Moderate low-carbohydrate/high-fat diets (LC-HFDs) are widely used to induce weight loss in overweight subjects, whereas extreme ketogenic LC-HFDs are used to treat neurological disorders like pediatric epilepsy. Usage of LC-HFD for improvement of glucose metabolism is highly controversial: some studies suggest that LC-HFDs ameliorate glucose tolerance whereas others could not find positive effects of these diets or reported impaired insulin sensitivity. Here we investigate the effects of LC-HFDs on glucose and insulin metabolism in a well-characterised animal model: rats were fed isoenergetic or hypocaloric amounts of standard control diet, a high-protein, "Atkins-style" LC-HFD, or a low-protein, ketogenic, LC-HFD. Both LC-HFDs induced lower fasting glucose and insulin levels associated with lower pancreatic β-cell volumes. However, dynamic challenge tests (oral and intraperitoneal glucose tolerance tests, insulin-tolerance tests and hyperinsulinemic-euglycemic clamps) revealed that LC-HFD pair-fed rats exhibited impaired glucose tolerance, and impaired hepatic and peripheral tissue insulin sensitivity, the latter potentially being mediated by elevated intramyocellular lipids. Adjusting visceral fat mass in LC-HF groups to that of controls by reducing the intake of LC-HF diets to 80% of the pair-fed groups did not prevent glucose intolerance. Taken together, these data show that lack of dietary carbohydrates leads to glucose intolerance and insulin resistance in rats despite causing a reduction in fasting glucose and insulin concentrations. Our results argue against beneficial effects of LC-HFDs on glucose and insulin metabolism, at least under physiological conditions. Therefore, use of LC-HFDs for weight loss or other therapeutic purposes should be balanced against potentially harmful metabolic side effects.