RESEARCH ARTICLE Identification of Recent Tuberculosis Exposure Using QuantiFERON-TB Gold Plus, a Multicenter Study Sandra Pérez-Recio,a Natàlia Pallarès,b Maria D. Grijota-Camino,a Adrián Sánchez-Montalvá,c Laura Barcia,d Silvia Campos-Gutiérrez,e Virginia Pomar,f Ramón Rabuñal-Rey,g María Elvira Balcells,h Deniz Gazel,i Natalia Montiel,j* Diego Vicente,k Ivana Goic-Barišic,l,m Thomas Schön,n,o Jakob Paues,o Ivana Marekovic,p Juana Cacho-Calvo,q Aleksandra Barac,r,s Delia Goletti,t Mercedes García-Gasalla,u§ José María Barcala,v María Teresa Tórtola,w,x Luis Anibarro,d Isabel Suárez-Toste,y Esther Moga,z María J. Gude-Gonzalez,aa Rodrigo Naves,bb Tekin Karslıgil,i Tania Martin-Peñaranda,k Goran Stevanovic,r,s Matilde Trigo,cc Verónica Rubio,f^ _Ilkay Karaog  lan,dd Nazan Bayram,ee Fernando Alcaide,ff,gg Cristian Tebé,b Miguel Santin,a,hh on behalf of the ESCMID Study Group for Mycobacterial Infections (ESGMYC) and the Mycobacteria Study Group (GEIM) of the Spanish Society of Clinical Microbiology and Infectious Diseases (SEIMC) a Tuberculosis Unit, Department of Infectious Diseases, Bellvitge University Hospital-Bellvitge Biomedical Research Institute (IDIBELL), L’Hospitalet de Llobregat, Barcelona, Spain b Biostatistics Unit (UBiDi), Bellvitge Biomedical Research Institute (IDIBELL), L’Hospitalet de Llobregat, Barcelona, Spain c Department of Infectious Diseases, PROSICS, Vall d'Hebrón University Hospital, Barcelona, Spain Citation Pérez-Recio S, Pallarès N, Grijota- d Tuberculosis Unit, Department of Internal Medicine, Complexo Hospitalario Universitario de Pontevedra, Camino MD, Sánchez-Montalvá A, Barcia L, Campos-Gutiérrez S, Pomar V, Rabuñal-Rey R, Pontevedra, Spain Balcells ME, Gazel D, Montiel N, Vicente D, e Microbiology and Infection Control Service, University Hospital of Canary Islands, Tenerife, Canary Islands, Spain Goic-Barišic I, Schön T, Paues J, Marekovic I, f Infectious Diseases Unit, Department of Internal Medicine, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain Cacho-Calvo J, Barac A, Goletti D, García- Gasalla M, Barcala JM, Tórtola MT, Anibarro L, g Infectious Diseases Unit, Department of Internal Medicine, Hospital Lucus Augusti, Lugo, Spain Suárez-Toste I, Moga E, Gude-Gonzalez MJ, h Department of Infectious Diseases, School of Medicine, Pontificia Universidad Católica de Chile, Santiago, Chile Naves R, Karslıgil T, Martin-Peñaranda T, i Department of Medical Microbiology, Faculty of Medicine, Gaziantep University, Gaziantep, Turkey Stevanovic G, Trigo M, Rubio V, Karao glan _I, Bayram N, Alcaide F, Tebé C, Santin M, on j Department of Microbiology, Hospital Costa del Sol, Marbella, Málaga, Spain behalf of the ESCMID Study Group for k Department of Microbiology, Hospital Universitario Donostia, San Sebastian, Spain Mycobacterial Infections (ESGMYC) and the l Department of Clinical Microbiology, University Hospital of Split, Split, Croatia Mycobacteria Study Group (GEIM) of the Spanish Society of Clinical Microbiology and m School of Medicine, University of Split, Split, Croatia Infectious Diseases (SEIMC). 2021. Identification n Department of Infectious Diseases, Kalmar County Hospital, Kalmar, Linköping University, Sweden of recent tuberculosis exposure using o Division of Infectious Diseases, Department of Biomedical and Clinical Sciences, Linköping University, Linköping, Sweden QuantiFERON-TB gold plus, a multicenter study. Microbiol Spectr 9:e00972-21. https://doi p Department for Clinical and Molecular Microbiology, University Hospital Centre Zagreb, Zagreb, Croatia .org/10.1128/Spectrum.00972-21. q Department of Microbiology, Getafe University Hospital, Getafe, Madrid, Spain Editor Kileen L. Shier, Quest Diagnostics r Clinic for Infectious and Tropical Diseases, Clinical Center of Serbia, Belgrade, Serbia Nichols Institute s Faculty of Medicine, University of Belgrade, Belgrade, Serbia Copyright © 2021 Pérez-Recio et al. This is an t Istituto Nazionale per le Malattie Infettive Lazzaro Spallanzani-IRCCS, Rome, Italy open-access article distributed under the terms of the Creative Commons Attribution 4.0 u Infectious Diseases Unit, Department of Internal Medicine, Hospital Universitari Son Llàtzer-IdISBa, Palma de International license. Mallorca, Illes Balears, Spain Address correspondence to Miguel Santin, v Infectious Diseases Unit, Hospital Universitario de Jerez, Jerez de la Frontera, Cádiz, Spain msantin@bellvitgehospital.cat, or w Department of Microbiology, Vall d’Hebrón University Hospital, Barcelona, Spain msantin@ub.edu. x Department of Genetics and Microbiology, Universitat Autònoma de Barcelona, Barcelona, Spain *Present address: Natalia Montiel, Department of Microbiology, Hospital Universitario Puerta y Pneumology Service, University Hospital of Canary Islands, Tenerife, Canary Islands, Spain del