Long-Term Outcomes after Rotational Atherectomy for Calcified Chronic Total Occlusion versus Nonchronic Total Occlusion Coronary Lesions

Background The role of rotational atherectomy (RA) in contemporary percutaneous coronary intervention (PCI) is expanding to include certain chronic total occlusion (CTO) lesions. However, the long-term outcome of RA in CTOs is still unclear. Objective To investigate in-hospital and long-term outcomes after RA for CTO compared to non-CTO calcified lesions. Moreover, this report evaluates the role of the elective RA approach in calcified CTOs. Methods and Results This study enrolled 812 patients (869 lesions; CTO, n = 80 versus non-CTO, n = 789). The mean age of the study population was 73.1 ± 8.6 years, the baseline characteristics were comparable in both groups. Balloon-resistant CTO lesions represented the main indication for RA in CTO patients (61.2%). The mean J-CTO score was 2.42 ± 0.95. The angiographic success rate was lower in CTO patients (88.8% vs 94.9%; p = 0.022). In-hospital major adverse cardiac events (MACE) rate was comparable in both groups (CTO 8.8% vs 7.0% in non-CTO;p = 0.557). At two-year follow-up, a higher target lesion failure (TLF) was observed in CTO group (25.5% vs 15.1%, log rank p = 0.041), driven by higher cardiac mortality while the clinically driven target lesion revascularisation (TLR) was comparable between the study groups. Elective RA for CTO had a shorter procedural time and lower rate of dissection (7.5% vs 25%; p = 0.030) compared to bail-out RA with similar long-term outcomes. Conclusion Compared to non-CTO, RA for CTO can be performed with a high procedural success rate and comparable in-hospital outcomes. Apart from higher cardiac mortality in the CTO group, the long-term outcomes are comparable in both groups. Elective RA is a feasible and beneficial approach to be used in CTO intervention.


Introduction
Chronic total occlusion (CTO) is encountered in approximately 15-20% of all percutaneous coronary interventions (PCI) [1]. Calcifed CTO lesions are not uncommon, since moderate or severe calcifcation is present in more than half of CTO lesions [2].
Indeed, wire crossing through the CTO lesion is not the only challenging step during CTO PCI as balloon-resistant calcifed CTO is another difculty and it includes balloon uncrossable and undilatable lesions. Balloon uncrossable lesions are those lesions that cannot be crossed with the smallest available balloons after successful guidewire crossing, and it ranges from 6% to 9% of all CTOs, while balloon undilatable lesions are those that cannot be expanded despite using highpressure balloons and it represents about 12% [3].
Rotational atherectomy (RA) became an indispensable tool for treating calcifed coronary lesions and it is increasingly applied for the preparation of certain calcifed CTOs. Elective application of RA in calcifed CTOs is not a common approach and using RA in CTO is usually limited to balloon-resistant calcifed lesions. However, RA in calcifed CTOs could be benefcial when used electively as being preprocedurally or early intraprocedurally planned [4].
Altogether, heavily calcifed balloon-resistant CTO lesions are among the most challenging coronary lesions as they often require sophisticated techniques and unfortunately are associated with lower success rates and more complications [2]. Te feasibility and in-hospital outcomes of RA in CTOs were reported in some studies [5,6]. However, data regarding the long-term outcomes of a calcifed CTO lesion treated with RA, particularly when compared with non-CTO, are lacking. Furthermore, the role of elective RA in CTO lesions is unclear.

Study Design and Patients.
Tis is a retrospective analysis of patients who underwent RA at a single centre (Heart Centre, Segeberger Kliniken, Bad Segeberg, Germany) between 2004 and 2018. We analysed 869 RA procedures for 812 patients; 789 non-CTO lesions (n � 732 patients); and 80 CTO lesions (n � 80 patients). Baseline, procedural, inhospital, and long-term follow-up data were compared. Written informed consent was obtained from all patients for analysis of their anonymized data, and data collection were approved by the local ethics committee.

Rotational
Atherectomy. RA was applied in heavily calcifed lesions according to the operator's discretion as either an elective or a bail-out strategy (including balloonresistant CTO lesions). In CTO lesions, wire exchange (to Rota-Wire) was performed by using a microcatheter, which either crossed the lesion or was advanced as far as possible into the lesion (in cases when a microcatheter could not completely cross the lesion). A procedure was regarded as a RA as soon as the Rota-Wire was advanced through the target lesion. RA was performed using the Rotablator or Rotapro systems (Boston Scientifc Scimed, Inc., Maple Grove, MN, USA). Te burr size was chosen up to 0.7 of the vessel diameter for non-CTO RA but a smaller burr size was used for CTO lesions [7]. Prior to RA, patients were treated with 325-500 mg aspirin orally and an oral loading dose of a P2Y12 inhibitor (clopidogrel in patients presenting with chronic coronary syndrome/prasugrel, or ticagrelor in patients presenting with acute coronary syndrome) and periprocedural anticoagulation with either UFH or bivalirudin was routinely administered. Post-PCI dual antiplatelet therapy and duration were applied according to the presentation (chronic coronary syndrome or acute coronary syndrome) as recommended in the guidelines.

