IGRAs for Diagnosis of Tuberculosis: 2010 Update Nira Pollock, M.D., Ph.D. Division of Infectious Diseases Beth Israel Deaconess Medical Center Boston,

IGRAs for Diagnosis of Tuberculosis: 2010 Update

Nira Pollock, M.D., Ph.D.Division of Infectious DiseasesBeth Israel Deaconess Medical CenterBoston, MA

May 1, 2010

Problems with the PPD

False positives• Recent BCG vaccine• non-TB mycobacteria (NTM)

False negatives• 25-30% patients with active pulmonary TB

initially negative• Newborn/elderly, immunosuppression,

renal failure, acute non-TB infection, etc• unable to distinguish active disease from

past exposure

Interferon-gamma Release Assays (IGRAs):

basic concepts• Expose T cells (isolated, or within whole blood) to:

– TB antigens (in peptide form), vs– positive control antigen (“mitogen,” e.g.

phytohemagglutinin A), vs – negative control (e.g. saline)

• Incubate overnight :– T cells (both CD8 and CD4) previously sensitized to

these TB antigens in vivo release IFN-– Mitogen stimulates cells non-specifically to release IFN-

as control for general T-cell anergy– Saline control defines level of background (should be

low)• Quantify amount of IFN-produced under each condition

IGRAs: basic concepts, cont.

• Theory: overnight incubation detects sensitized “effector” T cells, i.e. already activated in vivo (longer incubation could activate resting “central memory” T cells also)– ? Primarily CD4+ (Mack et al, 2009)

• Like PPD, IGRAs are unable to distinguish between LTBI and active disease

Mack et al, TBNET consensus statement; Eur Respir J 2009

Quantiferon-TB-Gold (Cellestis, Inc.)FDA-approved May 2005 for detection of LTBI and TB disease

** must incubate cells with antigen within 12 hours of collection

Quantiferon-TB Gold

• Peptide antigens used in assay simulate two proteins specific to Mycobacterium tuberculosis complex (MTBC: M. tuberculosis, M. bovis, M. africanum, M. microti, M. canettii):

– ESAT-6, CFP-10 (genes coding for both are found within MTBC RD1 region, which is deleted in M. bovis BCG strain)

– Eliminates false-positives due to BCG vaccination and to almost all NTM (**exceptions: M. kansasii, M. marinum, M. szulgai)

3rd generation: QFT-Gold In Tube (IT)

• FDA-approved October 2007 • NOW FORMALLY REPLACING 2nd GENERATION—

company no longer making prior version!

• Specimen collection: draw whole blood directly into three proprietary 1 mL blood collection tubes: – 1) TB-specific Ag (dried onto wall of tube)– 2) Nil (-) control– 3) Mitogen (+) control (dried onto wall of tube).

– TB-specific peptide antigens: ESAT-6, CFP-10, TB7.7. Goal of adding extra antigen: increase sensitivity. (Like ESAT-6/CFP-10, TB 7.7 is not present in BCG strains and most NTM.)

QFT-Gold IT (continued)• Must not under or over-fill tube. Shake 10x vigorously after draw.

Keep at room temp.• Must put at 37ºC within 16h of collection.• Incubate upright at 37ºC for 16-24h • Tubes can then be held at 2-27ºC for up to 3 days prior to

centrifugation (so can ship at room temp). • Centrifuge 15’ to separate plasma from cells, remove >150 L

plasma to assay (can store spun tube or plasma at 4ºC for 28 days).• Quantify IFN- in plasma by ELISA, as for QFT-G

• IT test format allows o/n incubation at site of draw (e.g. hospital or clinic), vs central testing center

• QFT-G IT is being done at the Hinton State Lab (contact them to obtain tubes and arrange submission); also offered at e.g. Quest

