Conversion and Reversion Rates of Tuberculosis Screening Assays in Patients With Rheumatic Diseases and Negative Baseline Screening Under Long-Term Biologic Treatment

Background: To determine the conversion and reversion rates of tuberculosis (TB) screening tests (Tuberculin Skin Test-TST, Interferon Gamma Release Assay-IGRA: T-SPOT.TB) during biologic treatment in patients with rheumatic diseases and negative baseline screening. Methods: This was a long-term, longitudinal cohort study of 50 patients with rheumatic diseases and negative baseline TB screening (TST: < 5 mm, negative T-SPOT.TB) treated with tumor necrosis factor inhibitors (TNFi) or other non-TNFi biologics. Patients were rescreened at a mean time of 1.4 (first rescreening) and 6.9 (second rescreening) years from baseline, with both assays. The conversion (negative to positive) and reversion (positive to negative) rate was calculated for each TB screening test. Results: Fifty patients (mean age = 60 years) with various rheumatic diseases (rheumatoid arthritis: n = 24, spondyloarthropathies: n = 23, other: n = 3) were enrolled. During the first phase (baseline to first rescreening), all patients were treated with TNFi while during the second phase (first to second rescreening), TNFi (54%) and non-TNFi (46%) were used. Fifteen patients (30%) displayed conversion of at least 1 screening assay during follow-up (10 at the first and 5 at the second rescreening). This conversion rate was higher with TST (n = 11, 22% or 3.47/100 patient-years) compared to T-SPOT.TB (n = 4, 8% or 1.74/100 patient-years). Among the 10 converters at the first rescreening, 5 received isoniazid (INH) preventive therapy and 5 did not; an equal number of patients (3/5, 60%) reverted to negative with or without INH therapy. None of the patients developed active TB during follow-up (6.9 ± 1.0 years). Conclusions: Approximately one-third of patients with rheumatic diseases and negative baseline TB screening developed conversion of at least 1 screening test during long-term biologic treatment. This occurred most often with TST and was usually a transient event. These findings do not support routine serial TB retesting in biologic-treated patients with rheumatic diseases in the absence of TB risk factors.

Currently available assays for the diagnosis of LTBI include the tuberculin skin test (TST) and the interferon-gamma (IFN-γ) release assays (IGRAs) T-SPOT.TB (Oxford Immunotec, Oxford, UK) and QuantiFERON-TB Gold In Tube (QFT-GIT; Cellestis, Carnegie, Victoria, Australia). Compared to TST, IGRAs have shown similar sensitivity but higher specificity for the detection of LTBI, while both are negatively affected by immunosuppressive therapy [5]. Based on findings from several studies, recent guidelines recommend their use rather than TST as the diagnostic test of choice for individuals 5 years or older [6].
Tumor necrosis factor-α inhibitors (TNFi) were the first class of biologic disease-modifying anti-rheumatic drugs (bDMARDs) that were used in patients with rheumatic diseases, and they are currently licensed for different types of inflammatory diseases including rheumatoid arthritis (RA), spondyloarthropathies (SpA), inflammatory bowel disease (IBD) and psoriasis. The introduction of TNFi in clinical practice was followed initially by an increase in the TB cases in patients with rheumatic diseases undergoing this type of treatment [7][8][9], whereas cases of TB reactivation have been described less often in patients treated with non-TNFi bDMARDs [10] and the newer targeted synthetic DMARDs (tsDMARDs) [11], especially in high-incidence areas [12]. Following the initial reports of TNFi-induced TB reactivation, universal screening with TB screening tests of all patients with rheumatic diseases starting therapy with biologics has been employed and has proved to be efficacious in substantially decreasing the incidence of TB reactivation [13].
Despite these encouraging results, there are still a number of unresolved issues regarding TB screening in patients with rheumatic diseases. First, the optimal use of one or the other TB screening test has not been clarified. Both the World Health Organization (WHO) and the American College of Rheumatology (ACR) recommend screening with either TST or IGRA without preferentially advocating one technique over the other [14]. Experts in the field support the implementation of dual screening with both tests in patients with rheumatic diseases, because this approach has been found to increase sensitivity [15,16], and it is also the practice for our unit [17]. Secondly, despite the recent ACR guidelines for annual rescreening of patients with RA treated with biologics and with negative baseline screening who have risk factors for TB exposure [14], the real-life data to support such a strategy are lacking.
Regardless of the rescreening approach employed, the rate of conversion and reversion of TB screening assays remains an issue. Recent data even with the IGRAs have shown that discordance in serial screening is not rare, especially in people with borderline positive results [18] as well as health care workers (HCWs) [19] and patients with rheumatic diseases [20][21][22][23].
Thus, the aim of our longitudinal cohort study was to evaluate the long-term rate of TB screening test conversion and reversion during biologic treatment. For this we utilized our previously published cohort of patients with rheumatic diseases and with negative screening at baseline [20].

