Abstract
Background
The Pipeline Flex embolization device with Shield technology (PED Shield, Medtronic, Irvine, CA, USA) is a flow diverter designed to reduce thrombogenicity through phosphorylcholine surface modification. Although in vitro and in vivo studies have demonstrated the thrombogenicity reducing effects of this technology, its effectiveness in real-world clinical practice remains unclear. This study aims to compare the number of post-procedure diffusion-weighted imaging (DWI)-positive lesions between PED Flex(Medtronic, Irvine, CA, USA) and PED Shield to assess the impact of surface modification.
Methods
This retrospective cohort study included patients with unruptured intracranial aneurysms treated with PED Flex or PED Shield between April 2016 and March 2024 at a single institution. Propensity score matching was performed to control for confounders, and the number of post-procedure DWI-positive lesions was evaluated as the primary outcome.
Results
In total, 148 procedures (132 patients, 138 aneurysms) were included, with 68 (46%) treated with PED Flex and 80 (54%) with PED Shield. Propensity score matching resulted in 47 matched pairs. After matching, the median number of post-procedure DWI-positive lesions was nine (interquartile range (IQR): 3–17) in the PED Flex group and three (IQR: 1–6) in the PED Shield group, with a significantly lower number in the PED Shield group (regression coefficient β = −10.70 [95% confidence interval (CI): −16.23 to −5.16), P<0.001).
Conclusion
After adjusting for confounders, the PED Shield group had significantly fewer post-procedure DWI-positive lesions than the PED Flex group, suggesting that phosphorylcholine surface modification technology may reduce thrombogenicity in real-world clinical practice.
Keywords: Aneurysm, Flow Diverter, Embolic, Complication, MRI
WHAT IS ALREADY KNOWN ON THIS TOPIC
The Pipeline Flex embolization device with Shield technology (PED Shield) is a flow diverter designed to reduce thrombogenicity through phosphorylcholine surface modification. However, its effectiveness in reducing thrombogenicity in real-world clinical practice remains unclear. Post-procedure diffusion-weighted imaging (DWI)-positive lesions are considered a potential marker for in-human thrombogenicity.
WHAT THIS STUDY ADDS
The PED Shield group had significantly fewer post-procedure DWI-positive lesions than the PED Flex group, suggesting that phosphorylcholine surface modification technology may reduce thrombogenicity in real-world clinical practice.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
The findings of this study support the ongoing development of next-generation flow diverters incorporating various surface modification technologies.
Introduction
Flow diverters (FDs) have greatly improved treatment outcomes for unruptured intracranial aneurysms and are now widely used in clinical practice.1,4 These devices function by redirecting blood flow away from the aneurysm sac through their high mesh density, promoting thrombosis within the aneurysm. However, the high metal coverage of FDs can also lead to thrombus formation on the stent surface, making thromboembolic events one of the most concerning complications after FD implantation. These complications can occur even with appropriate dual antiplatelet therapy (DAPT). Additionally, the hemorrhagic risks associated with potent antiplatelet therapy remain a significant concern.
To address the issue of thromboembolic complications, various FDs with surface modifications have been developed in recent years.5,7 The Pipeline Flex embolization device with Shield technology (PED Shield; Medtronic, Irvine, CA, USA) was the first FD to incorporate surface modification technology and was introduced several years ago. This modification, known as “Shield technology,” reduces material thrombogenicity by covalently bonding phosphorylcholine to the device surface. Both in vitro and in vivo studies have demonstrated that this surface modification reduces the thrombogenicity of the device.8 9 In clinical research, the efficacy and safety of the PED Shield have been established in single-arm studies.10,12 However, studies directly comparing thromboembolic complications between the PED Shield and non-surface-modified devices, such as the PED Classic and PED Flex(Medtronic, Irvine, CA, USA), remain limited. Recently, a few retrospective cohort studies have been published comparing the PED Classic/PED Flex with the PED Shield.13,16 However, none of these studies demonstrated a significant reduction in thromboembolic events with the PED Shield, leaving its impact in real-world clinical practice uncertain.
Small, silent infarcts caused by thromboemboli are frequently observed on diffusion-weighted imaging (DWI) following endovascular treatment for intracranial aneurysms. Previous studies suggest that flow diversion treatment is associated with a higher rate of DWI-positive thromboembolic events than are conventional treatments such as coil embolization.17 The number of DWI-positive lesions after endovascular procedures for aneurysms has been linked to symptomatic ischemic complications and is considered a potential marker of thrombogenicity.18 As a result, it has been adopted as an outcome measure in studies assessing the ability of surface-modified FDs to reduce thrombogenicity.13 19
This study aims to evaluate whether surface modification reduces thrombogenicity in real-world clinical practice by comparing the number of post-procedure DWI-positive lesions between the PED Shield and PED Flex.
