Minimally Invasive Direct Coronary Artery Bypass Graft Surgery or Percutaneous Coronary Intervention for Proximal Left Anterior Descending Artery Stenosis: A Meta-Analysis Salil V. Deo, MS, MCh, Vikas Sharma, MS, MCh, Ishan K. Shah, MBBS, Patricia J. Erwin, BLS, Lyle D. Joyce, MD, and Soon J. Park, MD Division of Cardiovascular Surgery, Mayo Clinic Rochester, Rochester, Minnesota; Department of Cardiovascular Surgery, Wockhardt Adventist Heart Institute, Surat Gujarat, India; Mayo Clinic Libraries, Mayo Clinic Rochester, Rochester, Minnesota; Division of Surgery, University of Minnesota, Minneapolis–St. Paul, Minnesota; and Division of Cardiovascular Surgery, University Hospitals, Case Western Reserve University, Cleveland, Ohio
Background. We conducted a metaanalysis comparing early and midterm cardiovascular adverse events associated with minimally invasive direct coronary artery bypass graft surgery (MIDCABG) and percutaneous coronary intervention (PCI), with a focus on drug-eluting stents (DES). Methods. A systematic literature review (MEDLINE, EMBASE, Scopus, and so forth) yielded 12 studies (7 randomized controlled trials; 5 observational) pooling more than 2,000 patients. A random effect, inverse variance metaanalysis was conducted, and a subgroup analysis of the PCI-DES cohort was performed. Events were compared as risk ratios using a 95% confidence interval (CI). Heterogeneity of results was evaluated by Eggers I2 test. Results are presented as early (0 to 1 year) and midterm (2 to 5 years). Results. Midterm mortality in the PCI and MIDCABG groups (3.6% and 2.6%, respectively) was comparable (1.24, 95% CI: 0.66 to 2.33; p [ 0.5; I2 [ 0%). Risk of early restenosis was lower in the MIDCABG cohort compared
with PCI (0.40, 95% CI: 0.16 to 0.99; p [ 0.05; I2 [ 57%). Although the early risk of recurrence of angina was comparable, over time it was 61% (43% to 74%) lower for MIDCABG patients (p < 0.001). Midterm results on analysis of the entire cohort demonstrated an increased risk for target vessel reinterventions (3.84, 95% CI: 2.7 to 5.5; p < 0.001) in the PCI cohort. Subgroup analysis revealed that the PCI-DES cohort (4 studies; 456 patients) had a higher risk of recurrent angina (risk ratio 3.4, 95% CI: 1.9 to 6.2; p < 0.001; I2 [ 0%) and target vessel reinterventions (risk ratio 4.16, 95% CI: 2.7 to 6.6; p < 0.001; I2 [ 0%) at midterm follow-up (2 to 5 years). Conclusions. Survival rates are comparable after either MIDCABG or PCI for proximal LAD disease. However, even the use of DES was associated with significantly higher rates of angina recurrence and the need for target vessel reintervention as compared with MIDCABG.
T
stents (DES). Therefore, we conducted a metaanalysis comparing clinical events associated with PCI and MIDCABG in the treatment of p-LAD disease.
herapeutic intervention for coronary artery disease consists of percutaneous intracoronary stenting (PCI) or coronary artery bypass graft surgery (CABG). Compared with surgery, PCI is a less invasive procedure associated with lower morbidity. Minimally invasive direct coronary artery bypass (MIDCABG) is a recent introduction, where the grafting is performed through a small thoracotomy incision. It is associated with a shorter hospital stay and earlier recovery than conventional surgery. Results comparing MIDCABG and PCI in proximal left anterior descending artery (p-LAD) disease are conflicting, especially with the recent addition of drug-eluting
(Ann Thorac Surg 2014;-:-–-) Ó 2014 by The Society of Thoracic Surgeons
Material and Methods Inclusion Criteria
Accepted for publication Jan 28, 2014.
Original articles (randomized controlled trials and observational studies) comparing PCI and MIDCABG for p-LAD disease (from 2000 to May 2013) were identified. The search was limited to (1) human subjects, (2) original articles, and (3) English language. Only studies reporting clinically relevant endpoints (eg, mortality, recurrence of angina, or reinterventions during follow-up) were included. Editorials and review articles were excluded.
Presented at the Poster Session of the Fiftieth Annual Meeting of The Society of Thoracic Surgeons, Orlando, FL, Jan 25–29, 2014.
Technique of MIDCABG
Address correspondence to Dr Deo, Adventist Wockhardt Heart Institute, Surat Gujarat 395001, India; e-mail: [email protected].
MIDCABG consisted of harvest of the left internal thoracic artery (LITA) and its subsequent grafting to the
Ó 2014 by The Society of Thoracic Surgeons Published by Elsevier Inc
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2014.01.086
2
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
Abbreviations and Acronyms BMS CI DES LAD LITA MIDCABG
= = = = = =
p-LAD
=
RCT RR TVR
= = =
bare metal stent confidence interval drug-eluting stent left anterior descending artery left internal thoracic artery minimally invasive direct coronary artery bypass graft surgery proximal left anterior descending artery randomized controlled trial risk ratio target vessel reintervention
LAD. That was most commonly accomplished through a left thoracotomy, in some cases, robotically assisted. Sequential grafting of the LITA was occasionally performed; however, only the LAD and its branches was grafted in all included studies.
Percutaneous Intervention Percutaneous coronary intervention was performed in the routine manner. Studies used a relatively uniform strategy of stent deployment with at least 90% patients having undergone either bare-metal stent (BMS) or DES insertion. A separate subgroup analysis of articles reporting outcome with DES was also conducted.
Studied Endpoints Clinical endpoints compared were duration of hospital stay, mortality, recurrence of angina, myocardial infarction, and need for reintervention. All results were assimilated into two periods: early (0 to 1 year) and midterm (2 to 5 years).
Search Strategy We conducted a systematic search of articles on EMBASE, Scopus, Web of Science, and MEDLINE using the terms “single vessel disease,” “left anterior descending artery,” “percutaneous intervention,” “drug-eluting stent,” “coronary artery bypass,” “minimally invasive coronary artery bypass,” and “MIDCABG” in various combinations. Two of the authors (V.S., S.V.D) independently screened abstracts and retrieved the full text articles. All full text articles were then evaluated independently to conform to the inclusion criteria. Discrepancies were resolved by consensus. Data were obtained from the studies using a prespecified data abstraction form.
Statistical Analysis Statistical analysis for binary data was performed using the “meta” package for R 3.0.1 [1, 2]. The risk ratio (RR) was implemented as the effect estimate. In the inverse weighted model, each study contributes a percentage of the final pooled estimate [3]. That is presented in each forest plot under the column of weight (W). As per Bate’s correction, 0.5 was added to each cell of the two-
Ann Thorac Surg 2014;-:-–-
by-two table in case the study or control arm had zero events [4]. The cutoff for the p value is 0.05; data are presented with 95% confidence intervals (CI). To compare the hospital stay, studies were pooled using the standardized mean difference as the effect estimate. Results are presented as a forest plot, depicting each individual risk ratio (RR) as well as the overall composite effect estimate. A risk ratio with its 95% CI less than 1 would favor MIDCABG. A random effects model has been implemented as both randomized controlled trial (RCT) and retrospective studies have been pooled [3]. Heterogeneity is defined as the variation among studies that would contribute to the overall results. It was evaluated using the Eggers I2 test and further stratified into low (25% to 49%), moderate (50% to 74%), and high (>75%) heterogeneity [3]. A subgroup analysis was performed after excluding the observational studies and compared to the overall effect. A separate subgroup analysis was conducted of studies implementing DES as their technique of percutaneous intervention. Publication bias was excluded by a visual inspection of the contour-enhanced funnel plot symmetry and determined using the Begg-Mazumdar rank correlation test. The results are presented as per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines for metaanalysis of observational studies [5, 6].