Patients
As previously reported, between October 2009 and December 2013, 247 patients with rheumatic diseases had been screened for LTBI before starting a TNFi in our unit of a tertiary referral hospital serving patients from across the country (Clinical Immunology-Rheumatology Unit, 2nd De-partment of Medicine and Laboratory, Hippokration General Hospital, Athens, Greece). Greece, according to the latest WHO data, is considered a low-TB-incidence country with an annual TB incidence rate of 4.5/100,000 [24]. Screening was performed with the TST, T-SPOT.TB, and QFT-GIT [20]. All patients were HIV negative at baseline.
Seventy patients with negative baseline screening by all methods were reevaluated approximately 1 year later (first rescreening) while receiving TNFi treatment [20]. Among them, 20 were not available for a second rescreening: 2 had died, 5 denied rescreening, and 13 were lost to follow-up. Thus, 50 patients who were available for a second rescreening approximately 5 years later were included in the study.
During follow-up, patient demographics, history of new TB contact, history of Bacillus Calmette-Guérin (BCG) vaccination (based on the presence of BCG scar), and past and current immunosuppressive or disease-modifying therapy (glucocorticoids, conventional synthetic/csDMARDs and bDMARDs) were recorded. Physical examination, chest-computed tomography (CT) scan, or any other appropriate evaluation were performed in all patients who exhibited conversion of 1 TB screening test in order to rule out active TB. Patients who were considered as high risk for TB development after TST and/or T.SPOT-TB conversion were given LTBI therapy according to the decision of the treating physician.
All patients gave informed consent prior to rescreening, and the extension of the study was approved by the Hospitals' Institutional Review Board.

TB screening tests (TST-IGRAs)
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. A diameter of transverse induration ≥ 5 mm was considered positive. Blood for IGRAs was drawn the same day just before the TST. The T.SPOT-TB assay was performed and interpreted according to the manufacturer's instructions [20]. No invalid or indeterminate results were recorded in any of the tested samples. TST conversion was defined as the change of TST from < 5 mm to ≥ 5 mm while the opposite was regarded as reversion.
Since the beginning of the study, the QFT-GIT assay that was used for the baseline and first rescreening [20] has been replaced by the newer QuantiFERON-TB Gold Plus (QFT-Plus, Qiagen) assay which was used in the second rescreening. Since the TB antigens used in these 2 assays are not the same, the QFT results were not included in the current analysis. Nevertheless, we observed that 7 patients demonstrated a QFT conversion (14%, 4 at the first and 3 at the second rescreening). Among the 4 early converters, 1 received INH preventive therapy and remained positive while all 3 patients who did not receive INH therapy, reverted to negative at the second rescreening).

Statistical Analysis
All statistical analyses were performed with SPSS (IBM SPSS Statistics for Windows, v. 20.0. Armonk, NY: IBM Corp). Continuous variables were expressed as mean ± SD or median (IQR) and dichotomous variables were expressed as absolute and percentage (%) values. Chi square or Fisher's exact test was used for comparison of dichotomous and Mann-Whitney or t test for continuous variables. Statistical significance was set at P < 0.05.

