Phase-rectified signal averaging for intrapartum electronic fetal heart rate monitoring is related to acidaemia at birth.

OBJECTIVE: Recent studies suggest that phase-rectified signal averaging (PRSA), measured in antepartum fetal heart rate (FHR) traces, may sensitively indicate fetal status; however, its value has not been assessed during labour. We determined whether PRSA relates to acidaemia in labour, and compare its performance to short-term variation (STV), a related computerised FHR feature. DESIGN: Historical cohort. SETTING: Large UK teaching hospital. POPULATION: All 7568 Oxford deliveries that met the study criteria from April 1993 to February 2008. METHODS: We analysed the last 30 minutes of the FHR and associated outcomes of infants. We used computerised analysis to calculate PRSA decelerative capacity (DC(PRSA)), and its ability to predict umbilical arterial blood pH ≤ 7.05 using receiver operator characteristic (ROC) curves and event rate estimates (EveREst). We compared DC(PRSA) with STV calculated on the same traces. MAIN OUTCOME MEASURE: Umbilical arterial blood pH ≤ 7.05. RESULTS: We found that PRSA could be measured in all cases. DC(PRSA) predicted acidaemia significantly better than STV: the area under the ROC curve was 0.665 (95% CI 0.632-0.699) for DC(PRSA), and 0.606 (0.573-0.639) for STV (P = 0.007). EveREst plots showed that in the worst fifth centile of cases, the incidence of low pH was 17.75% for DC(PRSA) but 11.00% for STV (P < 0.001). DC(PRSA) was not highly correlated with STV. CONCLUSIONS: DC(PRSA) of the FHR can be measured in labour, and appears to predict acidaemia more accurately than STV. Further prospective evaluation is warranted to assess whether this could be clinically useful. The weak correlation between DC(PRSA) and STV suggests that they could be combined in multivariate FHR analyses.

Antenatal cardiotocogram quality and interpretation using computers.

OBJECTIVE: To test the application in practice of computerized fetal heart rate (FHR) analysis in pregnancy. DESIGN: Randomized distribution of subjects with computerized analysis automatically revealed or concealed. SETTING: A district general hospital and a teaching hospital outside London. SUBJECTS: 2869 pregnant women studied within a year. OUTCOME MEASURES: Quality and duration of the cardiotocogram; quantitative measurement of FHR variation; number of stillbirths. RESULTS: With interactive advice to the operator, records were of improved quality (up to 28% without signal loss) with potentially much reduced recording time. The short-term FHR variation measured in the last records before intervention is reported for the first time. CONCLUSION: The benefits of using the computers include improvement in record quality and saving of time. In addition, where interpretation depended on estimation of FHR variation there was prima facie evidence of observer misinterpretation; visual analysis was unreliable. A larger trial is now required with more rigorous constraints on intervention.

Computerized analysis of normal fetal heart rate pattern throughout gestation.

OBJECTIVE: To analyze the evolution of computerized cardiotocography (cCTG) parameters throughout gestation in a large archive of traces from healthy fetuses. METHODS: This was a cross-sectional study of the first cCTG record from 4412 singleton fetuses with good pregnancy outcome. Normal ranges of cCTG parameters for 25 to 42 weeks were derived from analysis of only one cCTG record per fetus, and the relationship between the parameters and gestational age was investigated. RESULTS: Fetal heart rate (FHR) accelerations, short- and long-term variation overall, duration of episodes of high and low variation and variation in high episodes increased with advancing gestation. In contrast, maternal perception of fetal movements, basal FHR, variation in low episodes and the time until criteria for normality were met decreased with advancing gestation. Gestational age-related changes in FHR variation were less evident at the lowest percentiles. Episodes of high FHR variation were detected in most fetuses, even at 25 weeks. Opposite trends of basal FHR and variation were observed at 42 weeks. Large decelerations and the frequency and duration of low episodes were also higher at 42 weeks. CONCLUSIONS: The characteristics of the normal FHR pattern are quite defined from early on in gestation, follow a continuous trend with advancing gestation and change abruptly at 42 weeks. Gestational age-related changes are less obvious at the lowest percentiles.

The value of the short-term fetal heart rate variation for timing the delivery of growth-retarded fetuses.

