Immunopathologic studies over the past two decades have demonstrated that rapidly progressive glomerulonephritis (RPGN) can result from glomerular deposition of anti-GBM antibody, immune complexes, or from some as yet undefined mechanism that does not involve glomerular antibody deposition. The latter process may be cell mediated and resembles a small vessel vasculitis. Most cases of idiopathic RPGN are not accompanied by pathogenic glomerular immunoglobulin deposition. Recent experimental studies of immune mechanisms of glomerular injury have identified several new processes that can induce damage to the capillary wall sufficient to result in crescentic glomerulonephritis (GN). These include direct effects of anti-GBM antibody alone and of the complement C5b-9 (membrane attack) complex, nephritogenic effects of inflammatory effector cells that involve reactive oxygen species and glomerular halogenation, and injury mediated by sensitized lymphocytes independently of antibody deposition. Macrophages have been shown to participate in both intracapillary and extracapillary fibrin deposition and crescent formation as well as to mediate capillary wall damage. The role of resident glomerular cells and cell-cell interactions in glomerulonephritis is still under active investigation. Despite these several advances in understanding immune injury to the glomerulus, therapy for RPGN remains largely empiric. Although the prognosis in RPGN has clearly improved over time, no form of disease-specific therapy has been clearly shown yet to be beneficial in a controlled study. Interpretation of the existing literature on therapy is complicated by the availability of only historical rather than concurrent controls, lack of attention to several variables known to affect disease outcome, and uncertainty regarding bias in favor of reporting positive results. Available data suggests that optimal outcomes may be achieved in anti-GBM nephritis by treatment with steroids, immunosuppression and plasma exchange, particularly when therapy is directed at patients with mild but rapidly progressive disease before oliguria or severe azotemia develop. Pulse steroids are probably the most cost-effective therapy for the idiopathic form of RPGN, but treatment with cytotoxic agents should be considered if clinical or histologic evidence of vasculitis is present.