The prognostic impact of unplanned invasive coronary angiography following coronary artery bypass grafting

Abstract

Open in new tabDownload slide

OBJECTIVES

Myocardial ischaemia following coronary artery bypass grafting (CABG) is a potentially devastating complication. Nevertheless, the incidence, aetiology and prognostic relevance of unplanned coronary angiography (uCAG) remain understudied. We aimed to investigate the prevalence and outcome of patients undergoing urgent, uCAG in the postoperative period following CABG.

METHODS

We screened all patients undergoing isolated elective CABG in an academic referral centre between 2016 and 2021 and identified patients undergoing uCAG within 30 days of surgery. For uCAG patients, a distinction was made between patients undergoing re-revascularization (REV) and patients receiving conservative management (CON). The primary outcomes were 30-day mortality and unadjusted and adjusted long-term survival. Secondary outcomes were the indication for and prevalence of uCAG and urgent revascularization.

RESULTS

Of the 1918 patients undergoing isolated CABG, 78 individuals needed uCAG (4.1%), of whom 45 underwent immediate revascularization (REV group; 2.3% overall, 57% within the uCAG group, median age 69.9 years) and 33 were treated conservatively (CONS group; 1.7% overall, 42% within the uCAG group, median age 69.1 years). Patients undergoing uCAG (n = 78) had a higher 30-day mortality than patients not undergoing uCAG (n = 1840, 30-day mortality: 9.0% vs 0.4%, P < 0.001). Long-term survival was significantly decreased in patients undergoing uCAG in both unadjusted (hazard ratio 2.20, 95% confidence interval 1.30–3.73) and EuroSCORE-, age- and sex-adjusted models (hazard ratio uCAG 2.03, 95% confidence interval 1.16–3.56).

CONCLUSIONS

Unplanned postoperative coronary angiography is performed in 4.1% of isolated CABG procedures, and patients in need of such urgent invasive evaluation are subjected to decreased short- and long-term survival.

INTRODUCTION

Although elective coronary artery bypass grafting (CABG) is generally associated with favourable outcomes, a subset of patients requires unplanned postoperative coronary angiography (uCAG) after which urgent re-revascularization may be necessary. The main indication for uCAG is suspected periprocedural myocardial ischaemia (PMI) [1].

Recent reviews and expert consensus statements underscore the complexities in distinguishing between patients at risk of PMI and those with uncomplicated courses [2–4]. Indeed, PMI has a markedly varying aetiology, which can be divided into irreversible and reversible causes. However, the definition of PMI as a diagnostic entity varies [5], and several contradicting international and intersocietal definitions have been proposed [6–8]. When PMI is suspected, coronary angiography (CAG) is indicated to distinguish between irreversible and reversible causes and to potentially intervene therapeutically.

In the current study, we aim to examine the prevalence, underlying reasons and prognostic relevance of uCAG and urgent revascularization following CABG in patients with suspected PMI.

MATERIALS AND METHODS

Study design and ethical approval

The institutional review board approved the ethical protocol of the study and waived the need for informed consent due to its observational nature (METC 2022-3137, 1 February 2022). We adhered to the STROBE guidelines (please see Supplementary Material checklist).

Data collection

Patient and procedural characteristics were retrieved from our institutional database between 1 January 2016 and 31 December 2021. All electrocardiography (ECG) examinations were re-evaluated, in conjunction with serial biomarker measurements (MB-isoenzyme of creatine kinase, Roche Cobas mass assay, Roche Diagnostics, Basel, Switzerland, with a reference concentration of <4.9 µg/l). Finally, vital status was collected until 26 April 2024 (and for 94.0% complete, n = 114 patients with missing follow-up data).

Patient selection

Patients were included in the current study when undergoing isolated coronary surgery, either elective or urgent. Patients were excluded when undergoing emergency surgery or in the presence of preoperative mechanical support devices. Finally, patients undergoing hybrid revascularization procedures [i.e. ‘planned’ postoperative CAG to perform hybrid percutaneous coronary intervention (PCI) after successful grafting of the left anterior descending artery (LAD)] were excluded from the uCAG group and allocated to the reference group.

Study groups

The study group comprised of patients undergoing a uCAG (in the local centre or in referring centres, for which the data were retrieved), within 30 days after CABG. These uCAG patients were then categorized into a group undergoing immediate revascularization (REV group) or conservative treatment (CONS group). Patients not undergoing uCAG were used as a reference group.

