Xpert MTB/RIF: Technology Update

TB diagnostics and laboratory services: Actions for care delivery and sustainability

Annecy, 17-19 April 2012

Catharina Boehme

Conflict of interest

� FIND is a non-profit foundation devoted to developing

diagnostic tools for poverty-related diseases.

� In this role, FIND has development partnerships with � In this role, FIND has development partnerships with

industry, including with Cepheid.

� FIND has no financial benefits in any form.

Accuracy of Xpert MTB/Rif for detection of pulmonary

TB and Rif resistance: a metanalysis

� Included 18 studies (10,224 specimens)

� Pooled sensitivity: 90.4% [95% CI 89-91]

� Pooled specificity: 98.4% [95% CI 98-99]

� Pooled sensitivity for RIF: 94.1% [92-96]

� Pooled specificity for RIF: 97% [96-98]

Xpert performance in detecting extrapulmonary TB

Chang et al., J Infect 2012Chang et al., J Infect 2012

Tortoli et al., ERJ 2012

Reference Tissue Node CSF Gastric Pleural Urine Stool Comment

Ligthelm et al., JCM

2011

28/29

(96.6%)

FNAB material added to PBS

then treated SR:sample = 2:1

Hillemann et al., JCM

2011

20/29

(69%)

7/8

(87.5%)

5 /5

(100%)

2/2

(100%)

SR:sample = 3:1

Teo et al., JCM 2011 2/3

(66%)

4/4

(100%)

SR:sample = 2:1

Non-respiratory specimens - Xpert sensitivity

compared to culture

(66%) (100%)

Vadwai et al., JCM

2011

Biopsy

54/70

(77%)

1/3

(33%)

Body fluids

16/21

(76%)

SR:sample = 2:1

Many patients already treated

before biopsy

Miller et al., JCM

2011

Smear positive

4/4 (100%)

Smear negative

3/4 (75%)

Total

7/8 (88%)

SR:sample = 3:1

Zeka et al., JCM 2011 Smear positive

4/4 (100%)

Smear negative

21/44 (48%)

Total

25/48 (52%)

SR:sample = 3:1 Smear

negative

Causse et al., JCM

2011

Total

39/41 (95% )

Cobas TaqMan 78% sensitive

and 98% specific

Friedrich et al., JCM

2011

5/20

(25%)

SR:sample = 2:1

Moure et al., JCM

2012

5/12

(42%)

24/34

(71%)

2/2

(100%)

2/3

(67%)

7/26

(27%)

2/3

(67%)

2/2

(100%)

SR:sample = 3:1 Smear

negative

Tortoli et al., ERJ

2012

Total

188/238 (79%)

Xpert performance in pediatric TB

Baseline characteristics

Nr of

participants

Median

age

HIV

infection

Specimens

Nicol et al,

Lancet ID 2011

452 1.6 years 24% Induced

Rachow et al, 164 5.8 years 51.2% Induced and

Rachow et al. 2012

Rachow et al,

CID 2012

164 5.8 years 51.2% Induced and

spontaneous

Sensitivity and specificity for pediatric TBdetection

Sensitivity in

C+ (95 CI)

Specificity in

C- (95 CI)

Nicol et al,

Lancet ID 2011

76%

[64-87]

99%

[98 - 100]

Rachow et al,

CID 2012*

75%

[55-89]

100%

[99 - 100]

Sensitivity Specificity

*4/47 (8.5%) Xpert positive among highly probable TB

Cost-effectiveness

Citation Results suggest

Andrews JR et al.

AIDS 2012

CE for screening pre-ART

Abimbola TO et al.

JAIDS 2012

CE at reducing early mortality during first 6 mo ART

Vassal A et al.

PLoS Med 2011

CE compared to smear & clinical diagnosis S-TB

PLoS Med 2011

Meyer-Rath G et al. Johannesburg

HE²RO Policy Brief 2011

Cost modeling: incremental cost, higher # cases

diagnosed; do not capture benefits or opportunity costs

Theron G et al.

Eur Respir J 2011

Low resource settings smear & Xpert combined highest

accuracy and lowest cost-of-diagnosis

� In depths CE assessments in various diagnostic algorithms / clinical settings needed.

�However, currently available CE studies sufficiently favorable; unlikely that subsequent

studies will show an overall cost disadvantage.

�Rather, additional studies likely to serve only to better define the when/how.

Assay refinements 2010-2011

ADF 1

• Evaluation studies (99% RIF sensitivity & 100% specificity)

• Manufacturing scale up (drop to 96% RIF specificity)

ADF 2

• May 2010

• Demonstration studies (95% RIF sensitivity & 98% specificity)

EGM / STAG

ADF 3

• Oct 2010 (software change)

• Goal: RIF specificity increase

ADF 4

Cartridge

• Q4 2011 (software, Probe B and fluidics change)

