The use of the tuberculin skin test (TST) to diagnose

Transcription

1 Annals of Internal Medicine Article Prognostic Value of a T-Cell Based, Interferon- Biomarker in Children with Tuberculosis Contact Mustafa Bakir, MD; Kerry A. Millington, DPhil; Ahmet Soysal, MD; Jonathan J. Deeks, PhD; Serpil Efee; Yasemin Aslan, SRN; Davinder P.S. Dosanjh, DPhil; and Ajit Lalvani, DM Background: Enzyme-linked immunospot (ELISpot) assay is an increasingly widely used, T-cell based, interferon- release assay for diagnosing tuberculosis infection, but whether positive results are prognostic of active tuberculosis is not known. Objective: To determine whether ELISpot results predict the development of active tuberculosis among persons with recent tuberculosis exposure. Design: Longitudinal cohort study of children and adolescents with tuberculosis contact recruited from October 2002 to April Setting: Community-based contact investigations in Turkey. Patients: 908 children and adolescents with recent household tuberculosis exposure. Intervention: Enzyme-linked immunospot assay, incorporating early secretory antigenic target-6 and culture filtrate protein-10, and tuberculin skin test were done at baseline. Measurements: Incidence rates ratios of progression to active tuberculosis for contacts with positive tuberculin skin test and ELISpot results, and relative incidence rates comparing contacts with positive and negative test results. Results: Isoniazid preventive therapy was given to 688 (76%) contacts according to local guidelines. Fifteen contacts developed active tuberculosis over 1201 person-years of follow-up. Of 381 contacts with positive ELISpot results, 11 developed active tuberculosis over 536 person-years of follow-up (incidence rate, 21 per 1000 person-years [95% CI, 10.2 to 36.7 per 1000 person-years]), a statistically significant 3- to 4-fold increased risk for progression relative to ELISpot-negative contacts. Of 550 contacts with positive tuberculin skin test results, 12 developed active tuberculosis over 722 person-years of follow-up (incidence rate, 17 per 1000 personyears [CI, 8.6 to 29.0 per 1000 person-years]). Limitation: Only 3 of the 15 incident cases were confirmed by culture. Conclusion: Positive ELISpot results predict subsequent development of active tuberculosis in recent tuberculosis contacts. Although tuberculosis contacts with positive ELISpot results have an incidence rate of tuberculosis similar to that of contacts with positive tuberculin skin test results, ELISpot testing could allow more focused targeting of preventive therapy to fewer contacts. Ann Intern Med. 2008;149: For author affiliations, see end of text. The use of the tuberculin skin test (TST) to diagnose latent tuberculosis infection was established and validated in large studies, which observed that persons with a positive TST result had a higher risk for active tuberculosis than those with a negative TST result (1 7). However, diagnostic specificity of the TST is confounded in bacille Calmette Guérin (BCG) vaccinated persons, and sensitivity is reduced in vulnerable populations, including immunocompromised persons and children (8). The advent of T-cell interferon- release assays (IGRAs) may offer a realistic alternative to the TST (9 15). Currently, there are only 2 forms of IGRAs using antigens specific to Mycobacterium tuberculosis: Either interferon- secretion is measured in whole blood by the enzyme-linked immunosorbent (ELISA) assay, or interferon- secreting T cells are enumerated by the enzyme-linked immunospot (ELISpot) assay. Much published data indicate that IGRAs are more specific than the TST (12, 15, 16), and cross-sectional studies have shown good correlation of IGRA results with tuberculosis exposure in contacts (17 22). Treatment of recent tuberculosis contacts with positive IGRA results, as recommended by some national guidelines (23, 24), will only provide clinical benefit if these persons are at increased risk for progression to tuberculosis compared with those with negative results. Thus, there is a widely recognized, urgent need for large longitudinal studies with clinical outcomes to determine the prognostic value of positive IGRA results for development of tuberculosis (12, 15, 23 28). A recent report of predominantly immunocompetent adults with tuberculosis contact who declined preventive therapy showed that a higher proportion of those who initially had a positive result by wholeblood ELISA (QuantiFERON-TB Gold In-Tube, Cellestis, Carnegie, Victoria, Australia) developed active tuberculosis during 2 years of follow-up compared with contacts with positive TST results, using a 5-mm cutoff point (29). However, this important observation, which was based on 6 incident tuberculosis cases, cannot be extrapolated to ELISpot because of substantial discordance between the 2 assays (12) or to contact investigations involving high-risk populations (30). We conducted a longitudinal cohort study to determine the prognostic value of positive interferon- ELISpot results for development of active tuberculosis in a key highrisk population: children and adolescents with recent See also: Print Editors Notes Related article Web-Only Conversion of graphics into slides 2008 American College of Physicians 777

2 Article Prognostic Value of ELISpot in Tuberculosis Context Interferon- release assays, such as enzyme-linked immunospot (ELISpot), are increasingly used to help diagnose latent tuberculosis infection, but information about the ability of these tests to predict development of tuberculosis among exposed persons is scant. Contribution This longitudinal study followed 908 children and adolescents with recent household exposure to tuberculosis, most of whom received preventive therapy. Of 381 children with positive ELISpot results, 11 subsequently developed active tuberculosis compared with 12 of 550 children with positive tuberculin skin test results. Implication Household contacts with positive ELISpot results have about a 3- to 4-fold increased risk for progression to active tuberculosis compared with contacts with negative ELISpot results. Compared with the tuberculin skin test, ELISpot could allow more focused targeting of preventive therapy