1 - pai2009

Novel and I mprovedTechnologies forTuberculosi s Diagnosi s:Progress and Chal lengesMadhukar Pai, MD, PhDa,*, Jessica Minion, MDa,Hojoon Sohn, MPHa,b, Alice Zwerling, MSca, Mark D. Perkins, MDbDespite substantial success in implementingstandardizedcareandimprovingratesofcureinrecent years, the global burdenof tuberculosis(TB) remainsenormous. Lackof rapidandaccu-rate diagnosis and case detection are majorobstacles to TB control. TB diagnosis, even today,continuestorelyheavilyontoolssuchasdirectsmear microscopy, solid culture, chest radiog-raphy, and tuberculin skin testing: tools that oftenperform poorly, and require infrastructurefrequently unavailable in the periphery of thehealthsystemwherepatientsfirstseekcare.Thelimitations of the existing diagnostics toolboxhavebeenexposedbythehumanimmunodefi-ciencyvirus(HIV) epidemic1,2andbytheemer-gence of multidrug-resistant TB(MDR-TB) andextensively drug-resistant TB (XDR-TB). Diag-nostic delays and health systemfailures oftenresult inmissedor latediagnoses, withseriousconsequences for TB patients.3In the past few years, there has been an unprec-edentedlevel of interest andactivityfocusedonthedevelopment of newtoolsfor TBdiagnosis,largelybecauseofagenciessuchastheFounda-tionfor InnovativeNewDiagnostics (FIND), theStopTBPartnershipsNewDiagnosticsWorkingGroup(NDWG), the Global Laboratory Initiative(GLI) (another Stop TB Partnership WorkingGroup), the World Health Organization (WHO),and the Special Program for Research andTraininginTropical Diseases(TDR).2,46Fundingagencies such as the Bill & Melinda Gates Founda-tion, the Global Fund to Fight AIDS, TB and Malaria(GFATM),andUNITAIDhaveprovidedthemuch-neededresourcesandimpetustopushthenewtoolsagenda, inkeepingwiththeGlobal PlantoStop TB.7Thisarticlereviewstheexistingevidencebaseof TBdiagnostics, describes newtechnologiesandtheprogressmadeintheirdevelopmentandFinancial and competing interests disclosure: Madhukar Pai, Hojoon Sohn and Mark Perkins are affiliated withtheFoundationforInnovativeNewDiagnostics(FIND),Geneva.FINDisanonprofitagencythatworkswithseveral industrypartners indevelopingandevaluatingnewdiagnostics for neglectedinfectious diseases.Madhukar Pai andMarkPerkins arecoregroupmembers of theStopTBPartnerships NewDiagnosticsWorkingGroup.Theauthorshavenofinancialinvolvementwithanyorganizationorentitywithafinancialinterestinorfinancialconflictwiththesubjectmatterormaterialsdiscussedinthisarticleapartfromthosedisclosed. This work was supported in part by Canadian Institutes for Health Research (CIHR) grantMOP-89,918andMOP-81,362, andEuropeanCommissiongrantTBSusgent(FP7-HEALTH-2007-B). MadhukarPai is supported by a CIHRNew Investigator Award. These funding agencies had no role in thedevelopmentofthisarticle.aDepartment of Epidemiology, Biostatistics & Occupational Health, McGill University, 1020 Pine Avenue West,Montreal, H3A1A2,CanadabFoundation forInnovative NewDiagnostics,AvenuedeBude ,161202 Geneva, Switzerland*Correspondingauthor.E-mailaddress: [email protected](M.Pai).KEYWORDS

Tuberculosis Diagnostics Newtools

SensitivityandspecificityClin Chest Med 30 (2009) 701716doi:10.1016/j.ccm.2009.08.0160272-5231/09/$ see front matter 2009 Elsevier Inc. All rights reserved.chestmed.theclinics.comevaluation,andendswithareviewofcost-effec-tiveness and modeling studies of the potentialimpact of new diagnostics in TB control.THE EVIDENCE BASE OF TB DIAGNOSISAlthoughprimarydiagnostictrialsareperformedto generate data on test accuracy and operationalperformance, systematicreviewsandmeta-anal-yses provide the best synthesis of currentevidence on any given diagnostic test. In thepastfewyears,morethan30systematicreviewsand meta-analyses have been published onvarious TB tests.8These reviews have synthesizedthe results of more than 1000 primary studies andhaveprovideduseful insightsintothediagnosticaccuracy