Angiographic Findings of the Multicenter Randomized Study With the Sirolimus-Eluting Bx Velocity Balloon-Expandable Stent (RAVEL)
Sirolimus-Eluting Stents Inhibit Restenosis Irrespective of the Vessel Size
Background— Restenosis remains the major limitation of coronary catheter-based intervention. In small vessels, the amount of neointimal tissue is disproportionately greater than the vessel caliber, resulting in higher restenosis rates. In the Randomized Study With the Sirolimus-Eluting Bx Velocity Balloon-Expandable Stent (RAVEL) trial, ≈40% of the vessels were small (<2.5 mm). The present study evaluates the relationship between angiographic outcome and vessel diameter for sirolimus-eluting stents.
Methods and Results— Patients were randomized to receive either an 18-mm bare metal Bx VELOCITY (BS group, n=118), or a sirolimus-eluting Bx VELOCITY stent (SES group, n=120). Subgroups were stratified into terciles according to their reference diameter (RD; stratum I, RD <2.36 mm; stratum II, RD 2.36 mm to 2.84 mm; stratum III, RD >2.84 mm). At 6-month follow-up, the restenosis rate in the SES group was 0% in all strata (versus 35%, 26%, and 20%, respectively, in the BS group). In-stent late loss was 0.01±0.25 versus 0.80±0.43 mm in stratum I, 0.01±0.38 versus 0.88±0.57 mm in stratum II, and −0.06±0.35 versus 0.74±0.57 mm in stratum III (SES versus BS). In SES, the minimal lumen diameter (MLD) remained unchanged (Δ −0.72 to 0.72 mm) in 97% of the lesions and increased (=late gain, ΔMLD <−0.72 mm) in 3% of the lesions. Multivariate predictors for late loss were treatment allocation (P<0.001) and postprocedural MLD (P= 0.008).
Conclusions— Sirolimus-eluting stents prevent neointimal proliferation and late lumen loss irrespective of the vessel diameter. The classic inverse relationship between vessel diameter and restenosis rate was seen in the bare stent group but not in the sirolimus-eluting stent group.
Received April 25, 2002; revision received July 18, 2002; accepted July 29, 2002.
Restenosis remains the major limitation of coronary catheter-based intervention.1 In stented vessels, the major contributor to restenosis is neointimal proliferation, which is a ubiquitous, local, vascular reaction to catheter-induced vessel injury.2 Vessel diameter is an established predictor of angiographic outcome after catheter-based intervention, with a higher restenosis rate in smaller vessels.3 This is because of the disproportionately greater amount of neointimal tissue relative to the vessel caliber.4 Although coronary stents provide major benefits versus simple balloon angioplasty by inhibiting acute vessel closure, early vessel recoil, and late vessel constriction, they stimulate neointimal proliferation. Therefore, restenosis rates in small vessels may be similarly high with these 2 treatment modalities.5,6⇓ Inhibition of neointimal proliferation by local pharmacological interventions is a promising concept. Sirolimus (rapamycin) is an immunosuppressive drug approved for the prevention of renal transplant rejection. It also has potent antiproliferative and antimigratory effects on vascular smooth muscle cells.7 Recent clinical experience with sirolimus-eluting coronary stents has shown excellent results, with 0% restenosis at 4-month,8 6-month,9 and 12-month follow-up.10 At the time of these pilot studies, sirolimus-eluting stents were only available in a 3.0-mm or 3.5-mm diameter, limiting treatment to relatively large vessels. In the Randomized Study With the Sirolimus-Eluting Bx Velocity Balloon-Expandable Stent (RAVEL) trial, a smaller sirolimus-eluting stent with a diameter of 2.5 mm was available, and it allowed smaller vessels to be stented. This small sirolimus-eluting stent was used in 18% of patients.11 The present study investigated the relationship between angiographic outcome and vessel diameter for sirolimus-eluting stents compared with bare metal stents.
Patients and Stent Implantation
The patient population and stent implantation technique have been described in detail elsewhere.11 The 238 patients enrolled in the RAVEL trial had a single de novo lesion of a native coronary artery.
Patients were randomized (double-blind) for implantation of either an 18-mm uncoated bare metal Bx VELOCITY stent (BS), or a sirolimus-eluting Bx VELOCITY balloon-expandable stent (Cordis Corp, Johnson & Johnson) (SES). All drug-eluting Bx VELOCITY stents contained 140 μg sirolimus/cm2 (±10%). Total sirolimus content was 153 μg (±10%) on the 6-cell stent (2.5 and 3.0 mm in diameter) and 180 μg (±10%) on the 7-cell stent (3.5 mm in diameter). This difference in content was due to the differences in the surface area of the two stents. Stent implantation was performed in the conventional manner after predilation. Postdilatation was performed as necessary to achieve a residual stenosis below 20% with TIMI grade III flow. Patients received aspirin (at least 100 mg) indefinitely with either clopidogrel (75 mg daily) or ticlopidine (250 mg, twice daily) for 8 weeks.
