You are here

  • Home
  • Archive
  • Volume 7, Issue 3
  • Thromboembolic complications with Pipeline Embolization Device placement: impact of procedure time, number of stents and pre-procedure P2Y12 reaction unit (PRU) value

Article Text

Article menu

Download PDFPDF

New devices
Original research
Thromboembolic complications with Pipeline Embolization Device placement: impact of procedure time, number of stents and pre-procedure P2Y12 reaction unit (PRU) value
  1. Lee A Tan,
  2. Kiffon M Keigher,
  3. Stephan A Munich,
  4. Roham Moftakhar,
  5. Demetrius K Lopes
  1. Department of Neurosurgery, Rush University Medical Center, Chicago, Illinois, USA
  1. Correspondence to Dr Lee A. Tan, Department of Neurosurgery, Rush University Medical Center, 1725 W. Harrison St. Suite 855, Chicago, IL 60612, USA; lee_tan{at}rush.edu

Abstract

Background Thromboembolic events after Pipeline Embolization Device (PED) placement remain a feared complication among neuroendovascular surgeons. This study aimed to investigate potential risk factors for thromboembolic events in patients undergoing PED placement.

Methods Medical records of patients who underwent PED placement from April 2011 to August 2013 were reviewed. Variables including pre-procedure P2Y12 reaction unit (PRU) value, procedure time, number of PEDs deployed and perioperative neurovascular complications were recorded. Multivariate analysis was performed to identify risk factors for perioperative thromboembolic complications.

Results Seventy-four patients were identified. Six patients (8.1%) had changes in neurological status after PED placement including five (6.8%) thromboembolic complications and one (1.4%) delayed intracranial hemorrhage; 50.9% of patients had diffusion-weighted imaging (DWI) changes on post-procedural MRI. Longer procedure time (>116 min) and multiple PED placements (>1) were statistically significant risk factors for symptomatic thromboembolic events (p<0.01). A pre-procedural PRU value >208 had an OR of 11.32 (95% CI 0.06 to 212.57) for symptomatic thromboembolic complications, but the result was not statistically significant.

Conclusions DWI changes on MRI occurred at a much higher rate than new neurological symptoms following PED placement. Longer procedure time and multiple PED deployment are associated with higher risks of new neurological changes due to thromboembolic events. There was a trend for an increased risk of a symptomatic thromboembolic event in patients with pre-procedural PRU values >208. Reloading (clopidogrel 600 mg) patients with preoperative PRU >208 was safe and may have a protective effect on thromboembolic events.

  • Platelets
  • Flow Diverter
  • Complication
  • Embolic
  • Stent

https://doi.org/10.1136/neurintsurg-2014-011111

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    With recent advancements in neuroendovascular technology, flow-diverting stents such as the Pipeline Embolization Device (PED; ev3, Irvine, California, USA) provide a new therapeutic option in the treatment of complex cerebral aneurysms. The PED is a tightly braided, self-expanding, cobalt, chromium and platinum stent with 30–35% metal surface area coverage which promotes aneurysm thrombosis while maintaining blood flow to perforating branches. Since FDA approval in 2011, the PED has gained tremendous popularity and is becoming an indispensable tool for neuroendovascular surgeons in treating complex intracranial aneurysms. The higher metal surface area provided by these stents acts as a scaffold to facilitate endothelialization and subsequent total intraluminal reconstruction of the diseased vessel while maintaining the patency of the perforating vessels. However, endovascularization of the luminal surface occurs in a delayed fashion and, despite the routine use of dual antiplatelet therapy, stent stenosis and thromboembolic events after PED placement remain feared complications with up to 14% symptomatic cerebral infarctions and 52% diffusion-weighted imaging (DWI) changes on MRI reported from other single-center experiences.1 ,2

