In intracranial arterial dissections (IAD), the evidence with the use of antithrombotic agents is limited (1). For its counterpart, cervical arterial dissections (CAD), the Cervical Artery Dissection in Stroke Study (CADISS) failed to show differences in ipsilateral stroke or death, and angiographic recanalization rates between anticoagulation (AC) and antiplatelets (AP) (2). In addition, a more recent randomized trial (TREAT-CAD) did not show non-inferiority of aspirin to vitamin K antagonists (3). Based on the results from CADISS and TREAT-CAD, the use of aspirin as the standard therapy over anticoagulation in CAD patients is weak (1), and both anticoagulants and antiplatelets have been prescribed (4).

While the evidence for use of antithrombotic agents in IAD is limited, experts feel that AP have a better risk/benefit ratio over AC due to risk for subarachnoid hemorrhage (SAH) (4). In cases where endovascular reconstructive or deconstructive therapy (5) is planned to prevent rupture of dissecting aneurysms, AP would also be preferred. However, clinicians may still prefer to use AC to prevent embolization from fresh thrombus (2) and to promote recanalization or arterial healing. Evidence regarding such questions is lacking.

This study aimed to compare the differences in outcomes between antithrombotic modalities (AC vs. AP therapy) in terms of combined hemorrhagic and ischemic clinical outcomes and arterial outcomes.

Patients that presented with symptoms due to acute intracranial arterial dissections and used antithrombotics were enrolled from an institutional registry of cervicocephalic dissections (6). The diagnosis of intracranial dissections was based on the presence of below imaging findings (7–9) involving the intracranial arteries; luminal pearl and string sign (stenosis and dilatation); luminal stenosis with intimal flap/double lumen, or intramural hematoma (at-suppression T1-weighted MR or MR angiogram source images); luminal fusiform aneurysmal dilatation of the arterial trunk not located at an arterial branching point; luminal occlusion with visible intimal flap/double lumen, or associated with a pearl-and-string sign. From this registry, patients who presented between 2010 and 2021 with onset of symptoms to presentation within 31 days, diagnosed with IAD without SAH, and prescribed antithrombotics for its treatment were selected. Patients were excluded if (1) they underwent emergent endovascular or surgical reperfusion, (2) did not use anti-thrombotic medication during hospitalization and at discharge due to any reason; (3) underwent hemicraniectomy or suffered significant hemorrhagic transformation limiting use of antithrombotics, or (4) were transferred to another hospital (Figure 1). Ethics approval was obtained from the Ajou University Hospital International Review Board (AJOUIRB-MDB-2021-674), and the study was performed in accordance with the ethical standards of the 1964 Declaration of Helsinki and its later amendments. The board waived the need for obtaining patient consent due to retrospective study nature.

The current study results showed no differences in clinical outcomes evaluated as a 3-month composite primary endpoint between AC and AP for symptomatic unruptured IAD patients. A higher rate of early and late endovascular reconstructive/deconstructive treatment in the AP group was seen, but it was not supported by differences in early and late vascular outcomes, most likely representing predilection to use AP before endovascular repair. However, crossover from AC to AP was not infrequent, and early recanalization and aneurysmal changes were more common in the crossover group, likely resulting in the crossover of antithrombotic regimen. While the best antithrombotic regimen is still unclear and needs future studies, tailored antithrombotic regimens may be used for IAD with serial angiographic imaging to achieve optimal results.

In detailed analysis of primary endpoints, there were no differences in ischemic stroke related outcomes according to antithrombotic choice. Both the overall rate of ischemic END (5.8%) and ischemic stroke (3.5%) were low in the current study, and comparable to randomized controlled trials of CAD, which failed to show differences in outcomes between AC and AP. The CADISS trial recruited patients in whom ischemic stroke was present in 77.6%, and reported a 5.2% rate of any stroke or TIA at 1 year, and 2.4% rate of ipsilateral stroke at 1 year (2). The TREAT-CAD trial recruited patients in whom ischemic stroke was present in 52.1%, and reported a 3.6% rate of ischemic stroke at 3 months (3). To our knowledge, the current study used the largest dataset to compare the results of AC and AP therapy for IAD. While the current registry is retrospective in design and underpowered due to low ischemic events to draw conclusions regarding antithrombotics and ischemic outcomes, based on the current results, it is likely that the absolute differential effect of AC and AP on ischemic outcome would be very low.

In unruptured IAD, there is risk for subsequent rupture and SAH unlike CAD, which can be devastating. In this study, apart from a higher rate of early and late endovascular reconstructive/deconstructive treatment performed in the AP group, there were no differences in arterial changes, or new SAH or ICH according to antithrombotics. The higher rate of endovascular reconstructive/deconstructive treatment in the AP group is reasonable, as antiplatelets may have been used prior to flow diverter stents or stent-assisted coiling. Two cases of SAH occurred during admission, showing that rupture of dissecting aneurysms predominate in the early phase of dissections (18). Furthermore, their location of dissection was V4 segment of the vertebral artery and the intracranial ICA, possibly showing predilection for rupture to occur in areas of transition from cervical to intracranial arteries occur (7). Apart from this, subsequent hemorrhage was rare. It should be noted that while early aneurysmal changes were overall numerically more frequent, late aneurysmal changes were overall not uncommon, demonstrating the importance of repeat imaging.

