Pulsed-wave tissue Doppler echocardiography for the analysis of fetal cardiac arrhythmias




Rhythm analysis of the fetal heart is hampered by the inability to routinely obtain electrocardiographic recordings of the fetus. Doppler studies of fetal cardiac tissue movements, assessing cardiac movements both qualitatively and quantitatively, have recently been described. We used a conventional high-resolution ultrasound system to obtain rhythm data from pulsed-wave tissue Doppler signals of the fetal heart in normal cardiac rhythm and in a variety of fetal cardiac arrhythmias.


Fifty-five fetuses with normal (sinus) rhythm, 45 fetuses with rhythm disturbances and two neonates (one with arrhythmia and one with normal sinus rhythm) were studied. Using a conventional high-resolution ultrasound system equipped for fetal studies, but without specific tissue Doppler hardware or software, we performed pulsed-wave tissue Doppler echocardiography (PW-TDE) of atrioventricular valve ring excursions to study the atrial and ventricular mechanical actions. In the neonates, electrocardiograms were also recorded.


PW-TDE in normal fetuses shows a typical pattern of tissue motion parallel to the long axis of the heart and in the opposite direction to the blood flow, both in systole and diastole. This pattern is easily obtained from the tricuspid valve annulus in normal sinus rhythm and shows characteristic changes in various fetal arrhythmias.


PW-TDE of atrioventricular valve annulus movement patterns may prove to be a valuable additional tool for assessing fetal cardiac arrhythmias. Copyright © 2011 ISUOG. Published by John Wiley & Sons, Ltd.


Fetal arrhythmias, particularly sustained tachycardia or bradycardia, may lead to fetal cardiac failure, hydrops and death. Therefore, precise evaluation of the type and mechanism of a fetal arrhythmia is mandatory in order to define prenatal treatment options and prognosis. This task may be challenging because the fetal electrocardiogram (ECG) does not allow beat-to-beat analysis of atrial and ventricular actions, and other methods such as fetal magnetocardiography are not widely available. At present, the study of fetal cardiac rhythm depends on M-mode echocardiography and pulsed Doppler ultrasound1–3. Using these methods electrophysiological events are inferred from observed mechanical events4.

Tissue Doppler echocardiography (TDE) has been used to study global and regional wall motion5–7. Recently it has also been applied in the human fetus, using both conventional ultrasound systems8–11 and dedicated tissue Doppler ultrasound systems12–17. The imaging modalities used so far have included analysis of color-mapped duplex images to study fetal regional tissue velocities and arrhythmias14, 15. Using pulsed-wave TDE (PW-TDE), contraction and relaxation velocities have been measured and ventricular pressures estimated8, 11, 13.

Here, we present our experience in the application of PW-TDE to study normal and abnormal fetal cardiac rhythm.


Ultrasound system and settings

Based on initial experience with fetal TDE11 we used PW-TDE to study fetal cardiac rhythm. Several different high-resolution ultrasound systems (HDI5000 and iU22, with curved array transducers C7-4 and C5-2 (Philips, Bothell, WA, USA); Voluson 730 Expert and E8 Expert, with curved array transducer RAB8-4 (GE Medical Systems, Zipf, Austria); and S2000 with 10v4 transducer (Siemens, Erlangen, Germany)) equipped for prenatal or neonatal studies, but without specific tissue Doppler probes or software, were used. Fetal TDE settings were used as described previously11. Briefly, presets for each of the ultrasound systems were developed based on the standard pulsed Doppler presets. This included selecting a low Doppler receive gain, reducing the pulse-repetition frequency and minimizing the wall filter. Subsequently, color Doppler or PW Doppler signals of tissue motion, as opposed to blood flow, were displayed. Typical peak velocities for TDE are in the range of up to 10–15 cm/s. The receive gain was reduced to about one quarter compared with the typical blood-flow Doppler preset. The presets were invoked at the beginning of every examination. Sweep speed was adjusted according to a required high temporal resolution (highest possible sweep speed) or the need to record a longer ‘rhythm strip’ (medium to lowest sweep speed).