    Among other risk factors, the variable therapeutic responses to the standard dual antiplatelet regimen have been linked to thromboembolic complications after PED placement.1 ,3 The most common and widely used dual antiplatelet regimen is a combination of aspirin and clopidogrel. Aspirin irreversibly inactivates the cyclo-oxygenase (COX) enzyme, which in turn decreases prostaglandin and thrombaxane production and therefore decreases platelet activation. Studies have shown that aspirin has a relatively uniform response profile with about 95% target enzyme inhibition with appropriate dosing. Aspirin Reaction Units (ARUs) are sometimes used to measure aspirin activity with a commonly accepted therapeutic range of 350–549, while off-aspirin ARU values are usually in the range of 550–700. Unlike aspirin, clopidogrel has been found to have a variable response profile with a high percentage of low or non-responders, which is thought to be due to the metabolism and pharmacokinetics of clopidogrel. Clopidogrel is a prodrug that is converted into its active form in the liver by the cytochrome P450 system. Most of the absorbed clopidogrel is hydrolyzed into an inactive metabolite by carboxylase and only 10–15% is converted into the active form. The active metabolite irreversibly binds to P2Y12 receptors (ADP receptors) on the platelet surface, which prevents further ADP granule secretion from platelets and prevents amplification of platelet activation and aggregation.4 Factors contributing to the variable levels of active metabolite generation may include variable intestinal absorption, drug–drug interaction and P450 enzyme gene polymorphism.

    VerifyNow (Accumetrics, San Diego, California, USA) is a point-of-care platelet function test that measures the degree of P2Y12 receptor inhibition after stimulation with ADP. This assay has been found to correlate strongly with light transmittance aggregometry (LTA), which is the historical reference method for platelet reactivity in assessing P2Y12 receptor inhibition. VerifyNow results are reported in P2Y12 reaction units (PRUs), with a lower PRU value corresponding to a higher degree of P2Y12 receptor inhibition and, hence, a decreased likelihood of platelet activation and aggregation. Evidence from multiple prospective randomized studies in the cardiac literature has shown that individually tailored antiplatelet therapy guided by platelet function assays is associated with a better clinical outcome.5–8 Specifically, the GRAVITAS (Gauging Responsiveness With A VerifyNow P2Y12 Assay: Impact on Thrombosis and Safety) trial has shown that a peri-procedural PRU of <208 is associated with a lower rate of stent thrombosis and adverse cardiac events.7 Such a prospective study does not yet exist in the neuroendovascular literature. Routine measurement of platelet reactivity before PED placement has not been widely implemented in clinical practice by neurointerventionalists. Even though a recent retrospective study found a pre-procedure PRU value of >240 to be an independent predictor for thromboembolic complications in patients undergoing PED treatment, there is not yet a consensus on the optimal pre-procedural PRU value. This study aims to investigate the relationship between pre-procedure PRU and other potential risk factors for thromboembolic events in patients undergoing PED placement for treatment of cerebral aneurysms.

    Methods

    Medical record review

    The medical records of patients who underwent PED placement by the two senior authors (DKL, RM) from April 2011 to August 2013 were reviewed. Specific variables including age, gender, smoking status, aneurysm characteristics, pre-procedure antiplatelet regimen, pre-procedure PRU value measured by VerifyNow, pre-procedure ARU, procedure time, number of PEDs deployed and perioperative neurovascular complications up to 4 weeks after the PED procedure were recorded and analyzed. Perioperative thromboembolic complications were assessed by both clinical examination and MRI with DWI sequences performed within 24 h after the procedure. The MRIs were reviewed by a board-certified neuroradiologist and a neuroendovascular surgeon. All neurological deficits evident on neurological examinations up to 4 weeks after PED placement were documented. Major events were defined as symptoms that were persistent at 4 weeks and minor events were considered symptoms that were significantly improved or resolved at follow-up.

    Dual antiplatelet therapy protocol

    Dual antiplatelet therapy (clopidogrel 75 mg daily and aspirin 325 mg daily) was started 5 days prior to the planned procedure. Alternatively, some patients were loaded with clopidogrel 600 mg and aspirin 325 mg at least 2 h prior to surgery. PY2P12 receptor inhibition was assessed prior to the procedure by measuring PRU values using VerifyNow and the aspirin activity was assessed by measuring ARU. We used the most widely accepted cut-off values of ARU <550 and PRU <230 to determine whether a patient had an adequate response to the dual antiplatelet regimen (prior to the newest cardiac guideline of PRU <2087). All patients were also heparinized to a goal activated clotting time (ACT) of 250 s during PED deployment.

