Author + information
- Received November 29, 2016
- Revision received February 23, 2017
- Accepted March 23, 2017
- Published online June 19, 2017.
- Christoph Knappich, MDa,
- Andreas Kuehnl, MD, MPH, MBAa,
- Pavlos Tsantilas, MDa,
- Sofie Schmid, MDa,
- Thorben Breitkreuzb,
- Michael Kallmayer, MDa,
- Alexander Zimmermann, MD, MHBAa and
- Hans-Henning Eckstein, MD, PhDa,∗ ()
- aDepartment of Vascular and Endovascular Surgery, Klinikum rechts der Isar, Technical University of Munich, Munich, Germany
- bAQUA-Institut für Angewandte Qualitätsförderung und Forschung im Gesundheitswesen GmbH, Göttingen, Germany
- ↵∗Address for correspondence:
Univ.-Prof. Dr. Hans-Henning Eckstein, Department of Vascular and Endovascular Surgery, Klinikum Rechts der Isar, Technical University of Munich, Ismaninger Strasse 22, 81675 Munich, Germany.
Objectives The aim of this study was to analyze the association between intraprocedural and periprocedural variables and in-hospital stroke or death rate after carotid artery stenting.
Background In Germany, all open surgical and endovascular procedures on the extracranial carotid artery must be documented in a statutory nationwide quality assurance database.
Methods A total of 13,086 carotid artery stenting procedures for asymptomatic (63.9%) or symptomatic carotid stenosis (mean age 69.7 years, 69.7% men) between 2009 and 2014 were recorded. The following variables were analyzed: stent design, stent material, neurophysiological monitoring, periprocedural antiplatelet medication, and use of an embolic protection device. The primary outcome was in-hospital stroke or death. Major stroke or death, any stroke, and death, all until discharge, were secondary outcomes. Adjusted relative risks (RRs) were assessed using multilevel multivariable regression analyses.
Results The primary outcome occurred in 2.4% of the population (1.7% in asymptomatic and 3.7% in symptomatic patients). The multivariable analysis showed an independent association between the use of an embolic protection device and lower in-hospital rates of stroke or death (adjusted RR: 0.65; 95% confidence interval [CI]: 0.50 to 0.85), major stroke or death (adjusted RR: 0.60; 95% CI: 0.43 to 0.84), and stroke (adjusted RR: 0.57; 95% CI: 0.43 to 0.77). Regarding the occurrence of in-hospital death, there was no significant association (adjusted RR: 0.78; 95% CI: 0.46 to 1.35). None of the outcomes was associated with stent design, stent material, neurophysiological monitoring, or antiplatelet medication.
Conclusions The use of an embolic protection device was independently associated with lower in-hospital risk for stroke or death, major stroke or death, and stroke.
First performed in 1953 by Michael E. DeBakey (1), carotid endarterectomy (CEA) has been considered the standard therapy for stenosis of the extracranial portion of the carotid artery for more than half a century. With the first balloon angioplasty having been described in 1980 (2), a cornerstone for the emergence of a novel therapy was laid. In the past 20 years, the implantation of stents (3) and the use of embolic protection devices (EPDs) (4) have helped endovascular therapy mature as an alternative treatment, especially for patients with high operative risk (5).
Current guidelines recommend consideration of carotid artery stenting (CAS) as an alternative to CEA in asymptomatic patients, if the center complication rate has been demonstrated to be <3%, and especially if CEA would be technically challenging. Furthermore, it should be considered in symptomatic patients with high surgical risk or as an alternative to CEA with a center complication rate demonstrated to be <6% (6–8).
To date, the major randomized controlled trials (RCTs) have predominantly addressed the comparison of CAS and CEA (5,9–13). However, the distinct procedural technique as well as the periprocedural regimen in groups assigned to CAS regularly was left to the discretion of individual physicians.
