Rationale: In the absence of active tuberculosis, a positive tuberculin skin test (TST) or interferon-γ release assay (IGRA) result defines latent infection with Mycobacterium tuberculosis, although test results may vary depending on immunodeficiency.
Objectives: This study compared the performance of TST and IGRAs in five different groups of immunocompromised patients, and evaluated their ability to identify those at risk for development of tuberculosis.
Methods: Immunocompromised patients with HIV infection, chronic renal failure, rheumatoid arthritis, solid-organ or stem-cell transplantation, and healthy control subjects were evaluated head-to-head by the TST, QuantiFERON-TB-Gold in-tube test (ELISA), and T-SPOT.TB test (enzyme-linked immunospot) at 17 centers in 11 European countries. Development of tuberculosis was assessed during follow-up.
Measurements and Main Results: Frequencies of positive test results varied from 8.7 to 15.9% in HIV infection (n = 768), 25.3 to 30.6% in chronic renal failure (n = 270), 25.0% to 37.2% in rheumatoid arthritis (n = 199), 9.0 to 20.0% in solid-organ transplant recipients (n = 197), 0% to 5.8% in stem-cell transplant recipients (n = 103), and 11.2 to 15.2% in immunocompetent control subjects (n = 211). Eleven patients (10 with HIV infection and one solid-organ transplant recipient) developed tuberculosis during a median follow-up of 1.8 (interquartile range, 0.2–3.0) years. Six of the 11 patients had a negative or indeterminate test result in all three tests at the time of screening. Tuberculosis incidence was generally low, but higher in HIV-infected individuals with a positive TST (3.25 cases per 100 person-years) than with a positive ELISA (1.31 cases per 100 person-years) or enzyme-linked immunospot result (1.78 cases per 100 person-years). No cases of tuberculosis occurred in patients who received preventive chemotherapy.
Conclusions: Among immunocompromised patients evaluated in this study, progression toward tuberculosis was highest in HIV-infected individuals and was poorly predicted by TST or IGRAs.
Clinical trial registered with www.clinicaltrials.gov (NCT 00707317).
Patients with immunodeficiencies are particularly vulnerable to progression from latent infection with Mycobacterium tuberculosis to active disease. Tuberculin skin test and interferon-γ release assays are recommended for risk assessment for the future development of tuberculosis, but knowledge of their performance in immunocompromised patients is limited.
Tuberculin skin test and interferon-γ release assay responses vary substantially among different groups of patients with immunodeficiencies and are poor predictors for the development of tuberculosis in immunocompromised patients.
Tuberculosis is a leading cause of morbidity and mortality worldwide, and is the 10th cause of all recorded deaths (1). In the absence of an effective vaccine, tuberculosis prevention relies on early case finding, infection control measures, and preventive chemotherapy of individuals latently infected with Mycobacterium tuberculosis (2). In clinical practice, latent infection with M. tuberculosis (LTBI) is defined by the presence of an adaptive immune response against antigens of M. tuberculosis in individuals without evidence of active tuberculosis, and is either determined in vivo by the tuberculin skin test (TST) or ex vivo by interferon-γ release assays (IGRAs) (3). At present, two IGRAs are commercially available: the QuantiFERON-TB-Gold in-tube test (Qiagen, Hilden, Germany), an ELISA; and the T-SPOT.TB test (Oxford Immunotec, Abingdon, UK), an enzyme-linked immunospot assay (ELISPOT). A positive TST or IGRA result is assumed to identify individuals with the highest risk of progression to tuberculosis (4). In most cases in patients from low-incidence countries, tuberculosis is caused by reactivation of LTBI, and could have been avoided by preventive chemotherapy following targeted TST or IGRA testing (5). However, apart from individuals from defined tuberculosis risk groups, immunodiagnosis for LTBI and preventive chemotherapy is not generally recommended, because the positive predictive value of the TST and IGRAs for progression toward tuberculosis in the general population is very low (4, 6), especially in the absence of prior M. tuberculosis exposure, and preventive chemotherapy is not efficacious outside of risk groups (7).
The risk for tuberculosis is directly related to exposure with M. tuberculosis and patients with immunodeficiencies are particularly vulnerable to progression from LTBI to active disease (8, 9). These include HIV-infected individuals, transplant recipients, patients with chronic renal failure, or patients undergoing tumor necrosis factor antagonist therapies. In general, the risk for tuberculosis in immunocompromised persons is influenced by the underlying immunologic mechanisms and degree of immunodeficiency, and by the duration of and temporal relationship with previous M. tuberculosis exposure (8, 9). Because timing of exposure outside of contact tracing is generally unknown, a positive immunodiagnostic test result is considered as a proxy of prior M. tuberculosis exposure. In addition, the results of immunodiagnostic testing may depend on the nature and extent of the immunodeficiency and may not equally predict the risk for tuberculosis among different groups of immunocompromised patients. Because a positive test in the clinical setting is a finding on which a clinician needs to act, increased knowledge of performance and limitations of the currently available immunodiagnostic tests and their ability to predict tuberculosis in immunocompromised patients is needed.
In previous studies, the evaluation of immunodiagnostic tests for LTBI was mostly restricted to individual patients groups (10–13) or individual assays (14–19). Our study included five different groups of immunocompromised patients from 17 centers in 11 European countries, and tested all three immunodiagnostic tests in parallel. The two main objectives were to characterize the performance of three immunodiagnostic tests for LTBI in a patient population with various etiologies of immunodeficiency, and to assess the risk of developing tuberculosis in patients with a positive immunodiagnostic test at the time of screening.
Study participants were recruited from 17 European healthcare facilities; 18 years of age and older; and diagnosed with HIV infection, chronic renal failure, rheumatoid arthritis, solid-organ transplantation, or stem-cell transplantation. All local ethics committees approved the study and written informed consent was obtained from all individuals. Patients were recruited from June 1, 2008 to May 31, 2011 in a consecutive manner as part of their routine care, and followed up for the development of tuberculosis. We additionally included adult control patients at the same facilities from noninfectious disease departments who did not have an immunocompromizing clinical condition, and had a low risk for M. tuberculosis exposure.