Procedures

Surgical revascularization is performed using 3 approaches in our institution: on-pump multivessel CABG with cardioplegic arrest, off-pump CABG (OPCAB) and minimally invasive direct CABG (MIDCAB, off-pump, robotic-assisted). Multivessel CABG is performed in eligible patients with multivessel disease. OPCAB is performed in patients with an urgent indication for surgery with single-vessel disease, or in patients with multivessel disease who are ineligible for cannulation and/or aortic clamping. Finally, MIDCAB is reserved for elective patients with single-vessel disease (Supplementary Material S1).

Outcomes and definitions

The primary outcomes were 30-day mortality and unadjusted and adjusted long-term survival of uCAG patients and the reference CABG group. The secondary outcomes were the indication for and prevalence of uCAG and immediate re-revascularization. Definitions are presented in Supplementary Material S1.

Indication for uCAG

Patients who were postoperatively admitted to the intensive care unit underwent immediate ECG evaluation. If ischaemic changes suggestive of transmural ischaemia were present on ECG (ST-segment elevation or de novo left bundle branch block), uCAG was performed. In case of non-specific findings [such as ST-segment depression (not being reciprocal depression), T-wave inversion or in the absence of ECG abnormalities], biomarkers were awaited. In general, an absolute MB-isoenzyme of creatine kinase threshold of >100µg/l (corresponding to >20 times the reference value) would prompt angiographic evaluation, in agreement with a recent expert consensus statement [3].

Statistical analysis

The distribution of continuous variables was assessed using the Shapiro–Wilk or Kolmogorov–Smirnov test, depending on the sample size. Then, Student’s t-tests or Mann–Whitney U tests were applied to continuous variables, depending on the distribution, and chi-square tests or Fisher’s exact text (in case of count <5) to categorical variables.

The prognostic relevance of uCAG was explored in unadjusted and European System for Cardiac Operative Risk Evaluation (EuroSCORE)-, age- and sex-adjusted Cox-regression models, expressed in (adjusted) hazard ratios (HRs) with corresponding 95% confidence intervals (CIs). For sensitivity purposes, landmark analysis was performed before and after the timepoint of 30 days. The log-rank test was applied to compare the survival of patients undergoing re-CABG, PCI and conservative treatment within the uCAG cohort. For additional sensitivity analyses, patients undergoing uCAG were divided as having no abnormalities (no new occlusions/stenoses) or having new abnormalities (graft and/or anastomotic failure, or new native coronary stenosis/occlusion). Finally, a propensity score-matched sample of patients was compared to confirm the robustness of results (one-to-one matching, calliper set at 0.01).

Data analysis was performed using IBM SPSS Statistics version 28 (IBM, Armonk, USA) and the R statistical environment (R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Study groups

Out of 1918 patients undergoing isolated non-emergency CABG in the study period, 95 patients underwent postoperative CAG within 30 days after surgery, of which 17 CAGs were planned preoperatively (hybrid revascularization; LAD grafting followed by ‘planned’ PCI of non-LAD targets), and 78 patients (4.1%) underwent ‘unplanned’ (urgent) coronary angiography (uCAG, the study group). Consequently, the reference group comprised 1840 patients.

As predefined, uCAG patients were divided into conservatively treated patients (CONS group, n = 33, 1.7% overall, 43% within uCAG patients) versus patients undergoing revascularization (either PCI or re-CABG, REV group, n = 45, 2.3% overall, 57% within uCAG patients). Within the latter group, re-CABG was performed in 12 patients (26.7%) and PCI in 33 patients (73.3%, Fig. 1).

Baseline characteristics

Table 1 presents the baseline characteristics of the reference group (not undergoing uCAG, n = 1840) and the uCAG group (n = 78, for which a subdivision in the CONS and REV group is presented in Supplementary Materials S2 and S3). There were no statistically significant differences in baseline characteristics between the reference and uCAG group.

Procedural characteristics

Table 2 presents the procedural characteristics of the reference and uCAG groups. Within the uCAG group, 51 patients underwent conventional on-pump CABG (65.3%), 20 patients underwent MIDCAB (25.6%) and 7 patients underwent OPCAB (9.0%, Table 2, further details in Supplementary Material S3).

