TNF-alpha antagonist therapy modify the tuberculin skin test response

DOI 10.1007/s00296-010-1424-3 ORIGINAL ARTICLE TNF-alpha antagonist therapy modify the tuberculin skin test response Tulin Cagatay Zeki Kilicaslan Penbe Cagatay Munevver Mertsoylu Ziya Gulbaran Reyhan Yildiz Leyla Pur Sevil Kamali Ahmet Gul Received: 18 November 2009 / Accepted: 10 March 2010 Ó Springer-Verlag 2010 Abstract Tumour necrosis factor-alpha (TNF-a) antagonist drugs have been associated with increased risk of tuberculosis (TB). Tuberculin skin test (TST) is the most frequently used tool for identification of latent TB infection. We herein aimed to analyse the effect of TNF-a antagonists on the TST responses in a prospective study. The study group consisted of 182 patients (99 female, 83 male) who received TNF-a antagonists for various rheumatic disorders. All T. Cagatay (&) Z. Kilicaslan M. Mertsoylu Z. Gulbaran R. Yildiz L. Pur Istanbul Faculty of Medicine, Department of Pulmonary Diseases, Istanbul University, Capa, 34390 Istanbul, Turkey e-mail: tulcagat@istanbul.edu.tr Z. Kilicaslan e-mail: zkaslan@istanbul.edu.tr M. Mertsoylu e-mail: munevvermertsoylu@yahoo.com Z. Gulbaran e-mail: ziyagulbaran@yahoo.com R. Yildiz e-mail: drreyhanyildiz@yahoo.com L. Pur e-mail: leylapur@yahoo.com P. Cagatay Cerrahpasa Faculty of Medicine, Department of Biostatistics, Istanbul University, Istanbul, Turkey e-mail: penbecag@istanbul.edu.tr S. Kamali A. Gul Division of Rheumatology, Department of Internal Medicine, Istanbul Faculty of Medicine, Istanbul University, Istanbul, Turkey e-mail: kamalis@istanbul.edu.tr A. Gul e-mail: agul@istanbul.edu.tr patients were evaluated with TST along with other parameters on the day of referral and on the 12th month visit. For those patients with a response of \5 mm induration at the initial evaluation, the TST was repeated to observe the booster effect. Out of 182 patients, 87 patients (48%) had a negative (0 4 mm) and 95 (52%) had a positive (C5 mm) TST response at initial evaluation. The TST responses were converted from negative at initial visit to positive at 1-year repeat in 26 (30%) patients. A significant increase was observed in the diameters of TST that were repeated on the first year of TNF-a antagonist treatment (9.15 ± 0.55) compared to their initial diameters (6.60 ± 0.51) (P \ 0.001). Increased TST responses in patients receiving TNF-a antagonists may be associated with the restoration of suppressed immune reactivity against TB antigens with the decreased disease activity. The meaning of TST conversion in the definition of latent TB infection and the need for chemoprophylaxis in these patients remains to be answered by further studies. Keywords Tuberculin skin test (TST) Tumour necrosis factor (TNF)-a antagonist drugs Latent tuberculosis infection Rheumatoid arthritis Ankylosing spondylitis Introduction Tuberculosis (TB) is among the most frequent causes of death from infection in humans, accounting for around 1.6 million deaths annually [1]. Reactivation of a latent Mycobacterium tuberculosis infection can cause an important morbidity and mortality, especially among immunocompromised individuals [2]. Tumour necrosis factor-alpha (TNF-a) antagonist drugs (infliximab,

etanercept and adalimumab) are being frequently used in the treatment of various inflammatory rheumatic or other multisystem disorders, such as rheumatoid arthritis, ankylosing spondylitis and inflammatory bowel diseases; and an increased risk of TB and other infectious diseases has been documented in patients receiving TNF-a antagonist treatments. Therefore, all patients who will be treated with these agents have been screened for TB infection, and identification of the latent TB infection has again become an important topic of investigation [3, 4]. Tuberculin skin test (TST) is still the most frequently used and inexpensive test for identification of latent TB infection. TST is a classical delayed-type hypersensitivity (DTH) response to the intradermal injection of tuberculin antigen, and it represents a cutaneous T cell-mediated immune response. However, the accuracy and reliability of TST are significantly affected by immunosuppressive therapy and/or the activity of the underlying inflammatory disorders [5, 6]. At present, there is no standard method to diagnose latent TB infection reliably in patients receiving immunosuppressant