The study included a cross-sectional part where data on demographic and clinical parameters, and data related to the risk of M. tuberculosis exposure were obtained through a structured questionnaire. In addition, all participants had the three separate tests for LTBI administered (3). Experimental and clinical data were recorded electronically and transmitted to the coordinating center, where they were assessed for inconsistencies or missing entries. In the prospective part of the study, follow-up information on the occurrence of tuberculosis after initial testing was actively sought and collected by personal patient contact by the treating physician.
The TST according to the Mantoux technique (20) and the two IGRAs (the ELISPOT, T-SPOT.TB; and the ELISA, QuantiFERON-TB-Gold in-tube) were performed as described in the online supplement. The laboratory technicians were fully masked to the M. tuberculosis exposure status of the participants. The TST could be performed a maximum of 30 days before the IGRAs.
M. tuberculosis exposure was defined by a reported history of either exposure to M. tuberculosis, active tuberculosis, tuberculosis treatment, LTBI, or chemoprophylaxis for LTBI, or being a resident in a high tuberculosis incidence country for at least 1 year. For stratified analyses of subgroups, CD4 cell count was dichotomized at 200 cells per microliter (21). Duration of dialysis was dichotomized at 5 years, and the time after solid-organ transplantation at 1 year. In addition, a drug score was determined as specified in the Methods section of the online supplement to quantify the level of immunosuppression in transplant recipients. In patients with rheumatoid arthritis, severity was classified according to the disease activity score (DAS; group I, <3.3; group II, 3.3–5.1; group III, >5.1) (22).
The agreement between the proportions of positive tests was assessed by Cochran-Q test for matched samples (23). Agreement between tests was assessed by the Kappa statistics (24). The association between test results and M. tuberculosis exposure was assessed by logistic regression in which confounding was deemed present if the relationship between test result and exposure changed by more than 10% after inclusion of an exploratory variable. The risk of tuberculosis by initial test results was assessed using the Kaplan-Meier estimate in survival analysis assessing patient follow-up data at the earliest date of the last clinical assessment, diagnosis of tuberculosis, death, or end of follow-up (either 1, 2, or 5 yr after testing). Only patients with at least 30 days of follow-up were included in the analysis to consider very early tuberculosis as a screening failure rather than a valid case of incident tuberculosis.
A total of 1,537 patients with immunocompromizing medical conditions (Figure 1) and 211 immunocompetent control subjects were enrolled. Causes of immunodeficiency included chronic HIV infection (n = 768), chronic renal failure (n = 270), rheumatoid arthritis (n = 199), solid-organ transplantation (n = 197; 134 renal, 41 lung, 17 liver, 4 renal-pancreas, and 1 renal-liver transplants), and stem-cell transplantation (n = 103) (Table 1). Information on the distribution of patients by country is given in Table E1 in the online supplement.
HIV | CRF | RA | SOT | SCT | Control Subjects | |||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
n = 768 | n = 270 | n = 199 | n = 197 | n = 103 | n = 211 | |||||||
n | % | n | % | n | % | n | % | n | % | n | % | |
Female | 203 | 26.4 | 97 | 35.9 | 154 | 77.4 | 83 | 42.1 | 41 | 39.8 | 134 | 63.5 |
Age (median, IQR) | 40.8 | 33.5–48.4 | 62.5 | 49.0–73.3 | 55.7 | 45.4–63.6 | 56.5 | 46.4–63.9 | 57.1 | 44.8–64.6 | 23.3 | 20.1–43.8 |
White | 709 | 92.3 | 261 | 96.7 | 192 | 96.5 | 191 | 97.0 | 97 | 94.2 | 208 | 98.6 |
Immigrant | 101 | 13.2 | 42 | 15.6 | 14 | 7.0 | 22 | 11.2 | 12 | 11.7 | 6 | 2.8 |
Years since immigration (median, IQR) | 9.0 | 3.0–16.0 | 22.5 | 11.5–38.5 | 20.0 | 8.0–30.0 | 38.5 | 26.0–42.0 | 33.0 | 21.5–40.5 | 27.0 | 23.0–32.0 |
Mycobacterium tuberculosis exposure* | 187 | 24.3 | 66 | 24.4 | 87 | 43.7 | 66 | 33.5 | 24 | 23.3 | 0 | 0.0 |
History of exposure to M. tuberculosis | 44 | 5.7 | 28 | 10.4 | 19 | 9.6 | 31 | 15.7 | 3 | 2.9 | 0 | 0.0 |
History of active TB | 62 | 8.1 | 19 | 7.0 | 9 | 4.5 | 11 | 5.6 | 2 | 1.9 | 0 | 0.0 |
History of TB treatment | 58 | 7.6 | 13 | 4.8 | 6 | 3.0 | 6 | 3.1 | 1 | 1.0 | 0 | 0.0 |
History of LTBI | 13 | 1.7 | 1 | 0.4 | 20 | 10.1 | 31 | 15.7 | 0 | 0.0 | 0 | 0.0 |
History of LTBI chemoprophylaxis | 8 | 1.0 | 1 | 0.4 | 14 | 7.0 | 6 | 3.1 | 18 | 17.5 | 0 | 0.0 |
>1 yr in high TB–incidence country | 94 | 12.2 | 28 | 10.4 | 55 | 27.6 | 10 | 5.1 | 3 | 2.9 | 0 | 0.0 |
Valid results in all three tests† | 635 | 82.4 | 245 | 90.7 | 164 | 82.4 | 145 | 73.6 | 69 | 67.0 | 197 | 93.4 |
TST, ELISA, and ELISPOT test results were available from cross-sectional analysis of 1,495 (97.3%), 1,532 (99.7%), and 1,503 (97.8%) patients, and 211 (100%), 208 (98.6%), and 209 (99.1%) control subjects, respectively. The percentages of indeterminate and positive test results are shown in Figures 2A and 2B, respectively. The highest percentage of indeterminate results in the ELISA was observed among solid-organ and stem-cell transplant recipients (Figure 2A; 20.3 and 20.4%, respectively), whereas indeterminate results in the ELISPOT assay were most frequently observed in HIV-infected patients and stem-cell transplant recipients (11.2 and 14.6%, respectively). Indeterminate test results of IGRAs in these patient groups were related to the extent of immunosuppression (see Table E2). The percentage of indeterminate results among patients with chronic renal failure or rheumatoid arthritis was only marginally different from that observed in healthy control subjects (Figure 2A). Overall, failure to adequately react toward the positive control stimulus was the cause of indeterminate results in 85.0% of ELISA and 50.4% of ELISPOT samples, whereas excess reactivity in the negative control was the reason for indeterminate results in all other samples.
