Tibolone is used to treat climacteric complaints and prevent osteoporosis. These beneficial effects are exerted via its 3␣-and 3-hydroxymetabolites. Undesirable stimulation of the breast and endometrium is not apparent. Endometrial stimulation is prevented by the progestogenic activity of its ⌬ 4 -ene metabolite. The enzymes responsible for the formation of these active metabolites are unknown. Human aldo-keto reductase (AKR)1C isoforms have been shown to act as 3␣/3-hydroxysteroid dehydrogenases (HSDs) on 5␣-dihydrotestosterone (5␣-DHT). We show that AKR1Cs also efficiently catalyze the reduction of the ⌬ 5(10) -3-ketosteroid tibolone to yield 3␣-and 3-hydroxytibolone. Homogeneous recombinant AKR1C1, AKR1C3, and AKR1C4 gave similar catalytic profiles to those observed with 5␣-DHT. AKR1C1 catalyzed exclusively the formation of 3-hydroxytibolone, AKR1C3 showed weak 3/3␣-HSD activity, and AKR1C4 acted predominantly as a 3␣-HSD. Whereas AKR1C2 acted as a 3␣-HSD toward 5␣-DHT, it functioned exclusively as a 3-HSD on tibolone. Furthermore, strong substrate inhibition was observed for the AKR1C2 catalyzed reduction of tibolone. Using NAD ϩ , the 3-hydroxymetabolites were efficiently oxidized by homogeneous recombinant AKR1C2 and AKR1C4. However, because of potent inhibition of this activity by NADPH, AKR1Cs will probably act only as 3-ketosteroid reductases in vivo. Molecular docking simulations using crystal structures of AKR1C1 and AKR1C2 explained why AKR1C2 inverted its stereospecificity from a 3␣-HSD with 5␣-DHT to a 3-HSD with tibolone. The preference for AKR1C1 and AKR1C2 to form 3-hydroxytibolone, and the preference of the liver-specific AKR1C4 to form 3␣-hydroxytibolone, may explain why 3-hydroxytibolone is the major metabolite in human target tissues and why 3␣-hydroxytibolone is the major circulating metabolite.