In this review, we summarize the available information on the short- and long-term effects of pregnancy on the course of multiple sclerosis (MS). Published studies that used established criteria for the diagnosis of MS were given more weight than studies in which the criteria for diagnosis were unstated or unclear. Population-based studies were emphasized more than clinic-based studies, unless the clinic base was well defined and thought to be reasonably representative of the MS population in the geographic area. For completeness, small studies were also included but weighted accordingly in our overall conclusions. Methodologic limitations and biases inherent in the study methods are discussed. We conclude that patients with relapsing MS have an increased risk of relapse during the initial 6-month postpartum period. This increased risk does not seem to have a detrimental effect on the rate of developing sustained disability. In fact, a full-term pregnancy may increase the time interval to reaching a common disability endpoint-walking with the aid of a cane or crutch--or to having a secondarily progressive course. Evidence indicates that pregnancy may alter T-lymphocyte functions and cause clinically relevant consequences. The specific biochemical mechanisms responsible for these observations, however, remain undefined. Because of limitations of current knowledge, our conclusions are tentative at best. The data are most applicable to patients with relapsing-remitting MS in its early stages. MS is an unpredictable disease and is only one of many factors that patients must consider when a pregnancy is contemplated.