Bringing doctors together

or

A free, private network for verified physicians

Leading voices

Dr. Dorairaj Prabhakaran

General Cardiology

View

Dr. Trey Woods

General Cardiology · Emory Healthcare

View

Dr. Rishi Wadhera

General Cardiology · BIDMC

View

Dr. Robert Harrington

General Cardiology · WCM

View

Dr. Mike Gaziano

General Cardiology · BWH

View

Dr. Tyler Gluckman

General Cardiology · Providence

View

Dr. Cayla Pichan

General Cardiology · UT Southwestern

View

Dr. Sanjiv Kaul

General Cardiology · OHSU

View

Dr. Emily Lau

General Cardiology · MGH

View

Dr. Sarah Ashley

General Cardiology · Emory University Hospital

View
  • Discover your community

  • Stay current

  • A profile that represents you

  • Find your people

About RoonPrivacyTerms

    Overview

    Acute Limb Ischemia After Percutaneous Coronary Intervention for Stable Coronary Artery Disease.

    Apr 23, 2025DOI: 10.1161/jaha.124.040026Journal ArticleAHA

    Written by: Monil MajmundarWan-Chi ChanVivek BhatMohammed AbualenainKunal N PatelGokul RamaniGaurav ParmarKamal Gupta

    Find these authors on Roon

    VB
    Dr. Vivek Bhat
    General Cardiology · SUNY Upstate Medical University

    METHODS

    The data that support the findings of this study are available from the corresponding author on reasonable request.We derived our study population from the Nationwide Readmissions Database (NRD).9 The NRD is drawn from the Healthcare Cost and Utilizations Project (HCUP) State Inpatient Databases, containing data from US hospitals in 28 different states. Representing ≈58.7% of all hospitalizations in the United States, it uses a deidentified unique number that enables tracking each patient and their readmissions across hospitals within a calendar year. We included all hospitalizations in the NRD from January 1, 2016, to December 31, 2020. We did not include data before 2016 to avoid mixing International Classification of Diseases, Ninth Revision, Clinical Modification (ICD‐9‐CM) codes with International Classification of Diseases, Tenth Revision, Clinical Modification (ICD‐10‐CM) codes.This study was deemed exempt from review by the University of Kansas Medical Center institutional review board, and informed consent was not required, as the NRD contains deidentified patient information with prior ethical committee approval. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines (Table S1).We included all patients undergoing PCI with stable CAD. For this, we searched the NRD using appropriate ICD‐10‐CM codes for all patients who underwent PCI within the study period. From this initial sample, we excluded patients undergoing concurrent percutaneous procedures (transcatheter aortic valve replacement, extracorporeal membrane oxygenation, and percutaneous mitral valve repair/replacement), then excluded patients undergoing PCI for acute coronary syndrome.From this final sample, we compared the characteristics of patients with post‐PCI ALI with those who had ALI during the index admission. Figure 1 depicts the patient selection and exclusion process.ACS indicates acute coronary syndrome; CAD, coronary artery disease; ECMO, extracorporeal membrane oxygenation; NRD, Nationwide Readmissions Database; PCI, percutaneous coronary intervention; and TAVR, transcatheter aortic valve replacement.Our primary outcome was the yearly incidence of ALI and predictors of ALI among patients with stable CAD who underwent PCI.
    Secondary outcomes were in‐hospital mortality, in‐hospital amputation, composite of in‐hospital mortality and amputation, and unplanned all‐cause and ALI‐related readmissions at 30 days and 6 months. For post‐discharge outcomes, only patients discharged alive were included. For unplanned all‐cause readmissions, we excluded elective readmissions. Other outcomes included length of stay and cost of the index hospitalization. Tables S2 and S3 list the ICD‐10‐CM codes used to identify secondary outcomes.
    Continuous variables are presented as mean±SD or median and interquartile range (IQR) according to their distribution and categorical variables as percentages. We compared continuous variables between ALI and non‐ALI groups using the Student t test or Wilcoxon rank sum test, as applicable, and compared categorical variables using χ2 test.
    We used logistic regression for categorical variables to evaluate trends in the incidence of ALI by calendar year, adjusting for age, sex, and Elixhauser Comorbidity Index to determine the P value for trend. We constructed a multivariable hierarchical logistic regression model to assess predictors of ALI and outcomes such as in‐hospital mortality, in‐hospital major amputation, composite of in‐hospital mortality or major amputation, length of stay, and cost of index admission. A multivariable Cox regression model was utilized for 30‐day and 6‐month all‐cause readmission and 30‐day and 6‐month readmission due to ALI. Multivariable regression models were adjusted for age, sex, comorbidities, mechanical circulatory support (MCS), primary payer, hospital size, and teaching hospital status. Model performance was assessed using receiver operating characteristic curves. The cumulative incidence rate of 30‐day and 6‐month all‐cause readmission and 30‐day and 6‐month readmission due to ALI was derived from the Kaplan–Meier method, and log‐rank test was used for comparison between ALI and non‐ALI. A Kaplan–Meier graph was constructed for 6‐month all‐cause readmission and 6‐month readmission due to ALI.
    All P values were 2‐sided, with P<0.05 considered statistically significant. All statistical analyses were conducted using appropriate weighting, stratifying, and clustering samples to obtain national estimates, using SAS software version 9.4 (SAS Institute Inc.), and R Statistical Software version 4.3.3 (R Foundation for Statistical Computing).

