Introduction and historical perspectives

Open spina bifida (myelomeningocele) is a specific type of neural tube defect (NTD) resulting from a failure of closure of the caudal region of the neural tube early in embryogenesis distinct from anencephaly, which results when the cranial portion of the neural tube fails to close. The diagnosis and management of open spina bifida has evolved dramatically over the past century. The development of shunts for the treatment of hydrocephalus in the 1950s was a revolutionary breakthrough. Improved survival meant that children were living long enough to develop urinary complications; hence, progress was made in the 1960s with the introduction of urinary diversion procedures. Before the 1960s, the favoured approach for postnatal treatment was postponement of surgery until 2 years of life, with the belief that the strongest would survive and would therefore be the best candidates for surgical intervention . With the advancements of the 1950s and 1960s, this Darwinian approach came to be challenged as evidence of benefit of immediate neonatal intervention within 12–48 h of life. The development of Lorber’s criteria in the 1970s for triaging treatment generated a great ethical debate of triage versus immediate intervention for all and selective intervention . The 1980s and 1990s saw improvements in shunt technology and management, clean intermittent catheterisation protocols to reduce renal complications, and the development of spina bifida clinics to manage the medical, physical, and social needs of children, families, and adults with spina bifida. Over the past 25 years, improvements in prenatal diagnosis and advancements have been made in fetal therapy, with the publication of a randomised-controlled trial demonstrating benefit of fetal surgical repair of myelomeningocele .

Although progress in the field of myelomeningocele diagnosis and treatment has revolutionised the medical treatment of open spina bifida, the ethical debate surrounding the postnatal treatment of myelomeningocele that emerged in the 1970s has evolved significantly and is now complicated by issues surrounding prenatal diagnosis, including availability, economic feasibility, and selection for invasive fetal surgery and termination of pregnancy.

Epidemiology

The prevalence of spina bifida is known to have geographic, temporal, race, and ethnic variation . An estimated 3.5 cases per 10,000 live births occurred in the USA between 2004 and 2006, so the actual prevalence of myelomeningocele has been difficult to ascertain because of the availability of prenatal diagnosis and elective termination of pregnancy . Johnson et al. recently conducted a meta-analysis of prenatally diagnosed cases of anencephaly and spina bifida. It included retrospective or prospective cohort studies and studies using birth defects registry data conducted during or after 1990. Significant heterogeneity was found between studies . The investigators identified 15 studies limited to spina bifida and estimated the overall rate of termination of pregnancy to be 63%, with an individual study range of 31–97% . Rates of termination varied based on geographic location with termination of pregnancy being more frequent in Europe than the USA (66% v 50%). Five of the studies provided information on isolated myelomeningocele and reported a 56% termination of pregnancy rate for isolated spina bifida .

Epidemiology

The prevalence of spina bifida is known to have geographic, temporal, race, and ethnic variation . An estimated 3.5 cases per 10,000 live births occurred in the USA between 2004 and 2006, so the actual prevalence of myelomeningocele has been difficult to ascertain because of the availability of prenatal diagnosis and elective termination of pregnancy . Johnson et al. recently conducted a meta-analysis of prenatally diagnosed cases of anencephaly and spina bifida. It included retrospective or prospective cohort studies and studies using birth defects registry data conducted during or after 1990. Significant heterogeneity was found between studies . The investigators identified 15 studies limited to spina bifida and estimated the overall rate of termination of pregnancy to be 63%, with an individual study range of 31–97% . Rates of termination varied based on geographic location with termination of pregnancy being more frequent in Europe than the USA (66% v 50%). Five of the studies provided information on isolated myelomeningocele and reported a 56% termination of pregnancy rate for isolated spina bifida .

Risk factors

Several risk factors have been studied in association with spina bifida; however, only 28% of cases of spina bifida have attributable risk factors, and most of these associations are weak and have not been replicated in subsequent studies . The strongest established risk factor for an NTD is a family history. Several studies have estimated that women with a history of a child with an NTD carry a 3–8% risk of and NTD, including spina bifida, in a subsequent pregnancy, a risk that is consistent with most multifactorial inherited conditions .

Folic acid and genes related to folate metabolism have been identified to play a strong role in neural tube development. Folic acid deficiency has been associated with a two- to eight-fold increased risk of NTDs . Although women with pre-gestational diabetes have been traditionally considered to be at risk for diabetic embryopathy, including an NTD, a report by Correa et al. using the National Birth Defects Prevention study, a population-based cohort study of 10 birth defects surveillance programmes in the USA, found no association between type I or type II diabetes with increased risk of isolated spina bifida . The investigators showed the rarity of isolated congenital defects in women with pre-existing diabetes, and the study was likely underpowered to detect an association between pre-existing diabetes and isolated spina bifida, as there were only two cases of isolated spina bifida among type 1 or type 2 diabetics .