Methods
Study design and settings
This was a retrospective cohort study using a prospectively maintained database of consecutive patients treated for intracranial aneurysms with FDs at a single center in Japan from April 1, 2016 to March 31, 2024. We identified consecutive patients with intracranial aneurysms treated with either the PED Flex or PED Shield. Patients who were unable to undergo postoperative MRI within 48 hours of the index procedure were excluded. This study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for observational studies. The study protocol was approved by the Ethics Committee of Kyoto University Hospital (ID R2088), which waived the requirement for informed consent because of the retrospective study design.
Data collection
Baseline characteristics were collected for all patients, including age, sex, medical history (hypertension, hyperlipidemia, diabetes mellitus), smoking history, aneurysm location, aneurysm dimensions (neck size, maximal diameter), and the presence of a branch-incorporated aneurysm.
Antiplatelet regimen and anticoagulation
According to our institutional protocol, patients scheduled for treatment with FDs received DAPT consisting of aspirin (100 mg) and clopidogrel (75 mg) for at least 14 days before the procedure. The antiplatelet response was monitored using VerifyNow (Accumetrics, San Diego, CA, USA), and patients with clopidogrel resistance (>230 platelet reactivity units (PRU)) were transitioned to prasugrel. The definition of clopidogrel resistance was based on previous studies.20 21 We also included patients with >230 PRU before treatment (despite transitioning to prasugrel) in the analysis. DAPT was continued for at least 6 months after the procedure, followed by indefinite single antiplatelet therapy. Intraoperatively, patients received intravenous heparin (bolus of 50 U/kg), with the activated clotting time monitored to maintain a target value of >250 s. Heparin was discontinued after the procedure without reversal.
Procedure and MRI protocol
All procedures were performed under general anesthesia using a standard transfemoral approach. All PEDs were deployed through a Marksman (Medtronic, Irvine, CA, USA) or Phenom 27 (Medtronic, Irvine, CA, USA) microcatheter with a coaxial support system. FD implantation was performed by a neuroendovascular specialist (AI) with >10 years of experience in intracranial stent placement and approximately 30 PED Flex cases completed before the start of this study. The selection of PED type was exclusively PED Flex before June 1, 2019 and entirely PED Shield thereafter. When a single PED was insufficient to fully cover the aneurysm neck, additional devices were deployed. Adjunctive coiling was generally performed for aneurysms located in the subarachnoid space to prevent delayed rupture.
At our institution, postoperative magnetic resonance imaging (MRI) is routinely performed after FD implantation. In the present study, images were typically obtained within 48 hours of the procedure using a 3.0-Tesla MRI scanner (MAGNETOM Skyra; Siemens Healthineers, Erlangen, Germany) with a 64-channel head coil. The imaging parameters for DWI were as follows: echo planar imaging-based three-direction trace imaging sequence; axial acquisition; repetition time/echo time, 5000 ms/61 ms; b-value, 1000 s/mm2; echo train length, 70; field of view, 220×220 mm; slice thickness, 3 mm; slice gap, 1 mm; resolution, 0.69×0.69 mm; averaging factor of 2; and trace-weighted imaging. DWI-positive lesions were evaluated independently by two board-certified neuroendovascular surgeons (RA and AI), each with >10 years of experience in brain MRI interpretation. Contiguous DWI-positive lesions spanning multiple MRI slices were counted as one lesion. In cases of disagreement between the two investigators, a third board-certified neuroendovascular surgeon (SM) reviewed the data, and a final consensus was reached among all three.
Study outcomes
The primary outcome was the number of DWI-positive lesions detected within 48 hours of the index procedure on 3.0-Tesla MRI. The secondary safety outcomes were ischemic complications, hemorrhagic complications, and technical events. Ischemic complications were defined as any symptomatic ischemic event related to the procedure within 30 days, including transient ischemic attacks. Hemorrhagic complications were defined as any intracranial hemorrhage detected on postoperative computed tomography (CT) within 30 days. Technical events included fish mouthing, twisting, and device replacement due to deployment failure. The secondary efficacy outcomes were the rates of adequate aneurysm occlusion, defined as O'Kelly–Marotta grades22 C and D, at 6 and 12 months post-procedure.