QFT-G IT, continued

• Quantification of IFN-: ELISA, as for QFT-G– Results readout: positive, negative, or

indeterminate – Ideally, lab should report absolute value result in

IU/mL, so that clinician can evaluate how close absolute value is to the cutoff

– Lab should also report reason for indeterminate• Low mitogen response: insufficient or

dysfunctional lymphocytes, reduced lymphocyte activity due to prolonged specimen transport, improper specimen handling

• High background in nil control: heterophile antibodies (interfering human anti-mouse antibodies), intrinsic IFN-gamma secretion (? recent vaccination, ? just true for some people--1-2% of population per Cellestis website)

QFT-G IT results interpretation

Note: for QFT-G this value was >50%; seems that new cutoff would generate more positives

Note: for QFT-G nil cutoff was 0.7 IU/mL; seems that new cutoff of 8.0 would generate a lot fewer indeterminates..

T-Spot.TB (Oxford Immunotec; Elispot technology)

FDA-approved July 2008

• in vitro diagnostic test based on an enzyme-linked immunospot (ELISPOT) method

• enumerates M.tuberculosis-sensitized effector T cells responding to stimulation with a combination of peptides simulating ESAT-6 and CFP10 antigens, by capturing interferon-gamma (IFN-γ) in the vicinity of T cells from which it was secreted

Each spot=one reactive effector T-cell

T-Spot.TB

TSpot.TB results interpretation

• Positive: (ESAT-6-Nil) and/or (CFP-10-Nil) are > 8 spots. – (***note: this cutoff used to be 6 spots)

• Negative: both (ESAT-6-Nil) and (CFP-10-Nil) are < 4 spots.– (includes values less than zero).

• Borderline (equivocal): highest (TB antigen-Nil) spot count is 5, 6 or 7 spots– Collect a new specimen and retest

• Indeterminate: – nil control count is >10 spots, OR– mitogen control count is <20 spots and (TB Ag-nil) counts are <4

spots

Doing T-Spot in MA

• Oxford Immunotec has a testing facility in Marlborough (since July 2009): CLIA/CAP certified

• Specimens (blood only) must be shipped at room temp day of draw and have 32 hours to reach testing center (package insert says 8 hours, but Oxford has validated longer time frame); contact Oxford for details (tubes, shipping)

Assessing the accuracy of IGRAs

General principles used to date:

– Sensitivity: approximated by measuring proportion of positive tests in patients with culture-confirmed active TB

– Specificity: approximated by measuring proportion of negative tests in patients with low risk for TB infection

Problem: no confirmatory test exists for diagnosis of LTBI or culture-negative TB disease (no gold standard!)

QFT-G IT package insert (Jan 2009)

• Sensitivity in culture-confirmed active TB (all with <8days treatment prior to testing):– Japanese study (n=92): QFT-G IT 93.5%, QFT-G 83.7%– Australian study (n=27): QFT-G IT 88.9%, QFT-G 74.1%– US study (n=44): QFT-G IT 84.1%, QFT-G 77.3%

Overall sensitivity: QFT-G IT 89%, QFT-G 81%

• Specificity in subjects at low reported risk for TB infection (US study; subjects had no reported TB risk factors, and none had BCG history): Overall specificity: QFT-G IT 99.2%, QFT-G

99.8%, TST 99.1%

QFT-G IT package insert (Jan 2009)

• Cautions that the performance of the USA format of QFT-G IT has not been extensively evaluated in:– Individuals who have impaired or altered immune function such as

HIV infection/AIDS, s/p transplantation managed with immunosuppressive treatment, patients on immunosuppressive drugs (e.g. corticosteroids, methotrexate, azathioprine, cancer chemotherapy)

– Patients with the following clinical conditions: diabetes, silicosis, chronic renal failure, hematological disorders (e.g., leukemia and lymphomas), and other specific malignancies (e.g., carcinoma of the head or neck and lung).

– Individuals younger than age 17 years

– Pregnant women

Review of TSpot.TB FDA approval document/PI (July 2008)

• Sensitivity in culture-confirmed active disease (n=183): 95.6% using >6 spots, 90.7% using >8 spots.

