Treatment in Latent Tuberculosis Uveitis—Is Immunosuppression Effective or Is Conventional 3- or 4-Drug Antituberculosis Therapy Mandatory?

Background/Aims: Controversy exists regarding 3- or 4 drug antituberculosis therapy (conventional ATT) in uveitis patients having latent tuberculosis (LTB), especially while initiating therapy with corticosteroids and/or other immunosuppressants. Methods: We performed a monocentral retrospective analysis of posterior uveitis patients with latent TB. Latent TB was diagnosed, in case of a positive QuantiFERON®-TB-Gold test and normal chest imaging, after ruling out other causes of infectious and noninfectious uveitis. Patients with active TB were excluded. From 2016 to 2020 we included 17 patients. Ophthalmological evaluation consisted of Best corrected visual acuity (BCVA), slit lamp examination, fundoscopy, OCT, and fluorescein- and indocyaningreen- angiography before and at months 3, 6, 12, 24, and the last follow-up after treatment. Results: Initially, all patients had active posterior uveitis with occlusive (n = 5 patients) and nonocclusive retinal vasculitis (n = 12 patients). Mean follow up was 28 ± 15 months. Therapy was started with systemic corticosteroids (mean prednisolone equivalent 71.3 mg/d) and already after 3 months it could be tapered to a mean maintenance dosage of 8.63 mg/d. Additional immunosuppressive treatment with cs- or bDMARDs was initiated in 14 patients (82%) due to recurrences of uveitis while tapering the corticosteroids <10 mg per/day or because of severe inflammation at the initial visit. While being on immunosuppression, best corrected visual acuity increased from 0.56 logMAR to 0.32 logMAR during follow-up and only three patients had one uveitis relapse, which was followed by switch of immunosuppressive treatment. As recommended, TB prophylaxis with 300 mg/d isoniazid was administered in 11 patients for at least 9 months while being on TNF-alpha-blocking agents. No patient developed active tuberculosis during immunosuppressive therapy. Conclusion: Mainly conventional ATT is strongly recommended—as monotherapy or in combination with immunosuppressives—for effective treatment in patients with uveitis due to latent TB. Although in our patient group no conventional ATT was initiated, immunosuppression alone occurred as an efficient treatment. Nevertheless, due to possible activation of TB, isoniazid prophylaxis is mandatory in latent TB patients while being on TNF-alpha blocking agents.


Introduction
Tuberculosis (TB) is a worldwide spread bacterial infectious disease. It is caused by various types of mycobacteria (Mycobacterium tuberculosis complex; it most frequently affects the lungs as pulmonary tuberculosis and might involve also any part of the eye with or without other primary foci in the body [1].

Patients and Methods
We performed a retrospective analysis of all posterior uveitis patients with LTBI from 2016 to 2020 at the Department of Ophthalmology at the University Medical Center Hamburg-Eppendorf, Germany. We included 17 patients (mean age 54 ± 16.82 years; with a range from 34 to 82 years of age; female:male = 6:11) with posterior uveitis associated with occlusive or nonocclusive retinal vasculitis, diagnosed according the SUN criteria [12][13][14][15]. Additionally, the origin of our patients was evaluated (Figure 1).
The primary question was whether immunosuppressive treatment without conventional ATT is effective in LTBI patients with posterior uveitis. In addition, we ruled out if activation of TB occurred under immunosuppression, when only isoniazid prophylaxis (300 mg/d) was added, in case of initiating TNF-alpha-blocking therapy.
Secondary endpoints were the number of uveitis recurrences while tapering the systemic corticosteroids or immunosuppressives; the dosage of systemic corticosteroids; the change in best-corrected visual acuity (BCVA), the number of patients in whom additional immunosuppression was required, the number and severity of adverse events (AEs), and the number of patients requiring dose reduction or discontinuation of systemic corticosteroids or immunosuppression due to adverse events.

