Induction failure in granulomatosis with polyangiitis: a nationwide case-control study of risk factors and outcomes.

OBJECTIVE: To identify characteristics of granulomatosis with polyangiitis (GPA) associated with induction failure, describe salvage therapies and their efficacy. METHODS: We conducted a nationwide retrospective case-control study of GPA with induction failure between 2006 and 2021. Each patient with induction failure was randomly paired to three controls matched for age, sex and induction treatment. RESULTS: We included 51 patients with GPA and induction failure (29 men and 22 women). At induction therapy, median age was 49 years. Twenty-seven patients received intravenous cyclophosphamide (ivCYC) and 24 rituximab (RTX) as induction therapy. Patients with ivCYC induction failure more frequently had PR3-ANCA (93% vs 70%, P = 0.02), relapsing disease (41% vs 7%, P < 0.001) and orbital mass (15% vs 0%, P < 0.01) compared with controls. Patients with disease progression despite RTX induction therapy more frequently had renal involvement (67% vs 25%, P = 0.02) with renal failure (serum creatinine >100 µmol/l in 42% vs 8%, P = 0.02) compared with controls. After salvage therapy, remission was achieved at 6 months in 35 (69%) patients. The most frequent salvage therapy was switching from ivCYC to RTX (or vice versa), showing an efficacy in 21/29 (72%). Remission was achieved in nine (50%) patients with inappropriate response to ivCYC, while in patients with progression after RTX induction, remission was achieved in four (100%) who received ivCYC (with or without immunomodulatory therapy), but only in three (50%) after adding immunomodulatory therapy alone. CONCLUSION: In patients with induction failure, characteristics of GPA, salvage therapies and their efficacy vary according to induction therapy and failure modality.

Lower Relapses Rate With Infliximab Versus Adalimumab in Sight-Threatening Uveitis: A Multicenter Study of 330 Patients.

PURPOSE: To compare the relapse rate of sight-threatening noninfectious uveitis (NIU) in patients treated with infliximab (IFX) or adalimumab (ADA). DESIGN: Observational retrospective multicenter study. METHODS: A total of 330 patients (median age, 36 years; interquartile range, 27-54), 45.2% men) with sight-threatening NIU (ie, retinal vasculitis and/or macular edema) treated with anti-tumor necrosis factor [TNF]-α agents (IFX intravenously at 5 mg/kg at weeks 0, 2, 6, and every 4 to 6 weeks or ADA subcutaneously at 80 mg, then 40 mg every 2 weeks). Data were obtained retrospectively from patients' medical records. Main outcome measures were relapse rate, complete response of NIU, corticosteroid sparing effect, and safety. RESULTS: Main etiologies of uveitis included Behçet disease (27%), idiopathic juvenile arthritis (5.8%), and sarcoidosis (5.5%). The estimated relapse rate at 6 months after introduction of biological agents was 13% (95% CI = 0.009-0.16). IFX was associated with less relapse risk than ADA (hazard ratio [HR] = 0.52, 95% CI = 0.36- 0.77, P = .001). ADA and IFX were comparable in terms of complete response rate of NIU as well as corticosteroid-sparing effect. Behçet disease was associated with higher odds of complete response (HR = 2.04, 95% CI = 1.16 -3.60, P = .01] and lower relapse rate (HR = 0.53, 95% CI = 0.33-0.85, P = .009) than other causes of NIU with anti-TNF-α agents. CONCLUSIONS: In sight-threatening NIU, IFX seems to be associated with a lower relapse rate than ADA.

Anti-Tumor Necrosis Factor α versus Tocilizumab in the Treatment of Refractory Uveitic Macular Edema: A Multicenter Study from the French Uveitis Network.