OBJECTIVE: To assess the clinical value of the short-term fetal heart rate variation (STV) for timing the delivery of severely growth-retarded fetuses, many associated with pre-eclampsia. DESIGN: Retrospective cohort study. SETTING: John Radcliffe Maternity Hospital, Oxford, UK. POPULATION: Two hundred and fifty-seven fetuses with a birthweight less than third percentile and a last computerised cardiotocography performed within 24 h of delivery. METHODS: Analysis of the relationship between antepartum STV and the perinatal outcome. MAIN OUTCOME MEASURES: Stillbirth rate and the acid-base status at birth. RESULTS: There were no stillbirths or neonatal deaths (NNDs) within 24 h in the study population. Decreasing STV was correlated with earlier deliveries (P < 0.001), lower birthweight (P < 0.001), lower umbilical artery pH at birth (P < 0.001), worse acid-base status at birth (P < 0.001) and worse postnatal outcome (P < 0.002). The STV was able to predict the presence or absence of acidaemia and metabolic acidaemia (area under the receiver operating characteristic curve 0.70 and 0.75, respectively, P < 0.001). The risk of metabolic acidaemia increased as the antepartum STV decreased, the optimum cutoff level being < or = 3.0 milliseconds (positive and negative predictive values 64.6 and 86.6%). An STV < or = 3.0 milliseconds was associated with markedly higher rate of metabolic acidaemia and early NNDs compared with an STV > 3.0 milliseconds (54.2 versus 10.5% and 8.3 versus 0.5%, respectively; P < 0.001). The deaths of the former group were all due to extreme prematurity and very low birthweight. CONCLUSIONS: The antepartum STV is an important marker of perinatal outcome in severely growth-retarded fetuses. Timing the delivery of the most preterm and small fetuses remains a difficult task.

Short-term fetal heart rate variation, decelerations, and umbilical flow velocity waveforms before labor.

OBJECTIVE: To determine the value of computerized antepartum measurements of short-term fetal heart rate (FHR) variation and decelerations as a predictor of outcome, and to compare these with measurements of the umbilical artery flow velocity waveform. METHODS: Data were collected from 3563 high-risk patients for measurements of FHR variables (15,702 records) and outcome (intrauterine death or metabolic acidemia on delivery). Detailed analyses were made on 89 patients with one or more records in which short-term FHR variation was 3 milliseconds or less. Umbilical artery flow velocity waveforms were measured concurrently on 2441 occasions in 991 patients. RESULTS: More than half of the patients with FHR variation of 3 milliseconds or less were identified before 31 weeks. When short-term FHR variation exceeded 3 milliseconds, there were no intrauterine deaths and only one instance of metabolic acidemia on cesarean delivery. When FHR variation fell below 2.6 milliseconds, 34% of the subjects had metabolic acidemia on cesarean or intrauterine death. The appearance or absence of decelerations was an unreliable guide to outcome. The correlation between FHR variation and the umbilical artery flow velocity waveform was low (r = 0.32). The population studied contained some patients without placental vascular disease or, in five with pre-eclampsia, without abnormal umbilical artery velocity waveforms but with grossly reduced FHR variation. CONCLUSION: Reduction in short-term FHR variation, as measured by computer, reliably predicts fetal outcome and is a more general measure of fetal well-being than are umbilical artery flow velocities.

Approximate entropy, a statistic of regularity, applied to fetal heart rate data before and during labor.

