Incidence and outcomes of high bleeding risk patients with type 1 and type 2 myocardial infarction in a community-based cohort: Application of the Academic Research Consortium High Bleeding Risk Criteria

Abstract

Background and Aims:

The incidence and outcomes of high bleeding risk (HBR) patients in a community cohort according to the Academic Research Consortium (ARC) criteria is not known. We hypothesized that HBR is common and associated with worse outcomes for all-comers with myocardial infarction.

Methods:

We prospectively collected all patients with cardiac troponin T >99^th percentile upper limit of normal (≥0.01 ng/mL) in Olmsted County between 2003–2012. Events were retrospectively classified as type 1 myocardial infarction (T1MI), type 2 myocardial infarction (T2MI), or myocardial injury. Patients were further classified as HBR based on the “ARC-HBR definition.” Outcomes included all-cause mortality, cardiovascular mortality, recurrent MI, stroke, and major bleeding.

Results:

2,419 patients were included in the final study; 1,365 were classified as T1MI and 1,054 as T2MI. Patients were followed for a median of 5.5 years. ARC-HBR was more common in T2MI than T1MI (73% vs 46%, p<0.001). Among patients with T1MI, HBR was associated with higher all-cause mortality (HR 3.7, 95% CI 3.2–4.5, p<0.001), cardiovascular mortality (4.7, 3.6–6.3, p<0.001), recurrent MI (2.1, 1.6–2.7, p<0.001), stroke (4.9, 2.9–8.4, p<0.001), and major bleeding (6.5, 3.7–11.4, p<0.001). For T2MI, HBR was similarly associated with higher all-cause mortality (HR 2.1, 95% CI 1.8–2.5, p<0.001), cardiovascular mortality (2.7, 1.8–4.0, p<0.001), recurrent MI (1.7, 1.1–2.6, p=0.02) and major bleeding (HR 15.6, 3.8–63.8, p<0.001).

Conclusion:

HBR is common among unselected patients with T1MI and T2MI and is associated with increased overall and cardiovascular mortality, recurrent cardiovascular events, and major bleeding on long-term follow up.

Introduction

Anti-platelet therapy reduces the risk of recurrent ischemic events in patients with atherothrombotic heart disease and current guidelines recommend lifelong aspirin for patients with established atherosclerotic coronary artery disease and 12-months of dual antiplatelet therapy (DAPT) for patients with acute coronary syndrome (ACS) regardless of whether percutaneous coronary intervention (PCI) is performed.¹ Although effective at reducing ischemic events, anti-platelet medications increase the risk of major bleeding events, which are associated with worse outcomes.^2,3 This is particularly problematic for patients with a baseline high bleeding risk (HBR).

The optimal management of HBR patients with ischemic heart disease represents a major challenge. HBR patients have been largely excluded from PCI trials, leaving limited data to guide current antiplatelet strategies. Moreover, HBR has not been well-defined and multiple heterogeneous bleeding risk scores have been used in trials, making it challenging to compare studies.^4–8 As a result, the Academic Research Consortium for High Bleeding Risk (ARC-HBR) proposed a standardized definition of HBR based on current literature and expert consensus.⁹ They defined HBR as an annual risk of ≥4% of Bleeding Academic Research Consortium (BARC) 3 or 5 or ≥1% annual risk of an intracranial hemorrhage (ICH).¹⁰ While the ARC-HBR has been validated in several studies of patients undergoing PCI, no studies have examined this definition among unselected patients with acute myocardial infarction (MI), in particular type 2 MI (T2MI). Studies of PCI populations are highly selective and exclude patients at the highest bleeding risk spectrum, making it difficult to generalize the results to a real-world population that clinicians encounter in practice. We elected to study the ARC-HBR among an unselected community-based cohort to provide a more accurate assessment of the true incidence and outcomes associated with this definition. Accordingly, we reviewed all patients with an MI in a defined geographic region over a 10-year period. We determined the incidence and outcomes of HBR, divided into type 1 MI (T1MI) and T2MI. We further analyzed the predictors of mortality and major bleeding among patients with HBR.

