Challenges in Diagnosing Latent Tuberculosis Infection in Patients Treated with Tumor Necrosis Factor Antagonists

Testing methods

In Canada, LTBI testing is indicated when there is an increased risk of developing TB disease¹⁹. There are 2 main methods to diagnose LTBI: the tuberculin skin test (TST) and the newer interferon-γ release assays (IGRA). The TST performed by the Mantoux method (Table 1) is used in Canada, the United States, and the UK^19,20,21. It involves intradermal injection of purified protein derived (PPD) from Mycobacterium tuberculosis bacteria. Injection site induration within 48 to 72 hours caused by manifestation of a recall immunological response to TB antigens suggests past or present infection^19,20,21. The 2-step TST is mainly used for persons who will have routine serial screening for LTBI, such as healthcare workers¹⁹. IGRA quantifies interferon-γ (IFN-γ) production from T cells and is based on the fact that T cells already sensitized to TB antigens produce high IFN-γ levels once reexposed¹⁹.

Screening recommendations in Canada, the US, and the UK (Table 1)

Immunosuppressed persons [e.g., those with end-stage renal disease (ESRD), diabetes, human immunodeficiency virus (HIV), corticosteroid therapy, or TNF antagonist therapy] are included in the at-risk population discussed in the 2007 Canadian Tuberculosis Standards¹⁹, the 2010 Canadian Tuberculosis Committee (CTC) updated recommendations²², the 2000 US Recommendations²⁰, and the 2010 US updated guidelines for using IGRA²³. However, the UK National Institute for Health and Clinical Excellence (NICE) Guidelines (2006)²¹ do not reference immunosuppressed populations. The following all provide focused recommendations for patients taking or candidates for TNF antagonists: the Canadian Tuberculosis Standards and CTC update recommendations, the 2005 British Thoracic Society Guidelines, the 2010 European TBNET Consensus Statement, the 2007 Health Protection Agency position statement on the use of IGRA (which supplements the NICE Guidelines), and the Centers for Disease Control and Prevention (CDC)^{19,22,24,25,26,27}.

The CTC states that an IGRA test may be used if the initial TST result is negative and there is concern about LTBI in an immunocompromised person²². If 2 consecutive indeterminate IGRA results occur, clinicians should suspect anergy and are advised to rely on history, clinical features, and other laboratory results.

TNF antagonist manufacturer’s recommendations for TB/LTBI screening

Labels for TNF antagonists approved in Canada indicate all patients be evaluated for both active and latent TB before starting therapy. The etanercept label recommends evaluation according to the Canadian Tuberculosis Standards, while the labels for adalimumab, infliximab, and certolizumab indicate screening with TST. Only the adalimumab label states that a ≥ 5 mm induration size, for LTBI testing, is considered positive even for patients vaccinated with Bacille Calmette-Guérin (BCG), a consideration that is contrary to the Canadian and British TB guidelines^19,24. All labels warn of possible false-negative TST in immunocompromised or severely ill patients. In addition, labels for infliximab and certolizumab recommend taking a TB medical history.

All TNF antagonist labels recommend that if LTBI is diagnosed, treatment with anti-TB drugs should be started before initiating TNF antagonist therapy. In the ideal world, completion of LTBI treatment may be the preferred option. However, in practice this could be difficult to accomplish, so some experts suggest 4 weeks of LTBI therapy before initiating TNF antagonist treatment²⁷. All TNF antagonist labels, except for certolizumab, state that anti-TB therapy should be decided by consulting a TB expert if patients have a negative LTBI test yet still have a high risk of TB infection. All manufacturers advise that patients with active TB should not receive TNF antagonist therapy.

All TNF antagonist labels suggest appropriately monitoring for TB disease during and after TNF antagonist treatment, including patients who tested negative for LTBI. Although new infections are rare in low TB-prevalence countries such as Canada and the United States, they are of greater concern in countries with a high burden of illness²⁸. It is important to note that it is possible that medications other than TNF antagonist therapies (including emerging therapies or new indications) may have an effect on the expression of TB.

Practical strengths and weaknesses

Advantages include low cost, simplicity, and validation in large cohorts¹⁹. Testing, administration, and reading techniques can lead to variability in test results; thus, testing should be performed by trained health professionals¹⁹.

