Objective: To compare ease of recording and reliability of ultrasonographic approaches used to time fetal heart atrial and ventricular contractions.
Methods: Seventeen consecutive fetuses seen at our fetal cardiology unit for possible fetal cardiac arrhythmia were included in this study. The same ultrasonographer obtained M-mode tracings of atrial and ventricular free walls, atrial wall and opening of the aortic valves, a peak of the mitral valve, and the opening of the aortic valves; and Doppler signals of flow-velocity waveforms in the outflow tract of the left ventricle and simultaneous flow-velocity waveforms in the aorta and superior vena cava. The outcome measures were rate of successful attempts and intra- and interobserver reliability coefficients.
Results: Valid recordings were made for all patients with one M-mode (atrial and ventricular free walls) and two Doppler (intraventricular, superior vena cava, and ascending aorta) approaches. Atrioventricular intervals were significantly longer with M-mode compared with Doppler ultrasonography. Reliability coefficients were excellent (at least 0.89) for all intraobserver measurements. Comparisons of atrioventricular and ventriculoatrial interval measurements made by two observers gave the following intraclass correlation coefficients (95% confidence interval): atrioventricular = M-mode: 0.87 (0.79, 0.91), left ventricular outflow: 0.93 (0.89, 0.96), superior vena cava–aorta: 0.98 (0.97, 0.99); ventriculoatrial = M-mode: 0.79 (0.67, 0.87), left ventricular outflow: 0.97 (0.95, 0.98); superior vena cava–aorta: 0.99 (0.98, 0.99).
Conclusion: Fetal atrioventricular intervals measured indirectly from M-mode or Doppler tracings were equally reliable when measured by the same observer; the Doppler approaches had better correlation between measurements made by two different observers.
Ultrasonographic assessment of fetal arrhythmia has been described extensively.1–3 With it, atrial and ventricular depolarizations are identified indirectly by their mechanical (M-mode) or hemodynamic (Doppler) consequences. On M-mode tracings, atrial, ventricular, and valvular movements indicate atrial or ventricular contractions. Doppler signals usually advocated are flow-velocity waveforms recorded in the lower part of the outflow tract of the left ventricle, displaying inflow through the mitral valve followed by ejection toward the aortic valve.4 Another Doppler approach is the simultaneous recording of flow velocities in the superior vena cava and ascending aorta.5 We recently found in the ovine fetus that in the superior vena cava, contrary to the left ventricular outflow approach, the retrograde a wave caused by atrial contraction constantly was visible, even at heart rates over 200 beats per minute.6
A MEDLINE search using the term “fetal arrhythmia” from 1990 to January 2000 found 661 articles. None of those publications assessed the relative ease of making echocardiographic recordings suitable for arrhythmia investigation or the reliability of measurements. The aim of the present investigation was to compare the rate of successful attempts to record fetal M-mode and Doppler tracings suitable for timing atrial and ventricular contractions and to assess the reliability of the measurements.
Doppler recordings are more reliable than M-mode tracings for timing atrioventricular contractions of the fetal heart.
Fetal Cardiology Unit, Pediatric Cardiology Division, Department of Pediatrics, Sainte-Justine Hospital, University of Montreal, Montreal, Canada.
Address reprint requests to: Jean-Claude Fouron, MD, Saint Justine Hospital, Fetal Cardiology Unit, 3175, Cote Ste. Catherine, Montreal, Quebec H3T 1C5, Canada. E-mail: firstname.lastname@example.org
Received February 14, 2000. Received in revised form May 30, 2000. Accepted June 22, 2000.
Materials and Methods
Seventeen consecutive women referred to our unit for possible fetal cardiac arrhythmia and whose fetuses were found in sinus rhythm were included in this study. All gave informed consent to participate. Enrollment covered a 5-month period (August 1997 to January 1998). The mean maternal age was 30 years (range 24–39). Gestational ages varied from 20 to 34 weeks (mean 25 weeks). Morphologically, fetal hearts were normal in all subjects. Attempts were made by the same ultrasonographer (FP) to make the following tracings from all fetuses. Using the M-mode approach, atrial and ventricular free wall movements were recorded simultaneously from a four-chamber oblique view. It was difficult to identify the beginning of atrial and ventricular free wall contractions, so the peaks of the contraction waves corresponding to the smallest ventricular transverse diameters were taken as points of reference (Figure 1). With a five-chamber view, attempts were made to record movements of the aortic valve and the posterior wall of the left atrium (Figure 2). The a wave corresponded to the peak of atrial contraction, and the opening of the aortic valve was taken as the marker of the beginning of ventricular contraction. On the same five-chamber view, aortic and mitral valve movements also were recorded simultaneously. The timing of atrial contractions was given by late diastolic displacement of the posterior mitral valve leaflet.