Percentages of positive test results among all patients with valid results in all three assays showed substantial differences among the groups (Figure 2B). In general, positive tests were most frequently observed in patients with chronic renal failure (25.3–30.6%) and rheumatoid arthritis (25.0–37.2%), which contrasted with results from HIV-infected patients and solid-organ transplant recipients, respectively, where only 8.7–15.9% and 9.0–20% of tests were positive. The percentage of positive tests in HIV-infected patients was lowest in patients with low CD4 T-cell counts, whereas the effect of the level of immunodeficiency in other groups was less evident (see Table E3). When comparing the results of the three tests in each group, the percentage of positive test results among HIV-infected patients and solid-organ transplant recipients was lower when using the TST as compared with ELISA and ELISPOT (P < 0.001 for HIV-infection; P = 0.002 for solid-organ transplant recipients). Among stem-cell transplant recipients, positive tests were generally less frequent and were only observed using the IGRAs. The percentage of positive TST results was significantly higher as compared with that of IGRAs in patients with rheumatoid arthritis (P = 0.003), whereas no difference between the tests was found in patients with chronic renal failure (P = 0.121) (Figure 2B).
When comparing results of the three different tests, the agreement between the two IGRAs was higher than that between either IGRA and the TST (see Table E3). Agreement between ELISA and ELISPOT was substantial in patients with chronic renal failure (K=0.65) and rheumatoid arthritis (K=0.77), and moderate in HIV-infected individuals (K=0.50), and solid-organ (K=0.58) and stem-cell transplant recipients (K=0.41).
The association between self-reported evidence of prior M. tuberculosis exposure and test result was not confounded by age, sex, or group-specific characteristics of immunosuppression for any of the three tests, with the exception of patients with rheumatoid arthritis in which the DAS score confounded this association (see Table E4). Moderate associations of positive IGRA results with exposure were found in HIV-infected patients, which reached statistical significance for the ELISPOT (odds ratio [OR], 2.0; 95% confidence interval [CI], 1.2–3.4), and near significance for the ELISA (OR, 1.6; 95% CI, 0.9–3.1). In patients after solid-organ transplantation, a positive ELISA was significantly associated with exposure (OR, 2.2; 95% CI, 1.0–4.6). A significant association of each of the three tests with exposure existed for patients with rheumatoid arthritis. After adjusting for DAS score, the OR was 2.7 (95% CI, 1.4–5.1) for TST, 4.4 (95% CI, 2.1–9.4) for ELISA, and 4.9 (95% CI, 2.4–10.2) for ELISPOT. Conversely, although the percentage of positive test results in patients with chronic renal failure and rheumatoid arthritis was similarly high (Figure 2B), a positive TST or IGRA in patients with chronic renal failure was not associated with evidence of prior M. tuberculosis exposure.
When analyzing the number of positive tests in the three assays in relation to exposure status, the percentage of cases with three negative tests decreased along a gradient of likely exposure, and was highest in individuals nonexposed to M. tuberculosis (see Table E5). Likewise, the percentage of individuals with three positive tests increased with increasing exposure and ranged from 4.9% in nonexposed to 13.4% in cases where exposure was likely highest. Of note, this also held true for cases where only the two IGRAs were positive. In contrast, the percentage of individuals with only one positive test was rather similar across the groups (see Table E5).
Of the 1,537 patients, a total of 1,464 (95.3%) had an assessment of tuberculosis on follow-up and were included in the prospective part of the study. Eighteen patients were excluded because of a follow-up of less than 30 days. The number of patients with either a positive or negative test result included in the analysis for tuberculosis incidence was 1,404 for the TST, 1,342 for the ELISA, and 1,310 for the ELISPOT. A total of 11 patients developed active tuberculosis in the median follow-up time of 1.8 years (interquartile range [IQR], 0.2–3.0). Of these, 10 patients were HIV-infected and one was a solid-organ transplant recipient (Table 2). No cases of tuberculosis were observed among patients with chronic renal failure, rheumatoid arthritis, stem-cell transplantation, or control subjects. All patients who developed tuberculosis were born in a medium- or high-prevalence country for tuberculosis and had not received preventive chemotherapy after testing. Among HIV-infected patients, viral load was detectable in all cases (median, 20,593 copies per microliter; IQR, 82.5–49,819 copies per microliter), and median CD4 T-cell counts were 302 per microliter (IQR, 196–370 per microliter) (Table 2). Six out of the 11 patients either had a negative or indeterminate test result at the time of testing and only two had positive test results in all three assays (Table 2).
