Impact of Lesion Preparation Technique on Side Branch Compromise in Calcified Coronary Bifurcations: A Subgroup Analysis of the PREPARE-CALC Trial

Objectives To analyze the impact of different techniques of lesion preparation of severely calcified coronary bifurcation lesions. Background The impact of different techniques of lesion preparation of severely calcified coronary bifurcation lesions is poorly investigated. Methods We performed an as-treated analysis on 47 calcified bifurcation lesions treated with scoring/cutting balloons (SCB) and 68 lesions treated with rotational atherectomy (RA) in the PREPARE-CALC trial. Compromised side branch (SB) as assessed in the final angiogram was the primary outcome measure and was defined as any significant stenosis, dissection, or thrombolysis in myocardial infarction flow <3. Results True bifurcation lesions were present in 49% vs. 43% of cases in the SCB and RA groups, respectively. After stent implantation, SB balloon dilatation was necessary in around one-third of cases (36% vs. 38%; p = 0.82), and a two-stent technique was performed in 21.3% vs. 25% (p = 0.75). At the end of the procedure, the SB remained compromised in 15 lesions (32%) in the SCB group and 5 lesions (7%) in the RA group (p = 0.001). Large coronary dissections were more frequently observed in the SCB group (13% vs. 2%; p = 0.02). Postprocedural levels of cardiac biomarkers were significantly higher in patients with a compromised SB at the end of the procedure. Conclusions In the PREPARE-CALC trial, side branch compromise was more frequently observed after lesion preparation with SCB as compared with RA. Consequently, in calcified bifurcation lesions, an upfront debulking with an RA-based strategy might optimize the result in the side branch.


Introduction
Coronary calcification is encountered in around one of five percutaneous coronary intervention (PCI) procedures, and this rate is expected to further increase in an aging PCI population [1]. Severe forms of calcification impair balloon and stent delivery and increase the likelihood of stent under expansion and malapposition [2,3].
On the other hand, coronary bifurcation lesions account for 15-20% of PCIs [4,5]. As both the main and side branches are interpolated in a complex bifurcation core segment [6], PCI of coronary bifurcation lesions is associated with higher periprocedural complication rates as compared to nonbifurcation lesions [7].
Side branch compromise is an early described complication during PCI of bifurcation lesions [8]. e presence of initial side branch stenosis [9] and the bifurcation angle [10] are two well-known predictors of side branch compromise. In addition, main vessel plaque can cause carina shift leading to impairment of side branch flow [11]. Plaque composition including the presence of calcification together with plaque dimensions and the degree of stenosis in the main branch have also been reported as important predictors of side branch compromise [12,13].
In the PREPARE-CALC (the Comparison of Strategies to PREPARE Severely CALCified Coronary Lesions) randomized trial, we compared lesion preparation strategies of severely calcified coronary lesions using a scoring or cutting balloon (SCB) versus rotational atherectomy (RA) [14]. A trend was observed towards higher rates of side branch compromise in the SCB group in the total trial population. e current substudy of PREPARE-CALC aims to investigate the impact of different lesion preparation strategies of severely calcified coronary lesions located at a bifurcation segment on procedural and short-term outcomes.

Trial Design and Patient Population.
e trial design and study population have been described in detail elsewhere [14]. Briefly, PREPARE-CALC was an investigator-initiated, randomized controlled trial enrolling patients with documented myocardial ischemia and severely calcified native coronary lesions undergoing PCI. Between September 2014 and October 2017, two hundred patients with documented myocardial ischemia and severe calcification of the target native coronary lesion as defined by cineangiography (radiopacities noted without cardiac motion before contrast injection generally compromising both sides of the arterial lumen [15]) were enrolled in two high-volume centers in Germany.
Patients were randomized 1 : 1 to an initial strategy of lesion preparation using RA versus SCB followed by implantation of a new-generation sirolimus-eluting stent with bioabsorbable polymer (Orsiro; Biotronik AG, Bülach, Switzerland). Principal exclusion criteria were myocardial infarction (MI) within 1 week, decompensated heart failure, target lesions in coronary artery bypass grafts, in-stent restenosis, and target vessel thrombus. An independent clinical events committee, blinded to treatment assignment, adjudicated all major adverse events.
e study was approved by the local ethics committees of the participating centers, and each patient provided written informed consent for inclusion in the trial.