Treatment
In collaboration with the Department of Infectious Diseases at the University Medical Center Hamburg-Eppendorf, we developed an immunosuppressive regime without primary 3-to 4-drug-antituberculostatic therapy based on the rationale that uveitis in latent TB is mainly caused by an immunological reaction.
The primary question was whether immunosuppressive treatment without conventional ATT is effective in LTBI patients with posterior uveitis. In addition, we ruled out if activation of TB occurred under immunosuppression, when only isoniazid prophylaxis (300 mg/d) was added, in case of initiating TNF-alpha-blocking therapy.
Secondary endpoints were the number of uveitis recurrences while tapering the systemic corticosteroids or immunosuppressives; the dosage of systemic corticosteroids; the change in best-corrected visual acuity (BCVA), the number of patients in whom additional immunosuppression was required, the number and severity of adverse events (AEs), and the number of patients requiring dose reduction or discontinuation of systemic corticosteroids or immunosuppression due to adverse events.

Treatment
In collaboration with the Department of Infectious Diseases at the University Medical Center Hamburg-Eppendorf, we developed an immunosuppressive regime without primary 3-to 4-drug-antituberculostatic therapy based on the rationale that uveitis in latent TB is mainly caused by an immunological reaction.
At first, all patients received oral prednisolone equivalent at an initial dose of usually 1-2 mg/kg bodyweight at the primary visit, followed by weekly tapering to a maintenance dosage of 7.5 mg/d, respectively, 5.0 mg/d (Figure 2). In case of relapse while patients were on systemic corticosteroids at a dosage of ≥10 mg prednisolone equivalent immunosuppression with conventional Disease-Modifying Drugs (csDMARDs), methotrexate (MTX) or azatioprine (Aza) were initiated. In case of further relapse after 3 months of treatment with csDMARDs, the biological DMARD (bDMARD) adalimumab was applied additionally (because the TNF-alpha blocking agent adalimumab (ADA) is approved for posterior uveitis in Germany). In uveitis patients with severe ocular inflammation, csDMARDs or adalimumab were initiated immediately. In all patients who are on adalimumab and having no contraindications against MTX, low-dose methothrexate (5 mg weekly) was added to avoid the formation of antidrug-antibodies against adalimumab; in addition, for every patient who was on ADA treatment and, in some cases with other immunosuppressants, isoniazid (INH)-prophylaxis (300 mg/d) was added, because the administration of TNFalpha blockers and also other csDMARDs increase the risk of TB, especially by reactivating LTBI. The guidelines of the German society of rheumatology recommends prophylaxis with INH (300 mg/d) for at least 9 months starting 4 weeks prior to the initiation of adalimumab if possible [16,17].  Figure 2). In case of relapse while patients were on systemic corticosteroids at a dosage of ≥10 mg prednisolone equivalent immunosuppression with conventional Disease-Modifying Drugs (csDMARDs), methotrexate (MTX) or azatioprine (Aza) were initiated. In case of further relapse after 3 months of treatment with csDMARDs, the biological DMARD (bDMARD) adalimumab was applied additionally (because the TNF-alpha blocking agent adalimumab (ADA) is approved for posterior uveitis in Germany). In uveitis patients with severe ocular inflammation, csDMARDs or adalimumab were initiated immediately. In all patients who are on adalimumab and having no contraindications against MTX, low-dose methothrexate (5 mg weekly) was added to avoid the formation of antidrug-antibodies against adalimumab; in addition, for every patient who was on ADA treatment and, in some cases with other immunosuppressants, isoniazid (INH)-prophylaxis (300 mg/d) was added, because the administration of TNFalpha blockers and also other csDMARDs increase the risk of TB, especially by reactivating LTBI. The guidelines of the German society of rheumatology recommends prophylaxis with INH (300 mg/d) for at least 9 months starting 4 weeks prior to the initiation of adalimumab if possible [16,17].  Efficacy of treatment was assessed by standard ophthalmic examination techniques, including measurement of visual acuity (VA in logMAR) in both eyes (VA of hand motion and counting fingers was quantified with the Freiburg Visual Acuity Test [18]), slit-lamp examination, indirect binocular ophthalmoscopy, fluorescein and/or indocyanin green angiography, and Optical Coherence Tomography (OCT) imaging to determine disease severity. The examinations were performed before and in months 3, 6, 12, 18, and 24 after treatment initiation; in patients with a follow up of more than 24 months, a last follow-up visit was added. Efficacy of treatment was assessed by standard ophthalmic examination techniques, including measurement of visual acuity (VA in logMAR) in both eyes (VA of hand motion and counting fingers was quantified with the Freiburg Visual Acuity Test [18]), slit-lamp examination, indirect binocular ophthalmoscopy, fluorescein and/or indocyanin green angiography, and Optical Coherence Tomography (OCT) imaging to determine disease severity. The examinations were performed before and in months 3, 6, 12, 18, and 24 after treatment initiation; in patients with a follow up of more than 24 months, a last follow-up visit was added.

