To examine cardiac structural and functional changes in twin–twin transfusion syndrome (TTTS), relative to Quintero stage, as a means of evaluating the spectrum of cardiomyopathy in TTTS.
To examine cardiac structural and functional changes in twin–twin transfusion syndrome (TTTS), relative to Quintero stage, as a means of evaluating the spectrum of cardiomyopathy in TTTS.
This was a cross-sectional, retrospective study of 42 consecutive cases of TTTS referred to a single fetal therapy center. Quintero stages were assigned by standard criteria. Presence of ventricular hypertrophy, cardiomegaly, atrioventricular valve regurgitation (AVVR), ventricular systolic dysfunction and right ventricular outflow tract obstruction on fetal echocardiography were noted. The Doppler myocardial performance index (MPI), an index of global ventricular function, was calculated for both ventricles in subjects with adequate Doppler data. We compared cardiac changes across Quintero stages.
There was no cardiomyopathy observed in donor twins. The majority of subjects presented at Quintero Stage I (n = 14), II (n = 14) or III (n = 11), with fewer at Stages IV (n = 2) or V (n = 1). As early as Quintero Stages I and II, a significant proportion of recipient twins had ventricular hypertrophy (17/28, 61%), AVVR (6/28, 21%) or quantitative abnormalities in either right (12/24, 50%) or left (14/24, 58%) ventricular function. Increasing prevalence of biventricular systolic dysfunction and cardiomegaly accompanied advancing Quintero stage.
Changes in cardiac structure and function not reflected in Quintero staging occur in recipient twins early in the evolution of TTTS. Incorporation of cardiac findings into assessment of TTTS severity may prove useful in stratification of risk and treatment selection. Copyright © 2007 ISUOG. Published by John Wiley & Sons, Ltd.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Although cardiac changes in the recipient twin in twin–twin transfusion syndrome (TTTS) are well described1–4, in many centers characterization of these changes is not well integrated into the management of these patients. Ventricular systolic dysfunction, cardiac chamber enlargement, ventricular hypertrophy and atrioventricular valve regurgitation (AVVR) are often seen in the recipient twin of affected pregnancies and, in advanced cases, right ventricular (RV) outflow tract obstruction has been observed5. Recently, quantitative methods have been utilized to characterize cardiac changes in TTTS, among them the Doppler myocardial performance index (MPI)6, 7, an index of global systolic and diastolic function8, 9. However, the progression, or natural history, of cardiomyopathy seen in TTTS has not been fully characterized. In addition, it is not known whether careful fetal cardiac evaluation can identify the early changes in cardiac structure or function that precede the more dramatic changes that have been described1–5.
In clinical practice, the severity of TTTS is usually characterized by the staging system proposed by Quintero et al.10. Advancing Quintero stage has been associated with higher perinatal mortality11, 12, and has served as part of the inclusion criteria in a recent National Institute of Health (NIH)-sponsored, multicenter, randomized trial of amnioreduction vs. fetoscopic laser therapy for TTTS13. Although preliminary studies have suggested that cardiac changes may be present even in early Quintero stages7, cardiac findings are not incorporated into the Quintero assessment of TTTS severity. These observations led us to consider the spectrum of findings in TTTS cardiomyopathy, and how they relate to severity of TTTS as described by Quintero stage. It was the purpose of this study, therefore, to examine cardiac structural and functional changes in TTTS relative to Quintero stage, as a means of describing the early cardiac manifestations and progression of the cardiomyopathy seen in TTTS. Understanding that it is uncertain whether the temporal evolution of TTTS will progress stepwise through successive Quintero staging12, we made the assumption that a detailed, cross-sectional comparison of cardiovascular findings between Quintero stages may provide insight into the spectrum and, perhaps, the sequence of appearance, of cardiac findings in these pregnancies. We hypothesized that careful fetal cardiac assessment would identify abnormalities in cardiac structure and/or function, particularly in the recipient twin, even in the early stages of TTTS.