Methods

Participants

We have previously reported on the incidence of T1MI and T2MI in a community cohort as part of the Rochester Epidemiology project (REP).¹¹ We now report the incidence of HBR in this population. The REP was established in 1966 to link the medical records of residents within Rochester and Olmsted County, Minnesota.¹² This isolated region is well-suited for epidemiologic studies because care is provided chiefly through the Mayo Clinic and Olmsted Medical Center, allowing for a comprehensive review of care to residents of the county. According to the 2010 US census, Olmsted County had a population of 144,268 (48.8% males) with a median age of 36 years.¹³ Demographics are similar to the rest of the upper Midwest United States: 24% of the population were under age 18 years and 15.9% were above age 65 years. The population was predominantly white (83.6%), with 6.9% black or African American, 6.6% Asian, 5.2% Hispanic or Latino, and 2.5% two or more races. The median household income was $74,880 and 7.1% of the population was in poverty. The design of the current study allowed for prospective collection of all events with cardiac troponin T (cTnT) elevation above the 99^th percentile upper limit of normal (cTnT ≥0.01 ng/mL) in Olmsted County over a 10-year period, as previously described.¹¹

Collected Data

Data was collected between January 1, 2003 and December 31, 2012. Patients with a prior MI were excluded from the study. We collected information on patient demographics and clinical comorbidities through chart review. Comorbidities were determined by the International Classification of Diseases, Ninth Revision (ICD-9) and International Classification of Diseases, Tenth Revision (ICD-10) diagnosis codes.

We collected information on all patients who had a serum cTnT concentration above the 99^th percentile upper reference limit (≥0.01 ng/mL) in a clinic, community, or hospital within Olmsted County. The cTnT was measured using a sandwich electrochemiluminescence immunoassay (Elecsys 2010, Roche Diagnostic Corp). The delta cTnT was defined as the difference (rise or fall) in cTnT concentrations between two measurements less than one week apart. We prospectively collected death certificates, autopsy reports, obituary reports, and electronic death certificates, and the cause of death was determined after review of full medical records by the coroner. We then divided cases into cardiac and noncardiac causes based on the American Heart Association categories for cardiovascular deaths.¹⁴ This study was approved by the Mayo Clinic Institutional Review Board.

Definition of MI

The definition of MI was based on recommendations from the Task Force for the Fourth Universal Definition of Myocardial Infarction, and was defined as a rise and fall in cTnT with at least one concentration >99^th percent upper reference limit with at least one of the following: (1) ischemic symptoms; (2) new/presumed new ECG changes; (3) new imaging evidence of ischemia, such as regional wall motion abnormalities; or (4) identification of intracoronary thrombus on angiogram or autopsy.¹⁵

T1MI was defined as MI from a primary atherothrombotic coronary event. T2MI was defined as MI secondary to a decrease in myocardial oxygen supply or increase in myocardial oxygen demand not attributable to acute atherothrombosis. Type 4 or 5 MI were defined as procedure-related MIs. Cases in which the cause of the cTnT elevated occurred in the absence of clinical evidence of acute myocardial ischemia and which did not fit a previous definition were classified as myocardial injury. These included cases such as heart failure exacerbations, post-defibrillator therapy, myocarditis/pericarditis, and primary noncardiac pathologies such as sepsis, critical illness, and renal failure.

Classification of MI

All cases of MI were retrospectively reviewed by two independent cardiologists (C.E.R and R.G) to determine the cause of cTnT elevation, as previously described.¹¹ Information reviewed included clinical notes, electrocardiography, coronary angiography, echocardiography, computerized tomography, and magnetic resonance imaging if available. Patients were then characterized as TIMI, T2MI, type 4 MI, type 5 MI, or myocardial injury. T2MI was further subcategorized into those secondary to anemia versus those secondary to other causes. Any difference in classification was resolved through consensus. Furthermore, classification of comorbidities was performed by a single reviewer.

Exclusion criteria

We excluded patients who withdrew consent for research (n=11), patients with incomplete records or unclear cause of troponin rise (n=29), patients with previous MI (n=543), patients with type 4 or 5 MI (n=17), and patients with myocardial injury (n=2441).

Definition of HBR

The classification of HBR among patients with MI was based on the consensus definition proposed by the “Academic Research Consortium for High Bleeding Risk,” as detailed in Supplemental Table 1.⁹ Patients were defined as HBR if at least 1 major or 2 minor criteria were met at the time of the event. Anemia and thrombocytopenia were determined by admission labs. Long-term use of oral NSAIDs or steroids were determined through medication records. The rest of the HBR criteria were determined through review of ICD-9 and ICD-10 diagnosis codes. During follow up, cause of death was determined after review of patient records, death certification and autopsy were performed. Bleeding outcomes and stroke were classified according to the BARC classification based on ICD-9 and −10 diagnosis codes.¹⁰

Statistical analysis

Data are reported as mean ± standard deviation for continuous variables and number (%) for categorical variables. Comparisons were made using the one-way analysis of variance (ANOVA) for continuous variables and Pearson’s chi-squared test for categorical variables. Cumulative incidence curves were constructed to model outcomes of interest, including all-cause mortality, cardiovascular mortality, recurrent MI, stroke, and major bleeding. Death was a competing risk for recurrent MI, stroke and major bleeding, and non-cardiovascular death was a competing risk for cardiovascular mortality. In the presence of competing risks, these curves represent the marginal event rates for the event of interest. Gray’s test was used to make comparisons between curves.

