OR WAIT 15 SECS
Open fetal surgery is improving outcomes in fetuses with a variety of fatal or severely disabling disorders. Developments in minimally invasive procedures hold great promise for in utero treatment, as do studies in gene therapy.
|Jump to:||Choose article section... Who is referred for fetal intervention? Disorders amenable to fetal intervention When surgery can help When to use fetal medical therapy Prenatal gene therapy Putting it together|
Open fetal surgery is improving outcomes in fetuses with a variety of fatal or severely disabling disorders. Developments in minimally invasive procedures hold great promise for in utero treatment, as do studies in gene therapy.
Until recently, a disorder diagnosed prenatally could be managed only after birth. Some fetuses are too ill to survive to infancy, however, or they may suffer irreversible organ damage if disorders are left untreated throughout gestation. Advances in prenatal testing and imaging techniques now permit diagnosis of many birth defects and fatal developmental disorders as early as the first trimester of pregnancy. Continued animal research, and careful attention to its relevance for the human organism, has enriched our understanding of the natural history of these disorders and led to development of fetal surgical techniques to improve survival and postnatal function. Fetal surgery allows us to alter the natural history of these disorders in ways not possible in the postnatal period.
Improvements in our understanding of fetal pathology, technical abilities, and postoperative care have ushered in the modern period of fetal intervention. Procedures generally are used on human fetuses only after their benefits have been clearly documented in animal models of similar disorders. Until recently, intervention was limited to fetuses with dismal prospects for postnatal survival. Now, however, fetal therapy is applied in limited circumstances to at least one disorder that, while not fatal, has a high morbiditymyelomeningocele. Because prenatal surgery is associated with considerable risk of preterm labor, however, its full potential has not yet been realized. Continued research and improved prevention of preterm labor should expand the spectrum of surgically treatable fetal disorders. Fetal medical therapy alone is beneficial in some disorders.
Congenital disorders that are amenable to in utero therapy typically are diagnosed in the second trimester during the maternal evaluation. Table 1 lists the disorders most appropriate for referral to a center performing fetal intervention. Pregnant women should be referred as soon as possible after diagnosis, since prompt intervention often is key. At our center, a thorough evaluation, including sonography, ultrafast fetal magnetic resonance imaging, and genetic counseling provide further information to the team of fetal surgeons, obstetricians, neonatologists, nurses, and social workers who assess each patient. When the evaluation is complete, we meet with the family to present the team's findings, discuss options, and formulate a plan of care. We admit patients who elect open fetal intervention for tocolytic therapy (to inhibit uterine contractions and premature labor); they remain hospitalized for five to seven days after the procedure. An outpatient tocolytic regimen consisting of subcutaneous terbutaline or oral nifedipine typically is continued until delivery by cesarean section near term. Fetoscopic treatments usually have shorter hospital courses and sometimes do not require tocolysis.
Congenital diaphragmatic hernia
Congenital cystic adenomatoid malformation
Twin-twin transfusion syndrome
Twin-reversed arterial perfusion sequence
Amniotic band syndrome
Congenital high-airway obstruction
Congenital adrenal hyperplasia
Multiple carboxylase deficiency
Severe combined immunodeficiency syndrome
Disorders now considered appropriate for fetal surgery include congenital diaphragmatic hernia, mass lesions of the lung, other life-threatening mass lesions, myelomeningocele, obstructive uropathy, and twin disorders.
Congenital diaphragmatic hernia. In utero surgery was initially directed at fetal lung lesions. A series of animal experiments beginning in the 1980s was followed by surgery on midgestational fetuses with less than a 10% chance of postnatal survival.1,2 Congenital diaphragmatic hernia, believed to result from failure of the pleuroperitoneal membrane to fuse, occurs in about one in 2,000 live births. Despite improvements in postnatal management, mortality for congenital diaphragmatic hernia remains nearly 60%.3 Lowest survival rates are for fetuses who have large, left-sided lesions with herniation of the left lobe of the liver into the thorax. Untreated, these fetuses develop mediastinal shift and severe bilateral pulmonary hypoplasia, making postnatal ventilation extraordinarily difficult.
