Heart transplantation is the only cure for patients with terminal cardiac failure, but the supply of allogeneic donor organs falls far short of the clinical need1,2,3. Xenotransplantation of genetically modified pig hearts has been discussed as a potential alternative4. Genetically multi-modified pig hearts that lack galactose-α1,3-galactose epitopes (α1,3-galactosyltransferase knockout) and express a human membrane cofactor protein (CD46) and human thrombomodulin have survived for up to 945 days after heterotopic abdominal transplantation in baboons5. This model demonstrated long-term acceptance of discordant xenografts with safe immunosuppression but did not predict their life-supporting function. Despite 25 years of extensive research, the maximum survival of a baboon after heart replacement with a porcine xenograft was only 57 days and this was achieved, to our knowledge, only once6. Here we show that α1,3-galactosyltransferase-knockout pig hearts that express human CD46 and thrombomodulin require non-ischaemic preservation with continuous perfusion and control of post-transplantation growth to ensure long-term orthotopic function of the xenograft in baboons, the most stringent preclinical xenotransplantation model. Consistent life-supporting function of xenografted hearts for up to 195 days is a milestone on the way to clinical cardiac xenotransplantation7.