Unplanned Postoperative Catheterization during Admission for Coronary Artery Bypass Grafting is neither Cheap nor Benign, but May Rescue Patients

Abstract

Objective:

Literature describing outcomes of myocardial ischemia after coronary artery bypass grafting (CABG) are sparse. We hypothesized these patients had more complications and incurred higher costs of care.

Methods:

Using adult cardiac surgery data (ACSD) and cardiac catheterization (CathPCI) data from the Virginia Cardiac Services Quality Initiative (VCSQI), we identified patients who underwent unplanned cardiac catheterization after CABG from 2018–2021. ACSD were matched to CathPCI data examining earliest in-hospital catheterization. Patients not requiring catheterization served as the control group.

Results:

We identified 10,597 patients who underwent isolated CABG of which 41/10,597 underwent unplanned cardiac catheterization. 21/41 (51%) received PCI, most commonly for NSTEMI (n=7, 33%) and STEMI (n=6, 29%). Postoperative cardiac arrest occurred in 14 patients (40%). In patients who underwent PCI, 14 (67%) had a single lesion, 4 (19%) had two lesions, and 3 (14%) had three lesions. The LAD (38%) was the most frequently intervened upon vessel. Patients who underwent catheterization were more likely to require balloon pump support (26% vs. 11%), have prolonged ventilation (57% vs. 20%), renal failure (17% vs 7.1%), and reintubation [(37% vs. 3.8%), all p<0.04]. There was no statistical difference in operative mortality (4.9% vs. 2.3%, p=0.2) or Failure to Rescue (4.9% vs. 1.6%, p=0.14). Total costs were higher in patients who underwent unplanned catheterization ($81,293 vs $37,011, p<0.001).

Conclusions:

Unplanned catheterization after CABG is infrequent but associated with more complications and a higher cost of care. Therefore, determination of an association with operative mortality in patients with suspected ischemia after CABG requires additional study.

Central Message:

We found that unplanned catheterization after CABG is infrequent. It is associated with more complications and a higher cost of care, but does not appear to increase operative mortality

Introduction

More than 200,000 coronary artery bypass grafts (CABG) are performed annually in the United States.¹ Despite studies demonstrating the benefits of CABG in alleviating angina, the risk of recurrent symptoms is approximately 6–8%/year.² This has mainly been attributed to the progression of native coronary disease or graft failure. Up to 12% of graft defects can be detected within 1 week postoperatively.^3,4 However, not all patients with graft failure present with clinical signs of ischemia.^5,6 The extent of the myocardial territory impacted plays a significant role in a patient’s symptoms. Nevertheless, several studies have shown that graft occlusion leads to an increased risk of myocardial infarction (MI) or death.^7,8

The causes of graft failure are multifactorial with the primary culprits changing over time. Acute graft failure (AGF, <30 days) is more likely to occur due to thrombosis, technical concerns, or poor outflow. In comparison, early (1–24 months) or late (>24 months) graft failure favor neointimal hyperplasia or atherosclerosis as the likely etiology.^9,10 While common mechanisms for AGF that may result in myocardial ischemia have been documented, few studies have looked at the patient characteristics and postoperative outcomes in these patients. Many studies examining graft failure effectively include those who experience AGF, however, the few that comment on these patients do so in the context of a scheduled angiogram or in fact exclude them from the primary analysis due to higher rates death, MI, or revascularization that may otherwise skew results. ^4,5 This suggests that AGF patients are unique and understudied. Upon our review, no recent studies have specifically looked into patient characteristics and postoperative outcomes in this population. Furthermore, examining outcomes in those who experience in-hospital myocardial ischemia may provide insight into the impact of prompt intervention on outcomes.

Thus, the aim of our study was to evaluate the postoperative outcomes in patients who experienced clinically significant myocardial ischemia during their index admission after CABG and subsequently underwent an unplanned cardiac catheterization. We hypothesized that patients who underwent an unplanned cardiac catheterization after CABG had a higher incidence of postoperative complications and incurred higher costs of care.

Methods

Patient Data

The Virginia Cardiac Services Quality Initiative (VCSQI) includes 17 hospitals in Virginia and contains 99% of all adult cardiac surgeries in the region. Owing to the deidentified nature of the quality database, this study was exempt from review by the University of Virginia’s Institutional Review Board (Protocol #23305, deemed exempt July 14, 2021). Patient demographic and outcomes data were obtained from the STS database using a sample including adult patients who underwent an isolated CABG surgery in the VCSQI. Each patient was matched to their corresponding entry in our statewide coronary catheterization database (CathPCI) by date of birth, first name and last name. Patients with catheterizations immediately after surgery (before discharge or within 30 days) were identified. Patients who underwent planned postoperative catheterizations were excluded to effectively exclude hybrid coronary revascularizations.

