Intervention versus standard medical treatment in patients with symptomatic occlusion of the internal carotid artery: a randomised oxygen-15 PET study

In patients with transient ischaemic attack (TIA) or minor disabling ischaemic stroke associated with an occlusion of the internal carotid artery (ICA), the risk of recurrent stroke has been reported 5% to 6% per year [1]. This risk may be as high as 12% per year in case of a demonstrated compromised flow to the brain [2–4]. In these ‘high-risk’ patients, revascularization surgery may be considered, but evidence that interventional treatment reduce the risk of stroke in comparison with the standard treatment consisting of antithrombotic medication and control of vascular risk factors is lacking.

A recent study has shown that increased oxygen extraction fraction (OEF) measured by positron emission tomography (PET) is still an independent predictor of subsequent stroke under current medical treatment [5]. However, the Carotid Occlusion Surgery Study (COSS) included patients with symptomatic ICA occlusion and poor cerebral haemodynamics as demonstrated by increased OEF measured by PET and found no benefit of extracranial/intracranial (EC/IC) bypass surgery [4, 6]. This trial was stopped prematurely as the 2-year risk of ipsilateral stroke was between 20% and 25% in both surgical and non-surgical groups (p = .78). In this study, the proportion of patients who had a perioperative stroke (within 30 days after surgery) was 15% [4].

Other therapeutic strategies in patients with symptomatic ICA occlusion aiming to improve cerebral perfusion and reducing the risk of recurrent stroke are endarterectomy or endovascular treatment of a significant stenosis in one of the cerebropetal arteries that are part of collateral pathways [2, 7–10], or tapering of antihypertensive medication [11, 12].

These procedures have lower perioperative risks than an EC/IC bypass operation, but so far, their clinical effect has not been compared to standard treatment. Whether the application of these therapeutic strategies improves cerebral perfusion as compared with standard medical treatment is unclear.

The aim of this randomised pilot study was to investigate the haemodynamic effects as measured by oxygen-15 PET of endarterectomy or endovascular treatment of stenosed cerebropetal arteries, or tapering of antihypertensive medication in comparison with the standard medical treatment alone in patients with a symptomatic ICA occlusion.

Complete PET data sets were obtained for 23 patients (mean age 64 ± 11 years; 19 men). Of 41 eligible patients, 14 failed to complete the initial PET study and were excluded and four others did not complete the PET study after 3 months (Figure 1). Twelve patients were assigned to intervention (eight stenting or surgery and four tapering of antihypertensive medication) and 11 to standard treatment. Clinical characteristics and presence of stenosis in cerebropetal arteries and collateral pathways are shown in Table 1. Twenty-two patients had had recurrent ischaemic symptoms after the ICA occlusion had been demonstrated, but before inclusion in the study. The time between last symptoms and first PET scan was on average 38 ± 23 days in both treatment groups. At baseline, ten patients were characterized as haemodynamic stage 0, 11 as stage 1 and two as stage 2. Figure 2 shows parametric images of a typical patient with stage 2 haemodynamic failure. Patients with a symptomatic ICA occlusion had a mean CBF in the ipsilateral MCA territory of 32.2 ± 5.6, which was significantly lower (mean difference −5.2 mL/min/100 mL, 95% CI −9.8 to −0.6) than that in healthy subjects (mean CBF MCA territory of 37.4 ± 3.2) [18].

The mean time between PET scans at baseline and follow-up was 98 ± 10 (range 77 to 119) days. All patients were treated according to their randomisation. Randomisation occurred on average 5 ± 9 days after the baseline PET scan. Of the 12 patients assigned to intervention (Figure 1), eight patients underwent stenting or surgery after a mean time of 15 ± 10 (range 1 to 29) days following the baseline PET scan. In four patients, antihypertensive medication was tapered, which resulted in a mean increase in systolic blood pressure of 38 ± 9 mmHg. Intervention was complicated by a non-fatal myocardial infarction in one patient who underwent CEA of a contralateral ICA stenosis; none of the procedures were complicated by a stroke or hyperperfusion syndrome.

Mean values of CBF, CBV and OEF were similar at baseline and after 3 months, both in patients treated with intervention and in those assigned to standard treatment (Table 2), but CBF changes varied in individual patients (Figure 3). No differences in absolute CBF, CBV and OEF changes after 3 months were observed between patients assigned to intervention and those assigned to standard treatment (Table 2).