PURPOSE: To analyze the factors associated with response (control of ocular inflammation and corticosteroid-sparing effect) to biologics (anti-tumor necrosis factor [TNF]-α agents and tocilizumab) in patients with refractory uveitic macular edema (ME). DESIGN: Multicenter, retrospective, observational study. PARTICIPANTS: Adult patients with uveitic ME refractory to systemic corticosteroids, disease-modifying antirheumatic drugs, or both. METHODS: Patients received anti-TNF-α agents (infliximab 5 mg/kg at week 0, 2, 6, and every 4-6 weeks [n = 69] and adalimumab 40 mg/2 weeks [n = 80]) and tocilizumab (8 mg/kg every 4 weeks intravenously [n = 39] and 162 mg/week subcutaneously [n = 16]). MAIN OUTCOME MEASURES: Analysis of complete and partial response rates, relapse rate, low vision (visual acuity in at least 1 eye of ≥ 1 logarithm of the minimum angle of resolution), corticosteroid-sparing effect, and adverse events at 6 months. RESULTS: Two hundred four patients (median age, 40 years [interquartile range, 28-58 years]; 42.2% men) were included. Main causes of uveitis included Behçet's disease (17.2%), birdshot chorioretinopathy (11.3%), and sarcoidosis (7.4%). The overall response rate at 6 months was 46.2% (21.8% of complete response) with anti-TNF-α agents and 58.5% (35.8% of complete response) with tocilizumab. In multivariate analysis, treatment with tocilizumab (odds ratio, 2.10; 95% confidence interval [CI], 1.06-4.06; P = 0.03) was associated independently with complete response of uveitic ME compared with anti-TNF-α agents. Anti-TNF-α agents and tocilizumab did not differ significantly in terms of relapse rate (hazard ratio, 1.00; 95% CI, 0.31-3.18; P = 0.99) or occurrence of low vision (odds ratio, 1.02; 95% CI, 0.51-2.07; P = 0.95) or corticosteroid-sparing effect (P = 0.29). Adverse events were reported in 20.6% of patients, including serious adverse events reported in 10.8% of patients. CONCLUSIONS: Tocilizumab seems to improve complete response of uveitic ME compared with anti-TNF-α agents.

Development and Validation of a Bayesian Network for Supporting the Etiological Diagnosis of Uveitis.

The etiological diagnosis of uveitis is complex. We aimed to implement and validate a Bayesian belief network algorithm for the differential diagnosis of the most relevant causes of uveitis. The training dataset (n = 897) and the test dataset (n = 154) were composed of all incident cases of uveitis admitted to two internal medicine departments, in two independent French centers (Lyon, 2003-2016 and Dijon, 2015-2017). The etiologies of uveitis were classified into eight groups. The algorithm was based on simple epidemiological characteristics (age, gender, and ethnicity) and anatomoclinical features of uveitis. The cross-validated estimate obtained in the training dataset concluded that the etiology of uveitis determined by the experts corresponded to one of the two most probable diagnoses in at least 77% of the cases. In the test dataset, this probability reached at least 83%. For the training and test datasets, when the most likely diagnosis was considered, the highest sensitivity was obtained for spondyloarthritis and HLA-B27-related uveitis (76% and 63%, respectively). The respective specificities were 93% and 54%. This algorithm could help junior and general ophthalmologists in the differential diagnosis of uveitis. It could guide the diagnostic work-up and help in the selection of further diagnostic investigations.

Characteristics of Adult Patients with Idiopathic Retroperitoneal Fibrosis and Assessment of Risk of Relapse at Diagnosis.

OBJECTIVES: To compare adult patients' characteristics suffering from idiopathic retroperitoneal fibrosis between "relapse-free" and relapsing patients at the diagnosis and identify factors associated with relapse at initial presentation. METHODS: We conducted a retrospective multicentric study in four hospitals in Eastern France, from 1993 to 2020, of adult patients suffering from idiopathic retroperitoneal fibrosis. We analyzed clinical, biological, and radiological features at diagnosis and during a forty-month follow-up. RESULTS: Of 47 patients suffering from retroperitoneal fibrosis, 21 patients had idiopathic retroperitoneal fibrosis. Among them, 13 experienced one or more relapses during follow-up. At diagnosis, clinical characteristics, relevant comorbidities, biological and radiological features were similar between groups. Smoking cessation seems associated with decreased relapse risk (p: 0.0624). A total of 8 patients developed chronic renal failure during follow-up. Ureteral infiltration at diagnosis was associated with evolution to chronic renal failure (p: 0.0091). CONCLUSION: No clinical, biological, or radiological features could predict relapse at retroperitoneal fibrosis diagnosis, but smoking cessation may prevent relapse.

Quality of life in patients with uveitis: data from the ULISSE study (Uveitis: cLInical and medico-economic evaluation of a Standardised Strategy for the Etiological diagnosis).