Indication and timing of unplanned coronary angiography

The primary indications for uCAG were ischaemic ECG changes, notably ST-segment elevation or de novo left bundle branch block, representing 75.6% (n = 59) of the uCAG cohort, without a statistically significant difference between the CONS (75.7%) and REV (75.5%) groups (P = 0.984, Table 3).

Other uCAG reasons comprised isolated increases of cardiac biomarker concentrations (CONS: 18.1%, n = 6 versus REV: 6.6%, n = 3, P = 0.116) and ventricular arrhythmia (CONS: 3.0%, n = 1 versus REV: 11.1%, n = 5, P = 0.856). uCAG was performed after a median of 0 days [0–1.0] and 0 [0–2.5] days in the CONS and REV groups, respectively (P = 0.705, Table 3).

Outcomes of unplanned coronary angiography

In uCAG patients (n = 78), graft and/or anastomotic failure was observed in 48 patients (61.5% in uCAG, CONS 27.3%, REV 86.7%, P = 0.001), a new native coronary stenosis or occlusion in 10 patients (all: 12.7%, CONS: 15.2%, REV: 11.1%, P = 0.442) and no evidence of a new occlusion was observed in 19 patients (24.3%, CONS: 57.6%, REV: 0, P = 0.001, Table 4, 1 patient with missing data). Supplementary Material S4 presents the specific findings per subcategory. In addition, Supplementary Material S5 presents the procedures performed in the PCI and re-CABG groups (only 1 affected vessel was treated in all cases).

Primary outcome: clinical and prognostic relevance of unplanned coronary angiography

Thirty-day mortality in the overall cohort was 0.7% (14/1918). Patients undergoing uCAG had a significantly higher 30-day mortality as compared to patients not undergoing uCAG [0.4% (n = 7) vs 9.0% (n = 7), P < 0.001]. Cause-specific 30-day mortality is reported in Supplementary Material S6.

Median follow-up for survival was 4.6 years (interquartile range 3.4–5.9 years). Overall, the 5-year survival of the entire CABG cohort was 90.3% (95% CI 88.7–91.9%). Survival was significantly decreased in patients undergoing uCAG (HR 2.20, 95% CI 1.30–3.73, P = 0.003, Fig. 2A). One-, 3- and 5-year survival were reported in Fig. 2A and Supplementary Material S7. The uCAG association for survival persisted after adjustment for EuroSCORE II, age and sex (HR uCAG 2.03, 95% CI 1.16–3.56, P = 0.013).

In an exploratory landmark analysis before and after the timepoint of 30 days, we found that the difference between uCAG and the reference group was primarily present during the first 30 days (univariable HR uCAG first 30 days: 23.80, 95% CI 8.35–67.86, P < 0.001, multivariable HR uCAG first 30 days: 20.36, 95% CI 6.73–61.64, P < 0.001), after which the association between uCAG and survival was attenuated (univariable HR uCAG beyond 30 days: 1.23, 95% CI 0.60–2.50, P = 0.569, multivariable HR uCAG beyond 30 days: 1.15, 95% CI 0.55–2.42, P = 0.713, Fig. 2B).

Sensitivity analyses

Nineteen patients had no abnormalities upon uCAG (Table 4), while 59 patients had evidence of new stenoses (either native coronary or graft/anastomotic stenosis/occlusion). Long-term survival of patients without abnormalities on uCAG did not differ from the reference group (HR 1.03, 95% CI 0.26–4.17, P = 0.963), while the survival decrease was even more pronounced in patients with new abnormalities upon uCAG (HR 2.66, 95% CI 1.51–4.73, P < 0.001).

Propensity score matching yielded 76 pairs of patients, without statistically significant baseline or procedural characteristics (Supplementary Material S8). Thirty-day mortality was persistently increased in the uCAG group [0 vs 6 (7.9%), P = 0.028, Supplementary Material S8]. Overall long-term survival yielded similar results as the overall sample, although this difference did not reach statistical significance (Supplementary Material S9).

Clinical and prognostic relevance of therapeutic strategy

No statistically significant difference in 30-day mortality was observed between the CONS and the REV groups [CONS: 6.0% (n = 2), REV: 11.1% (n = 5), P = 0.692]. There was no statistically significant difference in the long-term survival between patients in the CONS and REV groups (log-rank P-value = 0.125, Fig. 3), but these analyses may have been underpowered.