Measures of screening performance

TST can deliver false-positive and false-negative results. Because of the potential for serious isoniazid (INH) hepatotoxicity reported with treatment of LTBI in patients with RA receiving methotrexate or sulfasalazine and TNF antagonists²⁹, LTBI tests should maintain sufficient specificity to minimize false-positives and potential adverse effects of unnecessary LTBI treatment. False-positive TST can be caused by previous BCG vaccination, sensitization to environmental nontuberculous mycobacteria (NTM), or inappropriate induration cutoff thresholds³⁰. In a recent metaanalysis, BCG taken by 12,728 subjects after 1 year of age yielded 41.8 false-positive TST reactions per 100 vaccinees³¹. BCG vaccination was discontinued in Canada as TB incidence rates began to fall. Children in Quebec, Newfoundland and Labrador, and Prince Edward Island were routinely BCG-vaccinated from the 1940s to the 1970s. Vaccination of healthcare workers was discontinued in the 1970s in Alberta, British Columbia, Manitoba, New Brunswick, Nova Scotia, Ontario, and Prince Edward Island. Many First Nations and Inuit populations continue to be vaccinated¹⁹. Worldwide, BCG continues to be widely used (BCG World Atlas; McGill University, Public Health Agency of Canada; http://bcgatlas.org).

False-negative TST can result from immune suppression caused by immunosuppressive therapies, coexisting illness (HIV, cancer, chronic renal failure, recent viral illness such as mononucleosis, mumps, and measles), or inappropriate induration cutoff thresholds^19,31.

Effect of anergy on screening performance

A consistent definition of anergy is absent from the literature; however, one possible clinically relevant definition is T cell hyporesponsiveness to previously exposed antigens due to immunosuppression caused by illness or drugs that may produce false-negative TST and IGRA results.

Anergy testing was developed to more accurately identify anergy in immunocompromised populations in order to better interpret negative TST results. It involves skin testing with a panel of control antigens, for example, Candida albicans and mumps. The CDC defines anergy as no reaction to these control antigens³². Anergy testing limitations include no standard convention for classifying positive response, the fact that validity in diagnosing TB has not been demonstrated, and the lack of reproducibility and standardized protocol³³. Thus anergy testing is not recommended by Canadian and US TB guidelines^19,20,33.

To maximize TST sensitivity, Canadian and US TB guidelines indicate a reduced induration threshold that defines a positive TST for the immunosuppressed population versus the general population. In most clinical situations in Canada, the criteria is 10 mm induration size¹⁹. However, ≥ 5 mm defines a positive reaction for immunosuppressed patients (e.g., HIV infection, organ transplant, corticosteroid therapy, TNF antagonist therapy). In the US, the threshold for persons with no TB risk factors is ≥ 15 mm, but for immunosuppressed persons, the cutoff is ≥ 5 mm^20,26. Contrary to the North American standards^19,20,26, the 2010 European TBNET Consensus Statement²⁷ recommends a TST cutoff of ≥ 10 mm for immunocompromised conditions.

Some studies show that a blunted or reduced recall response, often called anergy, can produce false-negative TST results in patients with RA not on immunosuppressive therapy^34,35,36,37. Sezer and colleagues observed that untreated patients with early-stage RA had statistically significant smaller induration diameter compared to treated patients with RA and AS, and healthy subjects³⁴. This less robust response to PPD may be due to the T cell’s active role in developing RA, thus a blunted response to PPD³⁴. In addition, data confirm that RA monocytes have lower PPD-antigen presentation capacity³⁷. Patients with RA receiving immunosuppressive therapy may exhibit attenuated delayed-type hypersensitivity (DTH)^{38,39,40,41,42}. A higher dose of methotrexate is associated with PPD anergy and a higher steroid dose was a predictor of negative PPD⁴². It has been reported that high doses of prednisone (≥ 15 mg/day) can cause false-negative TST reactions⁴³. A decrease in memory T4 cells was found in anergic patients, which may contribute to the decrease in antigen reactivity⁴¹. In contrast, there is weak evidence that supports the absence of anergy in patients with RA^44,45. However, the evidence supporting anergy in patients with RA is confounded by concomitant immunosuppressive medications and disease severity.

