At its introduction, it was hoped that continuous observation and analysis of fetal heart rate patterns would guide labor management so that the hypoxic fetus could be identified and delivered promptly, thereby reducing intrapartum stillbirth, neonatal neurologic injury, and subsequent neurodevelopmental delay. Unfortunately, this promise has never been realized,1 perhaps in part because of the lack of an ideal framework for analysis and interpretation of fetal heart rate patterns.
Recently, the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) in collaboration with the American College of Obstetricians and Gynecologists and the Society for Maternal-Fetal Medicine convened a consensus workshop that proposed an updated classification system and definitions for fetal heart rate tracings (Box 1).2 In this classification scheme, category I fetal heart rate patterns are normal and thought to predict a normal fetal acid-base status at the time of observation; category II fetal heart rate patterns are indeterminate; and category III patterns are abnormal and are considered to be most predictive of abnormal fetal acid-base status.2
One of the goals of the NICHD workshop was to suggest research priorities for electronic fetal heart rate monitoring, two of which were a estimation of the relative frequency of the fetal heart rate categories, and validation of the assumptions regarding fetal status with the new fetal heart rate categories. To date, few descriptive data are available to address these questions.
The objectives of this study were to estimate the amount of time spent in each of the NICHD fetal heart rate categories during labor and during the last 2 hours of labor in term singleton pregnancy and to estimate the relationship between the duration of time spent in each category and short-term neonatal outcomes.
MATERIALS AND METHODS
We performed a review of the intrapartum fetal heart rate characteristics of all patients with a singleton, nonanomalous fetus in term labor at 10 Intermountain Healthcare hospitals between March 1, 2007, and June 30, 2009. Patient inclusion criteria included: singleton pregnancy; gestational age 37.0 weeks or greater; in labor (either spontaneous or induced) with fetal heart rate monitoring for at least 120 minutes; and delivery during the monitoring period. Patients whose fetus or newborn had a structural or genetic abnormality were excluded. Patients admitted for scheduled cesarean delivery were excluded, even if they were monitored for more than 2 hours before surgery.
All study hospitals used electronic fetal monitoring in labor, and all labor and delivery nurses were specifically trained using standardized protocols in the reading of fetal heart rate monitor tracings. Before beginning work on an Intermountain Hospital labor and delivery suite, nurses complete a 4-hour classroom session in basic fetal heart rate monitoring. Within the first year, they must then successfully complete a 2-day course in fetal heart rate monitoring that was developed and is sponsored by the Association of Women's Health, Obstetric and Neonatal Nurses; this course includes a day of lecture materials and a day of skills demonstration and testing. Thereafter, each nurse must document continuing education in fetal heart rate interpretation. This ongoing training consists of attending at least two fetal heart rate Strip Review classroom presentations and completing at least 10 online self-study sessions using Intermountain's “Strip of the Month” each year.
All intrapartum fetal heart rate monitor tracings were read by a labor and delivery nurse at least every 20 minutes and the fetal heart rate characteristics (Box 2) entered into a bedside computer workstation. Whenever a change in fetal heart rate patterns developed (eg, onset of variable decelerations or change in fetal heart rate baseline), the nurse entered the new characteristics into the workstation, even if fewer than 20 minutes had passed since the last entry. If there were no further changes, the next assessment and data entry was 20 minutes later. If there were additional fetal heart rate changes observed sooner, the new characteristics were entered at the time of change. Thus, the duration of time of each characteristic and, therefore, each fetal heart rate category was tracked in real time rather than in blocks of time. The fetal heart rate tracing and the entered data were then stored in a systemwide enterprise data warehouse and linked to maternal and short-term neonatal data.
Fetal heart rate characteristics on all patients were retrieved from the data warehouse, and software was developed to convert the data on fetal heart rate characteristics into the appropriate NICHD fetal heart rate category at each reading. Because nursing entries were timed, the duration of time in fetal heart rate category could then be calculated.