Statistical Analysis
In this study, only 4 patients had a full set of data, so our main problem for statistical analysis was the missing data for follow up visits, because traditional statistical methods would discard all patients with at least one missing value. To efficiently deal with this problem, we analyzed the changes in best-corrected visual acuity (BCVA in logMar) and in systemic corticosteroid treatment via the mixed regression model by Pinheiro and Bates (2006). This method allows us to use all data available, and not only nonmissing time trajectories, and to perform multivariate analysis with interaction terms. The overall effect of time or interaction between a covariate and time was assessed with the ANOVA method. In cases where it was considered necessary, we repeated the estimation using robust regression method of Koller (2016) [19]. The pairwise comparisons were performed, and p values were adjusted via Tukey adjustment method. A p-value of ≤0.05 was considered statistically significant.
All analysis were performed with R Core Team (2019) [20].

Ocular Involvement
Before treatment initiation, all 17 LTBI patients were diagnosed with active posterior uveitis (chorioretinitis) associated with occlusive retinal vasculitis (n = 5 patients) or nonocclusive retinal vasculitis (n = 12 patients). Mean follow-up was 28.0 ± 15.0 months. Altogether, 30 eyes were affected, in 13 patients it was both eyes, while 4 patients had only unilateral involvement.

Corticosteroids
Treatment was initiated with systemic corticosteroids (1-2 mg/kg bodyweight prednisolone equivalent orally or intravenously) in 16 patients (94%). One patient did not tolerate the systemic steroids and received monotherapy with adalimumab instead. During follow-up, in only 3 patients (19%) was corticosteroid monotherapy was effective. Figure 3 demonstrates the mean dosage of corticosteroids at treatment initiation and the mean decrease during a follow-up time of two years.
The mean initial dose of systemically applied corticosteroids in our patients was 71.03 mg/d (standard (std.) error 1.695). At month 3, there was a significant decrease to a mean dosage of 8.63 mg/d (std. error 1.85, p < 0.001) and a final mean dosage of 2.83 mg/d at month 24.
After a follow-up of 6 months, treatment was changed from azathioprine to MTX due to side effects (elevated liver enzymes) in patient no. 8.
At month 12, one patient's (pat.no. 12) therapy was switched from MTX to azathioprine (due to severe hair loss) and in two patients (pat.no. 13,14) from MTX to adalimumab in combination with low-dose MTX (5 mg/week); in patient 13, this had to be done due to a uveitis relapse, and in patient 14, low-dose MTX was added to prevent the formation of antibodies against adalimumab.
At month 18, another patient (pat.no. 7) received additional low-dose MTX to his ADA treatment for reducing the risk of formation of antidrug-antibodies against ADA.
At month 24, patient no. 4 stopped low-dose corticosteroid treatment on his own because of personal problems due to weight gain. Patient no.1 stopped MTX treatmentalso on his own-due to gastrointestinal problems; then, he restarted it upon our request and reported no further problems.
At the last visit, after 36 months, in patient no. 2, azathioprine treatment was switched to adalimumab in combination with low-dose MTX (due to persisting macular edema). After 48 months, pat. no. 1 was able to stop all immunosuppressive treatments due to complete remission. In 10 patients (59%), immunosuppression in addition to corticosteroid treatment has to be started at the primary visit: in four patients (pat.no. 2, 3, 6, 10), azathioprine in a dosage of 150 mg/d orally; in one patient (pat.no. 11), adalimumab 40 mg every 2 weeks subcutaneously; and in four patients (pat. no. 1,12,15,16), methotrexate (MTX) 10-25 mg/week subcutaneously. One patient (pat.no. 17) was already on cyclosporine A (CSA 3 mg/kg bodyweight daily) because of an earlier corneal transplant. Another patient (pat.no. 14) did not tolerate the systemic steroids and received monotherapy with adalimumab instead. In six patients (pat.no. 4,5,7,8,9,13), corticosteroid monotherapy was effective-at least for the first 3 months.
At the 3-month visit, in three patients who had a uveitis relapse while tapering the corticosteroid monotherapy, additional therapy with azathioprine (pat.no. 8, 13) or adalimumab in combination with low-dose MTX (5 mg/week) (pat.no. 7) was initiated.
After a follow-up of 6 months, treatment was changed from azathioprine to MTX due to side effects (elevated liver enzymes) in patient no. 8.
At month 12, one patient's (pat.no. 12) therapy was switched from MTX to azathioprine (due to severe hair loss) and in two patients (pat.no. 13,14) from MTX to adalimumab in combination with low-dose MTX (5 mg/week); in patient 13, this had to be done due to a uveitis relapse, and in patient 14, low-dose MTX was added to prevent the formation of antibodies against adalimumab.
At month 18, another patient (pat.no. 7) received additional low-dose MTX to his ADA treatment for reducing the risk of formation of antidrug-antibodies against ADA.
At month 24, patient no. 4 stopped low-dose corticosteroid treatment on his own because of personal problems due to weight gain. Patient no.1 stopped MTX treatment- One patient (pat.no. 16) was lost for follow-up after 3 months, two patients (pat.no. 9, 17) were lost for follow-up after 6 months, and one patient (pat.no. 15) after 12 months.

