Article Text

Extended report
Frequent conversion of tuberculosis screening tests during anti-tumour necrosis factor therapy in patients with rheumatic diseases
  1. Chrisoula Hatzara,
  2. Emilia Hadziyannis,
  3. Anna Kandili,
  4. Christos Koutsianas,
  5. Anastasia Makris,
  6. Georgios Georgiopoulos,
  7. Dimitrios Vassilopoulos
  1. 2nd Department of Medicine and Laboratory, National University of Athens Medical School, Hippokration General Hospital, Athens, Greece
  1. Correspondence to Dr Dimitrios Vassilopoulos, 2nd Department of Medicine and Laboratory, National University of Athens Medical School, Hippokration General Hospital, 114 Vass. Sophias Ave, Athens 115 27, Greece; dvassilop{at}


Objectives To determine the rate of tuberculosis (TB) screening test conversion during anti-tumour necrosis factor (TNF) therapy in rheumatic patients with negative baseline screening.

Methods This was a prospective study of rheumatic patients with negative baseline TB screening (tuberculin skin test (TST): <5 mm, and negative T-SPOT.TB, QuantiFERON-TB Gold In Tube (QFT-GIT) and chest X-ray) treated with anti-TNF agents. All patients underwent re-screening for TB with all assays 1 year later. Factors associated with TB test conversion were analysed and compared between ‘converters’ and ‘non-converters’.

Results Seventy patients (mean age 50.6±15.5 years) with rheumatic disease (33 with rheumatoid arthritis, 33 with spondyloarthropathies and 4 with other conditions) were enrolled. Patients were treated with different anti-TNFs (27 with adalimumab, 14 etanercept, 16 infliximab, 8 golimumab, 5 certolizumab pegol) for 1 year. Twenty patients (29%) displayed conversion of at least one screening assay 12 months after anti-TNF therapy: conversion of TST occurred in 9 (13%), T-SPOT.TB in 7 (10%) and QFT-GIT in 5 (7%). Only one patient had concomitant conversion of more than one screening test. Univariate and multivariate analysis revealed that only infliximab was associated with a decreased rate of TB screening assay conversion (OR 0.048, 95% CI 0.004 to 0.606, p=0.017). No patient (40% received isoniazid therapy) developed active TB during follow-up (27±12 months).

Conclusions Approximately one third of patients with negative baseline TB screening develop conversion of at least one screening test during anti-TNF treatment. These findings should be considered when designing re-screening strategies and contemplating latent TB therapy.

  • Rheumatoid Arthritis
  • Anti-TNF
  • Tuberculosis

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Interferon-gamma (IFN-γ) release assays (IGRAs) including the T-SPOT.TB (Oxford Immunotec, Oxford, UK) and QuantiFERON-TB Gold In Tube (QFT-GIT; Cellestis, Carnegie, Victoria, Australia) assays are established tests for tuberculosis (TB) screening in rheumatic patients starting therapy with biologic agents, particularly anti-tumour necrosis factor (anti-TNF).1 ,2 There is ongoing discussion whether IGRAs should be used alone, interchangeably or in combination with the traditional tuberculin skin test (TST) for TB screening in patients starting biologics, with most experts supporting a dual screening strategy regardless of the presence of TB risk factors.3 ,4

Although the value of IGRAs for initial screening has been recognised, it remains uncertain what their role is for monitoring patients during biologic therapy (especially anti-TNF). This mainly applies to patients with negative assays at screening and TB risk factors who could be either exposed to Mycobacterium tuberculosis during therapy or have a reactivation of a previously undiagnosed latent infection. According to the most recent American College of Rheumatology (ACR) recommendations, patients with rheumatoid arthritis (RA) who receive biologic therapies and have negative baseline screening with either the TST or IGRAs, should be re-screened annually if there are ongoing or future risk factors for TB exposure.2 So far there have been few prospective studies examining the performance of these assays as monitoring tools for TB during biologic therapy.

The aim of our prospective study was to determine the rate of TB test conversion (TST and both IGRAs) during anti-TNF therapy in rheumatic patients with negative baseline screening.