• Goal: RIF specificity & sensitivity increase; decreased error rate

ND

• Further refinements as part of post-marketing surveillance

EGM / STAG

� Increased error rate >5% seen with G3; primarily Signal Loss Errors

� Probe B sensitive to annealing temperature fluctuations; occasional false Rif

resistants seen

Probe Ct at 64C - 68C anneal

n = 4 per temp

Stabilized probe B wild-type hybrids during temperature

increase above the optimum PCR anneal temperature

Reasons for introduction of G4 assay

n = 4 per temp

25.0

26.0

27.0

28.0

29.0

30.0

31.0

32.0

33.0

34.0

35.0

64C 67C 68C

FAM_Ct

CF3_Ct

CF4_Ct

CF5_Ct

CF6_Ct

Modifications incorporated in G4 assay

Summary of modifications

Fluidics Modified fluidics to improve robustness and reduce error rate

Assay

Settings

Modified background subtraction range

Modified valid rpoB Ct range for probes D and E

Modified valid SPC Ct range

Modified thresholds for probes A, C, D and E Modified thresholds for probes A, C, D and E

PCR

Cycling

Shortened PCR2 anneal time

Beads Added fluorescent tracer for refined probe check control

Probe B Modified molecular beacon sequence/quencher

� Xpert MTB/Rif G4 including updated ADF released December

2011.

Error Rate Significantly Reduced

Xpert non-determinant (ND) rate

Xpert ND Rate for

Case detection

Xpert ND Rate for

Rifampicin susceptibility

G3 5.4%

(34/627)

0.5%

(1/210)

G4 0.4% 0.9% G4 0.4%

(5/1252)

0.9%

(6/692)

Chi-Square Test

p-value

<0.001 0.69

� Significant reduction of non-reportable results for Xpert

MTB/RIF G4 assay (G4).

� No 5011 errors seen for G4.

Accuracy of G4 Assay Maintained

Overall Accuracy of Xpert MTB/Rif G3 assay versus G4 assay

Sensitivity in

C+

Specificity in S-

C-

Rif-R

Sensitivity

Rif-S

Specificity

G3 82.6%

(142/172)

[76 - 87]

98.8%

(242/245)

[96 - 100]

100.0%

(12/12)

[76 - 100]

100.0%

(127/127)

[97 - 100]

G4 86.0% 98.0% 100.0% 99.2%

(148/172)

[80 - 90]

(240/245)

[95 - 99]

(12/12)

[76 - 100]

(126/127)

[96 - 100]

McNamar’s

Test

p-value

0.51 0.88 0.82 0.95

� High sensitivity and specificity for TB case detection and RIF

susceptibility testing. No significant differences compared to G3.

� G4 assay unlikely to generate false-resistant RIF susceptibility calls

in the event of annealing temperature abnormalities.

Prototyping remote monitoring tool

Monitor Remote

Deployments

Support QC programs

Consumable logistics

Monitor disease and

resistance

Response

Support National

reporting

Xpert Test

Result Portal Support

reader

Medical

Reader

Cepheid

Xpert Test

Monitor

Overcoming hurdles to rollout: Towards remote

calibration

� Aims to eliminate need for annual module replacement.

� Recalibration of optical system, verification of the thermal

system and conduct of a series of system-level tests to

ensure full system functionality within specifications.

� The calibration kit includes five calibration cartridges &

software.software.

� Expected release: September 2012.

User runs calibration

software & loads the

calibration cartridges

Cepheid analyzes data

Activation code is provided

to user to update calibration

Certificate issued

Customer

Service notified

for follow-up

Data sent by internet or by CD

Service network and warranty

� Authorized service partners now present in 86 of the 145 HBDC countries.

� Progen (South Africa) is now Cepheid RSA, acquisition Feb 2012.

� Intense support to Labindia, to improve local service.

�HBDC test ordering only from Cepheid HBDC France and from Cepheid RSA

(for deliveries in RSA and neighboring countries).

� General warranty on systems is 12 months, from shipping date; warranty for

HBDC systems shipped from now on will be extended to 24 m (for GeneXpert

IV models only).

� Footprint reduction

� Hazard compliable

Packaging changes to reduce waste and shipping cost

Xpert MTB/RIF – 50 Pack example

� Hazard compliable

� Recyclable

� Shipping cost reduction

x5

Increased experience in trainings

Summary 2011/2012

2011

Sessions

2012 Q1

Sessions

Total

Trainees

Region /Organization

Remote

trainings50 20 109

Malawi Chipungu/ Mozambique/ FIND LabIndia,/ Orchids Diagnostics Nepal/ MSF

Kenya / MSF Kyrgyzstan CDC/ Right to care UCT/ intermedico Canada/CDC kemri/,

Caroga/, Iqaluit Canada/ Institute of Lisboa, Reference Lab of Conakry Guinea, Projuktitrainings

(Web, Tel, …)

50 20 109 Caroga/, Iqaluit Canada/ Institute of Lisboa, Reference Lab of Conakry Guinea, Projukti

Bangladesh, 'Waltraud Wernhart'-association , Germany, LATAM Distributors, CAREC

Trinidad, Tobago, MRC Gambia, PATH, St Johns Hospital for INDIA, Hospital Bichat

France for Cameroon, Belarus Tolosina, Rochembiocare, rocarsystem, CIDRZ, SPD,

LABINDIA, NRL Kyrgysthan

Mauren’s

training

sessions

20 3 52

WHO (11) FIND (11) Pointe Care ZMBW (3), MSF (5) KNCV (2), Myanmar Consultant (1),

end-user Senegal (1), CAROGA (2) , Fajarmasmurni Indonesia (2), MEDDENT Papua New