Mar, Cádiz, Spain. z Department of Immunology, Hospital de la Santa Creu i Sant Pau, Biomedical Research Institute Sant Pau § Present address: Mercedes García-Gasalla, (IIB Sant Pau), Barcelona, Spain Infectious Diseases Unit, Department of aa Department of Microbiology, Hospital Lucus Augusti, Lugo, Spain Internal Medicine, Hospital Universitari Son Espases-IdISBa, Palma de Mallorca, Illes Balears bb Program of Immunology, Institute of Biomedical Sciences, Faculty of Medicine, Universidad de Chile, Santiago, Chile Spain. cc Microbiology Department, Complexo Hospitalario Universitario de Pontevedra, Pontevedra, Spain ^ Present address: Verónica Rubio, Department dd Department of Infectious Diseases and Clinical Microbiology, Faculty of Medicine, Gaziantep University, of Internal Medicine, Hospital Municipal de Gaziantep, Turkey Badalona, Barcelona, Spain. ee Department of Pulmonary Diseases, Faculty of Medicine, Gaziantep University, Gaziantep, Turkey Received 28 August 2021 Accepted 5 October 2021 ff Department of Microbiology, Bellvitge University Hospital—Bellvitge Biomedical Research Institute (IDIBELL), Published 10 November 2021 L’Hospitalet de Llobregat, Barcelona, Spain Volume 9 Issue 3 e00972-21 MicrobiolSpectrum.asm.org 1 Pérez-Recio et al. gg Department of Pathology and Experimental Therapy, University of Barcelona, L’Hospitalet de Llobregat, Barcelona, Spain hh Department of Clinical Sciences, University of Barcelona, L’Hospitalet de Llobregat, Barcelona, Spain ABSTRACT We investigated whether the difference of antigen tube 2 (TB2) minus anti- gen tube 1 (TB1) (TB22TB1) of the QuantiFERON-TB gold plus test, which has been postu- lated as a surrogate for the CD81 T-cell response, could be useful in identifying recent tu- berculosis (TB) exposure. We looked at the interferon gamma (IFN-g) responses and differences in TB2 and TB1 tubes for 686 adults with QFT-plus positive test results. These results were compared among groups with high (368 TB contacts), low (229 patients with immune-mediated inflammatory diseases [IMID]), and indeterminate (89 asylum seekers or people from abroad [ASPFA]) risks of recent TB exposure. A TB22TB1 value .0.6 IUml21 was deemed to indicate a true difference between tubes. In the whole cohort, 13.6%, 10.9%, and 11.2% of cases had a TB2.TB1 result in the contact, IMID, and ASPFA groups, respectively (P = 0.591). The adjusted odds ratios (aORs) for an association between a TB22TB1 result of .0.6 IUml21 and risk of recent exposure versus contacts were 0.71 (95% confidence interval [CI], 0.31 to 1.61) for the IMID group and 0.86 (95% CI, 0.49 to 1.52) for the ASPFA group. In TB contact subgroups, 11.4%, 15.4%, and 17.7% with close, frequent, and sporadic contact had a TB2.TB1 result (P = 0.362). The aORs versus the close subgroup were 1.29 (95% CI, 0.63 to 2.62) for the frequent subgroup and 1.55 (95% CI, 0.67 to 3.60) for the sporadic subgroup. A TB22TB1 difference of .0.6 IUml21 was not associated with increased risk of recent TB exposure, which puts into question the clinical potential as a proxy marker for recently acquired TB infection. IMPORTANCE Contact tuberculosis tracing is essential to identify recently infected people, who therefore merit preventive treatment. However, there are no diagnostic tests that can determine whether the infection is a result of a recent exposure or not. It has been suggested that by using the QuantiFERON-TB gold plus, an interferon gamma (IFN-g) release assay, a difference in IFN-g production between the two antigen tubes (TB2 minus TB1) of .0.6 IUml21 could serve as a proxy marker for recent infection. In this large mul- tinational study, infected individuals could not be classified according to the risk of recent exposure based on differences in IFN-g in TB1 and TB2 tubes that were higher than 0.6 IUml21. QuantiFERON-TB gold plus is not able to distinguish between recent and remotely acquired tuberculosis infection, and it should not be used for that purpose in contact tuberculosis tracing. KEYWORDS QuantiFERON-TB gold plus, diagnosis, latent tuberculosis infection, tuberculosis-specific CD8 T cells C ontact tracing is central to tuberculosis (TB) control and prevention by helping to iden- tify and treat people who have been recently infected (1). The tuberculin skin test (TST) and, more recently, interferon gamma (IFN-g) release assays (IGRAs) have been used to detect TB infection among contacts of TB cases. However, both tests have a poor ability to predict progression from latent to active TB, with the IGRAs being, at best, only slightly better than the TST (2–5). Presently, in contact tracing, except for converters, there is no method to con- firm whether a positive IGRA result is due to recent exposure or exposure that occurred sev- eral years earlier. Consequently, all