In a non-randomized prospective manner we studied 100 fetuses of between 15 and 40 weeks' gestation that were referred for cardiac evaluation. In 55 fetuses, cardiac anatomy and function were normal; in another 45 fetuses, cardiac arrhythmias were present. In several of the normal fetuses both PW-TDE (at the level of the atrioventricular (AV) valve annulus and, if possible, including the AV valve leaflets) and PW Doppler of the blood flow through the AV valves were recorded simultaneously to study the temporal correlation between flow and tissue signals. A fetus with second-degree atrioventricular block (AVB) (type Wenckebach) was also studied using the same approach the day after delivery. A normal neonate was examined using PW-TDE with concurrent ECG.

All pregnant women had given informed consent to participate, and the study was approved by the Institutional Review Board.

Fetal studies

All fetuses had a complete fetal echocardiographic examination before entry in the study that demonstrated normal cardiac anatomy and included M-mode assessment of heart rate and AV synchronicity.

Subsequently, axial excursions of the ventricular wall at the AV valve annulus of the tricuspid valve were recorded using PW-TDE. After invoking the PW-TDE presets, as described above, the heart was imaged in an apical (or close to apical) insonation angle (less than 30°). The pulsed Doppler sample volume was adjusted in size and placed over the area covering the entire valve annulus excursion during systole and diastole. To obtain the tissue motion signals, the pulse repetition frequency was lowered (to about ± 15 cm/s), the wall filter was set to minimum and the receive gain was reduced to remove blood-flow signals. If necessary, the acoustic output was also reduced. Data acquisition was performed during fetal and voluntary maternal apnea and in the absence of fetal body movements.

The resulting motion curves permit qualitative assessment of the excursions caused by atrial and ventricular contraction and measurement of the intervals of the mechanical action segments. Intervals were measured using either the internal software of the ultrasound system or offline from digitally recorded PW-TDE images.

The normal fetuses were each studied once, whereas most fetuses with cardiac dysrhythmias were analyzed repeatedly.


PW-TDE signals could be acquired in all fetuses almost independently of fetal position. In the absence of fetal body movements, a recording of good quality was obtained within 2 min. Examples of PW-TDE recordings in fetuses with normal (sinus) rhythm and in fetuses with rhythm disturbances are shown in Figures 1–7 and S1–S4.

Figure 1.

Normal fetal pulsed-wave tissue Doppler echocardiography (PW-TDE) tracing with insonation of the right atrioventricular valve annulus from the base of the heart, where systolic wall movements appear below the baseline. During each cardiac cycle there is a systolic movement towards the apex of the heart (S′, below baseline) and two diastolic movements away from the apex (E′, A′, above baseline). The intervals denote (from left to right) the first (E′) and second (A′) diastolic excursions, the isovolumic contraction time (ICT) between the end of A′ and the onset of S′, and the isovolumic relaxation time (IRT) between the end of S′ and the subsequent E′. In this recording, a medium sweep speed was used to record a longer ‘rhythm strip’. For analysis of intervals the highest sweep speed should be selected.

Figure 2.

Fetal pulsed-wave tissue Doppler echocardiography of premature atrial contractions (PACs). (a) Non-conducted PAC. Two normal sequences of wall motions (early diastole (E′) and late diastole (A′) below baseline; systole (S′) above baseline) are followed by an early atrial event (*) which is a PAC that does not lead to a subsequent S′ because it is blocked in the atrioventricular node. The interval between two normal atrial activations (A′–A′) is indicated by a white bar. The interval between the pre-ectopic and postectopic A′ (red bar plus one white bar) is shorter than the expected interval between two normal beats (two white bars). (b) Conducted PAC. Two normal sequences are followed by an early atrial event (*) that leads to an associated ventricular contraction (&). In conducted PAC the interval between the pre-ectopic and postectopic A′ (red bar plus one white bar) is also shorter than the expected interval between three normal beats (non-compensatory pause).

