Background
Stage I twin-twin transfusion syndrome presents a management dilemma. Intervention may lead to procedure-related complications while expectant management risks deterioration. Insufficient data exist to inform decision-making.
Objective
The aim of this retrospective observational study was to describe the natural history of stage I twin-twin transfusion syndrome, to assess for predictors of disease behavior, and to compare pregnancy outcomes after intervention at stage I vs expectant management.
Study Design
Ten North American Fetal Therapy Network centers submitted well-documented cases of stage I twin-twin transfusion syndrome for analysis. Cases were retrospectively divided into 3 management strategies: those managed expectantly, those who underwent amnioreduction at stage I, and those who underwent laser therapy at stage I. Outcomes were categorized as no survivors, 1 survivor, 2 survivors, or at least 1 survivor to live birth, and good (twin live birth ≥30.0 weeks), mixed (single fetal demise or delivery between 26.0-29.9 weeks), and poor (double fetal demise or delivery <26.0 weeks) pregnancy outcomes. Outcomes were analyzed by initial management strategy.
Results
A total of 124 cases of stage I twin-twin transfusion syndrome were studied. In all, 49 (40%) cases were managed expectantly while 30 (24%) underwent amnioreduction and 45 (36%) underwent laser therapy at stage I. The overall fetal mortality rate was 20.2% (50 of 248 fetuses). Of those managed expectantly, 11 patients regressed (22%), 4 remained stage I (8%), 29 advanced in stage (60%), and 5 experienced spontaneous previable preterm birth (10%) during observation. The mean number of days from diagnosis of stage I to a change in status (progression, regression, loss, or delivery) was 11.1 (SD 14.3) days. Intervention by amniocentesis or laser therapy was associated with a lower risk of fetal loss ( P = .01) than expectant management. The unadjusted odds of poor outcome were 0.33 (95% confidence interval, 0.09–01.20), for amnioreduction and 0.26 (95% confidence interval, 0.09–0.77) for laser therapy vs expectant management. Adjusting for nulliparity, recipient maximum vertical pocket, gestational age at diagnosis, and placenta location had negligible effect. Both amnioreduction and laser therapy at stage I decreased the likelihood of no survivors (odds ratio, 0.11; 95% confidence interval, 0.02–0.68 and odds ratio, 0.07; 95% confidence interval, 0.01–0.37, respectively). Only laser therapy, however, was protective against poor outcome in our data (odds ratio, 0.29; 95% confidence interval, 0.07–1.30 for amnioreduction vs odds ratio, 0.12, 95% confidence interval, 0.03–0.44 for laser), although the estimate for amnioreduction suggests a protective effect.
Conclusion
Stage I twin-twin transfusion syndrome was associated with substantial fetal mortality. Spontaneous resolution was observed, although the majority of expectantly managed cases progressed. Progression was associated with a worse prognosis. Both amnioreduction and laser therapy decreased the chance of no survivors, and laser was particularly protective against poor outcome independent of multiple factors. Further studies are justified to corroborate these findings and to further define risk stratification and surveillance strategies for stage I disease.
Introduction
Twin-twin transfusion syndrome (TTTS) occurs in a minority of monochorionic twin pregnancies, and when present is associated with substantial risk for perinatal death, preterm birth, and neurologic impairment among survivors. While advanced-stage disease is associated with a significant burden of morbidity and mortality, insufficient data exist regarding the behavior of TTTS when diagnosed in its earliest stage. It is widely accepted that early disease can demonstrate a variable course, with some cases spontaneously resolving and others experiencing rapid deterioration.
Perhaps related to its variable natural history, controversy exists regarding optimal management of Quintero stage I TTTS. While a randomized controlled trial has provided level I evidence that selective fetoscopic laser photocoagulation offers effective treatment across all stages of disease, uncontrolled experiences have yielded conflicting results regarding whether laser therapy benefits specifically apply to stage I TTTS. Furthermore, a recent systematic review of current literature suggested that as an initial strategy for stage I TTTS, conservative management (observation) and laser photocoagulation were associated with essentially equivalent outcomes and were superior to amnioreduction.
Better defining the natural course of stage I disease could help to identify patients who are more likely to progress to advanced stages, as well as characterize how rapidly that progression may occur.
Our goal was to better define the behavior of stage I TTTS through a multicenter retrospective observational study design. The specific aims of this study were to investigate the natural history of stage I TTTS after diagnosis, to compare pregnancy outcomes of those managed expectantly with those managed at stage I by amnioreduction or laser therapy, and to evaluate demographic and sonographic features for their ability to predict disease behavior.
