LAUSR

PURPOSE To investigate clinical and biological factors influencing recurrences of severe toxoplasmic retinochoroiditis (TRC) confirmed by aqueous humor analysis. DESIGN Retrospective case series. METHODS Retrospective analysis of 87 subjects with severe TRC, proven by positive Goldmann-Witmer coefficient (GWC), Toxoplasma gondii (T. gondii) immunoblot, or T. gondii-specific polymerase chain reaction (PCR) in aqueous humor. Cases with immunosuppression or retinal scars without previous recorded episode were excluded. Time-dependent, clinical, treatment-related, and biological factors were explored by univariate and multivariate shared frailty survival analyses. RESULTS Among 44 included subjects (age, 40.4 ± 17.6 years; follow-up, 8.3 ± 2.7 years), 22 presented recurrences. There was 0.11 recurrence/patient/year and mean disease-free interval was 5.0 ± 2.9 years. The risk of recurrence was higher immediately after an episode (P < .0001). Among recurrent cases, the risk of multiple recurrences was higher when the first recurrence occurred after longer disease-free intervals (P = .046). In univariate analysis, the recurrence risk declined with higher number of intense bands on aqueous T. gondii immunoblot (P = .006), and increased when venous vasculitis was present initially (P = .019). Multivariate analysis confirmed that eyes with more intense bands on immunoblot had fewer recurrences (P = .041). There was a near-significant risk elevation after pyrimethamine/azithromycin treatment (P = .078 and P = .054, univariate and multivariate). Intravenous corticosteroid administration, oral corticosteroid administration, aqueous GWC, and T. gondii PCR did not influence recurrences (P = .12, P = .10, P = .39, and P = .96, respectively). CONCLUSIONS Recurrences of severe TRC are not random and may be influenced by clinical and biological factors possibly related to blood-retinal barrier alterations. These results may contribute to identifying biomarkers for TRC reactivation.