    RESULTS

    A total of 2 493 785 patients in the NRD underwent PCI during the study period. We finally included 629 656 patients who underwent PCI for stable CAD (Figure 1). Of this sample, 3456 (0.55%) had ALI. The majority (n=2865, 82.9%) experienced ALI of the lower extremity (LE), with 12.0% (n=414) having ALI of the upper extremity (UE). A total of 55.7% (n=1994) of the ALI group were men, compared with 64.9% (n=406 177) of the non‐ALI group (P<0.001). The mean±SD age of patients with ALI was 66.7±11.9 years, compared with 67.9±11.9 years in the non‐ALI group (P<0.001).
    Comorbidity burden differed significantly between both ALI and non‐ALI cohorts. The most significant difference was noted in the burden of peripheral arterial disease (PAD), with 90.7% of the ALI group having PAD compared with just 23.6% in the non‐ALI group (P<0.001). Further, the ALI cohort had a higher percentage of patients with a history of heart failure, chronic liver disease, and coagulopathy. Patients with ALI had a higher median Elixhauser Comorbidity Index, higher MCS placement (eg, Impella, intra‐aortic balloon pump, or left ventricular assist device), and a higher percentage of admissions to a large or teaching hospital. Finally, the ALI cohort had more patients who were insured by Medicaid or uninsured and a lower median household income per zip code. Table 1 summarizes the baseline characteristics of ALI and non‐ALI cohorts.
    Baseline Characteristics of PCI in Stable CAD, Stratified by Presence of ALI
    ALI indicates acute limb ischemia; CABG, coronary artery bypass grafting; CAD, coronary artery disease; IABP, intra‐arterial balloon pump; IQR, interquartile range; LVAD, left ventricular assist device; MCS, mechanical circulatory support; PCI, percutaneous coronary intervention; and TIA, transient ischemia attack.
    The yearly incidence of ALI varied, with an overall upward trend from 0.49% in 2016 to 0.59% in 2020 (P=0.02).
    PAD was the most powerful predictor of ALI on multivariable regression analysis, with patients exhibiting a 53 times greater risk of ALI (odds ratio [OR], 53.03 [95% CI, 43.76–64.28]; P<0.001). Weaker predictors included being female, having chronic liver disease or coagulopathy, a greater Elixhauser Comorbidity Index, requiring MCS, being uninsured, and being admitted at a large and/or teaching hospital. Other traditional cardiovascular comorbidities did not individually confer increased risk of post‐PCI ALI. The receiver operating characteristic curve C statistic was 0.89. Table 2 summarizes the predictors of post‐PCI ALI.
    Predictors of ALI After PCI in Stable CAD
    ALI indicates acute limb ischemia; CABG, coronary artery bypass grafting; CAD, coronary artery disease; IABP, intra‐arterial balloon pump; IQR, interquartile range; LVAD, left ventricular assist device; MCS, mechanical circulatory support; OR, odds ratio; PCI, percutaneous coronary intervention; and TIA, transient ischemia attack.
    Patients with ALI had a 1.6 times greater risk of in‐hospital mortality (12.9% versus 3.9%, P<0.0001), a 4.7 times greater risk of in‐hospital major amputation (4.6% versus 0.3%, P<0.0001), and a 2.1 times greater risk of in‐hospital composite of mortality and major amputation (16.9% versus 4.2%, P<0.001), compared with those without ALI. Further, they had a 9.3 times greater risk of 30‐day readmission for ALI (1.9% versus 0.1%, P<0.001) and an 8.9 times greater risk of 6‐month readmission for ALI (4.8% versus 0.2%, P<0.001). Finally, patients with ALI had a significantly higher risk of 30‐day all‐cause readmissions, and their index hospitalizations involved a longer LOS and greater cost. Table 3 summarizes the outcomes of ALI and non‐ALI cohorts, while Figure 2 compares ALI‐related (A) and all‐cause (B) readmissions between ALI and non‐ALI cohorts.
    Clinical Outcomes in ALI and Non‐ALI Cohorts
    ALI indicates acute limb ischemia; HR, hazard ratio; LOS, length of stay; and OR, odds ratio.
    Denominators represent the percentage of patients discharged alive, and the cumulative incidences are derived from the Kaplan–Meier survival estimate.
    Multivariable regression models adjusted for age, sex, comorbidities, mechanical circulatory support, primary payer, hospital size, and teaching hospital.
    ALI indicates acute limb ischemia.
    There was no significant difference in in‐hospital mortality (13.2% versus 12.9%, P=0.41), in‐hospital major amputation (1.1% versus 5.3%, P=0.34), or composite of in‐hospital mortality and major amputation (14.3% versus 17.4%, P=0.93) between patients with UE ALI alone compared with those with LE ALI alone (Table 4).
    Outcomes in Patients With Post‐PCI ALI of the UE and LE Alone
    ALI indicates acute limb ischemia; HR, hazard ratio; LE, lower extremity; OR, odds ratio; PCI, percutaneous coronary intervention; and UE, upper extremity.
    Denominators represent the percentage of patients discharged alive and the cumulative incidences are derived from the Kaplan–Meier survival estimate. Patients with both UE ALI and LE ALI were excluded from this comparison.
    Multivariable regression models adjusted for age, sex, comorbidities, large bore device, primary payer, hospital size, and teaching hospital.
    A total of 40.9% (n=1414) patients with post‐PCI ALI underwent embolectomy, while 11.6% (n=402) underwent thrombolysis. Among those undergoing embolectomy, 66.4% (n=939) underwent surgical embolectomy, while the remainder (33.6%, n=475) underwent endovascular embolectomy.
    Patients undergoing embolectomy had higher in‐hospital mortality (15.8% versus 10.3%), while those undergoing thrombolysis had higher in‐hospital major amputation (6.6% versus 9.8%), with a similar composite of in‐hospital mortality and amputation (18.6% versus 21.3%). All‐cause readmission was similar between the t groups, but patients undergoing thrombolysis were more likely to be readmitted for ALI at 6 months. Table 5 summarizes the outcomes of ALI in patients undergoing embolectomy and thrombolysis.
    Outcomes in Patients With Post‐PCI ALI Treated With Thrombolysis or Embolectomy
    ALI indicates acute limb ischemia; LOS, length of stay; and PCI, percutaneous coronary intervention.
    Denominators represent the percentage of patients discharged alive, and the cumulative incidences are derived from the Kaplan–Meier survival estimate.