Country of Birth/Actual Residence | Sex | TB (mo after Testing) | TB Type | TB Diagnosis | TST (mm) | ELISPOT (SFC) | ELISA (U/ml) | Chemoprevention | HIV CDC | HAART | HIV Load | CD4 T Cells/μl | ||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Offered | Done | |||||||||||||
HIV | Argentina/Italy | Male | 1.1 | Pulmonary | Clinical* | Positive (20) | Positive (18) | Positive (2.7) | No | No | C2 | No | 25,285 | 492 |
HIV | Bulgaria/Bulgaria | Male | 1.3 | Extrapulmonary | Culture confirmed | Positive (12) | Positive (16) | Positive (0.39) | No | No | C3 | Yes | 83,176 | 245 |
HIV | Portugal/Portugal | Male | 2.0 | Pulmonary | Culture confirmed | Positive (5) | Positive (125) | Negative (0.15) | Yes | No | A3 | No | 38,700 | 22 |
HIV | Portugal/Portugal | Female | 2.5 | Pulmonary | Clinical* | Negative (0) | Negative (0) | Negative (0) | No | No | A2 | No | 15,900 | 371 |
HIV | Portugal/Portugal | Male | 8.4 | Both | Culture confirmed | Negative (0) | Negative (1) | Indeterminate (0.13) | Yes | No | C3 | Yes | 120,000 | 50 |
HIV | Portugal/Portugal | Female | 21.8 | Both | Culture confirmed | Negative (0) | Negative (4) | Indeterminate (0) | Yes | No | A1 | No | 38,200 | 333 |
SOT | Croatia/Germany | Female | 25.2 | Extrapulmonary | PCR confirmed | Negative (0) | Negative (0) | Indeterminate (0) | No | No | NA | NA | NA | Unknown |
HIV | Portugal/Portugal | Male | 33.6 | Both | Culture confirmed | Negative (0) | Negative (1) | Negative (0.09) | No | No | A2 | No | <50 | 369 |
HIV | Eritrea/Germany | Male | 35.4 | Pulmonary | Clinical* | Negative (0) | Negative (0) | Negative (0.12) | No | No | C3 | Yes | 60 | 271 |
HIV | Argentina/Italy | Male | 52.5 | Pulmonary | Culture confirmed | Positive (10) | Indeterminate (49)† | Positive (0.7) | No | No | A2 | No | 10,071 | 263 |
HIV | Eritrea/Germany | Male | 56.7 | Pulmonary | Culture confirmed | Negative (0) | Positive (9) | Negative (0.09) | No | No | A3 | Yes | 90 | 354 |
The incidence of tuberculosis in HIV-infected patients with a positive test was generally higher as compared with a negative test (see Table E6). Two years after testing, the incidence was higher in patients with a positive TST result (3.26 cases per 100 person-years) as compared with the ELISPOT (1.80 per 100 person-years) and the ELISA (1.37 per 100 person-years) (see Table E6). Similar results were obtained after 1 and 5 years of follow-up (see Table E6).
Finally, the influence of preventive chemotherapy administered after testing on progression to tuberculosis was analyzed for all patients after 1, 2, and 5 years (Table 3). Information on preventive chemotherapy by patient group and country is given in Table E1. Among patients, 21.3% (47 of 196) of TST positives, 26.0% (56 of 215) of ELISA positives, and 24.0% (54 of 247) of ELISPOT positives received preventive chemotherapy. No case of tuberculosis occurred in these individuals. Among patients who had not received preventive chemotherapy, the incidence of tuberculosis in individuals with a negative test was expectedly low (0.15, 0.11, and 0.17 per 100 person-years after 2 yr for TST, ELISA, and ELISPOT, respectively). In contrast, the incidence of tuberculosis after a positive TST, ELISA, or ELISPOT was 1.15, 0.71, and 0.88 per 100 person-years after 2 years, respectively. Similar results were obtained after 1 and 5 years of follow-up (Table 3). Taken together, although the incidence of tuberculosis was higher in patients with positive test results, IGRAs were not superior in identifying patients at risk for developing tuberculosis when compared with the TST.
Test Result* | Prophylaxis† | N | 1 yr | 2 yr | 5 yr | |||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
PY at Risk | TB Cases | Incidence‡ | PY at Risk | TB Cases | Incidence‡ | PY at Risk | TB Cases | Incidence‡ | ||||
TST | Negative | No | 1,133 | 1084.2 | 2 | 0.18 (0.05–0.74) | 1951.5 | 3 | 0.15 (0.05–0.48) | 3076.4 | 7 | 0.23 (0.11–0.48) |
Negative | Yes | 75 | 73.0 | 0 | 0 | 110.1 | 0 | 0 | 131.2 | 0 | 0 | |
Positive | No | 149 | 144.5 | 3 | 2.08 (0.67–6.44) | 261.5 | 3 | 1.15 (0.37–3.56) | 382.4 | 4 | 1.05 (0.39–2.79) | |
Positive | Yes | 47 | 46.7 | 0 | 0 | 82.8 | 0 | 0 | 105.0 | 0 | 0 | |
ELISA | Negative | No | 1,079 | 1036.6 | 2 | 0.19 (0.05–0.77) | 1864.4 | 2 | 0.11 (0.03–0.44) | 2895.2 | 5 | 0.17 (0.07–0.41) |
Negative | Yes | 48 | 46.6 | 0 | 0 | 77.8 | 0 | 0 | 95.1 | 0 | 0 | |
Positive | No | 159 | 151.8 | 2 | 1.32 (0.33–5.27) | 282.4 | 2 | 0.71 (0.18–2.83) | 462.4 | 3 | 0.65 (0.21–2.01) | |
Positive | Yes | 56 | 55.4 | 0 | 0 | 85.6 | 0 | 0 | 100.5 | 0 | 0 | |
ELISPOT | Negative | No | 1,024 | 976.2 | 2 | 0.20 (0.05–0.82) | 1752.8 | 3 | 0.17 (0.06–0.53) | 2718.3 | 6 | 0.22 (0.10–0.49) |
Negative | Yes | 39 | 37.3 | 0 | 0 | 64.2 | 0 | 0 | 84.7 | 0 | 0 | |
Positive | No | 193 | 183.3 | 3 | 1.64 (0.53–5.07) | 342.0 | 3 | 0.88 (0.28–2.72) | 556.7 | 4 | 0.72 (0.27–1.91) | |
Positive | Yes | 54 | 53.9 | 0 | 0 | 86.3 | 0 | 0 | 103.7 | 0 | 0 |
We have evaluated currently available immunodiagnostic tests for the identification of LTBI and their value in assessing the risk for progression to active tuberculosis in a large cohort of patients with different etiologies of immunodeficiencies in Europe. As most important findings we identified substantial differences in the frequencies of positive M. tuberculosis–specific immune responses among patients with HIV infection, rheumatoid arthritis, chronic renal failure, and solid-organ or stem-cell transplants, whereas less prominent variations were observed between the TST, the QuantiFERON-TB-Gold in-tube, and the T-SPOT.TB test within these groups. Of note, among immunocompromised patients, the risk for the development of tuberculosis was clearly highest in HIV-infected individuals.