    In patients where either the pre-procedural PRU and/or ARU are not therapeutic, it is our practice to reload them with clopidogrel 600 mg and/or aspirin 325 mg immediately after the procedure before the ACT has normalized. If a patient is a hyporesponder to both clopidogrel and aspirin (ie, ARU >550 and PRU >230), it is also our practice to start an intra-arterial loading dose of abciximab (0.25 mg/kg) at the time of PED placement, followed by a 12 h infusion (0.125 μg/kg/min or 10 μg/min). In our practice, during the period of this study we did not cancel or postpone any intervention based on the results of the platelet activation test.

    Statistical analysis

    Multivariate analysis was performed to identify risk factors for both symptomatic thromboembolic complications and DWI changes on MRI. All reported p values are two-tailed; p<0.05 was considered statistically significant. All statistical analysis was done using ‘R’ statistical software (http://www.r-project.org).

    Results

    Seventy-four patients who received PED treatment during the study period were identified. The mean age was 60 years (range 28–85), with 12 men and 62 women. Among them, 17 were smokers and 57 were non-smokers. There were 58 patients with anterior circulation aneurysms and 16 patients with posterior circulation aneurysms. The mean size of the aneurysms was 9.87 mm; 46 patients had aneurysms <10 mm and 28 patients had aneurysms >10 mm. The mean pre-procedure PRU value was 210 (range 2–450), the mean ARU value was 446.5 (range 339–665), the mean procedure time was 103 min (range 12–364) and the mean number of PEDs used was 1.62 (range 1–8). A single PED was used in 50 patients while 24 patients received ≥2 PEDs during their treatment.

    Six patients (8.1%) had changes in neurological status after aneurysm repair, five of which (6.8%) were due to thromboembolic complications and one (1.4%) was due to a delayed intracranial hemorrhage 1 week after the procedure. The five thromboembolic complications included two major and three minor neurological deficits. The two major thromboembolic complications consisted of hemiparesis in both patients and the three minor thromboembolic complications consisted of one patient with transient dizziness and double vision, another patient with transient partial visual loss and a third patient with transient left hand apraxia. The major complication due to delayed hemorrhage caused left side hemi-neglect in one patient. During the study period, 57 patients underwent MRI within 24 h after the procedure; 29 (50.9%) had DWI changes on post-procedural MRI (table 1).

    View this table:
    Table 1

    Pre-procedure PRU and peri-procedural complications up to 4 weeks after PED placement

    Multivariate analysis was performed to identify potential risk factors for peri-procedural neurological events and odd ratios were calculated for each variable (table 2). Longer procedure time (>116 min) and multiple PED placements (>1) were identified as statistically significant risk factors for symptomatic thromboembolic complications (p<0.01; figures 1 and 2). Female gender appeared to be a risk factor for DWI changes on MRI (p=0.03) but not for symptomatic stroke (p=0.59). Although a PRU value >208 did not reach statistical significance as an independent risk factor for symptomatic stroke (p=0.06, figure 3) or DWI changes (p=0.79, figure 4), it demonstrated a trend toward an increase in ischemic events with neurological changes. No other variable studied conferred a risk for neurological events with statistical significance.

    View this table:
    Table 2

    Risk factors and thromboembolic complications up to 4 weeks after PED placement

    Figure 1
    Figure 1

    Box plot for procedure time and thromboembolic complications.

    Figure 2
    Figure 2

    Box plot for number of Pipeline Embolization Devices (PEDs) deployed and thromboembolic complications.

    Figure 3
    Figure 3

    Box plot for P2Y12 reaction unit (PRU) and symptomatic stroke.

    Figure 4
    Figure 4

    Box plot for P2Y12 reaction unit (PRU) and diffusion-weighted imaging (DWI) changes.

