Directional atherectomy before stenting versus stenting alone in percutaneous coronary interventions: A meta-analysis

International Journal of Cardiology 112 (2006) 178 – 183 www.elsevier.com/locate/ijcard

Directional atherectomy before stenting versus stenting alone in percutaneous coronary interventions: A meta-analysis Giampaolo Niccoli *, Luca Testa, Rocco Mongiardo, Annalisa Ricco, Flavia Belloni, Enrico Romagnoli, Antonio Maria Leone, Francesco Burzotta, Carlo Trani, Mario A. Mazzari, Antonio G. Rebuzzi, Filippo Crea Institute of Cardiology, Catholic University, Largo F. Vito 1-00168 Rome, Italy Received 6 June 2005; received in revised form 11 August 2005; accepted 28 August 2005 Available online 21 November 2005

Abstract Plaque debulking before stenting is still controversial. We performed a meta-analysis of 12 randomized and non-randomized trials comparing directional coronary atherectomy (DCA) before stenting versus stenting alone. Angiographic end points were acute gain, late loss and angiographic restenosis rate. Clinical end points were early major adverse cardiac events [MACEs: death, Q-wave myocardial infarction (MI), non-Q-wave MI], late MACEs (death, Q-wave MI) and target lesion revascularization (TLR). Data are expressed as odds ratio (OR) with 95% confidence intervals (CI) or weighted mean difference (WMD) with 95% CI, as appropriate. A total of 1216 patients undergoing DCA before stent and 1484 patients undergoing stent alone have been included. DCA before stent was associated to a better acute gain compared to stenting alone (WMD 0.23, [0.18 – 0.28]; p < 0.0001), to a striking reduction of angiographic restenosis rate (OR of 0.67, [0.54 – 0.84], p = 0.0003) and to a significantly lower rate of late TLR (OR 0.73 [0.59 – 0.91], p = 0.006). Late loss did not differ between the two groups (WMD 0.00 [ 0.08 and 0.08], p = 0.98). We found a higher rate of early MACEs for the combined approach (OR 1.87 [1.16 – 3.02], p = 0.01), with similar prevalence of late MACEs (OR 0.83 [0.65 – 1.06], p = 0.13). In conclusion, this metaanalysis demonstrates that DCA before stenting is superior to stenting alone with regard to acute angiographic results and TLR with a similar prevalence of late MACEs. The higher prevalence of early MACEs with DCA before stenting, however, is disturbing and probably related to distal embolization. D 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Directional coronary atherectomy; Stenting; Meta-analysis

1. Introduction Plaque debulking before stenting has been considered a promising strategy in the management of complex coronary lesions which are usually associated with high restenosis rate. When directional coronary atherectomy (DCA) alone is performed, restenosis seems to be mainly due to negative remodelling which is known to be prevented by stent implantation [1]. When stenting alone is performed residual plaque burden has been demonstrated to be associated to

* Corresponding author. Tel.: +39 06 35402795; fax: +39 06 3055535. E-mail address: [email protected] (G. Niccoli). 0167-5273/$ - see front matter D 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ijcard.2005.08.036

neointimal proliferation [2] and plaque shift may impair flow in side branches. Observational studies of DCA associated to stent were encouraging with 6-month restenosis rate ranging from 6.8% to 11% [3]. Yet, a recent large randomized trial (the AMIGO trial) [4] failed to demonstrate the superiority of the combined approach compared to stenting alone with regard to restenosis rate. However, in this trial restenosis rate in the stenting alone group was unexpectedly low and the rate of a suboptimal angiographic result post DCA was unexpectedly high. We performed a meta-analysis of randomized and nonrandomized trials comparing debulking before stenting versus stenting alone in order to better define the possible role of DCA in percutaneous coronary interventions.