Statistical analysis
Patient characteristics were described using the median and IQR for continuous variables and the number and percentage for categorical variables. Continuous variables were compared using the Mann–Whitney U test, while categorical variables were compared using the Chi-squared test or Fisher’s exact test. The study team identified clinically important covariates, including age, sex, hypertension, hyperlipidemia, diabetes mellitus, smoking history, aneurysm location (intradural vs extradural, and anterior vs posterior circulation), aneurysm maximum diameter, branch-incorporated aneurysm, and clopidogrel resistance. These variables were selected based on prior associations with ischemic complications following flow diversion treatment or because they were considered important demographic factors. A logistic regression model incorporating these covariates was used to develop a propensity score for PED selection. We used the Matching package in R (version 4.3.3) to perform one-to-one nearest-neighbor matching without replacement. Matching was based on the estimated propensity score (ie, the predicted probability from the logistic regression model), using a caliper width of 0.2 SD of the raw propensity score. A greedy matching algorithm in random order was applied, rather than optimal matching. Covariate balance was assessed by calculating the absolute standardized mean difference (ASMD), with an ASMD of >0.2 considered indicative of a meaningful difference. In both the unmatched and matched cohorts, the primary outcome was expressed as the regression coefficient (β) and 95% CI using analysis of covariance (ANCOVA). Secondary outcomes were reported as odds ratios (ORs) with 95% CIs using logistic regression analysis. Adjunctive coiling, multiple PED implantation and repeat flow diversion was not included in the propensity score model because its inclusion worsened the covariate balance after matching. Instead, it was adjusted for in post-matching analyses (ie, ANCOVA and logistic regression), because it was considered a potential confounder. To mitigate small-sample bias and correct for separation in logistic regression, particularly for rare events or outcomes with quasi-complete separation, Firth’s bias reduction method was applied.23 24
Two types of sensitivity analyses were performed to assess the robustness of the primary and secondary analyses. First, we conducted an inverse probability of treatment-weighting analysis to estimate the average treatment effect. This method assigns weights to each patient based on the inverse probability of receiving treatment, calculated as one for exposed patients and as propensity score / (1 − propensity score) for unexposed patients. Second, we performed an ANCOVA for the primary outcome, with the primary outcome as the dependent variable and all covariates, along with the type of PED, as independent variables. The β value and 95% CIs were calculated. Third, we conducted multivariable logistic regression analyses for the secondary outcomes, with the secondary outcome as the dependent variable and all covariates, along with the type of PED, as independent variables. Adjusted ORs and 95% CIs were calculated. Firth’s bias reduction method was also applied in the sensitivity analysis.
A two-sided P value of <0.05 was considered statistically significant. All analyses were performed using RStudio (version 2024.04.2, Build 764) running on R version 4.3.3 and JMP Pro software version 17 (SAS Institute, Cary, NC, USA).
Results
Patient flow
From April 1, 2016 through March 31, 2024, the database recorded 134 patients and 140 aneurysms treated with PEDs (figure 1). Two patients were excluded: one due to an inability to undergo MRI because of a cochlear implant and another who declined to participate in the study. The remaining 132 patients, 138 aneurysms, and 148 PED procedures were included in the final analysis.
Figure 1. Study flowchart. PED, pipeline embolization device; MRI, magnetic resonance imaging.
Baseline characteristics
Overall, the median age was 63 years (IQR=52–71), and 88% of patients were female. Hypertension, hyperlipidemia, and diabetes mellitus were present in 52%, 14%, and 26% of patients, respectively. Most patients were nonsmokers (64%). Among the aneurysms, 98% were located in the internal carotid artery: 34% in the cavernous segment, 46% in the paraclinoid segment, 15% at the posterior communicating artery bifurcation, and 3% at the anterior choroidal artery bifurcation. The remaining 2% were located in the posterior circulation, all of which were vertebral artery aneurysms. The median maximal aneurysm diameter was 12 mm (IQR=8.5–17.7), and the median neck diameter was 5.7 mm (IQR=4.4–7.9). Of the included patients, 68 (46%) were treated with PED Flex and 80 (54%) with PED Shield. PRU>230 before treatment was observed in 5% of patients. Adjunctive coiling was performed in 45% of cases. Multiple PEDs were used in 13% of cases. Repeat flow diversion was performed in 7% of cases. The median number of DWI-positive lesions was 5 (IQR=1–10).
Matching and covariate balance
Baseline characteristics of the covariates before and after matching are shown in table 1. Before matching, patients treated with PED Shield had lower rates of diabetes, higher rates of hyperlipidemia, and there was a greater proportion of smokers, and fewer cases of clopidogrel resistance. Additionally, aneurysms treated with PED Shield had larger maximum diameters, had wider necks, and were located in different anatomical regions. The discrimination ability of the estimated propensity scores, measured by the area under the receiver operating characteristic curve, was 0.76 (95% CI: 0.68 to 0.83). One-to-one propensity score matching created 47 matched pairs. After matching, baseline covariates were well balanced between the two groups, with an ASMD of <0.2, except for the adjunctive coiling and multiple PED implantation, which remained imbalanced. These variables were therefore adjusted for in subsequent analyses (eg, ANCOVA and logistic regression).
Table 1. Baseline characteristics of patients before and after propensity score matching.