Because RCTs are prone to selection bias and referral bias, the aim of the present investigation was to assess the associations between technical factors of CAS, such as stent design, material of the implanted stent, use of an EPD, as well as further periprocedural measures such as intraprocedural neurophysiological monitoring, and periprocedural antiplatelet medication, and the in-hospital rate of stroke or death after CAS under routine conditions in Germany.
This secondary data analysis is based on the statutory nationwide quality assurance registry databank, operated by the Institute for Applied Quality Improvement and Research in Health Care (AQUA Institute). The basic methods have been described elsewhere (14,15). In short, between 2009 and 2015, the AQUA Institute was commissioned and authorized by the German Federal Joint Committee (legal basis: §91 German Social Security Code Book [SGB] V ) to develop and implement external quality assurance in the German health care system pursuant to §137a SGB V. The AQUA Institute is also mandated for data validation, data analysis, and publication of annual quality reports (17).
In accordance with the German Federal Joint Committee directive concerning the measures of trans-sectoral and inpatient quality assurance (18,19), reporting of pre-defined and uniform documentation is compulsory for all procedures performed to treat extracranial carotid artery stenosis. These documentation reports include data on patients with statutory insurance, private insurance, patients without health care insurance, and self-payers (19). Because of legal obligations, the data collection covers nearly all (99.8% in 2014) CEA operations and CAS procedures performed in German hospitals registered under §108 SGB V.
In 2014, our working group was granted access to the quality assurance data by the German Federal Joint Committee, pursuant to §137a Abs. 10 SGB V. This study was approved by the ethics committee of the Technical University of Munich and was performed in accordance with the Good Practice of Secondary Data Analysis (20) and the Strengthening the Reporting of Observational Studies in Epidemiology statement (21). Nonanonymous patient-level data hosted by the AQUA Institute were accessed only by use of so-called controlled remote data processing. Conformity with German data protection laws was supervised by staff members of the AQUA Institute (T.B.).
All patients undergoing CAS for asymptomatic or symptomatic internal carotid artery stenosis between 2012 and 2014 were included. Exclusion criteria were emergency interventions for crescendo transient ischemic attack (TIA) or stroke in evolution, acute internal carotid artery occlusion, recurrent stenosis, tandem stenosis, carotid aneurysms, symptomatic internal carotid artery coiling, and symptomatic low-grade (<50%) stenosis with ulcerated plaque morphology. Inclusion and exclusion criteria are provided in Figure 1. Overall, 13,086 patients were available for the analysis. Among them, 8,360 were asymptomatic, and 4,726 patients were treated for symptomatic stenosis. Because complete follow-up until hospital discharge is mandatory, in-hospital follow-up was available for all patients.
With respect to their comorbidities, patients were categorized using the physical status classification system of the American Society of Anesthesiologists (22).
The primary outcome (dependent variable) was any stroke or death until discharge from hospital. The secondary outcomes were major stroke or death, any stroke (alone), and death (alone), all until hospital discharge. Applying the modified Rankin scale (23), stroke symptoms with scores of 0 to 2 were graded as minor and those with scores ≥3 as major.
To calculate the adjusted relative risk (RR) and 95% confidence interval (CI) for the independent variables (stent design, type of stent, use of an EPD, intraprocedural neurophysiological monitoring, periprocedural antiplatelet therapy), a multilevel Poisson regression model (24–27) was applied. To account for confounding and clustering of patients within centers, age, sex, American Society of Anesthesiologists stage, type of index event in symptomatic patients (amaurosis fugax, TIA, minor stroke, major stroke), ipsilateral and contralateral degree of stenosis, pre- and post-procedural assessment by a specialist in neurology, and hospital volume (empirical quintiles regarding the annual number of CAS procedures performed: 1 to 2, 3 to 6, 7 to 12, 13 to 26, and ≥27 per year) were entered as fixed-effect factors (Online Table 1 for variables used in the multivariable regression analysis). Anonymized hospital site codes valid in the treatment year were entered into the model as a random effect (random intercept only).