    DISCUSSION

    In our analysis of a large, nationwide sample of patients with stable CAD undergoing PCI, the incidence of post‐PCI ALI was low but increased slightly during the study period. PAD was, by far, the most robust predictor of ALI. Patients with ALI, expectedly, had far worse clinical outcomes than those without ALI. This included higher risk of in‐hospital mortality, in‐hospital major amputation, and ALI‐related readmissions. Approximately half of those with post‐PCI ALI underwent embolectomy or received thrombolysis, with the former having higher in‐hospital mortality and the latter having higher in‐hospital major amputation (Central Illustration).
    Interventionalists must be prepared to anticipate and successfully manage post‐PCI ALI, given its limb‐ and life‐threatening potential. Our study fulfills a significant lacuna in the literature, with the largest analysis of ALI after PCI for stable CAD to date.
    We found that ALI occurred in 0.55% of PCIs for stable CAD. Given an estimated yearly incidence of ALI of 10 per 100 000 people in the general population,10 this represents a 55‐fold greater incidence in patients undergoing PCI for stable CAD. Prior analyses of the NRD have estimated an incidence of 0.31% in PCIs for acute myocardial infarction (AMI),11 and 1.53% in PCIs for AMI complicated by cardiogenic shock.12 Other reports have mostly implicated arterial puncture closing devices, reporting an incidence of ALI in 0.1% to 0.7% of PCIs.5, 6, 7, 8 While arterial puncture closing devices reportedly improve patient satisfaction and reduce time to ambulation after PCI, meta‐analyses have reported them to confer a 2.1 times greater risk of leg ischemia compared with compression alone, with an incidence of 0% to 2%.13, 14 Attributed mechanisms included acute thromboembolism, arterial dissection (either flow‐limiting or dissection that gets thrombosed), suture‐ mediated injury, and embolization of collagen plugs or bioabsorbable footplates, depending on the device used.5 Other causes of post‐PCI ALI, not related to arterial puncture closing devices, included extensive aortic dissection after PCI,5, 8 and atheromatous emboli,15 related to the manipulation of wires, catheters, guides, and placement of other devices. Our analysis captures all of these possible causes, although we did not have access to the details behind each case of post‐PCI ALI.
    We found that the strongest predictor of post‐PCI ALI was PAD, which conferred a >50 times greater risk of ALI. This is consistent with analyses of PCI for AMI and AMI complicated by cardiogenic shock, which also reported that PAD was the strongest predictor for ALI, albeit to a much lower extent.11, 12 Another common finding was that MCS was a predictor of ALI; in analyses of PCI for AMI and complicated by cardiogenic shock, the authors reported that the use of microaxial mechanical support/left ventricular assist devices was associated with post‐PCI ALI.11, 12 These findings across studies highlight the need for greater vigilance while using MCS, particularly in patients with PAD. Other predictors included being female and having a higher comorbidity index. This is consistent with larger reports of vascular events related to arterial puncture closing devices, which reported the same predictors.7 However, the female predominance contrasts with large studies of the general population and of patients with PAD.10, 16 Interestingly, we found that cardiovascular risk factors (diabetes, hypertension, hyperlipidemia, smoking, prior revascularization) were not independently strongly predictive of post‐PCI ALI, despite them being known risk factors for PAD in the general population.10 Regardless, these risk factors are well known to contribute to PAD, with their poor control known to cause ALI over PAD.10 Therefore, it would be logical to recommend smoking cessation and aggressive medical management before undertaking elective PCI.