OBJECTIVE: To determine whether approximate entropy (ApEn), a new statistic of regularity, when applied to fetal heart rate (FHR) data antepartum or in labor, would offer an advantage over standard statistics of variation in predicting outcome. METHODS: A large data base of antepartum FHR records collected in clinical practice over 10 years was available. Two data sets in labor were stored on disk in small computers interfaced to fetal monitors on-line. Outcomes were assessed using blood gas values on delivery and Apgar scores. RESULTS: Antepartum, when the most favorable form of ApEn was used on 769 good-quality FHR records, the correlation with measurement of short-term variation was high. This was especially true when the fetal pulse interval variation fell below the normal range (less than 6 milliseconds short-term; r = 0.93) and in 20 other records with sinusoidal variation (r = 0.96). Approximate entropy varied with fetal sleep cycles and took longer to calculate than FHR variation. During the last hour of labor, in 319 records, there was no significant correlation between umbilical artery base deficit values on delivery and ApEn measurement. In 871 additional good-quality records of fetuses with normal outcome, the mean (+/- standard error [SE]) ApEn (0.95 +/- 0.005) was significantly greater than in 22 records (0.88 +/- 0.028) from fetuses with abnormal outcome (umbilical artery base deficit more than 12 mmol/L and Apgar score of 3 or less at 1 minute). However, consideration of the frequency distributions of these measurements showed that ApEn did not discriminate between normal and abnormal outcomes. The SD of fetal pulse intervals rose in labor whereas ApEn values fell, confirming that this new statistic of regularity differs from standard statistics of variation. CONCLUSION: Approximate entropy offered no advantage over measurement of short-term FHR variation antepartum, and neither measurement predicted outcome in labor.

Computerized fetal heart rate analysis in labor.

Observer variation in visual analysis of fetal heart rate (FHR) records is reportedly high, but can be avoided by computerized numerical analysis. The FHRs of 394 women in labor at 37 or more weeks' gestation were recorded on-line and analyzed to examine how different patterns related to outcome, as judged by umbilical arterial base deficit or Apgar score on delivery. The range of normality and the diversity of patterns of those delivered without acidemia were great. Late decelerations were of poor prognostic value. There was an increase in FHR variation during labor averaging 40%. In this preliminary study, conventional attributes of the FHR, alone or in combination, did not predict metabolic acidemia. Epidural analgesia in 240 women was identified as a confounding variable that significantly affected FHR patterns without influencing the condition of the infant at birth. It was associated with a higher FHR, less FHR variation and fewer decelerations, primiparity, longer labors, more operative deliveries, and a threefold greater cesarean rate. The rise in basal FHR, perhaps due to a rise in maternal temperature, may partly explain the high intervention rate in those without fetal acidemia.

Tracking poles with an autoregressive model: a confidence index for the analysis of the intrapartum cardiotocogram.

Clinicians often rely upon the cardiotocogram, a display of the fetal heart rate and maternal uterine activity (UA) over time, as a means of monitoring fetal health during labour. Fetal health can be monitored adequately only when the signal quality of the cardiotocogram is good. We propose an automated assessment of UA signal quality in order to create a confidence index for subsequent analysis of the intrapartum cardiotocogram. We use an autoregressive (AR) model of the UA to estimate the power at the contraction frequency, with high power indicative of "good" UA signal quality. 5th, 10th, and 15th-order AR models are used to assess the signal quality of 12 intrapartum UA traces as "good/medium" or "poor". We compare our results to two experts' visual assessments of signal quality. The 10th-order model exhibits the highest percent agreement rate of 62%. It also exhibits the most balanced false positive and false negative rates, where "good" or "medium" signal quality is considered a positive and "poor" signal quality a negative. The 10th-order model can therefore be used as a confidence index to reduce the errors made in the identification of uterine contractions in the UA trace and in the subsequent analysis of the cardiotocogram as a whole.

Relation of fetal heart rate signals with unassignable baseline to poor neonatal state at birth.

Electronic fetal heart rates (FHR) are used to monitor fetal health during labour. The paper records are visually assessed by clinicians, but automated alternatives are being developed. Interpretation, visual or computerised, depends on assigning a baseline to identify key features such as accelerations and decelerations. However, when the FHR is unstable the baseline may be unassignable, making conventional analysis unreliable. Such instability may reflect on fetal health. If true, these segments should not be discarded but quantified, for which we have developed a numerical method. In 7,568 labours, the association between unassignable baseline and umbilical arterial blood pH ≤ 7.05 at birth (evidence of poor health) was studied retrospectively. We found a consistent increase of the risk for acidaemia with longer intervals of unassignable baseline. This is detectable at the end of the first stage of labour, but stronger at the end of the second stage: in the last 30 min of labour, the odds ratios (with respect to baseline assignable throughout this period) increased from 1.99 (15 min unassignable) to 4.9 (30 min unassignable). Computerised analysis of the FHR becomes unreliable when the baseline cannot be assigned; however, this pattern is itself a pathological feature associated with acidaemia at birth.

