should igras replace the tst?a-mundial-tb_2013_jrot… · • routine tb screening is a cornerstone...

Should IGRAs replace the TST?

Jim Rothel

Melbourne, Australia

Disclaimer: I am a consultant to Cellestis, a QIAGEN company

Presentation Outline

2

Is the TST / QFT dual strategy useful?

Conclusions

Should IGRAs replace the TST?

Comparative performance of QFT and the TST

Peer reviewed IGRA publications over time: Still rapidly growing - As of March 2013, > 800 on QFT

Source: PubMed

0

100

200

300

400

500

600

700

800

2004 2005 2006 2007 2008 2009 2010 2011 2012

Total

Performance of IGRAs and the TST

Clear consensus from developed country studies:

• IGRAs are more specific than the TST • Unaffected by BCG and most NTM

• IGRAs are more sensitive for active TB

• IGRA results are more closely related to measures of LTBI risk

• IGRAs are less affected by immunosuppression

• QFT is more cost-effective in almost all settings (in developed countries)

However, one area of superior performance is still debated by some

• Progression to Active TB

Progression to Active TB

Is it sensible to require QFT to have a high PPV for progression before

being widely used?

• That’s what’s being expected by some

But, QFT is a measure of MTB infection

• A necessity for progression, not a cause

• The “requirement” for a high PPV seems misconceived

Progression to active TB is dependent on many factors:

• Being infected with MTB

• Time since infection

• Infectious dose

• Immune status

• Nutritional status

• Co-morbidities

• Etc. Etc.

But despite the inappropriate expectation – does QFT detect those who will

progress?

?

“Predictive value of interferon-gamma release

assays and tuberculin skin testing for predicting

progression from latent TB infection to disease

state: a meta-analysis.”

Diel R, Loddenkemper R, Nienhaus A. Chest. 2012

6

7

PPV= 1.5%

Diel R, Loddenkemper R, Nienhaus A. Chest. 2012 .

(N = 6,385)

8

PPV= 2.4%

Diel R, Loddenkemper R, Nienhaus A. Chest. 2012

(N = 3,391)

Progression to Active TB

9

PPV= 2.7%

Diel R, Loddenkemper R, Nienhaus A. Chest. 2012

(N = 5,194)

10

PPV= 6.8%

Diel R, Loddenkemper R, Nienhaus A. Chest. 2012

(N = 1,436)

Progression to Active TB

IGRA TST p

All studies PPV 2.7% 1.5% <0.0001

High risk populations PPV 6.8% 2.4% <0.0001

Progression to Active TB

12

NPV= 99.7%

n=12,154

Diel R, Loddenkemper R, Nienhaus A. Chest. 2012

13

NPV= 99.4%

n=8,618

Diel R, Loddenkemper R, Nienhaus A. Chest. 2012

Progression to Active TB

IGRA TST p

All studies PPV 2.7% 1.5% <0.0001

High risk populations PPV 6.8% 2.4% <0.0001

All studies NPV 99.7% 99.4% <0.01

Progression to Active TB

What can we learn from predictive studies?

• IGRAs (QFT) have a very high NPV for progression

o In most populations, a negative result can be trusted

• IGRAs (and the TST) alone are not predictors of progression to active TB

o They tell us if a person is infected - a needed condition for progression

• Progression is related to risk factors

o Perhaps the most important being recent infection, immune status and co-

morbidities

• In those with these risk factors (e.g. recent contacts) the rate of progression

if IGRA positive is high (much higher than if TST positive)

o In those with older (“remote”) infection the rate of progression is much lower

With QFT we have a test that detects those truly infected – we now have

to work out how to best use it → Targeted testing

Role of QFT for TB control

Targeted testing and treatment for LTBI is the key

Priority targets should include:

Close contacts

Immunosuppressed

Major medical co-morbidities

At risk HCWs

Immigrant screening

How does QFT compare with TST in

these target groups?

