Twin-reversed arterial perfusion sequence: pre- and postoperative cardiovascular findings in the ‘pump’ twin




To assess cardiovascular findings in twin-reversed arterial perfusion (TRAP) sequence pre- and post-therapy and compare these findings to traditional obstetric markers, defined as acardius to pump twin weight ratio and presence of polyhydramnios.


This was a retrospective review of 27 cases of TRAP sequence diagnosed between 2004 and 2008. Echocardiographic data included indexed cardiac output and functional and anatomic parameters. Ultrasound reports were reviewed for acardius to pump twin weight ratio and polyhydramnios. We assessed the relationship between cardiac output and the remaining cardiac/obstetric variables obtained pre- and post-treatment.


Twenty-three subjects had complete echocardiographic data sets at initial evaluation (mean gestational age, 20.4 ± 2.5 weeks) and, of these, post-treatment echocardiographic evaluation was available in 10. Six of seven (86%) pump twins with elevated indexed cardiac output had significant cardiovascular compromise. Most fetuses with abnormal cardiac output or right ventricular dysfunction normalized post-therapy. There was no relationship between cardiac output and obstetric markers.


Elevated indexed cardiac output is strongly associated with cardiovascular compromise. Traditional obstetric prognosticators do not correlate with cardiovascular derangements. In pump twins with cardiac compromise, postoperative cardiovascular status improves acutely. Given this analysis, we conclude that assessment of cardiovascular findings should be incorporated into the management and treatment of TRAP sequence. Copyright © 2009 ISUOG. Published by John Wiley & Sons, Ltd.


Twin-reversed arterial perfusion (TRAP) sequence is a rare and potentially lethal condition affecting approximately 1% of monozygotic twin gestations1. A normal twin, or ‘pump’ twin, perfuses its severely malformed cotwin with deoxygenated blood via retrograde flow in direct arterioarterial anastamoses between the umbilical arteries of each twin. The anomalous twin, or acardius, then returns further deoxygenated blood back to the pump twin through a direct venovenous anastamosis. Untreated, pump twin mortality is 50%1, usually owing to prematurity or high-output cardiac failure. Assessment of disease severity and prognosis in TRAP has historically been characterized by a number of sonographic indicators, including the acardius to pump twin weight (AC : PT) ratio or presence of polyhydramnios1. Others have described the use of umbilical artery resistance2 and pulsatility index ratios3 to assess the hemodynamic burden of the acardius; however, none of these studies attempted to directly relate cardiovascular findings on echocardiography to clinical status.

Cardiovascular decompensation in the pump twin is thought to occur as a result of a sustained volume overload due to the perfusion demands placed on it by the acardius. Based upon this hypothesis, radiofrequency ablation (RFA) of the umbilical cord is performed to interrupt blood flow to the acardius, theoretically reducing excess pump twin workload. However, despite the perceived hemodynamic burden in TRAP sequence and the benefits of therapeutic intervention, little data are available to describe the range and degree of cardiovascular findings specific to this condition.

The aim of this study was to describe the fetal echocardiographic findings in TRAP, compare these findings with historic indicators of disease severity (AC : PT ratio and polyhydramnios), and examine acute cardiovascular changes following RFA.


Between January 2004 and April 2007, 27 patients with TRAP sequence were referred to the Fetal Care Center of Cincinnati for the purpose of RFA of the acardius. All patients received a fetal echocardiogram for assessment of pump twin cardiac anatomy and function. In 23 patients, complete echocardiographic data sets were available; these subjects constitute the study population. All data collection and storage were done with the approval of the institutional review board and in congruency with the Health Insurance Portability and Accountability Act.

Obstetric ultrasound reports were reviewed to ascertain the presence of polyhydramnios and determine the AC : PT ratio. Polyhydramnios was defined as a deepest vertical pocket greater than 6 cm. The pump twin weight was determined by the formula of Hadlock et al.4. The acardius weight (in grams) was determined by ultrasonography as previously described1:

equation image

where length (in cm) represents the longest linear dimension of the acardius. A high-risk AC : PT ratio was identified as greater than 70%. Reverse flow was confirmed in the acardius's umbilical artery in all cases.

