Despite the fact that PCI indications have grown in the last years, the best revascularisation treatment for the patient with multiple vessel disease (MVD) is still controversial.
I - Balloon angioplasty (POBA) versus coronary surgery (CABG)
Before the introduction of coronary stents, several randomised trials had compared balloon angioplasty (POBA) to coronary surgery (CABG) (1-6). Those studies concluded that :
- late mortality and MI rates do not differ between the two modalities of treatment in the overall population,
- patients randomised to surgery have less recurrence of angina and need for further revascularisation procedures in a 5-8 years follow-up time(7-8)
While the same trials, however, suggested that
- the mortality rate in the subgroup of patients with diabetes was higher after POBA than after surgery, at least when the internal mammary artery was used.
The availability of coronary stents has made PCI procedures safer, with a reduction in the need for emergency CABG and has allowed better immediate angiographic results with a lower rate of restenosis. Furthermore, the introduction of GP IIb/IIIa inhibitors reduced the rate of periprocedural myocardial infarction (MI) and improved prognosis, in particular in diabetic patients treated with stent implantation(9-11).
II - PCI with BMS implantation versus CABG
On the other hand, surgical techniques have improved with the routine use of at least one and often multiple arterial grafts, beating heart surgery, mini-thoracotomy, avoiding sternotomy with the associated risks, and reducing pulmonary complications.
Four major randomised trials, which compared PCI with BMS implantation to CABG, have been published in the last few years.
1) The ARTS trial(12) enrolled 1205 low risk patients with MVD. At 1 year follow-up there were no significant differences in terms of mortality, stroke and MI, even though stented patients suffered of a higher rate of repeated revascularisation (16.8% vs. 3.5%, p<0.001) determining a better event-free survival rate in CABG patients (87.8% vs. 73.8%, p<0.001).
A sub analysis of ARTS(13) showed that patients undergoing CABG had a higher rate of complete revascularisation (84.1% vs. 70.5%, p<0.001) and that at 1 year patients treated with stents and incomplete revascularisation underwent CABG more often than those with complete revascularisation and stent implantation (10.0% vs. 2.0%, p<0.05), while no difference was noted in the surgical arm.
Among diabetic patients, those treated with stenting showed the lowest 1 year event free survival rate (63.4%) compared both to diabetic patients treated with CABG (84.4%, p<0.001) and to non diabetic patients treated with stenting (76.2%, p<0.04) but the difference was only caused by a higher rate of repeated revascularisation(14).
Similar results(15) were present at 3 years follow up when the event free survival rate was lower in stented diabetic patients compared to the surgery arm (52.7% vs. 81.3%, p<0.001). On the contrary, diabetic and non diabetic patients treated with CABG showed a similar rate of events (81.3% vs. 83.3%). However, the risk of cerebrovascular accidents (CVA) in diabetic patients treated with CABG was significantly higher than in patients treated with stent (4.2% vs. 0%), showing that diabetes is an independent risk factor for CVApost-CABG as documented in other studies(16).
At 5 years there was no difference in mortality between the stent and CABG arms (8.0% vs. 7.6%, p=0.83) in the overall population with a non significant increase in mortality after stenting in the diabetic patients (13.4% vs. 8.3%, p=0.27)(17). Another sub analysis of ARTS(18) showed no significant difference in terms of mortality, MI and CVA between patients with stable and unstable angina treated with PCI or CABG, with a need for repeat revascularisation significantly higher in those treated with stent in both groups.
2) The ERACI II trial(19) randomised 450 selected patients with MVD and indication to coronary revascularisation with stent or CABG.
At a mean follow up of 18 months, survival was higher in the stent arm compared to CABG (96.9% vs. 92.5%, p<0.017), despite a rate of repeat revascularisation four times higher in patients treated with stent (16.8% vs. 4.8%, p<0.002). It is important to emphasise that this study was criticised because of the excessively high 30-day mortality and MI in the surgery arm (5.7%) and because of the coil stent used which showed to be one of those with the highest risk of restenosis.
At 3 years follow up no differences were seen in terms of major adverse cardiac events (MACE) and MI between diabetic and non diabetic patients with a higher rate of repeat revascularisation confirmed in the diabetic PCI arm (36.4% vs. 27%, p<0.042).
At 5 years follow up, no survival benefit with any of the two revascularisation procedures was demonstrated but patients treated with CABG had better freedom from repeat revascularisation procedures (92.4% vs. 71.5%) and from MACE (76.4% vs. 65.3%; p=0.013) (20).
3) The SoS trial(21) involved 988 patients with MVD but with wider inclusion criteria compared to ARTS and ERACI II, including patients with low left ventricular ejection fraction (Table 1).
At a mean follow up of 2 years patients treated with PCI suffered a higher rate of repeat revascularisation (21% vs. 6%, p<0.0001), with a lower mortality in the CABG arm (2% vs. 5%, p=0.01).
In the diabetic subgroup of 142 patients there was a higher rate of repeat revascularisation in the stent arm with similar incidence of death and Q wave MI in both groups (10% vs. 12%, p ns). Recently, a sub analysis of SoS trial compared the 1 year outcomes of patients with and without acute coronary syndromes(22) and found that patients of both subgroups undergoing surgery had similar clinical outcomes, with less need for repeat revascularisation (ACS, 15.5% vs. 7.1%, p=0.04; non-ACS, 18% vs. 3.2%, p<0.001) but higher cost compared to PCI.