Considering the devastating nature of aneurysmal rupture, for unruptured IADs that do not present with ischemia, thrombotic risk should be stratified, and AC, or possibly antithrombotics, may be avoided in populations with low thrombotic risk. One example could be pure dilatation morphologies of dissections. It is generally accepted that steno-occlusive luminal morphology (8) is associated with ischemic risk in IAD. However, as a recent study claimed that stenotic segments more frequently occur in ruptured IVAD (19), luminal morphology-based decisions may not be sufficient. We have recently shown that higher arterial pulse wave velocity is associated with ischemia (6). Another study showed that a proximal dominant intramural hematoma in comparison to a distal one is associated with ischemic stroke (20). Individualized antithrombotics selection may be tailored based on such knowledge. Antithrombotics may be also selected according to whether the intimal flap ruptures and whether flow patency is maintained. Dissections with intact intima with or without limited flow patency were reported to more frequently present with ischemic symptoms, while healing was more common (21). Such findings may guide antithrombotic therapy, while these findings have to be further confirmed in IAD.

While current guidelines advocate the use of AP over AC in IAD due to better risk/benefit ratio (4), the current results show that AC may be safely used according to the clinician’s discretion and considering the potential advantages of each modality. While the treatment policy analysis shows no difference between AC and AP in rates of early and late arterial changes other than endovascular reconstructive/deconstructive treatment, crossovers from AC to AP were not uncommon. There was a higher rate of early recanalization and aneurysmal changes in the crossover group, which was likely the cause of crossover. There may be situations in which a specific regimen is preferred in a case-by-case basis. First, AC may be more preferable in patients long dissecting segments or dissecting occlusions, possibly able to cause worse outcomes due to perfusion failure (22). Endovascular treatment may also be limited by risk of intraprocedural arterial rupture in such patients (23). Recanalization through AC may be helpful for these patients because there still may be minimal arterial flow, sparing the critical penumbra slower than embolic occlusions. Second, AP may be preferable to AC if there is risk for aneurysmal enlargement. We have experienced patients that underwent early aneurysmal enlargement while under AC, and switching to AP resulted in its regression. Supplementary Figure 2A shows an example of successful regression of the aneurysmal segment after switching to antiplatelet agents. If anticoagulation is used, high-resolution arterial imaging may aid arterial follow-up because it can visualize potential arterial positive remodeling and aneurysmal changes that may not be visualized due to thrombosed pseudolumen (24). In some cases, oral AC may aid late arterial healing of stenosis of arterial dissecting segments (Supplementary Figure 2B).

The current study is inherently limited by the retrospective study design. First, baseline differences were observed between those with AC and AP, with higher number of ischemic stroke and TIA presentations, steno-occlusive arterial morphology, and vascular risk factor in the AC group. In some patient that presented with unruptured VAD, antiplatelets may have seen used as a premedication prior to flow diverter stenting or stent-assisted coiling (25, 26). Given this, the subsequent risk of ischemia may be biased to be higher in the AC group. Propensity score matching was performed in this regard. Patients that underwent mechanical thrombectomy or those that did not use antithrombotics were also excluded from the current study, as these patients would usually present with large infarcts and clinicians tend to avoid using AC in this population. After excluding such patients and propensity score matching, there were no differences in composite clinical outcomes between AC and AP groups. Second, the primary composite clinical outcome is prone to bias, as patients were treated to reduce hemorrhagic complications or ischemic complications in an individual basis. Thus, the differences in treatment outcomes according to thrombotic regimens are expected to decay. The high rate of transition in medical therapies is an example. Meanwhile, we believe that the current study results can guide clinicians through pros and cons of individualized antithrombotic therapy for IAD. Third, the description of arterial outcomes may be rather subjective, as various angiographic modalities were used. The outcome arterial healing in late vascular outcomes may be subjective, as it encompasses a wide range of arterial improvements such as normalization to mild improvements in luminal diameter. Small changes in appearance may be affected by different gantry angles and slice thicknesses. In the current study, as aneurysmal enlargement or recanalization were separately evaluated, arterial healing was used as a more inclusive terminology to represent the overall direction of vascular repair processes.

In conclusion, in an observational study of symptomatic IAD patients, there were no differences in composite clinical outcomes between AC and AP. Crossovers from AC to AP was common and associated with early vascular changes. Both AC and AP may be tailored based on clinician’s decision and repeat arterial imaging.