Using the Doppler technique, flow-velocity waveforms through the mitral valve were recorded with ejection waves toward the aorta by placing the sample volume in the lower part of the outflow tract of the left ventricle. On these tracings, atrial contraction corresponded to the start of the Doppler a wave, and ventricular contraction to the beginning of the ventricular ejection wave going in the opposite direction (Figure 3). After taking a real-time picture of the four chambers of the heart in a vertical position, a 90° rotation of the transducer allowed a sagittal view of the superior vena cava and aorta closely related to each other. Flow velocity through those two vessels was recorded simultaneously by widening the Doppler sample volume (Figure 4). With that approach, the atrioventricular interval was calculated from the beginning of the retrograde venous a wave to the beginning of the aortic ejection wave.
To evaluate the ease with which an adequate tracing could be recorded, a maximum of five minutes was allowed to find the ideal fetal position needed for each ultrasonographic approach. The success rate was calculated as the proportion of attempted examinations that allowed detection. Differences between the mean values obtained with each approach were analyzed by analysis of variance. Results of each observer were considered individually in those analyses. A paired t test ruled out systematic differences between observers for each approach. Intra- and interobserver reliabilities were estimated using intraclass correlation coefficients with a one-way random effect model. The 95% confidence interval, F statistic, and significance value were calculated for each coefficient. The level of significance was set at P < .05. Analyses were done with SPSS version 8.0 (SPSS Inc., Chicago, IL).
No M-mode recordings of simultaneous mitral and aortic valve movements could be made within the 5-minute limit imposed by the protocol. Atrial free wall and aortic valve movements could be recorded in only six of 17 women (35%). Those two approaches were excluded from further analysis. Recordings of atrial and ventricular contractions by M-mode, of Doppler blood-flow velocities in and out of the left ventricle, and velocity waveforms within the aorta and superior vena cava could be made in all women within the time limit.
The mean and standard deviations of measurements taken by the two observers are shown in Table 1. No significant difference was found between observers, ruling out systematic errors caused by differences in reading techniques. Significant differences were found between the three approaches for both observers and for both measurements (atrioventricular and ventriculoatrial). M-mode values for atrioventricular intervals were significantly higher than Doppler values. Values from left ventricular outflow also were significantly higher than superior vena cava–aorta values. An opposite distribution was found for the ventriculoatrial intervals.
Data on reliability of the measurements taken with the three ultrasonographic approaches are given in Table 2. The intraobserver reliability was excellent for all three ultrasonographic approaches (intraclass correlation coefficients ≥ 0.89). Interobserver results showed lower intraclass correlation coefficients values for M-mode (atrioventricular 0.87, ventriculoatrial 0.79) than for left ventricular outflow (atrioventricular 0.93, ventriculoatrial 0.97) and superior vena cava–aorta (atrioventricular 0.98, ventriculoatrial 0.99).