Efficacy of Immunosuppression: Relapses of Uveitis/Steroid-Sparing Effect
During corticosteroid monotherapy initiation of an additional treatment with csD-MARDs was necessary in three patients (pat.no. 7, 8, 13) at month 3 because of a uveitis relapse while tapering the corticosteroids to ≤10 mg prednisolone equivalent daily. At month 18, one patient (pat.no. 13) needed to be switched to ADA (with low-dose MTX) because of recurrence of uveitis during azathioprine treatment for 9 months. At the last visit, at month 36, in patient no. 2, azathioprine was switched to adalimumab in combination with low-dose MTX due to persisting macular edema.
Pat.no. 1 had recurrence of his uveitis due to stopping immunosuppression on his own after 9 months, and on the patient's request he received high-dose corticosteroids in addition to MTX. After 18 months, he was able to stop all immunosuppressants.
Altogether, during corticosteroid monotherapy, there were three uveitis relapses; during treatment with csDMARDs, two patients' treatments had to be switched due to recurrence of uveitis and in one patient due to chronic persisting CME. No relapse occurred in all patients who were on ADA. Another sign for the efficacy of immunosuppression is its steroid-sparing effect; at treatment initiation, patients had a mean dosage of systemic corticosteroids of 71.03 mg/d and, at the end of follow-up, the dosage was decreased significantly by 68.27 mg to a mean of 2.83 mg/d (Figures 2 and 4).

Side Effects of Systemic Therapy
One patient reported a single episode of pressure in the throat after taking azathioprine. The patient continued the medication and the pressure subsided completely (pat.no. 6).
Due to elevated liver enzymes, azathioprine was switched to MTX (pat.no. 8, 10) and to adalimumab in two patients (pat.no. 2, 13). During MTX treatment, one patient developed severe hair loss; therefore, the medication was changed to azathioprine (pat.no. 12). Another patient (pat.no. 1) stopped MTX on his own due to gastrointestinal problems and continued it later without further problems. One patient reported reactivation of dermatological herpes simplex virus (HSV) infection under adalimumab. He was treated with systemic aciclovir and immunosuppression could be continued with no further HSV recurrence (pat.no. 7).