This was a cross-sectional, retrospective, observational analysis of 42 consecutive cases of TTTS referred to the Fetal Care Center of Cincinnati for evaluation between the center's inception in March 2003 and November 2005. Fetal echocardiographic data and clinical records were reviewed in all subjects. All data collection and storage was carried out under approval of the institutional review board of Cincinnati Children's Hospital Medical Center to insure proper handling of protected health information under the Health Information Portability and Accountability Act. All patients underwent a detailed fetal anatomical survey, placental evaluation and Doppler evaluation of flow patterns in the ductus venosus, umbilical vein and umbilical artery. Sonographic presence (or absence) of a fetal bladder, evidence of hydrops fetalis, fetal weight and maximum/minimum amniotic fluid vertical pockets were determined in each fetus. The diagnosis of TTTS was defined as the presence of oligo-/polyhydramnios sequence in a monochorionic, diamniotic twin gestation with like-sex twins, a single placental mass and a thin intertwin membrane. Oligohydramnios in the donor twin was defined as deepest vertical amniotic fluid pocket ≤ 2 cm. Polyhydramnios in the recipient twin was defined as deepest vertical amniotic fluid pocket > 8 cm, or > 6 cm if prior amnioreduction had been performed after 20 weeks' gestation. The definition of polyhydramnios, which deviates slightly from published definitions14, was taken from the inclusion criteria for the NIH study being conducted at our center during the study period13. In each case, the Quintero stage was assigned using established criteria10 identified on initial assessment.
Complete fetal echocardiograms were obtained in all subjects according to published standards15 using either Vivid 7 (General Electric Corp., Milwaukee, WI, USA) or Sequoia (Siemens Corp., Mountain View, CA, USA) ultrasound systems. In both twins, thickness of the left ventricular (LV) free wall, interventricular septum and RV free wall was measured by either M-mode or two-dimensional (2D) echocardiography16, 17; ventricular hypertrophy was defined when at least two walls had a thickness above the normal 95% CI17. End-diastolic dimensions of RV and LV were measured by either M-mode or 2D echocardiography and compared with published normal values for gestational age16, 17. Overall cardiac size was assessed as the ratio of the cardiac area in the four-chamber view to the thoracic cross-sectional area at the same level. Cardiomegaly was defined as a cardiac/thoracic area ratio > 0.3518. The presence or absence of holosystolic AVVR (tricuspid or mitral) was noted by color Doppler flow imaging. Severity of AVVR was graded semiquantitatively as mild (narrow jet ≤ atrial length), moderate (narrow jet atrial length), or severe (wide jet atrial length). Ventricular systolic function was assessed by shortening fraction, and classified as either normal (≥ 28%) or depressed (< 28%).
In order to assess quantitatively RV and LV function, the ventricular MPI was calculated for both RV and LV in a subset of patients with adequate pulsed Doppler evaluation of both atrioventricular inflow and ventricular outflow, as described previously8, 9 (Figure 1). LV-MPI was measured by simultaneous pulsed-wave Doppler sampling of the LV inflow and outflow. In addition, the LV isovolumetric relaxation (IRT) and contraction (ICT) times were measured as indicated in Figure 1. RV-MPI was measured from separate images of tricuspid inflow and pulmonary outflow Doppler; RV-MPI was only calculated when the heart rate difference between tricuspid and pulmonic Doppler tracings was ≤ 5 bpm. MPI values in TTTS were compared with values obtained in 17 control fetuses with normal fetal echocardiograms performed for family history of congenital heart disease. An abnormally elevated MPI was defined as an elevation in MPI > 2 SDs above the mean value for controls. Differences in both RV- and LV-MPI between the recipient twin and its cotwin (donor twin) were calculated. The presence or absence of RV outflow tract obstruction was evaluated by 2D ultrasound and pulsed-wave and color flow Doppler.
