Review Article
557
The History of Fetal Therapy Kenneth J. Moise, Jr., MD1
UT Health School of Medicine, Houston, Texas Am J Perinatol 2014;31:557–566.
Abstract Keywords
► fetal surgery ► fetal intervention ► maternal–fetal surgery ► hemolytic disease of the fetus/newborn ► fetal myelomeningocele ► twin–twin transfusion ► fetal diaphragmatic hernia
Address for correspondence Kenneth J. Moise, Jr., MD, Department of Obstetrics, Gynecology and Reproductive Sciences, Division of MaternalFetal Medicine, UT Health School of Medicine, 6410 Fannin, Suite 700, Houston, TX 77030 (e-mail: [email protected]).
The Fetal Treatment Center founded by Michael Harrison is credited as the birthplace of fetal surgery. His trainees in pediatric surgery subsequently founded fetal centers throughout the United States. In Europe, the advent of minimally invasive fetal surgical techniques led to the establishment of treatment centers led predominantly by perinatologists. More recently, perinatologists in North America have begun to play a greater role in the field of fetal intervention. Intrauterine transfusion for the treatment of hemolytic disease of the fetus/newborn was the first successful fetal intervention. Although not subjected to the rigors of clinical trials, this treatment has withstood the test of time. Interventions for other fetal disease states such as twin–twin transfusion and repair of fetal myelomeningocele were investigated in animal models followed by randomized clinical trials before widespread adoption. Tracheal occlusion for diaphragmatic hernia is still currently being investigated as the next promising step in fetal intervention.
Michael Harrison is widely regarded as the “father of fetal surgery.”1 In a YouTube interview, he tells the story how during his first month of his general surgery internship at Massachusetts General Hospital he assisted in a neonatal surgery to correct a diaphragmatic hernia.2 He describes the surgery as a “beautiful procedure” however the infant died the following day from pulmonary hypoplasia. He goes on to say that he realized “that the only way to make that lung bigger is to fix something before birth so it has room to grow.” And so began Harrison’s 30-year adventure in the treatment of the unborn. He established the Fetal Treatment Center at the University of San Francisco (UCSF) in the early 1980s. The first successful open surgery for fetal lower urinary tract obstruction was performed in 1981; open surgical resection of fetal congenital adenomatoid malformation followed in 1984 with the first open repair of a fetal diaphragmatic hernia in 1989.3–5 Harrison protégés and their trainees went on to establish fetal treatment centers at many of the major children’s hospitals throughout the United States including those in Philadelphia, Boston, Cincinnati, Nashville, Houston, St Louis, and Denver. Pediatric surgeons initially took the lead
on these endeavors. Maternal–fetal medicine (MFM) specialists, the physicians that often made the initial prenatal diagnosis of fetal anomalies, were relegated to support roles at these centers. Across the pond in Britain and Europe, the picture was much different. Open fetal surgery to correct congenital anomalies was felt to be associated with an unacceptably high rate of maternal morbidity. However, with the advent of minimally invasive laser intervention for severe twin–twin transfusion King’s College in London became the epicenter for fetal therapy. As an obstetrician with expertise in fetal medicine, Kypros Nicolaides mentored other obstetricians in this emerging field. Fetal centers in Hamburg, Paris, Leuven, Barcelona, and Leiden soon emerged. A collaboration of European fetal surgery groups (Eurofoetus) developed to enroll patients in multicentered trials. Back in the United States, MFMs began to play more of an active role in fetal intervention. In August, 2004, an advisory committee that included pediatric surgeons and MFM that were active in the field attended a workshop sponsored by the National Institutes of Health and the Office of Rare Diseases.6 The term Maternal–Fetal therapy was introduced as the
received October 21, 2013 accepted after revision November 27, 2013 published online February 25, 2014
Copyright © 2014 by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York, NY 10001, USA. Tel: +1(212) 584-4662.
DOI http://dx.doi.org/ 10.1055/s-0033-1364191. ISSN 0735-1631.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
1 Department of Obstetrics, Gynecology and Reproductive Sciences,
History of Fetal Therapy
Moise
pregnant woman reaps no direct benefit from fetal surgery yet she incurs substantial risks. Issues of preclinical trials, patient consent and the need for multicenter clinical trials were addressed. Led by four perinatologists, the North American Fetal Treatment Network (NAFTNet) was established in October, 2004 to mirror the role of the Eurofoetus network.7 Meanwhile, laser centers for the treatment of severe twin– twin transfusion were led mostly by MFMs—many who considered twin-to-twin transfusion syndrome (TTTS) to be an obstetrical disease as it was related to placentation. Soon the term fetal intervention began to replace the term fetal surgery because TTTS treatments were performed more frequently than any other fetal procedures. MFMs began to play greater roles in the leadership of fetal centers often serving as codirectors with their pediatric surgical colleagues. In the most recent development of fetal centers, children’s hospitals without already established on-site delivery services have developed special delivery units and even women’s pavilions to allow for on-site obstetrical delivery for those pregnant women with a fetus diagnosed with a congenital anomaly. Many questions still lie ahead for the world of fetal intervention, particularly in the United States.8 These will become especially important with ever shrinking financial resources. Should we follow the model established by our European colleagues where some smaller countries only have one designated center? Clearly, the geographic distribution of fetal centers must be considered in the United States to allow for easy patient access. How many centers should exist so that patient numbers are not diluted thus allowing opportunity for education and research to further the field? Should centers be designated as centers of excellence based on experience and outcomes much like current transplant centers?
What is a Fetal Center? The definition of a fetal center is notably lacking from the literature. A recent committee opinion from the American College of Obstetricians and Gynecologists (ACOG) addresses this topic and noted that fetal centers exist in many forms. They are free standing centers (usually in conjunction with a children’s hospital) or centers can exist within established obstetrical or pediatric departments. The majority of fetal centers represent multidisciplinary clinics where pregnant women with fetal anomalies can receive diagnostic services, pregnancy management, prenatal consultation with pediatric subspecialists and coordinated planning for delivery and postdelivery neonatal care. These might be considered tier one centers. Often needle-based fetal intervention procedures are offered such as intrauterine transfusions or shunt placement. A tier two center would offer more advanced therapies such as laser ablation for twin–twin transfusion. Finally, a third tier center would offer the full gamut of accept fetal interventions including open fetal surgery for repair of spina bifida and for the removal of other rare fetal lesions. ACOG recommends that the fetal intervention team include MFM specialists to protect the interests of the pregnant woman.9 A recent task force statement on MMC repair proposed some of American Journal of Perinatology
Vol. 31
No. 7/2014
the components of an experienced fetal care team (►Table 1).10 Perhaps, the most important members of the team are the clinical care coordinators. These are often nurses with obstetrical or pediatric background, genetic counselors, social workers, or midwives. They serve as the “glue” to be sure that all the diverse subspecialists are on the same page for coordinated care. Other manuscripts in this special edition of the American Journal of Perinatology will highlight the evolving and expanded role of members of the fetal team. Tier three centers should also exhibit a commitment to education. Currently, four United States centers offer clinical fellowships in fetal intervention. Tier three centers should also be involved in research to further the development of new fetal interventions. Such research includes enrollment in clinical trials, submission of data to national registries and experimental work in animal models. The remainder of this chapter will elucidate the history and the basis for the treatment of the four most common conditions associated with fetal intervention: hemolytic disease of the fetus/newborn, twin–twin transfusion, myelomeningocele (MMC), and diaphragmatic hernia.
Hemolytic Disease of the Fetus and Newborn The story of the conquest of hemolytic disease of the fetus and newborn (HDFN) teaches us important lessons in fetal intervention. When solutions to this enormous clinical challenge were first being conceived and investigated, the internet was not yet even a concept and FAX machines were years in the future. Pioneering innovators in different parts of the world developed new concepts many times on parallel tracks. Descriptions of the newborn affected by HDFN can be found as far back as Hippocrates. In 1932, Diamond et al described the neonatal disease and proposed the term erythroblastosis fetalis based on the finding of circulating erythroblasts (nucleated red blood cells).11 A decade later, Levine et al12 were able to demonstrate a causal relationship between RhD antibodies and HDFN. Exchange transfusion in the neonate was introduced as the first effective therapy for those fetuses that survived to birth.13 However, 30% of fetuses continued to die in utero. Premature delivery was attempted to prevent these cases however a clinical trial in England failed to find that this resulted in improved neonatal outcome as prematurity itself was often associated with neonatal death.14 A more targeted diagnostic approach was developed based on the work by Bevis in England and later Liley in New Zealand.15,16 These investigators analyzed the bilirubin content of amniotic fluid to determine when imminent fetal death might occur. Unaware of the previous work of Liley on the other side of the world, Freda at Columbia-Presbyterian Hospital in New York began a specialized Rh antepartum clinic to test the idea that serial amniocenteses could be used to predict the severity of HDFN.17 He was able to demonstrate a reduction in the perinatal mortality from 30 to 9% over the course of 5 years. Bevis is credited with the first attempt at fetal intervention for HDFN.18 Using radiopaque dye injected into the mother and fluoroscopy to outline the placental vasculature, he
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
558
Table 1 Components of a fetal center Specialized services • Fetal MRI • Fetal echocardiography • Advanced ultrasound imaging Team members • Maternal–fetal medicine specialist • Pediatric surgeon • Other pediatric surgical specialists—neurosurgery, urology, plastic, craniofacial, and orthopedic • Pediatric specialists—neonatology, cardiology, radiology, nephrology, and neurology • Social work • Child life specialist • Obstetrical anesthesiologist • Pediatric anesthesiologist • Care coordinator • Genetic counselor Functional components • Designated leadership team • Regular multidisciplinary meetings to discuss cases • Fetal board or ethics committee • Quality control procedures • Research • Commitment to education • Long-term follow-up clinics Abbreviation: MRI, magnetic resonance imaging.
