Fetal Atrial Flutter: Electrophysiology and Associations With Rhythms Involving an Accessory Pathway.

BACKGROUND: Atrial flutter (AFl) accounts for up to one third of all fetal tachyarrhythmias and can result in premature delivery, hydrops, and fetal death in 10% of cases; however, the electrophysiology of AFl in utero is virtually unstudied. METHODS AND
RESULTS: In this observational study, we reviewed 19 fetal magnetocardiography studies from 16 fetuses: 15 fetuses (21-38 weeks' gestation) referred with an echocardiographic diagnosis of AFl and 1 fetus (20 weeks' gestation) referred with a diagnosis of tachycardia that was shown by fetal magnetocardiography to have transient AFl in addition to atrioventricular reciprocating tachycardia. Thirteen fetuses showed AFl during the fetal magnetocardiography session, including 4 that presented prior to the third trimester. Five fetuses had incessant AFl; all but 1 of the others with AFl showed additional significant rhythms. Specifically, AFl showed a strong association with rhythms involving an accessory pathway: atrioventricular reciprocating tachycardia, blocked reentrant premature atrial contractions, and ventricular preexcitation. The observed initiations and terminations of AFl most often involved reentrant premature atrial contractions. Spontaneous termination of AFl showed AFl cycle length oscillations. Nine fetuses with 2:1 AFl also showed periods of 4:1 conduction or variable conduction that oscillated between 2:1 and 4:1; however, 3:1 AFl was relatively rare.
CONCLUSIONS: Fetal AFl can occur as early as midgestation and is often accompanied by atrioventricular reciprocating tachycardia and other rhythms associated with an accessory pathway. The findings depict critical differences in the electrophysiology of AFl in the fetus versus the neonate.

Introduction

Fetal tachyarrhythmia is an uncommon condition that occurs in 0.4% to 0.6% of all pregnancies.1 Atrial flutter (AFl) accounts for 26% to 29% of all fetal tachyarrhythmias2, 3 and is defined as a rapid regular atrial rate of 300 to 600/min, accompanied by variable atrioventricular conduction.2 AFl can occur with structurally normal hearts or with congenital heart disease, including atrioventricular septal defect, Ebstein's malformation, hypoplastic left heart syndrome, and pulmonary atresia.4, 5, 6 Although the incidence of hydrops fetalis is similar in sustained AFl and atrioventricular reciprocating tachycardia (AVRT),2 the overall mortality of fetal AFl approaches 10% and is higher than that of AVRT, perhaps due to the higher incidence of congenital heart disease.2

Due to the difficulty of recording the fetal ECG, the electrophysiology of AFl in utero has not been investigated, except in small case studies. In this retrospective study, we utilized fetal magnetocardiography (fMCG), the magnetic analog of electrocardiography, to characterize the heart rate and rhythm patterns of fetuses presenting with AFl. We demonstrate a remarkably high incidence of AVRT and other rhythms associated with an accessory pathway.

Methods

The study cohort comprised pregnant women referred with a diagnosis of fetal AFl to the Biomagnetism Laboratories at the Department of Medical Physics, University of Wisconsin‐Madison from 2002 to 2015. We also included 1 case in which the referral diagnosis was AVRT but AFl was also seen during the fMCG session. The study was approved by the University of Wisconsin Health Sciences Institutional Review Board. Informed consent was obtained from each participant.

The fMCG was recorded using a 37‐channel (Magnes; 4D Neuroimaging, Inc., San Diego, CA) or a 21‐channel (Model 624; Tristan Technologies, San Diego, CA) superconducting quantum interference device (SQUID) biomagnetometer, housed in a magnetically shielded room. The fMCG was recorded in 10‐minute segments with a total recording time ranging from 20 to 100 minutes. Signal processing was used to remove maternal interference.

AFl was defined by a nearly constant atrial rate of >300/min, variable atrioventricular conduction with ratio >1:1, and abrupt onset and termination. AVRT was defined by heart rate >200/min with low baseline variability, 1:1 atrioventricular conduction, and abrupt onset and termination. We documented the cycle lengths and the percent time in AFl, AVRT, and the other observed rhythms.