The Quintero staging system typically assigns Stage III TTTS in one of two scenarios: (1) in the absence of end-diastolic flow in the umbilical artery, most frequently seen in the donor twin; (2) in the presence of reversed flow in the ductus venosus during atrial systole or pulsatility in the umbilical vein, most frequently seen in the recipient twin. As the cardiovascular etiology of these findings (increased placental resistance in the former, elevated central venous pressure in the latter) differs, we divided recipient twins with Stage III TTTS into two groups for comparison of cardiac findings: (1) recipient twins to which Stage III was assigned due to atrial systolic reversals in the ductus venosus or umbilical venous pulsatility, with (n = 1) or without (n = 4) absent/reversed flow in the umbilical artery; (2) recipient twins to which Stage III was assigned due to absent/reversed diastolic flow in the umbilical artery in either donor or recipient twin, without venous Doppler abnormalities.
Continuous, normally distributed variables are expressed here as mean ± 1 SD; non-normally distributed data are reported as median and range. Differences in proportion of abnormal cardiac findings between subgroups of Stage III recipient twins, and between recipient twins with/without prior amnioreduction were assessed by either chi-square analysis or Fisher's exact test as appropriate. Levene's test was performed to test for homogeneity of variance between groups. Because of some heterogeneity of variance, comparisons of RV- and LV-MPI between Quintero stages and versus controls were performed by non-parametric analysis of variance (Kruskal–Wallis) and Wilcoxon rank sums tests. Differences in RV- and LV-MPI, IRT, ICT and ejection time between recipient twins and their cotwins were assessed by paired t-tests. The relationship between RV- and LV-MPI in recipient twins was assessed using Spearman's correlation analysis. Statistical significance was assigned to P-values < 0.05. No adjustment was made to the P-values to account for multiple tests. All analyses were performed using either SAS® Statview® 5.0 (SAS Institute Inc., Cary, NC, USA) or SPSS 12.0 (SPSS Inc., Chicago, IL, USA) software.
We studied 42 pregnancies at a median gestational age of 20 (range, 17–30) weeks. Twenty-one of 42 subjects (50%) had undergone amnioreduction prior to fetal echocardiography and Quintero staging. The median time between Quintero staging and fetal echocardiography was 0 (range, 0–11) days; 39/42 subjects underwent echocardiographic assessment within 2 days of staging. No patient underwent a therapeutic procedure between staging at the Fetal Care Center and fetal echocardiography. The majority of TTTS cases were Quintero Stages I–III, with 14 at Stage I, 14 at Stage II, 11 at Stage III, two at Stage IV and one at Stage V.
There was no significant hypertrophy, AVVR, ventricular systolic dysfunction, or cardiomegaly in any donor twin, and ventricular chamber size and wall thickness for donor twins were within published normal limits17 in all cases. Ventricular morphometric assessment for recipient twins is presented in Table 1. Both RV and LV cavity dimensions were at the lower end of the normal range, whereas ventricular wall thickness was, on average, elevated compared with published normal values17. The prevalence of cardiac findings in recipient twins at each Quintero stage is presented in Table 2. The small number of subjects at each Quintero stage precluded meaningful statistical analysis and so prevalence of findings is reported. Ventricular hypertrophy, dysfunction (RV and/or LV shortening fraction < 28%) and AVVR were prevalent at all Quintero stages; the development of cardiomegaly and severe ventricular systolic dysfunction (RV or LV shortening fraction < 20%) appeared with increasing frequency in higher Quintero stages. RV outflow tract obstruction was seen in only three Stage-III and in both Stage-IV recipient twins.