Materials and methods


Between October 2009 and December 2013, 247 patients with various rheumatic diseases were screened prior to biologic therapy initiation at the Clinical Immunology-Rheumatology Unit (2nd Department of Internal Medicine and Laboratory, Athens University School of Medicine) for latent TB with the TST, T-SPOT.TB and QuantiFERON-TB Gold In Tube assays, as has been previously described.4 ,5 Seventy (28%) patients with negative baseline TB screening by all methods who were available for retesting 1 year later were included in this prospective study while undergoing anti-TNF treatment (the flow chart of all patients screened for TB is shown in online supplementary figure S1).

Patients with history of TB, baseline chest X-ray (CXR) findings suggestive of old or active TB, previous treatment with anti-TB (including isoniazid (INH)) or biologic agents were excluded from the study. All patients signed an informed consent form prior to their participation in the study, and the study was approved by the Institutional Review Board.

One year after the initiation of anti-TNF therapy, all patients underwent repeat testing for TB with all of the above-mentioned tests. A standard questionnaire was completed for each patient, including basic demographic data (sex, country of origin and country of residence), rheumatic disease (type and duration), comorbid conditions, history of previous TB contact or BCG vaccination, and concurrent immunosuppressive therapy (glucocorticoids, disease-modifying anti-rheumatic drugs (DMARDs)) both at baseline and 1 year after anti-TNF treatment initiation. Patients who were switched to another anti-TNF or biologic agent due to inefficacy or intolerance to the first anti-TNF agent were also recorded.

Physical examination and appropriate laboratory evaluation were performed every 1–3 months according to established guidelines for patient monitoring during anti-TNF therapy. A CXR was obtained for each patient at baseline and at re-screening and was evaluated for the presence of previous or inactive TB, as previously described.4

Tuberculin skin test

The TST was performed by intradermal injection (Mantoux method) of 0.1 mL (2 IU) of purified protein derivative (PPD RT 23; Statens Serum Institute, Copenhagen, Denmark) according to standard guidelines.6 The TST was considered negative when the diameter of the transverse induration was <5 mm, as previously described.4 ,5

IFN-γ release assays

The QFT-GIT and T-SPOT.TB assays were performed just before TST application in order to avoid potential interference with the IGRA results. The results of both IGRAs were interpreted according to the most recent Centers for Disease Control and Prevention (CDC) criteria.7 Six (2.4%) patients with indeterminate results (QFT-GIT: n=4; T-SPOT.TB: n=2) were excluded from the study (see online supplementary figure S1).

Statistical analysis

Continuous variables are expressed as mean±SD or median and IQR, while nominal variables are presented in absolute and percentage values (%). The two-sided Fisher's exact test was used for comparison of categorical variables, and the Wilcoxon rank test for continuous variables. Binary logistic regression analysis was performed in order to identify factors associated with TB screening test conversion. Re-sampling techniques were implemented to address the small size of the study and validate our main results. Statistical analysis was conducted using SPSS Statistics V.17.0 (SPSS, Chicago, Illinois, USA) and STATA V.11.1 (StataCorp, College Station, Texas, USA). Statistical significance was set at p<0.05. Power considerations and details of statistical analysis are provided in the online supplementary text.


Baseline patient characteristics

The patient characteristics are summarised in table 1. Among the patients, 39 (56%) were over 50 years of age, and approximately half had RA (n=33, 47%) or spondyloarthropathies (n=33, 47%). Nineteen (27%) patients had been vaccinated with BCG during adolescence or early adulthood. Risk factors for TB including age >50 years, possible remote TB exposure, and birth or previous residence in a country with high TB prevalence were present in 44 patients (63%).

Table 1

Patient characteristics

At the time of the initial TB screening, 63% of patients were receiving DMARDs (75% methotrexate, 18% leflunomide) and 43% corticosteroids (mean daily dose 6.6 mg). There was no significant change in this treatment during anti-TNF therapy (data not shown).

Among these 70 patients who were finally included in the study, 27 (39%) were initially treated with adalimumab, 16 (23%) with infliximab, 14 (20%) with etanercept, 8 (11%) with golimumab and 5 (7%) with certolizumab pegol. During the 1 year follow-up, six (9%) were switched to another anti-TNF agent (three from infliximab to adalimumab, two from adalimumab to etanercept, and one from etanercept to adalimumab). There were no significant differences in major baseline characteristics among users of different anti-TNF agents (data not shown).

