The role of Doppler studies in predicting individual intrauterine fetal demise after laser therapy for twin–twin transfusion syndrome

Authors


Abstract

Objective

To investigate the role of Doppler studies in predicting individual fetal demise in patients scheduled for selective laser photocoagulation of communicating vessels (SLPCV) for twin–twin transfusion syndrome (TTTS).

Methods

Doppler studies of the umbilical artery, umbilical vein, ductus venosus, tricuspid valve regurgitation and middle cerebral artery were performed in the donor and recipient twins before and 24 hours after SLPCV. Results were analyzed cross-sectionally and longitudinally. As multiple comparisons were made, an a priori alpha rejection was set at P < 0.001.

Results

One hundred and ten consecutive patients were available for analysis. Overall fetal survival was 68.6% (151/220) with at least one survivor in 88.2% (97/110) of cases. Absent or reversed end-diastolic velocity in the umbilical artery of the donor twin was the only preoperative Doppler result predictive of intrauterine fetal demise (IUFD) (10/15, 66.7%, P < 0.001). Postoperatively, reversed flow during atrial contraction in the ductus venosus of the donor twin showed a trend towards prediction of IUFD of this fetus (4/5, 80%, P = 0.007). No other Doppler studies, including the longitudinal analyses, were predictive of IUFD.

Conclusions

Our data suggest that preoperative absent or reversed end-diastolic velocity in the umbilical artery may be useful in predicting individual fetal demise of the donor twin in TTTS patients scheduled for SLPCV. This may reflect the role of decreased individual placental mass that may be associated with some donor twins. The inability of other Doppler studies to predict individual IUFD may be explained preoperatively by the effect of the interfetal vascular connections on the individual Doppler signals and postoperatively by the effect of surgery or the timing of the assessment. Our findings may be important in patient counseling, in furthering understanding of the disease, and perhaps in improving surgical technique. Copyright © 2003 ISUOG. Published by John Wiley & Sons, Ltd.

Introduction

Twin–twin transfusion syndrome (TTTS) occurs in 10–15% of monochorionic multiple pregnancies. Untreated, it is associated with a high risk of perinatal morbidity and mortality1–3. The disease is thought to occur from chronic unbalanced transfusion between twins across placental vascular communications, resulting in a donor and a recipient twin. The donor twin is thought to be hypovolemic and therefore more likely to show Doppler abnormalities in the arterial system, whereas the recipient twin is thought to be hypervolemic and therefore more likely to show Doppler abnormalities in the venous system4.

We have proposed a sonographic staging classification of TTTS5 that has prognostic significance in patients treated with serial amniocenteses, but not for those treated with our selective laser photocoagulation of communicating vessels (SLPCV) technique. In particular, amniocentesis patients with critically abnormal Doppler studies (absent or reversed end-diastolic velocity in the umbilical artery, reversed flow in the ductus venosus, or pulsatile umbilical venous flow) (Stage III) or hydrops (Stage IV) are less likely to have a successful outcome and more likely to have neonatal neurological complications than are SLPCV patients6. Because surgery renders each fetus independent, and because the donor and the recipient twins are affected differently by the disease process, we hypothesized that individual outcomes could be predicted by Doppler studies. The aim of this study was to investigate whether preoperative or postoperative Doppler studies could have any prognostic value in the prediction of intrauterine fetal demise (IUFD) after SLPCV in TTTS patients. Additionally, changes in Doppler abnormality status pre- and postoperatively were explored for their ability to predict IUFD in this population.