All pertinent echocardiographic data were analyzed to obtain a complete cardiovascular assessment prior to RFA. When available, post-RFA echocardiographic data were analyzed to assess for acute changes in cardiovascular status. Ventricular systolic function was assessed by calculation of biventricular shortening fractions by M-mode sonography; normal right and left ventricular shortening fraction was defined as ≥ 28%5. The degree of atrioventricular valve regurgitation was graded semi-quantitatively as mild (narrow jet ≤ half the atrial length), moderate (narrow jet > half the atrial length), or severe (wide jet > atrial length)6. Combined cardiac output was derived through Doppler ultrasound scan of the aortic and pulmonary valves, using the following equation:

equation image

The combined (right + left ventricular) cardiac output was then indexed to the estimated fetal weight, known as the combined cardiac index (CCI), and compared to published normal values7, which range from 225 to 625 mL/min/kg. Morphometric parameters obtained were cardiothoracic area ratio (CTR), cardiac chamber dimensions and ventricular wall thicknesses. A CTR of ≥ 0.35 was considered elevated5. Cardiac chamber and wall dimensions were compared to published gestational age-based normal values8. Pulsed-wave Doppler evaluation of the umbilical vein and ductus venosus was performed to assess for abnormalities in venous flow. Abnormal venous Doppler was defined as either absent or reversed flow in the ductus venosus during atrial systole, or umbilical venous pulsation in the free umbilical cord.

Both obstetric and cardiac ultrasound examinations were performed using Siemens Sequoia (Siemens Medical Solutions, Malvern, PA, USA) or Voluson E8 (GE Healthcare, Milwaukee, WI, USA) ultrasound systems. In all instances, tracings were obtained at an angle of insonation of less than 30° to ensure accurate velocity data. All measurements were performed three times and averaged.

The fetal cardiovascular profile score (CVPS)9 was used as a composite assessment of fetal cardiovascular status for the pump twin. The CVPS utilizes fetal echocardiographic and ultrasound findings to assess fetal cardiovascular state. It consists of a 10-point scale that incorporates the presence or absence of hydrops, abnormal venous and arterial Doppler findings, cardiomegaly, atrioventricular valve regurgitation and cardiac dysfunction. There are 1–2 point deductions from the total score depending on the extent of cardiovascular abnormalities noted (Table 1). Significant cardiovascular compromise was defined as a CVPS ≤ 8.

Table 1. Cardiovascular profile score9
ParameterNormal (2 points)− 1 point− 2 points
  1. CTR, cardiothoracic area ratio; MR, mitral regurgitation; SF, ventricular shortening fraction; TR, tricuspid regurgitation.

Hydrops fetalisNoneAscites, pleural effusion,Skin edema
   or pericardial effusion 
Abnormal venous DopplerNormal venous DopplerDuctus venosus atrial systolic reversalUmbilical venous pulsations
CardiomegalyCTR ≤ 0.35CTR > 0.35 and < 0.50CTR ≥ 0.50
Abnormal myocardial functionSF > 0.28 and no valveSF < 0.28 or TR orTR plus dysfunction or any MR
  regurgitation semilunar valve regurgitation 
Abnormal arterial DopplerNormal umbilical arteryAbsent end-diastolic flowReversed end-diastolic flow in
  diastolic flow in the umbilical artery the umbilical artery

Our institution established RFA as its preferred treatment method based upon published data on associated premature delivery rates and clinical success rates10. Intrafetal RFA was performed under local analgesia and intravenous sedation with a 19-gauge LeVeen needle with 2-cm tines. With ultrasound guidance, the tines deployed completely within the fetal abdomen at the umbilical cord insertion site. RFA was started with 60 W of power for 60 s, and the power was increased by 20-W increments at 60-s intervals to 120 W or when impedance dropped. Cessation of flow within the acardius was confirmed by color Doppler ultrasonography. Patients were observed overnight. Intravenous magnesium sulfate was used as needed following the procedure for tocolysis. No patient received non-steroidal anti-inflammatory drugs that could potentially result in altered hemodynamics.

Statistical analysis

Correlation analysis was used to assess the relationship between CCI and AC : PT ratio, CTR, and cardiac chamber sizes/wall dimensions. Differences in CCI in pump twins with or without polyhydramnios or elevated AC : PT ratio were compared by Wilcoxon rank-sum test. Changes in cardiovascular function before and after RFA were compared by Wilcoxon signed-rank test or McNemar's test as appropriate. Statistical significance was defined as P < 0.05.


Twenty-three subjects had complete echocardiographic data sets at initial evaluation (mean gestational age, 20.4 ± 2.5 weeks). Twenty-one patients subsequently underwent treatment. Two patients did not receive treatment owing to a contraindicated triplet gestation. Of the patients receiving treatment, 20 received RFA. The remaining subject underwent umbilical cord coagulation and release because of monoamniotic gestation with cord entanglement. Comprehensive pre- and post-RFA echocardiographic evaluations were available in nine subjects. The echocardiographic parameters obtained on follow-up examination were identical to those obtained preoperatively. The median time elapsed between pre- and post-RFA echocardiography was 3 (range, 2–11) days. Preoperative clinical findings in all the subjects are summarized in Table 2. Polyhydramnios was present in 9/23 (39%) of the subjects. An AC : PT ratio of ≥ 70% was present in 12/23 (52%) of the subjects. Cardiovascular compromise (CVPS ≤ 8) was present in 7/23 (30%) of the subjects; no pump twin had abnormal left ventricular systolic function pre-RFA.