attempted to inject blood directly into the placenta. Several fetuses died; however, RhD negative red cells were detected in the circulation of one neonate. Sir William Liley is credited with the first successful intraperitoneal fetal transfusion (IPT).19 A visiting young geneticist who had recently worked with missionaries in Nigeria noted that red cells had been successfully transfused into the peritoneal cavity of neonates with sickle cell disease. Normal appearing red blood cells were later noted on peripheral blood smears. On a previous occasion, Liley realized that he had accidentally entered the peritoneal cavity of a fetus at the time of amniocentesis. He recognized this because the yellow color of the ascitic fluid was more intense than that of previous amniotic fluid samples he had obtained. During the case, he injected air into the fetal peritoneal cavity and then injected radiopaque dye into the amniotic fluid as the needle was removed. A radiograph taken a short time later revealed an image of the fetal peritoneal cavity with the ingested dye outlining the fetal gastrointestinal tract. Liley felt that access to the fetal peritoneal cavity could be accomplished for the infusion of red cells. Institutional review boards that we know today did not yet exist. A committee of three senior obstetricians (including Liley) was convened and a decision made to attempt an
Moise
intraperitioneal intrauterine transfusion. After three unsuccessful attempts that resulted in fetal demise, the fourth fetus was delivered at 34-week and 7-day gestation after undergoing two IPTs at 32 weeks and 1 day and 33 weeks and 4 days. Further innovations by Liley included the use of metal grids to aid in the location of the needle entry site into the fetus and even the use of multiple needles to immobilize the fetus. Several investigators challenged the IPT approach as it was not equivalent to the neonatal exchange transfusion that was so effective after birth. Direct access to the fetal circulation was needed. Investigators at Columbia University in New York were the first to attempt the direct transfusion of red cells to the fetus.18 Access to the fetal superior sagittal sinus was achieved in a rhesus monkey. A hysterotomy was then undertaken in a human case at 32-week gestation. Blood was successfully aspirated from the superior sagittal fetal vein however catheter access could not be achieved; a perinatal loss occurred after a premature delivery several days later. Subsequently, several investigators attempted to transfuse the anemic fetus by direct access at the time of hysterotomy using the fetal femoral artery, saphenous vein, and internal jugular vein.20–22 All resulted in premature delivery and perinatal loss. Adamsons et al23 attempted to place a long-standing indwelling catheter into the fetal peritoneal cavity for the periodic infusion of red cells. Four attempts failed, but a fifth procedure performed in Brazil at 24-week gestation resulted in a live born infant 8 weeks later. Liley came to the New York on sabbatical to introduce the IPT method to American physicians. Early attempts at IPT involved the introduction of 14- or 16-gauge Tuohy needle into the fetal peritoneal cavity followed by injection of a radiopaque dye under continuous fluoroscopy. An epidural catheter was then placed through the needle and the needle withdrawn. Blood could then be infused through the catheter followed by its removal. The late 1970s saw the introduction of real-time ultrasound to direct the transfusion needle. A Boston group reported a 53% rate of survival for nonhydropic fetuses and 7% survival with hydrops. After the introduction of ultrasound for needle guidance, a survival rate of 62% was reported; however, survival in the hydropic fetus remained poor with a salvage rate of only 29%.24 No further refinements in the IPT method occurred until 1985 when intramuscular paralytic agents were introduced to cause cessation of fetal movements.25 Smaller diameter needles could now be used and indwelling catheters were no longer necessary. Hydrops fetalis continued to be associated with poor fetal survival as it was realized that intraperitoneal blood was not absorbed in these sick fetuses. Once again, fetal interventionists would attempt to gain direct access to the fetal circulation. In 1981, Rodeck et al26 performed the first successful intravascular fetal transfusion (IVT) using a fetoscope to guide the transfusion needle into a placental plate vessel. One year later, Bang et al27 performed the first ultrasound-guided IVT using the intrahepatic portion of the umbilical vein. In an attempt to evaluate fetuses for in utero infection with toxoplasmosis, French investigators were able to successfully access the umbilical vessels for sampling.28 The intravascular American Journal of Perinatology
Vol. 31
No. 7/2014
559
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
History of Fetal Therapy
History of Fetal Therapy
Moise
intrauterine transfusion was born and became the standard method of intervention for HDFN. An initial fetal hematocrit could now be obtained and a specific amount volume of packed red cells infused to achieve a desired final hematocrit value. Minimal refinements in the procedure followed. Some fetal interventionalists found that the intrahepatic portion of the umbilical vein was a better access point for the IVT.29 Paralytic agents injected at the start of the transfusion were noted to decrease the chance for fetal movement that could interfere with a successful procedure.30 A step-wise approach to the correction of severe fetal anemia in the early second trimester fetus improved fetal survival.31 In a recent series of over 1,400 IUTs in almost 500 fetuses, a survival rate of 91% in nonhydropic fetuses and 83% in hydropic ones has been reported from a national referral center in the Netherlands.32
Twin-to-Twin Transfusion Syndrome A description of the “third circulation” in the placentae of monochorionic twins can be traced back to placental studies in the late 1800s.33 Benirschke and Kim34 suggested that vascular placental connections were responsible for the clinical phenotype of twin-to-twin transfusion. Subsequent injection studies in the placentae of patients with monochorionic twins complicated by oligohydramnios-polyhydramnios sequence indicated a predominance of unidirectional flow through vascular anastomoses from the “donor” fetus to the “recipient.”35 Untreated TTTS presenting in the second trimester was associated with a perinatal mortality of up to 70%.36 Progressive polyhydramnios would often lead to premature rupture of membranes with delivery before a viable gestational age; dual fetal demise often occurred as well. Amnioreduction of the polyhydramnios relieved maternal symptoms but did not improve survival.37 TTTS rarely complicates monoamniotic twins. This led investigators to study the role of septostomy (a purposeful penetration of the intertwin membrane) to equalize the amniotic fluid volumes in both twins’ sacs.38 A randomized clinical trial between amnioreduction and septostomy however failed to show a therapeutic benefit to septostomy.39 The neodymium:yttrium-aluminum-garnet (NG-YAG) laser can transmit energy through a liquid medium. In 1983, DeVore et al40 began to explore if this modality could be used in an amniotic fluid environment for fetal intervention. A hysterotomy was performed in two lambs and YAG laser energy transmitted through a 600 µm quartz fiber was used to sever the tail, limb, and umbilical cord in an avascular fashion. In two additional cases, these investigators introduced a fetoscope by laparotomy and inserted the laser fiber just adjacent to it. Coagulation and excision of various tissues was successfully undertaken. Further work was undertaken in pregnant ewes. Two underwent hysterotomy and placental vessels of varying diameters were coagulated and then transected to be sure that the vascular flow had been disrupted. In two additional ewes, a laparotomy was undertaken and the laser fiber was introduced into the amniotic cavity through the side port of a pediatric cystoscope. Placental vessels were American Journal of Perinatology
Vol. 31
No. 7/2014
visualized and lasered. Subsequently, these vessels were retrieved and histologic evaluation indicated complete occlusion of their lumen with minimal damage to surrounding placental tissues. In a letter dated June 19, 1984, De Lia contacted Bernische and inquired as to what species of animals might best represent human placentation.41 He felt that the treatment of TTTS with laser had the “potential of being the most commonly performed fetal surgery.” This led to an investigation in rhesus monkeys targeting the vascular connections between the two lobes of their placentae.42 The first case was a failure when the laser fiber was introduced through a separate incision. Subsequent cases involved attachment of the fiber to the side of the fetoscope with a suture. In only one of these cases was the planned placental vascular coagulation successful. Other cases were complicated by stillbirth, abruption, infection, uterine hypertonicity, and inadvertent coagulation of vessels on the wrong placental disk. De Lia went on to extensively study the shared vasculature of placentae from monochorionic twin gestations (J. E. De Lia, MD, personal communication, 1990). The first three cases of laser for TTTS in human pregnancies were reported by De Lia in 1990.43 All fetuses survived the initial surgery at 18.5, 22, and 22.5 weeks. In the first two cases, preterm premature rupture of the membranes (PPROM) with delivery occurred 8.5 and 12 weeks after the surgery. The third case was delivered 6.5 weeks after the laser procedure because of severe maternal preeclampsia. Four of the six fetuses survived the neonatal period. In the United States, De Lia in Salt Lake City (later in Milwaukee) and Quintero in Detroit (later in Tampa) began to offer laser therapy for TTTS. Considerable doubt remained in the perinatal community regarding the benefit of laser therapy. How could one possibly visualize all of the possible anastomoses between the twins? Complications of PPROM, abruption, and even a maternal death were known to occur. Amnioreduction was routinely employed as part of the laser surgery; it alone had proved therapeutic in some reported cases of TTTS. In his first large clinical series of 26 patients, De Lia reported that 35% of cases had dual survivors, 31% had one survivor, and 35% had no survivors.44 A second center at the Harris Birthright Centre at King’s College in London adopted this new therapy. Nicolaides and Ville described their experience with 45 cases with an equal incidence of dual survivors, single survivors, and no survivors.45 Soon thereafter, a European consortium (Eurofoetus) was formed and worked with the Karl Storz Corporation to develop improved fetoscopic instrumentation specific for laser therapy. Between 1999 and 2002, a randomized-clinical trial was undertaken in Europe to compare laser therapy to the gold standard of amnioreduction.46 The study was stopped by the data safety monitoring committee after 50% of the planned enrollment when 142 patients had been randomized. Patients in the laser group were significantly more likely to have at least one survivor at 28 days of neonatal life (76 vs. 56%, p ¼ 0.009) and to have surviving infants free of “neurological abnormalities” at the age of 6 months (52 vs. 31%, p ¼ 0.003). A subsequent meta-analysis confirmed the improved outcome with laser therapy.47 These studies led to fetoscopic-
