Percutaneous Coronary Intervention of Complex Calcified Lesions With Drug-Coated Balloon After Rotational Atherectomy

Objectives: We investigated the safety and efficacy of PCI using drug‐coated balloon (DCB) after rotational atherectomy (rotablation) in a retrospective single center study in patients with calcified de novo coronary lesions. The majority of patients had an increased risk for bleeding. Background: DCB has been effective in the treatment of in‐stent restenosis, small vessels, and bifurcations. DCB enables short one month dual antiplatelet treatment. No published data exist on the use of DCB after rotablation. Methods: 82 PCIs were performed in 65 patients (mean age 72 ± 10 years) using rotablation followed by DCB treatment. The median follow‐up time was 17 months. 82% of the patients had at least one risk factor for bleeding such as oral anticoagulation. 32% had an acute coronary syndrome. Median duration of dual antiplatelet treatment was 1 month. Results: MACE (the composite of cardiovascular death, ischemia‐driven target‐lesion revascularization [TLR] or non‐fatal myocardial infarction) occurred in 14% and 20% of the patients at 12 and 24 months, respectively. The rate of ischemia‐driven TLR was 1.5% at 12 months and 3.0% at 24 months. No acute closure of the treated vessel occurred. Bailout stenting was needed in 10% of the PCIs. The incidence of significant bleeding was 9% at 12 months. Conclusions: This is the first study to show that PCI using DCB after preparation of calcified lesions with rotablation is safe and effective. This novel strategy may be considered especially in patients with a bleeding risk such as those using an oral anticoagulant.

Rotational atherectomy (RA, “rotablation”) for the mechanical debulking of calcified coronary plaques before balloon angioplasty (plain‐old balloon angioplasty, POBA) or stenting was introduced in 1986.[1] In spite of early enthusiasm, the restenosis rate was as high as 30–40% after RA alone or followed by POBA, and 23–43% after stenting with a bare metal stent (BMS).[2] , [3] , [4] , [5] Routine RA of calcified lesions did not demonstrate a reduction in late lumen loss as compared with standard balloon predilation before drug eluting stent (DES) implantation.[6] The current guidelines suggest that the use of RA should be kept as a bailout tool in selected cases and DES should be favored over BMS after RA, if there is no contraindication for 6‐ to 12‐month use of dual‐antiplatelet treatment (DAPT).[7] , [8] , [9] The indications of RA are balloon‐undilatable or uncrossable lesions, chronic total occlusions, heavy diffuse calcification as well as calcified ostial and bifurcated lesions.[7]

Whereas restenosis has become quite a rare phenomenon in the era of modern DES,[10] bleeding after PCI has remained an important clinical problem in certain patient groups such as those with oral anticoagulation, advanced age, kidney failure, or prior bleeding. The current guidelines recommend that in patients with a bleeding risk the choice between BMS and DES should be planned individually, and the duration of DAPT should be kept as short as possible (1 month after both BMS and DES, European Society of Cardiology [ESC] recommendation Class IIa, evidence Level C).[9]

PCI using a balloon coated with paclitaxel, that is, drug‐coated balloon (DCB) has been shown to be safe and effective in the treatment of in‐stent restenosis. For this indication DCB has a Class IA ESC recommendation.[9] There is a growing body of evidence that DCB may also be used in the treatment of de novo coronary artery lesions with TLR rates of only 1–5%.[11] , [12] , [13] DCB has shown special potential in some anatomic scenarios. These include small coronary vessels, diffuse disease and bifurcations where stenting may yield suboptimal results.[14] , [15] , [16] , [17] The major advantage of DCB PCI is the short DAPT duration (1 month in stable patients). Furthermore, in case of severe bleeding DAPT can be stopped immediately, which is usually not possible after stenting. In this study, we investigated the efficacy and safety of PCI using DCB after RA in consecutive series of 65 patients with calcified lesions, the majority having a bleeding risk.