Table 2. Clinical characteristics of pump twins
SubjectGA (weeks)CCI (mL/min/kg)PolyhydramniosAC : PT ratioCVPSHigh CTRAVVRVentricular dysfunctionAbnormal venous DopplerHydrops
  • *

    Pump twins with significant cardiac compromise.

  • Not assessed.

  • AC : PT ratio, acardius to pump twin weight ratio; AVVR, atrioventricular valve regurgitation; CCI, indexed combined cardiac output; CTR, cardiothoracic area ratio; CVPS, cardiovascular profile score; GA, gestational age.


There was a highly significant positive correlation between the CCI and CTR (r = 0.81, P < 0.0001, Figure 1). Six of the seven pump twins with elevated CCI had significant cardiovascular compromise (Figure 1). There was a significant positive correlation between CCI and transverse right atrial dimension (r = 0.67, P < 0.001) and transverse ventricular dimensions (left ventricle: r = 0.57; P < 0.01; right ventricle: r = 0.54; P < 0.01). There was no significant difference in CCI between subjects with or without polyhydramnios (592 ± 147 vs. 559 ± 289 mL/min/kg, P = 0.26) nor with or without a high AC : PT ratio (634 ± 253 vs. 544 ± 227 mL/min/kg, P = 0.34). There was no significant correlation between CCI and AC : PT ratio.

Figure 1.

Cardiothoracic area ratio (CTR) vs. indexed combined cardiac output (CCI) in 23 pump twins with TRAP sequence. The upper limit for CTR (······) and the published normal limits for CCI7 (vertical lines: ----, mean value; ____, 5th and 95th percentiles) are shown. Amongst pump twins with elevated CCI, six out of seven presented with evidence of significant cardiovascular compromise. ◊, subjects with significant cardiovascular compromise; ▴, subjects without cardiovascular compromise.

Of 20 subjects undergoing RFA, complete pre- and post-RFA data were available in 10 pump twins (Table 3). There was sudden intrauterine demise of one pump twin with hydrops (CVPS = 5) prior to follow-up imaging. At the discretion of the managing physicians, follow-up fetal echocardiography was not performed in the remaining nine cases. There was no evidence of significant cardiovascular compromise pre-RFA in any of these 10 cases.

Table 3. Cardiovascular findings in pump twins pre- and post-radiofrequency ablation
SubjectGA (weeks)CCI (mL/min/kg)CVPSHigh CTRAVVRVentricular dysfunctionAbnormal venous DopplerHydrops
  • *

    Pump twins with significant preoperative cardiac compromise.

  • AVVR, atrioventricular valve regurgitation; CCI, indexed combined cardiac output; CTR, cardiothoracic area ratio; CVPS, cardiovascular profile score; GA, gestational age.


There was a significant mean difference in CCI (730 ± 266 vs. 535 ± 100 mL/min/kg, P < 0.05) between pre- and post-RFA examinations. There was no significant difference in group means for CTR, left or right ventricular shortening fraction, or in the frequency of atrioventricular valve regurgitation or abnormal ductus venosus or umbilical venous Doppler.

Following RFA, the CVPS was improved or fully normalized in four of six pump twins defined as having significant cardiovascular compromise (CVPS ≤ 8) (Figure 2). The two pump twins whose CVPS decreased post-RFA both exhibited umbilical venous pulsations on follow-up ultrasound. One fetus demonstrated moderate constriction of the ductus arteriosus. This infant was also found to have mosaic Turner syndrome postnatally. The second fetus had the highest CCI and lowest CVPS (5) pre-operatively, and the acardius's cord was found to have recanalized on subsequent follow-up ultrasound scan.

Figure 2.

Pre- and post-radiofrequency ablation (RFA) cardiovascular profile score (CVPS) in pump twins. Each patient is represented pre- and postoperatively by an identical symbol. Significant cardiovascular compromise was defined as a CVPS ≤ 8 (----).

In pump twins with low right ventricular shortening fraction (< 28%), three of four (75%) normalized post-RFA (Figure 3). Among pump twins with abnormally elevated CCI preoperatively (n = 6), three decreased to normal and two to near-normal levels post-RFA (Figure 4).

Figure 3.

Pre- and post-radiofrequency ablation (RFA) right ventricular shortening fraction (RVSF) in pump twins. Each patient is represented pre- and postoperatively by an identical symbol. RVSF normalized in three of four pump twins post-RFA. ----, lower limit of normal RVSF.

Figure 4.

Pre- and post-radiofrequency ablation (RFA) indexed combined cardiac output (CCI) in pump twins. Each patient is represented pre- and postoperatively by an identical symbol. CCI normalized (or nearly normalized) post-RFA in five of six pump twins with preoperative elevation. ----, upper limit of normal for CCI7.