Methods

Patients undergoing SLPCV for the treatment of TTTS were considered eligible for the study. Preoperative assessment consisted of a comprehensive ultrasound examination including fetal anatomy, biometry, amniotic fluid volume, placental location and umbilical cord insertions. Color and pulsed Doppler studies were performed 6–24 h before surgery and within 24 h after surgery. Doppler waveform analysis included the umbilical artery (UA) and vein (UV), middle cerebral artery (MCA), ductus venosus (DV) and tricuspid valve regurgitation (TR). For color and pulsed Doppler the energy output levels were lower than 50 mW/cm2 spatial peak temporal average, and the high-pass filter was set at the lowest level. Doppler measurements were obtained in the absence of fetal breathing and/or movements. Patients in this study have also been partially included in previous data sets5, 7–9.

Critically abnormal Doppler studies were defined as follows: absent or reversed end-diastolic blood flow velocities in the umbilical artery (UA-AREDV), pulsatility index of the middle cerebral artery (MCAPI) < 5th percentile, reversed blood flow during atrial contraction in the ductus venosus (DV-RF), pulsatile umbilical venous flow (PUVF) at the placental cord insertion or umbilical loop, and TR.

The severity of the syndrome at the time of presentation was assigned according to the staging classification previously reported5. The study was approved by the institutional review board of St. Joseph's Hospital in Tampa, Florida and each patient gave informed written consent.

The study outcome was IUFD. Cross-sectional analysis was performed on pre- and postoperative Doppler findings. Longitudinal analysis was performed in fetuses that were alive 24 h after surgery and in whom both pre- and post-Doppler studies had been performed. Statistical analysis was conducted using SPSS v10.0 for Windows 98 (SPSS, Chicago, IL, USA). Chi-square and Fisher's exact tests were used as appropriate. As multiple comparisons were made while conducting the above analysis, an a priori alpha rejection level was set at P < 0.001. A separate analysis was conducted to quantify the overall difference in the percentage of IUFDs between donor and recipient twins. The P-value for this secondary analysis was set at < 0.05.

Results

Perinatal outcome

One hundred and ten consecutive patients with TTTS who underwent SLPCV at the Florida Institute for Fetal Diagnosis and Therapy, St. Joseph's Women's Hospital between October 1997 and February 2001 were included in the study. The median gestational age at the time of the procedure was 20.7 (range, 16.1–25.7) weeks. The median gestational age at the time of delivery was 33 (range, 16.8–40.4) weeks. The overall perinatal survival rate was 68.6% (151/220). Of the 110 pregnancies, both fetuses survived in 49.1% (54/110), only one fetus survived in 39.1% (43/110), and neither survived in 11.8% (13/110) of cases. Therefore, at least one fetus survived in 88.2% (97/110) of the pregnancies.

IUFD of the donor twin occurred in 29.1% (32/110) of patients, whereas IUFD of the recipient twin occurred in 17.3% (19/110) of the patients. This difference between IUFD of donors or recipients was statistically significant (P = 0.003). Of patients with a single IUFD, 62% of donor twins (20/32) and 79% (15/19) of recipient twins died within the first 24 h after surgery, leaving 90 donor twins (81.8%) and 95 recipient twins (86.4%) for postoperative Doppler assessment 24 h after surgery. TTTS did not recur in 109/110 cases (99%). With dual demise cases excluded, TTTS did not recur in 96/97 cases (99%).

Doppler findings: cross-sectional analyses

Table 1 shows the preoperative Doppler findings for donor and recipient twins. Successful Doppler assessments of the vessels in the donor twin ranged from 100% (110/110) for the UA to 73.6% (81/110) for TR, with an overall assessment of 92.2% (507/550) of Doppler parameters in this fetus. The corresponding value for successful Doppler assessments in the recipient twin was 96.9% (533/550), ranging from 100% (110/110) for the UA to 92.7% (102/110) for TR. Donor twins were more likely to show critically abnormal Dopplers in the UA, while the rest of the Doppler parameters were more affected in the recipient twins. These differences between twin status were statistically significant (P < 0.001). Postoperative Doppler assessment was achieved in 92.6% of parameters (417/450) of available donors and in 93.8% of parameters (446/475) of available recipients. Postoperatively, critically abnormal venous Dopplers were more likely to be present in the recipient twin than in the donor twin (Table 2).

