Background Adjunctive stenting has increasingly become an acceptable option for the endovascular treatment of unruptured aneurysms. The Nationwide Inpatient Sample (NIS) was used to compare US in-hospital outcomes related to coiling with and without adjunctive stenting for unruptured aneurysms.
Methods Hospitalizations for coiling of unruptured cerebral aneurysms from 2004 to 2008 were identified in the NIS by extracting ICD-9-CM codes for the diagnosis of unruptured aneurysm (437.3) and intracranial stenting (00.65) with coiling (39.52, 39.79 or 39.72) of cerebral aneurysms. All patients with a diagnosis of subarachnoid hemorrhage (430) and/or intracerebral hemorrhage (431) were excluded. Mortality and discharge to a long-term facility were compared between stent and non-stent patient groups using multivariate regression analysis.
Results Patients treated with stent-assisted coiling had an in-hospital mortality rate of 0.08–0.8% compared with a death rate of 0.5% (95% CI 0.3% to 0.7%) for patients who did not receive a stent during coiling (p=0.36). Patients in the stent group had a 3% rate of discharge to a care facility (95% CI 1.5% to 5.8%) compared with 5% (95% CI 4.5% to 5.6%) for those in the non-stent group (p=0.14). Patients treated with a stent had a similar likelihood of in-hospital mortality (adjusted OR, 2.12 (95% CI 0.32 to 7.11), p=0.34) and a lower likelihood of discharge to a long-term care facility (adjusted OR 0.59 (95% CI 0.24 to 1.16), p=0.16) compared with the non-stent group.
Conclusions Adjunctive stenting adds little in-hospital risk to the endovascular treatment of cerebral aneurysms. However, the need for dual antiplatelet therapy may predispose to delayed hemorrhagic complications and discontinuation of dual antiplatelet therapy may lead to delayed thromboembolic complications.
- cerebral aneurysm
- endovascular procedures
- outcome assessment
Statistics from Altmetric.com
Stenting allows for the coil embolization of wide-necked aneurysms that might otherwise not be amenable to endovascular therapy. In these cases, the adjunctive stent serves as a mechanical scaffold for the support of aneurysm coils, prevents coil protrusion into the parent vessel and may also allow for denser coil packing of the aneurysm. These benefits notwithstanding, the use of an adjunctive stent, however, may be associated with a higher risk of complications. Piotin et al recently reported a large series of 216 patients in which the periprocedural mortality rate in patients treated with adjunctive stenting was 6.0%, considerably higher than the 1.2% rate in their practice for procedures without stents (p=0.002).1 In order to further evaluate the periprocedural risk of adjunctive stenting, we used the US Nationwide Inpatient Sample (NIS) database from 2004 to 2008 to determine and compare outcomes of coiling of intracranial aneurysms with and without a stent.
International Classification of Diseases 9th Revision Clinical Modification (ICD-9-CM) diagnostic codes were used to identify all cases of unruptured aneurysm (ICD-9 437.3) from the NIS hospital discharge database (Healthcare Cost and Utilization Project of the Agency for Healthcare Research and Quality, Rockville, Maryland, USA).2 The NIS is the largest all-payer inpatient care database in the USA that analyzes national trends in healthcare utilization outcomes based on data extracted from a large stratified sample of US community hospitals, accounting for approximately 20% of annual hospitalizations in the USA. The NIS has been previously used to study hospitalizations associated with both unruptured and ruptured cerebral aneurysms.3 ,4 As stent-assisted coiling procedures were uncommon before 2004, only NIS data from 2004 to 2008 were analyzed. National estimates of disease incidence were calculated using the discharge-level weighting factors unique to each hospital and year. As part of the Healthcare Cost and Utilization Project (HCUP) data use agreement, any tabulated data containing 0–10 national estimate results were concealed to protect patient privacy.
All unruptured aneurysm discharge records (ICD-9 code 437.3) associated with endovascular repair or occlusion of head and neck vessels (ICD-9 procedure code 39.72), other endovascular repair of other vessels (39.79) or other repair of aneurysm (39.52) were included for further study. Unruptured aneurysms associated with subarachnoid hemorrhage (430) or intracerebral hemorrhage (431) were excluded from further analysis. Patients meeting these inclusion and exclusion criteria were then subdivided into stent or non-stent groups depending on whether they underwent percutaneous insertion of intracranial vascular stent(s) (00.65). Patient data (age, gender, race, admission type, number of diagnoses upon discharge) and hospital data (bed size, teaching status, and number of total discharges) were extracted for all selected patient records. In-hospital mortality, discharge to a long-term care facility, length of stay and total hospital charges were the primary outcomes of the study.
