Predictive value of cardiovascular parameters in stages 1 and 2 of twin-to-twin transfusion syndrome


  • Funding sources: None

  • Conflicts of interest: None declared



The Quintero staging of twin-to-twin transfusion syndrome (TTTS) does not include a comprehensive cardiovascular assessment. The aim of this study is to assess the predictive value of the myocardial performance index (MPI) and the Children's Hospital of Philadelphia (CHOP) score on recipient survival in Quintero stages 1 and 2 TTTS.


The cohort study was based on prospectively collected data between May 2008 and February 2013 in a population of stages 1 and 2 TTTS. Comparisons between groups were carried out using Student's t-test and χ2-test. A stepwise ascending multivariate logistic regression model was then built.


A total of 73 pregnancies in stages 1 and 2 of Quintero's classification were treated with laser. Rates of recipient fetal losses were higher when MPI was above 0.43 ms (71.4% vs 28.6%, p = 0.022). Rate of CHOP score above 5 was higher in the fetal loss group (28.6% vs 5.1%, p = 0.022). After adjustment for Quintero stages 1 or 2, the risk of recipient loss rate is higher according to CHOP score [OR 7.6; 95% confidence interval (CI) 1.3–43.5] or MPI value (OR 3.7; 95% CI 1.0–13.9).


The CHOP score and MPI are correlated with the recipient survival in stages 1 and 2 TTTS. © 2014 John Wiley & Sons, Ltd.


The twin-twin transfusion syndrome (TTTS) complicates 10% of monochorionic diamniotic twin pregnancies.[1-3] It accounts for 12% of mortality and 17% of perinatal morbidity in all twin pregnancies.[3-5] Untreated, severe forms are associated with a survival rate of less than 20%.[6, 7] Fetoscopy-guided laser photocoagulation of placental anastomoses significantly improves survival and is currently the reference treatment.[8] Nevertheless, this is not a procedure to be undertaken lightly because it can itself give rise to maternal morbidity[8, 9] and fetal morbidity and mortality.[8, 9] It is therefore essential to have reliable prognostic factors, in order to be able to identify among cases of TTTS, those pregnancies that might benefit from a laser treatment.

Quintero's classification is the most widely used system for categorizing TTTS.[10] Notwithstanding, it now appears to have significant shortcomings, and its prognostic value is contentious,[8, 11] especially because it does not integrate early-stage cardiovascular parameters. Fetal cardiovascular modifications in the early stages of the disease have been described, in particular in the recipient fetus.[11-15] The recipient may develop a specific cardiomyopathy because of the increased post-load and transplacental anastomotic passage of angiotensin (produced by the donor), giving rise to poorer myocardial compliance.[15, 16]

Other factors for TTTS that take account of fetal cardiac function have been proposed.[14, 17-22] The Children's Hospital of Philadelphia (CHOP) cardiovascular score comprises 12 parameters based on umbilical Doppler of the donor, venous Doppler, and cardiac function analysis of the recipient. The CHOP score is expressed as a score ranging from 0 to 20: the higher the score, the worse the cardiac function. The Tei index or myocardial performance index (MPI) evaluates global, systolic, and diastolic myocardial function.[23] It has the advantage of being independent of the gestational age,[18, 24, 25] ventricular geometry and fetal heart rate.[16, 23] The MPI value increases when ventricular dysfunction is present.[20] Its major utility is to allow straightforward and non-invasive assessment of right heart function. This is crucial,[14, 20] because the right ventricle is responsible for the bulk of perfusion during fetal life.[20] Studies evaluating the prognostic value of the CHOP score and MPI in TTTS are scarce. To the best of our knowledge, none has looked in depth to their relevance in the lower stages of Quintero's classification.

The principal objective of our study was to evaluate the prognostic value of the CHOP cardiovascular score and Tei index for the risk of recipient fetal loss in the stages 1 and 2 of Quintero's classification. The secondary objective was to study the prognostic value of these tools for global survival of both fetuses.


This was a single-center cohort study carried out in monochorionic diamniotic twin pregnancies complicated by TTTS at an early stage (Quintero 1 or 2) and treated by laser at the Centre Médico-Chirurgical and Obstétrical (CMCO) between July 2008 and October 2012. This is a center of reference for the diagnosis and the treatment for TTTS in the east of France. All the data were prospectively collected as part of routine care and then retrospectively analyzed to test our hypothesis.

