Endovascular treatment with the Enterprise stent versus the Neuroform or Low-Profile Visualized Intraluminal Support stent for unruptured aneurysms

Using a nationally representative multihospital database, this study compared key clinical and economic outcomes of endovascular treatment with the Enterprise stent versus its comparators, the Neuroform and LVIS stents. While there were no major differences in patient outcomes or complication rates between the stent groups, patients treated with Enterprise had 28–38% lower odds of all-cause, CV-related and UIA-related readmission compared with the comparator group. In the bivariate analysis, this difference was associated with a corresponding lower cost; however, the cost benefit was not statistically significant after adjustment for hospital clustering in the GEE model.

Few studies have compared outcomes among patients with UIA undergoing stent-assisted coiling using the Enterprise stent versus the LVIS or Neuroform stent, and these studies have reported mixed results. When comparing outcomes among 190 patients treated for UIA with the Enterprise stent or LVIS stent, Ge et al. observed no significant between-group differences in clinical outcomes (modified Rankin Scale [mRS] score: 0–2; 90.9 vs 94.9%, p = 0.379) or procedure-related complications (16.3 vs 10.9%, p = 0.298) [10]. Although the authors reported significantly higher rates of complete or near-complete occlusion in the LVIS group versus the Enterprise stent group (96.9 vs 88.4%, p = 0.048), this difference did not adjust for variation in patient clinical characteristics including aneurysm size, morphology and location. In another study by Kadkhodayan et al. comparing aneurysms treated with the Enterprise stent (98 patients) and Neuroform stent (160 patients) using prospective registry data, the Enterprise stent was associated with a statistically superior deployment rate (difficult deployment rate, 1.7 vs 15.9%, p < 0.0001) and occlusion rate (87.3 vs 73.0%, p = 0.0058) [8]. The authors reported no significant difference in hemorrhagic complications between the two groups; however, thromboembolic complications were higher for Enterprise stent patients compared with Neuroform patients (8.7 vs 1.4%, p = 0.0021). In a single-center study of 130 aneurysm patients treated with Enterprise stent and Neuroform stent, Durst et al. found that use of the Neuroform stent was a predictor for retreatment (p = 0.034), but did not observe any difference in occlusion rates between the two groups [11]. In another single-center study comparing the Enterprise, Neuroform, and LVIS (including LVIS Jr) stents, Wang et al. reported all three stents to be safe and effective for the treatment of cerebral aneurysms, with some variation seen in terms of initial complete occlusion, progressive occlusion and complication rates among the three groups [12].

Though the above studies provide useful information, the generalizability of their findings is limited by the single-center nature of these studies together with a lack of adjustment for patient characteristics in outcome assessments. A more recent study by Mokin et al. (2019) compared outcomes among 659 aneurysm patients undergoing endovascular coiling using Enterprise, Neuroform, and LVIS stents across multiple centers [15]. In unadjusted bivariate analyses, the results from the study indicated LVIS (including LVIS Jr.) group to have the highest rate of complete occlusion (RRGS 1) at baseline (64.4 vs 47.6 vs 56.2%; p = 0.008) and follow-up (83.9 vs 67.5 vs 78%; p = 0.004) as compared with Enterprise and Neuroform group. However, in multivariate adjustment, no significant difference in complete occlusion at baseline (p = 0.098) and follow-up (p = 0.113) was observed by type of stent in the study [15], reflecting that the significant differences observed in their bivariate analyses dissipates when adjusted for study covariates. Though these reports provide useful information, the results are difficult to generalize to a real-world setting given important limitations in study design including limited sample and largely single-center or few center experience, along with the fact that these studies did not adjust statistical comparisons for important patient and provider characteristics known to influence outcome.

In our study, we did not observe any significant difference in discharge status among patients who underwent endovascular treatment using the Enterprise stent versus comparator group stents. From bivariate comparison, patients in the Enterprise stent group did appear to have shorter LOS compared with the comparator group, though this difference was not significant when adjusting for hospital clustering.

Readmissions is a key marker of long-term treatment effectiveness when considering secondary data sources such as the PHD, which do not include tailored study end point variables; however, readmissions for cardiovascular and UIA reasons are a clinically important measure of success for patients and healthcare providers. Specifically, UIA-related readmissions could represent a differential in retreatment procedures being performed across the study devices. Patients in the Enterprise stent group had significantly lower rates of readmission compared with the Neuroform or LVIS comparator group. The approximately 38% lower likelihood of UIA-related readmission in the subsequent 12-month period postendovascular procedure for Enterprise stent patients versus the comparator group may reflect a clinical benefit accrued with the use of Enterprise stent. Our results are consistent with the finding of Durst et al. identifying use of Neuroform stent as a predictor of retreatment (compared with Enterprise stent) [9]. The authors alluded to differences in stent design, specifically to Neuroform’s open-cell design and an associated increased potential for coil herniation, as a likely cause of higher retreatment rates in the Neuroform group. It should be acknowledged that our comparator group included patients who underwent procedures using either the Neuroform stent or LVIS stent, unlike Durst et al. [9], who directly compared the Neuroform stent to the Enterprise stent. Further research is needed to understand individual differences when comparing use of the Enterprise stent versus Neuroform stent, and the Enterprise stent versus LVIS stent, in a real-world setting.

