We analyzed data for the 12-month period after renal biopsy was done in 130 patients with systemic lupus erythematosus to examine whether renal biopsy provides useful information on the nephritis of systemic lupus erythematosus beyond that clinically available. A stepwise linear regression analysis was used to construct a linear before biopsy model that predicted the change in renal function 12 months after biopsy. The model included serum creatinine, patient age, 24-hour urine protein, a laboratory index of renal activity, antibodies to DNA, urinalysis protein, change in inverse creatinine from 6 weeks before biopsy, and urine light chain protein, and had a squared multiple correlation coefficient (R2) of 0.246. Four prospectively chosen renal biopsy variables (glomerular cell counts, percent of sclerotic glomeruli, percent of glomeruli with crescents, and interstitial fibrosis) resulted in a 0.079 improvement in R2 (p less than or equal to 0.012). Both the percent glomerular sclerosis (p less than or equal to 0.0032) and subendothelial deposits shown by electron microscopy (p less than or equal to 0.0026) added significantly to the predictive power of the before biopsy model. Histologic classification did not add significantly to the before biopsy model. The renal biopsy information increased the power of a linear regression model to predict the effect of 12 months of treatment of active lupus nephritis.