G. Niccoli et al. / International Journal of Cardiology 112 (2006) 178 – 183

2. Methods 2.1. Search strategy Studies published within July 2005 comparing DCA before stenting versus stenting alone were identified by Medline entering the keywords ‘‘directional coronary atherectomy or debulking and stenting’’. Furthermore the Transcatheter Cardiovascular Therapeutics web site was extensively searched for this topic in the device directory. Topol’s interventional cardiology textbook chapter on directional coronary atherectomy was also analysed [5]. To avoid publication bias we also included unpublished studies reported between 1998 and July 2005 at Scientific sessions of the American Heart Association, Transcatheter Cardiovascular Therapeutics and American College of Cardiology. No restriction was applied with regard to the language. 2.2. Identified studies Twelve studies comparing DCA before stenting versus stenting alone were identified and were all included in the present analysis [4,6 –16]. General characteristics are listed in Table 1. Patients with a recent acute myocardial infarction or with an ejection fraction < 35% were excluded from all these studies. Angiographic exclusion criteria from all the studies were: unprotected left main (with the exception of Park et al.’s study [7]), severe vessel tortuosity, presence of severe coronary calcifications or of angiographically detectable thrombus in the target lesion, total occlusion. The use of another debulking device was an exclusion criterion except for one study [6]. 2.3. End points Angiographic end points were acute gain, final minimal lumen diameter, late loss and angiographic restenosis rate as previously defined [4]. Clinical end points were early major adverse cardiac events [MACEs: death, Q-wave myocardial

179

infarction (MI), non-Q-wave MI], late MACEs (death, Qwave MI) and target lesion revascularization (TLR). Number of events was obtained from raw data or derived from rates given in tables and text. All events were calculated from the total population of patients given at each time point during follow-up. 2.4. Statistical analysis Data tables were constructed in duplicate from primary sources. For dichotomous variables, data are expressed as odds ratio (OR) and 95% confidence interval (95% CI) according to a fixed effect model which yields narrower CI in the presence of heterogeneity between studies with respect to the random effect model. For continuous variables, data are expressed as fixed effect weighted mean difference (WMD) with 95% CI, which has the advantage of summarising results in the same unit of the variable. The extent of the heterogeneity was tested by the Mantel Haenszel method. Analyses have been performed using the Revman 4.2.2 freeware package program.

3. Results 3.1. Patient characteristics A total of 1216 patients undergoing DCA before stent (age 59 T 9.9) and 1484 patients undergoing stent alone (age 59 T 11.2) have been included in these studies. The mean follow-up was 12.4 months. Most studies included both stable and unstable angina patients. The prevalence of diabetic patients ranged from 3% to 37% (Table 1). All included studies reported QCA data for the initial procedure. In one study only QCA was not performed at follow-up [6]. 3.2. Procedural characteristics All procedures were performed through the femoral approach using 10 or 8 French compatible atherectomy

Table 1 General characteristics of included studies Study

Year

Study type

Target lesion

N of patients

N of lesions

Diabetes (%)

Mean FU (months)

Moussa et al. [11] Kobayashi et al. [12] Bramucci et al. [13] Karvouni et al. [10] Muramatsu et al. [8] Ahmed et al. [6] Park et al. [7] Airoldi et al. [9] Kawamura et al. [15] DESIRE [14] AMIGO [4] Kim et al. [16]

1998 1998 1998 2000 2000 2000 2001 2003 2004 2001 2004 2004

C-C C-C C-C C-C C-C C-C C-C C-C C-C CRT CRT CRT

Ostial and bifurcation LAD with >3 mm diameter Proximal non-tortuous artery Bifurcation Ostial Ostial vein graft Unprotected left main Ostial LAD Broad inclusion subset Broad inclusion subset Broad inclusion subset Ostial LAD

145 200 194 62 74 320 127 117 133 501 753 86

150 200 188 62 74 340 127 117 133 501 753 86

13 13 15 3 NA 32 16 16 30 NA 16 18

18 NA 14 NA 5.8 12 25.5 5.9 6.1 6 12 18.6

C-C = case control study, CRT = controlled randomized trial, FU = follow-up, LAD = left anterior descending, NA= not available, N = number.

180

G. Niccoli et al. / International Journal of Cardiology 112 (2006) 178 – 183

Table 2 Angiographic data from single studies Acute gain (S.D.)

Moussa et al. [11] Kobayashi et al. [12] Bramucci et al. [13] Karvouni et al. [10] Muramatsu et al. [8] Ahmed et al. [6] Park et al. [7] Airoldi et al. [9] Kawamura et al. [15] DESIRE [14] AMIGO [4] Kim et al. [16]

Late loss (S.D.)

