Although animal experiments play an important role in biomedical research, their use is ethically challenging. Primarily in Europe, North America and Australasia ethics committees are set up to control the animal use in science. Project approval is usually decided on a case-by-case basis with focus on ensuring that the animals are caused a minimum of harm relative to the possibility of achieving beneficial results. Even though rules in this area are reasonably uniform there seems to be significant room for differences, individual and culturally based, between ethics committees concerning how the rules are applied. Our aim was to conduct a review of empirical studies of the different kinds of animal ethics committees in order to clarify what is known about their operation and highlight information which is missing in their evaluation. Our main findings are that there is a significant variation in process and outcomes of decision-making at individual and group levels which cause inconsistency between decisions. Different approaches have been suggested to improve the reliability of ethical review but no evidence to support any of them. More empirical studies are needed.