Defnitions.
A CTO was defned as a coronary occlusion with thrombolysis in myocardial infarction (TIMI) 0 fow for ≥3 months. Te duration of CTO was either certain (angiographically confrmed) or likely (clinically confrmed) [8]. For CTO lesions, the J-CTO score was calculated as one point was assigned to each of the following variables: previously failed attempt, blunt stump, calcifcation, bending >45°, and occlusion length ≥20 mm [9]. Calcifcation was assessed angiographically according to the Mintz classifcation [10].
Angiographic success was defned as RA followed by stent implantation with residual stenosis <30% and TIMI 3 fow in the target vessel at the end of the procedure. Te procedural success was defned as angiographic success with the absence of in-hospital periprocedural complications (allcause death, periprocedural MI, recurrent angina requiring target vessel revascularisation (TVR) with PCI or emergency coronary artery bypass graft (CABG), and tamponade requiring pericardiocentesis or surgery) [11]. Bail-out RA was defned as a procedure, where the decision to perform RA was made during the PCI after the failure of balloon-based angioplasty. Elective RA was either preprocedurally or early intraprocedurally planned, based on angiographic or intravascular imaging but before any attempt of balloon angioplasty. Early generation drug-eluting stents (DESs) represent an early experience of DESs with thick struts platform and polymer that elutes antiproliferative medication, including the sirolimus-eluting CypherTM stent (Cordis, Miami Lakes, FL, USA) and the paclitaxel-eluting Taxus Liberte TM stent (Boston Scientifc, Boston, MA, USA). While new-generation drug-eluting stents are characterized by thinner stent strut platforms, polymers with improved biocompatibility and limus-based antiproliferative agents, including the cobalt-chromium everolimus-eluting XienceTM stent (Abbott Vascular, Santa Clara, CA, USA), the platinum-chromium everolimus-eluting PromusTM stent (Boston Scientifc, Natick, MA, USA), as well as the sirolimus-eluting Orsir-oTM stent (Biotronik, Bülach, Switzerland). Te periprocedural myocardial infarction (MI) (type 4a MI) was defned according to the fourth universal defnition of MI [12]. Coronary perforation was defned and classifed according to Ellis classifcation [13]. Slow fow is defned as TIMI fow grade ≤2 at the end of the procedure [14].

Follow-Up.
Clinical follow-up was obtained either by on-site clinical visits or scripted telephone calls with the patients or their general practitioners (our centre applies a routine angiographic and clinical follow-up after complex interventions). In-hospital MACE was defned as the composite of all-cause mortality, periprocedural MI and TVR, and two-year target lesion failure (TLF) was defned as the composite of cardiac death, target vessel MI, and clinically driven target lesion revascularisation (TLR). Te target lesion was defned as the treated segment including the 5 mm margin proximal and distal to the stent [15].

Statistical Analysis.
Qualitative variables are summarized as frequencies and percentages, while quantitative variables are summarized as mean ± SD or median (25 th -75 th quartiles), depending on variables distribution. For continuous data, Student's t-test or nonparametric tests (Mann-Whitney U) were used for the comparison according to data distribution, while chi-square or Fischer's exact tests were used for the comparison of qualitative data. Survival curves were created using the Kaplan-Meier method and compared using log rank test. Univariate analysis was performed using Cox regression analysis, the hazard ratio (HR) and the 95% confdence interval (CI) were presented. Multivariate Cox regression analyses were performed using entry criteria of p < 0.1 in the univariate analysis. Statistical analysis was performed using SPSS v. 23 software (IBM, Armonk, NY, USA).