Results The initial search criteria yielded 312 articles. Exclusion of duplicates and further refinement of Medical Subject Headings yielded 193 abstracts, which were evaluated for inclusion. The detailed PRISMA flow diagram is presented in Figure 1. Two articles were excluded to avoid patient duplication. A total of 14 articles [7–19] discussing the results of 12 studies (seven RCTs and five observational studies) fulfilled the selection criteria. Two articles presented results of the same randomized controlled trial at a follow-up period of 6 months and 5 years [9, 10]. Etienne and colleagues [7, 8] have presented two articles discussing a retrospective analysis of data from two institutions. Drenth and associates [16, 20] have also presented their data initially in 2002 and then results of a longer follow-up in 2004. Blazek and colleagues [21] presented data regarding results of BMS at the end of 10 years of follow-up. As other articles failed to present data with such a long follow-up period, it was excluded from the metaanalysis. Herz and coworkers [22] was excluded because the study included patients with triple vessel disease. Care was taken to ensure nonduplication of data during analysis. The detailed patient demographics are presented in Tables 1 and 2. As the majority of patients were part of RCTs, patient population between cohorts is relatively well matched. In four articles [10, 11, 15, 16], angiographic evaluation was performed at the end of the follow-up period. Drug-eluting stents were implemented as the interventional strategy in five studies (463 patients). Sirolimus- or paclitaxel-eluting stents were used for
Ann Thorac Surg 2014;-:-–-
revascularization in all studies [7, 8, 12, 14, 15]. Antiplatelet agent therapy before and after stent implantation was as per institutional protocol. All the RCTs naturally had well-matched baseline data. Ben-Gal and colleagues [15] have utilized MIDCABG for patients with occluded LAD vessels, as have others. BenGal and colleagues [15] have also incorporated 48% patients with triple vessel disease who underwent only LAD intervention for various reasons. Shirai and coworkers [19] have retrospectively compared the unpublished results of the Patency, Outcomes and Economics of MIDCABG trial with demographically matched patients from their institutional database who underwent PCI; the investigators admit that angiographic data regarding the Patency, Outcomes and Economics of MIDCABG trial cohort was not available while selecting matched patients. The MIDCABG was performed through a small left thoracotomy in nine studies. In these patients, the LITA was harvested either with the help of thoracoscopic instruments or under direct vision. Two investigators [17, 19] have not explicitly presented the method of MIDCABG, whereas ministernotomy with direct LITA harvest was the preferred technique for Kim and associates [13]. Surgery was performed in an off-pump manner without cardiopulmonary bypass. The use of antiplatelet agents after surgery varied per institutional protocol. Percutaneous intervention was performed in the conventional manner. In the five studies implementing DES,
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
3
either paclitaxel-eluting or sirolimus-eluting stents were used. The use of antiplatelet agents varied between studies.
Length of Hospital Stay A pooled analysis from five studies (1,069 patients) demonstrated that hospital stay was significantly shorter after PCI as compared with MIDCABG (p ¼ 0.001; Fig 2). Two articles [12, 18] presented data as median (range); hence, they were not pooled in this analysis; however, both reported a reduced duration of stay in the PCI cohort.
Early Angiographic Patency Early postprocedure angiographic evaluation was performed in four studies. MIDCABG has a slight advantage over PCI with regard to the risk of restenosis (RR 0.44, 95% CI: 0.16 to 0.99; I2 ¼ 57.3%; p ¼ 0.046). In this analysis, only Thiele and associates [12] incorporated DES, whereas all others implemented BMS.
Mortality Early mortality was presented in 10 studies (993 MIDCABG patients and 1589 PCI patients; Fig 3). Mortality was comparable at 1.3% and 1.7% in the MIDCABG and PCI cohorts, respectively (RR 1.17, 95% CI: 0.53 to 2.57; p ¼ 0.7; I2 ¼ 0%; Fig 1). Midterm mortality was presented in six studies; incidence of mortality (3.6% MICABG; Fig 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart demonstrating procedure of the systematic review process. Reproduced from Moher et al [5].
4
Variable
Cisowski et al 2002 Drenth et al 2002a
Country Poland Netherlands Study design RCT RCT Duration 12 months 48 months Cohort BMS MIDCABG BMS MIDCABG No. of pts. 50 50 51 51 Age, years 53 10 54 9.1 61 1.3 60 1.6 Men 84 82 75 78 DM 8 6 18 24 HTN 52 56 33 16 Smokers 52 48 30 37 Prior AMI . . 18 24 Unstable angina 10 8 . . EF . . . . MIDCABG technique Thoracotomy Thoracotomy
Iakovou et al 2002 United States Prospective OBS 39 months BMS MIDCABG 441 119 63 12 62 12 68 71 22 17 54 55 46 56 22 22 68 60 52 12 48 7 Exact method not specified
a Drenth, 2004, presented updated 4-year results of the same randomized controlled trial. (range).
b
Diegler et al 2002b
Reeves et al 2004
Germany United Kingdom RCT RCT 6 months 12 months BMS MIDCABG BMS MIDCABG 110 110 50 50 62 10.2 61 10 54 (49–61)c 58 (53–67)c 72 77 86 70 34 25 . . 72 71 . . 25 25 . . 45 45 . . . . . . 62 15 63 11 . . Thoracotomy Thoracotomy
Shirai et al 2004
Kim et al 2005
United States Prospective OBS 92 months BMS MIDCABG 429 152 63 11 61 12 65 73 24 26 53 59 19 15 43 37 . . 51 11 53 11 Exact method not specified
South Korea RCT 24 months BMS MIDCABG 50 50 61 12 63 12 60 70 20 15 55 55 45 55 22 22 65 55 51 11 49 13 Ministernotomy
Thiele, 2005, presented updated 5-year results of the same randomized controlled trial.
c
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
Table 1. Clinical Characteristics of Studies Implementing Bare Metal Stents for Percutaneous Coronary Intervention
Age is median
Values are percent, unless otherwise indicated. AMI ¼ acute myocardial infarction; DM ¼ diabetes mellitus; observational; PCI ¼ percutaneous coronary intervention;
EF ¼ ejection fraction; HTN ¼ hypertension; MIDCABG ¼ minimally invasive coronary artery bypass graft surgery; pts. ¼ patients; RCT ¼ randomized controlled trial.
OBS ¼
Ann Thorac Surg 2014;-:-–-
Ann Thorac Surg 2014;-:-–-
Table 2. Clinical Characteristics of Studies Implementing Drug-Eluting Stents for Percutaneous Intervention Variable
Ben Gal et al 2006
Thiele et al 2009
Etienne et al 2009
Etienne et al 2013
South Korea RCT 6 months DES MIDCABG 119 70 60 9 61 9 64 64 37 48 50 55 40 45 21 22 50 42 52 8 51 9 Thoracotomy Paclitaxel 20%, sirolimus 80%
Israel PM OBS 24 months DES MIDCABG 83 83 38.6%a 38.6%a 80 80 30 29 47 55 . . 31 42 . . 7.2%d 1.2%d Thoracotomy Sirolimus 100%
Germany RCT 12 months DES MIDCABG 65 65 66 (59–72)b 66 (59–71)b 69 71 28 25 83 85 14 18 23 23 . . 65 (60–66) 65 (60–70) Thoracotomy Sirolimus 100%
Belgium Retro OBS 24 months DES MIDCABG 154 154 62c 63c . . 62 23 71 55 43 56 27 16 24 18 2%d 3%d Thoracotomy Paclitaxel 59%, sirolimus 41%
Belgium Retro PM OBS 60 months DES MIDCABG 196 260 64 12 63 12 48 34 60 21 67 57 46 54 21 12 21 15 1%d 3%d Thoracotomy Paclitaxel 59%, sirolimus 41%
a b Ben Gal et al presented data as patient age greater than 70 years. Thiele et al presented data as median (interquartile range). Etienne et al presented percentage of patients with ejection fraction less than 30%.
c
Etienne et al presented data only as mean.
d
Ben Gal et al and
Values are percent, unless otherwise indicated. AMI ¼ acute myocardial infarction; DES ¼ drug-eluting stent; DM ¼ diabetes mellitus; EF ¼ ejection fraction; HTN ¼ hypertension; MIDCABG ¼ minimally invasive direct coronary artery bypass graft surgery; OBS ¼ observational; PCI ¼ percutaneous coronary intervention; PM ¼ propensity matched; RCT ¼ randomized controlled trial; Retro ¼ retrospective.
