STUDY LIMITATIONS
Our study has several limitations to be acknowledged. Our data was derived from a retrospective review in a single institution. A multi-center approach, incorporating a larger sample size, would be more adequately powered to analyze more subtle differences between outcomes. Our patients were not randomized to a given revascularization strategy. Although the impact of this may be somewhat mitigated by the fact that the treatment for each patient was not selected by a single practitioner, but rather by a multidisciplinary heart team, the fact that the overwhelming majority of patients with TVD in our institution were treated by either CABG or multi-vessel PCI reflects an intrinsic selection bias for those individuals offered HCR.
Due to the differences in SYNTAX scores between patient groups, our data cannot be used to comment on the superiority of one revascularization strategy over another. The data does, however, afford us the opportunity to evaluate the effectiveness of a ‘real-world’ heart team approach to the management of patients with TVD. Residual SYNTAX score calculations in our CABG and PCI-first HCR patients assumed patency of all surgical bypass grafts; although suboptimal, it was not deemed appropriate to undertake formal angiography in asymptomatic patients for the sole purpose of calculating a residual SYNTAX score.
We were able to achieve 100% follow up for our primary end-point, namely, all-cause death at 8 years. We utilized the National Death Index as our primary source for 8-year survival data; though this methodology may somewhat underestimate true mortality (30). Patients that underwent multi-vessel PCI had a numerically higher amount of death than CABG or HCR, however this did not meet significance. We cannot exclude the possibility of a type II error, as our study size was limited. With respect to the other long-term outcomes that were reported, it is important to note that there were 15 patients in the CABG arm, 5 in the PCI group, and 3 in the HCR cohort, that were all lost to follow-up. Furthermore, we are unable to correct for the use of DAPT which may confound the long-term results of MACE.
Our institution has extensive experience with minimal-access, robotic-assisted LIMA to LAD grafting, as well as with complex PCI. Our results may, therefore, not be directly generalizable to other institutions who have less experience with either revascularization strategy. Similarly, the low perioperative stroke rate observed in our sternotomy CABG cohort may not necessarily be translatable to other patient populations, given that we perform the majority of these cases ‘off-pump,’ a practice pattern not currently adopted by most surgical centers in the United States (31). Despite our center’s surgical experience, only 100 patients were treated with HCR, which reflects our selectiveness of the revascularization strategy.