Visual Acuity
The mean best-corrected visual acuity (BCVA) of all patients at baseline was 0.56 logMar (Logarithm of the Minimum Angle of Resolution) The standard deviation between the eyes was 0.51 (LogMar) and within the eyes 0.29 (LogMar). The overall effect of time was significant (p = 0.045). Visual acuity (VA) at month 3 and at month 18 increased significantly in comparison to VA before treatment initiation. Other pairwise comparisons during treatment at months 6, 12, and 24 were not statistically significant (p > 0.05) (Figure 5).
At the end of the observation period, the mean BCVA of all patients was 0.32 logMar.

Visual Acuity in Patients with Occlusive Retinal Vasculitis versus Patients with Nonocclusive Retinal Vasculitis
Baseline BCVA in the nonocclusion group was 0.435 (Std. error. 0.18) and in the occlusion group, 0.8, p = 0.103. The BCVA in the nonocclusion group was 0.18 and in the occlusion group, 0.9. The differences between the groups were significant using robust approach ( Figure 6). Due to the ischemic nature of occlusive retinal vasculitis, the VA prognosis is known to be worse compared with patients with nonocclusive retinal vasculitis. In our group, five patients had occlusive retinal vasculitis (pat.no. 3, 12, 13, 16, 17) and 12 patients had nonocclusive retinal changes. The cases with occlusive retinal vasculitis (occlusion group) have worse vision trajectory. The overall difference was significant, F (1,28) = 7.96, p = 0.012. They had negative changes when compared with cases with nonocclusive retinal vasculitis (nonocclusion group), interaction F (5117) = 3.22, p = 0.0092.
All intervals during treatment were significantly different to baseline in cases with no occlusion, p < 0.05. No significant change was observed in the occlusion group, p > 0.05.

Visual Acuity
The mean best-corrected visual acuity (BCVA) of all patients at baseline was 0.56 log-Mar (Logarithm of the Minimum Angle of Resolution) The standard deviation between the eyes was 0.51 (LogMar) and within the eyes 0.29 (LogMar). The overall effect of time was significant (p = 0.045). Visual acuity (VA) at month 3 and at month 18 increased significantly in comparison to VA before treatment initiation. Other pairwise comparisons during treatment at months 6, 12, and 24 were not statistically significant (p > 0.05) (Figure 5).
At the end of the observation period, the mean BCVA of all patients was 0.32 logMar.  litis. In our group, five patients had occlusive retinal vasculitis (pat.no. 3, 12, 13, 16, 17) and 12 patients had nonocclusive retinal changes. The cases with occlusive retinal vasculitis (occlusion group) have worse vision trajectory. The overall difference was significant, F (1,28) = 7.96, p = 0.012. They had negative changes when compared with cases with nonocclusive retinal vasculitis (nonocclusion group), interaction F (5117) = 3.22, p = 0.0092. All intervals during treatment were significantly different to baseline in cases with no occlusion, p < 0.05. No significant change was observed in the occlusion group, p > 0.05.  (Note: Due to some severe cases, we also repeated the analysis with robust estimation. The trend of the occlusion group having worse vision was confirmed by this method as well. The difference at baseline, which lacked statistical significance, was found significant using robust approach).

Other Complications
Seven patients (pat.no. 2, 9,10,11,12,13,16) showed retinal scarring. In one patient (pat.no. 2), the scarring was caused by vitrectomy, which had been performed even before treatment initiation, and only one patient (pat.no. 17) developed secondary glaucoma during the course of disease, which remained stable during locally applied brimonidine therapy.

Study Design
Different study groups [9][10][11][12] have concluded that 4-drug ATT is necessary to reduce the recurrences of uveitis. Bansal et al. [12] performed a retrospective interventional case series comparing two groups. Group 1 received 4-drug ATT and group 2 received corticosteroid monotherapy only. Their data demonstrated that uveitis relapse was statistically significantly less likely in the first group. They therefore concluded that 4-drug ATT reduced the risk of recurrence compared with corticosteroid monotherapy. However, they did not apply any additional immunosuppressive treatment in case of uveitis relapse

Other Complications
Seven patients (pat.no. 2, 9,10,11,12,13,16) showed retinal scarring. In one patient (pat.no. 2), the scarring was caused by vitrectomy, which had been performed even before treatment initiation, and only one patient (pat.no. 17) developed secondary glaucoma during the course of disease, which remained stable during locally applied brimonidine therapy.

