Three isoproteic (crude protein content: 56%) diets with different fat levels (11%, 20%, and 26%) were fed to triplicate groups of triploid brown trout (initial average body weight of 1.5 kg), reared in seawater. At the end of 3 months of feeding, fish fed the high-fat (HF) diet were split into two groups: a triplicate group of fish received the low-fat diet and another triplicate group was kept unfed for a further 2-month period. Fish initially fed the low-fat diet during the first period were continued to be fed the same diet. Fish fed the medium-fat (MF) diet during period 1 were eliminated for period 2. At the end of each period, comparative whole body analyses, sensory and instrumental (texture and colour) analyses were made on fresh and smoked fillets. During the first period, increasing dietary fat level had no significant effect on growth or feed utilisation, but increased whole body (14.6% to 17.9%, on wet weight basis) and muscle (8.3% to 11.0%) fat content. During the second period, the fish fed the low-fat diet had similar growth performance irrespective of previous nutritional history, whereas starvation led to significant loss of weight and fillet yield. Whole body fat content did not differ between groups (around 15%) at the end of period 2. In fish initially fed the HF diet, both starvation and feeding a low-fat diet led to a reduction in muscle lipid content. Sensory analyses revealed few differences between treatments, in terms of visual colour aspects, for both cooked and smoked fillets at the end of period 1. A positive relationship between instrumental colour analyses (L*, a*, b* values) and dietary fat levels was observed, but no difference was observed for instrumental texture measurements. At the end of period 2, a significant increase in parameters of colour was observed in unfed fish. Although both feed withdrawal and feeding a low-fat diet 2 months before slaughtering led to a reduction in fat content, starvation had the disadvantage of leading to significant weight loss.