Nonthrombotic internal jugular venous stenosis may facilitate cerebral venous thrombosis

Abstract Aims To explore the effect of nonthrombotic internal jugular venous stenosis (IJVS) exerted on cerebral venous thrombosis (CVT). Methods Patients with imaging confirmed CVT were enrolled into this real‐world case–control study consecutively from January 2018 through April 2021, and were divided into CVT and IJVS‐CVT groups, according to whether or not with non‐thrombotic IJVS. Chi‐square and logistic regression models were utilized for between‐group comparison of thrombotic factors. Results A total of 199 eligible patients entered into final analysis, including 92 cases of CVT and 107 cases of IJVS‐CVT. Chi‐square revealed that thrombophilic conditions were found in majority of CVT, while only minority in the IJVS‐CVT group (83.7% vs. 20.6%, p < 0.001). Multivariate logistic regression indicated that most identified thrombophilia were negatively related to IJVS‐CVT (all p < 0.05), including oral contraceptive use (β = −1.38), hyperhomocysteinemia (β = −1.58), hematology (β = −2.05), protein C/S deficiency (β = −2.28), connective tissue disease (β = −1.18) and infection (β = −2.77). All recruited patients underwent standard anticoagulation, 10 cases in IJVS‐CVT group also received jugular angioplasty for IJVS correction. Most participants obtained alleviations during 1‐year follow‐up. However, both clinical and imaging outcomes in IJVS‐CVT group were not as good as those in CVT group (both p < 0.05). Moreover, 8 cases with CVT and 7 cases with IJVS‐CVT were rehospitalized for CVT recurrences and underwent customized treatment. Conclusion Nonthrombotic IJVS may be one of the risk factors of CVT. Anticoagulation might need to be suggested for IJVS patients.


| INTRODUC TI ON
Cerebral venous thrombosis (CVT) refers to the thrombotic stenoocclusion of dural sinuses and/or cerebral veins, mainly affecting the young and middle-aged adults with a particular predilection for females. 1 Patients with CVT usually present with a wide constellation of nonspecific symptoms, such as headache, blurred vision, tinnitus, etc.
Note that various risk factors may predispose to CVT, including infective or noninfective, and inherited or acquired. 2 Or alternatively, the etiologies that mediated the CVT formation can also be categorized according to the classic Virchow triad, that is, endothelial damage, blood hypercoagulability or hemodynamic changes. 3 Although the increasing availability of modern neuroimaging modalities has advanced the early recognition of CVT with highly decreased mortality, 4 CVT misdiagnosis and missed diagnosis still remain as main issues in clinical settings at present, owing to its nonspecific manifestations and various causes.
Such situations place the patients at an increased risk of long-lasting illness, treatment delay, and even repeated relapse. 5 However, it is of utmost necessity to detect CVT and its origins as well as to explore the potential causes of CVT besides the well-known etiologies.
Notably, one prior study conducted by our team demonstrated that the narrowing internal jugular vein (IJV), detected by color Doppler flow imaging (CDFI), may disturb intracranial venous hemodynamics and therefore promote CVT occurrence. 6 However, as we all know that the values of CDFI may be varied from different operators for its operation dependence. Herein, we aimed to further establish the association between nonthrombotic IJVS and CVT based on neuroradiological imaging evidence to provide references for customized treatment and spark future studies with prospective directions.

