Clinical manifestations

Presentation and course

Cerebral angiography may be complicated by a diverse range of neurologic deficits. Onset is typically acute; it commonly occurs during the procedure. Delayed complications such as abscess formation may also occur after acute procedures (106). Anterior circulation deficits usually manifest as either a unilateral motor deficit (eg, hemiparesis, hemiplegia, fine motor "clumsiness") or a sensory deficit (eg, hemineglect, hypesthesia, paresthesia, dysproprioception, and dysgraphesthesia). Various forms of aphasia may result from injury to the dominant hemisphere. In contrast, posterior circulation insults during cerebral angiography are often nonlateralizing (with the exception of occlusions of the posterior inferior cerebellar artery arising from the ipsilateral injected vertebral artery), which can produce a Wallenberg syndrome. Such posterior circulation ischemic events may manifest as bilateral and asymmetric upper or lower extremity paresis, hemisensory deficits, ataxia, dyskinesia, multiple cranial neuropathies, "cerebellar" dysarthria, vertigo, and visual field deficits. Frequently, there is also an associated alteration of consciousness, due to involvement of the reticular activating system.

More unusual neurologic deficits may be poorly localizing. Cortical blindness may ensue following selective injection of contrast medium into the vertebrobasilar circulation (97; 41). A wide spectrum of generally transient neurologic deficits has also been reported, including amblyopia, hemianopsia, visual agnosia, visual hallucination, seizures, bulbar syndrome, and quadriparesis (96).

Transient anterograde amnesia has also been described (35). Transient cognitive dysfunction, such as various types of apraxias (eg, constructional, mathematical, writing), have also been described (50; 35). Ocular complications occur occasionally and usually manifest as transient monocular loss of visual acuity or field of vision. Central retinal artery thrombosis, which has an associated high incidence (up to 40%) of permanent visual deficit, may also occur (52).

A great number (40% to 60%) of periangiographic neurologic complications are transient ischemic attacks that usually resolve within minutes to hours (25; 49; 20; 65). Such events are presumed secondary to the temporary embolic occlusion of a vessel, followed by spontaneous recanalization.

Prognosis and complications

Discussion of the spectrum and significance of complications directly or indirectly resulting from a stroke related to cerebral angiography is beyond the scope of this review. These complications are essentially the same as those characteristic for ischemic stroke in general.

Clinical vignette

A 58-year-old man developed the acute onset of hemiparesis and language difficulties during angiography for evaluation of carotid stenosis. Examination revealed global aphasia and right hemiparesis. Emergent intraarterial thrombolysis with tissue plasminogen activator was performed for an occlusion of the left middle cerebral artery. After 30 minutes of thrombolytic therapy, recanalization of the left middle cerebral artery was achieved with simultaneous resolution of the prior neurologic deficits.

Biological basis

Etiology and pathogenesis

Various etiologies may be invoked in cerebral ischemia resulting from cerebral angiography. Although several pathophysiologic mechanisms are readily apparent, others may have indirect effects that contribute to the development of a stroke. Several studies have included abrupt worsening of a patient's neurologic status within 24 to 72 hours after angiography as an angiographic complication (25; 98; 40; 39). However, the precise causal relationship is often not evident. It is particularly difficult to determine whether a postangiographic ischemic event is actually attributable to the procedure in patients with recurrent transient ischemic attacks and in those with severely stenotic atherosclerotic vessels.

The majority of events are likely thromboembolic (see Table 1). Emboli can result from foreign-body material dislodging from the catheter tip (catheter fraying and fracture), guide-wire fracture, or atheromatous plaque disruption (74). More commonly, however, thromboemboli are due to fibrin and platelet thrombus accumulation on the distal portion of a catheter (50). Inadvertent injection of an air embolus also has been reported since the earliest clinical applications of this technique (50; 69). However, this complication is rarely encountered with current angiographic techniques, mainly due to implementation of a closed flush system and rigorous emphasis on meticulous elimination of atmosphere within the syringe. Furthermore, a relatively large volume of air must be injected to produce significant cerebral branch occlusions. This has been indirectly supported by a couple of studies using transcranial Doppler monitoring of the M1 segment of the middle cerebral artery during diagnostic cerebral angiography, which have shown that despite introducing innumerable atmospheric microemboli with each injection, no clinical sequelae were noted (57; 69).

