Objective
We sought to compare intrauterine risks with postnatal outcome in monochorionic pregnancies operated by fetoscopic laser surgery for twin-to-twin transfusion syndrome.
Study Design
A cohort of 602 consecutive cases was analyzed. Unexpected prenatal adverse events were identified when a fatal or potentially fatal event occurred that could have been avoided by timely delivery.
Results
The prospective risk of an unexpected adverse event dropped from 16.8% (95% confidence interval [CI], 13.6–20.5%) to 0% (95% CI, 0–11%) between 26-36 weeks. At 32 weeks, the residual risk was 1 in 17 (95% CI, 1/28–1/11). The perinatal rate of death or severe brain lesions dropped from 35% (25-47%) in infants delivered at 26-28 weeks down to 3% (1-6%) at 34-36 weeks.
Conclusion
Our results did not identify an optimal cut-off for elective preterm delivery in laser-operated twin-to-twin transfusion syndrome. Perinatal morbidity appears low from ≥32 weeks and the decision for elective delivery should be based upon medical history, parental demand, and expert assessment.
Despite significant improvements in the overall prognosis, fetoscopic laser coagulation for twin-to-twin transfusion syndrome (TTTS) carries a high risk of postoperative complications such as fetal demise, miscarriage, anemia-polycythemia sequence (TAPS), or recurrence. In most cases, these complications occur shortly after surgery and are therefore expected, if not predictable. The consequence of these complications is an overall reduced survival rate compared to uncomplicated monochorionic pregnancies.
In uncomplicated monochorionic pregnancies, high rates of late stillbirths have prompted a policy of elective preterm birth as early as 32 weeks. Such policy, however, has not been supported by the results of other studies. Following laser surgery for TTTS, the management and timing of delivery may consider 2 opposite options: (1) that these pregnancies are still at high risk up until late gestation because of possible late unpredictable complications; or (2) that surgery has reduced the likelihood of such late events and that they could be managed as dichorionic pregnancies.
Optimal management, therefore, involves a balance between the risk of intrauterine adverse events and the consequences of elective preterm birth. Moreover, newborns following TTTS have been shown to carry a high rate of neurological impairment. Therefore, in the absence of relevant decision-making results, one may favor the reduction of unnecessary preterm birth or favor the prevention of potential late unpredictable complications. In this study, we analyzed the determinants of the balance of these risks for the timing of delivery following fetoscopic laser surgery by analyzing the time-wise occurrence of unexpected severe complications concurrently with the postnatal mortality and morbidity in a single tertiary care center over a 6-year period.
Materials and Methods
Patients
All consecutive cases referred to our center for TTTS operated upon by selective laser coagulation of anastomotic vessels from 16-26 weeks were reviewed over a 6-year period (June 2004 through September 2010). Surgery was performed under fetoscopic control as described elsewhere. Patients were discharged 48 hours after surgery and followed up until delivery on a weekly basis in our institution or the referral center, allowing for early detection and accurate timing of any subsequent complication. We considered the time of the follow-up visit as a proxy for event time unless calendar time was precisely known. Data including demographic variables, follow-up reports, and potential complications were recorded prospectively using a dedicated database for prenatal (Astraia Software GmbH; Astraia, Munich, Germany) and postnatal (FileMaker Pro v5; Filemaker, Inc., Santa Clara, CA) follow-up.
Prenatal outcomes
Since the purpose of this study was to yield pragmatic decision-making results regarding the timing of delivery, we considered postoperative prenatal events as potentially expected or unexpected regardless of the underlying mechanisms of each type of complication. Unexpected avoidable events are defined as any potentially fatal event that would have been efficiently avoided by timely elective delivery, regardless of gestational age. On the contrary, expected complications are defined as those known to occur at a high rate, directly related to surgery and occurring within early postoperative follow-up, such as intrauterine fetal demise (IUFD) of 1 twin <7 days following surgery. Each complication was classified as expected or unexpected according to the following predefined rules: unexpected complications included TAPS, recurrence of TTTS, IUFD of 1 twin occurring >7 days after surgery without severe intrauterine growth restriction, IUFD of both twins, chorioamnionitis, abruptio placentae, severe vaginal bleeding during follow-up, or acute fetal distress during follow-up. Any other complication (eg, IUFD of 1 twin <7 days following surgery, IUFD in the context of severe growth restriction, or prenatal brain damage requiring cord coagulation) was considered as “expected” or “anticipated.” Similarly, to adhere to our definition of unexpected avoidable complications, preterm premature rupture of the membranes and miscarriage were not included since they cannot be prevented by a timely delivery (discussed below).
Postnatal outcomes
Postnatal follow-up was conducted until 28 days of life or discharge in all cases. Early neurological damage was defined as intraventricular hemorrhage stage >3 or cystic periventricular leukomalacia on transfontanellar ultrasound or magnetic resonance imaging. A composite outcome was considered as death or brain damage diagnosed by 28 days of life or discharge from the neonatal intensive care unit.
