During their life cycle as farmed animals, there are several situations in which fish are subjected to handling and confinement. Netting, weighing, sorting, vaccination, transport and, at the end, slaughter are frequent events under farming conditions. As research subjects, fish may also undergo surgical procedures that range from tagging, sampling and small incisions to invasive procedures. In these situations, treatment with anaesthetic agents may be necessary in order to ensure the welfare of the fish. The main objective of this paper is to review our knowledge of the effects of anaesthetic agents in farmed fish and their possible implications for welfare. As wide variations in response to anaesthesia have been observed both between and within species, special attention has been paid to the importance of secondary factors such as body weight, water temperature and acute stress. In this review, we have limited ourselves to the anaesthetic agents such as benzocaine, metacaine (MS-222), metomidate hydrochloride, isoeugenol, 2-phenoxyethanol and quinaldine. Anaesthetic protocols of fish usually refer to one single agent, whereas protocols of human and veterinary medicine cover combinations of several drugs, each contributing to the effects needed in the anaesthesia. As stress prior to anaesthesia may result in abnormal reactions, pre-anaesthetic sedation is regularly used in order to reduce or avoid stress and is an integral part of the veterinary protocols of higher vertebrates. Furthermore, the anaesthetic agents that are used in order to obtain general anaesthesia are combined with analgesic agents that target nociception. The increased use of such combinations in fish is therefore included as a special section. Anaesthetic agents are widely used to avoid stress during various farming procedures. While several studies report that anaesthetics are effective in reducing the stress associated with confinement and handling, there are indications that anaesthesia may in itself induce a stress response, measured by elevated levels of cortisol. MS-222 has been reported to elicit high cortisol release rates immediately following exposure, while benzocaine causes a bimodal response. Metomidate has an inhibitory effect on cortisol in fish and seems to induce the lowest release of cortisol of the agents reported in the literature. Compared to what is observed following severe stressors such as handling and confinement, the amount of cortisol released in response to anaesthesia appears to be low but may represent an extra load under otherwise stressful circumstances. Furthermore, anaesthetics may cause secondary adverse reactions such as acidosis and osmotic stress due to respiratory arrest and insufficient exchange of gas and ions between the blood and the water. All in all, anaesthetics may reduce stress and thereby improve welfare but can also have unwanted side effects that reduce the welfare of the fish and should therefore always be used with caution. Finally, on the basis of the data reported in the literature and our own experience, we recommend that anaesthetic protocols should always be tested on a few fish under prevailing conditions in order to ensure an adequate depth of anaesthesia. This recommendation applies whether a single agent or a combination of agents is used, although it appears that protocols comprising combinations of agents provide wider safety margins. The analgesic effects of currently used agents, in spite of their proven local effects, are currently being debated as the agents are administrated to fish via inhalation rather than locally at the target site. We therefore recommend that all protocols of procedures requiring general anaesthesia should be complemented by administration of agents with analgesic effect at the site of tissue trauma.