Article Text

Download PDFPDF

Measurement of neonatal heart rate using handheld Doppler ultrasound
  1. Amanda Dyson1,
  2. Michele Jeffrey1,
  3. Martin Kluckow1,2
  1. 1Royal North Shore Hospital, Neonatal Intensive Care Unit, Sydney, Australia
  2. 2University of Sydney, Sydney, Australia
  1. Correspondence to Dr Amanda Dyson, Grace Centre for Newborn Care, The Children's Hospital at Westmead, Hawkesbury Road, Westmead, NSW 2145, Australia; amandadyson{at}


Objective This pilot study aimed to determine whether handheld Doppler ultrasound is feasible and reliable for measuring neonatal heart rate (HR) when compared with ECG.

Setting Stable newborns were recruited from the neonatal intensive care unit and postnatal ward between July 2014 and January 2015 at Royal North Shore Hospital, Sydney, Australia.

Interventions Each newborn had their HR recorded every 15 s over 145 s using four different modalities: ECG, counted audible Doppler (AD) over 10 s, pulse oximetry (PO) and the Doppler display (DD).

Outcome measures The correlation and variation between each modality and ECG.

Results 51 newborns with a median gestational age of 38 weeks (27–41) and a mean weight of 2.78 kg (0.82 to 4.76) with a median postnatal age of 3 days (0–87) were studied. There was a mean difference of 0.69 bpm (95% CI −2.9 to +1.5) between AD-HR and ECG-HR with good correlation between modalities (r=0.94, p<0.01). The median time to achieve AD-HR was 3 s (1–45). The mean difference between DD-HR and ECG-HR was 5.37 bpm (95% CI −12.8 to +2.1) with moderate correlation (r=0.37, p=0.04). The mean difference between PO-HR and ECG-HR was 0.49 bpm (95% CI −1.5 to +0.51) with good correlation (r=0.99, p<0.01). The variability between AD-HR and ECG-HR decreased with decreasing weight.

Conclusions AD-HR correlates well with ECG-HR. Further research in the delivery room is recommended before using AD-HR in this area.

  • heart rate
  • neonate
  • Doppler
  • neonatal resuscitation
View Full Text

Statistics from

What is already known on this topic?

  • There are many modalities for newborn heart rate (HR) monitoring.

  • There are a number of disadvantages to current modalities in terms of time taken to produce a measurable HR and accuracy.

  • Handheld Doppler is a reliable tool for detecting fetal HR in delivery suites and is familiar to healthcare providers.

What this study adds?

  • Handheld Doppler is reliable in monitoring newborn HR when compared with ECG.

  • Handheld Doppler is quicker than other modalities at providing a recordable HR.

  • Handheld Doppler could easily be translated into a new method for determining neonatal HR.


Determining the need for resuscitation at birth primarily relies on the practitioner's clinical assessment of neonatal heart rate (HR) and respiratory effort.1 A neonate's response to resuscitation is also assessed by changes in their HR. While respiratory effort and colour can be assessed visually, there are a number of ways in which HR is assessed clinically. In well neonates, this is usually performed by auscultating the precordium or feeling for pulsations of the umbilical cord.1–3 It has been demonstrated that both auscultation and palpation underestimate the HR when compared with ECG-HR by a mean of −14 and −21 bpm, respectively.4 This is likely to lead to unnecessary intervention for otherwise well babies. Alternative methods of monitoring neonatal HR use pulse oximetry (PO) or continuous ECG. It has been shown that ECG is quicker at detecting HR at birth than PO.5 ,6 The median time taken to achieve a reliable HR with ECG once it had been attached was 2 s compared with that by pulse oximeter, which was 23 s.6 As a result of this, the recently updated Neonatal 2015 International Consensus on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care guideline recommends the use of ECG as an adjunct to clinical assessment and PO at birth7 It has been suggested that PO may underestimate the true HR at birth when compared with ECG over the first 7 min of life.8 ECG is generally considered to be the gold standard for HR monitoring as a result of which we have used this modality as the comparator for HR monitoring in this study.8

Having an audible sound that can be heard by the whole neonatal resuscitation team may be beneficial in guiding neonatal resuscitation and reducing the uncertainty associated with one individual being responsible for counting, averaging and communicating HR to the team. It is a familiar tool that is present in the delivery suite as it is commonly used by obstetric teams during labour to detect and monitor fetal HR.9 Handheld Doppler use for measuring newborn HR has been described in the literature but needs further investigation before being used clinically.10 We postulated that the use of handheld Doppler may be of value in improving the ease, speed and accuracy of HR monitoring in the immediate newborn period. The primary aim of this study was to examine the accuracy and feasibility of using handheld Doppler to measure neonatal HR when compared with ECG and PO.


Study population

Patients were recruited from the Neonatal Intensive Care Unit and Post Natal Ward of our perinatal centre between July 2014 and January 2015. Any stable neonate born at our institution during the study period was eligible to be included in the study. Neonates with known congenital heart disease were excluded as were critically unwell or unstable neonates requiring ventilation or inotropic support.

