A 60-hour high-carbohydrate (high-CHO) diet or a 36-hour low-carbohydrate (low-CHO) diet was followed by 24 healthy women in a cross-over design to modify liver glycogen content. Thereafter each subject was given a high-sucrose breakfast, a high-protein, high-fibre breakfast or no breakfast. The two different breakfasts evoked larger plasma glucose responses following the low-CHO diet than when following the high-CHO diet. When the two breakfasts followed the same pre-period diet, no significant differences were observed. We conclude that the composition of the previous diet influences the postprandial response to meals and that a standardised diet shortly before test meal studies is of importance for the results in this type of studies.