Abstract
Background/purpose
Percutaneous coronary intervention (PCI) is the preferred strategy for treatment of saphenous vein graft (SVG) disease. However, there remains ongoing debate on whether drug-eluting stents (DES) or bare-metal stents (BMS) should be used during SVG-PCI.
Methods/materials
We performed a meta-analysis of randomized controlled trials (RCTs) comparing DES and BMS for SVG-PCI. The primary end point was major adverse cardiac events (MACE), defined as composite of all-cause death, myocardial infarction (MI) or repeat revascularization. Secondary end points included individual MACE components, cardiac death and stent thrombosis (ST).
Results
Six RCTs including 1582 patients (50% receiving DES) met inclusion criteria. MACE occurred in 31% (244/797) patients receiving DES and 36% (281/785) patients receiving BMS (median follow-up, 12–35 months). There was no significant difference in MACE between DES and BMS (Odds Ratio (OR) 0.62, 95%CI 0.36–1.09, p = 0.10, I 2 = 77%). However, for individual components of MACE, DES was associated with a significant reduction in repeat revascularization (OR 0.53, 95%CI 0.29–0.97, p = 0.04, I 2 = 73%). There was no difference in all-cause death (OR 1.30, 95%CI 0.77–2.20, p = 0.33, I 2 = 40%), MI (OR 0.68, 95%CI 0.38–1.25, p = 0.22, I 2 = 56%), cardiac death (OR 1.08, 95%CI 0.45–2.64, p = 0.86, I 2 = 42%) or ST (OR 0.89, 95%CI 0.37–2.17, p = 0.80, I 2 = 35%) between stents.
Conclusions
Although there was no significant difference in MACE, DES is associated with a reduction in repeat revascularization compared with BMS in pooled randomized trials for SVG-PCI. The high occurrence of MACE in both stent platforms highlights the need for novel therapeutic approaches to improve clinical outcomes following SVG intervention.
Summary
We performed a meta-analysis of randomized controlled trials comparing DES and BMS for SVG-PCI. There was no significant difference in MACE between DES and BMS. However, for individual components of MACE, DES was associated with a significant reduction in repeat revascularization. The high occurrence of MACE in both stent platforms highlights the need for novel therapeutic approaches to improve clinical outcomes following SVG intervention.
Highlights
- •
No difference in MACE between DES and BMS in SVG-PCI at longest follow-up
- •
DES associated with reduced repeat revascularization compared with BMS
- •
SVG-PCI is associated with high failure rate regardless of stent type.
1
Introduction
Saphenous vein graft (SVG) failure following coronary artery bypass graft (CABG) surgery remains an ongoing concern with about 40% of grafts being occluded at 10 years [ ]. Due to the high operative risk associated with patients undergoing redo-CABG [ ], percutaneous coronary intervention (PCI) and stenting is the preferred therapeutic option for patients presenting with progressive SVG disease and now represents ~6% of all PCI performed in the United States [ ].
The unique micro-environment of accelerated intimal proliferation and hyperplasia in venous conduits [ ] means that SVG PCI is associated with a significantly increased and persistent risk of late failure compared with native-artery intervention [ ]. In recent years, there has therefore been a shift towards the use of drug-eluting stents (DES) over bare-metal stents (BMS) during SVG PCI [ ], aiming to reduce in-stent restenosis and target lesion revascularization. Previous meta-analyses support this approach and have demonstrated that use of DES is associated with improved clinical outcomes following SVG PCI [ , ]. However, these results were influenced by relatively short duration of clinical follow-up and inclusion of a high proportion of observational data. However, recent data has suggested that the benefit of DES may not be maintained with the passage of time, possibly due to late ‘catch-up’ restenosis [ ]. This has led to uncertainty on whether DES or BMS should be used during SVG-PCI.
We therefore performed an updated systematic review and meta-analysis of randomized trials to evaluate the clinical outcomes for patients with SVG lesions undergoing PCI using DES compared with BMS.