Figure 3.

Fetal pulsed-wave tissue Doppler echocardiography of premature ventricular contractions (PVCs). In the same fetus, PVCs are documented by different techniques and various degrees of severity. (a) Ventricular wall movement (systolic (S′), below baseline) commences to the left with a PVC (*) followed by a long postectopic pause (post-PVC), which shows separate early (E′) and late (A′) diastole. Subsequently, there are three normal beats, probably in sinus rhythm (nl SR), with fused E′ and A′, followed by another PVC (*). The interval preceding the PVC (dashed red bar) plus the postectopic interval (solid red bar) are as long as the interval between two normal sinus beats (white bar). This is a compensatory postectopic pause, which is consistent with PVC. (b) M-mode echocardiography in the same fetus demonstrates a PVC (*) that is not preceded by an atrial activation (a). The atrial contractions are quite regular, and there is a compensatory pause after the ectopic beat (dashed and solid red lines equal two V–V intervals). (c) Later in the same fetus, PVCs (*) are seen that alternate with normal atrioventricular activation (S′) in a bigeminal manner. The interval preceding an ectopic beat is shorter than the subsequent interval. Atrial activations are seen in a regular manner above the baseline; E′ and A′ are fused in the short diastole preceding the PVC and separated in the longer postectopic diastole.

Figure 4.

Fetal pulsed-wave tissue Doppler echocardiography in supraventricular tachycardia (SVT). (a) At a heart rate of 205 bpm there is a 1 : 1 association of atrial (below baseline) and ventricular (above baseline) activations. At such fast heart rates, early diastole (E′) and late diastole (A′) always coincide. (b) In this example (with opposite orientation compared to panel (a)), to the left there is a run of SVT with a heart rate of 214 bpm that ends abruptly after a ventricular activation (arrow). Subsequently, a normal sequence of E′, A′ and S′ (systolic movement) is seen. This demonstrates an underlying re-entry mechanism of the SVT that terminates in the accessory pathway. If termination in atrioventricular re-entry occurred at the atrioventricular node, one would expect no ventricular motion at the end of the run.

Figure 5.

M-mode and pulsed-wave tissue Doppler echocardiography tracings in a fetus with ventricular tachycardia, both obtained from the same dorso-anterior fetal position, in postnatally confirmed long-QT syndrome. (a) In a four-chamber plane the M-mode line crosses through the left atrial (LA) wall, the crux and the right ventricular (RV) wall, indicating a constant and normal atrial rate. Below the crux of the heart the much faster ventricular rate is seen. a, normal atrial contractions; s, tachycardic ventricular contractions. (b) In a four-chamber plane the sample volume is placed at the lateral hinge point of the tricuspid valve. A′ denotes regular atrial contractions, occurring at a constant rate of 114/min. S′ indicates ventricular contractions at a rate of about 220/min. The alternating weaker signals represent ventricular contractions that follow passive diastolic filling only, resulting in reduced preload. Note the greater signals from the excursion of the atrioventricular valve ring, caused by atrial and ventricular contraction, in comparison with the M-mode tracing in (a).

Figure 6.

Pulsed-wave tissue Doppler echocardiography of the tricuspid valve annulus in a fetus with second-degree atrioventricular block, type Wenckebach. (a) A′ denotes the late diastolic wall excursion caused by atrial contraction. S′ denotes systolic wall excursion. There is progressive lengthening of the conduction time in successive cardiac cycles (shown in milliseconds for the first three cycles) until eventually a ventricular response is missing (*). The first following A′ is smaller because of the high-end diastolic ventricular volume that results from longer diastole and skipped ventricular emptying. (b) Longer tracing showing the arrhythmia in the neonate.

Figure 7.