Clinical and ultrasonographic data were collated using the computerized medical dossier system (DIAMM®, Micro6®).

Each mother underwent detailed obstetrical ultrasonography along with biometric, morphological, and Doppler studies in using either Voluson E8 or Voluson 130 ultrasound systems.

The diagnosis of TTTS was defined by the presence of oligohydramnios sequence in a monochorionic diamniotic twin gestation with like-sex twins, a single placental mass and a thin intertwine membrane. The Quintero stage was established in all cases.[10] First stage is defined as an oligamnios in the donor twin (deepest vertical amniotic fluid pocket ≤2 cm) and a polyhydramnios in the recipient twin [deepest vertical amniotic fluid pocket >8 cm before 20 weeks' gestation (WG) or >10 cm after 20 WG], the latter being the criteria used in Europe and the bladder of the donor fetus being still visible. The second stage is defined by the absence of the donor bladder. It should be noted that laser was used for Quintero stage 1 TTTS only when there was maternal discomfort from the polyhydramnios (uterine contractions, abdominal pain, and breathing discomfort). Cardiac assessment comprised measurement of the Tei index (for our analysis, we were interested in the Tei index values for the right heart only) and the CHOP cardiovascular score. The technique for measuring the Tei index was described by Hernandez-Andrade et al.[19] The Doppler was placed in such a way that tricuspid and pulmonary valve movements were obtained as distinct images. The isovolumetric contraction time (ICT) was obtained by measuring the time lapse between the end of the tricuspid valve closing click and the beginning of the pulmonary valve opening click, ejection time (ET) between the beginning of the pulmonary valve opening click and the end of its closing click, and the isovolumetric relaxation time (IRT) between the end of the pulmonary valve closing click and the beginning of the tricuspid valve opening click. The MPI was then calculated by applying the formula (ICT + IRT)/ET. Estimation of the CHOP cardiovascular score was based on the sum of the 12 cardiovascular parameters defined by Rychik[14] in 2007. For all analyses, the thresholds applied for both factors are those found in the literature. Accordingly, for the CHOP cardiovascular score, we retained the threshold of 5[26] and for the MPI that of 0.43 ms[19, 27]: above these thresholds, the value was considered abnormal.

All patients included underwent laser photocoagulation. The surgery was performed by two experienced operators with 1.3 or 2.0 mm fetoscopes (the outer diameter of the sheath was 10 French) according to gestational age, under local anesthesia and antibiotic coverage. First, the method used was the selective coagulation. Since 2010, patients were randomized to photocoagulation technique: either the selective technique described by Quintero et al.[28] or the Solomon technique described by Lopriore et al.[29]

An amnioreduction was performed after the procedure to obtain a normal quantity of amniotic fluid in the recipient twin (residual pocket <4 cm). Usually, the patients left the hospital the day after surgery and were subjected to weekly then bi-monthly ultrasound monitoring.

After delivery, each placental mass was freshly recuperated and forwarded without preparation by the maternity units in which the infants issuing from these pregnancies were born. Amniotic membranes were peeled from the surface of the placental plate, the umbilical cords sectioned beneath the clamps, the placenta rinsed, and excess blood squeezed out along the length of the umbilical cords. The umbilical arteries and veins were catheterized. For prime visualization of the anastomoses, veins were injected with eosin and arteries with methylene blue until all the peripheral branches were optimally rendered. The cord was then clamped in order to maintain optimal filling of the vessels. The identification of residual anastomoses was processed by visual inspection, and then, a digital picture was taken using a Sony DSC-W220 digital camera (12.1 megapixels), and the image tool software (for Window v.3.0) was used to measure the diameter of anastomoses on the surface of the chorionic plate. This evaluation was performed by a single operator who knew the outcome of the pregnancies because only placentas with double survival and double intrauterine death were analyzed. The principal assessment criterion was defined as the loss of the recipient fetus. Fetal loss is defined by the death of the recipient twin occurring during pregnancy or at early neonatal period (<7 days); the donor fetus is alive or not. The secondary endpoints were the loss of one or both twins, and the study of the correlation between the CHOP score and the right ventricular (RV) Tei index.

Statistical analysis was performed using spss software (version 16, SPSS, Inc, Chicago, IL, USA). All analyses were carried out with a type 1 or α risk set at 5%.