and role of various tests (Table 1).8Theyhavealsoplayedakeyroleinrecentpolicystatements and guidelines on TBdiagnostics.9However, muchof theexistingevidencebaseisfocused on test accuracy (ie, sensitivity and spec-ificity). There are limited data on outcomes such asaccuracy of diagnostic algorithms (rather thansingletests)andtheirrelativecontributionstothehealth care system, incremental value of newtests, effect of newtests on clinical decision-makingandtherapeuticchoices, cost-effective-nessinroutineprogrammaticsettings,andeffecton patient-centeredoutcomes.8Recently,NDWGlaunched a comprehensive Web site resourceEvidence-basedtuberculosis diagnosis, avail-ableat http://www.tbevidence.org/ (Fig. 1). ThisWeb site is the most comprehensive single sourceof evidencesyntheses, policies, guidelines, andresearch agendas on TB diagnosis.IMPROVEDAND NEW TECHNOLOGIES: WHATSIN THE PIPELINE?New diagnostics pipeline for TB is rapidly expand-ing.In2008,theStopTBPartnershipsRetoolingTask Force (RTF) and NDWG produced a detailedbrochure on diagnostic tools in the pipeline, mainlyto provide guidance to National TB Programs(NTPs), andfor fundingandtechnical agenciesthat may wish to support the development, evalu-ation, or implementationof newtools.10Fig. 2shows the pipeline; the tools are stratified asWHO-endorsed, Toolsinlate-stagedevelop-ment/evaluation,orToolsinearlyphasedevel-opment. The figure not only describes thevarioustestsbutalsoprovidessomeinformationonthecommercial kitsavailable,trainingrequire-ments, and estimated costs.10A more exhaustivelist of various TB technologies was publishedby Perkins and Cunningham.1Some of thetechnologies are described in greater detail insubsequent sections.OPTIMIZED SMEAR MICROSCOPYAlthough much work is being done to develop newdiagnostics, in most resource-limited countriesdirect sputumsmear microscopy remains theprimary means for diagnosis of TB. Given theknown limitations of smear microscopy, consider-ableefforthasbeengiventoidentifyingmethodsthat can optimize the yield and accuracy of smearmicroscopy.1114These include light-emittingdiode (LED)-based fluorescence microscopy(FM), use of sputumprocessing methods, andoptimizationofspecimencollectionforsame-daydiagnosis15(ie, front-loadedmicroscopy). Fig. 3provides an overviewof the major commercialLEDtechnologies for microscopy.16Although pub-lisheddataarelimited(reviewedbyMinionandcolleagues16), LEDtechnologies seemto be prom-ising in settings in which FM has not been feasible,and a WHO policy on LED microscopy is expectedinNovember2009. Mobilephone-basedmicros-copy17andautomateddetectionsystemsusingimageprocessing18areother novel approachesthat have been proposed, although the useof these approaches is yet to be adequatelyvalidated.IMPROVEDAND NEWER CULTURE METHODSAutomatedLiquidCulturesAutomatedliquidculturesystemssuchasBacT/ALERTMP(bioMerieuxInc, Durham, NC, USA)and BD BACTEC MGIT (Becton Dickinson,Sparks, MD, USA) arecurrently consideredthegold-standardapproachforisolatingmycobacte-ria. Meta-analyses have shown that liquid systemsaremoresensitivefordetectionof mycobacteriaand may increase the case yield by 10%compared with solid media.19,20They also reducethe delaysin obtainingresultsto daysrather thanweeks.Useofliquidmediafordrugsusceptibilityresults in even greater time savings. However,liquidsystems are prone tocontamination andrequirestringent qualityassurancesystemsandtraining standards. In addition, they are moreexpensive and require equipment investments,though MGIT can also be used as a manualsystem. Traditionally, liquid culture has alwaysbeen used in tandem with solid media to maximizeyield and allowexamination of colony morphology.FIND projects demonstrated the feasibility of usingliquidculture as a stand-alone methodif rapidspeciesconfirmationispossiblethroughtheuseof rapid antigen