Article neonatology -to-Twin Transfusion: Part 3. Mortality and Neurodevelopmental Outcomes Following Intervention Ona M. Faye-Petersen, Objectives After completing this article, readers should be able to: MD,* Timothy M. Crombleholme, MD† 1. Compare and contrast the results of amnioreduction and laser therapy treatment of twin-to-twin transfusion syndrome (TTTS). 2. Discuss neurodevelopmental outcomes of delivered following treatment of TTTS. Author Disclosure Drs Faye-Petersen and Abstract Crombleholme have The high morbidity and mortality rates of twin-to-twin transfusion syndrome (TTTS) disclosed no financial generally are related to sequelae of cardiovascular dysfunction or vascular disruption. relationships relevant Neurologic sequelae associated with TTTS are emerging concerns for survivors. to this article. This A variety of clinical interventions, including amnioreduction, microseptostomy of the commenatry does not inter-twin membrane, and fetoscopic laser photocoagulation of placental anastomo- contain a discussion ses, have been used alone or in sequence to reduce the rates of mortality and morbidity. Because many of these specialized interventional procedures are performed of an unapproved/ at select centers in the , women may be treated at considerable distance investigative use of a from their primary obstetric care institutions and later return to deliver at their local commercial product/ facilities. Neonatologists may be unfamiliar with the relative efficacies and outcomes of device. the interventional procedures. In this review, we present a focused summary of the neurodevelopmental outcomes associated with these antenatal treatments.

Introduction Chronic twin-to-twin transfusion syndrome (TTTS) is a gestational condition in which a sustained net imbalance of blood volume transfuses between of a diamniotic monochorionic (DiMo) via placental anastomoses. It occurs in approximately 10% to 20% of all monochorionic twin gestations and has an 80% to 100% mortality rate, (1) if severe and left untreated, particularly if it is detected before 20 weeks of gestation. (2)(3) The pathogenesis of TTTS and its fetal and neonatal sequelae are discussed in Part 1 of this series on TTTS in this issue of NeoReviews. The clinical staging schemas of Quintero and associates (4)(5) and modifications (6)(7)(8) and various treatment modal- ities for TTTS (including amnioreduction [AR], microseptostomy of the inter-twin membrane, and nonselective and selective fetoscopic laser photocoagulation) are described in Part 2 in this issue. Also included in Part 2 are the efficacies of the various treatments and the pathologic examination of the DiMo . In this review (Part 3), we discuss mortality and the neurodevelopmental outcomes of infants whose gestations were treated by antenatal AR and fetal laser photocoagulation interventions.

Clinical Interventional Procedures in TTTS: Comparative Analysis of Fetal/Neonatal Survival The Eurofoetus trial (9) was the first prospective randomized trial to compare the efficacy and safety of treatment of TTTS with laser therapy or serial AR. Women presenting between 15 and 26 weeks’ gestation with polyhydramnios in the recipient twin and oligohydramnios in the donor twin participated in the trial. The patients were staged

*Associate Professor of Pathology and Obstetrics and Gynecology; Head, Microdissection Laboratory, Division of Anatomic Pathology, The University of Alabama at Birmingham, Birmingham Ala. †Director, Fetal Care Center of Cincinnati; Professor of Surgery, Pediatrics, and Obstetrics and Gynecology, University of Cincinnati College of ; Division of Pediatric General, Thoracic, and Fetal Surgery, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio.