Periprocedural Complications and In-Hospital Outcome.
In-hospital outcomes and complications are outlined in Tables 3 and 4. Te angiographic success was achieved in 71 CTO RA procedures (88.8% vs 94.9%; p = 0.022), and procedural success rate was 80% in CTO versus 90.5% in non-CTO RA procedures (p = 0.003).
In-hospital MACE was comparable in both groups (8.8% vs 7%; p = 0.557). However, the incidence of slow fow (7.5% vs 2.9%; p = 0.030), coronary perforation (8.8% vs 2%; p < 0.001), and cardiac tamponade (5.0% vs 1.8%; p = 0.054) were higher in the CTO group. Coronary perforation (66.7%) was the main cause of cardiac tamponade, followed by the insertion of temporary pacemaker (27.8%). Coronary perforation occurred mainly because of high pressure infation of the balloons/stents (43.5%) followed by PCI wire perforation in 30.4%. Seven wire perforations were detected; 2 in CTO group (one due to Pilot 150 guidewire and the other due to the Rotawire), while in non-CTO group 5 wire perforation were detected (2 from the Rotawire, 2 by Whisper, and one by Pilot 50 guidewires). All wire-induced perforations were caused by unintentional pushing of the wires to very distal and/or in small branches.

Long-Term Outcome.
At two-year follow-up, we observed a higher TLF in the CTO group (25.5% vs 15.1%; log rank p = 0.041), that was mainly driven by higher cardiac     (Figure 1). Te 2-year clinically driven TVR was also comparable in both the study groups (CTO: 16.7% vs non-CTO: 10.5%; log rank p = 0.428). On multivariate regression analyses, the presence of CTO lesion, chronic kidney disease (CKD), periprocedural MI, and reduced LV-EF were independently associated with the two-year TLF. Table 5 lists the uni-and multivariate regression analyses of 2-year TLF predictors.

Discussion
Te main fndings of our study are as follows: (1) RA for CTO lesions was achieved with a high success rate. However, the overall complications were higher than non-CTO RA, (2) long-term follow-up revealed higher TLF in the CTO patients after RA, and (3) elective RA could be performed in calcifed CTO lesions with shorter procedure time and less dissections than bail-out procedures but with similar longterm outcome. Rotational atherectomy has a unique mechanism for treating calcifed lesions that difers from other tools such as orbital atherectomy, lithotripsy, and cutting balloons. It utilizes a rotating diamond tipped burr with a constant, circular orbit that ablates in the forward direction, in other words it creates its own pathway [16]. Terefore, RA appears to be the most suitable plaque modifying tool for treating balloon resistant CTO lesions, particularly balloon uncrossable lesions that prevent passage of the smallest available balloon, since the other available tools require lesion crossing frst.
Most available data on RA in CTO compared the feasibility and in-hospital outcomes in rotablated CTOs versus conventional (non-rotablated) CTO PCI. However, such comparison might be inappropriate and biased, as patients who need RA are usually older, have more comorbidities and more calcifed long lesions (higher J-CTO score) [17]. Moreover, all rotablated CTOs had a successful wire crossing (according to studies defnition), so it is not appropriate to compare the success rate of rotablated CTOs with other CTO PCIs that include lesions with a failure of wiring. Our study compared the outcome in CTO and non-CTO calcifed lesions that required RA, and so the baseline characteristics were nearly similar in both groups.
Te procedural success rate after RA for CTOs varies in the literature from 77% to 95.6% [6,18,19]. In our analysis, CTO RA showed lower success rate than non-CTO and it was mainly driven by the presence of more slow fow at the end of the procedure. However, it is comparable to the success rate in the previously mentioned studies. Te high rate of slow fow in our cohort could be attributed to the aggressive RA (using large burr und higher speed), so we recommend to use smaller burr (preferably 1.25 mm) and lower speed (140000 instead of 170000 RPM) in CTO RA to avoid slow fow. Additionally, the use of rotational atherectomy cocktail (nitro-glycerine, verapamil, and heparin) with performing short runs (20-30 s) might reduce the rate of slow fow. It is noteworthy that early generation DESs were more used in non-CTO group, while the newer generation DESs were more implanted in CTO patients, this discrepancy attributed to the low number of CTO patients who required RA in this time period (era of early generation DESs) in our centre (only 1.4% of patients with CTO Vs 21% with non-CTO were treated by RA before 2007). Additionally, the advancement of CTO tools and increasing the experience in dealing with such complex CTO lesions synchronised with the development of newer generation DES.
Te complication rate was clearly higher in the CTO group in the form of coronary perforation, slow fow, and tamponade. Interestingly, the main causes of these complications were not directly related to RA; coronary perforation occurred mainly after a high-pressure infation of balloons/stents, and due to PCI wire perforation, while direct burr-induced perforation only occurred in 4 cases (2 cases in each group). Te perforation rate in our CTO group was higher than that mentioned in other previous studies by Azzalini et al. [18] and Brinkmann et al. [19] that could be related to heavily calcifed lesions in older patients in our cohort and the aggressive lesion preparation so that it is preferred to avoid very high pressure infation of the balloons/stents after rotablation in order to avoid perforation. Additionally, temporary pacemaker insertion before RA was the second common cause of cardiac tamponade after coronary perforation. Elective RA is not a common approach for preparation of CTO lesions and the most common indication to rotablate the CTO is the presence of balloon-resistant lesions. However, in our study, 38.7% of CTO RA were performed electively due to the presence of angiographic severe calcifcation [10]. Te main benefts of elective RA versus bail-out were shorter procedure time and lower rate of coronary dissection. However, these benefts were not refected on the long-term outcome as both approaches had similar two-year TLF rate. Tese results are in the same line with the study by Allali et al. [20] that compared planned versus bail-out RA for heavily calcifed coronary lesions.
Data about long-term outcomes after RA in CTOs are scarce, only two studies compared the long-term outcome in rotablated CTO versus conventional CTO PCI, both included a relatively small number of CTO patients [18,21]. To the best of our knowledge, our study is the frst to report the long-term outcome after RA in CTO compared to non-CTO and to evaluate the role of elective RA in calcifed CTO lesions.
Our study reported a higher 2-year TLF rate in CTO patients that was driven by higher cardiac mortality. Although both the study groups had similar baseline characteristics, the CTO patients had a trend toward lower LV-EF. Moreover, higher cardiac mortality could be related to the higher rate of periprocedural complications in the CTO group, higher perforation, and tamponade rate in addition to a trend toward higher periprocedural MI. Te higher rate of perforation and tamponade might be associated with poor long-term survival as reported by Stathopoulos et al. [22]. Furthermore, slow fow at the end of CTO PCI procedure is associated with higher MACE according to the report by Wang et al. [14] and as previously mentioned, we can limit these complications by less aggressive balloon/stent dilation after RA and proper selection of burr size and speed. A multicentre study to identify a survival advantage score to predict the outcome after CTO PCI, reported that the association between three variables (multivessel disease, Canadian cardiovascular society (CCS) score ≥2, and previous MI) in CTO patients predict poor long-term outcome after CTO revascularisation [23]. In our CTO cohort, multivessel disease was present in 68.8% and about 58.7% had CCS ≥ 2, and 21.3% reported previous MI. Tis could be another explanation for higher cardiac mortality in CTO patients even after revascularisation.
It is worth mentioning that 2-year clinically driven TLR, clinically driven TVR, and target vessel MI were comparable in both groups which are considered important long-term efectiveness endpoints of RA for calcifed CTO lesions.  Te major limitation of this study is its nonrandomized design in which operator bias and unmeasured confounders may prohibit defnitive conclusions. Tis study is also retrospective and has been performed at a single centre. Moreover, the relatively small size of CTO RA group might have afected the statistical comparison of in-hospital complications and the long-term outcome during followup. However, our report still provides relevant data on the real-world utilization and outcomes of RA in a cohort of CTO patients.