In certain cases, direct or indirect neurotoxicity of the injected contrast agent is responsible for the transient neurologic deficits (72). Several animal studies have shown that contrast hyperosmolality is capable of producing transient breakdown of the blood-brain barrier (42; 51; 50; 84). Breakdown of the blood-brain barrier may produce a myriad of local biochemical and physiologic derangements of both glia and neurons, either directly by chemotoxicity of the contrast agent or indirectly by sudden changes in extracellular ionic concentration. Such derangements may produce seizures, membrane potential dysfunction (eg, axonal hyperpolarization preventing action potential transmission), and excitotoxic synaptic transmission.

Intraarterial injection of contrast agent may also produce other important local and systemic effects, which can indirectly contribute to the risk of periangiographic stroke. For example, systemic hypersensitivity reactions and contrast agent–induced vasovagal episodes can produce profound hypotension, which may contribute to regional cerebral ischemia in an already compromised cerebrovascular system. Contrast agents also may produce paradoxical activation or inhibition of various coagulation factors. For example, ionic contrast agents tend to inhibit in vitro clotting, whereas nonionic agents have either no effect or a procoagulant effect in vitro (09; 94; 37). There is, however, direct and indirect evidence that both types of contrast agents may have procoagulant effects in vivo (eg, platelet activation) that may increase the risk of thromboemboli (41; 94; 37).

There are also important patient factors that likely increase the risk of stroke associated with cerebral angiography. Numerous studies that have demonstrated that patients with cerebrovascular disease are more likely to experience neurologic sequelae related to the procedure. A series of carotid stenting cases noted that severe aortic arch calcification and target lesion ulceration may be associated with an increased risk for magnetic resonance diffusion-weighted imaging-detected embolic events (56; 86).

Table 1. Possible Direct Mechanisms Responsible for Periangiographic Stroke

Table 2. Possible Indirect Mechanisms Responsible for Periangiographic Stroke

Table 3. Factors Predisposing to Periangiographic Stroke

Epidemiology

The true incidence of stroke associated with cerebral angiography (ie, the occurrence of an abrupt neurologic deficit, which is presumably ischemic in etiology, during an arbitrary interval of time around the performance of the procedure) has been difficult to establish. Several limiting factors have been recognized. Every study specifically designed to evaluate periangiographic neurologic complications has used a heterogeneous group of patients with both cerebrovascular occlusive disease and other types of both cerebrovascular and nonvascular central nervous system diseases. These studies generally fail to properly define the severity of encountered cerebrovascular occlusive disease and do not use methods to control for the natural history of these diseases. For example, some series may contain large numbers of patients with unusually severe and unstable cerebrovascular occlusive disease (eg, critical carotid artery stenosis, crescendo transient ischemic attacks, stroke in evolution), who are statistically more likely to suffer either a transient or permanent ischemic neurologic deficit during the arbitrarily defined periangiographic period due to the natural history of that disease (rather than due to complications arising from the angiogram). Inclusion of such patients without adjusting for the natural history of the disease, therefore, will introduce bias in establishing the actual risk of complication related to the procedure. A few investigators have acknowledged this potential confounding variable (40). In addition, prior studies have failed to correlate lateralizing signs and symptoms to specific vessels and vascular territories injected during the angiogram, thus, uncoupling the essential pathoetiologic link necessary to establish a causal relationship between cerebral angiography and stroke. Finally, the rapid evolution of techniques, technology, and indications for diagnostic and therapeutic neuroangiography make it difficult to make a "static" assessment of risk for complication because these various aspects of the procedure and practice are changing in ways that cannot be easily adjusted for both retrospective and prospective analyses.

