In the general population, high-sensitivity C-reactive protein (hsCRP), a marker of inflammation, is relatively stable over time and independently predicts cardiovascular events. Systemic lupus erythematosus (SLE), a chronic inflammatory disease, is strongly associated with coronary artery disease (CAD). The objective of this study was to determine the variability and correlates of hsCRP in patients with SLE. Two cohorts from the University of Toronto Lupus Clinic, one with newly diagnosed and the other with prevalent SLE for 4 or more years, were selected. HsCRP was measured on serially collected samples, and hsCRP levels were ranked according to quartiles of cardiovascular risk. Correlates of hsCRP were determined using multivariate regression modelling with analysis of repeated measures. Among 58 patients in the inception cohort, over time, 36 (62%) moved from one hsCRP risk quartile to another. Among 414 patients in the prevalent cohort, 294 (71.0%) moved from one risk quartile to another. In both cohorts, within-patient variance comprised the majority of total variance in hsCRP levels. In multivariate regression analysis, hsCRP increased with age (P = 0.002), postmenopausal status (P = 0.03), smoking (P = 0.007) and presence of infection (P = 0.0001) and decreased with use of immunosuppressives (P = 0.02). There is marked variability of hsCRP level over time in SLE, regardless of disease duration. This variability is due to age and SLE treatment, menopausal status, smoking and the occurrence of infection. The variability of hsCRP in SLE casts doubt over its usefulness as an independent predictor of CAD risk in this disease and potentially in other chronic inflammatory diseases.