Thyroid hormones have long been known to stimulate energy expenditure partly via loss of metabolic efficiency. The mechanism underlying the loss in metabolic efficiency observed, however, is not yet understood. An important candidate gene responsible for thyroid hormone induced thermogenesis was identified in 1997 with the discovery of skeletal muscle-uncoupling protein 3 (UCP3), a protein with approximately 60 % homology to the brown adipose tissue uncoupling protein 1 (UCP1). This short review summarizes our presentation held at the 'Thyroid and Sports' meeting; it does not aim to provide a concise overview of the available literature at this topic. Although induction of the UCP3 gene and increased protein expression during hyperthyroidism has been shown, there are no convincing data that increased UCP3 levels account for the increase in thermogenesis in the hyperthyroid state in humans. In contrast to cell and animal studies using ectopic overexpression of UCP3 as a model, induction of UCP3 in humans does not result in any apparent mitochondrial uncoupling. Hence, the primary physiological role of UCP3 may not be mitochondrial uncoupling, but uncoupling may occur as a side effect of a more pivotal role played by UCP3. Recently, UCP3 has been hypothesized to export fatty acid anions and/or lipid peroxides away from the mitochondrial matrix to prevent mitochondria from the harmful effects of peroxidized lipids. The present review aims to provide an overview of studies testing the feasibility of this unconventional function of UCP3.