		Intervention n = 12	Standard n = 11	Unadjusted difference in change (95% CI)	Adjusted difference in change *(95% CI)
		Mean ± SD	Mean ± SD
CBF MCA (mL/min/100 mL)	Baseline	31.3 ± 5.9	33.1 ± 5.4
	3 months	31.6 ± 4.9	33.5 ± 4.7
	Change	0.34 ± 6.2	0.42 ± 4.1	−0.1 (−4.7 to 4.5)	−1.0 (−7.4 to 5.4)
CBF ACA (mL/min/100 mL)	Baseline	32.6 ± 5.8	35.2 ± 5.5
	3 months	32.0 ± 4.2	35.0 ± 5.7
	Change	−0.56 ± 6.2	−0.28 ± 4.9	−0.3 (−5.2 to 4.6)	−1.6 (−8.4 to 5.2)
CBV MCA (mL/100 mL)	Baseline	3.0 ± 0.6	3.3 ± 0.9
	3 months	3.5 ± 1.4	3.2 ± 0.6
	Change	0.60 ± 1.7	−0.09 ± 1.2	0.7 (−0.6 to 2.0)	0.09 (−1.7 to 1.9)
CBV ACA (mL/100 mL)	Baseline	3.0 ± 0.6	3.4 ± 1.0
	3 months	3.5 ± 1.3	3.1 ± 0.7
	Change	0.49 ± 1.6	−0.23 ± 1.3	0.7 (−0.6 to 2.0)	0.08 (−1.6 to 1.8)
OEF MCA (%)	Baseline	46.4 ± 9.7	47.5 ± 5.5
	3 months	46.3 ± 7.9	48.4 ± 5.8
	Change	−0.08 ± 6.1	0.85 ± 6.8	−0.9 (−6.5 to 4.7)	3.1 (−4.1 to 10.3)
OEF ACA (%)	Baseline	44.4 ± 11.2	46.1 ± 6.2
	3 months	44.2 ± 7.3	47.2 ± 5.2
	Change	−0.17 ± 7.4	1.1 ± 6.3	−1.3 (−7.3 to 4.7)	3.1 (−4.6 to 10.8)

Asterisk ‘*’ indicates adjusted for age and sex; measurements were performed ipsilateral to the side of the ICA occlusion; CBF, cerebral blood flow; CBV, cerebral blood volume; OEF, oxygen extraction fraction; MCA, middle cerebral artery; ACA, anterior cerebral artery.

The eight patients treated with surgery or stenting of a stenosis in a collateral artery had a mean CBF of 30.8 ± 5.4 mL/min/100 mL at baseline and 31.8 ± 5.2 mL/min/100 mL after 3 months (mean difference 0.97, 95% CI −4.6 to 6.6). In the 11 patients who received standard medical treatment alone, mean CBF was 33.1 ± 5.4 mL/min/100 mL at baseline and 33.5 ± 4.7 mL/min/100 mL after 3 months (mean difference 0.42, 95% CI −2.4 to 3.2). The change in CBF did not differ between these two groups of patients (adjusted mean difference of change −0.02 mL/min/100 mL, 95% CI −7.2 to 7.1).

This study shows that patients with recent symptoms of cerebral ischaemia associated with occlusion of the ICA on average have a lower CBF in the hemisphere ipsilateral to the occlusion than healthy subjects. Oxygen-15 PET showed that interventions aimed at improving cerebral perfusion caused a borderline significant increase in CBF in patients who had an impaired haemodynamic state of the brain at presentation, but improvement in CBF was not different between interventional and standard treatment.