| Variables | Before matching | After matching | ||||
|---|---|---|---|---|---|---|
| PED flex (n=68) | PED shield (n=80) | ASMD | PED flex (n=47) | PED shield (n=47) | ASMD | |
| Age, years | 65 (55–73) | 62 (51–69) | 0.19 | 61 (55–71) | 63 (52–70) | 0.06 |
| Female sex | 61 (90%) | 69 (86%) | 0.11 | 44 (94%) | 43 (92%) | 0.08 |
| Hypertension | 35 (51%) | 42 (53%) | 0.02 | 23 (49%) | 22 (47%) | 0.04 |
| Diabetes mellitus | 22 (32%) | 16 (20%) | 0.28 | 12 (26%) | 14 (30%) | 0.04 |
| Hyperlipidemia | 3 (4%) | 17 (21%) | 0.52 | 2 (4%) | 1 (2%) | 0.12 |
| Smoking | 18 (27%) | 35 (44%) | 0.37 | 15 (32%) | 16 (34%) | 0.05 |
| Maximum AN diameter, mm | 15.7 (10.5–20.8) | 10.8 (6.0–16.3) | 0.50 | 12.6 (10.1–19.7) | 10.9 (6.6–20.4) | 0.04 |
| AN neck diameter, mm | 6.9 (5.4–8.4) | 4.8 (4.0–7.0) | 0.50 | 6.7 (5.1–7.8) | 4.9 (4.3–7.4) | 0.17 |
| AN location | ||||||
| Cavernous | 30 (44%) | 20 (25%) | 0.41 | 17 (36%) | 17 (36%) | <0.001 |
| Paraclinoid | 31 (46%) | 34 (43%) | 0.06 | 24 (51%) | 23 (49%) | 0.04 |
| Posterior communicating artery | 6 (9%) | 16 (20%) | 0.32 | 5 (11%) | 6 (13%) | 0.07 |
| Anterior choroidal artery | 1 (1%) | 4 (5%) | 0.20 | 1 (2%) | 1 (2%) | <0.001 |
| Ophthalmic artery | 0 (0%) | 3 (4%) | 0.28 | 0 (0%) | 0 (0%) | <0.001 |
| Vertebral artery | 0 (0%) | 3 (4%) | 0.28 | 0 (0%) | 0 (0%) | <0.001 |
| Anterior circulation | 68 (100%) | 77 (96%) | 0.28 | 47 (100%) | 47 (100%) | <0.001 |
| Intradural | 38 (56%) | 60 (75%) | 0.41 | 30 (64%) | 30 (64%) | <0.001 |
| Branch-incorporated AN | 7 (10%) | 23 (29%) | 0.48 | 6 (13%) | 7 (15%) | 0.06 |
| PRU>230 | 5 (7%) | 2 (3%) | 0.23 | 2 (4%) | 1 (2%) | 0.12 |
| Adjunctive coiling | 19 (28%) | 48 (60%) | 0.68 | 14 (30%) | 23 (49%) | 0.40 |
| Multiple PED implantation | 6 (9%) | 13 (16%) | 0.23 | 3 (6%) | 11 (23%) | 0.49 |
| Repeat flow diversion | 8 (12%) | 3 (4%) | 0.30 | 2 (4%) | 1 (2%) | 0.12 |
Values shown are median (IQR) or number (%). ASMD of >0.2 was considered a meaningful difference.
AN, aneurysm; ASMD, absolute standardized mean difference; PED, Pipeline embolization device; PRU, platelet reactivity units.
Primary outcome
The primary outcome before and after matching is shown in table 2. In the unmatched cohort, the number of DWI-positive lesions was significantly lower in the PED Shield group than in the PED Flex group (median: 3 vs 9; β = −9.88 (95% CI: −13.65 to −6.11), P<0.001). After matching, this difference remained significant, with fewer DWI-positive lesions in the PED Shield group than in the PED Flex group (median: 3 vs 9; β = −11.40 (95% CI: −17.28 to −5.52), P<0.001).
Table 2. Comparison of primary and secondary outcomes before and after propensity score matching.
| Outcomes | Before matching | After matching | ||||||
|---|---|---|---|---|---|---|---|---|
| PED Flex (n=68) |
PED Shield (n=80) |
Effect size (95% CI) |
P value | PED Flex (n=47) |
PED Shield (n=47) |
Effect size* (95% CI) |
P value | |
| Primary outcome | ||||||||
| Number of DWI-positive lesions† | 9 (4–17) | 3 (1–6) | −9.88 (−13.65 to −6.11) | <0.001 | 9 (3–17) | 3 (1–6) | −11.40 (−17.28 to −5.52) | <0.001 |
| Secondary safety outcomes‡ | ||||||||
| Ischemic complications§ | 5 (7%) | 0 (0%) | 0.07 (<0.001 to 0.67) | 0.01 | 4 (9%) | 0 (0%) | 0.10 (<0.001 to 1.00) | 0.05 |
| Hemorrhagic complications | 8 (12%) | 2 (3%) | 0.19 (0.04 to 0.94) | 0.04 | 4 (9%) | 1 (2%) | 0.26 (0.02 to 2.69) | 0.27 |
| Technical events | 8 (12%) | 1 (1%) | 0.09 (0.01 to 0.78) | 0.03 | 6 (13%) | 1 (2%) | 0.24 (0.03 to 1.37) | 0.14 |
| Secondary efficacy outcomes† | ||||||||
| 6 month follow-up | n=66 | n=69 | n=46 | n=44 | ||||
| OKM grade C and D | 43 (65%) | 64 (93%) | 6.85 (2.42 to 19.40) | <0.001 | 33 (72%) | 41 (93%) | 4.86 (1.26 to 25.61) | 0.03 |
| 12 month follow-up | n=67 | n=60 | n=46 | n=42 | ||||
| OKM grade C and D | 51 (76%) | 56 (93%) | 4.39 (1.38 to 14.00) | 0.01 | 39 (85%) | 40 (95%) | 2.97 (0.59 to 23.20) | 0.23 |
The effect sizes for the primary and secondary outcomes were calculated after adjusting for the covariates adjunctive coiling and multiple PED implantation.