Study Design
Different study groups [9][10][11][12] have concluded that 4-drug ATT is necessary to reduce the recurrences of uveitis. Bansal et al. [12] performed a retrospective interventional case series comparing two groups. Group 1 received 4-drug ATT and group 2 received corticosteroid monotherapy only. Their data demonstrated that uveitis relapse was statistically significantly less likely in the first group. They therefore concluded that 4-drug ATT reduced the risk of recurrence compared with corticosteroid monotherapy. However, they did not apply any additional immunosuppressive treatment in case of uveitis relapse while tapering the corticosteroids. In contrast, in 82% of our patients (14 patients), additional immunosuppression was necessary.
Ang et al. compared uveitis patients who received only ATT for 9 months to patients who discontinued the medication before [10,11]. However, they did not specify the exact medications that were given to the patients, but no additional immunosuppressive therapy was initiated. Only Tomkins-Netzer et al. compared LTBI uveitis patients who received ATT to patients who were treated with corticosteroids or, in 9% of these patients, with corticosteroids in association with immunosuppressants. However, this publication did not differentiate between patients receiving corticosteroid monotherapy or treatment combinations. In addition, the authors did not mention which immunosuppressants were administered in their patients. Altogether, uveitis relapses were statistically significantly less likely in patients treated with ATT (Tomkins-Netzer) [9].
In our study, only patients in whom systemic corticosteroids were not sufficient or severe inflammation was already present at the primary visit received additional medication. This potentially avoids serious side effects in 20% of the patients because serious adverse reactions to antituberculosis drugs are common [21] and immunosuppressants usually have less-severe side effects and are better tolerated than regular ATT. In addition, ATT is commonly recommended for 3-9 months in patients with LTBI [22,23]. There are no official recommendations for treatment duration of ATT in uveitis patients with LTBI. So, in real life, ATT application for 9 months and more seems to be the consensus in such uveitis patients. Serious side effects of ATT are observed in about 7% of patients, including hepatotoxicity (6%) and skin rash (0.6%). Gastrointestinal side effects, ocular toxicity, angioedema, and other side effects are comparatively rare with less than 0.4%. However, early diagnosis of ocular toxicity, which primarily affects the optic nerve, is crucial to prevent the potential loss of function [24][25][26]. Isoniazid 600-750 mg/d can cause neurological toxicity and hepatitis. Although much less frequent, hepatotoxicity can lead to liver transplantation or even to death [26,27]. The risk of hepatotoxicity due to isoniazid treatment in patients older than 65 years may be increased 3-up to 5-fold [28]. Rifampicin may cause gastrointestinal reactions or thrombocytopenic purpura and, rarely, shortness of breath, shock, acute hemolytic anemia, and acute renal failure. Pyrazinamide can lead to arthralgia and hepatitis. Ethambutol may cause a dose-related retrobulbar neuritis.

Methods of TB Testing
Bansal et al. and Ang et al. [10][11][12] used the tuberculin skin test (TST) with purified protein derivation (PPD) or recombinant purified protein derivates. Intradermal injection leads to delayed-type hypersensitivity between 48 and 72 h. The test is susceptible to placement errors, reading errors, false negatives in anergic patients, false positives in Bacille Calmette-Guérin (BCG)-vaccinated patients, the booster phenomenon (where repeated testing induces a positive result), and noncompliance, as it requires two visits to the physician. The test has a low specificity and sensitivity (reported sensitivity of 71% (95% CI (confidence interval), 65-74) and specificity of 66% (95% CI, 46-86) [12].
However, the gold standard in TB testing is the QuantiFERON ® -TB Gold test. After presentation of mycobacteria-specific antigens (ESAT-6, CFP-10, TB 7.7) the interferon production of specific T-lymphocytes of the patient is measured. The antigens used are largely specific to the Mycobacterium tuberculosis complex, but not to the BCG vaccine strain. Thus, an earlier vaccination does not lead to a positive test result. The QuantiFERON ® -TB Gold test has a specificity of 99% [29]. Therefore, TB testing, which we used in this study design, is clearly superior. Alternatively, the T-SPOT.TB test (T-Spot) can be used. It uses the Enzyme-linked immunospot (ELISPOT) methodology to similarly count the T cells that have been sensitized by TB infection. Sensitivity and specificity vary between different studies. The sensitivity of QuantiFERON ® of the new generation is higher than that of the older generation (94% vs. 81%), thus reaching the sensitivity of T-SPOT ® .TB [7].