to fewer contacts. The Editors household exposure to tuberculosis. Because tuberculosis infection in children is usually recently acquired, it serves as a key marker for M. tuberculosis transmission in the general population and triggers extensive contact investigations to identify infectious adult source cases. Moreover, prompt isoniazid preventive therapy in infected children reduces the risk for active tuberculosis, which is associated with high morbidity and mortality (31, 32). Validation of IGRAs for diagnosis of asymptomatic tuberculosis infection by determining the predictive value of positive test results in this population is therefore a global public health priority. METHODS Study Participants All adults with newly diagnosed sputum smear positive pulmonary tuberculosis at the 7 government-funded tuberculosis clinics on the Anatolian side of Istanbul over 18 months (from October 2002) were asked whether they had children living in the household, and we invited them to participate in the study. A total of 443 patients had 1 or more children in the household (household contacts). All agreed to participate and gave written informed consent on behalf of their children (17). When the index case was not the parent, consent was given by the child s grandparents or legal guardian. Contacts were included if they were 16 years or younger. There were no exclusion criteria for enrollment. Follow-up comprised 6 monthly symptom reviews and continued until incident tuberculosis was diagnosed or 2 years of follow-up was completed. We contacted parents or guardians who did not attend clinical follow-up appointments by telephone, and we assessed the health status of the child by using a standardized questionnaire. In addition, all children had a repeated ELISpot assay at 6-month follow-up. The institutional review board of Marmara University School of Medicine, Istanbul, Turkey; the Turkish Ministry of Health, Ankara, Turkey; and the World Health Organization Steering Committee on Research Involving Human Subjects, Geneva, Switzerland, granted ethical approval for the study. Children were vaccinated intradermally with BCG Pasteur 1173-P2 from age 2 to 3 months and were given a booster vaccination in the first year of primary school and at age 6 to 7 years, as recommended by Turkish Ministry of Health guidelines. We documented BCG vaccination by the number of BCG vaccination scars. We diagnosed and excluded prevalent cases of active tuberculosis from our cohort by history, physical examination, chest radiography, and microbiological investigations at enrollment. We administered a 6-month course of isoniazid preventive therapy to 3 groups of contacts on the basis of age and TST results in accordance with the Turkish Ministry of Health guidelines (33). The guidelines include all children younger than 6 years, regardless of baseline or subsequent TST results; children 6 years or older with a positive result in the first TST (that is, 10 mm in unvaccinated children and 15 mm in vaccinated children); and children 6 years or older with a negative result from the first TST but a converted TST result (that is, induration that increased by 6 mm from the first TST result and second TST induration 10 mm in unvaccinated children and 15 mm in vaccinated children). Parents or guardians, who were provided with tablets at 2 monthly intervals, administered isoniazid to their children. We questioned all parents or guardians who returned for follow-up about adherence to preventive treatment, and all reported full adherence. Ex Vivo Interferon- ELISpot Assays We performed ELISpot assays, using peptides spanning early secretory antigenic target-6 (ESAT-6) and culture filtrate protein-10 (CFP-10) (10 g/ml final concentration of each peptide; Louisiana State University Health Sciences Center Core Laboratories, New Orleans, Louisiana) and recombinant ESAT-6 (10 g/ml; Veterinary Laboratories Agency, Weybridge, United Kingdom) and recombinant CFP-10 (10 g/ml; Lionex Diagnostics and Therapeutics, Braunschweig, Germany), as described elsewhere (17, 21). The peptide assay was subsequently commercialized into the regulatory-approved T-SPOT.TB (Oxford Immunotec, Oxford, United Kingdom). Because the diagnostic potential of ESAT-6 and CFP-10 derived peptides and recombinant antigens has been assessed in published work, we determined the prognostic value of both forms of these M. tuberculosis specific antigens December 2008 Annals of Internal Medicine Volume 149 Number 11

3 Prognostic Value of ELISpot in Tuberculosis Article We counted and scored ELISpot plates by using an automated ELISpot counter (AID, Strassberg, Germany), using predefined size and intensity spot settings. We scored responses to peptides as positive if duplicate test wells contained a mean of 5 or more spot-forming cells (SFCs) more than the mean of the negative control wells and if this number was at least twice the mean of the negative control wells. We used this predefined threshold in all of our previous studies (17, 20 22, 34 38). We scored responses to antigen as positive if duplicate test wells contained a mean of at least 10 SFCs more than the mean of the negative control wells and if this number was at least twice the mean of the negative control wells. For simplicity, cells stimulated with ESAT-6 or CFP-10 peptides are referred to as ELISpot. TST We administered TST by the Mantoux method using 0.1 ml (2 tuberculin units) of purified protein derivative RT23 (Statens Serum Institut, Copenhagen, Denmark). The study pediatricians, who were blinded to ELISpot results, did the test and read and interpreted the results. They noted the cutaneous appearance of peau d orange in all participants, confirming intradermal inoculation of purified protein derivative. Induration was measured after 72 to 96 hours with a ruler. For our analysis, we scored a positive TST response if the induration diameter was 5 mm or greater, as recommended by the American Thoracic Society and Centers for Disease Control and Prevention guidelines (39). Assessment of Outcome Each child contact was to be clinically followed up every 6 months for 2 years, but we asked parents to return with the child immediately for further clinical