Other factors that have been associated with NTDs include hot tub use, obesity, and exposure to certain medications. Duong et al. found hot tub use of greater than 30 mins to be associated with a 2.91-fold increased risk of anencephaly; however, the risk of spina bifida was not significantly greater than controls . Obesity has been associated with the development of spina bifida. Werler et al. showed a relative risk of 2.8 (95% CI 1.1 to 6.7) for women with a pre-pregnancy body mass index over 32 kg/m .

Recently, Agopian et al. investigated the proportion of neural tube defects attributable to a known risk factor, and found that the adjusted odds ratio for a first- or second-degree family history was greatest at 6.99 (95% CI 3.87 to 12.60) compared with the adjusted odds ratio for obesity of 1.79 (95% 1.51 to 2.93); however, owing to its prevalence, obesity carried the greatest population attributable risk, accounting for 9.89% of cases that could be attributed to any studied risk factor. Folic acid antagonists, such as aminopterin, methotrexate and possibly trimethoprim, have also been associated with increased rates of fetal anomalies, including NTDs . Many anti-epileptic drugs are associated with congenital anomalies; valproic acid and carbamazepine have been shown to have the strongest association with NTDs .

Prevention

The greatest progress in spina bifida management has been in prevention. Results of the first randomised-controlled trial (RCT) evaluating folic acid for the prevention of NTDs in women with a previous history of a pregnancy affected by a NTD was published in 1981 . Although the study was quite small by modern standards, with only 44 compliant women in the treatment group and 51 women in the placebo group, the study showed a significant benefit of 4 mg of folic-acid supplementation in the preconceptional period to prevent NTD recurrence . Ten years later, the landmark Medical Research Council Vitamin Study was released confirming these results , followed by the publication of a second landmark RCT by Czeizel et al. showing benefit of folic acid in the periconceptional period for the prevention of first time NTDs. By 1993, a large public health effort was under way in the USA through the collaboration of the Centers For Disease Control, Health Services Administration, Food and Drug Administration, and National Institutes of Health. This effort involved a three-stage approach for the prevention of NTDs : (1) dietary supplementation of 0.4 mg of folic acid for low-risk women and 4 mg of folic acid for high-risk women of childbearing age; (2) improvements in dietary habits; and (3) fortification of the US food supply.

The third step was completed in 1998 when the USA mandated folic-acid fortification of all flour products . Many countries have introduced mandatory folic-acid fortification, with an increase from 33 countries in 2004 to 54 countries in 2007 . Studies of the USA, Canada and Chile have estimated a reduction in rates of spina bifida from 31–53% after mandatory fortification with folic acid . In 2008, the Centers For Disease Control reported an estimated US cost benefit of $312–425 million dollars annually with folic acid fortification . Despite the dramatic decline in NTDs with folic acid supplementation and fortification, more recent data in the post-fortification era suggest significant racial and ethnic disparities continue to exist in the prevalence of NTDs particularly among Hispanic women .

Several gene variants involved in the folate metabolism pathway have been identified in association with abnormal neural tube closure in human beings . Of recent discovery has been the complicated interaction between maternal and embryonic genes and the role of regulators of neural tube closure that are not involved in folate metabolism . The complex gene–gene and gene–environment interactions remain an active area of research that will continue to elucidate avenues for prevention, particularly in the post-folate fortification era.

Prenatal diagnosis

Prenatal diagnosis of spina bifida was incorporated into routine prenatal care via screening through maternal serum alpha-fetal protein (MSAFP) and the use of ultrasound. Fetal magnetic resonance imaging (MRI) is also being used in co-ordination with ultrasound to improve prognostic prediction and prenatal counselling.

Serum screening

Maternal serum alpha-fetal protein is a sensitive screening tool for NTDs, with a screening sensitivity of about 80–85% when the screen positive cut-off is set at 2.5 multiples of the median . With improving detection rates of NTD by ultrasound, the role of routine MSAFP screening has been challenged, and is useful for early diagnosis in women who may desire the option of termination of pregnancy. In low-risk women, where ultrasound is only used for indications and is not part of routine antenatal care, Dashe et al. showed that the use of MSAFP did not improve the detection of neural tube defects before delivery; however, when an NTD was detected, it was more likely to be diagnosed before 24 weeks (77% v 40%; P < 0.01), and on average the diagnosis occurred 8 weeks sooner than women who did not undergo screening with MSAFP .

