Reproducibility of Interpretation of Fetal Heart Rate Tracings

Review Article

Reproducibility of Interpretation of Fetal Heart Rate Tracings

Corresponding authorDr. Daniel M. Avery Jr, MD, Professor of Obstetrics & Gynecology and Professor of Community and Rural Medicine, College of Community Health Sciences, The University of Alabama, 850 5th Avenue East, Tuscaloosa, Alabama 35401, Tel: 205-348-1366; Fax: 205-348-4429; Email:
Interpretation of fetal heart rate (FHR) tracings has been difficult because of the lack of agreement in definitions and nomenclature. Studies have shown that the interpretation of FHR tracings can be unreliable. The Eunice Kennedy Shriver Institute of Child Health and Human Development NICHD), the American Congress of Obstetricians and Gynecologists (ACOG) and the Society for Maternal-Fetal Medicine (SMFM) developed a new three-tiered classification system of fetal heart rate abnormalities and a system for interpreting these abnormalities. ACOG published classifications and recommendations based on the appearance of the fetal heart tracing to provide some basis of decision making.
Materials and Methods: An online survey was presented by email to various obstetricians randomly distributed across the United States. Respondents were asked if they were aware of the new ACOG fetal heart rate (FHR) monitoring interpretation guidelines and used the terms “Category I, II, and III” terminology when identifying and interpreting FHR tracings. The survey focused on the actual interpretation of seven carefully chosen, 10-minute strips which the physician was asked to label as a Category I, II, or III tracing.
Results: Based on survey results, 96.5% of physicians reported having knowledge and awareness of ACOG’s categorization of fetal heart rate tracings, and 25.6% of those physicians acknowledged using these criteria when classifying the tracings. There is still significant variability in how physicians practicing obstetrics interpret FHR tracings. Interpretation of FHR tracings is most consistent when the tracing is normal
Conclusions: The findings of this study reinforce the hypothesis that even with the NICHD/ACOG criteria, there is low reproducibility of interpretation of fetal heart rate tracings. When confronted with seven fetal heart rate tracings, physicians practicing obstetrics in the United States showed that there is variability in interpretation of fetal heart tracings. This difficulty may exist because the guidelines are not clear-cut and there is room for subjectivity when labeling tracings as a specific category

and implementing further medical management. The ACOGguidelines for categorization of fetal heart rate tracings stillleave some question as how to definitely label each tracing. Also striking was the low adoption of the three-tiered system in practice.


The detection of the fetal heart beat dates back to 1650 with the French physician Marsac [1]. Fetal heart rate monitoringduring labor was first described by Evory Kennedy, a British physician in 1833 [1]. The fetoscope to detect the fetal heartbeat was invented by David Hillis in 1917 and Joseph DeLee in 1922 [1]. Phonocardiography to detect the fetal heart rate wasdeveloped in 1931 [1]. In 1958 Edward Hon introduced continuous fetal heart rate (FHR) monitoring to identify heart ratepatterns associated with hypoxic changes during labor that caused cerebral palsy and stillbirths so that the baby could bedelivered expeditiously when these patterns were present [2- 5]. Continuous FMR monitoring was proposed as a screeningtest for asphyxia to reduce perinatal morbidity and mortality [3,6-8]. Fetal heart rate monitoring has increased the numberof cesarean sections and operative vaginal deliveries but has made no impact on the incidence of cerebral palsy [4,5,9]. In1970, intra-partum asphyxia was demonstrated not to be a major cause of cerebral palsy and only accounts for 10% of thecases [4,6]. The incidence of cerebral palsy has remained stable over time at an incidence of 2 per 1,000 live births [6].
FHR monitoring is the most commonly used obstetric procedurein this country [5]. Continuous electronic fetal heart ratemonitoring consists of an electrode attached directly to the fetal head or the maternal abdominal ultrasound to detect thefetal heart rate plus an external tocodynamometer to identify contractions [7]. Electronic fetal monitoring has increasedover time and most women in labor (84%) today undergo FHR monitoring despite no demonstrated benefit [5]. Ananth et alin 2013 have published data that suggests that FHR monitoring may decrease neonatal morbidity and low 5 minute APGARscores [5]. Defensive medicine and fear of litigation have also increased the rate of cesarean section [9].