Quantitative Coronary Angiographic Analysis
Coronary angiograms were obtained in multiple views after intracoronary injection of nitrates. Quantitative analyses by edge-detection techniques were performed by an independent core laboratory (Cardialysis BV) blinded to treatment allocation. Reference diameter (RD), minimal luminal diameter (MLD), and degree of stenosis (as percentage of diameter) were measured before dilatation, at the end of the procedure, and at a 6-month follow-up. Restenosis was defined as >50% diameter stenosis at follow-up. Late loss was defined as MLD after the procedure minus MLD at follow-up.
The target lesion was defined as the stent segment plus 5 mm proximal and 5 mm distal to the edge of the stent. The vessel segment was defined as the segment bounded by side branches proximal and distal to the stent segment (Figure 1).
The accuracy of the method has been reported in detail.12 Given the accuracy of quantitative coronary angiography for MLD measurements, we used 2 standard deviations12 as the cut-off point for the classification of late loss indicating whether MLD was unchanged (no loss, ΔMLD −0.72 to 0.72 mm), reduced (late loss, ΔMLD >0.72 mm), or larger (late gain, ΔMLD <−0.72 mm, “negative late loss”) at follow-up.13
Both groups were stratified according to their vessel diameter. Vessel diameter was defined as the baseline RD in the vessel segment analysis before intervention. The terciles for the RD were calculated and used as cut-off points for subgroup definition.
Sample Size Estimation and Statistical Analysis Based on Late Loss
A sample size of 95 in each group had 87% power to detect a difference in means of 0.25 mm (the difference between a bare stent late loss mean, μB, of 0.80 mm and a sirolimus stent late loss mean, μS, of 0.55 mm), assuming that the common standard deviation is 0.55 using a 2-group t test with a 0.05 1-sided significance level. The sample size was increased to 110 in each group to account for noncompliance to 6-month angiographic follow-up.
Data are presented as mean±SD or proportions. For comparison of continuous data, a 2-tailed Student’s t test was performed. A value of P<0.05 was considered significant. To identify the factors that might be related to late lumen loss, linear regression analyses were performed. Predictors were chosen by stepwise linear regression using an entry criterion of 0.20 and a stay criterion of 0.05.
The 238 patients were randomly assigned (SES, n=120; BS, n=118). There were no significant differences with regard to procedural success (96.6% versus 93.1%), stents per patient (1.0±0.3 versus 1.1±0.3), and nominal stent diameter (3.06±0.34 mm versus 3.10±29 mm; SES versus BS, respectively).
Before the procedure, RD (2.60±0.54 mm versus 2.64±0.52 mm) and MLD (0.94±0.31 mm versus 0.95± 0.35 mm) were similar in both groups. After the procedure, there were also no meaningful differences (postprocedural RD, 2.62±0.44 mm versus 2.68±0.45 mm; postprocedural MLD, 2.43±0.41 mm versus 2.41±0.40 mm; SES versus BS, respectively). At follow-up, the SES group showed a larger MLD (2.42±0.49 mm versus 1.64±0.59 mm, P<0.001) and lower late lumen loss (−0.01±0.33 mm versus 0.80±0.53 mm, P<0.001). Binary restenosis was 0.0% in the SES group and 26.6% in the BS group (P<0.001).
Figure 2 illustrates the relation between postprocedural MLD and MLD at follow-up. In the SES group, the MLD (Figure 2A) remained basically unchanged; late loss was seen in 1 lesion and late gain was seen in 4 lesions (3%). In contrast, lumen reduction over time was seen in approximately half of the BS patients (n=55, 47%), and no late gain was seen. A similar pattern was found for the mean diameter over the entire length of the stent (Figure 2B).
Subgroups were stratified according to their RD (Figure 3). There were no significant differences in baseline patient and lesion characteristics in the SES and BS subgroups. There were also no significant differences in procedural parameters (Table 1⇓).
Analysis of the strata revealed a higher proportion of diabetic patients in small and intermediate vessels. The stent implantation procedure showed a decreasing balloon to artery ratio (stratum I versus stratum III: P<0.001 in both, BS and SES group) and increasing inflation pressure from stratum I to stratum III (stratum I versus stratum III: P<0.01 SES group; P=0.22 BS group).