NeoReviews Vol.9 No.9 September 2008 e393 neonatology twin-to-twin transfusion syndrome

according to Quintero criteria: (4)(5) 52% were Stage I detected a trend in adverse outcome affecting the recip- or II, 47% were Stage III, and 1% were Stage IV. Enroll- ient twin in one treatment arm and recommended to the ment was halted after a planned interim analysis revealed Data Safety Monitoring Board that the trial be stopped a significantly higher likelihood of survival of at least one to allow biostatistical analysis of the adverse trend. Re- twin to 28 days of age (76% versus 56%, Pϭ0.009) and to sults of the NIH TTTS trial showed no statistically 6 months of age (76% versus 51%, Pϭ0.002) in the laser significant difference in overall neonatal survival to 30 group compared with the AR group. In addition, more postnatal days (60% versus 43% pϭNS) or neonatal sur- infants were alive without neurologic abnormalities de- vival of one or both twins in the same (75% tected on neuroimaging studies in the laser group (52% versus 65%, pϭNS) in cases of severe TTTS treated by versus 31%, Pϭ0.003). The overall survival in the laser either AR or SFLP. Despite these overall results, a statis- arm was 57%, which was consistent with survival rates in tically significant worse fetal survival was observed previous reports of nonselective fetoscopic laser treat- among recipient twins in treated by SFLP ment (53%). (10)(11) This rate is significantly lower, compared with those treated by AR. This apparent co- however, than the survival reported with selective feto- nundrum can be accounted for by recipient fetal losses in scopic laser photocoagulation (SFLP) (64% to 68%). the SFLP arm being balanced by increased treatment (12)(13) Of particular concern is the poor survival ob- failures among recipients in the AR arm. These results served in the AR arm of 39%, which is significantly lower suggest that, in these highly selected cases of severe than previously reported (60% to 65%). (13)(14)(15) TTTS, neither treatment is superior to the other. Once (16) Antenatal, peripartum, and neonatal care was pro- TTTS reaches this degree of severity, the mortality vided by the referring hospital, and lack of standardiza- among recipients is considerable, but the losses may tion may explain some of these differences. (17) The occur at different times, depending on treatment. The decreased survival in the AR group may reflect the higher impact of TTTS severity on fetal survival is supported pregnancy termination rate in the AR group (16 versus 0 further by the significantly worse fetal survival among in the laser group). The terminations were requested recipient twins in Stages III and IV compared with those after the diagnosis of severe fetal complications. It would in Stage II. One of the strongest predictors of recipient be instructive to know whether these women were of- demise is echocardiographic evidence of TTTS cardio- fered cord coagulation as a means of rescuing one baby. myopathy. The losses of fetal recipients treated by SFLP (13) Reliable assessment of neurologic outcome is criti- usually occur within 24 hours of the procedure. In con- cal when assessing efficacy of treatment for TTTS. Al- trast, the recipients treated by AR are not lost following though the rate of abnormality on neurologic imaging the procedure, but there is progressive TTTS cardiomy- was lower in the laser group (7% versus 17%), long-term opathy, as reflected by more recipients in the AR arm neurodevelopmental assessment has revealed no differ- meeting criteria to be declared treatment failures. In ence in outcome between survivors treated by fetoscopic every case, findings in the recipient twin met criteria for laser and those treated by AR. treatment failure. Taken together, these data suggest a The National Institutes of Health (NIH)-sponsored disproportionate impact of TTTS cardiomyopathy on TTTS trial is the only other prospective randomized trial recipient survival in advanced stages of TTTS no matter comparing survival among those receiving AR versus what treatment they receive. SFLP. (18) This trial differed from the Eurofoetus trial in Recently, Rossi and D’Addario (19) reported a several important aspects. First, to qualify for the NIH Cochrane review of TTTS with a meta-analysis that Trial, the TTTS had to fail to respond to a qualifying included data from both the Eurofoetus and NIH trials. amniocentesis. The rationale for this requirement was to The conclusion drawn from this analysis was that SFLP eliminate those who were more likely to respond to AR, of TTTS is preferred over AR when it is available and AR the so-called “single amnio paradox.” Second, patients is preferred when SFLP is not available. The results of were candidates only if the TTTS presented earlier than this analysis likely are skewed toward fetoscopic laser 22 weeks of gestation, and no Stage I patients were based on the small numbers of individuals included from candidates for the trial. These two requirements were the NIH trial (nϭ40) compared with the number in- substantially different from the Eurofoetus trial in which cluded from the Eurofoetus trial (nϭ142). women were randomized into the trial up to 26 weeks of AR is readily available, less costly, and less invasive; gestation, and 52% of those entered were Stage I. (9) laser therapy is only available at select institutions and The NIH study was stopped early, after 42 women requires specialized training. Although it makes sense to were randomized, when the Trial Oversight Committee use AR where treatment options give similar results, it e394 NeoReviews Vol.9 No.9 September 2008 neonatology twin-to-twin transfusion syndrome