Pulsed-wave tissue Doppler echocardiography of complete fetal atrioventricular block. The filled white arrows denote ventricular contractions (above baseline) and the open arrows with a thin white border indicate atrial contractions (below baseline). The dashed open arrows show atrial activations that are obscured by ventricular activation, while the open arrows with a thick white border show atrial activity immediately after ventricular emptying, leading to a larger amplitude of the atrial motion. If atrial activation occurs later after systolic wall excursion (S′), separation of wall motion during early (E′) and late (A′) diastole is seen (#).

Normal pattern of PW-TDE

There were 55 fetuses in normal (probably sinus) rhythm. In these subjects, PW-TDE showed a typical pattern of valve ring movements in systole and diastole. Both the right and left ventricular walls showed the same excursion (Figure 1), but the best signal-to-noise ratio was usually obtained from the right ventricular wall at the level of the AV valve annulus; right ventricular PW-TDE was used for further rhythm analyses.

Wall motion during early (E′) and late diastole (A′) was always directed away from the apex. There was an interval between the end of A′ and the start of systolic motion (S′) during which isovolumic contraction (ICT) took place. Sometimes there was a discernible motion directed towards the apex during ICT. Another tissue motion between the end of A′ and the beginning of S′ was also present when AV conduction was blocked (see an example of AV block II, type Wenckebach, below), indicating a possible atrial origin in these cases (see below, Figure 6). S′ appeared as a broad-velocity complex directed opposite to E′ and A′. In normal sinus rhythm and even in tachycardia, S′ was followed by a short interval with no or very little wall motion, considered to be the isovolumic relaxation time (IRT; Figure 1).

Diastolic PW-TDE patterns depended on the fetal heart rate. At heart rates above 130 bpm we observed a fusion of E′ and A′ wall movements, similar to the flow signals in PW Doppler studies of blood flow across the AV valves.

Peak diastolic wall velocities were usually larger than systolic wall velocities, ranging in the right ventricle from 3 to 7 cm/s at 15 weeks' gestation to 7 to 13 cm/s at term.

Correlation with blood-flow Doppler

By increasing the Doppler sample volume size, placing it towards the ventricular cavity and changing the Doppler settings to also detect blood flow, both the wall motion and the blood flow at right ventricular inflow or left ventricular inflow and outflow regions could be recorded. AV valve annulus movements, as well as AV inflow and, in the left ventricle, aortic outflow, were imaged in the same tracing (Figure S1). The temporal relationship of PW-TDE and conventional blood-flow Doppler signals was depicted in such tracings, confirming the interpretation of the PW-TDE signals.

Correlation of PW-TDE with the ECG in a neonate

Simultaneous recording of tissue movements using PW-TDE and ECG in a neonate confirmed a correlation between the two. A′ coincided with the P-wave, E′ equaled the early diastolic tissue motion and S′ occurred with and following the QRS complex (Figure S2).

PW-TDE in fetal arrhythmias

There were 45 fetuses with a cardiac arrhythmia, including premature atrial contractions (PACs) (n = 28), atrial bigeminal ectopic beats (n = 3), premature ventricular contractions (PVCs) (n = 2), supraventricular tachycardia (SVT) (n = 5), ventricular tachycardia (n = 1), second-degree AVB, type Wenckebach (n = 1) and complete AVB (n = 5).

In PACs the early atrial activation was recorded using PW-TDE. If a PAC was blocked in the AV node, there was no subsequent systolic excursion S′, and the interval between pre-ectopic and postectopic atrial activation was shorter than the expected interval between two normal beats (non-compensatory pause; Figures 2a and S3). In conducted PACs the premature atrial activation was followed by a ventricular response. Conducted PACs showed an early atrial activation with associated ventricular response (Figure 2b), but also a non-compensatory pause.