Eighty two PCIs were performed in 65 patients using rotablation followed by DCB treatment between March 2011 and December 2013 in a single center. The consecutive patients treated with RA followed by DCB using iopromide as a carrier agent (SeQuent Please, B. Braun) were included in the study. The patient characteristics are shown in Table [NaN] . The mean age was 72 ± 10 years, 68% were male and 37% were diabetic. Moreover, 25% had suffered prior myocardial infarction, 47% had reduced ejection fraction (<50%) and 31% had moderate to severe renal failure. Importantly, 82% of the patients had a risk factor for bleeding (listed in Table [NaN] ). All patients were markedly symptomatic before PCI; 71% had CCS 3–4 symptoms and 32% presented with an acute coronary syndrome (ACS).

Baseline Characteristics of Patients

1 *Medical or surgical; **Chronic Kidney Disease Epidemiology Collaboration (CKD‐EPI) formula; 1Data available on 56 patients; 2Data available on 36 patients.

All patients were on aspirin (100 mg per day). If it was not in use beforehand, 250–500 mg was given i.v. or p.o. immediately before PCI. Clopidogrel (loading dose 600 mg p.o) followed by 75 mg per day was used as an ADP blocker. Ticagrelol and prasugrel were not used. Twelve percent of the patients did not receive clopidogrel at all during or after PCI. These were patients who had a contraindication for DAPT such as recent severe bleeding or urgent upcoming non‐cardiac surgery. Elective patients received enoxaparin 0.5 mg/kg i.v. before PCI. ACS patients were administered enoxaparin 0.75–1.0 mg/kg × 2 s.c. except those on warfarin and INR >2.0 in whom warfarin was continued without subcutaneosly (s.c.) administered enoxaparin. Nitroprusside was given intracoronarily (i.c.) if needed to prevent slow flow before and immediately after rotablation runs. In the procedures of the RCA and dominant LCX theophylline infusion (0.05–0.1 mg/kg/min) was used during rotablation to prevent bradycardia. Temporary pacemaker was not used. Fifteen percent of the procedures were performed via the radial access.

The criteria for the use of RA were as follows; the lesion was either 1) not crossable by balloon; 2) not adequately dilatable with non‐compliant balloon or cutting balloon; or 3) heavily calcified. RA was performed using the Rotablator (Boston Scientific, Maple Grove, MN). The final burr‐vessel ratio was 0.5–0.7. The speed of the burr ranged between 140.000 and 180.000 rpm.[7] After RA, predilatation of the lesion was done using a semi‐compliant, non‐compliant, or cutting balloon at the operator's discretion. The largest predilation balloon diameter was chosen so that it was the same as the estimated diameter of the target coronary segment. Finally, the DCB treatment was performed according to the German consensus statement.[18] Briefly, the DCB‐vessel diameter ratio was 0.8–1.0 and the balloon was approx. Two millileter longer in the both ends than the treated lesion. DCB was kept dilated for 30–60 s at least at the nominal pressure. Several DCBs were applied to cover the whole stenotic segment, if needed. After removal of the guidewire, the final angiograms were taken using at least two projections. In case a flow limiting dissection (TIMI <3) or significant recoil (>30%), a bailout BMS was placed using the spot stenting strategy.

The postoperative care of the patients and the follow‐up was done according to normal local practices. The majority of the patients had the follow up visit in a community health center. No routine control angiograms were obtained. The central hospital and all community health centers in the region use a common electronic medical record allowing access to all patient data and laboratory parameters during the follow‐up. Data on the cause of deaths were obtained from national population registry. Occurrence of MACE, defined as the composite of cardiovascular (CV) death, non‐fatal MI, and ischemia‐driven TLR, was studied at 12 and 24 months from the index procedure. The median follow‐up time was 17 months. Bleeding events were collected according to BARC Consortium criteria.[19] BARC bleeding types 2–5 were considered significant. The study protocol was accepted by the local Ethics Committee for clinical studies.

Altogether, 107 lesions were treated in 65 patients. A total of 82 lesions were treated with RA followed by DCB (Table [NaN] ). In the vast majority of patients (78%) only one lesion was treated with the RA + DCB strategy. The most often used final burr and DCB sizes were 1.75–2.0 and 3.0–3.5 mm, respectively. Thus, large coronary branches were treated in this study (Table [NaN] ). In addition, the target lesions were long as the most often utilized DCB lengths were 20 mm (36%) and 30 mm (37%).