assessment if the child developed intercurrent symptoms. The study pediatricians, who were blinded to the ELISpot results, diagnosed incident tuberculosis on the basis of clinical, radiologic, and microbiological criteria. Two independent clinicians separately assessed the clinical and radiologic evidence for culture-negative incident cases, and in each case, the diagnosis was further confirmed by a documented successful clinical and radiologic response to antituberculosis treatment. We treated children with a diagnosis of tuberculosis with 6 months of standard chemotherapy, and we extended the continuation phase of isoniazid and rifampicin to 10 months for children with miliary tuberculosis. We treated children with a diagnosis of multidrug-resistant tuberculosis with second-line agents on the basis of antibiotic susceptibility results. Statistical Analysis We defined each child s entry into the study as the date first examined and tested with the ELISpot and TST (baseline). The end point of the study for each child contact was development of active tuberculosis or the last follow-up assessment (whether by telephone or a clinic visit). We compared differences in baseline characteristics of participants who developed active tuberculosis and those who did not by using the chi-square test for binary variables and the Mann Whitney test for continuous variables. We used Poisson regression to estimate incidence rates of progression to active tuberculosis per 1000 person-years of followup, together with 95% CIs. We estimated incidence rate ratios to compare the prognostic value of test-positive with test-negative results, both unadjusted and adjusted for preventive therapy. We compared differences in SFCs at baseline by using the Mann Whitney test. We did all analyses in Stata, version 9.1 (StataCorp, College Station, Texas). Role of the Funding Source The Wellcome Trust and the United Nations (International) Children s Fund/United Nations Development Programme/World Bank/World Health Organization Special Programme for Research and Training in Tropical Diseases provided funding for the study. The funding sources had no role in the study design, data collection, data analysis, data interpretation, or writing of the report. RESULTS Participant Characteristics and Follow-up The final cohort comprised 908 children whom we followed for 1201 person-years; thus, the mean duration of follow-up was 1.3 years (Figure). There were 456 (50%) male children. The mean age of the cohort was 7.5 years (range, 1 month to 16 years). Of the final cohort, 728 (80%) children had been vaccinated with BCG. All 908 children in the analytic data set had complete baseline ELISpot and TST results available. There were 381 (42%) children with positive ELISpot results and 550 (61%) with positive TST results. The prevalence of positive TST results was statistically significantly higher among BCG-vaccinated contacts than among unvaccinated contacts (462 of 728 vs. 88 of 180; P 0.001). We lost 63 children to follow-up after clinical assessment at 6 months; withdrawal rates did not differ according to ELISpot status (6% of contacts with positive ELISpot results vs. 7% with negative results). Of the cohort, 270 children completed 24 months of follow-up with 6 monthly clinical reviews and 560 children did not attend clinical follow-up after 12 months. These 560 children were followed every 6 months by a study nurse who administered a standardized telephone questionnaire. Fifteen contacts developed active tuberculosis during follow-up. Development of active tuberculosis was not related to sex (incidence in girls vs. boys, 9 of 452 (2.0%) vs. 6 of 456 (1.3%); P 0.43), although children with incident cases tended to be younger (median age, 4 years [interquartile range, 1 to 8 years] vs. 8 years [interquartile range, 4 to 11 years]; P 0.074). Incident Tuberculosis Cases Table 1 shows demographic characteristics and test results of the 15 patients with incident tuberculosis. Table 2 shows clinical manifestations at diagnosis: At enrollment, 2 December 2008 Annals of Internal Medicine Volume 149 Number

4 Article Prognostic Value of ELISpot in Tuberculosis Figure. Study flow diagram. Index patients with sputum smear positive pulmonary tuberculosis ( n = 443)* Child contacts enrolled at central study clinic ( n = 1024) Child contacts enrolled ( n = 965) Child contacts ( n = 908) Contacts excluded ( n = 59) Index patient did not have sputum smear positive pulmonary tuberculosis: 6 Blood samples not available: 4 ELISpot assay results not available: 35 Prevalent cases: 14 Never attended follow-up ( n = 57) ELISpot-positive and TST -positive: 337 (100% ) ELISpot-positive and TST -negative: 44 (100% ) ELISpot-negative and TST -positive: 213 (100% ) ELISpot-negative and TST -negative: 314 (100% ) Received IPT : 317 (94% ) Received IPT : 18 (41%) Received IPT : 140 (66% ) Received IPT : 146 (46%) Incident cases ( n = 10 [3%]) Incident cases ( n = 1 [2%]) Incident cases ( n = 2 [1%]) Incident cases ( n = 2 [1%]) 451 person-years of follow-up 86 person-years of follow-up 272 person-years of follow-up 451 person-years of follow-up IR, 22.2 (95% CI, )** IR, 11.7 (95% CI, )** IR, 7.4 (95% CI, )** IR, 5.1 (95% CI, )** Study flow diagram detailing the follow-up of 908 children with complete baseline results for enzyme-linked immunospot (ELISpot) and tuberculin skin test (TST ). IPT isoniazid preventive therapy; IR incidence rate. * Index patients and child contacts were recruited at the 7 government-run tuberculosis clinics in the Anatolian side of Istanbul. Pediatric Infectious Disease Clinic at Marmara University Hospital, Istanbul. When sputum microscopy and culture reports for all 443 index cases were obtained and checked, 4 contacts had index cases who did not have sputum smear positive results and 2 contacts had index cases whose sputum grew nontuberculosis atypical mycobacteria. Two contacts were removed because of loss of ELISpot plates, and 33 contacts were removed because of an episode of bacterial contamination of peptide pool reagents. Twenty (5%) contacts had positive ELISpot results and 37 (7%) contacts had negative ELISpot results at recruitment. Isoniazid preventive therapy was