Abstract

Background:

Flow diversion stents (FDS), particularly the Pipeline Embolization Device (PED), have been widely used for intracranial aneurysm (IA) treatment. However, data on PED deployment in small-caliber parent vessels (< 2 mm) remain limited. This study aimed to evaluate the technical feasibility, safety, and efficacy of PED for IAs arising from small-caliber vessels.

Methods:

This was a multicenter retrospective study enrolling 71 eligible patients with IA located in parent vessels < 2 mm who underwent PED implantation. The primary safety endpoint included procedure-related death, symptomatic stroke, intracranial hemorrhage and asymptomatic in-stent stenosis; the primary efficacy endpoint was complete aneurysm occlusion [O’Kelly-Marotta (OKM) grade D] on follow-up digital subtraction angiography (DSA). Binary logistic regression analyses were used to identify factors associated with complications and delayed IA occlusion.

Results:

The patient age was 55 (47, 59) years, with 37 (52.1%) males. The follow-up duration was 9 (7, 15) months. Postoperative complications included ischemic events (16.9%), hemorrhagic events (4.2%), and asymptomatic in-stent stenosis (12.7%). The 1-year complete occlusion rate was 52%, whereas the long-term occlusion rate gradually increased to 80% with follow-up beyond 12 months. Univariable analysis showed PED length was associated with both delayed IA occlusion [odds ratio (OR) = 0.91, 95% confidence interval (CI) = 0.83–1.00; p = 0.040] and ischemic complications (OR = 1.36, 95% CI = 1.11–1.66; p = 0.003) and also revealed that pre-existing parent vessel stenosis was strongly associated with asymptomatic in-stent stenosis (OR = 13.50, 95% CI = 2.80–65.04; p = 0.001); multivariable analysis confirmed PED length as an independent predictor of ischemic complications (OR = 1.31, 95% CI = 1.06–1.63; p = 0.013).

Conclusion:

PED deployment in small-caliber parent vessels (<2 mm) is technically feasible and clinically viable, with favorable IA occlusion rates and an acceptable safety profile. These findings support the off-label use of PEDs in this challenging cohort, provided rigorous patient selection, optimal procedural planning and close follow-up are implemented.

Introduction

Intracranial aneurysm (IA) constitutes a prominent cerebrovascular pathology, with a prevalence of approximately 3–5% in the general population (1). They carry a substantial risk of rupture, which can result in subarachnoid hemorrhage (SAH), associated with a mortality rate of roughly 50% and significant morbidity among survivors (2, 3). Traditional therapeutic strategies, such as microsurgical clipping and endovascular coiling, have long served as the mainstays for IA management; however, these approaches are often constrained in complex cases like wide-necked or fusiform IA, where incomplete IA occlusion, recurrence, and procedural complications remain common challenges (4, 5). The emergence of flow diversion stent (FDS), notably the Pipeline Embolization Device (PED, Medtronic), has ushered in a paradigm shift in endovascular therapy. FDS facilitate endoluminal reconstruction by redirecting hemodynamics, promoting IA thrombosis, and inducing neointimal proliferation over the stent, thereby achieve higher long-term IA occlusion rates compared to conventional endovascular approaches endovascular techniques (6, 7).

Initially approved by the U. S. Food and Drug Administration for large or giant wide-necked IA located in the proximal internal carotid artery (ICA), the PED has increasingly been used off-label for a broader range of IA, including those in distal anterior and posterior circulation segments (8, 9). However, deployment of the PED in small-caliber vessels (mean diameter <2 mm) presents unique challenges, given the device’s nominal minimum diameter of 2.5 mm. This mismatch between vessel and stent diameters raises concerns about potential vessel injury, acute in-stent thrombosis, delayed vessel occlusion, and higher rates of ischemic complications arising from the anatomical fragility and tortuosity of distal vessels, as well as hemorrhagic complications, such as delayed IA rupture and distal parenchymal hemorrhage induced by hemodynamic redirection following PED placement (10, 11). Early studies evaluating FDS use in vessels ≤ 2.5 mm have reported procedural complication rates of 9–10%, symptomatic stroke incidences of approximately 7–8%, and mid-term follow-up complete occlusion rates of 70–75%. These findings suggest efficacy comparable to that in larger vessels but with heightened periprocedural risks, which warrant meticulous patient selection and antiplatelet therapy management (12).