• Specificity (used individuals with no TB risk factors and negative TST) (n=306):

97.1% using >6 spots, 99.0% using >8 spots (i.e. if “equivocal” (5,6, or 7 spots) are counted as negative).

• “Equivocal” or “borderline” result (TB Ag-nil = 5,6, or 7 spots): represents the area of overlap between results obtained for culture-confirmed positive samples and low risk TB negative samples

• Note: Oxford Immunotec website (4/12/10) quotes 95.6% sensitivity and 97.1% specificity, but these are for >6 spot cutoff, whereas current version uses >8 spot cutoff and equivocal range.

TSpot.TB FDA approval/PI (July 2008): clinical studies

• Goal: include subjects from all major risk groups indicated for TB screening by CDC guidelines (including those with potential for false positive/negative TST)

• TSpot.TB vs TST evaluated in typical candidates for routine LTBI screening, with various risk of exposure and progression (n=1403) (NOTE: used >6 spot cutoff)

• Included 328 HIV+, 229 recent contacts, 122 drug-induced immunosuppression, 97 IVDU, 108 DM, 195 ESRD. Many BCG-vaccinated and foreign-born. 93 children/adolescents.

TSpot.TB FDA approval/PI clinical studies:aggregate results (not by clinical subgroup)

• After controlling for the other variables, positive results for both T-SPOT.TB and TST were significantly associated with history of prior TB infection.

• A positive result for T-SPOT.TB was significantly associated with contact with infectious TB and birth in a TB endemic country; no such association observed for TST.

• A positive TST was associated with BCG vaccination; no such association observed for T-SPOT.TB

• A negative TST was associated with being immunocompromised; no such association observed for T-SPOT.TB

• TSPOT.TB results were not impacted by age

TSpot.TB FDA approval document/PI (continued)

• Notes theoretical cross-reaction (false-positive test) with M. kansasii, M.szulgai, M. marinum, M. xenopi, M gordonae (latter two not mentioned in QFT-G IT PI). However, actual data obtained in a very small # of patients—12 with MAC (all negative with TSpot), 1 with xenopi (positive), 4 with gordonae (all positive), 1 with kansasii (positive). (note: no marinum..)

• “The performance of this test has not been adequately evaluated with specimens from individuals younger than age 17 years, in pregnant women and in patients with hemophilia”

Direct comparisons of QFT-G IT, TSpot.TB, and TST: meta-analysis

Diel et al, Chest 2010• Evaluated comparative sensitivity in studies of subjects with active

TB confirmed by culture and/or PCR and/or histologic evaluation, treated for <2 weeks

• Evaluated comparative specificity for LTBI in studies of subjects who were healthy, native residents of low-incidence countries without any previously known exposure to TB, irrespective of BCG vaccination status.

• Evaluated indeterminate rates (though no apparent distinction between indeterminates due to high background, vs low mitogen)

• Included studies that evaluated immunosuppressed subjects

• Note: cutoff for TSpot.TB+ was >6 spots in all included studies, which as discussed is different than FDA-approved version

Diel et al metaanalysis, cont• Pooled sensitivities in active TB:

– TST: 69.9%– QFT-G IT: 81%.

• Note that in studies done in developing countries, sensitivity was 74.3%, vs 84.5% in developed countries. (Is this difference due to HIV co-infection, malnutrition, or other factors?)