R version 3.2.1 (R Foundation for Statistical Computing, Vienna, Austria) was used for data processing and statistical analysis, with extension packages gmodels and lme4 to calculate cross-classified tables, chi-square tests, and multivariable regression analyses. Variable codes were extracted from the codebooks provided by the AQUA Institute and harmonized over the time period from 2012 to 2014.
Patient characteristics are listed in Table 1. The mean age of the whole cohort was 69.7 ± 9.3 years, without a relevant difference between asymptomatic and symptomatic patients (0.08 years). At 69.7%, the majority of patients were men. About 98% of patients ranked in American Society of Anesthesiologists stages I to III. More than 90% of all patients had severe carotid stenosis, referring to a degree of at least 70% according to the NASCET (North American Symptomatic Carotid Endarterectomy Trial) criteria. In symptomatic patients, TIAs and minor strokes were the most common symptoms, amounting to 28.7% and 28.6%, respectively. Major strokes or symptoms according to amaurosis fugax occurred less frequently. The most common pre-operative diagnostic method was duplex ultrasound, preceding more than 90% of all CAS procedures. Transcranial Doppler ultrasonography was performed in symptomatic patients more than twice as often as in asymptomatic patients (22.7% and 54.0%, respectively). A pre-procedural neurological assessment was conducted more frequently in symptomatic patients (68.1% in asymptomatic patients, 89.1% in symptomatic patients). Post-procedural neurological assessment was performed less frequently (58.0% in asymptomatic patients, 80.9% in symptomatic patients).
Periprocedural and intraprocedural management
The results concerning periprocedural and intraprocedural management are shown in Table 2. Intraprocedural monitoring using electroencephalography, transcranial cerebral oximetry, somatosensory evoked potentials, or other methods was performed in about 38% of all patients. Transcranial cerebral oximetry was the method applied most frequently. Irrespective of the indication group, a combination of angioplasty and stent placement was the most common procedure (90.4%). EPDs were used in 60.4% of all patients. The rate of protection system application was higher in asymptomatic (67.1%) than in symptomatic (48.4%) patients. In more than one-half of the patients, closed-cell stents were delivered. Most of the implanted stents consisted of nitinol. As far as perioperative antiplatelet therapy is concerned, more than 80% of patients were treated with dual medications.
Univariate and multivariable regression analysis
The primary outcome occurred in 317 of 13,086 patients (2.4%; 1.7% in asymptomatic patients and 3.7% in symptomatic patients). The mortality rate was 0.4% and 0.9% in asymptomatic and symptomatic patients, respectively. In-hospital strokes were recorded in 1.3% and 2.7% of asymptomatic and symptomatic patients, respectively (Table 3).
Results of the univariate analysis are listed in Table 4. The only significant association with a lower rate of primary outcome events was found for the application of an EPD (RR: 0.55; 95% CI: 0.44 to 0.69; p < 0.001). The other variables (stent design, stent type, neurophysiological monitoring, periprocedural antiplatelet medication) were not associated with lower rates of in-hospital stroke or death.
With regard to the primary and secondary outcomes, the results obtained in the multivariate regression analysis are shown in Figure 2. As already indicated by the univariate analysis, the adjusted analysis confirmed a significantly lower rate of in-hospital stroke or death in patients undergoing CAS with EPDs (RR: 0.65; 95% CI: 0.50 to 0.85; p = 0.001). The use of an EPD was evenly associated with a significantly lower rate of any in-hospital stroke (RR: 0.57; 95% CI: 0.43 to 0.77; p < 0.001), whereas no association was shown with the occurrence of death until discharge (RR: 0.78; 95% CI: 0.46 to 1.35; p = 0.381). Regarding the combined secondary outcome of in-hospital major stroke or death (Online Figure 1), the use of an EPD was associated with a lower rate (RR: 0.60; 95% CI: 0.43 to 0.84; p = 0.003).