Screening by ultrasound

As the availability and use of ultrasound for aneuploidy screening in the first trimester has increased, several investigators have recently examined first-trimester ultrasound findings of open NTDs. Small biparietal diameter (BPD) has been associated with open spina bifida from 11–13 weeks . Bernard et al. evaluated BPD at 11–14 weeks, showing that 50% of cases of open spina bifida have a BPD below the fifth centile. Subsequently, this group showed that the combination of first-trimester BPD below the fifth centile with maternal serum markers MSAFP, and free hCG, improves the detection rate for open spina bifida to 70%, with a fixed false positive rate of 10% . Abnormalities of the posterior fossa in the first trimester have recently become of interest. Intracranial translucency, originally described by Chaoui et al. was reported to be the normal fluid-filled area of the fourth ventricle that is seen in the mid-sagittal view of the fetus during nuchal translucency screening ( Fig. 1 ). In a retrospective study of 191 fetuses, 100% of fetuses with a normal intracranial translucency were normal and open spina bifida could be excluded. In addition, further exploration of the intracranial anatomy in the first trimester has shown an overall reduction in cerebrospinal fluid, with corresponding reduction in the area of the cerebral ventricular system . First-trimester examination of the posterior fossa has been further refined with the use of three-dimensional ultrasound. Cisterna magna width and measurements of the brainstem compared with brainstem-occipital bone distance, have been identified as potential first-trimester screening signs for open spina bifida .

Ultrasound of a normal fetus in the mid-sagittal plane at 12 weeks gestation. The normal intracranial translucency of the fourth ventricle (IT) is present. The thalamus (T), midbrain (M), brainstem (B), medulla oblongata (MO) and choroid plexus of the fourth ventricle (CP) are also noted.

Diagnosis

The routine use of ultrasound for the detection of fetal anomalies has been debated since the publication of the RADIUS trial in 1994 . Ultrasound has consistently been identified as a sensitive tool for the detection of NTDs, and is considered the gold standard for the prenatal diagnosis of NTDs. RADIUS and EUROFETUS showed sensitivities of 88% and 89%, respectively, for the detection of NTDs. Since that time, advancements in ultrasound technology have improved these detection rates . Recent studies have reported detection rates by ultrasound, ranging from 96–100% .

Historically, amniotic fluid and acetylcholinesterase levels were used to confirm or exclude the presence of an open NTD in women with positive MSAFP screening. With advances in ultrasound technology, however, the need for amniocentesis for this purpose has become obsolete .

Sonographic signs

Abnormal development of the neural tube leads to the development of a Chiari II malformation resulting from herniation of the cerebellar vermis, fourth ventricle, and medulla oblongata through the foramen magnum into the upper cervical canal. The classic cerebellar and cranial findings seen on ultrasound associated with the Chiari II malformation were described by Nicolaides et al. in 1986 as the banana sign ( Fig. 2 ) and lemon sign ( Fig. 3 ). The banana sign describes the bowing of the cerebellum as it is pulled deep into the posterior fossa, obliterating the cisterna magna, and the lemon sign describes the scalloping of the frontal bones as a result of this caudad shift of the intracranial contents.

The banana sign. The cerebellum is pulled into the posterior fossa (arrow). Note the cisterna magna is not visible as the space is obliterated by the downward displacement of the cerebellum.

The lemon sign. Note the scalloping of the frontal bones (arrow). Also note the cranial biometry demonstrates microcephaly with the head circumference measuring less than 2.3% and biparietal diameter 2.6%.

The sonographic cerebellar signs associated with open spina bifida have been found to be the most reliable diagnostic signs. Van den Hof et al. identified that 95% of fetuses with an open myelomeningocele having cerebellar abnormalities on ultrasound regardless of gestational age compared with the lemon sign, which was present in 98% of fetuses 24 weeks or less but in only 13% of fetuses over 24 weeks. At 24 weeks or less, the cerebellar abnormality most commonly seen is the banana sign (72%) but, at over 24 weeks, absence of the cerebellum is most common (81%) .

Ventriculomegaly is a finding consistent with the progressive development of hydrocephalus seen in open spina bifida. Despite nearly 100% of neonates born with an open myelomeningocele having evidence of hydrocephalus at the time of birth, only 70% will develop hydrocephalus during fetal life . Although the development of the Chiari II malformation is obstructive in nature predisposing fetuses with open spina bifida to the development of ventriculomegaly, these fetuses often have a head size that is smaller than normal for gestational age. Nicolaides et al. showed that 61% of fetuses with an open myelomeningocele have a biparietal diameter below the fifth centile and 26% have a head circumference below the fifth centile.