We measured the PR, QRS, and QTc intervals in sinus rhythm, the PR, RP, and QRS intervals in AVRT, and the QRS interval during AFl. The QTc interval during tachycardia was difficult to measure due to overlap of the P and T waves. Cardiac time intervals were compared with those of normal fetuses from a reference database. Measurements exceeding the 95% prediction interval were considered to be prolonged.

Actocardiography was used to characterize the fetal heart rate patterns and to assess the effects of fetal movement on heart rate, rhythm, and conduction. Using autocorrelation to detect fetal QRS complexes, ventricular heart rate tracings were computed from the RR intervals, and actograms (tracings of fetal activity derived from movement‐related changes in signal amplitude) were derived from the instantaneous QRS amplitudes.7 Atrial heart rate tracings were also computed in subjects with large P‐waves.

Results

Nineteen fMCGs were performed on 16 singleton fetuses at 20 to 38 (mean 28) weeks’ gestation (Table). One fetus (#16) had Ebstein's anomaly; 1 had an HCN4 channelopathy (#12) with bradycardia; all others had structurally normal hearts. Nine fetuses were on medication at the time of fMCG (Table).

Table 1

Summary of fMCG Results

Fetus #	Gestational Age (wks)	Medication	Rhythm	Percent Time	RR (ms)	PR (ms)	QRS (ms)	QTc (ms)	Duration (s)
1	24	Digoxin, amiodarone	AFl 2:1	100	296	108	40	287	1200
2a	28	—	AFl 2:1	94	269	98	56	301	2400
			AFl Var	6	280/472
2b	29	Digoxin, metoprolol	AFl 2:1	38	257	111	63	276	3000
			AFl Var	2	279/480
			SR	60	407	94	61	395
2c	32	Digoxin, metoprolol	SR	100	457	94	52	441	2400
3	26‐6/7	—	AFl 2:1	74	269	73	60	366	3000
			AFl 4:1	6	530
			AFl Var	20	273/502
4	30	Digoxin	AFl 2:1	96	308	79	50	377	3000
			AFl 3:1	<1	432
			AFl Var	3	293/488
5	35‐5/7	—	AFl 2:1	48	261	94	54	417	3600
			AFl 4:1	428	504
			AFl Var	48	274/480
6	32‐5/7	Digoxin	AFl 2:1	47	282	39	36	369	2400
			AFl 4:1	<1	452
			AFl Var	18	283/469
			SR	34	440	99	40	501
			AVRT	<1	221	94	36
7	30‐2/7	—	AFl 2:1	<1	301	81	44	414	4800
			SR	99	430	96	42	442
8	35‐3/7	Digoxin, Synthroid	AFl 2:1	82	288	63	48	382	3100
		(levothyroxine sodium)	AFl 4:1	<1	531
			AFl Var	8	291/550
			BAT	5	494/781
			SR	4	401	109	61	346
9	36‐5/7	—	AFl 2:1	55	255	98	38	305	6000
			AFl 3:1	<1
			AFl 4:1	4	478
			AFl Var	15	279/504
			BAT	3	469/761
			AVRT	25	221	127	44	436
10	24‐3/7	Digoxin, sotalol, amiodarone	AFl 2:1	7	273	50	48	392	2400
			AFl 4:1	11	488
			BAB,BAT	70	741 500/800
			AVRT	14	238	111	46	275
11	21‐5/7	Digoxin	AFl 2:1	6	280	111	83	486	4900
			AFl 4:1	<1	513
			BAT	22	458/769
			SR	70	424	98	56	419
			AVRT	<1	284	159	54	342
12	36‐2/7	—	AFl 2:1	2	296	78	38	349	3000
			SR	98	569	116	48	533
13a	20	Digoxin	AFl 2:1	1	230	86	46	444	6000
			BAT	7	449/772
			SR	74	520	106	61	416
			AVRT	18	232	119	52	430
13b	20‐4/7	Digoxin	SR	29	440	108	65	425	6000
			AVRT	71	253	150	58	443
14	31‐4/7	—	BAT	2	453/786				2400
			CAB	11	269/448
			SR	87	422	108	48	417
15	29‐1/7	Digoxin, magnesium	SR	100	444	100	54	473	4800
16			SR	100	505	182	94	449	5720

