Valensise Herbert, Facchinetti Fabio, Vasapollo Barbara, Giannini Flavio, Monte Ilaria Di, Arduini Domenico
Department of Obstetrics and Gynecology, University of Rome Tor Vergata, Rome, Italy.
Eur J Obstet Gynecol Reprod Biol. 2006 Apr 1;125(2):185-92. doi: 10.1016/j.ejogrb.2005.06.034. Epub 2006 Feb 3.
To ascertain the diagnostic ability of a computerized fetal heart rate (FHR) analysis system in the identification of patients at risk of fetal distress in labour.
Three hundred and two healthy post-term pregnancies were enrolled in a retrospective, cross-sectional study and subdivided into two groups, with (n=42) or without (n=260) fetal distress in labour. The last computerized FHR recording before onset of labour was analyzed.
The two groups showed a significant difference only in FHR baseline and in percentage of small accelerations on total. The multivariate analysis showed that only the percentage of small accelerations was significantly related to the labour outcome. A higher diagnostic accuracy was obtained with use of neural network analysis, which allowed a sensitivity of 56%, specificity 91%, positive predictive value 53% and negative predictive value 92% with an overall accuracy of 86%.
The increase in FHR baseline and in small FHR accelerations can be major factors in the prediction of subsequent fetal distress in healthy term fetuses. Use of neural networks seems to further improve the ability of computerized FHR analysis in the prediction of intrapartum distress.
确定计算机化胎儿心率（FHR）分析系统在识别分娩时胎儿窘迫风险患者中的诊断能力。
302例健康过期妊娠纳入一项回顾性横断面研究，并分为两组，一组有分娩时胎儿窘迫（n = 42），另一组无分娩时胎儿窘迫（n = 260）。分析分娩开始前最后一次计算机化FHR记录。
两组仅在FHR基线和小加速占总数的百分比方面存在显著差异。多变量分析显示，只有小加速的百分比与分娩结局显著相关。使用神经网络分析获得了更高的诊断准确性，其灵敏度为56%，特异度为91%，阳性预测值为53%，阴性预测值为92%，总体准确率为86%。
FHR基线增加和FHR小加速增加可能是预测健康足月胎儿随后发生胎儿窘迫的主要因素。使用神经网络似乎进一步提高了计算机化FHR分析预测产时窘迫的能力。