Outcomes measured included STS operative mortality and major morbidities including postoperative renal failure, stroke, reoperation for any reason, deep sternal wound infection, and prolonged postoperative ventilation greater than 24 hours. Postoperative renal failure is defined as Kidney Disease: Improving Global Outcomes (KDIGO) stage 2 acute kidney injury or greater or new requirement for dialysis after surgery. Relevant data about postoperative catheterization is reported where relevant, including, but not limited to, vessels involved and interventions performed. Patients who underwent catheterization were stratified by whether or not they underwent PCI.

Cost data is provided by The National Uniform Billing Committee (NUBC) Universal Billing Data Set (UB-04). UB-04 data files are submitted to the VCSQI via the Virginia Hospital and Healthcare Association for every inpatient admission. This data set contains total hospital charges for the admission as well as the individual charges for operating room, radiology, laboratory, blood bank, critical care and acute care costs. Based on cost-to-charge fillings with the Centers for Medicare and Medicaid Services, these charges are then converted to estimated costs for the total hospital charges. These total charge cost data are then algorithmically matched with clinical data from the STS ACSD based on medical record numbers, admission and discharge dates, and other patient identifiers. Those left unmatched are then manually addressed.

Statistical Analysis

The reporting of statistical results adhered to the guidelines provided in the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) and TRIPOD (Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis) statements. Continuous variables are presented as median (Q1, Q3) or mean (standard deviation) as appropriate, while categorical variables are presented as n (%). Standard STS and CathPCI variable definitions were utilized. Missing data were imputed for variables, except creatinine, with less than 10% missingness using standardized approaches with sex-specific median for continuous variables and lower risk mode for categorical variables.

Differences between baseline covariates in patients who underwent unplanned cardiac catheterization and the control group were compared using chi-squared or Fisher tests for categorical measures as appropriate. Student t-test or Mann-Whitney U tests were utilized for continuous measures as appropriate based on normality. Rstudio^™ and the R programming language were used for statistical analysis.¹¹ Specific packages and a sample of the code used to perform the analysis can be found in the appendix. A cutoff of less than .05 for P values was used for statistical significance.

Results

Baseline Characteristics: Unplanned Cardiac Catheterization vs Remaining CABG Patients

A total of 10,597 patients who underwent isolated CABG between 2018–2021 were identified in the VCSQI database. All patients had cardiac catheterization data in the statewide Cath PCI database, with 1559 patients having both preoperative and postoperative catheterization data available (Figure 2). Ultimately, 41 patients (0.39%) underwent an unplanned postoperative cardiac catheterization during their index admission, with 21 patients requiring PCI.

Figure 2:

Consort diagram for unplanned cardiac catheterization after CABG group and control CABG group. (STS: Society of Thoracic Surgery, PCI: Percutaneous coronary intervention, CABG: Coronary artery bypass grafting)

Most unplanned catheterizations occurred during the first week postoperatively (36/41, 88%) with the majority performed within the first 48 hours (Figure 3). The unplanned catheterization group more frequently had hypertension (100% vs. 91%, p=0.045), cerebrovascular disease (41% vs 24%, p=0.038), and needed their CABG in an urgent fashion (73% vs. 61%, p=0.025). Baseline characteristics were fairly similar between groups as detailed in Table 1. This includes age (65% vs. 66%), female sex (27% vs. 24%), white race (71% vs. 81%), diabetes (59% vs. 55%), left main coronary artery disease (9.8% vs. 10%), prior myocardial infarction (66% vs. 56%), dialysis dependency (0% vs. 2.9%), and cross-clamp time [(74 mins vs 71), all p≥0.3]. There was a trend in the unplanned catheterization group towards a lower ejection fraction (50% vs. 56%, p=0.053) and hematocrit (38.5 vs. 40.2, p=0.075) as well as higher predicted risk of mortality (1.4% vs. 1.0%, p=0.063) and morbidity or mortality (11% vs. 8%, p=0.087). However, these differences were not statistically significant.

Figure 3:

Timing of catheterization after CABG: A) Time from CABG to catheterization in days. B) Rates of early catheterization C) Incidence of mortality according to time from CABG to catheterization in hours.