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
560
directed laser therapy becoming accepted as the standard treatment for advanced cases of severe TTTS. Many fetal interventionists from around the world visited centers in Europe and returned home to offer laser therapy. Evidence soon became available regarding a learning curve until expertise was achieved.48 The most recent data from the combined experience of two centers in the United States indicate that the overall rate of survival at the age of 30 days of at least one neonate is 91% and the survival of both neonates of the twin pair is 67%.49 In 2013, the Society for Maternal– Fetal Medicine endorsed laser as the treatment of choice for advanced stages of TTTS.50 Subsequent modifications in the technique have included a change from nonselective to selective to sequential, selective coagulation of placental anastomoses.51,52 The latter technique has resulted in a higher rate of donor survival. In cases of an anterior placenta with no clear avascular window for percutaneous fetoscopic access, laparoscopic assistance has been reported to allow entry of the fetoscope through the posterior uterus.53 In addition, a Solomon technique creating a complete dechorionation of the placenta using laser energy has proved successful in reducing the incidence of recurrent TTTS and twin anemia-polycythemia sequence.54
Myelomeningocele Until this time, fetal intervention was only considered for lethal fetal diseases. Consideration for fetal MMC repair opened new controversies regarding treatment to prevent significant life-long morbidity. More importantly, the history of MMC repair would set a standard for the scientific approach to the assessment and later implementation of a new fetal therapy. As is so often the case, an experiment of nature would point to the possibility of progressive neurologic damage in fetuses with MMC. In the case of congenital hemimyelocele, there is a duplication of the lower spinal cord in association with a typical appearing MMC. These neonates were shown to demonstrate neurologic compromise of the lower extremity ipsilateral to the exposed hemicord while the covered portion of the cord was associated with normal lower extremity function.55 A further investigation by Osaka et al56 found that in cases of myeloschisis, the Chiari II malformation was notably absent from human embryos up to the age of 55 days although it was present in later fetal life. These human observations led to animal experimentation to see if in utero repair of MMC could avert neurologic damage at birth. Michejda57 surgically created a spinal defect in a rhesus monkey model at days 110 to 125 of gestation (term, 160– 164 days). Immediate repair was undertaken with allogeneic bone paste and skin closure in five animals; repaired animals at birth exhibited normal neurological function and normal spinal cord morphology on histology. In three control animals, the MMC lesion was allowed to remain open; paraplegia with urinary incontinence and somatosensory loss below the level of the induced lesion was observed when these animals were born. Heffez et al58 proposed a “two-hit” hypothesis for the neurological damage of MMC—an initial congenital mye-
Moise
lodysplasia followed by progressive intrauterine injury due to either direct trauma or exposure to the toxic effects of amniotic fluid or both. This group surgically created an MMC in a rat model on day 18 of gestation (term, 22 days) with repair the following day. When compared with controls, these pups exhibited normal neurological function and normal spinal cord histology. Sentinel work in the ovine model is probably cited as the most significant contribution to lead the way to human innovative therapy. Meuli et al59 surgically created an MMC in 12 fetal sheep on day 75 of gestation (term, 150 days). At 100 days, seven fetuses remained alive and all underwent repair of their MMC. Surviving fetuses were delivered by C-section at 145 days and underwent urologic examination and evaluation of somatosensory-evoked potentials. Although some reduction in hind limb strength was noted, evoked potentials were normal. No stool or urinary incontinence was noted. Investigators at Vanderbilt University would be the first to venture into in utero repair of MMC in human fetuses using endoscopy.60 Three endoscopic ports were inserted into uterus and carbon dioxide insufflated into the amniotic cavity. A maternal split-thickness skin graft was then used to cover the MMC. Two perinatal deaths occurred while the other two infants delivered 5.6 and 11.7 weeks after the procedure—both exhibited mild somatosensory deficits at the age of 6 and 30 months. Because of the high perinatal mortality, no further attempts at minimally invasive repair were undertaken. A new approach using an open hysterotomy was developed.61 Three fetuses underwent standard surgical repair and were subsequently delivered at 33-, 34-, and 36-week gestation. One hysterotomy dehiscence with prolapse of the fetal arm into the maternal peritoneal cavity necessitated delivery. Only one of the three infants required a ventriculoperitoneal (V–P) shunt. The authors concluded that open fetal surgery for MMC repair was feasible. In an initial series of 29 patients reported from Vanderbilt with follow-up to the age of 6 months, neonates that had undergone in utero repair were matched to historic controls.62 Ventricular shunts for the treatment of hydrocephalus were needed in 59% of fetal repair cases as compared with 91% of controls. Postnatal magnetic resonance imaging (MRI) revealed the presence of the Chiari II malformation in 38% of fetal repair cases as compared with 91% of controls. A second series of 50 fetal repair cases reported by Children’s Hospital of Philadelphia (CHOP) indicated a perinatal survival rate of 94%.63 Ventricular shunts were required in 67% of thoracic, 44% of lumbar, and 20% of sacral lesions as compared with 100, 88, and 68% of historic controls. Four the United States centers began to offer fetal repair of MMC. A total of 12 cases were completed at the University of California at San Francisco (UCSF), 54 cases at CHOP, 170 cases at Vanderbilt, and 10 cases at the University of North Carolina. In January, 1999, investigators at UCSF applied to the National Institutes of Health (NIH) for a single-center study of in utero MMC repair under the Research Project Grant Program (RO1) mechanism (S. Adzick, e-mail personal communication, 2011). The second submission of the grant received a favorable score and was scheduled for funding in American Journal of Perinatology
Vol. 31
No. 7/2014
561
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
History of Fetal Therapy
History of Fetal Therapy
Moise
December, 2000. In July, 2000, the National Institute of Child Health and Human Development (NICHD) hosted a consensus conference on fetal MMC repair. A recommendation for a multicentered randomized-trial led to the original RO1 application by UCSF being tabled and a new U10 RFA was issued. Three centers (UCSF, Vanderbilt, and CHOP with George Washington University as the independent data monitoring center) qualified. At a meeting held in conjunction with the 19th annual meeting of the International Fetal Medicine and Surgery Society, an agreement was reached to call for a moratorium on fetal MMC repair at other fetal centers not participating in the randomized trial. The official award notice was announced by the NICHD in March, 2002, and the first patient was enrolled in February, 2003. The Management of Myelomeningocele (MOMS) trial began in the spring, 2002, with a predicted period of enrollment period of 3 years.64 A total of 200 patients were to be enrolled —100 randomized to in utero repair of MMC and 100 randomized to standard neonatal repair by the same neurosurgeons. Over 1,000 patients were screened over the course of 9 years to identify the 183 patients that were randomized. Fetal inclusion criteria were the presence of a fetal spinal defect between T12 and S1 in association with a Chiari II malformation, no evidence of kyphosis, gestational age at randomization of 19 weeks to 25 weeks and 6 days, normal karyotype, and no other major congenital anomalies. Maternal exclusions included a body mass index (BMI) more than 35 kg/m2, maternal age younger than 18 years, history of preterm labor, previous uterine incision in the upper segment of the uterus, insulin-dependent diabetes mellitus, red cell or platelet alloimmunization, a history of viral infections (hepatitis and human immunodeficiency virus), or refusal to accept blood products. Patients randomized to fetal surgery had to agree to remain in the city where they underwent fetal surgery with an accompanying companion until delivery. The trial was halted at the fourth interim analysis by the data safety monitoring committee when both the primary (1 year) and secondary outcomes (at 30 months) were found to be improved in the in utero repair group. At 1 year, 40% of infants in the fetal repair group required V–P shunting as compared with 80% in the postnatal repair group. At the age of 30 months, 42% of the children in the fetal repair group walked independently as compared with 21% in the neonatal repair group. Prematurity however complicated 80% of cases in the fetal repair group but was only noted in 15% of patients whose offspring underwent neonatal repair. The three centers that enrolled patients in the MOMS trial began to routinely offer the procedure within weeks of the New England Journal of Medicine report. Other fetal centers with experience in fetal interventions soon began to implement programs for fetal MMC repair. Most worked in collaboration with one of the MOMS trial centers to develop their protocols. In January 2013, the Committee on Obstetric Practice of ACOG issued a committee opinion recommending that pregnant women with a fetal MMC should be made aware of the findings of the MOMS trial and receive counseling regarding the option of fetal repair.65 More recently, a task force on fetal MMC repair proposed strict adherence to the MOMS criteria American Journal of Perinatology