Normally distributed continuous variables are displayed as mean and standard deviation, and categorical variables are presented as ratios and percentages. Comparisons for normally distributed continuous variables were made using Student's t-tests and for categorical variables using χ2-tests. Univariate analysis was performed in order to identify variables that might explain the outcome. A multivariate logistic regression model was then constructed after manual selection of variables. Odds ratio (OR) and their associated confidence intervals (CIs) were computed for each predictor. Finally, we used a non-parametric analysis to study the relationship between both cardiovascular factors studied.

This study was approved by the French Data Protection Authority (CNIL) under number 1672926 v.0.


The study involved 73 pregnancies complicated by early-stage TTTS: 29 in Quintero stage 1 and 44 in Quintero stage 2. Laser treatment took place at a median gestational age of 19.2 WG (+/−2.7). The median gestational age at delivery was 33 WG (+/−5.4). In our cohort, 14 recipient fetuses died, two in Quintero stage 1 and 12 in Quintero stage 2. Among these recipient fetuses, seven died in the 24 h following the laser surgery; there were two additional intrauterine fetal deaths of undetermined origin later on (one 31 days after the laser, and for the other, this information is missing). Four died in the context of premature rupture of the membranes and/or chorioamnionitis between 6 and 54 days after laser. There was one early neonatal death after spontaneous preterm birth at 28 WG (90 days after the laser) (Figure 1). With regard to the donor fetuses, we observed eight losses in Quintero stage 1 group and 13 losses in Quintero stage 2 group. In ten pregnancies (13.7%), both fetuses died, and 15 (20.5%) suffered single fetus losses: the donor twin died in 11 (15.1%) cases and the recipient fetuses in four cases (5.5%) (Figure 1).

Figure 1.

Flow chart of patients studied. FD, fetal death; PRM, premature rupture of membranes; PI, pregnancy interruption

The characteristics of the cohort based on the recipient fetus outcome (loss/living) are shown in Table 1. Both populations were comparable in terms of maternal age, parity, gestational age at laser treatment, placental localization, and the presence or absence of residual placental anastomoses.

Table 1. Univariate comparison between the group of living recipient fetuses and recipient fetal losses
VariablesLiving recipient fetuses n = 59Recipient fetal loss n = 14p-value
  1. SD, standard deviation; MPI, myocardial performance index; CHOP, Children's Hospital of Philadelphia; AV, arterio-venous anastomoses.

Mean maternal age (±SD)30.07 (+/−5.58)29.86 (+/−4.57)0.896
Mean gestational age on laser (±SD)20 (+/−2.67)19.52 (+/−3.11)0.559
Posterior placental localization: n (%)40 (67.8%)7 (50%)0.173
Anterior placental localization: n (%)19 (32.2%)7 (50%) 
Quintero stage 1: n (%)27 (45.8%)2 (14.3%)0.028
Quintero stage 2: n (%)32 (54.2%)12 (85.7%) 
Right MPI ≤0.43: n (%)37 (62.7%)4 (28.6%)0.022
Right MPI >0.43: n (%)22 (37.3%)10 (71.4%) 
CHOP score ≤5: n (%)56 (94.9%)10 (71.4%)0.022
CHOP score >5: n (%)3 (5.1%)4 (28.6%) 
None residual anastomoses: n (%)21 (60%)4 (57.1%)0.603
Residual anastomoses ≥1: n (%)14 (40%)3 (42.9%) 
None residual AV anastomoses: n (%)25 (71.4%)5 (71.4%)0.66
Residual AV anastomoses ≥1: n (%)10 (28.6%)2 (28.6%) 
Selective technique of laser coagulation36 (61%)8 (57.1%)0.59
Solomon technique of laser coagulation23 (39%)6 (42.9%) 

Among the 73 pregnancies, 44 (60.3%) were treated with the selective technique and 29 (39.7%) with the Solomon technique. The selective coagulation is associated with a higher rate of fetal loss but without statistically significant difference: 57.1% versus 42.9%, p = 0.59, Table 1.

On univariate analysis, it emerged that there was a significantly higher recipient mortality rate in Quintero stage 2 TTTS compared with stage 1 fetuses: 85.7% versus 14.3% (p = 0.028). Similarly, the risk of loss of this twin was significantly higher when the MPI value was superior to 0.43: 71.4% versus 28.6% (p = 0.022). The proportion of CHOP cardiovascular score above 5 is higher in the fetal loss group than in living recipient fetuses group: 28.6% versus 5.1% (p = 0.022), Table 1.

