Mycobacterium tuberculosis infection in immunocompromised hosts: A diagnostic challenge

The number of cases of tuberculosis (TB) in the United States declined by 45% between 1993 and 2006.1 This decline occurred disproportionately among United States–born persons (66%) compared to foreign-born individuals, among whom the number of cases actually increased by 5%. In solid-organ transplant recipients, the incidence of Mycobacterium tuberculosis (MTB) infection worldwide ranges between 0.35% and 15%, depending on the level of endemicity.2 This rate can be 36- to 74-fold greater than that in the general population.2 The majority of MTB cases after transplantation represent reactivation of latent tuberculosis infection (LTBI), although acquisition of primary TB infection post-transplant and transmission of MTB via the transplanted organ are infrequent possibilities. Active TB in the posttransplantation setting can pose a number of challenges: atypical presentation leading to a delay in diagnosis, therapy associated with drug interactions and toxicities, graft rejection, and formidable morbidity and mortality. This is of particular concern in both patients with end-stage liver disease awaiting liver transplantation and liver transplant recipients, given the hepatotoxicity of several anti-TB medications and their associated drug interactions with immunosuppressants. BCG, Bacille Calmette-Guérin; CFP-10, Culture Filtrate Protein-10; ESAT-6, Early Secretory Antigen Target-6; HIV, human immunodeficiency virus; IFN-γ, interferon-γ; IGRA, interferon-γ release assay; LTBI, latent tuberculosis infection; LTT, lymphocyte transformation test; MTB, Mycobacterium tuberculosis; QFT-TB, QuantiFERON TB; TB, tuberculosis; TST, tuberculin skin test. In this issue of Liver Transplantation, Holty et al.3 present a systematic review and meta-analysis of TB in liver transplant recipients. They included cases of active MTB infection following liver transplantation as well as transplant candidates or recipients who received ≥6 months of therapy for LTBI prior to transplantation. The prevalence of active TB in liver transplant recipients was 1.3%. Extrapulmonary infection was seen in 67%, multiorgan involvement was seen in 27%, and 31% died. Tuberculin skin test (TST) results were as follows: a) Among 38 cases of posttransplant active TB with known pretransplant TST status, 37% were TST-positive, 53% were negative, and 11% were anergic; among 17 of these patients who had a posttransplant TST, only 35% were positive. b) In 3 studies with 2972 patients who underwent liver transplantation, 31% of the patients had a pretransplant TST, and 13% of these were positive. None of the TST-positive patients who received isoniazid therapy for at least 6 months developed active TB, whereas 7 TST-positive patients developed active disease without such therapy (mean follow-up, 54 months). Short-term mortality in patients with active posttransplant TB was 31%, with 65% of deaths directly attributed to MTB infection. Because reactivation of LTBI represents the majority of posttransplant cases of active TB, it is prudent to screen and identify such patients before transplantation for consideration of appropriate therapy for LTBI. Until recently, TST was the only test available for LTBI diagnosis. As noted previously, TST was notoriously insensitive: 64% of patients with posttransplant active TB were either TST-negative or anergic prior to transplantation, and 65% of patients had a negative posttransplant TST in the setting of active TB.3 TST has additional limitations: variable sensitivity and specificity based on the cutoff values used, interobserver variability, the need for a return visit in 48 to 72 hours, false-positive reactions in the setting of exposure to nontuberculous mycobacteria and prior Bacille Calmette-Guérin (BCG) vaccination, cutaneous anergy with underlying immunosuppression, and the need for repeat testing if the first TST is negative (recall immunity–boosting).4 The specificity of TST in BCG-vaccinated populations has been estimated to be about 56%.5 Therefore, a recommendation of LTBI therapy for all patients with a positive TST would lead to unwarranted hepatotoxic drug exposure when the test is false-positive. This is particularly relevant in patients with end-stage liver disease or after liver transplantation.2 QuantiFERON TB (QFT-TB) Gold (Cellestis, Victoria, Australia) uses an enzyme-linked immunosorbent assay to measure production of IFN-γ by circulating T-cells in whole blood in response to MTB-specific antigens (Early Secretory Antigen Target-6 [ESAT-6] and Culture Filtrate Protein-10 [CFP-10]). QFT-TB Gold In-Tube (Cellestis) is similar to the