Volume 48, Issue 4 p. 470-475
Original Paper
Free Access

Can middle cerebral artery peak systolic velocity predict polycythemia in monochorionic–diamniotic twins? Evidence from a prospective cohort study

M. Fishel-Bartal

M. Fishel-Bartal

Department of Obstetrics and Gynecology, Sheba Medical Center, Tel-Hashomer, Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel

Search for more papers by this author
B. Weisz

B. Weisz

Department of Obstetrics and Gynecology, Sheba Medical Center, Tel-Hashomer, Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel

Search for more papers by this author
S. Mazaki-Tovi

S. Mazaki-Tovi

Department of Obstetrics and Gynecology, Sheba Medical Center, Tel-Hashomer, Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel

Search for more papers by this author
E. Ashwal

E. Ashwal

Department of Obstetrics and Gynecology, Sheba Medical Center, Tel-Hashomer, Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel

Search for more papers by this author
B. Chayen

B. Chayen

Department of Obstetrics and Gynecology, Sheba Medical Center, Tel-Hashomer, Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel

Search for more papers by this author
S. Lipitz

S. Lipitz

Department of Obstetrics and Gynecology, Sheba Medical Center, Tel-Hashomer, Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel

Search for more papers by this author
Y. Yinon

Corresponding Author

Y. Yinon

Department of Obstetrics and Gynecology, Sheba Medical Center, Tel-Hashomer, Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel

Correspondence to: Prof. Y. Yinon, Department of Obstetrics and Gynecology, Sheba Medical Center, Tel-Hashomer, 52621 Tel Aviv, Israel (e-mail: [email protected])Search for more papers by this author
First published: 11 December 2015
Citations: 23

ABSTRACT

Objective

The antenatal diagnosis of twin anemia–polycythemia sequence (TAPS) in monochorionic–diamniotic (MCDA) twin pregnancies is based on elevated peak systolic velocity in the middle cerebral artery (MCA-PSV) in the donor twin and decreased MCA-PSV in the recipient twin. However, the association between these parameters and polycythemia has not yet been established. The aim of this study was to determine whether MCA-PSV can predict polycythemia in MCDA pregnancies.

Methods

This was a prospective cohort study of MCDA pregnancies recruited at 14–18 weeks' gestation from a single tertiary care center between January 2011 and June 2014. Fetal MCA Doppler waveforms were recorded every 2 weeks from 18 weeks' gestation until delivery. Only those with an MCA-PSV measurement within 1 week of delivery were included in the analysis. Neonatal hematocrit level was determined in all twins from venous blood obtained within 4 h of delivery. Polycythemia was defined as a hematocrit of > 65%, and anemia as a hematocrit of < 45%. TAPS was diagnosed when an intertwin hemoglobin difference of > 8 g/dL and reticulocyte count ratio of > 1.7 were observed.

Results

Of 162 MCDA pregnancies followed during the study period, 69 had an MCA-PSV measurement within 1 week of delivery and were included in the study. Twenty-five neonates were diagnosed with polycythemia and nine twin pairs met the criteria for TAPS. In a pooled analysis, MCA-PSV was negatively correlated with neonatal hematocrit (P = 0.017, r = −0.215) and was significantly higher in anemic fetuses than in normal controls (1.15 multiples of the median (MoM) vs 1.02 MoM, respectively; P = 0.001). However, MCA-PSV was similar among polycythemic and normal fetuses (0.95 MoM vs 1.02 MoM, respectively; P = 0.47). Intertwin difference in MCA-PSV (delta MCA-PSV) was positively correlated with intertwin hematocrit difference (P = 0.002, r = 0.394). Moreover, twin pregnancies with an intertwin hematocrit difference of > 24% had a significantly greater delta MCA-PSV than did those with an intertwin hematocrit difference of ≤ 24% (delta MCA-PSV, 19 vs 5 cm/s; P < 0.001).

Conclusions

MCA-PSV is not significantly decreased in polycythemic MCDA twins. However, delta MCA-PSV is associated with a large intertwin difference in hematocrit, and its use may be better than conventional methods for the risk assessment of TAPS. Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.