The strength of this large prospective cohort study is that all three immunodiagnostic tests were evaluated in parallel in patients from five groups of different etiologies of immunodeficiencies, which allowed for a direct comparison of results within one study. Unlike skin testing, IGRAs are less influenced by immunosuppression in HIV-infected individuals and in patients after solid-organ or stem-cell transplantation. Similarities in test performance in these two patient groups may result from the fact that both immunodeficiencies primarily affect T cells, and lower percentages of positive test results appear to be caused by either an HIV-related overall decrease in CD4 T-cell counts (11, 25, 26) or by a drug-induced inhibition of T-cell functionality (27). In contrast, higher percentages of positive test results were found in patients with chronic renal failure or rheumatoid arthritis where immunodeficiency is multifactorial and not primarily acting on T cells (28). Patients with rheumatoid arthritis were striking in that the highest percentage of positive test results were found with the TST. However, because between-test agreement was higher among IGRAs, and both IGRAs showed a strong association with prior M. tuberculosis exposure, the TST seems to be less specific and seems to measure different subpopulations of patients.
The proportion of indeterminate IGRA results was expectedly lowest in immunocompetent control subjects, whereas up to 20% of indeterminate results were found in patients. Subgroup analyses indicated that a higher percentage was associated with markers of immunodeficiency, such as CD4 T-cell counts less than 200 cells per microliter, or recent transplantation attendant with higher levels of immunosuppression with multiple drugs in solid-organ and stem-cell transplant recipients. The findings in HIV-infected patients are in line with previous reports suggesting that HIV-positive individuals with less than 200 CD4+ T cells per microliter have impaired IGRA response (10, 11, 17, 18, 26). The proportion of indeterminate results in this study was lowest among patients with rheumatoid arthritis and renal failure, which was generally in line with indeterminate result rates in other studies in these patient groups (12, 29–34). Together with the highest percentage of positive test results, this suggests that immunodiagnostic assays in these two patient groups are least affected by the underlying etiology of immunodeficiency.
Tuberculosis incidence rates found after positive tests are in line with estimates from systematic reviews of predictive value of IGRAs for incident active tuberculosis, which have predominantly included nonimmunocompromised individuals during contact tracing (6, 35). Up to now, only a few studies exist that have specifically analyzed progression in immunocompromised patients, and most studies have only used one IGRA without comparison with TST (36). Most studies did not report incidence rates, but the percentage of tuberculosis cases among untreated patients with positive IGRA results ranged from 7 to 20% in HIV-infected patients (17, 19, 37), 5.6% in patients after renal transplantation (15), and 2.6% in patients before tumor necrosis factor antagonist therapy of which only a small number had rheumatoid arthritis (38). In our multicenter European study, incident cases of tuberculosis were lower and were almost exclusively found among HIV-infected patients. Most of these patients did not receive antiretroviral therapy and those who did still had detectable levels of viral replication confirming that antiretroviral therapy alone reduces the risk for the development of tuberculosis in HIV-infected patients substantially (39, 40). Although positive test results were generally more frequently observed with IGRAs as compared with the TST, progression to tuberculosis also occurred in patients with a negative result in any of the three tests. In half of the tuberculosis cases, LTBI was not detectable by any of the tests at the time of screening and positive results in all three tests were only observed in 2 of 11 patients who subsequently developed tuberculosis. Results from the current study further show that a higher percentage of positive test results in a given group of immunocompromised patients does not indicate a substantially higher risk for progression to tuberculosis. This is illustrated by the fact that no cases of active tuberculosis were observed among patients with renal failure or patients with rheumatoid arthritis not receiving preventive chemotherapy, although these were the patient groups with the highest percentage of positive test results. In contrast, the risk of developing tuberculosis was highest for patients with HIV infection, although the frequency of positive test results was substantially lower. These findings indicate that in immunocompromised patients, none of the three tests is sufficient to assess the risk of progression to tuberculosis.
A recent study in military recruits indicated that individuals where all three tests were positive had a higher epidemiologic risk of prior infection, whereas individuals with only one positive test were suggested to be likely false-positive (41). In line with these findings, the percentage of individuals with positive test results in all three assays or with two positive IGRAs in our study was also highest in cases with highest likelihood of exposure. The fact that the percentage of patients where only one test was positive was less strikingly associated with exposure variables may be considered as a hint toward false-positive results, but this may also be influenced by variable effects of immunodeficiency on immune reactivity in vivo and in vitro.
It is unclear which test should be preferred for immunodiagnostic testing. When cross-sectionally comparing test results obtained by IGRAs and TST, IGRAs generally had a higher rate of positivity and were more strongly associated with M. tuberculosis exposure. However, results from the longitudinal part of this study may suggest superiority of the TST for risk assessment in HIV-infected patients. However, given the considerably high percentage of positive tests and the low number of tuberculosis cases on follow-up, there was no striking difference between the tests and neither IGRA nor TST was adequately able to predict those at risk. In addition, although progression rates were low in patients with negative test results, physicians caring for immunocompromised patients must be aware that a negative result with any of the currently available tests does not rule out the future risk of developing tuberculosis.
The results from this study should have implications for refining future screening policies in immunocompromised patients in countries with low tuberculosis incidence. The low progression rates among test-positive individuals emphasize the limitations of current recommendations (7, 42) to screen all patients with chronic renal failure or with rheumatoid arthritis for LTBI in the absence of additional risk factors for tuberculosis. Because at least 50% of HIV-infected patients who developed tuberculosis had negative test results at the time of screening, both the positive and negative predictive values of any of the immunodiagnostic tests were very poor in low-incidence countries. Because development of tuberculosis was almost exclusively observed in patients with ongoing HIV replication, the data are consistent with lower tuberculosis rates associated with effective antiretroviral therapy, and it may be suggested that HIV-infected patients with positive immunodiagnostic test results do not have an increased risk of developing tuberculosis when HIV-replication is suppressed to undetectable levels by antiretroviral therapy. This emphasizes that the effect of antiretroviral therapy needs to be carefully studied and addressed when future policies for tuberculosis prevention are made for HIV-infected patients in countries of low tuberculosis prevalence.