positive results are considered recent infections and all of them are given preventive therapy to avoid leaving high-risk subjects untreated. In recent years, the importance of CD81 T-cells in the immune response against Mycobacterium tuberculosis infection has been recognized (6, 7). It has been shown that the CD81 T-cell response is higher in recent contacts of TB patients than in other groups (8). In this regard, the QuantiFERON-TB gold plus (QFT-plus), which includes an additional antigen tube (TB2) that can elicit IFN-g production by both CD41 and CD81 T-cell responses, pro- vides a surrogate marker of the CD81 T-cell response (9). In two previous studies (10, 11), an antigen tube 2 minus antigen tube 1 (TB22TB1) value .0.6 IUml21 in contacts was associated with sleeping in the same room as index cases and with European origin (10), and a higher quantitative TB2 IFN-g response was found in contacts living in the same Volume 9 Issue 3 e00972-21 MicrobiolSpectrum.asm.org 2 Identifying Recent Tuberculosis Exposure TABLE 1 Main characteristics of 686 individuals with positive QuantiFERON-TB gold plus results No. (%) or other value as indicated forb: Characteristica Contacts (n = 368) IMID patients (n = 229) ASPFA (n = 89) P value Gender Male 182 (49.5) 131 (57.2) 63 (70.8) 0.001 Female 186 (50.5) 98 (42.8) 26 (29.2) Age (yr) Mean (6SD) 45.4 (616.9) 55.8 (613.6) 28.5 (611.9) ,0.001 Median (IQR) 44.6 (32.1–55.5) 58.0 (44.4–66.7) 24.5 (20.1–33.1) ,0.001 BCG vaccination status Yes 195 (53) 43 (18.8) 12 (13.5) ,0.001 No 90 (24.4) 28 (12.2) 17 (19.1) Unknown 83 (22.6) 158 (69.0) 60 (67.4) Incidence of TB in country of birth $25  105 154 (41.8) 17 (7.4) 83 (93.3) ,0.001 ,25  105 214 (58.2) 212 (92.6) 6 (6.7) Type of contact Close 202 (54.9) Frequent 104 (28.3) Sporadic 62 (16.8) QFT-plus positive at baseline Yes 353 (95.9) No 15 (4.1) Contact with a smear-positive case Yes 285 (77.4) No 75 (20.4) Unknown 8 (2.2) IMID #10 mg PDN or equivalent within 8 wk Yes 53 (23.1) No 176 (76.9) DMARDs within 8 wk Yes 98 (42.8) No 131 (57.2) aBCG, bacillus Calmette-Guérin; DMARDs, disease-modifying antirheumatic drugs; IQR, interquartile range; PDN, prednisone; QFT-plus, QuantiFERON-TB gold plus; TB, tuberculosis. bASPFA, asylum seekers and people from abroad; IMID, immune-mediated inflammatory diseases. room as index cases (11), data highly suggestive of newly acquired infection. However, these studies included small numbers of individuals with TB22TB1 values .0.6 IUml21, and they did not compare contacts with groups without risk of recent exposure. In the present study, we investigated whether the QFT-plus could be used to iden- tify individuals recently exposed to TB and thereby act as a surrogate marker for newly acquired infection. This was done by identifying TB2, TB1, and TB22TB1 IFN-g levels in individuals with different risks of recent TB exposure. The rationale was that if an asso- ciation exists between differences in IFN-g production in TB2 and TB1 and time elapsed since primary infection, the QFT-plus should be able to stratify the groups by their background exposure. Specifically, we anticipated a gradient in the proportions of indi- viduals with TB22TB1 IFN-g values .0.6 IUml21 across the groups (with the highest rates in contacts and lower rates in individuals without risk of recent TB exposure). RESULTS In total, we included 686 adults with positive QFT-plus results, with 368 in the contact group, 229 in the patients with immune-mediated inflammatory diseases (IMID) group, and 89 in the asylum seekers or people from abroad (ASPFA) group. Their main characteristics are summarized in Table 1. Compared with the contact and IMID groups, the ASPFA group pre- dominantly included men (49.5%, 57.2%, and 70.8%, respectively; P = 0.001) and its members Volume 9 Issue 3 e00972-21 MicrobiolSpectrum.asm.org 3 Pérez-Recio et al. FIG 1 TB12Nil and TB22Nil (A) and TB22TB1 (B) values for IFN-g responses in the QuantiFERON-TB gold plus by group. Responses are shown using dot and box plots with whiskers, where the horizontal lines indicate the median values and the first and third quartiles. Within-group comparison (TB12Nil and TB22Nil): contacts, P = 0.083; IMID patients, P = 0.220; and ASPFA, P = 0.214. Between-group comparison: TB12Nil, P = 0.084; TB22Nil, P = 0.095; and TB22TB1, P = 0.935. ASPFA, asylum seekers and people from abroad; IMID, immune- mediated inflammatory disease; TB1, antigen tube 1 of the QFT-plus; TB2, antigen tube 2 of the QFT-plus. were younger (median ages 44.6, 58.0, and 24.5 years, respectively; P , 0.001). Of note, 58.2% of contacts were from countries with ,25 TB cases per 105 population. TB1 and TB2 interferon-c responses. Respectively, the median IFN-g concentrations for TB1 minus Nil (TB12Nil) and TB2 minus Nil (TB22Nil) were 2.74 and 2.58 