Comparative Performance of Tuberculin Skin Test, QuantiFERON-TB-Gold In Tube Assay, and T-Spot.TB Test in Contact Investigations for Tuberculosis*

Transcription

1 CHEST Original Research INFECTIOUS GRANULOMATOUS DISEASE Comparative Performance of Tuberculin Skin Test, QuantiFERON-TB-Gold In Tube Assay, and T-Spot.TB Test in Contact Investigations for Tuberculosis* Roland Diel, MD, MPH; Robert Loddenkemper, MD, FCCP; Karen Meywald-Walter, MD; Rene Gottschalk, MD; and Albert Nienhaus, MD, MPH Rationale: Mycobacterium tuberculosis (MTB)-specific interferon- release assays (IGRAs) are an alternative or adjunct to the tuberculin skin test (TST) in identifying recent contacts with latent tuberculosis infection (LTBI), but there are scarce data directly comparing performance of the tests. Objective: To evaluate the agreement between both IGRAs and to determine which contacts were most likely to represent LTBI, the QuantiFERON-TB-Gold In Tube assay (QFT) and the T-Spot.TB test (T-Spot) were compared in TST-positive persons recently exposed to pulmonary tuberculosis cases. Methods: Prospectively enrolled close contacts (n 812) of 123 culture-confirmed tuberculosis source cases underwent IGRA testing using standardized collected data. Factors independently influencing the risk of MTB infection and their interactions with each other were evaluated by multivariate analysis. Results: Five variables were found to significantly predict a positive IGRA test result (age, source case acid-fast bacilli positive and/or coughing, cumulative exposure time, foreign origin). There was excellent agreement between the two IGRAs (93.9%, 0.85), with QFT finding 30.2% of contacts positive and T-Spot finding 28.7%. Assuming positivity to both IGRAs as true infection, sensitivity of the TST at > 10 mm was 72% and at > 15 mm was 39.7%. The use of either IGRA as a replacement for the TST would decrease the number of LTBI suspects to be investigated by approximately 70%. Conclusions: IGRAs are a more accurate indicator of the presence of LTBI than the TST. Both QFT and T-Spot appear to be valuable public health tools, showing excellent agreement with each other. (CHEST 2009; 135: ) Key words: independent risk factor; interferon- release assay; latent infection; tuberculin skin test; tuberculosis Abbreviations: AFB acid-fast bacilli; BCG bacille Calmette-Guerin; CI confidence interval; IGRA interferon- release assay; LTBI latent tuberculosis infection; MTB Mycobacterium tuberculosis; OR odds ratio; QFT QuantiFERON-TB-Gold In Tube assay; T-Spot T-Spot.TB test; TST tuberculin skin test In low-incidence countries, identification and treatment of individuals recently infected with Mycobacterium tuberculosis (MTB) complex is central to tuberculosis control. 1 Until recently the detection of latent tuberculosis infection (LTBI) could only be based on tuberculin skin testing (TST), but with the advent of highly specific interferon- release assays (IGRAs), alternative tests are now available. 2 While the negative predictive value of the TST in excluding LTBI may be high, 3 the test is intrinsically confounded by crossreactive immune responses due to bacille Calmette- Guerin (BCG) vaccination or infection with nontuberculous mycobacteria. In contrast, IGRAs are unaffected by BCG vaccination and most nontuberculous mycobacteria and are at least as sensitive as the TST for detecting active tuberculosis. 4,5 However, no gold standard exists for LTBI, and thus evaluation of tests for LTBI is usually conducted by measuring association of a population s test results with their risk factors for infection. To evaluate tests for LTBI, access to contacts of persons with pulmonary tuberculosis is required. Numerous studies 6,7 have shown that the risk of becoming infected with MTB is largely determined by the frequency and duration of exposure to airborne droplet nuclei in poorly ventilated indoor 1010 Original Research

