Between 2012 and 2022, 4355 patients underwent CABG, with 94 (2.2%) receiving clinically indicated CAG before discharge. CAG was unsuccessful in 2 patients, resulting in 92 patients (79.3% male, 66 ± 10 years old) being included in the study.

Symptomatic early CGF was identified in 55 patients (59.8%), with a total of 81 findings. The most frequent finding was stenosis of the distal anastomosis (n = 27, 29.3%), followed by occlusion (n = 18, 19.6%), kinking (n = 13, 14.1%), stenosis of the bypass conduit (n = 9, 9.8%), reduced bypass flow (n = 7, 7.6%), stenosis at the level of the Y‑graft anastomosis (n = 3, 3.3%), bypass thrombosis (n = 2, 2.2%) and tenting due to a too short graft (n = 2, 2.2%) (Fig. 1; Tab. 1). Among the Y‑grafts, 12 patients (46.2%) had problems located at the distal anastomosis, with problems predominantly affecting vessels revascularising the anterior and/or lateral territory (n = 18, 69.2%).

Causes of symptomatic early coronary graft failure in patients undergoing clinically indicated in-hospital coronary angiography after coronary artery bypass graft surgery

Baseline characteristics were comparable across both groups (Tab. 2). Most patients had three-vessel disease (n = 56, 60.9%), good baseline LV function (n = 57, 62.0%), and underwent elective surgery (n = 60, 65.2%) mostly after non-ST-elevation myocardial infarction (n = 25, 27.2%) and with a median EuroSCORE II of 1.9 (1.2–4.0).

Procedural parameters were also mostly similar (Tab. 3). The left internal mammary artery (LIMA) alone was used in 9 cases (9.8%), while both LIMA and right internal mammary artery (RIMA) were used in 26 (28.3%). Among these, 5 (5.4%) were in situ and 21 (22.8%) had a Y-graft configuration. For arterial plus venous grafts, LIMA and saphenous vein graft (SVG) were used in 45 cases (48.9%) and LIMA, RIMA and SVG in 5 (5.4%). Of these, LIMA-SVG(Y) was used in 4 (4.3%) cases and a LIMA-RIMA(Y) and SVG in 1 case (1.1%). Concomitant surgical interventions were performed in 32 patients (34.8%). A Y-graft configuration was found significantly more often in patients with symptomatic early CGF (36.4% vs 16.2%; p = 0.035).

Potential indications for CAG included new ECG abnormalities (n = 78, 84.8%), echocardiographic changes (n = 26, 28.3%), biomarker elevation (n = 81, 88%) and prolonged inotropic, vasopressive and/or mechanical support (n = 40, 43.5%) (Tab. 4). There were no significant differences between patients with and without symptomatic early CGF in overall ECG abnormalities (p = 0.103), ischaemic changes (p = 0.196), rhythm disorders (p = 0.389) and conductions disorders (p = 0.514). Postoperative TTEs were performed in an almost equal number of cases (48.6% vs 47.3%; p = 0.897) with a similar rate of new RWMA (27.0% vs 29.1%; p = 0.691). In addition, cardiac biomarker elevation was comparable (86.5% vs 89.1%; p = 0.706), with no significant differences in median peak TnT, peak CK, ∆TnT or ∆CK (Tab. 4). Haemodynamic status 12 h post-CABG was similar, with prolonged support observed in 40.5% versus 45.5% of patients (p = 0.641).

At the time of in-hospital recatheterisation, 42 (45.7%) patients underwent revascularisation with a total of 47 interventions performed (18.9% without early CGF vs 63.6% with early CGF) (Table S1, Electronic Supplementary Material). Overall, PCI was the preferred treatment option (n = 38, 80.9%). In subgroup comparisons, PCI was used particularly in bypass (n = 10, 55.6%) or anastomotic stenosis cases, including distal (n = 15, 55.6%) and Y‑graft anastomoses (n = 2, 66.7%). Re-CABG was used mainly in more complex cases like bypass kinking (n = 5, 41.7%). Conservative management was applied in cases with limited revascularisation options or milder failures such as reduced bypass flow (n = 4, 54.1%) (Tab. 1).

The median follow-up after in-hospital recatheterisation was 33 (11–60) months for patients with symptomatic early CGF and 60 months (38–60) for patients without symptomatic early CGF (p = 0.005). Five-year follow-up data were available in 54 patients (58.7%), with MACE observed in 26 (28.3%), of whom 20 (76.9%) had early CGF. Patients with symptomatic early CGF showed a significantly lower MACE-free survival rate over a median follow-up period of 33 months (p = 0.023), although all-cause mortality did not differ significantly (p = 0.255) (Fig. 2). Multivariate analysis showed that venous graft integration (p = 0.005), Y‑graft configuration (p = 0.002) and prolonged inotropic use (p = 0.032) were significantly associated with symptomatic early CGF (Table S2, Electronic Supplementary Material).

Event-free follow-up. Event-free survival of patients with (blue line) and without (red line) symptomatic early coronary graft failure. MI myocardial infarction