Confidence intervals of respiratory mechanical properties derived from transfer impedance.
Confidence intervals of respiratory mechanical properties derived from transfer impedance.
复制标题
从传输阻抗得出的呼吸机械特性的置信区间。
DOI：
--
复制
发表时间：
1991
影响因子：
3.3
通讯作者：
C. Gallina
中科院分区：
文献类型：
--
作者：
M. Rotger;R. Peslin;E. Oostveen;C. Gallina
在10名健康受试者中研究了从4至32 Hz测量的呼吸传递阻抗（Ztr）得出的参数的短期个体内变异性。根据Lutchen和杰克逊（J. Appl. Physiol. 62：403-413，1987），将相应的95%置信区间（CI 0）与从单组数据（CIL）计算的置信区间进行比较。用DuBois等人的六系数模型（J.Appl.Physiol.8：587-594，1956）分析Ztr，该模型包括气道阻力（Raw）和惯性（Iaw）、组织阻力（Rti）、惯性（Iti）和顺应性（Cti）以及肺泡气体压缩性（Cg）。Iaw的变异性最低（Cio = 11.1%），紧随其后的是Raw（14.3%）和Cti（14.8%），Rti和Iti的变异性最大（分别为24.6%和93.6%）。使用排除Iti的更简单模型，显著降低了Iaw（P <0.01）和Rti（P <0.05）的变异性，但导致Raw和Iaw系统性降低以及Rti增加。除了两种模型的Raw和更简单模型的Iaw外，CIL大于Cio。无论采用何种模型，Rti和Iaw的两组置信区间之间均存在高度相关性，而Raw则无相关性。这表明前者系数的变异性主要反映了实验噪声，而后者的变异性主要是由于生物变异性。
Short-term intraindividual variability of the parameters derived from respiratory transfer impedance (Ztr) measured from 4 to 32 Hz was studied in 10 healthy subjects. The corresponding 95% confidence intervals (CIo) were compared with those computed from a single set of data (CIL) according to Lutchen and Jackson (J. Appl. Physiol. 62: 403-413, 1987). Ztr was analyzed with the six-coefficient model of DuBois et al. (J. Appl. Physiol. 8: 587-594, 1956), which includes airway resistance (Raw) and inertance (Iaw), tissue resistance (Rti), inertance (Iti), and compliance (Cti), and alveolar gas compressibility (Cg). The lowest variability was seen for Iaw (CIo = 11.1%), closely followed by Raw (14.3%) and Cti (14.8%), and the largest for Rti and Iti (24.6 and 93.6%, respectively). Using a simpler model, where Iti was excluded, significantly decreased the variability of Iaw (P less than 0.01) and Rti (P less than 0.05) but was responsible for a systematic decrease of Raw and Iaw and increase of Rti. Except for Raw with both models and Iaw with the simpler model, CIL was greater than CIo. Whatever the model, a high correlation between both sets of confidence intervals was found for Rti and Iaw, whereas no correlation was seen for Raw. This suggests that the variability of the former coefficients mainly reflects experimental noise, whereas that of the latter is largely due to biological variability.