Univariate Cox proportional hazards models and multivariable with adjustments for age, sex, and comorbidities were used to assess the effect of ARC-HBR status on all-cause mortality, cardiovascular mortality, recurrent MI, stroke, and major bleeding episodes. All comorbidities from table 2 were included in the multivariable models. Due to there being a limited number of events for the bleeding outcome, eight variables were selected based on their association with HBR and clinical relevance. All Cox model proportional hazards assumptions were assessed using plots of the scaled Schoenfeld residuals against time. No sensitivity analyses were conducted. Statistical significance was defined as a 2-tailed P value <0.05. Statistical analyses were performed using R (version 4.1.2).

Table 2.

Baseline characteristics including demographics, comorbidities, inpatient revascularization, and discharge medications according to myocardial infarction (MI) subset and presence of high bleeding risk (HBR).

	Type 1 MI (n = 1365)			Type 2 MI (n=1054)
Demographics	HBR (n=628)	No HBR (n=737)	P-value	HBR (n=769)	No HBR (n=285)	P-value
Age	77.1 ± 12.6	61.1 ± 12.5	<0.001	77.4 ± 13.6	63.8 ± 16.8	<0.001
Female	303 (48.2%)	201 (27.3%)	<0.001	438 (57.0%)	130 (45.6%)	<0.001
BMI	26.8 ± 8.9	28.6 ± 17.8	0.07	27.9 ± 17.8	26.5 ± 10.2	0.37
Comorbidities
Prior CABG	94 (15.0%)	55 (7.5%)	<0.001	70 (9.1%)	10 (3.5%)	0.002
Prior PCI	96 (15.3%)	76 (10.6%)	0.006	49 (6.4%)	13 (4.6%)	0.27
Number stents	0.2 ± 0.8	0.2 ± 0.6	0.12	0.09 ± 0.5	0.05 ± 0.3	0.18
Number DES	0.06 ± 0.4	0.06 ± 0.4	0.89	0.03 ± 0.2	0.02 ± 0.2	0.7
Number BMS	0.2 ± 0.7	0.1 ± 0.5	0.07	0.06 ± 0.4	0.03 ± 0.2	0.17
Hypertension	513 (82.1%)	453 (62.1%)	<0.001	547 (71.5%)	169 (59.7%)	<0.001
Hyperlipidemia	373 (59.6%)	417 (57.1%)	0.36	275 (35.9%)	84 (29.7%)	0.06
Diabetes mellitus	185 (29.6%)	128 (17.5%)	<0.001	124 (16.2%)	26 (9.2%)	0.004
Prior stroke	36 (5.7%)	11 (1.5%)	<0.001	16 (2.1%)	5 (1.8%)	0.74
CKD	39 (6.2%)	5 (0.7%)	<0.001	126 (16.4%)	15 (5.3%)	<0.001
Liver disease	12 (1.9%)	17 (2.3%)	0.61	29 (3.8%)	10 (3.5%)	0.84
Heart failure	21 (3.3%)	5 (0.7%)	<0.001	78 (10.1%)	8 (2.8%)	<0.001
COPD	107 (17.0%)	52 (7.1%)	<0.001	134 (17.4%)	29 (10.2%)	0.004
Current smoker	47 (7.5%)	122 (16.6%)	<0.001	57 (7.4%)	37 (13.0%)	0.08
Cancer	160 (25.5%)	65 (8.8%)	<0.001	200 (26.0%)	29 (10.2%)	<0.001
PUD	25 (4.0%)	17 (2.3%)	0.07	41 (5.3%)	11 (3.9%)	0.33
Procedures
PCI during admission	190 (30.3%)	357 (48.4%)	<0.001	40 (5.2%)	18 (6.3%)	0.48
CABG during admission	111 (17.7%)	117 (15.9%)	0.374	52 (6.8%	9 (3.2%)	0.03
Discharge antiplatelet or anticoagulation
Aspirin	448 (71.3%)	666 (90.4%)	<0.001	597 (77.6%)	232 (81.4%)	<0.001
Clopidogrel	323 (51.4%)	534 (72.5%)	<0.001	98 (12.7%)	74 (26.0%)	<0.001
Ticagrelor	10 (1.6%)	25 (3.4%)	0.04	1 (0.1%)	2 (0.7%)	0.12
Oral Anticoagulation	109 (17.4%)	58 (7.9%)	<0.001	190 (24.7%)	49 (17.2%)	0.01
No antiplatelet or OAC	56 (8.9%)	23 (3.1%)		93 (12.1%)	30 (10.5%)	<0.001
1 agent	252 (40.1%)	172 (23.3%)		348 (45.3%)	125 (43.9%)
2 agents	273 (43.5%)	486 (65.9%)		186 (24.2%)	124 (43.5%)
3 agents	47 (7.5%)	56 (7.6%)		12 (1.6%)	6 (2.1%)
Duration of dual therapy	304.5±130.1	174.6±256.4		117.6±164.4	107.4±156.6	<0.001
Other discharge medications
Statin	487 (77.5%)	680 (92.3%)	<0.001	400 (52.0%)	170 (59.6%)	0.03
Other lipid lowering	108 (17.2%)	188 (25.5%)	<0.001	83 (10.8%)	31 (10.9%)	0.97
Calcium channel blocker	119 (18.9%)	97 (13.2%)	0.003	162 (21.1%)	77 (27.0%	0.04
Beta blocker	538 (85.7%)	677 (91.9%)	<0.001	563 (73.2%)	203 (71.2%)	0.52
Isosorbide mononitrate	414 (65.9%)	613 (83.2%)	<0.001	223 (29.0%)	82 (28.8%)	0.94
Diuretics	469 (74.7%)	391 (53.1%)	<0.001	548 (71.3%)	172 (60.4%)	<0.001