AIMS: To assess vision-related (VR-QOL) and health-related quality of life (HR-QOL) in a large series of patients with de novo uveitis at baseline and 6-month follow-up. METHODS: Non-inferiority, prospective, multicentre, cluster randomised controlled trial registered under the Unique Identifier: NCT01162070. VR-QOL and HR-QOL were assessed by the 25-item National Eye Institute Visual Function Questionnaire (VFQ-25) and the Medical Outcomes Study 36-item Short Form Survey (SF-36). RESULTS: At inclusion, 466 patients completed the VFQ-25. The mean composite score was 80.0 (±16.7). In multivariate analysis, higher age, female sex and insidious onset were significantly associated with lower QOL. At 6 months, 138 patients completed the VFQ-25, with a significantly higher mean composite score of 82.6 (±16.7). SF-36 mental component was 42.9 (±11.3) and physical component was 47.2 (±8.5) at inclusion (n=425). HR-QOL improvement at 6 months was not clinically significant. CONCLUSION: QOL seems relatively well preserved in this cohort; only VR-QOL improved significantly at 6 months, especially in patients with low initial visual acuity.

Evaluating the cost-consequence of a standardized strategy for the etiological diagnosis of uveitis (ULISSE study).

MAIN OBJECTIVE: To prospectively assess the cost-consequence of a standardized diagnostic strategy as to compared to an open one for the etiological diagnosis of uveitis. DESIGN: This was a prospective, non-inferiority, multicentre, randomized controlled trial. METHODS: We included all consecutive patients with uveitis who had visited at least one of the Departments of Ophthalmology. In the standardized group, patients had a minimal work-up regardless of the type of uveitis (including evaluation of the CBC, ESR, C-reactive protein, tuberculin skin test, syphilis serology and chest X-ray). Depending on ophthalmological findings, further investigations could be performed. In the open strategy, ophthalmologists were free to order any kind of investigation. The main outcome was the mean cost per patient of each strategy. RESULTS: 903 uveitis patients were included from January, 2010 to May, 2013. The mean cost per patient of the standardized strategy was 182.97 euros [CI 95% (173.14; 192.80)], and the mean cost per patient of the open strategy was 251.75 euros [CI 95% (229.24; 274.25)]. Therefore, the mean cost per patient of the standardized strategy was significantly lower than the mean cost per patient of the open strategy (p<0.001). There were significantly fewer visits (p<0.001), fewer radiological procedures (p<0.004) and fewer laboratory investigations (p<0.001) in the standardized group. CONCLUSION: A standardized strategy is a cost-saving approach for the etiological diagnosis of uveitis.

Tocilizumab for the Management of Corticosteroid-Resistant Mild to Severe Graves' Ophthalmopathy, a Report of Three Cases.

Purpose: To discuss the use of tocilizumab in mild to severe Graves' ophthalmopathy as corticosteroid-adjunctive therapy. Methods: Retrospective case reports.Results: Three patients with corticosteroid-resistant or advanced diplopia-associated Graves' ophthalmopathy were subsequently treated with monthly intravenous tocilizumab at a dose of 8 mg/kg. None reported a past or present history of dysthyroidism. The adjunction of interleukin-6-receptor monoclonal antibody treatment was associated with a significant improvement in ocular symptoms, notably diplopia and proptosis, and functional prognosis in all patients, with one relapse approximately two months after the end of the treatment.Conclusion: These clinical reports confirm the relative efficacy and tolerability profile of intravenous tocilizumab in severe or corticosteroid-resistant Graves' ophthalmopathy.

Osseous sarcoidosis: A multicenter retrospective case-control study of 48 patients.

OBJECTIVE: To describe the clinical presentation, distribution of lesions, treatment, and outcomes of osseous sarcoidosis. METHODS: A French retrospective multicenter study of patients with biopsy-proven sarcoidosis analyzed patients with 1) a biopsy-proven granuloma without caseous necrosis, and either 2) osseous clinical manifestations, or 3) abnormal osseous imaging. Sarcoidosis patients with osseous involvement (cases) were compared with 264 age- and sex-matched sarcoidosis patients with no osseous manifestations (controls). RESULTS: In the osseous sarcoidosis group (n=88), forty-two (48%) patients had osseous-related symptoms involving the axial (69%) and/or appendicular (58%) skeleton. On imaging, the most commonly affected bones were in the spine (52%), pelvis (42%), hands (22%) and femur (19%). Compared with controls, cases had higher rates of mediastinal (93% vs. 47%) and extra-thoracic lymph node involvement (66% vs. 21%), pulmonary (90% vs. 65%) and cutaneous involvement (44% vs. 23%) (all P<0.0001), and hypercalcemia (8.5% vs. 2%, P=0.014). Spleen/liver and gastrointestinal involvement were less frequent in the osseous sarcoidosis group (29% vs. 45%, and 1% vs. 17%, respectively, P<0.0001). Response rates to with glucocorticoids alone, glucocorticoids plus methotrexate or glucocorticoids plus hydroxychloroquine were 23/44 (52%), 9/13 (69%) and 4/6 (67%), respectively. CONCLUSION: In patients with osseous sarcoidosis the spine and pelvis were the most commonly affected bones. Compared with controls, cases with osseous sarcoidosis have higher rates of thoracic and extra-thoracic lymph node involvement, pulmonary and cutaneous involvement, and hypercalcemia. Most patients with osseous sarcoidosis had a good response to glucocorticoids in combination with methotrexate or hydroxychloroquine.