– TSpot.TB: 87.5%. • Majority of studies done in developed countries; sensitivity in that

subgroup was 88.5%

• Pooled specificities in low-risk subjects:– QFT-G IT (5 studies): 99.2%– TSpot.TB (3 studies): 86.3%

• Pooled rates of Indeterminates:– QFT-G IT: 2.1%. In immunosuppressed subgroup: 4.4%– TSpot.TB: 3.8%. In immunosuppressed subgroup: 6.1%

IGRA performance in specific groups of interest

e.g. contacts of active TB cases• Overall consensus, IUATLD NAR meeting,

Vancouver, 2007: overall both IGRAs performing well (and comparably) in contact investigations

• Tspot.TB and QFT-G (including IT version) results correlate better than TST results with exposure to MTB1, 2

• Direct comparison TSpot.TB vs QFT-IT vs TST, 20093: both IGRAs appeared to indicate LTBI more accurately than TST, and IGRAs agreed well

• Suggests that IGRAs may be as or more sensitive than TST for recently acquired infection (in immunocompetent)

(1 Richeldi, AJRCCM 2006; 2 Arend et al, AJRCCM 2007; 3 Diel et al, Chest 2009)

1Kimura et al 1999; Chapman et al, 2002; Carrara et al, 2004, Dheda et al 2005

e.g. HIV Data mixed: can use IGRAs, but watch for

indeterminates, particularly at low CD4• QFT-G IT: e.g. Brock et al, 2006, Denmark: %

indeterminates correlated with low CD4 (24% in pts with CD4<100).

• ELISPOT assays1: overall perform better than TST. E.g. Dheda et al, 2005: T-Spot.TB in HIV-positive pts w/o other TB risk factors: technical performance independent of CD4 count. However, another study found more indeterminates with Tspot vs QFT-G (Stephan et al, 2008)

• Tspot may be more sensitive than QFT-G in this population (Mandalakas et al, 2008, small study in S. African patients)

HIV, continued• Diagnosis of active TB in HIV+:

– QFT-G IT might be a sensitive tool for detection/prediction of active TB in HIV+ (Aichelburg et al, CID 2009), or NOT… (Aabye et al, PLoSONE 2009)

– Cattamanchi et al, BMC ID 2010: TSpot.TB in 236 HIV+ active TB suspects in Uganda; mean CD4 of 49. 126 patients diagnosed with active TB by culture. 10% of subjects had insufficient mononuclear cell counts for TSpot assay. Of remainder:

• 25% had indeterminate results• IGRA sensitivity was 73% • Proportion of positive test results was similar across CD4 count

strata• **IGRA results did not meaningfully alter the probability of active

TB in patients with negative sputum smears

– **If IGRA sensitivity might be lower in HIV+ subjects (vs immunocompetent) with active TB (recall also Diel metaanalysis), what does this mean re: sensitivity for LTBI?

“Immunocompromised” patientsIGRAs (vs TST) do allow optimization of experimental conditions in vitro,

e.g. incubation time or adjustment of cell numbers, allowing potential for higher sensitivity. However, studies are as usual limited by lack of gold standard for LTBI.

• Overall: IGRAs seem to work, but true sensitivity for LTBI unknown.• In earlier studies, QFT-G had higher rate of indeterminates (low mitogen

control) than TSpot.TB (Ferrara et al, AJRCCM 2005 (Italy); Piana et al, AJRCCM 2006 (Italy); Ferrara et al, Lancet 2006 (Italy))

• More recent metaanalysis (Diel et al, 2010): rates of indeterminates (note: reason for indeterminate not defined) in “immunosuppressed” subgroups:

– QFT-G IT: 4.4%– TSpot.TB: 6.1%

• Occasional case reports of IGRAs being used to help with Dx of active TB in TST-negative immunosuppressed patients

• Disturbing case report of person who was QFT-G negative before liver transplant AND in setting of post-transplant active (Cx-positive) pulmonary TB (Codeluppi et al, 2006)

e.g. health care workers (HCW): depends where you are and what question you ask. For example:

• Japan (Harada, 2006): QFT-G vs TST – 95% s/p BCG. 93% TST>10mm, vs 10% +QFT-G. +QFT-G results were a/w

LTBI risk factors, while +TST results were not.• Rural India (Pai, 2005): QFT-IT vs TST

– 50% positive by either test, 31% by both• Russia (Drobniewski, 2007): QFT-IT

– QFT-IT was positive in 8.7% of medical/non-medical students, 39.1% of all doctors/nurses, 46.9% of TB doctors and nurses