The primary outcome was compared using analysis of covariance, and the regression coefficient (β) was calculated.
Secondary safety and efficacy outcomes were compared using logistic regression analysis, and the OR was calculated.
Firth’s bias reduction method was used.
DWI, diffusion-weighted imaging; OKM, O'Kelly–Marotta; PED, Pipeline embolization device.
Secondary safety outcomes
The secondary safety outcomes before and after matching are shown in table 2. In the unmatched cohort, ischemic complications were significantly lower in the PED Shield group than in the PED Flex group (0% vs 7%; OR=0.07 (95% CI: <0.001 to 0.67), P=0.01). Hemorrhagic complications were also significantly lower in the PED Shield group (3% vs 12%; OR=0.19 (95% CI: 0.04 to 0.94), P=0.04). Additionally, technical events were significantly lower in the PED Shield group than in the PED Flex group (1% vs 12%; OR=0.09 (95% CI: 0.01 to 0.78), P=0.03).
After matching, ischemic complications remained significantly lower in the PED Shield group than in the PED Flex group (0% vs 9%; OR=0.10 (95% CI: <0.001 to 1.00), P=0.05). However, there was no significant difference in hemorrhagic complications between the groups (2% vs 9%; OR=0.26 (95% CI: 0.02 to 2.69), P=0.27). Similarly, no significant difference was observed in technical events between the groups (2% vs 13%; OR=0.24 (95% CI: 0.03 to 1.37), P=0.14).
Secondary efficacy outcomes
The secondary efficacy outcomes before and after matching are shown in table 2. In the unmatched cohort, the adequate aneurysm occlusion rate at 6 months was higher in the PED Shield group than in the PED Flex group (93% vs 65%; OR=6.85 (95% CI: 2.42 to 19.40), P<0.001). Similarly, at 12 months, the adequate aneurysm occlusion rate remained higher in the PED Shield group than in the PED Flex group (93% vs 76%; OR=4.39 (95% CI: 1.38 to 14.00), P=0.01).
After matching, the adequate aneurysm occlusion rate at 6 months remained significantly higher in the PED Shield group than in the PED Flex group (93% vs 72%; OR=4.86 (95% CI: 1.26 to 25.61), P=0.03). However, at 12 months, there was no significant difference between the groups (95% vs 85%; OR=2.97 (95% CI: 0.59 to 23.20), P=0.23).
Sensitivity analysis
Table 3 presents the sensitivity analysis results for outcomes, which were consistent with the primary analysis.
Table 3. Sensitivity analysis of primary and secondary outcomes.
| Outcomes | IPTW | ANCOVA and MVLR* | ||
|---|---|---|---|---|
| Effect size (95% CI) | P value | Effect size (95% CI) | P value | |
| Primary outcome | ||||
| Number of DWI-positive lesions | −12.28 (−18.97 to −5.59) | <0.001 | −10.12 (−14.39 to −5.85) | <0.001 |
| Secondary safety outcomes | ||||
| Ischemic complications† | 0.08 (<0.001 to 0.56) | 0.007 | 0.15 (0.001 to 1.71) | 0.14 |
| Hemorrhagic complications | 0.19 (0.04 to 0.97) | 0.045 | 0.29 (0.03 to 1.88) | 0.24 |
| Technical events | 0.22 (0.03 to 1.83) | 0.16 | 0.24 (0.03 to 1.78) | 0.15 |
| Secondary efficacy outcomes | ||||
| 6 month follow-up | ||||
| OKM grade C and D | 7.01 (2.30 to 21.38) | <0.001 | 7.38 (2.14 to 31.5) | 0.003 |
| 12 month follow-up | ||||
| OKM grade C and D | 5.42 (1.56 to 18.83) | 0.007 | 4.00 (0.94 to 16.98) | 0.06 |
The primary outcome was compared using analysis of covariance, and the regression coefficient (β) was calculated. Secondary safety and efficacy outcomes were compared using multivariable logistic regression analysis, and the odds ratio was calculated.
Firth’s bias reduction method was used.
ANCOVA, analysis of covariance; DWI, diffusion-weighted imaging; IPTW, inverse probability of treatment-weighting; MVLR, multivariable logistic regression; OKM, O'Kelly–Marotta; PED, Pipeline embolization device.
In the inverse probability of treatment-weighting analysis, the number of DWI-positive lesions was significantly lower in the PED Shield group (β = −12.28 (95% CI: −18.97 to −5.59), P<0.001). The PED Shield group also had fewer ischemic complications (OR=0.08 (95% CI: <0.001 to 0.56), P=0.007) and fewer hemorrhagic complications (OR=0.19 (95% CI: 0.04 to 0.97), P=0.045). However, there was no significant difference in technical events between the two groups (OR=0.22 (95% CI: 0.03 to 1.83), P=0.16). The PED Shield group demonstrated significantly higher adequate aneurysm occlusion rates at both 6 months (OR=7.01 (95% CI: 2.30 to 21.38), P<0.001) and 12 months (OR=5.42 (95% CI: 1.56 to 18.83), P=0.007).