2 settings. Sputum positivity of source cases is an important factor for infectivity, and the risk of contracting MTB infection is highest for contacts sitting, sleeping, or working closely to an infectious person. 8,9 Although acid-fast bacilli (AFB) smearnegative patients have a lower infectious potential than smear-positive cases, molecular epidemiologic studies 10 have shown that they also effectively transmit tuberculosis but at approximately 20% of the rate for smear-positive patients. Transmission of tuberculosis bacilli is dramatically increased with the index patient coughing, especially in patients with cavitary pulmonary tuberculosis. 11 Additionally, to increase pretest predictability of MTB transmission, fixed cutoffs for aggregate exposure time of a contact to their index case have been proposed. These are 8 h of exposure to an AFB smear-positive source case, 12 and 40-h exposure in closed rooms, irrespective of smear status. 13 However, the above parameters are all based on TST. Two different IGRAs measuring responses to the MTB-specific proteins ESAT-6 and CFP-10 are now commercially available in many parts of the world: QuantiFERON-TB Gold In Tube assay (QFT) [Cellestis; Carnegie, VIC, Australia] and T-Spot.TB (T-Spot) [Oxford Immunotec; Abingdon, UK]. A recent metaanalysis found that both IGRAs were both more specific than the TST and at least as sensitive for detecting active TB (as a surrogate for LTBI) as the TST. 14 Additionally, the ELISPOTbased T-Spot test appeared more sensitive, but less specific, than the whole-blood, enzyme-linked immunosorbent assay-based QFT. However, only limited studies directly comparing QFT (In Tube version) with T-Spot and the TST in the same individuals have been reported, predominantly in cases of tuberculosis disease, for monitoring the *From the School of Public Health (Dr. Diel), Heinrich Heine University of Düsseldorf, Düsseldorf; German Central Committee Against Tuberculosis (Dr. Loddenkemper), Lungenklinik Heckeshorn, HELIOS, Klinikum Emil von Behring, Berlin; Public Health Department Hamburg-Central (Dr. Meywald- Walter), Hamburg; Institute of Medical Virology (Dr. Gottschalk), University Hospital, Johann Wolfgang Goethe University, Frankfurt am Main; and Institution for Statutory Accident Insurance and Prevention in the Health and Welfare Services (Dr. Nienhaus), Hamburg, Germany. Material and laboratory performance of the T-Spot.TB was sponsored by Oxford Immunotec. The authors have no conflicts of interest to disclose. Manuscript received September 3, 2008; revision accepted October 10, Reproduction of this article is prohibited without written permission from the American College of Chest Physicians ( org/misc/reprints.shtml). Correspondence to: Roland Diel, MD, MPH, Assistant Professor, School of Public Health c/o Institute of Medical Sociology, University of Düsseldorf, Postbox , Düsseldorf, Germany; Roland.Diel@uni-duesseldorf.de DOI: /chest effect of treatment 20 or in contacts of one single source case. 21 Thus, one aim of the present study was to generate information on their relative performance characteristics when used for screening recent contacts. We undertook a prospective community-based study of contacts with recent exposure to people with infectious tuberculosis, assessing the use of T-Spot head to head with QFT. We used a study design based on the National Institute for Clinical Excellence tuberculosis guidelines, 22 which recommends retesting those with a positive TST result using an IGRA as a confirmatory test. The study took place in Hamburg, with 1.8 million residents, the secondlargest city in Germany and 1 of the 16 German Federal States, of which Hamburg has the highest tuberculosis incidence rate (10.8 per inhabitants in 2005). 23 It also has the highest proportion of foreign residents, 14.2%, compared with a German national average of 8.8%. 24 Materials and Methods Study Design and Inclusion Criteria All contacts of culture-confirmed pulmonary tuberculosis cases that were reported to the Hamburg Public Health Department from December 2006 until February 2008 and who met the inclusion criteria of close contact (see following) were prospectively enrolled in the study. The study was embedded in our routine contact investigation work according to the legal mandate of the German Infection Protection Act and approved by the local ethics committee. All individuals were informed of the nature of the study and agreed to participate. All contacts were tested by the TST at least 8 weeks after identification of their respective source case. Definition of Close Contacts Close