Expert opinion on the use of biological therapy in non-infectious uveitis.

INTRODUCTION: Conventional immunosuppressive drugs, anti-TNF alpha treatments and biotherapies are increasingly being used in non-infectious uveitis. AREAS COVERED: The present work was led by a multidisciplinary panel of experts, including internal medicine specialists, rheumatologists and ophthalmologists, and proposes an extensive review on the use of biological agents in non-infectious uveitis. EXPERT OPINION: In case of dependency to steroids or sight-threatening disease, conventional immunosuppressive drugs (methotrexate, azathioprine and mycophenolate mofetil) and/or biological therapies such as anti-TNF alpha treatments (adalimumab, infliximab) can be used to achieve and maintain disease quiescence. Interferon is an efficient immunomodulatory drug that can be proposed as second-line therapy in specific indications (eg. refractory macular edema, sight-threatening Behçet's uveitis). Other biologics, especially tocilizumab, are showing promising results. Local treatments (steroids, sirolimus etc.) can be used as adjuvant therapies in case of unilateral relapse. Therapeutic response must always be evaluated by clinical examination and appropriate ancillary investigations.

Contribution of diagnostic tests for the etiological assessment of uveitis, data from the ULISSE study (Uveitis: Clinical and medicoeconomic evaluation of a standardized strategy of the etiological diagnosis).

PURPOSE: ULISSE is the only study that prospectively assessed the efficiency of a standardized strategy, compared to an open strategy for the etiologic diagnosis of uveitis. Our aim was to evaluate the diagnostic yield of the tests prescribed in the ULISSE study to clarify their relevance. METHODS: ULISSE is a non-inferiority, prospective, multicenter and cluster randomized study. The standardized strategy is a two-steps strategy: in the first step, common standard tests were performed, and in the second step, tests were guided by the clinical and anatomic type of uveitis. We reported the relevance of the diagnostic tests used in the standardized strategy, as well as the profitability of the tests that were prescribed to more than twenty patients in each group. Based on diagnostic criteria, either an ophthalmologist, or an internist, established the profitability of a test by considering whether the test lead to a diagnosis or not. RESULTS: Among the 676 patients included (standardized 303; open 373), a diagnosis was made for 152 (50.4%) in the standardized group and 203 (54.4%) in the open group. The most common entities were HLA-B27 associated uveitis (22%), spondyloarthritis (11%), sarcoidosis (18%), tuberculosis (10.7%) and herpes virus infections (8.5%). Among the first step's systematic tests, tuberculin skin test was the most contributive investigation (17.1%), followed by chest X-ray (8.4%), C reactive protein and ESR (6.6% and 5.1%), complete blood count (2.2%) and VDRL (2.0%). The second step's most often contributive tests were: HLA B27 (56.3%), chest-CT (30.3%) and angiotensin converting enzyme (ACE) (16.5%). HLA B27 and ACE were significantly more contributive in the standardized group than in the open group. Immunological tests were never contributive. Among the free investigations, or among the investigations guided by clinical or paraclinical findings, the most often contributive tests were: Quantiferon® (24%), electrophoresis of serum protein (7.8%) and sacroiliac imagery (46.4%). Intracellular serologies (1.7%), serum calcium (2.1%) and hepatic tests (3.3%) were exceptionally contributive. Among the third intention tests, labial salivary gland biopsies were contributive in 17.9% of cases, but the profitability of other invasive investigations (anterior chamber tap, vitrectomy, bronchoscopy and lumbar puncture) or specialized imagery (18F-FDG PET, Brain MRI) could not be determined since these test were rarely performed. CONCLUSION: Only a few diagnostic tests are useful for the etiological assessment of uveitis. They are often cheap, simple, more often guided by the clinical findings, and lead to an etiological diagnosis in most patients. On the other hand, some tests are never or exceptionally contributive, such as immunological tests or intracellular serologies. Further studies are required to evaluate the profitability of third intention imagery and invasive investigations.

