CHAPTER 186 Hydrocephalus in Children
Approach to the Patient
Recently, however, some evidence suggests that the incidence of pediatric hydrocephalus is decreasing. The number of first shunt insertions for children younger than 17 years decreased substantially in Canada between 1991 and 2000.1 The decrease may be partially related to a decline in the number of children with spina bifida. Case control studies of the effects of folic acid have shown a significant reduction in the incidence of neural tube defects, which have a high association with hydrocephalus.2 Furthermore, after a significant increase in the incidence of intraventricular hemorrhage between the 1970s and 1980s, there has been a marked decrease as experience with managing very preterm infants has grown. This too likely contributes to a decreased incidence of hydrocephalus in children.3 Societal decisions about how to treat very premature infants or those with significant malformations diagnosed in utero4 may also influence the incidence of hydrocephalus in live births.
Despite these trends, the burden of this illness remains large. Using the Healthcare Cost Utilization Project Kids’ Inpatient Database, a cross-sectional survey was performed in 1997, 2000, and 2003. Each year there were almost 40,000 admissions, approximately 400,000 hospital days, and between $1.4 billion and $2 billion in hospital charges for pediatric hydrocephalus. This accounted for 3.1% of all pediatric hospital charges. In addition, the children identified in this cross-sectional study had an increasing frequency of comorbidities.5 Clearly, pediatric hydrocephalus represents a huge burden of illness for children and is part of the daily lives of neurosurgeons in general and those treating children in particular. As a further testament to the frequency of hydrocephalus, the Hydrocephalus Clinical Research Network, a newly formed cooperative clinical trial group consisting of four pediatric neurosurgical centers, accumulated almost 1000 shunt procedures in the first 8 months of data acquisition.
Presentation
The decision to treat a child with ventriculomegaly can be very difficult. Once a shunt has been implanted, it is very difficult to determine whether it can be removed. The use of adjunctive measures, such as intracranial pressure monitoring,6 magnetic resonance spectroscopy,7 and the magnetic resonance measurement of cerebral blood flow,8 has been reported in difficult cases, but the decision to treat is usually based on observation over time. Progressively increasing head size, enlarging ventricles, or progressive symptoms are the most common measures and form the most solid basis for making the decision to treat.
Disease-Specific Considerations
Congenital Hydrocephalus
The majority of children with hydrocephalus present at or soon after birth. Many of them have aqueduct stenosis, Dandy-Walker malformation, holoprosencephaly, or other more generalized malformations of brain development. Aqueduct stenosis in males may be X-linked.9 In these children, the hydrocephalus is usually quite severe, and they have the clinical finding of adducted thumbs (Fig. 186-1). There may be other affected males in the family or a maternal history of spontaneous abortion.
Hydrocephalus Associated with Myelomeningocele
A newborn with myelomeningocele undergoes closure of the spinal defect and then observation for the development of hydrocephalus. In the past, 80% of children were thought to require ventriculoperitoneal shunt placement, but reduced rates of shunt placement have recently been reported.10 The most common manifestations of hydrocephalus in these children are increasing head circumference, splitting sutures, and full fontanelle, but some children may develop a large pseudomeningocele at the myelomeningocele repair site or a cerebrospinal fluid leak. These manifestations are often thought to be related to hydrocephalus and require the placement of a shunt. The importance of hydrocephalus in this population is emphasized by a multicenter trial funded by the National Institutes of Health that aims to randomize 200 fetuses to in utero or postnatal myelomeningocele closure. Based on suggestive preliminary data,11 the trial is testing the hypothesis that in utero closure can reduce the need for shunt placement by reducing the incidence of Chiari II malformation, the major cause of progressive hydrocephalus in this patient population.
Arachnoid Cyst
Midline and posterior fossa arachnoid cysts in newborns commonly cause obstructive hydrocephalus. These may occur in the suprasellar area (Fig. 186-2), quadrigeminal cistern (Fig. 186-3), or cerebellopontine angle. Endoscopic fenestration of the cyst rather than treating the ventricular system may relieve obstruction and reestablish normal flow.
Posthemorrhagic Hydrocephalus
Intraventricular hemorrhage in premature newborns is common and is related to the degree of prematurity and the birth weight.12 The probability of developing posthemorrhagic hydrocephalus depends on the grade of intraventricular hemorrhage. An overall 40% incidence of ventriculomegaly has been reported,13 but the incidence can be as high as 70% in patients with grade IV intraventricular hemorrhage.14 In managing these children, it should be recognized that a significant rate of arrest or resolution of this type of hydrocephalus has been reported.15 In these tiny infants, ventriculoperitoneal shunt insertion can be difficult and has a high rate of complications. A number of options for delaying shunt insertion have been used, including serial lumbar punctures or treatment with furosemide (Lasix) and acetazolamide (Diamox). None of these measures has been shown to reduce the incidence of long-term hydrocephalus in randomized trials.16,17 Temporizing with either a subgaleal shunt or a ventricular reservoir until the child reaches a weight of 1500 to 2000 g is a common practice. The proportion of children who receive such a temporizing measure and go on to permanent ventriculoperitoneal shunting is approximately 70% to 90%.18 An aggressive approach to reduce hydrocephalus after premature intraventricular hemorrhage was recently attempted using drainage, irrigation, and fibrinolytic therapy. Although a promising pilot study showed a reduced requirement for shunt surgery,19 a prospective randomized trial was stopped early because of an increased rebleed rate in the treatment group.20 Despite that, the 2-year follow-up showed a reduction in death or severe disability.21
Hydrocephalus Associated with Brain Tumors
The tendency for children’s brain tumors to occur in the posterior fossa and midline leads to a high incidence of associated hydrocephalus. Management with preoperative shunt placement is no longer common practice, and most surgeons opt to remove the tumor and monitor for the development of hydrocephalus. Recently, third ventriculostomy performed before tumor removal was reported to reduce the risk of hydrocephalus significantly.22 The criticism of this approach is that some of these third ventriculostomies may be unnecessary because a proportion of children will not develop progressive hydrocephalus after tumor removal. A validated patient score for predicting the development of hydrocephalus in these children before tumor resection has been reported (Table 186-1).23 Based on age, papilledema, severity of hydrocephalus, metastatic disease, and estimated preoperative tumor type, the chance of developing hydrocephalus can now be predicted before resection of the tumor (Table 186-2). Evaluating these factors allows a more informed discussion with patients and families and possibly the selective use of endoscopic third ventriculostomy before tumor surgery. External ventricular drain insertion at the time of tumor removal is common for tumors within the fourth ventricle but may be avoided in cerebellar hemispheric tumors. Following surgery for tumors in the lateral ventricle that are associated with hydrocephalus, the surgical tract may lead to postoperative decompression of the hydrocephalus into the subdural space. When this collection persists as a subdural hygroma, it may require treatment with a subdural shunt.