Cesarelli M, Romano M, Bifulco P
Department of Electronic and Telecommunications Engineering, University of Naples "Federico II", Italy.
Comput Biol Med. 2009 Feb;39(2):106-18. doi: 10.1016/j.compbiomed.2008.11.010. Epub 2009 Feb 3.
Concise indexes related to variability of foetal heart rate (FHR) are usually utilised for foetal monitoring; they enrich information provided by cardiotocography (CTG). Most attention is paid to the short term variability (STV), which relates to activity and reaction of autonomic nervous control of foetal heart. There is not a unique method to compute short term variability of the FHR but different formulas have been proposed and are employed in clinical and scientific environments: this leads to different evaluations and makes difficult comparative studies. Nine short term variability indexes: Arduini, Dalton, Organ, Sonicaid 8000, Van Geijn, Yeh, Zugaib a modified version of Arduini index and Standard Deviation were considered and compared to test their robustness in CTG applications. A large set of synthetic foetal heart rate series with known features were used to compare indexes performances. Different amounts of variability, mean foetal heart rate, storage rates, baseline variations were considered. The different indexes were in particular tested for their capability to recognise short term heart rate variability variation, their dependence on heart rate signal storage rate (as those provided by commercial cardiotocographic devices), on mean value of the foetal heart rate and on modifications of the floatingline, such in case of accelerations or decelerations. Concise statistical parameters relative to indexes scores were presented in comparative tables. Results indicate that although the indexes are able to recognise STV variation, they show substantial differences in magnitude and some in sensibility. Results depend on the frequency used to acquire and store FHR data (depending on devices); in general, the lower is data rate the more degraded are the results. Furthermore, results differently depend on FHR mean, some for their intrinsic definition; differences arise also in correspondences of accelerations and decelerations. Our results demonstrate that only indexes which refer directly to differences in FHR values, such as Organ and SD indexes, not show dependence on FHR mean. The use of the Standard Deviation index may provide efficient information while showing independence from the considered variables. Indexes performance in case of real cardiotocographic signals were also presented as examples.
与胎儿心率（FHR）变异性相关的简明指标通常用于胎儿监测；它们丰富了产时胎心监护（CTG）所提供的信息。其中，短期变异性（STV）最受关注，它与胎儿心脏自主神经控制的活动和反应有关。计算FHR短期变异性并没有一种独特的方法，而是提出了不同的公式并应用于临床和科研环境中：这导致了不同的评估结果，使得比较研究变得困难。本研究考虑并比较了九个短期变异性指标：阿尔杜尼指数、道尔顿指数、奥根指数、索尼赛德8000指数、范盖恩指数、叶指数、祖盖布指数（阿尔杜尼指数的修正版）和标准差，以测试它们在CTG应用中的稳健性。使用了大量具有已知特征的合成胎儿心率序列来比较各指标的性能。研究考虑了不同程度的变异性、平均胎儿心率、存储速率、基线变化。特别测试了不同指标识别短期心率变异性变化的能力，它们对心率信号存储速率（如商业产时胎心监护设备所提供的）、胎儿心率平均值以及浮动基线变化（如加速或减速情况下）的依赖性。比较表中列出了与指标得分相关的简明统计参数。结果表明，尽管这些指标能够识别STV变化，但它们在大小和某些敏感性方面存在显著差异。结果取决于用于采集和存储FHR数据的频率（取决于设备）；一般来说，数据速率越低，结果越差。此外，结果对FHR平均值的依赖性也不同，有些是由于其内在定义；在加速和减速的对应关系中也会出现差异。我们的结果表明，只有直接参考FHR值差异的指标，如奥根指数和标准差指数，不显示对FHR平均值的依赖性。使用标准差指数可能会提供有效的信息，同时显示出对所考虑变量的独立性。还给出了实际产时胎心监护信号情况下各指标的性能示例。