DISCUSSION

The current study was the first to investigate the long-term impact of uCAG following CABG. Our main findings are that unplanned postoperative invasive coronary angiography is performed in 4.1% of patients undergoing CABG, and 57% of these patients undergo immediate rescue revascularization, either through PCI or re-CABG, for which graft failure is the main indication. In our all-comer CABG patient cohort, with a mean 5-year survival rate of 90.3%, patients undergoing postoperative uCAG are subjected to increased short-term mortality and decreased long-term survival as a result of this early risk, reflecting the dismal prognosis associated with PMI, for which uCAG can be considered a surrogate. These findings highlight the need for further research and interventions to improve detection of periprocedural ischaemia and reduce the risk of mortality.

Omran et al. [9] reported a postoperative uCAG rate of 3.4% in a contemporary cohort of CABG patients, while Alqahtani et al. [1] found a uCAG prevalence of 4.4%, both of which seem in line with our findings. Remarkably, a recently published registry—based on the Virginia Cardiac Services Quality Initiative—found a surprisingly low 0.4% uCAG rate in over 10 000 patients undergoing CABG between 2018 and 2021 [10]. This disparity highlights the difficulty in studying these patient groups, as the varying indications to perform uCAG may confound findings and conclusions. For example, the former 2 studies by Omran et al. and Alqahtani et al. identified an independent association between uCAG and 30-day mortality, while the study incorporating VCSQI data did not. Given the exceptionally low rate of uCAG in the latter study, the absence of a difference may be the consequence of a relatively low event rate (i.e. the power of the study), or because of an inappropriately high threshold to perform uCAG. Nevertheless, none of these studies assessed the impact of uCAG on long-term outcomes beyond 30 days.

In our cohort, the main indications for performing uCAG were ischaemic ECG abnormalities (>75% of uCAGs), followed by isolated biomarker concentration increases and haemodynamic instability secondary to ventricular arrhythmia. These findings show that decision-making mimics ‘standard’ acute cardiac care, and underline the importance of ECG. Still, the rescue revascularization rate in patients with such true ischaemic ECG changes (ST elevation and/or left bundle branch block de novo) was ‘only’ 58% (34/59 patients), while 42% of patients were managed conservatively. Although ECG abnormalities in the absence of angiographic findings may be secondary to inadequate myocardial protection, air embolism and/or distal embolization, a conservative management was sometimes even indicated despite proven ischaemia and graft- or anastomotic failure. These treatment outcomes are in sharp contrast with the management of contemporary ‘primary’ ST-elevation myocardial infarction (STEMI) patients, in which acute PCI is performed in >90% of patients and in up to 97% of STEMI patients in Western Europe [11]. The explanation for this difference may lie in the difference between aetiologies. In primary STEMI patients, STEMI usually originates from a thrombotic vessel occlusion secondary to plaque rupture, for which PCI is an excellent treatment. In contrast, STEMI directly following CABG may be caused by technical and mechanical issues as well [12]. For example, if anastomotic graft failure as a result of suture displacement leads to a total occlusion of a vessel, PCI is futile and re-CABG may be indicated.

In addition, the indication for reintervention and the choice of treatment modality may differ between LAD targets and non-LAD targets. In our cohort, OPCAB and MIDCAB were primarily performed in patients with single-vessel disease (i.e. LAD disease, for which a left internal mammary artery (LIMA)–LAD was performed, Supplementary Material S1). Based on the prognostic relevance of a patent LIMA–LAD anastomosis, the threshold to perform uCAG was relatively low, as any suspicion of ischaemia would point to a potential LIMA–LAD problem. Consequently, the threshold to perform uCAG may have been lower in patients undergoing OPCAB and MIDCAB in our centre. Moreover, in the case of confirmed graft failure to the LAD, reintervention is definitely indicated and should be performed.

The 30-day mortality was 0.4% in the reference group, while it was significantly increased in patients undergoing uCAG (7/78, 9.0%). This difference confirms that patients in need of uCAG constitute a high-risk group, in line with previous studies that found such patients to have an increased risk of morbidity and mortality [1]. This may be the consequence of an elevated baseline risk, but this explanation was refuted by Alqahtani et al. [1] in adjusted analyses with correction for clinically relevant preoperative parameters. As such, uCAG should be seen as a marker of a high-risk population, where it can be assumed that the mortality rate is the consequence of periprocedural ischaemia itself, previously associated with low cardiac output syndrome and postcardiotomy shock, the use of temporary mechanical circulatory support and all their related complications. Furthermore, both postoperative PCI and re-CABG carry substantial procedural risks and may predispose to additional complications in an already weakened patient.