All statistical analyses were performed using JMP V.9 software (SAS Institute). Continuous data for stent and non-stent groups (age, length of stay and total charges) were presented as medians and interquartile ranges and compared using the Wilcoxon non-parametric signed-rank test. Categorical data (gender, race, in-hospital mortality and discharge status) were presented as relative frequencies and compared using the Pearson χ2 test. Statistical significance was defined as p<0.05. ORs of in-hospital mortality and discharge to care facility were calculated using multivariate logistic regression analysis after adjusting for patient and hospital characteristics found to be significantly different between stent and non-stent patient groups.
The national estimate of patient, procedural and hospital characteristics of the stent and non-stent groups are shown in table 1. Between 2004 and 2008 an estimated 27 921 endovascular repair procedures were performed on the US population among patients with unruptured aneurysms without a diagnosis of subarachnoid or intracerebral hemorrhage. Of these procedures, an estimated 1316 (4.7%) involved the placement of a stent. Significant differences between the stent and non-stent groups were observed for gender, race and admission source (table 1). The estimated number of women was 1039 (79%) in the stent group and 19 992 (75%) in the non-stent group (p=0.0017). The median age was 58 years (IQR 48–67) in the stent group and 57 years (IQR 48–66) in the non-stent group (p=0.37). Hospital characteristics of total discharges, bed size and ownership and teaching status were significantly different between the stent and non-stent groups.
National estimates of the frequencies of major adverse outcomes are shown in table 2. No significant differences in mortality were observed (p=0.06) between patients in the stent (≤10 (0.08–0.8%)) and non-stent groups (25 (0.5%)). The estimated number of patients who were discharged to a care facility was significantly higher in the non-stent group than in the stent group (5% vs 3%, p=0.0009). Length of stay was not significantly different between the stent and non-stent groups (p=0.19).
Multivariate logistic regression analysis was performed to adjust for patient and hospital characteristics found to be significantly different between the stent and non-stent patient groups. Unadjusted and adjusted ORs of in-hospital mortality and discharge to care facility are shown in table 3. After adjusting for differences in gender, age, admission source, hospital discharge volume, bed size and ownership and teaching status, patients in the stent group showed a similar likelihood of in-hospital mortality (OR 2.12, 95% CI 0.32 to 7.11, p=0.34) and discharge to a long-term care facility (OR 0.59, 95% CI 0.24 to 1.16, p=0.16) as patients in the non-stent group.
We found no significant increase in adverse outcomes in patients who received endovascular aneurysm treatment with stent-assisted coiling compared with those treated without adjunctive stenting. Patients who underwent stenting did not show higher rates of in-hospital mortality, discharge to care facility or length of hospitalization than patients who did not receive a stent. The likelihood of in-hospital mortality or discharge to a care facility was similar between the stent and non-stent groups.
These current data are in contrast to early reports of operators' experience with adjunctive stenting that were associated with high rates of adverse events (6–9% rate of thromboembolic complications and 2–4.5% death rate).5–8 There may be several reasons for these disparate findings. First, the data from these early experiences with stenting originated from small case series. Second, early generation stents were difficult to deploy,6 ,8 which might have contributed to higher morbidity and mortality. Third, an operator learning curve was probably present in early stenting cases resulting in a greater number of complications compared with more recent practice.
Recent studies of large samples of stent-treated aneurysms have reported widely variable permanent morbidity rates (2.8–5.6%) and mortality rates (0.9–2%).9–11 In a study by Mocco et al, the reported mortality rate of 2% appears to be spuriously elevated as a result of a disproportionate number of deaths in the subarachnoid hemorrhage cohort (12%) compared with the unruptured cohort (0.8%).12 Piotin et al reported significantly higher rates of complications (16/216, 7.4%) and deaths (13/216, 6.0%) in the stented group compared with the coiled control group (42/1109 (3.8%) complications, 13/1109 (1.2%) deaths).1 In this latter study the use of antiplatelet drugs and guidewire exchange maneuvers may explain the increased incidence of hemorrhagic complications and vessel perforations. Furthermore, thromboembolic complications also occurred more frequently in the stented patients. The study by Piotin et al also included both balloon-expandable and self-expandable stents, the former being responsible for much of the morbidity and mortality. For self-expanding stents the risk of procedural morbidity was 6.0% and mortality was 4.0%.
