We evaluated 522 patients with idiopathic, recurrent calcium nephrolithiasis using a comprehensive clinical and laboratory protocol, and obtained additional laboratory measurements during their subsequent years of treatment in our program. In 57 patients, a new calcium stone ultimately formed during treatment (relapse), whereas 189 others have been free of recurrence during at least two years (average 4.3 +/- 2.2 [SD] years) of follow-up. Compared with the patients who remained stone-free, the patients with relapse (1) had a shorter interval between the time they entered our program and the time their last recurrent, pretreatment stone formed; (2) excreted more calcium *in mg/kg of body weight pr 24 hours) in their urine during treatment (2.79 +/- 1.08 versus 2.39 +/- 0.98 [SD] for relapse and stone-free); and (3) increased their urine volume less during treatment compared with pretreatment values (delta in liters per 24 hours was -0.02 +/- 0.48 versus 0.23 +/- 0.54 for relapse and stone-free). The two groups were otherwise the same. All comparisons used only data obtained prior to relapse. A discriminant function using only these three characteristics correctly identified 72 percent of patients with relapse and 67 percent of those who remained stone-free.