Several prospective and retrospective studies have reported a highly variable incidence of transient and permanent neurologic complications related to cerebral angiography in the range of 1% to 12% (29; 27; 25; 49; 33; 20; 38; 70; 101; 65; 91; 107; 40; 39; 44; 103; 13; 67). In contrast, several other studies (both prospective and retrospective) have reported either low (less than 0.3%) or negligible neurologic complications associated with cerebral angiography (05; 75; 92; 66; 80). Reported permanent neurologic deficits resulting from angiography vary from 0% to 5.4% (103; 67). Some investigators have attempted to use metaanalysis techniques to obtain an average incidence of complications, but these studies eliminate many previous works that do not meet certain entry and validation requirements (eg, retrospective studies, inadequate follow-up, and poor definition of complications). These meta-analyses suggest that the average rate of neurologic complication associated with cerebral angiography may be between 2% and 4% (65; 40).

Dawkins and colleagues provided a single center review of angiographic complications. They noted clinical complications in 0.79% of 2924 angiographic procedures, including 0.41% significant puncture-site hematomas, 0.34% transient neurologic events, and 1 nonfatal contrast reaction. Asymptomatic technical complications occurred in 0.44%, including 3 groin dissections and 10 cervical dissections. Emergency procedures were associated with higher complication rates (15).

A variety of patient subgroup characteristics have been identified that appear to be associated with an increased risk of neurologic complications (29; 27; 50; 49; 77; 20; 70; 98; 91; 40; 39; 66). Most of these studies indicate with various degrees of statistical power that patients with significant atherosclerotic disease of the carotid and vertebral arteries are probably at highest risk for both transient and permanent ischemic complications related to cerebral angiography. A large prospective study of 2899 individuals found that neurologic complications of cerebral angiography were significantly more common in people with cardiovascular disease (103). A metaanalysis of 3 published prospective studies revealed that the incidence of neurologic angiographic complications was also much higher in patients with a history of transient ischemic attacks or stroke compared to those with aneurysms, subarachnoid hemorrhage, or arteriovenous malformations (12).

Two studies have also specifically addressed the risk of cerebral angiography in patients suffering from migraine headaches (49; 91). Patients with complicated migraines were significantly at risk of not only having a severe migraine headache precipitated by cerebral angiography but also of developing significant neurologic deficits (49). Another study, however, did not show any increased risk of periangiographic complication associated with migraine (91).

Other important risk factors include elevated creatinine (higher than 1.2 mg/dL), an increased number of catheters used (more than 1), an increased volume of contrast agent, the experience of the neuro-angiographer, the length and difficulty of the procedure, and the patient's age (25; 77; 33; 20; 70; 44; 103; 56; 85). One study noted the safety of carotid angiography performed by vascular surgeons (87).

Prevention

The prevention of ischemic complications resulting from cerebral angiography has been one of the major motivating forces for the technical advancement and refinement of this diagnostic procedure. Improvements in risk reduction first occurred with the numerous technical advances and refinements in direct puncture and percutaneous catheterization techniques, catheter and guide-wire technology, and contrast agents early in the "modern era" of cerebral angiography. Further refinements in catheter and guide-wire materials and the establishment of training standards for the then newly defined specialty of neuroradiology in the late 1960s and early 1970s also probably contributed to the reduction of complications.