Although our study found no significant differences in the rates of periprocedural complications of coil treatment with adjuvant stenting compared with coil treatment alone, disparities in delayed complications are probably not captured in our analysis. A majority of delayed complications following adjuvant stenting are a result of the thrombogenic potential of the stent device and include delayed thrombosis, delayed hemorrhage and delayed stenosis.5 ,7 ,8 In a series of 156 patients with follow-up, Fiorella et al found nine cases (5.8%) of moderate or severe delayed in-stent stenosis with focal neurological symptoms occurring up to 3 months after the procedure, including two parent vessel occlusions.13 Studies by Yahia et al and Mocco et al demonstrated similar rates of delayed thromboembolic events associated with stent-assisted coil treatment (3/65 (4.6%) and 7/213 (3%), respectively).14 ,15 In the study by Mocco, all seven cases (3%) of delayed thrombotic events were related to cessation of dual antiplatelet therapy. Delayed complications may therefore contribute to the disparity between our study and those with higher reported rates of stent-related complications.
Antiplatelet medication might be expected to cause occasional delayed hemorrhagic complications, but this has not yet been reported. Two large meta-analyses of multiple trials of aspirin and dual platelet therapy in patients with cardiac disease have shown that aspirin alone increases the risk for both intracranial and extracranial hemorrhage.16 ,17 Serious bleeding rates associated with aspirin use range between 1.5% and 2.5%,17 and dual antiplatelet therapy further increases the risk of all types of bleeding up to 40–50%.16
The Neuroform device (Stryker, Kalamzoo, Michigan, USA) was approved by the United States Food and Drug Administration in 2002 and was the first stent marketed for aneurysm treatment in the USA. The Enterprise device (Codman, Miami Lakes, Florida, USA) was approved by the FDA in 2007. The Enterprise was an improvement on deployment,10 as were newer versions of the Neuroform device.18 Our study spans hospitalizations that occurred between 2004 and 2008 and would therefore include patients treated with both of these devices, but we are unable to obtain specific data regarding which stents were used from the NIS database. Balloon-assisted coil treatment is an alternative to stenting in many cases. Because balloon-assisted coil treatment does not incur significant additional periprocedural risk19 and has not been reported to have any of the delayed complications associated with stenting, it may be a safer alternative.
We did not evaluate patients with subarachnoid hemorrhage who underwent coil treatment of an aneurysm with stenting because there are very few such patients in the NIS. The use of adjunctive stenting for ruptured aneurysms has been limited by the need for antiplatelet therapy. A recent systematic review of patients with ruptured aneurysms showed that stent-assisted coiling can be performed with high degrees of technical success, but adverse events appear more common and clinical outcomes are probably worse than those achieved without stent assistance.20 In this review, clinically significant intracranial hemorrhagic complications occurred in 8% of patients and clinically significant thromboembolic events occurred in 6% of patients.
This study has a number of limitations. First, we acknowledge that some coding inaccuracies undoubtedly occur which can affect the retrospective evaluation of an administrative database.21 Second, many patient and procedural details are not available in the NIS, including potential variation in technique (types of stents used, technique for stent placement, anticoagulation regimens) and possibly differences in patient selection for stent-assisted coiling versus traditional coiling or primary surgical intervention. Furthermore, the NIS does not contain detailed patient information such as inpatient morbidity, specific neurological deficits or medical complications that may have contributed to a poor outcome. While we performed multivariate regression to minimize patient- and hospital-specific confounders in the stent and non-stent groups, our study is retrospective and patients were not treated in a randomized manner. There is therefore significant potential for selection bias that might affect outcomes of patients who were stented or not. Finally, we were unable to analyze either short- or long-term outcomes that occurred after hospital discharge or the incidence of minor complications that did not result in death or discharge to a care facility. Additional prospective studies are warranted to analyze patient outcomes further following stenting procedures.
The adjunctive use of a stent adds little in-hospital risk to the endovascular treatment of cerebral aneurysms. However, we acknowledge that the need for dual antiplatelet therapy may predispose to delayed hemorrhagic complications and discontinuation of dual antiplatelet therapy may lead to delayed thromboembolic complications.
Competing interests JSM receives an investigator-initiated research grant from General Electric Healthcare. HJC is the site PI of a registry study for Cordis Endovascular. GL receives an educational grant from EV3 and gives expert testimony in relation to EV3/Coviden. DFK is a RSNA deputy editor. These disclosures are unrelated to the current study. RJM and APN have no competing interests.
Ethics approval This study used the Nationwide Inpatient Sample, a de-identified patient database. Therefore, this study did not require IRB review in accordance with the Code of Federal Regulations, 45 CFR 46.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Data were generated from the Nationwide Inpatient Sample, a publically-available database.
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.