Table 1:

Baseline characteristics of unplanned cardiac catheterization group and CABG patient population (LIMA: left internal mammary artery, RIMA: right internal mammary artery)

Variable	N	Unplanned Catheterization, N = 411	All CABG, N = 10,5561	p-value2
Preoperative
Age	10,596	65 (60, 72)	66 (59, 73)	0.4
Female sex	10,597	11 (27%)	2,505 (24%)	0.6
Body Surface Area (m2)	10,488	2.88 (2.75, 2.97)	2.88 (2.78, 3.00)	0.5
Race
White	10,127	29 (71%)	8,138 (81%)	0.11
Black/African American	9,698	8 (20%)	1,433 (15%)	0.4
Asian	9,716	3 (7.5%)	408 (4.2%)	0.2
Native Hawaiian / Pacific Islander	9,635	0 (0%)	30 (0.3%)	>0.9
American Indian / Alaskan Indian	9,635	0 (0%)	28 (0.3%)	>0.9
Primary payor	10,485
Medicare		18 (44%)	5492 (53%)
Commercial Health Insurance		13 (32%)	3,381 (32%)
Medicaid		2 (4.8%)	534 (5.1%)
Health Maintenance Organization		0 (0%)	266 (2.5%)
Military Health		1 (2.4%)	187 (1.8%)
Charitable Care/Foundation Funding		0 (0%)	47 (0.5%)
Correctional Facility		0 (0%)	23 (0.2%)
Indian Health Service		0 (0%)	1 (<0.1%)
Non-U.S. Plan		0 (0%)	4 (<0.1%)
None / self		5 (12%)	374 (3.6%)
Other		0 (0%)	9 (<0.1%)
Diabetes	10,492	24 (59%)	5,215 (50%)	0.3
Hypertension	10,492	41 (100%)	9,469 (91%)	0.045
Cerebrovascular Disease	10,492	17 (41%)	2,500 (24%)	0.038
Peripheral arterial disease	10,492	3 (7.3%)	1,447 (14%)	0.4
Tobacco use	10,492			>0.9
Never smoker		15 (37%)	4,145 (40%)
Former smoker		17 (41%)	3,984 (38%)
Current some day smoker		0 (0%)	172 (1.6%)
Current every day smoker		9 (22%)	2,109 (20%)
Smoking status unknown		0 (0%)	22 (0.2%)
Chronic lung disease	10,491			0.7
None		24 (59%)	6,904 (66%)
Mild		8 (20%)	1,771 (17%)
Moderate		2 (4.9%)	496 (4.7%)
Severe		3 (7.3%)	538 (5.1%)
Unknown		2 (4.9%)	383 (3.7%)
Dialysis dependency	10,492	0 (0%)	307 (2.9%)	0.6
Liver disease	10,487	1 (2.4%)	357 (3.4%)	>0.9
Prior myocardial infarction	10,595	27 (66%)	5,961 (56%)	0.5
Previous MI Timing	5,899			0.4
≤6 Hours		0 (0%)	82 (1.4%)
> 6 to 24 hours		1 (3.7%)	127 (2.2%)
1 to 7 Days		17 (63%)	2,920 (50%)
8 to 21 Days		4 (15%)	797 (14%)
>21 Days		5 (19%)	1,946 (33%)
Prior PCI	3,558	13 (100%)	3,172 (89%)	0.4
Previous PCI Interval	3,098			>0.6
≤6 Hours		0 (0%)	75 (2.4%)
> 6 Hours		11 (100%)	3,012 (98%)
Previous CABG	3,483	0 (0%)	81 (2.3%)	>0.9
Diseased Vessels	10,470			0.6
One		0 (0%)	364 (3.5%)
Two		9 (22%)	2,106 (20%)
Three		32 (78%)	7,952 (76%)
Left Main Coronary Artery Stenosis	10,497	4 (9.8%)	1,050 (10%)	>0.9
Heart Failure	10,490	16 (39%)	3,217 (31%)	0.4
NYHA Classification	5,263			0.4
Class I		2 (8.7%)	598 (11%)
Class II		8 (35%)	1,303 (25%)
Class III		5 (22%)	862 (16%)
Class IV		2 (8.7%)	317 (6.0%)
Not documented		6 (26%)	2,160 (41%)
Left Ventricular Ejection Fraction (%)	10,388	50 (43, 60)	56 (48, 60)	0.053
Preoperative Creatinine	10,478	0.90 (0.79, 1.20)	1.00 (0.80, 1.20)	0.2
Preoperative Hematocrit	10,479	38.5 (35.4, 41.6)	40.2 (36.3, 43.6)	0.075
Preoperative Platelets	10,471	214,500 (173,750, 285,000)	217,000 (179,000, 260,000)	0.6
INR	10,066	1.00 (1.00, 1.06)	1.00 (1.00, 1.10)	0.14
Predicted Risk of Mortality	10,492	0.014 (0.007, 0.025)	0.010 (0.006, 0.019)	0.063
Predicted Risk of Morbidity or Mortality	10,492	0.11 (0.06, 0.17)	0.08 (0.06, 0.14)	0.087
Intraoperative
CABG Status	10,597			0.025
Elective		9 (22%)	3,791 (36%)
Urgent		30 (73%)	6,454 (61%)
Emergent		1 (2.4%)	294 (2.8%)
Emergent Salvage		1 (2.4%)	17 (0.2%)
Cross-clamp time (Mins)	9,906	74 (47, 96)	71 (54, 90)	0.9
Cardiopulmonary Bypass time (Mins)	10,164	96 (71, 124)	94 (74, 117)	0.8
Off-Pump CABG	10,497	2 (4.9%)	389 (3.7%)	0.3
LIMA Used	10,490	38 (92.4%)	10,122 (97.2%)	0.3
Pedicle	10,490	37 (90%)	9,579 (92%)	0.3
Skeletonized	10,490	1 (2.4%)	543 (5.2%)	0.3
RIMA Used	10,490	3 (7.3%)	644 (6.1%)	0.2
Pedicle	10,490	2 (4.9%)	412 (3.9%)	0.2
Skeletonized	10,490	1 (2.4%)	232 (2.2%)	0.2
Radial Artery Graft Used	10,497	2 (4.9%)	886 (8.5%)	0.6
Radial artery harvest and preparation time (Mins)	453	39 (38, 40)	55 (43, 69)	0.12
Vein Graft Used	10,497	35 (85%)	9,592 (92%)	0.15
Saphenous Vein Harvest and Preparation Time (Mins)	3,424	57 (42, 86)	55 (41, 72)	0.5

