This study assessed whether certain clinicopathologic variables could explain the impact of race on outcome in 86 patients who had severe lupus nephritis and were available for long-term follow-up after participating in a prospective, controlled, clinical trial. Fifty-four (63%) patients were white, 21 (24%) were black, and 11 (13%) were categorized as other. The proportion of patients with anti-Ro, anti-nRNP, and anti-Sm was significantly greater among black patients. Biopsies with segmental active proliferative and necrotizing lesions that involved >or=50% of glomeruli +/- membranous glomerulonephritis (class III >or=50%+/-V) were significantly more common (white 44%, black 76%, other 36%; P < 0.05) and diffuse proliferative glomerulonephritis +/- membranous glomerulonephritis (class IV+/-V) was less common (white 54%, black 24%, other 64%) among black patients. Attainment of a remission was greatest among white patients (white 52%, black 29%, other 27%; P = 0.09). Features that were predictive of a remission were white race, baseline serum creatinine, and class IV+/-V lesions. Patient survival at 10 yr (white 81%, black 59%, other 73%; P = 0.029) and renal survival at 10 yr (white 68%, black 38%, other 61%; P = 0.015) were significantly poorer in black patients. Predictors of ESRD were serum creatinine, the presence of anti-Ro antibodies, class III >or=50%+/-V lesions, and failure to achieve a remission. In conclusion, racial differences were observed in the serologic and histologic features at presentation, response to treatment, and outcome of patients with severe lupus nephritis. In a population of patients with severe lupus nephritis, black patients were significantly more likely to have a serologic profile and renal lesions that were associated with more aggressive renal disease and resulted in worse outcomes than white patients.