INTRODUCTION

Monochorionic–diamniotic (MCDA) twin pregnancies share a single placenta and nearly all have intertwin anastomoses connecting the fetal circulations, which may cause intertwin blood exchange1, 2. In most cases intertwin blood transfusion is balanced. However, an unbalanced net transfusion may occur and lead to various complications. Twin-to-twin transfusion syndrome (TTTS) is a well-established complication, with a reported incidence of 10–15% in MC twin pregnancies, characterized by severe amniotic fluid discordance, although similar hemoglobin levels are usually observed3, 4. Another recently described clinical complication of MCDA twin pregnancy is twin anemia–polycythemia sequence (TAPS). TAPS is caused by the slow net transfer of blood from one fetus to the other via a few tiny, and mostly unidirectional, anastomoses5, 6. It is characterized by large intertwin hemoglobin differences in the absence of amniotic fluid discordance7. TAPS can occur spontaneously in approximately 3–5% of ‘uncomplicated’ MC twin pregnancies8-10 or iatrogenically following laser ablation in up to 13% of cases11-14.

The antenatal diagnosis of TAPS is based on measurement of the middle cerebral artery peak systolic velocity (MCA-PSV). Diagnosis requires elevation of the MCA-PSV over 1.5 multiples of the normal median (MoM) in one twin with a concordant decrease of the MCA-PSV in the cotwin15-17. However, there is a debate regarding the correct threshold for the antenatal detection of fetal polycythemia. A MCA-PSV MoM of < 0.8 has been suggested by some experts12, whereas others have proposed a more conservative threshold of < 1.0 MoM as evidence of polycythemia15.

The postnatal diagnosis of TAPS is based on an intertwin hemoglobin difference of > 8 g/dL with donor reticulocytosis or typical placental angioarchitecture with small vascular anastomoses17.

The correlation between MCA-PSV and fetal anemia is well established18, 19. However, the use of MCA-PSV for the diagnosis of fetal polycythemia has not yet been validated20. The aim of this study was to determine whether fetal MCA-PSV can predict polycythemia in MCDA twin pregnancies.

METHODS

This was a prospective cohort study of women with an MCDA twin pregnancy who were recruited at 14–18 weeks' gestation from a single tertiary care center between January 2011 and June 2014. The study was approved by the local institutional ethics committee.

All MCDA twin pregnancies referred to our fetal medicine unit underwent a complete and thorough evaluation. Chorionicity was determined by first-trimester ultrasound examination. Each assessment included standard biometry to confirm gestational age and the size of each fetus, assessment of amniotic fluid volume in each sac, an anatomical survey of each fetus to rule out morphological anomalies and Doppler flow measurement (umbilical artery, MCA-PSV and ductus venosus). MCA-PSV measurements were obtained as previously described, with 1.5 MoM used as a cut-off value for diagnosing moderate or severe fetal anemia18. The MCA was identified using color or power Doppler ultrasound. The angle of insonation was kept as close to 0° as possible and never exceeded 30°. The sample volume was placed close to the internal fetal carotid artery. All ultrasound examinations were performed by one of four fetal medicine specialists (B.W., B.C., S.L., Y.Y.). In cases of uncomplicated MCDA twin pregnancy, subsequent routine antenatal evaluation was scheduled for every 2 weeks. If there was evidence of intrauterine growth restriction (IUGR), discordant fetal growth, discordant fluid volumes or discordant Doppler flow in the MCA-PSV, fetal surveillance was intensified accordingly. At our center, measurement of MCA-PSV is performed routinely in all MCDA twin pregnancies from 18 weeks' gestation until delivery. Only patients who had an MCA-PSV measurement within 1 week of delivery were included in our analysis. Of note, delayed cord clamping was not employed at our center during the study period.

Data collection

Data were obtained from our departmental electronic medical charts. The following demographic and obstetric variables were recorded: maternal age, parity, mode of conception, pregnancy complications, mode of delivery, gestational age at delivery and birth weight. The neonatal hematocrit was recorded from venous blood obtained within 4 h of delivery.

Definitions

Gestational age at delivery was determined from the last menstrual period (LMP) and was verified by the first ultrasound gestational-age measurement. If the LMP-estimated due date was not consistent with the due date obtained from first-trimester ultrasound growth measurements (i.e. difference > 7 days in the first trimester), then the latter date was used to define gestational age. In cases of in-vitro fertilization, gestational age was determined according to the date of embryo transfer. Polycythemia was defined as venous hematocrit > 65%, and anemia as hematocrit < 45%21. Diagnosis of TAPS was based on the postnatal criteria of an intertwin hemoglobin difference of > 8 g/dL and elevated reticulocyte count in the anemic twin15, 17.

