Interventional & Peripheral Cards · Duke University
METHODS
The data source for this project was the American College of Cardiology NCDR Chest Pain‐MI Registry. The NCDR Chest Pain‐MI Registry is a national quality improvement registry that collects information from participating hospitals on patients with AMI in the United States. Data on demographics, medical history, in‐hospital treatment, and in‐hospital outcomes during AMI hospitalization are collected locally by trained data abstractors and submitted to a centralized, secure database operated by the American College of Cardiology. Quality assurance protocols and annual external audits are implemented in order to ensure data accuracy.14 These data will not be made publicly available. Waiver of written informed consent and authorization for this study were granted by Chesapeake Research Review Incorporated.This study included all admissions for patients 18 years and older hospitalized with AMI between January 1, 2019, and March 31, 2023. Patients with type II MI or supply/demand mismatch are excluded from the registry. Patients were excluded if they filled out a short‐from Data Collection Form, which lacked information regarding PAD diagnosis, or were transferred to another hospital within 24 hours. For patients with multiple identifiable admissions, only the index admission was included (Figure S1).The NCDR Chest Pain‐MI Registry defines PAD based on the 2014 American College of Cardiology/American Heart Association Key Data Elements definition, described as atherosclerotic disease involving the infrarenal aorta, iliac, or infrainguinal arteries, as well as history of carotid, renal, mesenteric, or aortic interventions, excluding upper extremity or intracranial vascular diseases.15 The registry collects detailed characteristics including patient demographics (age, sex, race, weight), presentation signs and symptoms (heart failure, heart rate, systolic blood pressure, MI type), medical history (hypertension, diabetes, smoking status, dyslipidemia, prior cardiovascular events and interventions), laboratory results (initial hemoglobin, creatinine clearance, initial troponin ratio), and home medications (aspirin, clopidogrel, warfarin, beta blockers, angiotensin‐converting enzyme inhibitors, angiotensin receptor blockers, aldosterone blockers, statins, and nonstatin lipid‐lowering agents) (Table S1).The primary end points of this study were in‐hospital all‐cause mortality and major bleeding events. Major bleeding events in the registry were defined as suspected bleeding requiring red blood cell transfusion or a procedural intervention, or a decrease in hemoglobin level of >3.0 g/dL.16 In exploratory analyses evaluating secondary outcomes, we assessed cardiac arrest, cardiogenic shock, heart failure, stroke, and new requirement for dialysis.
Crude assessments comparing the cohorts with and without PAD were conducted using chi‐square tests for categorical variables and Wilcoxon rank‐sum test for continuous variables. Standardized differences were used to assess for imbalances in baseline characteristics, with a standardized difference of >10% considered statistically significant.
Generalized estimating equations with logistic regression and adjustments for clustering were used to assess in‐hospital outcomes, with odds ratios (ORs) and 95% CIs used to quantify risk associations, using the group without PAD as a reference. For adjustment in the model, we used the NCDR adjusted mortality and major bleeding models, which include demographics, clinical presentation, medical history, laboratory results, home medications, insurance status, and hospital characteristics (Table S1).17, 18, 19, 20 Tests for interaction with PAD were performed on age (≥65 years versus <65 years), gender, race, AMI type (STEMI versus NSTEMI), revascularization (either percutaneous coronary intervention [PCI] or coronary artery bypass grafting [CABG] versus no revascularization), and presence of multivessel disease.
Besides initial troponin ratio, which was missing in 5.6% of admissions, all other covariates had <3% missing data; missing data were addressed using multiple imputation with 5 imputed data sets employing various fully conditional specification methods for different types of variables. All statistical analyses were performed using SAS version 9.4 software.
RESULTS
The final cohort included 493 740 admissions from 670 hospitals. Of these, 312 638 (63.3%) were for NSTEMI and 181 102 (36.7%) were for STEMI. Patients had a diagnosis of PAD in 36 274 (7.4%) of cases. The cohort with PAD had a higher mean age (70.4 years versus 65.0 years) and higher prevalences of hypertension (92.9% versus 74.6%), dyslipidemia (85.4% versus 60.7%), end‐stage renal disease on dialysis (8.6% versus 2.1%), malignancy (18.1% versus 11.6%), diabetes (57.0% versus 35.3%), and former smoking (42.7% versus 27.4%). The cohort with PAD also had higher rates of prior AMI (36.0% versus 17.8%), PCI (45.2% versus 22.2%), CABG (27.3% versus 9.2%), stroke (18.8% versus 7.1%), heart failure (33.3% versus 11.6%), and atrial fibrillation (19.2% versus 9.4%). Patients with PAD had a higher proportion of NSTEMI (78.1% versus 62.2%) and a lower proportion of STEMI (21.9% versus 37.9%) (Table 1).