Criteria for the design of fetal heart rate analysis systems.

Criteria are described for the automated analysis of fetal pulse intervals, fetal movements and of uterine contractions measured externally, antenatally and interactively on-line, for implementation on a personal computer interfaced to an appropriate fetal monitor, and tested on 10,000 records. Measurements of short and longer term fetal heart rate variation are compared; both are required to identify sinister records. Recall and display of records acquired on the same patient over several weeks has proved a useful diagnostic aid.

Limitations of antenatal fetal heart rate monitors.

Erroneous or doubtful decelerations in fetal heart traces were present in 111 of 1000 consecutive antenatal clinical records obtained by monitors with autocorrelation. The incidence was 20% in fetuses less than 30 weeks of gestational age. Their elimination reduced the number of "decelerative" records by 42%. Erroneous or doubtful accelerations were also present in 11% of records. These errors are caused by the fetal heart rate monitor and may contribute to the high intraobserver and interobserver variation on visual analysis. They can be detected by computer analysis.

The advantages of computerized fetal heart rate analysis.

A brief account is given of the advantages of computerized analysis of human fetal heart rate (FHR) traces antenatally, based on clinical use over 8 years. Accuracy is greater. Results are presented quantitatively and consistently. The numerical measures of the FHR pattern are related to other objective measures of fetal health, e.g. initial compensated hypoxaemia or terminal acidaemia. Computerized analysis has shown that changes in FHR variation are a better guide than the presence or absence of large decelerations. Recording time is used better. Synoptic displays of data over 4 weeks show significant trends in fetal heart rate variation and movements with time. And the problems generated by the limitations of fetal heart rate monitors are identified to exclude spurious information.

System 8000: computerized antenatal FHR analysis.

SYSTEM 8000 is a computerized system for antenatal fetal heart rate (FHR) analysis, with interaction online to ensure good quality recording and to minimize the time required to obtain the necessary information (based on fetal movements and tocodynamometer readings as well as FHR). The equipment consists of a Personal Computer with hard disk, interfaced to a fetal monitor. Software is written in C. An extensive definition is given on most of the functions and parameters as calculated by the system, e.g. record quality, uterine contraction peaks, basal heart rate, variation, decelerations and accelerations. System 8000 is designed to take account of the episodic changes in FHR and fetal movements characteristic of sleep states. Their presence naturally affects the mean FHR variation calculated over, say 20-30 mins. But, as the amplitude of these episodes is attenuated in association with growth retardation and hypoxaemia, the measurements of variation decline. In practice inter- and intra-observer variation is greatest in assessing FHR variation. Yet a decrease in variation is the most reliable index of fetal deterioration. The system measures FHR variation accurately and reproducibly, as well as fetal movements. In the synoptic display these two equally important fetal variables are given their rightful prominence. It has been interesting to observe how rarely the basal FHR changes in fetuses suffering progressive respiratory insufficiency, and how extreme tachycardia (a basal rate greater than 170/min) tends to reflect other fetal problems such as infection or maternal pyrexia.

Improvements in computerized fetal heart rate analysis antepartum.

The continued development of a computerised system for measuring the pattern of the antepartum fetal heart rate (FHR) is described. Previous work had established that measurement of FHR variation objectively detects chronic fetal hypoxaemia and the onset of metabolic acidaemia antepartum. The normal centiles were calculated for the amplitude of long-term FHR variation, in episodes of high and low variation, week by week from 24-42 weeks gestation. Reference to these (automatically by the computer) improved discrimination between normal and questionable records in 38% of records, with a small saving of time. Two types of sinusoidal rhythm were described (slow, 1 in 2-5 minutes, incidence 0.16% of subjects; and faster, 2-5 per minute, incidence 0.025%) with methods for their detection. Both may be of sufficient amplitude to induce an episode of high FHR variation. The different effects of maternal steroid (betamethasone or dexamethasone) administration of FHR variation were compared, and the clinical consequences considered. The frequency distribution of basal FHR in normal and abnormal records was measured, and the effects on basal FHR outside the normal range (120-160 bpm) on FHR variation described. Adjustment of the FHR baseline was undertaken when, exceptionally, large abrupt changes in heart rate occurred during a record. The duration and frequency of FHR record acquisition in clinical practice were reviewed, and new policies recommended. With adequate safeguards, measurement by a computer offers reliable objective information from which fetal health may be assessed, more objectively and accurately than by visual inspection.