In the analysis of covariance, the number of DWI-positive lesions remained significantly lower in the PED Shield group (β = −10.12 (95% CI: −14.39 to −5.85), P<0.001). In the multivariable logistic regression analysis, no significant differences were observed between the two groups for ischemic complications (OR=0.15 (95% CI: 0.001 to 1.71), P=0.14), hemorrhagic complications (OR=0.29 (95% CI: 0.03 to 1.88), P=0.24), or technical events (OR=0.24 (95% CI: 0.03 to 1.78), P=0.24). The PED Shield group had a significantly higher adequate aneurysm occlusion rate at 6 months (OR=7.38 (95% CI: 2.14 to 31.5), P=0.003), but there was no significant difference between the two groups at 12 months (OR=4.00 (95% CI: 0.94 to 16.98), P=0.06).
Discussion
In this propensity score-matched retrospective cohort study, the number of DWI-positive lesions following intracranial aneurysm treatment was significantly lower with the surface-modified PED Shield than with the non-surface-modified PED Flex. Because DWI-positive lesions have been suggested as potential markers of thrombogenicity, these findings confirm that the phosphorylcholine surface modification technology’s ability to inhibit thrombogenesis—previously demonstrated only in vitro and in vivo—also holds true in real-world clinical practice. There was a trend toward a lower rate of symptomatic ischemic complications in the PED Shield group; however, the difference was not statistically significant. There was no significant difference in hemorrhagic complications or technical events between the two groups. The adequate aneurysm occlusion rate was significantly higher in the PED Shield group at 6 months, but no difference was observed between the two groups at 12 months.
In two previous studies directly comparing ischemic complications after PED implantation with and without surface modification, the PED Shield did not significantly reduce ischemic complications.14 15 One possible explanation is that these studies primarily targeted small aneurysms with an average size of <6 mm. Prior research has shown that ischemic complication rates after flow diversion treatment are lower in small aneurysms than in large aneurysms,4 and the number of DWI-positive lesions post-procedure also decreases with smaller aneurysm sizes.25 26 Because the previous studies focused on patients with small aneurysms, ischemic complications in the non-surface-modified PED group were already low, minimizing the difference between the PED Shield and non-surface-modified PED groups. As a result, the sample size may have been insufficient to detect a significant difference. Another study comparing DWI-positive lesions after flow diversion treatment found no significant differences in the rate or number of DWI-positive lesions between the PED Shield and PED Flex groups.13 Although that study used propensity score matching to adjust for confounders, it did not account for aneurysm size, a key factor associated with DWI-positive lesions. This limitation, along with sample size constraints, may have contributed to the inability to demonstrate significant results favoring the PED Shield. By contrast, our study had a median maximum aneurysm diameter of 12 mm, with a predominance of large aneurysms, which are associated with a higher risk of ischemic complications and DWI-positive lesions. Therefore, the PED Shield’s ability to reduce thrombogenicity may have been more pronounced in this cohort. Supporting this hypothesis, a study comparing the proportion of patients with three or more DWI-positive lesions after treatment with PED Flex and PED Shield showed that the PED Shield group had significantly fewer such lesions on univariable analysis.27 The cohort in that study had a mean maximum aneurysm diameter of approximately 12 mm, similar to our cohort, further suggesting that the PED Shield’s surface modification is more effective in reducing thrombogenicity in larger aneurysms.
Our study has several strengths compared with previous research. First, it minimized potential confounding by performing propensity score matching using covariates previously reported to be associated with DWI positivity, including age, sex, aneurysm size, aneurysm location, and other key factors potentially influencing outcomes. Additionally, sensitivity analyses using alternative statistical methods were conducted to enhance internal validity and confirm the robustness of the primary analysis results. Second, high-resolution DWI with a 3 mm slice thickness was performed using 3T MRI, improving sensitivity for detecting ischemic lesions and ensuring more reliable findings. Third, first-generation PEDs with different delivery systems were excluded from the analysis. Because the PED Flex and PED Shield share the same delivery system, stent structure, and strand materials—differing only in the presence or absence of surface modification—this study allowed for a more direct comparison of the effects of surface modification.
Limitations
Our study has several limitations. First, as a retrospective cohort study conducted at a single center with a small sample size, there is a risk of selection bias, information bias, and limited external validity. Prospective studies or randomized controlled trials would be ideal to address these issues. Instead, a large-scale, multicenter retrospective study with appropriate adjustments for confounding factors would be more feasible. Second, postoperative DWI-positive lesions were evaluated at a single time point within 48 hours of the procedure, and new ischemic lesions on follow-up MRI, including DWI and fluid attenuated inversion recovery sequences, were not assessed. Third, although propensity score matching was used to adjust for as many confounders as possible, the potential influence of unmeasured confounders cannot be ruled out. One such factor is the learning curve, which encompasses procedural technique, patient selection, and device selection. However, because all PED placements in this study were performed by a single experienced operator who had completed approximately 30 PED Flex cases before the study began, it is unlikely that the learning curve significantly affected the results. Another potential confounder is that surface modification improves device deployment by reducing metal-to-metal friction and friction between the device and the microcatheter. While this modification is designed to reduce thrombogenicity, the improved deployability may also help minimize embolic events during deployment, potentially reducing the incidence of DWI-positive lesions. However, in this study, it was not possible to separately evaluate the effects of reduced thrombogenicity and improved deployability.
