Insufficient information is available about the effectiveness of high-dose intravenous corticosteroid treatment in non-renal SLE to determine whether it provides better management of the disease. Issues of steroid preparation, dose and frequency have not been addressed. Significant short-term complications appear to be minimal; the question of long-term effects on avascular necrosis may not be answerable since most patients also receive oral corticosteroids at some time during their course. No large randomized controlled trial of pulse IV-MP in lupus nephritis is available. Ideally, such a trial would include randomization with stratification according to important prognostic variables. Previous studies of other agents have been hampered by lack of agreement about both the identification of relevant prognostic variables and the selection of 'significant' measures of outcome. The observation that patients with recent deterioration in glomerular filtration rate are more likely to improve after IV-MP may form the basis for stratification in a controlled trial of the comparative efficacy of IV-MP and other forms of treatment. However, lack of improvement in glomerular filtration rate among patients without recent deterioration in function may reflect the inadequacy of our outcome measures rather than a true lack of beneficial effect for the patient.