The hypothesis that lymphocytotoxic antibodies are associated with neuropsychiatric involvement in systemic lupus erythematosus (NP-SLE) is re-evaluated in this study. In an unselected cohort of 98 women with SLE a cross-sectional study has been performed to analyse associations among standardised clinical, neurological, and neuropsychological assessments and lymphocytotoxic antibodies measured by microcytotoxicity assay. Fifty patients showed objective clinical evidence of continuing or past NP-SLE and 54 patients had cognitive impairment. In accordance with previous observations 44% (24/54) of the cognitively impaired group did not have clinically detectable evidence of NP-SLE. Although lymphocytotoxic antibodies were found to be only marginally more prevalent in those patients with a clinical diagnosis of NP-SLE than in those without (32% v 23%), these antibodies were significantly associated with cognitive impairment (chi 2 = 5.42; p less than 0.02). No association was detected between lymphocytotoxic antibodies and either overall systemic disease activity or other organ system involvement, suggesting that the association between lymphocytotoxic antibodies and cognitive dysfunction in SLE is specific.