Fetal magnetocardiography (fMCG) results of 16 fetuses studied in 19 sessions, listing for each session the observed rhythms with the percent time present, cycle lengths (RR), and waveform interval measurements. The total duration of the fMCG recordings is shown in the last column. The fetuses are ordered by total percent time in AFl during the first session. Serial sessions for fetuses #2 and #13 are listed consecutively in chronological order with the different session indicated by a suffix (a, b, c). Fetus #16 had Ebstein's anomaly. AFl Var indicates atrial flutter with variable AV conduction with RR interval oscillating between the values shown in column 6; AVRT, atrioventricular reciprocating tachycardia; BAB, blocked atrial bigeminy due to reentrant premature atrial contractions (PACs); BAT, blocked atrial trigeminy due to reentrant PACs with oscillating RR interval (column 6); CAB, conducted atrial bigeminy due to conducted PACs from an automatic focus with oscillating RR interval (column 6); SR, sinus rhythm.

The observed rhythms and their prevalence are compiled in Table. The rhythms included sinus rhythm, AFl, AVRT, and complex atrial ectopy. AFl showed conduction ratios of 2:1, 3:1, and 4:1, and variable conduction in which the ratio cycled between 2:1 and 4:1 in a regular pattern. Complex atrial ectopy showed 2 forms. The most common was premature atrial contractions (PACs) with fixed coupling interval, presumed to be reentrant PACs, which resulted in atrial bigeminy/trigeminy or couplets. The other was conducted PACs with a longer, variable coupling interval, presumed to arise from an automatic focus, which resulted in atrial bigeminy.

Two fetuses (#15 and #16) referred with a diagnosis of AFl showed only sinus rhythm during the fMCG study and another (#14) showed frequent ectopy but no tachycardia. The remaining 13 showed AFl. Two fetuses (#2 and #13) were studied in multiple sessions. An interesting and notable finding is that 4 of 14 fetuses (29%) presented with AFl during the second trimester, including 2 that had brief periods of AFl prior to 22 weeks gestation (#13 and #11).

Nine fetuses showed periods of 4:1 conduction (Figure 1B) and/or variable conduction that oscillated between 2:1 and 4:1 (Figures 1C and 3B). In all 7 fetuses with sustained 2:1 and 4:1 AFl, the mean RR interval in 4:1 AFl was less than twice the RR interval in 2:1 AFl, implying that the AFl rate was faster during 4:1 than 2:1 conduction.

Figure 1

Notable rhythm patterns. A, Short burst of AFl with variable conduction in fetus #13, the youngest fetus in the cohort, studied at 20 weeks. The last several beats show reentrant PACs. B, Transition from 4:1 AFl to 2:1 AFl with variable preexcitation in fetus #11 at 21‐5/7 weeks. The first beat of 2:1 AFl is aberrantly conducted. The flutter waves are obscure after the transition; however, they are visible at the termination of AFl, shown in Figure 2E, which occurs ≈60 seconds later. C, AFl with variable conduction ratio in fetus #5, showing Wenckebach‐like phenomenon and fractionation of the flutter wave. The RR intervals are slightly longer than twice the PP intervals, and the PR intervals progressively increase prior to the nonconducted beat, which results in an effective conduction ratio of 8:3. Variable preexcitation is present. D, Trigeminal rhythm involving paroxysms of AFl‐like rhythm initiated and terminated by nonconducted reentrant PACs in fetus #9. The coupling time from the initiating PAC (asterisks) to the first AFl wave is longer than the flutter cycle length. E, AVRT showing ST depression (inverted projection) and QRS/T discordance in fetus #10 at 24‐3/7 weeks. The T‐wave showed cyclical amplitude and morphology variations of uncertain origin with a period of ≈3 seconds. The termination of this episode of AVRT is shown in Figure 2B. AFl indicates atrial flutter; AVRT, atrioventricular reciprocating tachycardia; PACs, premature atrial contractions.

