OVERALL GOAL
To assess the relationship between cerebral hemodynamics and cognitive function in patients undergoing treatment for unilateral carotid artery occlusion with extracranial-intracranial (EC-IC) bypass in the Carotid Occlusion Surgery Study (COSS).

SPECIFIC AIM
To determine whether patients who have symptomatic, unilateral carotid artery occlusion and stage II cerebral hemodynamic failure as measured by increased oxygen extraction fraction (OEF) on positron emission tomography (PET) will have better cognitive function following EC-IC bypass compared with an identical group of patients treated with best medical therapy alone.

PRIMARY HYPOTHESIS
Among COSS patients who survive stroke-free after 2 years, hemisphere-specific and global neurocognitive function will be significantly better among patients who have undergone EC-IC bypass than those in the medically-treated group. 

SECONDARY HYPOTHESES
1. Among COSS patients who survive stroke-free after 2 years, the subset of surgical patients with normal OEF as documented by the COSS 1-month post-operative PET will demonstrate greater improvements in neurocognitive function than those in the medically-treated group.

2. At the 2-year time point, cognitive change will correlate with change in disability and quality of life measures.

BACKGROUND
Cognitive impairment, cerebral hypoperfusion, and relevance to COSS.

The notion that cerebral hypoperfusion due to carotid artery disease can produce cognitive impairment was first proposed by Dr. C. Miller Fisher in 1951, based on a necropsy case of a man with a progressive dementia who had bilateral carotid artery occlusion (Fisher 1951).  Based on this and other cases, Fisher postulated that both unilateral and bilateral carotid occlusion can produce a dementia state, and proposed for the first time that restoration of blood supply could potentially reverse the condition.  Between 1951 and 1985, when the multi-center EC/IC bypass collaborative study group published a failure of the bypass procedure to reduce stroke risk (1985), many reports appeared in the cerebrovascular literature attempting to correlate carotid occlusive disease with cognitive impairment.

Most evidence from that period suggested that cognitive impairment can occur with carotid disease, independent of cerebral infarction, and regardless of whether there is unilateral or bilateral disease.  (Bakker, Klijn et al. 2000; Rao 2001)  Yamauchi showed that among patients with unilateral carotid disease both right and left hemisphere functions were impaired(Yamauchi, Fukuyama et al. 1996). The possibility that surgical revascularization could improve cognitive function was approached only in the form of non-randomized case series, however. No randomized clinical trial has ever addressed the question. Nielson et al, for example, found that 23 of 33 patients undergoing EC-IC bypass had impaired neurocognitive scores preoperatively and were improved at 3 months after surgery. (Nielsen, Hojer-Pedersen et al. 1986) In their series, however, only the left-hemisphere injured patients showed impairment at baseline and improvement with treatment. No hemodynamic measures were acquired in that study, and no control population was evaluated. In a similar case series of 38 patients undergoing EC-IC bypass, Drinkwater et al found minor improvements in neuropsychological tests of processing speed and memory but no global improvement. (Drinkwater, Thompson et al. 1984) Again, no cerebral hemodynamic data or control data were provided. Binder et al. compared 12 EC-IC bypass surgery patients with 7 ICA occlusion patients treated medically. Both the surgical and medical groups improved on neurobehavioral measures, but there was no difference between the 2 groups in cognitive status 2 months after surgery. Longer-term outcomes were not reported, and again, no hemodynamics or control groups were included. (Binder, Tanabe et al. 1982) Younkin et al evaluated CBF with a 133Xe inhalation technique in a series of 44 patients undergoing EC-IC bypass. Neuropsychological scores increased at 3 months and further increased at 9 months.(Younkin, Hungerbuhler et al. 1985) Despite the positive long-term improvements in neurocognition, however, CBF was shown to increase to a maximum at 3 months after the operation, then decline to near pre-operative levels by 9 months. The authors proposed that the clinical improvements were due to recovery from stroke deficits and were not explained by improved CBF, a conclusion similarly reached in a study by Meyer et al (Meyer, Lotfi et al. 1990). Finally, a single case was reported by Tsuda et al, who described global cognitive impairment in a patient with an occlusion in the left ICA siphon.(Tsuda Y 1994) Following left EC-IC bypass rCBF measured by 123I-IMP Single Photon Emission Tomography (123I-IMP SPECT) was restored to nearly normal levels over the course of 2 to 9 months. Improvement in cognition occurred over the first 2 months after bypass, and the improvement was sustained for 3 years. In sum, these data are suggestive but not conclusive as to whether cerebral hemodynamic failure and recovery correlates with neurocognitive change.