The study has some limitations, of which most are inherent to the nonrandomized observational study design. The low number of cases with tuberculosis may be confounded by the fact that patients with the highest presumed risk for the development of tuberculosis were potentially more likely to have received preventive chemotherapy. Although formal randomization of receiving preventive therapy is ethically questionable, techniques to adjust for this treatment-by-indication bias, such as inverse probability weighting (43), could not be reliably used given the small proportion of patients receiving preventive therapy and the limited information of confounding variables. In addition, as in the setting of contact tracing, comparisons of progression rates toward tuberculosis among different groups of immunocompromised patients would be least biased if patients were harmonized for presumptive exposure, and if exogenous reinfection rates were either absent or identical. However, in this observational study, the time of exposure is largely unknown, which is caused by the fact that risk assessment in immunocompromised patients in a clinical routine setting is not primarily guided by recent contact. One of Comstock´s studies in military recruits showed a high predictive value of a positive TST in the first year, whereas rates were increasing over 5 years of follow-up in those with baseline negative tests (44). In our study, we did not perform serial testing or specifically collect information on reexposure after testing. Therefore, we could not exclude that infection with M. tuberculosis occurred after screening, although incidence rates for tuberculosis were similar in individuals with HIV infection at the different time points of follow-up. Finally, this study was not powered to identify regional differences in the role of different assays to assess development of tuberculosis.
In conclusion, the performance of TST and IGRAs differs among patients with various etiologies of immunodeficiency. Immunocompromised patients at risk for developing tuberculosis are poorly identified by both TST and IGRAs. Better methods and biomarkers are urgently needed to specifically target preventive chemotherapy in immunocompromised individuals who would otherwise develop tuberculosis in the future.
Additional contributors in the TBNET study group “IGRAs in immunocompromised patients”: Alessandra Apostoli, Brescia, Italy; Michael Block, Borstel, Germany; Emanuele Borroni, M.S., Milano, Italy; Roxana S. Bumbacea, M.D., Bucharest, Romania; Luigi Codecasa, M.D., Milan, Italy; Gilda Cuzzi, P.H.N., Rome, Italy; Carmen C. Diaconu, Ph.D., Bucharest, Romania; Rumiana Drenska, M.D., Sofia, Bulgaria; Ivailo Elenkov, M.D., Sofia, Bulgaria; Jürgen Finke, M.D., Freiburg, Germany; Andrea Glaewe, R.N., Borstel, Germany; Andreas Günther, M.D., Kiel, Germany; Mara-Julia Hantsche, M.D., Lübeck, Germany; Anders Harju, Stockholm, Sweden; Gerhard Held, M.D., Homburg, Germany; Ulrich Hopt, M.D., Freiburg, Germany; Oya Kayacan, M.D., Ankara, Turkey; Winfried V. Kern, M.D., Freiburg, Germany; Lenka Krabbe, R.N., Borstel, Germany; Razvan Lungu, M.D., Bucharest, Romania; Sarah Leyking, M.D., Homburg, Germany; Lourdes Mateo, M.D., Badalona, Spain; John Mee, B.Sc., London, United Kingdom; Sonia Minguez, M.D., Badalona, Spain; Birgitta Nordmark, Stockholm, Sweden; Przemyslaw Pisarski, M.D., Freiburg, Germany; Jordi Puig, R.N., Badalona, Spain; Luca Richeldi, M.D., Modena, Italy; Jan Rupp, M.D., Lübeck, Germany; Urban Sester, M.D., Homburg, Germany; Norbert Südkamp, M.D., Freiburg, Germany; Yana Todorova, Ph.D., Sofia, Bulgaria; Cristina Tural, M.D., Badalona, Spain; Susanne Usadel, M.D., Freiburg, Germany; Gerd Walz, M.D., Freiburg, Germany; Thomas Watter, M.D., Freiburg, Germany; Heinrike Wilkens, M.D., Homburg, Germany; and Valentina Vanini, M.L.T., Rome, Italy.
1. | Lozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V, Abraham J, Adair T, Aggarwal R, Ahn SY, et al. Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 2012;380:2095–2128. |
2. | Erkens CG, Kamphorst M, Abubakar I, Bothamley GH, Chemtob D, Haas W, Migliori GB, Rieder HL, Zellweger JP, Lange C. Tuberculosis contact investigation in low prevalence countries: a European consensus. Eur Respir J 2010;36:925–949. |
3. | Mack U, Migliori GB, Sester M, Rieder HL, Ehlers S, Goletti D, Bossink A, Magdorf K, Hölscher C, Kampmann B, et al.; C. Lange; TBNET. LTBI: latent tuberculosis infection or lasting immune responses to M. tuberculosis? A TBNET consensus statement. Eur Respir J 2009;33:956–973. |
4. | Diel R, Goletti D, Ferrara G, Bothamley G, Cirillo D, Kampmann B, Lange C, Losi M, Markova R, Migliori GB, et al. Interferon-γ release assays for the diagnosis of latent Mycobacterium tuberculosis infection: a systematic review and meta-analysis. Eur Respir J 2011;37:88–99. |
5. | International Union Against Tuberculosis Committee on Prophylaxis. Efficacy of various durations of isoniazid preventive therapy for tuberculosis: five years of follow-up in the IUAT trial. Bull World Health Organ 1982;60:555–564. |
6. | Rangaka MX, Wilkinson KA, Glynn JR, Ling D, Menzies D, Mwansa-Kambafwile J, Fielding K, Wilkinson RJ, Pai M. Predictive value of interferon-γ release assays for incident active tuberculosis: a systematic review and meta-analysis. Lancet Infect Dis 2012;12:45–55. |
7. | Mazurek GH, Jereb J, Vernon A, LoBue P, Goldberg S, Castro K; IGRA Expert Committee; Centers for Disease Control and Prevention (CDC). Updated guidelines for using interferon gamma release assays to detect Mycobacterium tuberculosis infection—United States, 2010. MMWR Recomm Rep 2010;59:1–25. |
