In an effort to clarify the role of immunosuppressive drugs in the management of lupus nephritis, we pooled data from all published clinical trials in which patients had been randomly assigned to receive either prednisone alone or prednisone plus cyclophosphamide or azathioprine. The pooled analysis showed that patients receiving immunosuppressive drugs had less renal deterioration (P = 0.006), were less likely to have end-stage renal disease (P = 0.023), and were less likely to die from kidney disease (P = 0.024) than patients receiving steroids alone. There were no significant differences with respect to deaths from nonrenal causes or overall mortality. When cyclophosphamide and azathioprine were considered separately, both were associated with a 40 per cent reduction in the rates of adverse renal outcomes, although because the treatment groups were smaller, many of these differences were not statistically significant. A power analysis showed that a study of 100 high-risk patients (development of renal insufficiency in 50 per cent) would be needed to prove that an immunosuppressive agent is 50 per cent superior to steroids alone in preventing renal deterioration. We conclude that immunosuppressive drugs and steroids together are more effective than steroids alone in treating lupus nephritis and that published trials have reached false negative conclusions because of small sample sizes.