One of the most important preventive measures is the combination of strict adherence to meticulous technique (eg, double flush infusions within the catheter, smooth and atraumatic manipulations of guide wire and catheter, minimization of the time a catheter is positioned within a vessel) and the performance and supervision of the procedure by an experienced neuro-angiographer. Such recommendations are not only intuitively credible, but also have been supported by some studies specifically addressing these issues (77; 70; 19). Additional measures, such as continuously perfusing a positioned catheter with heparinized saline and using hand-injection technique, also probably contribute to enhancing the overall safety of both diagnostic and therapeutic cerebral angiography. Performance of the cerebral angiogram in a modern angiography suite with high-resolution digital subtraction capability (1024 x 1024 pixel matrix) has enabled more rapid and safer acquisition of digital images using hand injections of contrast agent (19). This technological advance has dramatically diminished the overall time of these procedures and has obviated the need to use an automated pressure injector (which may increase the risk of adverse events such as introduction of air emboli, subintimal injection and dissection, and overinjection of contrast media). Because previous studies have shown that duration of the procedure increases the risk of ischemic complication during diagnostic neuroangiography, it is likely that this technological advance has contributed to reducing complications. With the expanding number of endovascular specialists performing angiography for carotid revascularization, operator-related factors in the development of angiographic complications are of increasing importance. One study suggested equivalent safety when diagnostic angiography is performed by vascular surgeons (01). A large series of diagnostic angiography performed by cardiologists revealed 0/333 stroke complications (02). Another large series demonstrated that the risk for neurologic complications related to catheter-based diagnostic cerebral angiography can approach zero (99). Even as diagnostic cerebral angiography proliferates in new geographic regions, the safety rates remain consistent (08).

Periprocedural anticoagulation may reduce thromboembolic phenomena (31; 06; 95; 17; 37; 19; 32). Investigators performing coronary and peripheral vascular angiography demonstrated such use with a combination aspirin and heparin (31; 06; 95). Debrun and colleagues were the first to advocate routine prophylactic antithrombotics using aspirin and heparin (17). Currently, the most popular method of prophylactic anticoagulation for cerebral angiographic procedures is the use of moderate-dose bolus infusion of heparin. This has been particularly emphasized in neurointerventional cases that are often lengthy, when a guiding catheter is usually left within a vessel for extended periods of time (17; 32). The recommended anticoagulation regimen during diagnostic and interventional neuroangiography varies somewhat based on institution; an intravenous bolus of between 3000 units and 5000 units of heparin sulfate is administered initially, followed usually by repeat bolus administration of 1000 units every hour (32). Continuous flush infusion of heparinized normal saline (1000 to 3000 units/L) through either the diagnostic or guiding catheter also offers added protection by eliminating stagnation of blood (and, thus, preventing fibrin and platelet accumulation) on the distal tip and within the lumen of the catheter.

Use of an antiplatelet agent may provide additional protection from iatrogenic thromboemboli during cerebral angiography. The major disadvantage of this approach is the inability to rapidly reverse the antiplatelet action, which may lead to increased hemorrhagic complications (eg, groin hematoma, retroperitoneal hematoma) or the natural history of certain diseases (eg, intracranial aneurysms, arteriovenous malformations).

In elective cases, renal function studies before angiography may be used to minimize risk of contrast-induced nephropathy. The urgency of acute stroke triage, however, should not be delayed by waiting for these results (71).

Intraprocedural monitoring with transcranial Doppler ultrasound may also increase detection rates for downstream emboli, allowing refinement in technique or even immediate treatment (82).