drugs or having an immunocompromised condition. Most of the guidelines accept a 5-mm TST response as an indication of latent infection after the exclusion of active disease and recommend the treatment of latent TB infection with isoniazid or alternatives [7, 8]. Blockade of TNF-a action by these drugs can affect cellular immune responses and granuloma formation [9]. Though there are accumulating data about the association of these agents with an increased risk for TB infection, little is known about their effect on TST response [10]. Therefore, we herein aimed to analyse the effect of TNF-a antagonists on the TST responses in a prospective study. Materials and methods Patients The study group comprised those 182 patients (99 women, 83 men) who were followed-up at least for a year among a pool of 481 patients who were referred to the Department of Pulmonary Diseases for the evaluation of TB infection before starting a TNF-a antagonist agent. The underlying pathologies were rheumatoid arthritis in 82 patients (45.1%), ankylosing spondylitis in 51 (28%), psoriatic arthritis and other disorders in the remaining 49 (26.9%) patients. The Ethics Committee of Istanbul Faculty of Medicine approved the study protocol, and a written informed consent was obtained from all patients. All patients were evaluated with history, physical examination, chest radiograph and TST on the day of referral. Further investigations such as thorax CT, direct smear and cultures for mycobacteriae when a suspicious lesion was detected. The TST was done to the left forearm of the patients using at five Tuberculosis Control Units of tuberculin, and the response was evaluated as the diameter of induration at 72 h of injection. For those patients with a response of \5-mm induration, the TST was repeated 7 10 days later to observe the booster effect. Patients with a TST response of C5-mm induration and having no findings of active TB disease or a history of close contact with a patient with TB were given 5 mg/kg of isoniazid (maximum 300 mg/day) chemoprophylaxis for 9 months according to the local guideline for the treatment of latent infection. Patients receiving TNF-a antagonists were monitored with 2-monthly visits, and a repeat TST and chest radiograph were obtained from all patients as well as a detailed history of a new TB contact and clinical symptoms during the 12th month visit. Statistical analysis Statistical analyses were conducted using Unistat 5.1 software. All numerical values are reported as mean ± SE. The TST response variability as changes in induration diameters were compared with a non-parametric Wilcoxon test. The parameters which may potentially affect the TST responses were tested using a logistic regression analysis. P values \0.05 were considered as significant. Results Demographic and clinical characteristics of the patients are shown in Table 1, and the mean diameters of the first and second TST responses in all patients are shown in Table 2. None of the patients developed active TB disease during the follow-up. Out of 182 patients, 87 patients (47.8%) had a negative (0 4 mm) and 95 (52.2%) had a positive (C5 mm) TST response at the initial evaluation. A total of 156 patients including all patients with a negative TST response and patients with an abnormal chest radiograph consistent with past TB or a history of close contact were given isoniazid chemoprophylaxis for 9 months. After 1 year exposure to TNF-a antagonist agents, 61 out of 182 patients (33.5%) had a negative and 121 (66.5%) had positive TST responses (Table 2). In a subgroup analysis of patients with rheumatoid arthritis, 48 out of 82 patients (58.5%) had a negative TST response at the initial visit, and the TST responses were converted to positive in 12 of them (25%) after 1-year treatment with TNF-a antagonist drugs (Table 3). In 51 patients with ankylosing spondylitis, 17 (33.3%) patients had a negative TST response, and the TST responses were converted to positive