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
Country Study design Study duration Cohort No. of patients Age, years Men DM HTN Smokers Prior AMI Unstable angina EF MIDCABG technique Type of DES
Hong et al 2005
5
6
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
Ann Thorac Surg 2014;-:-–-
Fig 2. Forest plot pooling together the duration of hospital stay for both cohorts. The minimally invasive direct coronary artery bypass graft surgery (MIDCAB) group had a significantly longer stay compared with the percutaneous coronary intervention (PCI) cohort (p ¼ 0.001). (CI ¼ confidence interval; SMD ¼ standardized mean difference.)
2.6% PCI) was comparable in both groups (RR 1.24, 95% CI: 0.65 to 2.33; p ¼ 0.50). The funnel plot of pooled studies demonstrates absence of publication bias (p ¼ 0.216; Fig 4).
angina (p < 0.001; I2 ¼ 0%). The RR of recurrent angina was 2.56 (95% CI: 1.75 to 3.84) in the PCI cohort compared with the MIDCABG cohort.
Myocardial Infarction Recurrence of Angina Early recurrent angina was present in 16.5% and 20.2% patients in the MIDCABG and PCI cohorts, respectively (p ¼ 0.34; Fig 5). However, at the end of 2 to 5 years of follow-up, while 93.2% of the MIDCABG group were symptom free, only 79% of the PCI cohort were free of Fig 3. The forest plot shows the pooled results of mortality between the two cohorts, and demonstrates that mortality in the two groups is comparable (A) during early follow-up and (B) at midterm follow-up. (CI ¼ confidence interval; MIDCAB ¼ minimally invasive direct coronary artery bypass graft surgery; PCI ¼ percutaneous coronary intervention; RR ¼ risk ratio.)
The incidence of early myocardial infarction was comparable in both cohorts (2.3% and 0.6% in the MIDCABG and PCI cohorts, respectively; p ¼ 0.08; Fig 6). The risk of myocardial infarction (RR 0.78, 95% CI: 0.36 to 1.69) at midterm follow-up was similar in both the MIDCABG cohort and the PCI cohort (p ¼ 0.28).
Ann Thorac Surg 2014;-:-–-
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
7
to 6.67) times higher in the DES cohort (21.5%) as compared with the MIDCABG cohort (4.8%; p < 0.001; I2 ¼ 0%; Fig 8).
Comment
Fig 4. Contour-enhanced funnel plot demonstrating absence of publication bias among included studies. (Dark gray area ¼ 0.1 > p > 0.05; medium gray area ¼ 0.05 > p > 0.01; light gray area ¼ p > 0.01.)
Target Vessel Reintervention At the end of 1 year of follow-up, target vessel reintervention (TVR) was required in 4.3% and 17.9% of the MIDCABG cohort and PCI cohort, respectively (p < 0.001; I2 ¼ 0%; Fig 7). In the pooled midterm analysis, 5.2% of MIDCABG patients and 21.5% of PCI patients needed TVR (p < 0.001; I2 ¼ 0%). Thus, the risk of TVR was 3.84 (95% CI: 2.7 to 5.55) times higher in the PCI cohort compared with the MIDCABG cohort at midterm follow-up.
Subgroup Analysis in Studies Implementing DES Five articles pooling four studies comparing MIDCABG and DES were analyzed. Two articles reported recurrence of angina at midterm; it was and 6.8% and 23.5%, respectively, in the MIDCABG cohort and DES cohort (p < 0.001; I2 ¼ 0%). At midterm follow-up, the risk for TVR (after pooling three studies) was 4.16 (95% CI: 2.70
Guidelines regarding PCI and CABG in multivessel coronary artery disease are relatively well established. However, limited data are available comparing MIDCABG and PCI for p-LAD disease, especially in the present era of DES. We therefore conducted a pooled analysis of more than 2,000 patients comparing the clinical outcomes in these cohorts. Our metaanalysis and systematic review of 12 studies (seven RCTs) demonstrate that recurrence of angina and the need for a reintervention is much lower in the MIDCABG cohort as compared with the PCI cohort at the end of 2 to 5 years of follow-up. This important difference is maintained even when comparing MIDCABG versus DES. Percutaneous intervention, being a less invasive procedure than CABG, is an attractive option for patients, especially as DES are reported to have lower restenosis rates compared with BMS. MIDCABG was introduced as a minimally invasive alternative to conventional CABG. The surgical approach is through a left thoracotomy or, occasionally, a ministernotomy. After harvesting the LITA either with the use of thoracoscopic instruments or under direct vision, an off-pump LITA-LAD anastomosis is constructed. The smaller skin incision, avoidance of complete sternotomy, and the off-pump technique definitely reduce morbidity of this procedure compared with routine CABG [23]. Important factors determining success after PCI are vessel-related factors. Our pooled analysis contains seven RCT; although patients were primarily matched Fig 5. Forest plots demonstrating recurrence of angina in the (A) early period and (B) intermediate period. (CI ¼ confidence interval; MIDCAB ¼ minimally invasive direct coronary artery bypass graft surgery; PCI ¼ percutaneous coronary intervention; RR ¼ risk ratio.)
8
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
Ann Thorac Surg 2014;-:-–-
Fig 6. Forest plot demonstrates the individual and pooled incidence of myocardial infarction in the minimally invasive direct coronary artery bypass graft surgery (MIDCABG) and percutaneous coronary intervention (PCI) cohorts. The incidence of acute myocardial infarction (AMI) was comparable at (A) early follow-up and (B) and midterm follow-up. (CI ¼ confidence interval; RR ¼ risk ratio.)
according to demographics, evaluation of angiographic data demonstrates comparability. However, as reported by other retrospective studies, patients with calcified, occluded, or highly tortuous vessels were recommended for MIDCABG. Such procedure selection is important
Fig 7. Forest plot demonstrates the pooled need for reintervention in the (A) early period and (B) intermediate period. (CI ¼ confidence interval; MIDCAB ¼ minimally invasive direct coronary artery bypass graft surgery; PCI ¼ percutaneous coronary intervention; RR ¼ risk ratio.)
in the “real world” to ensure optimal results for our patients. An important aspect of MIDCABG is appropriate training. Diegeler and coworkers [24] have reported an increase in patency rate from 96% to 99% from 1997 to
Ann Thorac Surg 2014;-:-–-
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
9
Fig 8. The forest plot presents the need for target vessel reintervention (TVR [midterm follow-up]) stratified by the type of drug-eluting stent (DES) used. As demonstrated by the lower part of the figure (DES ¼ 1), the risk of reintervention was higher with DES when compared with minimally invasive direct coronary artery bypass graft surgery (MIDCABG). (CI ¼ confidence interval; PCI ¼ percutaneous coronary intervention; RR ¼ risk ratio.)