Comparative performance of QFT and TST

Probably the best way to compare performance is to use estimates from

independent meta-analyses

SENSITIVITY Meta-analyses QFT TST

Diel et al Chest 2010 – Developed country 84.5% 71.5%

Sester et al ERJ 2011 80% 68%

SPECIFICITY Meta-analyses QFT TST

Diel et al Chest 2010 99.2% ND

Pai et al 2008 (BCG vaccinated) ND 59%

Pai et al 2008 (not vaccinated) ND 97%

US CDC Guidelines 99% 86%

Diel et al ERJ 2011 99.4% 88.7%

NOTE: This is sensitivity for ACTIVE TB, sensitivity for LTBI likely much higher

Comparative performance of QFT and TST

For the European setting, let us assume the following estimates:

QFT Sensitivity for active TB 83.0% Specificity 99.4%

TST Sensitivity for active TB 71.5% Specificity 78%*

(* TST specificity based o a 50% BCG vaccination rate)

What does this mean for test accuracy?

How does QFT and TST compare in accuracy? QFT Se. 83%, Sp. 99.4%. TST Se. 71.5%, Sp. 78%

@ given

Prevalence

TEST PPV (%) NPV (%) Accuracy (%)

1% QFT 58.3% 99.8% 99.2%

TST 3.2% 99.6% 77.9%

5% QFT 87.9% 99.1% 98.6%

TST 14.6% 98.1% 77.7%

25% QFT 97.9% 94.6% 95.3%

TST 52.0% 89.1% 76.4%

50% QFT 99.3% 85.4% 91.2%

TST 76.5% 73.2% 74.8%

How does QFT and TST compare in accuracy? QFT Se. 83%, Sp. 99.4%. TST Se. 71.5%, Sp. 78%

@ given

Prevalence

TEST PPV (%) NPV (%) Accuracy (%)

1% QFT 58.3% 99.8% 99.2%

TST 3.2% 99.6% 77.9%

5% QFT 87.9% 99.1% 98.6%

TST 14.6% 98.1% 77.7%

25% QFT 97.9% 94.6% 95.3%

TST 52.0% 89.1% 76.4%

50% QFT 99.3% 85.4% 91.2%

TST 76.5% 73.2% 74.8%

How does QFT and TST compare in accuracy? QFT Se. 83%, Sp. 99.4%. TST Se. 71.5%, Sp. 78%

@ given

Prevalence

TEST PPV (%) NPV (%) Accuracy (%)

1% QFT 58.3% 99.8% 99.2%

TST 3.2% 99.6% 77.9%

5% QFT 87.9% 99.1% 98.6%

TST 14.6% 98.1% 77.7%

25% QFT 97.9% 94.6% 95.3%

TST 52.0% 89.1% 76.4%

50% QFT 99.3% 85.4% 91.2%

TST 76.5% 73.2% 74.8%

How does QFT and TST compare in accuracy? QFT Se. 83%, Sp. 99.4%. TST Se. 71.5%, Sp. 78%

@ given

Prevalence

TEST PPV (%) NPV (%) Accuracy (%)

1% QFT 58.3% 99.8% 99.2%

TST 3.2% 99.6% 77.9%

5% QFT 87.9% 99.1% 98.6%

TST 14.6% 98.1% 77.7%

25% QFT 97.9% 94.6% 95.3%

TST 52.0% 89.1% 76.4%

50% QFT 99.3% 85.4% 91.2%

TST 76.5% 73.2% 74.8%

Cost-effectiveness of IGRAs, TST and chest x-ray

for TB screening

A number of papers have been published investigating the cost-effectiveness of

TB screening methods. Key populations studied are: • Contacts Mol Diagn Ther. 2008;12:235-51.

• Healthcare workers Hosp Infect. 2011;78:152-4.

• Prisons Epidemiol Infect. 2013; 3:1-11. [Epub]

• Hemodialysis patients Nephrol Dial Transplant. 2012 [Epub]

• School students Mol Diagn Ther. 2012;16:181-90

• Employees Am J Infect Control. 2011;39:e67-72

• Rheumatoid arthritis patients Mol Diagn Ther. 2010;14:367-73

• Elderly Mol Diagn Ther. 2010;14:229-36

• In almost every situation, QFT has been shown as the most efficient

screening method from both a health and a cost perspective

• Some studies have reported the TST/QFT dual strategy as the most cost

effective, but these papers have used debatable estimates of sensitivity and

specificity in their analysis. – NOTE: The UK’s NICE guidelines now report QFT alone as the most cost effective

Is the TST / QFT dual strategy useful?

Is the TST / QFT dual strategy useful?

