The implementation of systematic pregnancy screening programmes, and the increased use and improving quality of medical imaging techniques, have lead to earlier detection and better understanding of the natural history of fetal anomalies. Where most fetal conditions are adequately treatable after birth, some disorders progress during fetal life and can lead to severe morbidity or fetal and neonatal demise. This inherently raises the question of prenatal therapy. Some fetal conditions are amenable for fetal surgical intervention, part of them by minimal access. We provide an overview of the rationale for, the technical aspects of, and (if available) the outcomes of the most common minimally invasive prenatal therapies. These include intrauterine transfusion, fetal cardiac procedures, interventions for lower urinary tract obstruction, thoracic and pulmonary pathology, fetoscopic laser of placental vessels for twin-to-twin transfusion syndrome, and selective reduction in complicated monochorionic twin pregnancies.
Introduction
Fetal therapy became a clinical reality after the introduction of high-resolution ultrasound. It has been boosted by the implementation of systematic pregnancy screening programmes that have led to the early detection and better understanding of the natural history of fetal anomalies. Only a limited number of fetal anomalies are progressive in nature, so that, when left untreated, they lead to fetal and neonatal demise, or severe morbidity. Prenatal intervention aims at halting or reversing the progression of these diseases and improving the postnatal outcome. Where some fetal conditions such as open neural tube defects, large solid lung masses or sacrococcygeal teratomas are still treated through maternal laparotomy and hysterotomy, we have witnessed an increased introduction of minimally invasive fetal therapeutic procedures over the past 2 decades. In this chapter, we present an overview of the rationale for, the technical aspects of, and outcomes of the most common minimally invasive prenatal therapeutic procedures. These procedures are reserved for well-selected fetuses at the most severe end of the spectrum, and are only offered in a limited number of centres. Optimal patient selection, expert counselling, rigorous clinical protocols, and experienced multidisciplinary teams are strict criteria for fetal therapy programmes.
Intrauterine transfusion
Intrauterine transfusions are carried out for severe fetal anaemia and thrombocytopaenia. Technical aspects of the procedure are identical for both conditions. Given the relative rarity of fetal thrombocytopaenia, we refer for issues on current diagnosis and management to an excellent recent review and restrict our discussion to fetal anaemia. The incidence of rhesus D (RhD) hemolytic disease haemolytic disease has decreased dramatically since the routine administration of anti-D immunoglobulins. As a consequence, other causes of fetal anaemia, such as feto–maternal haemorrhage twin anaemia polycythaemia sequence, parvovirus B19 infections and haemolytic disease, resulting from other than RhD red blood cell antigens, have gained in relative importance. The severity of anaemia in a fetus at risk can be assessed by non-invasive measurement of peak systolic velocity in the middle cerebral artery. Severe anaemia, mandating antenatal intervention, is present when the peak systolic velocity is consistently over 1.5 multiples of the median and shows an increasing trend.
Intrauterine transfusions are usually carried out under local or locoregional anaesthesia, in an outpatient setting ( Fig. 1 ). The umbilical vein is punctured under ultrasound guidance with a 20 or 22 gauge amniocentesis needle, typically either in its intrahepatic portion or at the placental cord root. A free cord loop can also be used, but it is more fragile, and the insertion point may bleed on withdrawal of the needle. Intra-peritoneal transfusions are only rarely carried out, as the absorption of red blood cells from the peritoneal cavity is slow and unpredictable. At the start of the transfusion, a blood sample is drawn for haemoglobin and haematocrit measurement. Pancuronium can be administered to immobilise the fetus. O Rhesus negative blood with a haematocrit of 80%, which has been screened for the most common infectious diseases and irradiated to prevent graft versus host reaction, is transfused. The amount to be transfused can be calculated from the estimated feto–placental blood volume, the initial fetal haematocrit, and the target haematocrit, which is usually around 45%. The final haematocrit is verified at the end of the procedure. In case of severe anemia, a single transfusion might not allow the target to be achieved without overfilling the fetus, and multiple transfusions with a few days interval may be necessary. After a successful transfusion, the peak systolic velocity in the middle cerebral artery decreases immediately. In case of persistent haemolysis, the fetal haematocrit drops around 1% per day. Therefore, repeated transfusions at 2–3 weeks interval are required in Rhesus allo-immunisation. The last transfusion is usually given around 34 weeks of gestation to allow for a delivery around 36–37 weeks.
In experienced hands, intrauterine transfusions are successful in over 97% of cases. The rate of severe complications is 3% per procedure in RhD and the procedure-related fetal loss rate is 1.6%. The latter increases to 10% in transfusions carried out before 22 weeks of gestation. Overall survival after intrauterine transfusions for RhD is 75%, when the fetus is hydropic at the first transfusion and over 90% in non-hydropic fetuses. A large follow-up study on neurologic outcomes after intrauterine transfusions for RhD shows a cerebral palsy rate of 2.1% and an overall risk for developmental delay of 4.8%. Neurologic impairment is more likely if hydrops was present, which stresses the importance of close follow up in pregnancies with rhesus immunisation.