administered on the basis of age and TST result and was interpreted in accordance with Turkish Ministry of Health guidelines. Eighteen contacts who had positive ELISpot and negative TST results at recruitment and 49 contacts who had negative ELISpot and TST results at recruitment were given IPT because their TST converted. In total, 688 contacts received IPT, of whom 41 were exposed to index cases with multidrug-resistant tuberculosis. Thirteen incident cases received IPT: 6 had positive ELISpot and TST results, 4 had positive ELISpot and negative TST results (2 of whom had converted TST results), and 3 had negative ELISpot and negative TST results (2 of whom had converted TST results). None of the incident cases had negative ELISpot and positive TST results. ** Per 1000 person-years of follow-up. all patients with incident tuberculosis were asymptomatic, with a normal chest radiograph, and had no history of tuberculosis. We detected 8 cases of incident tuberculosis at scheduled follow-up visits, and the remaining 7 patients presented to clinics outside of scheduled follow-up appointments. The proportion of incident cases detected by December 2008 Annals of Internal Medicine Volume 149 Number 11

5 Prognostic Value of ELISpot in Tuberculosis Article Table 1. Demographic Characteristics and ELISpot and TST Results of Incident Tuberculosis Cases Patient Number Age at Recruitment Sex Received BCG Vaccination, n Baseline Induration, mm TST ELISpot Isoniazid Preventive Therapy Follow-up Induration, mm Spot-Forming Cells per Million PBMC* Baseline Result 6-Month Follow-up Result 66 4 mo Female Negative Positive Yes y Female Positive Positive Yes y Female Positive Positive Yes y Male Positive Positive Yes y Female Negative Negative Yes mo Female Positive Positive Yes y Female Positive Yes y Male Negative Positive Yes y Female Positive Yes y Male Positive Positive Yes y Male Positive Positive Yes y Male Positive Positive Yes y Female Positive Positive Yes y Female Negative Negative No y Male Positive Negative No BCG bacille Calmette Guérin; ELISpot enzyme-linked immunospot; PBMC peripheral blood mononuclear cells; TST tuberculin skin test. * Summated spot-forming cells from 6 pools of peptides derived from early secretory antigenic target-6 (ESAT-6) and culture filtrate protein-10 (CFP-10). The ELISpot result was positive when the test wells contained 20 spot-forming cells per million PBMC than negative control and this number was at least twice the negative control. Patient 218 had negative ESAT-6 peptide results and positive recombinant ESAT-6 results. Patients 218 and 472 had negative CFP-10 peptide results and positive recombinant CFP-10 results. Patients 215 and 573 had positive CFP-10 peptide results and negative recombinant CFP-10 results. All other peptide or antigen results were concordant. Patients 134, 135, 284, 472, 510, and 573 completed their isoniazid regimen before developing active tuberculosis. Patients 134, 135, and 138 and patients 215 and 218 lived in the same household and were exposed to the same index case. Patients 134, 135, and 138 were exposed to the same index case with tuberculosis and were resistant to isoniazid, rifampicin, ethambutol, and streptomycin. Patient 134 was resistant to isoniazid and rifampicin but sensitive to ethambutol and streptomycin. Patient 138 was resistant to isoniazid, rifampicin, ethambutol, and streptomycin. Patient 135 had a negative culture. positive ELISpot results and diagnosed at scheduled clinical follow-up visits (5 of 11 cases) was no greater than the corresponding proportion of ELISpot-negative incident cases diagnosed at scheduled clinical follow-up (3 of 4 cases). All 15 patients with incident tuberculosis had clinical and radiologic findings strongly suggestive of tuberculosis; 3 of the patients also had positive cultures for M. tuberculosis from clinical specimens (Table 2). All patients with incident tuberculosis responded fully to antituberculosis treatment, including those treated with second-line drugs for multidrug-resistant tuberculosis, with complete clinical and radiologic resolution. At recruitment, 11 patients with incident tuberculosis had positive ELISpot results and 12 had positive TST results. Ten patients with incident tuberculosis had both positive ELISpot and TST results, 1 had a positive ELISpot result and a negative TST result, 2 had negative ELISpot results and positive TST results, and 2 had negative ELISpot and TST results (Table 1). Two of 4 patients with incident tuberculosis who initially had negative ELISpot results had positive results when retested 6 months later. Although the quantitative definition of ELISpot conversion is evolving because of the natural biological variability around the threshold for a positive result, both converted results were strongly positive (376 and 140 SFCs per million peripheral blood mononuclear cells, respectively). Of the 3 patients who had negative TST results, all were retested 2 to 4 months later, and 1 had a converted result to 15-mm induration (Table 1). Incidence of Tuberculosis Of the 381 contacts who had positive ELISpot results, 11 progressed to active tuberculosis within 536 personyears of follow-up (incidence rate, 20.5 per 1000 personyears [95% CI, 10.2 to 36.7 per 1000 person-years]). Of the 550 contacts with positive TST results, 12 progressed to tuberculosis in 722 person-years (incidence rate, 16.6 per 1000 person-years [CI, 8.6 to 29.0 per 1000 personyears]) (Table 3). Thus, ELISpot detected a similar number of incident cases of tuberculosis from a smaller group of positive test results. Moreover, a significantly higher proportion of contacts had positive TST results than positive ELISpot results (550 of 908 vs. 381 of 908, respectively; P 0.001), which is consistent with the higher specificity of ELISpot. Children with positive ELISpot results were 3.4 times more likely than those with negative results to develop active disease (P 0.036) (Table 3). Children with positive TST results were 2.7 times more likely than those with negative results to develop disease, but this difference was not significant (P 0.131) (Table 3). We prescribed 688 children isoniazid preventive therapy. Because we administered isoniazid in part on the basis of TST results, interpreted in accordance with the Turkish 2 December 2008 Annals of Internal Medicine Volume 149 Number