Bifurcation Subgroup.
We performed a post hoc astreated analysis of lesions located at a bifurcation segment. Lesions which did not involve a bifurcation were excluded. Coronary bifurcation lesions were defined as coronary artery narrowing occurring adjacent to and/or involving the origin of a significant side branch. A significant side branch was defined as having a diameter >2 mm or that supplies a significant myocardial area [16,17]. e decision of performing lesion preparation using RA or SCB in the bifurcation (only main branch, only side branch, or both) was left to the operator's discretion.
Patients in whom a crossover from a balloon-based strategy to rotational atherectomy was performed were included in the RA group. In addition to lesions and procedural characteristics collected during the study period, two interventional cardiologists reviewed all coronary angiograms and index procedures to collect more details concerning the anatomy, the treatment steps as well as the final angiographic result of the bifurcation lesions.
Compromised side branch at the end of the procedure was the primary outcome measure of the analysis and was defined as one or more of the following: any significant stenosis (>70% diameter stenosis), dissection of the side branch, and/or final thrombolysis in myocardial infarction (TIMI) flow <3. We also compared the evolution of cardiac biomarkers between patients with a compromised side branch and those without a compromised side branch at the end of the procedure.
Baseline, postprocedural, and follow-up coronary angiograms of the target lesions of the main trial were digitally recorded and assessed off-line by the quantitative angiographic core laboratory (ISA Research Center, Munich, Germany) with an automated edge detection system (QAngio, version 7.3; Medis Medical Imaging Systems, the Netherlands) by independent personnel blinded to treatment allocation [14]. Results concerning blinded QCA of target lesions involving a bifurcation are reported in this analysis.

Statistical Methods.
Statistical analysis was performed using Stata SE 14 (StataCorp LP, Texas, USA). Quantitative variables are summarized as mean ± SD or median (interquartile range (IQR)) and are compared by the 2-sided unpaired t-test. Categorical variables are summarized as frequencies and proportions and are compared by the chisquare or Fisher's exact tests. All probability values were two-tailed, and a p value of <0.05 was considered significant.

Study Population.
e study flowchart is illustrated in Figure 1, and baseline characteristics are summarized in Table 1. Of 278 lesions in the 200 patients enrolled in the PREPARE-CALC trial, we identified two study groups for the present analysis: the SCB group including 47 bifurcation lesions in 43 patients (mean age 75.0 ± 6.6 years; 79% males) and the RA group with 68 lesions located at bifurcations in 61 patients (mean age 74.8 ± 6.5 years, 77% males). In the RA group, 11 patients with 14 lesions had been initially randomized to an SCB strategy, but during the index procedure bailout RA was required. More patients in the SCB group had a history of previous myocardial infarction compared to the RA group (39% vs. 16%; p � 0.008). Other clinical characteristics were balanced between both groups.