IUml21 in the contact group (P = 0.083), 1.76 and 1.83 IUml21 in the IMID group (P = 0.220), and 2.59 and 2.78 IUml21 in the ASPFA group (P = 0.214). There were no significant differences for TB12Nil (P = 0.084) or TB22Nil (P = 0.095) between groups (Fig. 1 and Table S1 in the supplemental material). TB1 and TB2 interferon-cresponse among contacts. When we restricted the anal- ysis to the 368 contacts, the respective median IFN-g concentrations for TB12Nil and TB22Nil were 2.68 and 2.36 IUml21 for close contacts (P = 0.785), 3.03 and 3.47 IUml21 for frequent contacts (P = 0.016), and 2.54 and 2.58 IUml21 for sporadic contacts (P = 0.118). No significant differences were observed for TB12Nil (P = 0.646) or TB22Nil (P = 0.371) when comparing these three subgroups (Fig. 2 and Table S2). Next, we analyzed the 360 contacts for whom the sputum smear microscopy result of the index case was known. The median IFN-g concentrations for TB12Nil and Volume 9 Issue 3 e00972-21 MicrobiolSpectrum.asm.org 4 Identifying Recent Tuberculosis Exposure FIG 2 TB12Nil and TB22Nil (A) and TB22TB1 (B) values for IFN-g responses in the QuantiFERON-TB gold plus by extent of exposure. Responses are shown using dot and box plots with whiskers, where the horizontal lines indicate the median values and the first and third quartiles. Within-group comparison (TB12Nil and TB22Nil): close contacts, P = 0.785; frequent contacts, P = 0.016; and sporadic contacts, P = 0.118. Between-group comparison: TB12Nil, P = 0.646; TB22Nil, P = 0.371; and TB22TB1, P = 0.053. TB1, antigen tube 1 of the QFT-plus; TB2, antigen tube 2 of the QFT-plus. TB22Nil were 2.73 and 2.47 IUml21 among smear-positive contacts (P = 0.031), com- pared with 2.75 and 2.65 IUml21 among smear-negative contacts (P = 0.906). There were no significant differences for TB12Nil (P = 0.655) or TB22Nil (P = 0.730) when comparing these two groups (Fig. 3). Given that the conversion of QFT-plus from a negative to a positive result in contact tracing denotes newly acquired infection, we also looked at the IFN-g concentrations among the 15 contacts that showed conversion. The median IFN-g concentrations were 1.50 IUml21 for TB12Nil and 1.12 IUml21 for TB22Nil. Among these, 11 contacts showed conversion in both tubes (8 presented conversion outside the borderline range), 2 showed conversion in only TB1, and 2 showed conversion in only TB2 (all four within the borderline range). Although 8 of the 15 had higher IFN-g concentrations in the TB2 tube than in the TB1 tube at the second test, the TB22TB1 was .0.6 IUml21 in only one case (Table S3). TB22TB1 value >0.6 IUml21 as a predictor of greater risk of recent exposure. The proportion of cases with IFN-g concentrations higher in TB2 than in TB1 (TB2.TB1) was 44.6% (43.5% in the contact group, 46.3% in the IMID group, and 44.9% in the Volume 9 Issue 3 e00972-21 MicrobiolSpectrum.asm.org 5 Pérez-Recio et al. FIG 3 TB12Nil and TB22Nil (A) and TB22TB1 (B) values for IFN-g responses in the QuantiFERON-TB gold plus by sputum smear microscopy result of the index case. Responses are shown using dot and box plots with whiskers, where the horizontal lines indicate the median values and the first and third quartiles. Within-group comparison (TB12Nil and TB22Nil): P = 0.031 for smear positive and P = 0.906 for smear negative. Between-group comparison: P = 0.655 for TB12Nil, P = 0.730 for TB22Nil, and P = 0.457 for TB22TB1. TB1, antigen tube 1 of the QFT-plus; TB2, antigen tube 2 of the QFT-plus. ASPFA group; P = 0.796). The respective proportions among the 368 contacts were 39.1% of close contacts, 50.0% of frequent contacts, and 46.8% of sporadic contacts (P = 0.162). When using the .0.6 IUml21 criterion (TB22TB1 . 0.6 IUml21), these proportions decreased (13.6% in the contact group, 10.9% in the IMID group, and 11.2% in the ASPFA group; P = 0.591). The proportions also decreased when we analyzed the con- tact subgroups (11.4% among close contacts, 15.4% among frequent contacts, and 17.7% among sporadic contacts; P = 0.362). A TB22TB1 value .0.6 IUml21 was not associated with a greater risk of recent TB exposure. Versus the contact group, the odds ratios (ORs) were 0.78 (95% confidence interval [CI], 0.47 to 1.30) for the IMID group and 0.81 (95% CI, 0.39 to 1.66) for the ASPFA group in the univariate regression analysis. This lack of association remained af- ter adjusting for covariates (i.e., age, gender, incidence of TB in the country of birth [,25  105 or $25  105], and European origin) (Table 2). Similarly, a TB22TB1 value .0.6 IUml21 was not associated with a closer exposure. Versus the close contact Volume 9 Issue 3 e00972-21 MicrobiolSpectrum.asm.org 6 Identifying Recent Tuberculosis Exposure TABLE 2 Utility of a TB22TB1 IFN-g value .0.6 IUml21 for predicting a higher likelihood of recent