Vol. 31
No. 7/2014
for patient selection and strict adherence to the clinical care protocol used in the trial.10 Centers were also encouraged to receive guidance from established MOMS centers. We are still awaiting the final chapter in the story of fetal MMC repair. Minimally invasive therapy with a patch technique is being developed in animal models.66 Human cases using multiple uterine punctures for a fetoscopic approach have been reported.67 A tissue engineering approach is on the horizon. 68
Congenital Diaphragmatic Hernia Congenital diaphragmatic hernia (CDH) can usually be diagnosed in the second trimester of pregnancy when ultrasound reveals a fetal stomach at the level of the heart. Neonatal intensive care units use methods such as oscillatory ventilation, inhaled nitric oxide, and extra corporeal membrane oxygenation (ECMO) to achieve survival rates of greater than 70%; however, cases diagnosed prenatally were reported in the early 1990s to be associated with a “hidden mortality” due to fetal demise with survival rates of less than 60%.69 Michael Harrison at UCSF led the initial efforts to improve this statistic. He and his coinvestigators created a fetal lamb model by placing inflatable balloons in the thoracic space to mimic the space-occupying lesion of the abdominal viscera in CDH. When the balloon remained inflated, the newborn lambs died soon after birth because of pulmonary hypoplasia. Deflation of the balloons by day 120 gestation (term, 140 days) revealed normal pulmonary development.70 This group of investigators then surgically created a diaphragm defect in fetal lambs at 100 days of gestation; newborn lambs again showed pulmonary hypoplasia at birth. Attempts to repair the defect at day 120 resulted in fetal demise in all cases—this was thought to be secondary to compromised venous return to the heart as a result of the acute displacement of the intrathoracic viscera back into the peritoneal cavity. Creation of an abdominoplasty with a silastic patch solved this problem and resulted in survival of the lambs with normal pulmonary function.71 A series of primate experiments followed with development of a specialized stapling device to prevent bleeding at the time of the hysterotomy. In the 1990, after six consecutive failures, Harrison group reported the first successful in utero repair of CDH in a human pregnancy.5 Further experience with fetal repair indicated that when the liver was herniated into the thoracic cavity, reduction of the liver resulted in compromise of the intrahepatic portion of umbilical venous return and subsequent fetal death. The NIH and March of Dimes subsequently funded a nonrandomizedtrial of hysterotomy for open fetal CHD repair in “liver-down” cases as compared with standard postnatal repair.72 Fifty-five cases of CDH were referred to UCSF; four fetuses underwent in utero repair. Their outcomes were compared with seven cases treated at UCSF with neonatal repair in the same time period. Survival in the fetal repair group was 75 and 86% in the postnatal repair group; mean gestational age at birth was 32 weeks compared with 38-week gestation. On the basis of these results, a moratorium on further open in utero repair of CDH was enacted.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
562
A new approach to the treatment of fetal CDH was needed. Studies in 1948 by Jost and Policard73 in the rabbit demonstrated that pulmonary fluid was produced by the fetal lungs and entered the amniotic cavity through the trachea. Five rabbit fetuses underwent tracheal ligation and exhibited an increase in alveolar size with histological features of the lungs similar to controls. Later, Alcorn et al74 demonstrated in a sheep model that tracheal occlusion (TO) results in increased lung size with thin alveolar walls and a deficit of type II pneumocytes. However, it was another error of nature that would lead to a novel approach for in utero therapy for CDH. Jay Wilson would be in the medical library at Boston Children’s Hospital to copy an article about CDH (J. Wilson, oral personal communication, 2011). On the back page of his article, he happened to see an abstract of a subsequent article that described a case of Fraser syndrome—a genetic syndrome associated with bilateral renal agenesis in conjunction with tracheal atresia.75 He knew from previous animal investigations that anhydramnios from absent urine production should have been associated with lethal pulmonary hypoplasia. Yet, the congenital tracheal atresia had resulted in large lungs. His group studied this concept in fetal lambs.76 Three animals underwent bilateral nephrectomy, three underwent TO alone, and three underwent nephrectomy and TO. Fetal lung volumes in both TO groups were four times larger than controls and five times larger than the lungs of the animals in the nephrectomy alone group. More importantly, the lungs in the TO groups appeared to be histologically mature with the increase in size because of cell multiplication. A single fetal lamb underwent the surgical creation of a diaphragm hernia with a simultaneous TO. At birth, this lamb’s lungs were enlarged and had displaced the herniated abdominal viscera from the thorax. This led the authors to propose TO as a possible fetal treatment for CDH at the Section on Surgery at the 1992 annual meeting of the American Academy of Pediatrics. What followed was an extensive series of experiments in ovine, rabbit, and rat models that confirmed the therapeutic effects of TO in cases of fetal CDH.77 A critical issue remained—who was an appropriate candidate for this innovative therapy? Metkus et al78 was the first to describe the use of ultrasound to calculate the cross-sectional area of the lung on the contralateral side to the diaphragmatic hernia. This was compared with an internal control—the head circumference —and called the lung-to-head ratio (LHR). Universal neonatal mortality was noted when the LHR was less than 0.6; however, a survival rate of 61% was associated with an LHR between 0.6 and 1.35 and was 100% with an LHR more than 1.35. As the normal fetal head circumference increases in size more rapidly than the lung size with advancing gestational age, Jani et al79 introduced the concept of the observed-toexpected ratio to correct for gestational age. As technology advanced, fast spin MRI became part of the routine evaluation of the fetus with CDH.80 Calculated lung volumes have not proven any more predictive of neonatal survival than the LHR. However, MRI has proven beneficial in determining the amount of liver that has herniated into the thoracic cavity, a factor that is predictive of poor neonatal survival.81
Moise
Early efforts to perform TO in the fetal lamb model involved placement of an intratracheal foam insert or an endotracheal tube with a foam cuff with and without a magnetically activated valve.82 Subsequently, investigators at UCSF successfully applied an extratracheal clip using a minimally invasive approach in a fetal lamb model.83 With this background work completed, eight human fetuses with severe CDH were treated with TO using a polymeric foam (n ¼ 2), a spring-loaded aneurysm clip (n ¼ 1), and hemoclips (n ¼ 5) using a maternal hysterotomy for access.84 Unfortunately, all infants delivered prematurely between 27 and 34 weeks gestation, six infants died, one experienced unrelated-bowel necrosis after discharge, and only one had normal respiratory function at the age of 20 months. A novel approach was next proposed by Deprest et al in Leuven, Belgium.85 A special fetoscope was introduced into the trachea of fetal lambs and a detachable endovascular balloon used for occlusion. The UCSF group moved quickly to implement this method in two human fetuses with right-sided CDH; both infants were delivered by an ex utero intrapartum therapy procedure with good outcomes.86 Between April, 1999, and July, 2001, an NIH-funded randomized-clinical trial of TO was undertaken at UCSF in fetuses with a left-sided CDH, liver herniated into the chest, and LHR less than 1.4.87 A total of 13 patients were randomized to routine neonatal management and 11 patients underwent TO. Interestingly, the first two patients in the TO group underwent fetoscopic placement of extratracheal clips; subsequently, the data safety monitoring committee allowed a change in TO methodology to an intraluminal balloon. Neonatal survival in the TO group was 77% as compared with 73% in the postnatal treatment group. TO patients delivered 6 weeks earlier than the standard treatment group. Additional studies in the CDH fetal lamb model demonstrated that 15 days after TO was sufficient to stimulate lung growth. If the tracheal balloon was removed several days before delivery, recovery of type II pneumocytes occurred.88 This data in the lamb model led to a proposal to maintain TO for 6 to 8 weeks in the human fetus with release at 34-week gestation. The FETO (fetal TO) task force was formed through an alliance of fetal centers in London, Barcelona, and Leuven to undertake a feasibility trial of fetoscopic-guided TO for the treatment of fetal diaphragmatic hernia.89 Fetuses with either a left- or right-sided CDH with “liver-up” and an LHR of
557
The History of Fetal Therapy Kenneth J. Moise, Jr., MD1
UT Health School of Medicine, Houston, Texas Am J Perinatol 2014;31:557–566.
