In the present report we describe 4 previously healthy women who developed cryptococcal pneumonia during pregnancy, and 1 pregnant woman with cryptococcal meningitis. These cases illustrate a previously uncharacterized spectrum of cryptococcal disease. We also discuss 24 patients previously reported who had cryptococcal meningitis during pregnancy. Finally, we review the available data for each therapeutic option and present an algorithm for management based on appraisals of disease severity and risk to the unborn fetus. This report emphasizes the need for heightened awareness of cryptococcosis in the differential diagnosis of pneumonia, chest pain, and hypoxemia in the pregnant patient, but at present, there are insufficient epidemiologic data to determine whether incidences of pulmonary or disseminated cryptococcosis actually increase during pregnancy. The risk of congenital cryptococcosis to the unborn fetus is low, and the most likely mechanism whereby neonates acquire invasive fungal pulmonary infection is through aspiration. While it is unclear whether there is any real increased risk of spontaneous abortion or premature labor, the data indicate that overall fetal outcome depends on effective treatment of maternal infection. For patients with dense air-space consolidation, progressive pulmonary disease, or dissemination, antifungal therapy is necessary. Optimal treatment is determined by the acuity and severity of the clinical presentation. Amphotericin B (approximately 1 g) with or without flucytosine represents the choice for initial treatment of the more acutely ill patient with disseminated or progressive pulmonary cryptococcosis who requires hospitalization (whether during or after pregnancy). Oral fluconazole appears to be safe and effective alternative therapy after delivery for the less severely ill patient who can be managed on an outpatient basis. While the use of fluconazole during pregnancy generally appears safe in terms of fetal outcome (49, 58), the class C status and single report of fetal malformation (62) preclude confident recommendation for its use during pregnancy. The risks and benefits of this effective and generally less toxic drug should be discussed with the parents and weighed against the use of amphotericin B. For pregnant women with limited pulmonary cryptococcosis (segmental or nodular infiltrates) and no evidence of dissemination, we recommend close follow-up without antifungal therapy similar to the recommendation for normal hosts with minimal disease. However, it is important to note that there is no extensive experience upon which to base this recommendation for pregnant individuals (45, 55, 103, 108). It is prudent to use frequent physical examinations (for example, every 1-2 months), combined with chest roentgenograms and serum cryptococcal antigens to monitor progression and/or development of disease in both the mother and child for approximately 6 months postpartum. In conclusion, cryptococcosis during pregnancy presents a special challenge to the clinician. A balanced therapeutic approach holds great promise for successful maternal and fetal outcomes.