contacts were defined as people who had been exposed to persons with culture proven pulmonary tuberculosis during their infectious stage (see online supplement for details). Questionnaire Sociodemographic and clinical data were derived from interviews of the individuals by trained public health staff using a standardized questionnaire (see online supplement). IGRAs and Mantoux TST TST was administered by the Mantoux method using 0.1 ml (2 tuberculin units) of purified protein derivative RT23 (Statens Serum Institute; Copenhagen, Denmark). TST reaction was scored as positive if induration diameter was 5 mm, according to the German guidelines, 25 irrespective of BCG vaccination status. Induration was read by trained and well-experienced public health nurses. All TST-positive contacts had blood drawn for both QFT and T-Spot on the day of reading the TST. IGRA testing was performed and interpreted according to manufacturer instructions (see online supplement). CHEST / 135 / 4/ APRIL,

3 Statistical Analysis Statistical analysis were performed as described in the online supplement. Results Study Population A total of 2,004 close contacts of patients with culture-confirmed pulmonary tuberculosis were enrolled into the study. Six contacts with history of a TST 5 mm who had been subjects in previous contact tracing exercises and two contacts who had been treated for active tuberculosis were excluded. Seven contacts did not return for TST reading. Of the remaining 1,989 contacts (none with a history of a prior TST), 842 were TST positive ( 5 mm) and had blood drawn for QFT and T-Spot. For 22 individuals, the T-Spot result could not be determined due to inability to isolate sufficient lymphocytes. Of the remaining 820 contacts, T-Spot was indeterminate in 7 contacts (7 of 820 contacts, 0.9%) and QFT in 1 contact (1 of 820 contacts, 0.1%). None of the contacts reported or were known to have any medical reasons for immunosuppression, especially being seropositive for HIV, undergoing hemodialysis, currently being treated with immunosuppressive medication, having a malignant disease or diabetes mellitus, or recently been immunized with live vaccines. Thus, any relationship to indeterminate results could not be evaluated. In summary, 812 TST positive ( 5 mm) contacts of a total of 123 source cases, 57 AFB-smear positive and 66 negative, all culture positive, with complete results available for QFT and T-Spot, formed the study population. Characteristics of the contacts are shown in Table 1. The average number of close contacts per source case was 6.6 and varied from 1 to 18 individuals; the mean cumulative exposure time ( SD) was hours (data not shown); 321 contacts (39.5%) were household or intimate contacts, 291 contacts (36.1%) were coworkers, 87 contacts (10.7%) were pupils or teachers at traditional or vocational schools, 51 contacts (6.3%) were health-care workers, 44 contacts (5.4%) were nonintimate friends, 11 contacts (1.7%) were copatients in a hospital, and 5 contacts (0.6%) were members of sports clubs. Only 215 of all contacts (26.5%) recalled an episode of their respective source case coughing. More than half of the contacts (n 453, 55.8%) were BCG vaccinated, with the rate of vaccination nearly equivalent for foreign born and those of German origin (51.9% vs 48.1%, respectively). Table 1 Description of the Study Population Variables No. % Age, yr Gender Female Male BCG No Yes Country of birth Germany Other country Mantoux TST, mm Smear status of source case Negative Positive Household or intimate contact No Yes Close contact to coughing index No Yes Contact time, h Total QFT and T-Spot Results Two hundred forty-five of the 812 TST-positive contacts (30.2%) were QFT positive and 233 contacts (28.7%) were positive by T-Spot (Table 2). The positive rate increased for both QFT and T-Spot with increasing TST cutoffs. BCG vaccination was highly associated with the probability of having a negative IGRA result (93 T-Spot positive among 453 BCG vaccinates vs 140 T-Spot positive among 359 unvaccinated; 99 of 453 vs 146 of 359, respectively, for the QFT; p for each IGRA). Agreement between the two IGRAs was high (762 of 812; 93.9%) with a value of (95% confidence interval [CI], 0.78 to 0.92). Discordant results between the two IGRAs were less frequent for those with a TST 15 mm (1.3% vs 6.2% for all participants). As shown in Figure 1, the majority of contacts with IGRA discordant results had responses close to the cutoffs of both tests, with the level of response for the test that was negative generally higher than for those contacts negative by both tests. Univariate analysis revealed that contacts who reported coughing of their source case were significantly more likely to be IGRA positive (QFT: 107 of 215 [49.8%] vs 138 of 597 [23.1%]; T-Spot: 99 of Original Research

