Summary Many predictive equations for basal metabolic rate (BMR) based on anthropometric measurements, age, and sex have been developed, mainly for healthy Caucasians. However, it has been reported that many of these equations, used widely, overestimate BMR not only for Asians, but also for Caucasians. The present study examined the accuracy of several predictive equations for BMR in Japanese subjects. In 365 healthy Japanese male and female subjects, aged 18 to 79 y, BMR was measured in the post-absorptive state using a mask and Douglas bag. Six predictive equations were examined. Total error was used as an index of the accuracy of each equation's prediction. Predicted BMR values by Dietary Reference Intakes for Japanese (Japan-DRI), Adjusted Dietary Reference Intakes for Japanese (Adjusted-DRI), and Ganpule equations were not significantly different from the measured BMR in either sex. On the other hand, Harris-Benedict, Schofield, and Food and Agriculture Organization of the United Nations/World Health Organization/United Nations University equations were significantly higher than the measured BMR in both sexes. The prediction error by Japan-DRI, Adjusted-DRI, and Harris-Benedict equations was significantly correlated with body weight in both sexes. Total error using the Ganpule equation was low in both males and females (125 and 99 kcal/d, respectively). In addition, total error using the Adjusted-DRI equation was low in females (95 kcal/d). Thus, the Ganpule equation was the most accurate in predicting BMR in our healthy Japanese subjects, because the difference between the predicted and measured BMR was relatively small, and body weight had no effect on the prediction error.

Summary Several cross-sectional studies in Pima Indians and Caucasians have indicated that obese individuals with type 2 diabetes have a higher basal metabolic rate (BMR) than healthy, obese individuals. However, no study has investigated this comparison in Japanese subjects, who are known to be susceptible to type 2 diabetes due to genetic characteristics. Thirty obese Japanese adults with pre-type 2 diabetes ( n ϭ 7) or type 2 diabetes ( n ϭ 13) or without diabetes ( n ϭ 10) participated in this study. BMR was measured using indirect calorimetry. The relationships between residual BMR (calculated as measured BMR minus BMR adjusted for fat-free mass, fat mass, age, and sex) and biomarkers including fasting glucose, glycosylated hemoglobin (HbA 1c ), fasting insulin, homeostasis model assessment of insulin resistance (HOMA-R), triglycerides, and free fatty acids were examined using Pearson's correlation. BMR in diabetic subjects adjusted for fat-free mass, fat mass, age, and sex was 7.1% higher than in non-diabetic subjects. BMR in diabetic subjects was also significantly ( p Ͻ 0.05) higher than in non-diabetic subjects. There was a significant correlation between residual BMR and fasting glucose ( r ϭ 0.391, p ϭ 0.032). These results indicate that in the Japanese population, obese subjects with type 2 diabetes have higher BMR compared with obese non-diabetic subjects. The fasting glucose level may contribute to these differences.

To express intensity of physical activity, energy expenditure is often divided by either body weight, resting metabolic rate, or fat-free mass. These calculations are used widely as the physical activity index. However, it is unclear how body size influences the valid estimation of intensity of various kinds of activities. In the present study, we investigated whether these indices are able to adjust for body size when calculating energy expenditure in various kinds of activities. In addition, we examined to what extent the error of index is introduced by differences in body size. Resting metabolic rates and energy expenditure during sitting light work, 4 lifestyle and 7 ambulant activities were measured in the postabsorptive state using indirect calorimetry in 71 healthy Japanese adults. We regarded an index as an inappropriate adjustment for body size when there was a significant correlation between it and body weight. Energy expenditure normalized by body weight correlated with body weight in all sedentary states; when normalized by lying resting metabolic rate it correlated with body weight in 3 ambulant activities; when normalized by sitting resting metabolic rate it correlated with body weight in 2 lifestyle and 5 ambulant activities; and when normalized by fat-free mass it correlated with only 1 ambulant activity. The indices caused errors in estimates of activity intensity of less thanϮ10% when body weight was more than 10 kg above average. In conclusion, the body weight-normalized index was inappropriate for sedentary activities and the other three indices were inappropriate for ambulant activities. However, the use of any of these indices introduces an error in the estimate of total energy expenditure of considerably less thanϮ10% for body weights within the normal range.