No abstract

Short-term variability (STV) has now been widely accepted as an index of autonomic nervous system integrity in fetuses and neonates. It may be quantified by sampling the electrocardiogram, and computing the Differential Index (DI) after detecting the QRS complexes by computer (1). This paper presents a computationally more efficient method for quantifying STV that correlates excellently with the DI. Further, STV results so obtained are correlated with neonatal outcome and clinical variables. METHOD OF QUANTIFICATIONRather than using the rapidly varying electrocardiogram, this method is based on the relatively slowly varying instantaneous heart rate (IHR) waveform. Information on the inter-beat intervals, from which the DI is computed, is also present in the IHR waveforms, since IHR is the reciprocal of inter-beat intervals. In general, however, exact instants of time when a heart beat (or QRS complex) occurred cannot be determined from the instantaneous heart-rate waveform. This problem is circumvented as described below.We define variability during some fixed time window as rhe corrected maximum deviation of the IHR from the heart rate trend during that period. The trend is determined by linear regression, and the maximum absolute deviation of IHR from this trend is corrected for the step-like characteristics of the IHR waveform. This corrected deviation is associated with the instant of time at the center of the time window, which is empirically chosen to be 4 seconds. In real time, the window and its center keep shifting and a continuous plot of STV is possible STV computed in this manner was correlated with the DI computed from the ECG for several hours of recordings and correlation between them was found to be excellent. The computational time involved, however, was two orders of magnitude less, since ECG sampling and QRS detection were bypassed. OBSERVATIONSA population of 45 patients internally monitored during labor was included in this retrospective study. The baseline STV was categorized into a) a low variability group, in which STV was less than or equal to 2.75 b.p.m., and b) a high variability group, in which STV was more than 2.75 b.p.m. In the low-variability population, 28.7Vo were found to have Hobel's Newborn Risk Scores (2) of >75 and 22.7% had one-minute Apgar scores 5 5 . The corresponding figures for the high-variability group were 4.3% and 4.3%. Of the sub-population with Hobel's Newborn Risk Score >75, 86% were correctly predicted by low STV; however, the over-all false alarm rate was 33.3%.These data provide the basis for further prospective evaluation of Baseline Short-term Variability as a predictor of fetal well-being.ACKNOWLEDGEMENT