6 Article Prognostic Value of ELISpot in Tuberculosis Table 2. Clinical Manifestations of All Incident Tuberculosis Cases Patient Number Culture Sample Culture Result Symptoms Suggestive of Tuberculosis* Physical Signs Consistent with Tuberculosis Chest Radiography Supportive of Tuberculosis Chest CT Supportive of Tuberculosis Type of Disease Time from Recruitment to Diagnosis, mo 66 Gastric Negative Yes No Yes Yes Pulmonary 5 Yes 134 Gastric Positive No Yes Yes Yes Pulmonary 28 Yes 135 Gastric Negative Yes Yes Yes Yes Pulmonary 28 Yes 138 Skin lesion Positive Yes Yes Yes Yes Miliary 18 Yes 174 Gastric Negative No No Yes Yes Pulmonary 6 Yes 176 Gastric Negative Yes Yes Yes Yes Miliary 3 Yes 215 Gastric Negative Yes No Yes Yes Pulmonary 5 Yes 218 Gastric Negative Yes No Yes Yes Pulmonary 17 Yes 284 Gastric Negative Yes No Yes Pulmonary 7 Yes 290 Gastric Negative Yes No Yes Yes Pulmonary 7 Yes 472 BAL Positive Yes No Yes Yes Pulmonary 14 Yes 510 Gastric Negative Yes Yes No Yes Pulmonary 3 Yes 573 BAL Negative Yes Yes Yes Yes Pulmonary 12 Yes 608 Gastric Negative Yes No Yes Yes Pulmonary 24 Yes 874 Gastric Negative Yes No Yes Yes Pulmonary 18 Yes Clinical and Radiographic Response to Therapy BAL bronchoalveolar lavage; CT computed tomography. * Symptoms included cough for more than 7 days, chest pain, wheezing, pyrexia for more than 7 days, night sweats, malaise, and weight loss. Physical signs included chest crackles, decrease in chest sounds, rales, erythematous nodular skin abscess, prolonged chest sounds, and wheezing. Abnormal chest radiography findings consistent with active tuberculosis included lung consolidation, cavitation, mediastinal enlargement, linear fibrotic opacities, paracardiac infiltration, parahilar infiltration, hilar enlargement, fibronodular infiltration, blunting of the phrenic sinus, and pleural effusion. Abnormal chest CT findings consistent with active tuberculosis included lung consolidation, mediastinal lymphadenopathy, cavitation, calcified nodules, miliary nodules, characteristic lung infiltration, and pleural effusion. Patient 284 did not have CT. Ministry of Health guidelines, and because agreement between TST and ELISpot results was high ( 0.65), a higher proportion of contacts with positive ELISpot results received isoniazid therapy than did those with negative results (353 of 381 (93%) vs. 335 of 527 (64%); P 0.001). Thus, administration of isoniazid preventive therapy is partially related to the test result and could confound the incidence rate. Adjustment for treatment mildly increased the relative incidence rates between positive and negative test results (Table 3). Because a high proportion (71%) of contacts exposed to index cases infected with M. tuberculosis susceptible to isoniazid received preventive therapy, including 10 of the 15 cases of incident tuberculosis, we had limited data on contacts who had not received preventive therapy to estimate the incidence rate. In these untreated contacts, 54 had positive ELISpot results, of whom 4 progressed to active tuberculosis within 93 person-years of follow-up (incidence rate, 43.0 per 1000 person-years [CI, 11.7 to per 1000 person-years]). Contacts who had positive ELISpot results were statistically significantly more likely than those with negative results to develop active tuberculosis (incidence rate ratio, 10.6 [CI, 1.19 to 95.2]; P 0.034). Of the 83 contacts who had positive TST results, 3 progressed to tuberculosis within 114 person-years of follow-up (incidence rate, 26.2 per 1000 person-years [CI, 5.4 to 76.6 per 1000 person-years]). Of the 333 contacts who had positive ELISpot results to recombinant ESAT-6/CFP-10 antigen, 11 progressed to active tuberculosis within 462 person-years of follow-up (incidence rate, 23.8 per 1000 person-years [CI, 11.9 to 42.6 per 1000 person-years]). The incidence rate in contacts who had positive results to purified protein derivative was similar to that in contacts who had positive results to the nontuberculosis control antigen streptokinase streptodornase, and children who had positive ELISpot results to these antigens were no more likely than those who had negative results to develop active disease (Table 3). Of the 337 contacts who had positive results for both tests, 10 progressed to tuberculosis within 451 person-years of follow-up (incidence rate, 22.2 per 1000 person-years [CI, 10.6 to 40.8 per 1000 person-years]) (Figure). The low number of cases of incident tuberculosis with discordant test results (positive ELISpot and negative TST results [1 case from 44 contacts] and negative ELISpot and positive TST results [2 cases from 213 contacts]) resulted in poor estimates of incidence rates with very wide CIs for these test combinations. Magnitude of Baseline ELISpot Results and Incidence of Tuberculosis Within ELISpot-positive contacts, the size of the baseline interferon- response to peptides did not significantly differ between contacts who developed tuberculosis and those who remained well (median SFCs per million peripheral blood mononuclear cells, 176 [interquartile range, 88 to 1332] vs. 182 [interquartile range, 70 to 534]; P 0.27). This was also true for responses to recombinant antigen (median SFCs per million peripheral blood mononuclear cells, 246 [interquartile range, 200 to 918] vs. 264 [interquartile range, 124 to 526]; P 0.183) December 2008 Annals of Internal Medicine Volume 149 Number 11