Model and analysis
The analysis of prenatal outcomes was conducted using time-to-event methods. For a given patient, prenatal outcome was any of 3 competing events: (1) occurrence of an unexpected complication; (2) spontaneous labor before the occurrence of any potential unexpected complication; or (3) elective delivery by induction or cesarean section before the occurrence of any potential unexpected complication. Of note, miscarriage is included in option 2 since it is defined as labor <24 weeks. Only the first chronologically unexpected event is considered in case of multiple events during the prenatal period. We considered date of pregnancy as the time origin. Cumulative incidences are computed for each of the 3 possible competing outcomes. Prospective risks and event rate per week of gestation were computed for pregnancies >26 weeks of gestation without unexpected complications. Prospective risks were computed as the ratio of the total number of adverse events following the onset of week k divided by the number of pregnancies at risk at onset of week k . Event rates were computed as the ratio of the number of events during week k divided by the number of pregnancies at risk at onset of week k . We also considered prospective risks and event rates in ongoing pregnancies with 1 and 2 survivors at 26 weeks separately. Confidence intervals (CIs) for proportions were computed using the exact binomial distribution. All analyses were conducted using R v2.13 with package mstate.
Results
Over the study period, 648 cases of TTTS underwent operation between 16-26 weeks of gestation. Forty-six cases were lost to follow-up, leaving 602 cases for analysis. Quintero stages were distributed as follows: 14%, stage 1; 38%, stage 2; 45%, stage 3; and 3%, stage 4. Median gestational age at treatment was 21 weeks (interquartile range, 19–23 weeks), with 17% of cases ≤18 weeks’ gestation. At the time of surgery, the placenta was considered anterior in 47% of cases with 19% of transplacental fetoscopic entries. These demographic findings did not differ significantly in the 46 cases lost to follow-up, except for a significant difference in the proportions of Quintero stages with an increase in higher stages (5%, 41%, 45%, and 9% for Quintero stages 1-4, respectively, in the group lost to follow-up, P = .045 for the comparison with complete cases, Fisher exact test). The overall survival according to 0, 1, or 2 twins at birth was 79 (13%), 189 (32%), and 334 (55%), respectively, yielding an 87% rate of survival of at least 1 twin. Loss of 1 twin was related to IUFD in 142 (24%) cases and cord coagulation was performed as a second-line treatment for subsequent complications in 47 (8%) cases. Overall loss of both twins was related to miscarriage in 31 (5%) cases and dual fetal demise in 48 (8%) cases.
Postnatal outcomes
In all, 523 pregnancies with at least 1 live twin delivered from ≥24 weeks, yielding 857 newborns. Between birth and discharge from the neonatal intensive care unit, 10/189 (5%) and 44/668 (7%) twins died in the groups of single-survivor and dual-surviving twins at birth, respectively. Within the overall population of twins surviving at 28 days, the rate of early brain damage defined as intraventricular hemorrhage grade ≥3 or cystic periventricular leukomalacia was 23/803 (3%). Table 1 and Figure 1 present the rates of postnatal composite outcome defined by death or early brain damage at 28 days across gestational age at delivery: death or early brain damage occurred in 55% (34-74%) of infants delivered between 24-26 weeks and decreased rapidly to 5% (1-15%) for those delivered ≥36 weeks.
| Gestational age at delivery (2-wk blocks) | n pregnancies delivered | n fetuses delivered | Death or early brain injury at 28 d, % (95% CI) |
|---|---|---|---|
| 24–25 | 12 | 20 | 55 (34–74) |
| 26–27 | 39 | 69 | 35 (25–47) |
| 28–29 | 55 | 88 | 10 (5–18) |
| 30–31 | 80 | 140 | 7 (4–13) |
| 32–33 | 127 | 214 | 5 (3–8) |
| 34–35 | 175 | 283 | 4 (2–7) |
| ≥36 | 35 | 43 | 5 (1–15) |
Prenatal events
Unexpected fatal or potentially fatal events occurred in 196/602 (33%) pregnancies initially undergoing operation and comprised dual fetal demise in 48 (8%) cases, unanticipated fetal death >7 days following surgery in 8 (1%) cases, recurrence of TTTS in 51 (8.5%) cases, anemia or TAPS in 48 (8%) cases, chorioamnionitis in 24 (4%) cases, abruptio placentae in 13 (2%) cases, and spontaneous acute fetal distress in 20 (3.5%) cases. Because some patients had experienced 2 subsequent adverse events, complications do not add up to the total number of unexpected complications. In those cases, only the first chronological event is recorded to determine the overall rate of unexpected events. The following visit was used as a proxy for event time in 50/196 (26%) cases. On the pregnancy time line, the incidence of prenatal unexpected severe complications rises steadily from 16 weeks up to 34 weeks, reaching a total incidence of 33%, and is stable thereafter ( Figure 2 ), meaning that the rate of complications past that time is low. However, 1 patient presented with dual fetal demise at 34 weeks in an otherwise uncomplicated pregnancy, 1 patient was delivered in emergency because of overt chorioamnionitis at 34 weeks, and 1 patient presented with late recurrence at 35+3 weeks despite initial laser surgery at 21 weeks. The stacked incidences of unexpected complications, spontaneous delivery, and elective delivery in Figure 2 demonstrate how unexpected events compete against delivery either elective or spontaneous. The area defining the incidence of spontaneous labor increases from ≥16 weeks with a rate of miscarriage defined by spontaneous labor <24 weeks of about 5% as presented earlier. At 32 weeks, 16% of patients will have delivered spontaneously prior to any unexpected adverse event. Because of the impact of our preventive practice aiming to deliver pregnancies electively no later than 34-35 weeks, the rate of elective delivery by induction or cesarean section shows a sharp increase at around 32-33 weeks. Overall, by the end of prenatal follow-up, 33% of patients had experienced an unexpected complication, 25% had delivered spontaneously without unexpected complications (including 5% <24 weeks), and 42% delivered electively without unexpected complications.