    Discussion

    In our cohort, six neurological complications occurred after PED placement, representing 8.1% of the patients. Five of the six neurological complications (6.8%) were due to thromboembolic events and DWI changes were detected in 50.9% of patients on post-procedural MRI; both of these are consistent with previously reported complication rates.1 ,2 ,9 ,10 By performing multivariate analysis, we identified two statistically significant risk factors for neurological changes due to thromboembolic events: longer procedure time (>116 min) and multiple PED deployment (>1 stent). It is intuitive that a longer procedure time may confer a higher risk for thromboembolic events, given that there is a longer catheter time and mechanical manipulation, so more risk for activation of platelets and initiation of coagulation cascade. Longer procedure time also often indicates that a particular case is more technically challenging, either with difficult access or complex aneurysm morphologies. Multiple PED deployment is also identified as a risk factor. This is not surprising given that cases requiring multiple PEDs are often more complex with larger aneurysms, and multiple PEDs equate to more thrombogenic surface for platelet activation. In our cohort, the average number of PEDs used in patients with aneurysms <10 mm was 1.19 compared with an average of 2.32 PEDs used in patients with aneurysms >10 mm. Furthermore, when overlapping PEDs are used, the overlapped region will probably have a metal to artery ratio of >30% (since each PED has ∼30% metal coverage), and this may be an additional risk factor for mechanical obstruction of jailed perforators.

    It is noteworthy that DWI changes on post-procedural MRI were present in over half of the patients (50.9%) with PED placement. A recent study by Heller et al2 reported a similar rate of DWI changes (52%) after PED placement in a cohort of 23 patients. The authors did not identify the inadequate response to clopidogrel as a risk factor for post-procedure DWI changes. The presence of the high rate of DWI changes led the authors to postulate that the commonly observed hemorrhagic complications after PED placement may be due to hemorrhagic conversion of ischemic areas. However, our data do not seem to support this theory since there was only one hemorrhagic complication (1.4%) in our cohort despite the high rate of DWI changes (50.9%). Also of interest is the fact that our patient with the hemorrhagic complication, consistent with previous reports,11 ,12 had a combination of very low PRU of 67 (hypersensitive to clopidogrel) and DWI changes. After this case we become more concerned about dual antiplatelet therapy in clopidogrel hyper-responders with DWI changes. This is only an observation at this point and needs to be further investigated. Our data support the finding by Heller et al that the pre-procedural PRU and rate of DWI changes on postoperative MRI did not have any statistically significant correlation. It is also interesting to note that female gender was a statistically significant risk factor for DWI changes after PED placement in our cohort. We do not have a definitive explanation for the increased risk of DWI changes; however, it has been known in the cardiac literature that female gender is an independent risk factor for strokes in patients with atrial fibrillation.13 Several mechanisms have been proposed to explain this phenomenon including hormonal therapy, menopause and sex differences in prothrombotic and inflammatory biomarkers, platelet aggregation and hypercoagulable states.13 We suspect that similar factors may have contributed to the higher risk of DWI changes after PED placement in women.

    Of the 74 patients, there were 35 (47.3%) with pre-procedure PRU <208 (mean=106) and 39 (52.7%) with PRU >208 (mean=290). The percentage of clopidogrel hyporesponders was 52.7% (using PRU cut-off of 208), which is slightly higher than previously reported values (32.9–42.9%).14 ,15 This slight discrepancy can be explained by the fact that the previous studies were using higher PRU cut-off values than 208. Although there was a trend towards more neurological changes in the group with PRU >208, it was not statistically significant (p=0.06). This result is slightly different from a recent report by Delgado Almandoz et al1 in which the authors found PRU >240 to be a statistically significant independent risk factor for thromboembolic events in their cohort of 44 patients. This may be due to the fact that our study is underpowered, although a p value of 0.06 is very close to approaching statistical significance. In addition, the incidence of ischemic events in our patients with PRU >208 may also have been lowered due to our routine of reloading these patients with clopidogrel 600 mg and aspirin 325 mg immediately after the procedure and before the ACT had normalized. However, a limitation in this analysis is the fact that we do not reassess platelet reactivity after the additional loading dose. While this may improve the pharmacodynamic effect of the drug, studies have shown that many patients may still not demonstrate a measurable effect of the drug after the additional dose.16 Therefore, even though our data did not reveal PRU >208 as a statistically significant risk factor for symptomatic thromboembolic complications, from our clinical experience and the available literature we strongly believe that the pre-procedure PRU value does have a significant effect on the rate of symptomatic strokes after PED procedures. We also feel that the practice of checking the preoperative PRU value and reloading patients after the procedure is a safe strategy and may have prevented more thromboembolic events in our series. This could explain why there was no statistical significance in our series compared with previous reports1 and the cardiology literature. This is only an observation that needs to be further investigated. We also did not check PRU levels after reloading to confirm the effect of the oral reloading dose of 600 mg clopidogrel.