4 Table 2 Results of IGRA Tests by Diameter of TST TST Results, mm QFT Positive T-Spot Positive Both Positive Discordant No. % No. % No. % No. % All subjects * 1.3 All ( 5) Unvaccinated All ( 5) BCG vaccinated * 2.6 All ( 5) *The one person with a TST 15 mm and a discordant result in the IGRA was positive by QFT and negative by T-Spot. [46.9%] vs 134 of 597 [22.4%]; p for each). There was also a significant relationship between the positivity rate for both IGRAs and cumulative exposure time of contacts (p ), and contacts of AFB-positive source cases were more likely to be IGRA positive than those of AFB-negative source cases (p ; Table 3). In contacts of AFBpositive source cases with exposure 40 h, the positivity rate for either of the two IGRAs was more than six times higher than that of persons in the short exposure ( 8 h) category; and in contacts of AFBnegative source cases, the positivity rate was more than twice that of persons in the short exposure category of positivity. QFT (IU/mL of IFN-γ) n=31 n= n=548 n= T-Spot.TB (No. spots) Figure 1. Comparison of QFT and T-Spot.TB results for all 812 contacts. The gray circle represents 358 subjects with QFT responses 0.1 ml and T-Spot results of 0 spots. The gray square represents 52 subjects with QFT responses 10 IU/mL and T-Spot responses of 30 spots. Performance of the Mantoux TST in Relation to IGRA Results Most contacts with a TST response 15 mm were also IGRA positive (68 of 77, 88.3% for QFT; and 67 of 77, 87.0% for T-Spot), suggesting high specificity for the TST when using this cutoff. However, only 56.7% (97 of 171) and 54.4% (93 of 171) of the contacts with TST responses of 11 to 15 mm and 14.2% (80 of 564) and 12.9% (73 of 564) with TST responses of 6 to 10 mm were positive by QFT and T-Spot, respectively, suggesting diminishing specificity with lower cutoffs (Fig 2). If a positive result for both IGRAs were to be assumed as representative of true infection, the sensitivity of the TST would be only moderate at a 10-mm cutoff (153 of 214, 71.5%) and very low at a 15-mm cutoff (67 of 214, 31.3%). Independent Predictors of Test Positivity Multiple logistical regression analysis revealed significant associations between positive IGRA results and increasing age, foreign origin, AFB smear positivity, source case coughing, and exposure time (Table 4). The chance of having a positive QFT or T-Spot result increased by 30% (p 0.003) and 20% (p 0.012), respectively, with each 10-year period of age. Foreign origin led to more than a twofold increase in likelihood of a positive QFT (p 0.001) or T-Spot (p 0.001) result. Positive QFT or T-Spot results were approximately twice as likely if the source case was AFB smear positive (p for both tests), and up to four times more likely if the index person had been coughing (p for both). Contact time as a single factor was associated with a higher chance of IGRA positivity only if CHEST / 135 / 4/ APRIL,

5 Table 3 QFT and T-SPOT.TB Results by Smear Status of Source Case and Contact Time QFT Positive T-Spot Positive Source Case Contact Time, h No. No. % p Value for Trend No. % p Value for Trend All All Smear negative All Smear positive All exposure time to the respective source case was at least 40 h (p 0.001). However, among contacts of AFB smear-positive source cases, the likelihood of a positive result increased by a factor of 4.4 (p 0.001) for QFT, and 3.2 (p 0.001) for T-Spot, for those contacts with a cumulative exposure time 8 h when compared to similar contacts of AFB smear-negative cases. This association strengthened if an exposure time of 8 h had been exceeded (odds ratios [ORs], 9.0 and 6.3, respectively; p for both tests) [Table 5]. Combining the variables