Transmaternal fetal electrocardiograms, still based on an average QRS complex, do not yet allow detailed analysis of individual cardiac cycles. Evaluation of the chronology of fetal atrial and ventricular contractions routinely is achieved indirectly by ultrasonography. Such measurements are an essential part of the surveillance of fetuses at risk of atrioventricular block with maternal anti-Ro antibodies. In postnatal life, precise determination of atrioventricular relationship in cases of supraventricular tachycardia is a major element on which the choice of an appropriate antiarrhythmic agent is based.7 The same is true in fetal life.3 In the present study, M-mode and Doppler investigations had excellent detection rates of reliable tracings suitable for timing fetal atrioventricular contractions. However, M-mode, which is widely used for that purpose in fetal cardiology, had greater interobserver variability. M-mode–derived atrioventricular intervals were greater than Doppler measured intervals. Those observations are likely related to the fact that the reference point for ventricular contraction was conveniently taken at the smallest ventricular transverse diameter, which actually corresponds to the end of the mechanical event; the exact points at which myocardial fibers start or stop to shorten are difficult to identify on fetal M-mode tracings. Another systematic source of variation, this one common to all ultrasonographic approaches, comes from the fact that isometric contraction times and electromechanical delays are included in atrioventricular intervals measured, causing them to be constantly longer than the PR interval of electrocardiograms.6
Compared with fetal M-mode, Doppler tracings usually have better resolution and sharper contours, allowing easy identification of beginnings and ends of hemodynamic events. Values measured with the Doppler technique have been shown experimentally to be closely related to the PR and RP intervals of surface electrocardiograms.6 Unfortunately, the left ventricular outflow approach can be used only when the rhythm is sinusal with a heart rate not exceeding 160–170 beats per minute; tachycardia causes an overlap of the e and a waves of flow-velocity waveforms through the mitral valve. That approach is still valid to rule out first- and second-degree blocks in fetuses with heart rates within normal range.
In this study, we took advantage of the close relationship between the aorta and superior vena cava and of the fluid-filled fetal lungs, which allowed easy ultrasonographic access to those two vessels. Blood-flow velocity within the vena cava close to the heart is normally influenced by mechanical events of the cardiac cycle. Two forward waves are observed, the first corresponding to atrial filling during ventricular systole and the second to ventricular inflow of blood during the early part of diastole; those two forward waves are followed by a small reverse wave caused by atrial contraction that occurs late in diastole. Widening of the Doppler sample volume allowed simultaneous recording of the reverse venous a wave and the ejection wave through the aorta (Figure 4). The fact that the venous a wave remains visible regardless of heart rate makes that approach highly useful. The simultaneous Doppler velocimetry of inferior vena cava and abdominal aorta was advocated previously as an alternative method of investigation of fetal arrhythmia.8 We elected not to include that approach in our study because of a diastolic forward flow velocity component in the abdominal fetal aorta, which renders the identification of the retrograde venous a wave and precise measurements of atrioventricular intervals difficult, if not impossible.
Actual data from the different ultrasonographic approaches are not identical, so follow-up studies on the same fetus or comparative investigations between fetuses should always be done with the same approach.
1. Allan LD, Crawford DC, Anderson RH, Tynan M. Evaluation and treatment of fetal arrhythmias. Clin Cardiol 1984;7:467–73.
2. Kleinman CS, Copel JA, Weinstein EM, Santulli TV Jr, Hobbins JC. In utero diagnosis and treatment of fetal supraventricular tachycardia. Semin Perinatol 1985;9:113–29.
3. Jaeggi E, Fouron JC, Fournier A, van Doesburg NH, Drblik SP, Proulx F. Ventriculoatrial time interval measured on M-mode echocardiography: A determining element in the diagnosis, treatment and prognosis of fetal supraventricular tachycardia. Heart 1998;79:582–7.
4. Strasburger JF, Huhta JC, Carpenter RJ Jr, Garson A Jr, Mc-Namara DG. Doppler echocardiography in the diagnosis and management of persistent fetal arrhythmias. J Am Coll Cardiol 1986;7:1386–91.
5. Reed KL, Appleton CP, Anderson CF, Shenker L, Sahn DJ. Doppler studies of vena cava flows in human fetuses. Insights into normal and abnormal cardiac physiology. Circulation 1990;81:498–505.
6. Dancea A, Fouron JC, Miro′ J, Skoll A, Lessard M. Correlation between electrocardiographic and ultrasonographic time-interval measurements in the fetal lamb heart. Pediatr Res 2000;47:324–8.
7. Pieper SJ, Stanton MS. Narrow QRS complex tachycardias. Mayo Clin Proc 1995;70:371–5.
8. Chan FY, Woo SK, Ghosh A, Tang M, Lam C. Prenatal diagnosis of congenital fetal arrhythmias by simultaneous pulsed Doppler velocimetry of the fetal abdominal aorta and inferior vena cava. Obstet Gynecol 1990;76:200–4.