OBJECTIVE: To compare the antimetabolites methotrexate and 6-mercaptopurine as single-agent medical abortifacients using clinical and immunohistochemical analyses. METHODS: Twenty-seven women with gestations less than 7 weeks from the last menstrual period (LMP) were randomized to receive intramuscular methotrexate, 50 mg/m2, or oral 6-mercaptopurine, 200 mg. Forty-six additional women received methotrexate after randomization was discontinued. Women returned at 2-week intervals. Those without fetal cardiac activity were followed until complete abortion. Those with fetal cardiac activity were considered failures and underwent suction abortions. Tissue collected at the time of suction abortion was analyzed with the cell-proliferation immunohistochemical assay Ki-67. RESULTS: All 12 women in the 6-mercaptopurine group had fetal cardiac activity at follow-up and underwent suction abortion; therefore, this arm of the study was discontinued. Six of the 61 women who received methotrexate had fetal cardiac activity at follow-up and also underwent suction abortion. Fetal cardiac activity was present after methotrexate in three of 55 women at less than 6 weeks from the LMP and in three of six between 6 and 7 weeks from the LMP (P < .01). Women who aborted after methotrexate started bleeding on day 19 (standard deviation [SD] 7.8), bled for 9 days (SD 4.0), and used minimal pain medications. Tissues exposed to methotrexate showed decreased Ki-67 activity compared with tissues exposed to 6-mercaptopurine (P = .003). CONCLUSION: In oral doses of 200 mg, 6-mercaptopurine did not induce early abortion. A single intramuscular dose of methotrexate used without prostaglandins induced abortion in most women at gestational ages of less than 6 weeks. Ki-67 activity was lower in a small sample of fetal tissues exposed to methotrexate than in tissues exposed to 6-mercaptopurine.