would be prudent to move promptly to laser therapy if outcome in who do not have rigorous studies can prove that this therapy has better TTTS. The incidence of severe neurodevelopmental ab- short- and long-term outcomes in the setting of ad- normalities in monochorionic twins without TTTS is 6%. vanced disease. TTTS survivors who develop neurologic handicap and For patients who respond to AR, the overall survival mental retardation do not always have abnormal neona- rate has been 88%. (20) In those cases in which echocar- tal ultrasonography results. Similarly, not all children diographic progression is detected despite AR, the over- who have abnormal ultrasonography findings have clini- all survival rate when SFLP is performed is 80%. The cally significant neurodevelopmental deficits. In one difference in survival between responders to AR and small study that followed TTTS survivors for a mean of those who progress to SFLP are not statistically signifi- 6.2 years (range, 4 to 11 years), the incidence of cerebral cantly different, suggesting that survival was not compro- palsy was 26% (5/19 infants) in the group treated by mised by an initial trial of AR before progressing to serial AR. All of these children had abnormal mental SFLP. development in addition to motor deficits. Of note, three of the five children had normal findings on neonatal head Neurodevelopmental Outcomes in TTTS ultrasonography. In the combined cohort of children Versus Outcomes Following Interventions whose mothers had been treated with AR or conservative Although much attention has focused on the effect of treatment, 22% (5/23) who did not have treatment on survival in TTTS, the neurologic morbidity or abnormal mental development had mild speech delay among survivors frequently is underappreciated. The and required special education. One limitation to this International Amnioreduction Registry tracked 223 and other studies is the lack of a comparable conserva- women who had TTTS diagnosed before 28 weeks’ tively treated cohort group. Given the improved survival gestation and were treated with serial aggressive AR. (16) of TTTS babies who receive AR and other treatment Of those infants who survived to 4 weeks of age and modalities, however, it is unlikely that such a cohort ever underwent clinically indicated cranial ultrasonography, will be available for comparison. 24% of recipient (26/109 scanned) and 25% of donor Studying infants from pregnancies complicated by twins (22/88 scanned) had abnormal findings. Findings TTTS and treated with AR, Mari and associates (16) included severe intraventricular hemorrhage, ventricular detected a rate of cerebral palsy of 4.7% (2 of 42 infants) dilation, cerebral echogenic foci, cerebral cysts, and in those children who survived to more than 24 months periventricular leukomalacia among other less common of age. One reason for the lower incidence of cerebral lesions. Eighty infants died before reaching 4 weeks of palsy than in the study by Lopriore and colleagues (22) age, and how many of these would have had abnormal may be related to the latter study group having more imaging if cranial ultrasonography had been performed is severe disease, with all the patients diagnosed before unknown. Among patients in the TTTS Registry from 28 weeks’ gestation versus up to 33 weeks’ gestation in Australia and New Zealand, most of whom had been the study by Mari and colleagues. Of note, in the Mari treated with AR, the rate of abnormal cranial ultrasonog- study, nine survivors had mild speech or motor delay. raphy findings was similar at 27.3%. (13) The rate of Wee and associates (23) studied the long-term neuro- periventricular leukomalacia in this group was 10.8%, logic outcome of 52 children from 31 TTTS pregnancies which is particularly important due to the association of who survived to more than 18 months, most of whose this lesion with cerebral palsy. In another small series of mothers had been treated with AR. The comparison was patients treated with AR, the rate of abnormal neonatal a regional cohort of term and preterm infants, with most cranial ultrasonography findings was as high as 58%. (21) born very preterm. In addition, the TTTS babies were It is important to recognize, however, that neuroimaging compared with matched singleton and twin control does not always correlate with neurodevelopmental out- groups. The mean intelligence quotient (IQ) of TTTS come. An who has normal findings on head ultra- survivors was significantly lower than the comparison sonography and magnetic resonance imaging can be cohort, due primarily to a 13-point IQ reduction in those neurodevelopmentally devastated, and an infant who has children born before 33 weeks’ gestation. There was no evidence of leukoencephalomalacia on imaging studies difference in the rate of cerebral palsy (5.8% for TTTS can be neurodevelopmentally intact. versus 4.9% for very preterm twins versus 3.3% for very Only a few studies have reported longer-term neuro- preterm singletons) or behavioral test results in the developmental outcome. When interpreting these stud- TTTS survivors. This was a small study, however, and not ies, it is important to appreciate the neurodevelopmental sufficiently powered to demonstrate differences in cere-