Prehospital and In‐Hospital Characteristics of Patients With and Without Peripheral Artery Disease
Other includes all races except for White, Black, and Asian.
Regarding in‐hospital variables, patients with PAD had higher rates of heart failure on admission (23.8% versus 11.2%), multivessel disease (71.2% versus 53.9%), and left main stenosis ≥50% (18.9% versus 8.5%). Cardiogenic shock occurred more frequently among patients with PAD (4.3% versus 3.6%). Patients with PAD had higher rates of femoral access (55.9% versus 43.3%). The rates of PCI (58.8% versus 69.7%) and stent placement (52.8% versus 65.2%) were lower in patients with PAD, and CABG rates were similar (7.7% versus 8.3%) (Table 1). Mechanical support device use was similar between groups, with 5.13% in patients with PAD versus 5.04% in patients without PAD (Table S2).
Patients with PAD had higher use of home aspirin (64.1% versus 37.0%), clopidogrel (35.6% versus 10.0%), and beta blockers (60.0% versus 33.7%). They also more frequently used statins (68.7% versus 41.5%), warfarin (4.5% versus 1.7%), apixaban (9.2% versus 3.8%), and rivaroxaban (3.6% versus 1.4%) (Table S3).
Patients with PAD experienced higher rates of mortality after presenting with AMI (overall: 8.23% versus 4.87%, adjusted OR [aOR], 1.25 [95% CI, 1.19–1.31]), with higher mortality rates after both STEMI and NSTEMI (STEMI: 15.87% versus 7.55%, aOR, 1.46 [95% CI, 1.35–1.58]); NSTEMI: 6.03% versus 3.18%, aOR, 1.14 [95% CI, 1.08–1.21]), interaction P<0.001). Patients without PAD <65 and ≥65 had increased odds of mortality compared with counterparts with PAD (<65: 6.16% versus 2.88%, aOR, 1.42 [95% CI, 1.30–1.56]); ≥65: 9.09% versus 6.76%, aOR, 1.18 [95% CI, 1.12–1.25]) interaction P=0.0223). (Table 2) There was a higher mortality risk in patients with PAD who underwent PCI or CABG, whereas there was no significant difference in mortality between patients with and without PAD who did not undergo revascularization (revascularization: 6.55% versus 3.70%, aOR, 1.49 [95% CI, 1.40–1.59]); nonrevascularized: 11.78% versus 9.23%, aOR, 1.04 [95% CI, 0.97–1.12]) interaction P<0.001) (Table 2, Figures S2 and S3). There were no significant interactions for mortality risk based on sex (interaction P=0.984) or race (interaction P=0.836). Causes of death were similar between cohorts, with higher rates among patients with PAD of death due to heart failure (5.2% versus 3.6%), other cardiovascular reason (9.8% versus 7.3%), and infection (3.2% versus 2.6%) (P<0.001) (Table 3).
Unadjusted and Adjusted Odds of Mortality and Major Bleeding Events in PAD Versus Non‐PAD With Subgroup Analysis Based on Age, MI Type, and Revascularization (PCI or CABG)
Multivariate adjustment was performed based on the NCDR (National Cardiovascular Data Registry) Chest Pain‐MI risk models for mortality and major bleeding. CABG indicates coronary artery bypass grafting; MI, myocardial infarction; OR, odds ratio; PAD, peripheral artery disease; PCI, percutaneous coronary intervention; and STEMI, ST‐segment–elevation myocardial infarction.
Interaction P values for mortality: The effect of PAD on mortality varied by age (P=0.0223), STEMI status (P<0.001), and revascularization status (P<0.001). There was no significant difference by sex (P=0.984) or race (White vs composite of Black, Asian, and Other, P=0.836).
Interaction P‐values for major bleeding events: The effect of PAD on major bleeding varied by revascularization status (P<0.001). There was no significant interaction with age (P=0.413), sex (P=0.836), race (P=0.267), or STEMI vs non‐STEMI (P=0.923).