Statistical analysis

Data analysis was performed with SPSS v 21.0 software (IBM SPSS Statistics for Windows, Version 21.0, IBM Corp., Armonk, NY, USA). Student's t-test and the Mann–Whitney U-test were used to compare continuous variables with and without normal distribution, respectively, between the groups. The chi-square and Fisher's exact tests were used for categorical variables. Spearman's correlation coefficient was used to assess the correlation between fetal MCA-PSV and neonatal hematocrit and between intertwin differences in MCA-PSV (delta MCA-PSV) and intertwin differences in hematocrit (delta hematocrit); P < 0.05 was considered to indicate a statistically significant difference.

RESULTS

During the study period, 162 MCDA twin pregnancies were followed at our center, of which 69 had an MCA-PSV measurement within 1 week of delivery and were included in our analysis. Of 69 MC pregnancies, four were complicated by intrauterine fetal death of one of the twins following laser therapy for TTTS and three underwent selective fetal reduction owing to selective IUGR Type II, resulting in 131 neonates available for evaluation (Figure   1). The demographic and clinical characteristics of the study population are shown in Table   1.

Details are in the caption following the image
Flowchart of inclusion of study population of monochorionic–diamniotic (MCDA) twin pregnancies with measurement of fetal middle cerebral artery peak systolic velocity (MCA-PSV) 1 week before delivery. IUFD, intrauterine fetal death.
Table 1. Demographic and clinical characteristics of 69 monochorionic–diamniotic twin pregnancies
Characteristic Value
Maternal age (years) 32.4 (22–43)
Parity 2.4 (1–8)
Mode of conception
Spontaneous 55 (79.7)
Ovulation induction 4 (5.8)
In-vitro fertilization 10 (14.5)
Pregnancy complication
Uncomplicated 30 (43.5)
TTTS 17 (24.6)
Selective IUGR 13 (18.8)
TAPS 9 (13.0)
GA at delivery (weeks) 33.6 (24.6–38.3)
Delivery < 32 weeks 11 (15.9)
Birth weight (g) 1878 (460–3445)
Mode of delivery
Cesarean section 35 (50.7)
Vaginal delivery 34 (49.3)
Male:female ratio 1:1
  • Data are given as median (range) or n (%).
  • GA, gestational age; IUGR, intrauterine growth restriction; TAPS, twin anemia–polycythemia sequence; TTTS, twin-to-twin transfusion syndrome.

Of the 69 pregnancies included in our study, 30 (43%) were uncomplicated, 17 (25%) were complicated by TTTS, 13 (19%) by selective IUGR and nine (13%) met the criteria of TAPS. Of the 17 cases with TTTS, three were categorized as Stage I, nine as Stage II and five as Stage III; 14 were treated by selective fetoscopic laser ablation. The neonatal hematological characteristics of the nine cases of TAPS are described in Table   2. TAPS in two of the cases developed after laser ablation and in seven it developed spontaneously. In six cases, TAPS was diagnosed prenatally based on MCA-PSV; four were managed expectantly and two were treated with intrauterine blood transfusion.

Table 2. Neonatal hematological measurements in nine twins with twin anemia–polycythemia sequence
Polycythemic twin Anemic twin
Hb (g/dL) Hct (%) Hb (g/dL) Hct (%) RC (%)
24 73 9.2 28 25
21 71 10.5 36 20.3
18.8 66 10.5 37 18
24 78 10.3 36 18.9
26 85 11.2 39 17
25.7 85 11.2 29 25
19 66 10.8 37 N/A
22.3 76 8.2 31 24
19.5 68 11.5 32 N/A
  • Hb, hemoglobin; Hct, hematocrit; N/A, not available; RC, reticulocyte count.

Of the 131 neonates evaluated, 84 (64%) had a normal hematocrit, 22 (17%) were anemic and 25 (19%) were polycythemic.