Table 1. Preoperative critically abnormal Doppler in donor and recipient twins
Doppler parameterDonor (n (%))Recipient (n (%))P
  1. UA-AREDV, absent or reversed end-diastolic velocity in the umbilical artery; MCAPI, middle cerebral artery pulsatility index < 5th percentile; DV-RF, reversed flow during atrial contraction in the ductus venosus; PUVF, pulsatile umbilical vein flow; TR, tricuspid valve regurgitation.

UA-AREDV15/110 (13.6)1/110 (0.9)< 0.001
MCAPI27/102 (26.5)55/105 (52.4)< 0.001
DV-RF2/108 (1.9)30/109 (27.5)< 0.001
PUVF4/106 (3.8)33/107 (30.8)< 0.001
TR1/81 (1.2)37/102 (36.3)< 0.001
Table 2. Postoperative critically abnormal Doppler in donor and recipient twins
Doppler parameterDonor (n (%))Recipient (n (%))P
  1. UA-AREDV, absent or reversed end-diastolic velocity in the umbilical artery; MCAPI, middle cerebral artery pulsatility index < 5th percentile; DV-RF, reversed flow during atrial contraction in the ductus venosus; PUVF, pulsatile umbilical vein flow; TR, tricuspid valve regurgitation.

UA-AREDV4/90 (4.4)1/94 (1.1)0.16
MCAPI32/81 (39.5)25/91 (27.5)0.09
DV-RF5/90 (5.5)25/95 (26.5)< 0.001
PUVF4/89 (4.5)21/95 (22.1)< 0.001
TR4/67 (6.0)24/71 (33.8)< 0.001

Preoperative UA-AREDV of the donor twin was associated with a significantly higher risk of IUFD of this fetus (10/15, 66.7% vs. 22/95, 23.2%, P < 0.001). The remaining preoperative Doppler parameters did not predict IUFD in either twin (Table 3). Preoperative critically abnormal Dopplers in one twin were not associated with IUFD of the other twin (data not shown).

Table 3. Cross-sectional analysis: intrauterine fetal demise (IUFD) of donor and recipient twins according to preoperative critically abnormal Doppler
Preoperative Doppler parameterIUFD of donor (n (%))IUFD of recipient (n (%))
AbnormalNormalPAbnormalNormalP
  1. UA-AREDV, absent or reversed end-diastolic velocity in the umbilical artery; MCAPI, middle cerebral artery pulsatility index < 5th percentile; DV-RF, reversed flow during atrial contraction in the ductus venosus; PUVF, pulsatile umbilical vein flow; TR, tricuspid valve regurgitation.

UA-AREDV10/15 (66.7)22/95 (23.2)< 0.0010/119/109 (17.4)0.64
MCAPI 8/27 (29.6)23/75 (30.7)0.929/55 (16.4) 8/50 (16.0)0.96
DV-RF 0/232/106 (30.2)0.355/30 (16.7)13/79 (16.5)0.97
PUVF 1/4 (25.0)29/102 (28.4)0.889/33 (27.3) 7/74 (9.5)0.017
TR 0/121/80 (26.3)0.558/37 (21.6) 9/65 (13.8)0.31

Postoperative Doppler studies showed a trend of DV-RF in the donor towards prediction of IUFD (4/5, 80% vs. 16/85, 18.4%, P = 0.007). The remaining Doppler parameters did not predict IUFD in either twin (Table 4). Critically abnormal postoperative Doppler findings in one twin were not associated with an increased likelihood of IUFD of the other twin (data not shown).