Recommended Therapy of Latent Tuberculosis
The presumed pathogenesis of uveitis related to latent TB infection consists of inflammation caused by presence of low numbers of bacteria within the eye with or without superimposed immune reaction to mycobacterial or ocular antigens [5]. This suggests that immunosuppression might be the superior therapeutic strategy in that case. This thesis is also supported by the WHO guidelines for latent tuberculosis.
The main principle in guiding testing and treatment for LTBI is that the benefit outweighs the risk to the individual, while the WHO guidelines strongly recommend systematic testing and treatment of LTBI for people living with HIV, adult and child contacts of pulmonary TB cases, patients initiating antitumor necrosis factor inhibition treatment, patients receiving dialysis, patients preparing for organ or hematological transplantation, and patients with silicosis. Treatment and testing in any other group in countries with low incidence of TB (<10 cases per 100,000 population per year) is not generally recommended. In 2019, 54 countries had a low incidence of TB, mostly in the American and European region, plus a few countries in the Eastern Mediterranean and Western Pacific regions [2].
In contrast to tuberculosis requiring treatment, the germ populations in LTBI are probably very small. This assumption is based on the results of animal studies; however, it is unclear to what extent the bacterial burden of the individual can be reduced in LTBI, as there are no suitable methods to accurately quantify the intracellular pathogens. The probability of spontaneous resistance mutation of M. tuberculosis in the presence of LTBI in humans is considered very low. Therefore, monotherapy of LTBI is sufficiently safe (Exception: assumed infection by INH-or RMP-resistant pathogens) [30].

Efficacy of Preventive Therapy for LTBI
A sufficient number of controlled trials are available to assess the efficacy of preventive therapy with INH, and each has been analyzed in a Cochrane meta-analysis for both HIVpositive [32] and HIV-negative individuals [33]. Treatment dosage varied among the trials: INH 300 mg or 600 mg for daily regimens; 600 mg or 900 mg for twice weekly regimens [32,33].
The number needed to treat (NNT) chemo-preventively to prevent tuberculosis ranges from 30 to 89 in immunocompetent individuals and from 14 to 80 in immunocompromised individuals [18]. It is possible that therapeutic regimens based on Rifamycine derivatives are even more effective than INH monotherapy for the preventive treatment of LTBI [18].
As the recommendations state, preventive LTBI therapy is necessary when initiating treatment with antitumor necrosis factor [17] and the SAFEBIO study [18] demonstrated that even csDMARDs and high-dose systemic corticosteroids can reactivate TBC; all patients with adalimumab and some patients with corticosteroids and/or csDMARDs in our study received 300 mg isoniazid daily for at least 6 months. In total, 11 patients were medicated with preventive INH therapy in addition to immunosuppressive treatment and none of our study patients disclosed TB reactivation.

