PROGNOSTIC MARKERS IN PATIENTS WITH ANTINEUTROPHIL CYTOPLASMIC AUTOANTIBODY-ASSOCIATED MICROSCOPIC POLYANGIITIS AND GLOMERULONEPHRITIS

The purpose of this study was to determine the prognostic value of cli nical, laboratory, and pathologic features at the time of presentation on patient and renal survival in patients with antineutrophil cytopla smic autoantibody (ANCA)-associated microscopic polyangiitis and glome rulonephritis (excluding Wegener's granulomatosis). One hundred seven ANCA-positive patients with necrotizing and crescentic glomerulonephri tis, including 69 with evidence for microscopic polyangiitis, were eva luated for this study. The relative risk of death was calculated for t he following potential prognostic indicators: (1) ANCA pattern; (2) pu lmonary hemorrhage at onset; (3) presence of extrarenal manifestations versus renal limited disease; and (4) treatment with corticosteroids and cyclophosphamide (intravenous or oral), compared with corticostero ids alone. Cox's proportional hazard model was used to assess the pred ictive value of the following variables on renal survival: (1) age; (2 ) race; (3) pulmonary symptoms at onset of disease; (4) renal patholog y; (5) ANCA pattern; and (6) peak serum creatinine values obtained nea r the time of renal biopsy. Patients were followed prospectively for 2 .5 yr (range, 5 days to 12 yr 2 months). There were 12 disease-related deaths and 46 patients who reached ESRD. The relative risk (and 95% c onfidence interval) of patient death was 8.65 (3.36, 22.2) times great er in patients who presented with pulmonary hemorrhage, and 3.78 (1.22 , 11.70) times greater in patients with cytoplasmic ANCA compared to t hose with perinuclear ANCA. The relative risk of pulmonary hemorrhage was no different by ANCA pattern. The risk of death was 5.56 times low er in the cyclophosphamide-treated patients versus those treated with corticosteroids alone. The predictors of renal survival were entry ser um creatinine value (P = 0.0002), race (African Americans having a wor se outcome compared with Caucasians, P = 0.0008), and the presence of arterial sclerosis on kidney biopsy (P = 0.0076) when controlling for age, ANCA pattern, microscopic polyangiitis versus glomerulonephritis alone, and pulmonary involvement. Pathology indices such as glomerular necrosis, glomerular crescents, glomerular sclerosis, and interstitia l sclerosis were not predictive of renal survival when controlling for entry serum creatinine value, race, and arterial sclerosis. However, in the subgroup of patients with a peak creatinine value of less than or equal to 3.0 mg/dL (N = 29), increased interstitial sclerosis was a predictor of a poor renal outcome (P = 0.04).