Table 4. Cross-sectional analysis: intrauterine fetal demise (IUFD) of donor and recipient twins according to postoperative critically abnormal Doppler
Postoperative Doppler parameterIUFD of donor (n (%))IUFD of recipient (n (%))
AbnormalNormalPAbnormalNormalP
  1. UA-AREDV, absent or reversed end-diastolic velocity in the umbilical artery; MCAPI, middle cerebral artery pulsatility index < 5th percentile; DV-RF, reversed flow during atrial contraction in the ductus venosus; PUVF, pulsatile umbilical vein flow; TR, tricuspid valve regurgitation.

UA-AREDV2/4 (50.0)15/86 (17.4)0.100/18/93 (8.6)0.75
MCAPI6/32 (18.8) 8/49 (16.3)0.771/25 (4.0)8/66 (12.1)0.24
DV-RF4/5 (80.0)16/85 (18.8)0.0072/25 (8.0)7/70 (10.0)0.76
PUVF0/418/85 (21.2)0.305/21 (23.8)5/74 (6.8)0.025
TR1/4 (25.0) 8/63 (12.7)0.483/24 (12.5)3/47 (6.4)0.38

Doppler findings: longitudinal analyses

Table 5 shows the percentages of IUFD in the donor twin relative to sequential Doppler changes. Ninety donors were alive at 24 h after the procedure. There was a trend for UA-AREDV at either pre- or postoperative examination towards prediction of IUFD of the donor twin (7/13, 53.8% vs. 10/77, 12.9%, P = 0.002). No other Doppler changes were predictive of IUFD in this fetus.

Table 5. Longitudinal analysis: intrauterine fetal demise (IUFD) of the donor according to pre- and postoperative sequential Doppler changes
Doppler parameterIUFD (n (%)) with sequential Doppler change:
N–NP–NN–PP–PAny
  1. Doppler sequence changes: N–N, normal Doppler results pre- and postoperatively; P–N, postoperative normalization of a previously abnormal Doppler; N–P, postoperative worsening of a previously normal Doppler; P–P, abnormal Doppler results pre- and postoperatively; Any, an abnormal Doppler result either pre- or postoperatively. IUFD, intrauterine fetal demise; UA-AREDV, absent or reversed end-diastolic velocity in the umbilical artery; MCAPI, middle cerebral artery pulsatility index < 5th percentile; DV-RF, reversed flow during atrial contraction in the ductus venosus; PUVF, pulsatile umbilical vein flow; TR, tricuspid valve regurgitation.

UA-AREDV10/77 (12.9)5/9 (55.6)1/2 (50.0)1/2 (50.0)7/13 (53.8)
MCAPI 5/31 (16.1)2/12 (16.7)4/23 (17.4)2/9 (22.2)8/44 (18.2)
DV-RF18/85 (21.1) —4/5 (80.0) —4/5 (80.0)
PUVF16/81 (19.8)0/10/4 —0/5
TR 6/51 (11.8) —1/4 (25.0) —1/4 (25.0)

Table 6 shows the percentages of IUFD of the recipient twin relative to Doppler changes. Ninety-five recipients were alive at 24 h after the procedure. PUVF in either the pre- or postoperative examination was associated with a tendency towards IUFD of the recipient twin (7/32, 21.8% vs. 1/61, 1.6%, P = 0.002).

Table 6. Longitudinal analysis: intrauterine fetal demise (IUFD) of the recipient according to pre- and postoperative sequential Doppler changes
Doppler parameterIUFD (n (%)) with sequential Doppler change:
N–NP–NN–PP–PAny
  1. Doppler sequence changes: N–N, normal Doppler results pre- and postoperatively; P–N, postoperative normalization of a previously abnormal Doppler; N–P, postoperative worsening of a previously normal Doppler; P–P, abnormal Doppler results pre- and postoperatively; Any, an abnormal Doppler result either pre- or postoperatively. IUFD, intrauterine fetal demise; UA-AREDV, absent or reversed end-diastolic velocity in the umbilical artery; MCAPI, middle cerebral artery pulsatility index < 5th percentile; DV-RF, reversed flow during atrial contraction in the ductus venosus; PUVF, pulsatile umbilical vein flow; TR, tricuspid valve regurgitation.