Efficacy of Immunosuppressive Therapy in Uveitis Patients with LTBI
The mean best-corrected visual acuity (BCVA) at baseline was 0.56 ± 0.51 logMar and increased significantly to 0.36 logMar at the last follow-up. The mean initial dose of systemically applied corticosteroids in our patients was 71.03 mg/d (standard (std.) error 1.695) and, at month 3, there was a significant decrease to a dose of 8.63 mg/d (std. error 1.85, p < 0.001) and to a dose of 2.83 mg/d at month 24. In our study, three patients had only one uveitis relapse while they were on corticosteroid monotherapy, during treatment with csDMARDs. In only two patients did treatment have to be switched due to recurrence of uveitis; in patients who were on ADA, no relapse occurred. Altogether, five relapses occurred during immunosuppressive treatment in 17 patients during a mean follow-up at 28 months; thus, 71% of our patients had no relapse. Ang et al. [7] only included patients that did not have a recurrence of uveitis 1 year after 4-drug ATT therapy, which was administered for at least 6 months. In their study about the duration of ATT, they achieved that only 14.4% of patients who applied corticosteroid monotherapy had no recurrence of their uveitis while 24% of patients who received 4-drug ATT did not have a recurrence after 6 months after completing the therapy [8]. Bansal et al. found that patients treated with 4-drug ATT and systemic corticosteroids had a 16% chance of a uveitis relapse while patients treated with corticosteroids alone had a likelihood of 47% [12]. The patient groups of Tomkins-Netzer et al. [9] disclosed a 30% rate of uveitis relapse in the ATT-treated group, and in 48% of patients who received corticosteroid as monotherapy or in combination with immunosuppressants.

Conclusions
In our study, 18% of patients improved with corticosteroid monotherapy alone, sparing those patients additional side effects caused by stronger immunosuppression. A total 82% of patients received immunosuppressive therapy with cs-or bDMARDs in addition to systemic corticosteroids, when corticosteroid monotherapy was not sufficient. Eleven patients (65%) received additional isoniazid prophylaxis with 300 mg/d, while six patients did not receive any antituberculostatic therapy. Our patients had significant improvement of BCVA. Of our patients, 71% had no uveitis relapse; in only five patients, a recurrence of uveitis occurred during the follow-up of 28 months, which made a treatment switch necessary. The mean dosage of systemic corticosteroids could be significantly reduced to a mean of 2,76 mg/d at the end of follow-up. Therapy with corticosteroids and immunosuppressives with additional INH prophylaxis proved to be a safe treatment option in patients with uveitis in latent tuberculosis. No patient developed active tuberculosis during the follow-up period.
Our data demonstrate that additional 3-to 4-drug ATT seems not to be necessary for efficient uveitis treatment in patients with latent tuberculosis. Especially, when comparing the recurrence rates with ATT-treatment studies [8][9][10][11][12], we could achieve the lowest rate with 71% of relapse-free patients. This study further supports the thesis that uveitis in latent tuberculosis is rather caused by immunoreactions to mycobacterial or ocular antigens in the presence of only low numbers of tuberculosis bacteria and not by infection.
It is most important to keep in mind that uveitis patients with LTBI must have INH prophylaxis with 300 mg/d for at least 9 months in case of initiating TNF-alpha-inhibiting treatment [14,18]. Whether INH prophylaxis should be initiated in every patient diagnosed with LTBI is also controversial, because in low TB burden countries, patients without risk factors to develop active TB, such as an HIV infection, are not likely to develop active TB. Most probably, INH prophylaxis is not necessary in every patient with uveitis posterior due to latent TB. Further, in our study, six LTBI patients did not receive INH prophylaxis and none of them developed active TB.
Interestingly, the WHO guidelines on the therapy for LTBI in countries with a low prevalence of TB only recommend Isoniazid or Rifampicin monotherapy or a combination of both, instead of a 3-or 4-drug ATT [18]. However, until today, this therapy regime has not been widely recognized by ophthalmologists, maybe due to the fear of TB activation. So, since Isoniazid monotherapy (300-750 mg/d) [30] for 6-9 months is sufficient treatment for LTBI itself, the administration of INH might also have an impact on the uveitis.
Since this cohort is rather small, further investigations are needed to support this thesis. Institutional Review Board Statement: The study was approved by the medical ethics committee of the Ärztekammer Hamburg, Germany and followed the tenets of the declaration of Helsinki for research involving human subjects. Ethical committee approval No. 2020.10344-BO-ff.
Informed Consent Statement: Informed consent for publication was obtained from all participants of the study using institutional consent forms.

Conflicts of Interest:
The authors declare no conflict of interest.