8. | Sester M, Bumbacea D, Duarte R, Lange C. Tuberculosis in the immunocompromised host. Eur Respir Monogr 2012;58:230–241. |
9. | Horsburgh CR Jr, Rubin EJ. Clinical practice. Latent tuberculosis infection in the United States. N Engl J Med 2011;364:1441–1448. |
10. | Luetkemeyer AF, Charlebois ED, Flores LL, Bangsberg DR, Deeks SG, Martin JN, Havlir DV. Comparison of an interferon-gamma release assay with tuberculin skin testing in HIV-infected individuals. Am J Respir Crit Care Med 2007;175:737–742. |
11. | Leidl L, Mayanja-Kizza H, Sotgiu G, Baseke J, Ernst M, Hirsch C, Goletti D, Toossi Z, Lange C. Relationship of immunodiagnostic assays for tuberculosis and numbers of circulating CD4+ T-cells in HIV infection. Eur Respir J 2010;35:619–626. |
12. | Triverio PA, Bridevaux PO, Roux-Lombard P, Niksic L, Rochat T, Martin PY, Saudan P, Janssens JP. Interferon-gamma release assays versus tuberculin skin testing for detection of latent tuberculosis in chronic haemodialysis patients. Nephrol Dial Transplant 2009;24:1952–1956. |
13. | Hadaya K, Bridevaux PO, Roux-Lombard P, Delort A, Saudan P, Martin PY, Janssens JP. Contribution of interferon-γ release assays (IGRAs) to the diagnosis of latent tuberculosis infection after renal transplantation. Transplantation 2013;95:1485–1490. |
14. | Bartalesi F, Vicidomini S, Goletti D, Fiorelli C, Fiori G, Melchiorre D, Tortoli E, Mantella A, Benucci M, Girardi E, et al. QuantiFERON-TB Gold and the TST are both useful for latent tuberculosis infection screening in autoimmune diseases. Eur Respir J 2009;33:586–593. |
15. | Kim SH, Lee SO, Park JB, Park IA, Park SJ, Yun SC, Jung JH, Kim YH, Kim SC, Choi SH, et al. A prospective longitudinal study evaluating the usefulness of a T-cell-based assay for latent tuberculosis infection in kidney transplant recipients. Am J Transplant 2011;11:1927–1935. |
16. | Manuel O, Humar A, Preiksaitis J, Doucette K, Shokoples S, Peleg AY, Cobos I, Kumar D. Comparison of quantiferon-TB gold with tuberculin skin test for detecting latent tuberculosis infection prior to liver transplantation. Am J Transplant 2007;7:2797–2801. |
17. | Aichelburg MC, Rieger A, Breitenecker F, Pfistershammer K, Tittes J, Eltz S, Aichelburg AC, Stingl G, Makristathis A, Kohrgruber N. Detection and prediction of active tuberculosis disease by a whole-blood interferon-gamma release assay in HIV-1-infected individuals. Clin Infect Dis 2009;48:954–962. |
18. | Brock I, Ruhwald M, Lundgren B, Westh H, Mathiesen LR, Ravn P. Latent tuberculosis in HIV positive, diagnosed by the M. tuberculosis specific interferon-gamma test. Respir Res 2006;7:56. |
19. | Kim YJ, Kim SI, Kim YR, Wie SH, Park YJ, Kang MW. Predictive value of interferon-γ ELISPOT assay in HIV 1-infected patients in an intermediate tuberculosis-endemic area. AIDS Res Hum Retroviruses 2012;28:1038–1043. |
20. | Mantoux C. L'intradermo-reaction de la tuberculine et son interpretation clinique. Presse Med 1910;2:10–13. |
21. | Centers for Disease Control (CDC). Revision of the CDC surveillance case definition for acquired immunodeficiency syndrome. Council of State and Territorial Epidemiologists; AIDS Program, Center for Infectious Diseases. MMWR Morb Mortal Wkly Rep 1987;36:1S–15S. |
22. | Wells G, Becker JC, Teng J, Dougados M, Schiff M, Smolen J, Aletaha D, van Riel PL. Validation of the 28-joint Disease Activity Score (DAS28) and European League Against Rheumatism response criteria based on C-reactive protein against disease progression in patients with rheumatoid arthritis, and comparison with the DAS28 based on erythrocyte sedimentation rate. Ann Rheum Dis 2009;68:954–960. |
23. | Cochran WG. The comparison of percentages in matched samples. Biometrika 1950;37:256–266. |
24. | Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159–174. |
25. | Elzi L, Schlegel M, Weber R, Hirschel B, Cavassini M, Schmid P, Bernasconi E, Rickenbach M, Furrer H; Swiss HIV Cohort Study. Reducing tuberculosis incidence by tuberculin skin testing, preventive treatment, and antiretroviral therapy in an area of low tuberculosis transmission. Clin Infect Dis 2007;44:94–102. |
26. | Aabye MG, Ravn P, PrayGod G, Jeremiah K, Mugomela A, Jepsen M, Faurholt D, Range N, Friis H, Changalucha J, et al. The impact of HIV infection and CD4 cell count on the performance of an interferon gamma release assay in patients with pulmonary tuberculosis. PLoS ONE 2009;4:e4220. |
27. | Sester U, Wilkens H, van Bentum K, Singh M, Sybrecht GW, Schäfers HJ, Sester M. Impaired detection of Mycobacterium tuberculosis immunity in patients using high levels of immunosuppressive drugs. Eur Respir J 2009;34:702–710. |
28. | Bumbacea D, Arend SM, Eyuboglu F, Fishman JA, Goletti D, Ison MG, Jones CE, Kampmann B, Kotton CN, Lange C, et al. The risk of tuberculosis in transplant candidates and recipients: a TBNET consensus statement. Eur Respir J 2012;40:990–1013. |
29. | Ponce de Leon D, Acevedo-Vasquez E, Alvizuri S, Gutierrez C, Cucho M, Alfaro J, Perich R, Sanchez-Torres A, Pastor C, Sanchez-Schwartz C, et al. Comparison of an interferon-gamma assay with tuberculin skin testing for detection of tuberculosis (TB) infection in patients with rheumatoid arthritis in a TB-endemic population. J Rheumatol 2008;35:776–781. |
30. | Chen DY, Shen GH, Chen YM, Chen HH, Hsieh CW, Lan JL. Biphasic emergence of active tuberculosis in rheumatoid arthritis patients receiving TNFα inhibitors: the utility of IFNγ assay. Ann Rheum Dis 2012;71:231–237. |
31. | Inanc N, Aydin SZ, Karakurt S, Atagunduz P, Yavuz S, Direskeneli H. Agreement between Quantiferon-TB gold test and tuberculin skin test in the identification of latent tuberculosis infection in patients with rheumatoid arthritis and ankylosing spondylitis. J Rheumatol 2009;36:2675–2681. |