With the refinement of positron emission tomography (PET) in the 1980s, it became possible to quantify cerebral hemodynamics in terms of oxygen and glucose utilization, and cerebral autoregulatory capacity. (Olsen, Larsen et al. 1983; Sette, Baron et al. 1989; Powers 1991) Unlike methods that measured cerebral blood flow (CBF) alone, the newer PET method allowed functionally important hypoperfusion resulting from carotid occlusive disease to be distinguished from hypoperfusion due simply to reduced metabolic demands of damaged (infarcted) tissue. Using quantitative PET methods, cerebral hemodynamic status could be defined more completely. Current hemodynamic theory holds that as perfusion pressure falls CBF is maintained by autoregulatory vasodilation of cerebral arterioles. When cerebral perfusion pressure falls low enough, as might occur in some cases of severe, chronic ICA stenosis or occlusion, the arterioles maximally dilate. This state is referred to as “exhausted cerebrovascular reserve,” and can be imaged by PET as increased cerebral blood volume relative to CBF (increased CBV/CBF ratio). At this so-called “stage-I” hemodynamic failure, pharmacological vasodilatory challenge with acetazolamide injection or CO2 inhalation fails to dilate the arterioles further and may be associated with increased stroke risk. (Yonas, Smith et al. 1993; Kuroda, Houkin et al. 2001; Markus and Cullinane 2001; Marshall, Rundek et al. 2003) As perfusion pressure falls further, CBF begins to decrease and a metabolic compensation occurs as oxygen extraction fraction (OEF) increases to maintain tissue oxygen metabolism. This “stage II” hemodynamic failure, sometimes referred to as misery perfusion (Baron, Bousser et al. 1981), is thought to represent a vulnerable hemodynamic state in which any further drop in perfusion will produce ischemia. Until recently, what had not been established was whether stage-II hemodynamic failure can be shown to be associated with cognitive impairment and whether rescue from this state can improve or preserve cognitive function. Sasoh et al showed that among 25 patients with mild stroke or TIA, unilateral ICA or MCA occlusion, and increased OEF by PET, verbal IQ, performance IQ and full scale IQ were all impaired before EC-IC bypass. Furthermore, the scores significantly increased after bypass(Sasoh, Ogasawara et al. 2003). Because of the similarities in design to COSS, this study supports the likelihood that the RECON patient population will behave in a similar fashion. RECON now provides the unique opportunity to test these notions in a randomized clinical trial.

Some additional support for the idea that revascularization can improve cognitive status can be derived from studies assessing cognition before and after carotid endarterectomy (CE). In a systematic review of the literature, Lunn et al categorized 16 of 28 studies as demonstrating a benefit of CE for cognitive improvement. (Lunn, Crawley et al. 1999) A majority of the positive studies were conducted before 1984, however, and few analyzed cerebral hemodynamics directly. Recently, Heyer et al showed that patients undergoing endarterectomy improved on some elements of neuropsychometric testing, including executive and motor functions, but decreased or did not improve on others.(Heyer, Adams et al. 1998) No hemodynamic measures were assessed. In another study Hillis et al showed that in patients with high-grade carotid stenosis, regional hypoperfusion was associated with language deficits or hemispatial neglect for left and right hemisphere stroke respectively, and that return to normal MR perfusion patterns after endarterectomy was associated with improved cognition. (Hillis, Wityk et al. 2002). In our own experience, we have presented a case of a 30 year old man with right sided large vessel occlusive disease (Moya Moya) who underwent a revascularization procedure in the right hemisphere. He had developed cognitive decline over many years, including slow processing speed, verbal and visual-spatial memory impairment and visual-spatial neglect. Following right-sided revascularization (frontal Burr holes), his blood flow improved by Single Photon Emission Tomography (SPECT) and performance on neurocognitive tests improved as well, particularly for tasks traditionally associated with right hemisphere function. Figure 1 (below) shows the improvement in copying of the Rey Complex Figure, one of the neuropsychological tests to be included in the battery for this protocol.

Figure 1

Like the EC-IC bypass studies, CE studies have varied widely in the timing of cognitive assessment, the specific neuropsychological tests performed, and the use or absence of control populations.  These variables are crucial to determine whether there is truly a physiological link between cognitive impairment and hemodynamic failure in carotid occlusive disease.  All of these variables can be addressed in the context of this partially blinded, randomized, controlled clinical trial.

The COSS trial.
The COSS trial is an NINDS-sponsored, multicenter, randomized, controlled clinical trial designed to test the hypothesis that surgical anastomosis of the superficial temporal artery to the middle cerebral artery (EC-IC bypass) when added to the best medical therapy can reduce subsequent ipsilateral ischemic stroke at 2 years in patients with symptomatic internal carotid artery (ICA) occlusion and Stage II hemodynamic failure as measured by increased oxygen extraction fraction (OEF) on positron emission tomography (PET). (Grubb 2003) The primary endpoint in the COSS trial is subsequent ipsilateral fatal or non-fatal stroke. Although quality of life and functional scales of activities of daily living are analyzed as secondary endpoints, no formal assessment of cognitive function is made.

As an ancillary study to the COSS trial, RECON (Randomized Evaluation of Carotid Occlusion and Neurocognition) represents a unique opportunity to conduct an adequately-powered, randomized, controlled clinical trial to evaluate the role of hemodynamic failure as an independent determinant of cognitive function in large vessel disease. We hope to determine not only whether chronic cerebral hemodynamic failure is associated with cognitive impairment, but more importantly, whether treatment of hemodynamic failure can improve cognitive function or protect against further cognitive decline. RECON cognitive outcomes will be assessed in patients who have not reached COSS endpoints (ie, no recurrent stroke) so that recurrent stroke will not confound the assessment of cognitive status at the end of the 2-year study period. In addition, because COSS includes a baseline PET scan to determine OEF status, we will know the patients’ hemodynamic status. Finally, cognitive outcome measures will be correlated with the functional scales assessed by COSS to verify the clinical relevance of cognitive change. Positive results in the RECON study would be important for the COSS trial because if COSS fails to show the predicted 40% reduction in recurrent stroke rate at 2 years, the efficacy of the EC/IC bypass may still be demonstrated by virtue of its ability to improve or preserve cognitive function. This would help to establish the important principle that neurological outcomes in carotid artery disease should be considered more that just recurrent stroke. If both COSS and RECON have positive results, COSS will not only be supported, but bypass may be viewed as having a beneficial effect beyond prevention of recurrent stroke, to date the only primary outcome measure used in carotid artery treatment trials. Other revascularization treatments for carotid artery disease such as extracranial and intracranial stent-angioplasty, encephalodural arterial synangiosis (EDAS), and even hypertensive therapy in acute large vessel stroke may then be evaluated for cognitive outcomes as well.