|Dimension||Z-score (mean ± SD)|
|RV end-diastolic dimension||− 0.9 ± 0.7|
|LV end-diastolic dimension||− 0.5 ± 0.6|
|RV free wall thickness||2.8 ± 1.8|
|IVS thickness||2.4 ± 1.5|
|LV free wall thickness||2.8 ± 1.8|
|Cardiac finding||Prevalence of cardiac finding (n (%))|
|Stage I||Stage II||Stage III||Stage IV||Stage V|
|Ventricular hypertrophy||8/14 (57)||9/14 (64)||6/11 (55)||1/2 (50)||1/1 (100)|
|Atrioventricular valve regurgitation||2/14 (14)||4/14 (28)||5/11 (45)||2/2 (100)||1/1 (100)|
|Moderate or severe tricuspid regurgitation||1/14 (7)||3/14 (21)||4/11 (36)||2/2 (100)||1/1 (100)|
|Ventricular systolic dysfunction||3/14 (21)||3/14 (21)||6/11 (55)||2/2 (100)||1/1 (100)|
|Severe ventricular systolic dysfunction (SF < 20%)||0/14 (0)||0/14 (0)||3/11 (27)||2/2 (100)||1/1 (100)|
|Cardiomegaly||1/14 (7)||1/14 (7)||2/11 (18)||2/2 (100)||1/1 (100)|
|Elevated RV-MPI*||3/10 (30)||9/14 (64)||5/8 (63)||NA||1/1 (100)|
|Elevated LV-MPI*||6/11 (55)||9/13 (70)||4/9 (44)||1/1 (100)||1/1 (100)|
RV-MPI was measured in 33/42 recipient twins and 23/42 donor twins; LV-MPI was measured in 35/42 recipient twins and 32/42 donor twins. The MPI in donor twins was normal in all cases except one, in which the RV-MPI was elevated (0.58) in the setting of absent end-diastolic flow in the donor twin's umbilical artery. The MPI for both RV and LV in recipient twins is presented in Figure 2. When compared with normal fetuses, a significant number of recipient twins displayed abnormalities in both RV- and LV-MPI, despite being at an early Quintero stage (Stages I and II). The mean RV-MPIs in Stage I (0.50 ± 0.07), Stage II (0.52 ± 0.13) and Stage III (0.70 ± 0.25) fetuses were all significantly elevated compared with control fetuses (0.32 ± 0.08, P < 0.001 in all cases) (Figure 2b). The RV-MPIs in Stage III recipient twins was also higher than that in recipient twins at Stages I or II (P < 0.05). The mean LV-MPIs in Stage I (0.48 ± 0.11) and Stage II (0.50 ± 0.09) fetuses were also elevated compared with control fetuses (0.33 ± 0.05, P < 0.001, Kruskal–Wallis) (Figure 2a). The mean LV-MPI in Stage III fetuses (0.52 ± 0.27) was also elevated, but this did not achieve statistical significance compared with control fetuses. The MPI values in recipient twins at Stages IV and V were all significantly elevated (> 0.70, data not shown); however, the small number of MPI values in the groups precluded meaningful statistical comparison.
The mean difference in RV- and LV-MPI between recipient twins and their donor cotwins was 0.27 ± 0.15 and 0.20 ± 0.17, respectively (both P < 0.0001). The intertwin difference in RV- and LV-MPI did not increase significantly across Quintero Stages I–III. In recipient twins, RV-MPI correlated strongly with LV-MPI (r = 0.65, P < 0.0001) (Figure 3), suggesting that ventricular dysfunction may occur concomitantly in the RV and LV.
Comparison of IRT, ICT and ejection times between recipient twins and donor twins was performed. As there was a significant difference in cardiac cycle length between recipient twins and donor twins (421 ± 22 vs. 410 ± 22 ms, P < 0.01), ICT, IRT and ejection time were corrected for heart rate by dividing them by the square root of the cycle length. Heart rate-corrected IRT (80 ± 13 vs. 58 ± 9 ms, P < 0.0001) and ICT (66 ± 28 vs. 34 ± 9 ms, P < 0.0001) were both prolonged in recipient twins compared with donor twins. There was no significant difference in ejection time between donor and recipient twins.
Of the 11 cases of Stage III TTTS, six were staged solely due to absent/reversed flow in the umbilical artery of the donor twin (with normal Doppler flow patterns in the recipient twin), and five were staged due to abnormal ductus venosus and/or umbilical venous flow patterns in the recipient twin with (n = 1) or without (n = 4) absent/reversed flow in the umbilical artery of the donor twin. Although group numbers were small, there was a trend towards higher prevalence of abnormal cardiac findings in those Stage III recipient twins with venous Doppler abnormalities (Table 3).
