Collateral flow is an important feature of cerebral circulation when major arteries are occluded, which varies depending on individuals. Collateral flow in acute ischemic stroke is known as a predictor of acute thrombolytic and endovascular therapy outcome as well as long-term prognosis (1–5). Patients with a cervical large artery atherosclerosis (LAA) have a more extensive cerebral collateral circulation and a better functional outcome at 90 days than those with a cardioembolic (CE) stroke (6). Previous studies have also suggested that LAA stroke is associated with better collateral flow than CE stroke (6–9). Theoretically, LAA in humans develops over decades, which might promote the gradual development of cerebral collateral flow. For example, patients with a severe (71–99%) stenosis have a better collateral flow than those with a moderate (51–70%) stenosis (6). In contrast, since CE stroke is not accompanied by chronic cerebral hypoperfusion, the chances of collateral artery formation and recruitment are less likely to occur in these patients.

Although the importance of collateral flow has attracted more attention over the past decade, studies on the factors which determine collateral flow are not well understood. The potential relationship of collateral flow with history of cardiovascular risk factors, such as hypertension, congestive heart failure, hyperlipidemia, and diabetes might exhibit complex interactions. Calleja et al. have reported that diabetes is significantly associated with poorer collateral flow on admission (1).A history of hypertension has also been more frequently found among patients with poor collateral flow (2, 10, 11). Menon et al. have found that metabolic syndrome, hyperuricemia, and age are associated with poor collateral flow in patients with acute ischemic stroke (12). In case of coronary collateral flow, eGFR is an independent affecting factor (13).

We hypothesize that factors related to collateral flow in patients with acute ischemic stroke caused by large vessel occlusion (AIS-LVO) are different according to stroke etiology. In this study, we investigated the extensive physiologic, laboratory, and imaging parameters which have been reported to be possibly related with collateral flow in patients with AIS-LVO. We further analyzed the results according to stroke etiology with the assumption that stroke subtype may influence the collateral development.

In this study, we identified that elevated initial SBP was associated with poor cerebral collateral flow and more severe symptoms in the CE group. In the contrary, these associations with SBP were not observed in the LAA group. To the best of our knowledge, these results are novel findings. With the correlation analysis, initial SBP elevation was associated with poor mCTA score and higher NIHSS score in the CE group. According to our regression results, poor collateral flow can be predicted with initial SBP in the CE group, but not in the LAA group. Moreover, poor collateral flow was associated with lower ASPECT score in the CE group. This means that collateral flow is more important for predicting the clinical manifestation and ischemic tissue damage of the CE group. However, the number of patients is too small, so conclusion should be cautious especially in the LAA group.

The patients with AIS-LVO due to LAA had a better collateral flow than CE stroke. The association of stroke subtype and collateral flow has been reported by previous studies (6–8).

Previous studies have also reported elevated mean SBP levels in patients with AIS-LVO showing that patients with LAA had higher initial SBP than patients with CE (10, 16). Kim et al. have found that patients with non-CE have higher baseline SBP than patients with embolic stroke (17). The sustained better collateral flow with blood pressure elevation in the LAA group might be explained by the gradual development of collaterals. Primary collaterals refer to arterial segments of the circle of Willis. In the event of large intracranial artery occlusions, these primary collaterals can provide immediate diversion of cerebral blood flow (CBF) to ischemic regions through existing anastomoses. Secondary collaterals include the ophthalmic artery and leptomeningeal arteries as well as other anastomoses between the distal, small-caliber arteries. Tertiary collaterals refer to newly developed micro vessels through angiogenesis at the periphery of ischemic regions. Meanwhile, secondary collaterals might be anatomically present, although up regulating the capacity of these alternative routes to support CBF might require time (18). Although angiogenesis, sprouting of new capillaries from postcapillary venules, is active during chronic hypoxia (19), arteriogenesis is triggered by fluid shear stress rather than by hypoxia. The growth of cerebral collaterals is correlated with the rising intravascular flow rate. Thus, elevated SBP may increase the intravascular flow rate, which in turn facilitates the development of cerebral collaterals through arteriogenesis. Patients with LAA might have adapted to elevated SBP due to chronic hypoperfusion, resulting in a good collateral flow. Previous studies have found that lower SBP is associated with less hemodynamic improvement during follow-up in patients with atherosclerotic internal carotid artery (ICA) or MCA disease (20). In patients presenting with AIS-LVO with proximal MCA occlusion, initial recruitment of collaterals occurs rapidly (21). But in the CE cases, the event of a large artery occlusion is abrupt and there is no time for the development of tertiary collaterals. Therefore, a limited recruitment of collateral channels and ongoing cerebral edema may cause blood pressure elevation in the CE group to maintain cerebral perfusion pressure. But higher blood pressure can also facilitate further tissue damage through cerebral edema and hemorrhagic transformation (22, 23). In case of futile recanalization after endovascular thrombectomy, cerebral edema volume is increased by same mechanism (24).

