Variables Description
(a)
The following covariates were considered:
(1) Demographic variables including sex, country of birth (dichotomic variable: “Spaniard” or “not Spaniard”), age (at the beginning of each treatment course, categorized in “<17”, “17 to 29”, “30 to 55”, “>55” years old), and calendar time [year at the beginning of each treatment course, categorized in: “before 2000” (before the availability of any biological drug), “2000 to 2007” (only one anti-TNF therapy was available), and “2008 to 2015” (more anti-TNFs and second line biological therapies became available)].
(2) Disease-related variables, including laterality of the uveitis (“unilateral”, “bilateral”), course (“acute”, “recurrent”, “chronic”), location (“anterior”, “intermediate”, “posterior” and “panuveitis”), diagnosis (“primary intermediate uveitis”, “ophthalmologic chorioretinitis”, “idiopathic panuveitis”, “Behçet disease”, “primary vascular retinitis”, “Vogt-Koyanagi-Harada syndrome”, and “others”), and association with systemic disease (such as Behçet disease or sarcoidosis; “yes”, “no”). Furthermore, we collected information regarding visual acuity and inflammatory state at the beginning of each treatment course, including BCVA (using the decimal Snellen scale and categorizer in: “≥0.5”, “<0.5 to ≥0.1”, and “<0.1”), worst corrected visual acuity (categorizer in: “≥0.5”, “<0.5 to ≥0.1”, and “<0.1”), presence of cells in the anterior chamber (ranked using an ordinal scale ranging from 0 to 4+), keratic precipitates (KP; “yes”, “no”), ocular hypertension (OHT; it was defined as an intraocular pressure reading greater than 21 mmHg as measured by a Goldmannapplanationtonometer; “yes”, “no”), vitreous haze (ranked using an ordinal scale ranging from 0 to 4+), snowballs (“yes”, “no”), retinal vasculitis (“yes”, “no”), cystoid macular edema (CME ; “yes”, “no”), chorioretinitis (“yes”, “no”), papillitis (“yes”, “no”), and epiretinal membrane (ERM; “yes”, “no”).
(3) Pharmacological-related variables including treatment with topical corticosteroids (“yes”, “no”), and dose of oral corticosteroid at the beginning of each treatment course (prednisone-equivalent doses of alternative corticosteroids2 were calculated for these evaluations when necessary; categorized in quartiles); elapsed time from the first visit in our clinic to the date the first ISD therapy was prescribed (categorized in “before first visit in our clinic”, “during the first month”, “2 to 6 months”, “7 to 12 months”, and “> 12 months”); and number of previous courses (continuous variable).
(4) We also assessed the influence of a previous discontinuation due to clinical improvement, inefficacy and/or ADRs.
(b)
Regarding the diagnosis of CME and ERM, and the grading of anterior chamber inflammation and vitreous haze, it is important to take into account that there was a change in methodology during the period of the study: regarding the formed, in 2005 it was introduced in our clinic the use of optical coherence tomography (OCT) to aid in the diagnosis of both CME and ERM. Before the use of OCT, CME was clinically diagnosed through the presence of macular thickening or cysts. Definitive diagnosis was performed with the aid of a fluorescein angiography (FA), which was used in case there was a clinical suspicion of CME based on the findings of the indirect ophthalmoscopy or an unexplained decrease of visual acuity. FA was also performed when clinical signs of vasculitis [including perivascular whitish cuffs (sheathing), calibre changes of vessels, perivascularoedema, intrarretinalhaemorrhages or the presence of cotton-wool spots, and defined by vessel leakage and late staining of the vessel walls] or papillitis (including optic dischyperaemia, swelling, and blurring of disk margins, and defined by leakage of the papillary vessels) were observed in the indirectophthalmoscopy.Therefore not all patients at baseline or during follow-up underwent a FA, only those with a clinical suspicion of any of the referred complications.
ERM was clinically diagnosed thorough the observation of sheen or abnormal reflectivity of the macular surface, with presence of distorted and straightened retinal vessels.
Regarding the grading of anterior chamber inflammation and vitreous haze, before 2005, they were performed using the Uveitis Scoring System3. After 2005, we adopted the Standardization of Uveitis Nomenclature workshop ranking4.
To account for the effect of these temporal changes in ISDs discontinuation, we created a new dichotomous variable, “OCT availability/SUN criteria” (“yes”, ”no”; for those treatment courses that started on or after January the 1st, 2005, or before that date, respectively).
(See references at the Supplementary References file).