In a pooled analysis, MCA-PSV was negatively correlated with neonatal hematocrit (P = 0.017, r = −0.215; Figure   2a). As expected, MCA-PSV was significantly higher in anemic fetuses than in normal fetuses (1.15 MoM vs 1.02 MoM; P = 0.001). However, MCA-PSV was similar between polycythemic and normal fetuses (0.95 MoM vs 1.02 MoM; P = 0.47; Figure   2b). Of the 25 twins with polycythemia, 16 (64%) had MCA-PSV < 1 MoM prior to delivery and only three (12%) had MCA-PSV < 0.8 MoM. In the TAPS group, six of the polycythemic twins had MCA-PSV < 1 MoM and two of them had MCA-PSV < 0.8 MoM. Severe polycythemia with hematocrit > 70% was found in 12 twins, of whom three (25%) had an MCA-PSV > 1 MoM. In one of these three cases the neonatal hematocrit was 85%, with an MCA-PSV before delivery of 1.19 MoM, while the cotwin had a hematocrit at birth of 39%, with an MCA-PSV before delivery of 1.61 MoM.

Details are in the caption following the image
(a) Spearman's correlation between middle cerebral artery peak systolic velocity (MCA-PSV) multiples of the normal median (MoM), measured 1 week before delivery, and neonatal hematocrit in 131 monochorionic–diamniotic twins (r = −0.215, P = 0.017). (b) Boxplot of fetal MCA-PSV MoM according to anemic, normal and polycythemic neonatal hematocrit. Boxes with internal lines represent median and interquartile range and whiskers are range.

There was a positive correlation between the intertwin difference in MCA-PSV (delta MCA-PSV) and the intertwin difference in neonatal hematocrit (delta hematocrit) (P = 0.002, r = 0.394; Figure   3a). Moreover, twin pregnancies with an intertwin hematocrit difference of more than 24% (correlating with a hemoglobin difference of > 8 g/dL, as hematocrit is usually defined as three times the value of hemoglobin) had significantly higher delta MCA-PSV than did twin pregnancies with an intertwin hematocrit difference of ≤ 24% (delta MCA-PSV, 19 vs 5 cm/s; P < 0.001; Figure   3b). The performance of prenatal MCA-PSV in predicting neonatal anemia and polycythemia in MCDA twins and of delta MCA-PSV in predicting TAPS is shown in Figure   4. MCA-PSV in predicting neonatal anemia (area under the receiver–operating characteristics curve (AUC), 0.687 (95% CI, 0.547–0.827), P = 0.005; Figure   4) had a moderate performance, whereas delta MCA-PSV in predicting TAPS (AUC, 0.871 (95% CI, 0.757–0.985), P < 0.001; Figure   4) had a good performance. A relatively poor performance was seen for MCA-PSV in predicting neonatal polycythemia (AUC, 0.617 (95% CI, 0.505–0.728), P = 0.07; Figure   4).

Details are in the caption following the image
(a) Spearman's correlation between intertwin difference in middle cerebral artery peak systolic velocity (delta MCA-PSV) and intertwin difference in neonatal hematocrit (delta hematocrit) in 131 monochorionic–diamniotic twins (r = 0.394, P = 0.002). (b) Boxplot of delta MCA-PSV in twins with elevated delta hematocrit (> 24%) and those with normal delta hematocrit (≤ 24%). Boxes with internal lines represent median and interquartile range and whiskers are range.
Details are in the caption following the image
Receiver–operating characteristics (ROC) curves for accuracy of fetal middle cerebral artery peak systolic velocity (MCA-PSV), measured within 1 week of delivery in monochorionic–diamniotic twin pregnancies, for predicting neonatal anemia (area under ROC curve (AUC) = 0.687 (95% CI, 0.547–0.827); P = 0.005) (image) and neonatal polycythemia (AUC = 0.617 (95% CI, 0.505–0.728); P = 0.070) (image) and ROC curve for accuracy of intertwin difference in MCA-PSV for predicting twin anemia–polycythemia sequence (AUC = 0.871 (95% CI, 0.757–0.985); P < 0.001) (image).

DISCUSSION

In this study, we aimed to determine whether fetal MCA-PSV could predict polycythemia in MCDA twin pregnancies. Our findings indicate that, as expected, anemic MCDA twins have elevated MCA-PSV before delivery. However, MCA-PSV was not lower among polycythemic MCDA twins. Moreover, the intertwin difference in MCA-PSV was positively correlated with the intertwin difference in hematocrit after birth, and twins with a large intertwin hematocrit difference had a significantly greater intertwin discordance in MCA-PSV before delivery than did twins with a normal hematocrit difference.