Dexamethasone and fetal heart rate variation.

OBJECTIVE: To determine the effect of maternal administration of dexamethasone on fetal heart rate and its variation. DESIGN: Retrospective analysis of computerised data derived from cases studied over three years. SETTING: High risk pregnancy unit, John Radcliffe Hospital, Oxford. SUBJECTS: Twenty-eight pregnant women, at 27 to 32 weeks of gestation, to whom dexamethasone was given to accelerate pulmonary maturation in the expectation of preterm delivery. METHODS: Dexamethasone (two doses of 12 mg intramuscularly, 12 h apart) was given on 51 occasions at weekly intervals (one to four occasions per patient). Complete data were available for cardiotocograph analysis from computerised measurement of fetal heart rate variables for two days before and four days after dexamethasone and, in 19 women, measurements of umbilical arterial flow velocity waveforms before and after dexamethasone. RESULTS: In 10 pregnancies without fetal distress there was a highly significant (P < 0.01) transient rise in short term fetal heart rate variation after dexamethasone administration, from means (SE) 6.4 (0.28) to 9.8 (0.4) ms. In 18 pregnancies with subsequent delivery for fetal distress (abnormal fetal heart rate pattern) and high umbilical arterial resistance index [mean 0.93 (0.06 SE)], the rise in short term fetal heart rate variation was less (P < 0.01), from mean (SE) 5.4 (0.26) to 6.1 (0.48) ms. In a further case of discordant twin pregnancy, the larger twin continued to respond to dexamethasone administrations with a rise in fetal heart rate variation for five weeks; the smaller twin, with maintained tachycardia and reduced umbilical arterial end-diastolic flow velocity, failed to respond after the first two weeks. CONCLUSION: The results show that maternal dexamethasone administration normally causes a rise in fetal heart rate variation for up to a day. This rise is reduced in pre-eclampsia or intrauterine growth retardation, associated with a reduction in umbilical flow, perhaps because of a consequential lower concentration of steroid in the fetus. The results contrast with those for betamethasone which has been reported to reduce fetal heart rate variation.

Short-term variation in abnormal antenatal fetal heart rate records.

In a retrospective study the relation of reduced fetal heart rate variation to fetal acidemia was analyzed with a computerized system for numeric analysis. Between 1983 and 1987, 78 pregnancies were identified in which at least one record of the fetal heart rate had very low long-term variation. The outcome was analyzed to determine the numeric criteria of fetal heart rate variation that most efficiently detect a fetus that will die (preterminal) or is dying (terminal). Because fetal compromise was found on occasion to be associated with a slow sinusoidal fetal heart rate rhythm that increased measures of long-term variation. It was necessary to define a new index of short-term fetal heart rate variation (the 1/16 minute epoch-epoch variation). This was closely related to long-term variation (r = 0.9) but provided better detection of preterminal records as judged by metabolic acidemia at delivery or intrauterine death.

Computerized analysis of episodic changes in fetal heart rate variation in early labor.

Fetal heart rate variation in early labor was measured by computerized analysis in cyclic episodes of low or high variation in 136 women at 37 to 42 weeks' gestation. The amplitude (mean +/- SE) in episodes of low variation was 20.6 +/- 0.4 milliseconds; in high variation it was 57.3 +/- 1.1 milliseconds. The duration (mean +/- SE) of low episodes (24.3 +/- 1.3 minutes) was less than that of high episodes (45.1 +/- 2.7 minutes) but was sometimes greater than 1 hour. In episodes of low variation the amplitude was less than 5 beats/min long term in 11% and less than 2.5 milliseconds (pulse intervals) short term in 8%; these measures would be predictive of intrauterine death if persistent before birth. We conclude that the use of decreased fetal heart rate variation as a diagnostic sign of acute fetal hypoxemia in labor is incorrect, because changes of this size occur randomly as a consequence of fetal sleep states. There was no significant correlation between fetal heart rate variation over the last hour of labor and umbilical arterial base deficit on delivery.