AFB positivity of the source case and coughing also led to an additive effect markedly increasing the likelihood of IGRA positivity by 10.6 times for QFT (p 0.001) and 8.4 times for T-Spot (p 0.001). After controlling for confounding, being a household or intimate contact as a single factor did not reach statistical significance. Among contacts of foreign origin, the chance of having a positive IGRA more than doubled (ORs, 2.3 and 2.6 for QFT and T-Spot, respectively; each p 0.001), while those BCG vaccinated were 2.5 times less likely to be IGRA positive (OR, 0.4; p for both tests). % Test Positive QFT T-Spot.TB n=564 n=171 n= TST induration (mm) Figure 2. IGRA-positive rates stratified by TST induration. Correlation Between the Cutoffs Agreement between the two IGRAs was high (93.9%; 0.852) using the cutoffs of 0.35 IU/mL (QFT) and six spots (T-Spot) suggested by the manufacturers (Table 6). However, at these cutoffs discordant results were not balanced, with QFTpositive/T-Spot-negative results occurring more often than the inverse profile (31 subjects and 19 subjects, respectively). Discordant results were found to be balanced (22 for each profile) and their number minimized when cutoffs of 0.55 IU/mL for QFT and seven spots for T-Spot were used (agreement 94.6%; 0.865). Optimizing statistical concordance, the highest value (0.895) was obtained using a cutoff of 0.6 IU/mL for QFT and nine spots for T-Spot; however, the gain in statistical agreement ( ) was only 4.8% (from 85.2 to 89.5%) over that using the manufacturers cutoffs. Discussion This study compared the two commercially available IGRAs (QFT and T-Spot) in a large cohort of TST-positive ( 5 mm) close contacts of active tuberculosis cases. Positive results from both IGRAs were strongly associated with measures of exposure and infection risk, confirming results of previous studies 26 investigating the tests individually. The overall level of agreement between QFT and T-Spot was excellent (93.9%), suggesting that any differences in sensitivity between the IGRAs may be due to small fluctuations around the cutoffs. Analysis of varying cutoffs revealed use of 0.6 IU/mL for QFT and nine spots for T-Spot would provide a relatively small increment in overall agreement of 1.7% to 95.9%, insufficient for suggesting alteration of the company recommended cutoffs Original Research

6 Table 4 Adjusted ORs for QFT and T-Spot.TB QFT Positive* T-Spot Positive* Variables No. (%) OR (95% CI) p Value No. (%) OR (95% CI) p Value Age, yr 1.3 ( ) ( ) (21.4) 3 (21.4) (30.1) 25 (26.9) (23.5) 50 (23.5) (31.2) 76 (32.1) (34.0) 66 (31.1) (34.9) 13 (30.2) Gender (NS) (NS) Female 109 (28.8) (28.2) 1 Male 133 (30.7) 1.2 ( ) 126 (29.1) 1.1 ( ) BCG No 145 (40.4) (39.0) 1 Yes 97 (21.4) 0.4 ( ) 93 (20.5) 0.4 ( ) Country of birth Germany 86 (21.9) 1 78 (19.9) 1 Other 156 (37.1) 2.3 ( ) 155 (36.9) 2.6 ( ) Smear status of source case Negative 88 (22.6) 1 89 (22.7) 1 Positive 154 (36.5) 2.1 ( ) 144 (34.3) 1.9 ( ) Household or intimate contact (NS) (NS) No 115 (23.4) (22.8) 1 Yes 127 (39.6) 1.4 ( ) 121 (37.7) 1.5 ( ) Close contact to coughing index No 136 (22.8) (22.4) 1 Yes 106 (49.3) 4.0 ( ) 99 (46.0) 3.5 ( ) Contact time, h (13.9) 1 32 (13.4) (19.9) 1.8 ( ) 30 (19.2) 1.7 ( ) (42.6) 5.7 ( ) 171 (42.6) 4.9 ( ) *For both QFT and T-SPOT.TB, gender was excluded from the final model. NS not significant. A metaanalysis 14 comparing the two IGRAs suggested QFT was more specific than T-Spot but less sensitive for detecting active tuberculosis. Our results are in discord with this analysis because we found more contacts positive by QFT than T-Spot. However, we were evaluating both tests performance in people at risk of having LTBI, the primary target for IGRAs, rather than those with active Table 5 Interaction Between Smear Status and Cough and Between Smear and Contact Time QFT Positive* T-SPOT.TB Positive* Interaction* No. (%) OR (95% CI) No. (%) OR (95% CI) Smear/cough of source case Negative/no cough 59 (19.2) 1 59 (19.2) 1 Positive/no cough 77 (26.6) 2.0 ( ) 75 (25.9) 2.0 ( ) Negative/cough 30 (35.3) 2.5 ( ) 30 (35.3) 2.4 ( ) Positive/cough 76 (58.5) 10.6 ( ) 69 (53.1) 8.4 ( ) p Value Smear/contact time Negative/ 8 h 5 (4.8) 1 6 (5.7) 1 Negative/ 8 40 h 4 (5.6) 1.1 ( ) 5 (7.0) 1.2 ( ) Negative/ 40 h 80 (37.0) 10.2 ( ) 78 (36.1) 6.9 ( ) Posositive/ 8 h 28 (21.1) 4.4 ( ) 26 (19.5) 3.2 ( ) Positive/ 8 40 h 27 (31.8) 9.0 ( ) 25 (29.4) 6.3 ( ) Positive/ 40 h 98 (48.5) 18.6 ( ) 93 (46.0) 12.9 ( ) p Value *Two separate final models were calculated containing either the single variables or the combined variable smear and cough of source case or smear and contact time. All models contained age, country of birth, and household or intimate contact. CHEST / 135 / 4/ APRIL,

