OBJECTIVE To study the neuropsychiatric (NP) manifestations in a large cohort of southern Chinese patients with systemic lupus erythematosus (SLE) according to the new 1999 American College of Rheumatology (ACR) case definitions and their clinical associations. METHOD Patients with SLE who were followed from 1984 to 2000 were retrospectively reviewed. Patients with NP manifestations were ascertained and classified by at least 2 rheumatologists, with the collaboration of neurologists and psychiatrists. The association of NP manifestations with other clinical features and autoantibodies was studied by statistical analysis. RESULTS Five hundred eighteen patients with SLE were studied. The female to male ratio was 7.8 to 1 and the mean age of disease onset was 29.5 +/- 12.0 years (range 9-80). The mean duration of followup was 7.3 +/- 6.7 years (range 0.3-23.0). Ninety-six patients (19%) had 133 NP events and the mean number of events per patient-year of followup was 0.035. In decreasing order of frequency. these events were: seizure disorder (28%), cerebrovascular disease (19%), acute confusional state (14%), psychosis (11%), myelopathy (8%), mood disorder (6%), headache (4%), movement disorder (2%), cranial neuropathy (3%), demyelinating syndrome (1.5%), anxiety disorder (1.5%), mononeuritis multiplex/mononeuropathy (1.5%), aseptic meningitis (1%), and polyneuropathy (1%). Cognitive dysfunction was not classified because of the lack of standard neuropsychological testing for every patient. Univariate analysis revealed that NP-SLE was associated with a positive lupus anticoagulant (LAC) (p = 0.001), a strongly positive IgG anticardiolipin (aCL) (p = 0.01). leukopenia (p = 0.01), lymphopenia (p = 0.03), thrombocytopenia (p = 0.03), and pulmonary involvement (p = 0.03). Multivariate analysis showed that a strongly positive IgG aCL [RR 3.1 (1.3-7.7), p = 0.01] and a history of cyclophosphamide treatment [RR 4.3 (2.1-9.0), p < 0.001] were independently associated with NP manifestations in our cohort. Among the NP features, cerebrovascular disorder was particularly associated with the presence of LAC [OR 3.3 (1.4-8.0), p = 0.01] and a strongly positive IgG aCL [OR 3.1 (1.1-8.2), p = 0.031]. CONCLUSION The point prevalence of overt NP manifestations in our cohort of patients with SLE was 19%. This percentage was likely higher if subtle cognitive dysfunction was included. Seizure and cerebrovascular disorders were the most common NP features. The presence of antiphospholipid antibodies was significantly associated with NP manifestations, especially cerebrovascular disorders.