As healthcare dollars are increasingly scrutinized, it is important to understand cost differences associated with device selection from both a payer and provider perspective. Our results showed that endovascular procedure costs were US$1756 lower in the Enterprise stent group versus the comparator stent group. When examining the total cost of care, calculated as the sum of index endovascular procedure costs and 12-month all-cause inpatient admission costs, Enterprise stent use was associated with a cost savings of US$22,129 compared with the comparator stents. When limiting the total cost of care to include only the index procedure and 12-month UIA-related readmission costs, this cost difference was US$3789 in favor of the Enterprise stent group. Notably, these differences approached significance in the GEE model. These results indicate a potential economic benefit associated with use of the Enterprise stent versus the Neuroform or LVIS stents.

It is important to note that the population of patients who underwent endovascular treatment with the Enterprise stent in the PHD generally had higher rates of specific co-morbidities such as stroke/TIA, peripheral vascular disease and depression as well as higher CCI scores compared with the comparator group before matching. Further research is needed to better understand the underlying reason (if any) for this difference in patient clinical profiles between the two groups, especially the observation of higher co-morbidity burden among patients treated with the Enterprise stent.

This study had a few limitations. Though we used a robust analytical technique that implemented propensity matching, unmeasured confounders could have influenced the study results. In the PHD, readmission is only captured if the patient returns to the same hospital and not if the patient is readmitted to another hospital (irrespective of whether that hospital contributes data to the PHD or not). As a result, the rates of readmission observed in our study could underestimate true rates of readmission among patients undergoing endovascular procedures. Given the lack of differential coding for wide-necked aneurysms, we were unable to ascertain the proportion of such patients in our study. Also important, due to limitations in the data and patient volume we included multiple generations of both the Neuroform and LVIS stents in our comparator group; further research is needed to understand the differences between outcomes associated with Enterprise and each of these respective competitor stents. Moreover, as medical device features evolve, it would be important to study differences among the latest version of these devices. For example, the Enterprise 2 stent was designed to have an improved vessel wall conformability, with one study reporting the gap between vessel wall and stent for Enterprise 2 stent to be half of the gap observed for the first generation of Enterprise stent [16]. Similarly, multiple generations of the Neuroform device have been launched over the years ever since the device was first approved in 2002 [17]. The most recent version of this device, in other words, the Neuroform Atlas has been said to have lower profile delivery, better scaffolding, improved trackability and higher conformability as compared with prior generations of the device [17]. As with LVIS, there are multiple generations including LVIS and LVIS Jr, with the latter designed to allow deployment to smaller and more tortuous vessels [18]. As more data become available for the three study devices, researchers could perform a comparative analysis focused on these recent versions. Lastly, coding errors during claims processing could have influenced the study results.

To the best of our knowledge, this is the first study to compare outcomes that assess the effectiveness of UIA treatment among patients who underwent endovascular coiling procedures with the Enterprise stent versus the Neuroform or LVIS stent in a real-world setting. Our results suggest that use of the Enterprise stent is associated with a significant decrease in hospital readmission for cardiovascular and UIA reasons and a similar safety profile compared with the Neuroform and LVIS stents. The results for LOS and cost outcomes were mixed, with results from bivariate analyses suggesting significantly lower LOS and cost (index admission and total cost) among Enterprise stent patients as compared with competitor stent, while the results from GEE for LOS and cost indicated no significant difference among the study cohorts. Further research is needed to examine if lower rates of inpatient admissions among Enterprise stent cohort translate into significant economic savings as compared with Neuroform/LVIS stents.

All the authors contributed to the conception and design of the study. E Kottenmeier, SHY Lee and R Khanna were responsible for the collection and assembly of data, and data analysis and interpretation. All the authors were responsible for the manuscript writing and final approval of the manuscript.

This study was sponsored by Cerenovus. R De Leacy is on the scientific advisory board of Cerenovus, stroke advisory board/consultant of Siemens, consultant for Imperative Care, Penumbra and Medical Metrics and has equity in Q’Apel Medical, Synchron and Endostream. E Kottenmeier, SHY Lee and R Khanna are employees of Johnson and Johnson. AM Spiotta has consulting agreements with Penumbra, Cerenovus, Terumo and Stryker. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

The authors acknowledge Superior Medical Experts for manuscript editing services.

In the United States, retrospective analyses of the Premier Healthcare Database data are considered exempt from institutional review board (IRB) oversight as dictated by Title 45 Code of Federal Regulations, Part 46 of the USA, specifically 45 CFR 46.101(b)(4) (http://www.hhs.gov/ohrp/humansubjects/guidance/45cfr46.html).

Data may be obtained from a third party and are not publicly available.