tuberculosis infectiona Crude analysis Adjusted analysisb Variable No. with QFT-plus (+) OR (95% CI) P value OR (95% CI) P value Whole cohort (n = 686) Group (likelihood of recent infection) Contacts (high) 368 1 1 IMID patients (low) 229 0.78 (0.47–1.30) 0.339 0.71 (0.31–1.61) 0.406 ASPFA (indeterminate) 89 0.81 (0.39–1.66) 0.556 0.86 (0.49–1.52) 0.604 Contacts with known sputum smear result for the index case (n = 360) Type of contact Close 201 1 1 Frequent 100 1.47 (0.74–2.94) 0.269 1.29 (0.63–2.62) 0.490 Sporadic 59 1.77 (0.81–3.89) 0.153 1.55 (0.67–3.60) 0.307 aASPFA, asylum seekers or people from abroad; CI, confidence interval; IMID, immune-mediated inflammatory diseases; OR, odds ratio; TB1, antigen tube 1 of the QFT-plus; TB2, antigen tube 2 of the QFT-plus. bAdjusted for age, gender, incidence of tuberculosis in the country of birth (, or $25  105), and European origin (yes or no) for the whole cohort and contacts, and also adjusted for the sputum smear result of the index case (positive or negative acid-fast bacillus stain) for contacts. subgroup, ORs were 1.47 (95% CI, 0.74 to 2.94) for frequent and 1.77 (95% CI, 0.81 to 3.89) for sporadic subgroups in the univariate regression analysis. This lack of associa- tion also remained after the corresponding adjusted analyses for the same variables plus the sputum smear result of the index case (Table 2). DISCUSSION The introduction of the QFT-plus test has been accompanied by the expectation that this would allow newly acquired TB infection to be distinguished from remotely acquired TB infection. However, in this study, we could not validate previous findings suggesting that a TB22TB1 value .0.6 IUml21 served as a proxy marker for recent infection. In one study (11), 14 of 105 (13.3%) infected contacts had a TB22TB1 value .0.6 IUml21, and the quantitative TB2 IFN-g response was higher in contacts living in the same room as index cases than in other contacts. In an Italian study performed in a cohort of 119 TB contacts, 18 of 68 (26.5%) QFT-plus-positive contacts had a TB22TB1 value .0.6 IUml21. This revealed an independently positive association between a TB22TB1 value .0.6 IUml21 and both European origin and sleeping in the same room as the index case, consistent with newly acquired infection from the index case (10). However, the chief limitation of these two prior studies was the absence of comparison between contacts and groups with low or zero risk of recent exposure. Indeed, coupled with the small numbers of individuals with TB22TB1 values .0.6 IUml21 (14 and 18 in each study, respectively), this was a major limitation on their ability to obtain definitive results about a hypothetical association between TB22TB1 and risk of recent TB expo- sure. A laboratory-based data study (12) found that contacts and individuals with peri- odic checks by occupational health services had a higher proportion (33%) of TB22TB1 values .0.6 IUml21 than did patients tested before immunosuppressant therapy (11%). However, the consistency of the results was again limited by the small number of participants with a TB22TB1 value .0.6 (18 contacts and 2 patients tested before immunosuppressant therapy). In our study, to provide a contrast with data for individuals at high risk of recent ex- posure, we included a control group of individuals without known risk of recent expo- sure (IMID group) and another group of individuals with indeterminate risk (ASPFA group). We included the ASPFA group because it is representative of those commonly screened for TB infection in Europe, and as such, the findings could have real-life impli- cations for clinical management. Our results show that samples from individuals with high, low, and indeterminate risk of recent exposure behave in the same manner in terms of the IFN-g response and differences in TB1 and TB2 tubes. Two additional anal- yses support this observation. First, only 1 of the 15 converters (which supposedly indi- cates recently acquired infection) had a TB22TB1 value .0.6 IUml21, and second, Volume 9 Issue 3 e00972-21 MicrobiolSpectrum.asm.org 7 Pérez-Recio et al. there was no association between the smear status of the index case and a TB22TB1 value .0.6 IUml21. Given that a positive sputum smear in the index case implies a greater risk of transmission, our findings support even further the lack of correlation between a higher IFN-g response in the TB2 tube and recent TB exposure. Consistent with our data, a cross-sectional study of 31 recent contacts with positive QFT-plus results (from a total of 412 TB contacts) failed to show a distinct CD81 T-cell response (13). In another prospective study of 492 TB contacts that assessed the QFT- plus value for predicting incident TB, the receiver operating characteristic curves were similar for the TB1 and TB2 results (14). Therefore, a difference between the TB2 and TB1 results (as a proxy for the CD81 T-cell response) failed to discriminate progressors from nonprogressors, indirectly providing evidence of a lack of discriminative value for recent TB exposure (14). To the best of our knowledge, this is the largest study to have investigated the dif- ference in IFN-g response between TB1 and TB2 QFT-plus tubes in individuals with dif- ferent risks of recent TB exposure. Contrary to previous studies, our results argue against the potential clinical utility of a TB22TB1 value .0.6 IUml21 as a proxy for recent TB infection. This study has limitations that deserve comment. First, the retrospective design may have jeopardized the accuracy of data on contact status and immunosuppression in the IMID group, which may have led to some cases being misclassified. Second, the actual hours and places of exposure to the index case were only available in a small number of cases, and these data were discarded for the analyses. This limitation was palliated, at least in part, by the fact that the concentric circle approach already accounts for the place of exposure and stratifies contacts for more/less than 6 h and for daily versus not daily. Third, the optimal control group would have been healthy individuals with positive QFT-plus results and no evidence of recent contact. However, it was not feasible to identify such patients because screening is not routinely per- formed in otherwise healthy people with no risk of developing active TB. In conclusion, a difference of .0.6 IUml21 in the TB22TB1 value was not associated with a higher risk of recent TB exposure, which puts into question the potential clinical utility of this measure as a proxy that can identify recently acquired TB infection. MATERIALS AND METHODS Study design and participants. We conducted a retrospective multicenter study to evaluate the IFN-g responses in QFT-plus tubes, including the differences between the TB1 and TB2 tubes (i.e., the TB22TB1 value), in people who tested positive for TB infection between 1 June 2016 and 31 May 2018. The study was conducted at 19 centers in 7 countries (11 hospitals in Spain, 2 in Sweden, 2 in Croatia, and 1 each in Chile, Turkey, Serbia, and Italy). We included individuals $18 years of age who were (i) assessed as a part of a TB contact investiga- tion, (ii) screened for TB infection prior to therapy with biologic agents because of immune-mediated inflammatory diseases (IMID), or (iii) asylum seekers or people from abroad (ASPFA). We only included contacts of microbiologically confirmed pulmonary TB patients. The extent of exposure was defined, according to the concentric circles approach (15), as close when the contact was $6 h daily, frequent when the contact was ,6 h daily, and sporadic when it was not daily. For patients screened before bio- logical therapy, we only included those without known recent exposure to a patient with TB. We excluded individuals with HIV infection, advanced chronic liver disease, end-stage kidney disease, solid organ or hematopoietic stem cell transplantation, corticosteroid treatment (.10 mg/day of prednisone or equivalent), biologic agents, or chemotherapy for cancer in the 8 weeks before testing or with suspi- cion of active TB. According to the risk of recent TB exposure, participants were considered (i) high risk, which included contacts of TB cases, (ii) low risk, which included patients screened before therapy with biolog- ics, and (iii) indeterminate risk, which included ASPFA. Conversion of the QFT-plus test was defined has having a positive test that followed a negative test performed 8 to 12 weeks apart, according to the manufacturer’s instructions (16). A more conservative definition of conversion (outside the borderline range) (17) was established as an increase in the IFN-g concentration from ,0.20 IUml21 to .0.99 IUml21 in the TB1, TB2, or both tubes. Data collection and analysis. We collected data on demographics, reasons for testing, bacillus Calmette-Guérin (Mycobacterium bovis BCG) vaccination status, and QFT-plus results. Study data were collected and managed using REDCap (Research Electronic Data Capture) tools hosted at Bellvitge Biomedical Research Institute (IDIBELL) (18). Due to the exploratory aim of the study, no formal sample Volume 9 Issue 3 e00972-21 MicrobiolSpectrum.asm.org 8 Identifying Recent Tuberculosis Exposure size calculation was determined based on a single hypothesis test. The final number of included partici- pants was defined by the available participants in each center. The data analysis proceeded as follows. First, we compared TB1 and TB2 IFN-g concentrations within and between recent exposure risk groups (i.e., high, low, and indeterminate) and within and between each contact type (i.e., close, frequent, and sporadic). Second, we compared the proportions of cases in which IFN-g production was higher in TB2 than in TB1 between groups by risk of recent exposure and type of contact. Finally, we compared the TB22TB1 values between the same groups. Given that the intrinsic variability of the QFT-plus could produce differences between the two tubes, we reduced this bias by treating any difference in TB22TB1 values .0.6 IUml21 as a real difference. We chose this cutoff (TB22TB1 . 