7 Prognostic Value of ELISpot in Tuberculosis Article DISCUSSION We found that child tuberculosis contacts with positive ELISpot results had a statistically significant 3- to 4-fold increased risk for progression to active tuberculosis compared with contacts with negative ELISpot results. The incidence rate of tuberculosis in contacts who had positive ELISpot results was similar to that of contacts who had positive TST results, but ELISpot detected a similar number of incident cases from fewer contacts with positive test results. Enzyme-linked immunospot is a quantitative assay; thus, interpretation of ELISpot results is not restricted to a binary read-out. The size of the baseline interferon- response did not statistically significantly differ between incident cases and contacts who did not develop tuberculosis, suggesting that although a positive ELISpot result is prognostic of progression to tuberculosis, the magnitude of this response does not further refine the risk for progression. This is in contrast to the TST, in which the size of the tuberculin reaction correlates with the risk for subsequent progression to tuberculosis (1 7). None of the incident cases had a history of tuberculosis, and tuberculosis was excluded at recruitment by the absence of symptoms and normal radiologic and clinical examinations. Thus, a positive ELISpot result at recruitment reflected asymptomatic M. tuberculosis infection and not active tuberculosis or relapse of tuberculosis. We evaluated all household members clinically and with chest radiography, and we treated any secondary cases of tuberculosis promptly at the time of recruitment of child contacts. Hence, reexposure of child contacts in the household would have been minimal, although we cannot exclude reinfection from exposure in the community after recruitment. Very few prospective studies have assessed the relationship between IGRA results and clinical outcome, and to our knowledge, our study is the first to focus on a key high-risk group. In a small study of 24 household contacts of sputum smear positive pulmonary tuberculosis in Ethiopia, 6 of the 7 incident cases of tuberculosis were positive to recombinant ESAT-6 in a 5-day interferon- research ELISA compared with 3 of 17 contacts who did not develop active tuberculosis (40). In a more recent household contact study in Germany, a significantly higher proportion of untreated household contacts with positive ELISA results progressed to active tuberculosis than contacts who had positive TST results when using the 5-mm threshold (P 0.003) (29). However, this proportion was not significantly greater than the proportion of contacts who had positive TST results and progressed to tuberculosis when using a 10-mm threshold (P 0.1). Hill and coworkers (41) identified 26 incident cases during 2 years of followup of household contacts and, although contacts did not receive isoniazid preventive therapy, neither TST (using a 10-mm threshold) nor ELISpot (using a substantially higher threshold for positive results than that used in our current and previous studies and in the commercial kit) was prognostic of subsequent progression to tuberculosis. The reason for the lack of prognostic power of both TST and ELISpot in the African study is unclear but may relate Table 3. Incidence Rates of Tuberculosis and Incidence Rate Ratios among Child Contacts, by ELISpot and TST Results at Recruitment Test ELISpot Result TST Result ESAT-6/CFP-10 PPD SKSD* Positive Negative Positive Negative Positive Negative Positive Negative Total results, n Incident tuberculosis cases, n Time at risk, person-years Unadjusted analysis Incidence rate per 1000 person-years (95% CI) 20.5 ( ) 6.0 ( ) 12.6 ( ) 12.3 ( ) 13.2 ( ) 11.3 ( ) 16.6 ( ) 6.3 ( ) Incidence rate ratio 3.41 ( ) 1.03 ( ) 1.16 ( ) 2.65 ( ) (95% CI) P value Adjusted analysis Incidence rate ratio 3.86 ( ) 1.09 ( ) 1.14 ( ) 3.28 ( ) (95% CI) P value CFP-10 culture filtrate protein-10; ELISpot enzyme-linked immunospot; ESAT-6 early secretory antigenic target-6; PPD purified protein derivative; SKSD streptokinase streptodornase; TST tuberculin skin test. * A nontuberculosis control antigen. Response to TST was scored positive if the induration diameter was 5 mm, in accordance with the American Thoracic Society and Centers for Disease Control and Prevention guidelines. Adjusted for isoniazid preventive therapy administered to contacts exposed to drug-sensitive tuberculosis. 2 December 2008 Annals of Internal Medicine Volume 149 Number

8 Article Prognostic Value of ELISpot in Tuberculosis to the highly endemic setting, in which community transmission outside households may cause a substantial proportion of tuberculosis cases, even in household contacts (42). Our study has several limitations. Only 3 of the incident cases in this study were culture-positive; however, clinical specimens from children are difficult to obtain and are usually culture-negative (43, 44). For the culturenegative cases, the clinical and radiologic findings strongly suggested active tuberculosis, and all responded well to antituberculosis treatment. The number of incident cases in our study was small and the CIs around the point estimates were wide, but the risk for subsequent tuberculosis in the recent child contact who had positive ELISpot results was statistically significant. Because a high proportion of contacts and incident cases received isoniazid preventive therapy, we had very limited data on contacts who had not received preventive therapy. As a result, our estimates of incidence rates in untreated contacts who had positive ELISpot and TST results were less robust but did suggest a stronger prognostic value of each test result in this group. Given the ethical imperative for prompt isoniazid preventive therapy in young children exposed to tuberculosis, it may not be possible to robustly assess whether IGRAs are more strongly prognostic of subsequent tuberculosis than TST in the absence of preventive treatment. Despite administration of isoniazid, 13 children with incident cases developed tuberculosis, 3 of whom were exposed to isoniazid-resistant M. tuberculosis. Reasons for progression to disease in the remaining 10 children include the fact that isoniazid preventive therapy for 6 months is only 60% effective (31, 32), and adherence to the therapy may not have been complete in all children, even though parents or guardians reported full adherence of their children when they returned to collect more tablets every 2 months. Although there may theoretically have been potential bias in terms of greater failure to complete the isoniazid regimen among those who had negative TST results (interpreted according to guidelines from the Turkish Ministry of Health) and among those who received BCG vaccination, all parents were informed that a negative TST result did not rule out infection and