Descriptive statistics were performed on fetal heart rate category data for all patients from the entire monitoring period and from the last 2 hours of monitoring before delivery. Proportions were compared using chi-square testing with Yates' correction. P<.01 was considered statistically significant. The investigation was approved by the Intermountain Healthcare institutional review board.
A total of 48,444 women met criteria for inclusion. The number of study patients from the 10 hospitals ranged from 1,982 to 7,239 during the 28-month sampling period. The majority of the patients were white, married, and parous. The mean maternal age was 27.0 years, the mean time monitored was 513.1 minutes, the median time monitored was 432.4 minutes, and the range of monitoring time was 120.0–5,578.3 minutes. The demographics of the entire group are detailed in Table 1.
Labor was spontaneous in 25,590 (52.8%) patients, induced for a medical or obstetric indication in 10,704 (22.1%), and electively induced in 12,150 (25.1%). Oxytocin was used, either for induction or augmentation of labor, in 24,825 (51.2%) patients.
Overall, category I fetal heart rate patterns were observed at some point in 48,191 (99.5%) of tracings, category II patterns were found in 40,758 (84.1%) of tracings, and category III patterns were seen in 54 (0.1%) of tracings. A large majority, 83.6%, of patients had a mix of categories during their labor. Only 7,686 (15.9%) patients had monitor strips with category I fetal heart rate patterns exclusively, 253 (0.5%) had only category II patterns, and no patients (0.0%) had only category III fetal heart rate.
Examining the entire duration of monitoring in all patients, the fetal heart rate pattern was classified as category I 77.9% of the time, as category II 22.1% of the time, and as category III 0.004% of the time. Segmenting the study population by parity, nulliparas spent a slightly smaller proportion of time in category I and slightly greater time in category II during their labors compared with multiparas (Table 2).
In the 2 hours before delivery, category I tracings became less common, and category II and category III tracings became more common. Among all patients in the last 2 hours of monitoring, category I tracings were present 60.9% of the time, category II tracings were present 39.1% of the time, and category III tracings were seen 0.006% of the time (Table 2).
The group whose entire labor was spent in fetal heart rate category I had good outcomes. Only 290 (3.8%) had 1-minute Apgar scores less than 7, and 46 (0.6%) had 5-minute Apgar scores less than 7. There were 249 (3.2%) neonatal intensive care unit (NICU) admissions in this group. Only 14 (0.2%) neonates had low 5-minute Apgar scores and NICU admission. Outcomes were similar in the group whose last 2 hours of labor was solely in fetal heart rate category I with an increased incidence of low 1-minute Apgar score but no difference in the other neonatal outcomes (Table 3).
To assess whether the amount of time spent in category II was related to short-term neonatal outcome, patients were divided into subgroups based on the amount of time that was spent in category II during the last 2 hours before delivery. Patients who spent the entire last 2 hours of labor in category I were compared with those who spent increasing time in category II, divided into quartiles of 1–25%, 26–50%, 51–75%, and 76–100% of the last 2 hours in fetal heart rate category II.
There was no difference between these subgroups with respect to maternal age, race, marital status, or tobacco use. However, with increasing time in category II above 25% of the last 2 hours of labor, nulliparity and medical–obstetric complications were more common (both P<.01; Table 4).
Increasing time in category II in the last 2 hours of labor also was associated with increased likelihood of NICU admission or Apgar score less than 7 but not until more than 50% of the time was spent in fetal heart rate category II. In the 51–75% category II quartile, only the likelihood of a low 1-minute Apgar score and the likelihood of NICU admission were increased (both P<.01). However, in the 76–100% category II quartile, there was an increased likelihood of 1-minute Apgar score less than 7, 5-minute Apgar score less than 7, NICU admission, and the combination of NICU admission and 5-minute Apgar score less than 7 (all P<.001; Table 4).