Abbreviations: ACE-I = Angiotensin-converting enzyme inhibitor; ARB = Angiotensin II receptor blockers; BMI = body mass index; BMS = bare metal stent; CI = confidence interval; CABG = coronary artery bypass graft; CKD = chronic kidney disease; COPD = chronic obstructive pulmonary disease; DES = drug eluting stent; HBR = high bleeding risk; MI = myocardial infarction; OAC = oral anticoagulation; PCI = percutaneous coronary intervention; PUD = peptic ulcer disease.

Results

Study population

We prospectively collected 5,460 patients with at least one cTnT >99^th percentile upper reference limit in Olmsted County over a 10-year period. We excluded 583 patients including 543 with a previous MI, 29 with incomplete records or an unclear etiology, and 11 who withdrew research consent. We reviewed the remaining 4,877 patients and excluded an additional 2,441 patients determined to have myocardial injury and 17 patients with type 4 or 5 MI, yielding a final study population of 2,419 unique patients (Figure 1). The mean age of included patients was 70.8±15.5 years and 1,072 (44.3%) were female.

Figure 1.

Consort diagram demonstrating the inclusion and exclusion criteria for the final study population.

T1MI was the underlying etiology in 1,365 (56.4%) patients and T2MI in 1,054 (43.6%) patients. Baseline characteristics according to MI type are shown in Table 2.

Outcomes according to bleeding risk

The breakdown of major and minor criteria for HBR for the T1MI and T2MI groups are shown in Table 1. Of the total population, 1,397 (57.8%) met criteria for HBR. 628 (46.0%) patients with T1MI were HBR, as compared to 769 (73.0%) patients with T2MI (p<0.001) (Table 2). The number of patients meeting major and minor criteria for HBR among patients with T1MI was 404 (64.3%) and 375 (59.7%), respectively, as compared to 580 (75.4%) and 387 (50.3%) among patients with T1MI (both p<0.001).

Table 1.

Number of patients with type 1 myocardial infarction (T1MI) and type 2 MI (T2MI) meeting individual components of the Academic Research Consortium-High Bleeding Risk (ARC-HBR) criteria.

Criteria	T1MI (N = 628)	T2MI (N = 769)	P-value
Major
Anticipated use of long-term oral anticoagulation	38 (6.1%)	30 (3.9%)	0.06
Severe or end-stage CKD (eGFR <30 mL/min)	134 (21.3%)	244 (32.7%)	<0.001
Spontaneous bleeding requiring hospitalization or transfusion in the past 6 mo or at any time, if recurrent	34 (5.4%)	69 (9.0%)	0.01
Hemoglobin <11 g/dL	147 (23.4%)	334 (43.4%)	<0.001
Moderate or severe baseline thrombocytopenia† (platelet count <100 × 109/L)	20 (3.2%)	71 (9.2%)	<0.001
Previous spontaneous ICH (at any time)	22 (3.5%)	21 (2.7%)	0.41
Previous traumatic ICH within the past 12 months	1 (0.2%)	4 (0.5%)	0.26
Presence of a bAVM	0 (0.0%)	0 (0.0%)	N/A
Moderate or severe ischemic stroke within the past 6 months	29 (4.6%)	26 (3.4%)	0.24
Chronic bleeding diathesis	19 (3.0%)	30 (3.9%)	0.38
Liver cirrhosis with portal hypertension	2 (0.3%)	0 (0.0%)	0.12
Active malignancy (excluding nonmelanoma skin cancer) within the past 12 months	59 (9.4%)	70 (9.1%)	0.85
Nondeferrable major surgery on DAPT	7 (1.1%)	11 (1.4%)	0.60
Recent major surgery or major trauma within 30 d before PCI	4 (0.6%)	12 (1.6%)	0.11
Minor
Age ≥75 years	416 (66.2%)	503 (65.4%)	0.74
Moderate CKD (eGFR 30–59 mL/min)	335 (53.3%)	348 (45.3%)	0.003
Spontaneous bleeding requiring hospitalization or transfusion within the past 12 mo not meeting the major criterion	2 (0.3%)	0 (0.0%)	0.12
Hemoglobin 11–12.9 g/dL for men and 11–11.9 g/dL for women	160 (25.5%)	149 (19.4%)	0.006
Long-term use of oral NSAIDs	3 (0.5%)	0 (0.0%)	0.06
Long-term use of oral steroids	6 (1.0%)	2 (0.3%)	0.09
Any ischemic stroke at any time not meeting the major criterion	117 (18.6%)	110 (14.3%)	0.03