Conclusion
Compared to non-CTO, RA in CTO is feasible with a high success rate and similar in-hospital outcomes but with higher periprocedural complications. Apart from higher cardiac deaths in CTO patients after RA due to higher clinical and procedural complexities, the long-term outcomes were comparable. Elective RA in CTO can shorten the procedure time and decrease the incidence of dissection in comparison with the bail-out RA.

Data Availability
Te data used to support the fndings of this study will be available from the corresponding author upon request.

Additional Points
Perspectives. What is known? RA for preparation of calcifed CTO lesions is feasible with high procedural success rate. What is new? Up to our knowledge, this is the frst study to compare the long-term outcome after RA in CTO versus non-CTO calcifed lesions and to evaluate the role of elective RA in calcifed-resistant CTO lesions. Tis study emphasis on the importance of RA in certain CTO lesions and gives insights into the long-term outcomes compared to the conventional RA for calcifed non-CTOs. Moreover, it sheds the light on elective RA for CTOs and its advantages over bail-out RA.

Disclosure
Tis research was performed as part of the employment (as clinical fellow in interventional cardiology) in Heart Centre, Segeberger Kliniken, Bad Segeberg, Germany.

Conflicts of Interest
M. Abdel-Wahab and G. Richardt report receiving an institutional research grant from Biotronik and St. Jude Medical. G. Richardt is an advisory board member for Abbott Vascular and Boston Scientifc and reports receiving lecture fees from Abbott Vascular, Biotronik, Boston Scientifc and Edwards Lifesciences. R. Toelg reports receiving honoraria for advisory board membership and lectures from Abbott Vascular and Biotronik. R. Hemetsberger received speaker's honoraria from Boston Scientifc. A. Allali is a proctor for Boston Scientifc. Te other authors declare that they have no conficts of interest.