The current study describes detailed cardiac findings in pump twins with TRAP sequence, and relates these findings to their clinical status as well as to other—perhaps more widely used—indicators of disease severity such as AT : PT ratio and polyhydramnios. Our data also demonstrate, in a smaller cohort of pumps twins with post-RFA data, the cardiovascular benefit of RFA in the short-term follow-up therapy. To our knowledge, this is the largest study describing comprehensive cardiovascular findings pre- and post-RFA in TRAP sequence to date.

The pathophysiology in TRAP sequence consists of arterioarterial and venovenous chorioangiopagus with a pump twin perfusing a severely malformed cotwin with either no, or a very primitive and dysfunctional, heart. This arrangement results in a variable high cardiac output state in the pump twin, which is obligated to provide cardiac output to both its own systemic and placental circulation, as well as the vascular circulation of the acardius. In addition, the venous return of the acardius is added to the pump twin's circulation. Other investigators have attempted to quantify the degree of flow to the acardius by assessment of the relative resistance in the umbilical artery of pump twin and acardius2 or pulsatility index ratios3. However, we are not aware of prior reports utilizing quantitative echocardiography to assess cardiovascular status in TRAP.

The current study demonstrates a direct relationship between CCI, cardiothoracic area ratio, and cardiac morphometric measurements. Our data indicate that increases in global heart size appear to occur as a result of combined atrial and ventricular enlargement. Individual chamber dimensions increase uniformly in direct relation to elevations in cardiac output, a finding indicating excess volume load as the cause in pump twin heart failure.

Our data further suggest that elevated CCI is associated with a high incidence of cardiovascular compromise—as assessed by CVPS—due to the presence of ventricular systolic dysfunction, atrioventricular valve regurgitation, cardiomegaly, hydrops, or abnormal Doppler velocimetry. This makes intuitive sense, as the fetal heart operating at progressively higher outputs would be expected to reach it functional limits. Pathologic elevations in cardiac filling pressure related to ventricular volume loading likely manifest as end-diastolic flow reversals in the ductus venosus and umbilical venous pulsations.

It is interesting to note that, in the current study cohort, traditional obstetric prognosticators such as the AC : PT ratio and polyhydramnios1 did not correlate with cardiovascular derangements. Although an AC : PT ratio > 70% has been associated with a high risk of adverse pregnancy outcome, including premature delivery and rupture of membranes, hydramnios, and fetal ‘heart failure’1 it does not appear to closely reflect the hemodynamic burden placed on the pump twin. Given the poor correlation between AC : PT ratio and CCI, we speculate that identification of pump twins at a higher risk of heart failure may be improved by fetal echocardiography, by detection of elevations in either cardiothoracic area ratio and/or indexed combined cardiac output (Figure 1). It may be clinically useful, for example, to utilize pump twin cardiovascular data for stratification of patients into various treatment strategies, based on the level of suspicion of impending cardiovascular compromise. Specifically, patients with lower pump twin CCI may benefit from expectant management, whereas pump twins with high CCI may benefit from more invasive therapy aimed at acardius cord occlusion.

Echocardiography of pump twins following RFA demonstrated that the predominant changes are noted in fetuses with elevated CCI or ventricular dysfunction preoperatively. The inconsequential change noted in the CCI of pump twins that exhibited normal CCI before treatment suggests that the excess workload demanded by the acardius in those pregnancies was not significant. Following RFA of the acardius, acute ‘volume unloading’ of the pump twin is observed, as evidenced by lower CCI. In addition, post-RFA fetal echocardiography demonstrated that ventricular systolic function has the capability of rapid recovery. Improvements in compromised fetuses' venous Doppler flow patterns also suggest that, in some cases, loading conditions were altered sufficiently to normalize cardiac filling pressures.

This study was limited by its retrospective nature and, despite representing a relatively large cohort for a study examining a rare fetal condition, its sample size is modest. There is also a potential referral bias, as all cases were referred to our fetal care center specifically for RFA therapy. Owing to this bias, we were not able to collect serial cardiac data in an untreated population, nor were we able to evaluate fetuses undergoing alternative therapies. In addition, pre- and post-RFA data were not obtained in all cases. These factors may limit the ability to generalize our results.

In conclusion, quantitative fetal echocardiography reveals that changes in CCI are associated with a high incidence of cardiovascular compromise. Therefore, our data suggest that fetal echocardiography can play an important adjunctive role in the clinical assessment and stratification for therapy in TRAP sequence. Moreover, many of these cardiovascular derangements normalize following RFA of the acardius umbilical cord. Additional, possibly multicenter, studies should be performed to better define the role of echocardiography in evaluation, therapeutic stratification, and follow-up of TRAP sequence and to elucidate the cardiac natural history and response to various treatment methods.