Angiographic and Procedural Characteristics.
Angiographic and procedural characteristics are shown in Table 2. Lesions were mostly located in the LAD (60% in the SCB group and 63% in the RA group). Side branches were larger than 2 mm in 41 lesions (87%) in the SCB group and 54 lesions (79%) in the RA group; the angle between the side and main branch was most commonly less than 70°(62% vs. 66%; p � 0.62). Lesions were mostly classified as "1.1.0" and "1.1.1" according to the Medina classification (80.9% of lesions). Although a true bifurcation lesion (Medina x.x.1) was described in 49% of lesions in the SCB group and 43% in the RA group (p � 0.50), lesion preparation of the side branch with SCB or RA was performed in only 6.3% and 2.9%, respectively (p � 0.38).
A single-stent technique was used in most lesions (68% vs. 75%; p � 0.89), and in around one-third of cases, an   intervention in the side branch with balloon dilatation or kissing-balloon inflation was necessary (36% vs. 38%; p � 0.82). Proximal optimization technique (POT) was performed in 60% and 62% among SCB and RA groups, respectively. Culotte (12 lesions) and T-and-protrusion (9 lesions) stenting were the most frequently performed twostent techniques. e number of implanted stents per lesion was higher in the SCB group (1.85 ± 0.92 vs. 1.50 ± 0.68; p � 0.02). A compromised side branch at any time during the procedure was documented in 53% of lesions in the SCB group and 41% in the RA group (p � 0.20). At the end of the procedure, residual side branch compromise-the primary outcome measure of this study-was observed in 15 lesions (32%) in the SCB group and in 5 lesions (7%) in the RA group (p � 0.001).
Details of blinded core lab quantitative coronary angiographic analysis of the target lesions involving bifurcations are listed in Table S1. In general, angiographic characteristics of treated lesions, acute lumen gain after the procedure, and late lumen loss at 9 months were not significantly different between the two groups. Table 3. Fluoroscopy time and procedural duration were significantly higher in the RA group. Large coronary dissections were more frequently observed in the SCB group (13% vs. 2%; p � 0.02). In-hospital outcome was similar between both groups. Protocol-defined periprocedural myocardial infarction occurred in only one patient in the RA group.

Procedural and In-Hospital Events. Procedural and inhospital events are shown in
We further analyzed the evolution of cardiac biomarkers in patients with (n � 20) and without (n � 84) a compromised side branch at the end of the procedure until 24 hours postprocedure.
e median value of CK-MB was significantly higher at 16 hours post-PCI in patients with a compromised side branch (208 U/L vs. 39.5 UL; p � 0.04) with a trend towards higher troponin T levels (0.756 ng/ml vs. 0.101; p � 0.08) (Figure 2).

