The present study assessed the effect of separate reduction of each energy-delivering nutrient – protein, fat and carbohydrate – on glucose tolerance and insulin response in a strict carnivore: the domestic cat (Felis catus). Three isoenergetic, home-made diets with the following energetic distribution, low protein (LP): protein 28 % of metabolisable energy; fat 43 %; nitrogen-free extract 29 %; low fat: 47, 27 and 25 %; low carbohydrate (LC): 45, 48 and 7 %, were tested in a 3 × 3 Latin square design. Nine healthy normal-weight cats were randomly assigned to each of the diets in a random order at intervals of 3 weeks. At the end of each testing period, intravenous glucose tolerance tests were performed. Plasma glucose concentrations and area under the glucose curve showed no differences. Area under the insulin curve was lower when cats were fed the LP diet, and the second insulin peak tended to be delayed when the LC diet was fed. In contrast to other studies, in which energy sources were elevated instead of being reduced, the present trial contradicts the often suggested negative impact of carbohydrates on insulin sensitivity in carnivores, and shows that reducing the dietary carbohydrate content below common amounts for commercial foods evokes an insulin-resistant state, which can be explained by the cats' strict carnivorous nature. It even points to a negative effect of protein on insulin sensitivity, a finding that corresponds with the highly gluconeogenic nature of amino acids in strict carnivores.