2
Methods
2.1
Data sources and search strategy
A digital literature search was performed through the MEDLINE and EMBASE databases for the period up to 1 September 2017. Medical Subject Heading (MeSH) keywords including ‘saphenous vein’ and ‘drug eluting stents’ were used where available and the search was limited to English language and randomized controlled trials. Reference lists of eligible articles and previous meta-analyses were reviewed for further potential citations. Manual searches were also performed through presentations and abstracts of recent major international cardiovascular conferences. The study protocol was prospectively registered with the PROSPERO international register (CRD42017077683) and fully adhered to the PRISMA statement (Preferred Reporting Items for Systematic Reviews and Meta-Analyses). An example search strategy is presented in Table S1 in the Data Supplement.
2.2
Study selection
Study characteristics for inclusion were as follows: (1) randomized controlled clinical trial, (2) involvement of saphenous vein graft lesions and (3) comparison of clinical outcomes between DES and BMS. Studies arising from observational registry data or trials that included non-SVG lesions were excluded. We evaluated outcomes for each trial with preference for longest reported clinical follow-up.
2.3
Data items and collection process
Data items to be collected were specified prior to the literature search. Two investigators (F.J.H. and J.N.) independently conducted the literature search and performed data extraction for study design, patient baseline demographics, procedural characteristics and clinical outcomes. Extracted data were verified by the senior author (A.J.B.) with any discrepancies resolved by consensus. Risk of bias within individual studies was evaluated according to the Cochrane Collaboration Assessment for risk of bias in included studies.
2.4
Clinical end points
The primary end point of this study was major adverse cardiac events (MACE), defined as a composite of death, myocardial infarction (MI) or repeat revascularization. Death from any cause was prioritized and if not reported, cardiac death was used. Variations in definition included one trial specifying target-vessel MI rather than any MI [ ], while repeat revascularization was defined as target vessel revascularization (TVR) in five trials [ ] and target lesion revascularization (TLR) in one [ ]. Secondary end points included the individual components of MACE, as well as cardiac death and definite or probable stent thrombosis. MI included both procedural and non-procedural MI. Procedural MI was uniformly defined as an elevation of creatine kinase-myocardial band (CK-MB) >3 times upper limit of normal. The definition of non-procedural MI required an elevation in cardiac biomarkers (CK-MB or troponin) above the upper limit of normal, although the thresholds used and the necessity of electrocardiographic changes for MI diagnosis varied between trials. TVR was defined as any new revascularization (either PCI or CABG) procedure in the target vessel, and was performed in association with angina or objective evidence of myocardial ischemia (i.e. ischemia-driven) in one trial [ ]. TLR was defined in one trial as any repeat revascularization procedure involving the vessel supplied by the target venous graft in the presence of angiographic restenosis (≥50% on quantitative coronary angiography) with either angina or a positive functional study corresponding to the region served by the target graft, or diameter stenosis ≥70% at follow-up angiography in the absence of documented clinical or functional ischemia [ ]. Stent thrombosis was defined according to the Academic Research Consortium definition in four trials [ , , , , ], with one trial, published prior to the ARC definition, using the definition by Iakovou and colleagues [ ].
2.5
Statistical analysis
Data were analyzed by random-effects modeling for the primary end point and individual secondary end points at longest follow-up. Sensitivity analyses were performed to assess differences between median length of follow-up (<24 months versus ≥24 months), patients with or without diabetes mellitus, and by graft age at time of PCI (≤13 years versus >13 years). Alongside the Dersimonian and laird method, the Hartung-Knapp method was also applied in sensitivity analyses for a more conservative summary estimate [ ]. Summary statistics are reported as pooled odds ratios (ORs) with 95% confidence intervals (CIs). Statistical heterogeneity was quantified using the I 2 statistic. Heterogeneity was quantified as low, moderate, or high based on I 2 values of 25%, 50% and 75%, respectively [ ]. Meta-regression analyses were performed to evaluate the association between demographic variables and the individual trial ORs (log-transformed). Quantitative evidence of small study effects was assessed using Harbord’s modified test [ ]. A 2-sided p value of <0.05 was considered significant. Statistical analyses were performed using Stata MP 13.0 (Stata Corp LP, College Station, TX) and the metan suite of commands.
