The spectrum of diseases associated with necrotizing glomerulonephritis and its prognosis.

Necrotizing glomerulonephritis (NGN) represents small-vessel vasculitis in the kidney. To assess the diseases associated with necrotizing glomerular changes and their prognosis we studied all 32 patients who had this histologic finding on kidney biopsy from 1969 to 1982 and compared them to those patients who had crescentic, diffuse, or focal and segmental glomerulonephritis without necrosis (n = 29). The diseases associated with NGN were systemic lupus erythematosus (n = 6/15), Henoch-Schönlein purpura (n = 3/4) Goodpasture's syndrome (n = 4/7), Wegener's granulomatosis (n = 6/6), polyarteritis (n = 4/5), infective endocarditis (n = 2/3), and idiopathic rapidly progressive glomerulonephritis (n = 7/21). Necrotizing glomerulonephritis occurred significantly more often in the vasculitides than in all the other disorders put together. The most difficult diagnosis problem occurred in patients with renal disease and pulmonary hemorrhage (n = 9), in three of whom diagnosis was uncertain even after autopsy (two autopsies done within one month and one within three months of presentation). A fourth patient had a linear staining for IgG along the glomerular basement membrane (GBM) on kidney biopsy but was subsequently diagnosed as having Wegener's granulomatosis. Comparison of patients with without NGN revealed no difference in outcome (death or dialysis) one year after biopsy (38% v 43%) or in serum creatinine levels one year later (4.6 v 4.8 mg/dL). The prognostic effect of NGN was not obscured by unequal distribution of other adverse prognostic factors in the two groups. The most important prognostic characteristics we identified for outcome were serum creatinine at biopsy (chi 2 = 24.0, P less than .0004) and the sum of activity and chronicity indexes on biopsy (chi 2 = 12.7, P = .0004). These variables were similarly distributed in patients with and without necrosis, mean serum creatinine levels at biopsy being 4.3 v 4.2 mg/dL and sum of indexes 7.8 v 8.0. Other factors such as clinical diagnosis and therapy were not important prognostically and therefore could not explain our results. We conclude that NGN in patients with active proliferative glomerulonephritis has multiple causes. Diagnostic difficulties occurred in those with anti-GBM-negative pulmonary hemorrhage. The appearances of small-vessel vasculitis in the kidney did not appear to have prognostic significance.