TB screening test conversion during anti-TNF therapy

Conversion of at least one assay was observed in 20 patients (29%), referred to as ‘converters’, during anti-TNF therapy (table 1). Nine patients (13%) displayed conversion of the TST, seven (10%) of T-SPOT.TB and five (7%) of QFT-GIT (table 2). The detailed characteristics of the patients who converted are shown in table 2. If stricter criteria for test conversion were applied, as has been proposed,8–10 then 15 patients (21%) would have been considered as converters. More specifically, TST conversion defined as a TST increase >6 mm above the baseline value to a diameter of >10 mm8 was seen in eight patients (11%), T-SPOT.TB conversion, defined as an increase from <6 spots/well at baseline to ≥9 spots/well,9 in five patients (7%), and QFT-GIT conversion, defined as an IFN-γ concentration (after specific TB antigen stimulation) of >0.7 IU/mL or even 1 IU/mL,9 ,10 in two (3%) patients.

Table 2

Characteristics of converters

Only one patient converted more than one TB screening test (5%). This patient (no. 20, table 2), a 32-year-old male with adult-onset Still's disease receiving treatment with prednisolone (∼10 mg/day), methotrexate and etanercept, converted both the TST and QFT-GIT tests. There was no obvious TB contact during anti-TNF treatment, but the patient was born in a high prevalence TB area. Interestingly, although he had a TST conversion from 0 to 15 mm, his IFN-γ level after specific TB antigen stimulation was 0.44 IU/mL, which is close to the cut-off value (0.35 IU/mL) of the QFT-GIT assay.

Another patient (no. 19, table 2), who was a healthcare worker, had a definite contact with a patient with active TB during anti-TNF treatment and displayed conversion only of the T-SPOT.TB assay.

Among the patients who converted, eight (40%) received INH prophylaxis (300 mg/day orally) for 9 months. During the follow-up period (mean±SD 27±12 months, median 25 months), none of the patients developed active TB.

Factors associated with TB test conversion

In table 1, the various characteristics of converters (n=20) and non-converters (n=50) are compared. There was no statistically significant difference between the two groups in terms of sex, age, diagnosis, previous BCG vaccination, TB risk factors or concomitant DMARD or steroid therapy. Among patients who displayed conversion of at least one TB screening assay, fewer were treated with infliximab (5%) and more with etanercept (35%) compared to ‘non-converters’ (30% and 14%, respectively). This difference was statistically significant for infliximab (p=0.028) but not for etanercept (p=0.094).

Similar results were obtained when a univariate analysis was performed examining different factors that could be associated with TB test conversion such as age >50 years, sex, diagnosis of RA, possible previous TB exposure (possible remote TB exposure or birth/residence in a high TB prevalence country), previous BCG vaccination and treatment (steroids or DMARDs, anti-TNF agents) (table 3). Infliximab (p=0.05) and etanercept (p=0.054) use were associated with TB test conversion.

Table 3

Univariate and multivariate analysis of factors associated with TB screening conversion during anti-TNF therapy

Multivariate analysis (table 3) showed that the use of infliximab was protective against TB test conversion (model 1: OR 0.048, 95% CI 0.003 to 0.606, p=0.017; model 2: OR 0.031, 95% CI 0.002 to 0.414, p=0.009; see online supplementary text for details). These results were further confirmed through permutation: converters in our study had almost 92% decreased odds (model 1: OR 0.087, 95% CI 0.002 to 0.925, p=0.0396; model 2: OR 0.05, 95% CI 0.001 to 0.537, p=0.005) of being treated with infliximab among the anti-TNF agents after adjustment for various confounders. Possible previous TB exposure was also associated with TB test conversion in one of the models (model 2: OR 7.243, 95% CI 1.093 to 47.993, p=0.04).

There was no difference in the results when the analysis was performed according to the anti-TNF agent that the patients were most or ever exposed to during the treatment period, in order to account for patients who switched anti-TNF therapy (data not shown). Similar results were obtained when the analysis was carried out only in patients who had not switched anti-TNF agent during the 1-year follow-up period (n=64, data not shown).