Differential diagnosis

Acute neurologic deterioration during or immediately after cerebral angiography is most likely due to thromboembolism or neurotoxicity of the contrast agent. Stroke mimics should also be considered, however. Periprocedural seizures may occur in those patients who have medically intractable seizures (who are often referred for Wada provocative testing in evaluation for epilepsy surgery) and cerebral arteriovenous malformations. Fatal stroke has also been reported after Wada testing (28). Theoretically, such patients may be more sensitive to the neurotoxic effects of the contrast medium, leading to precipitation of seizure activity during angiography. Vasovagal episodes may produce either nonfocal or focal neurologic deficits depending. Patients with various types of both congenital and acquired alterations in the major brachiocephalic arteries (eg, carotid dissection, atherosclerosis of the internal carotid artery), circle of Willis (eg, hypoplasia of the A1 segment, atherosclerotic stenosis of the M1 segment), and pial vasculature (eg, arteriosclerosis, dysautoregulation) are at risk for this complication. Fortunately, the majority of such cases are reversible when adequate cardiac output is reestablished. Various etiologies for vasovagal events during cerebral angiography have been suggested, including pain, psychological stress, mechanical stimulation of the carotid sinus, and idiosyncratic reaction to contrast media. Micturition syncope has also been anecdotally reported to have occurred during angiography and is likely a variant of vasovagal events. Severe hypoglycemia from both natural and iatrogenic etiologies may produce neurologic deficits. Usually, these episodes are readily identified due to prior knowledge of the patient's history (eg, insulin-dependent diabetes, commonly associated medications) and associated signs and symptoms (eg, diaphoresis, palpitations, double vision, alteration in the level of consciousness). Finally, a variety of cardiac arrhythmias (eg, paroxysmal atrial tachycardia, ventricular tachycardia, sick sinus syndrome, aberrant reentry tachycardias) may produce hemodynamically significant reductions in cardiac output, resulting in either regional or global cerebral ischemia. Bradyarrhythmias (eg, sick sinus syndrome, digitoxin toxicity) and tachyarrhythmias (eg, atrial flutter with supraventricular tachycardia, ventricular tachycardia) are not uncommon in the older population of patients referred for cerebral angiography. Most cardiac arrhythmias can be prevented or recognized early enough to minimize their clinical impact. Diagnosis of a cardiac arrhythmia during cerebral angiography is usually easily recognized because it is generally standard practice for patients to be continuously monitored by ECG during the procedure.

Diagnostic workup

Further evaluation of a periangiographic stroke frequently depends on the circumstances surrounding the event and the suspected etiology. Strokes occurring during the procedure should be evaluated by repeat injection into the suspected vascular distribution of the affected part of the brain in order to rule out a major thromboembolic event. Identification of major branch occlusions in this setting may be an indication for superselective intraarterial thrombolysis. Similarly, foreign-body emboli (eg, avulsed catheter fragment), although presently a rare complication of modern cerebral angiography, may be retrieved using neurointerventional techniques. Symptomatic catheterization-induced dissection of the carotid or vertebral arteries should also be identified because such vascular injuries are usually best treated with extended systemic anticoagulation. Although also uncommon, both catheter-induced and contrast media-induced vasospasm that is hemodynamically significant should be identified, as these events may be treated effectively by systemic and local pharmacologic interventions.

Particularly severe lateralizing and global events should always prompt an emergent CT examination of the head to exclude the possibility of acute intracranial hemorrhage (particularly intraparenchymal hematoma), as these events occasionally occur in certain patients undergoing cerebral angiography, such as those with cerebral arteriovenous malformations, dural fistulae with cortical venous drainage, intracranial aneurysm, and both chronic and acute hypertensive disease.

Hyperacute MRI, including diffusion and perfusion weighted imaging, may provide additional useful prognostic information regarding the extent and potential reversibility of ischemic injury (90).

Management

Despite implementation of the previously outlined preventive measures, strokes still occur in association with cerebral angiography. A variety of options are currently available in the therapeutic armamentarium to manage such ischemic events; however, many of these therapies are considered experimental or have not been fully tested for efficacy. Every therapeutic option also subjects the patient to some potentially adverse events. Consequently, specific management strategies must be individualized; the suspected etiology, pathophysiology, and prognosis and severity must be considered in weighing the possible risks and benefits of a particular therapeutic intervention.

Because many deficits have been shown to be transient in nature, conservative, supportive management is frequently all that is necessary. Ischemic events occurring on the angiography table can be quickly evaluated for major vascular occlusion and then monitored closely for a short period of time to assess for evidence of reversibility, stability, or progression. Systemic etiologies for cerebral oligemia (eg, hypotension, cardiac arrhythmia) and alternate mechanisms for either focal or global neurologic dysfunction (eg, seizures, adverse pharmacologic reaction) should also be evaluated.