¹n (%); Median (IQR)

²Fisher’s exact test; Wilcoxon rank sum test; Pearson’s Chi-squared test

Postoperative Outcomes: Unplanned Cardiac Catheterization vs. Remaining CABG patients

Postoperative complications are detailed in Table 2. The unplanned catheterization group had higher rates of postoperative cardiac arrest (40% vs. 4.8%), reintubation (37% vs 3.8%), prolonged ventilation (57% vs. 20%), and reoperation [(17% vs. 1.5%), all p<0.001]. They were more likely to require intra-aortic balloon pump (IABP) support (26% vs. 11%, p=0.008), particularly in the postoperative setting (56% vs. 2.5%, p<0.001). They also more frequently required catheter based assist devices (7.3% vs. 0.2%, p<0.001). The unplanned catheterization group was more likely to develop renal failure (17% vs. 7.1%, p=0.037), had a longer hospital length of stay (18 days vs. 8, p<0.001), and a more costly hospital admission ($81,293 vs. $37,011, p<0.001). However, the groups had similar operative mortality (4.9% vs. 2.3%, p=0.2) and failure to rescue (4.9% vs. 1.6%, p=0.14). Table 4 details the pre-, intra-, and postoperative characteristics of patients who experienced a major morbidity in those who underwent catheterization versus the control CABG cohort. Most baseline characteristics were comparable and operative mortality (9.5% vs. 16%, p=0.6) was similar.

Table 2:

Postoperative outcomes for unplanned cardiac catheterization group and CABG patient population (IABP: intra-aortic balloon pump)

Variable	N	Unplanned Catheterization, N = 411	All CABG, N = 10,5561	p-value2
Operative mortality	10,597	2 (4.9%)	244 (2.3%)	0.2
Major morbidity	10,493	21 (51%)	1,044 (10.0%)	<0.001
Failure to rescue	10,597	2 (4.9%)	169 (1.6%)	0.14
Reoperation	3,898	6 (17%)	58 (1.5%)	<0.001
Prolonged ventilation	3,896	20 (57%)	764 (20%)	<0.001
Renal failure	3,853	6 (17%)	272 (7.1%)	0.037
Stroke	3,856	1 (2.9%)	125 (3.3%)	0.7
Cardiac arrest	3,861	14 (40%)	183 (4.8%)	<0.001
Reintubation	10,592	15 (37%)	403 (3.8%)	<0.001
Atrial fibrillation	3,861	16 (46%)	2,501 (65%)	0.015
Sepsis	3,853	1 (2.9%)	103 (2.7%)	0.6
IABP	7,761	9 (26%)	825 (11%)	0.008
When IABP inserted	801			<0.001
Preoperative		3 (33%)	619 (78%)
Intraoperative		1 (11%)	153 (19%)
Postoperative		5 (56%)	20 (2.5%)
Catheter based assist device used	10,490	3 (7.3%)	76 (0.7%)	0.002
Preoperative		0 (0%)	38 (0.4%)	<0.001
Intraoperative		0 (0%)	18 (0.2%)
Postoperative		3 (7.3%)	20 (0.2%)
Support type	10,490			<0.001
Left ventricular		2 (4.9%)	67 (0.6%)
Right ventricular		0 (0%)	8 (<0.1%)
Biventricular		1 (2.4%)	1 (<0.1%)
ICU initial length of stay	1,934	77 (48, 191)	66 (37, 101)	0.13
Hospital length of stay (days)	10,479	18.0 (9.0, 27.0)	8.0 (6.0, 12.0)	<0.001
Total Cost (USD)	6,542	$81,293 ($52,062, $161,259)	$37,011 ($30,074, $46,815)	<0.001