Fetal Heart Rate Tracing Categories
In 2008, the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), the American Congressof Obstetricians and Gynecologists (ACOG), and the Society for Maternal-Fetal Medicine (SMFM) developed a newthree-tiered classification of fetal heart rate abnormalities and a system for interpreting these abnormalities [2]. CategoryI FHR tracings are normal tracings which are not associated  with fetal asphyxia [2]. They include a baseline heart rate between110-160, moderate variability defined as “fluctuations in the baseline heart rate that are irregular in amplitude andfrequency of 6-25 bpm”, no late or variable decelerations, but possible early decelerations, and possible accelerations [2].Category III FHR tracings are abnormal and indicative of hypoxic  risk to the fetus and possible acidemia [2,10]. They includeeither no baseline variability or the presence of recurrent late decelerations, variable decelerations, bradycardia, ora sinusoidal pattern [2].
Category II FHR tracings are indeterminate and include a wide variety of possible tracings that do not fit in either Category Ior Category III [2,10]. The classification of Category II tracings includes the following: bradycardia with variability, tachycardia,minimal variability, no variability with no recurrent decelerations, marked variability, absence of induced accelerationseven after fetal stimulation, recurrent variable decelerations with minimal or moderate baseline variability, prolonged decelerationslasting more than two minutes, but less than ten minutes, recurrent late decelerations with moderate variability,variable decelerations with other characteristics such as slow return to baseline, overshooting the baseline, or ‘shoulders’[2].
Reduced FHR variability is the most reliable indicator of fetal compromise [7]. A flat FHR tracing with no variability may reflectneurologic damage to the fetus that has already occurred [7]. According to a study by Jackson et al, Category I and CategoryII patterns are common in labor and Category III are unusual [3]. Perinatal morbidity is associated with an increase intime in Category II during the last two hours of labor [3].
Management of Fetal Heart Rate Tracing Categories
The management of FHR abnormalities is based upon the classification of the FHR tracing and clinical situation of thepatient and her risk factors. Category I FHR tracings are considered  “normal” and are not typically associated with fetalcomplications, such as academia. These can be managed either through continuous monitoring or through periodic monitoring.ACOG suggests that these patients be monitored every 30 minutes during the first stage of labor and every 15 minutesduring the second stage of labor and changes in management  are only necessary if the category of the FHR changes.
Category II FHR tracings are indeterminate and contain many possibilities and management is typically determined by whichof the possibilities exist. These tracings require closer supervision,  more frequent evaluation, documentation and correctionof abnormalities by conservative management and intrauterine resuscitation [2]. Accelerations and moderate variabilitysuggest normal acid-base balance [2].
Category III FHR tracings are abnormal and these tracings have been associated with adverse neurologic abnormalities,although the predictive value is poor [2,10]. When intrauterine  resuscitation of these abnormalities fails, delivery shouldbe expedient [2]. Studies are lacking to demonstrate how soon the delivery should be effected. The traditional “decision-to-incision time” of thirty minutes to perform a cesarean section has not been proven [2].
Efficacy and Reliability of Fetal Heart Rate Monitoring
There is controversy over the efficacy of fetal heart rate monitoring [10]. Interpretation of FHR tracings is difficult because of lack of agreement in definitions and nomenclature [7,8]. In a low risk pregnancy there is no data demonstrating that FHR monitoring is superior to intermittent auscultation [10]. The  majority of women in this country are monitored during labor with no known benefit [5]. Studies have shown that the interpretation of FHR tracings is unreliable [10]. There is both inter-observer (21%) and intra-observer (22%) variability in FHR interpretation [10,11]. When the FHR tracing is not normal, the significance of the tracing is difficult to ascertain [11]. Obstetricians interpret FHR tracings similarly only 29% of the time [10]. The false positive rate is 99% [9]. Even when an obstetrician reviews a tracing he has previously interpreted, 21% of the time, he interprets it differently later [10]. Interpretation of FHR tracings are most consistent when the tracing is normal [3,10]. If the neonatal outcome is known, obstetricians interpret tracings differently [10]. So, there is variability in interpretation of FHR tracings [10].
