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

Abstract

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.

Keywords: antituberculosis therapy; immunosuppression; latent tuberculosis; posterior uveitis; retinal vasculitis; uveitis.

Figure 1

Shows the distribution of countries where the patients originated: Germany, 29.4% (n = 5); Russia, 11.8% (n = 2); Turkey, 11.8% (n = 2); Afghanistan, 5.88% (n = 1); Bahrain, 5.88% (n = 1); Brazil, 5.88% (n = 1); Ghana, 5.88% (n = 1); Lebanon, 5.88% (n = 1); Portugal, 5.88% (n = 1); Romania, 5.88% (n = 1); Saudi Arabia, 5.88% (n = 1).

Figure 2

Shows the patients’ daily demand of systemic corticosteroids (mean dosage in mg/d and their 95% confidence intervals (CI)) at the primary visit FUO (71.03 mg/d) and the mean decrease at FU1 (3-month visit: 8.63 mg/d), FU2 (6-month visit: 6.72 mg/d), FU3 (12-month visit: 5.09 mg/d), FU4 (18-month visit: 4.5 mg/d), and FU5 (24-month visit:2.83 mg/d); m = months.

Figure 3

Shows the patients’ daily demand of systemic corticosteroids (mean dosage in mg/day and their 95% confidence intervals) in contrast with the initial visit FUO (71.03 mg/d) and the mean decrease at FU1 (month 3 visit: −62.48 mg/d), FU2 (month 6 visit: −64.31 mg/d), FU3 (month 12 visit: −65.94 mg/d), FU4 month 18 visit: −66.53 mg/d), and FU5 (month 24 visit: −68.27 mg/d); m = months. p value *** 0.001.

Figure 4

Shows the different treatment options and the switch of therapy of each patient during the follow up in months (m). Each color represents a different therapy; the larger dots indicate a switch of immunosuppression. Patient no. 9, 15, 16, 17 did not have complete follow-up.

Figure 5

The estimates and their 95% confidence intervals in BCVA (logMar) from baseline. A total 157 BCVA data were used for the analysis. The overall effect of time was significant, F (5122) = 2.35, p = 0.045. Visual acuity at FU1 (month 3) and FU4 (month 18) was significantly different from initial visual acuity when performing pairwise comparisons between all time intervals and adjusting p-values. Other pairwise comparisons were not statistically significant p > 0.05. The data were analyzed via mixed regression modeling @pinheiro2006mixed. (Note: Some severe cases with bad vision caused decreased mean values for visual acuity, which led to large standard errors and, thus, less statistical significance. For this reason we applied robust mixed regression as well to confirm the findings. In this model, all mean values during treatment were statistically significantly different from baseline but the magnitude of change was somewhat less than estimated with the nonrobust method. All changes during treatment were estimated to be of around 1 Snellen line (−0.1 LogMar change)). FU1: 3-month visit (3 m), FU2: 6-month visit, (6 m) FU3: 12-month visit 12 m), FU4: 18-month visit (18 m), FU5: 24-month visit (24 m), BCVA: best-corrected visual acuity, logMAR: Logarithm of the Minimum Angle of Resolution. p value * 0.05 ** 0.01.

Figure 6

Shows the BCVA in patients with nonocclusive retinal vasculitis (green line) compared patients with occlusive retinal vasculitis (red line). The cases with occlusion have worse vision trajectory. The overall difference was significant, F (1,28) = 7.96, p = 0.012, where cases with occlusion had overall worse vision. Furthermore, they have negative changes compared with cases without occlusion (interaction, p = 0.0092). The data were analyzed with the ANOVA method. All postoperative intervals were significantly different to baseline in cases with no occlusion, p < 0.05. No significant change was observed in the group with occlusion, p > 0.05. Baseline BCVA in the nonocclusion group was 0.435 (Std. error. 0.18) and in the occlusion group, 0.8, p = 0.103. The differences between the groups at postoperative visits were all significant, with occlusion group having worse vision, 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).

Figure 7

(A–C) OCT of the right eye (OD) of patient no. 13 showing cystoid macular edema before treatment (A), after 4 months while being on 30 mg/d corticosteroids and azathioprine 150 mg/d (B), and at the last follow-up after 41 months during adalimumab 40 mg every 2 weeks sc, methotrexate 10 mg/weekly, and 1 mg/d prednisolone (C). Macular edema regressed completely over time.

Figure 8

(A–D). Fundus OD of patient no. 13. (A) Active occlusive retinal vasculitis before initiating anti-inflammatory therapy. (B) Demonstrates the fundus appearance at month 24 after intravitreal Dexamethason and Bevacizumab injections and panretinal laser coagulation, due to CME and retinal neovascularizations resulting from massive retinal ischemia. BCVA was 0.1 logMar and the patient received immunosuppressive treatment with low-dose corticosteroids (7.5 mg/d), adalimumab (40 mg sc every other week), and low-dose MTX (5 mg weekly). (C,D) Fluoresceine angiography (FAG) of OD at the initial visit showing severe occlusive retinal vasculitis at the central posterior pole (C) and massive ischemia with nonperfusion areas and retinal neovascularizations at the posterior segment (D).

Figure 9

(A–D) Shows the retina of the left eye (OS) of patient 13 before and after therapy. (A) (fundus picture) and (B) (fluoresceine angiography (FAG) show occlusive retinal vasculitis. (C) (fundus picture) and (D) (FAG) display the OS after treatment, showing a normal fundus appearance and FAG. The BCVA was 0.0 logMar.