Woven EndoBridge aneurysm treatment is associated with high feasibility,great safety,and very good efficacy also in long-term follow-up

Dear Editor,

We read with great interest the paper “The Woven EndoBridge device, an Effective and Safe Alternative Endovascular Treatment of Intracranial Aneurysm—Systematic Review” by Mantilla et al. ¹ Intrasaccular devices were introduced in clinical practice in Europe in 2010 and have been increasingly used in the management of intracranial aneurysms, singularly wide-neck bifurcation aneurysms. While several intrasaccular devices are currently available for clinical practice (WEB, Microvention, Aliso Viejo, CA, USA; Artisse, Medtronic, Irvine, CA, USA; Contour, Stryker, Fremont, CA, USA), only the WEB device was properly evaluated by several Good Clinical Practice (GCP) studies conducted in Europe, the USA, and Asia.^2,3 Moreover, a large number of prospective and retrospective, single, and multicenter series have also been published conducting several meta-analyses. Since 2016, at least 10 meta-analyses dedicated to WEB (not including Mantilla's) have been published and are referenced in Pubmed.

Mantilla et al. are providing another review of the literature dedicated to WEB. However, their work presents numerous errors. To start, the number of patients/aneurysms is different in the Abstract (1705/1224) and in the Results section (1599/1705), the latter being the correct numbers if we consider the population listed in Table 1 of Mantilla's paper. In addition, their analysis presents calculation errors that have tremendous consequences for the understanding of the readers. For example, the rate of successful WEB placement is reported at 28.1% (abstract and text), but the recalculation (1649/1705) conducts to a more realistic number (97.7%) in line with the results reported in the GCP series: 96.4% in European studies (WEBCAST, French Observatory, WEBCAST-2) and 98.7% in WEB-IT.^2,3

Another basic methodological error is encountered several times: when, in a paper, numbers were not available for a given analysis, the total number of patients (denominator) used for data calculation was not adjusted accordingly. For instance, Mantilla et al. report a rate of adequate occlusion at follow-up of 49.7% (848/1705), which is in fact (taking into account the missing data) 61.2% (848/1386). This percentage is still below the results reported in the European studies (79.1%) and in WEB-IT (84.6%), probably due to several factors: the selection of patients and aneurysms treated with WEB is probably more refined in clinical studies than in clinical practice; the centers participating in clinical studies are also usually selected among those having a large experience of endovascular aneurysm treatment.

Another important factor (reported by the authors in the Limitations section) is that they have defined adequate occlusion based on the BOSS (Bicêtre Occlusion Scale Score) as BOSS 0, 0ʹ, and 1, excluding 2. Adequate occlusion was initially defined using the Montreal scale as a complete aneurysm occlusion or neck remnant. By using a different definition the authors make it impossible to compare the results of their systematic review to other papers of the literature.

Sometimes, the calculation error favors the WEB treatment. For instance, the percentage of thromboembolic (TE) complications is reported to be 6.5% (111/1704) but is, in fact, 6.9% (111/1599). By the way, it is still lower compared to what is reported in European studies (14.4%) and WEB-IT (9.4% including TIAs). This higher rate of TE events is probably related to a more extensive report in relation to the design and methodology of the studies. In European studies and in WEB-IT, all TE events were reported including TIAs (2.7% in WEB-IT) and events not associated with a clinical worsening. In WEB-IT 2 out of 14 TE events were associated with clinical worsening, whereas in European studies five out of 24 TE events were associated with sequelae. Unfortunately, in Manilla et al. paper the percentage of the different types of TE events (permanent/transient, with or without sequelae) is not reported making a more relevant analysis impossible. Moreover, we have also to consider that the meta-analysis is mixing GCP series, in which adverse events were independently evaluated by a Clinical Event Committee (CEC), and the prospective or retrospective series, in which adverse events were self-reported and analyzed by the centers themselves.

Similarly, the rate of retreatment is not 3.9% (67/1705) but 7.6% (67/879) and this number is similar to what was reported in European studies (6.9% at 1 year) and WEB-IT (5.6%). A similar limitation is also encountered for anatomical results that are self-evaluated in most retrospective and prospective series but are analyzed by an independent core lab in GCP studies.

Finally, some numbers are also erroneously reported from the published series listed in Table 1 of Mantilla's paper. For instance, the number of adequate occlusions at follow-up in the three European GCP studies is reported to be 43 but is, in fact, 121. ⁴ As a consequence, the numbers in Table 2 of Mantilla's paper are also probably erroneous. However, it is not possible to analyze this point as the names of the series are surprisingly missing.

All these errors will indubitably mislead the readers in their perception of WEB aneurysm treatment results. Will a physician consider as acceptable a treatment with a success rate of 28.1%? Is a rate of 49.7% of adequate occlusion acceptable at follow-up?

Finally, the authors outline the need for further research “to optimize device selection, standardize classification systems, and enhance long-term evaluation.” This final conclusion outlines a great lack of knowledge from the authors who have apparently completely missed the recent publications of five-year follow-up in both European studies and WEB-IT.^4,5 These recent publications have demonstrated the unparalleled safety and great stability of aneurysm occlusion after WEB aneurysm treatment. We share with the authors the necessity to institutionalize high-quality evaluation of new devices introduced in the INR field with single-arm GCP studies accurately evaluating safety and efficacy, RCTs comparing new devices to devices currently used in clinical practice, and properly conducted meta-analyses. We hope that other intrasaccular devices (Contour and Artisse) will benefit from similar stringent evaluations than the ones conducted for WEB.