In addition to the RA population, evidence supports the existence of anergy in patients with ESRD on hemodialysis, HIV, or lupus^46,47,48. A significantly higher proportion of patients with ESRD had no TST induration versus controls⁴⁷. Similarly, HIV-positive patients were significantly less likely to be TST-positive compared to HIV-negative patients⁴⁶. Patients with active systemic lupus erythematosus had lower skin response to PPD than inactive or control patients⁴⁸.

Scientific basis of IGRA

Currently, 2 IGRA types, QuantiFERON^®-TB Gold In-Tube (QFT-GIT) and T-SPOT^®.TB [a simplified variant of the enzyme-linked immunospot (ELISPOT)], are approved in many countries. QFT-GIT is the latest generation test; its predecessors include QuantiFERON^®-TB (QFT) and QuantiFERON^®-TB Gold (QFT-G), approved in 2001 and 2005, respectively. QFT-GIT involves collecting whole blood and incubating it with proteins specific to M. tuberculosis but absent from BCG (unlike the TST) and most NTM. Lymphocytes from TB-infected patients will recognize M. tuberculosis-specific proteins and respond by releasing IFN-γ, which is quantified. T-SPOT.TB is similar to QFT-GIT, except that quantified peripheral blood mononuclear cells are incubated and enumerated.

Practical strengths and weaknesses

Although initial material and ancillary costs are greater than TST, IGRA can improve LTBI detection specificity in a BCG-vaccinated population. IGRA testing for the most part requires 1 appointment, while TST requires a second appointment 48 to 72 hours later. Additionally, IGRA uses a negative control and a positive mitogen-based control, minimizing additional visits due to inconclusive results¹⁹. However, a repeat test is required if an inclusive IGRA result occurs. The labels for QFT-GIT and T-SPOT.TB state that if the response to TB antigens is negative but the response to the mitogen is low, the result is considered indeterminate. A study by Greenberg and colleagues demonstrated an increase in indeterminate results due to low mitogen responses in patients with RA using the earlier generation IGRA (GFT)⁴⁴. Such a result can provide additional information because it could be due to poor specimen preparation, insufficient lymphocytes (for QFT), reduced lymphocyte activity due to improper specimen handling, incorrect filling/mixing of the mitogen tube, or inability of the patient’s lymphocytes to generate IFN-γ (one facet of anergy).

Measures of screening performance

A recent metaanalysis identified 38 studies published prior to March 31, 2008, that examined sensitivity and specificity of QFT-G, QFT-GIT, and T-SPOT.TB⁴⁹. Eight studies were excluded because they included noncommercial assays, leaving 30 studies for analysis. Studies assessing sensitivity had patients with microbiologically confirmed active TB but who were not immunocompromised. Studies assessing specificity had healthy patients with a low risk of TB and no known exposure to TB. The IGRA type used in the 30 studies was 11 QFT-G only, 5 QFT-G and T-SPOT.TB, 2 QFT-GIT and T-SPOT.TB, 4 QFT-GIT only, and 6 T-SPOT.TB only. Pooled sensitivity of the 22 QFT-G/QFT-GIT and 13 T-SPOT.TB studies was 76% (95% CI 72% to 80%) and 90% (95% CI 86% to 93%), respectively. Pooled specificity was 98% (95% CI 96% to 99%) for all QFT-G/QFT-GIT studies and 93% (95% CI 86% to 100%) for T-SPOT.TB. Sensitivity estimates for TST were heterogeneous, with a pooled estimate of 77% (95% CI 71% to 82%). Specificity estimates for TST in non-BCG-vaccinated populations were consistently high, with a pooled estimate of 97% (95% CI 95% to 99%). Specificity was low in BCG-vaccinated populations, 59% (95% CI 46% to 72%) and highly heterogeneous⁴⁹. Overall, the pooled T-SPOT.TB sensitivity was higher than that of the QFT-G and QFT-GIT assay⁴⁹. This has also been shown in rheumatic patients scheduled for TNF antagonist treatment and in immunocompromised patients^50,51. It is important to note that the lack of a gold standard for LTBI diagnosis is a limitation in the studies that assessed TST and IGRA specificity and sensitivity. In the absence of a gold standard, populations with active TB and low risk of TB were used by Pai and colleagues as surrogate standards to estimate accuracy⁴⁹.