NeoReviews Vol.9 No.9 September 2008 e395 neonatology twin-to-twin transfusion syndrome

bral palsy. Still, these researchers appropriately raise the management (1) and reduced morbidity and its severity issue that studies evaluating long-term neurologic out- in survivors. come in TTTS need to consider that most TTTS preg- nancies are delivered very preterm as well as the fact that Summary twins generally are more likely to experience neurologic The survival of one or both twins of a gestation compli- compromise. cated by TTTS is dependent on the gestational week of Even fewer studies have examined the long-term out- onset of TTTS, the rapidity of its progression, the cardiac come of survivors of TTTS treated with intrauterine laser status of the recipient, the gestational week of clinical photocoagulation therapy. Banek and colleagues (24) intervention, and at birth. Both AR and reported that in 89 such children, 78% showed normal SFLP are important means of increasing the duration of development at a median age of 22 months. Eleven time for intrauterine development and, thereby, improv- percent had minor neurologic abnormalities, including ing outcomes for the infants, but the use of MRI may strabismus, mildly delayed motor development, or mildly become a key diagnostic means to enable earlier detec- abnormal speech. The remaining 11% suffered significant tion and mitigation of the deleterious effects of TTTS. neurologic deficiencies, including cerebral palsy, hemi- However, central nervous system lesions or functional paresis, and spastic quadriplegia. Of note, significantly deficits may not be detectable at the time of birth or in more children in the neurologically impaired groups the newborn period for a twin from a gestation with were born very preterm. Also of importance, two infants TTTS. Thus, as for the risks for long-term cardiac or from the most severely affected group had abnormal renal dysfunction, long-term follow up to detect neuro- scan results before laser treatment. The findings of logic sequelae should be communicated to the pediatri- this study are consistent with those of Sutcliffe and cians and staff caring for these infants. associates, (25) who reported a cerebral palsy rate of 9% in children after in utero treatment with laser therapy for TTTS. Graef and coworkers, (26) in a report of 167 References TTTS survivors who had been treated by fetoscopic laser, 1. Crombleholme TM. The treatment of twin-twin transfusion syndrome. Semin Pediatr Surg. 2003;12:175–181 found normal neurodevelopmental testing results in 2. Luks FI, Carr SR, De Paepe ME, Tracy TF Jr. What–and 86.8% of cases, with 7.2% of infants having minor neur- why–the pediatric surgeon should know about twin-to-twin trans- ologic deficiencies and 6% having major neurologic defi- fusion syndrome. J Pediatr Surg. 2005;40:1063–1069 ciencies such as cerebral palsy, hemiparesis, and quadri- 3. Harkness UF, Crombleholme TM. Twin-twin transfusion syn- drome: where do we go from here? Semin Perinatol. 2005;29: plegia. These findings were not unlike those in follow-up 296–304 of monochorionic twins without TTTS, and the most 4. Quintero RA, Morales WJ, Allen MH, Bornick PW, Johnson severely affected children were delivered prior to PK, Kruger M. Staging of twin-twin transfusion syndrome. J Peri- 28 weeks’ gestation, suggesting an important influence natol. 1999;19:550–555 of gestational age on neurodevelopmental outcome. 5. Quintero RA, Dickinson JE, Morales WJ, et al. Stage-based treatment of twin-twin transfusion syndrome. Am J Obstet Gynecol. Similarly, Ortqvist and colleagues (27) reported the neu- 2003;188:1333–1340 rodevelopmental outcome of 114 survivors treated in the 6. Taylor MJ, Govender L, Jolly M, Wee L, Fisk NM. Validation of Eurofoetus trial in which 13.2% had evidence of a major the Quintero staging system for twin-twin transfusion syndrome. neurodevelopmental abnormality. However, there was Obstet Gynecol. 2002;100:1257–1265 7. Duncombe GJ, Dickinson JE, Evans SF. Perinatal characteristics no difference between those who had been treated by and outcomes of pregnancies complicated by twin-twin transfusion laser and those treated by AR. Perinatal factors, including syndrome. Obstet Gynecol. 2003;101:1190–1196 gestational age at delivery and Apgar score, correlated 8. Michelfelder E, Gottliebson W, Border W, et al. Early manifes- with adverse outcome. tations and spectrum of recipient twin cardiomyopathy in twin-twin Of note, earlier and more sensitive antenatal detection transfusion syndrome: relation to Quintero stage. Ultrasound Ob- stet Gynecol. 2007;30:965–971 of central nervous system injury (1)(28) has been re- 9. Senat MV, Deprest J, Boulvain M, Paupe A, Winer N, Ville Y. ported recently with the adjunct use of magnetic reso- Endoscopic laser surgery versus serial amnioreduction for severe nance imaging (MRI) techniques. MRI has enabled the twin-to-twin transfusion syndrome. N Engl J Med. 2004;351: detection of cerebral venous sinus and urinary collecting 136–144 10. De Lia JE, Kuhlmann RS, Harstad TW, Cruikshank DP. system dilatation and more clearly delineated central Fetoscopic laser ablation of placental vessels in severe previable nervous system lesions. The improvements in antenatal twin-twin transfusion syndrome. Am J Obstet Gynecol. 1995;172: detection may result in further improvements in clinical 1202–1208 e396 NeoReviews Vol.9 No.9 September 2008 neonatology twin-to-twin transfusion syndrome