Baseline Characteristics of Lesions

2 DCB, drug eluting balloon; DES, drug eluting stent; BMS, bare metal stent; POBA, plain old balloon angioplasty; LAD, left anterior descending artery; RCA, right coronary artery; LCX, left circumflex artery; RPD, right posterior descending artery; RPL, right posterior lateral artery. * 2 lesions were treated with BMS and 1 with POBA after rotablation.

Size of Largest Rotablator Burr and the Diameter and Length of DCBs Used

LAD was the most common target vessel (46%). Ten left main (LM) lesions were treated with the RA + DCB strategy (12%). 98% of the lesions were de novo (only one procedure was done for restenosis). In third of the cases, cutting balloon was used for lesion preparation after RA and before DCB. The mean number of DCBs used per lesion was 1.1. The median duration of DAPT was 1 month and the mean duration 2.7 months. Significant bleeding (BARC2‐5) occurred in 9% of patients by 12 months. Bailout stent was needed in 10% of the procedures.

Figure [NaN] shows a typical calcified lesion in the proximal LCX of an 84‐year‐old female presenting with non‐ST elevation MI. The lesion was treated with RA followed by DCB (Fig. [NaN] A–C). Angiogram obtained 26 months later shows an excellent long term angiographic result (Fig. [NaN] D).

The overall MACE rate was 14% both in stable and ACS patients at 12 months and 20% at 24 months in the whole patient population (Fig. [NaN] A). At 24 months, the MACE rate appeared to be higher in ACS patients than in stable patients. The MACE comprised mostly of CV death and non‐fatal MI both at 12 and 24 months (Fig. [NaN] B). CV deaths doubled from 12 to 24 months. In stable patients the CV death rate was lower both at 12 and 24 months than in ACS patients (4.5% and 9.1% vs. 9.5% and 19.0%, respectively).

Ischemia‐driven TLR rate was very low during the study period: 1.5% and 3.1% at 12 and 24 months, respectively (Fig. [NaN] B). The Figure [NaN] C shows the components of MACE at 12 and 24 months in stable CAD and ACS populations. In the patients who underwent PCI for a de novo LM lesion using the RA + DCB strategy, there was one ischemia‐driven revascularization (8%) at 5 months and one CV death (8%) at 12 months. One bailout BMS was used in the LM PCIs (10%).

This is the first study to show that PCI of complex calcified lesions using RA followed by DCB treatment is safe. Proximal and long coronary lesions were treated in this study. RA followed by DCB offers many theoretical advantages over stenting of calcified lesions. Most importantly, DCB requires only short term DAPT (usually 1 month) which is beneficial in patients having a bleeding risk. Furthermore, since no metallic material is left in the coronary artery, DAPT may be stopped immediately in case of severe bleeding. Moreover, there is no risk of application of an undersized stent. Positive remodeling of the treated vessel has been demonstrated after DCB treatment.[20] , [21] , [22] Finally, PCI with DCB is often simpler and less time consuming than stenting especially in bifurcations.

In studies using DES after RA in complex calcified lesions, the cumulative TLR rate has varied between 2 and 12%.[5] , [23] , [24] , [25] , [26] , [27] , [28] In the only RCT (ROTAXUS trial) assessing routine RA before Taxus DES implantation, the TLR rate was high (12% at 9 months) but this is likely explained by the fact that control angiographies tend to increase the TLR rate.[6] The differences in the patient population and lesions treated may also contribute to the variation in the results. In our study, ischemia‐driven TLR rate was low (1.5% and 3.0% at 1 and 2 years, respectively), which corroborates the previous registry‐based studies using DCB in de novo lesions.[11] , [12] , [13] , [21] , [29] Lesions in the large vessels were treated, which is likely an important factor behind the low clinically indicated TLR rate. Also, the lack of routine noninvasive testing and follow‐up angiographies in part explain the low clinically indicated TLR rate in this study.

