Patterns and prevalence of cognitive dysfunction in systemic lupus erythematosus

Abstract

Objective: Systemic lupus erythematosus (SLE) is a chronic autoimmune disease, in which cognitive dysfunction is common, but poorly understood. This study aims to characterize the prevalence and patterns of cognitive dysfunction in SLE. Method: SLE patients (n = 95) and demographically matched healthy controls (n = 48) underwent cross-sectional cognitive testing using the 1-hr conventional neuropsychological test battery recommended by the American College of Rheumatology for use in SLE. We used standard deviations (SD) from the healthy control group to define impairment. For each cognitive test we compared SLE and control groups using independent samples t-tests (or alternatives when needed). We performed cluster analysis using a machine learning algorithm to look for patterns of cognitive dysfunction. Results: The SLE group performed significantly worse than healthy controls on every cognitive test. The largest differences were in the domains of verbal fluency, working memory and attention, while fine motor and psychomotor speed were the least affected domains. As expected, the prevalence of cognitive dysfunction varied depending on the SD cut-off used, with 49% of participants being >1.5 SD below the healthy control mean in at least two cognitive domains. Heat mapping showed variability in the pattern of dysfunction between individual patients and cluster analysis confirmed the presence of two clusters of patients, which were those significantly impaired versus those having preserved cognition. Conclusions: Cognitive dysfunction is common in SLE but markedly heterogeneous across both cognitive domains and across the SLE group. Cluster analysis supports the use of a binary definition of cognitive dysfunction in SLE.