Despite these insights, evidence specifically focusing on PED placement in parent vessels < 2 mm remains scarce. Most existing data are derived from small single-center retrospective series or systematic reviews, underscoring the need for additional research on mid- to long-term safety and angiographic outcomes (13). The present multicenter retrospective study aims to assess the safety and efficacy of PED deployment for IA originating from small parent vessels (< 2 mm) by analyzing clinical and radiological outcomes. This work thereby enhances the evolving understanding of off-label PED use in this challenging IA subset.

Materials and methodsStudy design

This retrospective multicenter study was approved by the ethics committees of all participating centers and conducted in compliance with the principles of informed consent, with all patients providing written informed consent prior to treatment (Approval Number: KY2025-345-02).

From the prospectively maintained institutional IA databases of three comprehensive stroke centers, we retrospectively identified patients who underwent PED Flex placement, with at least one PED Flex deployed in a parent vessel with a mean diameter <2 mm. Patients with intracranial tumors, cerebrovascular malformations, traumatic or infectious dissections, polycystic kidney diseases, rheumatoid autoimmune diseases, ruptured IA, severe intracranial atherosclerotic stenosis/occlusion or those receiving FDS treatments at other institutions were excluded because these conditions may substantially affect intracranial vascular morphology, hemodynamics, or treatment-related risks. To isolate and accurately evaluate the safety and efficacy of PED for IA, these potential confounding factors were excluded from the study cohort. The study period included cases treated between 2018 and 2022. Treatment decisions were made on a case-by-case basis by a multidisciplinary endovascular team consisting of dual-trained neurosurgeons and neurointervenionalists. The Measurements of proximal and distal parent vessel diameters, aneurysm neck, and maximum IA diameter, were performed based on 3D rotational digital subtraction angiography (DSA) images with standard calibration and mean vessel diameters were computed for each case. All imaging evaluations were independently conducted by two neurosurgeons with more than 5 years of neurointerventional experience. Any discrepancies between the two readers were resolved through consensus discussion. A third senior neurosurgeon with more than 5 years of experience supervised the entire imaging review process and took responsibility for the accuracy and validity of the final measurements. Both anterior and posterior circulation IA meeting the primary inclusion criteria were included. Data collection encompassed relevant patient demographics (age, sex), comorbidities (hypertension, hyperlipidemia, diabetes, coronary artery disease, prior stroke, and history of SAH), IA characteristics (type, location, size and parent vessel diameter), clinical presentation, immediate and delayed complications, and follow-up outcomes (both radiological and clinical).

Antiplatelet therapy

CYP2C19 genotyping and antiplatelet responsiveness testing were consistently performed across all participating centers. All patients received standard dual antiplatelet therapy (DAPT), consisting of 100 mg aspirin and 75 mg clopidogrel daily, initiated at least 5 days preoperatively and continued for 3–12 months postoperatively. Clopidogrel was switched to ticagrelor (90 mg twice daily) in patients identified as clopidogrel-resistant via CYP2C19 genotyping. For acutely ruptured IA treated emergently, patients received a loading dose of DAPT (300 mg aspirin and 300 mg clopidogrel) preoperatively, followed by an intravenous glycoprotein IIb/IIIa inhibitor (tirofiban; weight-adjusted dosage) administered for 12–24 h. Subsequent maintenance was continued with oral standard DAPT. The duration of antiplatelet therapy was adjusted according to postoperative DSA follow-up findings such as lifelong aspirin administration or discontinuation after a specified duration.