Initial fetal surgery focused on repair of the diaphragmatic defect. Tracheal occlusion has been more successful, however. It induces in utero pulmonary growth through retention of lung fluid and expansion of lung volume.4 The procedure, which follows the evaluation outlined in Figure 1, consists of placing occlusive hemoclips across the trachea at 25 to 26 weeks' gestation. Most often, this maneuver is accomplished via an open maternal laparotomy and hysterotomy, exposure of the fetus's neck, and circumferential dissection of the trachea, though endoscopic techniques are being introduced.5 Definitive repair of the diaphragmatic defect is deferred until after the infant is born.
A fetus with tracheal occlusion is delivered using the EXIT (ex utero intrapartum treatment) procedure, in which the fetus is maintained on placental circulation until the tracheal clip is removed and an airway established. EXIT procedures may also be used to establish an airway in infants with massive cervical teratoma, cystic hygroma, and congenital high-airway obstruction syndrome. Once the baby has a secure airway and is ventilated, the umbilical cord is divided.
Mass lesions of the lung. Such lesions as congenital cystic adenomatoid malformation, bronchopulmonary sequestration, and fetal hydrothorax can lead to pulmonary hypoplasia, mediastinal shift, and impaired cardiac function. Untreated, these lesions can progress to fetal cardiovascular failure (hydrops), which is almost always fatal.6 Fetal hydrops is evidenced by ascites, pericardial effusion, and scalp or skin edema.
Congenital cystic adenomatoid malformation is a benign cystic mass of the lung formed by overgrowth of the terminal respiratory bronchioles. While 20% of lesions become smaller during gestation, others continue to enlarge and require intervention if hydrops develops.7 Large congenital cystic adenomatoid malformations with a unilocular cyst and associated hydrops can be treated by percutaneous placement of a thoracoamniotic shunt. After the shunt decompresses the cyst, resection can be delayed until the postnatal period. Predominantly solid congenital cystic adenomatoid malformations with fetal hydrops require a timely fetal thoracotomy and resection of the affected lobe. In the future, less invasive proceduresablation of thoracic masses by laser, cryotherapy, or radio-frequency ablationmay further improve outcomes by lowering the risk of preterm labor.
A bronchopulmonary sequestration is a nonfunctional segment of lung with a systemic vascular supply that may or may not have a connection to the tracheobronchial tree. Like some congenital cystic adenomatoid malformations, most bronchopulmonary sequestrations spontaneously regress. Those that enlarge can cause hydrops: Pleural fluid accumulates, leading to a tension hydrothorax. The fluid can be evacuated by a thoracoamniotic shunt.
Massive fetal hydrothorax may also develop secondary to chylothorax and be decompressed by percutaneous shunts. The shunt can be removed after birth, since the chylothorax is typically transient and usually requires no further intervention.8
Additional mass lesions. Sacrococcygeal teratomas, which arise from the coccyx, are the most common tumors of the newborn. Infants who develop sacrococcygeal teratomas in the neonatal period have a good prognosis, but mortality for those with a prenatal diagnosis is 30% to 50%.9,10 Tumor mass can lead to excess amniotic fluid (polyhydramnios) and premature birth. Lethal hemorrhage from these highly vascular tumors may result from rupture either spontaneously or at the time of delivery. In addition, sacrococcygeal teratomas may become massive, diverting blood flow from its usual course (vascular steal), leading to high-output cardiac failure, hydrops, and death in utero. The object of fetal surgery is to eliminate or reduce vascular steal and normalize fetal hemodynamics. Following maternal hysterotomy, the fetal buttocks are exposed and the anorectal sphincter identified; then the tumor is wrapped at its base with a tourniquet. The mass is then separated (down to its base) from the surrounding normal tissues by a finger fracture technique to permit ligation of the vascular pedicle. Close monitoring of fetal hemodynamics is required. Minimally invasive techniques for tumor ablation or occlusion of the tumor's blood supply are being developed.
While most pericardial teratomas can be treated after birth, these mass lesions occasionally induce fetal hydrops with pericardial effusion. Aspiration may lessen the effects of the tumor, though open surgical resection remains an option if needed to maintain fetal viability. 11
Myelomeningocele, the most common and severe form of spina bifida, refers to protrusion of the meninges and spinal cord through open vertebral arches. Although great strides have been made in preventing this disorder through folic acid supplementation, spina bifida still affects nearly one in 2,000 infants born alive, and we know little about how to treat this devastating malformation.12 In addition to being permanently paralyzed, individuals with myelomeningocele often are limited by mental retardation, bowel and bladder dysfunction, and orthopedic disabilities. Significant morbidity and mortality comes from hydrocephalus, the Arnold-Chiari II malformation, and spinal cord tethering at the site of surgical repair. Until recently, the only way to manage prenatally diagnosed myelomeningocele was either to perform an abortion or to close the spinal canal surgically at birth and provide lifelong supportive care.