that BCG vaccination did not adequately protect against disease. Our finding that ELISpot is prognostic of active tuberculosis holds regardless of the extent of adherence to isoniazid preventive therapy in the study population. A positive interferon- ELISpot result is a useful and valid marker of latent tuberculosis infection because it predicts the subsequent development of active tuberculosis, which suggests that contacts with a diagnosis of latent tuberculosis infection on the basis of ELISpot results could benefit from preventive therapy. Although the risk for progression to active tuberculosis in contacts with a positive ELISpot result was similar to that in contacts with a positive TST result, if contacts who had positive test results were targeted with preventive treatment, statistically significantly more contacts who had positive TST results than those who had positive ELISpot results would need to be treated to prevent a similar number of incident cases. Thus, ELISpot testing could allow more focused targeting of preventive therapy to fewer contacts. Moreover, contacts awareness of the risk for progression may improve adherence to the preventive treatment regimen (45). Confirming whether the prognostic power of ELISpot is significantly higher than that of TST will require larger studies identifying more cases of clinical disease. There is potential to further improve the prognostic value of T-cell biomarkers by measuring interferon- responses to additional novel M. tuberculosis specific antigens (46) or by measuring additional cytokines or chemokines (47 49). Although our study is generalizable because it was conducted in unselected child household tuberculosis contacts in community-based contact investigations, similar studies are now required to validate and quantify the prognostic value of ELISpot in other high-risk groups, including persons coinfected with HIV (50). From Marmara University School of Medicine, Istanbul, Turkey; Tuberculosis Immunology Group, National Heart and Lung Institute, Imperial College London, London, United Kingdom; and University of Birmingham, Edgbaston, Birmingham, United Kingdom. Note: Drs. Bakir and Millington contributed equally to this work. Acknowledgment: The authors thank the children who took part in the study and their parents. They also acknowledge the crucial support of the Istanbul Association for the Fight Against Tuberculosis and the physicians and nurses of the 7 government-run tuberculosis clinics in the Anatolian side of Istanbul. Grant Support: By the Wellcome Trust and the United Nations (International) Children s Fund/United Nations Development Programme/ World Bank/World Health Organization Special Programme for Research and Training in Tropical Diseases, a UK Department of Health Senior Fellowship in Evidence Synthesis (Dr. Deeks), and the Sir Halley Stewart Trust (Dr. Dosanjh s PhD studentship). Dr. Lalvani is a Wellcome Senior Research Fellow in Clinical Science, and Dr. Millington was a Wellcome Trust Prize PhD student. Both are members of the Wellcome Trust-funded Centre for Respiratory Infection, Imperial College London. Potential Financial Conflicts of Interest: Consultancies: A. Lalvani (Oxford Immunotec Ltd. [nonexecutive director from 2003 to 2007]). Stock ownership or options (other than mutual funds): A. Lalvani (Oxford Immunotec Ltd.), University of Oxford (Oxford Immunotec Ltd.). Patents received: A. Lalvani (T-cell based diagnosis of tuberculosis infection.) Patents pending: K.A. Millington (T-cell based diagnosis of tuberculosis infection.), D.P.S. Dosanjh (T-cell based diagnosis of tuberculosis infection.), A. Lalvani (T-cell based diagnosis of tuberculosis infection.) Reproducible Research Statement: Study protocol: Not available. Statistical code and data set: Available to academic investigators from Dr. Lalvani ( , a.lalvani@imperial.ac.uk) after agreement is made by written request. Requests for Single Reprints: Ajit Lalvani, DM, Tuberculosis Immunology Group, Department of Respiratory Medicine, National Heart and Lung Institute, Imperial College London, Norfolk Place, London W2 1PG, United Kingdom; , a.lalvani@imperial.ac.uk December 2008 Annals of Internal Medicine Volume 149 Number 11

9 Prognostic Value of ELISpot in Tuberculosis Article Current author addresses and author contributions are available at References 1. Ferebee SH, Mount FW. Tuberculosis morbidity in a controlled trial of the prophylactic use of isoniazid among household contacts. Am Rev Respir Dis. 1962;85: [PMID: ] 2. Comstock GW, Livesay VT, Woolpert SF. The prognosis of a positive tuberculin reaction in childhood and adolescence. Am J Epidemiol. 1974;99: [PMID: ] 3. Fine PE, Sterne JA, Pönnighaus JM, Rees RJ. Delayed-type hypersensitivity, mycobacterial vaccines and protective immunity. Lancet. 1994;344: [PMID: ] 4. BCG and vole bacillus vaccines in the prevention of tuberculosis in adolescence and early adult life. Bull World Health Organ. 1972;46: [PMID: ] 5. Fine PE, Bruce J, Ponnighaus JM, Nkhosa P, Harawa A, Vynnycky E. Tuberculin sensitivity: conversions and reversions in a rural African population. Int J Tuberc Lung Dis. 1999;3: [PMID: ] 6. Horsburgh CR Jr. Priorities for the treatment of latent tuberculosis infection in the United States. N Engl J Med. 2004;350: [PMID: ] 7. Menzies D. Interpretation of repeated tuberculin tests. Boosting, conversion, and reversion. Am J Respir Crit Care Med. 1999;159: [PMID: ] 8. Huebner RE, Schein MF, Bass JB Jr. The tuberculin skin test. Clin Infect Dis. 1993;17: [PMID: ] 9. Starke JR. Interferon-gamma release assays for diagnosis of tuberculosis infection in children. Pediatr Infect Dis J. 2006;25: [PMID: ] 10. Richeldi L. An update on the diagnosis of tuberculosis infection. Am J Respir Crit Care Med. 2006;174: [PMID: ] 11. Barnes PF. Diagnosing latent tuberculosis infection: turning glitter to gold [Editorial]. Am J Respir Crit Care Med. 2004;170:5-6. [PMID: ] 12. Menzies D, Pai M, Comstock G. Meta-analysis: new tests for the diagnosis of latent tuberculosis infection: areas of uncertainty and recommendations for research. Ann Intern Med. 2007;146: [PMID: ] 13. Lalvani A, Millington KA. T cell-based diagnosis of childhood tuberculosis infection. Curr Opin Infect Dis. 2007;20: [PMID: ] 14. Schluger NW. The diagnosis of tuberculosis: what s old, what s new. Semin Respir Infect. 