Hospital Variability Drives Inconsistency in Antiplatelet use following Coronary Bypass

Abstract

Background:

Continuation of dual antiplatelet therapy (DAPT) following coronary artery bypass grafting (CABG) after acute myocardial infarction is recommended by current guidelines. We sought to evaluate guideline adherence over time and factors associated with post-operative DAPT within a regional consortium.

Methods:

Isolated CABG patients from 2011–2017 that had a myocardial infarction within 21 days prior to surgery were included. Patients were stratified by DAPT prescription at discharge and by time period, early (2011–2014) vs. late (2015–2017). Hierarchical regressions were then performed to evaluate factors influencing DAPT use after CABG.

Results:

A total of 7,314 patients were included with an overall rate of DAPT utilization of 31.2% that increased from 29.6% in the early to 33.4% in the late era (p<0.01). There was significant variability in hospital rates of DAPT (Range: 9.5%−92.1%) and hospital level changes over time (26% increased, 11% decreased and 63% remained stable). After adjustment for clinical factors, era was not associated with DAPT use but treating hospital remained significantly associated with DAPT use. Other clinical factors associated with increased DAPT utilization included off-pump surgery (OR 4.48, p<0.01), and prior percutaneous coronary intervention (OR 2.02, p<0.01) while atrial fibrillation (OR 0.39, p<0.01) was associated with decreased utilization.

Conclusions:

Dual antiplatelet use has increased between 2011 and 2017, driven primarily by evolving patient demographics. Significant hospital level variability drives inconsistency in DAPT utilization. Efforts to promote DAPT use for patients treated with CABG after myocardial infarction in concordance with current guidelines should be targeted at the hospital level.

Patients with acute myocardial infarction (AMI) are at significant risk for recurrent ischemic events and cardiovascular mortality.(1) Dual antiplatelet therapy (DAPT) has become a mainstay of antithrombotic therapy for secondary prevention following AMI and has been shown to reduce the risk of future cardiovascular events.(2,3) There is high utilization of DAPT in AMI patients treated with a noninvasive strategy and for those treated with percutaneous coronary intervention (PCI).(4,5) However, the rate of DAPT in patients who undergo surgical revascularization is lower.(6)

One of the factors contributing to this discrepancy is that the benefit of DAPT after coronary artery bypass grafting (CABG) is debated.(7) Recent multicenter, randomized evidence suggests that in elective CABG patients, DAPT improves vein graft patency.(8) In addition, retrospective studies and post-hoc analyses of AMI cohorts demonstrate a benefit of DAPT after CABG in this setting.(6,9,10) However, other investigations have found contradictory results and given the limited prospective, randomized data, debate on the appropriateness of DAPT after CABG remains.(11,12) Despite this being an area of active investigation, the 2016 American College of Cardiology (ACC)/ American Heart Association (AHA) Guidelines Focused Update on the Duration of Dual Antiplatelet Therapy in Patients with Coronary Artery Disease included a class I recommendation for completion of 1 year of DAPT after CABG in the setting of AMI.(13)

To evaluate factors associated with DAPT use after CABG in the setting of AMI, we analyzed rates of DAPT utilization and changes over time within a multicenter regional quality consortium. Specifically, we sought to determine the relative impact that clinical factors, individual surgeon, and hospital level practices have on guideline-directed antiplatelet agent use in the post-operative setting.

Patients and Methods

Patient Data

The Virginia Cardiac Services Quality Initiative (VCSQI) includes 19 hospitals and surgical groups in the region. Registry data includes 99% of all adult cardiac surgery in the region and methodologies for clinical data acquisition have been described previously.(14) Standard Society of Thoracic Surgeons (STS) definitions were used for all variables.(15) Institutional STS data is voluntarily submitted by each center and compiled in a central database to be used for quality improvement. Due to the de-identified nature of the quality database, this study was exempt from review by the University of Virginia Institutional Review Board.

All patients who underwent isolated CABG between July 1, 2011 and June 30, 2017 were extracted from the VCSQI database. Only patients with AMI within 21 days of surgery were included. Patients were excluded if data on discharge antiplatelet agents were missing. Patients were stratified by DAPT use or non-use and whether their operation occurred in the early (2011–2014) or late (2015–2017) era. These years were chosen to allow for similar size group comparisons between eras and specific sensitivity analyses were performed to evaluate the impact of the 2016 ACC/AHA guidelines. Dual antiplatelet therapy was defined as the use of aspirin and any adenosine diphosphate receptor inhibitor.