Myelomeningocele became the first nonlethal disorder to which fetal surgery was applied, following careful animal experimentation in fetal lambs. Because of the significant risks of prenatal intervention, at our institution we offer fetal surgery only to a mother whose fetus has an Arnold-Chiari malformation, mild or moderate ventriculomegaly, normal leg movements, no apparent clubbing of the feet, normal karyotype, and no accompanying lethal anomalies, as shown in the flowchart (Figure 2). By limiting interventions to those with the Arnold Chiari malformation, we target infants most likely to suffer from hydrocephalus or life-threatening brainstem symptoms and to require frequent postnatal surgical intervention. The spinal defect is closed between 21 and 25 weeks' gestation, using techniques similar to those of postnatal repair.
Reports from two institutions indicate that in selected patients fetal repair may improve neurologic function and reduce morbidity and mortality from hydrocephalus and the Arnold Chiari II malformation by reversing hindbrain herniation.13,14 While the infants who were studied require long-term follow-up, the ramifications of these outcomes are significant. Ascent of the hindbrain caused by normalized cerebral spinal fluid hydrodynamics may reduce hydrocephalus, making it unnecessary to place ventricular shunts. The location of the hindbrain in a more normal anatomic position should reduce the symptomatic sequelae of the Arnold-Chiari malformation and need for subsequent surgery. Reduced incidence of club feet and other orthopedic anomalies should limit the need for surgical intervention, enhance the possibility of walking, and promote independence. We await assessment of the impact of prenatal intervention on bowel and bladder continence, sexual function, ambulation, and mental capacity.
Fetal obstructive uropathies in either the upper or the lower urinary tract, leading to dilation, occur in about one in 100 pregnancies.1 About half the cases of upper-tract disease are attributed to ureteropelvic junction obstruction, usually resulting from abnormal wall thickening. Other causes of upper tract disease are vesicoureteral reflux (33%), vesicoureteral junction obstruction (9% to 14%), and posterior urethral valves (2% to 9%).15
In male fetuses, lower tract obstruction generally results from developmental abnormalities of the urethra, most commonly posterior urethral valves or urethral atresia. In female fetuses, lower tract obstruction is generally associated with developmental anomalies of the cloaca or is a component of syndromic abnormalities. In addition, several genetic disorders (such as megacystis-microcolon syndrome) and chromosomal aneuploidies (such as trisomy 21 and 18) can be associated with lower urinary tract obstructions. Unilateral hydronephrosis is a typical prenatal sonographic feature of upper tract disease; distended fetal bladder, bilateral hydronephrosis, and decreased amniotic fluid volume characterize significant lower tract obstruction.
In utero therapy is intended to prevent or reverse some of the prenatal anatomic and pathologic changes that can lead to complications or death after birth. Obstruction can lead to bladder wall hypertrophy and hyperplasia, loss of bladder compliance, and progressive compression of the renal parenchyma with dysplasia and dysfunction. The primary cause of death in prenatally diagnosed obstructive uropathy is not renal disease, however, but pulmonary hypoplasia.16 When urine fails to fill the amniotic space, progressive oligohydramnios ensues, resulting in poor pulmonary growth, limb deformities, abnormal abdominal- wall development, and Potter facies.
To be a candidate for prenatal intervention for obstructive uropathy, as shown in Figure 3, the fetus should be male and have a lower tract obstruction but no other anomalies that would adversely affect the prognosis. In addition, an ultrasound should reveal recent onset of oligo/anhydramnios or reduced amniotic fluid volume. In female fetuses the obstruction is likely part of a cloacal malformation and intervention is unlikely to improve outcome. Fetuses selected for possible surgical intervention are further evaluated by serial analysis of urine obtained by fine-needle bladder drainage (vesicocentesis). A minimum of three bladder drainages are performed at 48- to 72-hour intervals, and sodium, chloride, osmolality, calcium, b-2 microglobulin, and total protein are measured at each interval. A pattern of decreasing hypertonicity and values that gradually fall below established thresholds indicates the potential for renal salvage and identifies fetuses most likely to benefit from in utero intervention.17 With severe renal damage (dysplasia), urine hypertonicity remains unchanged on serial sampling.