Stent design, material of the stent, use of intraprocedural neurophysiologic monitoring, and perioperative intake of antiplatelet medication were not associated with risk for the combined primary outcome stroke or death until discharge, nor the individual components (Figure 2). The application of neurophysiologic monitoring showed a trend toward higher rates of primary outcome events (RR: 1.25; 95% CI: 0.96 to 1.62; p = 0.093).
Application of an EPD
This study shows that application of an EPD was independently associated with significantly lower rates of in-hospital stroke or death. This finding confirms some results from the published research. Because of the low rates of periprocedural clinical events, RCTs investigating the impact of EPDs on the periprocedural risk for stroke or death are lacking. RCTs using surrogate markers as endpoints were not able to detect a positive effect of EPD use (28,29). However, retrospective investigation of registry data (30,31) and review of published studies (32) indicated an advantage of EPD application. Similarly, in a more recent and comprehensive review of 23,461 procedures, the risk for a periprocedural stroke was lower among patients who received protected carotid angioplasty or stenting compared with those who were treated without EPDs (RR: 0.62; 95% CI: 0.54 to 0.72) (33). Another review including 54,713 patients sought to identify risk factors for perioperative stroke or death after CAS. The insertion of an EPD was associated with a lower rate of stroke or death within 30 days (RR: 0.57; 95% CI: 0.43 to 0.76) (34). Similar to the present investigation, a major drawback of these reviews was that no information on the specific type of EPD (i.e., proximal vs. distal) used was available. Although prospective studies showed an advantage of proximal compared with distal protection on the occurrence of surrogate events (35–37), a retrospective analysis of registry data involving 10,246 patients did not show a significant difference (1.5% vs. 2.4%; p = 0.16) in in-hospital stroke or death rates after CAS using proximal or distal protection, respectively (38). However, this study shows that the use of an EPD is associated with a lower risk for in-hospital stroke or death, not only within clinical trials or registries but also in real-world practice.
In this investigation, the implantation of closed-cell and semi-closed-cell stents showed trends toward lower rates of in-hospital stroke or death compared with open-cell stents, but with a broad CI. The effect of stent design on periprocedural rates of stroke or death was the object of a retrospective analysis conducted by Schillinger et al. (39) including 1,684 patients. Amounting to 3.1% for patients treated with closed-cell stents and 2.4% for those undergoing CAS with open-cell stents, the combined stroke or death rates did not differ significantly (p = 0.38) (39). In contrast to the present study, an investigation of the Vascular Registry of the Society for Vascular Surgery of 4,337 CAS procedures found trends toward lower rates of in-hospital stroke (odds ratio [OR]: 0.781; 95% CI: 0.462 to 1.320; p = 0.36) and death (OR: 0.515; 95% CI: 0.218 to 1.213; p = 0.13) favoring CAS with insertion of open-cell stents. Regarding the occurrence of the combined outcome of death, stroke, or TIA, the use of open-cell stents compared with those with closed-cell design was associated with a significantly lower risk (OR: 0.674; 95% CI: 0.460 to 0.987; p = 0.04) (40). The distinct reasons for the different results regarding stent designs remain unclear. To find out which patient subgroups might benefit from each stent design, further trials considering additional parameters, such as carotid morphology and plaque stability, are necessary.
The results of the present analysis show that stents made of nitinol are not related to a significantly different rate of primary and secondary outcome events compared with those consisting of stainless steel or other materials. The evidence regarding the impact of different stent materials on perioperative stroke or death rates is sparse. A retrospective study of 178 patients detected no significant difference in stroke rates (3.3% vs. 2.2%) for patients who underwent CAS with stainless steel and nitinol stents, respectively (41).
