Pediatric Cardiology

, Volume 32, Issue 6, pp 754–758

Predictors of ECMO Support in Infants with Tachycardia-Induced Cardiomyopathy


  • Jack Christian Salerno
    • Division of Pediatric CardiologySeattle Children’s Hospital
  • Stephen Paul Seslar
    • Division of Pediatric CardiologySeattle Children’s Hospital
  • Terrence Ung Hoong Chun
    • Division of Pediatric CardiologySeattle Children’s Hospital
  • Mina Vafaeezadeh
    • Division of Pediatric Cardiac SurgerySeattle Children’s Hospital
  • Andrea Rae Parrish
    • Division of Pediatric CardiologySeattle Children’s Hospital
  • Lester Cal Permut
    • Division of Pediatric Cardiac SurgerySeattle Children’s Hospital
  • Gordon Alan Cohen
    • Division of Pediatric Cardiac SurgerySeattle Children’s Hospital
    • Division of Pediatric Cardiac SurgerySeattle Children’s Hospital
Original Article

DOI: 10.1007/s00246-011-9961-4

Cite this article as:
Salerno, J.C., Seslar, S.P., Chun, T.U.H. et al. Pediatr Cardiol (2011) 32: 754. doi:10.1007/s00246-011-9961-4


The development of tachycardia-induced cardiomyopathy (TIC) is related to the rate and duration of supraventricular tachycardia (SVT). Infants may be more susceptible to TIC because early symptoms might be unrecognized. Extracorporeal membrane oxygenation (ECMO) may improve outcome in patients with SVT and TIC; however, clinical predictors of infants who require ECMO support have not been determined. The purpose of this study was to identify predictors of the need for ECMO in infants with SVT and TIC. Sixteen infants <6 months of age who experienced resolution of TIC following control of arrhythmia were identified. Three patients (19%) required ECMO support. Comparisons were made between patients who required ECMO and those who did not. The groups were similar with respect to age at presentation, type of SVT, rate of SVT, and degree of ventricular dysfunction. However, patients requiring ECMO had increased median M-mode-derived left ventricular end diastolic dimension (LVED) z-score when compared to the medically managed patents (+2.8 vs. 0.0, P = 0.009). No patient in the medically managed group had an LVED z-score >2.3. Infants presenting with SVT and TIC with LVED z-score >2 are at increased risk for requiring ECMO support and early use of ECMO should be considered.




Tachycardia-induced cardiomyopathy (TIC) is a known complication of supraventricular tachycardia (SVT) and is characterized by ventricular dysfunction with clinical manifestations of heart failure that are reversible with normalization of heart rate. The development of TIC is related to the rate and the duration of SVT [3, 17]. In newborns, reentry SVT is the commonest cause of TIC. It is well known that infants are more likely to present with congestive heart failure than older patients [6] due to delay in recognition, which might place them at greater risk of an adverse outcome. The prognosis and recovery of cardiac function after achieving control of SVT is excellent. Although medical management is the mainstay of arrhythmia treatment in pediatric patients, in some cases pharmacologic therapy might not be effective.

Several centers have reported the use of extracorporeal membrane oxygenation (ECMO) to support patients with refractory, life-threatening arrhythmias in neonates [16] and older patients [4, 15, 18]. ECMO facilitates myocardial recovery by establishing a period of hemodynamic stability, during which antiarrhythmic medication can be optimized [16]. Although ECMO-related neurologic complications occur in as many as 11% of neonates [5], there is evidence that early initiation of mechanical cardiac support can improve survival in patients with reversible ventricular dysfunction [2, 15]. Although life-threatening arrhythmia is an uncommon indication for initiating mechanical cardiopulmonary support, ECMO might provide an effective therapeutic option in selected patients in whom conventional therapy fails. The purpose of this study was to determine echocardiographic predictors of the need for ECMO in infants with TIC.


The Seattle Children’s Hospital institutional review board approved this retrospective review. Using medical records, all infants <6 months of age who were hospitalized with SVT and TIC between November 2002 and February 2008 were identified. Patients with structural heart disease or previous cardiac surgery were excluded from analysis. TIC was defined as systolic ventricular dysfunction (shortening fraction of <30%) associated with sustained or incessant tachycardia in patients with otherwise structurally normal hearts. Shortening fraction calculations were based on echocardiographic data obtained on admission prior to pharmacologic treatment. Data collected at the time of admission included severity of clinical presentation, age, arterial blood gas results, mechanism of SVT, antiarrhythmic therapy, and echocardiographic data. One patient with TIC and a single coronary artery that coursed between the great vessels was excluded from the analysis.