prior assay but uses tubes precoated with antigens so that blood can be directly collected into them. It also includes a third MTB-specific antigen, TB7.7. T-SPOT.TB (Oxford Immunotec, Oxford, United Kingdom) uses an enzyme-linked immunospot method to detect individual activated T-cells that secrete IFN-γ in response to MTB-specific antigens (ESAT-6 and CFP-10). IGRAs have been touted to possess several advantages over TST: Being in vitro assays using MTB-specific antigens, they eliminate interobserver variability, the need for a return visit, and cross-reactivity with BCG and non-TB mycobacteria. IGRAs seem to be unaffected by boosting and lack issues with anergy, and their results have a rapid turnaround time. In a recent meta-analysis, the specificity of the 3 available IGRAs was 93% to 99%.5 Sensitivity was comparable between QFT-TB Gold (78%), QFT-TB Gold In-Tube (70%), and TST (77%); T-SPOT.TB appeared to have better sensitivity (90%) overall.5 There are disadvantages associated with IGRAs, however: the necessity of an equipped laboratory, careful blood handling and incubation within 12 to 16 hours after collection, unclear interpretation of indeterminate results, higher cost, a lack of defined cutoff values (in contrast to TST) based on the underlying risk of MTB infection, and no prospective data regarding future risk of active TB in those with positive IGRA test results.6 T-SPOT.TB has been shown to be cost-effective.7, 8 IGRAs are currently offered by several commercial, reference, and public health laboratories. Accessibility to a laboratory performing IGRAs can vary on a regional basis in the United States. The utility of IGRAs in immunocompromised patients is not clearly defined because of a paucity of controlled trials. In a Japanese study of immunocompromised patients (60% were BCG-vaccinated), QFT-TB Gold was positive in 78% (versus 50% for TST) with active TB.9 T-SPOT.TB was compared with TST in 2 Italian studies in which immunosuppressed hematology patients were known contacts of smear-positive MTB cases.10, 11 T-SPOT.TB was positive twice as often as TST in this population (35% versus 18%)11; TST positivity diminished with increasing immunosuppression (from 26% to 14.5%), whereas T-SPOT.TB results remained unaffected.10 QFT-TB Gold In-Tube revealed a stronger association with risk factors for LTBI (odds ratio = 24) compared to TST in immunosuppressed patients with autoimmune diseases.12 Neither corticosteroids nor disease-modifying antirheumatic drugs significantly affected response to QFT-TB Gold In-Tube; however, receipt of tumor necrosis factor-α inhibitors seemed to attenuate release of IFN-γ in these assays.12 In human immunodeficiency virus (HIV)–infected patients, the performance of T-SPOT.TB was independent of the CD4 T-cell count, in sharp contrast to TST results13, 14 Clinical trial data on the role of IGRAs in transplant candidates or recipients are noticeably sparse. A prospective study compared the results of QFT-TB Gold In-Tube with simultaneously obtained TST among 153 patients with chronic liver disease awaiting liver transplantation.15 Overall agreement between the 2 tests was 85% (24% had a positive TST, and 22% had a positive QFT-TB Gold In-Tube assay). Prior BCG vaccination (in 82% of the cohort) was not associated with a discordant result. Twelve patients (7.8%) had an indeterminate result with the QFT-TB Gold In-Tube assay; all were TST-negative. A low lymphocyte count and a high Model for End-Stage Liver Disease score were significantly associated with an indeterminate QFT-TB Gold In-Tube result.15 A recent study evaluated 48 patients awaiting liver transplantation with TST, T-SPOT.TB, and the lymphocyte transformation test (LTT) using tuberculin as a stimulus.16 Four patients were positive by T-SPOT.TB, 6 were positive by TST, and 28 were positive by LTT. All 4 patients with positive T-SPOT.TB were also positive by TST and LTT and were considered to have LTBI. Several findings from this study are worth noting: a) None of the patients with LTBI received specific therapy, and none had evidence of active TB during follow-up; 2 of 4 patients were 21 and 17 months post–liver transplantation. b) In comparison with a group of 95 healthy volunteers, the level of reaction to the mitogen-positive control in the T-SPOT.TB assay was significantly lower in patients awaiting liver transplantation. This is a potential reflection of the high incidence of anergy (or negative TST) seen during pretransplant evaluation.3 c) T-SPOT.TB results from frozen and fresh T-cells did not differ. Two isolated reports of negative IGRAs in transplant recipients have been published. In the first report, a liver transplant