Intrauterine transfusion
Intrauterine transfusions are carried out for severe fetal anaemia and thrombocytopaenia. Technical aspects of the procedure are identical for both conditions. Given the relative rarity of fetal thrombocytopaenia, we refer for issues on current diagnosis and management to an excellent recent review and restrict our discussion to fetal anaemia. The incidence of rhesus D (RhD) hemolytic disease haemolytic disease has decreased dramatically since the routine administration of anti-D immunoglobulins. As a consequence, other causes of fetal anaemia, such as feto–maternal haemorrhage twin anaemia polycythaemia sequence, parvovirus B19 infections and haemolytic disease, resulting from other than RhD red blood cell antigens, have gained in relative importance. The severity of anaemia in a fetus at risk can be assessed by non-invasive measurement of peak systolic velocity in the middle cerebral artery. Severe anaemia, mandating antenatal intervention, is present when the peak systolic velocity is consistently over 1.5 multiples of the median and shows an increasing trend.
Intrauterine transfusions are usually carried out under local or locoregional anaesthesia, in an outpatient setting ( Fig. 1 ). The umbilical vein is punctured under ultrasound guidance with a 20 or 22 gauge amniocentesis needle, typically either in its intrahepatic portion or at the placental cord root. A free cord loop can also be used, but it is more fragile, and the insertion point may bleed on withdrawal of the needle. Intra-peritoneal transfusions are only rarely carried out, as the absorption of red blood cells from the peritoneal cavity is slow and unpredictable. At the start of the transfusion, a blood sample is drawn for haemoglobin and haematocrit measurement. Pancuronium can be administered to immobilise the fetus. O Rhesus negative blood with a haematocrit of 80%, which has been screened for the most common infectious diseases and irradiated to prevent graft versus host reaction, is transfused. The amount to be transfused can be calculated from the estimated feto–placental blood volume, the initial fetal haematocrit, and the target haematocrit, which is usually around 45%. The final haematocrit is verified at the end of the procedure. In case of severe anemia, a single transfusion might not allow the target to be achieved without overfilling the fetus, and multiple transfusions with a few days interval may be necessary. After a successful transfusion, the peak systolic velocity in the middle cerebral artery decreases immediately. In case of persistent haemolysis, the fetal haematocrit drops around 1% per day. Therefore, repeated transfusions at 2–3 weeks interval are required in Rhesus allo-immunisation. The last transfusion is usually given around 34 weeks of gestation to allow for a delivery around 36–37 weeks.
In experienced hands, intrauterine transfusions are successful in over 97% of cases. The rate of severe complications is 3% per procedure in RhD and the procedure-related fetal loss rate is 1.6%. The latter increases to 10% in transfusions carried out before 22 weeks of gestation. Overall survival after intrauterine transfusions for RhD is 75%, when the fetus is hydropic at the first transfusion and over 90% in non-hydropic fetuses. A large follow-up study on neurologic outcomes after intrauterine transfusions for RhD shows a cerebral palsy rate of 2.1% and an overall risk for developmental delay of 4.8%. Neurologic impairment is more likely if hydrops was present, which stresses the importance of close follow up in pregnancies with rhesus immunisation.
Fetal cardiac procedures
Despite improvements in neonatal (surgical) care and the development of dedicated follow-up programmes for infants with congenital heart disease, the outcome of fetuses with hypoplastic left heart syndrome remains poor. Postnatal surgery, which results in a far from optimal single ventricle Fontan-type circulation has a considerable mortality rate, leading to a total survival of less than 65%. Moreover, one-half of long-term survivors have poor neurodevelopmental outcome, which may in part have an antenatal origin. Indeed, a preferential return of oxygenated blood towards the right ventricle and lower body rather than towards the left ventricle and the brain may lead to suboptimal brain oxygenation in utero .
