Cerebral Revascularization After the Carotid Occlusion Surgery Study

What Candidates Remain, and Can We Do Better?

Timothy G. White, MD; Hussam Abou-Al-Shaar, MD; Jung Park, MD, PhD; Jeffrey Katz, MD; David J. Langer, MD; Amir R. Dehdashti, MD


Neurosurg Focus. 2019;46(2) 

In This Article


Prior randomized controlled trials have demonstrated no benefit to surgical revascularization in patients with symptomatic cerebrovascular atherosclerotic steno-occlusive disease.[1,10] However, optimal medical management, which includes optimization of atherosclerotic and stroke risk factors as well as DAPT, still fails to prevent strokes in a small number of patients. These patients may present with perfusion-dependent examinations or recurrent TIAs and/or strokes.[4,11] No randomized controlled trial has investigated the role of bypass in this patient group. Although the COSS did stratify patient risk by using hemodynamic imaging modalities, it did not specifically investigate patients in whom best medical therapy failed.[4] The purpose of this retrospective analysis was to demonstrate the safety and perioperative risks associated with revascularization in theoretically high-risk patients in whom medical management has failed and who showed definitive impaired cerebrovascular reserve. Here we demonstrate that when done at a high-volume cerebrovascular center with specified neuroanesthesia, this procedure can be done with a reasonable level of morbidity and may represent an effective salvage procedure in certain patients. In this series perioperative ischemic events occurred in 9% of patients (3/35), with only 1 patient developing an irreversible neurological deficit.

The original 1985 extracranial-intracranial bypass trial randomized 137 patients with symptomatic carotid or intracranial steno-occlusive disease but failed to stratify patients by risk prior to randomization.[7] Ultimately, no benefit was found with surgical intervention at an average follow-up of 55.8 months, with stroke rates in the medical arm of 29% versus 31% in the surgical arm. Following the publication of these results, validated criteria for cerebral hemodynamic impairment were developed via the STLCOS.[3,5] At that time, studies on bypass hemodynamics demonstrated that revascularization was able to correct impaired hemodynamics in patients.[11] That being said, the increased OEF used to randomize patients in the COSS may not have been an adequate predictor of hemodynamic impairment.[2]

The STLCOS established that PET OEF was able to predict stroke risk in patients with symptomatic cerebrovascular steno-occlusive disease and stage II hemodynamic impairment. In patients with misery perfusion and increased OEF, the stroke risk approached 26.5% at 2 years, compared to 5.6% in patients without impaired perfusion, which led to a predicted benefit of 40% when developing the COSS and determined patient number to power the study.[3,5,7] However, in the medical arm of the COSS the stroke risk amounted to 22% over the 2-year study period, with the medical arm faring far better than predicted.[10] Given that the predicted stroke risk used for the COSS was based on studies done in the 1990s, further studies have investigated whether, with newer medical therapy, misery perfusion still predicts stroke risk.[12] More recently, Yamauchi et al. demonstrated that misery perfusion on PET OEF still predicts increased stroke risk in patients with symptomatic cerebrovascular steno-occlusive disease.[13] The authors established that even with the most recent optimal medical management, increased OEF alone (as used in STLCOS and COSS) leaves patients at increased risk of stroke. Therefore, stricter management of these patients is required, and there may still be a role for revascularization in these highest-risk patients. In this study we mainly used SPECT imaging combined with a Diamox challenge in order to perform risk stratification in patients with hypoperfusion and to determine which patients have decreased cerebrovascular reserve.

Ultimately, it was not the goal of this series to validate diagnostic criteria. However, all patients were selected to be possible candidates for bypass based on their clinical presentation combined with hemodynamic imaging in the form of SPECT with and without Diamox, QMRA NOVA, MR perfusion, or CT perfusion. Therefore, further studies would be justified to define the optimal modality to determine impaired cerebral hemodynamics and define misery perfusion.

The outcomes of the COSS were driven by two main factors: the impressively good outcomes of the medically managed patients and the perioperative morbidity of revascularization surgery. Perioperative stroke risk in the COSS was 14%, which was similar to that of the original extracranial-intracranial bypass trial, and these perioperative ischemic events made up the majority of stroke risk in this study group.[10] Once past the initial perioperative period the surgical patients maintained a significantly lower stroke risk than those in the medical arm. Therefore, decreasing perioperative morbidity and ensuring operative safety is important going forward. In this series we demonstrate that at a high-volume center with specialized neuroanesthesia, perioperative risk in high-risk patients was maintained at less than 10%. Analysis of the COSS data predicted that to show benefit, surgical risk would need to be roughly 50% of what it was. That being said, this patient population should be considered somewhat separate from the COSS population because, for reasons that have been previously discussed, optimal medical therapy had failed.

Indeed, although the indications for cerebral revascularization for patients with symptomatic steno-occlusive cerebrovascular disease have waned after the publication of the COSS, they have not gone away altogether. However, literature addressing the individual patient populations that may benefit the most has waned as well. This series represents a unique assessment of perioperative risk in high-risk patients not clearly addressed by the COSS. Clearly, further research into this patient group is warranted in order to assess the risk of further stroke after initial therapy as well as to determine optimal treatment modalities.