Table 1 Characteristics of the patients receiving TNF-a antagonist drugs for 1 year n % Mean ± SE Age 182 40.45 ± 1.09 Sex Female 99 54 Male 83 46 Smoking habit Yes 102 55 No 80 45 Cigarette Packages/year 82 13.63 ± 1.53 Rheumatic disease Rheumatoid arthritis 82 45 Ankylosing spondylitis 51 28 Psoriatic arthritis and others 48 27 Disease duration (years) (n = 182) 11.17 ± 0.51 Other drug use duration (months) 26.11 ± 0.89 (n = 182) Tuberculosis contact Yes 20 10.9 No 162 89.1 History of TB Regular treatment Yes 7 3.8 No 175 96.2 BCG Yes 160 88 No 22 12 Use of isoniazid Yes 156 85.7 No 26 14.3 Hepatotoxicity Yes 17 9.3 No 165 90.7 Chest X-ray Lesion (I) Yes 97 53.3 No 85 46.7 in 7 (41.1%) of them (Table 3). In total, the TST responses were converted from negative at initial visit to positive at 1-year repeat in 26 (29.8%) patients, and the TST responses stayed negative in 61 patients (70.2%) after 1-year exposure to TNF-a antagonists. The mean TST response of patients with rheumatoid arthritis was found to be smaller than those of patients with ankylosing spondylitis, but the difference between those patients was not statistically different (Table 3). We observed a significant increase in the induration diameters of TST responses that were repeated on the first Table 2 Tuberculin skin test (TST) results of the patients who received TNF-a antagonists for 1 year year of TNF-a antagonist treatment (9.15 ± 0.55) compared to their initial diameters (6.60 ± 0.51) (P \ 0.001, Table 2). We found that the majority of patients with a 0-mm induration at their first TST (n = 81) had also a negative result in their second TSTs (n = 54). When those patients with a negative TST in the last screen were excluded from the analysis, the remaining patients second TST response diameters were again found to be significantly increased compared to their first diameters (Table 2). The increase of TST diameter at first year visit compared to their initial values was significant in all subgroups of patients (Table 3). Logistic regression analysis revealed no effect of age, sex, smoking, underlying inflammatory condition, type and duration of immunosuppressive treatments, past TB history, positive family history or a close contact, BCG vaccination or isoniazid chemoprophylaxis when we grouped the patients as those with an increased TST response and others. Discussion n Mean ± SE P TST TST-0 182 6.60 ± 0.51 \0.001 TST-1 year 182 9.15 ± 0.55 Excluding TST 0 mm (1 year) TST-0 128 9.38 ± 0.58 \0.001 TST-1 year 128 13.01 ± 0.47 TST-0: Induration diameter of the tuberculin prior to TNF antagonist treatment TST-1 year: Induration diameter of the tuberculin after 1 year use of TNF antagonist treatment In this study, we prospectively investigated the effect of TNF-a antagonist drugs on the diameter of TST induration in a group of patients with inflammatory rheumatic conditions, and we observed a significant increase in the diameter of TST following a 1-year exposure to TNF-a antagonist drugs. Reactivation of latent M. tuberculosis infection is a major complication of TNF-a antagonist treatment, but its mechanism is not fully understood. Identification of those patients with a latent TB infection is very important before starting a TNF-a antagonist treatment. TST is still the most commonly used and inexpensive tool for the investigation of latent TB infection, although interferon-gamma (IFN-c) release assays (IGRAs) have started to be used in recent years for providing more information in the diagnosis of

Table 3 Comparison of the tuberculin skin test (TST) results in subgroups of patients with different rheumatic disorders Rheumatic disorder TST-0 (%) TST-1 year (%) Rheumatoid arthritis (n = 82) TST \ 5 mm 48 (58.5%) 36 (43.9%) TST C 5 mm 34 (41.5%) 46 (56.1%) Induration diameter (mm) (mean ± SE) 5.20 ± 0.75 7.15 ± 0.82 (P \ 0.01) Ankylosing spondylitis (n = 51) TST \ 5 mm 17 (33.3%) 10 (19.6%) TST C 5 mm 34 (66.7%) 41 (80.4%) Induration diameter (mm) (mean ± SE) 8.59 ± 1.01 11.65 ± 0.97 (P \ 0.001) Psoriatic arthritis and others (n = 49) TST \ 5 mm 22 (45%) 15 (30.6%) TST C 5 mm 27 (55%) 34 (69.4%) Induration diameter (mm) (mean ± SE) 6.88 ± 0.94 9.90 ± 1.04 (P \ 0.001) latent infection especially in those patients with an immunocompromised condition [4, 11 13]. According to the recent recommendations, a chemoprophylaxis with isoniazid for up to 9 months is advised for patients with a tuberculin test diameter of C5 mm and who will receive TNF-a antagonist drugs. TNF-a antagonist drugs are mostly used in patients with an autoimmune or autoinflammatory disorder who have not responded to standard antiinflammatory or immunosuppressant drugs. The delayed type hypersensitivity is considered to be suppressed in immunocompromised