Abbreviations: bAVM = brain arteriovenous malformation; CKD = chronic kidney disease; DAPT = dual antiplatelet therapy; eGFR = estimated glomerular filtration rate; HBR = high bleeding risk; ICH = intracranial hemorrhage; NSAID = nonsteroidal anti-inflammatory drug; PCI = percutaneous coronary intervention.

In T1MI

Among patients with T1MI, there were 638 (47%) deaths during a median follow-up of 5.5 (interquartile range 3.1–7.9) years. Death was most common within the first 3 months after MI (Figure 2A). On univariate analysis, HBR was associated with higher all-cause mortality (HR 3.7, 95% CI 3.2–4.5, p<0.001) and cardiovascular mortality (HR 4.7, 95% CI 3.6–6.3, p<0.001). Patients with HBR similarly had an elevated risk of MI (HR 2.1, 95% CI 1.6–2.7, p<0.001), stroke (HR 4.9, 95% CI 2.9–8.4, p<0.001), and major bleeding (HR 6.5, 95% CI 3.7–11.4, p<0.001) at 6 years. The cumulative incidence curves for patients with T1MI according to bleeding risk are shown in Figure 2. At discharge, patients with HBR were less likely to be prescribed aspirin, clopidogrel, and/or ticagrelor (all p<0.05). They were less likely to be discharged on dual therapy (43.5% vs 65.9%, p<0.001) and were more likely to be discharged on no antiplatelet or oral anticoagulation (8.9% vs. 3.1%, p<0.001).

Figure 2.

Cumulative incidence curves demonstrating the estimated (A) all-cause mortality, (B) cardiovascular mortality, (C) future myocardial infarction, (D) future stroke, and (E) major bleeding events for patients with type 1 myocardial infarction.

In T2MI

Among patients with T2MI, 766 (73%) patients died during follow-up. Death was most common within the first 3 months after MI (Figure 3A). On univariate analysis, HBR was associated with higher all-cause mortality (HR 2.1, 95% CI 1.8–2.5, p<0.001) and cardiovascular mortality (HR 2.7, 95% CI 1.8–4.0, p<0.001). HBR was similarly associated with an increased risk of MI (HR 1.7, 95% CI 1.1–2.6, p=0.02), stroke (HR 22.8, 95% CI 3.1–165.6, p=0.002), and major bleeding (HR 15.6, 95% CI 3.8–63.8, p<0.001) at 6 years. The wide CI around the HR for stroke is due to only 1 stroke in the non-HBR group (42 in the HBR group). The cumulative incidence curves for patients with T2MI according to bleeding risk are shown in Figure 3. At discharge, patients with HBR were less likely to be prescribed aspirin or clopidogrel (both p<0.001). They were also less likely to be discharged on dual therapy (24.2% vs. 43.5%, p<0.001).

Figure 3.

Cumulative incidence curves demonstrating the estimated (A) all-cause mortality, (B) cardiovascular mortality, (C) future myocardial infarction, (D) future stroke, and (E) major bleeding events for patients with type 2 myocardial infarction.

Outcomes for subsets of T1MI

We further divided the T1MI population into those who underwent revascularization with PCI and/or CABG and those who were medically managed and compared outcomes using univariable analysis. Patients who underwent PCI or CABG were less likely to meet criteria for HBR compared to patients that did not undergo PCI/CABG (38% vs 56%, p<0.001). For both patients undergoing PCI/CABG and those who were managed with medical therapy alone, HBR remained associated with increased all cause and cardiovascular mortality as well as MI, stroke and major bleeding. The results are summarized in Supplemental Table 2.