Statistical Analysis

Continuous variables with skewed distributions are presented as median [interquartile range (IQR)], normally distributed continuous variables as mean ± standard deviation, and categorical variables as count (percentage). Wilcoxon rank sum test was used for skewed continuous variables, independent t test for normally distributed continuous variables and the Chi-Square test was utilized for categorical variables. Hierarchical logistic regression with a generalized linear regression model was used to analyze the effect of center and surgeon level clustering on dual antiplatelet therapy use with both variables treated as random effects. Aggregate effect size for treating hospital and surgeon practice was calculated by summation of estimates and standard errors for each level. Operative year, as well as variables known to influence anti-thrombotic medication use from prior studies and other variables thought by the authors to be important for the decision making in prescribing DAPT were selected a priori and included in the models for risk-adjustment. These are listed in Table 2. Additional risk adjusted analyses were performed to evaluate the impact of the 2014 ACC/AHA Guideline for the Management of Non-ST Elevation Acute Coronary Syndrome and the 2016 ACC/AHA Guideline Focused Update on Duration of Dual Antiplatelet Therapy by evaluating adjusted DAPT use before and after these guideline releases.(13,16) SAS version 9.4 (SAS Institute, Cary, NC) statistical software was used for analysis with a statistical threshold 0.05 set for significance.

Table 2:

Results of hierarchical model predicting risk-adjusted dual antiplatelet use.

	OR	p-value
Off-pump surgery	4.48 (2.77–7.23)	<0.0001
Prior percutaneous coronary intervention	2.02 (1.63–2.49)	<0.0001
Aggregate hospital effect	1.93 (1.05–3.55)	0.0026
ADP Inhibitor with 5 days of Surgery	1.81 (1.38–2.38)	<0.0001
Peripheral arterial disease	1.72 (1.31–2.27)	0.0001
Prior stroke	1.52 (1.10–2.09)	0.0110
Aggregate surgeon effect	1.24 (0.68–2.27)	0.1784
Major morbidity	1.21 (0.84 −1.75)	0.2981
ST elevation myocardial infarction	1.12 (0.85–1.48)	0.4369
Predicted risk of morbidity or mortality	1.01 (1.00–1.02)	0.0109
Heart failure	0.99 (0.77–1.28)	0.9478
Patient age	0.98 (0.98–0.99)	0.0024
Post-operative length of stay	0.96 (0.94–0.98)	0.0003
Year	0.90 (0.81–1.00)	0.0505
Preoperative dialysis	0.76 (0.44–1.32)	0.3300
Transfusion	0.72 (0.56–0.93)	0.0103
Post-operative atrial fibrillation	0.64 (0.50–0.81)	0.0003
History of atrial fibrillation/flutter	0.39 (0.28–0.55)	<0.0001

(C-statistic 0.881) (ADP: adenosine diphosphate; CI: Confidence Interval; OR: Odds Ratio PCI: Percutaneous coronary intervention)

Results

A total of 7,314 patients who underwent isolated CABG in the setting of a recent AMI were included in the analysis after excluding 167 (2.2%) patients with missing discharge medication data (Figure 1). Of these, 4,164 (56.9%) underwent surgery in the early era and 3,149 (43.1%) in the late era (1). In the late era, patients were more likely to have diabetes (50.5% vs. 47.6%, p=0.014), hypertension (86.8% vs. 85.0, p=0.031) and heart failure (31.2% vs. 24.0%, p<0.001) and were more likely to have had a PCI (29.6% vs. 26.2%, p=0.001) in the past. Although rates of other guideline-directed medications including statin therapy and post-operative beta blockers were >95% throughout the study period, the overall rate of post-operative DAPT utilization was 31.2%. In 2011, DAPT use was 16.8% and increased over the subsequent 3 years. DAPT use peaked in 2014 at 35.4% and plateaued for the remainder of the study period (Figure 2).

