The reproducibility of measurements of postprandial energy expenditure (PEE) was determined by assessing, for 345 min, the diet-induced thermogenesis (DIT) of a 4055-kJ mixed meal in 10 lean young men on four occasions. The DIT was determined by using either a 20-min premeal baseline metabolic rate (method 1) or a single resting metabolic rate (RMR) measured on a separate day (method 2). The DIT, in kilojoules, for a given subject calculated by method 1 varied significantly between occasions, CV 15.4% (P < 0.05). The D peak (PEE peak minus baseline) did not vary significantly between the four trials (CV 13%). The results obtained by using method 2 were much more reproducible for both the DIT (CV 4.2%) and the D peak (CV 4.8%). Thus, the reproducibility of the PEE measurement is enhanced by using an RMR measured on a separate day and by using the postprandial peak value instead of the DIT.