2003;18: [PMID: ] 15. Pai M, Zwerling A, Menzies D. Systematic review: T-cell-based assays for the diagnosis of latent tuberculosis infection: an update. Ann Intern Med. 2008;149: [PMID: ] 16. Pai M, Riley LW, Colford JM Jr. Interferon-gamma assays in the immunodiagnosis of tuberculosis: a systematic review. Lancet Infect Dis. 2004;4: [PMID: ] 17. Soysal A, Millington KA, Bakir M, Dosanjh D, Aslan Y, Deeks JJ, et al. Effect of BCG vaccination on risk of Mycobacterium tuberculosis infection in children with household tuberculosis contact: a prospective community-based study. Lancet. 2005;366: [PMID: ] 18. Shams H, Weis SE, Klucar P, Lalvani A, Moonan PK, Pogoda JM, et al. Enzyme-linked immunospot and tuberculin skin testing to detect latent tuberculosis infection. Am J Respir Crit Care Med. 2005;172: [PMID: ] 19. Zellweger JP, Zellweger A, Ansermet S, de Senarclens B, Wrighton-Smith P. Contact tracing using a new T-cell-based test: better correlation with tuberculosis exposure than the tuberculin skin test. Int J Tuberc Lung Dis. 2005;9: [PMID: ] 20. Richeldi L, Ewer K, Losi M, Bergamini BM, Roversi P, Deeks J, et al. T cell-based tracking of multidrug resistant tuberculosis infection after brief exposure. Am J Respir Crit Care Med. 2004;170: [PMID: ] 21. Ewer K, Deeks J, Alvarez L, Bryant G, Waller S, Andersen P, 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: [PMID: ] 22. Lalvani A, Pathan AA, Durkan H, Wilkinson KA, Whelan A, Deeks JJ, et al. Enhanced contact tracing and spatial tracking of Mycobacterium tuberculosis infection by enumeration of antigen-specific T cells. Lancet. 2001;357: [PMID: ] 23. Division of Tuberculosis Elimination, National Center for HIV, STD, and TB Prevention, Centers for Disease Control and Prevention (CDC). Guidelines for using the QuantiFERON-TB Gold test for detecting Mycobacterium tuberculosis infection, United States. MMWR Recomm Rep. 2005;54: [PMID: ] 24. National Collaborating Centre for Chronic Conditions. Tuberculosis: Clinical Diagnosis and Management of Tuberculosis, and Measures for Its Prevention and Control. London: Royal Coll of Physicians; World Health Organization. Global Tuberculosis Control: Surveillance, Planning, Financing. Geneva: World Health Organization; WHO publication WHO/HTM/TB/ Sterling TR, Haas DW. Transmission of Mycobacterium tuberculosis from health care workers. N Engl J Med. 2006;355: [PMID: ] 27. Andersen P, Doherty TM, Pai M, Weldingh K. The prognosis of latent tuberculosis: can disease be predicted? Trends Mol Med. 2007;13: [PMID: ] 28. Pai M, Dheda K, Cunningham J, Scano F, O Brien R. T-cell assays for the diagnosis of latent tuberculosis infection: moving the research agenda forward. Lancet Infect Dis. 2007;7: [PMID: ] 29. Diel R, Loddenkemper R, Meywald-Walter K, Niemann S, Nienhaus A. Predictive value of a whole blood IFN-gamma assay for the development of active tuberculosis disease after recent infection with Mycobacterium tuberculosis. Am J Respir Crit Care Med. 2008;177: [PMID: ] 30. Stout JE, Menzies D. Predicting tuberculosis: does the IGRA tell the tale? [Editorial]. Am J Respir Crit Care Med. 2008;177: [PMID: ] 31. Marais BJ, Gie RP, Schaaf HS, Beyers N, Donald PR, Starke JR. Childhood pulmonary tuberculosis: old wisdom and new challenges. Am J Respir Crit Care Med. 2006;173: [PMID: ] 32. Hsu KH. Thirty years after isoniazid. Its impact on tuberculosis in children and adolescents. JAMA. 1984;251: [PMID: ] 33. Türkiye Cumhuriyeti Sağlık Bakanlığı A. Türkiye de tüberkülozun kontrolü için başvuru kitabı. Ankara, Turkey: Rekmay; Accessed at on 23 September Lalvani A, Pathan AA, McShane H, Wilkinson RJ, Latif M, Conlon CP, et al. Rapid detection of Mycobacterium tuberculosis infection by enumeration of antigen-specific T cells. Am J Respir Crit Care Med. 2001;163: [PMID: ] 35. Liebeschuetz S, Bamber S, Ewer K, Deeks J, Pathan AA, Lalvani A. Diagnosis of tuberculosis in South African children with a T-cell-based assay: a prospective cohort study. Lancet. 2004;364: [PMID: ] 36. Pathan AA, Wilkinson KA, Klenerman P, McShane H, Davidson RN, Pasvol G, et al. Direct ex vivo analysis of antigen-specific IFN-gamma-secreting CD4 T cells in Mycobacterium tuberculosis-infected individuals: associations with clinical disease state and effect of treatment. J Immunol. 2001;167: [PMID: ] 37. Chapman AL, Munkanta M, Wilkinson KA, Pathan AA, Ewer K, Ayles H, et al. Rapid detection of active and latent tuberculosis infection in HIV-positive individuals by enumeration of Mycobacterium tuberculosis-specific T cells. AIDS. 2002;16: [PMID: ] 38. Lalvani A, Nagvenkar P, Udwadia Z, Pathan AA, Wilkinson KA, Shastri JS, et al. Enumeration of T cells specific for RD1-encoded antigens suggests a high prevalence of latent Mycobacterium tuberculosis infection in healthy urban Indians. J Infect Dis. 2001;183: [PMID: ] 39. Targeted tuberculin testing and treatment of latent tuberculosis infection. This official statement of the American Thoracic Society was adopted by the ATS Board of Directors, July This is a Joint Statement of the American Thoracic Society (ATS) and the Centers for Disease Control and Prevention (CDC). This statement was endorsed by the Council of the Infectious Diseases Society of America. (IDSA), September 1999, and the sections of this statement. Am J Respir Crit Care Med. 2000;161:S [PMID: ] 40. Doherty TM, Demissie A, Olobo J, Wolday D, Britton S, Eguale T, et al. Immune responses to the Mycobacterium tuberculosis-specific antigen ESAT-6 signal subclinical infection among contacts of tuberculosis patients. J Clin Microbiol. 2002;40: [PMID: ] 41. Hill PC, Jackson-Sillah DJ, Fox A, Brookes RH, de Jong BC, Lugos MD, et al. Incidence of tuberculosis and the predictive value of ELISPOT and Mantoux tests in Gambian case contacts. PLoS ONE. 2008;3:e1379. [PMID: ] 42. Verver S, Warren RM, Munch Z, Richardson M, van der Spuy GD, Borgdorff MW, et al. Proportion of tuberculosis transmission that takes place in households in a high-incidence area. Lancet. 2004;363: [PMID: 2 December 2008 Annals of Internal Medicine Volume 149 Number