The first line of treatment for infants with SVT at our institution is a beta-blocker. When tachycardia is incessant, we frequently initiate an intravenous beta-blocker. If only partial control of tachycardia is obtained with adequate beta-blockade, we add flecainide or sotalol as second-line therapy. When there is evidence of significant concomitant cardiomyopathy or hemodynamic compromise, amiodarone is used. ECMO support was initiated when patients presented in or developed evidence of cardiogenic shock and were unable to maintain adequate cardiac output. Venoarterial ECMO was initiated utilizing right internal jugular and common carotid cannulation. ECMO support was discontinued when TIC resolved or adequate control of arrhythmia was achieved.

Data are presented as median (range), median (interquartile range), or number (percentage). Statistical analysis of categorical variables was conducted with the Fisher’s exact test. Tests of association involving quantitative numerical measurements were conducted with the Mann–Whitney U-test. Analyses were performed using STATA version 10.1, with P values <0.05 considered to be statistically significant. All reported P values are two-sided.


During the study period, 192 infants were admitted and treated for symptomatic dysrhythmia. Sixteen (8%) of these patients met study criteria and were included in analyses. Median age at admission for the study group was 21 days (range: 0–133 days). Tachycardia (n = 6; 38%) and respiratory distress (n = 5; 31%) were the most common presenting symptoms (Table 1). During the same period, 152 patients received ECMO support at our institution. Three of the 16 (19%) study patients required ECMO support, representing 2% of overall ECMO patients. One patient required ECMO support immediately after presentation following cardiopulmonary resuscitation. Another was placed on ECMO for intractable hypotension while being loaded with amiodarone. In the final patient, ECMO support was initiated 50 h after admission for persistent acidosis and inadequate control of atrial dysrhythmia despite amiodarone and beta-blocker therapy.
Table 1

Patient characteristics

Baseline characteristics (n = 16)

 Age (days)

21 (0–133)

 Weight (kg)

4.0 (2.4–9.3)

 Male gender

12 (75%)

 Admission pH

7.28 (7.0–7.4)

Presenting features


6 (38%)

 Respiratory distress

5 (31%)

 Poor feeding

4 (25%)

 Cardiac arrest

1 (6%)

Presenting diagnosis

 Reentrant supraventricular tachycardia

12 (75%)

 Automatic atrial tachycardia

3 (19%)

 Reentrant supraventricular tachycardia aa + automatic atrial tachycardia

1 (6%)

 Maximum heart rate

280 (170–300)

 Left ventricular shortening fraction (%)

22 (3.0–31.2)

 LVED z-score

+0.7 (−2.6 to +4.3)

Medication required for treatment

 Beta blocker

10 (63%)


1 (6%)


1 (6%)

 Amiodarone + beta-blocker

4 (25%)

 Length of hospitalization (days)

4.5 (1–30)

Study patients were divided into two groups for comparison: those requiring ECMO (n = 3) and those not requiring ECMO (n = 13). The groups were similar with respect to age and weight at presentation, maximum rate of SVT, and degree of systolic ventricular dysfunction (Table 2). Univariate analysis of clinical variables collected at time of admission identified increased left ventricular end diastolic dimension (LVED) z-score [14] as a statistically significant risk factor for the need of ECMO support (+2.8 vs. +0.1; P = 0.023; Fig. 1).
Table 2

Comparison of patients requiring ECMO versus those who did not require ECMO


No ECMO (n = 13)

ECMO (n = 3)


Age (days)

29 (1–77)

16 (0–133)


Weight (kg)

4.3 (2.4–6.8)

3.6 (2.4–9.3)


Male gender

9 (69%)

3 (100%)


Intubated prior to antiarrhythmics

3 (23%)

2 (67%)


Reentrant supraventricular tachycardia

11 (85%)

1 (33%)


Automatic atrial tachycardia

2 (15%)

1 (33%)

Reentrant supraventricular tachycardia + automatic atrial tachycardia

0 (0%)

1 (33%)

Maximum heart rate

280 (175–300)

270 (200–300)


Left ventricular shortening fraction (%)

23 (9.5–31.2)

16.1 (3.0–22.0)


LVED z-scorea

0.0 (−2.6 to +2.3)

2.8 (+2.5 to +4.3)


Inotropic therapy prior to antiarrhythmics

2 (15%)

1 (33%)


Wolff–Parkinson–White syndrome

1 (8%)

0 (0%)


Length of hospitalization (days)

3 (1–30)

23 (21–23)


aLeft ventricular end diastolic diameter
Fig. 1

Patients with TIC who required ECMO support had significantly higher LVED z-scores than those who did not require support. All patients with LVED z-score >2.3 required ECMO support

The average duration of ECMO support was 125 h (76–160 h). No patients experienced ECMO-related complications. All patients survived. The median length of hospitalization was 4.5 days (23 days for ECMO patients vs. 3 days for non-ECMO patients; P = 0.025). Resolution of TIC following control of underlying arrhythmia was observed in all patients.