Diversion therapy with a vesicoamniotic shunt is the prenatal intervention of choice for relieving the obstruction, reversing the oligohydramnios, preventing further renal damage, and improving outcome. Using ultrasound guidance, the surgeon generally places the shunt percutaneously from the lower bladder to the amniotic space. This procedure is effective, but close follow-up of patients is necessary because the shunts tend to dislodge.
Posterior urethral valve obstruction may also be treated with cystoscopic ablation. In this procedure, a fetoscope is passed into the distended bladder and into the proximal urethra to identify the source of the obstruction. If the diagnosis of a valve obstruction (as opposed to an atresia) can be confirmed, disruption with a laser or other form of energy is technically possible.18 This approach has been used experimentally; though results have been mixed, it holds great promise for relieving the obstruction without using a catheter. Compared with catheterization, the laser approach also may allow for more normal physiologic bladder storage and voiding cycling, which may play an important role in long-term bladder function. 19
Twin disorders. Fetoscopic procedures have also proved useful with two complications of twin pregnancies, twin-twin transfusion syndrome and twin-reversed arterial perfusion sequence. In twin-twin transfusion syndrome, abnormal placental vascular connections between monozygotic twins in a common chorionic sac can lead to severe growth discordance, hydrops, and fetal demise. When the condition is diagnosed before 26 weeks' gestation, mortality is 60% to 100% for both twins, and survivors have a high incidence of neurologic injury.20 Most cases can be treated with serial amnioreduction with or without a microseptostomy between the amniotic membranes, but more severe cases require an invasive procedure.21,22 In some cases, placental vessel ablation with a Nd:YAG laser fiber successfully interrupts the abnormal vasculature, reversing the pathophysiology and salvaging both fetuses. Selective feticide by cord ligation with preservation of the less compromised twin is the only option in other pregnancies, however.
In twin-reversed arterial perfusion sequence, one twin has no heart and is perfused by a normal "pump" twin. Sometimes the pump twin develops cardiomyopathy and cardiac failure and can be salvaged only by occlusion of the cord to the twin without a heart. This may be accomplished by either bipolar cauterization or ligatures.
Fetoscopic cutting and cauterization has also been used in severe cases of amniotic band syndrome.23 In this uncommon disorder, constricting bands of amnion wrap around the fetus and may cause vascular and lymphatic occlusion. Tight constrictions can result in amputation of the affected digit or extremity. Fetoscopic release may be warranted in the second trimester, if the bands threaten life or limb.
Several inborn errors of metabolism and other conditions are appropriate for fetal medical therapy.
Congenital adrenal hyperplasia has presentations ranging from ambiguous genitalia in females or virilization in males to life-threatening adrenal insufficiency. The most common enzyme defect is in 21-hydroxylase, resulting in accumulation of 17-hydroxyprogesterone in the amniotic fluid. The purpose of prenatal therapy is to prevent masculinization of female fetuses by adrenal suppression, which is accomplished by giving the mother dexamethasone. Since masculinization can begin early in the first trimester, mothers at risk because they have given birth to a child with this disorder should begin steroid therapy before the eighth week of the pregnancy.24 These mothers can then undergo chorionic villus sampling and measurement of amniotic fluid molecular markers at nine to 12 weeks' gestation to determine if the fetus is female and affected, in which case steroids are continued. If the fetus is male or unaffected, the treatments can be discontinued.