• Denmark (Soberg, 2007): QFT-G vs TST– ID dept employees: 34% TST+, 1% QFT-G+. 89% of TST+ were BCG-

vaccinated. • Urban US (Pollock, 2008): QFT-G

– In TST+ newly hired employees with increased risk of having LTBI (large PPD, residence in highly endemic area, recent or remote contact, conversion, CXR findings c/w old TB, patient care): 28% QFT-G+, 70% QFT-G-

• Many more….mostly descriptive (TST results vs IGRA results)

Patients approaching TNF-alpha blocker therapy

The problem: many have underlying diseases or are on immunosuppressive medications which can compromise TST sensitivity. But how sensitive are the IGRAs in this group? Again, limited by lack of a gold standard.

• E.g. Laffitte et al, Br J Dermatol 2009; retrospective study of TST vs T-Spot.TB in 50 patients with psoriasis considering TNF-alpha blocker (in Switzerland)– Positive TSpot was strongly a/w presumptive Dx of LTBI (by risk

factors), while TST was not– 20% of subjects had positive TST and negative TSpot and were NOT

treated for LTBI; no reactivation detected with median f/u of 64 weeks (but note, small numbers overall)

• E.g. Diel et al, Pneumologie 2009 (German recommendations): due to expectation of false negative AND false positive TST in these patients, they recommend “highly specific” IGRA instead (but what about IGRA sensitivity??)

Children• Lewinsohn, Lobato, and Jereb, Curr Opin Pediatrics 2010:

– Overall, performance of IGRAs equivalent or superior to that of the TST, but evidence supports usage of IGRAs in children aged 5 years or older only (insufficient evidence re: performance in younger kids, and sensitivity poorly defined in that group)

– In kids >5, IGRAs preferred over TST when specificity is paramount or when patients might not return for TST reading

– Kids <5: TST preferred

• E.g. Bianchi et al, Pediatr Infect Dis J 2009:– QFT-G IT was positive in 15 of 16 (93.8%) children with active pulmonary TB – Among IGRA+ children (excluding active TB), TST- were significantly younger

than TST+ children (so could IGRA be more sensitive than TST in younger kids?)

Are CFP-10, ESAT-6, +/- TB7.7 sufficient for comprehensive detection of

LTBI? • Overall: in contact investigations,

sensitivity of IGRA=TST, and IGRAs correlate better with TB exposure

• For active TB, sensitivity of IGRAs = or > to TST

• Could IGRAs be sensitive to recent/active infection, but not remote infection? (Pollock et al, ICHE 2008)

Relying on IGRAs for making clinical decisions: how much caution should

we use at this point?

• if we base Tx decisions on IGRA results alone, many individuals with clinical risk factors historically considered suggestive of true LTBI will suddenly be exempt from treatment. Is this good or bad? – AND, some of these risk factors have

historically been associated with increased reactivation risk (e.g. PPD>15 mm, recent immigration from high risk country, various CXR findings)

• But can IGRAs actually distinguish those at higher reactivation risk? Should we only care about the IGRA+?

Studies of predictive value of IGRAs

Hard to do studies of predictive value of positive IGRA for development of active TB—typically, ethically would need to consider treatment of LTBI if IGRA+…….

• E.g. Diel et al, AJRCCM 2008, Germany: evaluated rates of progression to active TB in close contacts (immunocompetent) within 2 years of contact screening. – 11% of contacts were QFT-G IT+, vs 40% TST+ (>5 mm). – 41 QFT-G IT+ subjects refused LTBI treatment; 6 (14.6%) developed

active TB. 219 TST+ subjects refused treatment; 5 (2.3%) progressed to active TB. Concluded that QFT-G IT is a more accurate indicator of LTBI than the TST and provides at least the same sensitivity for detecting those who will progress to active TB.