Treatment strategy

All procedures were performed under general anesthesia. Vascular access was obtained via the right common femoral artery. A guiding catheter was then navigated to the ipsilateral ICA or vertebral artery, based on the IA’s location. Subsequently, a comprehensive evaluation of the parent vessel and IA was conducted to formulate the treatment strategy. Recorded anatomical characteristics included IA type, neck, maximum IA diameter, parent vessel diameter, and planned PED length (with a minimum 3-mm safety margin extending proximally and distally beyond the IA neck). PED deployment was performed using a triaxial system, consisting of a guiding catheter, intermediate catheter, and either a Marksman™ (Medtronic) or Phenom-27™ (Medtronic) microcatheter advanced over a micro-guidewire. Following PED deployment, immediate postprocedural DSA was acquired in magnified views for all patients to confirm adequate device positioning across the IA neck and optimal vessel wall apposition. Additionally, standard angiographic views were obtained to exclude thromboembolic or hemorrhagic complications. Intravenous heparin (50–70 U/kg) was administered after general anesthesia if one or more of the following conditions were present: preoperative platelet function test failing to meet the inhibition standard, emergency patients requiring flow diversion device implantation, prolonged surgical duration, tortuous blood vessels, severe stenosis of the parent vessel, multiple intracranial arterial stenosis, or poor local stent apposition requiring massage to achieve optimal apposition. An additional 1,000 U of heparin was given every 1 h during the operation. The operator could use tirofiban instead of heparin. Regularly, tirofiban was administered by intravenous pump at a rate of 0.1 mL•kg/h during the operation under the same indications as heparin, and the administration was discontinued 12 h after the operation. For patients with severe intraoperative vascular spasm that did not resolve spontaneously after 5–10 min of observation, 0.5 mg of nimodipine was slowly injected intraoperatively. In addition, oral or enteral nimodipine (60 mg every 4 h for 21 consecutive days) was administered to patients with aneurysmal SAH. Hemostasis at the femoral access site was achieved with an arterial closure device.

Outcomes

The primary clinical safety endpoint included procedure-related death, symptomatic ischemic stroke, asymptomatic in-stent stenosis or occlusion, and procedure-related intracranial hemorrhage. Symptomatic ischemic stroke and procedure-related intracranial hemorrhage were defined as new focal neurological deficits consistent with territorial ischemia or hemorrhage, confirmed by postoperative CT or MRI. In-stent stenosis was defined as any reduction in the contrast-filled lumen of the parent vessel on follow-up DSA. On DSA images, ISS is visualized as a discernible gap between the contrast-opacified vessel lumen and the inner wall of the PED (14). Clinical outcomes were assessed using the modified Rankin Scale (mRS) score, with evaluations performed at discharge and during scheduled follow-up visits (mRS score: 0 = completely asymptomatic, 1 = symptomatic but without significant functional impairment, able to perform all daily activities and work, 2 = mild disability, unable to perform all activities but can manage personal affairs without assistance, 3 = moderate disability, requiring some help but able to walk independently, 4 = moderately severe disability, unable to walk independently, requiring assistance in daily life, 5 = severe disability, bedridden, incontinent, completely dependent on others for daily activities, 6 = death). The primary efficacy endpoint was complete occlusion of the target IA, as determined by the O’Kelly-Marotta (OKM) grading scale, on follow-up DSA. 1-year aneurysm occlusion rate was defined as complete aneurysm occlusion (OKM class D) documented on the patient’s DSA within the 12-month follow-up. Final aneurysm occlusion rate was defined as complete aneurysm occlusion (OKM class D) documented on the patient’s last available DSA follow-up. Delayed aneurysm occlusion was defined as persistent contrast filling within the aneurysm sac on DSA performed at ≥12 months after PED implantation.