32. | Greenberg JD, Reddy SM, Schloss SG, Kurucz OS, Bartlett SJ, Abramson SB, Bingham CO III. Comparison of an in vitro tuberculosis interferon-gamma assay with delayed-type hypersensitivity testing for detection of latent Mycobacterium tuberculosis: a pilot study in rheumatoid arthritis. J Rheumatol 2008;35:770–775. |
33. | Passalent L, Khan K, Richardson R, Wang J, Dedier H, Gardam M. Detecting latent tuberculosis infection in hemodialysis patients: a head-to-head comparison of the T-SPOT.TB test, tuberculin skin test, and an expert physician panel. Clin J Am Soc Nephrol 2007;2:68–73. |
34. | Winthrop KL, Nyendak M, Calvet H, Oh P, Lo M, Swarbrick G, Johnson C, Lewinsohn DA, Lewinsohn DM, Mazurek GH. Interferon-gamma release assays for diagnosing Mycobacterium tuberculosis infection in renal dialysis patients. Clin J Am Soc Nephrol 2008;3:1357–1363. |
35. | Diel R, Loddenkemper R, Nienhaus A. Predictive value of interferon-gamma release assays and tuberculin skin testing for predicting progression from latent TB infection to disease state: a meta-analysis. Chest 2012;142:63–75. |
36. | Chee CB, Sester M, Zhang W, Lange C. Diagnosis and treatment of latent infection with Mycobacterium tuberculosis. Respirology 2013;18:205–216. |
37. | Soborg C, Ruhwald M, Andersen PH, Ravn P. 6-year follow-up of 522 HIV-positive individuals screened for Mycobacterium tuberculosis infection in Denmark. Eur Respir J 2014;44:540–543. |
38. | Jung YJ, Lyu J, Yoo B, Lee CK, Kim YG, Yang SK, Byeon JS, Kim KJ, Ye BD, Lee KH, et al. Combined use of a TST and the T-SPOT.TB assay for latent tuberculosis infection diagnosis before anti-TNF-α treatment. Int J Tuberc Lung Dis 2012;16:1300–1306. |
39. | Lawn SD, Badri M, Wood R. Tuberculosis among HIV-infected patients receiving HAART: long term incidence and risk factors in a South African cohort. AIDS 2005;19:2109–2116. |
40. | Karo B, Haas W, Kollan C, Gunsenheimer-Bartmeyer B, Hamouda O, Fiebig L; German ClinSurv HIV Study Group. Tuberculosis among people living with HIV/AIDS in the German ClinSurv HIV Cohort: long-term incidence and risk factors. BMC Infect Dis 2014;14:148. |
41. | Mancuso JD, Mazurek GH, Tribble D, Olsen C, Aronson NE, Geiter L, Goodwin D, Keep LW. Discordance among commercially available diagnostics for latent tuberculosis infection. Am J Respir Crit Care Med 2012;185:427–434. |
42. | European Centre for Disease Prevention and Control. Use of interferon-gamma release assays in support of TB diagnosis. Stockholm: ECDC; 2011. |
43. | Hermans SM, Manabe YC, Kiragga AN, Hoepelman AI, Lange JM, van Leth F. Risk of tuberculosis after antiretroviral treatment initiation: a comparison between efavirenz and nevirapine using inverse probability weighting. Antivir Ther 2013;18:615–622. |
44. | Comstock GW, Edwards LB, Livesay VT. Tuberculosis morbidity in the U.S. Navy: its distribution and decline. Am Rev Respir Dis 1974;110:572–580. |
Present addresses: *Pulmonary Diseases Department, Uludağ University School of Medicine, Bursa, Turkey; †ICREA Infection Biology Laboratory, Department of Experimental and Health Sciences, Universitat Pompeu Fabra, Barcelona, Spain; ‡University of Rome “Tor Vergata,” Rome, Italy; and §Marmara University School of Medicine, Department of Intensive Care, Istanbul, Turkey.
|| A complete list of additional contributors in the TBNET study group “IGRAs in Immunocompromised Patients” may be found before the beginning of the References.
Author Contributions: M.S. made a substantial contribution to the conception and design of the work and to the acquisition, analysis, and interpretation of data for the work; wrote the manuscript; critically revised the manuscript for important intellectual content; and gave final approval of the current version to be published. F.v.L. made a substantial contribution to the conception and design of the work and to the interpretation of data for the work, performed statistical analysis, wrote the manuscript, critically revised the manuscript for important intellectual content, and gave final approval of the current version to be published. P.R. made a substantial contribution to the conception and design of the work and to the interpretation of data for the work, critically revised the manuscript for important intellectual content, and gave final approval of the current version to be published. C.L. made a substantial contribution to the conception and design of the work and to the acquisition, analysis, and interpretation of data for the work; wrote the manuscript; critically revised the manuscript for important intellectual content; and gave final approval of the current version to be published. All other authors made a contribution to the acquisition of the data for the work, critically revised the manuscript for important intellectual content, and gave final approval of the current version to be published. All authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
For the purpose of this study QuantiFERON Gold-in-tube tests kits and T-SPOT.TB tests kits were donated to TBNET by the respective manufacturers, Cellestis Ltd. (now Qiagen, Hilden, Germany) and Oxford Immunotec (Abingdon, UK). Both companies had no role in the idea, design, and conduction of the study and had no access to study results at any time.
Originally Published in Press as DOI: 10.1164/rccm.201405-0967OC on October 10, 2014
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
Author disclosures are available with the text of this article at www.atsjournals.org.