Materials and Methods
The North American Fetal Therapy Network is a consortium of 30 medical institutions in the United States and Canada with shared interest and expertise in fetal therapy and other forms of multidisciplinary care for complex fetal disorders. Participating centers retrospectively identified pregnancies complicated by stage I TTTS receiving care from January 2005 through November 2014 with documented antenatal events and pregnancy outcomes. A database (Access, Microsoft, Seattle, WA) was constructed to collect demographic, sonographic, echocardiographic, multivessel Doppler velocimetry, prenatal intervention, complication, and delivery information. Internal controls were constructed within the database to ensure recruitment of only Quintero stage I TTTS (monochorionic diamniotic twin gestation, maximum vertical pocket of the donor sac ≤20 mm and maximum vertical pocket of the recipient sac ≥80 mm, both bladders visualized, no evidence of critically abnormal Doppler waveforms, no evidence of hydrops, both fetuses alive). Exclusion criteria included stage ≥II TTTS, polyhydramnios or oligohydramnios insufficient to make the diagnosis of stage I, major congenital anomalies of one or both fetuses, monoamniotic twins, twin reversed arterial perfusion, selective fetal growth restriction, higher order multifetal gestations, gestational age at diagnosis of ≥30 weeks, and twins that had undergone an in utero procedure (amnioreduction or septostomy) prior to the diagnosis of stage I TTTS at the North American Fetal Therapy Network center. Each center individually collected site-specific data for all qualifying cases. Deidentified data were pooled into a common database at the coordinating center for analysis. Institutional review board approval was obtained at each participating center.
Intervals were calculated from diagnosis at stage I to a change in status (regression or progression in stage, intervention, complications such as preterm premature rupture of membranes, termination, and delivery) as documented by follow-up ultrasounds or events captured within the database. The estimated date of delivery was used to anchor dating. Each record was evaluated by 2 investigators (S.P.E. and R.S.M.) operating independently.
Cases were retrospectively divided into 3 management strategies: those managed expectantly until progression and those treated at stage I with either amnioreduction or laser photocoagulation. Fetal survival rates were expressed as the percentage of no survivors, 1 survivor, 2 survivors, and at least 1 survivor to live birth. To better characterize infant health and prognosis (strongly determined by gestational age at delivery) we categorized pregnancy outcomes as good (dual live twin delivery ≥30.0 weeks), mixed (loss of 1 fetus or delivery between 26.0-29.9 weeks), and poor (double fetal demise or delivery <26.0 weeks). Outcomes were analyzed by initial management strategy.
Characteristics of women according to expectant management and intervention were compared using t tests for continuous variables and χ 2 tests for categorical variables. Multivariate multinomial logistic regression analysis was employed to compare outcome groups, independent of confounders. Women with 2 survivors or good outcome (defined above) were the referent in all models as we were interested in features that reduced the likelihood of less than optimal outcomes.
Potential confounders were maternal and fetal characteristics thought to have a potential impact on outcomes or management strategy including maternal age, nulliparity, gestational age, cervical length, placenta location, and recipient maximum vertical pocket at diagnosis of stage I. Multivariate multinomial logistic regression models were constructed using nulliparity, gestational age, placenta location, and recipient maximum vertical pocket. Statistical significance was set at alpha <0.05.
Materials and Methods
The North American Fetal Therapy Network is a consortium of 30 medical institutions in the United States and Canada with shared interest and expertise in fetal therapy and other forms of multidisciplinary care for complex fetal disorders. Participating centers retrospectively identified pregnancies complicated by stage I TTTS receiving care from January 2005 through November 2014 with documented antenatal events and pregnancy outcomes. A database (Access, Microsoft, Seattle, WA) was constructed to collect demographic, sonographic, echocardiographic, multivessel Doppler velocimetry, prenatal intervention, complication, and delivery information. Internal controls were constructed within the database to ensure recruitment of only Quintero stage I TTTS (monochorionic diamniotic twin gestation, maximum vertical pocket of the donor sac ≤20 mm and maximum vertical pocket of the recipient sac ≥80 mm, both bladders visualized, no evidence of critically abnormal Doppler waveforms, no evidence of hydrops, both fetuses alive). Exclusion criteria included stage ≥II TTTS, polyhydramnios or oligohydramnios insufficient to make the diagnosis of stage I, major congenital anomalies of one or both fetuses, monoamniotic twins, twin reversed arterial perfusion, selective fetal growth restriction, higher order multifetal gestations, gestational age at diagnosis of ≥30 weeks, and twins that had undergone an in utero procedure (amnioreduction or septostomy) prior to the diagnosis of stage I TTTS at the North American Fetal Therapy Network center. Each center individually collected site-specific data for all qualifying cases. Deidentified data were pooled into a common database at the coordinating center for analysis. Institutional review board approval was obtained at each participating center.