detection tests for speciation.Pai et al 702There are currently 3 manufacturers of these rapidtests, which detect the TB-specific protein MPT64inalateral flowformat (eg, Capilia-TB, TAUNS,Numazu, Japan).21In 2007, WHOreleased a policystatement on the use of liquid culture systems andonspeciesconfirmationthroughantigendetec-tion.22TheWHOpolicyrecommendsphasedim-plementation of these systems as a part ofa country-specific comprehensive plan for labora-torycapacitystrengthening, andaddresseskeyissues, including biosafety, customer support,staff training, maintenanceof infrastructureandequipment, specimentransport, andreportingofresults.22UnconventionalandNewerCultureMethodsBecause commercial automated liquid culturesare expensive and may require sophisticatedinstrumentation, several researchers haveproposed unconventional and novel culture-based approaches for TB diagnosis and drugresistance testing. These approaches includemicroscopic observation drug-susceptibility test(MODS),23thin-layer agar (TLA),24andthedirectnitratereductaseassay(NRA),25alsoknownastheGriess method.Recent reviews havesumma-rizedtheir characteristicsandpotential role.2628Althoughthesemethods are promisingas theyallowtheuseof inexpensivematerialsandgiveturnaroundtimessimilar toliquidculture, thesetests are not well standardized, and requireextensivetrainingandoptimizationbeforeroutineclinical use. These methods all require routinespecimen processing, the most burdensomecomponent of mycobacterial culture, beforedirect inoculationwithsputum.As for all culture-based methods, qualityassurance is critical to minimize contaminationandtoensurebiosafetystandardsarefollowed.Appropriate quality-control systems are oftenlacking, recommended equipment (such asbiosafetycabinets) may be unavailable,and strictadherencetoinfectioncontrol practicesisinfre-quently enforced in resource-limited settings.Someof thesenovel culture-basedassayshaveattempted to address laboratory safety issuesinherentintheculturingofMycobacteriumtuber-culosis by sealing the inoculated cultures in trans-parent plates or tubes, and relying on visualinspectionof typical colonymorphology(MODSandTLA) orcolorchanges(Griess) toidentifyTBgrowth. Although disposal of the biohazardousmaterial remainsaconcern, minimizingtheneedfordirecthandlingand manipulationofmycobac-terial cultures by laboratory technologists is animportant advantage.MOLECULAR TESTSNucleic acid amplification tests (NAATs) havebeeninusefor manyyears, althoughtheir usehasbeenlargelyrestrictedtohigh-incomecoun-tries. For example, the 2009updatedguidelineonuseof NAATsbytheUSCentersforDiseaseControl andPrevention(CDC) statesthat NAAtestingbeperformedonat least onerespiratoryspecimen from each patient with signs and symp-toms of pulmonary TB for whom a diagnosis of TBisbeingconsideredbut hasnot yet beenestab-lished, andfor whomthetest result wouldaltercasemanagement or TBcontrol activities, suchascontact investigations.29Clearly, thisrecom-mendation is focused on high-income settingsthat have the resources to implement theseguidelines.As demonstrated in several meta-analyses,existingNAATshavehighspecificity,butmodestand variable sensitivity, especially in smear-nega-tiveandextrapulmonaryTB.3033Several newerNAATshavebeendevelopedrecently, including2technologiescodevelopedwithFIND,theloop-mediated isothermal amplification (Eiken ChemicalCoLtd,Tokyo,Japan),asimplifiedmanualNAATdesigned for peripheral laboratory facilities,34andthe Xpert MTB/RIFassay (Cepheid, Sunnyvale,CA, USA), afullyautomatedNAATplatformthatcandetectTBandrifampinresistance.35Bothofthese tests are formatted for use outside referencecenters, toreplaceorsupplementmicroscopyathealthcentersanddistrict hospitals. Thesetestshave shown great promise in early studies,althoughpublishedevidenceisstill