Effect of anti-TNF therapies on IFN-γ responses

IFN-γ responses measured by IGRAs in vitro could be influenced directly by anti-TNF therapies. This was assessed by comparing mitogen-induced IFN-γ responses (corresponding to the maximum IFN-γ response) at baseline and at the end of the 1 year of treatment (figure 1). In the entire patient population there was a small but statistically significant decrease in IFN-γ response (from 18.8±7.9 IU/L to 16.2±8.6 IU/L, p=0.034) after anti-TNF treatment. This decrease was mainly seen in patients who had received infliximab therapy (p=0.01).

Figure 1

Interferon- γ (IFN-γ in IU/L) levels measured after mitogen stimulation by the QuantiFERON-TB Gold In Tube assay in patients at baseline (solid bars) and after 1 year of different anti-tumour necrosis factor therapies (white bars) (see ‘Materials and methods’ section for details).


This is the first prospective study in the literature examining the conversion rate of all available TB screening assays (TST, two IGRAs) in rheumatic patients receiving anti-TNF therapy. Our study shows that a significant proportion of patients convert at least one TB screening assay during 1 year of anti-TNF therapy (21–29%, depending on the criteria used for conversion definition). These results have important implications for the use of these assays as monitoring tools in daily clinical practice.

A number of previous studies have indicated that despite appropriate baseline screening for TB, some patients can develop TB during anti-TNF treatment.11–14 This could be due either to new M. tuberculosis exposure during therapy (most commonly) or reactivation of an undiagnosed latent TB infection (‘false negative’ baseline TST or IGRAs). This risk in rheumatic patients receiving anti-TNF therapy has been estimated to range from 0.2% to 4%.12–14 In an older study where only the TST was used for screening, there were six cases of TB among 1059 patients (0.6%) treated with infliximab, whereas in a very recent study, there were four cases of TB among 1893 patients (0.2%) treated with golimumab who were screened both by the TST and QFT-GIT assays.13 In a similar study although from a highly endemic area (Taiwan), there were six cases of TB among 150 patients (4%) with negative baseline screening with the TST and QFT-G assays.14 Five of these six patients developed TB after 20–24 months of treatment, indicating that they were most likely newly acquired infections.

Recently, both the ACR2 and an expert group in an updated consensus statement on the use of biologic agents in rheumatic patients,1 have recommended repeat TB screening for patients receiving anti-TNF agents who have risk factors for TB and negative baseline screening with the TST or IGRAs. Furthermore, the ACR has recommended that re-screening should be performed annually.2 However, in both cases the level of evidence was C, indicating the absence of well-performed randomised or non-randomised studies supporting this recommendation.1 ,2

There are as yet few longitudinal studies examining the performance of re-screening in patients receiving anti-TNF agents.14–22 The rate of conversion ranged from 0% to 37% for the TST,15–20 from 4% to 12% for QFT-G,14 ,17 from 0% to 12% for QFT-GIT18–20 ,22 and from 0% to 10.5% for T-SPOT.TB21 (duration of follow-up: 3 months to >3 years). In general, the rate of conversion was higher in high TB prevalence countries. For example, for the TST the conversion rate was 0–12% in low prevalence countries (Italy,17 ,20 Austria19) increasing to 25–37% in high prevalence areas (Turkey,18 Taiwan,15 Korea16). None of these studies examined all three assays in parallel and no risk factors for assay conversion were identified with the exception of one study where administration of anti-TNFs for more than 3 years was associated with TST conversion.16

There is increasing concern regarding the clinical significance of these IGRA test conversions during retesting. Most of the available data come from large scale studies in healthcare workers where conversion rates ranged from 0.7% to 14.4% for QFT-GIT and from 3.9% to 4.9% for T-SPOT.TB.23 The majority of these positive IGRAs were negative at retesting 6 months later in a recent US study (76–77%).24 Owing to these variations, which usually occur around the cut-off value for each assay, stricter criteria for test positivity for retesting have been proposed but not yet validated in clinical practice.8–10

Our study showed that approximately 10% (7–13%) of patients convert one individual test after 1 year of anti-TNF treatment. When the stricter criteria for test conversion were applied, the rate ranged between 3% and 11%. Test conversion occurred more frequently with the TST (13%, or 11% by the strict criteria), followed by T-SPOT.TB (10% or 7%, respectively) and QFT-GIT (7% or 3%, respectively). Interestingly, only the patient born in a highly endemic area had more than one test converted (TST and QFT-GIT).