Restenosis rate (%)

Stent alone

DCA + stent

Stent alone

DCA + stent

Stent alone

DCA + stent

2.3 2.27 2.22 1.98 1.6 2.02 2.9 2.13 2.15 1.87 2.06 2.5

2.98 2.84 2.43 2.3 2.2 2.42 3.1 2.4 2.61 2.1 2.1 2.8

1.02 (0.75) NA 1.04 (0.71) 1.19 (0.85) 1.2 (0.9) NA NA 1.05 (0.84) 0.7 (0.83) NA 0.69 (0.66) 1.5 (0.7)

0.91 (0.94) NA 0.82 (0.67) 1.03 (0.79) 1.2 (1.1) NA NA 0.8 (0.63) 0.93 (0.93) NA 0.79 (0.76) 1.7 (0.9)

21 23 31 43 43 NA 23 28 18.8 16 22 28

11 6 6 28 20 NA 8 11 12.5 15 26 37

(0.38) (0.46) (0.46) (0.36) (0.65) (0.68) (0.55) (0.5) (0.87) (0.41) (0.42) (0.7)

(0.52) (0.5) (0.41) (0.5) (0.65) (0.68) (0.55) (0.5) (0.78) (0.4) (0.47) (0.5)

NA= not available.

devices. All non-randomized studies included high risk, complex lesions: one study included unprotected left main [7], two included ostial lesions [8,9], one included bifurcation lesions [10], one included both ostial and bifurcation lesions [11], one included lesions in the proximal left anterior descending [12], one study included saphenous vein graft (SVG) only [6], two included an high proportion of B2/C type lesions [13,15]. Two randomized studies included a high percentage of favourable lesions [4,14].

At a mean follow-up of 12.4 months the combined approach compared to stent alone was associated to a striking reduction of angiographic restenosis rate (OR of 0.67, [0.54 – 0.83], p = 0.0003) (Fig. 2). Late loss did not differ between the two groups (WMD 0.00, [ 0.08 and 0.08], p = 0.98). A significant heterogeneity was found for all the end points described above ( p < 0.001 for acute gain and angiographic restenosis rate and p = 0.03 for late loss). 3.4. Clinical end points

3.3. Angiographic end points Angiographic data from single studies are summarised in Table 2. The combined approach of DCA before stent compared to stent alone was associated to a better acute gain (WMD 0.23, [0.18 – 0.28]; p < 0.0001) (Fig. 1) and to a bigger final MLD (WMD 0.23, [0.19 – 0.27]; p < 0.0001).

The combined approach of DCA before stent was associated to a significantly lower rate of late TLR (OR 0.73, [0.59 – 0.91], p = 0.006) (Fig. 3). The combined approach compared to stent alone was associated to a higher rate of early MACEs (OR 1.87, [1.16 – 3.02], p = 0.01) (Fig. 4), mainly due to higher rate of early MI

Fig. 1. Comparison of acute gain between DCA before stenting versus stenting alone. Data are expressed as weighted mean difference and 95% CI. The number of patients for each group is shown near to the study.

G. Niccoli et al. / International Journal of Cardiology 112 (2006) 178 – 183

181

Fig. 2. Comparison of angiographic restenosis between DCA before stenting versus stenting alone. Data are expressed as OR and 95% CI. The number of patients for each group is shown near to the study.

(OR 1.61 [1.09 –2.37], p = 0.02), with similar rate of late MACEs (OR 0.83, [0.65 –1.06]; p = 0.13). A significant heterogeneity was observed for TLR ( p = 0.04) but not for early or late MACEs.

4. Discussion In this meta-analysis, DCA before stenting compared to stenting alone is associated to a greater acute gain and to a lower angiographic restenosis and TLR rates. A disturbing finding is the higher rate of early MACEs in patients treated by DCA before stenting.

Three possible mechanisms may explain the lower angiographic restenosis rate and lower TLR in patients treated by DCA before stent. First, previous studies have demonstrated that a major determinant of restenosis is stent size. In a recent study by Kawamura et al. a strategy of DCA and stenting compared to stenting alone allowed selection of larger stent and better stent expansion at lower pressure than a strategy of stenting alone, resulting in a better acute gain at intravascular ultrasound (IVUS) [15]. Second, Prati et al. demonstrated that residual plaque burden after stenting measured by IVUS is a major determinant of restenosis; thus, DCA might reduce post stenting plaque burden by mechanically cutting plaque material [2]. Third, the rounded

Fig. 3. Comparison of TLR between DCA before stenting versus stenting alone. Data are expressed as OR and 95% CI. The number of patients for each group is shown near to the study.

