Eide I, Zahlsen K
Statoil Research Centre, Trondheim, Norway.
Arch Toxicol. 1996;70(7):397-404. doi: 10.1007/s002040050291.
The paper describes experimental and statistical methods for toxicokinetic evaluation of mixtures in inhalation experiments. Synthetic mixtures of three C9 n-paraffinic, naphthenic and aromatic hydrocarbons (n-nonane, trimethylcyclohexane and trimethylbenzene, respectively) were studied in the rat after inhalation for 12h. The hydrocarbons were mixed according to principles for statistical experimental design using mixture design at four vapour levels (75, 150, 300 and 450 ppm) to support an empirical model with linear, interaction and quadratic terms (Taylor polynome). Immediately after exposure, concentrations of hydrocarbons were measured by head space gas chromatography in blood, brain, liver, kidneys and perirenal fat. Multivariate data analysis and modelling were performed with PLS (projections to latent structures). The best models were obtained after removing all interaction terms, suggesting that there were no interactions between the hydrocarbons with respect to absorption and distribution. Uptake of paraffins and particularly aromatics is best described by quadratic models, whereas the uptake of the naphthenic hydrocarbons is nearly linear. All models are good, with high correlation (r2) and prediction properties (Q2), the latter after cross validation. The concentrations of aromates in blood were high compared to the other hydrocarbons. At concentrations below 250 ppm, the naphthene reached higher concentrations in the brain compared to the paraffin and the aromate. Statistical experimental design, multivariate data analysis and modelling have proved useful for the evaluation of synthetic mixtures. The principles may also be used in the design of liquid mixtures, which may be evaporated partially or completely.
本文描述了吸入实验中混合物毒代动力学评估的实验和统计方法。研究了三种C9正链烷烃、环烷烃和芳烃（分别为正壬烷、三甲基环己烷和三甲基苯）的合成混合物在大鼠吸入12小时后的情况。根据统计实验设计原则，使用混合设计在四个蒸汽水平（75、150、300和450 ppm）下混合这些碳氢化合物，以支持一个具有线性、相互作用和二次项的经验模型（泰勒多项式）。暴露后立即通过顶空气相色谱法测量血液、大脑、肝脏、肾脏和肾周脂肪中的碳氢化合物浓度。使用PLS（偏最小二乘法）进行多变量数据分析和建模。去除所有相互作用项后获得了最佳模型，这表明碳氢化合物在吸收和分布方面不存在相互作用。石蜡尤其是芳烃的摄取情况最好用二次模型描述，而环烷烃的摄取几乎是线性的。所有模型都很好，具有高相关性（r2）和预测特性（Q2），后者是交叉验证后的结果。与其他碳氢化合物相比，血液中芳烃的浓度较高。在浓度低于250 ppm时，与石蜡和芳烃相比，环烷烃在大脑中的浓度更高。统计实验设计、多变量数据分析和建模已被证明对合成混合物的评估有用。这些原则也可用于设计可能部分或完全蒸发的液体混合物。
