Centre for Medical and Exercise Physiology, School of Medicine, University of Wollongong, Wollongong, NSW, 2522, Australia.
Eur J Appl Physiol. 2021 Jan;121(1):193-208. doi: 10.1007/s00421-020-04515-1. Epub 2020 Oct 3.
In tachymetabolic species, metabolic rate increases disproportionately with body mass, and that inter-specific relationship is typically modelled allometrically. However, intra-specific analyses are less common, particularly for healthy humans, so the possibility that human metabolism would also scale allometrically was investigated.
Basal metabolic rate was determined (respirometry) for 68 males (18-40 years; 56.0-117.1 kg), recruited across five body-mass classes. Data were collected during supine, normothermic rest from well-rested, well-hydrated and post-absorptive participants. Linear and allometric regressions were applied, and three scaling methods were assessed. Data from an historical database were also analysed (2.7-108.9 kg, 4811 males; 2.0-96.4 kg, 2364 females).
Both linear and allometric functions satisfied the statistical requirements, but not the biological pre-requisite of an origin intercept. Mass-independent basal metabolic data beyond the experimental mass range were not achieved using linear regression, which yielded biologically impossible predictions as body mass approached zero. Conversely, allometric regression provided a biologically valid, powerful and statistically significant model: metabolic rate = 0.739 * body mass (P < 0.05). Allometric analysis of the historical male data yielded an equivalent, and similarly powerful model: metabolic rate = 0.873 * body mass (P < 0.05).
It was established that basal and resting metabolic rates scale allometrically with body mass in humans from 10-117 kg, with an exponent of 0.50-0.55. It was also demonstrated that ratiometric scaling yielded invalid metabolic predictions, even within the relatively narrow experimental mass range. Those outcomes have significant physiological implications, with applications to exercising states, modelling, nutrition and metabolism-dependent pharmacological prescriptions.
在代谢速度快的物种中，代谢率与体重不成比例地增加，这种种间关系通常通过幂函数关系进行建模。然而，种内分析则较少见，特别是对于健康人类，因此，人类代谢也可能按幂函数关系进行缩放的可能性被进行了研究。
通过呼吸测量法确定了 68 名男性（18-40 岁；56.0-117.1kg）的基础代谢率，这些男性是在五个体重类别中招募的。数据是在仰卧、体温正常、休息、充分休息、充分水合和吸收后状态下的参与者中收集的。线性和幂函数回归均被应用，且评估了三种缩放方法。还分析了历史数据库中的数据（2.7-108.9kg，男性 4811 人；2.0-96.4kg，女性 2364 人）。
线性和幂函数函数都满足统计要求，但不满足生物学上零质量起点的前提要求。使用线性回归无法在实验质量范围之外获得无质量依赖性的基础代谢数据，因为当体重接近零时，线性回归会产生生物学上不可能的预测。相反，幂函数回归提供了一个生物学上有效、强大且具有统计学意义的模型：代谢率=0.739体重（P<0.05）。对历史男性数据的幂函数分析产生了等效的、同样强大的模型：代谢率=0.873体重（P<0.05）。
研究结果表明，从 10-117kg 之间的人体基础和静息代谢率与体重呈幂函数关系，幂指数为 0.50-0.55。研究还表明，即使在相对较窄的实验质量范围内，比率缩放也会产生无效的代谢预测。这些结果具有重要的生理学意义，可应用于运动状态、建模、营养和代谢依赖的药物处方。