|Cardiac finding||Prevalence of cardiac finding (n)||P|
|AEDF in donor||Venous Doppler changes in recipient|
|Ventricular hypertrophy||3/6||3/5||> 0.999|
|Atrioventricular valve regurgitation||1/6||4/5||0.08|
|Moderate or severe tricuspid regurgitation||1/6||3/5||0.24|
|Ventricular systolic dysfunction||1/6||3/5||0.24|
In 50% (21/42) cases, amnioreduction had been performed prior to the Quintero staging and echocardiography at our center. The distribution of Quintero stages was not significantly different between those that underwent amnioreduction and those that did not (χ2 = 3.4). The proportion of cardiac findings was likewise not significantly different between these groups (Table 4).
|Cardiac finding||Prevalence of cardiac finding (n)||P|
|Prior amnio.||No prior amnio.|
|Atrioventricular valve regurgitation||7/21||7/21||NS|
|Moderate or severe tricuspid regurgitation||6/21||5/21||NS|
|Ventricular systolic dysfunction||9/21||6/21||NS|
This study demonstrates that significant changes in ventricular structure and function occur, even in the earliest stages of TTTS, in the recipient twin but not in the donor twin. Whereas donor twins uniformly had normal cardiac structure and function, in recipient twins there was a significant incidence of ventricular hypertrophy and AVVR and elevations in both RV- and LV-MPI, consistent with decrements in both systolic and diastolic ventricular function, as early as Quintero Stages I and II. As in previous reports1–5, more severe cardiomyopathy, evidenced by severe AVVR, cardiomegaly, ventricular systolic dysfunction and RV outflow tract obstruction was observed, but these appeared to occur, with increasing frequency, in more advanced stages of TTTS as defined by Quintero staging.
These findings are important, in that careful evaluation of cardiac structure and function may represent a means of early identification of TTTS in monochorionic twin pregnancies, when more advanced criteria for the diagnosis of TTTS are not yet present. Discrepancies in amniotic fluid volume and fetal size are not unique to TTTS; in contrast, cardiomyopathy being limited to the recipient twin does appear to be unique to TTTS1, 3, 4, 7. Therefore, early clinical diagnosis of recipient twin cardiomyopathy, specifically by quantitative identification of ventricular hypertrophy or elevated MPI, may allow clinicians to achieve early diagnosis of TTTS and thus initiate therapy at an earlier point in the disease process. Furthermore, identification of cardiac abnormalities in the recipient twin, with prevalence rates of 7–70% (Table 2), in cases of TTTS that have been graded as Quintero Stage I or II could allow clinicians to identify a more severe case of TTTS than would otherwise be suggested by Quintero staging alone. This may in turn lead physicians to consider more aggressive therapy, and could result in more favorable clinical outcomes for these high-risk pregnancies.
There has always been concern that spontaneous intrauterine fetal demise occurs, inexplicably, even in Quintero Stages I and II. Moreover, when demise of the recipient twin following laser treatment of TTTS occurs, it does not appear to correlate to preoperative Quintero stage or Doppler velocimetry19. These findings raise the possibility that some recipient twins are more compromised, perhaps by TTTS cardiomyopathy, than is revealed by ultrasound alone. It may be useful, therefore, to modify the Quintero staging of TTTS to include echocardiographic findings. In particular it may prove useful to modify staging in cases of Quintero Stage I or II TTTS when the recipient twin displays abnormalities in the MPI, significant ventricular hypertrophy or AVVR, as this may represent a more ‘advanced’ case of TTTS than that in cases of Stage I or II TTTS without recipient twin cardiac findings20. Our findings also suggest that within Stage III disease, the cardiac ‘phenotype’ of the recipient twin is different when there are alterations in venous Doppler flow patterns in the recipient twin, compared with when umbilical artery flow abnormalities in the donor twin (indicative of placental insufficiency) predominate. This concept may explain the wide deviation in both RV- and LV-MPI values seen in Stage III TTTS (Figure 2), with abnormally high ventricular MPI values in those cases with significant recipient twin cardiomyopathy (with venous Doppler changes), and more normal ventricular MPI when staging is based on abnormalities in donor twin umbilical artery flow alone (Table 3).