Abstract Keywords
► fetal surgery ► fetal intervention ► maternal–fetal surgery ► hemolytic disease of the fetus/newborn ► fetal myelomeningocele ► twin–twin transfusion ► fetal diaphragmatic hernia
Address for correspondence Kenneth J. Moise, Jr., MD, Department of Obstetrics, Gynecology and Reproductive Sciences, Division of MaternalFetal Medicine, UT Health School of Medicine, 6410 Fannin, Suite 700, Houston, TX 77030 (e-mail: [email protected]).
The Fetal Treatment Center founded by Michael Harrison is credited as the birthplace of fetal surgery. His trainees in pediatric surgery subsequently founded fetal centers throughout the United States. In Europe, the advent of minimally invasive fetal surgical techniques led to the establishment of treatment centers led predominantly by perinatologists. More recently, perinatologists in North America have begun to play a greater role in the field of fetal intervention. Intrauterine transfusion for the treatment of hemolytic disease of the fetus/newborn was the first successful fetal intervention. Although not subjected to the rigors of clinical trials, this treatment has withstood the test of time. Interventions for other fetal disease states such as twin–twin transfusion and repair of fetal myelomeningocele were investigated in animal models followed by randomized clinical trials before widespread adoption. Tracheal occlusion for diaphragmatic hernia is still currently being investigated as the next promising step in fetal intervention.
Michael Harrison is widely regarded as the “father of fetal surgery.”1 In a YouTube interview, he tells the story how during his first month of his general surgery internship at Massachusetts General Hospital he assisted in a neonatal surgery to correct a diaphragmatic hernia.2 He describes the surgery as a “beautiful procedure” however the infant died the following day from pulmonary hypoplasia. He goes on to say that he realized “that the only way to make that lung bigger is to fix something before birth so it has room to grow.” And so began Harrison’s 30-year adventure in the treatment of the unborn. He established the Fetal Treatment Center at the University of San Francisco (UCSF) in the early 1980s. The first successful open surgery for fetal lower urinary tract obstruction was performed in 1981; open surgical resection of fetal congenital adenomatoid malformation followed in 1984 with the first open repair of a fetal diaphragmatic hernia in 1989.3–5 Harrison protégés and their trainees went on to establish fetal treatment centers at many of the major children’s hospitals throughout the United States including those in Philadelphia, Boston, Cincinnati, Nashville, Houston, St Louis, and Denver. Pediatric surgeons initially took the lead
on these endeavors. Maternal–fetal medicine (MFM) specialists, the physicians that often made the initial prenatal diagnosis of fetal anomalies, were relegated to support roles at these centers. Across the pond in Britain and Europe, the picture was much different. Open fetal surgery to correct congenital anomalies was felt to be associated with an unacceptably high rate of maternal morbidity. However, with the advent of minimally invasive laser intervention for severe twin–twin transfusion King’s College in London became the epicenter for fetal therapy. As an obstetrician with expertise in fetal medicine, Kypros Nicolaides mentored other obstetricians in this emerging field. Fetal centers in Hamburg, Paris, Leuven, Barcelona, and Leiden soon emerged. A collaboration of European fetal surgery groups (Eurofoetus) developed to enroll patients in multicentered trials. Back in the United States, MFMs began to play more of an active role in fetal intervention. In August, 2004, an advisory committee that included pediatric surgeons and MFM that were active in the field attended a workshop sponsored by the National Institutes of Health and the Office of Rare Diseases.6 The term Maternal–Fetal therapy was introduced as the
received October 21, 2013 accepted after revision November 27, 2013 published online February 25, 2014
Copyright © 2014 by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York, NY 10001, USA. Tel: +1(212) 584-4662.
DOI http://dx.doi.org/ 10.1055/s-0033-1364191. ISSN 0735-1631.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
1 Department of Obstetrics, Gynecology and Reproductive Sciences,
History of Fetal Therapy
Moise
pregnant woman reaps no direct benefit from fetal surgery yet she incurs substantial risks. Issues of preclinical trials, patient consent and the need for multicenter clinical trials were addressed. Led by four perinatologists, the North American Fetal Treatment Network (NAFTNet) was established in October, 2004 to mirror the role of the Eurofoetus network.7 Meanwhile, laser centers for the treatment of severe twin– twin transfusion were led mostly by MFMs—many who considered twin-to-twin transfusion syndrome (TTTS) to be an obstetrical disease as it was related to placentation. Soon the term fetal intervention began to replace the term fetal surgery because TTTS treatments were performed more frequently than any other fetal procedures. MFMs began to play greater roles in the leadership of fetal centers often serving as codirectors with their pediatric surgical colleagues. In the most recent development of fetal centers, children’s hospitals without already established on-site delivery services have developed special delivery units and even women’s pavilions to allow for on-site obstetrical delivery for those pregnant women with a fetus diagnosed with a congenital anomaly. Many questions still lie ahead for the world of fetal intervention, particularly in the United States.8 These will become especially important with ever shrinking financial resources. Should we follow the model established by our European colleagues where some smaller countries only have one designated center? Clearly, the geographic distribution of fetal centers must be considered in the United States to allow for easy patient access. How many centers should exist so that patient numbers are not diluted thus allowing opportunity for education and research to further the field? Should centers be designated as centers of excellence based on experience and outcomes much like current transplant centers?
What is a Fetal Center? The definition of a fetal center is notably lacking from the literature. A recent committee opinion from the American College of Obstetricians and Gynecologists (ACOG) addresses this topic and noted that fetal centers exist in many forms. They are free standing centers (usually in conjunction with a children’s hospital) or centers can exist within established obstetrical or pediatric departments. The majority of fetal centers represent multidisciplinary clinics where pregnant women with fetal anomalies can receive diagnostic services, pregnancy management, prenatal consultation with pediatric subspecialists and coordinated planning for delivery and postdelivery neonatal care. These might be considered tier one centers. Often needle-based fetal intervention procedures are offered such as intrauterine transfusions or shunt placement. A tier two center would offer more advanced therapies such as laser ablation for twin–twin transfusion. Finally, a third tier center would offer the full gamut of accept fetal interventions including open fetal surgery for repair of spina bifida and for the removal of other rare fetal lesions. ACOG recommends that the fetal intervention team include MFM specialists to protect the interests of the pregnant woman.9 A recent task force statement on MMC repair proposed some of American Journal of Perinatology
Vol. 31
No. 7/2014
the components of an experienced fetal care team (►Table 1).10 Perhaps, the most important members of the team are the clinical care coordinators. These are often nurses with obstetrical or pediatric background, genetic counselors, social workers, or midwives. They serve as the “glue” to be sure that all the diverse subspecialists are on the same page for coordinated care. Other manuscripts in this special edition of the American Journal of Perinatology will highlight the evolving and expanded role of members of the fetal team. Tier three centers should also exhibit a commitment to education. Currently, four United States centers offer clinical fellowships in fetal intervention. Tier three centers should also be involved in research to further the development of new fetal interventions. Such research includes enrollment in clinical trials, submission of data to national registries and experimental work in animal models. The remainder of this chapter will elucidate the history and the basis for the treatment of the four most common conditions associated with fetal intervention: hemolytic disease of the fetus/newborn, twin–twin transfusion, myelomeningocele (MMC), and diaphragmatic hernia.
Hemolytic Disease of the Fetus and Newborn The story of the conquest of hemolytic disease of the fetus and newborn (HDFN) teaches us important lessons in fetal intervention. When solutions to this enormous clinical challenge were first being conceived and investigated, the internet was not yet even a concept and FAX machines were years in the future. Pioneering innovators in different parts of the world developed new concepts many times on parallel tracks. Descriptions of the newborn affected by HDFN can be found as far back as Hippocrates. In 1932, Diamond et al described the neonatal disease and proposed the term erythroblastosis fetalis based on the finding of circulating erythroblasts (nucleated red blood cells).11 A decade later, Levine et al12 were able to demonstrate a causal relationship between RhD antibodies and HDFN. Exchange transfusion in the neonate was introduced as the first effective therapy for those fetuses that survived to birth.13 However, 30% of fetuses continued to die in utero. Premature delivery was attempted to prevent these cases however a clinical trial in England failed to find that this resulted in improved neonatal outcome as prematurity itself was often associated with neonatal death.14 A more targeted diagnostic approach was developed based on the work by Bevis in England and later Liley in New Zealand.15,16 These investigators analyzed the bilirubin content of amniotic fluid to determine when imminent fetal death might occur. Unaware of the previous work of Liley on the other side of the world, Freda at Columbia-Presbyterian Hospital in New York began a specialized Rh antepartum clinic to test the idea that serial amniocenteses could be used to predict the severity of HDFN.17 He was able to demonstrate a reduction in the perinatal mortality from 30 to 9% over the course of 5 years. Bevis is credited with the first attempt at fetal intervention for HDFN.18 Using radiopaque dye injected into the mother and fluoroscopy to outline the placental vasculature, he
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
558
Table 1 Components of a fetal center Specialized services • Fetal MRI • Fetal echocardiography • Advanced ultrasound imaging Team members • Maternal–fetal medicine specialist • Pediatric surgeon • Other pediatric surgical specialists—neurosurgery, urology, plastic, craniofacial, and orthopedic • Pediatric specialists—neonatology, cardiology, radiology, nephrology, and neurology • Social work • Child life specialist • Obstetrical anesthesiologist • Pediatric anesthesiologist • Care coordinator • Genetic counselor Functional components • Designated leadership team • Regular multidisciplinary meetings to discuss cases • Fetal board or ethics committee • Quality control procedures • Research • Commitment to education • Long-term follow-up clinics Abbreviation: MRI, magnetic resonance imaging.