¹Median (IQR); n (%)

²Wilcoxon rank sum exact test; Wilcoxon rank sum test; Pearson’s Chi-squared test; Fisher’s exact test

Table 4:

Characteristics of patients with major morbidity

Variable	Early Catheterization, N = 211	No Catheterization, N = 1,0651	p-value2
Age	66 (61, 70)	68 (60, 74)	0.5
Female sex	8 (38%)	328 (31%)	0.5
White	14 (67%)	744 (72%)	0.6
Hypertension	21 (100%)	989 (93%)	0.4
Diabetes	12 (57%)	607 (57%)	>0.9
Current smoker	5 (24%)	283 (26.2%)	>0.9
Chronic lung disease			0.4
None	8 (38%)	581 (55%)
Mild	7 (33%)	205 (19%)
Moderate	1 (4.8%)	73 (6.9%)
Severe	2 (9.5%)	99 (9.3%)
Cerebrovascular Disease	10 (48%)	316 (30%)	0.2
Prior MI	16 (76%)	755 (71%)	0.8
Dialysis Dependent	0 (0%)	53 (5.0%)	0.6
Preoperative creatinine	0.99 (0.80, 1.30)	1.03 (0.82, 1.36)	0.3
Liver disease	0 (0%)	59 (5.5%)	0.6
Prior PCI	6 (29%)	55 (27%)	0.8
Prior CABG	0 (0%)	2 (1.0%)	>0.9
Cardiopulmonary bypass time	116 (75, 130)	100 (77, 128)	0.7
Prolonged ventilation	20 (95%)	764 (72%)	0.017
Reoperation	6 (29%)	58 (5.5%)	<0.001
Stroke	1 (4.8%)	127 (11.9%)	0.9
Renal failure	6 (29%)	274 (26%)	0.8
Operative Mortality	2 (9.5%)	169 (16%)	0.6

¹Median (IQR); n (%)

²Wilcoxon rank sum exact test; Wilcoxon rank sum test; Pearson’s Chi-squared test; Fisher’s exact test

Percutaneous Coronary Intervention vs Diagnostic Catheterization: Findings and Postoperative Outcomes.

Among those who underwent catheterization, 21/41 underwent PCI and 20/41 diagnostic catheterization. These catheterizations occurred at 7/17 VCSQI centers. The most common indication for cardiac catheterization in both groups was acute coronary syndrome (43% vs. 45%, p=0.9) and cardiac arrest (19% vs. 25%, p=0.7) with the most common indication for PCI being Non-ST Elevation myocardial infarction (n=7, 33%) and STEMI (n = 6, 29%). Locations of lesions intervened upon with PCI are detailed in supplemental table 1. Among them, 14 (67%) had a single lesion, 4 (19%) had two lesions, and 3 (14%) had three lesions. The LAD (38%), circumflex (29%), right coronary arteries (24%), and vein graft (19%) were more frequently intervened upon. No patients had evidence of in-stent re-stenosis in either group. Of the 20 diagnostic catheterization patients, only 1 patient went on to receive reintervention for myocardial ischemia (95% vs. 6.7%, p<0.001). Operative mortality was similar between the two groups (4.8% vs. 5%, p>0.9) and all deaths occurred within the first 48 hours postoperatively. Similarly, complication rates between the two groups were similar including major morbidity (62% vs 40%, p=0.2) and cardiac arrest (38% vs 30%, p>0.9).

Discussion

In this study, we examined the outcomes associated with clinically significant, in-hospital, postoperative myocardial ischemia after CABG. This is fortunately a rare event with an incidence of <0.5% in our study. When these events do occur, however, they significantly impact the postoperative course as evidenced by the high rates of postoperative cardiac arrest (40%), reintubation (37%), prolonged ventilation (57%), and reoperation (17%). This complication also more than doubles the cost of a patient’s hospitalization ($81,293 vs. $37,011). Despite this burden, there was no difference in operative mortality or failure to rescue. Furthermore, unplanned cardiac catheterization did not appear to increase failure to rescue compared to others with major morbidity. However, these rates were not low (9.5% vs. 16%). While the exact etiology of myocardial ischemia in these patients cannot be identified with the data available, it is presumably related to their native disease or factors pertaining to their recent operation - namely graft failure.

Why Does Graft Failure and Myocardial Ischemia Occur?