The clinical entity TAPS, characterized by the presence of an intertwin hemoglobin difference, has raised a need for the antenatal detection of polycythemia in MCDA twins. The prenatal diagnosis of TAPS is based on the suspicion of anemia in one twin with concurrent evidence of polycythemia in the cotwin. Antenatal criteria for the diagnosis of TAPS have been proposed, including an MCA-PSV of > 1.5 MoM for the donor twin and an MCA-PSV of < 0.8 MoM for the recipient twin12. Others have suggested a new cut-off level for MCA-PSV of < 1.0 MoM for the recipient twin15. However, none of these cut-off levels has been validated, and the current study is the first to examine the value of MCA-PSV measurement in predicting polycythemia in MCDA twins. Based on our findings, it seems that the absence of reduced MCA-PSV does not exclude polycythemia, as 36% of the polycythemic twins in our study had an MCA-PSV of > 1 MoM, and only 12% of the polycythemic twins had an MCA-PSV of < 0.8 MOM. It is of note that the recipient twin in one of the TAPS cases included in this study was born with severe polycythemia, with a hematocrit of 85%, but the predelivery MCA-PSV was 1.19 MoM. Therefore, TAPS cannot be excluded even if MCA-PSV is not decreased in either twin. As the absolute value of MCA-PSV seems to be inappropriate for a prenatal diagnosis of fetal polycythemia and TAPS, other parameters should be sought in order to improve our ability to detect TAPS antenatally. The data of the current study indicate that the difference in intertwin MCA-PSV (delta MCA-PSV) might serve as a reliable tool for a prenatal diagnosis of TAPS, as it was positively correlated with the intertwin hematocrit difference. The postnatal diagnostic criteria for TAPS include an intertwin hemoglobin difference of > 8 g/dL15, 17. Consequently, we divided the twins in our study according to an intertwin hematocrit difference of more or less than 24%, correlating to a hemoglobin difference of > or < 8 g/dL, and found significantly greater discordant MCA-PSV among twins with an intertwin hematocrit difference of more than 24%. These findings reinforce our hypothesis that intertwin discordance in MCA-PSV might be a better tool than the absolute value of MCA-PSV in the risk assessment of TAPS.

The antenatal detection of TAPS is crucial, as it affects management of the pregnancy. The antenatal management of TAPS is not definitive and depends mainly on gestational age. Most experts agree that fetuses ≥ 32 weeks' gestation should be delivered. However, if it is detected before 30 weeks some have advocated considering transfusion to the anemic twin to prolong gestation22. Nevertheless, the main challenge is to manage antenatally the polycythemic twin rather than the anemic one. It is of note that transfusion to the anemic twin carries the risk of worsening the polycythemia with hyperviscosity in the cotwin. Therefore, some have suggested performing intraperitoneal transfusion, which results in slower absorption, or prenatal exchange transfusion for the polycythemic twin23. Moreover, severe polycythemia is associated with vascular limb injury as well as cerebral ischemia and hemorrhage24, 25. Hence, a non-invasive tool that could determine the severity of polycythemia, similarly to MCA-PSV, in the prenatal assessment of fetal anemia, would be of great benefit for the antenatal management of MC pregnancies complicated by TAPS. Our data are too preliminary to decide whether intertwin delta MCA-PSV could serve not only as a tool to detect TAPS but also to determine the severity of polycythemia in the recipient twin.

The strength of this study is that it was a prospective single-center study in which all patients were managed uniformly using a standardized surveillance protocol. Moreover, all MCA-PSV measurements were obtained by one of four well-trained fetal medicine specialists. However, given the relatively small sample size of our study and since our cohort included all MC twins referred to our unit but only nine cases of TAPS, larger studies are needed in order to confirm our findings and to validate the role of intertwin delta MCA-PSV in the prenatal diagnosis of TAPS. Furthermore, such studies may define cut-off values of delta MCA-PSV for the detection of TAPS. In addition, all MCA-PSV measurements were performed within 1 week of delivery, while major hematological changes can occur in MC twins during the last week of pregnancy, including at delivery. Consequently, the MCA-PSV measured 1 week prior to delivery does not always correspond with the neonatal hematocrit level.

In conclusion, the current study indicates that MCA-PSV is not significantly lower among polycythemic MC twins. However, intertwin difference in MCA-PSV is positively correlated with intertwin hematocrit difference and therefore may prove to be superior to conventional methods for the prenatal detection of TAPS.