Multiple carboxylase deficiency results from decreased activity of biotin-dependent mitochondrial enzymes. Infants or young children with this disorder develop metabolic acidosis and dermatitis that are responsive to biotin supplementation. In mothers known to have borne children with multiple carboxylase deficiency, prenatal biotin supplementation can normalize serum and urinary organic profiles and improve neonatal well-being.19
Methylmalonic acidemia results from an inability to convert vitamin B12 to its active form. Prenatal diagnosis can be established by measuring maternal urinary methylmalonic acid excretion. Maternal therapy with intramuscular cyanocobalamin has been tried, but the overall benefit to the fetus is not clearly established.25
Ebstein's anomaly. Hydrops caused by fetal congestive heart failure in connection with Ebstein's anomaly has been successfully treated with medical therapy. In fetuses with severely impaired cardiac function with tricuspid regurgitation, maternal therapy with digitalis has resulted in a gradual improvement in fetal hemodynamics and a decrease in fetal cardiomegaly.26
Some genetic disorders cause permanent damage to the fetus in utero, making postnatal therapies inadequate. In addition, the fetus has a unique period of immune tolerance early in gestation when transplantation, with hematopoietic stem cells for example, entails little risk of rejection and immunosuppresion is not needed. Thus, prenatal gene therapy, while still experimental, has tremendous potential for treating a variety of disorders. (For details, see "Pediatrics and the Human Genome Project" in the May issue of Contemporary Pediatrics.)
Many significant challenges must be overcome before effective prenatal gene therapy can be performed. The disorder must be diagnosed early in gestation, the defective gene must be identifiable, and a functional form of the gene that can produce the gene product must be available. Next, the functional gene construct must be delivered to appropriate target cells and transferred into these cells, and the cells must produce enough gene product to replace the defective gene product. One strategy for gene therapy is to deliver the gene through viral vectors, and many potential viral vectors have been investigated for postnatal therapies. For this therapy to be effective over the lifetime of the patient, the functional gene must be placed permanently into the target cell genome. The long-term safety and efficacy of gene transfer has yet to be established.
In a different approach, hematopoietic stem cells that carry a functional form of the gene are transplanted into the fetus. If necessary, mononuclear cells can be obtained from the fetus and exposed ex vivo to a retrovirus containing the gene construct. The cells then are reinfused into the fetus, where they migrate through the body and differentiate into the appropriate cell type in the target organ. Low levels of engraftment, resulting in minimal production of the necessary gene product, significantly limit this strategy, however. In addition, manipulation of the fetus before onset of immunocompetence (early in the second trimester) can lead to death.
Three categories of genetic disorders are the focus of most research into prenatal gene therapy: inborn errors of metabolism, such as lysosomal storage diseases; immunodeficiencies, such as severe combined immunodeficiency syndrome; and hemoglobinopathies, such as thalassemias and sickle cell anemia. Several of the lysosomal storage diseases produce ongoing neurologic damage throughout gestation, and this damage is irreversible even with postnatal treatment. Unfortunately, gene strategies for targeting fetal brain tissue are still largely experimental and not ready for human application, though results so far are promising. X-linked, recessive, severe combined immunodeficiency syndrome (one of many forms of severe combined immunodeficiency syndrome) has been successfully treated with in utero transplantation of paternal T-cell-depleted stem cells.27 Weekly injections at 16, 17, and 18 weeks' gestation led to delivery of a baby with X-linked disease but a functional immune system. Injection, early in the second trimester, of bone marrow into fetuses with hemoglobinopathies is now being studied.
Fetal intervention is now being performed to treat many fatal or severely disabling disorders for which postnatal therapy is either too late or inadequate. While open fetal surgery is standard, minimally invasive procedures are being developed and may become increasingly common. In certain disorders, medical treatment alone may improve outcomes. Gene therapy holds great promise for many genetic defects, but clinical application awaits additional research.