Predictors of mortality and major bleeding among patients with T1MI

We assessed the predictors of mortality and major bleeding among patients with T1MI using multivariable analysis (Table 3). HBR was associated with a higher risk of all-cause mortality (HR 1.6, 95% CI 1.3–2.0, p<0.001), cardiovascular mortality (HR 1.8, 95% CI 1.3–2.6, p<0.001), and major bleeding (HR 6.7, 95% CI 3.6–12.5, p<0.001). There were no predictors of major bleeding aside from HBR on multivariable analysis. Bleeding events occurred early after MI with most occurring during the first month (Figures 2, 3 and S1).

Table 3.

Major predictors of all-cause mortality, cardiac mortality, and major bleeding among patients with type 1 myocardial infarction using multivariable analysis.

	All-cause mortality		Cardiovascular Mortality		Major Bleeding
Variable	Hazard Ratio (95% CI)	P value	Hazard Ratio (95% CI)	P value	Hazard Ratio (95% CI)	P value
HBR	1.57 (1.28–1.93)	<0.001	1.80 (1.28–2.53)	<0.001	6.69 (3.57–12.5)	<0.001
Prior CABG	1.22 (0.96–1.56)	0.10	1.38 (0.97–1.95)	0.08	1.64 (0.89–3.04)	0.11
Prior PCI	1.20 (0.71–2.04)	0.49	1.06 (0.48–2.32)	0.89	0.81 (0.19–3.54)	0.78
Age	1.06 (1.05–1.07)	<0.001	1.07 (1.05–1.08)	<0.001	1.0 (0.98–1.01)	0.67
Male sex	0.90 (0.76–1.07)	0.24	0.96 (0.74–1.26)	0.79	1.06 (0.65–1.73)	0.80
Hypertension	0.69 (0.57–0.85)	<0.001	0.59 (0.43–0.80)	<0.001	1.17 (0.64–2.16)	0.61
Diabetes mellitus	1.42 (1.25–1.84)	<0.001	1.58 (1.17–2.13)	0.003	0.68 (0.38–1.23)	0.20
Prior stroke	1.68 (1.16–2.43)	0.006	1.70 (0.97–2.97)	0.06	0.68 (0.16–2.84)	0.60
Hyperlipidemia	0.80 (0.67–0.95)	0.01	0.68 (0.68–1.17)	0.41
CKD	1.57 (1.10–2.25)	0.01	1.94 (1.22–3.09)	0.005
Liver disease	1.76 (1.08–2.88)	0.02	0.75 (0.24–2.41)	0.63
Heart failure	1.67 (1.04–2.68)	0.04	2.55 (1.45–4.51)	0.001
COPD	1.78 (1.45–2.18)	<0.001	1.69 (1.23–2.33)	0.001
Cancer	0.92 (0.76–1.11)	0.39	0.79 (0.58–1.07)	0.13

Abbreviations: CI = confidence interval; CABG = coronary artery bypass graft; CKD = chronic kidney disease; COPD = chronic obstructive pulmonary disease; HBR = high bleeding risk; MI = myocardial infarction; PCI = percutaneous coronary intervention; PUD = peptic ulcer disease, T2MI = type 2 myocardial infarction

Predictors of mortality and major bleeding among patients with T2MI

We similarly analyzed the predictors of mortality and major bleeding among patients with T2MI using multivariable analysis (Table 4). HBR was associated with a higher risk of all-cause mortality (HR 1.3, 95% CI 1.1–1.6, p=0.01) and major bleeding (HR 17.7, 95% CI 4.2–74.6, p<0.001), but was not associated with an elevated risk of cardiovascular mortality (HR 1.3, 95% CI 0.8–1.9, p=0.28).

Table 4.

Major predictors of all-cause mortality, cardiac mortality, and major bleeding among patients with type 2 myocardial infarction using multivariable analysis.

	All-cause Mortality		Cardiovascular Mortality		Major Bleeding
Variable	Hazard Ratio (95% CI)	P value	Hazard Ratio (95% CI)	P value	Hazard Ratio (95% CI)	P value
HBR	1.31 (1.08–1.61)	<0.001	1.28 (0.84–1.94)	0.25	17.7 (4.22–74.55)	<0.001
Prior CABG	1.09 (0.83–1.44)	0.53	1.63 (0.99–2.67)	0.05	1.35 (0.62–2.95)	0.45
Prior PCI	1.42 (0.80–2.53)	0.24	1.57 (0.55–4.47)	0.4	1.13 (0.26–4.96)	0.87
Age	1.03 (1.03–1.04)	<0.001	1.06 (1.05–1.08)	<0.001	1.00 (0.98–1.02)	0.98
Male sex	1.01 (1.03–1.04)	0.91	0.95 (0.70–1.28)	0.74	1.75 (1.07–2.86)	0.03
Hypertension	0.64 (0.55–0.76)	<0.001	0.66 (0.47–0.93)	0.02	0.66 (0.38–1.12)	0.12
Diabetes mellitus	1.28 (1.04–1.58)	0.02	1.17 (0.76–1.80)	0.49	0.76 (0.37–1.57)	0.46
Prior stroke	1.89 (1.16–3.07)	0.01	1.78 (0.65–4.86)	0.26	3.75 (1.16–12.16)	0.03
Hyperlipidemia	0.92 (0.78–1.08)	0.29	1.15 (0.84–1.56)	0.39
CKD	1.31 (1.07–1.60)	0.009	1.12 (0.72–1.75)	0.62
Liver disease	1.21 (0.81–1.82)	0.35	1.84 (0.84–4.04)	0.13
Heart failure	1.23 (0.95–1.60)	0.12	1.78 (1.12–2.80)	0.01
COPD	1.55 (1.29–1.86)	<0.001	1.45 (1.00–2.11)	0.05
Cancer	1.17 (0.99–1.39)	0.07	0.96 (0.68–1.36)	0.83