7 Table 6, Percentage of Agreement, Positive Results, and Discordant Results for Different Cut Points of Both IGRAs T-Spot QFT Cutoff, IU/mL Agreement, % Positive results, No Discordant results, No.* 35 vs vs vs vs 9 62 vs 6 68 vs 5 71 vs Agreement, % Positive results, No Discordant results, No.* 22 vs vs vs vs vs vs 6 46 vs Agreement, % Positive results, No Discordant results, No.* 19 vs vs vs vs vs vs 8 37 vs Agreement, % Positive results, No Discordant results, No.* 14 vs vs vs vs vs vs vs Agreement, % Positive results, No Discordant results, No.* 9vs64 13vs48 15vs31 17vs25 19vs17 21vs12 24vs Agreement, % Positive results, No Discordant results, No.* 8vs71 12vs55 14vs38 16vs32 18vs24 20vs19 22vs16 *QFT positive/t-spot.tb negative vs QFT negative/t-spot.tb positive. tuberculosis reported in the metaanalysis. Sensitivity of IGRAs has routinely been estimated in patients with active TB, as there is no gold standard for LTBI, but active tuberculosis patients have been reported to have reduced interferon- responses to tuberculosis antigens. 27 This, along with our findings, suggests that IGRA responses of active tuberculosis patients may not be representative of responses from people with LTBI. Assuming positivity to both IGRAs as true infection (ie, using this as a gold standard ), specificity of the TST was very poor (1,145 of 1,775; 64.5%) for our total cohort using the 5-mm cutoff recommended for contact tracing in Germany. Our findings therefore support the recommendations of the National Institute for Clinical Excellence and the German Central Committee Against Tuberculosis 28 of confirming positive TST reactions with an IGRA before offering treatment for LTBI. However, the practice of sequential testing, while perhaps slightly cheaper, 29,30 may not provide the most sensitive screening. We reported that some tuberculosis contacts can be QFT positive, but TST negative, and go on to active tuberculosis if not treated. 31 By using TST, even at a low cutoff of 5 mm, as the initial screening test, some infected individuals who would have been positive by an IGRA will remain undetected. Our results suggest that higher TST cutoffs 10 mm or 15 mm are not a viable option for screening contacts because they resulted in poor sensitivities of 72% and 39.7%, respectively (assuming IGRA positivity as true infection). Overall, both IGRAs demonstrated excellence performance in our study, but can our study results provide any insight into how to run contact tracing investigations? Preselection of contacts with the highest risk for contracting MTB infection is desirable in order to increase pretest probability of LTBI before contacts are screened 32 and to maximize cost-effectiveness of the process. 33 The highly specific IGRA tests offer new possibilities for following that objective. However, no study comparing the new IGRA tools with respect to the independent risk factors, prospectively following a larger community-based population of 1016 Original Research

8 close contacts, has yet been reported. Recently, Aissa and coworkers 34 provided a model for predicting the risk of infection among contacts of tuberculosis patients using the traditional TST. Using a cutoff of 10 mm among BCG-unvaccinated and at least 15 mm among vaccinated contacts to represent a true MTB infection in their study, eight independent risk factors were identified. While sensitivity of their model was acceptable, specificity and positive predictive values were only 0.34 and 0.36, respectively. However, if we had employed their TST model in our contact population, sensitivity would have been unacceptably low (112 of 214, 57%), assuming those positive to both IGRAs were truly infected. Of note, using a TST cutoff of 10 mm in BCG-vaccinated contacts instead of 5 mm would have misclassified 20 of 81 of those contacts (25%) as negative (Table 2). Our results confirm the overall importance of timedependent categorization of 8 h for contacts of AFB smear-positive source cases and for limiting testing of contacts of AFB smear-negative patients to those with 40-h cumulative exposure. However, the risk of MTB infection following short ( 8 h) but very close exposure to an AFB-positive source can be markedly increased. 