Hypoplastic heart can already be present in early pregnancy, but about 5% develop in the second trimester and are progressive ( Fig. 2 ). In these cases, which are often due to outlet valve obstruction, fetal balloon valvuloplasties may allow for intrauterine ventricular recovery and growth, and increase the chance of a postnatal biventricular repair. Despite initial disappointing attempts with prenatal balloon valvuloplasty for aortic valve stenosis, two groups pursued the idea of prenatal intervention for severe aortic stenosis. They defined criteria for intervention ( Table 1 ) and optimised techniques for needle-based access to the fetal heart. The technical aspects of the procedure are relatively standard and comparable between both groups. The procedure may require general maternal anaesthesia for optimal fetal positioning, but most are done under local or locoregional anaesthesia. Initially, the laparotomy rate in the study by McElhinney et al. was as high as 27%, but this dropped to 10% in their last 50 cases. For hypoplastic left heart syndrome, an 18 or 19 gauge needle is inserted into the fetal left ventricle at the level of the apex and in alignment with the left ventricular outflow tract. A guide wire and a catheter with a coronary dilatation balloon are advanced through the aortic valve, which is dilated to 120% of the valve annulus ( Fig. 2 ). Technical success, defined as successful inflation and dilatation, is achieved in 70% of cases. This can often be documented by the appearance of aortic regurgitation. Peri-operative complications are common. These include bradycardia necessitating fetal resuscitation (17–38%), haemopericardium (13%), ventricular thrombosis (15–20%) as well as fetal demise (8–13%). Significant left ventricular growth may be observed in utero , yet only 33–67% of the technically successful procedures end up in a postnatal biventricular repair. Given the recent introduction of this procedure in fetal medicine, long-term outcomes are not yet available.
| Arzt et al. ( n = 24) | McElhinney et al. ( n = 70) | |
|---|---|---|
| Inclusion criteria | Left ventricle length z-score > −3 | Left ventricle length z-score > −2 |
| Aortic arch reversed flow | Aortic arch reversed flow or 2 of the following | |
| Left to right shunt over foramen ovale | Left to right shunt over the foramen ovale | |
| Endocardial fibroelastosis | Monophasic inflow left ventricle | |
| Bidirectional flow pulmonary veins | ||
| Mitral valve z-score > −3 | ||
| Depressed left ventricle function but conserved antegrade or retrograde flow | ||
| Gestational age at valvuloplasty in weeks (range) | 26.6 (21.4–32.7) | 23.2 (20–31) |
| Mini-laparotomy rate (%) | 0 | 27 |
| Technical success rate a (%) | 70 | 74 |
| Postnatal biventricular repair rate (%) | 67 | 33 |
a Technical success defined as successful inflation of balloon.
Similar percutaneous cardiac balloon procedures have been proposed for hypoplastic left heart syndrome with a highly restrictive foramen ovale, and pulmonary atresia with intact ventricular septum. Data on these rarer cases are scarce, and virtually nothing has been published on longer-term outcomes.
Interventions for lower urinary tract obstruction
Fetal lower urinary tract obstruction (LUTO) occurs in about 2–3 per 10,000 pregnancies. Most of these cases are caused by isolated posterior urethral valves or urethral atresia, but associated structural or chromosomal anomalies are present in 25%. Untreated, the perinatal mortality of isolated LUTO is high, primarily resulting from pulmonary hypoplasia secondary to the oligohydramnion, but also from severe renal failure. Survivors have a high incidence of chronic renal failure, and up to 17% reach end-stage renal disease within 10 years. Vesico-amniotic shunting has been proposed to improve the outcome of these fetuses. Evidence-based criteria on which fetuses should be considered eligible for in-utero treatment are lacking. Most centres require the presence of severe oligohydramnios and acceptable fetal renal function, based on a variety of diagnostic criteria.
Vesico-amniotic shunting is a percutaneous procedure using a purpose-designed cannula and trocar with echogenic tip. The instrument is advanced through the fetal abdominal wall into the bladder to deploy a catheter allowing bladder drainage into the amniotic cavity. Feasibility is highly dependent on fetal position. In case of complete anhydramnios, amnioinfusion may be required to allow for deploying the external loop of the shunt inside the uterus. The most frequent complication of vesico-amniotic shunting is preterm rupture of the membranes, leading to a mean gestational age at delivery of 34–35 weeks. Shunt dislodgment, which occurs in 34%, and obstruction, may mandate a repeat procedure. More severe complications, such as iatrogenic gastroschisis and chorioamnionitis, leading to maternal sepsis, are rare. Some evidence shows that vesico-amniotic shunting improves neonatal survival, especially in fetuses with poor renal function. Overall survival ranges between 50 and 90%, and one in three survivors has end-stage renal disease at a mean follow up of 5 years. Survivors may also have other respiratory and growth problems, but their self-reported quality of life falls within the normal range.
More recently, intrauterine fetal cystoscopy has been proposed. The procedure, which can be carried out as early as 16 weeks of gestation, allows a more robust diagnosis. In case of urethral valves, definitive treatment by laser fulguration may be attempted. A recent systematic review of four fetal cystoscopy studies showed that the initial diagnosis of posterior urethral valves changed in 32% (six out of 19) of cases, typically towards urethral atresia. For postnatal survival, cystoscopic ablation of the valves was superior to expectant management yet comparable to shunting. Long-term follow-up data for these cohorts on urinary or pelvic floor function are not available yet.
