Scenario

An 8-year-old boy is currently an inpatient in your paediatric unit having had a sudden onset of left sided weakness after a roller-coaster ride. Brain MRI has shown this to be due to a right-sided pontine infarct; brain and neck magnetic resonance angiography (MRA) were normal. Assessment has not identified any risk factors for arterial ischaemic stroke (AIS). Transfer has been arranged to the tertiary centre, where the paediatric neurologists are planning to request diagnostic catheter cerebral angiography (CA) to exclude vertebral artery dissection. The child's parents are apprehensive about this and would like to know the risks of this procedure, especially as they have heard these potentially include neurological complications.

Structured clinical question

In a child with suspected cerebral arteriopathy (patient) undergoing diagnostic catheter angiography (intervention), what is the risk of neurological complications (outcome)?

Search strategy and outcome

Using the PubMed, Medline and EMBASE databases, the terms ‘CA’, ‘digital subtraction angiography’ and ‘cerebral arteriography’ were sequentially paired with the qualifiers ‘adverse effects’, ‘complications’, ‘side effects’ and ‘safety’. The results were filtered by age, to include subjects under the age of 19 years. In total this yielded 149 papers, the abstracts of which were reviewed. The majority of these featured adults and so were excluded. Papers detailing complications of interventional CA were also excluded. Four relevant studies remained, each published within the last 15 years and documenting the complications of diagnostic CA in children. We also identified several other papers which were of interest without directly contributing to our structured clinical question. These will be briefly mentioned to add context to our discussion.

Commentary

Although non-invasive cerebrovascular imaging techniques are available (eg, MRA), CA remains the gold standard for characterisation of cerebrovascular disease. Arteriopathy (an abnormality of the cerebral circulation) is identified in around 80% of children with AIS; in particular, vertebral artery dissection is the most commonly identified risk factor for posterior circulation AIS in children. Although MRA has a high diagnostic yield in the evaluation of childhood AIS,1 it is recognised that the posterior circulation may be difficult to evaluate comprehensively,2 hence the recommendation for CA in the patient described above. CA entails arterial puncture, usually of the femoral artery in the groin, and introduction of a catheter into the neck arteries. Comprehensive cerebrovascular evaluation entails selective catheterisation of the vertebral, internal and external carotid arteries. Potential complications may relate to the femoral arterial puncture (thrombosis, bleeding, dissection, pseudoaneurysm or fistula formation), catheter or wire damage to the neck vessels (dissection or vasospasm), or inadvertent introduction of gas or thrombus from the catheter into the cerebral circulation. The last two complications can result in permanent neurological deficit—a stroke.

The risks of interventional CA are likely to be different and be influenced by the underlying pathology, duration of the procedure, and the nature of the intervention among other factors. These are beyond the scope of this review.

Two paediatric studies from the early 1980s reported iatrogenic emboli in 14 of 3731 (0.4%),3 radiological complications (mainly of vasospasm) in 38 of 1869 (1.1%), and clinical complications including embolism, respiratory depression, bleeding and infection in 12 of 678 (1.8%).4 However, equipment and techniques have significantly improved in the intervening years. Further relevant considerations are changes in the indications for CA, especially with evolution in non-invasive cerebrovascular imaging modalities, and a more aggressive approach to characterisation and treatment of cerebrovascular disease in children.

Three recent, large-scale, prospective reviews tell us that the risks of CA are low in carefully selected adult patients. Two studies documented no permanent adverse neurological events following 2924 and 1715 diagnostic CA procedures, respectively.5 ,6 Willinsky et al 7 described 14 permanent neurological events in 2899 procedures (0.5%). In these reviews, transient neurological complications occurred in 0.06–0.9% of procedures, and non-neurological complications in 0.5–0.66%. These studies also tell us that adverse events are more likely to occur in older adults and those with cerebrovascular atheromatous disease, in emergency procedures and with prolonged procedures.

The four recent studies detailing outcomes in children undergoing diagnostic CA are summarised in table ? 8–11 In total, 653 CA procedures have been documented in these studies. Indications were suspected or known vascular malformation, AIS, intracranial haemorrhage, arterial stenosis or occlusion, vasculitis, venous thrombosis, and vascular tumour or trauma. Neurological adverse events were rare. One child with a complex dural arteriovenous fistula died following a posterior fossa varix rupture 3 h after an uneventful CA.10 One child with a cerebellar arteriovenous malformation (AVM) experienced a haemorrhage 12 h after an uneventful CA. This was treated surgically, but clinical outcome is not recorded.8 In both cases, it is unclear whether the CA was causally implicated in the haemorrhage. One further child with moyamoya syndrome experienced a transient neurological deficit following CA.8

Although our structured clinical question concerns neurological complications, other forms of complication may occur. From the 653 procedures documented in the studies above, non-neurological adverse events were reported in 19 cases, 18 of which represented complications occurring at the puncture site. There is evidence that these groin complications can be minimised by use of a small catheter, ultrasound guidance, minimising manipulation of the femoral artery, and applying prolonged manual pressure, followed by a pressure dressing.11 ,12 Carefully titrated intra-procedural heparin is also used by some operators in those children without known intracranial haemorrhage.10 No child had a reaction to the contrast agent, but this was reported occasionally in the adult studies mentioned (in five of 7538 procedures).5–7 Other potential adverse events not reported in these cohorts include those relating to general anaesthesia and to the effects of radiation exposure. The risks of these are likely to be small in diagnostic procedures, but could not be quantified by our literature review.

The major limitation of these data is the relatively small number of patients. Rarely occurring but serious adverse events may not have been identified. One study was retrospective, increasing the risk that not all cases were captured, or that adverse events were missed. Only one study carried out active surveillance for late-onset complications: Burger et al 10 performed a telephone follow-up of children under the age of 11 enquiring about lower limb extremity complications, and detected no complications at a mean interval of 28 months in the 47 of 71 cases who were reached. All of the papers describe procedures performed at single centres with specialist teams of neuroradiologists. The data may not be applicable to centres where CA is not performed frequently, however, the trend in medical practice is towards centralisation of these specialist skills. Despite these limitations, the published data provide reassurance concerning the safety of diagnostic CA in children.