1999. A small, 3.4%, risk of conversion to sternotomy is also present. Cremer and associates [25] evaluated 205 MIDCABG patients with coronary angiography and reported patency rates similar to those for conventional CABG. However, they did report instances of large unclipped collateral branches of the ITA being missed, inadvertent grafting of the ITA to the diagonal branches, and ITA kinking due to excessive length. Such events are definitely more likely when operating on a moving target within a small access area. Therefore, although newer techniques have made the procedure more appealing to the patient, it is important to understand that each new instrument has its own learning curve. A study comparing patency rates of single vessel bypass by conventional and MIDCABG methods demonstrates the importance of appropriate training [26]. The pooled articles did not present comparison of economic data. A recent cost-effectiveness analysis [27] demonstrates that stenting was beneficial for the first 2 years. From the third year on, given the higher rates of reintervention and recurrence of anginal symptoms in the PCI cohort, it no longer was the preferred option. The investigators also separately analyzed the use of DES and reached the same conclusion. An earlier metaanalysis on this topic pooled five RCTs predominantly containing patients undergoing PCI with BMS [28]. Our results are based on a pooled analysis of more than 2,000 patients. Importantly, given the increase in available data from then to now, we were able to stratify results according to follow-up (early and midterm) as well as stent type (BMS and DES). Both factors are extremely important in reaching an unbiased conclusion.
Study Strengths and Limitations As with all metaanalyses, our results are dependent on reported study-level data. Although some within study variation was present among the MIDCABG and PCI cohorts, overall patient matching was excellent. Additionally, a large proportion of our patients were part of
various RCTs. A very low heterogeneity in our pooled results demonstrates minimal between-study variations. In conclusion, survival rates are comparable after either MIDCABG or PCI for p-LAD disease. However, even the use of DES was associated with significantly higher rates of angina recurrence and the need for TVR as compared with MIDCABG.
References 1. Schwarzer G. Package “meta.” Meta-analysis with R, version 2.5-1, 2013. 2. Core Team R. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2013. 3. Higgins JPT, Green S, editors. Cochrane handbook for systematic reviews of interventions, version 5.1.0 [updated March 2011]. The Cochrane Collaboration, 2011. 4. Friedrich JO, Adhikari NK, Beyene J. Inclusion of zero total event trials in meta-analyses maintains analytic consistency and incorporates all available data. BMC Med Res Methodol 2007;7:5. 5. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ 2009;339:b2535. 6. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994;50:1088–101. 7. Etienne PY, Glineur D, Papadatos S, et al. Comparison of minimally invasive direct coronary artery bypass surgery with implantation of drug-eluting stents in patients with left anterior descending coronary artery disease. Innovations (Phila) 2009;4:340–4. 8. Etienne PY, D’Hoore W, Papadatos S, et al. Five-year followup of drug-eluting stents implantation vs minimally invasive direct coronary artery bypass for left anterior descending artery disease: a propensity score analysis. Eur J Cardiothorac Surg 2013 March 14 [E-Pub ahead of print]. 9. Thiele H, Oettel S, Jacobs S, et al. Comparison of bare-metal stenting with minimally invasive bypass surgery for stenosis of the left anterior descending coronary artery: a 5-year follow-up. Circulation 2005;112:3445–50. 10. Diegeler A, Thiele H, Falk V, et al. Comparison of stenting with minimally invasive bypass surgery for stenosis of the left anterior descending coronary artery. N Engl J Med 2002;347:561–6. 11. Cisowski M, Drzewiecki J, Drzewiecka-Gerber A, et al. Primary stenting versus MIDCAB: preliminary report— comparison of two methods of revascularization in single left
10
12.
13. 14.
15.
16.
17.
18.
19.
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
anterior descending coronary artery stenosis. Ann Thorac Surg 2002;74(Suppl):1334–9. Thiele H, Neumann-Schniedewind P, Jacobs S, et al. Randomized comparison of minimally invasive direct coronary artery bypass surgery versus sirolimus-eluting stenting in isolated proximal left anterior descending coronary artery stenosis. J Am Coll Cardiol 2009;53:2324–31. Kim JW, Lim DS, Sun K, et al. Stenting or MIDCAB using ministernotomy for revascularization of proximal left anterior descending artery? Int J Cardiol 2005;99:437–41. Hong SJ, Lim DS, Seo HS, et al. Percutaneous coronary intervention with drug-eluting stent implantation vs. minimally invasive direct coronary artery bypass (MIDCAB) in patients with left anterior descending coronary artery stenosis. Catheter Cardiovasc Interv 2005;64:75–81. Ben-Gal Y, Mohr R, Braunstein R, et al. Revascularization of left anterior descending artery with drug-eluting stents: comparison with minimally invasive direct coronary artery bypass surgery. Ann Thorac Surg 2006;82:2067–71. Drenth DJ, Winter JB, Veeger NJ, et al. Minimally invasive coronary artery bypass grafting versus percutaneous transluminal coronary angioplasty with stenting in isolated high-grade stenosis of the proximal left anterior descending coronary artery: six months’ angiographic and clinical follow-up of a prospective randomized study. J Thorac Cardiovasc Surg 2002;124:130–5. Iakovou I, Dangas G, Mehran R, et al. Minimally invasive direct coronary artery bypass (MIDCAB) versus coronary artery stenting for elective revascularization of the left anterior descending artery. Am J Cardiol 2002;90:885–7. Reeves BC, Angelini GD, Bryan AJ, et al. A multi-centre randomised controlled trial of minimally invasive direct coronary bypass grafting versus percutaneous transluminal coronary angioplasty with stenting for proximal stenosis of the left anterior descending coronary artery. Health Technol Assess 2004;8:1–43. Shirai K, Lansky AJ, Mehran R, et al. Minimally invasive coronary artery bypass grafting versus stenting for patients with proximal left anterior descending coronary artery disease. Am J Cardiol 2004;93:959–62.
Ann Thorac Surg 2014;-:-–-
20. Drenth DJ, Veeger NJ, Middel B, et al. Comparison of late (four years) functional health status between percutaneous transluminal angioplasty intervention and off-pump left internal mammary artery bypass grafting for isolated highgrade narrowing of the proximal left anterior descending coronary artery. Am J Cardiol 2004;94:1414–7. 21. Blazek S, Holzhey D, Jungert C, et al. Comparison of baremetal stenting with minimally invasive bypass surgery for stenosis of the left anterior descending coronary artery: 10-year follow-up of a randomized trial. J Am Coll Cardiol Intv 2013;6:20–6. 22. Herz I, Moshkovitz Y, Hendler A, et al. Revascularization of left anterior descending artery with drug-eluting stents: comparison with off-pump surgery. Ann Thorac Surg 2005;79:88–92. 23. Holzhey DM, Jacobs S, Mochalski M, et al. Seven-year follow-up after minimally invasive direct coronary artery bypass: experience with more than 1300 patients. Ann Thorac Surg 2007;83:108–14. 24. Diegeler A, Spyrantis N, Matin M, et al. The revival of surgical treatment for isolated proximal high grade LAD lesions by minimally invasive coronary artery bypass grafting. Eur J Cardiothorac Surg 2000;17:501–4. 25. Cremer J, Mugge A, Wittwer T, et al. Early angiographic results after revascularization by minimally invasive direct coronary artery bypass (MIDCAB). Eur J Cardiothorac Surg 1999;15:383–8. 26. Vicol C, Nollert G, Mair H, et al. Midterm results of beating heart surgery in 1-vessel disease: minimally invasive direct coronary artery bypass versus off-pump coronary artery bypass with full sternotomy. Heart Surg Forum 2003;6:341–4. 27. Rao C, Aziz O, Panesar SS, et al. Cost effectiveness analysis of minimally invasive internal thoracic artery bypass versus percutaneous revascularisation for isolated lesions of the left anterior descending artery. BMJ 2007;334:621. 28. Jaffery Z, Kowalski M, Weaver WD, Khanal S. A meta-analysis of randomized control trials comparing minimally invasive direct coronary bypass grafting versus percutaneous coronary intervention for stenosis of the proximal left anterior descending artery. Eur J Cardiothorac Surg 2007;31:691–7.