The causes of graft failure are numerous and include endothelial dysfunction, thrombosis, neointimal hyperplasia, atherosclerosis, vasospasm, oxidative stress, poor outflow, and surgical techniques.^6,10 Technical concerns and graft thrombosis are the main culprits in AGF.^10,12 However, one of the most significant predictors of graft failure can be evaluated prior to surgery - recipient vessel diameter. Goldman and colleagues demonstrated that coronary arteries >2.0mm on preoperative angiography had an 88% 10 year patency rate vs. 55% for those < 2.0mm. Although the standard is to graft the LIMA to the LAD, patency benefits of recipient vessels >2.0m held true even for saphenous vein grafts (SVG) with 10 year patency rates of 90%. Unfortunately in this study, data regarding recipient vessel diameter were not available for review. Regarding AGF, Goldman also found that patients with patent vein grafts at 1 week would go on to have a 6 year patency rate of 76%⁴. For reference, the 12–18 month patency rates of the PREVENT IV trial was 75%.¹³

Intraoperative conduit care also plays a significant role in AGF. Excessive manipulation of vein grafts has been shown to increase rates of graft failure.¹⁴ Saphenous vein harvest preparation times were similar between the unplanned catheterization group and CABG cohort (57 mins vs. 55, p=0.5). This would suggest that there was no unusual difficulty during the vein harvest in the unplanned catheterization group. In addition, open and no touch techniques during harvest have been investigated and shown to decrease the rate of graft failure, angina, repeat revascularization, MI and death.^{15 16 17} However, findings from newer trials such as the REGROUP and SUPERIOR SVG trial may refute some of the previous findings regarding rates of major adverse cardiac events with these minimal manipulation techniques.^{18 19} Nevertheless, there is some unavoidable risk associated with graft harvest as the vasovasorum is disrupted. This subjects grafts to ischemia reperfusion injury.²⁰

The use of arterial grafts, specifically the radial artery (RA) and bilateral internal mammary arteries, has been advocated for in CABG surgery.²¹ This stems from data suggesting these conduits have higher long term patency rates than SVGs, with one meta-analysis finding at least a 3-fold higher incidence of graft occlusion in SVG compared to arterial grafts.²² LIMA (92.4% vs 97.2%), RIMA (7.3% vs. 6.1%), and radial artery utilization rates [(4.9% vs. 2.2%), all p≥0.2] were similar between the unplanned catheterization and CABG control group, which may suggest that conduit selection may not have played as significant of a role in early postoperative myocardial ischemia. Furthermore, conduit preparation was fairly similar between these two groups with rates of LIMA harvest via pedicle technique being high in both (90% vs 92%, p=0.3). Harvesting the LIMA as a pedicle rather than skeletonizing it has been shown to protective against graft failure, and likely contributed to the overall low rate of AGF in this study.²³ Radial artery graft preparation times were also similar between the two groups (39 mins vs 55, p=0.12). However, it should be noted that the RIMA and radial artery were infrequently used in the unplanned catheterization cohort and likely underpowered to definitively comment on their impact on AGF. Nevertheless, in patients who went on to have a PCI, 7/21 patients (33%) had a lesion in a graft with 4/7 receiving a PCI in a vein graft. This coincides with the aforementioned long-term data that suggests higher rates of graft occlusion in SVG compared to arterial conduits.

One particular concern with using arterial grafts, however, is their predilection for vasospasm. This is particularly true in those with an atherosclerotic burden that may put the patient at risk for competitive flow as well as those with high risk for needing postoperative vasopressor support, which could increase the risk of AGF.²⁴ Patients in the unplanned catheterization group more frequently required hemodynamic support with an IABP (26% vs 11%, p=0.008) or other assist devices (7.3% vs. 0.7%, p=0.002), which could precipitate vasospasm. Strides have been made in using vasodilating agents to help mitigate this risk.

How Can We Reduce Graft Failure?

Several common place procedures can be modified to reduce the risk of graft failure. During graft preparation after it is harvested, grafts are frequently distended to assess for leakage. This has been shown to induce endothelial damage and increase the risk of graft thrombosis, and it is recommended to distend the graft to less than 100 mmHg.²⁵ Another common practice is to mark the graft with a skin marker to orient the graft and prevent kinking. However, these markers were designed for the skin. Isopropyl alcohol solvents and methylene blue dyes used in these markers have been shown to cause local smooth muscle dysfunction, which can precipitate graft failure.²⁶ Brilliant Blue FCF has shown promise as an alternative to limit endothelial and smooth muscle cell dysfunction.²⁷ Graft preservation solutions have also been compared. Buffered solutions have been associated with decreased vein graft failure and endothelial damage as compared to saline or heparinized autologous blood based solutions.²⁸ While this level of granularity is not available in our data, it should be noted by surgeons performing CABG surgery.

