Six hundred twelve eligible, previously untreated patients with active multiple myeloma and at least some data available for analysis were entered into a randomized trial (Southwest Oncology Group [SWOG] Phase III myeloma study 8229/30), in which the prognostic significance of pretreatment serum beta 2 microglobulin levels was evaluated. Because there was no statistically significant survival difference between the alternating and syncopating VMCP/VBAP regimens, it was possible to evaluate serum beta 2 microglobulin for the total population all together. The serum beta 2 microglobulin measurements showed the highest significance of any prognostic factor, both in the bivariate and multivariate regression analyses. The median survival was 36 months for the 322 patients with pretreatment serum beta 2 microglobulin values of less than 6 micrograms/mL, as compared with a median survival of 23 months for the 225 patients with a beta 2 level of greater than or equal to 6 mcg/mL (P less than .0001). The stepwise multiple regression model first contained serum beta 2 microglobulin, followed by serum albumin, serum calcium, age, and serum creatinine. Serum beta 2 microglobulin was highly correlated with stage: median values ranged from 3.7 micrograms/mL for stage IA, to 10.1 for stage IIIB. It was possible to stratify myeloma patients based on combinations of serum beta 2 microglobulin with both albumin and age, producing excellent separation of patients into low-, intermediate-, and high-risk categories. It is concluded that serum beta 2 microglobulin is the most powerful prognostic factor currently available for multiple myeloma and that it can be used alone or in combination with other variables for pretreatment stratification.