| Patient enrollment
Patients with magnetic resonance black-blood thrombus imaging (MRBTI) confirmed CVT were enrolled into this real-world casecontrol study consecutively from January 2018 through April 2021 in Xuanwu Hospital, Capital Medical University. Considering that IJV thrombosis (IJVT) may develop as a result of CVT extension, 5 or perhaps occur spontaneously with CVT as triggered by the same pathology, 7 cases with MRBTI-confirmed IJVT were rejected from this study. All eligible patients signed the informed consents prior to enrollment and data acquisition.
Demographics, clinical and neuroimaging data of all patients were obtained from our inpatient database. The conditions of bilateral IJVs in all patients enrolled were assessed into normal or stenotic conditions, according to the maps of head-and-neck contrast-enhanced magnetic resonance venography (CE-MRV) and/ or computerized tomography venography (CTV). Premised on the results mentioned above, patients were divided into CVT group (isolated CVT without nonthrombotic IJVS) and IJVS-CVT group (CVT combined with unilateral or bilateral nonthrombotic IJVS). Worthy of note is that IJVS is defined as the inner diameter at the stenosis segment less than 50% of the one at the adjacent normal segment, accompanied with abnormal collaterals. 8,9 All patients were strictly selected with the following inclusion and exclusion criteria:

| Clinical data collection and evaluation
Both diplopia and visual field defects were inducted into the category of visual disorders in this study. Ophthalmologists with extensive experience performed color fundus photography for all patients enrolled. The intensity of papilledema was graded using the Frisén scale.
The Frisén value ranges from a minimum score of 0 to a maximum of 5, and the severer papilledema is awarded with the correspondingly higher Frisén score. 10

| Neuroimaging assessment
Imaging spectrums of all patients enrolled were retrospectively reviewed from medical imaging system to validate the nonthrombotic IJVS and preclude the MRBTI-diagnosed IJVT or any conditions that might affect ICP. In our institution, magnetic resonance imaging (MRI) scanning was routinely performed in all patients with CVT except for those with claustrophobia and MR contraindications, based on which, the parenchymal lesions, such as venous ischemia or hemorrhagic transformation, were disclosed. A part of patients, especially the candidates for endovascular procedure, were also imaged with other diagnostic modalities, including digital subtract angiography (DSA). It was worth noting that all patients were positioned supine and their head were in neutral positions when being imaged.
Locations (transverse sinus, sigmoid sinus, superior or inferior sagittal sinus, straight sinus or cortical veins) and laterals (right, left or bilateral) of the thrombi in cerebral venous system were confirmed by MRBTI. In cases with CVT plus nonthrombotic IJVS, the stenotic types of IJVS (venous wall damage or external compression) and the patterns of extrinsic compression (osseous, arterial or surrounding tissue compression) were also marked according to the results of CE-MRV and/or CTV. Furthermore, to determine the relationship between ICH and accessory venous circulation, the intra-and extracranial collaterals were evaluated as well, and the extents of which were allocated as mild or severe degrees based on the venograms. In detail, the scoring scheme for dilated scalp veins, namely additional routes for intracranial venous drainage in regions with venous blockage, were assigned as no presence, mild dilation or severe dilatation through visual assessment. 15 Besides, the dimensions of extracranial vertebral collaterals were classified into mild or severe degrees. The former means the maximal cross-sectional area of the collaterals <25% of the adjacent normal IJV, while the latter ≥25%. 8,9 An illustrated map concerning the compensatory patterns of the intra-and extracranial collaterals was shown as Figure 1. Two experienced neuroradiologists who were blinded to our clinical data and study intention completed all neuroimaging assessments. The interobserver divergencies were resolved through a discussion, or perhaps, a third rater would persuade the two assessors and decide the final result if the discussion failed to reach a consistency.

| Follow-up study
All patients informed consent to this study were followed-up by outpatient clinic or detailed telephone-based interviews. Also, some patients with recurrent CVT were rehospitalized for opti-

| Statistical analysis
Statistical Package for Social Sciences (SPSS) software version 25.0 was applied for our data processing. Numerical variables following the normal distribution were expressed as mean ± standard deviation (M ± SD) and calculated by two independent samples T-test; otherwise, they were presented as median (interquartile range,

| Demographics
A total of 199 patients with MRBTI confirmed CVT were enrolled consecutively from January 2018 through April 2021, and were grouped into CVT (n = 92) and IJVS-CVT (n = 107) groups, depending on the absence or presence of nonthrombotic IJVS. Details were displayed in Table 1. No statistical differences were found in the gender ratio and BMI between the CVT group (M/F = 41/51, 25.7

