Previous studies of the in vitro effects of recombinant erythropoietin (rEPO) on T and B cells and studies of lymphocyte subsets in dialysis patients receiving rEPO therapy suggest that rEPO might augment immune responses. In the present study indices of autoimmunity (antinuclear antibody, anti-double-stranded DNA, and antiphospholipid antibody [immunoglobulins G and M]) were measured before, during, and after rEPO therapy in five systemic lupus erythematosus patients without renal failure (mean serum creatinine, 1.5 +/- 0.3 mg/dL). The rEPO therapy was self-administered by subcutaneous injection in doses ranging from 4,000 units once weekly to 3,000 units three times weekly for 3 to 7 months. On rEPO therapy, each patient experienced an increase in hematocrit. The mean baseline hematocrit increased from 32 +/- 2.0 to a peak of 42.2 +/- 3 (P < 0.001) at 3 to 7 months and then decreased to baseline values 1 to 2 months after rEPO was discontinued. During this time the indices of autoimmunity were not significantly changed by rEPO therapy. Systemic lupus erythematosus activity, assessed by serum C3, serum creatinine, urinalysis, and 24-hour proteinuria, also was unchanged by rEPO therapy. The rEPO therapy was generally well tolerated. However, one patient, who was also receiving replacement estrogen therapy and had high-titer antiphospholipid antibody, experienced episodes of thrombophlebitis while on rEPO therapy. In conclusion, we found no evidence that rEPO increases autoimmunity in systemic lupus erythematosus. However, we observed a temporal relationship between episodes of thrombophlebitis and rEPO therapy in a single patient with high-titer anticardiolipin antibody who was also receiving replacement estrogen therapy. These associations require further investigation.