attempted to inject blood directly into the placenta. Several fetuses died; however, RhD negative red cells were detected in the circulation of one neonate. Sir William Liley is credited with the first successful intraperitoneal fetal transfusion (IPT).19 A visiting young geneticist who had recently worked with missionaries in Nigeria noted that red cells had been successfully transfused into the peritoneal cavity of neonates with sickle cell disease. Normal appearing red blood cells were later noted on peripheral blood smears. On a previous occasion, Liley realized that he had accidentally entered the peritoneal cavity of a fetus at the time of amniocentesis. He recognized this because the yellow color of the ascitic fluid was more intense than that of previous amniotic fluid samples he had obtained. During the case, he injected air into the fetal peritoneal cavity and then injected radiopaque dye into the amniotic fluid as the needle was removed. A radiograph taken a short time later revealed an image of the fetal peritoneal cavity with the ingested dye outlining the fetal gastrointestinal tract. Liley felt that access to the fetal peritoneal cavity could be accomplished for the infusion of red cells. Institutional review boards that we know today did not yet exist. A committee of three senior obstetricians (including Liley) was convened and a decision made to attempt an
Moise
intraperitioneal intrauterine transfusion. After three unsuccessful attempts that resulted in fetal demise, the fourth fetus was delivered at 34-week and 7-day gestation after undergoing two IPTs at 32 weeks and 1 day and 33 weeks and 4 days. Further innovations by Liley included the use of metal grids to aid in the location of the needle entry site into the fetus and even the use of multiple needles to immobilize the fetus. Several investigators challenged the IPT approach as it was not equivalent to the neonatal exchange transfusion that was so effective after birth. Direct access to the fetal circulation was needed. Investigators at Columbia University in New York were the first to attempt the direct transfusion of red cells to the fetus.18 Access to the fetal superior sagittal sinus was achieved in a rhesus monkey. A hysterotomy was then undertaken in a human case at 32-week gestation. Blood was successfully aspirated from the superior sagittal fetal vein however catheter access could not be achieved; a perinatal loss occurred after a premature delivery several days later. Subsequently, several investigators attempted to transfuse the anemic fetus by direct access at the time of hysterotomy using the fetal femoral artery, saphenous vein, and internal jugular vein.20–22 All resulted in premature delivery and perinatal loss. Adamsons et al23 attempted to place a long-standing indwelling catheter into the fetal peritoneal cavity for the periodic infusion of red cells. Four attempts failed, but a fifth procedure performed in Brazil at 24-week gestation resulted in a live born infant 8 weeks later. Liley came to the New York on sabbatical to introduce the IPT method to American physicians. Early attempts at IPT involved the introduction of 14- or 16-gauge Tuohy needle into the fetal peritoneal cavity followed by injection of a radiopaque dye under continuous fluoroscopy. An epidural catheter was then placed through the needle and the needle withdrawn. Blood could then be infused through the catheter followed by its removal. The late 1970s saw the introduction of real-time ultrasound to direct the transfusion needle. A Boston group reported a 53% rate of survival for nonhydropic fetuses and 7% survival with hydrops. After the introduction of ultrasound for needle guidance, a survival rate of 62% was reported; however, survival in the hydropic fetus remained poor with a salvage rate of only 29%.24 No further refinements in the IPT method occurred until 1985 when intramuscular paralytic agents were introduced to cause cessation of fetal movements.25 Smaller diameter needles could now be used and indwelling catheters were no longer necessary. Hydrops fetalis continued to be associated with poor fetal survival as it was realized that intraperitoneal blood was not absorbed in these sick fetuses. Once again, fetal interventionists would attempt to gain direct access to the fetal circulation. In 1981, Rodeck et al26 performed the first successful intravascular fetal transfusion (IVT) using a fetoscope to guide the transfusion needle into a placental plate vessel. One year later, Bang et al27 performed the first ultrasound-guided IVT using the intrahepatic portion of the umbilical vein. In an attempt to evaluate fetuses for in utero infection with toxoplasmosis, French investigators were able to successfully access the umbilical vessels for sampling.28 The intravascular American Journal of Perinatology
Vol. 31
No. 7/2014
559
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
History of Fetal Therapy
History of Fetal Therapy
Moise
intrauterine transfusion was born and became the standard method of intervention for HDFN. An initial fetal hematocrit could now be obtained and a specific amount volume of packed red cells infused to achieve a desired final hematocrit value. Minimal refinements in the procedure followed. Some fetal interventionalists found that the intrahepatic portion of the umbilical vein was a better access point for the IVT.29 Paralytic agents injected at the start of the transfusion were noted to decrease the chance for fetal movement that could interfere with a successful procedure.30 A step-wise approach to the correction of severe fetal anemia in the early second trimester fetus improved fetal survival.31 In a recent series of over 1,400 IUTs in almost 500 fetuses, a survival rate of 91% in nonhydropic fetuses and 83% in hydropic ones has been reported from a national referral center in the Netherlands.32
Twin-to-Twin Transfusion Syndrome A description of the “third circulation” in the placentae of monochorionic twins can be traced back to placental studies in the late 1800s.33 Benirschke and Kim34 suggested that vascular placental connections were responsible for the clinical phenotype of twin-to-twin transfusion. Subsequent injection studies in the placentae of patients with monochorionic twins complicated by oligohydramnios-polyhydramnios sequence indicated a predominance of unidirectional flow through vascular anastomoses from the “donor” fetus to the “recipient.”35 Untreated TTTS presenting in the second trimester was associated with a perinatal mortality of up to 70%.36 Progressive polyhydramnios would often lead to premature rupture of membranes with delivery before a viable gestational age; dual fetal demise often occurred as well. Amnioreduction of the polyhydramnios relieved maternal symptoms but did not improve survival.37 TTTS rarely complicates monoamniotic twins. This led investigators to study the role of septostomy (a purposeful penetration of the intertwin membrane) to equalize the amniotic fluid volumes in both twins’ sacs.38 A randomized clinical trial between amnioreduction and septostomy however failed to show a therapeutic benefit to septostomy.39 The neodymium:yttrium-aluminum-garnet (NG-YAG) laser can transmit energy through a liquid medium. In 1983, DeVore et al40 began to explore if this modality could be used in an amniotic fluid environment for fetal intervention. A hysterotomy was performed in two lambs and YAG laser energy transmitted through a 600 µm quartz fiber was used to sever the tail, limb, and umbilical cord in an avascular fashion. In two additional cases, these investigators introduced a fetoscope by laparotomy and inserted the laser fiber just adjacent to it. Coagulation and excision of various tissues was successfully undertaken. Further work was undertaken in pregnant ewes. Two underwent hysterotomy and placental vessels of varying diameters were coagulated and then transected to be sure that the vascular flow had been disrupted. In two additional ewes, a laparotomy was undertaken and the laser fiber was introduced into the amniotic cavity through the side port of a pediatric cystoscope. Placental vessels were American Journal of Perinatology
Vol. 31
No. 7/2014
visualized and lasered. Subsequently, these vessels were retrieved and histologic evaluation indicated complete occlusion of their lumen with minimal damage to surrounding placental tissues. In a letter dated June 19, 1984, De Lia contacted Bernische and inquired as to what species of animals might best represent human placentation.41 He felt that the treatment of TTTS with laser had the “potential of being the most commonly performed fetal surgery.” This led to an investigation in rhesus monkeys targeting the vascular connections between the two lobes of their placentae.42 The first case was a failure when the laser fiber was introduced through a separate incision. Subsequent cases involved attachment of the fiber to the side of the fetoscope with a suture. In only one of these cases was the planned placental vascular coagulation successful. Other cases were complicated by stillbirth, abruption, infection, uterine hypertonicity, and inadvertent coagulation of vessels on the wrong placental disk. De Lia went on to extensively study the shared vasculature of placentae from monochorionic twin gestations (J. E. De Lia, MD, personal communication, 1990). The first three cases of laser for TTTS in human pregnancies were reported by De Lia in 1990.43 All fetuses survived the initial surgery at 18.5, 22, and 22.5 weeks. In the first two cases, preterm premature rupture of the membranes (PPROM) with delivery occurred 8.5 and 12 weeks after the surgery. The third case was delivered 6.5 weeks after the laser procedure because of severe maternal preeclampsia. Four of the six fetuses survived the neonatal period. In the United States, De Lia in Salt Lake City (later in Milwaukee) and Quintero in Detroit (later in Tampa) began to offer laser therapy for TTTS. Considerable doubt remained in the perinatal community regarding the benefit of laser therapy. How could one possibly visualize all of the possible anastomoses between the twins? Complications of PPROM, abruption, and even a maternal death were known to occur. Amnioreduction was routinely employed as part of the laser surgery; it alone had proved therapeutic in some reported cases of TTTS. In his first large clinical series of 26 patients, De Lia reported that 35% of cases had dual survivors, 31% had one survivor, and 35% had no survivors.44 A second center at the Harris Birthright Centre at King’s College in London adopted this new therapy. Nicolaides and Ville described their experience with 45 cases with an equal incidence of dual survivors, single survivors, and no survivors.45 Soon thereafter, a European consortium (Eurofoetus) was formed and worked with the Karl Storz Corporation to develop improved fetoscopic instrumentation specific for laser therapy. Between 1999 and 2002, a randomized-clinical trial was undertaken in Europe to compare laser therapy to the gold standard of amnioreduction.46 The study was stopped by the data safety monitoring committee after 50% of the planned enrollment when 142 patients had been randomized. Patients in the laser group were significantly more likely to have at least one survivor at 28 days of neonatal life (76 vs. 56%, p ¼ 0.009) and to have surviving infants free of “neurological abnormalities” at the age of 6 months (52 vs. 31%, p ¼ 0.003). A subsequent meta-analysis confirmed the improved outcome with laser therapy.47 These studies led to fetoscopic-