PVCs led to a long postectopic interval and wide separation of E′ and A′. Subsequent normal beats showed the normal E′–A′ pattern i.e. narrow or (in case of heart rates above about 130 bpm) fused E′ and A′. The sum of the intervals before and after a PVC was twice the interval between two normal sinus beats (compensatory pause after a PVC; Figure 3). This feature, however, could not be seen in sustained bigeminal PVCs (Figure 3c).

In SVT, E′ and A′ always coincided, and there was a 1 : 1 association of atrial and ventricular motion (Figure 4a). In the case of a re-entry mechanism of the SVT, the tachycardic run ended abruptly and the normal sequence resumed (Figure 4b). Using high sweep speeds, analyses of the ventriculo-atrial interval would be easily possible, allowing inferences of the nature of the tachycardia, as is possible using blood-flow Doppler18, 19.

PW-TDE of ventricular tachycardia with AV dissociation is shown in Figure 5. There were regular atrial activations, but much more rapid and dissociated ventricular contractions. This fetus had ventricular tachycardia and was found to have long QT syndrome with torsade de pointes tachycardia after birth.

In second-degree AVB (type Wenckebach) there was a progressive increase of the interval between A′ and S′ until a ventricular contraction was skipped, indicating final non-conduction after progressive lengthening of the conduction time. After a skipped ventricular beat the following diastolic peak velocity was decreased (Figures 6 and S4) because the ventricle had not emptied in between. In the affected case this pattern persisted in the neonate on the day after delivery.

In third-degree AVB there was complete dissociation between wall motions caused by atrial and ventricular contractions with a slow ventricular rate. In Figure 7, PW-TDE demonstrates a normal atrial rate but a slow ventricular rate. PW-TDE also revealed the different degrees of ventricular filling depending on the incidental timing of atrial and ventricular contractions. Immediately following a ventricular contraction, filling conditions were optimal with low ventricular volume and open AV valves, thus leading to a pronounced atrial contraction (A′) on PW-TDE.


Assessment of fetal arrhythmias remains challenging as conventional ECG is not available in the fetus. Recent improvements in signal filtering and processing of fetal ECGs acquired through the maternal abdomen have improved the display of signal-averaged tracings and may be useful for measuring time intervals20, but do not allow beat-to-beat analysis of atrial and ventricular activity. Other methods, such as magnetocardiography, have been used to analyze fetal cardiac rhythm21–24, but have considerable practical limitations and have therefore not gained widespread clinical use. At present, the commonly used techniques for studying fetal cardiac rhythm remain M-mode echocardiography and PW Doppler ultrasound1, 18, 25–29.

Echocardiography and Doppler ultrasound display mechanical events that follow the electrical activation of atrial and ventricular chambers in the fetus.

M-mode echocardiography has the advantage of a high temporal resolution, but it is dependent on fetal position and may require several and prolonged attempts to acquire appropriate tracings.

Blood-flow assessment by PW Doppler ultrasound can display diastolic and systolic flow events in one recording, such as in left ventricular inflow and outflow, at the border of the superior vena cava and the ascending aorta, or even further from the heart at the branch pulmonary arteries and veins or at the renal artery and renal vein18, 22, 24, 26, 27. These approaches are less dependent on the fetal position, but have the disadvantage of different PW propagation times that may interfere with the analysis of electromechanical coupling.

TDE, like M-mode, records tissue motion and is one step closer to the actual electromechanical basis than is blood-flow Doppler. Unlike M-mode, TDE signals contain information from both atrial and ventricular chambers in one continuous line8, 11, making it less equivocal and, in our opinion, often easier to read. In PW-TDE there is only one segment interrogated for tissue excursion; A′ can be obscured by the (stronger) S′ if they coincide (Figure 7). In this case, the M-mode would still show an atrial signal. In such a setting, M-mode and PW-TDE are clearly complementary.