Outcome of patients with sarcoidosis refractory to TNF antagonists: a case series.

Background: Tumor necrosis factor (TNF) antagonists have been reported as an efficient third-line therapy for sarcoidosis but there is no data regarding patients who do not respond to this treatment. Objective: To report the characteristics, the outcome and the response to therapy of patients with sarcoidosis resistant to TNF antagonists. Methods: Patients from the French STAT (Sarcoidosis Treatment with Anti-TNF) registry who were classified as non-responders and who were followed-up for >1 year were included. The response to further therapies was classified as complete response, or partial response, and the others were classified as non-responders. Results: Among the 132 patients from the registry, 14 were considered as non-responders to anti-TNF. Nine patients (66% of women; mean age 48 years) were analyzed. The mean number of organs involved was 4.2. Seven patients were previously treated with more than 2 immunosuppressive treatments. The mean duration of the anti-TNF treatment was 9 months (range, 3-24). After a mean follow-up duration of 58 months (median, 35; range, 19-128) a complete response was observed in 2/9 cases, a partial response in 5/9 cases, and 2/9 cases were considered as non-responders. In all but one patient, the immunosuppressant that allowed the clinical response had previously been used. Furthermore, the dosage was not necessarily increased to gain efficacy. Non-responders were treated by corticosteroids only because of their comorbidities or noncompliance. Conclusion: In patients who do not respond to TNF antagonists, previously used immunosuppressants may be useful. Excluding a differential diagnosis, assessing compliance and testing for anti-drug antibodies should be systematic. (Sarcoidosis Vasc Diffuse Lung Dis 2018; 35: 371-375).

Uveitis: Diagnostic work-up. A literature review and recommendations from an expert committee.

PURPOSE: Diagnosis of uveitis is difficult. Etiologic investigations should take into account the epidemiology of uveitis and should focus on the most severe forms of the disease and those which can be treated. This study was undertaken to establish recommendations for the diagnosis of uveitis. METHODS: Recommendations were developed by a multidisciplinary panel of 14 experts, including internists, ophthalmologists, and rheumatologists, and are based on a review of the literature and the results of the ULISSE study, which was the first prospective study to assess the efficacy of a standardized strategy for the etiologic diagnosis of uveitis. The following groups of patients are not included in these recommendations: children, immunocompromised patients, patients with severe retinal vasculitis, and those with specific eye diseases diagnosed by ophthalmologic examination only. RESULTS: Diagnosis should be guided by the medical history of the patient and physical examination. Serologic screening for syphilis is appropriate in all forms of uveitis. If uveitis is not diagnosed at this stage, investigations oriented by the anatomic characteristics of uveitis are proposed. These consist of assays for HLA-B27 (in unilateral acute anterior non-granulomatous uveitis), serum angiotensin-converting enzyme, interferon-gamma release, chest computed tomography (chronic uveitis), cerebral magnetic resonance imaging and anterior chamber tap with interleukin-10 analysis (intermediate or posterior uveitis in patients >40years-old). Other investigations prescribed in the absence of orientation are usually unhelpful. CONCLUSIONS: A strategy is proposed for the etiologic diagnosis of uveitis. The benefit of more invasive investigations remains to be determined.

Efficacy and safety of tumor necrosis factor antagonists in refractory sarcoidosis: A multicenter study of 132 patients.