Figure 1:

Consort diagram depicting derivation of study cohort and exclusion criteria. (CABG: Coronary Artery Bypass Grafting; MI: Myocardial Infarction)

Figure 2:

Trend in dual antiplatelet use over the study period.

Hospital level variability

There was substantial hospital level variability in the rates of DAPT utilization (Range: 9.5%−92.1%). Changes over time varied with 26% (5/19) of centers demonstrating a significant increase, 11% (2/19) demonstrating a significant decrease, and 63% (12/19) without significant change (Figure 3). The percent change in hospital-specific DAPT use varied from −44.6% to +217% over the time period of the study.

Figure 3:

Hospital level changes in dual antiplatelet therapy at discharge. A statistically significant increase (green) or decrease (red) in hospital level DAPT use is depicted by the colored lines. The relative hospital volume is signified by the size of the points. (CABG: Coronary Artery Bypass Grafting; DAPT: Dual antiplatelet use)

Comparison by DAPT status

Overall, 2285 (31.2%) of patients were discharged on DAPT. Patients receiving DAPT were younger (63.0±10.7 vs. 64.4±10.6, p<0.0001), had higher prevalence of peripheral arterial disease (17.5% vs. 12.8%, p<0.0001) and more likely to have undergone PCI in the past (37.1% vs. 23.4%, p<0.0001) (Table 1). In addition, those receiving DAPT were more likely to have had a ST-segment elevation MI (18.1% vs. 15.2%, p<0.0001), received ADP inhibitors preoperatively (22.4% vs. 16.5%, p<0.0001) and underwent off-pump surgery (12.7% vs. 3.7%, p<0.0001). Post-operative, in hospital outcomes by DAPT status are shown in Table 1.

Table 1:

Comparison of preoperative characteristics, operative variables and post-operative outcomes of patients stratified by dual antiplatelet utilization.

	No DAPT	DAPT	p-value
n	5029 (68.8%)	2285 (31.2%)
Preoperative Characteristics
Age	64.4 ± 10.6	63 ± 10.7	<0.0001
Sex (Female)	1335 (26.9%)	669 (29.3%)	0.0386
Hypertension	4258 (84.8%)	2006 (87.9%)	0.0005
Diabetes Mellitus	2428 (48.3%)	1139 (49.9%)	0.2184
Peripheral Arterial Disease	640 (12.8%)	399 (17.5%)	<0.0001
History of Atrial Fibrillation	284 (12.4%)	97 (8.2%)	0.0002
History of Heart Failure	1394 (27.7%)	586 (25.7%)	0.0643
History of Cerebrovascular Accident	411 (8.2%)	250 (11.0%)	0.0001
History of PCI	1175 (23.4%)	848 (37.1%)	<0.0001
Preoperative Hemodialysis	189 (3.8%)	78 (3.4%)	0.4664
Predicted Risk of Mortality	1.53% [0.79–3.35]	1.51% [0.75–3.25]	0.2863
Presentation and Operative Factors
ST Segment Elevation Myocardial Infarction	762 (15.2%)	412 (18.1%)	0.0016
ADP inhibitor within 5 days	829 (16.5%)	511 (22.4%)	<0.0001
Off-Pump Surgery	188 (3.7%)	289 (12.7%)	<0.0001
Cross Clamp Time	67 [52–84]	74 [56–94]	<0.0001
Cardiopulmonary Bypass Time	93 [73–116]	102 [79–126]	<0.0001
Post-operative Medications
Aspirin at Discharge	4878 (97%)	2285 (100%)	<0.0001
ADP Inhibitor at Discharge	75 (1.5%)	2285 (100%)	<0.0001
Discharge Anticoagulant	754 (15.0%)	114 (5.0%)	<0.0001
Post-operative Outcomes
Operative Mortality	34 (0.7%)	10 (0.4%)	0.2216
Major Morbidity	660 (13.1%)	218 (9.5%)	<0.0001
Post-operative Atrial Fibrillation	1192 (23.7%)	426 (18.7%)	<0.0001
Transfusion	1552 (30.9%)	603 (26.4%)	<0.0001
Renal Failure	120 (2.4%)	27 (1.2%)	0.0007
Stroke	69 (1.4%)	34 (1.5%)	0.695
Cardiac Arrest	68 (1.4%)	21 (0.9%)	0.1178
DSWI	15 (0.3%)	8 (0.4%)	0.7129
Reoperation for Bleeding	101 (2.0%)	36 (1.6%)	0.2057
Prolonged Ventilation (>24 Hours)	460 (9.2%)	139 (6.1%)	<0.0001
Post-operative Length of Stay	6 [5–8]	6 [4–8]	<0.0001
Readmission	535 (10.6%)	242 (10.6%)	0.9513