Reproducibility of interpretation of fetal heart rate tracings may be difficult. The variability and possible patterns of Category II tracings increase the difficulty of interpretation of FHR tracings [3]. Chiossi et al reported that the interpretation of FHR non-reassuring patterns were fraught with both uncertainty and fear of missing a decompensating pregnancy [9]. The quality of the FHR tracing and scaling used on the monitor also affect interpretation [7]. There is also variation in response to non-reassuring patterns depending on whether the interpretation is by a nurse, resident or attending obstetrician [9]. In reviewing perinatal deaths, approximately 50% of FHR interpretations have been questioned in one study [6]. There is also variation in FHR patterns depending on gestational age, maternal conditions, medications, etc [8]. In summary, physicians are usually in agreement with FHR tracings that are normal (Category I) and those that demonstrate severe fetal compromise (Category III) [8]. It is those tracings in Category II with all the possibilities that exist that pose the problem.
Training in FHR monitoring is an integral component of clinical obstetrical care that improves interpretation, communication and management skills, inter-observer consistency, emergency response, improved safety and fewer adverse events [6]. In a study by Ayres-de-Campos, baseline estimation which is an important aspect of analyzing FHR tracing patterns, can be reproduced with prior training [12]. Inter-observer agreement of FHR baseline can be improved with prior education [12]. Reliability can be improved with instruction and training in the classification system [8].
Materials and Methods
This cross-sectional study was approved by the Institutional Review Board of The University of Alabama. One of the study’s authors (DEH) selected 7 fetal heart rate (FHR) tracings in the course of his practice which contained no identifying information whatsoever. From those tracings, 10-minute segments were extracted for presentation to the subject physicians which were chosen to assure optimal clarity upon photocopying and electronic transmission. Eighteen hundred email addresses of physicians who identified themselves as obstetricians were  purchased from Integrated Medical Data, (300 Carnegie Center, Princeton, NJ) and an online survey was presented to them via email. Integrated Medical Data is a third party healthcare database provider. The subject physicians were asked if they were aware of the NICHD/ACOG FHR monitoring interpretation guidelines and whether they used the terms “Category I, II and III” when interpreting FHR tracings. The physicians were additionally asked to interprep each of the seven FHR tracings and label them as a Category I, II or III tracing.
Integrated Medical Data randomly distributed the online survey to those 1800 physicians via email. No statistical power testing was done prior to deciding on the number of physicians to sample as this was felt to be a descriptive study. Randomization was implemented using the Microsoft SQL function NEWID, which uses a mathematical random number generator to extract the records of the sample population. Respondents included physicians currently practicing obstetrics (N= 86). No incentives or benefits were provided to the physicians who participated in this research study. The emails sent to each physician contained a link to a constructed website containing the survey questionnaire. The internet-based website was run by WordPress using a plugin known as “Visual Form Builder” that allowed responses to be recorded and stored to a password-protected Excel spreadsheet without any trace (e.g. IP addresses) back to physician. Physicians were allowed 2 months to complete the survey before the survey was shut down, as to not incorporate spam responses.
Demographic information was collected and respondents were asked if they were aware of the NICHD/ACOG FHR monitoring interpretation guidelines and whether they used the terms “Category I, II, and III” when identifying and interpreting FHR tracings. They were then asked to interpret each of the seven carefully chosen, 10-minute fetal heart rate tracings presented alone with no associated clinical vignette including information such as gestational age or the stage of labor. Physicians were asked to label them as a Category I, II, or III tracing (Figure 1). Only physicians who responded to particular questions or tracings were included in the analysis where responses  were valid. The data were analyzed with SPSS Statistics Version21. Descriptive statistics are presented as frequencies and percentages.
Tracing 1
Tracing 2
Tracing 3
Tracing 4
Tracing 5
Tracing 6
Tracing 7