Figure 2

Initiation and termination of fetal AFI. A, Initiation of 2:1 atrial flutter by a conducted PAC (asterisk), probably reentrant due to negative polarity, with PR prolongation in fetus #11. The first few flutter waves (arrows) are visible due to the variable RR interval at onset, but thereafter the QRS complex substantially overlaps every other flutter wave. The sixth beat has the shortest RR and is aberrantly conducted; however, aberrancy was less common and less pronounced in AFl than in AVRT. B, Initiation of AFl with the beat immediately following AVRT in fetus #10. A reentrant PAC (asterisk) initiates AFl (arrows). Given the pause following the termination of AVRT, the slow AV conduction is somewhat surprising. The pause and slow AV conduction could be due to autonomic activity. C, Initiation of 4:1 atrial flutter during sinus rhythm in fetus #11. During the time between the last sinus P‐wave (asterisk) and the first regular flutter waves (arrows), the atrial rhythm is rapid and irregular with no P‐wave, suggesting that AFl is initiated by a fibrillation‐like rhythm. Notice that the RR intervals are relatively uniform throughout the transition. D, Termination of atrial flutter by AVRT in fetus #9. A modest, but abrupt, cycle length shortening occurs at the onset of AVRT (arrow) with no break in tachycardia between the rhythms. Although the AFl rhythm is regular, the RR interval at termination shows a short‐long‐long (S‐L‐L) oscillation pattern, presumably due to changes in AV conduction. E, Termination of AFl with AFl short‐long (S‐L) cycle length oscillations in fetus #11. Variable degrees of preexcitation are seen during AFl. AFl indicates atrial flutter; AVRT, atrioventricular reciprocating tachycardia; PACs, premature atrial contractions.

AVRT was seen in 5 fetuses, comprising 38% (5 of 13) of those that also showed AFl during fMCG and 27% (4 of 15) of those with a dominant presentation of AFl at the time of referral. In 4 fetuses the RR interval was substantially shorter in AVRT (217–238 ms; mean 224.3 ms) than in 2:1 AFl (255–282 ms; mean 272.5 ms). In 1 fetus (#13) the RR intervals were nearly the same (232 ms versus 230 ms). This fetus was the youngest in the cohort (20 weeks) and had the highest flutter rate and lowest percent time in AFl of all fetuses that showed AFl. The rhythm patterns of AVRT observed here were compatible with those reported in previous fMCG studies.8

The most common form of atrial ectopy was blocked atrial trigeminy due to reentrant PACs. This was present in 6 of 16 (38%) fetuses, including 4 of 5 with AFl and AVRT, 1 of 8 with AFl alone, and 1 of 3 that did not show AFl during fMCG. One fetus with blocked atrial trigeminy (#10) also had blocked atrial bigeminy, which resulted in bradycardia. Three fetuses with blocked atrial trigeminy (#9, 10, and 14) showed blocked atrial couplets, which are relatively rare but have been seen previously in fetuses with blocked atrial bi/trigeminy.9 One subject (#14) with blocked atrial trigeminy also showed atrial bigeminy due to conducted PACs with a relatively long, variable coupling interval.

Variable ventricular preexcitation was seen in 4 fetuses (Figures 1B, 1C, 1D, and 2E): 2 with AFl and AVRT; 2 with AFl alone.

Initiation and Termination of AFl

In several fetuses with intermittent AFl it was possible to observe the mechanisms of initiation and termination. The great majority of the initiation patterns involved reentrant PACs, including the examples of transient AFl shown in Figure 1A and 1E. Sustained AFl was observed to initiate with reentrant PACs from sinus rhythm (Figure 2A), blocked atrial trigeminy, and immediately after a pause following termination of AVRT (Figure 2B). AFl was also seen to initiate with a rapid, irregular atrial rhythm, resembling fibrillation (Figure 2C). AFl was observed to terminate to AVRT and sinus rhythm. Unlike the transitions from AVRT to AFl, the transitions from AFl to AVRT typically showed no break in tachycardia (Figures 2D and 4). Spontaneous termination of AFl to sinus rhythm showed AFl cycle length oscillations (Figure 2E).