Limitations

First, this was a retrospective analysis. Second, uCAG can only be seen as a surrogate for PMI, and does not necessarily imply that ischaemia was actually present. Third, different procedures were performed (CABG, OPCAB, MIDCAB), which may have affected outcomes, although patients were matched for procedures in sensitivity analyses. Then, the indication for uCAG was based on clinical judgement, which can vary between physicians. Consequently, group assignment could be affected by confounding by indication. In line of this limitation, 24.4% of patients did not show any abnormalities upon uCAG, which may artificially decrease the risk of adverse events in the uCAG group. Still, we have mitigated for this limitation by performing sensitivity analyses excluding such patients. Furthermore, the exact reasons for a conservative or invasive management could not be adequately retrieved because of the retrospective nature of the study. In general, a conservative treatment was used when there were no angiographic abnormalities or when the patient was stable and PCI would be futile (i.e. technical failure). Conversely, patients would be revascularized when they were unstable and/or lesions were amenable to PCI or re-CABG. Notably, follow-up was not complete for 100% of patients (i.e. 94.0% had missing data, n = 114 patients). Finally, although the single-centre design ensures high interval validity, external validity (or generalizability) may be affected as different centres apply varying protocols, which can differ in use of biomarkers [2], and indication thresholds to perform uCAG or diagnose PMI. Furthermore, our institution can be considered a high-volume centre for CABG, and results may not be generalizable to lower-volume centres.

CONCLUSION

Unplanned postoperative CAG was performed in 4.1% of patients undergoing CABG procedures and only half of these patients were eligible for urgent re-revascularization. Patients undergoing uCAG constitute a high-risk group, and are predisposed to increased early mortality, which results in decreased long-term survival.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

FUNDING

None received.

Conflict of interest: Samuel Heuts is supported by the Dutch Heart Foundation through the Dekker program. Arnoud W.J. van 't Hof reports that his institution received unrestricted grants from Abbott, Roche Medtronic, Boehringer Ingelheim and Astra Zeneca, unrelated to this work.

DATA AVAILABILITY

All data are available upon reasonable request to the corresponding author.

Author contributions

Samuel Heuts: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Visualization; Writing—original draft. Roberto Bova: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Visualization; Writing—original draft. Jamie L.R. Romeo: Conceptualization; Data curation; Formal analysis; Writing—review & editing. Pieter A. Vriesendorp: Conceptualization; Supervision; Validation; Writing—review & editing. Iwan C.C. van der Horst: Conceptualization; Supervision; Validation; Writing—review & editing. Patrique Segers: Conceptualization; Supervision; Validation; Writing—review & editing. Jos G. Maessen: Conceptualization; Supervision; Validation; Writing—review & editing. Can Gollman-Tepeköylü: Conceptualization; Supervision; Validation; Writing—review & editing. Arnoud W.J. van 't Hof: Conceptualization; Supervision; Validation; Writing—review & editing. Elham Bidar: Conceptualization; Supervision; Validation; Writing—review & editing. Arpad Lux: Conceptualization; Data curation; Investigation; Methodology; Supervision; Validation; Writing—original draft.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Frank A. Baciewicz Jr., Emilio Monguió and the other anonymous reviewers for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• CABG

  Coronary artery bypass grafting

 
• CAG

  Coronary angiography

 
• CI

  Confidence interval

 
• CK-MB

  MB-isoenzyme of creatine kinase

 
• CONS

  Conservative (group)

 
• ECG

  Electrocardiography

 
• EuroSCORE

  European System for Cardiac Operative Risk Evaluation

 
• HR

  Hazard ratio

 
• LAD

  Left anterior descending artery

 
• MIDCAB

  Minimally invasive direct CABG

 
• OPCAB

  Off-pump CABG

 
• PCI

  Percutaneous coronary intervention

 
• PMI

  Periprocedural myocardial ischaemia

 
• REV

  Revascularized (group)

 
• STEMI

  ST-elevation myocardial infarction

 
• uCAG

  Unplanned coronary angiography

Author notes