Diffuse alveolar hemorrhage (DAH) complicating systemic lupus erythematosus (SLE) remains a devastating pulmonary complication of this systemic disease. We conducted this study to review the clinicopathologic presentation and the effects of prior treatment, presence of infection, and current treatment on the survival and outcome of patients with DAH and SLE. We reviewed the records of 15 SLE patients who experienced 19 episodes of DAH over a 10-year period in a single tertiary care hospital. These patients were compared with 57 previously reported cases. The 19 episodes of DAH represented 3.7% of the 510 admissions occasioned by various complications of SLE. As previously reported, the majority (66%) were women with a median age of 27 years. The onset was often abrupt: < 3 days in 12 of the episodes. In 3 patients (20%), DAH was the initial manifestation of SLE, compared with 11% in the literature series. In the other patients in the present series, DAH appeared a median of 31 months following the diagnosis of SLE, versus 35 months in the literature series. In only 42% of the episodes in the present series, compared with 66% in the literature series, was hemoptysis present at the time of admission. However, hemoptysis eventually appeared in all 19 episodes. Temperature elevation (> 38 degrees C) was another inconsistent finding, found in only 5 episodes (26%) in the present series. The most constant concurrent systemic finding was lupus nephritis (14/15 patients). This represents a significant increase when compared with the literature series (29/48 patients). In 8 of 10 patients in whom lung tissue was available, pulmonary capillaritis accompanied the DAH. This represents a marked difference in the underlying histologic pattern when compared with the literature series. In those patients, 72% (31/43 patients) had bland pulmonary hemorrhage, and capillaritis was described in only 6 patients. The overall patient mortality rate was 53% in the current series and 50% in the literature series. Factors associated with an increased mortality in the present series include the following: mechanical ventilation (62%) versus no mechanical ventilation (0%); infection (78%) versus no infection (20%); and cyclophosphamide therapy for the acute DAH episode (70%) versus no cyclophosphamide therapy (20%). The incidence of infection in DAH and SLE (9/19 episodes) is far greater than previously reported (7/ 57 episodes). One possible explanation for this difference is the increased use of outpatient immunosuppressive therapy with monthly intravenous cyclophosphamide therapy for lupus nephritis. Eighteen DAH episodes in the present series were treated with intravenous methylprednisolone. When one combines both the current and literature series experience (16 episodes), the use of plasmapheresis does not improve survival. Of the 7 patients in the present series who survived all episodes of DAH, 6 remain alive a median of 50 months post episode and without recurrence of DAH. Diffuse alveolar hemorrhage is an uncommon but lethal complication of SLE. The survival rate remains unchanged from previous reports. The absence of hemoptysis should not exclude this diagnosis, particularly in those patients who experience an acute pulmonary syndrome with new radiographic infiltrates accompanied by falling hematocrit and the presence of a hemorrhagic bronchoalveolar lavage. Evidence for lupus nephritis is present in the great majority of cases. Most cases demonstrate the histologic pattern of pulmonary capillaritis. The mortality is adversely affected by the need for mechanical ventilation, either the presence of infection at the time of admission or the development of infection in the hospital, and the use of cyclophosphamide for treatment of the acute event.