2
Methods
2.1
Data sources and search strategy
A digital literature search was performed through the MEDLINE and EMBASE databases for the period up to 1 September 2017. Medical Subject Heading (MeSH) keywords including ‘saphenous vein’ and ‘drug eluting stents’ were used where available and the search was limited to English language and randomized controlled trials. Reference lists of eligible articles and previous meta-analyses were reviewed for further potential citations. Manual searches were also performed through presentations and abstracts of recent major international cardiovascular conferences. The study protocol was prospectively registered with the PROSPERO international register (CRD42017077683) and fully adhered to the PRISMA statement (Preferred Reporting Items for Systematic Reviews and Meta-Analyses). An example search strategy is presented in Table S1 in the Data Supplement.
2.2
Study selection
Study characteristics for inclusion were as follows: (1) randomized controlled clinical trial, (2) involvement of saphenous vein graft lesions and (3) comparison of clinical outcomes between DES and BMS. Studies arising from observational registry data or trials that included non-SVG lesions were excluded. We evaluated outcomes for each trial with preference for longest reported clinical follow-up.
2.3
Data items and collection process
Data items to be collected were specified prior to the literature search. Two investigators (F.J.H. and J.N.) independently conducted the literature search and performed data extraction for study design, patient baseline demographics, procedural characteristics and clinical outcomes. Extracted data were verified by the senior author (A.J.B.) with any discrepancies resolved by consensus. Risk of bias within individual studies was evaluated according to the Cochrane Collaboration Assessment for risk of bias in included studies.
2.4
Clinical end points
The primary end point of this study was major adverse cardiac events (MACE), defined as a composite of death, myocardial infarction (MI) or repeat revascularization. Death from any cause was prioritized and if not reported, cardiac death was used. Variations in definition included one trial specifying target-vessel MI rather than any MI [ ], while repeat revascularization was defined as target vessel revascularization (TVR) in five trials [ ] and target lesion revascularization (TLR) in one [ ]. Secondary end points included the individual components of MACE, as well as cardiac death and definite or probable stent thrombosis. MI included both procedural and non-procedural MI. Procedural MI was uniformly defined as an elevation of creatine kinase-myocardial band (CK-MB) >3 times upper limit of normal. The definition of non-procedural MI required an elevation in cardiac biomarkers (CK-MB or troponin) above the upper limit of normal, although the thresholds used and the necessity of electrocardiographic changes for MI diagnosis varied between trials. TVR was defined as any new revascularization (either PCI or CABG) procedure in the target vessel, and was performed in association with angina or objective evidence of myocardial ischemia (i.e. ischemia-driven) in one trial [ ]. TLR was defined in one trial as any repeat revascularization procedure involving the vessel supplied by the target venous graft in the presence of angiographic restenosis (≥50% on quantitative coronary angiography) with either angina or a positive functional study corresponding to the region served by the target graft, or diameter stenosis ≥70% at follow-up angiography in the absence of documented clinical or functional ischemia [ ]. Stent thrombosis was defined according to the Academic Research Consortium definition in four trials [ , , , , ], with one trial, published prior to the ARC definition, using the definition by Iakovou and colleagues [ ].
2.5
Statistical analysis
Data were analyzed by random-effects modeling for the primary end point and individual secondary end points at longest follow-up. Sensitivity analyses were performed to assess differences between median length of follow-up (<24 months versus ≥24 months), patients with or without diabetes mellitus, and by graft age at time of PCI (≤13 years versus >13 years). Alongside the Dersimonian and laird method, the Hartung-Knapp method was also applied in sensitivity analyses for a more conservative summary estimate [ ]. Summary statistics are reported as pooled odds ratios (ORs) with 95% confidence intervals (CIs). Statistical heterogeneity was quantified using the I 2 statistic. Heterogeneity was quantified as low, moderate, or high based on I 2 values of 25%, 50% and 75%, respectively [ ]. Meta-regression analyses were performed to evaluate the association between demographic variables and the individual trial ORs (log-transformed). Quantitative evidence of small study effects was assessed using Harbord’s modified test [ ]. A 2-sided p value of <0.05 was considered significant. Statistical analyses were performed using Stata MP 13.0 (Stata Corp LP, College Station, TX) and the metan suite of commands.
3
Results
A total of 35 citations were reviewed and screened, with 22 studies identified for potential inclusion. On further evaluation, reasons for exclusion included studies not being a comparison of clinical outcomes between DES and BMS (5 studies), studies that were not randomized (7 studies) and several studies referring to same study population (5 studies). A detailed flow chart of the search algorithm is presented in Fig. 1 .