TDE can be obtained either from color duplex images or, more globally, using PW Doppler of the AV valve annulus. Color TDE with dedicated equipment has been used for offline measurement of the segments of the mechanical heart cycle and to study fetal arrhythmias15, 30. Color TDE-derived measurements of AV conduction track ECG-PR intervals more closely than blood-flow Doppler-derived AV intervals31. Using a specialized analysis system, an assessment of regional velocities in fetuses with cardiac arrhythmias has been described previously15. This approach, however, requires multiple measurements of time intervals of unknown precision, and only a limited number of examples in the form of unprocessed data from fetal arrhythmias have been published15.

In the present study, we have demonstrated that high-resolution ultrasound systems for fetal imaging without specific hardware or software can be used for the recording of tissue motion and detailed characterization of fetal arrhythmias. Such systems are in widespread use in units faced with fetal arrhythmias18, 24, 26. Using this approach, detailed descriptions of normal PW-TDE recordings and examples of the common fetal arrhythmias were obtained. Diastolic and systolic wall motion velocities were within the range of those published previously8, 11. These findings may improve the ability to precisely analyze fetal arrhythmias and to select appropriate therapeutic options.

Measurements of short time-intervals, such as the PR interval, have also been recently attempted in the fetus because lengthening of PR intervals may precede high-degree AVB32. In principle, estimation of fetal PR intervals from blood-flow Doppler is affected by the delay between the electrical events, the resulting tissue movement and the onset of the flow propagation. This might lead to an overestimation of the true AV conduction time from blood-flow Doppler studies. There is also a possible confounding effect of fetal heart rate on these measurements. PW-TDE appears to be more accurate in estimating such short time-intervals from mechanical actions than is blood-flow Doppler assessment31. Recently, PW tissue Doppler has also been used to measure the isovolumic intervals and ejection time, and the myocardial performance index was calculated from these measurements33, 34.

In most clinical cases we did not measure all segments of the mechanical cardiac cycle quantitatively, but rather analyzed the pattern of atrial- and ventricular-derived valve annulus motions. This way, typical and highly characteristic patterns for the different arrhythmias became apparent. In some cases, such as in the AVB II (type Wenckebach), however, progressive lengthening of the mechanical segment encompassing the AV conduction time was also analyzed quantitatively, thus showing the potential of the technique for measuring time intervals.

Intriguing findings were the mechanical events seen in the interval between the end of A′ and the beginning of S′; this segment is thought to contain the ICT. In normal heart rhythm, there is a sharp tissue motion in the same direction as, and immediately preceding, the broad movement during S′. In our study we found a tissue movement towards the apex between the end of A′ and the beginning of S′ also in fetal AVB when it is readily seen in cycles that show no S′ following A′ (Figure 6). In fetuses with AVB, this motion was seen as a small, but discrete peak immediately following A′, but in the opposite direction (Figure 7). We speculate that in these cases the peak preceding S′ represents a recoil effect of atrial contraction on the valve annulus.

PW-TDE to study fetal cardiac rhythm is available on all diagnostic ultrasound systems currently used for fetal diagnosis. Initial teaching experience shows that the ability to obtain diagnostic PW-TDE tracings can be acquired quickly and its interpretation appears easier than that of M-mode, in part because in PW-TDE the atrial signals are usually clearer than those on M-mode. Therefore, PW-TDE may prove to be a valuable additional tool for assessing fetal cardiac arrhythmias.


The following supporting information may be found in the online version of this article:

equation image Figure S1 Simultaneous display of mitral inflow, aortic outflow and left ventricular tissue Doppler as well as tricuspid inflow and right ventricular tissue Doppler.

Figure S2 Simultaneous recording of pulsed-wave tissue Doppler echocardiography (tricuspid valve annulus) and the electrocardiogram in a healthy newborn.

Figure S3 Fetal pulsed-wave tissue Doppler echocardiography of premature atrial contractions.

Figure S4 Pulsed-wave tissue Doppler echocardiography of the tricuspid valve annulus in a fetus with second-degree atrioventricular block, type Wenckebach.