Discussion
e main findings of this study can be summarized as follows: (i) Comparing a strategy of balloon dilatation using a scoring/cutting balloon vs. an atheroablative method using rotational atherectomy in the main vessel of calcified lesions involving a bifurcation, the atheroablative strategy resulted in fewer compromised side branches at the end of the procedure (ii) Myocardial injury, as assessed by cardiac markers of necrosis was higher in patients with a compromised side branch at the end of the procedure Treatment of bifurcation lesions is a growing challenge in contemporary PCI because of the technical complexity, higher risk of periprocedural complications, and worse outcomes compared to nonbifurcation lesions [18]. In addition to the difficulties related to lesion preparation and device delivery, the presence of coronary calcification represents an additional challenge in treating bifurcation lesions. Fujino et al. demonstrated that the presence of coronary calcification at the site of a bifurcation assessed by optical coherence tomography is associated with a higher risk of side branch occlusion [19]. In this context, the main issues to provide optimal vessel patency are careful lesion preparation to "soften" the lesion, prevention of plaque shifting, and careful carina reconstruction [20].
In the single-arm prospective AGILITY (AngioSculpt Coronary Bifurcation Study) trial [21], the scoring balloon was used for the side branch dilatation prior to the deployment of a drug-eluting stent in the main vessel as a modified provisional strategy in true bifurcation lesions. In that study, the postscoring balloon dissection rate was 6% in the side branch and crossover to stent deployment of the affected side branch was required in only 10.8%. Despite the promising findings of this study, moderate to severe calcification was only present in 24.7% of the main vessel, whereas in the PREPARE-CALC trial severe coronary calcification was the main inclusion criterion. Furthermore, lesion preparation in our study was mainly performed in the main vessel rather than the side branch.
Several observational studies support the safety and effectiveness of RA in calcified bifurcation lesions, as high (>90%) success rates can be achieved, and the need for bailout side branch stenting is less than 20% [22,23]. Recently, Chambers et al. demonstrated in a series of patients undergoing lesion preparation with either orbital atherectomy or RA for severely calcified lesions similar low 30-day MACE rates when comparing patients with bifurcation versus nonbifurcation lesions [24].
Local factors such as hemodynamic forces play a major role in the formation of atherosclerosis and its regional distribution. Atherosclerosis has a predilection for the outer walls of bifurcations, where shear stress is lower and blood flow is turbulent, whereas the carina is typically spared from plaque precipitation. e mechanism of side branch closure due to calcium has not been fully elucidated, but a higher risk of carina shift due to reduced compliance of the wall opposing the side branch and a lower resistance encountered by the inflated balloon at the SB ostium could be a mechanism [25]. e differences in the mechanism of lesion preparation could therefore explain the superiority of RA over SCB in preventing side branch compromise observed in our study. Rotational atherectomy ablates hard plaque components while displacing and sparing soft tissues by the so-called differential cutting principle [26]. Although in current practice the role of RA has changed from aggressive debulking to plaque modification [27], the atheroablative effect of RA by reducing the volume of the plaque might still reduce the risk of carina shift during lesion preparation and after stent implantation.
Cutting and scoring balloons are designed to cut or fracture the continuity of fibrocalcific plaques without any atheroablative effect. erefore, the plaque compression during balloon inflation could result in carina shift and SB disturbance [28]. Moreover, in our analysis, we observed higher rates of coronary dissection in the SCB group (13% vs.  Figure 2: Evolution of median value of cardiac biomarkers in patients with (n�20) and without (n�84) compromised side branch at the end of the procedure. 6 Journal of Interventional Cardiology 2%; p � 0.02). e extension of coronary dissection from the treated main branch to the side branch might explain the more frequent side branch dissection and slow flow observed in that group. Although more compromised side branches were observed in the SCB group, the rate of SB stenting was not significantly different between the groups. is finding could be related to the higher number of stents implanted per lesion in the SCB group. Nevertheless, the analysis of cardiac biomarkers revealed that myocardial injury following PCI was more frequent in patients with a compromised side branch at the end of the procedure compared with those without SB compromise. Garcia-Garcia et al. showed in a recent analysis of pooled data from five coronary stent trials and one large registry that CK-MB elevation after PCI was independently associated with mortality at one year [29].

Study Limitations.
e PREPARE-CALC trial was powered to assess strategy success of SCB versus RA followed by DES implantation in calcified coronary lesions. e current analysis is a post hoc analysis consisting of about half of the patients originally enrolled in this trial; the results should therefore be considered hypothesis-generating as the groups were not adequately powered to detect a difference.
Another limitation arises from the definition of "significant side branch" which was based upon a subjective judgment of the operator as a branch that the operator does not want to lose. In addition, a dedicated bifurcation QCA was not performed. Compared to the high number of true bifurcation lesions, lesion preparation using SCB or RA in the side branch was performed in a very limited number of lesions (5 lesions overall) and this could affect the outcome in case of severe calcification in the side branch.
Finally, this analysis suffers from the same limitations as previously described for the PREPARE-CALC trial, i.e., the majority of cases in the SCB group were performed using scoring rather than cutting balloons. Patients with only stable coronary artery disease were included.

Conclusion
Comparing a strategy of SCB versus RA in severely calcified coronary bifurcation lesions, we observed a significantly higher rate of side branch compromise with an SCB-based strategy, which did not translate into worse short-term clinical outcome in this small cohort. Side branch compromise was associated with more extensive periprocedural myocardial injury. erefore, in calcified bifurcation lesions, an upfront debulking with an RA-based strategy might optimize the result of PCI in the side branch.

Data Availability
e data used to support the findings of this study may be released upon application to the steering committee of the PREPARE-CALC trial, who can be contacted at Heart Center Segeberger Kliniken GmbH, Am Kurpark 1, 23795 Bad Segeberg, Germany.