Hepatic steatosis often occurs in the context of insulin resistance, dyslipidaemia, type 2 diabetes mellitus and obesity. It is associated with a risk for development of more severe forms of non-alcoholic fatty liver disease (NAFLD), including steatohepatitis, fibrosis and cirrhosis. The amount of triglycerides in an intrinsically normal liver is not fixed, but can readily be modified by nutritional conditions. For instance, short term fasting increases hepatic triglyceride content (through increased release of fatty acids from the adipose tissue). Conversely, prolonged caloric restriction reduces liver fat content. In addition, many other (patho)physiological factors can modulate hepatic triglyceride storage, including alcohol, drugs and dietary composition. Importantly, the propensity to develop steatosis in a certain (patho)physiological state may be determined by genes and/or physical fitness. In this issue of The Journal of Physiology, Thyfault et al. (2009) report the results of a careful study on the biochemical and molecular differences in the livers of rats, which were selectively bred to robustly differ with respect to their intrinsic aerobic capacity. They show that sedentary rats with genetically determined low aerobic fitness develop steatosis and more severe hepatic fibrosis at natural death compared to sedentary rats with intrinsically high aerobic capacity. They also demonstrate that low aerobic capacity is linked to decreased mitochondrial content and mitochondrial oxidative capacity in liver. Isolated hepatic mitochondria from high capacity runners oxidized palmitate more completely (to CO2) and in greater amounts, whereas markers of hepatic peroxisomal activity, acyl CoA oxidase and catalase, were decreased in these animals. A key transcriptional factor for fatty acid synthesis, SREBP-1c, was lower in livers of high capacity runners. These remarkable observations require careful interpretation. The study design has some limitations. In particular, the two groups of rats differed not only with respect to their aerobic capacity, but also in terms of other factors that clearly affect hepatic triglyceride content. For instance, the rats with low aerobic capacity were 42% heavier and the adipocyte cell volume of their omental fat pad was 50% larger. Moreover, they were hyperinsulinaemic, hypertriglyceridaemic, and considerably less active than the high capacity runners. The authors postulate that reduced mitochondrial oxidation in hepatocytes of rats with low aerobic capacity is directly responsible for their propensity to store triglycerides in liver. However, it is conceivable that reduced muscle mitochondrial content, a putative determinant of insulin resistance, hyperinsulinaemia and obesity, indirectly caused hepatic triglyceride accumulation in these animals. The design of the current study does not allow precise conclusions as to which of these potential pathophysiological factors underlies their findings. It is likely that different mechanistic routes contribute. Indeed, the liver may not be merely an innocent bystander, passively adapting to the changes in fatty acid metabolism associated with muscle fitness. Rather, the results raise the possibility that there are active intrahepatic biochemical and molecular mechanisms associated with aerobic fitness that affect hepatic fatty acid metabolism, and, as a consequence, liver triglyceride content. This concept is also supported by a previous study of the authors, which documented that sudden cessation of daily physical activity in hyperphagic/obese rats resulted in stimulation of biochemical pathways known to initiate hepatic steatosis, including decreased hepatic mitochondrial oxidative capacity, increased hepatic expression of de novo lipogenesis proteins, and increased hepatic malonyl CoA levels (Rector et al. 2008). This perspective contrasts with the concept of the liver acting as an innocent bystander, passively storing fatty acids in the form of triglycerides. Rather, aerobic fitness may not be reflected by adaptations of skeletal muscle only, but also involve the liver (and perhaps other tissues as well). The findings reported by Thyfault and colleagues warrant further investigation of the mechanistic underpinnings of this putative muscle–liver axis, because deeper insight may open new alleys for the treatment of non-alcoholic fatty liver disease.