Abbreviations: CI = confidence interval; CABG = coronary artery bypass graft; CKD = chronic kidney disease; COPD = chronic obstructive pulmonary disease; HBR = high bleeding risk; MI = myocardial infarction; PCI = percutaneous coronary intervention; PUD = peptic ulcer disease, T2MI = type 2 myocardial infarction

Discussion

This is the first geographic cohort study to report the incidence and outcomes of HBR in an unselected population of all comers with acute MI using the ARC definition for HBR. The major findings of this study are: First, the incidence of HBR is higher than previous selected cohorts of patients undergoing PCI have suggested and comprises almost 75% of patients with T2MI and 50% of patients with T1MI. Second, HBR is associated both with an elevated risk of cardiac mortality, recurrent MI and stroke, and an elevated risk of major bleeding over 6 years. All-cause mortality in the HBR group was high, with >50% of those with T1MI and >75% of those with T2MI deceased over 6 years. Last, HBR was an accurate predictor of major bleeding events for patients with both T1MI and T2MI.

Incidence of HBR among patients with T1MI and T2MI

Previous studies of HBR using the ARC-HBR definition have mostly focused on patients undergoing PCI, with two studies of patients with suspected ACS. This study design typically excludes most patients with T2MI who were admitted with another primary diagnosis.^16–36 Patients undergoing PCI are already a selected population and would not capture patients with T1MI treated with medical therapy and those in whom PCI is not undertaken due to HBR.^37,38 As a result of these limitations, the estimated incidence of HBR according to the ARC-HBR standardized definition varies between 17% to 60% of patients undergoing PCI.^18,19,23,35 Our unselected community-based study design allowed estimation of the true incidence of HBR in ACS. We found the incidence of HBR to be higher than previously reported, with 73% of patients with T2MI and 46% of patients with T1MI meeting criteria for HBR. This is likely due to our broad inclusion criteria of all patients with raised troponin. Prior studies of HBR were primarily in patients admitted with an MI and would not capture severely ill patients admitted under internal medicine or intensive care with index non-cardiac admission.

HBR and mortality

HBR was associated with worse all-cause and cardiovascular mortality for patients with T1MI and T2MI, with most events happening early (within 3 months) after the index MI. As expected, HBR also predicted major bleeding events, the majority of which happened early. Similar to TRACER (randomized control trial [RCT] of vorapaxar [inhibitor of thrombin-induced activation] vs placebo), PLATO (RCT of clopdiogrel vs ticagrelor) and the Health Maintenance Organization Research Network Stent registry, we found that the risk of bleeding was highest in the first 30 days following MI, while highest risk of recurrent MI/cardiovascular death was in the first 2 months post MI.^39–41 Patients who meet HBR criteria are also at heightened risk of cardiovascular death/recurrent cardiovascular events, making the decision to prescribe anti-platelet therapy particularly challenging. Patients with T1MI and HBR were less likely to be prescribed DAPT, including P2Y12 inhibitors, at discharge in our study. In addition, both patients with T1MI and T2MI were less likely to be prescribed aspirin. Prior studies showed HBR to be a poor prognostic factor with all-cause mortality ranging from 4.7% at 1-year to 31% at 5-years.^17,22–25 The mortality observed in our population was considerably higher, with nearly 60% of T1MI and 75% of T2MI patients estimated to be deceased after 6 years, which is most likely due to our inclusive study design. Cardiovascular mortality in our community-based cohort was similarly higher, with nearly 25% of T1MI and 15% of T2MI patients with fatal cardiac events within 6 years. The mortality rates observed in our population are striking and emphasize the medical complexity and vulnerability of patients with HBR. Although previous studies have correlated HBR with worse outcomes, the true magnitude of this correlation has likely been underestimated due to exclusion of high-risk patients for PCI and omission of patients who experience MI in the context of another illness who are not captured in cardiology databases.