Causes of Death of Patients With and Without Peripheral Artery Disease Presenting With AMI
AMI indicates acute myocardial infarction; and PAD, peripheral artery disease.
There was an increased risk of major bleeding events among patients with PAD (9.80% versus 6.35%, aOR, 1.23 [95% CI, 1.17–1.28]). Those who underwent revascularization had an increased risk of major bleeding (10.42% versus 6.33%, aOR, 1.31 [95% CI, 1.25–1.37]), whereas there was no significant difference in aOR for major bleeding between nonrevascularized patients with and without PAD (8.51% versus 6.44%, aOR, 1.05 [95% CI, 0.97–1.13]) (interaction P<0.001). There were no significant interactions for major bleeding events with age (interaction P=0.413), sex (interaction P=0.836), or race (interaction P=0.267) (Table 2). The types of bleeding in patients with PAD were comparable to those in patients without PAD, as were in‐patient anticoagulant medications (Table S4). Access site bleeding was slightly higher in patients with PAD (1.0% versus 0.8%), as were gastrointestinal bleeding (1.6% versus 0.8%) and genitourinary bleeding (0.4% versus 0.2%) (Table 4). There was no significant interaction between PAD and multivessel disease for either mortality or major bleeding risks (interaction P=0.360).
Types of Bleeding Events Among Patients With and Without Peripheral Artery Disease Presenting With AMI
AMI indicates acute myocardial infarction; and PAD, peripheral artery disease.
Regarding secondary outcomes, patients with PAD experienced higher rates of in‐hospital cardiac arrest (6.4% versus 4.6%), cardiogenic shock (6.9% versus 4.8%), heart failure (8.0% versus 5.4%), new requirements for dialysis (2.5% versus 1.0%), and ischemic or hemorrhagic stroke (1.4% versus 0.9%) (all P values <0.001) (Figure).
PAD indicates peripheral artery disease.
DISCUSSION
Although prior research has shown poor long‐term outcomes in patients with PAD after AMI hospitalization, this study demonstrates worse in‐hospital outcomes as well, with higher rates of in‐hospital overall mortality, major bleeding events, cardiac arrest, cardiogenic shock, heart failure, new dialysis requirement, and stroke.2, 3, 4, 5, 6, 7, 8, 9, 10 The study's large‐scale, retrospective cohort design uses a large data set with diverse cohorts, improving the generalizability of the findings and enabling us to characterize the cohort with PAD with greater specificity than previous studies, including angiographic findings such as left main stenosis, multivessel disease, and left ventricular ejection fraction. Furthermore, we examined home medication regimens to assess the uptake of guideline‐directed medical therapy for patients with PAD, which inform strategies to improve secondary prevention. Lastly, we analyzed procedural details, including access site choice (radial versus femoral), PCI versus CABG use, and anticoagulation strategies, shedding light on contemporary interventional practices in this high‐risk population. Our study used the NCDR risk adjustment model while conducting subgroup analyses to identify where the increased risk was most pronounced. Beyond assessing mortality and major bleeding events, we evaluated causes of mortality and specific bleeding complications, offering a more nuanced understanding of PAD‐related risks. Subgroup analysis shows that the increased mortality risk is particularly pronounced in younger patients with PAD and in those presenting with STEMI. Additionally, the study highlights the increased risk of major bleeding events in patients with PAD, especially those undergoing revascularization. This study's findings emphasize the importance of preventative care, medical management, and tailored treatment strategies in the acute setting for patients with PAD experiencing AMI.
We suspect that endothelial dysfunction in PAD contributes to adverse outcomes during AMI. Impaired endothelial function reduces nitric oxide production, leading to diminished vasodilation, increased vascular resistance, and compromised blood flow, thereby exacerbating ischemia during AMI.21 Patients with PAD exhibit increased baseline expression of prothrombotic factors, including plasminogen activator inhibitor‐1, von Willebrand factor, and soluble endothelial protein C receptor.22, 23, 24 The dysfunctional endothelium also promotes a proinflammatory state by upregulating adhesion molecules and chemokines, such as macrophage chemoattractant peptide‐1, which enhance platelet activation, leukocyte adhesion, and infiltration, thereby accelerating atherosclerotic plaque formation and progression. Elevated oxidative stress in PAD further depletes nitric oxide bioavailability, exacerbating endothelial dysfunction and contributing to vascular inflammation and plaque instability.25, 26 Our data also demonstrate that patients with PAD are at higher risk for hemorrhagic complications, even after adjusting for antiplatelet therapy. Although this may partly reflect increased procedural complexity in patients with PAD, the magnitude of increased risk suggests that endothelial dysfunction itself appears to contribute to hemorrhagic risk. This question warrants further basic investigation. Collectively, these mechanisms contribute to increased mortality, ischemia, and hemorrhage during AMI in patients with PAD.