Current knowledge of cerebral haemodynamic and metabolic changes over time in medically treated patients with an ICA occlusion is based on small patient series, and both improvement and deterioration over time have been described. In one study, ten medically treated patients with symptomatic ICA occlusion and increased OEF underwent a second PET scan after 12 to 59 months [22]. Over time, patients showed on average an increase in CBF hemispheric (ipsilateral/contralateral) ratio and a decrease in OEF hemispheric ratio, but without changes in absolute values of CBF and OEF [22]. Another study of seven patients with symptomatic ICA occlusion and normal CBF and OEF at baseline found a significant decrease in CBF and increase in OEF after 24 to 64 months, implicating haemodynamic deterioration over time [23]. Other studies reported a bilateral increase in cerebrovascular reactivity measured by Transcranial Doppler (TCD) after CEA of a contralateral carotid stenosis in the presence of a symptomatic or asymptomatic ICA occlusion [9, 10, 24]. In contrast to these studies, we did not find a significant effect of treatment aimed at improving the flow state of the brain when compared to control patients in a randomised design. A possible explanation could be that the cerebral haemodynamic state also improved without intervention. Patients with haemodynamic compromise at baseline who received standard treatment showed a small non-significant increase in CBF. It has been shown before that patients with ICA occlusion and a compromised flow state of the brain may improve spontaneously over time [25]. In addition, in patients with a normal haemodynamic state of the brain (stage 0) at baseline who underwent an intervention, particularly those treated by tapering of antihypertensive medication, a further increase in CBF may not be expected because of preserved cerebral autoregulation.

A limitation of our randomised pilot study is that we did not select patients on the basis of demonstrated cerebral haemodynamic compromise. The design of the study was aimed at including patients at high risk of recurrent stroke by selection on clinical criteria: ischaemic symptoms of the brain after documented occlusion, excluding patients with eye symptoms alone and those with symptoms only before the ICA occlusion were documented without ongoing symptoms thereafter [26]. We had expected that these clinical criteria would have resulted in a much larger proportion of patients with haemodynamic stages 1 and 2. Because the average time between the last symptoms of patients and the baseline PET scan was 38 days, spontaneous improvement of a compromised flow state of the brain may have occurred before the PET scan took place. This may have contributed to the lack of a difference between the PET scan at baseline and after 3 months.

Another limitation might be heterogeneity in treating patients assigned to intervention. We have chosen for this randomization scheme rather than designing separate studies for each of the treatment options because all treatment options aim to improve cerebral perfusion. However, effects of tapering of antihypertensive medication on the haemodynamic state of the brain may well be different from those of surgery or stenting of stenosed arteries that may be important for collateral blood supply. Subgroup analysis of the present data comparing surgery or stenting of a stenosis in the contralateral ICA, ipsilateral ECA or VA with standard medical treatment alone did not show significant differences in CBF, but the number of patients in each group was small. Likewise, effects of surgery or stenting of the contralateral ICA on cerebral perfusion may be different than those of surgery or stenting of the ipsilateral ECA or VA. Furthermore, it has been described that both endarterectomy and stenting can be complicated by a hyperperfusion syndrome [27], but in none of the patients, this complication occurred.

The third limitation is that a relatively high number of patients had to be excluded because they did not have two successfully performed PET studies. Our failure rate was similar to the rate of 22% for quantitative studies in the St Louis Study [3, 6], and lower than in another PET study that reported a failure rate of 41% [28]. Finally, the sample size in our study was probably too small to detect differences between the two treatment strategies except when they were large.

Although the results of this pilot study are not conclusive, some important lessons can be learned. Patients with TIA or ischaemic stroke in the hemisphere ipsilateral to the ICA occlusion are at risk of subsequent cerebral ischaemia. In these patients with symptomatic ICA occlusion, the clinical benefit of CEA or endovascular treatment of an additional stenosis in the cerebropetal arteries or tapering of antihypertensive medication has not been shown by a randomised trial. The recently published results of the Carotid Occlusion Surgery Study randomised trial [4] that showed no benefit of STA-MCA bypass surgery over medical treatment despite improvement of cerebral haemodynamics in the surgical group reinforce the need for other therapies to improve outcome in patients with ICA occlusion at high risk of recurrent stroke. Oxygen-15 PET is useful to identify a subgroup of patients with haemodynamic compromise, but we could not demonstrate that stenting or surgery of stenosed cerebropetal arteries or tapering of antihypertensive medication improved the haemodynamic state of the brain to a large extent in comparison with standard treatment. Whether revascularization of collaterals or tapering of antihypertensive medication can reduce the risk of recurrent stroke can only be determined in a large randomised trial of patients with a symptomatic ICA occlusion at high risk of stroke.