Intervals were calculated from diagnosis at stage I to a change in status (regression or progression in stage, intervention, complications such as preterm premature rupture of membranes, termination, and delivery) as documented by follow-up ultrasounds or events captured within the database. The estimated date of delivery was used to anchor dating. Each record was evaluated by 2 investigators (S.P.E. and R.S.M.) operating independently.
Cases were retrospectively divided into 3 management strategies: those managed expectantly until progression and those treated at stage I with either amnioreduction or laser photocoagulation. Fetal survival rates were expressed as the percentage of no survivors, 1 survivor, 2 survivors, and at least 1 survivor to live birth. To better characterize infant health and prognosis (strongly determined by gestational age at delivery) we categorized pregnancy outcomes as good (dual live twin delivery ≥30.0 weeks), mixed (loss of 1 fetus or delivery between 26.0-29.9 weeks), and poor (double fetal demise or delivery <26.0 weeks). Outcomes were analyzed by initial management strategy.
Characteristics of women according to expectant management and intervention were compared using t tests for continuous variables and χ 2 tests for categorical variables. Multivariate multinomial logistic regression analysis was employed to compare outcome groups, independent of confounders. Women with 2 survivors or good outcome (defined above) were the referent in all models as we were interested in features that reduced the likelihood of less than optimal outcomes.
Potential confounders were maternal and fetal characteristics thought to have a potential impact on outcomes or management strategy including maternal age, nulliparity, gestational age, cervical length, placenta location, and recipient maximum vertical pocket at diagnosis of stage I. Multivariate multinomial logistic regression models were constructed using nulliparity, gestational age, placenta location, and recipient maximum vertical pocket. Statistical significance was set at alpha <0.05.
Results
Ten centers contributed a total of 180 cases for analysis. In all, 56 cases were excluded for incomplete documentation of prenatal or delivery events (n = 25) or had incomplete staging (n = 13), were not stage I (n = 13), had other exclusions such as monoamniotic twins (n = 3), or were >29.9 weeks (n = 2). A total of 124 included cases remained for analysis.
In all, 49 cases (40%) were initially managed expectantly while 30 (24%) underwent amnioreduction and 45 (36%) underwent laser therapy at stage I. There was no difference in maternal age or parity among the 3 groups. There were differences, however, in recipient maximum vertical pocket, cervical length, placenta location, and gestational age at diagnosis among the groups ( Table 1 ).
| Characteristic | Overall n = 124 | EM n = 49 | AR n = 30 | Laser n = 45 | P value |
|---|---|---|---|---|---|
| Maternal age, y (SD) | 30.0 (6.1) | 31.5 (6.6) | 29.6 (5.6) | 28.8 (5.7) | .133 |
| Nulliparity, n (%) | 51.4 (42.9) | 21 (45.7) | 13 (44.8) | 17 (38.6) | .774 |
| Recipient MVP, mm (SD) | 109 (24.4) | 100 (21.0) | 118.4 (25.6) | 112.6 (24.2) | .002 |
| Cervical length, mm (SD) | 35.0 (12.2) | 36.8 (10.1) | 28.8 (14.7) | 36.3 (11.8) | .047 |
| Placenta location n (%) | .005 | ||||
| Anterior | 50 (40.3) | 24 (48.9) | 17 (56.6) | 9 (20.0) | |
| Posterior | 51 (41.1) | 19 (38.8) | 10 (33.3) | 22 (48.8) | |
| Other | 23 (18.5) | 6 (12.2) | 3 (10.0) | 14 (31.1) | |
| GA at diagnosis, wk (SD) | 21.5 (2.7) | 20.8 (2.5) | 23.5 (2.9) | 21.1 (2.1) | .0001 |
| GA at delivery, wk (SD) | 30.7 (4.9) | 29.8 (5.8) | 30.4 (3.7) | 31.9 (4.4) | .117 |
| Diagnosis to delivery, d (SD) | 64.5 (40.2) | 64.0 (46.0) | 47.8 (32.1) | 75.6 (34.7) | .014 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
