Introduction

The intervention of cerebrospinal fluid (CSF) shunting for CSF accumulations has brought about long-term survival and avoidance of disabilities in children suffering from hydrocephalus. Unfortunately, the insertion of inert non-growing hardware in infants and children who are usually very active and can expect to grow 20 to 30 times their birth weight leads to a high rate of shunt complications. Some studies find that 60% or more shunts need revising after several years. The task of the Emergency Physician is to diagnose those complications, commence time critical treatments and refer to a neurosurgical service when appropriate.

Clinical presentation

The developmental stage of the child with a CSF shunt can cause considerable variation in the clinical presentation of shunt complications. The most obvious is the presence of an open anterior fontanelle in infants up to the age of approximately 9 to 18 months. Simply looking and feeling the fontanelle allows an estimation of intracranial pressure. A bulging fontanelle is a highly specific but not very sensitive sign of under-shunting; a sunken fontanelle can be a sign of overshunting.

The cranial sutures in children are not fused and can undergo diastasis due to raised intracranial pressure. The older a child is, the longer and higher the intracranial pressure has to be to cause diastasis. This also means that the head circumference will rapidly increase when there is inadequate shunting. Therefore it is important to measure the occipitofrontal circumference and compare it with previous records if available and plot it on growth centile charts. A head circumference that is rapidly crossing centiles in an upward fashion or a head circumference that is in the very high centile range especially when the other parameters, weight and length, are not, is an indicator of undershunting. Another implication of unfused cranial sutures is that because this allows an increase in cranial volume it retards the rise in intracranial pressure. This may be the reason that small infants present with more non-specific signs and symptoms than older children.

The most common question an emergency physician will have to answer when confronted with a child who has a CSF shunt is ‘Should I refer this patient to the neurosurgical service?’ A recent study by Piatt et al¹ attempted to quantify the power of various symptoms and sign to predict shunt malfunction or infection. Table 8.1.2 shows the strongly predictive signs and symptoms that allow referral to be made on the basis of that single finding. This may be done even before computerised tomography (CT) scanning, as the neurosurgeon may prefer to have the scan done locally to allow for easier comparison with previous scans (see discussion of CT scanning below). Table 8.1.3 shows the signs and symptoms with strong positive predictive power but not strong enough to warrant immediate referral if just one feature is present on its own. Thus in a patient with a ventriculoperitoneal shunt (VPS) who presents with fever alone, an initial general work up for a cause of the fever is warranted and then consideration for referral to neurosurgery made if no definite cause for the fever is found. However, fever with another feature listed in Table 8.1.3, such as headache, warrants early neurosurgical referral. The question of whether to send the child home is more difficult. The absence of any of the symptoms and signs listed in Tables 8.1.2 and 8.1.3 does not rule out shunt malfunction or infection. If concerned about shunt infection, this is much less likely if the patient is more than 6 months from the last shunt insertion or revision. However, this does not exclude shunt malfunction. Where a symptom or sign that is not of high predictive power is adequately explained by another diagnosis (e.g. vomiting with diarrhoea and recent contact with a case of gastroenteritis) CSF shunt complication is very unlikely. For cases where CSF shunt complications are neither ruled in or out on historical and examination findings, one must resort to investigations and/or observation and/or neurosurgical consultation.

Examination

For examination features of raised intracranial pressure see Chapter 8.2. Features of infection e.g. fever, rigors, lethargy, localised redness, swelling and peritonism (ventriculoperitoneal shunt) or pleurisy (ventriculopleural shunt) should be sought.

Investigations

On rare occasions, insertion of a needle into a shunt pumping chamber may be useful for both diagnostic and therapeutic reasons but should generally be performed by neurosurgical colleagues. The exception is in the rapidly deteriorating child, where contact with the neurosurgeon is delayed or where in remote locations after discussion with the treating neurosurgeon it is considered important to get some information to help with a decision to transport. In the case of a moribund child, an emergency physician may relieve the raised intracranial pressure by inserting a 25-gauge butterfly needle into the bulb/pumping chamber at 45 degrees to the skin under strict aseptic technique.² It is important not to advance too deeply as the needle can damage valve mechanisms so badly that replacement is necessary. The pressure can be measured with a similar technique to that used in a lumbar puncture and then CSF drained off until the pressure is 10 cmH₂O.

In the stable child, an X-ray of the entire shunt may demonstrate a disconnection or kinking causing blockage. The CT scan is the usual preferred method of imaging, because it provides clear images in a short space of time. However, difficulties associated with obtaining a CT scan need to be considered. One difficulty is keeping the child still long enough to obtain adequate images even with rapid CT scanners. In infants, firm wrapping and sucrose on a dummy or pacifier may be sufficient to keep them still. When these fail procedural sedation or general anaesthesia are required. This is less desirable because of anaesthetic risk, longer delays, use of resources, and the potential to temporarily obscure one of the important signs of raised intracranial pressure (ICP), decreased level of consciousness. Another consideration is the harmful effects of the radiation exposure. A protocol designed to minimise radiation exposure in children should be used and the head angulated to avoid radiation exposure to the eyes. This reduces the risk of premature cataract formation. These precautions may not be routine at institutions that do not frequently scan children and the clinician may have to ensure that these things are done. Some studies have examined magnetic resonance imaging (MRI) as an alternative to CT scanning; however, the scan takes longer, makes it difficult to access the child, often requires the child to be sedated or anaesthetised and often does not provide images as clear as a CT.

Enlarged ventricles on scanning may indicate undershunting due to obstruction. However, this may be chronic and comparison with previous CT scans is necessary for accurate interpretation. Obliteration of the perimesencephalic cistern is a particularly worrying sign and mandates urgent neurosurgical consultation. The scan may also show the reason for malfunction (e.g. catheter tip embedded in brain tissue) or it may show a ventriculitis when the lining of the ventricles enhances with contrast. Alternatively, neuroimaging may reveal small ‘slit-like’ ventricles. These are thought to be due to overshunting and may be non-compliant. This leads to sharp fluctuations in ICP, with very little change in CT appearance. The management of this complication is a neurosurgical challenge. The emergency physician needs to be aware that small ventricles do not mean a normal ICP. Nuclear medicine shunt function studies are also used in some centres for diagnosing shunt blockage. An abdominal ultrasound may be useful to identify a CSF pseudocyst.