4) The AWESOME trial(23) included 454 patients randomised to PCI or CABG with more severe features compared to the previous studies that is the presence of refractory myocardial ischaemia and at least one of 5 risk factors (previous cardiac surgery, MI within 7 days, left ventricular ejection fraction <35%, age >70 years, intra-aortic balloon pump needed for stabilisation); furthermore two registries were activated, one including 1650 patients where the choice of treatment was left to the doctor and another one, including 327 patients angiographically eligible for the trial, where the patient chose the treatment. At 3 years follow up no survival difference was noted between CABG and PCI patients in the trial and in the two registries. The same survival outcomes were found in the subgroup of diabetic patients(24) but with a higher rate of repeat revascularisation in the PCI arm.
The results of this study suggest that PCI represents a valid alternative to CABG in patients with refractory ischaemia and high surgical risk. In these patients the choice of treatment must be balanced between risks and benefits associated with the two techniques, taking into account both doctor and patient choices.
III – Conclusions of these four trials
Certainly, these trials have influenced the therapeutic attitude of the “real world” cardiologists who prefer the surgical approach for those patients with diffuse and severe coronary disease and consider the interventional approach for those patients with multiple but relatively focal vessel disease where the reduced invasiveness of PCI procedures is counterbalanced by the increased need for repeat revascularisation at follow up.
Diabetes is in fact associated with unfavorable results with both CABG and PCI(25). Cardiac surgery in diabetic patients shows a higher immediate mortality rate and a higher risk of post-operative complications such as renal failure and infections of the sternotomy site, with subsequent clinical and economical effects.
On the other hand, PCI procedures in this category of patients are associated with a higher risk of mortality and restenosis. These trials showed that percutaneous revascularisation can be a relatively safe alternative to surgery in a selected group of diabetic patients with multiple vessel disease, at the cost of a higher rate of ischemic recurrences.
Table 1. Summary of major trials comparing PCI vs. CABG in the stent era.
|SoS (21)||AWESOME(23)||ARTS II (29)|
|Number of pts||1205||450||988||454||607|
|Primary end-point||MACCE-free survival at 1 year||MACCE at 30 days||Rate of repeat revascularisation||Survival||MACCE-free survival at 1 year|
|Type of study||Randomised
PCI vs CABG
PCI vs CABG
PCI vs CABG
PCI vs CABG
|PCI registry with same inclusion criteria ARTS|
|3-vessel disease (%)||31||56||42||45||54|
|GP IIb/IIIa use (%)||NA||28||8||11||33|
MACCE at 1-3 year F-up
|26.2 vs 12.2||22.2 vs 18.6*||22.5 vs 12.4||52 vs 39**||10.4|
NA = not available
* Derived from data presented in the paper, including death, MI and repeat revascularisation.
**Derived from data presented in the paper, including death, unstable angina and repeat revascularisation
IV- PCI with DES versus CABG
Recently, drug eluting stents (DES) have been shown in multiple randomised trials to reduce restenosis and target lesion revascularisation but not mortality or MI compared to bare metal stents (BMS)(26), but these trials were conducted in patients who received single lesion treatment and a high average LV function. No randomised clinical trials comparing DES-assisted PCI to CABG have been published yet but several single center registries have reported the safety and efficacy of multiple vessel DES implantation(27-28).
The ARTS-II(29) is the first and still the only available trial comparing DES to BMS implantation in patients with MVD. This study was a single arm, 45-center, European study in which 607 patients with inclusion criteria matching the ARTS I trial were treated with sirolimus eluting stent. Outcome was compared to the ARTS I trial. The primary endpoint of this study was the freedom from major adverse cardiac and cerebrovascular events (MACCE) at one year compared to the ARTS-I-CABG arm. The results of this trial were more satisfactory in the ARTS II arm compared to ARTS-I-CABG and ARTS I-PCI (MACCE rate 10.4% vs. 12.2% vs. 26.2%, respectively) even if the ARTS-II patients had more unfavorable characteristics than the ARTS-I trial.
The availability of DES and the improved technical approach to the chronic total occlusions will reduce the indications to surgery for the treatment of MVD. This advantage will be more difficult to achieve in the diabetic population in which, even if the problem of restenosis is reduced with the use of DES(30), there is the important negative influence on prognosis deriving from the progression of atherosclerosis and the risk of spontaneous myocardial infarction (the risk of MI varies from 1-2% /year in non-diabetic patients with single vessel disease to 10-15% /year in diabetic patients with three vessel disease)(31). From this point of view, CABG is obviously superior to PCI because, substituting the native diseased vessel which has vulnerable plaques prone to spontaneous rupture with a o disease free arterial graft, reduces the risk of such event(32).
Another important aspect to evaluate will be the economic impact related to the use of PCI compared to CABG in multivessel patients that at the moment does not seem to be cost-saving(33).
V – Conclusions regarding diabetic patients
Next step in the future is to evaluate whether the gap between diabetic and non diabetic patients with multivessel disease will be overcome through both the improvement of secondary prevention strategies and the technological evolution which will allow treating with DES not only the significant coronary artery lesions but even possible not significant vulnerable coronary plaques.
In the meantime, three on going trials comparing DES to CABG in patients with MVD (SYNTAX, FREEDOM and CARDia) will provide more data to debate about the optimal therapeutic strategy in multivessel patients.
Given the lack of available data from dedicated randomised clinical trials, and the last European Society of Cardiology guidelines for PCI published in 2005, PCI in patients with multivessel disease and/or diabetes mellitus has a class of recommendation IIb with level C of evidence(34).
The content of this article reflects the personal opinion of the author/s and is not necessarily the official position of the European Society of Cardiology.