Antenatal cardiotocogram quality and interpretation using computers.

OBJECTIVE: To test the application in practice of computerized fetal heart rate (FHR) analysis in pregnancy. DESIGN: Randomized distribution of subjects with computerized analysis automatically revealed or concealed. SETTING: A district general hospital and a teaching hospital outside London. SUBJECTS: 2869 pregnant women studied within a year. OUTCOME MEASURES: Quality and duration of the cardiotocogram; quantitative measurement of FHR variation; number of stillbirths. RESULTS: With interactive advice to the operator, records were of improved quality (up to 28% without signal loss) with potentially much reduced recording time. The short-term FHR variation measured in the last records before intervention is reported for the first time. CONCLUSION: The benefits of using the computers include improvement in record quality and saving of time. In addition, where interpretation depended on estimation of FHR variation there was prima facie evidence of observer misinterpretation; visual analysis was unreliable. A larger trial is now required with more rigorous constraints on intervention.

A randomised controlled comparison of betamethasone with dexamethasone: effects on the antenatal fetal heart rate.

OBJECTIVE: To compare the effects of maternal administration of betamethasone and dexamethasone on fetal heart rate, using computerised numerical analyses, and to examine the association between changes in short term variation and the timing and indication for delivery. SETTING: John Radcliffe Hospital, Oxford. SAMPLE: Fifty-nine women with singleton pregnancies, who were at risk of delivery before 34 weeks of gestation, had received no steroids in the preceding week and could give informed consent. METHODS: Women were randomised on a double-blind basis to receive either betamethasone or dexamethasone. The fetal heart rate was recorded (60-minute duration at similar times of day) before steroid administration and on each of the following two days; changes were measured by computerised analyses. MAIN OUTCOME MEASURES: Changes in short term variation and long term variation of fetal heart rate or the number of fetal movements. Statistical analysis was nonparametric. RESULTS: Betamethasone and dexamethasone had no differential effects on the computerised cardiotocography. However, both steroids decreased baseline fetal heart rate, increased long term variation, increased short term variation, and decreased fetal movements on the first day after steroid administration, and decreased high fetal heart rate variation and decelerations on the second day. Among 13 women who were delivered within one week of steroid administration, smaller rises in short term variation on day one were associated with delivery for fetal indications. CONCLUSIONS: Both betamethasone and dexamethasone produced transient, unexplained changes in the fetal heart rate over the two days following steroid administration. Larger changes in short term variation were associated with fetal wellbeing. It is important to recognise that such changes are a physiological response of the human fetus to steroid administration.

Fetal heart rate patterns in term labor vary with sex, gestational age, epidural analgesia, and fetal weight.

OBJECTIVES: Our purpose was to analyze the previously unreported effect of fetal sex on the fetal heart rate in labor and to measure its magnitude in relation to the effects of other independent clinical variables. STUDY DESIGN: The last hour of the intrapartum heart rates of 1884 term singleton fetuses collected during routine clinical monitoring over 19 months in Oxford, United Kingdom, was analyzed by computerized techniques. The records were selected for completeness and continuity until within at least 30 minutes of delivery. A subset of records from earlier in labor and a separate archive of antepartum normal term records were also examined. RESULTS: Female fetuses had significantly faster heart rates than male fetuses (P <.0001). Epidural analgesia, weight percentile (adjusted for age and sex), parity, the duration of first and second stages of labor, and a fall in umbilical arterial blood pH at birth also independently modulated the fetal heart rate (all P <.0001). The effects of these independent variables on heart rate were additive, the most important being epidural analgesia as a cause of tachycardia. The effect of fetal sex was less in the first stage, 6 to 7 hours before delivery, and was not present before the onset of labor (in another 552 pregnancies at 37 to 38 weeks). CONCLUSIONS: The fetal heart rate response of female fetuses to normal labor differs from that of male fetuses. Computerized numeric analysis of intrapartum fetal heart rate patterns will need to take into account the multiple factors that influence the fetal heart rate to identify precisely which patterns predict clinical outcome.