0.6 IUml21) because this has been considered the threshold for CD81 T-cell response and, by using this, some studies have shown associations between TB22TB1 . 0.6 IUml21 and exposure in- tensity, proximity to the index case, and proximity to time of infection (10–12). We performed Mann-Whitney tests for comparisons between two groups and Kruskal-Wallis tests for comparisons among more than two groups. The Wilcoxon matched-pairs signed-rank test was used to compare two variables within the same group. We assessed the association of the TB22TB1 value with a higher probability of recent infection by crude and adjusted logistic regression analysis. First, we assessed the association of a TB22TB1 value .0.6 IUml21 and risk of recent exposure by univariate regression analysis, both in the whole cohort (contacts, IMID patients, and ASPFA) and in the contact subgroups (close, frequent, and sporadic), and second, we adjusted for variables potentially associated with changes in TB22TB1 results (age, gender, incidence of TB in the country of birth, and European ori- gin, for the whole cohort and contacts, and also adjusted for the sputum smear of the index case), regardless of their significance in the univariate analysis. All reported P values were calculated with sta- tistical significance set at a P value of less than 0.05. Odds ratios (ORs) and 95% confidence intervals (CIs) are reported. Statistical analyses were performed using R version 3.5.0 for Windows (19). Ethics. The study received ethical clearance from the Ethics Committees of all participating centers. The confidentiality of data was protected according to National and European Protection Data law (e.g., General Data Protection Regulation; EU, 2016/679). Data availability. Data are fully available at the REDCap tools hosted at Bellvitge Biomedical Research Institute (IDIBELL) upon request to the principal investigator and corresponding author Miguel Santin. SUPPLEMENTAL MATERIAL SUPPLEMENTAL FILE 1, PDF file, 0.2 MB. ACKNOWLEDGMENTS This research received financial support for publication from the Department of Clinical Sciences of the University of Barcelona. M. Santin and S. Pérez-Recio conceived and designed the study. S. Pérez-Recio, M. D. Grijota-Camino, A. Sánchez-Montalvá, L. Barcia, S. Campos-Gutiérrez, V. Pomar, R. Rabuñal-Rey, M. E. Balcells, D. Gazel, N. Montiel, D. Vicente, I. Goic-Barišic, T. Schön, J. Paues, I. Marekovic, J. Cacho-Calvo, A. Barac, D. Goletti, M. García-Gasalla, J. M. Barcala, M. T. Tórtola, L. Anibarro, I. Suárez-Toste, E. Moga, M. J. Gude-Gonzalez, R. Naves, T. Karslıgil, T. Martin-Peñaranda, G. Stevanovic, M. Trigo, V. Rubio, _I. Karaog  lan, N. Bayram, F. Alcaide, and M. Santin performed case reviews and data collection. M. Santin, S. Pérez-Recio, C. Tebé, and N. Pallarès performed data analysis and interpretation. M. Santin and S. Pérez-Recio wrote the original draft. S. Pérez-Recio, N. Pallarès, M. D. Grijota-Camino, A. Sánchez-Montalvá, L. Barcia, S. Campos-Gutiérrez, V. Pomar, R. Rabuñal-Rey, M. E. Balcells, D. Gazel, N. Montiel, D. Vicente, I. Goic-Barišic, T. Schön, J. Paues, I. Marekovic, J. Cacho-Calvo, A. Barac, D. Goletti, M. García-Gasalla, J. M. Barcala, M. T. Tórtola, L. Anibarro, I. Suárez-Toste, E. Moga, M. J. Gude-Gonzalez, R. Naves, T. Karslıgil, T. Martin-Peñaranda, G. Stevanovic, M. Trigo, V. Rubio, _I. Karaog  lan, N. Bayram, F. Alcaide, C. Tebé, and M. Santin reviewed and edited the manuscript and approved the final draft. REFERENCES 1. World Health Organization. 2014. The End TB Strategy. Global strategy Respir Crit Care Med 183:88–95. https://doi.org/10.1164/rccm.201006 and targets for tuberculosis prevention, care and control after 2015. A67/ -0974OC. 11. The official text approved by the 67th World Health Assembly, May 3. Rangaka MX, Wilkinson KA, Glynn JR, Ling D, Menzies D, Mwansa- 2014. World Health Organization, Geneva, Switzerland. Kambafwile J, Fielding K, Wilkinson RJ, Pai M. 2012. Predictive value of 2. Diel R, Loddenkemper R, Niemann S, Meywald-Walter K, Nienhaus A. interferon-g release assays for incident active tuberculosis: a systematic 2011. Negative and positive predictive value of a whole-blood interferon- review and meta-analysis. Lancet Infect Dis 12:45–55. https://doi.org/10 gamma release assay for developing active tuberculosis: an update. Am J .1016/S1473-3099(11)70210-9. Volume 9 Issue 3 e00972-21 MicrobiolSpectrum.asm.org 9 Pérez-Recio et al. 4. Muñoz L, Santin M. 2013. Interferon-g release assays versus tuberculin 11. Lee MR, Chang CH, Chang LY, Chuang YC, Sun HY, Wang JT, Wang JY. skin test for targeting people for tuberculosis preventive treatment: an 2019. CD8 response measured by QuantiFERON-TB gold plus and tuber- evidence-based review. J Infect 66:381–387. https://doi.org/10.1016/j.jinf culosis disease status. J Infect 78:299–304. https://doi.org/10.1016/j.jinf .2012.12.005. .2019.01.007. 