limited. FINDiscurrentlyevaluatingthesetestsinhigh-burdencountries.TheXpertMTB/RIFassayhasrecentlybeen CE marked with package insert data showinggreater than 95% detection of all TB patients.Lineprobeassays(LPAs) haverecentlybeenintroduced in many countries for molecular detec-tion of drug resistance from smear-positive speci-mens. Twocommercial LPAs areavailable: theINNO-LiPA Rif.TB (Innogenetics NV, Gent,Belgium) andGenoTypeMTBDRplus(HainLife-science GmbH, Nehren, Germany). Meta-analyseshave shown that LPAs are highly accurate, and theGenoTypeassay, inparticular, performswell forrapiddetectionof rifampinresistanceinsmear-positivesputum specimens.36,37In 2009, a newerassay(GenoTypeMTBDRsl assay)becameavail-able.38This assay allows the simultaneous detec-tion of the M tuberculosis complex and resistanceto fluoroquinolones or aminoglycosides/cyclicpeptides or ethambutol from smear-positivepulmonary specimens or cultureisolates. Thus,thecombineduseofGenoTypeMTBDRplusandNew Technologies for Tuberculosis Diagnosis 703Table1Summary of findings fromseveral systematic reviews onTB diagnostic testsDiagnostic Test Disease/SiteMajor Findings/Results of SystematicReviewsDiagnosis of active TBSputum smearmicroscopyPulmonary TB-FM is on average 10% more sensitivethan conventional microscopy. Spec-ificity of FM and conventionalmicroscopy is similar. FM is associatedwith improved time efficiency-Centrifugation and overnight sedi-mentation preceded by any ofseveral chemical methods (includingbleach) are more sensitive than directmicroscopy; specificity is unaffectedby sputum-processing methods-When serial sputum specimens areexamined, the mean incrementalyield or increase in sensitivity fromexamination of third sputum spec-imen ranges between 2% and 5%NAATs Pulmonary andextrapulmonary TBNAATs have high specificity and positivepredictive value. However, they havelower (and highly variable) sensitivityand negative predictive value for allforms of TB, especially in smear-negative and extrapulmonarydisease. In-house (home brew)NAATs produce highly inconsistentresults compared with commercial,standardized NAATsCommercial serologicantibody detectiontestsPulmonary andextrapulmonary TBSerologic tests for pulmonary andextrapulmonary TB produceinconsistent estimates of sensitivityand specificity; none of the assaysperforms well enough to replacemicroscopyADA TB pleuritis, pericarditis,peritonitisMeasurement of ADA levels in pleural,pericardial, and ascitic fluid has highsensitivity and specificity forextrapulmonary TBIFN-g TB pleuritis Pleural fluid IFN-g determination isa sensitive and specific test for thediagnosis of TB pleuritisPhage amplificationassaysPulmonary TB Phage-based assays have highspecificity but lower and variablesensitivity. Their performancecharacteristics are similar to sputummicroscopyAutomated liquidculturesPulmonary TB Automated liquid cultures are moresensitive than solid cultures; time todetection is more rapid than solidcultures(continued on next page)Pai et al 704Table1(continued)Diagnostic Test Disease/SiteMajor Findings/Results of SystematicReviewsDiagnosis of latent TBTST Latent TB infection-Individuals who have receivedBCG vaccination are more likely tohave a positive TST; the effect ofBCG on TST results is less after 15years; positive TST with indura-tions of greater than 15 mm aremore likely to be the result of TBinfection than of BCG vaccination-The effect on TST of BCG receivedin infancy is minimal, especially 10years after vaccination. BCGreceived after infancy producesmore frequent, more persistent,and larger TST reactions. NTMinfection is not a clinically impor-tant cause of false-positive TST,except in populations with a highprevalence of NTM sensitizationand a low prevalence of TBinfectionT-cellbased IGRAs Latent TB infection IGRAs have excellent specificity(higher than the TST), and areunaffected by prior BCGvaccinationDiagnosis