On multivariate analysis, after adjustment for the Quintero stage, the risk of recipient loss was significantly higher when the CHOP score was above 5: OR = 7.58 (95% CI 1.32–43.48), p = 0.023 (Table 2). This also applied when the MPI value was superior than 0.43: OR = 3.75 (1.01–13.89), p = 0.048 (Table 2).

Table 2. Prediction of recipient fetal loss using several univariate and multivariate models
Models predictive of recipient fetal lossOR95% CIp-value
  1. OR, odds ratio; CI, confidence interval; MPI, myocardial performance index; CHOP, Children's Hospital of Philadelphia; Inf., inferior; Sup., superior.

  2. MPI: ≥0.43 versus <0.43; CHOP: >5 versus ≤5; Quintero: stage 2 versus stage 1.

Univariate modelsMPI4.201.1814.930.027
Multivariate modelsCHOP7.581.3243.480.023

With regard to analysis in relation to the global outcome of pregnancy (loss of one or both twins/two living twins), on univariate analysis, there was no significant difference in global survival, either in relation to the Quintero stage or CHOP score. The rate of loss of at least one twin was higher when the MPI value was above 0.43: 46.9% versus 24.4%, p = 0.039. On multivariate analysis, after adjustment for the Quintero stage, the risk of loss of at least one twin was higher when the MPI value was above 0.43: OR = 2.73 (95% CI 1.01–7.41), p = 0.047.

Moreover, we found a significant positive correlation between the CHOP score and the RV Tei index in the two first stages of Quintero's classification with a coefficient of correlation of 0.25, p = 0.035.


We found that the CHOP score and MPI value are predictors of recipient fetus loss in lower-stage TTTS, after adjustment for the Quintero stage. Similarly, an increased MPI is associated with a higher risk of loss of at least one of the twins. This shows that the Quintero classification is insufficient for predicting outcome of recipient survival in stages 1 and 2.

The Quintero classification is the most widely used TTTS staging system, but its prognostic value is currently subject to debate,[11, 30, 31] especially because it does not incorporate cardiovascular parameters.[17, 32] The prognosis of stages 3 and 4 TTTS is poor and the utility of a laser treatment beyond doubt. In the early stages, the prognosis is dependent on cardiovascular modifications, which should be taken into account in order to enhance the Quintero classification. Fetal ultrasonography would appear to be of interest, particularly in stage 1 when the utility of laser treatment is controversial.

The cohort size in our series was fairly substantial given the prevalence of TTTS, but it did not allow, for reasons of statistical power, for a separate analysis of stages 1 and 2 of the Quintero classification. A degree of selection bias may be present in our population because solely pregnancies treated by laser were analyzed (some stage 1 pregnancies were not included); however, that does not appear to us to call into question the validity of our results. Ultrasonography along with CHOP cardiovascular score and TEI index measurements were conducted by the same operator, avoiding inter-observer variations. For our ultrasonographer, the mean time taken to measure the 12 parameters of the CHOP cardiovascular score was 15 min and the MPI 1 was 2 min.

Studies with the CHOP cardiovascular score are relatively scarce.[14, 26, 33] Our study has shown a significant increase in the risk of recipient fetal loss with a sevenfold higher risk when the CHOP score is above 5. To our knowledge, only a single study[26] has tested the prognostic value of this score in a population treated by laser but failed to find a significant link between the fetal survival rate (global and recipient fetus) and CHOP cardiovascular score. A study into exclusively Quintero classification stage 1 pregnancies did not find a correlation between a high CHOP score and an increased risk of disease progression or greater need to perform amnioreduction[33] nor gestational age at delivery. Neonatal survival was not analyzed. The CHOP cardiovascular score, by means of the 11 parameters that it measures in the recipient fetus, would probably allow analysis of the chronology of fetal cardiac damage. However, it is a score that is difficult to establish because it demands both an experienced ultrasonographer and a considerable working time: its total completion time has been estimated at 60 min.[14] This time can be reduced to 30–45 min with training and experience.[14] Routinely, obtaining the CHOP cardiovascular score raises the problem of the duration of ultrasonography. It is therefore necessary, before making it a feature of current practice, that its application and utility in the management of TTTS be justified.