Conservative management may incorporate anticoagulation or antiplatelet therapy in situations of suspected thromboembolic phenomena occurring during the procedure. In this particular setting, heparinization during (if not already prophylactically instituted) and shortly after the acute phase of a cerebral ischemic event not only offers the theoretical advantage of decreasing the risk of repeated thrombotic and embolic events, but may also actually enhance the activity of the patient's native thrombolytic system and minimize propagation of thrombus from an embolic nidus.

Acute focal cerebral ischemia also may benefit from standard medical management maneuvers used in the intensive care unit for neurosurgical patients with a variety of neurologic problems.

Intraarterial thrombolysis of a proximal embolus produced iatrogenically during cerebral angiography may be an effective therapeutic option in selected patients (07). This endovascular therapeutic technique requires a well-trained operator skilled in superselective micronavigation to position an infusion microcatheter into the occluded intracranial branch (eg, M1 segment of the middle cerebral artery). The occluded vessel is then continuously infused with a thrombolytic agent (eg, urokinase, tissue plasminogen activator) over a 1- to 2-hour time interval. Some investigators also have advocated performing some limited degree of mechanical disruption of the embolus (which increases the exposed surface area of the clot to both exogenous and intrinsic thrombolytic agents). This latter maneuver, however, is highly controversial because many investigators believe that such manipulation of thrombus may produce downstream emboli. Some of the most dramatic, excellent outcomes of this technique have been reported in the setting of an acute thromboembolic event occurring during either diagnostic or therapeutic neuroangiography (108; 18). Angioplasty and stenting may be performed with a wide range of reported procedural complication rates (14; 26).

Retrieval of the foreign-body emboli that may rarely occur from angiographic catheters and guide wires is now possible with a variety of endovascular snares that have been developed. Such devices designed for intracranial navigation must either be used with an existing microcatheter positioned into the cerebral vasculature or be incorporated into a microcatheter system that is then navigated into a targeted vessel.

Fortunately, acute symptomatic dissection during cerebral angiography is distinctly unusual. Most acute dissections during these procedures are actually localized subintimal tears that produce only partial compromise of the vascular lumen. These small dissections almost invariably heal within a period of 3 to 6 months and usually produce no significant clinical sequelae. Many of these lesions are treated with oral anticoagulation during this period. Rarely, these localized subintimal tears can produce a hemodynamically significant compromise of the vessel lumen, which also usually do not produce acute symptoms if adequate collateral circulation is present within the circle of Willis. However, if collateral reserve is inadequate, symptomatic oligemic (or "hemodynamic") cerebral ischemia ensues. These rare events may be amenable to endovascular interventional techniques. One technique described by Dion and colleagues for repairing acute dissection is using "low-pressure" angioplasty for tracking down acute intimal tears (21). Angioplasty with stent deployment may also be used to treat cervicocephalic arterial dissection. When traumatic carotid cavernous fistulas are created during endovascular thrombectomy for acute ischemic stroke, they may be successfully treated if conservative management fails (03).

Catheterization-induced vasospasm is usually asymptomatic and brief in duration. Occasionally, it may be severe enough to produce ischemic symptoms (more commonly in patients who already have compromised collateral reserve). In such situations, a variety of endovascular therapeutic options are available. Topical application of nitroglycerin paste has been used for a number of years to minimize catheter-induced vasospasm of the external carotid circulation. Intraarterial infusion of nitroglycerin into an affected brachiocephalic vessel is not recommended because of potential serious neurotoxicity. However, if topical nitroglycerin is ineffective, local intraarterial infusion of papaverine hydrochloride may be used. This technique has been used effectively for treatment of post-subarachnoid hemorrhage-induced vasospasm of the intracranial vasculature (55; 19) and has been anecdotally used for treatment of larger vessel vasospasm (19). Finally, low-pressure angioplasty with a silicone elastometer balloon may be used in analogous fashion to intracranial angioplasty for subarachnoid-induced vasospasm (19).

Special considerations

Pediatric age group

In children younger than 3 years of age, the rate of complications for cerebral angiography is comparable to rates reported for older children and lower than rates reported for adults (47; 63).