Some studies suggest that IGRA, when compared to TST, may improve specificity and sensitivity in diagnosing LTBI prior to or during TNF antagonist treatment in patients with rheumatic disease^{30,38,52,53,54,55,56,57,58,59}. Some data propose that IGRA reduces false-positive TST results associated with BCG and NTM and may minimize the number of false-negative TST results due to immunosuppression^{30,38,52,54,55,57,58,60}. It should be noted that IGRA tests may also generate false-positive results, which may limit the utility of these tests for detecting LTBI without the use of other complementary tests⁵³. Accordingly, some studies^58,59 suggest IGRA should replace TST, while others^{30,38,44,52,53,54,57,60,61} suggest concomitant tuberculin skin and IGRA testing is more effective in diagnosing LTBI in all patients. Inanc and colleagues propose retesting patients with QFT-G after starting TNF antagonist therapy if QFT-G results before starting TNF antagonists were negative or indeterminate⁵⁵. However, there are conflicting data on whether IGRA improve specificity and sensitivity and the role of these new methodologies remains under study.

Few data exist on patients with psoriasis regarding LTBI detection and use of IGRA. However, a study by Laffitte and colleagues screened for LTBI in patients with psoriasis before TNF antagonist treatment by comparing T-SPOT.TB versus TST, and found a positive T-SPOT.TB strongly associated with risk factors for LTBI⁶². This study represents the first report of the frequency of LTBI (20%) in patients with psoriasis prior to TNF antagonist therapy. The incidence reported here is considerably higher, and it is possible that this overestimation could be due to the small number of patients in the study; to the fact that 20% of patients came from or previously lived in a country with a high incidence of TB⁶²; or to false-positive results that may have occurred with the use of IGRA⁵³.

The effect of anergy on screening performance for patients treated with TNF antagonists

Clinical studies in patients receiving TNF antagonist therapy for inflammatory diseases, including RA, demonstrated a decrease in IFN-γ-releasing CD4+ T lymphocytes or a decrease in IFN-γ production^63,64. This decrease in IFN-γ due to TNF antagonist treatment could cause an indeterminate or false-negative IGRA result, thus Hamdi and colleagues recommend against using the ELISPOT assay to diagnose LTBI in patients receiving such treatment⁶³. Currently, there are not enough consistent data to evaluate the effect of TNF antagonists on TST.

Discordance between TST and IGRA results

Multiple studies demonstrate discordant TST and IGRA results^{30,38,44,52,53,54,58,61,62,65}. Discordant results in an RA population that includes candidates for TNF antagonist therapy or those currently receiving it can be explained by several factors: differences between study populations; a false-negative TST or IGRA caused by anergy; a false-positive TST due to BCG vaccination or NTM infection; an indeterminate IGRA that results from immunocompromised status and/or TNF antagonist treatment, further suppressing the immune response; poor sample preparation; or inconsistent laboratory technique^{19,30,38,44,52,54,57,61,65,66}. BCG vaccination also causes discordant results between TST and T-SPOT.TB in patients with psoriasis prior to TNF antagonist therapy⁶². Although evidence suggests that IGRA with TST, used as screening tools for LTBI, improve specificity and sensitivity in populations with prior BCG vaccination and possibly other select populations, anergy can cause indeterminate or false-negative IGRA and TST results. Thus, clinicians should refer to guidelines that recommend relying on patient history, clinical features, and other laboratory results to diagnose LTBI in immunocompromised patients¹⁹.

Individualizing assessment of LTBI in candidates for TNF antagonist therapy

It is important that LTBI diagnostic tests maintain sufficient specificity and sensitivity, to minimize unnecessary LTBI therapy with isoniazid and activation of TB in patients with LTBI who receive TNF antagonist treatment. Whether a TB specialist, rheumatologist, or dermatologist is responsible for LTBI screening depends on what is appropriate in the physician’s region and the resources that are available.

