Enthusiasm for delivering an alternative blood supply to the brain by surgical means has waxed and waned for over three decades. In situations where a major cerebral vessel is sacrificed for removal of macroscopic pathology (such as a skull base tumour or giant intracranial aneurysm), the need to replace lost cerebral blood flow (CBF) is obvious and often required as part of a staged surgical procedure.^1,2 In such cases, the need and type of surgical bypass graft (high or low flow) is dictated by the presence or absence of adequate collateral vascular pathways, the design of the circle of Willis, and any extracranial to intracranial (EC-IC) vascular connections.³ Detailed cranial angiography and observation of the clinical and physiological responses to temporary test occlusion of parent vessels provide the relevant information. When reliably practised, daunting cervical and cranial pathologies can be approached confidently with acceptable morbidity.⁴ Although simple in principle, techniques for assessing the need for surgical bypass procedures are not always practised resulting in incomplete treatments, and/or high surgical morbidity from cerebral strokes. Centralisation of expertise and adoption of a more appropriate referral of difficult pathology will serve to address the variation in practice for such cases.

CEREBRAL REVASCULARISATION IN CHRONIC VASCULAR OCCLUSION

A far more intense debate surrounds the indications for cerebral revascularisation in the treatment of chronic vascular occlusive disease. Extracranial atheromatous degenerative vascular disease is by far the commonest pathological entity falling into this group,^5,6 although intracranial atheroma and chronic inflammatory disease (Moya Moya) is significant in Eastern communities. Patients usually present with neurological events leading to investigation and identification of occlusive disease. However, unlike stenotic disease, occluded vessels do not attract a high risk of future embolic events.^5–7 Most patients achieve normal or near normal CBF by means of the aforementioned collaterals. Only a minority of those who survive the occlusion without an immediate major stroke exist in a chronic state of low CBF. Accurate neurological evaluation of the presenting symptoms (retinal v cerebral) is clearly important in determining the risk of further clinical events.⁸ However, from a clinical and pathophysiological standpoint, these patients should be considered for CBF assessment and cerebral bypass surgery—not for CBF replacement but for CBF augmentation.^9–11

WHY THE CONTINUED DEBATE?

Since the first description of an EC-IC bypass procedure by Yasargil,¹² much harm has been done in the wide propagation of the procedure without resort to a sound physiological or clinical substrate, upon which to base the procedure. Most patients presenting with symptomatic carotid occlusive disease accommodate the pathology well; hence, most do not need a cerebral bypass. Undiscerning surgical practice raised medical concerns, the result was the controversial multicentre Cooperative Study (EC-IC Bypass Study) that failed to observe any benefit from the bypass procedure.¹³ The widespread surgical practice of EC-IC bypass collapsed, largely as a result of withdrawal of the funding mechanisms supporting the operation. Those who were more thoughtful in the application of EC-IC bypass procedures dissented.^14–16 Why was there no selection of patients more suitable for the procedure, particularly those who were clinically symptomatic of cerebral hypoperfusion? Why were so many patients excluded from the study? Was the study cohort largely represented by patients in whom there was uncertainty? Indeed, was the uncertainty principle helpful in this regard? The Cooperative Study addressed only a part of contemporary surgical practice and failed to observe the basic physiological and clinical indications for the operation. A negative result was inevitable.

Some clinical doors remained open. Pioneers of selective EC-IC procedures have accumulated supporting and compelling information addressing a cohort of patients who attract high risks for future stroke, risks that reflect a chronic state of low CBF.^{9–11,17,18,19,20,21} Of the variety of techniques used to measure CBF, observations of baseline values are generally unhelpful. Activated studies measuring relative increases in CBF have greater predictive values for identifying those at risk.^17–21 Hence a number of cerebrovascular reactivity indices have evolved which, although different, point in the same direction—impaired cerebrovascular reactivity is a risk factor for future strokes. Those with negative reactivity on acetazolamide activated xenon computed tomography (CT) have a 12-fold increase in incidence of stroke,²² and those with an increased oxygen extraction fraction indicating exhausted cerebrovascular reserve (measured using positron emission tomography (PET)) have an increased stroke risk of 30%.^23,24 Likewise, severely impaired CBF reactivity to carbon dioxide (measured using transcranial Doppler flowmetry) also increases the risk by approximately 30%.^18,25 Two ongoing randomised trials have adopted different physiological criteria upon which to base surgery. The North American Carotid Occlusion Surgery Study (COSS)²⁶ is using PET derived oxygen extraction fraction, whereas the Japanese Extracranial to Intracranial Bypass Trial (JET)²⁷ is using acetazolamide CBF activation measured with either single photon emission CT or xenon CT imaging. JET has already demonstrated significant short term gains for the surgical group, and the imaging modalities adopted have the obvious advantage of far wider clinical availability than PET derived indices.

The key questions remaining are:

• What is the threshold for cerebral perfusion that defines the patient at risk?

• Which technique for CBF measurement should be used?

• Which cerebrovascular indices have the greatest sensitivity for prediction of stroke?

• What are the absolute benefits of surgery?

• What surgical procedure is best?

• What rates for poor surgical outcome can be accepted?

Without a universally accepted measurement for CBF, and without adopting a standardised surgical approach, gains for cerebral revascularisation will be dilute, and clinical benefits hard to identify. Similar concerns have already muddied the waters for treatment of other cerebrovascular conditions. Thus in the same way that carotid endarterectomy needs to meet vigorously defined surgical standards,²⁸ the same will be true for EC-IC bypass. Indeed, more so in the face of a technically far more demanding operation. Low flow bypasses can normalise an abnormal oxygen extraction fraction and regional CBF,^29–33 and reverse neurocognitive deficits.³⁴ Hence the technically more demanding high flow bypass (using a saphenous vein or radial artery graft) seems unnecessary when considering an operation to provide CBF augmentation. Less conventional methods for cerebral revascularisation, such as the use of vascular tissues applied directly to the cerebral cortex, require similar physiological and clinical scrutiny before their use can be justified.

CONCLUSION

With relatively small numbers of suitable cases, centralisation of expertise would be important for achieving a measure of surgical quality control. Once broad criteria for offering cerebral revascularisation have been accepted and parallel surgical standards defined and met, the thresholds for surgery in individual centres can be developed further. Adoption of a multidisciplinary team approach assists in accurate patient assessment and selection and management of vascular risk factors, and provides an environment for testing novel applications for revascularisation surgery.