35 Thus selection of contacts for screening has to be not only time based, but also activity based, and episodes of very close exposure should be carefully queried and documented. A usual approach for contact investigations is to focus highly on those exposed within the index case s home and family. 36 However, after controlling for confounding, household and intimate contact did not reach statistical significance for being an independent risk factor of infection. This suggests that some of the other exposure factors examined in our study have a comparatively greater importance. This is in line with the seemingly paradoxical results of Snider and coworkers, 37 who found a strong tendency for either all or none of the household contacts to be positive in a study involving 1,352 contacts of 796 source cases. It should also be noted that increasing age and foreign origin were associated with contacts being IGRA positive, despite efforts to exclude those with known prior exposure to active tuberculosis. This suggests that a proportion of the IGRA-positive subjects were due to prior MTB infection and not due to their recent exposure. Conclusions Following 812 immunocompetent contacts of infectious tuberculosis cases through routine contact tracing in a German metropolis suggests that the use of the highly MTB-specific IGRA tests offers the possibility of substantially reducing LTBI screening to those truly infected and provides a better basis for developing standardized contact investigation algorithms in terms of independent risk factors. ACKNOWLEDGMENT: The authors thank the staff of the office of tuberculosis control at the Public Health Department Hamburg-Central, without whom this study would not have been possible. Material and laboratory performance of the T-Spot.TB was sponsored by Oxford Immunotec. References 1 American Thoracic Society. Targeted tuberculin testing and treatment of latent tuberculosis infection. Am J Respir Crit Care Med 2000; 161:S221 S247 2 Rothel JS, Andersen P. Diagnosis of latent Mycobacterium tuberculosis infection: is the demise of the Mantoux test imminent? Expert Rev Anti Infect Ther 2005; 3: Berkel GM, Cobelens FG, de Vries G, et al. Tuberculin skin test: estimation of positive and negative predictive values from routine data. Int J Tuberc Lung Dis 2005; 9: Mori T, Sakatani M, Yamagishi F, et al. Specific detection of tuberculosis infection: an interferon- -based assay using new antigens. Am J Respir Crit Care Med 2004; 170: Ewer K, Deeks J, Alvarez L, et al. Comparison of T-cell-based assay with tuberculin skin test for diagnosis of Mycobacterium tuberculosis infection in a school tuberculosis outbreak. Lancet 2003; 361: Houk VN, Baker JH, Sorensen K, et al. The epidemiology of tuberculosis infection in a closed environment. Arch Environ Health 1968; 16: Sultan L, Nyka W, Mills C, et al. Tuberculosis disseminators: a study of the variability of aerial infectivity of tuberculosis patients. Am Rev Respir Dis 1960; 82: Marks SM, Taylor Z, Qualls NL, et al. Outcomes of contact investigations of infectious tuberculosis patients. Am J Respir Crit Care Med 2000; 162: Reichler MR, Reves R, Bur S, et al. Evaluation of investigations conducted to detect and prevent transmission of tuberculosis. JAMA 2002; 287: Behr MA, Warren SA, Salamon H, et al. Transmission of Mycobacterium tuberculosis from patients smear-negative for acid-fast bacilli. Lancet 1999; 353: Loudon RG, Spohn SK. Cough frequency and infectivity in patients with pulmonary tuberculosis. Am Rev Respir Dis 1969; 99: Kenyon TA, Valway SE, Ihle WW, et al. Transmission of multi-drug resistant Mycobacterium tuberculosis during a long airplane flight. N Engl J Med 1996; 334: Behr A, Hopewell PC, Paz EA, et al. Predictive value of contact investigation for identifying recent transmission of