National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, GA, USA.
J Nutr. 2013 Jun;143(6):1001S-10S. doi: 10.3945/jn.112.172882. Epub 2013 Apr 17.
The physiologic status of an individual may influence biomarkers of nutritional status. To help researchers with planning studies and interpreting data, we assessed the associations between common physiologic variables (fasting, inflammation, renal function, and pregnancy) and 29 biomarkers of diet and nutrition measured in blood or urine in a representative sample of the adult U.S. population (aged ≥ 20 y; pregnancy variable and iron indicators limited to women aged 20-49 y) participating in NHANES 2003-2006. We compared simple linear regression (model 1) with multiple linear regression [model 2, controlling for age, sex, race-ethnicity, smoking, supplement use, and the physiologic factors (and urine creatinine for urine biomarkers)] and report significant findings from model 2. Not being fasted was positively associated with most water-soluble vitamins (WSVs) and related metabolites (RMs). Some WSV, fat-soluble vitamin (FSV) and micronutrient (MN), and phytoestrogen concentrations were lower in the presence of inflammation (C-reactive protein ≥ 5 mg/L), whereas fatty acids and most iron indicators were higher. Most WSVs and RMs were higher when renal function was impaired [estimated glomerular filtration rate <60 mL/(min · 1.73 m(2))]. Most WSV, FSV and MN, and fatty acid concentrations were higher in pregnant compared with nonpregnant women, but vitamins A and B-12 and most iron indicators were lower. The estimated changes in biomarker concentrations with different physiologic status were mostly small to moderate (≤ 25%) and generally similar between models; renal function, however, showed several large differences for WSV and RM concentrations. This descriptive analysis of associations between physiologic variables and a large number of nutritional biomarkers showed that controlling for demographic variables, smoking, and supplement use generally did not change the interpretation of bivariate results. The analysis serves as a useful basis for more complex future research.
个体的生理状态可能会影响营养状况的生物标志物。为了帮助研究人员进行研究规划和数据分析，我们评估了常见生理变量（禁食、炎症、肾功能和妊娠）与 29 种血液或尿液营养生物标志物之间的关联，这些生物标志物来自于参与 NHANES 2003-2006 研究的美国成年人（年龄≥20 岁；妊娠变量和铁指标仅限于 20-49 岁的女性）的代表性样本。我们比较了简单线性回归（模型 1）和多元线性回归[模型 2，控制年龄、性别、种族、吸烟、补充剂使用以及生理因素（尿液生物标志物则控制尿肌酐）]，并报告了模型 2 中的显著发现。未禁食与大多数水溶性维生素（WSV）及其相关代谢物（RM）呈正相关。在存在炎症（C-反应蛋白≥5mg/L）的情况下，一些 WSV、脂溶性维生素（FSV）和微量营养素（MN）以及植物雌激素的浓度较低，而脂肪酸和大多数铁指标的浓度较高。当肾功能受损[估计肾小球滤过率<60mL/（min·1.73m2）]时，大多数 WSV 和 RM 的浓度较高。与非妊娠女性相比，妊娠女性的大多数 WSV、FSV 和 MN 以及脂肪酸的浓度较高，但维生素 A 和 B-12 以及大多数铁指标的浓度较低。不同生理状态下生物标志物浓度的估计变化大多较小到中等（≤25%），且在模型之间基本相似；然而，肾功能对 WSV 和 RM 浓度有几个较大的差异。这项关于生理变量与大量营养生物标志物之间关系的描述性分析表明，控制人口统计学变量、吸烟和补充剂使用通常不会改变对二元结果的解释。该分析为更复杂的未来研究提供了有用的基础。
