The aim of this study was to determine whether the different autoantibodies predict early damage in patients with systemic lupus erythematosus (SLE). The patients comprised a prospective inception cohort of 205 patients with SLE, 154 on follow-up for at least five years after diagnosis. Eight patients who died before the fifth year of disease course were included in analyses comprising survival. Organ damage was measured using the Systemic Lupus International Collaborating Clinics--American College of Rheumatology damage index (SDI). Endpoints were the development of some (SDI > or = 1) or severe (SDI > 2) damage at five years after diagnosis or the combined outcome 'SDI > or = 1 or death at five years'. Autoantibodies [anti-DNA, anti-Ro, anti-La, anti-Sm, anti-U1RNP, any anti-ENA and antiphospholipid (aPL)] were included in univariate and multivariate analysis. 'Age at diagnosis' was also included as an independent variable in multivariant analyses. Sapporo criteria were used to define aPL positivity. Eighty-four patients (54.5%) had accrued damage at five years, 17 patients (11.0%) having severe damage. Patients with aPL had damage in a higher proportion (63.2% any damage, 17.6% severe damage). Only aPL were related to damage in univariate analysis (P = 0.03). In logistic regression models, aPL were the only independent predictors of damage at five years (OR 1.94, 95% CI 1.01-3.73), severe damage at five years (OR 3.34, 95% CI 1.11-10.03) and increasing damage since diagnosis (OR 2.46, 95% CI 1.24-4.87). No autoantibody was a predictor of the outcome 'SDI > or = 1 or death at five years'. The conclusion was that aPL predict early damage in patients with SLE.