Wufuer et al. have reported that lower SBP is associated with good collateral flow in patients with ICA occlusion (11). But our study found no relationship between collateral flow grade in stroke patients with LAA and initial SBP. The inverse association between poor collateral flow and high SBP in the CE group was not prominent in the LAA group. This is because they already have elevated SBP. However, whether the variability of blood pressure affects the collateral flow remains unclear.

In this study, there were differences in the baseline premorbid statin use, HbA1c, eGFR, and echocardiographic parameters between the 2 groups. However, these findings were not associated with the collateral flow grade in multivariable regression analysis. Lee et al. have reported that premorbid use of statin in atrial fibrillation patients is associated with an excellent collateral flow (25). Unlike the previous report, there was no association between collateral flow and premorbid use of statin in our study. This might be due to differences in the baseline characteristics of enrolled patients. They included MCA M1 or M2 occlusion and excluded AIS-LVO due to LAA. Anatomical variations of circle of Willis such as agenesis of ACA and fetal PCA were not associated with the collateral flow grade in our study. A good collateral flow is more frequently associated with the development of collaterals in the anterior communicating artery, the posterior communicating artery, and the leptomeningeal collateral (11). Up to now, there is no confirmative conclusion regarding the effect of collaterals via the circle of Willis (26).

The strength of our study is that included patients were exclusively M1 total occlusion with mCTA identified within 24 h (263 ± 392 min), which can analyze the collateral flow in a strictly controlled environment. Second, all the echocardiographic parameters were investigated which might affect cardiac output and cerebral blood flow. Third, as for the strategy of acute stroke treatment, out data may have clinical implication on acute stroke management. Lowering blood pressure in LAA may disturb physiologic compensation of cerebral collateral flow especially at higher SBP range. In the contrary, CE stroke with high initial SBP has less possibility of having good collateral flow. Therefore, it is reasonable to perform early and rapid recanalization during hyperacute period and monitor the development cerebral edema during later period in these patients.

Our study has several limitations, mostly inherent to its retrospective design and relatively small sample size in a single center. Multicenter trials and larger-scale longitudinal study are warranted to examine whether blood pressure contributes to collateral flow. Second, this study examining blood pressure and collateral flow was cross-sectional, thus allowing us to detect associations without formulating a causal link. Third, the majority of patients with CE stroke etiology in our study had atrial fibrillation. Atrial fibrillation might be relatively more prevalent than other CE sources in patients with LVO. However, other major CE stroke etiology can also cause abrupt vessel occlusion, which is basically different from the mechanism of LAA. Lastly, although it is generally considered that digital subtraction angiography (DSA) is the golden standard for collateral evaluation. mCTA is a quick, easy and non-invasive method associated with clinical outcome, which are merits over DSA (27).

In conclusion, poor cerebral collateral flow in AIS-LVO was associated with elevated initial SBP. The association of collateral flow with initial SBP was significant in CE group, but not in LAA group. However, the number of patients is too small, so conclusion should be cautious especially in LAA group. Further study is needed to understand the mechanisms related to differential effects of SBP on cerebral collateral flow by stroke subtypes.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the Ethical Committee of the Samsung Seoul Medical Center. The patients/participants provided their written informed consent to participate in this study.

JS established the study concept, analyzed and interpreted the data, and wrote the manuscript. J-WC, W-KS, and OB established study database and made critical revision of the manuscript. G-MK established the study idea and database and made critical revision of the manuscript with intellectual contents. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.