The overall MACE rate (14%) at 12 months after RA + DCB was similar as compared to published registries using RA + DES.[23] , [24] , [25] , [26] , [27] , [28] Approximately, half of the MACE rate in our data was driven by CV deaths. As expected, the risk of CV death was higher in patients presenting with ACS than in patients with stable CAD. This is in concordance with the published registry data on the RA + DES strategy.[25] , [27] The stent thrombosis risk has been quite high (0.8–4.2%) after RA followed by DES implantation,[5] , [10] , [23] , [25] , [26] which is at least partly explained by the risk of malapposition of stent struts in a calcified vessels. The acute vessel closure rate after DCB treatment in de novo lesions has been shown to be low in registries, ranging between 0% and 0.5%.[11] , [29] Importantly, in our study no acute closure of the treated vessel occurred showing the safety of leaving the vessel without a stent after RA.

Previously, RA + DES has been successfully applied also for LM lesions. However, this approach results in a very high TLR rate of 13–25%.[30] , [31] , [32] There is a small registry trial in which de novo LM lesions were treated with the DCB only strategy with 93% angiographic success at 4 months.[33] In our data, there were ten (12%) cases in which the LM lesion was treated by the RA + DCB strategy. One ischemia‐driven TLR occurred at 5 months and one CV death at 12 months in this population. PCI using DCB in the unprotected LM lesion deserves further research especially in patients with a contraindication for prolonged DAPT or suboptimal anatomy for bifurcation stent techniques.

Rotablation causes greater injury to the vessel than normal angioplasty, which likely triggers exaggerated neointimal proliferation and explains the high incidence of restenosis.[2] , [3] , [4] Administration of lipophilic paclitaxel inhibits the proliferation and migration of smooth muscle cells during the critical first 1–2 weeks after intimal injury and allows the later positive remodeling of the vessel.[20] , [21] , [29] At concentrations used in DCBs, paclitaxel has less effect on endothelial cells and does not induce significant apoptosis, which contributes to the normal endothelialization of the target vessel segment.[20] Chronic inflammation in the DCB‐treated vessel is avoided owing to the absence of polymers and foreign metallic material. There may be a concern that DCB is not effective in calcified lesions. However, RA reduces the calcific burden, which probably helps paclitaxel penetrate vessel layers. Experimental studies suggest that atherectomy does not compromise vascular healing after DCB treatment.[34] Importantly, atherectomy followed by DCB is already an established method in the percutaneous treatment of calcified femoro‐popliteal lesions in man.[35]

To achieve a low rate of restenosis with DCB, it is essential that stenting is limited to bailout situations (flow‐limiting dissection [TIMI flow <3] or significant recoil >30%).[18] Moreover, the decision to stent the lesion primarily instead of the DCB strategy should be done after proper predilation (balloon‐vessel ratio 1.0). Bailout BMS after DCB treatment strongly increases the risk of restenosis. This is most evident in small vessels where the use of BMS after DCB has been shown to increase the risk of MACE and TLR by 19‐ and 31‐fold, respectively.[17] In a recent work, it was shown that dissections with no impact on blood flow heal after DCB treatment without clinical sequalae.[29] Thus, minor or moderate dissections are not harmful but, instead, may facilitate better drug penetration and later positive enlargement of the vessel.[29] In contrast, significant recoil (>30%) increases the likehood of restenosis after DCB PCI. Fortunately, recoil is less common in calcified rota‐ablated lesions (having mechanical support from calcification) than in more elastic fibroatheromas.

Whereas restenosis has become a rare phenomenon, bleeding has remained perhaps the most important clinical problem of the contemporary PCI, especially in patients with a high risk, for example, due to the use of OAC. A recent registry report from Denmark examined outcomes of 8700 hospitalized patients with atrial fibrillation and stable CAD. The study showed that compared to OAC monotherapy, the risk of bleeding was increased when aspirin (hazard ratio 1.5), clopidogrel (hazard ratio 1.8), or both aspirin and clopidogrel (hazard ratio 2.8) were added to the regimen.[36] In the WOEST study, 44% of patients with triple therapy had a bleed as compared to 19% with double therapy (OAC + clopidogrel) at 1 year.[37] Because these detrimental observations, there is growing interest to address shorter DAPT in patients with bleeding risk.