Statistical analysis

Statistical analyses were performed using SPSS software (version 25.0; IBM Corp., Armonk, NY, United States). Normally distributed continuous variables were expressed as mean ± standard deviation (SD). Non-normally distributed continuous variables were reported as median (interquartile range, IQR; Q1–Q3). Categorical variables were presented as counts (percentages). Continuous variables were compared using the two-tailed Student’ s t-test or Mann–Whitney U test, as appropriate based on data distribution. Categorical variables were compared using the chi-square test or Fisher’ s exact test. Variables with p < 0.05 in univariable analysis were included in multivariable logistic regression models to identify independent predictors of procedure-related complications and clinical efficacy outcomes, after adjusting for potential confounding factors. Besides, based on our experience, we estimate that the postoperative complication rate will not exceed 20% of the total number of patients. According to the widely accepted events-per-variable rule (≥10 events per variable), the ideal number of variables for the multivariable model was calculated as approximately 1.5. Therefore, we limited the number of variables included in the multivariable model to a maximum of 2 to maintain statistical reliability. A two-sided p < 0.05 was considered statistically significant.

ResultsPatient demographics and baseline characteristics

After initial screening, a total of 71 eligible patients were enrolled in this study. Figure 1 illustrates the study flowchart. The patient age was 55 (47, 59) years, with 37 patients (52.1%) being male. Baseline comorbidities included hypertension in 34 patients (47.9%), diabetes in 7 (9.9%), dyslipidemia in 6 (8.5%), coronary artery disease in 8 (11.3%), prior SAH in 11 (15.5%), and prior ischemic stroke in 6 (8.5%). Sixteen patients (22.5%) reported current smoking, and 12 (16.9%) reported regular alcohol consumption. Preprocedural modified Rankin Scale (mRS) scores were 0 in 48 patients (67.6%), 1 in 18 (25.4%), 2 in 2 (2.8%), and 3 in 3 (4.2%). Ten patients (14.1%) had a history of prior endovascular treatment for IA (Table 1).

The flowchart of this study. After screening, a total of 71 eligible patients were enrolled in this study. IA, Intracranial aneurysm.

VariablesNumber (proportion)/median (Q1, Q3)Age55 (47, 59)Male37 (52.1%)Hypertension34 (47.9%)Diabetes7 (9.9%)Dyslipidemia6 (8.5%)Coronary artery disease8 (11.3%)History of SAH11 (15.5%)History ischemic stroke6 (8.5%)Current smoking16 (22.5%)Regular alcohol consumption12 (16.9%)Pre-operation mRS048 (67.6%)118 (25.4%)22 (2.8%)33 (4.3%)History of endovascular treatment for IA10 (14.1%)

Baseline characteristics of the participants.

IA, Intracranial Aneurysm; SAH, subarachnoid hemorrhage; mRS, modified Rankin Scale score: 0 = completely asymptomatic, 1 = symptomatic but without significant functional impairment, able to perform all daily activities and work, 2 = mild disability, unable to perform all activities but can manage personal affairs without assistance, 3 = moderate disability, requiring some help but able to walk independently, 4 = moderately severe disability, unable to walk independently, requiring assistance in daily life, 5 = severe disability, bedridden, incontinent, completely dependent on others for daily activities, 6 = death.

Procedural and IA characteristics

IA characteristics included bifurcation IA in 39 patients (54.9%), IA with daughter sacs in 12 (16.9%), non-saccular IA in 34 (47.9%), symptomatic IA in 36 (50.7%), ruptured IA in 5 (7.0%) and multiple IA in 15 (21.1%). The IA neck was 6.8 (5.00, 9.79) mm, maximum IA diameter was 10.94 (6.93, 13.6) mm, and parent vessel diameter was 1.8 (1.62, 1.9) mm. Parent vessel stenosis was observed in 14 patients (19.7%) and 4 (5.6%) patients had intracranial arterial stenosis in vessels other than the parent artery. IA locations were as follows: A2 segment in 6 patients (8.5%), A3 segment in 14 (19.7%), M1-M2 segments in 33 (46.5%), M2-M3 segments in 8 (11.3%), P2 segments in 9 (12.7%), and P3 segment in 1 (1.4%). The PED length was 20 (20, 30) mm, PED diameter was 3 (2.5, 3.25) mm, and median diameter mismatch between the PED and parent vessel was 1.2 (0.92, 1.71) mm. Procedural details included deployment of multiple PEDs in 7 patients (9.9%), intraoperative balloon angioplasty in 11 (15.5%), and PED-assisted coiling in 16 (22.5%). Intraoperative medications administered included nimodipine in 17 patients (23.9%), heparin in 32 (45.1%), and tirofiban in 37 (52.1%) (Table 2).