Cardiac Time Intervals and Waveform Morphology

The cardiac time intervals in sinus rhythm were normal, except for 1 fetus (#11) with shortened PR and prolonged QRS during intermittent ventricular preexcitation, 2 fetuses that showed modest QTc prolongation (#12 and #6), 1 fetus with Ebstein's anomaly (#16) that showed marked PR (180 ms) and QRS (85 ms) prolongation, and the HCN4 subject (#12) with sinus bradycardia. Three fetuses showed very large P‐waves compatible with atrial hypertrophy. Two fetuses (#5 and #2) showed fractionated flutter waves (Figure 1C).

The 5 subjects with AVRT showed RP/RR ratios in the range 0.27 to 0.56 (mean 0.46). Two showed QRS prolongation with bundle branch block during AVRT. One showed ST depression with QRS/T discordance (Figure 1E).

Fetal Actocardiograms

Five fetuses had incessant AFl. Of these, 1 (#1) showed only 2:1 conduction with nearly constant heart rate (Figure 3A). This was perhaps the sickest patient, with moderate ventricular dysfunction, short inflow Doppler, and moderate to severe AV valve regurgitation. The others showed at least some degree of variable conduction (Figure 3B). Fetal movement had little effect on the AFl rate, but could enhance AV conduction.

Figure 3

Actocardiograms in incessant fetal AFI. A, Fetus #1 had incessant 2:1 AFl. The heart rate, plotted below on an expanded scale, was nearly constant throughout, showing little variation with fetal movement (double‐headed arrow). B, Fetus #5 showed highly variable atrioventricular conduction. The occasional periods of 2:1 AFl were strongly associated with fetal movement. AFl indicates atrial flutter.

The most complex actocardiograms were seen in fetuses with diverse intermittent rhythms (Figure 4). The data in Figure 4A encompasses a period when a number of rhythms were present intermittently: AVRT, AFl with 2:1, 4:1, and variable conduction, and a trigeminal rhythm due to blocked atrial couplets. The different rhythms usually had distinct heart rates and/or heart rate patterns that allowed them to be distinguished. Occasionally, however, the heart rate patterns showed deviations that caused them to resemble those of other rhythms (Figure 4). The data in Figure 4B are notable for the pronounced beat‐to‐beat fetal heart rate variability in AFl and AVRT. The atrial rate in AFl was relatively constant, implying that the heart rate variability in AFl was due to marked changes in AV conduction. Previously, we have attributed similar heart rate oscillations in fetal tachycardia to the existence of dual AV pathways.8

Figure 4

Diversity and intermittency of the fetal rhythms depicted by actocardiograms. In (A and B) the three panel shows the atrial heart rate (top), ventricular heart rate (middle), and actogram (bottom). A, Fetus #10 showed complex heart rate patterns due to alternation between intermittent AVRT, AFl with 2:1, 4:1, and variable conduction, and a trigeminal rhythm involving blocked atrial couplets (BAC). AFl was easily identified by the high atrial rate (>400/min). The conduction ratio varied between 2:1 and 4:1, as indicated in the ventricular heart rate tracing. AVRT showed the same tachycardic heart rate in both the atrial and ventricular tracings. BAT gave rise to prominent heart rate oscillations. Notice that during the last episode of AFl the flutter rate was not constant. It started out high and gradually declined, which is more typical of AVRT than AFl. The conduction ratio was initially 4:1 and improved with decreasing flutter rate. This fetus was relatively inactive, and the rhythm transitions were not strongly associated with fetal movement. B, Fetus #9 showed pronounced beat‐to‐beat heart rate variability during both AFl and AVRT. The first 40 seconds of the tracings showed predominantly 2:1 AFl with occasional isolated slow beats corresponding to 4:1 conduction. Notice that the slow beats were relatively uniform in cycle length, compared to the slow beats during AVRT in the second half of the tracing. The high heart rate variability in AFl was due to an irregular short‐long‐long oscillation pattern. The flutter rate was constant, implying that the variability was due to changes in AV conduction. The high heart rate variability was consistently attenuated by fetal movement (asterisks). The heart rate variability in AVRT is due to a regular short‐long pattern in which the RP (VA) interval is constant but PR (AV) oscillates, again implying that the variability is due to changes in AV conduction. Usually, the periods of AVRT and 2:1 AFl could be discriminated based on the higher heart rate during AVRT; however, the episode of AFl preceding the transition from AFl to AVRT at 515 seconds (arrow) has relatively high heart rate and variability that makes it appear to be a resumption of the prior episode of AVRT. The onset of AVRT was associated with fetal movement, as has been described previously.8 AFl indicates atrial flutter; AVRT, atrioventricular reciprocating tachycardia; BAT, blocked atrial trigeminy.