(ADP: Adenosine Diphosphate; DSWI: Deep Sternal Wound Infection; PCI: Percutaneous Coronary Intervention)

Risk-adjusted factors impacting DAPT use

The strongest clinical predictor of DAPT utilization by regression analysis was off-pump surgery (OR 4.48, p<0.0001). Other clinical factors strongly associated with increased DAPT use were prior PCI (OR 2.02, p<0.0001), and recent use of an ADP inhibitor (OR 1.81, p<0.0001, while a history of atrial fibrillation or flutter (OR 0.39, p<0.0001) was associated with decreased DAPT use (Table 2). The overall magnitude of the effect of the treating hospital was relatively important and statistically significant (OR 1.93, p=0.0026), whereas aggregate surgeon effect was not statistically significant (OR 1.24, p=0.1784). Due to the wide center level variability in DAPT use, the influence of each hospital’s practice ranged from odds ratios of 0.21 to 23.0 (Figure 4). In additional risk adjusted analyses, operation dates before or after the 2014 and 2016 guidelines did not significantly impact DAPT utilization (p=0.0634 and p=0.0507, respectively). Finally, due to the potential that discharge on anticoagulation or atrial fibrillation might impact these findings, a sensitivity analysis was performed excluding these populations. The model retains its excellent predictive ability with a C-statistic of 0.918 and while there are slight differences in the magnitude of the effect size, the variables that are significant on regression analysis are largely unchanged when these patients are excluded.

Figure 4:

Hospital specific adjusted odds ratios for dual antiplatelet use at discharge. (DAPT: Dual Antiplatelet Therapy).

Comment

In this analysis of a regional quality database containing over 7,000 patients who underwent CABG following AMI, we observed a relatively low rate of discharge on DAPT after CABG (31.2%). While there was a significant increase in DAPT over time, this plateaued in 2014. This change appears to be in part affected by evolving patient demographics, with a surgical population at higher overall ischemic risk and more patients presenting for CABG after previous PCI. Moreover, there was significant hospital level variability in DAPT use. Although anticipated clinical factors influenced medication choices, hospital culture or policy appears to have a dominant effect on patient treatment.

The primary finding of the study was that although there was an increase in the utilization of DAPT in post-AMI CABG patients, the overall penetration of this guideline-based therapy remains low. Further, neither the 2014 nor the 2016 ACC/AHA guidelines significantly influenced adjusted DAPT rates. These data suggest that there is a persistent gap between guideline-directed medical therapy and real world clinical practice. Although thought to be most important in patients treated with PCI, compliance with guideline-directed medical therapy significantly reduces cardiovascular events after CABG.(17,18) A review of guideline-directed medical adherence in 6 contemporary revascularization trials demonstrated poorer compliance rates in CABG patients compared to patients revascularized with PCI.(19) Given that these data are reported from well-executed clinical trials with significant oversight, one would expect an even greater deviation in real world practice. Therefore, more emphasis should be placed on ensuring that patients treated with CABG receive appropriate secondary prevention therapies to optimize long-term outcomes.(12) In order to promote guideline adherence, post-infarction care pathways need to be established when patients move from cardiology to surgical services in the setting of AMI. While there will certainly be a cohort of patients with clinical contraindications for DAPT and patients may have DAPT initiated at post-discharge follow up, approximately two thirds of the cohort was not receiving DAPT at discharge. Continuing DAPT for 1 year continues to be a class I recommendation, and the surgical community should take responsibility for ensuring that their patients receive optimal medical therapy at least at the time of discharge.(13)

