Rotational Atherectomy to Enable Sirolimus-Eluting Stent Implantation in Calcified, Nondilatable De Novo Coronary Artery Lesions

The sirolimus-eluting stent (SES) has demonstrated its angiographic and clinical efficacy in the suppression of coronary neointimal hyperplasia in a number of randomized clinical trials.^1–7 Observational studies have supported the trial results in patient and lesion subsets that were excluded from these trials.^8–14 The common denominator of all studies to date is that the targeted lesions were not subjected to any pretreatment other than predilatation. In particular, nondilatable calcified lesions were excluded from clinical trials and, to date, have not been studied in the context of drug-eluting stents (DES). Rotational atherectomy is the preferred treatment modality to debulk calcified lesions that do not lend themselves readily to stent implantation. It was the purpose of the present study to assess the mid-term clinical and angiographic outcomes of SES implantation after rotational atherectomy in native de novo calcified coronary artery lesions.

Methods

Between April, 2002, and May, 2005, 44 patients (mean age [± 1 SD] 67 ± 9 years) with a total of 47 native de novo coronary artery lesions underwent rotational atherectomy followed by SES (Cypher, Cordis, Miami Lakes, Florida) implantation at our institutions. Patients with a documented diagnosis of angina pectoris, unstable angina, or silent ischemia were included in this study if their baseline coronary angiogram showed a percent diameter stenosis of at least 70% (visual estimate), that could either not be crossed with a balloon catheter or could not be predilated with pressures up to 20 atm. All patients were informed about the interventional procedures and gave their written consent. According to standard care, patients were premedicated with 100 mg of aspirin, begun at least 12 hours before the procedure, and clopidogrel, administered either at 75 mg for at least 3 days before the intervention or as a loading dose of 300 mg before or immediately after the procedure. During the procedure, intravenous boluses of heparin were administered to maintain an activated clotting time in excess of 250 seconds. The use of glycoprotein 2b/3a receptor antagonists was left to the investigator’s discretion.

Rotational atherectomy. All study centers utilized the Rotablator^® rotational angioplasty system (Boston-Scientific, Natick, Massachusetts). A 0.009-inch guidewire, introduced by way of a 6–8 Fr arterial sheath and guiding catheter, was used to cross the lesion. A single burr was used per lesion, with burr size chosen to achieve an estimated burr/artery ratio between 0.5 and 0.7. Rotational atherectomy was performed in 20-second intervals at a speed of 170,000 rpm; care was taken to not let the speed drop by more than 5%. During activation of the rotablator, a saline coolant containing 5,000 IU of heparin, 5 mg of verapamil, and 1 mg of nitroglycerin per 500 ml was administered locally. The device was activated only when any changes in the electrocardiogram, induced by the preceding activation, had disappeared. Usually, 3–5 passes through the lesion were performed. Subsequent lesion dilation with a maximum pressure of 8 atm was performed at the investigator’s discretion, using a balloon shorter than the intended stent length and with a nominal diameter less than the reference vessel diameter.

Stent implantation. SES implantation followed standard interventional procedures.³ Stents were implanted with the intention that the total length of implanted stent(s) extended beyond the treated vessel segment by at least 5 mm proximally and distally. Overlapping of multiple stents by 2 to 4 mm was recommended. Stents were deployed with at least 12 atm. Postdilation of the stent(s) was mandatory in cases of bifurcational lesions (using the kissing-balloon technique¹¹) and left to the investigator’s discretion otherwise. Patients were discharged on a regimen of aspirin (100 mg/d indefinitely) and clopidogrel (75 mg/d for 6 months).

Quantitative coronary angiography. All coronary angiograms were analyzed independently by Mediolanum Cardio Research (Milan, Italy). The postintervention and follow-up analyses consisted of an “in-stent” analysis delimited by the margins of the implanted stent(s) and an “in-segment” analysis, which included the proximal and distal 5-mm vessel segments adjacent to the stent(s).

Follow-up. All patients were encouraged to return for clinical and angiographic follow-up at 6–8 months.

Definitions. Acute gain = minimum lumen diameter (MLD) postintervention minus MLD at baseline; Late loss = MLD postintervention minus MLD at follow-up; Loss index = late loss divided by acute gain; Binary restenosis = diameter stenosis ? 50% at follow-up; major adverse cardiac events (MACE) = cumulative incidence of all-cause deaths, (Q-wave and non-Q-wave) myocardial infarctions (MIs), and clinically driven target lesion revascularizations (TLRs) by either coronary artery bypass grafting or percutaneous catheter intervention.