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
560
directed laser therapy becoming accepted as the standard treatment for advanced cases of severe TTTS. Many fetal interventionists from around the world visited centers in Europe and returned home to offer laser therapy. Evidence soon became available regarding a learning curve until expertise was achieved.48 The most recent data from the combined experience of two centers in the United States indicate that the overall rate of survival at the age of 30 days of at least one neonate is 91% and the survival of both neonates of the twin pair is 67%.49 In 2013, the Society for Maternal– Fetal Medicine endorsed laser as the treatment of choice for advanced stages of TTTS.50 Subsequent modifications in the technique have included a change from nonselective to selective to sequential, selective coagulation of placental anastomoses.51,52 The latter technique has resulted in a higher rate of donor survival. In cases of an anterior placenta with no clear avascular window for percutaneous fetoscopic access, laparoscopic assistance has been reported to allow entry of the fetoscope through the posterior uterus.53 In addition, a Solomon technique creating a complete dechorionation of the placenta using laser energy has proved successful in reducing the incidence of recurrent TTTS and twin anemia-polycythemia sequence.54
Myelomeningocele Until this time, fetal intervention was only considered for lethal fetal diseases. Consideration for fetal MMC repair opened new controversies regarding treatment to prevent significant life-long morbidity. More importantly, the history of MMC repair would set a standard for the scientific approach to the assessment and later implementation of a new fetal therapy. As is so often the case, an experiment of nature would point to the possibility of progressive neurologic damage in fetuses with MMC. In the case of congenital hemimyelocele, there is a duplication of the lower spinal cord in association with a typical appearing MMC. These neonates were shown to demonstrate neurologic compromise of the lower extremity ipsilateral to the exposed hemicord while the covered portion of the cord was associated with normal lower extremity function.55 A further investigation by Osaka et al56 found that in cases of myeloschisis, the Chiari II malformation was notably absent from human embryos up to the age of 55 days although it was present in later fetal life. These human observations led to animal experimentation to see if in utero repair of MMC could avert neurologic damage at birth. Michejda57 surgically created a spinal defect in a rhesus monkey model at days 110 to 125 of gestation (term, 160– 164 days). Immediate repair was undertaken with allogeneic bone paste and skin closure in five animals; repaired animals at birth exhibited normal neurological function and normal spinal cord morphology on histology. In three control animals, the MMC lesion was allowed to remain open; paraplegia with urinary incontinence and somatosensory loss below the level of the induced lesion was observed when these animals were born. Heffez et al58 proposed a “two-hit” hypothesis for the neurological damage of MMC—an initial congenital mye-
Moise
lodysplasia followed by progressive intrauterine injury due to either direct trauma or exposure to the toxic effects of amniotic fluid or both. This group surgically created an MMC in a rat model on day 18 of gestation (term, 22 days) with repair the following day. When compared with controls, these pups exhibited normal neurological function and normal spinal cord histology. Sentinel work in the ovine model is probably cited as the most significant contribution to lead the way to human innovative therapy. Meuli et al59 surgically created an MMC in 12 fetal sheep on day 75 of gestation (term, 150 days). At 100 days, seven fetuses remained alive and all underwent repair of their MMC. Surviving fetuses were delivered by C-section at 145 days and underwent urologic examination and evaluation of somatosensory-evoked potentials. Although some reduction in hind limb strength was noted, evoked potentials were normal. No stool or urinary incontinence was noted. Investigators at Vanderbilt University would be the first to venture into in utero repair of MMC in human fetuses using endoscopy.60 Three endoscopic ports were inserted into uterus and carbon dioxide insufflated into the amniotic cavity. A maternal split-thickness skin graft was then used to cover the MMC. Two perinatal deaths occurred while the other two infants delivered 5.6 and 11.7 weeks after the procedure—both exhibited mild somatosensory deficits at the age of 6 and 30 months. Because of the high perinatal mortality, no further attempts at minimally invasive repair were undertaken. A new approach using an open hysterotomy was developed.61 Three fetuses underwent standard surgical repair and were subsequently delivered at 33-, 34-, and 36-week gestation. One hysterotomy dehiscence with prolapse of the fetal arm into the maternal peritoneal cavity necessitated delivery. Only one of the three infants required a ventriculoperitoneal (V–P) shunt. The authors concluded that open fetal surgery for MMC repair was feasible. In an initial series of 29 patients reported from Vanderbilt with follow-up to the age of 6 months, neonates that had undergone in utero repair were matched to historic controls.62 Ventricular shunts for the treatment of hydrocephalus were needed in 59% of fetal repair cases as compared with 91% of controls. Postnatal magnetic resonance imaging (MRI) revealed the presence of the Chiari II malformation in 38% of fetal repair cases as compared with 91% of controls. A second series of 50 fetal repair cases reported by Children’s Hospital of Philadelphia (CHOP) indicated a perinatal survival rate of 94%.63 Ventricular shunts were required in 67% of thoracic, 44% of lumbar, and 20% of sacral lesions as compared with 100, 88, and 68% of historic controls. Four the United States centers began to offer fetal repair of MMC. A total of 12 cases were completed at the University of California at San Francisco (UCSF), 54 cases at CHOP, 170 cases at Vanderbilt, and 10 cases at the University of North Carolina. In January, 1999, investigators at UCSF applied to the National Institutes of Health (NIH) for a single-center study of in utero MMC repair under the Research Project Grant Program (RO1) mechanism (S. Adzick, e-mail personal communication, 2011). The second submission of the grant received a favorable score and was scheduled for funding in American Journal of Perinatology
Vol. 31
No. 7/2014
561
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
History of Fetal Therapy
History of Fetal Therapy
Moise
December, 2000. In July, 2000, the National Institute of Child Health and Human Development (NICHD) hosted a consensus conference on fetal MMC repair. A recommendation for a multicentered randomized-trial led to the original RO1 application by UCSF being tabled and a new U10 RFA was issued. Three centers (UCSF, Vanderbilt, and CHOP with George Washington University as the independent data monitoring center) qualified. At a meeting held in conjunction with the 19th annual meeting of the International Fetal Medicine and Surgery Society, an agreement was reached to call for a moratorium on fetal MMC repair at other fetal centers not participating in the randomized trial. The official award notice was announced by the NICHD in March, 2002, and the first patient was enrolled in February, 2003. The Management of Myelomeningocele (MOMS) trial began in the spring, 2002, with a predicted period of enrollment period of 3 years.64 A total of 200 patients were to be enrolled —100 randomized to in utero repair of MMC and 100 randomized to standard neonatal repair by the same neurosurgeons. Over 1,000 patients were screened over the course of 9 years to identify the 183 patients that were randomized. Fetal inclusion criteria were the presence of a fetal spinal defect between T12 and S1 in association with a Chiari II malformation, no evidence of kyphosis, gestational age at randomization of 19 weeks to 25 weeks and 6 days, normal karyotype, and no other major congenital anomalies. Maternal exclusions included a body mass index (BMI) more than 35 kg/m2, maternal age younger than 18 years, history of preterm labor, previous uterine incision in the upper segment of the uterus, insulin-dependent diabetes mellitus, red cell or platelet alloimmunization, a history of viral infections (hepatitis and human immunodeficiency virus), or refusal to accept blood products. Patients randomized to fetal surgery had to agree to remain in the city where they underwent fetal surgery with an accompanying companion until delivery. The trial was halted at the fourth interim analysis by the data safety monitoring committee when both the primary (1 year) and secondary outcomes (at 30 months) were found to be improved in the in utero repair group. At 1 year, 40% of infants in the fetal repair group required V–P shunting as compared with 80% in the postnatal repair group. At the age of 30 months, 42% of the children in the fetal repair group walked independently as compared with 21% in the neonatal repair group. Prematurity however complicated 80% of cases in the fetal repair group but was only noted in 15% of patients whose offspring underwent neonatal repair. The three centers that enrolled patients in the MOMS trial began to routinely offer the procedure within weeks of the New England Journal of Medicine report. Other fetal centers with experience in fetal interventions soon began to implement programs for fetal MMC repair. Most worked in collaboration with one of the MOMS trial centers to develop their protocols. In January 2013, the Committee on Obstetric Practice of ACOG issued a committee opinion recommending that pregnant women with a fetal MMC should be made aware of the findings of the MOMS trial and receive counseling regarding the option of fetal repair.65 More recently, a task force on fetal MMC repair proposed strict adherence to the MOMS criteria American Journal of Perinatology