The basis for the low rate of DAPT use is multifactorial, however, in this case, the lack of universal acceptance of DAPT following CABG is most likely the primary cause. Although the benefits of aspirin for secondary prevention are clear, surgeons are unsure whether there is a benefit to adding a second antiplatelet agent.(11,20) Recent data from Zhao et. al. provides some of the strongest evidence for the efficacy of DAPT to improve vein graft patency.(8) This study was done in an elective CABG population and used ticagrelor rather than the more commonly prescribed clopidogrel.(5) Ticagrelor is known to have a more rapid onset with a higher peak platelet inhibition and therefore may be more efficacious than clopidogrel.(21) In addition to the biochemical advantage of ticagrelor, its superior clinical potency has been shown in a subgroup analysis of patients treated with CABG in the PLATO trial. In this study, ticagrelor in combination with aspirin significantly reduced cardiovascular mortality at 1-year compared to clopidogrel with aspirin.(22) Prior single institution and multi-center randomized controlled trials with clopidogrel as the second antiplatelet agent have suggested the benefit of DAPT to both improve early graft patency and limit native coronary artery disease progression.(23–26) Multiple pooled analyses have also demonstrated superiority of DAPT to aspirin, with the most recent meta-analysis including over 11,000 patients and demonstrating the protective effect of DAPT with decreased graft occlusion (Relative Risk: 0.79) and reduction in all-cause mortality (Relative Risk: 0.67).(27–29)

Previous studies have demonstrated a wide variation of knowledge and attitudes surrounding dual antiplatelet use after CABG at the surgeon level.(30,31) However, in this study, we found that the treating hospital plays an unexpectedly large role in determining the antiplatelet regimen that patients receive. Most likely, this reflects the fact that surgeons at the same institution tend to follow similar protocols. Hospital level variation is not a new observation and has been demonstrated by other multicenter groups in several areas including pre-CABG antiplatelet discontinuation and transfusion rates after cardiovascular procedures.(32–34) However, in our current study, inter-hospital consistency in the utilization of other guideline-directed medications, such as statins and aspirin, suggest this is not indicative of a hospital quality gap, but rather a difference of perception as to the appropriateness of DAPT in post-CABG patients.

Beyond the treating hospital, there are a number of clinical factors that influence which patients receive post-operative DAPT. Patients that had received an ADP inhibitor prior to surgery and those who likely have a separate indication for treatment, such as those with peripheral arterial disease or prior PCI, were more likely to receive DAPT. In addition, those who underwent off-pump CABG were much more likely to receive DAPT. Patients that were at higher risk for bleeding events, such as older patients, those with atrial fibrillation and therefore, an indication for anticoagulation, or those requiring post-operative transfusions were less likely to receive DAPT. These findings are generally in concordance with a single center study that found age, preoperative ADP use, preoperative dialysis, off-pump surgery and anticoagulant use to be predictors of DAPT following CABG.(31) These findings suggest that clinicians currently weigh the relative likelihood of ischemic and bleeding events to determine who would most likely benefit from DAPT. Clinical risk scores similar to those that exist for the decision to extend DAPT beyond 1 year after PCI may be beneficial for this decision after CABG.(35)

This study is limited by its retrospective nature, which incorporates an element of selection bias. Regression modeling was used to control for these factors. In addition, the specific agents used are not identifiable, and there are known differences between medications within the same class. We only included a select subset of patients who underwent CABG, those with a recent AMI and therefore, are unable to draw conclusions about the larger population that requires CABG in other settings or with concomitant valve surgery. Finally, conclusions drawn from this regional consortium may not be generalizable to other regions.

In conclusion, although the rate of dual antiplatelet use for patients treated with CABG after AMI has increased in recent years, this has plateaued and much of the change can be attributed to evolving patient characteristics. There is substantial hospital-level prescribing variability that impacts the secondary prevention therapy that patients receive. Multidisciplinary efforts are needed to ensure patients receive optimal, guideline-directed secondary prevention following AMI and should be targeted at the hospital level.

Abbreviations:

ACC

American College of Cardiology

AHA

American Heart Association

AMI

Acute Myocardial Infarction

CABG

Coronary Artery Bypass Grafting

DAPT

Dual Antiplatelet Therapy

IQR

Interquartile Range

PCI

Percutaneous Coronary Intervention

STS

Society of Thoracic Surgeons

VCSQI

Virginia Cardiac Services Quality Initiative