recipient developed pulmonary TB 2 months after transplantation.17 He had a negative TST and a negative QFT-TB Gold assay before transplantation and a negative QFT-TB Gold assay in the setting of active TB post-transplant. In the second report, 2 patients with active TB following renal transplantation tested negative with an IGRA (not specified).18 What is the role of IGRAs in various categories of immunosuppressed patients (patients with underlying malignancy, cancer chemotherapy, autoimmune diseases, receipt of tumor necrosis factor-α inhibitors, or HIV infection, candidates awaiting organ transplantation, and transplant recipients)? Are there true differences in performance among currently available IGRAs in this population? T-SPOT.TB appears to be more sensitive than QFT-TB Gold and QFT-TB Gold In-Tube.5, 19 Because T-SPOT.TB involves the counting of lymphocytes to include 250,000 cells in each well and detects IFN-γ production by individual activated T-cells, it may be more sensitive than QFT-TB Gold in patients with lymphopenia9, 10 and have fewer indeterminate results.9, 15, 19 This observation seems to be evident also in HIV-infected patients: the performance of T-SPOT.TB was independent of the CD4 T-cell count,13 and the positive response was consistently higher for T-SPOT.TB in HIV-infected patients with lower CD4 T-cell strata in comparison with QFT-TB Gold.14 Will the performance of both TST and an IGRA increase detection rates for LTBI in the immunosuppressed? If so, should the tests be conducted concurrently to minimize TST boosting the IGRA response?20 As with TST, should there be separate defined cutoffs for IGRAs in immunocompetent and immunocompromised hosts based on the risk of development of active TB? Are IGRAs useful for the diagnosis of active TB in the immunocompromised host? In anecdotal patients with active TB, IGRAs were negative in the setting of organ transplantation,17, 18 HIV infection,18 malnourishment,18 and other etiologies for immune compromise.9 With augmented immunosuppression in the immediate posttransplant period, it is not surprising to have negative T-cell–based assays in the presence of active TB. Whether T-cell reactivity to antigens in TST or IGRAs will return with the passage of time remains to be seen. In immunocompetent cohorts with active TB, the sensitivity of QFT-TB Gold ranges from 64%21 to 86%22; among those with extrapulmonary TB, 6 (86%) of 7 patients had a negative QFT-TB Gold assay.21 TST is similarly insensitive in patients with disseminated and miliary TB.23 With up to 67% of liver transplant recipients presenting with extrapulmonary TB (27% with multiorgan involvement),3 we cannot solely rely upon IGRAs or TST to establish a diagnosis of active TB; results should be interpreted in conjunction with other clinical and diagnostic findings. What should our diagnostic strategy for LTBI in solid-organ transplant candidates involve? One approach would be to subject every transplant candidate to TST, perform an IGRA if the TST is positive, and treat only those with a positive TST and a positive IGRA. This would minimize drug exposure in patients with a false-positive TST and would be cost-effective.8 With up to 64% of post–liver transplant active TB patients demonstrating negative pretransplant TST results,3 the logical strategy might be to perform both TST and an IGRA concurrently during pretransplant evaluation. Those with a positive IGRA should be considered for LTBI treatment (regardless of the TST result). If the TST is positive and the IGRA is negative, treatment should not be offered (unless there is convincing evidence of MTB infection such as recent TST conversion, prior history of LTBI or active TB, occupational exposure, or a chest radiograph consistent with old TB). Despite everything that has been said and done, it is important to remind ourselves that there is no available gold standard for LTBI diagnosis against which TST and IGRAs can be evaluated. All studies comparing TST with IGRAs have used indirect evidence such as microbiologically confirmed TB or predetermined risk factors for LTBI (for sensitivity analysis) and healthy, low-risk subjects without known exposure to TB (for specificity analysis). Prospective, randomized trials assessing the usefulness of IGRAs in various populations of immunosuppressed hosts, including transplant candidates and recipients, are urgently needed. Algorithms incorporating TST and each of the available IGRAs with predefined cutoff values for a positive test in normal and immunocompromised hosts might provide us with an accurate strategy to maximize LTBI diagnosis and therapy and prevent reactivation TB in this vulnerable population.