VariablesNumber (proportion)/median (Q1, Q3)Multiple PED deployment7 (9.9%)Intraoperative balloon angioplasty11 (15.5%)PED assisted with coils16 (22.5%)Nimodipine utilization17 (23.9%)Heparin utilization32 (45.1%)Tirofiban utilization37 (52.1%)Ischemic complication12 (16.9%)Procedure-related intracranial hemorrhage3 (4.2%)Asymptomatic in-stent stenosis9 (12.7%)Delayed occlusion of aneurysm16 (22.5%)Procedure-related death2 (2.8%)Bifurcation IA39 (54.9%)IA with daughter sac12 (16.9%)Non-saccular IA34 (47.9%)Symptomatic IA36 (50.7%)Ruptured IA5 (7.0%)Multiple IAs15 (21.1%)IA neck (mm)6.8 (5.00, 9.79)Maximum diameter of the IA (mm)10.94 (6.93, 13.60)Mean diameter of parent vessel (mm)1.8 (1.62, 1.9)Stenosis of parent vessel14 (19.7%)Stenosis of other intracranial arteries4 (5.6%)IA LocationA26 (8.5%)A314 (19.7%)M1-M233 (46.5%)M2-M38 (11.3%)P29 (12.7%)P31 (1.4%)Lenth of PED (mm)20 (20, 30)Diameter of PED (mm)3 (2.5, 3.25)Diameter difference between PED and parent vessel (mm)1.2 (0.92, 1.71)PED Covered ≥ 2 branches observed in DSA55 (77.5%)Incomplete wall apposition5 (7.0%)Duration of aspirin use post-operation (month)12 (6, 24)Duration of clopidogrel use post-operation (month)6 (3, 6)Clopidogrel replaced by ticagrelor2 (2.8%)Last follow-up time (month)9 (7, 15)Last follow-up mRS056 (78.9%)17 (9.9%)23 (4.2%)32 (2.8%)41 (1.4%)

Procedure-related and follow-up characteristics of the participants.

IA, Intracranial Aneurysm; PED, Pipeline Embolization Device; mRS, modified Rankin Scale score: 0 = completely asymptomatic, 1 = symptomatic but without significant functional impairment, able to perform all daily activities and work, 2 = mild disability, unable to perform all activities but can manage personal affairs without assistance, 3 = moderate disability, requiring some help but able to walk independently, 4 = moderately severe disability, unable to walk independently, requiring assistance in daily life, 5 = severe disability, bedridden, incontinent, completely dependent on others for daily activities, 6 = death; A2, A3: segment of the anterior cerebral artery; M1, 2, M3: segment of the middle cerebral artery; P: P2, P3: segment of the posterior cerebral artery.