The overall incidence of TIC has been reported to be as high as 10% in older patients being treated for focal atrial tachycardia, with cardiomyopathy occurring more frequently in patients exhibiting slower incessant tachycardia [9]. Although few published studies examined TIC in neonates, one group reported that 48% of neonates with SVT at their institution presented with evidence of TIC [7]. The overall mortality of this group of patients was 6%. The immature myocardium has a relatively high noncontractile content, a less mature sympathetic system, underdeveloped intracellular calcium regulatory mechanisms, and reduced functional reserve [12]. As a result, the immature heart tends to be less responsive to cardiotonic medications and might be more susceptible to TIC. The treatment of incessant SVT can be particularly challenging in the setting of compromised ventricular function. Although amiodarone has proven efficacy in children with incessant SVT, a recent randomized trial involving intravenous amiodarone demonstrated that adverse reactions, including hypotension, bradycardia, and atrioventricular block, are common [11]. Even though similar concerns exist with the use of beta-blockers in this patient population, it has been shown that the use of short-acting intravenous beta-blockers is associated with a low rate of clinically significant hypotension in children [1]. This report is similar to our clinical experience. As a result, we frequently use esmolol as initial therapy for incessant SVT and progress to amiodarone when there is evidence of hemodynamic compromise, such as hypotension or end-organ dysfunction.

Although initially developed for patients with severe respiratory failure, ECMO has been increasingly used to provide mechanical support for patients with predominately myocardial failure. During the past 20 years, ECMO has been used to provide mechanical cardiopulmonary support to more than 4,600 pediatric patients with life-threatening heart failure according to the Extracorporeal Life Support Organization registry [5]. Of these, 240 neonates and infants received ECMO support for cardiomyopathy of various etiologies, with a reported survival rate of 59%. However, the overall survival rate for infants in cardiogenic shock is only 38%, underscoring the impact severity of illness plays in ECMO survival. Additionally, survival is only 24% in neonates who experience cardiac arrest prior to initiation of ECMO support. Likelihood of ventricular recovery appears to improve with earlier initiation of ECMO support [2] before irreversible myocardial and other end-organ damage is sustained. Because ECMO-related neurologic complications occur in as many as 11% of neonates [5], appropriate determination of timing of support and patient selection is of paramount importance. In a canine model of TIC, left ventricular ejection fraction shows significant recovery within 24–48 h of return to normal rhythm and normalizes within 1–2 weeks [10]. Consistent with this finding, all infants in our study experienced normalization of systolic dysfunction with rhythm control, supporting an underlying diagnosis of TIC.

In our patient population, left ventricular dilation (LVED z-score >2) predicted the need for ECMO support, with a positive predictive value of 74% and a negative predictive value of 100%. In contrast to the inherent difficulties in objectively measuring left ventricular systolic function during SVT, measurement of left ventricular dimension is relatively straightforward and reproducible. The significance of left ventricular dilation predicting need for ECMO is consistent with the finding that the severity of LV dilation at the time of listing for heart transplantation is associated with outcome in infants with dilated cardiomyopathy, whereas severity of LV systolic dysfunction is not [13].

By defining clinically useful predictors such as increased LVED, we believe that we can identify patients at highest risk for hemodynamic collapse who would benefit from the use of ECMO as a bridge to recovery. We speculate that ECMO support may facilitate the use of necessary antiarrhythmic therapy due to its ability to support cardiac output. This is particularly the case with the use of amiodarone, which is known for its efficacy but might be limited by its detrimental hemodynamic effects [11]. The use of ECMO to successfully rescue a neonate with amiodarone-related cardiovascular collapse has been reported [8].

The small number of patients included in this retrospective study reflects the relatively rare nature of TIC in pediatric patients. Conclusions based on the data presented must be interpreted with this important limitation in mind. Undoubtedly, a larger, multi-institutional study would provide additional information that might be useful in identifying patients at greatest risk for requiring ECMO support due to TIC-related hemodynamic instability.


In this retrospective study of TIC, patients with normal or minimal left ventricular dilation were successfully managed with medical therapy alone, whereas those with increased left ventricular dilation required ECMO support. Although the study was limited to hospitalized patients, whereas most clinical management of patients with SVT occurs in the outpatient setting, the findings suggest that ECMO support should be available in centers that manage infants with arrhythmias. The early use of ECMO should be considered when treating infants with TIC and left ventricular dilation due to the possibility of hemodynamic collapse, which can be potentiated by antiarrhythmic therapy.


This work was partially funded by an educational grant from the Robert B. McMillen Foundation, Issaquah, Washington.

Copyright information

© Springer Science+Business Media, LLC 2011