of drug resistancePhage amplificationassaysRapid detection ofrifampicin resistanceWhen used on culture isolates,phage assays have high sensitivity,but variable and lower specificity.In contrast, evidence is lackingabout the accuracy of these assayswhen they are directly applied tosputum specimensLPAs: INNO-LiPA Rif.TB(LiPA)and GenoTypeMTBDR assaysRapid detection ofrifampicin resistanceLiPA is a highly sensitive and specifictest for the detection of rifampicinresistance in culture isolates. Thetest has lower sensitivity whenused directly on clinical specimens.The GenoType MTBDR assays haveexcellent sensitivity and specificityfor rifampicin resistance evenwhen directly used on clinicalspecimensColorimetricredox-indicatormethods and NRAsRapid detectionof rifampicinand isoniazidresistanceColorimetric methods and NRAs arehighly sensitive and specific forthe rapid detection of rifampicinand isoniazid resistance in cultureisolatesAbbreviations: ADA, adenosine deaminase; NTM, nontuberculous mycobacterial.Adapted from Pai M, Ramsay A, OBrien R. Evidence-based tuberculosis diagnosis. PLoS Med 2008;5(7):e156.New Technologies for Tuberculosis Diagnosis 705GenoType MTBDRsl allows the rapid detection ofXDR-TB. LPAs currently require routine specimenprocessing, DNA extraction, and conventionalpolymerase chain reaction analysis in a multiroomfacility, andarethuslimitedtouseinreferencelaboratories.In2008, WHOendorsedtheuseof LPAsforrapiddetectionofMDR-TBatthecountrylevel.39In 2009, UNITAID approved funding for a programcalled EXPAND-TB that will supply MDR-TBdiagnostics to high-burden countries.40Withanewgrant of US$61,482,085, theproject, ledFig.1. Home page of the Web site Evidence-basedTB Diagnosis, http://www.tbevidence.org. (Courtesy of the StopTB Partnerships New Diagnostics Working Group; with permission. Available at: http://www.tbevidence.org.)Fig. 2. Summary of new technologies by the RTF and NDWG. (From World Health Organization & Stop TB Partner-ship.Newlaboratorydiagnostictoolsfortuberculosiscontrol.Geneva:WorldHealthOrganization;2008;withpermission.)Pai et al 706bytheGLI inclosecollaborationwithFINDandtheGlobal DrugFacility, will expandtheuseofLPAs for rapid MDR-TB diagnosis.40A keycomponent of thisinitiativewill bethestrength-ening of laboratories in countries where LPAswill beintroducedinaphasedmanner, throughcollaboration between various partners.Strengthening of laboratory capacity is criticalfor thesuccessof thisprogram, andindeed, forthe successful implementation of any newTBtechnology.IMMUNE-BASED TESTSSerologic,AntibodyDetectionTestsSystematic reviews have reported strong evidencethat existing commercial serologic tests are of littleclinical value because of suboptimal accuracy andhighinconsistentresults.41,42Thiswasreaffirmedinarecentstudyof19commercial testsbyTDR/WHO, whichshowedsuboptimal performanceofall the rapid tests evaluated.43A more recentsystematic review examined the accuracy ofvariousin-house, purifiedantigensfor serodiag-nosis.44Althoughnoantigenachievedsufficientsensitivitytoreplacesputumsmear microscopy,this review helped identify several promisingpotential candidate antigens for an antibodydetection test for pulmonary TBin patients in-fectedanduninfectedwithHIV.Thiscomprehen-sive reviewalso showed that combinations ofselect antigensprovidedhighersensitivitiesthansingleantigens.44Several industryandacademicgroups are currently working on developingimproved serodiagnostic tests, especially forpoint-of-care (POC) use.AntigenDetectionTestsAntigendetectionhasthepotential toovercomesomeof thewell-recognizedproblemswithanti-bodydetectionassays, especiallyinpopulationsinfected with HIV. Although several antigendetectionassayshavebeenevaluated,detectionof urinary lipoarabinomannan (LAM) (a heat-stablelipoglycaninthemycobacterial cell wall)was considered a particularly