Periprocedural antiplatelet medication
The underlying data did not show a significant association between the periprocedural intake of antiplatelet medication and the occurrence of the primary or secondary outcome. The most probable reason for this surprising finding is that too few patients were not treated with antiplatelet medications perioperatively (about 2%). A recently published analysis of patients included in the ICSS (International Carotid Stenting Study) similarly did not show a significant effect of periprocedural therapy with any antiplatelet medication on the risk for stroke, myocardial infarction, or death within 30 days of CAS (RR: 0.62; 95% CI: 0.17 to 2.33; p = 0.48). However, dual antiplatelet therapy with aspirin and clopidogrel after risk adjustment was identified as an independent predictor of lower rates of stroke, myocardial infarction, or death within 30 days of CAS (adjusted RR: 0.59; 95% CI: 0.36 to 0.98; p = 0.04) (42). Unfortunately, the present investigation does not provide a separate analysis for different regimens of antiplatelet therapy.
Patients entered in the German carotid quality assurance database who underwent CAS under neurophysiological monitoring tended to show higher rates of in-hospital stroke or death, but the result was not statistically significant. The reason for this trend is likely to be found in confounding by indication, as intraoperative neuromonitoring at many centers is conducted selectively if pre-procedural diagnosis indicates an elevated perioperative stroke risk. Because the distinct reasons for the use of neurophysiological monitoring are unknown, adjustment was not possible in the multivariable regression analysis. Therefore, results concerning neurophysiological monitoring are likely to underlie a confounding by indication bias.
The present study had several limitations (14,15). First, the study design was retrospective and observational. Because patients were not randomized for the different procedural techniques and adjunct measures, selection bias as well as confounding by indication is possible. This implies that all results need to be interpreted as associations rather than causal relationships.
Second, follow-up data covered only the in-hospital period. Because most of the perioperative events presumably occur within the first days after CAS, a detection bias is considered to be low.
Third, all data in the database are self-reported by the attending physicians, and reporting bias cannot be ruled out. However, data reports were reviewed by the regional offices for quality assurance (Landesgeschäftsstellen für Qualitätssicherung) and the occurrence of suspect data induced a process of structured dialogue to clarify abnormalities systematically. Nonetheless, underreporting of adverse events is theoretically probable and might be the reason for the overall low rates of perioperative stroke or death reported in this registry (1.7% in asymptomatic patients and 3.7% in symptomatic patients). A review including 206 studies with a total of 54,713 patients stated the 30-day risk for stroke or death to be as high as 3.3% and 7.6% in asymptomatic and symptomatic patients, respectively (34). Although underreporting cannot be ruled out, any potential information bias can be considered homogeneous among the variables analyzed in this study.
Because no unequivocal definition for the stent design is provided in the database manual, classification as open-cell design or closed-cell design remains to some extent dependent on the assumption of the interventionalists.
Fourth, residual confounding cannot be excluded, because some possible confounders were not collected (e.g., information on the type of EPD or stent used, comorbidities, cardiovascular risk profile, routine medication, presence of restenosis, intraoperative heparin or protamine application, or the reasons for the application of a certain procedural technique).
Because information on the distinct type of EPD is lacking, this study does not contribute to answering the question of whether proximal or distal protection, with or without flow reversal, is associated with a better outcome.
Last, no information on cause of death is provided in the registry.
Despite the listed limitations, data collection was prospective, nationwide, and unselected. Because data reporting was compulsory, the present dataset represents real-world practice in Germany between 2012 and 2014. In conclusion, this investigation is among the largest studies to analyze the associations between different technical factors and adjunct measures and the risk for stroke or death in the periprocedural period after CAS. The use of an EPD was independently associated with a lower risk for stroke or death until discharge. To prove causality of this effect, it should be investigated in a randomized trial.
WHAT IS KNOWN? Systematic reviews indicated that the use of an EPD is associated with a lower risk for perioperative stroke or death after CAS.
WHAT IS NEW? This retrospective study confirmed an association of EPD application with a lower risk for in-hospital stroke or death in real-world practice.