Discussion

Our study is the first to comprehensively characterize the associations between AFl and rhythms involving an accessory pathway prior to birth. Intermittent AVRT was seen in 5 of 13 (38%) fetuses that showed AFl during the fMCG session. In addition to AVRT, our fetuses showed ventricular preexcitation and complex atrial ectopy due to reentrant PACs. Alternation of these rhythms with AFl resulted in complicated heart rate and rhythm patterns, underscoring the importance of the accessory pathway for providing a comprehensive explanation. Blocked atrial bi/trigeminy, including blocked atrial couplets, were seen in 6 of 16 (38%) fetuses, including 4 of 5 (75%) with AFl and AVRT. Thus, if AFl is noted to occur in conjunction with periods of complex atrial ectopy, the medical team should assess for the presence AVRT, given that its presence may influence therapy.

Most prior studies of fetal tachycardia have not reported on the co‐occurrence of AFl and AVRT or imply that the rate is low. For example, the large fetal tachycardia studies of Krapp and coworkers2 and Jaeggi and coworkers3 made no mention of co‐occurrence. van Engelen and coworkers6 reported that only 1 of 30 fetuses with AFl also showed episodes of AVRT. An early study of fetal tachycardia by Maxwell and coworkers10 involved 12 cases of AVRT, 8 of AFl, and 3 cases in which the rhythm varied between AVRT and AFl. The characteristics of the rhythms were not reported; however, the proportion of fetuses with AFl that showed AVRT alternating with AFl (27%) was similar to that in our study. Other reports of co‐occurrence have been largely confined to case studies.11

The association between AFl and accessory pathways was first highlighted by Till and Wren.12 In a cohort of 9 subjects with AFl in utero or at birth, 3 showed AVRT following dc cardioversion. In a postnatal transesophageal electrophysiologic study of 30 subjects with supraventricular tachycardia, Naheed and coworkers13 found that 22 had AVRT and 8 had AFl. Of the 8 with AFl, 5 (62.5%) had inducible AVRT. None, however, were noted to have spontaneous AVRT, and postnatal AVRT is generally uncommon in patients with prenatal AFl. Our study not only corroborates an association between AFl and accessory pathways in the fetus, but also demonstrates that accessory pathways in the fetus exhibit a greater propensity for spontaneous, natural conduction, compared to the neonate. This finding suggests that accessory pathways often become nonfunctional at late stages of fetal development.

The reason for the association between AFl and AVRT is unknown. Till and Wren12 noted that AVRT impairs cardiac function and may cause atrial dilatation, which may facilitate initiation and maintenance of AFl. Naheed and coworkers13 similarly speculated that simultaneous ventriculoatrial contraction, atrial distension, and functional atrioventricular valve incompetence from annular enlargement occur during AVRT, and may predispose the fetal or neonatal atrium to the development of intraatrial reentrant tachycardia. The fetuses in our study showed atrial rhythms with large P‐wave amplitude, fractionated flutter waves, and frequent atrial ectopy, which are suggestive of atrial dilatation and conduction impairment.

Nine fetuses with AFl (69%) showed periods of 4:1 AFl or AFl with variable conduction that oscillated between 2:1 and 4:1 AFl, whereas 3:1 AFl was relatively rare. A possible explanation for the rareness of 3:1 AFl is that the AV node may have 2 distinct levels of block, with the lower level having a lower conduction rate. In this circumstance, the overall conduction ratio will be a multiple of the conduction ratio of the upper level. If the conduction ratio of the upper level is 2:1, then the overall conduction ratio can be 2:1 or 4:1. A 3:1 conduction ratio is possible if the conduction ratio of the upper level is 1:1 or 3:2, but these are much less likely. This explanation is compatible with the observation that the heart rate in 4:1 AFl was slightly greater than half the rate in 2:1 AFl.