INTRODUCTION: The off-label use of TNF antagonists in refractory sarcoidosis is increasingly reported but data on their efficacy and safety are still insufficient. OBJECTIVE: To report on efficacy and safety of TNF antagonists in severe and refractory sarcoidosis. METHODS: Examination of retrospective demographic, clinical, therapeutic, and adverse event data on 132 sarcoidosis patients (58% women; mean (min-max) age = 45.5 (14-78) years) given TNF antagonists (mainly infliximab, 91%) and investigation of response-linked factors. RESULTS: The overall clinical response (complete and partial) rate was 64%. TNF-antagonist efficacy (i.e., significant decrease of the ePOST score) was noted in cases with neurologic, heart, skin, and upper respiratory tract involvements. No significant difference in efficacy was found between anti-TNF used alone and TNF with immunosuppressant. The use of anti-TNF allowed reducing prednisone dosage at end of follow-up (p < 0.001). Adverse events were observed in 52% of the patients; they included infections (36%) and allergic reactions (8%) and required treatment interruption in 31 cases (23%). When TNF antagonists were interrupted, 13 patients experienced relapses within 14 months on average (median follow-up: 20.5 months). CONCLUSION: TNF antagonists were efficacious in about two-thirds of patients with severe/refractory sarcoidosis but their use led to a high rate of adverse events.

Randomized Controlled Trial Evaluating a Standardized Strategy for Uveitis Etiologic Diagnosis (ULISSE).

PURPOSE: To prospectively assess the efficiency of a standardized diagnostic approach, compared to an open strategy, for the etiologic diagnosis of uveitis. DESIGN: Noninferiority, prospective, multicenter, clustered randomized controlled trial. METHODS: Consecutive patients with uveitis, who visited 1 of the participating departments of ophthalmology, were included. In the standardized group, all patients had a minimal evaluation regardless of the type of uveitis (complete blood count, erythrocyte sedimentation rate, C-reactive protein, tuberculin skin test, syphilis serology, and chest radiograph) followed by more complex investigations according to ophthalmologic findings. In the open group, the ophthalmologist could order any type of investigation. Main outcome was the percentage of etiologic diagnoses at 6 months. RESULTS: Nine hundred and three patients with uveitis were included from January 2010 to May 2013 and the per-protocol population comprised 676 patients (open 373; standardized 303). Mean age at diagnosis was 46 years. Anatomic distribution of uveitis was as follows: anterior (60.8% and 72.3%, P = .0017), intermediate (11.7% and 12.3%, P = .8028), posterior (17.8% and 8.2%, P = .0004), and panuveitis (15.3% and 15.2%, P = .9596). An etiologic diagnosis was established in 54.4% of cases in the open group and 49.5% in the standardized group (P = .2029). The difference between both strategies (standardized minus open) was -4.9% (95% CI [-12.5%; 2.6%]). There were more investigations in the open group than in the standardized group (5371 vs 3759, P < .0001). CONCLUSION: The standardized strategy appears to be an efficient diagnostic approach for the etiologic diagnosis of uveitis, although its noninferiority cannot be proved.

Efficacy and safety of TNF antagonists in ocular sarcoidosis: data from the French registry STAT.

Backgroung: This study investigated the efficacy and safety of TNF antagonists in sarcoid uveitis in unselected cases. Design: This is a multicentre study on patients with sarcoidosis who received TNF antagonists in pneumology and internal medicine departments in France. We present here the subgroup of patients with biopsy-proven sarcoid uveitis included in the nationwide registry STAT (Sarcoidosis treated with TNF AnTagonists). Results: Among the 132 patients included in this multicenter study, 18 patients with refractory uveitis were treated as a first-line TNF antagonist with infliximab (n=14), adalimumab (n=3) and certolizumab (n=1). Before anti-TNF initiation, the median duration of sarcoidosis was 42 months and 83% of the patients have been treated with at least one immunosuppressive drug. Six patients switched for a second-line TNF antagonist. After a mean time under treatment of 29 months, the treatment resulted in a significant decrease of the ophthalmic extrapulmonary Physician Organ Severity Tool (ePOST) (mean score: 4.2 vs. 2.6) scores and a steroid sparing effect (29.4±20.7 vs. 6.2±5.2 mg/d). Overall, the ophthalmic response, either complete or partial, was 67%. Nine patients (50%) presented adverse events, including severe infectious complications in 5 patients, which required anti-TNF treatment interruption in 6 cases (33%). Among the 7 responder patients who discontinued anti-TNF therapy, 71% relapsed. Finally, 12 patients (67%) could continue TNF antagonist treatment. Conclusions: TNF antagonists were efficient in 67% of biopsy-proven refractory sarcoid uveitis. Severe adverse events, mainly infectious complications, were frequent. The high frequency of relapses after anti-TNF-α discontinuation requires a close patient follow-up thereafter. (Sarcoidosis Vasc Diffuse Lung Dis 2017; 34: 74-80).