| Clinical features
As in Table 1 21.5%), and Frisén scores (median scores: 2.2 vs. 0.6). Additionally, the incidence of cerebral venous infarctions, irrespective of ischemia or hemorrhage, was mildly higher in the IJVS-CVT group than that in the CVT group, even though not reaching statistical difference (p > 0.05). The history of hypertension, diabetes mellitus, hyperlipidemia, smoking, and alcohol use were balanced between the two groups (all p > 0.05).
A vast majority of patients in the CVT group were implicated with clear thrombophilia, while such evidence was only found in minority cases in the IJVS-CVT group (83.7% vs. 20.6%, p < 0.001); 85 out of 107 patients (79.4%) in the IJVS-CVT group failed to identified with any routine risk factors of CVT, which indicated that IJVS might be closely associated with CVT formation. Specifically, statistical differences reached significant levels in almost all of the CVT risks as aforementioned between the two groups (all p < 0.05), aside from obstetric causes (p = 0.556). An illustrated diagram was displayed in Figure S1. For further test, as shown in Table 2 were considered caused by non-thrombotic IJVS, which was mainly secondary to external compression in this study. A corresponding forest plot is provided as Figure S2.
F I G U R E 1 Different patterns of the intra-and extracranial collaterals. A1 to E1 were the CTV maps and A2 to E2 were the CE-MRV pictures. The first three columns: the dilation patterns of the scalp veins (red arrows), no presence (A1and A2), mild dilation (B1 and B2) and severe dilatation (C1 and C2). The other two columns: the dilation patterns of the vertebral collaterals (red arrows), mild degree (D1 and D2) and severe degree (E1 and E2)

| Neuroimaging findings
As in Table 3

| Treatment
As for treatment patterns (Table 3)

| Follow-up outcomes
The average duration of clinical follow-up in this study was 13.0 ± 11.7 months in the CVT group and 12.5 ± 10.4 months in the IJVS-CVT group (Table 3)     Our study design successfully escaped from such dispute, as the IJVS-CVT cohort in this study was prevailingly without known CVTprovoked factors. For further determination, binary logistic regression was conducted and further confirmed that most thrombophilia conditions were negatively associated with IJVS-CVT. This may be explained that the enrolled IJVS patients in this study were mainly due to the age-dependent external compression, whereas patients with thrombophilia-triggered CVT were relatively young and without compressed IJVs.

| Probable mechanism for thrombosis in IJVS-CVT
According to the Virchow triad, entities that are responsible for endothelial dysfunction, blood hypercoagulability, and hemodynamic aberrations are the three elements for venous thrombus formation. 3 It is well-known that nonthrombotic IJVS is one of the cul-

| Thrombus distribution in CVT and IJVS-CVT
In this study, MRBTI was employed to display the direct visualization of thrombi. 24 Thanks to which we found that the thrombi in CVT were often distributed in venous sinuses, and only 35.9% involved cortical veins, namely cerebral cortical venous thrombosis (CCVT); conversely, the CCVT in the IJVS-CVT group accounted for up to 69.2%. We hence reasoned that nonthrombotic IJVS may probably impart higher thrombotic risk to the cortical veins rather than the venous sinuses, one possible explanation for which is that the impeded venous drainage caused by IJVS may more likely block the distal cerebral venous outflow with the ensuing delayed venous emptying within the distant cortical veins. Such a situation may render the cortical veins fill with static blood thereby contributing F I G U R E 2 Representative cases pertaining to the treatment patterns. A1-A4: a CVT patient with mild vertebral dilation was given with oral anticoagulation. The collaterals (red thin arrows) were fewer after (A2) than before (A1) medication, and the thrombi (white arrows) in the left transverse sinus (A3) were almost disappeared after therapy (A4). B1-B4: a CVT patient with straight sinus thrombosis underwent thrombolysis and mechanical thrombectomy. The straight sinus obtained almost complete recanalization after operation as detected by DSA (B2, white arrow) and MRBTI (B4, white arrow) maps. C1-C4: an IJVS-CVT patient underwent IJV balloon dilation. The stenotic IJV was corrected after balloon angioplasty on DSA map (C2, red thin arrow

