中文摘要：

利用自制的质子转移反应质谱（PTR-MS）探索了高血糖人群呼气特征挥发性有机物（VOCs）,利用改进的呼气分析系统对70名血糖含量正常志愿者与22名高血糖志愿者的呼气进行实时在线分析,数据统计分析采用Mann-Whitney U非参数检验和线性判别分析法。实验结果表明两组受试者呼气存在5种差异VOCs,该5种VOCs的质荷比（m/z）分别为61、31、120、104、82。其线性判别模型的受试者工作曲线（Re-ceiver operating curve,ROC）下面积达到0.926,灵敏度和特异度分别为86.4%和84.3%,该模型有一定的诊断价值。考察了两组志愿者呼气中丙酮含量的分布情况,统计分析结果显示两组的丙酮含量无显著性差异,丙酮含量与血糖浓度间也无显著相关性。

英文摘要：

In this paper,the detection of the high blood glucose（BG） volunteers characteristics VOCs in breath gas was performed with a homemade PTR-MS.An improved real-time online breath analysis system was used for the analysis of 70 normal volunteers and 22 volunteers who suffered from BG.The date were statistically analyzed by Mann-Whitney U test and linear discriminant stepwise regression.The five product ions of VOCs which showed significant differences between the two groups were identified.The mass to charge ratio（m/z） of the five VOCs were m/z 61,31,120,104 and 82,respectively.The five VOCs diagnostic values were evaluated by Receiver operating characteristic（ROC） curve.The area under ROC（AUC） was 0.926,and the specificity and sensitivity were 84.3% and 86.4%,respectively.This paper also focused on the investigation of the acetone concentration distribution in the two groups of volunteers breath gas,and the statistical analysis results showed that the acetone in two groups had no significant difference,and the acetone concentration and blood sugar concentration had also no significant correlation.