Conclusion
After adjusting for confounders, the PED Shield group had significantly fewer post-procedure DWI-positive lesions than the PED Flex group, suggesting that phosphorylcholine surface modification technology may reduce thrombogenicity in real-world clinical practice. No significant differences were observed between the two groups in terms of ischemic complications, hemorrhagic complications, or technical events. Additionally, while the occlusion rate at 6 months was higher in the PED Shield group than in the PED Flex group, no significant difference was observed at 12 months. These findings support the growing trend toward developing next-generation FDs incorporating various surface modification technologies.
Acknowledgements
We thank Angela Morben, DVM, ELS, from Edanz (https://jp.edanz.com/ac), for editing a draft of this manuscript.
Footnotes
Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Patient consent for publication: Not applicable.
Provenance and peer review: Not commissioned; externally peer reviewed.
Ethics approval: The study was approved by the Kyoto University Hospital institutional review board (ID R2088).
Data availability statement
Data are available upon reasonable request.
References
- 1.Brinjikji W, Murad MH, Lanzino G, et al. Endovascular treatment of intracranial aneurysms with flow diverters: a meta-analysis. Stroke. 2013;44:442–7. doi: 10.1161/STROKEAHA.112.678151. [DOI] [PubMed] [Google Scholar]
- 2.Kallmes DF, Brinjikji W, Cekirge S, et al. Safety and efficacy of the Pipeline embolization device for treatment of intracranial aneurysms: a pooled analysis of 3 large studies. J Neurosurg. 2017;127:775–80. doi: 10.3171/2016.8.JNS16467. [DOI] [PubMed] [Google Scholar]
- 3.Dandapat S, Mendez-Ruiz A, Martínez-Galdámez M, et al. Review of current intracranial aneurysm flow diversion technology and clinical use. J Neurointerv Surg. 2021;13:54–62. doi: 10.1136/neurintsurg-2020-015877. [DOI] [PubMed] [Google Scholar]
- 4.Hanel RA, Cortez GM, Lopes DK, et al. Prospective study on embolization of intracranial aneurysms with the pipeline device (PREMIER study): 3-year results with the application of a flow diverter specific occlusion classification. J Neurointerv Surg. 2023;15:248–54. doi: 10.1136/neurintsurg-2021-018501. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Li Y-L, Roalfe A, Chu EY-L, et al. Outcome of Flow Diverters with Surface Modifications in Treatment of Cerebral Aneurysms: Systematic Review and Meta-analysis. AJNR Am J Neuroradiol . 2021;42:327–33. doi: 10.3174/ajnr.A6919. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Zoppo CT, Mocco J, Manning NW, et al. Surface modification of neurovascular stents: from bench to patient. J Neurointerv Surg. 2024;16:908–13. doi: 10.1136/jnis-2023-020620. [DOI] [PubMed] [Google Scholar]
- 7.Ma L, Hoz SS, Al-Bayati AR, et al. Flow Diverters with Surface Modification in Patients with Intracranial Aneurysms: A Systematic Review and Meta-Analysis. World Neurosurg. 2024;185:320–6. doi: 10.1016/j.wneu.2023.12.132. [DOI] [PubMed] [Google Scholar]
- 8.Girdhar G, Ubl S, Jahanbekam R, et al. Thrombogenicity assessment of Pipeline, Pipeline Shield, Derivo and P64 flow diverters in an in vitro pulsatile flow human blood loop model. eNeurologicalSci. 2019;14:77–84. doi: 10.1016/j.ensci.2019.01.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Marosfoi M, Clarencon F, Langan ET, et al. Acute thrombus formation on phosphorilcholine surface modified flow diverters. J NeuroIntervent Surg. 2018;10:406–11. doi: 10.1136/neurintsurg-2017-013175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Martínez-Galdámez M, Lamin SM, Lagios KG, et al. Treatment of intracranial aneurysms using the pipeline flex embolization device with shield technology: angiographic and safety outcomes at 1-year follow-up. J Neurointerv Surg. 2019;11:396–9. doi: 10.1136/neurintsurg-2018-014204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Rice H, Martínez Galdámez M, Holtmannspötter M, et al. Periprocedural to 1-year safety and efficacy outcomes with the Pipeline Embolization Device with Shield technology for intracranial aneurysms: a prospective, post-market, multi-center study. J Neurointerv Surg. 2020;12:1107–12. doi: 10.1136/neurintsurg-2020-015943. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Chia GS, de Villiers L, Carraro do Nascimento V, et al. Safety and Clinical Effectiveness of Pipeline Shield Device for Intracranial Aneurysms in an Australian Cohort (SCOPE‐AUS) SVIN . 2022;2:1–8. doi: 10.1161/SVIN.121.000292. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Cortez GM, Benalia VHC, Sauvageau E, et al. Diffusion-weighted imaging lesions after intracranial aneurysm treatment with Pipeline Flex and Pipeline Flex with Shield technology: a retrospective cohort analysis. J Neurointerv Surg. 2024;16:385–91. doi: 10.1136/jnis-2023-020250. [DOI] [PubMed] [Google Scholar]