HBR and major bleeding events in T1MI

Since HBR predicts not only bleeding events but also patients at the highest risk for cardiovascular mortality and recurrent events, awareness of the interplay and risk profile underscores the importance of cardiology involvement in care of these patients. Our data provide long term CV mortality and bleeding rates in the HBR population, allowing clinicians clearer data on which to base decisions regarding DAPT therapies.

HBR and major bleeding events in T2MI

T2MI is an understudied population, and we were the first to demonstrate that the HBR score is predictive in this group. Patients with T2MI are more commonly older with higher rates of diabetes and renal failure, while these patients will also have higher risk of bleeding events due to their older age and higher comorbidity burden, in particular anemia, thrombocytopenia, and CKD.¹¹ Patients with T2MI commonly have coronary artery disease and may have a primary indication for antiplatelet therapy, however this will need to be balanced against their bleeding risk. However, the long-term treatment strategies for patients with T2MI and no CAD remains uncertain and there are no current guidelines.^{11,17,23,24,42}

Revascularization and Optimal Medical Therapy in T1MI

In our subgroup analysis of patients with T1MI, we found that HBR was associated with significantly higher rates of mortality, cardiovascular events, and major bleeding for patients who underwent revascularization or conservative management with optimal medical therapy. Interestingly, the hazard ratios were higher for all-cause mortality, cardiovascular mortality, myocardial infarction, and stroke among those treated medically while the hazard ratios were similar for major bleeding. This would suggest that patients at HBR could benefit from upfront revascularization; however, these patients were significantly more likely to be medically managed in our population, presumably due to the perceived necessity of DAPT after PCI. Future studies are needed to confirm that patients with HBR and T1MI do better with a revascularization strategy.

Clinical implications and future studies

Our community-based cohort demonstrated that HBR is more common than previously appreciated and present in more than half of the total population of all-comers with acute MI. HBR is a poor prognostic factor and is associated with worse all-cause mortality, cardiac mortality, recurrent ischemic events, and major bleeding episodes. While previously validated among patients undergoing PCI, we can now extend this validation to include all-comers with T1MI and T2MI.

The results of our findings emphasize that there is a need for improvement in the recognition and management of HBR patients with ischemic heart disease, as they are at elevated risk of both major bleeding as well as fatal cardiovascular events early after MI. We now recognize that drug eluting stents can be safely implanted with short duration of anti-platelets (1–3 months), and more recent trials have tested the efficacy of single anti-platelet therapy post PCI.^36,43–47 In our study of those with HBR and T1MI, almost half of patients with discharged on dual or triple therapy and the mean duration of dual therapy was around 10 months. Routine use of the ARC-HBR score may help identify patients who would be better served for anti-platelet monotherapy or shorter durations of DAPT, thus providing a lower risk of major bleeding while also providing protection against future cardiovascular events. Future RCTs are needed to determine the optimal use of anti-platelet agents among those with HBR, including the number of agents as well as duration of therapy. Future studies are also needed to confirm our findings among more diverse geographic and ethnic populations as well as with the newer 5^th generation high sensitivity troponin.

Limitations

This study was conducted in a single community cohort and the findings may not be generalizable to populations with different racial and ethnic backgrounds. Second, our data are from 2003–2012 and management of MI, PCI practice, and anti-platelet agents have advanced since this time. T1MI and T2MI were classified on the basis of clinical criteria and not all patients had coronary angiography, thus it is possible that some of T2MIs were unrecognized T1MIs. The study period was before the transition to the use of 5^th generation (high sensitivity) troponin assays, which may have changed the incidence of T1MI and T2MI. Due to the inherent nature of this retrospective cohort study of a single community as well as exclusion of a small number of cases with incomplete data, selection bias may be contributing to the results. Finally, the list of medications is based on prescriptions issued and did not capture patient adherence or changes made to medical therapy after discharge.

Conclusion

This epidemiologic study demonstrates the incidence and outcomes of HBR among an unselected population of patients with acute MI. HBR is prevalent among patients with T1MI and T2MI and is a strong predictor of bleeding risk and cardiac events in both populations. Our study suggests that the ARC-HBR definition should be used to identify patients who may benefit from single anti-platelet therapy or a shorter duration of DAPT. Future studies are needed to determine the optimal therapeutic strategies among this vulnerable population.

Highlights:

• Approximately 75% of patients with type 2 myocardial infarction and 50% of patients with type 1 myocardial infarction are at high bleeding risk.

• High bleeding risk is associated with elevated cardiac mortality, recurrent myocardial infarction, stroke, and major bleeding over 6 years.

• All-cause mortality is high among patients with high bleeding risk, with >50% of those with type 1 and >75% with type 2 myocardial infarction deceased over 6 years.