Background: Exposure to isotretinoin during pregnancy is associated with a high risk of major fetal malformations. Objective: Our purpose was to determine the reasons for and outcomes of in utero isotretinoin exposure. Methods: On the basis of 433 spontaneous reports, we describe the likely reasons these exposures occurred and the outcomes of these pregnancies. In our analysis of the outcomes of pregnancy, we separately consider the pregnancies known to us before their outcomes were determined as prospective cases. Results: Timing of conception in relation to initiation of therapy with isotretinoin was known for 396 women. Of these, 130 patients (33%) were already pregnant when they started isotretinoin. An additional 65 patients (16%) became pregnant in the first 3 weeks of isotretinoin use. Pregnancy outcomes were known on 409 pregnancies. Among these, 222 (54%) ended in elective abortion and 29 (7%) in spontaneous or missed abortion. Of 151 births, 72 (48%) were normal, 71 (47%) had congenital malformations, and 8 (5%) had abnormalities other than malformations. Of 94 prospectively ascertained pregnancies that ended in births, 28% had congenital malformations (95% confidence interval 19% to 37%). Exposure to isotretinoin during any time and for as little as one capsule within the first trimester have been associated with congenital malformations. Conclusion: The high proportion of exposures in already or recently pregnant women illustrates the importance of obtaining a negative pregnancy test before the initiation of isotretinoin therapy and of delaying the commencement of isotretinoin therapy until the onset of the next menstrual period. Furthermore, the importance of reliable contraceptive methods should be emphasized to patients when isotretinoin is prescribed. Young women seem to be at an especially high risk of pregnancy exposure to isotretinoin. There is a substantial risk of congenital malformation at all therapeutic doses of isotretinoin, even when the duration of exposure is brief.