• Vs. e.g. Kik et al, Eur Respir J 2009: looked at immigrants who were close contacts of smear + TB cases, all found to have TST >5 mm during contact investigation: followed for 2 years. – PPV for progression to TB disease was comparable and LOW for QFT-G

IT (2.8%), T-Spot TB (3.3%), TST>10 mm (3.1%), TST >15 mm (3.8%)

Predictive value of IGRAs, cont

• E.g. Hill et al, PLoS One 2008, The Gambia: risk of progression to active TB after positive ELISPOT (similar to TSpot) or TST in case contacts, over 2 year period. Noone got preventive therapy, per local guidelines. – Rates of progression in ELISPOT+ was similar to rates in

TST+. – Because initial ELISPOT and TST were each positive in just

over half of secondary cases, while 71% were initially positive by one or the other test, they concluded that positivity by either might be the best indication for preventive treatment.

– Note: there were clearly some NEW infections over study time period (discordant genotyping between index and secondary case isolates) so this really confuses this study.

• San Francisco IGRA experience--?? Not seeing spike in TB cases after switching to IGRA only for TB screening programs…..

Our clinical response to all this data: We feel great about the IGRA+. We’re just

not sure what to do with all the IGRA-…

• We don’t assume (for now) that a negative IGRA rules out LTBI. Perhaps, in future, we can be confident that it does—or, at least, that it rules out high baseline reactivation risk.

• Consider offering treatment to certain high-risk populations even with a negative IGRA result:

– 1. ***Patients with medical risk factors placing them at higher risk of TB reactivation if they do have LTBI, i.e. HIV, chronic oral steroid treatment, TNF-alpha blocker treatment, renal insufficiency, diabetes, some malignancies.

– 2. recent TB contact (debatable, given good IGRA performance in contact studies)

– 3. PPD conversion (>10 mm increase) in past 2 years (also debatable, given performance in contact studies)

– 3. Abnormal CXR potentially consistent with old TB in significant burden (e.g. large scar, nodule, after r/o with smear/culture)

FAQ: Do positive IGRA results turn negative with TB or LTBI

treatment?• Multiple studies on this topic: data mixed, but general

consensus is NO, not reliably.– E.g. local study: Pollock et al, ICHE 2009: HCW treated for LTBI

with 9 months INH still had positive QFT-G after treatment.

• Suggested approach to this issue based on current data:– IGRA results should not be used to assess the effectiveness of

recent or remote treatment courses for TB/LTBI: many (if not most) individuals will continue to test positive after standard therapy

– Do not assume that an individual who reports prior TB/LTBI therapy but still tests positive by IGRA has not been appropriately treated in the past

– Neither providers nor patients should expect reliable changes in IGRA results after standard treatment

Serial testing with IGRAs

• Primarily relevant to HCW or other individuals requiring annual screening

• Multiple issues to think about:– reproducibility of test results in a given individual

tested repeatedly over time, without intervening exposures to TB

– appropriate definition of reversion/conversion– optimal test cutoffs – (e.g. initially raised by Pai et al, 2006,2009,

India)

From QFT-G IT package insert:

• The magnitude of the measured IFN-g level cannot be correlated to stage or degree of infection, level of immune responsiveness, or likelihood for progression to active disease.

Reproducibility of IGRA results in serial testing

• E.g. Detjen et al, Clin Vaccine Immunol 2009: 27 S. African HCW, tested with QFT-G IT on day 1 (2 tests, by different operators) and day 3 (1 test). – 6/27 had discordant results of some kind– variability in the magnitude of IFN-gamma responses between assays

performed for a given individual– most variability seen in assays that were obtained from an individual on

two different days. – Conclusion: This intra-individual variability could influence

interpretation of serial measurements• E.g. Van Zyl-Smit, AJRCCM 2009: 26 S. African subjects; repeated

IGRAs (T.SpotTB, QFT-G IT) 4x over 21D prior to TST (to assess within-patient variability), and then again on days 3,7,28, 84 post-TST (to assess for boosting of IGRA by TST).– Pre-TST tests: 7/26 had spontaneous conversions/reversions (6 for

TSpot, 1 for QFT-G IT). 95% of variability was 3-spot or 80% IFN-gamma response variation on either side of baseline values—could be useful for interpreting conversions/reversions

Effect of TST on IGRA results• QFT-G IT package insert: in U.S. specificity study (**individuals

with no reported TB risk factors), a subset of subjects were retested 4-5 weeks after initial QFT-G IT/TST. Agreement between 2 QFT-G IT tests was 98.5% (out of 530 subjects, 5 went posneg, and 3 went negpos.)