UA-AREDV8/93 (8.6) —0/10/10/2
MCAPI2/31 (6.5)5/32 (15.6)0/81/16 (6.3)6/56 (10.7)
DV-RF7/63 (11.1)0/70/72/18 (11.1)2/32 (6.2)
PUVF1/61 (1.6)2/11 (18.2)1/5 (20.0)4/16 (25.0)7/32 (21.8)
TR3/40 (7.5)0/40/43/19 (15.8)3/27 (11.1)

Discussion

Our data suggest that Doppler studies may have a limited role in predicting postoperative IUFD in TTTS patients treated with SLPCV. The predictive role was limited to the donor twin and to the UA. No other Doppler parameters were significantly associated with a greater likelihood of IUFD in the donor or in the recipient twin. These results may have important implications in our understanding of the effect of surgery on the hemodynamic status of the fetuses and other potential factors that may determine in-utero survival.

Because our surgical technique (SLPCV) essentially converts the original monochorionic twin pregnancy into a ‘functional dichorionic’ gestation, we hypothesized that one could use Doppler to predict the individual outcome of the fetuses. This argument is based on the elimination of the untoward effect of IUFD on the co-twin by interrupting the vascular communications between the fetuses. To eliminate confounding effects relative to gestational age at delivery or neonatal complications, we chose in-utero demise as the outcome variable. However, two issues became apparent during the interpretation of the data. First, the fetuses and their Doppler signals are interdependent prior to surgery. Thus, normal or critically abnormal Doppler signals may depend on the preoperative interaction between the fetuses (e.g., a fetus may have a normal Doppler signal because of the presence of vascular communications!). Second, surgery may exert a profound effect on the hemodynamic status of the fetuses, such that postoperative Doppler signals may show improvement, worsening, or no change relative to the preoperative assessment. This could result in improvement or worsening in the status of either twin. Thus, an apparently healthier fetus may deteriorate after surgery, or an apparently sick twin may recover postoperatively. Because neither of these effects (preoperative fetal interaction/postoperative surgical effect) can be predicted before surgery, our finding of a limited value for Doppler in predicting IUFD after SLPCV is not surprising.

Overall, 31.4% of the fetuses died in-utero. It is conceivable that the excess IUFD (12%) of donors over recipients (29% vs. 17%) may represent the incidence of concomitant placental insufficiency in this fetus. Of patients with a single IUFD, 62% of donor twins and 79% of recipient twins died within the first 24 hours after surgery. This is likely to be a direct but unavoidable effect of surgery. The lack of untoward effect of IUFD of one twin on the co-twin attests to the complete separation of the vascular anastomoses between the fetuses.

Preoperative UA-AREDV was present in 13.6% (15/110) of donors. This finding was associated with an increased likelihood of postoperative IUFD of this fetus (66.7%, P < 0.001). UA-AREDV has been previously reported as an important predictor of fetal death, particularly of the donor twin, both in patients treated with amnioreduction10, 11 and in those treated by laser12, 13. Because laser introduces a change in the natural history of the disease, the value of Doppler in predicting IUFD of the donor twin in patients treated surgically may not necessarily be comparable to those treated symptomatically. Two papers have correlated fetal outcome with Doppler assessment in patients treated with laser. In a series of 132 cases of TTTS treated by endoscopic laser, Ville et al.12 found 12% of cases with UA-AREDV which were associated with an increased risk of perinatal death of donors (75%, P = 0.03). No other vessel was studied in this paper. Zikulnig et al.13 reported 121 cases of TTTS treated by endoscopic laser surgery, in which 19% of donor twins had UA-AREDV. Although there appeared to be a trend towards lower survival (43% vs. 61%, P = 0.16), they did not find this difference to be statistically significant. It should be noted that both studies assessed neonatal outcome, while our study refers to IUFD. Our results are in line with those of Zikulnig et al.13, who reported a 50% death rate within 24 h of the procedure in this subgroup (donors with UA-AREDV) of patients. Conversely, of our original 15 donors with UA-AREDV, 11 (73%) were alive 24 h after surgery. Of these, nine (82%) showed postoperative end-diastolic flow, but of these, five (55.6%) subsequently died in utero. Thus, 4/15 (26.7%) fetuses showed improvement in UA Doppler and survived. Technically, surgery cannot ‘add’ any additional placenta to the donor twin, except perhaps in cotyledons perfused by an artery and a vein of the donor also drained by a vein of the recipient (in which case, the vein of the recipient is lasered). Therefore, postoperative appearance of diastolic blood flow in fetuses with preoperative UA-AREDV suggests that preoperative UA-AREDV may not necessarily result from placental insufficiency, as commonly thought. Rather, UA-AREDV may indeed reflect hypotension secondary to the types of vascular anastomoses present, with or without placental insufficiency of the donor twin.