patients. Therefore, it may not be possible to obtain a reliable response to the skin tests [2]. An anergic reaction to delayed type hypersensitivity tests such as TST in patients with chronic arthritis may result from disease activity, immunosuppressive treatment, or both. Some investigators showed that there could be an improvement in delayed type hypersensitivity response in patients receiving disease modifying anti-rheumatic drugs, but the actual effect of TNF-a antagonists on TST has not yet been studied in detail [14]. In a recent study, Joven et al. [10] reported an increase in diameter of TST in a smaller group of patients receiving TNF-a antagonists. Our findings support Joven et al. s observations in a larger series of patients, which also allowed to do a subgroup analysis in patients with rheumatoid arthritis and ankylosing spondylitis. We found that a significant increase in the diameter of TST can be expected in patients who received anti-tnf-a agents for 1 year as well as a conversion from negative TST to positive in a considerable amount of patients (30%). None of the patients described a contact with a patient with TB and none developed a clinical, radiological or bacteriological sign in favour of an active TB infection during 1-year follow-up. At the same time, none of the patients received INH treatment for tuberculosis prophylaxy. These observations suggest that increase in the diameter of TST response may be associated with TNF-a antagonist treatment, and evaluation of latent TB infection should be repeated at certain intervals. The main proinflammatory cytokines in TB pathogenesis are TNF-a that are produced by activated macrophages and T cells, and IFN-c produced by CD4, CD8 and cd T cells as well as natural killer cells [15, 16]. Both of these cytokines participate in granuloma formation, and TNF-a synergizes with IFN-c in its tuberculostatic activity [17 19]. This can explain the higher incidence of TB seen in patients receiving anti-tnfa treatment. However, steady TNF-a production during chronic TB is not without cost to the host, resulting in cachexia and weight loss in advanced disease [20]. It is known that serum TNF-a levels are increased in patients with rheumatoid arthritis [21, 22], but there are different findings on the effects of TNF-a antagonists on serum TNF-a levels. There is no detailed information about the effect of TNF-a antagonists on the TST response in rheumatoid arthritis patients. It was reported that delayed type hypersensitivity response was decreased in rheumatoid arthritis patients examined by the multitest technique, and TST response was improved following treatment with gold salts [14]. Therefore, it can be suggested that elevated expression of proinflammatory cytokines including TNF-a can suppress the TST response, and the expected TST response can be restored after the anti TNF-a treatment. On the other hand, improvement of TST response may also be associated with the tapering of other immunosuppressive drugs including corticosteroids during the TNF-a antagonist treatment. In this study, we observed that TNF-a antagonist treatment does not decrease the delayed type hypersensitivity reaction, but on the contrary, an increased TST response can be seen in patients receiving these agents, which may suggest the restoration of suppressed immune reactivity against TB antigens with the decreased disease activity. No data could be reached for the effect of INH prophylaxis on TST diameter but theoretically it can be said

that INH has no effect on TST or even has a decreasing effect. Furthermore the effect of INH is out of question in this group of patients with negative TST and who are converted to a positive TST after a 1 year s TNF treatment as they did not receive INH at all during this period. Our second hypotheses may not have an impact upon the explanation of the conversion to a positive TST due to the diminution of the disease activation, leading to the vanishing of the immune suppression because in all the patient groups who are either immune suppressive or non-immune suppressive (RA, ankylosing spondylitis, psoriasis and others), all showed a conversion from negative to positive. Therefore, this conversion and TST diameter increase may be attributed to TNF-a antagonist treatment. We conclude that during the TNF-a antagonist treatment, a repeated TST may