182

G. Niccoli et al. / International Journal of Cardiology 112 (2006) 178 – 183

Fig. 4. Comparison of early MACEs (death, Q wave MI, non-Q wave MI) between DCA before stenting versus stenting alone. Data are expressed as OR and 95% CI. The number of patients for each group is shown near to the study.

cross section obtained after atherectomy has been associated with lower neointimal proliferation [17]. In our metaanalysis patients treated by DCA plus stenting exhibited a higher acute gain than patients treated by stent alone, thus suggesting that the utilization of larger stents played an important role in determining the lower restenosis or TLR rate, in spite of a similar late loss. Our study demonstrates a higher rate of early MACEs with a combined strategy than with stenting alone. Notably, the higher rate of early MACEs for the combined approach is mainly due to higher periprocedural MI. This may be due to various factors including distal thromboembolism, coronary spasm, hypotension and prolonged ischemia during atherectomy device positioning [18]. Furthermore, Dehemer et al. demonstrated that DCA causes more platelet aggregation than balloon angioplasty [19]. In a large cohort of patients Stone et al. demonstrated that myonecrosis after PCI is common in high risk patients after atheroablation and stenting [20]. Strategies to avoid early MI with DCA may include potent antiplatelet agents such the IIb – IIIa blockers or mechanical prevention of distal thromboembolism with distal protection. Interestingly, in the EPIC trial abciximab compared to placebo was associated to lower rate of periprocedural complications among unstable patients treated by DCA [21]. Conversely, distal filter protection is associated to reduced periprocedural MI in the setting of saphenous vein graft interventions, which are known to be associated with distal thromboembolism [22]. 4.1. Study limitation This study is limited by a striking discrepancy of the results among studies. Indeed, case control studies suggest a superiority of the combined strategy, while randomized

studies fail to show significant differences between the two strategies. The different results between case control studies and randomized trials may have several explanations. Firstly, expertise in performing DCA may be different in different trials. Centres involved in case control studies probably had higher volume of DCA procedures than centres involved in randomized trials. Secondly, lesion selection was different in case control and randomized studies. Indeed the former included high risk lesions, while the latter included predominantly non-complex lesions. Accordingly, restenosis rate in the stent arm of the AMIGO trial was unexpectedly low, thus suggesting selection of more favourable lesion. The advantages of DCA are probably greater when the risk of plaque shift at bifurcation lesions or ostial lesions is high or when large plaque burden could affect stent selection and apposition.

5. Conclusions In conclusion, our study demonstrates that a strategy of DCA and stenting is superior to stenting alone with regard to acute angiographic results, and TLR, especially when facing with complex lesion subsets, probably due to the possibility of employing bigger stents. The introduction of drug eluting stents might cast some shadows on the utilization of debulking devices. Yet, the SIRIUS [23] trial failed to confirm the zero restenosis rate observed in RAVEL [24] trial due to the inclusion of more complex lesions. Furthermore, bifurcation stenting with DES is still associated to a rather high side branch restenosis rate, which was 22% in the study by Colombo et al. [25]. Thus, a possible advantage of a combined strategy of DCA and stenting for the treatment of more complex lesions

G. Niccoli et al. / International Journal of Cardiology 112 (2006) 178 – 183

suggested by our meta-analysis might well be confirmed in the drug eluting stent era. Notably, multivariate analysis of the TAXUS study demonstrated that final acute gain remains a powerful predictor of restenosis [26].

References [1] Meine TJ, Bauman RP, Yock PG, Rembert JC, Greenfield Jr JC. Coronary artery restenosis after atherectomy is primarily due to negative remodeling. Am J Cardiol 1999;84(2):141 – 6. [2] Prati F, Di Mario C, Moussa I, et al. In-stent neointimal proliferation correlates with the amount of residual plaque burden outside the stent: an intravascular ultrasound study. Circulation 1999;99(8):1011 – 4. [3] Bersin RM, Simonton CA. Rotational and directional coronary atherectomy. Cathet Cardiovasc Interv 2003;58:485 – 99. [4] Stankovic G, Colombo A, Bersin R, et alAMIGO Investigators. Comparison of directional coronary atherectomy and stenting versus stenting alone for the treatment of de novo and restenotic coronary artery narrowing. Am J Cardiol 2004;93(8):953 – 8. [5] Safian RD. Coronary atherectomy: directional and extraction techniques. In: Topol EJ, editor. Textbook of interventional cardiology, Fouth edition . Saunders. p. 523 – 48 [Charter 25]. [6] Ahmed JM, Hong MK, Mehran R, et al. Comparison of debulking followed by stenting versus stenting alone for saphenous vein graft aortoostial lesions: immediate and one year outcomes. J Am Coll Cardiol 2000;35:1560 – 8. [7] Park SJ, Hong MK, Lee CW, et al. Elective stenting of unprotected left main coronary artery stenosis: effect of debulking before stenting and intravascular ultrasound guidance. J Am Coll Cardiol 2001;38:1054 – 60. [8] Muramatsu T, Tsukahara R, Ho M, et al. Efficacy of directional coronary atherectomy before stent implantation for coronary ostial lesions. J Invasive Cardiol 2000;12(9):440 – 5. [9] Airoldi F, Di Mario C, Stankovic G, et al. Clinical and angiographic outcome of directional atherectomy followed by stent implantation in de novo lesions located at the ostium of the left anterior descending coronary artery. Heart 2003;89(9):1050 – 4. [10] Karvouni E, Di Mario C, Nishida T, et al. Directional atherectomy prior to stenting in bifurcation lesions: a matched comparison study with stenting alone. Cathet Cardiovasc Interv 2001;53(1):12 – 20. [11] Moussa I, Moses J, Di Mario C, et al. Stenting after optimal lesion debulking (SOLD) registry. Angiographic and clinical outcome. Circulation 1998;98(16):1604 – 9. [12] Kobayashi Y, Moussa I, Akiyama T, et al. Low restenosis rate in lesions of the left anterior descending coronary artery with stenting following directional coronary atherectomy. Cathet Cardiovasc Diagn 1998;45(2):131 – 8. [13] Bramucci E, Angoli L, Merlini PA, et al. Adjunctive stent implantation following directional coronary atherectomy in patients with coronary artery disease. J Am Coll Cardiol 1998;32(7):1855 – 60.