Studies evaluating the utility of MPI in TTTS are also fairly scanty. To our knowledge, our cohort is the largest to date. We found that when the MPI value is above 0.43, the risk of recipient fetal loss is increased threefold. The studies found in the literature have principally examined modifications in the cardiac function of the recipient fetus[27, 34, 35] but without assessing the prognostic value of the Tei index on the obstetrical outcome.[14, 20, 36] Others have investigated the obstetrical outcome but have not found a genuine correlation.[36] The majority of these studies concur in respect of various findings. The Tei index is a new Doppler index evaluating systolic and diastolic cardiac function.[23] It is technically straightforward to obtain and has a high degree of reproducibility.[19] It has a key advantage: it allows ready measurement of right fetal heart function, a measurement that hitherto has been difficult to obtain.[14, 20] The value of the MPI is significantly higher in cases of TTTS than in its absence and in the recipient fetus compared with the donor fetus.[12, 20]

Most studies have evaluated global fetal cardiac function based on the MPI of right and left hearts[12, 15, 27, 35]; only a few authors have investigated the Tei index of the left ventricle alone.[25, 37] In our study, we sought to evaluate the MPI of the right heart in the recipient fetus. Indeed, the literature has clearly shown the importance of studying right heart function in the fetus.[14, 20] In 43.8% of the recipient fetuses in our cohort, the right MPI was abnormal (>0.43), reflecting the high prevalence of early vascular repercussions as early as Quintero stages 1 and 2, as has already been observed.[35] The mean MPI value in our cohort was 0.43 ms, which is less than that reported in the literature[14, 20, 34] probably because the gestational age was close to 20 weeks, earlier than that found in the literature.[14, 20, 25, 34, 35, 38] The Tei index reveals fetuses with cardiac dysfunction. These fetuses are more vulnerable, so it seems necessary to introduce a factor in the lower stages of Quintero's classification that takes account of this information. MPI appears to be a good one.

The focus of this article was the recipient twin loss. It would be useful to have also a test to predict the outcome of the donor, especially in our series where the donor loss rate is higher than the recipient loss rate in both stages 1 and 2. However, we have not analyzed the prediction of donor fetal loss on the basis of cardiovascular parameters because there is no rational pathophysiology that underlies this hypothesis. Moreover, in the CHOP score, only one item of 12 concerns the donor twin.

In our study, 60.3% of the pregnancies were treated by selective laser surgery, and only 39.7% were treated by Solomon technique. This difference can be explained by the fact that before 2010, the selective technique was preferentially used. Since 2010, patients were part of a randomized trial (selective technique vs Solomon technique), which was recently concluded.[39] In this study, there is no statistically significant difference between these two techniques (p = 0.59).

Concerning the evaluation of the placenta after birth, it makes sense to study only placenta with double survival and double intrauterine death. Indeed, when the laser is complicated by the loss of one of the twins early after the procedure, the placental architecture is altered given the necrotic degradation and maceration. In our study, we analyzed 42 placentas out of 58 cases (72.4%). We found residual anastomoses in 17 (40.5%) cases and residual arterio-venous anastomoses in 12 (28.5%) cases. These results are in line with the results found in other studies. Lewi et al.[40] found between 21% and 100% of residual anastomoses according to the outcome, and Lopriore et al.[41] found up to 33% of residual anastomoses.

Evaluation of the placenta after delivery does probably not reflect the physiology because the pressures used to fill the vessels are certainly higher than physiological pressure. Moreover, it is possible that anastomoses have not been visualized during the laser surgery either because they had a too small diameter or because they were still not present. The laser technique is imperfect. An alternative laser surgery technique is to use a technique in which the entire vascular equator is coagulated: the Solomon technique (vs selective coagulation technique).

The correlation we found between CHOP score and RV Tei index is probably related to the physiological relationship between both. Indeed, the CHOP score includes a simple visual assessment of Tei index by evaluation of tricuspid inflow patterns. As normal double peak inflow shifts into a single peak, the time between tricuspid valve closure and valve opening lengthens, thereby contributing to a higher Tei index value.


The Quintero classification is insufficient for determining the prognosis of recipient fetuses in TTTS. In stages 1 and 2, the incorporation of cardiovascular parameters predicted recipient fetal loss.


  • The Quintero's classification for twin-to-twin transfusion syndrome (TTTS) does not integrate cardiovascular parameters. Studies evaluating the prognostic value of the Children's Hospital of Philadelphia (CHOP) score and the myocardial performance index (MPI) in TTTS, especially in lower stages, are scarce.


  • The CHOP score and MPI are strongly correlated with recipient survival and can predict fetal loss. They may be used in order to improve evaluation of stages 1 and 2 of TTTS.