Actions can be taken in the postoperative setting to minimize the risk of graft failure. Initiating aspirin early in the postoperative course to prevent graft thrombosis is an established practice endorsed by multiple guidelines.^21,31 While the data available in our study precluded us from examining the timing of antiplatelet therapy initiation, rates of aspirin use at discharge were fairly high across the unplanned catheterization when not otherwise contraindicated. This was also the case for ADP inhibitors in PCI group (Table 3). Dual antiplatelet therapy has been shown to decrease graft occlusion rates compared to aspirin alone, with the tradeoff of a higher risk of bleeding events.³² As such, the 2021 ACC/AHA/SCAI guidelines gave a COR IIb, LOE B-R recommendation for the use of DAPT for one year in selected patients to improve vein graft patency.²¹

Table 3:

Diagnostic Catheterization versus PCI Procedure Details and Postoperative Outcomes

Variable	PCI, N = 211	Diagnostic, N = 201	p-value2
Procedure Details
Catheterization lab indication
Cardiac arrest	4 (19%)	5 (25%)	0.7
Acute coronary syndrome	9 (43%)	9 (45%)	0.9
Arrhythmia	4 (19%)	3 (15%)	>0.9
Left ventricular dysfunction	2 (9.5%)	5 (25%)	0.2
lesions with in stent restenosis	0 (0%)	0 (NA%)
Reintervention for Myocardial Ischemia	19 (95%)	1 (6.7%)	<0.001
PCI indication			>0.9
CAD	2 (10%)	0 (0%)
Stable angina	1 (5%)	0 (0%)
NSTE – ACS	7 (35%)	0 (0%)
STEMI – unstable (>12 hours of symptoms)	1 (5%)	0 (0%)
STEMI – Acute	5 (25%)	0 (0%)
Other	4 (20%)	0 (0%)
Lesions count per PCI			>0.9
1	14 (67%)	0 (NA%)
2	4 (19%)	0 (NA%)
3	3 (14%)	0 (NA%)
PCI vessel intervention
Left main coronary artery	3 (14%)	0 (NA%)	>0.9
Left anterior descending artery	8 (38%)	0 (NA%)	>0.9
Right coronary artery	5 (24%)	0 (NA%)	>0.9
Circumflex artery	6 (29%)	0 (NA%)	>0.9
Posterior descending artery	2 (9.5%)	0 (NA%)	>0.9
Ramus artery	1 (4.8%)	0 (NA%)	>0.9
LIMA graft used	20 (95%)	18 (90%)	0.6
Radial artery graft used	2 (9.5%)	0 (0%)	0.5
Vein graft used	15 (71%)	20 (100%)	0.021
Lesion in graft	7 (33%)	0 (0%)	0.009
PCI in vein graft	4 (19%)	0 (0%)	>0.9
Culprit lesion status			>0.9
Culprit Known	12 (60%)	0 (0%)
Culprit None	7 (35%)	0 (0%)
Culprit Unknown	1 (5.0%)	0 (0%)
Postoperative Outcomes
Postoperative catheter based assist device used	2 (9.5%)	1 (5%)	>0.9
Postoperative IABP	3 (14%)	2 (10%)	>0.9
Cardiac arrest	8 (38%)	6 (30%)	>0.9
Reoperation for bleeding	2 (10%)	1 (6.7%)	>0.9
Reoperation (other)	2 (10%)	4 (27%)	0.4
Major morbidity	13 (62%)	8 (40%)	0.2
Operative mortality	1 (4.8%)	1 (5.0%)	>0.9
Aspirin at discharge	19 (90%)	18 (95%)	>0.9
ADP inhibitor at discharge	20 (95%)	7 (37%)	<0.001

¹Median (IQR); n (%)

²Wilcoxon rank sum exact test; Wilcoxon rank sum test; Pearson’s Chi-squared test; Fisher’s exact test

(NSTEMI: non-ST elevation myocardial infarction)

Considerations if Postoperative Myocardial Ischemia Does Occur

The aforementioned strategies provide opportunities to reduce the risk of AGF and ultimately myocardial ischemia, however, it can still occur. Patients with AGF may be asymptomatic, but those with symptoms frequently present with recurrent angina, cardiogenic shock, myocardial infarction or sudden death. This is consistent with our findings as 40% of the unplanned catheterization group suffered from a cardiac arrest and approximately two-thirds of patients who underwent PCI suffered from acute coronary syndrome. Predictors of poor outcomes such as acute MI, repeat revascularization or death include age, hypertension, previous MI, and reduced ejection fraction.⁸ The unplanned catheterization group did have a higher incidence of hypertension (100% vs. 91%, p=0.045) and a trend toward a lower ejection fraction (50% vs. 56%), p=0.053). However, the majority of baseline characteristics were similar between the unplanned catheterization group and the control cohort in our study.