Characterisation of the spinal lesion

Ultrasound imaging of the spinal lesion in cases of spina bifida can be quite obvious when there is an open lesion with the formation of a cystic myelomeningocele. The location of bony dysraphism is more difficult in cases of a closed neural tube defect, or small myelomeningoceles, particularly lesions isolated to the sacral spine. Despite 100% sensitivity for the detection of open spina bifida on ultrasound in high-risk women, Lennon et al. reported that the spinal lesion was only visualised in 95.6% of cases. Although the intracranial findings play a critical role in the diagnosis of open spina bifida, characterising the spinal lesion is essential for outcome prediction. In 1996, Cochrane et al. described neonatal and childhood outcomes, including surgical morbidity, ambulatory patterns, bladder and bowel continence, and school performance, by lesion level, in 85 children diagnosed with open spina bifida and treated in the neonatal period from 1971–1981 . The investigators retrospectively evaluated the prenatal ultrasound images of 26 of these infants, and found that the lesion level could be correctly categorised as thoraco-lumbar (last intact level L2), lumbar (L3–4), and sacral (L5). In 2011, Biggio et al. examined prenatal lesion level by two-dimensional ultrasound for the prediction of ambulatory status by 2 years of age in 33 cases of prenatally diagnosed open spina bifida. The investigators found lesion level on prenatal ultrasound was concordant with postnatal diagnosis within two spinal levels in 92% of cases. No cases with thoracic lesions were ambulatory, 50% of cases with L1–L3 lesions were ambulatory and 100% of L4–S3 cases were ambulatory. The use of three-dimensional ultrasound may be more accurate for the localisation of lesion level compared with two-dimensional ultrasound, but a discrepancy exists with the clinical significance of the improved accuracy, which is generally in the region of one spinal level compared with postnatal findings .

Magnetic resonance imaging

The use of fetal MRI has become an important adjunct tool for perinatologists and paediatric neurosurgeons to aid in counselling and surgical planning for fetuses diagnosed with spina bifida on ultrasound . Mangels et al. demonstrated the ability to characterise accurately the level of myelomeningocele within one to two vertebral levels in 89% of cases using MRI. Important for patient counselling, characteristics of myelomeningocele size and level based on MRI correlate with paediatric outcomes, such as ambulation, bladder function, scoliosis and dysphagia . Compared with ultrasound, MRI seems to be equally accurate for the prediction of the highest affected vertebral level .

Fetal therapy

Prenatal repair of open spina bifida was first carried out in a human in 1997 and, by 2003, over 200 cases had been carried out . This early work demonstrated promising results, but given the invasive nature of the procedure, significant maternal and fetal risks became apparent. Without an RCT to weigh the risks and benefits, prenatal repair of open spina bifida remained controversial. The Management of Myelomeningocele Study (MOMS) was conducted to test the hypothesis that fetal myelomeningocele repair would result in better neurologic outcomes than traditional postnatal repair . Between 2003 and 2010, with the support of the maternal–fetal medicine and fetal surgery communities across the USA, a moratorium on prenatal myelomeningocele repair was placed, and the MOMS trial was able to enroll 183 eligible patients. Findings of the first 158 randomised before July 1, 2009 were reported. The investigators found a reduction in the rate of death, need for shunt by 12 months of age, and hindbrain herniation by 12 months in the group that received prenatal repair. Independent ambulation was doubled, and an improvement in vertebral level of function was two or more vertebral levels higher than traditional postnatal repair. On the other hand, the rates of preterm birth, uterine dehiscence, maternal blood transfusion at the time of delivery, placental abruption, pulmonary oedema, fetal bradycardia, and oligohydramnios were all increased in the prenatal treatment arm. Forty-six per cent of the prenatal repair group was delivered before 34 weeks, and 13% before 30 weeks; however, the perinatal death rate did not differ between groups . In response to the published outcomes of the MOMS trial, the American College of Obstetrics and Gynecology published a Committee Opinion in January 2013 advocating the benefit of prenatal myelomeningocele repair, but stressing the importance of proper patient selection. The College recommended the risks and benefits of prenatal repair should be discussed with any woman who qualifies based on the MOMS trial criteria, and prenatal myelomeningocele repair should be offered only at facilities with the expertise and infrastructure that can provide the high level of intense care required for these patients .

Investigation continues into further techniques to improve prenatal therapy for open spina bifida repair. Currently, minimally invasive surgical techniques using fetoscopy have proven to be inferior to open fetal surgery; however, with advancing technology, some of the technical dilemmas of the fetoscopic approach may improve over time . The more promising treatment and exciting area of innovation has been with a non-traditional approach using engineered tissues held in a flexible patch, sponge or hydro-gel matrix to create a water-tight protective seal over the defect using a single fetoscopic port or a percutaneous approach with an amniocentesis needle .