 99.2 99.4 95.8 99.9

S. aureus 65.3 99.2 99.3 95.6 91.3 99.8

K. pneumoniae 37.5 99.5 98.0 97.7 88.6 99.8

E. coli 38.6 99.7 99.0 98.7 89.5 99.7

H. influenzae 36.6 99.9 95.6 93.8 97.0 99.7

Other bacteria 29.6- 69.4 99.9 89.2-97.2 96.4-99.9 49.4-94.9 99.5-99.0

Key:> 90% > 80% < 70%

Implementability Comparison of PCR-Tests

Criterion FilmArray Pneumonia

Panel

Unyvero Pneumonia

Panel

Overall time to result 70 min 270 min

Footprint of the machine 3.2 sq. feet 7.4 sq. feet

Quality of Customer

service

Avg. user score 4/5

“Good communication and

service”

Avg. user score 3/5

“Problems usually

attributed to user”

Consumable logistics Superior performance on

5 different criteria

Ease of use Avg. user score 9/10

“Easy to use but

requirement for a sharps

bin is a negative”

Avg. user score 6/10

“Extra transfer step after

30 min is inconvenient”

Types of Technology

• PCR Based

• Real-time/qPCR

• Droplet PCR

• Isothermal amplification

• Microarray

• MALDI-TOF MS + PCR

• PNA-FISH

• Rapid phenotypic testing

• Mass spectroscopy

• Whole genome/next generation sequencing

• Transcriptomics

Antibiotic Prescribing in ICUs at UCLH, Cromwell and NNUH - 2017

• Commonest empirical antimicrobials

• Piperacillin/tazobactam ± Aminoglycoside

• Co-amoxiclav

• Ceftazidime

• Ciprofloxacin + Teicoplanin

Antimicrobial Prescribing for HAP/VAP

0

10

20

30

40

50

60

70

80

Empirical Not Empirical

Nu

mb

er

of

pat

ien

ts

Outside local guideline

Within local guideline

WP2: Observational Outcome Study: (n = 120 patients/4 Hospitals)

Characteristic Proportion

Gender Male 66%Female 34%

Age Children 28% (median age 1 year)

Adults 72% (median age 68 years)

Type of PneumoniaHAP 46%

VAP 54%

Reason for Admission

Medical 68%

Surgical 30%

Other 2%

Previous Antimicrobial Therapy (<15 days)Yes 84%

No 16%

Characteristic Median (days) Mean (days)

Time since admissionHospital 7.9 15.6

ICUs 4.0 7.4

Ventilation (n=93)

Total 11.2 11.3

Before Recruitment

3.1 7.7

After Recruitment

7.0 4.73

Predicted Activity of Antimicrobial Therapy to treat HAP/VAP by Diagnostic Result obtained by Routine Microbiology (RM) vs. FilmArray

Pneumonia resolved (n = 88)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Active therapy Inactive therapy Negative result

Pro

po

rtio

n o

f P

atie

nts

RM Filmarray

Pneumonia not resolved (n = 32)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Active therapy Inactive therapy Negative result

Pro

po

rtio

n o

f P

atie

nts

RM Filmarray

p= 0.005 p= 0.029

FilmArray Torch placed at Point of Care within 12 Critical

Care Units

Patient with suspected HAP/VAP at Participating ICU about to Receive Antibiotic

RANDOMISE466 patients

Treatment Arm

Test Respiratory Specimen using FilmArray in ICU

FilmArray Result+ INHALE Prescribing

Algorithm delivered to Treating Clinician

Control Arm

Standard of CareRespiratory

Specimen Sent to Micro Lab

Assess appropriateness of antibiotics at 24h

and clinical outcome at 14 Days

WP3 INHALE RCT

INHALE Prescribing Algorithm

WP4: Behavioral

StudyPatient Factors

Clinician Experience

Clinician Psychology

Social Factors

Contextual Factors

Consultant: I think time was a big thing because there were fewer people on at night, there is no boss. You’re racking around trying to get to see everyone with all the admissions coming in as

well. And then you see, oh, he’s had a temperature and the nurse is flapping around and keeps phoning me about it, oh, give him

some Augmentin and on we go. And perhaps I didn’t give it as much thought when I was

pressured or rushed.

Conclusions• Current methods for microbiological diagnosis of HAP and VAP are

slow and unsatisfactory

• Unyvero and FilmArray Pneumonia Panel tests performed well at pathogen detection

• Oxford Nanopore sequencing directly from sample showed tremendous promise

• 120-patient study suggests that more sensitive pathogen detection by FilmArray may improve patient Rx and outcomes

• FilmArray currently undergoing the multi-centre INHALE RCT where testing is performed at the point of care in the ICU

Funder’s Statement

• This presentation presents independent research funded by the National Institute for Health Research (NIHR) under its Programme Grants for Applied Research Programme (Reference Number RP-PG-0514-20018).

• The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health.”

UCL/UCLH

Dr. Vanya Gant

Mr Alp Aydin

Mr. Dewi Owen

Dr. Zaneeta Dhesi

Dr. Federico Ricciardi

Dr Julie Barber

Dr David Brealey

Ms. Alyssa Pandolfo

Prof. Robert Horne

University of East Anglia/Quadram

Institute

Prof. David Livermore

Dr. Justin O’Grady

Dr. Rossella Baldan

Dr. Hollian Richardson

Ms. Charlotte Russell

Ms. Juliet High

Mr. Antony Colles

Prof. Ann Marie Swart

The INHALE Study Team

& All Staff at Participating Hospitals

and Laboratories

Industrial Collaborators

bioMerieux Biofire

Oxford Nanopore Technologies

Curetis GmbH