• Interest in using the TST as the initial test and confirming positives using

QFT was first raised by the UK’s NICE guidelines

– They found this process was slightly less expensive than using QFT alone

– This guideline was NOT based on health-effectiveness

• The updated NICE guidelines now recommend IGRA only

– Largely based on updated performance data and practical issues with running

two tests

– Identified QFT alone as the least expensive method

• A small number of published cost-effectiveness studies find TST/QFT

strategy slightly cheaper than QFT alone

– Due to different models and or different performance estimates?

But cost is not the only parameter to consider – what about diagnostic

accuracy?

IGRAs versus TST in serial testing of HWs Nienhaus et al. Expert Rev. Anti Infect. Ther. 11(1), 37-48 (2013)

• Routine TB screening is a cornerstone in preventing TB in hospitals

• Clear advantage to using IGRAs for pre-employment screening

• Serial TB screening should be risk-based. Proposes 3 groups regarding frequency of screening

• High risk: regular contact with infectious TB or infectious material – screen annually, bi or tri-annually

• Medium Risk: regular contact with patients not known to have TB – screen routinely or only when exposed

• Low risk: no regular contact with patients – no routine screening

Nienhaus et al. Expert Rev. Anti Infect. Ther. 11(1), 37-48

(2013)

Missed by QFT? –

Unlikely

BCG false positive more

likely

Missed by TST

Effect of introducing IGRAs in HCW TB screening

• Less positives: reduction in number of CXRs and preventive treatment

• Low-intermediate incidence countries: 32 cross sectional studies and all but one noted lower prevalence of positive IGRA than +TST

• However, significant number of QFT+ / TST - results

Conclusion: TST followed by IGRA confirmation should not be done in low/intermediate incidence countries

954 close contacts

142 QFT-positive/

TST-positive

Chemoprophylaxis

RIF and/or INH

No active TB

51 QFT-positive

(49 TST-positive)

Mean follow-up >3.5 yr

Not treated

17 developed

active TB

343 TST

negative

5 QFT-positive

TST-negative

Not treated

2 developed

active TB

413 TST

positive

Not treated Not treated

No active TB No active TB

198 QFT-positive 756 QFT-negative

Predictive power of QFT for development of active TB Diel, Loddenkemper et al., AJRCCM, 27 August 2010

TST / QFT strategy would have

missed these two TB cases

Is the TST / QFT dual strategy useful?

• Some studies suggest it might be slightly cheaper

• However, sensitivity will be reduced compared with QFT alone

• Two different testing methods will have to be maintained

• What does the clinician do with someone TST positive / QFT negative?

• The US CDC specifically recommend against using the dual strategy

• UK NICE guidelines no longer recommend TST/QFT strategy

Should IGRAs replace the TST?

Should IGRAs replace the TST?

Some Facts:

• QFT is significantly more specific than TST, unaffected by BCG

• QFT is more sensitive in detecting active TB

• QFT results are closely related to measures of infection risk – TST results

are often not associated

• QFT has significantly higher PPV and NPV for progression than the TST

• QFT is less affected by immunosuppression

• QFT is objective, whereas TST interpretation is subjective

• QFT is more cost-effective in almost all developed country settings

Given the above, it’s hard to think of any situation where

QFT should not be preferentially used

CONCLUSIONS

Tackling LTBI is likely to provide the biggest gains in enhancing TB control

The TST has been useful, but has many performance problems

New tools now exist that have and should continue to improve TB control

• Development of enhanced IGRAs is already happening

Development of new drug regimens that simplify LTBI treatment (Rifapentene/INH) make testing for LTBI more practical

The benefit of LTBI screening and treatment lies in targeted use (testing of highest-risk groups)

Many thanks for your attention

HSIA - ARTHRITIS & RHEUMATISM 2012

5 different multicentre/multinational studies were conducted with Golimumab (J&J)

2,228 patients with RA, PsA or AS

34% BCG vaccinated

67% taking corticosteroids (prednisolone), 43% at time of testing

All screened with TST (different cut-offs), QFT and chest x-ray

Multivariate regression analysis looked at the variables:

• Methotrexate use

• Corticosteroid use

• Age (>65 years vs < 65 years)

• World region (Asia, Nth America, East Europe, West Europe, Latin America)

HSIA - ARTHRITIS & RHEUMATISM 2012

Limited agreement between QFT and TST

HSIA - ARTHRITIS & RHEUMATISM 2012

Low rate of QFT indeterminate results - 1.8%

Agreement between positive TST and QFT results

HSIA - ARTHRITIS & RHEUMATISM 2012

Test Positive %

TST 7.0

QFT 9.4

TST cut-off varied according to

local country guidelines.