11. Ville Y, Hyett J, Hecher K, Nicolaides K. Preliminary experi- amnioreduction (AR), selective fetoscopic laser photocoagulation ence with endoscopic laser surgery for severe twin-twin transfusion (SFLP), and intrafetal radiofrequency ablation (RFA). Am J Obstet syndrome. N Engl J Med. 1995;332:224–227 Gynecol. 2007;197(suppl):S202 12. Hecher K, Plath H, Bregenzer T, Hansmann M, Hackeloer BJ. 21. Denbow ML, Battin MR, Cowan F, Azzopardi D, Edwards Endoscopic laser surgery versus serial amniocenteses in the treat- AD, Fisk NM. Neonatal cranial ultrasonographic findings in pre- ment of severe twin-twin transfusion syndrome. Am J Obstet Gy- term twins complicated by severe fetofetal transfusion syndrome. necol. 1999;180:717–724 Am J Obstet Gynecol. 1998;178:479–483 13. Dickinson JE, Evans SF. Obstetric and perinatal outcomes 22. Lopriore E, Nagel HT, Vandenbussche FP, Walther FJ. Long- from the Australian and New Zealand twin-twin transfusion syn- term neurodevelopmental outcome in twin-to-twin transfusion drome registry. Am J Obstet Gynecol. 2000;182:706–712 syndrome. Am J Obstet Gynecol. 2003;189:1314–1319 14. Elliott JP, Urig MA, Clewell WH. Aggressive therapeutic 23. Wee LY, Taylor M, Watkins N, Franke V, Parker K, Fisk NM. amniocentesis for treatment of twin-twin transfusion syndrome. Characterisation of deep arterio-venous anastomoses within mono- Obstet Gynecol. 1991;77:537–540 chorionic placentae by vascular casting. Placenta. 2005;26:19–24 15. Pinette MG, Pan Y, Pinette SG, Stubblefield PG. Treatment of 24. Banek CS, Hecher K, Hackeloer BJ, Bartmann P. Long-term twin-twin transfusion syndrome. Obstet Gynecol. 1993;82:841–846 neurodevelopmental outcome after intrauterine laser treatment for 16. Mari G, Roberts A, Detti L, et al. Perinatal morbidity and severe twin-twin transfusion syndrome. Am J Obstet Gynecol. 2003; mortality rates in severe twin-twin transfusion syndrome: results of 188:876–880 the International Amnioreduction Registry. Am J Obstet Gynecol. 25. Sutcliffe AG, Sebire NJ, Pigott AJ, Taylor B, Edwards PR, 2001;185:708–715 Nicolaides KH. Outcome for children born after in utero laser 17. Fisk NM, Galea P. Twin-twin transfusion–as good as it gets? ablation therapy for severe twin-to-twin transfusion syndrome. Br J N Engl J Med. 2004;351:182–184 Obstet Gynaecol. 