Eighty-six physicians practicing obstetrics in the United States participated in the survey. The entire study population described themselves as having graduated from an obstetrics and gynecology residency (N=86). Based on survey results, 96.5% of physicians reported having knowledge and awareness of the NICHD/ACOG categorization of fetal heart rate tracings, while 26.5% of those physicians acknowledged using these criteria when classifying the tracings. See Table 1 for physician characteristics.

Table 1. Characteristics of the Study Population (Physicians Currently Practicing Obstetrics in the U.S.)

The respondent’s categorization of FHR tracings (N=86) is found in Table 2 which includes the frequency in which each tracing was categorized as a category I, II, or III. NA, no response; Data are n (%) unless otherwise specified

Table 2. Respondent’s Categorization of Fetal Heart Rate Tracings (N=86)

This data was reported in percentages that each category was chosen. Table 2 shows there was variability in classification of the tracings based on physician responses. However, tracing 4 was the only heart tracing to obtain 100% uniformity in response of Category I. Based on observation alone, it is noted that tracing 2, 6, and 7 showed the greatest variability in responses, whereas 1, 3, and 5 had less variability in response. From the survey as a whole, category 2 was the most frequently selected response (n=226), next was category 3 (n=185), and last was category 1 (n=172) with the fewest responses.


This study was a cross-sectional research project which aimed to observe to what extent variability in interpretation of FHR tracings among practicing obstetricians in the United States still exists in the presence of the 2008 three-tier categorization described above. Based on the survey results, there is still significant variability in interpretation despite the widespread awareness of the NICHD/ACOG guidelines among practicing obstetricians. Just as other authors have found, the greatest agreement was on what constitutes a normal or Category I tracing. It is interesting that the widespread knowledge of the three tier FHR categories is not correlated with widespread use in clinical practice. This attempt to bring a level of objectivity and reproducibility of the interpretation of FHR patterns has not met with the expected success and suggests that practitioners
do not find the guidelines immediately relevant or useful for managing their laboring patients. This difficulty may exist because the criteria, primarily for Category II tracings are not clear-cut and there is room for subjectivity when labelling tracings and then implementing further medical management. However, attempts to further delineate the wide variety of Category II patterns may end up generating an unwieldy system which is unhelpful in clinical practice.

In this study, there was no specific analysis of the reproducibility of interpretation among those physicians who actually claimed to use the NICHD/ACOG classification terminology in their daily practice. Perhaps, the issue is greater than just the knowledge or training on the classification system. Those who actually use the system most frequently may prove to have the greatest agreement in interpretations. This is another area where further study may be illuminating, beyond just determining the reasons for and characteristics of obstetricians who actually utilize the NICHD/ACOG categorization terminology in their daily practice. The characteristics of those who did not respond to this survey are unknown and significant selection bias cannot be ruled out as having a large impact on our results.

The data showed that category 2 was most commonly selected response regardless of years of experience and clinical setting. Category 2 may have been chosen often because it is the broadest classification. In addition, physicians have been taught to focus on category 1 and 3 because of their definitive criteria, and to choose category 2 as a default when the particular criteria in category 1 and 3 are not met. However, we cannot ignore the possibility that physicians are simply unaware of the classification requirements for each category as well as true meanings of the terms of absent beat-to-beat variability and recurrent late decelerations. These definitions alone can skew a clinician’s perspective on what the actual requirements of category 3.

Actual limitations which should be considered when interpreting this data are as follows. Although, surveys have been used to test obstetrician responses, they are not the most valid test that can be used. The management of obstetric cases requires an individualized approach based on many maternal and laboring variables that cannot be accounted for in a survey, such as, gestational age, maternal health, pregnancy complications, stage of labor, accessibility of operating rooms and anesthesia staff, prenatal history, and maternal and fetal responses to conservative management of fetal distress. Although these factors do not directly influence an interpretation of a fetal heart rate tracing, it may skew the clinician’s perspective on labeling a category 3 versus category 2. This was evidenced by the findings that tracings 6 and 7 showed large discrepancies between choosing category 2 or category 3.