Table 2 summarizes the key angiographic data. Vessel segment analysis showed similar preprocedural and postprocedural MLD in both treatment groups throughout corresponding strata.
Restenosis, Late Lumen Loss, and Vessel Size
At follow-up, the MLD was consistently larger in the SES groups. In all strata, the restenosis rate was 0% in the SES groups, with extremely low and consistently uniform late loss. In the BS strata, the classic inverse relationship between restenosis rate and vessel diameter was seen. Restenosis rate virtually doubled with decreasing vessel size from 20% in large vessels (stratum III) to 35% in small vessels (stratum I). The amount of late loss, however, was similar in the 3 groups (0.80 mm in stratum I, 0.88 mm in stratum II, and 0.74 mm in stratum III). Therefore, the observed increase in restenosis rate in smaller vessels in this series is driven largely by the relative amount of obstruction as a function of vessel diameter rather than being due to an absolute increase in neointimal hyperplasia in smaller vessels.
Vessel Segment Analysis
Vessel segment analysis revealed minimal late gain in both the MLD and RD over time in SES subgroups but not in BS groups (Table 2).
Target Lesion Analysis (Including Stent Segment and the Proximal and Distal Edges)
The SES subgroups showed minimal late loss at the stent segment (0.01±0.25 mm, 0.01±0.38 mm, and − 0.06±0.35 mm in strata I, II, and III, respectively) and proximal edges (0.04±0.34 mm, 0.08±0.42 mm, and 0.03±0.43 mm in strata I, II, and III, respectively), whereas the distal SES edges had minimal late gain (−0.05±0.29 mm, −0.14±0.31 mm, and −0.09±0.31 mm in strata I, II, and III, respectively). In contrast, the BS subgroups showed pronounced late loss in the stent segment and moderate late loss at the proximal and distal edges.
We investigated the relationship between vessel diameter and angiographic outcome 6 months after sirolimus-eluting stent implantation in patients in the RAVEL trial. The main findings of the study are that sirolimus-eluting stents prevent restenosis irrespective of vessel diameter and do not show the classic inverse relationship of vessel diameter to restenosis rate.
Quantitative coronary angiography convincingly demonstrates the absence of neointimal proliferation and restenosis in all patients treated with the sirolimus-eluting stent within the first 6 months, unlike those treated with bare metal stents. This truly remarkable finding creates a totally new paradigm in interventional cardiology and puts paid to the well-established existing paradigm, the classic inverse relationship between vessel diameter and restenosis rate.3
Prevention of Neointima Growth
Neointimal growth is a normal reaction to vascular injury. Smooth muscle cells are considered to be the main components of coronary artery neointima after stent implantation, and the severity of the reaction may be modulated by the extent of stent-induced vessel injury14 and the inflammatory reaction around the stent struts.15
Vessel injury is influenced by stent surface material, geometric configuration, implantation technique, and vessel size.16 Neointimal hyperplasia and persistent tissue proliferation are related to the degree of vessel injury (balloon/artery ratio×inflation pressure).17
In our patients, 2 stent configurations (6-cell and 7-cell designs) were used. Stent implantation technique varied with vessel size. In small vessels, a relatively higher balloon to artery ratio of 1.3 was achieved, whereas the balloon to artery ratio was lower (1.0) in large vessels. Conversely, the inflation pressure was lower in small vessels than in larger vessels (14 atm versus 16 atm).
In the present study, the effectiveness of the sirolimus-eluting stent was extremely strong and was affected neither by established risk factors for restenosis nor by stent configuration, balloon to artery ratio, or balloon pressure. Other than treatment allocation, the only independent predictor for late loss was the postprocedural MLD.
The very low late loss, which is consistently reported in all studies with sirolimus-eluting stents,8–10⇓⇓ raised concerns about late lumen enlargement. In the present study, there was evidence of late lumen gain (negative late lumen loss) in 3% of SES patients.
Furthermore, minimal but consistently negative late loss was seen at the distal edges of the stent. This phenomenon might be related to the downstream elution of the drug.
Although the finding of late lumen gain in a very small percentage of patients is interesting, it is worth noting that there have been no clinical events attributable to this phenomenon in the patients treated with the sirolimus-eluting stent at 1-year follow-up, or in the patients of Sousa et al10 for up to 2 years. Mechanistic angiographic analysis of the Sao Paulo series10 showed stable lumen dimensions with minimal late lumen loss between 4- and 12-month follow-up (in-stent MLD 2.90±05 mm at 4 months and 2.87±0.4 mm at 12 month; slow-release group) that matches well with the stable clinical result.