Background. We conducted a metaanalysis comparing early and midterm cardiovascular adverse events associated with minimally invasive direct coronary artery bypass graft surgery (MIDCABG) and percutaneous coronary intervention (PCI), with a focus on drug-eluting stents (DES). Methods. A systematic literature review (MEDLINE, EMBASE, Scopus, and so forth) yielded 12 studies (7 randomized controlled trials; 5 observational) pooling more than 2,000 patients. A random effect, inverse variance metaanalysis was conducted, and a subgroup analysis of the PCI-DES cohort was performed. Events were compared as risk ratios using a 95% confidence interval (CI). Heterogeneity of results was evaluated by Eggers I2 test. Results are presented as early (0 to 1 year) and midterm (2 to 5 years). Results. Midterm mortality in the PCI and MIDCABG groups (3.6% and 2.6%, respectively) was comparable (1.24, 95% CI: 0.66 to 2.33; p [ 0.5; I2 [ 0%). Risk of early restenosis was lower in the MIDCABG cohort compared
with PCI (0.40, 95% CI: 0.16 to 0.99; p [ 0.05; I2 [ 57%). Although the early risk of recurrence of angina was comparable, over time it was 61% (43% to 74%) lower for MIDCABG patients (p < 0.001). Midterm results on analysis of the entire cohort demonstrated an increased risk for target vessel reinterventions (3.84, 95% CI: 2.7 to 5.5; p < 0.001) in the PCI cohort. Subgroup analysis revealed that the PCI-DES cohort (4 studies; 456 patients) had a higher risk of recurrent angina (risk ratio 3.4, 95% CI: 1.9 to 6.2; p < 0.001; I2 [ 0%) and target vessel reinterventions (risk ratio 4.16, 95% CI: 2.7 to 6.6; p < 0.001; I2 [ 0%) at midterm follow-up (2 to 5 years). Conclusions. Survival rates are comparable after either MIDCABG or PCI for proximal LAD disease. However, even the use of DES was associated with significantly higher rates of angina recurrence and the need for target vessel reintervention as compared with MIDCABG.
T
stents (DES). Therefore, we conducted a metaanalysis comparing clinical events associated with PCI and MIDCABG in the treatment of p-LAD disease.
herapeutic intervention for coronary artery disease consists of percutaneous intracoronary stenting (PCI) or coronary artery bypass graft surgery (CABG). Compared with surgery, PCI is a less invasive procedure associated with lower morbidity. Minimally invasive direct coronary artery bypass (MIDCABG) is a recent introduction, where the grafting is performed through a small thoracotomy incision. It is associated with a shorter hospital stay and earlier recovery than conventional surgery. Results comparing MIDCABG and PCI in proximal left anterior descending artery (p-LAD) disease are conflicting, especially with the recent addition of drug-eluting
(Ann Thorac Surg 2014;-:-–-) Ó 2014 by The Society of Thoracic Surgeons
Material and Methods Inclusion Criteria
Accepted for publication Jan 28, 2014.
Original articles (randomized controlled trials and observational studies) comparing PCI and MIDCABG for p-LAD disease (from 2000 to May 2013) were identified. The search was limited to (1) human subjects, (2) original articles, and (3) English language. Only studies reporting clinically relevant endpoints (eg, mortality, recurrence of angina, or reinterventions during follow-up) were included. Editorials and review articles were excluded.
Presented at the Poster Session of the Fiftieth Annual Meeting of The Society of Thoracic Surgeons, Orlando, FL, Jan 25–29, 2014.
Technique of MIDCABG
Address correspondence to Dr Deo, Adventist Wockhardt Heart Institute, Surat Gujarat 395001, India; e-mail: [email protected].
MIDCABG consisted of harvest of the left internal thoracic artery (LITA) and its subsequent grafting to the
Ó 2014 by The Society of Thoracic Surgeons Published by Elsevier Inc
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2014.01.086
2
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
Abbreviations and Acronyms BMS CI DES LAD LITA MIDCABG
= = = = = =
p-LAD
=
RCT RR TVR
= = =
bare metal stent confidence interval drug-eluting stent left anterior descending artery left internal thoracic artery minimally invasive direct coronary artery bypass graft surgery proximal left anterior descending artery randomized controlled trial risk ratio target vessel reintervention
LAD. That was most commonly accomplished through a left thoracotomy, in some cases, robotically assisted. Sequential grafting of the LITA was occasionally performed; however, only the LAD and its branches was grafted in all included studies.
Percutaneous Intervention Percutaneous coronary intervention was performed in the routine manner. Studies used a relatively uniform strategy of stent deployment with at least 90% patients having undergone either bare-metal stent (BMS) or DES insertion. A separate subgroup analysis of articles reporting outcome with DES was also conducted.
Studied Endpoints Clinical endpoints compared were duration of hospital stay, mortality, recurrence of angina, myocardial infarction, and need for reintervention. All results were assimilated into two periods: early (0 to 1 year) and midterm (2 to 5 years).
Search Strategy We conducted a systematic search of articles on EMBASE, Scopus, Web of Science, and MEDLINE using the terms “single vessel disease,” “left anterior descending artery,” “percutaneous intervention,” “drug-eluting stent,” “coronary artery bypass,” “minimally invasive coronary artery bypass,” and “MIDCABG” in various combinations. Two of the authors (V.S., S.V.D) independently screened abstracts and retrieved the full text articles. All full text articles were then evaluated independently to conform to the inclusion criteria. Discrepancies were resolved by consensus. Data were obtained from the studies using a prespecified data abstraction form.
Statistical Analysis Statistical analysis for binary data was performed using the “meta” package for R 3.0.1 [1, 2]. The risk ratio (RR) was implemented as the effect estimate. In the inverse weighted model, each study contributes a percentage of the final pooled estimate [3]. That is presented in each forest plot under the column of weight (W). As per Bate’s correction, 0.5 was added to each cell of the two-
Ann Thorac Surg 2014;-:-–-
by-two table in case the study or control arm had zero events [4]. The cutoff for the p value is 0.05; data are presented with 95% confidence intervals (CI). To compare the hospital stay, studies were pooled using the standardized mean difference as the effect estimate. Results are presented as a forest plot, depicting each individual risk ratio (RR) as well as the overall composite effect estimate. A risk ratio with its 95% CI less than 1 would favor MIDCABG. A random effects model has been implemented as both randomized controlled trial (RCT) and retrospective studies have been pooled [3]. Heterogeneity is defined as the variation among studies that would contribute to the overall results. It was evaluated using the Eggers I2 test and further stratified into low (25% to 49%), moderate (50% to 74%), and high (>75%) heterogeneity [3]. A subgroup analysis was performed after excluding the observational studies and compared to the overall effect. A separate subgroup analysis was conducted of studies implementing DES as their technique of percutaneous intervention. Publication bias was excluded by a visual inspection of the contour-enhanced funnel plot symmetry and determined using the Begg-Mazumdar rank correlation test. The results are presented as per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines for metaanalysis of observational studies [5, 6].