5. Zellweger JP, Sotgiu G, Block M, Dore S, Altet N, Blunschi R, Bogyi M, 12. Pieterman ED, Liqui Lung FG, Verbon A, Bax HI, Ang CW, Berkhout J, Bothamley G, Bothe C, Codecasa L, Costa P, Dominguez J, Duarte R, Fløe Blaauw G, Brandenburg A, van Burgel ND, Claessen A, van Dijk K, Heron A, Fresard I, García-García JM, Goletti D, Halm P, Hellwig D, Henninger E, M, Hooghiemstra M, Leussenkamp-Hummelink R, van Lochem E, van Loo Heykes-Uden H, Horn L, Kruczak K, Latorre I, Pache G, Rath H, IHM, Mulder B, Ott A, Pontesilli O, Reuwer A, Rombouts P, Saegeman V, Ringshausen FC, Ruiz AS, Solovic I, Souza-Galvão ML, Widmer U, Witte P, Scholing M, Vainio S, de Steenwinkel JEM. 2018. A multicentre verification Lange C, TBNET. 2015. Risk assessment of tuberculosis in contacts by IFN- study of the QuantiFERON-TB gold plus assay. Tuberculosis (Edinb) 108: g release assays. A Tuberculosis Network European Trials Group Study. 136–142. https://doi.org/10.1016/j.tube.2017.11.014. Am J Respir Crit Care Med 191:1176–1184. https://doi.org/10.1164/rccm 13. Tsuyuzaki M, Igari H, Okada N, Suzuki K. 2020. Role of CD8 T-cell in .201502-0232OC. immune response to tuberculosis-specific antigen in QuantiFERON-TB 6. Rozot V, Patrizia A, Vigano S, Mazza-Stalder J, Idrizi E, Day CL, Perreau M, gold plus. J Infect Chemother 26:570–574. https://doi.org/10.1016/j.jiac Lazor-Blanchet C, Ohmiti K, Goletti D, Bart PA, Hanekom W, Scriba TJ, .2020.01.011. Nicod L, Pantaleo G, Harari A. 2015. Combined use of Mycobacterium tu- 14. Gupta RK, Kunst H, Lipman M, Noursadeghi M, Jackson C, Southern J, berculosis-specific CD4 and CD8 T-cell responses is a powerful diagnostic Imran A, Lozewicz S, Abubakar I. 2020. Evaluation of QuantiFERON-TB tool of active tuberculosis. Clin Infect Dis 60:432–437. https://doi.org/10 gold plus for predicting incident tuberculosis among recent contacts: a .1093/cid/ciu795. prospective cohort study. Ann Am Thorac Soc 17:646–650. https://doi 7. Lewinsohn DA, Heinzel AS, Gardner JM, Zhu L, Alderson MR, Lewinsohn .org/10.1513/AnnalsATS.201905-407RL. DM. 2003. Mycobacterium tuberculosis-specific CD81 T cells preferen- 15. Grupo de Estudio de Contactos de la UITB. 1999. Documento de Con- tially recognize heavily infected cells. Am J Respir Crit Care Med 168: senso sobre el estudio de contactos en los pacientes tuberculosos. Med 1346–1352. https://doi.org/10.1164/rccm.200306-837OC. Clin (Barc) 112:151–156. 8. Nikolova M, Markova R, Drenska R, Muhtarova M, Todorova Y, Dimitrov V, 16. Qiagen. 2016. QuantiFERON-TB gold plus (QFT-Plus) ELISA package insert. Taskov H, Saltini C, Amicosante M. 2013. Antigen-specific CD4- and CD8- Qiagen, Germantown, MD. www.quantiferon.com/irm/content/PI/QTF/ positive signatures in different phases of Mycobacterium tuberculosis PLUS/2PK-Elisa/UK.pdf. infection. Diagn Microbiol Infect Dis 75:277–281. https://doi.org/10.1016/j 17. Jonsson J, Westman A, Bruchfeld J, Sturegård E, Gaines H, Schön T. 2017. .diagmicrobio.2012.11.023. A borderline range for Quantiferon Gold In-Tube results. PLoS One 12: 9. Petruccioli E, Chiacchio T, Pepponi I, Vanini V, Urso R, Cuzzi G, Barcellini L, e0187313. https://doi.org/10.1371/journal.pone.0187313. Cirillo DM, Palmieri F, Ippolito G, Goletti D. 2016. First characterization of 18. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. 2009. the CD4 and CD8 T-cell responses to QuantiFERON-TB Plus. J Infect 73: Research electronic data capture (REDCap)—a metadata-driven method- 588–597. https://doi.org/10.1016/j.jinf.2016.09.008. ology and workflow process for providing translational research infor- 10. Barcellini L, Borroni E, Brown J, Brunetti E, Campisi D, Castellotti PF, matics support. J Biomed Inform 42:377–381. https://doi.org/10.1016/j.jbi Codecasa LR, Cugnata F, Di Serio C, Ferrarese M, Goletti D, Lipman M, .2008.08.010. Rancoita PM, Russo G, Tadolini M, Vanino E, Cirillo DM. 2016. First evalua- 19. R Core Team. 2020. R: a language and environment for statistical comput- tion of QuantiFERON-TB gold plus performance in contact screening. Eur ing. R Foundation for Statistical Computing, Vienna, Austria. https://www Respir J 48:1411–1419. https://doi.org/10.1183/13993003.00510-2016. .R-project.org. Volume 9 Issue 3 e00972-21 MicrobiolSpectrum.asm.org 10