Further, the use of a third party database provider to send the surveys is also a limitation due to our inability to verify that the responders were physicians practicing obstetrics. Finally, the ability of physicians to choose multiple options for clinical setting may have skewed the data addressing the effect of type of practice on categorization and malpractice suits. This study does provide some insight into the question of interpretation of FHR patterns and strongly suggests that reproducibility is generally low, except for normal FHR tracings. However, the small number of professionals who responded to the survey and smaller still, the number of physicians who actually use the categorization terminology in daily practice limits its generalizability.

Fundamentally, the NICHD/ACOG guidelines for categorization of fetal heart rate tracings still leave some questions as to how to definitely label each FHR tracing. This supports the authors’ assertions that tracing categorization can be subjective. Knowledge of, and application of the categorization system  re not directly correlated with each other and multiple other variables must be taken into account when managing a laboring patient. These and other factors all appear to lead to a low rate of reproducibility of interpretations of FHR tracings, particularly when the tracing is abnormal.


1. Chez BF, Harvey MG, Harvey CJ. Intrapartum Fetal Monitoring: Past, Present and Future. J Perinat Neonat Nurs. 2000, 14(3): 1-18.

2.American College of Obstetricians and Gynecologists. ACOG Practice Bulletin No. 116: Management of Intrapartum Fetal Heart Rate Tracings. Obstet Gynecol. 2010, 116(5): 1232- 1240.

3.Jackson M, Holmgren C, Esplin MS, Henry E, Varner MW. Frequency of Fetal Heart Rate Categories and Short-Term Neonatal Outcome. Obstet Gynecol. 2011, 118(4): 803-808.

4.Resnik R. Electronic Fetal Monitoring: The Debate Goes On… And On…And On. Obstet Gynecol. 2013, 121(5): 917-918.

5.Anath CV, Chauhan SP, Chen HY, D’Alton ME, Vintzileos AM. Electronic Monitoring in the United States. Obstet Gynecol 2013, 121(5): 927-933.

6.Pehrson C, Sorensen JL, Amer-Wahlin I. Evaluation and Impact of Cardiotocography Training Programmes: a Systematic Review. BJOG: An International Journal of Obstetrics and Gynecology 2011; 118: 926-935.

7.Cunningham FG, Leveno KJ, Bloom SL, Hauth JC, Rouse DJ,Spong CY, editors. Williams Obstetrics. 23rd ed. United States of America: McGraw-Hill Companies. Teratology and medications that affect the fetus. 2010, 315.

8.The National Institute of Child Health and Human Development Research Planning Workshop: Electronic Fetal Heart Rate Monitoring: Research Guidelines for Interpretation. JOGNN. 1997, 26(6): 635-640.

9.Chiossi G, Costantine M, Pfannenstiel JM, Hankins GDV, Saade GR, Wu ZH. Intervention of Fetal Distress Among Obstetricians, Registered Nurses, and Residents: Similarities, Differences, and Determining Factors. Obstet Gynecol 2011, 118: 809-817.

10.American College of Obstetricians and Gynecologists. ACOG Practice Bulletin No. 106: Intrapartum Fetal Heart Rate Monitoring: Nomenclature, Interpretation, and General Management Principles. 2010 Compendium of Selected Publications, 324-334.

11.van Geijn HP. Developments in CTG Analysis. Bailliere’s Clinical Obstetrics and Gynaecology. 1996, 10(2): 185-209.

12.Ayres-de-Campos D, Bernardes J, Marsal K, Nickelsen C, Makarainen L et al. Can the Reproducibility of Fetal Heart Rate Baseline Estimation Be Improved? Eur J Obstet Gynecol Reprod Biol. 2003, 112: 49-54.

Be the first to comment on "Reproducibility of Interpretation of Fetal Heart Rate Tracings"

Leave a comment

Your email address will not be published.