The purpose of this study was to determine the prevalence and correlates of fibromyalgia (FM) in a prospective, multiethnic systemic lupus (SLE) cohort. A total of 266 SLE patients with disease duration of < or = 5 years at study entry were evaluated longitudinally for the presence of FM (per ACR criteria). Sociodemographic factors, behavioral/psychological variables, clinical features, serologic factors (autoantibodies), and self-reported functioning (MOS SF-36) were ascertained in all patients. Subjects were evaluated at study entry and annually thereafter. The prevalence of FM was then calculated, as was the prevalence of FM-like manifestations (widespread pain with at least 6, but fewer than 11/18 tender points). Variables were evaluated for association with FM or FM-like manifestations by univariate and stepwise logistic regression analyses. FM was present in 14 patients (5%; 9/92 Caucasians (C), 4/109 African Americans (AA), 1/65 Hispanics (H)) and FM/FM-like manifestations in 35 (13%; 16 C, 9 AA, 10 H). There was no difference noted between those with and without FM with respect to gender, education level, income below poverty level, disease activity or damage. By stepwise logistic regression analyses, the strongest association with both FM and FM/FM-like manifestations was a self-reported history of anxiety or affective disorder (P = 0.0237, OR = 4.6 and P = 0.0068, OR = 3.4, respectively). Caucasian ethnicity was strongly associated with FM (P = 0.0066, OR = 7.5) and African American ethnicity was negatively associated with FM/FM-like (P = 0.0204, OR = 0.3). Poorer self-reported physical functioning was associated with FM/FM-like (P = 0.0443, OR = 0.96). FM and FM-like manifestations correlate best with the presence of Caucasian ethnicity, concomitant anxiety or affective disorder, and to a lesser extent with poorer self-reported physical functioning. African American ethnicity is negatively associated with the combination of FM and FM-like manifestations. Clinical measures of disease activity, disease damage, specific organ dysfunction, sociodemographic factors and serologic features are not correlated with FM in this early SLE cohort.