Postoperative Doppler assessments were limited by interim IUFDs. Although the difference did not reach statistical significance, donors with postoperative DV-RF had an increased tendency to die in utero (4/5, 80% vs. 16/85, 18.8% P = 0.007). Thus, it is possible that postoperative worsening of a donor's DV flow to critically abnormal levels may not be considered a benign or transient sign of increased preload14, but instead a potential marker for postoperative IUFD of the donor twin. This hypothesis needs further investigation.

PUVF was detected preoperatively in 30.8% of the recipient twins. IUFD occurred in nine of these recipients after surgery, this was not statistically significant (P = 0.017) at the <0.001 level. Demise of a recipient twin with preoperative PUVF could be explained by one of two mechanisms. First, the timing of the surgery may be such that it is performed beyond the point of no return in a terminally ill fetus. Alternatively, further hemodynamic overload of the recipient may result from intraoperative transfusion secondary to laser obliteration of recipient-to-donor before donor-to-recipient arteriovenous communications. Proving either of these hypotheses is virtually impossible. However, the latter concept may suggest that in patients with PUVF of the recipient twin, the donor-to-recipient arteriovenous communications should be obliterated first, followed by the recipient-to-donor communications, to prevent intraoperative overload of the recipient twin. Although DV-RF and TR were often detected, these parameters did not correlate with IUFD of the recipient (Tables 1 and 3).

The lack of predictive value of MCA Dopplers is perplexing. MCA evaluation was reported by Suzuki et al.15 to be a useful method to assess differential diagnoses between selective growth restriction in monochorionic twins and TTTS. They found a lower pulsatility index in the case of selective growth-restricted twins in comparison with normal or even higher impedance when TTTS was present. Our results do not support their findings, perhaps because of the small number of cases of TTTS (11) in their report. Although in our data a high number of fetuses, particularly recipients, had a MCAPI < 5th percentile, this parameter did not have any influence on fetal outcome. Others have also reported a decreased MCAPI in recipient twins16. We speculate that chronic blood exchange and imbalanced hemodynamic status may elicit a ‘brain-sparing effect’ as a compensating mechanism, but without the classical meaning and poor prognosis that has been reported in intrauterine growth restriction from placental insufficiency.

Our study adds a new dimension to the role of Doppler studies in the assessment of patients with TTTS. Doppler studies may help predict and understand postoperative outcomes and may point to the potential contribution of concomitant factors, such as placental insufficiency or cardiac overload, in the incidence of IUFD. Our results may also suggest the possibility of considering individualization of the lasering sequence, to decrease the likelihood of intraoperative transfusion and its potential deleterious effects on either of the twins. Because approximately 27% of donor twins with preoperative UA-AREDV may show postoperative improvement and survive after SLPCV, we would caution about resorting to umbilical-cord occlusion of this fetus as a primary management choice for these patients.

Ancillary