Page 87 - Read Online

Browne et al. Vessel Plus 2024;8:19 https://dx.doi.org/10.20517/2574-1209.2023.126 Page 3 of 18

Table 1. Graft failure rates from pooled and post hoc analyses of LIMA, radial artery, saphenous vein and RIMA grafts
Graft failure - n/N (%)
Included Saphenous vein Time of imaging
Study
trials Radial No- (years)
LIMA Conventional RIMA
artery touch
Individual patient meta-analyses
Gaudino et al., 7 387/4,006 21/152 (13.8) 172/8,740 (19.7) 61/265 1.02 (1.00, 1.03)*
[13]
2023 (9.7) (23.0)
Gaudino et al., 6 21/921 (2.3) 67/710 (9.4) 101/576 (17.5) 10/74 (13.5) 5.4 ± 2.4 †
[4]
2020
Network meta-analyses
[27] ‡
Deng et al., 2022 18 - 5.9% 8.6% 13.7% 10.8% 3.5 (1.5-5.4)
(2.4-10.0) (5.7-12.7) (9.8-18.8) (3.5-28.8)
Post hoc analysis
Alboom et al., 1 68/1,068 (6.4) 9/91 (9.9) 232/2239 (10.4) 22/82 (26.8) 1.13 ± 0.30
[7]
2022
† ‡
*Refers to median (IQR); refers to mean ± SD; refers to weighted mean (95%CI). CI: Confidence interval; LIMA: left internal mammary artery;
RIMA: right internal mammary artery.

were early single-center trials (i.e., few surgeons from high-volume centers), whereas all 7 trials included in
the 2023 analysis were multi-center trials (i.e., many surgeons from lower-volume centers). A volume-
outcome relationship has been reported for BIMA grafting using meta-regression, with centers performing
higher proportions of BIMA grafting associated with reduced long-term mortality . Therefore, the
[26]
discrepancy in rates of graft failure lends some support to the idea that a larger gradient in center or surgeon
experience exists in multi-center CABG trials, where results of surgeons pooled from around the world
(although more generalizable) are likely to be worse than results of surgeons from high-volume single-
centers often pioneering arterial grafting.

Network meta-analyses
An update was recently reported for a network meta-analysis (Deng et al.) of 18 randomized trials that
included 6,543 patients and 8,272 grafts with the aim of determining the second best conduit in CABG
surgery based on rates of graft failure [5,27] . After a mean angiographic follow-up of 3.5 years, rates of graft
failure were % (95%CI): 5.9% (2.4-10.0) for radial artery, 8.6% (5.7-12.7) for no-touch saphenous vein, 10.8%
(3.5-28.8) for RIMA, and 13.7% (9.8-18.8) for conventionally harvested saphenous veins. Compared with
conventionally harvested veins, rates of graft failure were lower using radial artery (incidence rate ratio,
0.56; 95%CI: 0.43-0.74) and no-touch saphenous veins (incidence rate ratio, 0.56; 95%CI: 0.44-0.70) but not
RIMA (incidence rate ratio, 1.06; 95%CI: 0.72-1.54). Similar results were reported in another network meta-
analysis (Yokoyama et al.) of 13 RCTs (3,728 patients and 2,773 grafts) indicating lower rates of graft failure
with radial artery and no-touch saphenous veins compared with conventionally harvested veins at
maximum angiographic follow-up and in sensitivity analyses restricted to trials with ≥ 3 and ≥ 5 years of
follow-up . These results suggest that the radial artery and no-touch saphenous veins may both be
[28]
considered the best second conduit for minimizing graft failure over the medium to long term.

Post hoc analysis
We recently completed a post hoc analysis (Alboom et al.) of the Cardiovascular Outcomes for People
[6]
[7]
Using Anticoagulation Strategies (COMPASS) CABG study . The analysis included 1,068 patients
(3,480 grafts) who underwent CABG surgery and had complete angiographic follow-up at 1 year. The
COMPASS CABG study was one of few that directly compared angiographic results of all the common
CABG conduits (LIMA, RIMA, radial artery, and saphenous vein) within a single study using systematic