WHAT IS NEXT? A prospective RCT is required to confirm the hypothesis obtained from this study. Ideally, this should also focus on a comparison of different EPD types.
For a supplemental table and figure, please see the online version of this article.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- carotid artery stenting
- carotid endarterectomy
- confidence interval
- embolic protection device
- odds ratio
- randomized controlled trial
- relative risk
- German Social Security Code Book
- transient ischemic attack
- Received November 29, 2016.
- Revision received February 23, 2017.
- Accepted March 23, 2017.
- 2017 American College of Cardiology Foundation
- Kerber C.W.,
- Cromwell L.D.,
- Loehden O.L.
- Kernan W.N.,
- Ovbiagele B.,
- Black H.R.,
- et al.
- Tendera M.,
- Aboyans V.,
- Bartelink M.L.,
- et al.
- Tsantilas P.,
- Kuehnl A.,
- Breitkreuz T.,
- et al.
- Kuehnl A.,
- Tsantilas P.,
- Knappich C.,
- et al.
- ↵Federal Ministry of Justice and Consumer Protection. Sozialgesetzbuch (SGB) Fünftes Buch (V)—Gesetzliche Krankenversicherung 2015. Available at: https://www.gesetze-im-internet.de/bundesrecht/sgb_5/gesamt.pdf. Accessed August 24, 2015.
- ↵Allgemeine Methoden im Rahmen der sektorenübergreifenden Qualitätssicherung im Gesundheitswesen nach §137a SGB V. Available at: https://www.sqg.de/sqg/upload/CONTENT/Hintergrund/Methodenpapier/AQUA_AllgemeineMethoden_Version_3-0.pdf. Accessed August 24, 2015.
- ↵Gemeinsamer Bundesausschuss. Richtlinie des Gemeinsamen Bundesausschusses gemäß §137 Abs. 1 SGB V i.V.m. §135a SGB V über Maßnahmen der Qualitätssicherung für nach §108 SGB V zugelassene Krankenhäuser. Richtlinie über Maßnahmen der Qualitätssicherung in Krankenhäusern—QSKH-RL. Gemeinsamer Bundesausschuss 2014. Available at: https://www.g-ba.de/downloads/62-492-935/QSKH-RL_2014-12-04-und-2014-06-19.pdf?. Accessed August 24, 2015.
- ↵Gemeinsamer Bundesausschuss. Richtlinie des Gemeinsamen Bundesausschusses nach §92 Abs. 1 Satz 2 Nr. 13 i.V.m. §137 Abs. 1 Nr. 1 SGB V über die einrichtungs- und sektorenübergreifenden Maßnahmen der Qualitätssicherung (Qesü-RL) 2014. Available at: https://www.kvwl.de/arzt/recht/kbv/rl/gba_qs_sektor.pdf. Accessed August 24, 2015.
- van Swieten J.C.,
- Koudstaal P.J.,
- Visser M.C.,
- Schouten H.J.,
- van Gijn J.
- Knol M.J.,
- Le Cessie S.,
- Algra A.,
- Vandenbroucke J.P.,
- Groenwold R.H.
- Zou G.
- Zahn R.,
- Mark B.,
- Niedermaier N.,
- et al.
- Kastrup A.,
- Groschel K.,
- Krapf H.,
- Brehm B.R.,
- Dichgans J.,
- Schulz J.B.
- Touze E.,
- Trinquart L.,
- Chatellier G.,
- Mas J.L.
- Montorsi P.,
- Caputi L.,
- Galli S.,
- et al.
- Bijuklic K.,
- Wandler A.,
- Hazizi F.,
- Schofer J.
- Cano M.N.,
- Kambara A.M.,
- de Cano S.J.,
- et al.
- Giri J.,
- Parikh S.A.,
- Kennedy K.F.,
- et al.
- Schillinger M.,
- Gschwendtner M.,
- Reimers B.,
- et al.