Of 30 therapy studies which distinguish between improved and non-improved renal function, 350 patients with rapidly progressive glomerulonephritis (RPGN) were evaluated. Pure descriptions of cases were not included. The cases of RPGN were divided into autoantibody-induced and non-autoantibody-induced groups. This latter group was subdivided into idiopathic and symptomatic RPGN. A further distinction was drawn between the different forms of symptomatic RPGN, but no separate evaluations were made, on account of the small numbers of cases. Therapies were divided into immunosuppression, anticoagulant therapy, pulse therapy, and therapeutic plasmapheresis. In autoantibody-induced RPGN, improved renal function was evidenced in only five cases out of 27. In contrast to this, 66% of the non-oliguric patients with creatinine levels greater than 6 mg/dl showed improved renal function after plasma separation. In non-autoantibody-induced RPGN, the least favourable results were shown by anticoagulant treatment, where improvement in renal function was produced in only 34% of the cases treated, and haemorrhagic complications occurred in 25%, about half of which had a fatal outcome. Under pulse therapy, 27 out of 38 patients (71%) showed improvement, as against 59 out of 93 (63%) under plasmapheresis. In contrast to the situation in autoantibody-induced RPGN, it is possible in non-autoantibody-induced RPGN to achieve therapy-induced improvement also in a high percentage of cases where terminal renal insufficiency is present, and even when dialysis treatment has just been commenced. The collected statistics for therapeutic results achieved in RPGN are compared and contrasted with two controlled studies which showed diverging findings.