Page 94 - Read Online

Page 10 of 18 Browne et al. Vessel Plus 2024;8:19 https://dx.doi.org/10.20517/2574-1209.2023.126

At present, conflicting evidence from mostly single-center observational studies suggests similar overall
rates of graft failure after skeletonized and pedicled IMA harvesting. However, recent reports of higher rates
of graft failure and increased risks of major adverse cardiovascular events associated with the use of
skeletonized IMA, warrant a more thorough evaluation in randomized trials.

Proximal RIMA graft configuration (in situ, free, or composite)
In situ vs. free
The RIMA is commonly used without removal from their proximal origin as an “in situ” graft (no proximal
anastomosis) or removed from their origin and grafted proximally to the aorta as a “free” graft or to another
graft, often as a Y-shaped “composite” graft (e.g., onto LIMA, radial artery, or saphenous vein grafts)
[Figure 2]. In a single-center retrospective observational study of 7,092 patients undergoing bilateral
internal mammary artery (BIMA) grafting with enrollment from 1972 to 2016, patient-level RIMA failure
rates after 15 years were: 9% in situ, 9% free from aorta, 11% composite from LIMA, and 23% composite
from saphenous vein grafts. The most important risk factor for graft failure was target vessel location (i.e.,
RIMA-to-LAD had lower rates of failure than RIMA to diagonal, left circumflex, or right coronary arteries).
After adjusting for target vessel location, the risks of graft failure with in situ vs. free RIMA were similar
regardless of the type of non-in situ configuration (i.e., free graft from aorta or composite graft from LIMA
or vein), leading the authors to conclude that long-term rates of RIMA graft failure were independent of
proximal configuration.

In our analysis of the COMPASS CABG dataset, we found that in situ RIMA failed more than twice as
frequently as RIMA from the aortic or composite configuration when grafted to the left circumflex artery,
although the difference was not significant (63% [10/16] vs. 24% [4/17]; odds ratio: 0.20; 95%CI: 0.03-1.03;
[7]
P = 0.054) .
In a single-center retrospective observational study of BIMA grafting in 5,766 patients (enrollment 1986-
2008), 10-year rates of graft failure were similar with in situ RIMA compared with free RIMA (11% [50/450]
vs. 9% [49/541]) . When RIMA was grafted specifically to the left circumflex artery, 13% (10/76) failed
[55]
using in situ compared with 6% (17/276) using a free RIMA graft. When grafted to the right coronary artery,
failure rates were higher with in situ compared with free RIMA grafting (26% [37/141] vs. 7% [4/58];
P = 0.02). We speculate that the high failure rate observed with in situ RIMA may be attributed to challenges
with reaching distant target vessels (e.g., inappropriate tension, mismatch in size between graft and the
target vessel, potential kinking along lengthy grafts, etc.).

Similar results were observed in a more recent single-center retrospective observational study of 282
patients who underwent CABG with BIMA grafting (enrollment 2000-2012), where 69 patients received an
in situ RIMA and 213 patients received a free RIMA grafted to the left circumflex territory to supplement
LIMA-LAD grafting . 5-year rates of patient-level graft failure, estimated from Kaplan-Meier curves, were
[36]
higher with in situ RIMA compared with free RIMA (19.7% vs. 3%, respectively; P = 0.01). Current
observational evidence suggests that a free proximal grafting configuration may be preferable to the in-situ
configuration for RIMA grafting to non-LAD target vessels. However, we note that the confidence in this
assertion should reflect the poor quality of the evidence.

In situ vs. composite
The use of the RIMA as a Y-composite, rather than in situ graft, facilitates sequential grafting (i.e., multiple
distal anastomoses) of the RIMA. As a result, complete coronary revascularization can more often be
achieved using BIMA without the need for additional conduits. In a single-center randomized controlled