Radial artery versus saphenous vein as a second coronary bypass conduit in septuagenarians

The recent article by Hayward and colleagues [1] reported mid-to-late RAPCO trial survival in septuagenarians. They found that patients who receive an internal-thoracic artery to the left anterior descending (LITA-LAD) had similar 10-year survival when the next best target is grafted with radial artery (RA) or saphenous vein (SVG). This result diverges from those in a related retrospective propensity-matched analysis where RA and SVG were compared as the second conduit of choice in LITA-LAD CABG [2]. There, RA was associated with significantly improved 6-year survival particularly after the third postoperative year, and this was true for both the younger and older sub-cohorts [2]. These discordant findings [1,2] may derive from more restrictive inclusion/exclusion criteria of the RAPCO trial [1] compared to a large all-inclusive retrospective series. We wish to communicate a number of observations that may elucidate further these divergent findings: The RA and SVG PAPCO groups’ number of grafts was similar at 3.2 and 3.3 grafts [1] yet no information on arterial versus venous use for the third and fourth grafts is given. The authors are leaders of RA use and given their propensity to maximize arterial grafts, it is likely that the total arterial grafts were similar for the SVG and RA groups. If true, this may explain the similar survival data. Multiple groups have reported that increasing arterial grafts results in improved survival. It is safe to assume that most to all patients had multi-vessel coronary disease. But, it is conceivable that the two- versus three-vessel disease may have differed for the study groups (data not provided), providing this serves as a surrogate measure of completeness of revascularization, a well-accepted determinant of mid-to-late CABG outcomes. The authors’ primary criterion for the ‘next best target other than the LAD’ for study conduit placement was the coronary target vessel size. Hence, because of the usual greater probability of right coronary dominance, it is quite likely that a disproportionately higher number of the study conduits (SVG or RA) may have been used to revascularize right coronary artery (RCA) targets as opposed to LAD diagonals or branches of the circumflex system. Ironically, multiple authors have reported that RA graft patency is generally worse when placed to the RCA even when ≥70% stenosis is present (e.g., [2,3]). Critically, this considered with point II above, puts forth the possibility that the SVG group may have had more radial grafts placed to targets where it is associated with superior patency when compared to the RA group. Such a scenario has significant implications on the survival comparison, and a more complete description of the overall RA and SVG graft placement is needed to allow a better interpretation of this study’s findings. We noted that RA had a larger number of protocol breaches compared to SVG (9 vs 2) that are uncorrected given the intent-to-treat analysis. Given the small sample size and low overall adverse events rate, it is helpful if a description of these events in breaches (if any) is provided, or if a complementary comparison with breaches removed is shown. We congratulate the authors on this thought-provoking paper and look forward to their response.