2001;108:1246–1250 18. Crombleholme TM, Shera D, Lee H, et al. A prospective, 26. Graef C, Ellenrieder B, Hecher K, Hackeloer BJ, Huber A, randomized, multicenter trial of amnioreduction vs selective feto- Bartmann P. Long-term neurodevelopmental outcome of 167 chil- scopic laser photocoagulation for the treatment of severe twin-twin dren after intrauterine laser treatment for severe twin-twin transfu- transfusion syndrome. Am J Obstet Gynecol. 2007;197:396 e1–e9 sion syndrome. Am J Obstet Gynecol. 2006;194:303–308 19. Rossi AC, D’Addario V. Laser therapy and serial amnioreduc- 27. Ortqvist LCS, Chevret S, Bussieres L, Staraci S, Huard F, Ville tion as treatment of twin-twin transfusion syndrome: a metaanalysis Y. Long-term neurodevelopmental outcome in twin-to-twin trans- and review of the literature. Am J Obstet Gynecol. 2008;198: fusion syndrome in the Eurofoetus trial. Am J Obstet Gynecol. 147–152 2006;195(suppl 1):S3 20. Cromblehome TM, Livingston JC, Polzin W, et al. Multimo- 28. Quarello E, Molho M, Ville Y. Incidence, mechanisms, and dality and sequential therapy for twin-twin transfusion syndrome patterns of fetal cerebral lesions in twin-to-twin transfusion syn- (TTTS): a stage and gestational age based approach to the use of drome. J Matern Fetal Neonatal Med. 2007;20:589–597

NeoReviews Vol.9 No.9 September 2008 e397 neonatology twin-to-twin transfusion syndrome

NeoReviews Quiz

6. The twin-to-twin transfusion syndrome (TTTS) trial sponsored by the National Institutes of Health has examined the effect of treatments such as amnioreduction and selective fetoscopic laser photocoagulation on the mortality of one or both twins. Of the following, the strongest predictor of demise of the recipient twin in TTTS is: A. Echocardiographic evidence of cardiomyopathy. B. Gestational age of twin at birth. C. Gestational week of clinical intervention. D. Gestational week of onset of TTTS. E. Rapidity of progression of TTTS.

7. The TTTS Registry from Australia and New Zealand has examined the effect of treatments such as amnioreduction and selective fetoscopic laser photocoagulation on the neurologic outcome among surviving twins. Of the following, the most common abnormality on cranial ultrasonography among twins treated for TTTS reported by this TTTS Registry is: A. Cerebral cysts. B. Cerebral echogenic foci. C. Intraventricular hemorrhage. D. Periventricular leukomalacia. E. Ventricular dilation.

e398 NeoReviews Vol.9 No.9 September 2008