On the other hand, a patient with definite TB exposure during treatment only converted the T-SPOT.TB assay, indicating that a dual screening strategy may be needed for re-screening as at baseline. If such a dual re-screening strategy is implemented (TST and one IGRA), then the conversion rate will range between 21% and 23% (17–19% with the stricter criteria), in which case approximately one in five patients with negative baseline screening would require INH treatment according to current guidelines.

The clinical significance of these TB test conversions is unclear at the moment. It remains to be clarified whether these results represent true positive conversions indicative of an underlying latent TB infection, or false positive results due to within-subject variations of the respective assays. In a recent study from Italy, among six patients who demonstrated QFT-GIT conversion at 12 months after anti-TNF treatment, only two (33%) remained positive 6 months later at retesting.20 The possibility that the TST (which contains the specific TB antigens included in the IGRAs) could induce a false positive IGRA result (‘boosting effect’) is unlikely given that retesting was performed 1 year later. In a systematic literature review of IGRA retesting, among 13 studies, five showed no boosting effect, while in the remaining, repeat IGRAs were performed close to TST administration (up to 4 months).25 Even in these studies, boosting occurred mainly in patients with positive baseline IGRAs. In the only long term study, no boosting of the T-SPOT.TB assay was seen up to 2 years after repeated TST administration.26

Despite the small number of patients included in this study, an attempt was made to identify risk factors associated with TB test conversion. Univariate and multivariate analysis showed that only the use of infliximab was associated with a decreased risk for TB test conversion, while etanercept seemed to increase this risk, but this change was not statistically significant. This is an interesting finding since it is well known that the risk for TB reactivation is regarded as generally being higher with infliximab compared to etanercept.27 Although there is not a clear explanation for this, it could be attributed to the suppressive effect that infliximab may have on IFN-γ responses measured in vitro. Indeed, when the mitogen-induced IFN-γ levels were compared before and after treatment, only infliximab led to a statistically significant reduced response.

Our study has certain limitations including its relatively small size and the absence of serial retesting after the 1-year time point. In any case, subsequent retesting would have been of limited value since 40% of the converters received INH treatment. Furthermore, the study was performed in a country with a high prevalence of past TB exposure and so the results may not apply to countries with low TB prevalence.4 Additionally, since our patients were purposely selected to have all three tests negative at baseline, we may have increased the risk for false positive results upon retesting.

In conclusion, our comparative prospective study shows that patients treated with anti-TNF agents display an annual individual conversion rate of TB screening tests of ∼10% that increases to ∼20% or even ∼30% if a dual (TST plus one IGRA) or triple (TST plus 2 IGRAs) retesting strategy, respectively, is adopted. Taking into account the risks associated with a missing diagnosis of new or reactivated TB in our patient population, we have adopted a dual screening and re-screening strategy (TST and one IGRA) that greatly increases the likelihood of diagnosing TB infection. However, prospective studies with large patient populations estimating the proportion of converters who eventually develop TB are needed before specific definite guidelines are issued.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Handling editor Tore K Kvien

  • Contributors  CH and EH: study design; acquisition, analysis, and interpretation of data; drafting the work or revising it critically for important intellectual content; and final approval of the version to be published. AK, CK, AM and GG: acquisition of data; revising the work critically for important intellectual content; and final approval of the version to be published. DV: conception and study design; acquisition, analysis, and interpretation of data; drafting the work and revising it critically for important intellectual content; and final approval of the version to be published. All authors agreed 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.

  • Funding This work was supported in part by research grants from the Hellenic Society for Rheumatology and the Special Account for Research Grants (S.A.R.G.), National and Kapodistrian University of Athens, Athens, Greece.

  • Competing interests None.

  • Ethics approval Hippokration General Hospital Ethics Committee approved this study.

  • Provenance and peer review Not commissioned; externally peer reviewed.