    Evidence regarding the use of alternative P2Y12 inhibitors such as prasugrel or ticargrelor in clopidogrel hyporesponders have been mixed, with results ranging from no increase in hemorrhagic complications to an increased rate of death.14 ,17 ,18 We therefore have not adapted these drugs as part of our routine antiplatelet protocol in our clinical practice. With the increasing availability of PRU testing, we believe that there will be mounting evidence to support personalized antiplatelet regimens in an effort to reduce perioperative thromboembolic events associated with PED procedures and stent placement in general.

    There are several limitations to this study. Given its retrospective nature, potential confounding factors and bias inherent to all retrospective studies are unavoidable. Perioperative complications are probably multifactorial, and there may have been risk factors that we have not identified or included in the statistical analysis. For instance, as with all new technologies, there may be a learning curve with PED placement and this may be a major confounding factor. Furthermore, there was an increased rate of complications in the posterior circulation with PED placement; however, given the relatively small number of patients with posterior circulation aneurysms treated by PED, the study did not have enough power to detect a statistically significant difference. Additionally, the patients in our cohort were treated by two different endovascular surgeons, and differences in technique and personal experience may also have confounded the results.

    Conclusion

    DWI changes on MRI after PED repair of cerebral aneurysms were identified in 50.9% of the cases studied. Longer procedure time (>116 min) and multiple PED deployment were associated with higher risks of neurological changes. A pre-procedural PRU value >208 had an increasing trend towards symptomatic ischemic events. The practice of reloading (clopidogrel 600 mg) patients with preoperative PRU >208 was safe and may have decreased the number of thromboembolic events in our series. No patients had thromboembolic complications with PRU <208. Low PRU values in patients with DWI changes may be associated with an increased risk of hemorrhage. Larger prospective studies are needed to further validate our findings and to improve patient safety with PED procedures.