The Importance of Late Loss as a Predictor of Restenosis
The classic inverse relationship between vessel diameter and restenosis rate was not seen in the sirolimus-eluting stent group. This offers new therapeutic options for small vessels, in which conventional stenting is of questionable value.5 This is especially true for diabetic patients, who often have small arteries because of diffuse coronary artery disease.18 In addition, they frequently have an exaggerated neointimal proliferative response that manifests as significantly greater late loss at the treatment site and a resultant 2-fold increase in in-stent restenosis in small vessels (44% versus 23%, P=0.002) as compared with nondiabetic patients with similar-sized vessels.19 In our study, diabetes mellitus did not attenuate the effectiveness of the sirolimus-eluting stent. These findings contrast markedly with what was seen in the bare stent group. Restenosis rates almost doubled from the tertile with the largest diameter vessels to the one with the smallest vessels (20% to 35%), whereas late loss increased only modestly (0.74 mm to 0.80 mm). This dramatic increase in restenosis rate is explicable on the basis of hydraulics. A late loss of 0.80 mm in a 3.0-mm diameter vessel versus a 2.0-mm diameter vessel results in a 46% versus a 64% obstruction. Late loss is the most sensitive and operator-independent assessment of the effect of drug-eluting stents and can be used to predict what the restenosis rate will be in vessels of different diameters. Simply reporting angiographic restenosis rates, which can be influenced by case selection and operator techniques, is no longer sufficient in the era of drug-eluting stents.
Sirolimus-eluting stents prevent neointimal proliferation and late lumen loss irrespective of the vessel size. The classic inverse relationship between vessel diameter and restenosis rate was seen in the bare stent group but not in the sirolimus-eluting stent group. This finding with the sirolimus-eluting stent has the potential to considerably expand the use of these stents in smaller vessels and to eliminate the present difference in reintervention rates between patients treated with coronary artery bypass surgery and stenting.20
Dr Regar is supported by a grant of the Deutsche Forschungsgemeinschaft. We thank Dr B. Firth for his critical review of the manuscript.
- ↵Foley DP, Melkert R, Serruys PW. Influence of coronary vessel size on renarrowing process and late angiographic outcome after successful balloon angioplasty. Circulation. 1994; 90: 1239–1251.
- ↵Doucet S, Schalij MJ, Vrolix MC, et al. Stent placement to prevent restenosis after angioplasty in small coronary arteries. Circulation. 2001; 104: 2029–2033.
- ↵Kastrati A, Schomig A, Dirschinger J, et al. A randomized trial comparing stenting with balloon angioplasty in small vessels in patients with symptomatic coronary artery disease. ISAR- SMART Study Investigators. Intracoronary Stenting or Angioplasty for Restenosis Reduction in Small Arteries. Circulation. 2000; 102: 2593–2598.
- ↵Marx SO, Marks AR. Bench to bedside: the development of rapamycin and its application to stent restenosis. Circulation. 2001; 104: 852–855.
- ↵Sousa JE, Costa MA, Abizaid A, et al. Lack of neointimal proliferation after implantation of sirolimus-coated stents in human coronary arteries: a quantitative coronary angiography and three-dimensional intravascular ultrasound study. Circulation. 2001; 103: 192–195.
- ↵Rensing BJ, Vos J, Smits PC, et al. Coronary restenosis elimination with a sirolimus eluting stent: first European human experience with six month angiographic and intravascular ultrasonic follow-up. Eur Heart J. 2001; 22: 2125–2130.
- ↵Sousa JE, Costa MA, Abizaid AC, et al. Sustained suppression of neointimal proliferation by sirolimus-eluting stents: one-year angiographic and intravascular ultrasound follow-up. Circulation. 2001; 104: 2007–2011.
- ↵Reiber JH, Serruys PW, Kooijman CJ, et al. Assessment of short-, medium-, and long-term variations in arterial dimensions from computer-assisted quantitation of coronary cineangiograms. Circulation. 1985; 71: 280–288.
- ↵Serruys PW, Luijten HE, Beatt KJ, et al. Incidence of restenosis after successful coronary angioplasty: a time- related phenomenon: a quantitative angiographic study in 342 consecutive patients at 1, 2, 3, and 4 months. Circulation. 1988; 77: 361–371.
- ↵Rogers C, Edelman ER. Endovascular stent design dictates experimental restenosis and thrombosis. Circulation. 1995; 91: 2995–3001.