- Top of page
- PATIENTS AND METHODS
- AUTHOR CONTRIBUTIONS
Despite the extensive implementation of BCG vaccination for newborn babies and for 7–10-year-old school children without a characteristic BCG scar in Taiwan, tuberculosis (TB) was prevalent in a notification rate of 74.6 per 100,000 populations in 2002 (1). The increased prevalence of active TB, which has been reported in patients with rheumatoid arthritis (RA), is thought to be associated with disease-related disturbances in immune function as well as treatment with immunosuppressive agents (2, 3). Recently, treatment with tumor necrosis factor α (TNFα) antagonists has been found to be associated with an increased TB risk (4–6). As noted, active TB in patients with RA receiving TNFα antagonists mainly appears to be due to reactivation of latent TB infection (LTBI) (7). Adalimumab is the first fully human anti-TNF monoclonal antibody for RA treatment (8). Prior to implementing routine TB screening in adalimumab RA clinical trials, the rate of developing TB disease increased (9). Guidelines have recommended that screening for TB exposure should be carried out before the initiation of TNFα antagonists and appropriate prophylaxis should be initiated before starting anti-TNF treatment if evidence of LTBI exists (10, 11). Recent studies demonstrated that isoniazid prophylaxis (INHP) could effectively prevent the reactivation of TB (12, 13).
Tuberculin skin test (TST) with an intradermal injection of purified protein derivative (PPD) from Mycobacterium tuberculosis culture filtrate has been widely used as a diagnostic tool for LTBI (14). However, the TST has some drawbacks, including variability in test application and reading, and low specificity as PPD presenting in nontuberculous mycobacteria (NTM) and in BCG strains (15, 16). The TST-positive rate (induration diameter ≥10 mm) was 47% in a group of patients ages 20–59 years and up to 64% in a group with household contact of TB in Taiwan. In addition, the annual incidence of disease caused by NTM was 21.5 per 100,000 patients in 2002 (17, 18). A recent diagnostic method, QuantiFERON-TB gold (QFT-G; Cellestis Limited, Carnegie, Victoria, Australia) is an in vitro test for M tuberculosis infection that measures interferon-γ (IFNγ) secreted by T cells stimulated with mixtures of synthetic peptides including the early secretary antigenic target-6 (ESAT-6) and culture filtrate protein 10 (CFP-10) that are specific to M tuberculosis (16, 19–21). The recent Centers for Disease Control and Prevention (CDC) guidelines suggest that QFT-G can replace TST under all circumstances in which the TST is currently used, including periodic testing (19, 22). Recent studies demonstrated that the IFNγ-based assay appeared promising for TB screening in patients with RA prior to anti-TNFα therapy (13, 23) and for serial testing of TB in health care workers (24). There are limited data on QFT-G performance in patients with RA as an immune-compromised population. To our knowledge, there are no reports on serial testing of TSTs and QFT-G assays in patients with RA receiving anti-TNFα therapy.
The aim of the present study was to investigate the performance of QFT-G assays combined with TSTs in periodic testing for TB in patients with RA treated with adalimumab in Taiwan. The results of TSTs and QFT-G assays and QFT-G conversions or reversions were evaluated. We also evaluated the diagnostic accuracy of QFT-G and TST in patients with RA who developed active TB disease after anti-TNFα therapy.
- Top of page
- PATIENTS AND METHODS
- AUTHOR CONTRIBUTIONS
RA patients often have compromised immune function as a result of underlying diseases, increased age, comorbidities, and the use of immunosuppressive agents (2, 3, 31); an increased risk of active TB disease has been reported in those treated with TNFα antagonists (4–6). In the present study, 4 (9.3%) of 43 RA patients developed active TB disease during 2-year adalimumab therapy. Similar to previous reports showing a high prevalence of extrapulmonary TB in RA patients receiving anti-TNF therapy (5, 6), all of our RA patients who developed active TB disease had extrapulmonary involvement. The high proportion (100%) of extrapulmonary involvement found in this study suggested that reactivation of LTBI may be one cause of active TB disease in RA patients receiving adalimumab therapy (32). Considering the longer interval (a range of 10–18 months) between starting adalimumab therapy and the onset of active TB disease, however, a recent exposure to TB infection is also a probable cause of TB in our patients, who were immunocompromised and lived in Taiwan, which has a high prevalence of TB. Additionally, all 8 patients who had positive TST results at baseline screening and completed INHP did not develop active TB disease. Our results were consistent with previous reports indicating the efficacy of INHP in preventing reactivation of LTBI (12, 13, 33).
Our data showing 81.4% negative TST results at baseline screening in patients with active RA were in accordance with the results reported by Ponce de Leon et al (70.6%) (34). Use of immunosuppressive agents and immune dysfunction related to RA may be the causes for the reported negative TST results in our study as well as other previous reports (34, 35). Additionally, a recent study demonstrated a significant reduction in the reactivity against M tuberculosis in RA patients during adalimumab therapy (36). Our results support previous reports claiming that the clinical utility of TST is limited in detecting LTBI in patients receiving immunosuppressive therapy (34, 35).