Results The initial search criteria yielded 312 articles. Exclusion of duplicates and further refinement of Medical Subject Headings yielded 193 abstracts, which were evaluated for inclusion. The detailed PRISMA flow diagram is presented in Figure 1. Two articles were excluded to avoid patient duplication. A total of 14 articles [7–19] discussing the results of 12 studies (seven RCTs and five observational studies) fulfilled the selection criteria. Two articles presented results of the same randomized controlled trial at a follow-up period of 6 months and 5 years [9, 10]. Etienne and colleagues [7, 8] have presented two articles discussing a retrospective analysis of data from two institutions. Drenth and associates [16, 20] have also presented their data initially in 2002 and then results of a longer follow-up in 2004. Blazek and colleagues [21] presented data regarding results of BMS at the end of 10 years of follow-up. As other articles failed to present data with such a long follow-up period, it was excluded from the metaanalysis. Herz and coworkers [22] was excluded because the study included patients with triple vessel disease. Care was taken to ensure nonduplication of data during analysis. The detailed patient demographics are presented in Tables 1 and 2. As the majority of patients were part of RCTs, patient population between cohorts is relatively well matched. In four articles [10, 11, 15, 16], angiographic evaluation was performed at the end of the follow-up period. Drug-eluting stents were implemented as the interventional strategy in five studies (463 patients). Sirolimus- or paclitaxel-eluting stents were used for
Ann Thorac Surg 2014;-:-–-
revascularization in all studies [7, 8, 12, 14, 15]. Antiplatelet agent therapy before and after stent implantation was as per institutional protocol. All the RCTs naturally had well-matched baseline data. Ben-Gal and colleagues [15] have utilized MIDCABG for patients with occluded LAD vessels, as have others. BenGal and colleagues [15] have also incorporated 48% patients with triple vessel disease who underwent only LAD intervention for various reasons. Shirai and coworkers [19] have retrospectively compared the unpublished results of the Patency, Outcomes and Economics of MIDCABG trial with demographically matched patients from their institutional database who underwent PCI; the investigators admit that angiographic data regarding the Patency, Outcomes and Economics of MIDCABG trial cohort was not available while selecting matched patients. The MIDCABG was performed through a small left thoracotomy in nine studies. In these patients, the LITA was harvested either with the help of thoracoscopic instruments or under direct vision. Two investigators [17, 19] have not explicitly presented the method of MIDCABG, whereas ministernotomy with direct LITA harvest was the preferred technique for Kim and associates [13]. Surgery was performed in an off-pump manner without cardiopulmonary bypass. The use of antiplatelet agents after surgery varied per institutional protocol. Percutaneous intervention was performed in the conventional manner. In the five studies implementing DES,
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
3
either paclitaxel-eluting or sirolimus-eluting stents were used. The use of antiplatelet agents varied between studies.
Length of Hospital Stay A pooled analysis from five studies (1,069 patients) demonstrated that hospital stay was significantly shorter after PCI as compared with MIDCABG (p ¼ 0.001; Fig 2). Two articles [12, 18] presented data as median (range); hence, they were not pooled in this analysis; however, both reported a reduced duration of stay in the PCI cohort.
Early Angiographic Patency Early postprocedure angiographic evaluation was performed in four studies. MIDCABG has a slight advantage over PCI with regard to the risk of restenosis (RR 0.44, 95% CI: 0.16 to 0.99; I2 ¼ 57.3%; p ¼ 0.046). In this analysis, only Thiele and associates [12] incorporated DES, whereas all others implemented BMS.
Mortality Early mortality was presented in 10 studies (993 MIDCABG patients and 1589 PCI patients; Fig 3). Mortality was comparable at 1.3% and 1.7% in the MIDCABG and PCI cohorts, respectively (RR 1.17, 95% CI: 0.53 to 2.57; p ¼ 0.7; I2 ¼ 0%; Fig 1). Midterm mortality was presented in six studies; incidence of mortality (3.6% MICABG; Fig 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart demonstrating procedure of the systematic review process. Reproduced from Moher et al [5].
4
Variable
Cisowski et al 2002 Drenth et al 2002a
Country Poland Netherlands Study design RCT RCT Duration 12 months 48 months Cohort BMS MIDCABG BMS MIDCABG No. of pts. 50 50 51 51 Age, years 53 10 54 9.1 61 1.3 60 1.6 Men 84 82 75 78 DM 8 6 18 24 HTN 52 56 33 16 Smokers 52 48 30 37 Prior AMI . . 18 24 Unstable angina 10 8 . . EF . . . . MIDCABG technique Thoracotomy Thoracotomy
Iakovou et al 2002 United States Prospective OBS 39 months BMS MIDCABG 441 119 63 12 62 12 68 71 22 17 54 55 46 56 22 22 68 60 52 12 48 7 Exact method not specified
a Drenth, 2004, presented updated 4-year results of the same randomized controlled trial. (range).
b
Diegler et al 2002b
Reeves et al 2004
Germany United Kingdom RCT RCT 6 months 12 months BMS MIDCABG BMS MIDCABG 110 110 50 50 62 10.2 61 10 54 (49–61)c 58 (53–67)c 72 77 86 70 34 25 . . 72 71 . . 25 25 . . 45 45 . . . . . . 62 15 63 11 . . Thoracotomy Thoracotomy
Shirai et al 2004
Kim et al 2005
United States Prospective OBS 92 months BMS MIDCABG 429 152 63 11 61 12 65 73 24 26 53 59 19 15 43 37 . . 51 11 53 11 Exact method not specified
South Korea RCT 24 months BMS MIDCABG 50 50 61 12 63 12 60 70 20 15 55 55 45 55 22 22 65 55 51 11 49 13 Ministernotomy
Thiele, 2005, presented updated 5-year results of the same randomized controlled trial.
c
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
Table 1. Clinical Characteristics of Studies Implementing Bare Metal Stents for Percutaneous Coronary Intervention
Age is median
Values are percent, unless otherwise indicated. AMI ¼ acute myocardial infarction; DM ¼ diabetes mellitus; observational; PCI ¼ percutaneous coronary intervention;
EF ¼ ejection fraction; HTN ¼ hypertension; MIDCABG ¼ minimally invasive coronary artery bypass graft surgery; pts. ¼ patients; RCT ¼ randomized controlled trial.
OBS ¼
Ann Thorac Surg 2014;-:-–-
Ann Thorac Surg 2014;-:-–-
Table 2. Clinical Characteristics of Studies Implementing Drug-Eluting Stents for Percutaneous Intervention Variable
Ben Gal et al 2006
Thiele et al 2009
Etienne et al 2009
Etienne et al 2013
South Korea RCT 6 months DES MIDCABG 119 70 60 9 61 9 64 64 37 48 50 55 40 45 21 22 50 42 52 8 51 9 Thoracotomy Paclitaxel 20%, sirolimus 80%
Israel PM OBS 24 months DES MIDCABG 83 83 38.6%a 38.6%a 80 80 30 29 47 55 . . 31 42 . . 7.2%d 1.2%d Thoracotomy Sirolimus 100%
Germany RCT 12 months DES MIDCABG 65 65 66 (59–72)b 66 (59–71)b 69 71 28 25 83 85 14 18 23 23 . . 65 (60–66) 65 (60–70) Thoracotomy Sirolimus 100%
Belgium Retro OBS 24 months DES MIDCABG 154 154 62c 63c . . 62 23 71 55 43 56 27 16 24 18 2%d 3%d Thoracotomy Paclitaxel 59%, sirolimus 41%
Belgium Retro PM OBS 60 months DES MIDCABG 196 260 64 12 63 12 48 34 60 21 67 57 46 54 21 12 21 15 1%d 3%d Thoracotomy Paclitaxel 59%, sirolimus 41%
a b Ben Gal et al presented data as patient age greater than 70 years. Thiele et al presented data as median (interquartile range). Etienne et al presented percentage of patients with ejection fraction less than 30%.
c
Etienne et al presented data only as mean.
d
Ben Gal et al and
Values are percent, unless otherwise indicated. AMI ¼ acute myocardial infarction; DES ¼ drug-eluting stent; DM ¼ diabetes mellitus; EF ¼ ejection fraction; HTN ¼ hypertension; MIDCABG ¼ minimally invasive direct coronary artery bypass graft surgery; OBS ¼ observational; PCI ¼ percutaneous coronary intervention; PM ¼ propensity matched; RCT ¼ randomized controlled trial; Retro ¼ retrospective.