There is a broad consensus as to the definition of a normal intrapartum fetal heart rate pattern: normal baseline, moderate variability, and absence of variable or late decelerations.3,4 Category I fetal heart rate patterns are assumed to be reflective of normal fetal oxygenation status, and bad outcomes are uncommon. Our data confirm this notion. Among patients whose fetal heart rate patterns in the last 2 hours before delivery were exclusively category I, only 0.2% (one in every 531) had low 5-minute Apgar scores and admission to the NICU.
Similarly, there is accord on abnormal fetal heart rate characteristics: absent variability, recurrent late or variable decelerations, prolonged bradycardia, and sinusoidal patterns.3,4 Fortunately, category III fetal heart rate patterns are a very rare event. In our population of patients in term labor, category III was identified in only 0.1% of patients (one in every 897).
Between the extremes, category II fetal heart rate patterns are very common, occurring in 84% of labors. The proportion of time spent in category II increases in the 2 hours before delivery. There is an increased likelihood of adverse short-term outcomes with increasing time in category II fetal heart rate patterns.
Despite the association between increasing time in category II and low Apgar scores and NICU admission, the huge majority of newborns who had predominantly category II tracings experienced no short-term morbidity. Thus, describing category II as indeterminate is appropriate. However, its high frequency and broad range of outcomes diminish the usefulness of category II fetal heart rate patterns as an indicator of fetal condition. This is reflected in the recommendations for management of category II tracings, which stress “evaluation, continued surveillance, initiation of appropriate corrective measures when indicated, and reevaluation.”5
Others have addressed the diversity found in the indeterminate category of fetal heart rate tracings. Larma and colleagues examined fetal heart rate characteristics in the last hour before delivery of 107 neonates with documented umbilical cord acidosis and compared them with those of a control group with normal blood gas results. They found differences in baseline rate, bradycardia, variability, and reactivity between acidotic newborns with and without hypoxic–ischemic encephalopathy. However, the fetal heart rate characteristics discriminated poorly between acidotic and nonacidotic newborns, because the frequency of accelerations, decelerations, variable decelerations, and severe variable decelerations were not different between the two groups.6
Parer and Ikeda have recognized that broad fetal heart rate categories will necessarily include a wider range of outcome severities. Basing their classification system on published outcome risks related to specific fetal heart rate characteristics, they proposed a color-coded, five-category scheme, essentially dividing the indeterminate category into three additional risk levels.7 Although their system is rather complex, a study of tracing reviews found good agreement between blinded, expert readers.8 Also noting that such a system lends itself well to computer-based interpretation, they showed that a software-based fetal heart rate assessment system correlated well with the human experts.8
The large number of patients, managed in both tertiary and community hospitals, is a strength of our study. The ongoing training in fetal heart rate monitoring provided to all labor and delivery clinical staff is an additional strength. Also, our data set links to specific patient outcomes rather than to literature-based risks. Our study has several disadvantages. One is the retrospective nature of the data. There is a possibility of distortion of our results through censoring by intervention, because there may have been expedited delivery in some cases with more prolonged nonreassuring category II patterns. This would likely decrease morbidity in the category II group but reflects a real-world assessment of the associations between fetal heart rate and outcome. More significantly, there is a lack of umbilical cord blood gas data; cord gas sampling is not required in routine deliveries in the Intermountain Healthcare system, and Apgar scores and NICU admission are used as functional markers for neonatal hypoxia.
The use of intrapartum fetal heart rate monitoring as a screening test has been rightly questioned.9 In our data set, there are more than 40 combinations of fetal heart rate characteristics that qualify a fetal heart rate pattern as category II. Combining so many different fetal heart rate patterns into a single “indeterminate” category surely detracts from the use of fetal heart rate monitoring as an indicator of fetal condition. Future research should focus on the subcharacterization of category II fetal heart rate tracings, the cumulative effect of the duration of unfavorable fetal heart rate characteristics, and the effects of other risk factors to differentiate those at highest risk from those at low risk for neonatal compromise and so to improve the predictive value of fetal heart rate monitoring.