good candidate,basedonearlystudies, especiallyinindividualsinfected with HIV.45Early proof-of-principledata and the attractiveness of a simple urine-based TB test led to rapid commercializationof thistest, initiallybyChemogenInc(Portland,ME, USA), and subsequently by InvernessMedical Innovations(Waltham, MA, USA), whichmarketed the test as Clearview TB enzyme-linked immunosorbent assay (ELISA). Subse-quent fieldstudiesinhigh-burdensettingshaveshown LAMperformance to be variable andsuboptimal, with lower sensitivity than ex-pected.46,47However, some emerging datasuggest that LAMmay performbetter in HIV-positive individuals with advancedimmunosup-pression.48Work is ongoing to improve andoptimize the performance of LAM detectionassays.Fig. 3. CommercialLEDproductscurrently availableforTBdiagnostics. (Adapted fromMinionJ,Sohn H,PaiM.Light emitting diode technologies for TB diagnosis: whats on the market? Expert Rev Med Devices2009;6(4):34145; with permission. Images have been reproduced with permission fromthe respective companies.)New Technologies for Tuberculosis Diagnosis 707Interferon-gReleaseAssaysUntilrecently,thediagnosisoflatenttuberculosisinfectiondependedsolelyonthetuberculinskintest (TST), a test with several limitations.49A majoradvance in recent times has been the developmentof T-cell-based interferon-g release assays(IGRAs). IGRAs are in vitro tests that are based oninterferon-g (IFN-g) release after T-cell stimulationby antigens (such as early secretedantigenic target6[ESAT6] andculturefiltrateprotein10[CFP10])thataremorespecifictoMtuberculosisthanthepurifiedproteinderivative(PPD). TwoIGRAsarecurrently available as commercial kits that areapproved by the US Food and Drug Administration(FDA) and CE marked for use in Europe: the Quan-tiFERON-TBGoldIn-Tube(QFT) assay(CellestisLtd., Carnegie, Australia), and the T-SPOT.TBassay (Oxford Immunotec, Abingdon, UK).Systematic reviews have reported strongevidencethatIGRAshavehighspecificitythatisunaffected by bacille Calmette-Gue rin (BCG)vaccination.50,51TST,incontrast,hashighspeci-ficity in populations who have not been vaccinatedwithBCGbutspecificityismodestandinconsis-tentinpopulationsvaccinatedwithBCG. Inlow-incidence settings, IGRA results correlate wellwithsurrogatesof TBexposure. Thehighspeci-ficity of IGRAs is proving to be useful in individualsvaccinated with BCG, particularly in countrieswhereTSTspecificity is compromisedby BCGvaccination after infancy or by multiple BCGvacci-nations.49A World Atlas of BCGPolicies andPractices (Fig. 4) has been compiled to help clini-ciansandpublichealthpractitionersbetterinter-pretTSTanddecideonpopulationsinwhichthemore-specific IGRAs may be more appropriatethan the TST.52For example, some countriesrecommendboosterBCGshotspostinfancyandinto adolescence, which can compromise thevalueof TST. IGRAsmaybeexcellentoptionsinthese populations. The Atlas provides informationon current and past policies on vaccination.Sensitivityof IGRAsandTSTisnot consistentacrosstestsandpopulations,butIGRAsseemtobe at least as sensitive as the TST(estimatedwithactiveTBasthesurrogatereferencestan-dard).51However, as pointed out by several inves-tigators,53,54thediagnosisof activeTBdependson microbiological detection of Mtuberculosis.Immune-basedtests, such as IGRAs andTST,do not directly detect M tuberculosis; they merelyindicateacellularimmuneresponsetorecent orremote sensitization with M tuberculosis. BecauseIGRAs cannot distinguish between latent andactiveTB,apositiveIGRAresultmaynotneces-sarilyindicateactiveTB. AnegativeIGRAresultwouldnot conclusivelyruleout activediseaseinanindividualsuspectedtohaveTB(similartotheresults of a TST).The use of IGRAs is steadily increasing in coun-tries with low or intermediate incidence. More thana dozen countries now have at least 1 guideline orstatement ontheuseof