DISCUSSION
HCR has not, as yet, become widely employed as a mainstream option for treating coronary artery disease, particularly for those patients who have more than just an isolated LAD lesion. Published data on HCR is relatively limited, with the majority of studies thus far being primarily single-center, non-randomized reports, incorporating only small to moderate patient numbers, with few exceptions (18, 19). At present, no randomized trials specifically compare HCR with CABG. To the best of our knowledge, there is only one other study to analyze propensity-matched data, comparing HCR to both CABG and multi-vessel PCI in patients with TVD (20). However, our study is thus far the only propensity-matched study comparing long-term clinical endpoints for all three revascularization methods.
For patients undergoing HCR, we prefer a MIDCAB-first approach, followed by interval PCI, typically within 4 to 6 weeks of surgery. This allows the surgical revascularization to be performed without concern for potential bleeding that may be associated with the dual antiplatelet therapy that is mandatory after PCI with DES. We have, nevertheless, previously described our satisfactory experience with MIDCAB in patients taking dual antiplatelet agents (21), results that have not necessarily been consistent amongst all surgical groups (22). More importantly, however, the MIDCAB-first approach allows the patency of the LIMA to LAD bypass graft to be interrogated during the subsequent PCI procedure.
A PCI-first HCR strategy was pursued in those patients who presented with a coronary syndrome in which the culprit lesion was deemed to be within one of the non-LAD vessels, or in those individuals in whom the angiographic severity and clinical import of at least one of the non-LAD stenoses was thought to be greater than that of the disease within the LAD itself. For these patients, subsequent LIMA to LAD grafting was undertaken on uninterrupted dual antiplatelet therapy.
Of the 100 patients in our final study population who underwent HCR, 72 had MIDCAB followed by PCI, whereas the remaining 28 had PCI prior to MIDCAB. We did not perform any cases of simultaneous MIDCAB and PCI, although this approach has been adopted by other investigators (23, 24). None of our MIDCAB patients required intraoperative conversion to sternotomy.
The primary end-point of our study, all-cause mortality at 8 years, was not significantly different across the three revascularization cohorts. Freedom from MACE at 8 years was also comparable between patient groups. However, despite having a higher SYNTAX score at baseline, those patients that underwent HCR ended up with a lower residual SYNTAX score did those who had multi-vessel PCI, reflective of a significant difference in the completeness of revascularization achieved. By contrast, HCR patients had similar residual SYNTAX scores to those patients managed by traditional CABG. Our data also shows that early complications from HCR were infrequent. The HCR cohort did benefit from a shorter length of hospital stay than did the CABG group.
The mean baseline SYNTAX score in our HCR cohort was 28.9 ± 10.6. In the SYNTAX trial, patients with TVD with similar, intermediate, SYNTAX scores demonstrated no significant difference in the incidence of MACE at 12 months, when the PCI group was compared to CABG (1). When this group was of patients was followed for 5 years, MACE rates were indeed found to be higher in the PCI cohort, but there was no significant mortality difference between treatment groups (25). Similarly, in diabetic patients with intermediate SYNTAX scores, the FREEDOM trial found that the composite end-point of death, MI and stroke at 5 years favored CABG over PCI, again with no significant mortality difference identified (2). These results suggest that, in the TVD patient population with intermediate SYNTAX scores, although mid-term survival is comparable across treatment arms, morbidity may be higher after PCI, particularly with respect to the increased incidence of repeat revascularization and new myocardial infarction. We suggest that it may be this group of patients that may be ideally suited to benefit from HCR.
The longer-term overall mortality rates in our patient groups were slightly lower than that observed in comparable individuals in some of the larger trials. It is our hypothesis that these low mortality rates are attributable, at least in part, to the low residual SYNTAX scores that were able to be achieved in each patient group, regardless of the technique of revascularization. There is good data to support the assertion that the completeness of revascularization (as quantified by the residual SYNTAX score) is a predictor of short-to-medium term MACE in patients with multi-vessel coronary artery disease undergoing PCI or CABG (26 – 28). Our group has previously reported that a low residual SYNTAX score after HCR was associated with 8-year survival that was not significantly different than that seen after traditional CABG in patients with double-vessel coronary disease (29). Our current results suggest that HCR is also associated with favorable long-term results in patients with TVD, provided that the operators can achieve a low residual SYTNAX score.
This study is also one of the first thus far to demonstrate that HCR may be considered a reasonable method of revascularization in TVD patients with left main-stem involvement. Of the 24 patients with left main coronary disease who underwent HCR, there was only 1 death (4.2%) at long-term follow up.
Sixteen patients who underwent HCR required repeat revascularization by 8-year follow-up. Of these, only 6 (37.5%) actually required a ‘target vessel revascularization’. The remaining 10 patients (62.5%) all required intervention for de-novo lesions. In those 6 HCR individuals that required target vessel revascularization, 3 procedures were performed due to problems with the LIMA to LAD graft, whereas 3 interventions were undertaken in non-LAD vessels. The incidence of target vessel revascularization was not significantly different between the HCR, CABG and multi-vessel PCI groups; in the CABG population, 2 patients required a re-intervention for the LIMA to LAD graft.