Interventions for fetal thoracic pathology
Pleural effusions
Fetal pleural effusions can either be isolated (primary hydrothorax or chylothorax) or secondary to other conditions, such as diaphragmatic hernia, bronchopulmonary sequestration, cardiac anomalies, fetal infections, metabolic, chromosomal or syndromal disorders. Mild-to-moderate primary hydrothorax, which does not lead to functional cardiac impairment, mediastinal shift or severe lung compression, has good outcomes when managed conservatively. Expectantly managed fetuses with hydrops and severe pulmonary hypoplasia, however, only survive in 25%.
Drainage of the effusion should theoretically alleviate intra-thoracic compression hence yield a better outcome. Such drainage can be achieved either by repeated thoracocentesis (the effusion usually reaccumulates within 2 days), pleuroamniotic shunting or pleurodesis. A recent systematic review showed that outcomes of thoracocentesis and shunting were similar, yet shunting is often preferred when more distant from term. The technical aspects of pleuroamniotic shunting are similar to vesico-amniotic shunting, and the same shunts are used. The lateral or posterior chest wall of the fetus are preferred for shunt insertion to avoid lesions to mammary gland and nipple ( Fig. 3 ). Shunt migration or obstruction leads to the need for repeated procedures in 10–20% of cases. Preterm premature rupture of membranes and preterm delivery are common, with an average gestational age at delivery after pleuroamniotic shunting of 34–35 weeks. Yinon et al. recently summarised a larger case study on shunting for hydrothorax. Neonatal survival was 55% in 206 hydropic fetuses, and 85% in 72 non-hydropic fetuses. Gestational age at birth and duration of shunting are both important predictors of outcome.
In an attempt to avoid thoraco-amniotic shunting or repeated thoracocentesis, fetal pleurodesis can be attempted (e.g. by injecting OK-432 ). OK-432 is an inactivated biological response modifier derived from Streptococcus pyogenes , and causes an inflammatory response through irritation of the pleura. Survival rate in the only two published studies is 75% (18 and 24%, respectively) in non-hydropic cases, and 21% (five out of 24) in hydropic fetuses. These results are not different from the natural history of the disease and are clearly inferior to what is seen with thoraco-amniotic shunting.
Pulmonary parenchymal lesions
Congenital cystic adenomatoid malformation of the lung (CCAM) arises from an overgrowth of the terminal respiratory bronchioles. In bronchopulmonary sequestration, non-functional pulmonary parenchyma is separated from the normal lung and receives its blood supply from the systemic circulation. CCAMs are classified as microcystic, macrocystic or mixed based on their appearance on ultrasound. Most commonly, CCAMs are small, and about 50% regress in the 3rd trimester of pregnancy. Limited lesions do not compromise cardiac function or lung development, and have favourable outcomes with conservative prenatal management. Some can even be managed conservatively postnatally.
More rare are the masses that grow very large and lead to pulmonary hypoplasia or hydrops, the latter being almost invariably fatal if left untreated. In large macrocystic CCAMs, therapeutic size reduction can be obtained by thoraco-amniotic shunting of the larger cyst(s). Cavoretto et al. summarised data on 24 non-hydropic fetuses who underwent thoraco-amniotic shunting at 27 weeks for CCAM causing severe mediastinal shift. Mean gestational age at delivery was 37 weeks and postnatal survival was 87.5%. For 50 hydropic cases, undergoing either thoracocentesis or thoraco-amniotic, survival was only 66%. Large microcystic CCAMs are considered not to be amenable for thoraco-amniotic shunting as mainly solid. Alternatively, they have been ablated using minimally invasive techniques (interstitial laser [ n = 5], radiofrequency ablation [ n = 1] or cyano-acrylate injection into the mass [ n = 1]). Of the six fetuses with reported neonatal outcomes, only two survived. Open fetal surgical lobectomy has a survival rate of 50%, yet at the cost of higher maternal morbidity. More recently, the disappearance of hydrops and even regression of the microcystic mass has been reported after maternal administration of glucocorticoids (betamethasone 12 mg, twice). Current reports on fetuses with microcystic CCAM and associated effusions or hydrops are presented in Table 2 . The survival of 86% ( n = 18 out of 21 fetuses) compares favourably to what is reported for other interventions.
| N | Hydrops (n) | Isolated effusion (n) | Mean gestational age at steroids in weeks (range) | Resolution of effusion (n) | Survival (n) | |
|---|---|---|---|---|---|---|
| Peranteau et al., 2007 | 5 | 2 | Two ascites, one scalp oedema | 22.9 (19.4–25.60) | 4 | 5 a |
| Morris et al., 2009 | 5 | 4 | One ascites | 25.4 (20.7–30.7) | 4 | 4 |
| Curran et al., 2009 | 11 | 9 | Two ascites | 24.3 (22.4–26.6) | 9 | 9 |
| Total | 21 | 15 (71.4%) | 6 (28.6%) | 24.3 (19.4–30.7) | 17 (81.0%) | 18 (85.7%) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
