Drug-eluting stents versus bare metal stents in saphenous vein graft disease: insights from a meta-analysis of 7,090 patients

E. NAVARESE (1), A. BUFFON(1), A. LUPI (2), M. SANSA (2), A. BONGO (2), S. DE SERVI (3) 1) Catholic University of Sacred Heart, ROME, ITALY; 2) Ospedale “Maggiore della Carità, NOVARA, ITALY; 3) Civic Hospital Legnano, Legnano, ITALY;

Clinicaltrials.gov Identifier: NCT01036048

Background

• Aortocoronary saphenous vein graft (SVG) disease represents the “Achilles’ heel” of CABG interventions due to SVG high failure rate • Drug-eluting stents (DES) are a major advance in interventional cardiology, but evidence for using these devices does not exist for many “off-label”

indications

• No clear indication is available about DES use in SVG disease, a common “off-label” indication for DES implantation in daily practice

Background

Histopathology of SVG degeneration the degeneration of SVG is quite a different phenomenon in comparison with native coronary artery atherosclerosis a)

Cross-section of an SVG implanted with a self-expanded metallic stent and developing late in-stent restenosis. The stent struts show a thick neointima developed within the stent lumen close to the struts (asterisk). The wires on the right-hand side (arrow) are close to a necrotic core and only a thin layer of healing neointima is observed in the lumen side of the stent (b) Atherosclerotic plaque and large thrombus protruding into the SVG lumen through the stent struts (asterisk).

a)

b) Ribichini, Histopathology of saphenous vein grafts, Clinical Science 2008

Aims and methods

The aim of this work was to perform a meta-analysis on DES vs BMS in SVG disease

• Inclusion criteria were: 1) randomized and/or non randomized

studies

2) studies reporting clinical outcomes as

overall death and/or acute myocardial infarction and/or target vessel revascularization

3) follow up period longer than 6 months • Odds ratios (ORs) were computed from individual studies and pooled according to a fixed effect (e.g. inverse variance weighting) or random effect model in case of statistical heterogeneity

Flow chart of the meta-analysis (N= 7090 patients) 1216 citations retrieved from data base searches 40 trials assessed according to the selection criteria 3 randomized controlled and 21 non-randomized studies 1176 titles/abstracts excluded because non-relevant 16 studies excluded accor-ding to explicit selection criteria:

•

9 lacking a control group

•

1 availability of another report from the same study with longer follow up

•

3 with SVG data mixed with other type of lesions

•

3 not reporting one of more of the pre specified endpoints

Vermeersch Brilakis Jeger Applegate Assali Bansal Brodie Chu Ellis Ge Gioia Goswami Kaplan Latib Lee Minutello Nair Okabe Ramana Shishehbor van Twisk Vignali Whorle Wilson

Quality table of included studies

Randomized Controlled Studies Selection bias

A A A

Performance bias

A A A

Attrition bias

A A A

Non-Randomized Observational Studies Detection bias

A A A

Selection

                    

Comparability

                    

Outcome

                     Randomized studies: the quality was appraised according to the Cochrane Collaboration , estimating separately the risk of selection, performance, detection, and attrition bias (expressed as low risk of bias [A], moderate risk of bias [B], high risk of bias [C] Non-randomized studies: The Newcastle Ottawa Scale for non randomized studies assigns star for three area of study quality: selection, comparability and outcome

Results: mortality

Favour DES Favour BMS

Results: myocardial infarction

Favour DES Favour BMS

Results: Target vessel Revascularization

Favour DES Favour BMS

Results: Meta-regression for TVR

Discussion (1)

• This meta-analysis offers an evidence summing up the whole literature from randomized and observational studies comparing DES vs BMS use in SVG disease • In the present analysis DES were found to significantly reduce rate of TVR but did reduce neither mortality in the meta-analysis of randomized studies nor rate of overall MI • Meta-regression highlighted that benefits from DES in the reduction of

TVR are more substantial in older SVGs

Discussion (2)

•

Death

The opposite finding on mortality raised from the randomized and observational studies suggests that a definitive conclusion in favour to DES use in this setting cannot be drawn yet •

MI

the benefit from DES use could be largely diluted by acute coronary syndromes arising from other previously untreated coronary lesions

To exhaustively address these controversial findings, randomized trials powered for mortality and MI with adequate long-term follow up are warranted