    Patients with post‐PCI ALI had a 12.9% risk of in‐hospital mortality and a 4.6% risk of in‐hospital major amputation, both significantly greater than patients without ALI. Readmission rates were high, with an all‐cause readmission rate of 23.1% at 30 days. These outcomes were not significantly different between UE ALI and LE ALI. Prior estimates report short‐term mortality rates of ALI at 15% to 25%,4, 10 short‐term amputation rates of ALI at ≈25%,10 and 30‐day readmission rates of critical limb ischemia at ≈23.6%.17 Therefore, our findings suggest that post‐PCI ALI is not significantly different than ALI in other settings. Further, these outcomes were not significantly different between UE ALI and LE ALI. While there are limited comparative data on outcomes, overall, UE ALI is typically caused by thromboembolism, while LE ALI is caused by atherosclerosis.18 Our findings of similar outcomes possibly suggest that the underlying pathomechanism of post‐PCI ALI is the same regardless of the extremity affected.
    Finally, we found that embolectomy was more frequently used than thrombolysis in post‐PCI ALI. Patients undergoing embolectomy had higher mortality, perhaps related to surgical risk, while those receiving thrombolysis had greater risk of amputation. This contrasts with prior evidence showing no significant differences in short‐term mortality or amputation,4, 19, 20 suggesting that our findings may require further investigation.

    LIMITATIONS

    Our study is not without limitations. First, we relied on the NRD, which is an administrative database with the inherent risk of erroneously coding some comorbidities. However, this was likely ameliorated by its stringent quality‐control measures, including verification of state‐level data and multiple audits of diagnostic codes. Second, the NRD only includes inpatient admissions. Thus, we may have missed PCIs for stable CAD performed on an outpatient basis. Third, we did not have data regarding the characteristics of PCI, such as the route of access, use of ultrasound, medications used, and timing of different aspects of the procedure. For example, knowing the access site may have provided further insight into the different mechanisms involved in post‐PCI ALI, as these are likely different between femoral and radial access. Fourth, while PAD emerged as the strongest predictor of ALI, we did not have information regarding other aspects of PAD, such as the anatomic characteristics or severity of PAD, and how they correlated with the characteristics of post‐PCI ALI. Similarly, MCS requirement was significantly associated with post‐PCI ALI, but we remain unsure of whether the ALI was in the limb used for MCS placement, in the limb used for PCI access, or in an unrelated limb. MCS placement due to an intraprocedural complication may have necessitated speedy placement, but with suboptimal technique, thus contributing to later complications. Thus, overall, we had limited information regarding the causes of post‐PCI ALI. Fifth, the clinical decision‐making involved with the treatment of ALI was unknown. Finally, the NRD does not provide racial or ethnic data; this may have provided further insight into predictors of post‐PCI ALI and disparities in care.

    CONCLUSIONS

    ALI is an uncommon complication of PCI in patients with stable CAD. PAD is the strongest predictor of ALI. Patients with post‐PCI ALI have greater mortality, risk of amputation, and likelihood of readmissions.

    Sources of Funding

    None.

    Disclosures

    None.

    Supporting information

    Tables S1–S3
    VBDr. Vivek Bhat

    Dr. Vivek Bhat

    Cardiology · General Non-Invasive Cardiology
    SUNY Upstate Medical University
    vivekb