1. Hedrick MH, Estes JM, Sullivan KM, et al: Plug the lung until it grows (PLUG): A new method to treat congenital diaphragmatic hernia in utero. J Pediatr Surg 1994;29 (5):612
2. DiFiore JW, Fauza DO, Slavin R, et al: Experimental fetal tracheal ligation reverses the structural and physiological effects of pulmonary hypoplasia in congenital diaphragmatic hernia. J Pediatr Surg 1994;29(2):248
3. Harrison MR, Adzick NS, Estes JM, et al: A prospective study of the outcome for fetuses with diaphragmatic hernia. JAMA 1994;271:382
4. Harrison MR, Adzick NS, Bullard KM, et al: Correction of congenital diaphragmatic hernia in utero VII: A prospective trial. J Pediatr Surg 1997;32(11):1637
5. Harrison MR, Mychaliska GB, Albanese CT, et al: Correction of congenital diaphragmatic hernia in utero IX: Fetuses with poor prognosis (liver herniation and low lung-to-head ratio) can be save by fetoscopic temporary tracheal occlusion. J Pediatr Surg 1998;33(7):1017
6. Adzick NS, Harrison MR, Crombleholme TM, et al: Fetal lung lesions: Management and outcome. Am J Obstet Gynecol 1998;179:884
7. MacGillivray TE, Harrison MR, Goldstein RB, et al: Disappearing fetal lung lesions. J Pediatr Surg 1993;28 (10):1321
8. Longaker MT, Laberge JM, Dansereau J, et al: Primary fetal hydrothorax: Natural history and management. J Pediatr Surg 1989;24(6):573
9. Flake AW, Harrison MR, Adzick NS, et al: Fetal sacrococcygeal teratoma. J Pediatr Surg 1986;21(7): 5631
10. Holterman AX, Filiatrault D, Lallier M, et al: The natural history of sacrococcygeal teratomas diagnosed through routine obstetric sonogram: A single institution experience. J Pediatr Surg 1998;33(10):899
11. Bruch SW, Adzick NS, Reiss R, et al: Prenatal therapy for pericardial teratomas. J Pediatr Surg 1997;32(7):1113
12. Lary JM, Edmonds LD: Prevalence of spina bifida at birth-United States, 1983-1990: A comparison of two surveillance systems. MMWR CDC Surveill Summ 1996; 45:15
13. Sutton LN, Adzick NS, Bilaniuk LT, et al: Improvement in hindbrain herniation demonstrated by serial fetal MRI following fetal surgery for myelomeningocele. JAMA 1999;282:1826
14. Bruner JP, Tulipan M, Paschall RL, et al: Intrauterine repair of myelomeningocele, "hindbrain restoration" and the incidence of shunt-dependent hydrocephalus. JAMA 1999;282:1819
15. Nguyen HT, Kogan BA: Upper urinary tract obstruction: Experimental and clinical aspects. Br J Urol 1998;81:13
16. Nakayama DK, Harrison MR, de Lorimer AA: Prognosis of posterior urethral valves presenting at birth. J Pediatr Surg 1986;21(1):43
17. Johnson MP, Corsi P, Bradfield W, et al: Sequential fetal urine analysis provides greater precision in the evaluation of fetal obstructive uropathy. Am J Obstet Gynecol 1995;173:59
18. Quintero RA, Hume R, Johnson MP, et al: Percutaneous fetal cystoscopy, and endoscopic fulguration of posterior urethral valves. Am J Obstet Gynecol 1995;172:206
19. Close CE, Carr MC, Burns MW, et al: Lower urinary tract changes after early valve ablation in neonates and infants: Is early diversion warrented? J Urol 1997;157:984
20. Bebbington MW, Wittmann BK: Fetal transfustion syndrome: Antenatal factors predicting outcome. Am J Obstet Gynecol 1989;160:913
21. Weiner CP, Ludomirski A: Diagnosis, pathophysiology, and treatment of chronic twin-to-twin transfusion syndrome. Fetal Diagnosis 1994;9:283
22. Ville Y, Hyett J, Hecher K, et al: Preliminary experience with endoscopic laser surgery for severe twin-twin transfusion syndrome. N Engl J Med 1995;32:224
23. Quintero RA, Morales WJ, Phillips J, et al: In utero lysis of amniotic bands. Ultrasound Obstet Gynecol 1997;10:307
24. Speiser PW: Prenatal treatment of congenital adrenal hyperplasia. J Urol 1999;162:534
25. Evans MI, Duquette DA, Rinaldo P, et al: Modulation of B12 dosage and response in fetal treatment of methylmalonic aciduria (MMA); Titration of treatment dose to serum and urine MMA. Fetal Diagn Ther 1997;12:21
26. Hsieh YY, Lee CC, Chang CC, et al: Successful prenatal digoxin therapy for Ebstein's anomaly with hydrops fetalis. A case report. J Reproductive Med 1998;43:710
27. Flake AW, Puck JM, Almieda-Porda G, et al: Successful in utero correction of X-linked recessive severe combined immuno-deficiency (X-SCID): Fetal intraperitoneal transplantation of CD34 enriched paternal bone marrow cells (EPPBMC). N Engl J Med 1996;335:1806
Scott Adzick, Danielle Walsh. Fetal intervention: Where we are, where we're going. Contemporary Pediatrics 2000;6:33.