Supplementary Material for: Influence of Hormonal Profile on Resting Metabolic Rate in Normal, Overweight and Obese Individuals

<b><i>Aims:</i></b> To investigate whether blood thyroid stimulating hormone (TSH), cortisol, insulin and glucose concentrations (plus glucose:insulin ratio; GIR) could improve the accuracy of resting metabolic rate (RMR) prediction in normal, overweight and obese persons. <b><i>Methods:</i></b> Predictive equations were developed and compared against indirect calorimetry measures for RMR in 217 weight-control clinic participants (n = 128 males and n = 89 females: ∼24% normal weight, ∼39% overweight and ∼37% obese). <b><i>Results:</i></b> Using the common accuracy criteria of the proportion of predicted RMR within ±10% of measured RMR, our equations (using age, height, weight and gender, plus the blood factors, both independently and in combination) were accurate ∼36-44% of the time, for the whole sample, and when separated by gender and weight class. Specifically, the addition of the blood hormone and glucose concentrations improved the accuracy of predicted RMR by only 1-8% (NS). <b><i>Conclusions:</i></b> Including blood TSH, cortisol, insulin, glucose and GIR into RMR prediction equations did not significantly improve estimation accuracy, which in any case only met a criterion of ±10% of the measured RMR ∼40% of the time. Further work to refine the prediction of RMR is still needed, and at present, direct measurements should be made wherever possible.

<b><i>研究目的：</i></b> 探讨血清促甲状腺激素（thyroid stimulating hormone, TSH）、皮质醇、胰岛素及葡萄糖浓度（辅以葡萄糖/胰岛素比值（glucose:insulin ratio, GIR））是否可提升正常体重、超重及肥胖人群静息代谢率（resting metabolic rate, RMR）预测的准确性。<b><i>研究方法：</i></b> 本研究纳入217名体重管理门诊参与者（男性128名，女性89名；其中约24%为体重正常者、约39%为超重者、约37%为肥胖者），构建静息代谢率预测方程，并将其与间接测热法测得的静息代谢率结果进行对比。<b><i>研究结果：</i></b> 以「预测静息代谢率与实测值偏差在±10%范围内」作为通用准确性评价标准，针对全样本及按性别、体重分层的亚组分析显示，本研究构建的方程（纳入年龄、身高、体重、性别，辅以血液指标，可单独使用或联合应用）的准确预测占比约为36%~44%。具体而言，添加血液激素及葡萄糖浓度仅使静息代谢率预测准确性提升1%~8%，差异无统计学显著性。<b><i>研究结论：</i></b> 在静息代谢率预测方程中纳入血清促甲状腺激素、皮质醇、胰岛素、葡萄糖浓度及GIR，并未显著提升预测准确性；且无论如何，该模型仅在约40%的情况下满足「预测值与实测静息代谢率偏差在±10%范围内」的标准。目前仍需开展进一步研究以优化静息代谢率的预测效果，现阶段应尽可能采用直接测量法。