The effect of PAD on in‐hospital mortality significantly differed by STEMI status, age, and revascularization status. STEMIs are characterized by transmural infarction of myocardium and represent a more significant physiologic stressor than NSTEMI, with higher overall rates of mortality. Because PAD impairs multiple organ systems, patients with PAD have reduced physiological reserve to withstand the acute stress of STEMI, making them more vulnerable to adverse outcomes.25, 26, 27, 28, 29, 30 Although there was increased absolute mortality >age 65, younger patients with PAD had higher rates of death relative to non‐PAD comparators. Younger patients with PAD represent a population often exposed to risk factors (eg, smoking, diabetes) at an early age with particularly aggressive and advanced atherosclerosis, whereas their counterparts without PAD have yet to experience age‐related vascular dysfunction. Furthermore, aging itself has been shown to cause endothelial dysfunction secondary to oxidative stress, cellular senescence, and telomere dysfunction. Aging seems to have partially equalized the risk between the 2 groups.31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42
Among revascularized patients, PAD was associated with increased adjusted odds of mortality and major bleeding events. Proceduralizing patients with PAD is typically riskier and more complex, as demonstrated here by higher rates of requiring femoral access, blood transfusion, multivessel disease, and left‐main disease.28, 43, 44 Some data suggest worse outcomes among proceduralized patients with PAD, such as a previous study showing increased mortality in patients with PAD who received mechanical circulatory support devices. Consequently, data suggest interventionalists are now placing mechanical circulatory support devices less frequently in patients with PAD, although mechanical circulatory support devices rates were similar in our study and not factored into the adjustment model.45 However, both nonrevascularized cohorts with and without PAD had high crude rates of mortality, suggesting that nonrevascularized patients may have been too ill or hemodynamically unstable to undergo invasive procedures.46 Our data raise the question of when a noninterventional approach to AMI in patients with PAD is preferential to an interventional approach.
These data suggest that physicians should be aware of the presence of PAD as a complicating comorbidity in patients with AMI with higher associated risk for clinical deterioration. Given their likelihood of decompensating, patients with PAD should have a lower threshold for cardiac critical care unit admission than their counterparts. Additionally, outcomes in this patient population may benefit from using transradial access, conservative iodinated contrast and heparin dosing, and adequate postprocedural access site care.47 Anticoagulant medications with shorter half‐lives may be preferable. Patients with PAD in our study were in fact more likely than patients without PAD to receive femoral access, possibly owing to stenosed radial arteries or decreased awareness among interventionalists. For patients with cardiogenic shock and AMI, early revascularization of the culprit artery has been shown to be safer than multivessel PCI in several randomized trials.48, 49, 50 Enhancing individualized risk stratification, particularly for younger patients and those with STEMI, may help identify those needing higher levels of care. Neither the Thrombolysis in Myocardial Infarction score nor the GRACE (Global Registry of Acute Coronary Events) score use history of PAD in risk stratification. Although the HEART score (History, ECG, Age, Risk factors, Troponin) accounts for PAD, it is grouped nonspecifically with other risk factors that do not carry a similar elevation in risk of death and bleeding.51, 52, 53, 54
Outpatient management of PAD is critical in preventing complications from the disease. The 2024 American College of Cardiology/American Heart Association guidelines for managing PAD recommend high‐intensity statin, as well as single antiplatelet therapy with aspirin or clopidogrel for symptomatic patients (Class I). Combining low‐dose rivaroxaban with low‐dose aspirin is advised for further reduction of major adverse cardiac and limb events in symptomatic patients or those who have undergone revascularization with a low risk of bleeding (Class I). For asymptomatic PAD patients, single antiplatelet therapy with aspirin may be considered to reduce major adverse cardiac event risk (Class IIa). Other Class I recommendations include treating hypertension to a goal of <130/80 mm Hg with angiotensin‐converting enzyme inhibitors or angiotensin receptor blockers, tobacco cessation, supervised exercise therapy, and using glucagon‐like peptide‐1 receptor agonists and sodium glucose transporter 2 inhibitors for diabetes treatment.55, 56, 57, 58
Although adherence to evidence‐based therapies was high, with 99.7% of patients with PAD receiving aspirin or clopidogrel and 68.7% prescribed a statin, there remain opportunities for optimization, particularly in tobacco cessation and guideline‐directed pharmacotherapy. Approximately one third of patients with PAD and AMI were active smokers, for example. Although most patients with PAD had a history of hypertension without end‐stage renal disease, only half were on an angiotensin‐converting enzyme inhibitor, angiotensin receptor blockers, or angiotensin receptor‐neprilysin inhibitor, despite their safety even in advanced chronic kidney disease. Of the patients with PAD in this study, only 3.6% were prescribed rivaroxaban, 2.6% a glucagon‐like peptide‐1 receptor agonist, and 3.1% an sodium glucose transporter 2 inhibitor.