An interesting finding of our study was the high rate (37.0%) of TST conversion observed among 27 patients who completed 12 months of therapy with adalimumab. The TST conversion might have been caused by NTM cross-reaction, delayed booster effect, and true recent TB infection. Although NTM infection might have caused the TST conversion, the influence was less than that of TB because the annual incidence (per 100,000 patients) of diseases caused by NTM was lower than TB in Taiwan (18), and the absolute impact of NTM on TST was very low (0.1–2.3%), even in populations with a high prevalence of NTM (37). In a recent report by Hatemi et al, the booster phenomenon was observed in 29% of patients with RA (38). Because the 2-step TST was not performed on any patient at baseline screening in the present study, the delayed booster response of repeated TST might be the probable cause of TST conversion in a predominantly BCG-vaccinated population (39–42), especially in patients who were given BCG after infancy (37). Moreover, recent studies have also demonstrated that TNFα inhibitors do not suppress the tuberculin skin testing, or affect the skin conversion rate (38, 43). Besides the NTM cross-reactivity or delayed booster effect, we think the other skin conversions might be due to true recent TB infection. Additionally, application technique and interreader variability in TST may have biased the results of this study. At the time of our investigation, no available guidelines recommended periodic TSTs for RA patients receiving TNFα inhibitors. Furthermore, therapeutic strategies of prophylactic therapy for patients with TST conversions are only recommended for children younger than 12 years, not for adults, by Taiwan's CDC. Although INHP was not administered to our patients with TST conversions, we closely monitored these patients and thoroughly assessed them to detect active TB disease early. We also learned from this study that standard 2-step TSTs should be performed in RA patients at baseline screening before starting TNFα inhibitors. This is to avoid mislabeling the true TST conversion as a booster phenomenon during periodic PPD testing. In addition, periodic TSTs should be considered in the followup for RA patients receiving adalimumab therapy, and any patient with TST conversion should receive prophylactic therapy for TB in accordance with current guidelines (44).
Of our 10 patients with TST conversion, only 2 patients had positive QFT-G results: one developed active TB disease and the other who did not develop active TB had a weak IFNγ response. None of the other 8 patients with TST conversion developed active TB disease. In agreement with previous studies reporting an increased risk of TB disease within 2 years of ESAT-responsive contacts (45) and those showing that in vitro blood assays with TB-specific antigens had better specificity than TSTs (23, 46, 47), our data supported the usefulness of QFT-G assays in detecting TB infection in BCG-vaccinated patients receiving adalimumab therapy. A higher specificity of QFT-G assay would help patients avoid unnecessary INHP and its associated toxic effects (48).
Using 5-mm induration diameter as the TST cutoff value in this study, the overall agreement between the TSTs and QFT-G assays was 60.0%, with a kappa coefficient of 0.215, indicating poor agreement. Our results were consistent with previous studies reporting an overall agreement of 70.2–97.6% between these tests, with kappa values of 0.25–0.79 (15, 49, 50), and with 1 study demonstrating even less agreement among BCG-vaccinated patients compared with nonvaccinated individuals (51). We speculate that the poor agreement could be ascribed to several factors, including the distinct immunologic mechanisms responsible for positive results of TSTs or QFT-G assays, the different antigens used in the 2 tests, and the presence of booster effects in TST.
Similar to the results of recent studies (13, 23, 52), 2 of our RA patients (5.7%) who received immunosuppressive agents, including adalimumab therapy, had an indeterminate QFT-G result. The inclusion of positive PHA control was allowed for the identification of RA patients who had an indeterminate QFT-G result due to their inability to evoke an in vitro immune response, which may indicate the effect of immunosuppressive agents (52, 53).
In this study, QFT-G reversion was observed in 2 RA patients who had completed INHP and in another 2 patients after the completion of anti-TB chemotherapy against their TB disease. This finding was supported by a progressive decrease in the frequency of M tuberculosis antigen-specific IFNγ-secreting T cells and the reversion to negative in the IFNγ release assay during anti-TB chemotherapy (54, 55). Our data indicate that QFT-G assays might aid in monitoring the efficacy of anti-TB therapy, with QFT-G reversion as a marker of successful treatment. However, the number of enrolled patients was small and therefore these data have to be confirmed in a larger study.
In conclusion, patients with RA receiving immunosuppressive therapy, which may lead to false-negative TST results, should undergo serial tests for LTBI and receive INHP in accordance with current guidelines. As suggested by Pratt et al (13), reliable results of QFT-G assays can prospectively help clinicians to accurately identify the group with high risk of developing TB disease and provide these patients with INHP prior to anti-TNF therapy. Although the sample size of this study was too small for us to come to definite conclusions, we still highly recommend that QFT-G assay be added as a part of the screening procedure for the detection of LTBI in patients with RA prior to the initiation of TNFα antagonist therapy.