Table 2 shows the prospective risks of unexpected adverse events by 1-week blocks in patients reaching 26 weeks without a prior unexpected adverse event. The initial number of ongoing pregnancies at risk at 26 weeks was 447 and dropped to 32 at ≥36 weeks. Conversely, the weekly rate of spontaneous labor increased rapidly from ≥35 weeks, reaching 61.5% (36-82%) at 38 weeks. The rate of unexpected adverse events was found stable from 26-33 weeks with a 2-3% rate of complications per week. Concurrently, the prospective risk of a subsequent unexpected adverse event at 26 weeks waned from 16.8% (95% CI, 13.6–20.5%) or 1 in 6 (95% CI, 1/7–1/5) to 0% (95% CI, 0–11%) at ≥36 weeks. At 32 weeks, the prospective risk was 1 in 17 (95% CI, 1/28–1/11), meaning that approximately 17 elective deliveries would be required to avoid 1 unexpected adverse event.
| Wk | At-risk per 1-wk block, n | Unexpected adverse events per 1-wk block, n | Rate of unexpected adverse events, % | Unexpected events in ongoing pregnancy, n | Prospective risk of unexpected adverse events, % | Spontaneous labor per 1-wk block, n | Rate of spontaneous labor, % |
|---|---|---|---|---|---|---|---|
| 26 | 447 | 9 | 2 | 75 | 16.8 | 6 | 1.3 |
| 27 | 431 | 10 | 2.3 | 66 | 15.3 | 11 | 2.6 |
| 28 | 410 | 14 | 3.4 | 56 | 13.7 | 9 | 2.2 |
| 29 | 387 | 4 | 1 | 42 | 10.9 | 12 | 3.1 |
| 30 | 367 | 8 | 2.2 | 38 | 10.4 | 8 | 2.2 |
| 31 | 343 | 13 | 3.8 | 30 | 8.7 | 15 | 4.4 |
| 32 | 296 | 6 | 2 | 17 | 5.7 | 14 | 4.7 |
| 33 | 251 | 8 | 3.2 | 11 | 4.4 | 15 | 6 |
| 34 | 187 | 2 | 1.1 | 3 | 1.6 | 6 | 3.2 |
| 35 | 70 | 1 | 1.4 | 1 | 1.4 | 5 | 7.1 |
| 36 | 32 | 0 | 0 | 0 | 0 | 2 | 6.2 |
| 37 | 25 | 0 | 0 | — | — | 3 | 12 |
| 38 | 13 | 0 | 0 | — | — | 8 | 61.5 |
Since the likelihood of fetal complications is strongly related to the number of fetuses, rates of unexpected event were broken down according to the number of live fetuses in ongoing pregnancies at 26 weeks. Table 3 and Figure 3 present the prospective risks according to the number of live fetuses at 26 weeks. The initial number of at-risk pregnancies at 26 weeks was N = 133/447 (30%) for pregnancies reaching 26 weeks as singletons and N = 314/447 (70%) for those reaching 26 weeks as twins. In pregnancies reaching 26 weeks as singletons the cesarean rate in eventless pregnancies was 75/119 (63%), whereas it was 221/253 (87%) in pregnancies reaching 26 weeks as twins and remaining eventless until delivery. Expectedly, the prospective risk of subsequent unexpected events was twice as high in twins compared to pregnancies reduced to singletons (19.4% [15-24%] and 10.5% [6-17%], respectively). However, Table 3 and Figure 3 show that in both cases the risk reached 0% by 36 weeks, although earlier in singletons. Nonetheless, >36 weeks, the results should be interpreted cautiously since the numbers of ongoing pregnancies are low in each group.