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
Country Study design Study duration Cohort No. of patients Age, years Men DM HTN Smokers Prior AMI Unstable angina EF MIDCABG technique Type of DES
Hong et al 2005
5
6
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
Ann Thorac Surg 2014;-:-–-
Fig 2. Forest plot pooling together the duration of hospital stay for both cohorts. The minimally invasive direct coronary artery bypass graft surgery (MIDCAB) group had a significantly longer stay compared with the percutaneous coronary intervention (PCI) cohort (p ¼ 0.001). (CI ¼ confidence interval; SMD ¼ standardized mean difference.)
2.6% PCI) was comparable in both groups (RR 1.24, 95% CI: 0.65 to 2.33; p ¼ 0.50). The funnel plot of pooled studies demonstrates absence of publication bias (p ¼ 0.216; Fig 4).
angina (p < 0.001; I2 ¼ 0%). The RR of recurrent angina was 2.56 (95% CI: 1.75 to 3.84) in the PCI cohort compared with the MIDCABG cohort.
Myocardial Infarction Recurrence of Angina Early recurrent angina was present in 16.5% and 20.2% patients in the MIDCABG and PCI cohorts, respectively (p ¼ 0.34; Fig 5). However, at the end of 2 to 5 years of follow-up, while 93.2% of the MIDCABG group were symptom free, only 79% of the PCI cohort were free of Fig 3. The forest plot shows the pooled results of mortality between the two cohorts, and demonstrates that mortality in the two groups is comparable (A) during early follow-up and (B) at midterm follow-up. (CI ¼ confidence interval; MIDCAB ¼ minimally invasive direct coronary artery bypass graft surgery; PCI ¼ percutaneous coronary intervention; RR ¼ risk ratio.)
The incidence of early myocardial infarction was comparable in both cohorts (2.3% and 0.6% in the MIDCABG and PCI cohorts, respectively; p ¼ 0.08; Fig 6). The risk of myocardial infarction (RR 0.78, 95% CI: 0.36 to 1.69) at midterm follow-up was similar in both the MIDCABG cohort and the PCI cohort (p ¼ 0.28).
Ann Thorac Surg 2014;-:-–-
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
7
to 6.67) times higher in the DES cohort (21.5%) as compared with the MIDCABG cohort (4.8%; p < 0.001; I2 ¼ 0%; Fig 8).
Comment
Fig 4. Contour-enhanced funnel plot demonstrating absence of publication bias among included studies. (Dark gray area ¼ 0.1 > p > 0.05; medium gray area ¼ 0.05 > p > 0.01; light gray area ¼ p > 0.01.)
Target Vessel Reintervention At the end of 1 year of follow-up, target vessel reintervention (TVR) was required in 4.3% and 17.9% of the MIDCABG cohort and PCI cohort, respectively (p < 0.001; I2 ¼ 0%; Fig 7). In the pooled midterm analysis, 5.2% of MIDCABG patients and 21.5% of PCI patients needed TVR (p < 0.001; I2 ¼ 0%). Thus, the risk of TVR was 3.84 (95% CI: 2.7 to 5.55) times higher in the PCI cohort compared with the MIDCABG cohort at midterm follow-up.
Subgroup Analysis in Studies Implementing DES Five articles pooling four studies comparing MIDCABG and DES were analyzed. Two articles reported recurrence of angina at midterm; it was and 6.8% and 23.5%, respectively, in the MIDCABG cohort and DES cohort (p < 0.001; I2 ¼ 0%). At midterm follow-up, the risk for TVR (after pooling three studies) was 4.16 (95% CI: 2.70
Guidelines regarding PCI and CABG in multivessel coronary artery disease are relatively well established. However, limited data are available comparing MIDCABG and PCI for p-LAD disease, especially in the present era of DES. We therefore conducted a pooled analysis of more than 2,000 patients comparing the clinical outcomes in these cohorts. Our metaanalysis and systematic review of 12 studies (seven RCTs) demonstrate that recurrence of angina and the need for a reintervention is much lower in the MIDCABG cohort as compared with the PCI cohort at the end of 2 to 5 years of follow-up. This important difference is maintained even when comparing MIDCABG versus DES. Percutaneous intervention, being a less invasive procedure than CABG, is an attractive option for patients, especially as DES are reported to have lower restenosis rates compared with BMS. MIDCABG was introduced as a minimally invasive alternative to conventional CABG. The surgical approach is through a left thoracotomy or, occasionally, a ministernotomy. After harvesting the LITA either with the use of thoracoscopic instruments or under direct vision, an off-pump LITA-LAD anastomosis is constructed. The smaller skin incision, avoidance of complete sternotomy, and the off-pump technique definitely reduce morbidity of this procedure compared with routine CABG [23]. Important factors determining success after PCI are vessel-related factors. Our pooled analysis contains seven RCT; although patients were primarily matched Fig 5. Forest plots demonstrating recurrence of angina in the (A) early period and (B) intermediate period. (CI ¼ confidence interval; MIDCAB ¼ minimally invasive direct coronary artery bypass graft surgery; PCI ¼ percutaneous coronary intervention; RR ¼ risk ratio.)
8
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
Ann Thorac Surg 2014;-:-–-
Fig 6. Forest plot demonstrates the individual and pooled incidence of myocardial infarction in the minimally invasive direct coronary artery bypass graft surgery (MIDCABG) and percutaneous coronary intervention (PCI) cohorts. The incidence of acute myocardial infarction (AMI) was comparable at (A) early follow-up and (B) and midterm follow-up. (CI ¼ confidence interval; RR ¼ risk ratio.)
according to demographics, evaluation of angiographic data demonstrates comparability. However, as reported by other retrospective studies, patients with calcified, occluded, or highly tortuous vessels were recommended for MIDCABG. Such procedure selection is important
Fig 7. Forest plot demonstrates the pooled need for reintervention in the (A) early period and (B) intermediate period. (CI ¼ confidence interval; MIDCAB ¼ minimally invasive direct coronary artery bypass graft surgery; PCI ¼ percutaneous coronary intervention; RR ¼ risk ratio.)
in the “real world” to ensure optimal results for our patients. An important aspect of MIDCABG is appropriate training. Diegeler and coworkers [24] have reported an increase in patency rate from 96% to 99% from 1997 to
Ann Thorac Surg 2014;-:-–-
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
9
Fig 8. The forest plot presents the need for target vessel reintervention (TVR [midterm follow-up]) stratified by the type of drug-eluting stent (DES) used. As demonstrated by the lower part of the figure (DES ¼ 1), the risk of reintervention was higher with DES when compared with minimally invasive direct coronary artery bypass graft surgery (MIDCABG). (CI ¼ confidence interval; PCI ¼ percutaneous coronary intervention; RR ¼ risk ratio.)
