Low-carbohydrate diets for the treatment of obesity are currently receiving widespread popularity, mainly in Anglo-Saxan countries. Several controlled trials have demonstrated larger weight loss (3-6 kg) during these diets compared to conventional low-fat diets. The weight differences between these diets were statistically significant after 6 months but not any more after 12 months. Possible reasons for the increased weight loss may be the initial loss of body water, later on the diminishment of food choices, the satiating properties of proteins and the anorectic effect of ketosis. Controlled studies have not demonstrated any increase in LDL-cholesterol but a favorable increase in HDL-C and a lowering of serum triglycerides. Even if these diets "work" in selected patients, they are problematic because the increase in protein ingestion may lead to increased intake of saturated fat, and intake of water soluble antioxidants, vitamins and of fiber may be too low. For these reasons, their protective effect on the "major killers " atherosclerotic diseases, hypertension, diabetes type 2 and some forms of cancer is questionable. Moreover, these diets are unpleasant on the long run, and social eating is impaired. For all these reasons they should not be generally recommended. Reasonable nutritional changes for the treatment of obesity should be suitable for lifelong use, and they should always be combined with increased physical activity and with behavioral measures.