Study protocol

Neonates had their HR monitored in four different ways simultaneously over a 145 s period: via ECG (display), PO (display) and using a handheld Doppler (recording both the Doppler display (DD) HR and counting audible Doppler (AD) HR over 10 s). All AD measurements were performed by a single investigator. A second investigator recorded the four measurements every 15 s providing us with 10 measurement time points per baby and 40 data points per baby. The position of the Doppler on the neonate's chest and the time taken to locate the HR using the Doppler were recorded. The Doppler was held over the sternum in the 2nd intercostal space (angle of Louis) and then moved to the left then right sternal edge until a regular AD sound was heard. This landmark was chosen as it should correspond to the junction of the ascending aorta and aortic arch and the branching of the main pulmonary artery.11 The pulse oximeter used was a Radical-7 Pulse Oximeter (Masimo, Irvine, California, USA). The averaging time for HR for the pulse oximeter is 2 s with a reported accuracy of 3–5 bpm in neonates depending on the presence or absence of motion. The ECG monitor used was a GE Marquette Hellige Eagle 1000 (GE Medical Systems, Waukesha, Wisconsin, USA) with a range of 15–300 bpm and an averaging time of 2 s. The HR was read off the monitor of each of these devices.

Handheld Doppler ultrasound

The Doppler used for this study was the Hadeco Smartdop 45 (Hadeco—2-7-11 Arima, Miyamae-ku, Kawasaki, 216-0003, Japan), which is a bidirectional handheld Doppler usually used for measuring arterial and venous blood flow in the extremities and also for detecting fetal HR. An 8 MHz probe was used for this study as the focal depth is approximately 1 cm, which we found to be better for detecting neonatal HR than the 2 MHz probe typically used for fetal HR monitoring, which has a focal depth of 10 cm. The measurement range of the Doppler in the mode we used was 60–220 bpm and the display HR was averaged over 4 s (stated accuracy of 5%12).


The Bland-Altman method was used to compare the difference between alternative ways of measuring HR with ECG. Each individual neonate's readings in each modality were averaged over the 145 s collection period. The difference in means between two modalities was compared using a Student's t-test and the Pearson r statistic was used to examine the correlation between modalities. Bland-Altman plots were then produced. This was performed using SPSS V.22.0 (IBM, Armonk, New York, USA).


Fifty-one patients were recruited; their demographic characteristics are described in table 1.

Table 1

Demographic characteristics of study participants

The median time taken to achieve measurable AD-HR sound was 3 s (1–45 s). The position in which this was achieved was in the 2nd intercostal space over the sternum in 70.5% of cases, the 2nd intercostal space over the left sternal edge in 25.5% and the 2nd intercostal space over the right sternal edge in 3.9%. The ECG, PO and counted AD had the fewest missing data points (5.7%, 4.9% and 5.9%, respectively), while the DD did not record 20.1% of data points and displayed a blank screen.

Correlation between modalities and ECG

All modalities correlated with ECG with statistical significance (table 2); however, the DD-HR was far less correlated than AD-HR or PO-HR.

Table 2

Pearson's correlation between modality and ECG as a comparator

Audible Doppler versus ECG

There was no evidence for within-patient differences between AD-HR and ECG-HR (mean difference 0.69 bpm, 95% CI −2.9 to +1.5, p=0.52). As demonstrated in figure 1, the SD of the differences from the mean was narrow with the limits of 1.96*SD of the mean being 14.4 to +15.8 bpm. The SD of the differences in HR between AD and ECG was lower in smaller neonates <1.5 kg when compared with larger babies (see table 3).

Table 3

Mean difference and SD of the difference between Doppler-HR and ECG-HR by weight

Figure 1

Bland-Altman plot comparing audible Doppler heart rate (HR) and ECG-HR. Mean difference between HR measurements and 1.96*SD of the difference are denoted by horizontal lines.

Doppler display versus ECG

There was no statistical evidence for within-patient differences between DD-HR and ECG-HR (mean difference 5.4 bpm, 95% CI −12.8 to +2.1, p=0.15). As demonstrated in figure 2, the SD of the mean difference between modalities was greater than with other modalities (1.96*SD mean=−50.2 to +39.4 bpm). The SD of the difference between modalities was greater in babies >1.5 kg (see table 3).

Figure 2

Bland-Altman plot of the difference between Doppler display and ECG. Mean difference between heart rate measurements and 1.96*SD of the difference are denoted by horizontal lines.

ECG versus pulse oximetry

There was no evidence for within-patient differences between PO-HR and ECG-HR (mean difference 0.49 bpm, 95% CI −1.5 to +0.51, p=0.33). The SD of the differences between modalities is narrower than other modalities as demonstrated in figure 3 (1.96*SD from mean=−6.41 to +7.4 bpm).

Figure 3

Comparison of heart rate (HR) between pulse oximetry (PO) and ECG. Mean difference between HR measurements and 1.96*SD of the difference are denoted by horizontal lines.