1999. A small, 3.4%, risk of conversion to sternotomy is also present. Cremer and associates [25] evaluated 205 MIDCABG patients with coronary angiography and reported patency rates similar to those for conventional CABG. However, they did report instances of large unclipped collateral branches of the ITA being missed, inadvertent grafting of the ITA to the diagonal branches, and ITA kinking due to excessive length. Such events are definitely more likely when operating on a moving target within a small access area. Therefore, although newer techniques have made the procedure more appealing to the patient, it is important to understand that each new instrument has its own learning curve. A study comparing patency rates of single vessel bypass by conventional and MIDCABG methods demonstrates the importance of appropriate training [26]. The pooled articles did not present comparison of economic data. A recent cost-effectiveness analysis [27] demonstrates that stenting was beneficial for the first 2 years. From the third year on, given the higher rates of reintervention and recurrence of anginal symptoms in the PCI cohort, it no longer was the preferred option. The investigators also separately analyzed the use of DES and reached the same conclusion. An earlier metaanalysis on this topic pooled five RCTs predominantly containing patients undergoing PCI with BMS [28]. Our results are based on a pooled analysis of more than 2,000 patients. Importantly, given the increase in available data from then to now, we were able to stratify results according to follow-up (early and midterm) as well as stent type (BMS and DES). Both factors are extremely important in reaching an unbiased conclusion.
Study Strengths and Limitations As with all metaanalyses, our results are dependent on reported study-level data. Although some within study variation was present among the MIDCABG and PCI cohorts, overall patient matching was excellent. Additionally, a large proportion of our patients were part of
various RCTs. A very low heterogeneity in our pooled results demonstrates minimal between-study variations. In conclusion, survival rates are comparable after either MIDCABG or PCI for p-LAD disease. However, even the use of DES was associated with significantly higher rates of angina recurrence and the need for TVR as compared with MIDCABG.
References 1. Schwarzer G. Package “meta.” Meta-analysis with R, version 2.5-1, 2013. 2. Core Team R. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2013. 3. Higgins JPT, Green S, editors. Cochrane handbook for systematic reviews of interventions, version 5.1.0 [updated March 2011]. The Cochrane Collaboration, 2011. 4. Friedrich JO, Adhikari NK, Beyene J. Inclusion of zero total event trials in meta-analyses maintains analytic consistency and incorporates all available data. BMC Med Res Methodol 2007;7:5. 5. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ 2009;339:b2535. 6. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994;50:1088–101. 7. Etienne PY, Glineur D, Papadatos S, et al. Comparison of minimally invasive direct coronary artery bypass surgery with implantation of drug-eluting stents in patients with left anterior descending coronary artery disease. Innovations (Phila) 2009;4:340–4. 8. Etienne PY, D’Hoore W, Papadatos S, et al. Five-year followup of drug-eluting stents implantation vs minimally invasive direct coronary artery bypass for left anterior descending artery disease: a propensity score analysis. Eur J Cardiothorac Surg 2013 March 14 [E-Pub ahead of print]. 9. Thiele H, Oettel S, Jacobs S, et al. Comparison of bare-metal stenting with minimally invasive bypass surgery for stenosis of the left anterior descending coronary artery: a 5-year follow-up. Circulation 2005;112:3445–50. 10. Diegeler A, Thiele H, Falk V, et al. Comparison of stenting with minimally invasive bypass surgery for stenosis of the left anterior descending coronary artery. N Engl J Med 2002;347:561–6. 11. Cisowski M, Drzewiecki J, Drzewiecka-Gerber A, et al. Primary stenting versus MIDCAB: preliminary report— comparison of two methods of revascularization in single left
10
12.
13. 14.
15.
16.
17.
18.
19.
DEO ET AL MIDCABG OR PCI FOR PROXIMAL LAD STENOSIS
anterior descending coronary artery stenosis. Ann Thorac Surg 2002;74(Suppl):1334–9. Thiele H, Neumann-Schniedewind P, Jacobs S, et al. Randomized comparison of minimally invasive direct coronary artery bypass surgery versus sirolimus-eluting stenting in isolated proximal left anterior descending coronary artery stenosis. J Am Coll Cardiol 2009;53:2324–31. Kim JW, Lim DS, Sun K, et al. Stenting or MIDCAB using ministernotomy for revascularization of proximal left anterior descending artery? Int J Cardiol 2005;99:437–41. Hong SJ, Lim DS, Seo HS, et al. Percutaneous coronary intervention with drug-eluting stent implantation vs. minimally invasive direct coronary artery bypass (MIDCAB) in patients with left anterior descending coronary artery stenosis. Catheter Cardiovasc Interv 2005;64:75–81. Ben-Gal Y, Mohr R, Braunstein R, et al. Revascularization of left anterior descending artery with drug-eluting stents: comparison with minimally invasive direct coronary artery bypass surgery. Ann Thorac Surg 2006;82:2067–71. Drenth DJ, Winter JB, Veeger NJ, et al. Minimally invasive coronary artery bypass grafting versus percutaneous transluminal coronary angioplasty with stenting in isolated high-grade stenosis of the proximal left anterior descending coronary artery: six months’ angiographic and clinical follow-up of a prospective randomized study. J Thorac Cardiovasc Surg 2002;124:130–5. Iakovou I, Dangas G, Mehran R, et al. Minimally invasive direct coronary artery bypass (MIDCAB) versus coronary artery stenting for elective revascularization of the left anterior descending artery. Am J Cardiol 2002;90:885–7. Reeves BC, Angelini GD, Bryan AJ, et al. A multi-centre randomised controlled trial of minimally invasive direct coronary bypass grafting versus percutaneous transluminal coronary angioplasty with stenting for proximal stenosis of the left anterior descending coronary artery. Health Technol Assess 2004;8:1–43. Shirai K, Lansky AJ, Mehran R, et al. Minimally invasive coronary artery bypass grafting versus stenting for patients with proximal left anterior descending coronary artery disease. Am J Cardiol 2004;93:959–62.
Ann Thorac Surg 2014;-:-–-
20. Drenth DJ, Veeger NJ, Middel B, et al. Comparison of late (four years) functional health status between percutaneous transluminal angioplasty intervention and off-pump left internal mammary artery bypass grafting for isolated highgrade narrowing of the proximal left anterior descending coronary artery. Am J Cardiol 2004;94:1414–7. 21. Blazek S, Holzhey D, Jungert C, et al. Comparison of baremetal stenting with minimally invasive bypass surgery for stenosis of the left anterior descending coronary artery: 10-year follow-up of a randomized trial. J Am Coll Cardiol Intv 2013;6:20–6. 22. Herz I, Moshkovitz Y, Hendler A, et al. Revascularization of left anterior descending artery with drug-eluting stents: comparison with off-pump surgery. Ann Thorac Surg 2005;79:88–92. 23. Holzhey DM, Jacobs S, Mochalski M, et al. Seven-year follow-up after minimally invasive direct coronary artery bypass: experience with more than 1300 patients. Ann Thorac Surg 2007;83:108–14. 24. Diegeler A, Spyrantis N, Matin M, et al. The revival of surgical treatment for isolated proximal high grade LAD lesions by minimally invasive coronary artery bypass grafting. Eur J Cardiothorac Surg 2000;17:501–4. 25. Cremer J, Mugge A, Wittwer T, et al. Early angiographic results after revascularization by minimally invasive direct coronary artery bypass (MIDCAB). Eur J Cardiothorac Surg 1999;15:383–8. 26. Vicol C, Nollert G, Mair H, et al. Midterm results of beating heart surgery in 1-vessel disease: minimally invasive direct coronary artery bypass versus off-pump coronary artery bypass with full sternotomy. Heart Surg Forum 2003;6:341–4. 27. Rao C, Aziz O, Panesar SS, et al. Cost effectiveness analysis of minimally invasive internal thoracic artery bypass versus percutaneous revascularisation for isolated lesions of the left anterior descending artery. BMJ 2007;334:621. 28. Jaffery Z, Kowalski M, Weaver WD, Khanal S. A meta-analysis of randomized control trials comparing minimally invasive direct coronary bypass grafting versus percutaneous coronary intervention for stenosis of the proximal left anterior descending artery. Eur J Cardiothorac Surg 2007;31:691–7.