Post-operative and follow-up outcomes

Postoperative complications included ischemic events in 12 patients (16.9%), procedure-related intracranial hemorrhage events in 3 (4.2%), asymptomatic in-stent stenosis in 9 (12.7%), delayed IA occlusion in 16 (22.5%), and procedure-related death in 2 (2.8%). Both patients suffered procedure-related death succumbed to fatal intracranial hemorrhage that occurred in the early postoperative period following PED implantation. The clinical progression of both cases was rapid, and thus neither patient underwent postoperative DSA evaluation. Based on the findings of emergency postoperative CT, we hypothesize that both events were attributable to delayed IA rupture after the PED procedure. The follow-up duration was 9 (7, 15) months, with a maximum follow-up duration of 60 months. Durations of postoperative antiplatelet therapy were 12 (6, 24) months for aspirin and 6 (3, 6) months for clopidogrel; clopidogrel was replaced with ticagrelor in 2 patients (2.8%). At the final DSA follow-up, the primary efficacy endpoint (OKM grade D, complete occlusion) was achieved in 80% of cases. For one-year clinical outcomes, assessed using the mRS: 56 patients (78.9%) achieved 0 score, 12 (16.9%) had 1–3 scores, 1 (1.4%) had 4 scores (Table 2). The IA occlusion curve (Figure 2) depicts the cumulative IA occlusion rate over time, showing a rapid increase within the first 12 months, followed by a plateau phase extending to 60 months, with the rate stabilizing at approximately 80%. This trend highlights the time-dependent efficacy of PED in small-caliber vessels, with the majority of IA occlusions occurring within the initial year. Detailed data on the number of patients at each follow-up time point and the DSA completion rate at each time point are provided in Supplementary Table 1.

The IA occlusion curve for PED deployed in small caliber vessels (< 2 mm) for IA treatment. This curve indicates that more than 50% of patients achieve IA occlusion within 1 year after PED deployment. As follow-up time extends, an increasing number of patients attain satisfactory efficacy outcomes, underscoring the clinical significance of long-term follow-up for such patients.

Factors associated with post-operative complications and delayed IA occlusion

To avoid collinearity, we excluded variables with variance inflation factor >10 from the logistic regression analysis. Univariable logistic regression analysis showed that IA neck [odds ratio (OR) = 1.10, 95% confidence interval (CI) = 1.00–1.21; p = 0.039] and PED length (OR = 1.36, 95% CI = 1.11–1.66; p = 0.003) were significantly associated with postoperative ischemic complications. In multivariable logistic regression analysis, PED length remained an independent predictor of ischemic complications (OR = 1.31, 95% CI = 1.06–1.63; p = 0.013) (Table 3). For delayed IA occlusion, univariable analysis revealed that PED length was a significant associated factor (OR = 0.91, 95% CI = 0.83–1.00; p = 0.040). Regarding asymptomatic in-stent stenosis, parent vessel stenosis was strongly associated with this complication (OR = 13.50, 95% CI = 2.80–65.04; p = 0.001). Given that only one variable reached statistical significance in the univariable analysis for delayed IA occlusion or asymptomatic in-stent stenosis, multivariable analysis was not deemed appropriate to avoid overfitting (Table 4).

CharacteristicsUnivariableAdjustedPOR95%CIPOR95%CIFemale0.4930.670.212.12////Age0.8151.000.971.04////Hypertension0.9150.940.32.94////Diabetes0.991NANANA////Dyslipidemia0.992NANANA////Coronary artery disease0.5330.500.064.41////History of ischemic stroke0.4512.000.3312.13////Current smoking0.7910.830.203.38////Regular alcohol consumption0.6790.710.143.64////IA locationA2refrefrefref////A30.992NANANA////M1-M20.2790.370.062.21////M30.1480.140.011.99////P1-P20.120.120.011.72////P30.998NANANA////Bifurcation IA0.6571.300.414.14////Non-saccular IA0.6351.320.424.13////IA with daughter sac02642.180.568.56Stenosis of parent vessel0.992NANANA////Multiple PED deployment0.1553.250.6416.48////PED covered ≥ 2 branches observed in DSA0.8230.8480.203.594////Incomplete wall apposition0.999NANANA////PED with coils0.2662.050.587.21////Intraoperative balloon angioplasty0.5891.500.346.52////Multiple IAs0.992NANANA////Nimodipine utilization0.6880.750.183.05////Heparin utilization0.8891.090.353.40////