Vol. 31
No. 7/2014
for patient selection and strict adherence to the clinical care protocol used in the trial.10 Centers were also encouraged to receive guidance from established MOMS centers. We are still awaiting the final chapter in the story of fetal MMC repair. Minimally invasive therapy with a patch technique is being developed in animal models.66 Human cases using multiple uterine punctures for a fetoscopic approach have been reported.67 A tissue engineering approach is on the horizon. 68
Congenital Diaphragmatic Hernia Congenital diaphragmatic hernia (CDH) can usually be diagnosed in the second trimester of pregnancy when ultrasound reveals a fetal stomach at the level of the heart. Neonatal intensive care units use methods such as oscillatory ventilation, inhaled nitric oxide, and extra corporeal membrane oxygenation (ECMO) to achieve survival rates of greater than 70%; however, cases diagnosed prenatally were reported in the early 1990s to be associated with a “hidden mortality” due to fetal demise with survival rates of less than 60%.69 Michael Harrison at UCSF led the initial efforts to improve this statistic. He and his coinvestigators created a fetal lamb model by placing inflatable balloons in the thoracic space to mimic the space-occupying lesion of the abdominal viscera in CDH. When the balloon remained inflated, the newborn lambs died soon after birth because of pulmonary hypoplasia. Deflation of the balloons by day 120 gestation (term, 140 days) revealed normal pulmonary development.70 This group of investigators then surgically created a diaphragm defect in fetal lambs at 100 days of gestation; newborn lambs again showed pulmonary hypoplasia at birth. Attempts to repair the defect at day 120 resulted in fetal demise in all cases—this was thought to be secondary to compromised venous return to the heart as a result of the acute displacement of the intrathoracic viscera back into the peritoneal cavity. Creation of an abdominoplasty with a silastic patch solved this problem and resulted in survival of the lambs with normal pulmonary function.71 A series of primate experiments followed with development of a specialized stapling device to prevent bleeding at the time of the hysterotomy. In the 1990, after six consecutive failures, Harrison group reported the first successful in utero repair of CDH in a human pregnancy.5 Further experience with fetal repair indicated that when the liver was herniated into the thoracic cavity, reduction of the liver resulted in compromise of the intrahepatic portion of umbilical venous return and subsequent fetal death. The NIH and March of Dimes subsequently funded a nonrandomizedtrial of hysterotomy for open fetal CHD repair in “liver-down” cases as compared with standard postnatal repair.72 Fifty-five cases of CDH were referred to UCSF; four fetuses underwent in utero repair. Their outcomes were compared with seven cases treated at UCSF with neonatal repair in the same time period. Survival in the fetal repair group was 75 and 86% in the postnatal repair group; mean gestational age at birth was 32 weeks compared with 38-week gestation. On the basis of these results, a moratorium on further open in utero repair of CDH was enacted.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
562
A new approach to the treatment of fetal CDH was needed. Studies in 1948 by Jost and Policard73 in the rabbit demonstrated that pulmonary fluid was produced by the fetal lungs and entered the amniotic cavity through the trachea. Five rabbit fetuses underwent tracheal ligation and exhibited an increase in alveolar size with histological features of the lungs similar to controls. Later, Alcorn et al74 demonstrated in a sheep model that tracheal occlusion (TO) results in increased lung size with thin alveolar walls and a deficit of type II pneumocytes. However, it was another error of nature that would lead to a novel approach for in utero therapy for CDH. Jay Wilson would be in the medical library at Boston Children’s Hospital to copy an article about CDH (J. Wilson, oral personal communication, 2011). On the back page of his article, he happened to see an abstract of a subsequent article that described a case of Fraser syndrome—a genetic syndrome associated with bilateral renal agenesis in conjunction with tracheal atresia.75 He knew from previous animal investigations that anhydramnios from absent urine production should have been associated with lethal pulmonary hypoplasia. Yet, the congenital tracheal atresia had resulted in large lungs. His group studied this concept in fetal lambs.76 Three animals underwent bilateral nephrectomy, three underwent TO alone, and three underwent nephrectomy and TO. Fetal lung volumes in both TO groups were four times larger than controls and five times larger than the lungs of the animals in the nephrectomy alone group. More importantly, the lungs in the TO groups appeared to be histologically mature with the increase in size because of cell multiplication. A single fetal lamb underwent the surgical creation of a diaphragm hernia with a simultaneous TO. At birth, this lamb’s lungs were enlarged and had displaced the herniated abdominal viscera from the thorax. This led the authors to propose TO as a possible fetal treatment for CDH at the Section on Surgery at the 1992 annual meeting of the American Academy of Pediatrics. What followed was an extensive series of experiments in ovine, rabbit, and rat models that confirmed the therapeutic effects of TO in cases of fetal CDH.77 A critical issue remained—who was an appropriate candidate for this innovative therapy? Metkus et al78 was the first to describe the use of ultrasound to calculate the cross-sectional area of the lung on the contralateral side to the diaphragmatic hernia. This was compared with an internal control—the head circumference —and called the lung-to-head ratio (LHR). Universal neonatal mortality was noted when the LHR was less than 0.6; however, a survival rate of 61% was associated with an LHR between 0.6 and 1.35 and was 100% with an LHR more than 1.35. As the normal fetal head circumference increases in size more rapidly than the lung size with advancing gestational age, Jani et al79 introduced the concept of the observed-toexpected ratio to correct for gestational age. As technology advanced, fast spin MRI became part of the routine evaluation of the fetus with CDH.80 Calculated lung volumes have not proven any more predictive of neonatal survival than the LHR. However, MRI has proven beneficial in determining the amount of liver that has herniated into the thoracic cavity, a factor that is predictive of poor neonatal survival.81
Moise
Early efforts to perform TO in the fetal lamb model involved placement of an intratracheal foam insert or an endotracheal tube with a foam cuff with and without a magnetically activated valve.82 Subsequently, investigators at UCSF successfully applied an extratracheal clip using a minimally invasive approach in a fetal lamb model.83 With this background work completed, eight human fetuses with severe CDH were treated with TO using a polymeric foam (n ¼ 2), a spring-loaded aneurysm clip (n ¼ 1), and hemoclips (n ¼ 5) using a maternal hysterotomy for access.84 Unfortunately, all infants delivered prematurely between 27 and 34 weeks gestation, six infants died, one experienced unrelated-bowel necrosis after discharge, and only one had normal respiratory function at the age of 20 months. A novel approach was next proposed by Deprest et al in Leuven, Belgium.85 A special fetoscope was introduced into the trachea of fetal lambs and a detachable endovascular balloon used for occlusion. The UCSF group moved quickly to implement this method in two human fetuses with right-sided CDH; both infants were delivered by an ex utero intrapartum therapy procedure with good outcomes.86 Between April, 1999, and July, 2001, an NIH-funded randomized-clinical trial of TO was undertaken at UCSF in fetuses with a left-sided CDH, liver herniated into the chest, and LHR less than 1.4.87 A total of 13 patients were randomized to routine neonatal management and 11 patients underwent TO. Interestingly, the first two patients in the TO group underwent fetoscopic placement of extratracheal clips; subsequently, the data safety monitoring committee allowed a change in TO methodology to an intraluminal balloon. Neonatal survival in the TO group was 77% as compared with 73% in the postnatal treatment group. TO patients delivered 6 weeks earlier than the standard treatment group. Additional studies in the CDH fetal lamb model demonstrated that 15 days after TO was sufficient to stimulate lung growth. If the tracheal balloon was removed several days before delivery, recovery of type II pneumocytes occurred.88 This data in the lamb model led to a proposal to maintain TO for 6 to 8 weeks in the human fetus with release at 34-week gestation. The FETO (fetal TO) task force was formed through an alliance of fetal centers in London, Barcelona, and Leuven to undertake a feasibility trial of fetoscopic-guided TO for the treatment of fetal diaphragmatic hernia.89 Fetuses with either a left- or right-sided CDH with “liver-up” and an LHR of
