In some cases tubal pregnancy resolves spontaneously. The risk of subsequent surgical intervention due to either tubal rupture or the entry criteria of the study varies from 0 per cent to 31 per cent. The major problem in nonsurgical treatment of tubal pregnancy is the absence of a parameter that reveals the threat of tubal rupture. In addition, data on the functional recovery of the fallopian tube are controversial. The scarcity of data on medical treatment with RU486, glucose 50 per cent, KCL, and actinomycin-D make proper evaluation impossible. Both MTX and prostaglandin treatment should be investigated further. Compiled data on prostaglandin treatment in cases of unruptured tubal pregnancy do not show better results than data on expectant management only. If, however, patients with initial serum hCG levels greater than 1000 mIU/ml or greater than 2500 mIU/ml are excluded from this therapy, the risk of tubal rupture diminishes. Side effects are minimal, especially if injection into the corpus luteum is omitted. Compiled data on MTX treatment in cases of unruptured tubal pregnancy show a crude risk of subsequent surgical intervention of 5 per cent. If patients with an initial serum hCG level exceeding 10,000 mIU/ml are excluded, the risk of tubal rupture is limited to 3 per cent. (The estimated risk of persistent trophoblastic activity after conservative surgical therapy is also 5 per cent.) Studies on the optimum MTX dosage, treatment scheme, and method of administration are still going on. Side effects are reversible and minimal. Theoretically, the local injection of MTX is more effective. Although often used to propagate a new way of treatment, fertility in the future is a questionable parameter in the evaluation of therapy. Fertility is influenced by so many factors other than the method of treatment that it can only be used for treatment evaluation in a case control or a randomized prospective study. Such a study has yet to be published. Besides the influence on future fertility, other results of treatment, such as morbidity, cost, and length of hospital stay should be taken into account.