• Van Zyl-Smit, AJRCCM 2009: 26 S. African subjects; after baseline IGRAs (T.SpotTB, QFT-G IT), repeated IGRAs on days 3,7,28, 84 post-TST (to assess for boosting of IGRA by TST).– **Post-TST tests: 8 subjects boosted above defined baseline

variability by day 7, but not day 3. 2 initially IGRA-negative subjects converted to IGRA-positive.

– Conclusion: safe to do QFT-G IT or TSpot within 3 days of performing TST (i.e. on day of TST read).

– Cohort as a whole showed some persistently elevated IFN-gamma responses up to day 84 after TST, though some individuals had returned to pre- TST levels by day 28.(So what are implications for long-term boosting effects, e.g. in those receiving annual testing?)

Effect of TST on IGRA, continued

• Review by van Zyl-Smit et al, PLoSOne 2009: 13 studies– Studies used different TU for TST, different time points for IGRAs

after TST, and varied re: initial TST/IGRA status of individuals– 5 studies concluded boosting of IGRA by TST did NOT occur; in

4/5, earliest timepoint of repeat IGRA was 28 days-9 months after TST. In 5th, IGRA was repeated only on day 3 after TST.

– 7 studies demonstrated TST-induced ‘‘boosting’’ of IGRA responses; in 5/7, repeat IGRA was done within 21 days after TST.

Conclusions: – Boosting more pronounced in IGRA-positive (i.e. sensitized)

individuals, but also occurred in a smaller but not insignificant proportion of IGRA-negative subjects

– Time frame of repeat IGRA is key. TST appeared to affect IGRA responses only after 3 days, and may be issue particularly between days 7-28; boosting effect may apparently persist for up to 3 months and then wane, but evidence for this is weak.

Preliminary (unpublished) data from a 4-site (U.S.) collaborative study of serial IGRAs in

HCW

• Longitudinal study of HCW undergoing routine testing for LTBI; overall low risk for TB acquisition at work– 15% born in high-burden country– 10% s/p BCG– 0.4% HIV, 3% DM, 2% “other

immunocompromise”

• Baseline 2-step TST, QFT-G IT, TSpot.TB– IGRAs done BEFORE placement of 1st TST– Repeat all 3 tests at 6, 12, and 18 months

Slides obtained from Dr. John Bernardo, BMC

Baseline Results in subjects with no prior (+) TST or LTBI treatment

n = 2083

TST QFT* T-SPOT*

Positive 43 (2.1) 76 (3.7) 108 (5.2)

Negative 2040 (97.9) 2007 (96.3) 1907 (91.6)

Borderline 68 (3.3)

* p < 0.0001 compared to the TST (borderline T-Spots categorized as negative)

6 month Follow-up

Conversion Reversion*

TST 6 / 1503 (0.4) 11 / 21 (52.4)

QFT-GIT 56 / 1516 (3.7) 28 / 56 (50)

T-SPOT 52 / 1473 (3.3) 47 / 85 (55.3)

Conversion = (-) baseline (+) 6 month

Reversion = (+) baseline (-) 6month

* Total Baseline Positive = 43 TST, 76 QFT-GIT, 108 T-SPOT

12 month Follow-upConversion Reversion

TST 1 / 362 (0.3) n/a

QFT-GIT 9 / 384 (2.3) 7 / 11 (63.6)