Patient Characteristics
Fifty patients with rheumatic diseases and negative baseline screening were included in the current study ( Figure 1 and Table 1). All patients (n = 50) had their first rescreening 1.4 years (mean) and their second rescreening 6.9 years (mean) from baseline,. The characteristics of the 50 patients at the time of the second rescreening are shown in Table 1. Their mean age was 60 years, 31 (62%) were women, and their mean disease duration was 15 years. Rheumatoid arthritis (RA) was the most common diagnosis (n = 24, 48%) followed by spondyloarthopathies (SpA, n = 23, 46%) and other rheumatic diseases (n = 3, 6%). All patients received TNFi during the first phase (baseline to first rescreening) as previously described [20] while during the second phase (first to second rescreening) almost all (47/50, 94%) were treated with bDMARDs (TNFi: 54%, non-TNFi: 46%), 2 (4%) received only csDMARDs, and 1 (2%) did not receive any additional therapy. At the time of the second rescreening, only 8 patients (16%) were receiving glucocorticoids with a mean prednisolone equivalent daily dose of 5 mg.
Regarding TB risk factors, 5 patients (10%) were from countries with high TB incidence while only 2 (4%) reported possible TB exposure during follow-up. Fifteen (30%) patients had been previously vaccinated with BCG (usually during their childhood or early adolescent period).

TB Screening Test Conversion and Reversion During Follow-Up
During the follow-up period (~ 6.9 years), 15 (30%) patients with negative baseline TB screening, showed conversion in at least 1 TB screening test (Figure 1). Their detailed characteristics are shown in Table 2.
Thus, only 2 (33%) patients who had a conversion at first rescreening remained positive at the end of follow-up (both with TST, persistent converters). None of these 50 patients developed active TB during the follow-up period.

Comparison Between Converters and Non-Converters
There were no statistically significant differences between converters (n = 15) and non-converters (n = 35) as shown in Table 3, although there was a trend for increasing age (62.2 ± 12.9 vs 57.9 ± 13.9 years) and a non-RA diagnosis (67% vs 46%) for converters vs non-converters, respectively.
Similarly, the conversion rate per 100 patient-years from baseline to first rescreening (70.9 patient-years) was 9.87 for TST and 5.64 for T-SPOT.TB, whereas the rates for the total duration of follow-up (288 patient-years) were 3.47 and 1.74 per 100 patient-years, respectively (OR = 2, 95% CI: 0.68-6.47,P = 0.3).

DISCUSSION
To our knowledge, this is the first prospective, clinical practice based, long-term (~7 years) study that evaluated the conversion and reversion rates of 2 different TB screening assays (TST and IGRA/T-SPOT.TB) during serial retesting in biologic-treated patients with rheumatic diseases and negative baseline screening. Our main findings are that in a low risk population, conversion of a TB assay during biologic treatment is common (occurring in 30% of patients), and it is more frequent with TST than an IGRA and is usually a transient event. Early converter -