- 14.El Naamani K, Mastorakos P, Yudkoff CJ, et al. Comparison of PED/PED Flex and PED Shield in the treatment of unruptured intracerebral aneurysms. J Neurosurg. 2024;140:436–40. doi: 10.3171/2023.7.JNS23981. [DOI] [PubMed] [Google Scholar]
- 15.Ramirez-Velandia F, Enriquez-Marulanda A, Filo J, et al. Comparison of Thromboembolic Events Between Pipeline Embolization Device (PED) Shield and PED/PED Flex: A Propensity Score-Matched Analysis. Neurosurgery. 2024;95:330–8. doi: 10.1227/neu.0000000000002883. [DOI] [PubMed] [Google Scholar]
- 16.García-Núñez J, D Vera D, Ortiz AF, et al. A Study on the Efficacy and Safety of Pipeline Shield Stents and Nonmodified Surface Flow-Diverting Stents in Patients with Intracranial Aneurysms. World Neurosurg. 2024;183:e781–6. doi: 10.1016/j.wneu.2024.01.028. [DOI] [PubMed] [Google Scholar]
- 17.Habtezghi AB, Ghozy S, Bilgin C, et al. DWI-Detected Ischemic Lesions after Endovascular Treatment for Cerebral Aneurysms: An Updated Systematic Review and Meta-analysis. AJNR Am J Neuroradiol. 2023;44:1256–61. doi: 10.3174/ajnr.A8024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kang D-H, Kim BM, Kim DJ, et al. MR-DWI-positive lesions and symptomatic ischemic complications after coiling of unruptured intracranial aneurysms. Stroke. 2013;44:789–91. doi: 10.1161/STROKEAHA.112.669853. [DOI] [PubMed] [Google Scholar]
- 19.Pierot L, Lamin S, Barreau X, et al. Coating (Coating to Optimize Aneurysm Treatment in the New Flow Diverter Generation) study. The first randomized controlled trial evaluating a coated flow diverter (p64 MW HPC): study design. J Neurointerv Surg. 2023;15:684–8. doi: 10.1136/neurintsurg-2022-018969. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Price MJ, Endemann S, Gollapudi RR, et al. Prognostic significance of post-clopidogrel platelet reactivity assessed by a point-of-care assay on thrombotic events after drug-eluting stent implantation. Eur Heart J. 2008;29:992–1000. doi: 10.1093/eurheartj/ehn046. [DOI] [PubMed] [Google Scholar]
- 21.Price MJ, Berger PB, Teirstein PS, et al. Standard- vs high-dose clopidogrel based on platelet function testing after percutaneous coronary intervention: the GRAVITAS randomized trial. JAMA. 2011;305:1097–105. doi: 10.1001/jama.2011.290. [DOI] [PubMed] [Google Scholar]
- 22.O’kelly CJ, Krings T, Fiorella D, et al. A novel grading scale for the angiographic assessment of intracranial aneurysms treated using flow diverting stents. Interv Neuroradiol. 2010;16:133–7. doi: 10.1177/159101991001600204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Firth D. Bias reduction of maximum likelihood estimates. Biometrika. 1995;82:667. doi: 10.1093/biomet/82.3.667-b. [DOI] [Google Scholar]
- 24.Heinze G, Schemper M. A solution to the problem of separation in logistic regression. Stat Med. 2002;21:2409–19. doi: 10.1002/sim.1047. [DOI] [PubMed] [Google Scholar]
- 25.Nariai Y, Takigawa T, Hyodo A, et al. Thromboembolic Events Detected by Diffusion-Weighted Magnetic Resonance Imaging after Flow Diverter Treatment: The Impact of Procedure Time. World Neurosurg. 2022;167:e1241–52. doi: 10.1016/j.wneu.2022.09.010. [DOI] [PubMed] [Google Scholar]
- 26.Heller RS, Dandamudi V, Lanfranchi M, et al. Effect of antiplatelet therapy on thromboembolism after flow diversion with the pipeline embolization device. J Neurosurg. 2013;119:1603–10. doi: 10.3171/2013.7.JNS122178. [DOI] [PubMed] [Google Scholar]
- 27.Kawamoto S, Ozaki T, Asai K, et al. Treatment Outcomes of PED for Unruptured Aneurysms of Internal Carotid Artery: Comparison of PED-Flex and PED-Shield. Neurol Med Chir(Tokyo) 2024;64:316–22. doi: 10.2176/jns-nmc.2024-0034. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data are available upon reasonable request.