To fully describe the clinical course of lupus nephritis in an African-American population, we report our experience with 54 patients seen at a large inner-city hospital over a period of 14 years. The patients were divided into five histopathologic groups. Group MES (n = 3) represented mesangial nephritis (World Health Organization [WHO] class II) and group FOC (n = 11) represented mild and moderate focal segmental proliferative glomerulonephritis (WHO class III). Group DIF (n = 9) included patients with severe segmental proliferative, diffuse proliferative, membranoproliferative, and membranous and severe superimposed proliferative lesions (WHO classes III, IV, and Vd). Group CRES (n = 9) combined all the patients with cellular crescents in more than 40% of the glomeruli and included patients in WHO classes III (severe), IV, and Vc and d. Group MEM (n = 22) represented membranous nephritis occurring alone or with superimposed mesangial or mild segmental proliferative lesions (WHO class Va and b). Groups DIF and CRES received intensive treatment with high-dose prednisone and cytotoxic drugs. Groups FOC and MEM received lower doses of prednisone, but half of the patients later received intensive treatment largely for severe systemic manifestations. The three patients in group MES remained well. End-stage renal failure (ESRF) developed in 11 of 18 patients in groups DIF and CRES combined, and in two of 22 patients in group MEM. Three of 11 patients in group FOC, five in groups DIF and CRES, and one in group MEM died. The actuarial 5- and 10-year survival rates were, respectively, 78% and 78% for FOC, 80% and 0% for DIF and CRES, and 100% and 100% for MEM (P < 0.03 v DIF/CRES). Five- and 10-year survival rates without ESRF were, respectively, 78% and 78% for FOC, 52% and 0% for DIF and CRES (P < 0.05), and 94% and 85% for MEM (P = 0.002 v DIF/CRES). Univariate proportional hazards regression analysis, uncontrolled for histopathologic groups, showed a significant association between ESRF and severe thrombocytopenia (P = 0.003), serum creatinine above 1.4 mg/dL at entry (P = 0.04), and severe systemic manifestations (P = 0.05). After controlling for histopathologic groups, only thrombocytopenia remained strongly associated with ESRF, both by univariate (P = 0.01) and multivariate (hazard ratio = 14.19, P = 0.05) analyses. We conclude that severe proliferative lupus nephritis in African-Americans has a poor prognosis. For mild and moderate focal proliferative nephritis and uncomplicated membranous lupus nephritis the prognosis is as good as in white patients. Severe thrombocytopenia predicts ESRF.