T-SPOT 3 / 356 (0.8) 10 / 16 (62.5)

Conversion = (-) baseline (-) 6 month (+) 12 month

Reversion = (-) baseline (+) 6month (-) 12 month

Some take-home points• IGRAs should not be used alone to exclude the Dx of

active TB– In particular, sensitivity in question for extra-pulmonary TB1

• IGRAs cannot distinguish between active and latent TB• IGRAs may remain positive even after appropriate

treatment of active or latent TB.• Sensitivity for diagnosis of LTBI is impossible to

calculate, given absence of a gold standard for this Dx. Exercise caution when interpreting negative IGRA results in individuals with major risks for TB reactivation.– “A negative result must be considered with the individual’s medical

and historical data relevant to probability of M. tuberculosis infection and potential risk of progression to tuberculosis disease, particularly for individuals with impaired immune function.” (QFT-G IT package insert, 2009)

1. Dewan et al, CID 2007

Some take-home points, cont.

• Specificity of IGRAs is very high, but occasionally you will see a patient with NO apparent TB risk factors and a positive IGRA result….– Check absolute value to see if they are close to cutoff for

positive– would repeat, if negative repeat again as tie-breaker..– Again, consider who should be tested in the first place, and who

shouldn’t• It is still not clear how well IGRAs will perform in serial

testing situations (e.g. HCW) or what the true impact of TSTs on subsequent IGRAs actually is. Can we trust conversions if IGRAs are used for annual testing in relatively low risk settings? Would those conversions be stable if we waited 6 months and retested?

December, 2005: CDC guidelines for use of QFT-G

• “CDC recommends that QFT-G may be used in all circumstances in which the TST is currently used, including contact investigations, evaluation of recent immigrants, and sequential-testing surveillance programs for infection control (e.g., those for health-care workers).”

• left open the possibility that "QFT-G sensitivity for LTBI might be less than that of the TST," while acknowledging that the lack of a confirmatory test would make this difficult to assess

• "each QFT-G result and its interpretation should be considered in conjunction with other epidemiologic, historic, physical, and diagnostic findings."

New CDC recs for use of IGRAs in development…coming in

2010!!??2010!!??• Likely to advocate broad use (including in annual testing), and

use in place of TST, rather than as confirmatory test.• My opinion: if we are going to make clinical decisions based on

IGRA results, then we need to focus on estimating IGRA sensitivity/NPV for LTBI and also potentially revisit the clinical guidelines regarding increased reactivation risk. What will we do with TST+/IGRA- individuals who:– Have various forms of relative immunocompromise, or are

going to become immunocompromised (e.g. by transplant, TNF-alpha blockers)?

– Are recent immigrants from endemic areas?– Have CXR findings consistent with past TB (and which CXR

findings, specifically, matter?)

MACET recommendations on use of IGRAs 6-13-08

• Recent contacts: IGRAs seem to perform well (good sensitivity and correlation with TB exposure); can use IGRA or TST

• Immunocompromised: two groups– Pre-immunocompromisation (awaiting transplant,

going on TNF-alpha blocker or steroids, etc): use both tests, Tx if either positive

– Already immunocompromised (including HIV): same recommendations

MACET recs 2008, cont.

• Recent immigrants: panel unable to reach consensus, as negative test does not appear to rule out LTBI, and some of this population could be recently infected. Clinical f/u after testing is optimal.

• HCW: same caveats as above. Agreed that either IGRA or TST could be used. Those with key reactivation risk factors who are IGRA negative should have clinical f/u if possible.

MACET recs 2008, continued

• Children: limited data, no recommendations (could update, given recent analyses suggesting good performance in kids>5)

• Low-risk individuals: given low pre-test probability, test results difficult to interpret. Best to NOT test with either IGRA or TST.

• Adults with recent BCG: IGRAs can be helpful given high specificity

• Active TB: can use IGRA to rule IN infection (either latent or active), but NOT to rule OUT active disease (given limits to sensitivity)