Agreement between QFT and TST

fair (kappa coefficient of 0.22).

BCG vaccination explained some of the TST positive discordance

HSIA - ARTHRITIS & RHEUMATISM 2012

781 (34.2%) out of the 2,282 BCG vaccinated

BCG vaccinated Non-BCG vaccinated

Test Positive %

TST 15.2

QFT 9.1

Test Positive %

TST 5.0

QFT 5.8

Supports TST being significantly affected by BCG vaccination

Screening results by geographic region

and prior BCG vaccination

HSIA - ARTHRITIS & RHEUMATISM 2012

Region Positive % %BCG

North America 9.3 11.8

Western Europe 9.3 11.8

Eastern Europe 16.0 81.0

Latin America 15.8 78.8

Asia 33.2 40.8

Multivariate analysis of factors associated with LTBI

HSIA - ARTHRITIS & RHEUMATISM 2012

• TST results less strongly associated with region of origin but also with BCG vaccination

• QFT positive results associated with indicators of TB risk

Age > 65 years

TB rate in region of origin

• MTX or corticosteroid use did not have an effect on QFT or TST result

Small trend for higher positive rate with QFT

HSIA - ARTHRITIS & RHEUMATISM 2012

Key messages:

• QFT positivity correlated with TB risk factors – more so than the TST

• Suggests higher sensitivity than TST in immunosuppressed

• MTX (10mg/day) or corticosteroids did not effect test result

• TST confounded by BCG vaccination

• Very low number of indeterminate results (1.8%)

Authors conclusion: “In the absence of a true gold standard for latent TB infection,

results of this comparison (…) suggest that the IGRA provides greater specificity and

possibly greater sensitivity than the TST.”

Renal dialysis patients

Rogerson et al. Am J Kidney Dis. 2013 61:33-43

Analyzed published studies employing IGRAs in ESRD patients

• 9 studies were identified

• 3 TST vs QFT (n=347)

• 2 TST vs T-Spot (n=418)

• 4 3-way studies (n=361)

• Results compared to TST by risk association:

• Positive QFT more strongly associated with clinical risk than TST

– Radiologic evidence of past TB (OR 4.29, p=0.001)

– Contact to TB (OR 3.36, p=0.001)

• Negative QFT associated with BCG (OR .30 p=0.002)

- Protective?

• T-Spot: no statistical differences for any clinical risk factor

QFT versus TST in renal dialysis patients

Rogerson et al. Am J Kidney Dis. 2013 61:33-43

Major conclusions

• “ELISA-IGRA likely to be a more accurate diagnostic tool for LTBI in ESRD”

• Consistent with previous systematic reviews of general population showing QFT results are better correlated with TB exposure and independent from prior BCG

• “Propose that the ELISA-IGRA should be the test of choice”

KOBASHI – INTERNAL MEDICINE 2012

• There are differing opinions in the medical community regarding use

of IGRAs to aid active TB diagnosis

• Most studies saying not to use IGRAS are from developing countries

• This study looked at differential diagnosis of active TB in Japan:

• 66 patients who required clinical differentiation of pulmonary TB

• 22 diagnosed with TB based on culture

• 44 had other pulmonary infections confirmed

• Compared TST, QFT-2G, QFT-3G and T-Spot.TB

KOBASHI – INTERNAL MEDICINE 2012

KOBASHI – INTERNAL MEDICINE 2012

The 3 patients without TB and QFT positive, were also T-Spot positive

• Two of these had M. kansasii confirmed and would be expected positive

T-Spot was presumably false-positive in 3 patients without TB

KOBASHI – INTERNAL MEDICINE 2012

QFT had good sensitivity and specificity for diagnosing active TB in this study

HOWEVER:

• This was a developed country study, with assumed low rate of LTBI

• In a high prevalence setting would expect poor specificity

• MANY of the 22 patients with TB would have been easily diagnosed without IGRA

• No need for IGRA in most cases of active TB diagnosis

CONCLUSIONS:

• QFT can aid diagnosis of active TB, when conventional methods are inconclusive

• Use of IGRAs for active TB should be selective

• QFT (or any IGRA) should not be used as a routine diagnostic for active TB

• Especially in moderate to high prevalence settings