Abstract
OBJECTIVES: The long-term outcome of psychosis in association with systemic lupus erythematosus (SLE) has been insufficiently characterised. We used a specialist centre cohort of patients with SLE and psychosis to investigate their clinical outcome and phenotypic and laboratory characteristics. METHODS: Retrospective cohort study of 709 SLE patients seen at a specialist centre between January 1978 and November 2018. Clinical, biochemical and immunological characteristics (Bonferroni corrected), and serum neuronal surface antibody profile using novel cell-based assays, were compared between patients with and without psychosis. RESULTS: Eighteen (18/709, 2.5%) patients developed lupus psychosis over a mean ± SD of 17.5 ± 11.0 years follow-up. Psychosis fully remitted in 66.7% (12/18) with a combination of antipsychotic (in 38.9%) and immunosuppressive therapy (methylprednisolone 72.2%, cyclophosphamide 55.6%, rituximab 16.7%, plasma exchange 27.8%, prednisolone 50%). Patients who developed lupus psychosis may be more likely to have anti-RNP antibodies (50.0% vs 26.5%) and less likely to have anti-cardiolipin antibodies (5.6% vs 30.0%), but this was not significant in our small sample. Neuronal surface autoantibody tests found GABABR autoantibodies in 3/10 (30.0%) lupus psychosis patients compared with only 3/27 (11.1%) in age- and sex-matched SLE controls using fixed cell-based assays (P =0.114). However, GABABR antibodies were not replicated using a live cell-based assay. NMDAR-antibodies were not detected with fixed or live cell assays in any samples. CONCLUSION: Lupus psychosis is rare but treatable. In this rare sample of eighteen patients from a 40-year cohort, no significant biomarker was found, but some preliminary associations warrant further exploration in a larger multicentre analysis.