Guinea (1), MOH of Kenya (2), Orchids Diagnostics(1), CDC TANZANIA (2), ZAMBART

PROJECT (2), ANRS France (1), Mozaic (1), Living Group Malawi (1), GBG Moldova(2), NRL

Tadjikistan (1)

On site

trainings7 6 181

South Africa (11), MSF (13), KNCV (8), Institute of Lisboa in Portugal (6)

Institute of Munich in Germany (5), NTRL (17), MACARE Philippines (8)

UNION congress (various, about 21), Algeria-Institut Pasteur (11), Cameroun-Institut

Pasteur (24), India-Labindia training (13) - FIND consultants (24) – Enduser india

(Dharavi, Indore, LNJ) (20)

� More than 340 persons trained

�Training guidelines now available for partners

Current HBDC Training Model

Training Type How/Where Comments

Remote Web / Tel

Sessions of 45 min

On Appointment.

Useful to accelerate the discovery of the system/tests

Can cover Installation, Software, Test Discovery, maintenance.

Free of charge.

At Cepheid

Toulouse

2 d training Prescheduled sessions. One per month or more.

Available Calendar; Booking required.

Cover all points in a detailed way including unit maintenance (module exchange).

Training is free of charge, meals are included.

Not included: Travel expenses to site and hotel expenses.

At customer’s 1 to 2 days On demand installation/training with partners. booking required. At customer’s

site by Cepheid

partners

1 to 2 days

training

On demand installation/training with partners. booking required.

Covers installation, training, use of the systems & of the Xpert MTB/RIF assay.

Training time is covered by Cepheid and paid to partners.

Not included: Travel and time expenses to go to training site and hotel/food

expenses. Partners are authorized to charge (with quotation approval from end

users) for covering these expenses.

At customer’s

central site by

Cepheid

(for large programs)

1 to 2 days

centralized

training

Centralized training for country(ies) program launch (multiple centers)

On demand course or session covering installation, use of the systems & of the

Xpert MTB/RIF assay.

Training time is covered by Cepheid.

Travel and time expenses to reach training site and hotel/food expenses. To be

quoted by Cepheid.

Panels for validation and EQA

� NHLS panel (dried, inactivated MTB culture spots on filter cards)

� GLI panel (artificial sputum spiked with heat inactivated MTB)

� Vircell panel (lyophilized artificial sputum spiked with heat inactivated MTB)

� CDC panel (DNA)

A multi-disease platform with expanding menu

Progress on Xpert HIV viral load testing

� The Xpert® HIV-1 Test is an in vitro diagnostic test designed for quantitation

of HIV-1 in human plasma over the range of 40 to 10,000,000 copies/ml.

� The test will also be intended for use in assessing and diagnosing newborn

infection (CE IVD labeling). infection (CE IVD labeling).

� The test is designed around detection of the HIV-LTR; as such, it is expected

to be able to detect novel and emerging HIV-1 groups, as well as HIV-2, but

quantitation claims for these targets will not necessarily be in the label.

� Expected release date for CE-IVD: Dec 2013.

Molecular assays for treatment

monitoring using propidium monoazide?

Comparison Ct mean values obtained from sputum

samples collected before starting treatment (t0) and

10-20 days after the beginning of anti-TB therapy (t1).

PMA treated Not treated

PPaolo Miotto, Andrea M. Cabibbe, Sara Bigoni, Alberto

Matteelli, Daniela M. Cirillo.

RETREATMENT OF CLINICAL SPECIMENS WITH PROPIDIUM

MONOAZIDE ALLOWS THE USE OF MOLECULAR ASSAYS FOR

MONITORING THE RESPONSE TO THERAPY IN TB PATIENTS

(Presented as oral presentation at the 32nd Annual Congress of

the European Society of Mycobacteriology, Lubeck – D, 26-29

June 2011)

Thank youThank youThank youThank you

23

Development study, Latvia

HBDC Warranty Program

� General warranty on Systems is 12 months/365 days, from shipping date

� Warranty can start from Installation date IF the installation qualification report is sent to

Cepheid immediately after installation (Unjustified delay between shipment and

installation might be rejected by Cepheid )

� Warranty covers parts and labor and shipping charges, but it does not cover calibration

� Warranty applies if defective parts can be collected back at Cepheid’s demand

� Cepheid may request some collaboration from the end user - for customs clearance for � Cepheid may request some collaboration from the end user - for customs clearance for

instance - in case the country of destination is not yet covered by a partner

� The actual test is NOT covered by a manufacturer warranty except if a recall procedure is

initiated or a Lot issue is identified and replacement is decided by Cepheid

� Warranty for HBDC systems shipped from now on will be extended to 48 months for

GeneXpert IV models only

� To date customers can purchase a complete one year warranty extension contract for

2 900 USD for a GX IV4 Modules. This includes the present Swap based calibration process

� Prices might be changed when the Remote Calibration process is implemented.