This study has several limitations. It relies on data from NCDR hospitals, which, although diverse and representative of the United States, may not fully capture the characteristics of nonparticipating hospitals or patients outside the United States. Data collection was restricted to variables included in the registry's standardized data collection form. PAD was assessed as a binary variable, preventing stratification by disease severity. Consequently, undiagnosed PAD may have been present in the cohort without PAD. Although our model adjusts for multiple confounders, residual confounding may persist due to unmeasured factors such as total length of stay, frailty, and disparities in care. PAD affects multiple organ systems, as do its upstream risk factors, including tobacco use and poorly controlled diabetes. Although our model adjusts for kidney function and common comorbidities, the cumulative burden of unmeasured or inadequately controlled conditions likely contributed to the elevated risk observed in patients with PAD. Caution is warranted in attributing the increased risk solely to PAD itself. The NCDR model accounts for insurance status but does not fully adjust for other aspects of socioeconomic status, limiting its ability to capture patients' broader social context. Finally, the study's retrospective design limits the ability to draw causal inferences. Further research is needed to identify strategies for reducing in‐hospital mortality in this high‐risk population.
CONCLUSIONS
This large‐scale, retrospective cohort study of 493 740 cases of AMI across 670 hospitals demonstrates increased risk of in‐hospital mortality, major bleeding events, cardiac arrest, cardiogenic shock, heart failure, new dialysis requirement, and stroke among patients with PAD hospitalized for AMI. These results emphasize the importance of individualized treatment strategies, enhanced risk stratification, and optimization of medical management for patients with PAD to improve their in‐hospital outcomes.
Sources of Funding
None.
Disclosures
Dr Faridi receives research funding from the National Institutes of Health/National Heart, Lung, and Blood Institute (K23HL161424). Dr Wang contributed to this work while an employee of Duke Clinical Research Institute. Dr Wang is now an employee of the Patient Centered Outcomes Research Institute. The perspective represented in this article does not necessarily represent Patient Centered Outcomes Research Institute'sviews. Dr Mena‐Hurtado is a consultant for Terumo and reports research grants NIH, Shockwave, and Merck. Dr Smolderen receives grants from J&J, Merck, Shockwave, and is a consultant for Happify, Cook, and Terumo. She receives funding from National Institutes of Health/National Heart, Lung, and Blood Institute (R01HL163640‐02S1 and R01HL163640‐01A1) and National Institutes of Health/National Center for Complementary and Integrative Health (R21AT012430‐01). Dr Rymer has received research funding from the National Heart, Lung, and Blood Institute, Novo Nordisk, Chiesi, Abiomed, and Pfizer. Dr Curtis has a contract with the American College of Cardiology for his role as Chief Scientific Advisor of the American College of Cardiology's National Cardiovascular Data Registry. He holds equity interest in Medtronic. Dr Secemsky receives research funding from the National Institutes of Health/National Heart, Lung, and Blood Institute (K23HL150290), Food and Drug Administration, Harvard Medical School's Shore Faculty Development Award, AstraZeneca, BD, Boston Scientific, Cook, CSI, Laminate Medical, Medtronic and Philips. He also is a consultant or speaker for Abbott, Bayer, BD, Boston Scientific, Cook, CSI, Endovascular Engineering, Inari, Janssen, Medtronic, Philips, and VentureMed. The remaining authors have no disclosures to report.