The options for treating graft failure include medical, interventional and surgical therapy. However, many of these patients often require repeat revascularization.^8,13 Several trials have examined PCI outcomes in patients who have had a prior CABG.³³ Multiple factors are to be considered when deciding between PCI and reoperation. Significant patient comorbidities, single graft or focal lesions, STEMI, acute or early graft failure, limited conduit availability, and a preserved LIMA-LAD anastomosis are among the factors that would favor PCI over reoperative CABG.⁹ 67% of patients in our study only had a single lesion that required PCI. However, 14% of patients had a PCI to the left main coronary artery and 38% to the left anterior descending. While the decision to pursue PCI versus reoperation is difficult, these patients may have otherwise benefitted from open revascularization. Data illustrating reoperation for graft revision are limited. Those who did undergo reoperative revision may be included in the six patients who had a reoperation for another cardiac reason. However, this cannot be said for certain. Many patients may not have had the appropriate disease pattern to be intervened upon at all – which is functionally representative of the diagnostic catheterization group. Only one patient in the diagnostic catheterization group would go on to have reintervention for myocardial ischemia – likely surgical revision. As discussed above, this could be due to progression of the native coronary disease, transient causes such as vasospasm, or other causes not captured by the ACSD.

In addressing graft thrombosis, PCI in grafts can be done with care. Shoaib and colleagues looked at patients with prior CABG who experienced NSTEMI and underwent PCI. They found that patients had similar rates of 30-day mortality regardless of whether the PCI was performed in the native coronary artery or the SVG³⁴. This is reflected in our study where operative mortality was similar in the PCI group and the diagnostic catheterization group, where the former, 19% of patients had a PCI in their graft and 35% presented with an NSTEMI.

Once appropriate measures have been taken to diagnose and treat patients with in-hospital myocardial ischemia after CABG, the focus turns to getting the patient back on track. As evidenced in Table 1, there was significantly higher rates of cardiac arrest, reintubation, prolonged ventilation, and reoperation. However, albeit infrequent to begin with, there was no statistically significant difference in operative mortality. Arguably more important, was the lack of difference in failure to rescue despite this clinically significant event. Strobel and colleagues have advocated for adding cardiac arrest to the failure to rescue definition given its lethality. In their study of >40,000 patients, cardiac arrest was associated with a 55% operative mortality. In this study, despite a 40% rate of cardiac arrest in the unplanned catheterization group, there was only a 4.9% rate of operative mortality. This would suggest that despite such a clinical instability in these patients, they may be able to recover fairly well with prompt intervention. ³⁵ As such, one could then argue that despite the more than two-fold increase in cost in the unplanned catheterization group, these dollars were well spent.

Limitations

Our study has several limitations. First, this data was reviewed in a respective fashion and may be subject to unmeasured confounding. Second, although there was no difference in operative mortality, the small sample size and even smaller number of operative deaths that occurred in the unplanned catheterization group does raise a concern for a Type II error and should be interpreted accordingly. Third, due to the lack of granularity of the data available, we cannot determine the exact etiology of myocardial ischemia that resulted in these unplanned cardiac catheterizations. Fourth, we can only partially speak to intraoperative conduit care and the timing of postoperative antiplatelet therapy initiation which we anticipate would impact AGF and subsequent myocardial ischemia.

Conclusions

Postoperative myocardial ischemia requiring an unplanned cardiac catheterization during the index admission for CABG is a rare event associated with increased costs and complications after CABG. However, it is a symptom rather than a cause, and may warrant cardiac catheterization to investigate the etiology. Prompt recognition and intervention may improve outcomes as these efforts do not appear to be associated with an increase in operative mortality or failure to rescue. See Figure 1 for a graphical abstract of the study. Several opportunities exist to reduce risk in the preoperative, intraoperative, and postoperative phases.

Figure 1:

Graphical Abstract

Central Picture:

Operative Mortality and Failure to Rescue in Patients Undergoing Unplanned Postoperative Catheterization During Index Admission for CABG versus CABG Control Group (CABG: Coronary Artery Bypass Grafting)

Perspective Statement:

There is an inherent risk of recurrent myocardial ischemia after CABG. In-hospital postoperative ischemia is a rare occurrence that may require repeat revascularization. Limited data exist examining outcomes in this sub-population of patients which require immediate attention.

Abbreviations:

STS

Society of Thoracic Surgeons

CABG

Coronary Artery Bypass Grafting

PCI

Percutaneous Coronary Intervention

SVG

Saphenous Vein Graft

AGF

Acute Graft Failure

RA

Radial Artery

MI

Myocardial Infarction

KDIGO

Kidney Disease: Improving Global Outcomes