IGRAs.55TheseincludetheUnitedStates, Canada, theUnitedKingdom,Fig. 4. WorldAtlas of BCGPolicies andPractices, http://www.bcgatlas.org. (Courtesyof AliceZwerling, MSc,Montreal, Canada; withpermission.)Pai et al 708Japan, France, Spain, Italy, Germany, Switzerland,Australia, theNetherlands, Denmark, theCzechRepublic, the Slovak Republic, Korea, andNorway. Intheseguidelines, 3mainapproacheshavebeenrecommendedfor theuseof IGRAs:(1) TSTshouldbereplacedby IGRA; (2) eitherTSTor IGRAmaybeused; (3) 2-stepapproachwith TST first, followedby IGRA. Although thebroad approach may follow 1 of these recommen-dations, some guidelines recommend more than 1approach, depending on the risk group tested. Forexample, subgroups such as children andimmunocompromised patients often receivedifferent recommendations fromother groups.Table 2 shows the approaches recommendedfor useof IGRAsinseveral low-incidencecoun-tries.55Asseeninthetable,thereisconsiderablediversity of howvarious countries currently recom-mend and use IGRAs. The 2-step approach seemstobethemost dominant strategyandthismaypartly be because of cost considerations.Despite the large number of publications onIGRAs, evidenceisstill limitedontheprognosticvalueofthesetests,andtheiraddedvalueinTBdiagnosis and control.51,56There is growingevidencethat theperformanceof IGRAs variesbetween countries with high and low incidence ofTB.57Their role, if any, seems to be limited in low-income countries with a high TB burden, althoughseveral fieldevaluationsareongoing, supportedby FIND and other agencies.57ImprovedSkinTestsAwell-recognizedlimitationof theconventionalTSTisthelackofspecificityofthePPD,acrudemixturewithalargenumberof potentiallycross-reactingantigens. Investigators workingonthisproblem have attempted to replace PPD with anti-gens (such as ESAT6) that are specific to M tuber-culosis. Small-scale, phase 1 trials of thisimproved skin test have shown promise, butfurther validation is needed.58,59Despite thelimitedevidenceonthesereagents, 1company(Masterpharm, Russia) is already marketingacommercial product calledDiaskintest (basedonESAT6/CFP10).60Itremainstobeseenif thisimprovedskintestreagentcansafelyreplacetheconventional PPD.POC TECHNOLOGIESThe ideal TB diagnostic test is a simple, low-tech-nology, POCtest that canberapidlyperformedand yield accurate results. In 2009, a groupincluding representatives fromMe decins SansFrontie` res, Treatment ActionGroup, PartnersinHealth, andother agencies, developedminimumtechnical test specificationsthat must drivethedevelopment of any newPOCTBtest (Table3).61No existingtestmeets allofthese specifica-tions, althoughtheXpert MTB/RIFassaymeetsmost of them. However, because of growinginterest in newtools and biomarkers, and theincreased availability of funding and grants,several agencies and groups are working ondevelopingPOCtestsforTB,includingimprovedserologic assays, detection of volatile organiccompounds in breath, hand-held moleculardevices, microchip technologies, and tests basedon platforms such as proteomics andmetabolomics.Recently, the X PRIZE Foundation receivedaplanninggrant fromtheBill &MelindaGatesFoundationtodevelopanXPRIZEfor effectivediagnosis of TB in the developing world.62Itremainstobeseenifsuchprize-basedcompeti-tionsfosterinnovationsthatdeliverthePOCtestthat will revolutionizeTBdiagnosis. Asignificantlimitationontheeffect of aPOCtest for TBisthat TB is a notifiable disease that requires patientTable 2Recommendations fromvarious countries that have guidelines on the use of IGRAsaGeneral Testing Approach CountriesTST should be replaced byIGRA (ie, only IGRA is used)Germany (anti-TNF-a therapy), Switzerland (anti-TNF-a therapy),Denmark (anti-TNF-a therapy, BCG-vaccinated contacts/adults)Either TST or IGRA maybe usedUnited States, France, Australia (refugees), Japan (QFT preferredin all groups except in children