Statistics. Continuous variables are presented as means ± 1 standard deviation or as medians plus interquartile range where appropriate. Nominal variables are presented as counts and percentages. Exact 95% confidence intervals (CIs) were calculated based on the binomial distribution. Comparisons between nominal variables were performed by Fisher’s exact test, with a two-sided p-value Results Procedural outcomes. All procedures were successfully completed as intended, with a single stent implanted in 35 lesions (74.5%) and multiple stents implanted in the remaining 12 lesions, for a total of 61 stents. Mean deployment pressure was 15.4 ± 2.4 atmospheres. High-pressure balloon postdilation of the stent was deemed necessary in 12 lesions (25.5%). Total implanted stent length amounted to 34 ± 15 mm. The most frequently used burr sizes were 1.50 mm (n = 20) and 1.75 mm (n = 15). Clinical outcomes. Patients were followed clinically for a mean of 7.7 months. In-hospital, acute stent thrombosis resulting in a Q-wave MI occurred in 1 patient. The patient, a 53-year-old diabetic woman with hypertension, hyperlipidemia and a history of prior percutaneous coronary intervention, received fibrinolytic therapy and recovered uneventfully; she died suddenly 7 months after the index treatment. Another 59-year-old male patient with 3-vessel coronary artery disease and a long history of prior surgical and percutaneous coronary interventions died during emergency coronary artery bypass grafting of the target vessel 3 months after the index treatment (the patient’s bypass surgery was counted as a TLR). TLRs by catheter intervention were required in 3 patients (6.8%; 95% CI, 1.4–18.7%). One of these patients experienced chest pain 6 hours postintervention and was found to have a distal-stent dissection that was covered with another SES. Two other patients presented with symptomatic 75% and 100% target-lesion in-stent restenosis at 5 and 4 months, respectively; the stenosis was retreated successfully with balloon angioplasty, and the total occlusion was covered with a paclitaxel-eluting stent following directional atherectomy. Thus, the cumulative incidence of MACE in our series was 11.4% (95% CI, 3.8–24.6%).

Quantitative coronary angiography. At the completion of the intervention, the MLD was increased from a mean of 0.75 mm to an in-stent mean of 2.50 mm, reflecting a mean acute gain of 1.74 mm. Angiographic follow-up at a mean of 7.0 months was obtained from 29 patients (66%) who had received treatment in 31 lesions. Of the remaining patients, 2 had died before their scheduled angiographic reassessment and 13 refused angiographic follow-up. With a median in-stent late loss of 0.31 mm and a median in-segment late loss of 0.22 mm, diameter stenosis at follow-up averaged 27.1% in-stent and 36.6% in-segment. Restenosis was observed within the stent in 3 lesions (9.7%; 95% CI, 2.0–25.8%), at the proximal stent margin in 2 lesions (6.5%; 95% CI, 0.8–21.4%), and at the distal stent margin in 1 lesion (3.2%; 95% CI, 0.1–16.7%). All restenotic lesions were singular and, except for 1 reocclusion and 1 12-mm in-stent restenosis, focal (mean length 7.5 ± 1.0 mm). Both proximal-margin restenoses were located at the site of the postinterventional MLD. Thus, in-segment binary restenosis occurred in a total of 6 lesions (19.4%; 95% CI, 7.5–37.5%). Diameter stenosis for these 6 restenotic lesions ranged from 55–100% (median, 60%); 2 of these lesions were symptomatic and required a TLR (see above). In-segment restenosis was found in 5 of 23 lesions (21.7%) treated with a single stent and in 1 of 8 lesions (12.5%) that had received multiple stents (p > 0.999). Baseline lesion length also had no impact on restenoses, with 3 (21.4%) occurring in 14 lesions ? 20 mm and 3 (17.6%) occurring in 17 lesions > 20 mm (p > 0.999). However, bifurcational lesions appeared to have an increased risk of restenosis (2 of 6 lesions [33.3%; 95% CI, 4.3–77.7%] versus 4 of 25 nonbifurcational lesions [16.0%; 95% CI, 4.5–36.1%]; p = 0.567).

Discussion

Clinical outcomes. In this study, rotational atherectomy enabled SES implantation for calcified de-novo coronary artery lesions in a cohort of 44 patients at high risk for major adverse events. All patients had nondilatable lesions, about 50% of which were longer than 20 mm, 30% of our patients had previously undergone coronary artery bypass grafting, 50% had a history of prior coronary intervention, and 93% presented with multivessel disease. The combined approach of rotational atherectomy and SES implantation was found to be efficacious and safe. Following the index intervention, 1 Q-wave MI secondary to acute stent thrombosis and 1 target lesion revascularization (TLR) occurred in-hospital. The cumulative in- and out-of-hospital incidences at a mean of 7.7 months of TLR and MACE of 9.1% (4 patients) and 11.4% (5 patients), respectively, compared favorably with those observed in randomized controlled trials involving the SES.^1–7 The clinical outcome of our patients is all the more remarkable since randomized controlled trials explicitly excluded nondilatable calcified lesions. The recently published Sirolimus-Eluting and Paclitaxel-Eluting Stents for Coronary Revascularization (SIRTAX) and Sirolimus versus paclitaxel-eluting stents in de novo coronary artery lesions (REALITY) trials both had a prevalence of about 35% of moderately to severely calcified lesions treated with SES.^5,6 However, these lesions were “suitable for stent implantation”5 or did not “preclude predilatation”6 and were on average shorter than in our study (17.0 mm in SIRTAX and 11.8 mm in REALITY). The 9-month and 12-month composite TLR and MACE rates for the SES in SIRTAX and REALITY were 4.8%, 6.0%, 6.2% and 10.7%, respectively. Although a strict comparison of these trials with our study is confounded by the more complex nature of the lesions pretreated with rotational atherectomy, we formally compared our TLR and MACE rates with those of the SIRTAX trial. The comparison yielded an odds ratio associated with the rotational atherectomy approach of 2.0 for both TLR and MACE. However, the 95% confidence intervals of the odds ratios of 0.7–6.0 for TLR and 0.7 to 5.3 for MACE indicated a lack of statistical significance as both include the value 1 (no increased odds). Thus, clinically, our patients did not appear to fare worse than those studied in the aforementioned randomized trials.