| Different symptoms between CVT and IJVS-CVT
According to our results, the dilated scalp veins and the relevant clinical symptoms, such as headache were more prevalent in CVT patients, whereas older ages, head noise, dizziness, insomnia, in addition to emotion changes and vertebral collaterals were more popular in IJVS-CVT patients. Correspondingly, the IJVS in this study were mainly caused by the compression of bones or stiffened arteries, both of which were also age-dependent. The above consisted with a notion that nonthrombotic IJVS tended to be acquired and develop with higher frequency during the aging process as revealed previously. 25 And the IJVS-triggered ICH may force the extension of collaterals in order to lighten the drainage burden of the stenotic jug- function as compensatory pathways to relieve the venous pressure and ICP. 29 Mimic with vertebral collaterals, the existence of highly dilated scalp veins in CVT could be considered as a surrogate marker for anomalous venous return. 31 Although the dilated scalp veins in CVT provided partial support for the obstructed venous reflow and thereby alleviated the drainage resistance, 15,32 and might probably decrease the risk of venous infarction, 33 they were still not strong enough to completely compensate the thrombosis-mediated ICH.
Therefore, the average ICP in the CVT group was obviously higher than that in the IJVS-CVT group in this study. asymptomatic and not all of them suffered from CVT. Given that, the strict patient selections, including collaterals must present around the deformed IJV and the IJVS without collaterals must be excluded from the participations, were adopted in this study.

| Different outcomes between CVT and IJVS-CVT
Noteworthy was that nonthrombotic IJVS with collateral formation was judged to be a pathological entity as previously emphasized. 37,38 Therefore, we believed that our imaging protocols might act to eliminate the influence from the anatomic jugular variants and the IJVS in this study was indeed a venous anomaly. Besides, our series of studies were already concerned on the relationship between IJVS and CVT based on ultrasound evaluation or case report, 6,39

| LIMITATI ON S AND FUTURE DIREC TIONS
Shortcomings are worthy noticing: Firstly, this study is only a mono-center real-world case-control study with a relatively short follow-up period. Secondly, it remains undetermined whether or not nonthrombotic IJVS-CVT should be granted with long-term anticoagulation. Thirdly, the cut-off value of the degree of IJVS on promoting CVT is still not clear. Last but not least, the accurate mechanisms of IJVS-induced CVT were not elaborated very clearly due to the shortage of auxiliary modalities for a specific hemodynamic study.
We hope findings from future studies can shed more light on the questions raised above.

| CON CLUS IONS
To sum up, nonthrombotic IJVS may be a noteworthy risk factor for CVT occurrence, possibly due to the obstructed venous outflow induced by the IJVS. Besides, patients with nonthrombotic IJVS may possibly benefit from continuous anticoagulation and IJVS correction, although further well-designed clinical trials with a large simple size are still needed to clarify these findings.

ACK N OWLED G M ENT
We thank all patients and doctors who participated in this study for their cooperation.

CO N FLI C T O F I NTE R E S T
XW, JY, DZ, YD, XJ, and RM report no conflicts of interest.

AUTH O R CO NTR I B UTI O N S
XW and JY completed the study and drafted this manuscript. RM, DZ, XJ, and YD revised the manuscript and edited English. RM contributed to the conception and design of this study and proposed the amendments.

CO N S E NT FO R PU B LI C ATI O N
The authors agree to publish.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data and materials related to the current study can be accessed from our corresponding author upon reasonable request.