Mycobacterium tuberculosis. Am J Respir Crit Care Med 1998; 158: Pai M, Zwerling A, Menzies D. Systematic review: T-cellbased assays for the diagnosis of latent tuberculosis infection; an update. Ann Intern Med 2008; 149: Dominguez J, Ruiz-Manzano J, De Souza-Galvao M, et al. Comparison of two commercially available interferon blood tests for immunodiagnosis of tuberculosis. Clin Vaccine Immunol 2008; 15: Lee JY, Choi HJ, Park IN, et al. Comparison of two commercial interferon gamma assays for diagnosing Mycobacterium tuberculosis infection. Eur Respir J 2006; 28: Goletti D, Vincenti D, Carrara S, et al. Selected RD1 peptides for active tuberculosis diagnosis: comparison of a interferon whole-blood enzyme-linked immunosorbent assay CHEST / 135 / 4/ APRIL,

9 and an enzyme-linked immunospot assay. Clin Diagn Lab Immunol 2005; 12: Vincenti D, Carrara S, Butera O, et al. Response to region of difference 1 (RD1) epitopes in human immunodeficiency virus (HIV)-infected individuals enrolled with suspected active tuberculosis: a pilot study. Clin Exp Immunol 2007; 150: Detjen AK, Keil T, Roll S, et al. Interferon- release assays improve the diagnosis of tuberculosis and nontuberculous mycobacterial disease in children in a country with a low incidence of tuberculosis. Clin Infect Dis 2007; 45: Chee CB, KhinMar KW, Gan SH, et al. Latent tuberculosis infection treatment and T-cell responses to Mycobacterium tuberculosis-specific antigens. Am J Respir Crit Care Med 2007; 175: Arend S, Thjjsen SF, Leyten EM, et al. Comparison of two interferon- assays and tuberculin skin test for tracing tuberculosis contacts. Am J Respir Crit Care Med 2007; 175: National Institute for Health and Clinical Excellence. Clinical guideline 33: tuberculosis; clinical diagnosis and management of tuberculosis, and measures for its prevention and control. Available at: Accessed September 1, Bericht zur Epidemiologie der Tuberkulose in Deutschland für Berlin, Germany: Robert Koch Institut, Statistische Ämter des Bundes und der Länder: Gebiet und Bevölkerung: Ausländische Bevölkerung. Wiesbaden, Germany: Statistisches Bundesamt, German Central Committee Against Tuberculosis. Richtlinien für die Umgebungs-untersuchungen bei Tuberkulose [Guidelines for environmental contact tracing in tuberculosis]. Gesundheitswesen 1996; 58: Richeldi L. An update on the diagnosis of tuberculosis infection. Am J Respir Crit Care Med 2006; 174: Hirsch CS, Toossi Z, Othieno C, et al. Depressed T-cell interferon- responses in pulmonary tuberculosis: analysis of underlying mechanisms and modulation with therapy. J Infect Dis 1999; 180: Diel R, Forssbohm M, Loytved G, et al. German Central Committee Against Tuberculosis. Empfehlungen für die Umgebungsuntersuchungen bei Tuberkulose [Recom-mendations for environmental contact tracing in tuberculosis]. Pneumologie 2007; 58: Diel R, Nienhaus A, Loddenkemper R. Cost-effectiveness of interferon- release assay screening for latent tuberculosis infection treatment in Germany. Chest 2007; 131: Wrighton-Smith P, Zellweger J-P. Direct costs of three models for the screening of latent tuberculosis infection. Eur Respir J 2006; 28: Diel R, Loddenkemper R, Meywald-Walter K, et al. Predictive value of a whole-blood IFN- assay for the development of active TB disease. Am J Respir Crit Care Med 2008; 177: Weiss S. Contact investigations: how do they need to be designed for the 21st century? Am J Respir Crit Care Med 2002; 166: Leung CC, Bailey WC, Yew WW. Screening contacts of tuberculosis: overcoming obstacles with an old tool. Am J Respir Crit Care Med 2008; 177: Aissa K, Madhi F, Ronsin N, et al. Evaluation of a model for efficient screening of tuberculosis contact subjects. Am J Respir Crit Care Med 2008; 177: Diel R, Seidler A, Nienhaus A, et al. Occupational risk of tuberculosis transmission in a low incidence area. Respir Res 2005; 6:35 36 Guidelines for the investigation of contacts of persons with infectious tuberculosis: recommendations from the National Tuberculosis Controllers Association and CDC. MMWR Morb Mortal Wkly Rep 2005; 54(RR-15): Snider DE Jr, Kelly GD, Cauthen GM, et al. Infection and disease among contacts of tuberculosis cases with drugresistant and drug-susceptible bacilli. Am Rev Respir Dis 1985; 132: Original Research