Quantitative coronary angiography. Except for lesion length, baseline and post-intervention quantitative angiographic parameters in our patients were consistent with those observed in randomized trials of the SES. Quantitative coronary angiography obtained from 29 patients (31 lesions) at a mean follow-up of 7 months revealed a median late loss of 0.31 mm in-stent and 0.22 mm in-segment. Although we chose to represent the late-loss distributions by their medians because of a pronounced right-skewed distribution of in-stent late loss in particular (less so in-segment late loss), both distributions are in the range of late losses after SES implantation observed previously. For instance, the SIRIUS investigators reported a mean in-stent late loss at 8 months of 0.17 (± 0.45) mm and a mean in-segment late loss of 0.24 (± 47) mm.² In E-SIRIUS corresponding mean late losses of 0.20 (± 0.38) mm and 0.19 (± 0.38) mm were found,³ and the SIRTAX investigators reported 0.12 (± 0.36) mm and 0.19 (± 0.45) mm, respectively.⁵ Compared with these trials, there appeared to be a trend towards an increased in-stent late loss in our study. Considering the fact that the mean reference vessel diameter and mean post-intervention MLD in our study were virtually identical with those found in E-SIRIUS (2.63 mm vs. 2.60 mm and 2.50 mm vs. 2.43 mm, respectively) and only about 6% less than in Sirolimus-Eluting Versus a Standard Stent in Patients at High Risk for Coronary Restenosis (SIRIUS) and SIRTAX, the increased late loss may be explained by the fact that, although rotational atherectomy provides a channel through the lesion, it does not completely abrogate all calcium. It is therefore conceivable that the residual calcium impairs the diffusion of sirolimus into the vessel wall and thus reduces its efficacy to a certain extent. Yet, compared with the late loss of about 1.1 mm observed after rotablation-mediated bare-metal stent implantation,¹⁵ it can be stated that the power of the SES to suppress neointimal hyperplasia appeared to be well in effect in rotablated lesions. Binary restenosis rates, both in-stent and in-segment, were markedly higher in our study (9.7% and 19.4%, respectively) than in randomized SES trials involving less complex lesions. Since restenosis rates tend to increase with lesion length also for the SES,¹⁶ the increased mean length of the lesions treated in our study (24.5 mm for lesions with angiographic follow-up) may by itself have accounted for the increased in-stent restenosis rate. Three lesions developed restenosis outside the stent margins (2 proximal, 1 distal), causing our in-segment restenosis rate to be twice as high as the in-stent restenosis rate. By way of an explanation, it must be appreciated that rotablation is always begun at a not clearly defined point proximal to the lesion. Thus, it is conceivable that proximal vessel segments injured by rotablation were not adequately covered with SES. The distal-margin restenosis appeared to be due to the progression of atherosclerosis. Rotational atherectomy before SES implantation did not seem to impact the morphology of restenotic lesions previously described as predominantly focal.¹⁷

Clinical relevance. Calcified lesions are known to be associated with an increased risk of post-stenting complications, such as dissections, vessel perforations, and a high calcium content frequently disallows complete stent expansion. The resulting impaired wall apposition increases the risk of stent thrombosis and restenosis.¹⁸ Nondilatable calcified lesions have been excluded from all clinical DES trials to date. They were shown in our study to be amenable to rotational atherectomy-mediated SES implantation with acceptable mid-term clinical results. Particular care is advisable to liberally stent the proximal portion of rotablated lesions.

Limitations

This is an observational study justified by the fact that the coronary lesions studied could not be treated without rotational atherectomy. The sample size of 44 patients (47 lesions) is fairly small, giving rise to wide 95% confidence intervals for proportions. Angiographic follow-up was obtained from only two-thirds of patients. At a mean of 7 months, angiographic follow up was 1 month short of the 8 months prespecified in the SIRIUS^2–4 and SIRTAX⁵ trials. Intravascular ultrasound to assess wall apposition of the stents was not performed.

Conclusions

This study showed that rotational atherectomy to enable SES implantation in long, nondilatable calcified coronary lesions is feasible and safe. The mid-term clinical outcome of our patients appeared not to be worse than that of patients undergoing SES implantation without rotational atherectomy for less complex lesions. Quantitative coronary angiography in a subset of patients underscored the need for extended proximal-lesion stent coverage.