The glycaemic index (GI) has been developed in order to classify food according to the postprandial glycaemic response. This parameter is of interest, especially for people prone to glucose intolerance; however, the effects of a low-GI (LGI) diet on body weight, carbohydrate and lipid metabolism remain controversial. We studied the effects of either a LGI or high-GI (HGI) diet on weight control and cardiovascular risk factors in overweight, non-diabetic subjects. The study was a randomized 5-week intervention trial. The thirty-eight subjects (BMI 27·3 (sem 0·2) kg/m2) followed an intervention diet in which usual starch was replaced ad libitum with either LGI or HGI starch. Mean body weight decrease was significant in the LGI group ( − 1·1 (sEM 0·3) kg, P = 0·004) and was significantly greater than in the HGI group ( − 0·3 (sEM 0·2) kg, P = 0·04 between groups). Hunger sensation scales showed a trend towards a decrease in hunger sensation before lunch and dinner in the LGI group when compared with the HGI group (P = 0·09). No significant increase in insulin sensitivity was noticed. The LGI diet also decreased total cholesterol by 9·6 % (P < 0·001), LDL-cholesterol by 8·6 % (P = 0·01) and both LDL-:HDL-cholesterol ratio (10·1 %, P = 0·003) and total:HDL-cholesterol ratio (8·5 %, P = 0·001) while no significant changes were observed in the HGI group. Lowering the GI of daily meals with simple dietary recommendations results in increased weight loss and improved lipid profile and is relatively easy to implement with few constraints. These potential benefits of consuming a LGI diet can be useful to develop practical dietetic advice.