INH therapy
The characteristics of the 15 patients who converted TST or T-SPOT.TB (grey boxes) are shown. This includes those who remained positive at both re-screenings ("persistent converters", n=2, #1-2)), those who reversed from positive to negative between the first and second rescreening without isoniazid (INH) therapy ("transient converters", n=3, #3-5), those who converted between the first and second rescreening ("late converters", n=5, #6-10), and finally the converters at the first rescreening who received INH therapy ("early converters on INH therapy", n=5, #11-15). * Expressed as prednisolone equivalent dose (in mg/day)  The comparison between those who did (n = 15) or did not (n = 35) convert during follow-up is shown.
TB screening with TST and/or IGRAs is currently recommended as a standard procedure for all patients with rheumatic diseases starting biologic or targeted synthetic (such as kinase inhibitors) therapy [14], because of the high incidence of TB reactivation that was noticed in the early years of TNFi therapies without appropriate TB screening for LTBI [7]. Nevertheless, in real life settings it has been shown that even among patients with negative baseline screening, TB can occur during biologic treatment. Its incidence varies significantly based mainly on the TB incidence in the respective population (0.2%-4%) [12,25,26]. In a recent study from a highly endemic area (South Africa), the rate of active TB was 1.2/100 patient-years among biologic-users; half of the active TB cases were due to new TB exposure and half were due to TB reactivation (indicative of TB screening failure) [12].
Although TB screening is now a routine procedure for all patients starting biologics, the value of routine serial retesting during long-term biologic therapy has not been studied adequately. Some scientific societies recommend rescreening only for high-risk patients [14,27,28] whereas others do not offer any specific recommendations [27]. Despite these recommendations though, in daily clinical practice the majority of rheumatologists (63%) in a recent survey, reported serial TB retesting of biologic-treated patients with rheumatic diseases (24% annually for all patients, 15% every 2 years and 24% only for those with TB risk factors) [29]. Interestingly, this practice was more common in the United States (70%) than outside the United States (31%) [29].
It should be noted that this high conversion rate occurred in a low-TB-risk patient population (only 2/50 patients had potential TB exposure during the follow-up period) residing in a low-TB-incidence country [24].
Data from different countries have shown that the conversion rate of IGRAs may not differ significantly between countries with low or high TB incidence [20,21,23,26,42] but as Kim et al recently reported, in high-incidence countries the rate of active TB development among converters is higher [41].
The potential practical implications of our findings are that the demonstration of a TB test conversion during serial retesting of biologic-treated patients with rheumatic diseases will require extensive evaluation (including sputum stains and cultures, CT of the chest, additional lab work etc.) for active TB of a large proportion of them. Furthermore, patients with a negative workup for active TB should receive appropriate preventive therapy (INH or rifampin) for 6 to 9 months with close clinical and laboratory monitoring, which would increase cost and the possibility of drug-related adverse events.
A second unresolved issue with serial TB retesting is whether these represent true or transient conversions. Most data so far are derived from serial retesting in otherwise healthy HCWs. These data have shown a high reversion rate (form positive to negative) for the IGRA assays, QFT (22-76%) [43,44], and T-SPOT.TB (77%) [44]. In a recent study, Moses et al estimated (using a Markov's model) that serial testing with QFT-GIT of HCWs in low-incidence areas like North America would result in a 24.6% false positive rate over a 10 year period [45]. Based on these data, the CDC currently does not recommend routine serial retesting for HCWs, except for those at increased risk for TB [46].
So far there have not been any similar data from serial retesting in patients with rheumatic diseases treated with biologics. There has been only 1 retrospective study of transplant patients who converted their QFT assay (n = 23/195, 12%) and had been retested (n = 7). Among these patients, 3/7 (43%) reverted to negative [47].
In our prospective study, 5 patients who converted and did not receive INH preventive therapy were followed for ~ 5 years. The decision not to administer INH preventive therapy was made by the treating physician based on the emerging data at the time showing a high frequency of IGRA reversion in HCWs, as discussed above [44]. None of these patients developed TB while 3 of them reverted to negative (reversion rate: 60%). Interestingly, this rate of reversion was the same for the 5 patients who had been treated with INH preventive therapy (3/5, 60%) by their physician. Despite the small number of retested patients this is a novel finding that needs to be discussed further.
Limitations of our study include its small size, the absence of serial retesting in a control group treated only with non-biologics, the setting of a low-TB-incidence area, and the low proportion of foreign-born individuals or immigrants from high-TB-burden countries. Thus, our results may not be generally applicable to patients with other co-morbidites such as HIV infection, end stage renal disease, solid organ or bone marrow or stem cell transplants, or for high-TB-incidence countries.
In conclusion, our long-term, real-life, prospective study showed a high conversion of TB screening tests (TST, IGRA) during serial retesting in a low-risk biologic-treated patient population with rheumatic diseases, which in the majority of cases was a transient phenomenon. Although the number of included patients was small, we believe that these real-life data do not support routine serial TB retesting in biologic-treated patients with rheumatic diseases, with the exception of high-risk patients or for those who present with suspicious clinical findings for TB.

POTENTIAL CONFLICTS OF INTEREST
The authors declare that there are no competing interests in this study.