Little is known about the spontaneous initiation and termination patterns of AFl. Even postnatal data are scarce due to the rarity of AFl and its often incessant nature. In this study, AFl was observed to initiate with atrial ectopy, or due to AV reentry or a rapid, irregular rhythm, resembling fibrillation. The ability of reentrant PACs to initiate and terminate AFl was remarkable, and further supports the association between AFl and accessory pathways.12 A termination pattern characterized by AFl cycle length oscillations was also seen. This pattern was observed by Ortiz and coworkers in a canine model and was attributed to changes in conduction in an area of slow conduction.14 The intermittency of the rhythms and the abruptness of the transitions between them, often with little change in cycle length, undoubtedly contribute to the difficulty of detecting them using echocardiography.

Another notable observation was that 4 of 14 fetuses (29%) presented with AFl during the second trimester, including 2 that presented prior to 22 weeks. Others have reported that the initial presentation of AFl occurs mainly during the third trimester.3, 15 They speculated that the paucity of presentation at <30 weeks’ gestation is due to the inability of the small, immature atrium to maintain a continuous atrial macro‐reentrant circuit. The relatively early detection of AFl among subjects in this study suggests that it is important for the medical team to consider AFl as a potential mechanism in the fetus that presents with tachycardia at any gestational age.

This finding, however, cannot be attributed to our use of fMCG. All of the subjects were referred with a diagnosis of AFl, except for 1 subject at 20 weeks that was referred with a diagnosis of fetal tachycardia and showed AVRT with only a few brief periods of AFl.

Prolonged, continuous monitoring by fMCG can provide a more accurate evaluation of complex, intermittent rhythms, including the percent time spent in each rhythm. Also, assessment of waveform morphology by fMCG can ascertain the degree to which conduction occurs through the accessory pathway versus the AV node. In this study, variable preexcitation during AFl was seen in 4 fetuses; however, none showed a sustained wide‐QRS rhythm. Krapp and colleagues reported that digoxin was used as first‐line therapy in 67.6% of cases. Conversion to sinus rhythm was achieved in 32 of 71 cases (45.1%) with digoxin treatment alone.2, 4, 5, 16 Recently, sotalol has been recommended as first‐line therapy, as in several published series it has been most effective in restoring sinus rhythm, even in the hydropic fetus.1, 3, 17, 18 Our finding that an accessory pathway may be present in some fetuses with AFl suggests that if AFl occurs in conjunction with supraventricular tachycardia, implying the possibility of preexcitation, sotalol might be a better choice over digoxin for treatment. For more refractory AFl with hydrops, intramuscular digoxin and/or amiodarone can successfully restore sinus rhythm or slow the ventricular rate to improve hemodynamics.19, 20 Amiodarone has been shown to slow the fetal heart rate in AFl; however, the conversion rate is low. Treatment strategies for fetal arrhythmias are described in the American Heart Association's recent scientific statement on fetal cardiac disease.21

Study Limitations

The study was observational. Patients came from multiple centers across the United States, which made it difficult to obtain follow‐up information. Therapy and the timing of the studies with respect to therapy were not controlled, which limited our ability to assess the effects of therapy on rhythm. The referral pattern was not preselected, and it is possible that cases referred represented more complex rhythm patterns, where fMCG could potentially be complementary to echo diagnosis. The referral of patients was limited to those who were stable and could travel to the Biomagnetism Laboratory. This likely reduced the number of subjects close to term, when atrial flutter is often noted, as well as the sickest patients with prolonged inpatient stays. The lower signal‐to‐noise ratio of fetal MCG, compared to that of postnatal ECG, limited the resolution of the P and T waves in the raw tracings, especially at early gestational ages.

Conclusions

Fetal AFl can occur as early as midgestation and is often accompanied by AVRT and other rhythms associated with an accessory pathway. The study validates the concept that the electrophysiology of the fetus and neonate show important differences, and further demonstrates the efficacy of fMCG for precise assessment of fetal rhythm.

Sources of Funding

This research was supported by the National Institutes of Health (grant number R01 HL63174) and Friede Springer Herzstiftung, Pacelliallee 55, 14195 Berlin.

Disclosures

None.