We have demonstrated that there is a good correlation between AD-HR and ECG-HR in a population of stable newborns of varying size and gestational age. The Doppler was quick and easy to use with a median time taken to obtain a measurable AD-HR of 3 s (1–45), which is quicker than the reported median time taken to obtain PO trace of 23 s.6 The Doppler provided a clear audible HR, which could theoretically be useful in providing a whole resuscitation team with HR information quickly rather than relying on an individual to auscultate, calculate and then communicate the HR to the team. The presence of an AD sound implies that there is cardiac output/blood flow, while the presence of an ECG HR does not always imply cardiac output (eg, in the rare case of electromechanical dissociation).

The variability between AD-HR and ECG-HR is a little higher than that demonstrated between PO-HR and ECG-HR (mean difference of 0.69 bpm compared with 0.48 bpm), which is not clinically significant. The mean difference that we found between AD-HR and ECG is less than that previously reported by another group (3.15 bpm) with a similar correlation found between the two modalities.13 There is previously reported evidence that the use of a cord clamp Doppler applied at delivery correlates well with palpated pulse HR for 1 min; however, umbilical artery pulsatility decreases after birth, may become unreliable14 and as previous mentioned palpated HR underestimates ECG-HR.4 While AD is quicker than PO at obtaining HR, it would not replace this modality in neonatal resuscitation as it does not provide information about oxygen saturation.

The SD of the difference between PO-HR and ECG-HR was smaller in our study than that reported in other similar studies.15 One group comparing HR values in newborns at birth using PO and ECG found a mean difference of 2 bpm with 95% confidence limits at ±26 bpm.15 It is possible that this higher level of correlation reflects the stability of our patient group; most newborns were either asleep or being held comfortably in a parent's or researcher's arms during data collection, while previous studies have performed this measurement in the delivery suite where there would be a greater variability in HR. While we did find that both AD-HR and DD-HR were harder to record during crying, it was still possible to obtain AD-HR in crying newborns. We postulated that holding the Doppler at the angle of Louis or directly to the right or left of this anatomical landmark may allow us to detect flow in one of the two outflow tracts of the heart; either the main pulmonary artery or ascending aorta. The angle of Louis corresponds to the level of the 2nd cervical rib where the ascending aorta becomes the aortic arch and where the main pulmonary artery branches.11 We found AD-HR over the sternum the majority of the time as this was the anatomical location in which we started holding the Doppler; however, we recognise that the location used to detect Doppler HR depends on the focal depth of the probe used. The ideal position and probe to use for newborn HR measurement will need to be the subject of future research. We found that over time the investigator using the Doppler became quicker at detecting a measurable HR, which may suggest that training is of benefit if this method were used clinically. As it is small and handheld, it should not be any more cumbersome than using a stethoscope. Ideally, it will be possible to have a device that is compatible with handheld obstetric Doppler ultrasounds that are already present in the majority of delivery rooms making it practical and cost-effective. This is a direction for future research.

There are a number of limitations to our study, including that the handheld Doppler used is not purpose built for measuring HR in newborns. It is quoted as reliable for monitoring HR 60–220 bpm12 and the focal depth of the probe (1 cm) may not be sufficient for detecting flow in the great vessels of larger babies. It may also partly explain why the Doppler did not calculate and display a reliable HR for our patients. We found that the differences in HR between AD/DD-HR and ECG-HR were less in patients weighing <1.5 kg, which is likely to reflect the focal depth of our Doppler probe. We recognise that having an accurate DD-HR reading would be a more practical application of this modality and optimising this use of Doppler HR monitoring will be the subject of future research. Another limitation of this study is that our patient population is more stable than that found in the delivery room where there is more variability of HR. The results of this study are therefore not currently generalisable to babies in the delivery suite as the reliability of this modality has not been tested in unstable newborns or those with a HR of less than 100 bpm.


AD-HR is a feasible and reliable way of measuring HR in newborn infants. The advantages of this modality are that it obtains HR quickly and provides a clearly heard audible heart beat, which could theoretically provide audible HR information for a whole resuscitation team within seconds. Future directions for research include the development of a purpose-built handheld Doppler for neonatal HR monitoring that is able to provide an accurate displayed HR as well as testing the reliability of this tool in the delivery suite. The use of handheld Doppler in the delivery room cannot be recommended until the results of further research testing its feasibility and reliability during resuscitation are available.


The Hadeco Smartdop 45 was kindly loaned to us by The Critical Group Pty ACN 002 628 846.


View Abstract


  • Contributors AD and MK designed this study in discussion with MJ. AD and MJ implemented the study protocol, performed the data collection and results analysis. All authors were involved in reviewing the final version of the manuscript.

  • Competing interests None declared.

  • Ethics approval This study was approved by the human research ethics committee of The Northern Sydney Local Health District (RESP/14/127).

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement Questions about the data should be directed to the corresponding author (Amanda Dyson).

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.