In this study, 82% of the patients had a risk factor for bleeding and 40% used OAC. Although 32% of patients had ACS, the DAPT duration was kept to minimum (median 1 month and mean 2.7 months). We found that 9% of the patients had a significant bleeding episode during 1 year follow‐up (BARC 2–5), which we think is acceptable considering the high risk profile of the study population. It is noteworthy that 12% of the PCIs using RA + DCB were done using aspirin monotherapy without clopidogrel. PCI without any ADP blocker or discontinuation of DAPT during the first four weeks after PCI is not possible even with the newest generation DES.[10] DCB may offer a novel strategy for the management of patients suffering from myocardial ischemia and having an active bleed or patients waiting for emergent or urgent surgery.

There are some important limitations in this retrospective registry study. Firstly, as rotablation is still quite a rare procedure, the number of patients is relative low. Secondly, due to the retrospective nature of the study, there was no routine clinical follow‐up or routine control angiographies. We could not have a control group for RA + DCB because 1) there is no established alternative method of RA; 2) BMS after RA leads to unethically high restenosis rates; and 3) during the time of the study there was no 3rd generation DES available. Finally, the duration of DAPT was at the discretion of the operator.

We conclude that the stentless PCI strategy using RA followed by DCB treatment is safe and efficient in calcified complex lesions of patients with a bleeding risk. These results warrant a randomized trial to confirm the role of DCB in comparison to stenting possible using 3rd generation DES in these problematic patients.

All authors listed meet the authorship criteria according to the latest guidelines of the International Committee of Medical Journal Editors. All authors are in agreement with the manuscript.

This study was supported by VTR grant of Kuopio University Hospital and a grant from the Finnish Cardiac Society.

Additional supporting information may be found in the online version of this article at the publisher's web‐site.

Graph: Enlargement of the treated vessel after RA followed by DCB treatment in 84‐year‐old female with universal atherosclerosis presenting with non‐ST elevation MI. DES was planned to be avoided due to a history of subdural hematoma 4 years earlier. (A) angiogram shows a complex lesion of the proximal LCX with two stenosis (arrows) with ostial involvement. The lesion was non‐crossable with 1.25 mm balloon. (B) After lesion preparation with a 1.75 mm burr and predilation with a cutting balloon (3 × 10 mm), 3 × 30 mm DCB (asterisk) was dilated at the lesion for 30 s. (C) The final result immediately after DCB treatment. (D) 26 months after the index procedure the longterm angiographic outcome is excellent. Enlargement the treated vessel segment is observed (arrowheads), which may be due to positive remodelling as well vasodilatation. See online data supplement for videofiles of the angiograms.

Graph: MACE (cardiovascular death, non‐fatal myocardial infarction, and ischemia‐driven TLR) rates at 12 and 24 months after rotablation followed by DCB treatment (n = 65). (A) The overall MACE rate was 14% both in stable and ACS patients at 12 months. The MACE rate was higher in ACS patients than in stable patients at 24 months. (B) The MACE comprised mostly of cardiovascular death and non‐fatal MI both at 12 and 24 months. In contrast, ischemia‐driven TLR was low during the study period (1.5% and 3.1% at 12 and 24 months, respectively). (C) MACE components at 12 and 24 months in stable CAD and ACS. Cardiovascular mortality was high at 24 months in this population with advanced CAD. DES, drug‐eluting stent; MACE, major adverse cardiovascular events; DCB, drug‐coated balloon; ACS, acute coronary syndrome; TLR, target‐lesion revascularization.

Graph: Video S1. Coronary angiogram of 84yr old female with universal atherosclerosis presenting with non‐ST elevation MI. DES was planned to be avoided due to a history of subdural hematoma 4 years earlier.

Graph: Video S2. After lesion preparation with a 1.75 mm burr.

Graph: Video S3. After predilation with a cutting balloon (3 × 10 mm).

Graph: Video S4. 3 × 30 mm DCB was dilated at the lesion for 30 s covering the ostium of theLCX.

Graph: Video S5. The final result.

Graph: Video S6. Angiogram 26 months after the procedure shows excellent longterm angiographic outcome.