    References

    1. ↵
      1. Delgado Almandoz JE,
      2. Crandall BM,
      3. Scholz JM,
      4. et al
      . Pre-procedure P2Y12 reaction units value predicts perioperative thromboembolic and hemorrhagic complications in patients with cerebral aneurysms treated with the Pipeline Embolization Device. J Neurointerv Surg 2013;5(Suppl 3):iii3–10.
      OpenUrlAbstract/FREE Full Text
    2. ↵
      1. Heller RS,
      2. Dandamudi V,
      3. Lanfranchi M,
      4. et al
      . Effect of antiplatelet therapy on thromboembolism after flow diversion with the Pipeline Embolization Device: clinical article. J Neurosurg 2013;119:1603–10.
      OpenUrlPubMedWeb of Science
    3. ↵
      1. Delgado Almandoz JE,
      2. Crandall BM,
      3. Scholz JM,
      4. et al
      . Last-recorded P2Y12 reaction units value is strongly associated with thromboembolic and hemorrhagic complications occurring up to 6 months after treatment in patients with cerebral aneurysms treated with the pipeline embolization device. AJNR Am J Neuroradiol 2014;35:128–35.
      OpenUrlAbstract/FREE Full Text
    4. ↵
      1. Bonello L,
      2. Tantry US,
      3. Marcucci R,
      4. et al
      . Consensus and future directions on the definition of high on-treatment platelet reactivity to adenosine diphosphate. J Am Coll Cardiol 2010;56:919–33.
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Bonello L,
      2. Camoin-Jau L,
      3. Armero S,
      4. et al
      . Tailored clopidogrel loading dose according to platelet reactivity monitoring to prevent acute and subacute stent thrombosis. Am J Cardiol 2009;103:5–10.
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Bonello L,
      2. Camoin-Jau L,
      3. Arques S,
      4. et al
      . Adjusted clopidogrel loading doses according to vasodilator-stimulated phosphoprotein phosphorylation index decrease rate of major adverse cardiovascular events in patients with clopidogrel resistance. J Am Coll Cardiol 2008;51:1404–11.
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Price MJ,
      2. Angiolillo DJ,
      3. Teirstein PS,
      4. et al
      . Platelet reactivity and cardiovascular outcomes after percutaneous coronary intervention: a time-dependent analysis of the Gauging Responsiveness With a VerifyNow P2Y12 Assay: Impact on Thrombosis and Safety (GRAVITAS) Trial. Circulation 2011;124:1132–7.
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Price MJ,
      2. Berger PB,
      3. Angiolillo DJ,
      4. et al
      . Evaluation of individualized clopidogrel therapy after drug-eluting stent implantation in patients with high residual platelet reactivity: design and rationale of the GRAVITAS trial. Am Heart J 2009;157:818–24.e1.
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Kan P,
      2. Siddiqui AH,
      3. Veznedaroglu E,
      4. et al
      . Early postmarket results after treatment of intracranial aneurysms with the pipeline embolization device: a US multicenter experience. Neurosurgery 2012;71:1080–8.
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Yu SC-H,
      2. Kwok C-K,
      3. Cheng P-W,
      4. et al
      . Intracranial aneurysms: midterm outcome of Pipeline embolization device—a prospective study in 143 patients with 178 aneurysms. Radiology 2012;265:893–901.
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Campo G,
      2. Parrinello G,
      3. Ferraresi P,
      4. et al
      . Prospective evaluation of on-clopidogrel platelet reactivity over time in patients treated with percutaneous coronary intervention. J Am Coll Cardiol 2011;57:2474–83.
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Sibbing D,
      2. Schulz S,
      3. Braun S,
      4. et al
      . Antiplatelet effects of clopidogrel and bleeding in patients undergoing coronary stent placement. J Thromb Haemost 2010;8:250–6.
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Cove CL,
      2. Albert CM,
      3. Andreoti F,
      4. et al
      . Female sex as an independent risk factor for stroke in atrial fibrillation: possible mechanisms. Thromb Haemost 2013;111(3). doi:10.1160/TH13-04-0347
    14. ↵
      1. Akbari SH,
      2. Reynolds MR,
      3. Kadkhodayan Y,
      4. et al
      . Hemorrhagic complications after prasugrel (Effient) therapy for vascular neurointerventional procedures. J Neurointerv Surg 2012;5:337–43.
      OpenUrlPubMed
    15. ↵
      1. Lee DH,
      2. Arat A,
      3. Morsi H,
      4. et al
      . Dual antiplatelet therapy monitoring for neurointerventional procedures using a point-of-care platelet function test: a single-center experience. Am J Neuroradiol 2008;29:1389–94.
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Bonello L,
      2. Armero S,
      3. Ait Mokhtar O,
      4. et al
      . Clopidogrel loading dose adjustment according to platelet reactivity monitoring in patients carrying the 2C19*2 loss of function polymorphism. J Am Coll Cardiol 2010;56:1630–6.
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Kastrati A
      . New anti-platelet agents: the end of resistance? Thromb Res 2012;130(Suppl 1):S53–5.
      OpenUrlPubMed
    18. ↵
      1. Nordeen JD,
      2. Patel AV,
      3. Darracott RM,
      4. et al
      . Clopidogrel resistance by P2Y12 platelet function testing in patients undergoing neuroendovascular procedures: incidence of ischemic and hemorrhagic complications. J Vasc Interv Neurol 2013;6:26–34.
      OpenUrlPubMed

    Footnotes

    • Contributors LAT, RM and DKL were responsible for the conception and design of the study; KMK and SAM were responsible for acquisition of data; LAT and DKL conducted the analysis and interpretation of data; LAT drafted the article; all authors were involved in critically revising the manuscript and final approval of the version to be published.

    • Competing interests RM is a consultant for Covidien. DKL has financial and research relationships with Covidien, Stryker and Penumbra. The other authors have no competing interests to disclose.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Data sharing statement The data included in this study came from our single institution and are not shared with any other party.