LTBI screening is discouraged in healthy persons where TB risk is low¹⁹. However, the Canadian TB Standards and TNF antagonist product labels mandate screening for LTBI and active TB in all immunocompromised patients¹⁹ and candidates for TNF antagonist therapy, respectively (Table 1, Figure 1). In such cases, the first test should be a TST. If the patient is immunosuppressed and was not BCG-vaccinated and the induration is ≥ 5 mm, then the TST is interpreted as positive. The patient should be considered to have LTBI after active TB is ruled out by chest radiograph and medical examination. If the patient has been BCG-vaccinated, a false-positive TST can occur and an IGRA may be of benefit^23,30, however, it is not recommended by the CTC²² and further evidence is needed. It is increasingly recognized that a TST may overestimate true LTBI⁶⁷. This has resulted in the recommendation to use IGRA to clarify a positive TST result in adults at low risk^22,23. Whether this applies to persons being considered for TNF antagonists is unknown but is under study⁶⁸. When clinical suspicion suggests a possible false-negative TST in an immunocompromised person (e.g., someone with HIV, organ transplant, ESRD, or TNF antagonist therapy), an IGRA may be used, with the T-SPOT.TB being the preferred test^19,22. If 2 indeterminate results occur, anergy is suspected and the clinician must rely on history, clinical features, and other laboratory results to make a decision. For the general nonimmunocompromised Canadian population with no TB risk, the TST induration cutoff is 10 mm¹⁹.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Evaluation of latent tuberculosis infection (LTBI) in candidates for tumor necrosis factor (TNF) antagonist therapy. *Tuberculin skin test (TST) positivity varies according to the population under study. For a normal population, a TST > 10 mm should be considered positive¹⁹. For a population at risk (including HIV infection; close contact of active contagious case; children suspected of having TB disease; abnormal chest radiograph with fibronodular disease; other immune suppression such as TNF antagonist therapy, chemotherapy), ≥ 5 mm should be considered positive^{19. †}Consider based on background risk. Interferon-γ release assay (IGRA) positivity varies according to the population under study. How different populations affect IGRA results is still unclear. ^‡May be beneficial, but further evidence is required. ?: Areas with few data, which will continue to evolve. BCG: Bacille Calmette-Guérin; INH: isoniazid; RMP: rifampicin.

Even if the baseline evaluation suggests no LTBI, TB may still develop while taking TNF antagonist therapy^38,61. Thus careful evaluation of new symptoms suggestive of active disease must be performed while patients receive TNF antagonist therapy¹⁹.

The risk of progressing to active TB in patients with RA is elevated, likely the result of both the underlying disease and more importantly, nonbiologic medications, particularly corticosteroids^9,69. According to the Canadian TB Standards, ≥ 5 mm is the TST cutoff for immunosuppressed patients¹⁹. Patients with moderate to severe RA may have reduced response to TST and potentially IGRA as a result of concomitant medications and in some instances as a result of their underlying disease process^{34,37,38,40,42,63,64,70}. Consideration should be given to the risk and benefit of the 5 mm cutoff, as the number of false-positives may be high in patients who are not immunosuppressed or anergic. However, according to the American College of Rheumatology, a TST is considered positive if the reaction is ≥ 5 mm for immunosuppressed patients with RA⁷¹.

For the patient with psoriasis, currently there is no evidence of anergy; however, the absence of data does not exclude the phenomenon in psoriasis and other diseases. Relatively speaking, patients with psoriasis are healthy, therefore a 5 mm TST cutoff is not reasonable. Although there is a lack of evidence from national guidelines, the standard of care followed in Canada for LTBI screening in patients with psoriasis who are candidates for TNF antagonist therapy follows the screening standard for the general population; a TST reaction is considered positive if the induration is > 10 mm and the patient should receive LTBI treatment¹⁹.

All TNF antagonist candidates should undergo LTBI screening since an increased TB risk exists with TNF antagonist therapy. Adherence to the TB guidelines and awareness of the reasons for false-positive or negative TST results and false-negative or indeterminate IGRA results are important for more accurate LTBI diagnosis and the appropriate LTBI therapy.