183

[14] Aizawa T, Tamai H, Asakura Y, et al. Clinical and angiographic results of the debulking and stenting in restenosis elimination (DESIRE) trial. Circulation 2001;104:2954. [15] Kawamura A, Asakura Y, Ishikawa S, et al. Stenting after directional coronary atherectomy compared with directional coronary atherectomy alone and stenting alone: a serial intravascular ultrasound study. Circ J 2004;68(5):455 – 61. [16] Kim YH, Hong MK, Lee SW, et al. Randomized comparison of debulking followed by stenting versus stenting alone for ostial left anterior descending artery stenosis: intravascular ultrasound guidance. Am Heart J 2004;148:663 – 9. [17] Murata T, Hiro T, Fujii T, et al. Impact of cross-sectional geometry of the post-deployment coronary stent on in-stent neointimal hyperplasia: an intravascular ultrasound study. Circ J 2002;66:489493. [18] Carrozza J, Baim J. Complications of directional coronary atherectomy: incidences, causes and management. Am J Cardiol 1993;72:47E – 54E. [19] Dehmer GJ, Nichols TC, Bode AP, et al. Assessment of platelet activation by coronary sinus blood sampling during balloon angioplasty and directional coronary atherectomy. Am J Cardiol 1997;80(7):871 – 7. [20] Stone GW, Mehran R, Dangas G, Lansky AJ, Kornowski R, Leon MB. Differential impact on survival of electrocardiographic Q-wave versus enzymatic myocardial infarction after percutaneous intervention: a device-specific analysis of 7147 patients. Circulation 2001;104(6):642 – 7. [21] Ghaffari S, Kereiakes DJ, Lincoff AM, et al. Platelet glycoprotein IIb/IIIa receptor blockade with abciximab reduces ischemic complications in patients undergoing directional coronary atherectomy. EPILOG Investigators. Evaluation of PTCA to improve long-term outcome by c7E3 GP IIb/IIIa receptor blockade. Am J Cardiol 1998 (Jul 1);82(1):7 – 12. [22] Stone GW, Rogers C, Ramee S, et al. Distal filter protection during saphenous vein graft stenting: technical and clinical correlates of efficacy. J Am Coll Cardiol 2002 (Nov 20);40(10):1882 – 8. [23] Moses JW, Leon MB, Pompa JJ. Sirolimus eluting stents versus standard stents in patients with stenosis in a native coronary artery. N Engl J Med 2003;349:1315 – 23. [24] Morice MC, Serruys P, Sousa JE. A randomized comparison of a sirolimus eluting stent with a standard stent for coronary revascularization. N Engl J Med 2002;346:1773 – 80. [25] Colombo A, Moses JW, Morice MC, et al. Randomized study to evaluate sirolimus-eluting stents implanted at coronary bifurcation lesions. Circulation 2004 (Mar 16);109(10):1244 – 9. [26] Stone GW, Ellis SG, Cox DA, et al. TAXUS-IV Investigators. A polymer-based, paclitaxel-eluting stent in patients with coronary artery disease. N Engl J Med 2004;350:221 – 31.