Our findings corroborate earlier reports which found that alterations in the ventricular MPI were present in the recipient of affected twin pregnancies6, 7. In comparison to these prior reports, our data are strengthened by a larger overall study cohort, and by substantially more observations of RV- and LV-MPI. Our cohort included a large number (n = 28) of Stage I and II TTTS, in contrast to the report by Barrea et al.6 (n = 7), supporting the important conclusion that cardiac changes occur with significant frequency even in ‘mild’ TTTS, as assessed by Quintero staging. We found a significant incidence of global LV dysfunction in early Quintero stages. In contrast to previous reports, our findings reveal that both diastolic and systolic dysfunction, as evidenced by progressive elevations in both IRT and ICT, account for the observed increases in LV-MPI in recipient twins. Our data also suggest that LV dysfunction usually accompanies developing RV dysfunction, a finding that has not been reported uniformly6.
The presence of biventricular hypertrophy and dysfunction is not unexpected, as systemic hypertension in the recipient twin is thought to represent one aspect of the pathophysiology of TTTS21, 22. As the fetal circulation is a parallel one, systemic hypertension is expected to produce alterations in the structure and function of both ventricles. Interestingly, changes in the LV-MPI, reflecting abnormalities in global LV function, have been documented in the setting of adult systemic hypertension, both in the presence and in the absence of LV hypertrophy23, 24. This suggests that alterations in the MPI occur early in the course of systemic hypertension, when ventricular afterload is increased, yet evidence of end-organ effects such as ventricular wall hypertrophy are not present. It is therefore possible that MPI assessment of global ventricular function in monochorionic multifetal gestations could serve as a very early indicator of cardiac involvement in developing TTTS, potentially occurring prior to ventricular hypertrophy, systolic dysfunction and AVVR.
There are several limitations to our observations. As with many single-center studies of TTTS, the sample size is modest, making generalization from our findings more difficult. While the ideal control group would consist of monochorionic twin gestations without TTTS, our control group consisted of singleton fetuses. However, the mean RV- and LV-MPI in our control group were the same as those in previous reports of singleton fetuses25, as well as those in monochorionic twins without TTTS26. We also compared the MPI in recipient twins to their cotwins, as a form of ‘internal’ control, and these data lend further credence to the abnormalities noted in the recipient twins. Although no subject underwent therapeutic intervention between Quintero staging and fetal echocardiography, 50% of our subjects had undergone amnioreduction prior to evaluation. This has the potential to serve as a confounding factor in comparing cardiac findings with Quintero stage, although we did not find significant differences in the distribution of Quintero stages or frequency of cardiac findings in cases in which there had been prior amnioreduction compared with cases in which there had not. We were also unable to evaluate the relationship of cardiac findings to outcome in our study population given the sample size and the relative heterogeneity of treatment modalities, which included serial amnioreduction, selective fetoscopic laser ablation of placental vascular anastomoses, laser septostomy and selective fetocide by radiofrequency ablation in certain advanced cases.
It is unclear whether the natural history of TTTS includes stepwise progression from one Quintero stage to the next12, and thus, by comparing cardiovascular findings between Quintero stages, we have not demonstrated conclusively the true natural history of this condition. In this era of fetal therapy, however, it is unlikely that a true natural history study can be performed because of the lack of cases that will not undergo intervention. Nonetheless, our findings are at least suggestive that the observed cardiac changes in recipient twins may evolve in the order outlined by Quintero staging. Further, longitudinal study is necessary to confirm these impressions.
In summary, it is clear that demonstrable, quantifiable changes in both RV and LV structure and function occur in recipient twins at even the earliest stages of TTTS as characterized by the Quintero staging system. Given the suspected pathophysiology of TTTS, it is likely that careful cardiac evaluation of ventricular hypertrophy and global function, by determination of the MPI, will allow clinicians to identify TTTS at a very early stage in its development. Large, multicenter studies of monochorionic twin pregnancies are needed to better characterize the early cardiac natural history in TTTS and its relation to clinical outcomes and prognoses.