Thirteen cases of disseminated infection with Mycobacterium avium-intracellulare (MAI) seen at the National Jewish Hospital and Research Center and 24 cases from the literature were analyzed to define clinical and therapeutic features of the disease. Disseminated MAI infection was a disease of immunocompromised and apparently normal hosts. It was acquired from the environment by unknown mechanisms, usually entering the body through the lungs and spreading to include the reticuloendothelial system, bones, and less commonly, the skin. Diagnosis was often delayed and required culture of tissue or secretions. Medical personnel must maintain a high index of suspicion for MAI disease, especially in immunocompromised hosts. These patients should be monitored carefully for evidence of MAI with frequent cultures of blood and bone marrow. Blood culture systems able to recover MAI promptly and reliably should be employed (52, 64). New diagnostic aids, such as the standardized preparation of PPD-B currently being prepared or tests for antibody to MAI, will help in differentiating MAI from other processes. If MAI is recovered, broad-spectrum therapy should be instituted. Response to combination antimicrobial chemotherapy in the patients surveyed in this report was gratifying. Over two-thirds of treated patients responded to therapy. New antimycobacterial agents such as ansamycin and thienamycin have been shown to have activity against MAI in vitro (40, 81, 92) and may further improve therapeutic efficacy. Studies of in vitro synergy, currently in progress in our laboratory, will also help define the optimal therapeutic regimen for each individual patient. While the patients presented in this report had a reassuring response to therapy, those who had many bacilli in the tissues had a poorer outcome. Patients with AIDS often have this lepromatous histology (37) and thus may respond more poorly than the patients in this report even when optimal therapy is employed. Careful monitoring of AIDS patients for MAI infection may permit earlier institution of therapy and improve the chances for control of the infection. Studies to assess the relationship of in vitro sensitivity to therapeutic response in these patients are currently underway in our laboratory. It is hoped that early institution of therapy and optimization of regimens according to in vitro sensitivity data will lead to decreased morbidity and mortality in all patients with MAI infection.