be beneficial since a conversion from negative to positive in as much as 30% of the patients at the 1-year follow-up of anti-tnf-a therapy can be observed. The meaning of this TST conversion in the definition of latent TB infection and the need for chemoprophylaxis in these patients remains to be answered by further studies. Conflict of interest statement References None. 1. World Health Organization (2007) WHO report 2007: global tuberculosis control: surveillance, planning, financing. WHO, Geneva 2. Sester U, Junker H, Junker H et al (2006) Improved efficiency in detecting cellular immunity towards M. tuberculosis in patients receiving immunosuppressive drug therapy. Nephrol Dial Transplant 21:3258 3268 3. Saliu OY, Sofer C, Stein DS et al (2006) Tumor-necrosis-factor blockers: differential effects on mycobacterial immunity. J Infect Dis 194:486 492 4. Dukes HC (2004) Infectious complications of treatment with biologic agents: infectious arthritis and immune dysfunction. Curr Opin Rheumatol 16:393 398 5. Vukmanovic-Stejic M, Reed JR, Lacy KE et al (2006) Mantoux test as a model for a secondary immune response in humans. Immunol Lett 107:93 101 6. Helliwell MG, Panayi GS, Unger A (1984) Delayed cutaneous hypersensitivity in rheumatoid arthritis: the influence of nutrition and drug therapy. Clin Rheumatol 3:39 45 7. Furst DE, Breedveld FC, Kalden JR et al (2003) Updated consensus statement on biological agents for treatment of rheumatoid arthritis and other immune mediated inflammatory diseases. Ann Rheum Dis 62(Suppl 2):2 9 8. Thoracic Society Standards of Care Committee (2005) BTS recommendations for assessing risk and for managing Mycobacterium tuberculosis infection and disease in patients due to start anti-tnf-a treatment. Thorax 60:800 805 9. Hamdi H, Mariette X, Godot V et al (2006) The RATIO (Recherche sur Anti-TNF et Infections Opportunistes) Study Group. Inhibition of anti-tuberculosis T-lymphocyte function with tumour necrosis factor antagonists. Arthritis Res Ther 8:R114 10. Joven BE, Almodóvar R, Galindo M et al (2006) Does antitumour necrosis factor a treatment modify the tuberculin TST response? Ann Rheum Dis 65:699 11. Hill PC, Jeffries DJ, Brookes RH et al (2007) Using ELISPOT to expose false positive skin test conversion in tuberculosis contacts. PLoSONE 2:183 12. Kunst H (2006) Diagnosis of latent tuberculosis infection: the potential role of new technologies. Respir Med 100:2098 2106 13. Codecasa LR, Ferrarese M, Penati V et al (2005) Comparison of tuberculin skin test and Quantiferon immunological assay for latent tuberculosis infection. Monaldi Arch Chest Dis 63:158 162 14. Smith MD, Smith A, O Donnell J et al (1989) Impaired delayed type cutaneous hypersensitivity in rheumatoid arthritis reversed by chrysotherapy. Ann Rheum Dis 48:108 113 15. Kerksiek KM, Pamer EG (1999) T cell responses to bacterial infections. Curr Opin Immunol 11:400 405 16. Ogawa T, Uchida H, Kusumoto Y et al (1991) Increase in tumor necrosis factor a and interleukin-6 secreting cells in peripheral blood mononuclear cells from subjects infected with Mycobacterium tuberculosis. Infect Immun 59:30021 30025 17. Bermudez LE, Kaplan G (1995) Recombinant cytokines for controlling mycobacterial infections. Trends Microbiol 3:22 27 18. Kindler V, Sappino AP, Grau GE et al (1989) The inducing role of tumor necrosis factor in the development of bactericidal granulomas during BCG infection. Cell 56:731 740 19. Flesch IEA, Kaufmann SHE (1990) Activation of tuberculostatic macrophage functions by interferon gamma, interleukin 4 and tumor necrosis factor. Infect Immun 58:269 271 20. Hussain R, Shiratsuchi H, Phillips M et al (2001) Opsonizing antibodies (IgG1) up-regulate monocyte proinflammatory cytokines tumour necrosis factor-a (TNF-a) and IL-6 but not antiinflammatory cytokine IL-10 in mycobacterial antigen-stimulated monocytes implications for pathogenesis. Clin Exp Immunol :210 218 21. Pachot H, Arnaud B, Marrote H et al (2007) Increased tumor necrosis factor-a mrna expression in whole blood from patients with rheumatoid arthritis: reduction after infliximab treatment does not predict response. J Rheumatol 34:2158 2161 22. Marotte H, Maslinski W, Miossec P (2005) Circulating tumour necrosis factor-a bioactivity in rheumatoid arthritis patients treated with infliximab: link to clinical response. Arthritis Res Ther 7:149 155