Whilst the prognostic value of serum beta-2 microglobulin (s-beta 2m) is well documented, the lack of a simple strategy for its use means that it is rarely ever measured in clinical practice. The prognosis associated with s-beta 2m at two different points in HIV infection, as defined by the CD4 count, was studied in a cohort of 111 men with haemophilia registered at the Royal Free Hospital School of Medicine, London. At CD4 counts of 0.5 and 0.2 x 10(9)/l, a raised s-beta 2m level was significantly associated with an increased risk of developing AIDS (P = 0.002 and 0.022 respectively, adjusted for the patient's age). Kaplan-Meier progression rates to AIDS by 4.5 years after a CD4 count of 0.5 x 10(9)/l were 57% (95% CI 32-82%) in those with s-beta 2m levels of 3 mg/l or more, but 20% (95% CI 4-36%) in those with s-beta 2m levels of less than 3 mg/l. By 3.5 years after a CD4 count of 0.2 x 10(9)/l, Kaplan-Meier progression rates to AIDS were 75% (95% CI 52-98%) in those with s-beta 2m levels of 3.8 mg/l or more, and 47% (95% CI 29-66%) in those with s-beta 2m levels of less than 3.8 mg/l. In the absence of acute viral infections, a raised s-beta 2m indicates those who will tend to progress to AIDS more rapidly than those with lower s-beta 2m levels and the same CD4 count. S-beta 2m levels in general are likely to be higher in haemophilia patients than in other, non-haemophilic risk groups. Whilst care should be taken, therefore, when applying our chosen cut-off values to non-haemophilic patients, our findings support the introduction of prophylaxis and antiviral therapies at a higher CD4 count in those with raised s-beta 2m levels relative to other patients in the same risk group whilst delaying treatment in those with lower CD4 counts, but relatively normal s-beta 2m levels.