Limb ischemia and bleeding in patients requiring venoarterial extracorporeal membrane oxygenation.

OBJECTIVE: Acute limb ischemia (ALI) and cannulation site bleeding are frequent complications of venoarterial (VA) extracorporeal membrane oxygenation (ECMO) and are associated with worse outcomes. The goals of this study were to assess our rates of ECMO-related ALI and bleeding and to evaluate the efficacy of strategies to prevent them, such as distal perfusion cannula (DPC) and ultrasound-guided cannulation. METHODS: This is a single-center retrospective cohort study of adult patients placed on peripheral VA-ECMO at a tertiary medical center between 2014 and 2018. ALI was defined as new ischemia of the extremity ipsilateral to arterial cannulation. Significant cannulation site bleeding was defined as excessive bleeding requiring intervention (eg, transfusion or reoperation). Univariate analyses were used to identify factors associated with ALI, bleeding, and in-hospital mortality. RESULTS: During the study period, 105 patients were placed on peripheral VA-ECMO (61.3% female; mean age, 54.9 ± 14.8 years). Nearly half (46.6%) had ECMO implantation in an extracorporeal cardiopulmonary resuscitation setting and 37 (44.0%) had a DPC. Average duration of support was 5.6 ± 5.0 days. Overall in-hospital mortality and death on ECMO support were 65.1% and 50%, respectively. ALI occurred in 21 (20%) and cannulation-related bleeding occurred in 24 (22.9%) patients who were treated with a total of 27 procedures, including thromboembolectomy (22.2%), vascular repair (18.5%), and fasciotomy (25.9%). On univariate analysis, cannulation in the operating room (odds ratio [OR], 0.25; 95% confidence interval [CI], 0.08-0.77; P = .02) was associated with decreased risk of ALI, whereas cannulation in the operating room (OR, 2.65; 95% CI, 1.09-6.45; P = .03) and cutdown approach (OR, 4.96; 95% CI, 2.32-10.61; P < .0001) were associated with increased risk of bleeding. Ultrasound-guided placement was associated with decreased risk of bleeding (OR, 0.81; 95% CI, 0.04-0.84; P = .03). DPC was not associated with either ALI (P = .47) or bleeding (P = .06). ALI (OR, 2.68; 95% CI 1.03-6.98; P = .04), age (OR, 1.94; 95% CI, 1.03-3.69; P = .04), and worse baseline heart failure (OR, 2.01; 95% CI, 1.02-3.97; P = .04) were associated with greater risk of in-hospital mortality. Ultrasound-guided cannulation (OR, 0.41; 95% CI, 0.20-0.87; P = .02) was associated with decreased risk of in-hospital mortality. CONCLUSIONS: ALI and significant bleeding are common occurrences after peripheral VA-ECMO cannulation. Whereas DPC placement did not significantly decrease risk of ALI, ultrasound-guided cannulation decreased the risk of bleeding. Cannulation in the operating room is associated with decreased risk of ALI at the expense of increased risk of bleeding. ALI, older age (≥65 years), and worse heart failure increased risk of in-hospital mortality.

The Comorbidity-Polypharmacy Score is an Objective and Practical Predictor of Outcomes and Mortality after Vascular Surgery.

BACKGROUND: The comorbidity-polypharmacy score (CPPS) was developed to quantify the severity of comorbidities of patients with geriatric trauma. CPPS is the sum of the number of medications and comorbidities, and is thus objective, user-friendly, and potentially adaptable to many clinical situations. We sought to understand if CPPS associates with outcomes and mortality after common vascular surgery procedures. METHODS: This is a retrospective single-center study. A total of 466 patients who underwent carotid endarterectomy, infrainguinal bypass, percutaneous lower extremity revascularization, or endovascular abdominal aortic aneurysm repair at a single medical center were included. CPPS were classified as mild, moderate, severe, and morbid based on scores of 0-7, 8-15, 15-21, and ≥21, respectively. End points were reinterventions, 30-day readmission, and mortality. We used chi-squared tests to analyze differences in categorical variables; Kruskal-Wallis tests to analyze differences in continuous variables; Kaplan-Meier estimation and Cox proportional hazard modeling to examine survival data; and receiver operator characteristic (ROC) curve analyses to assess sensitivity and specificity. RESULTS: The mean preoperative CPPS was 14.1 ± 6.1. Higher CPPS were associated with longer hospital and postoperative length of stay (P < 0.001). Severe and morbid CPPS categories had higher rates of ICU admission, reintervention, and 30-day readmission which did not reach statistical significance after correction for multiple comparisons. CPPS was independently associated with 1- and 5-year mortality in a multivariable Cox model (hazard ratio = 2.2, 95% confidence interval: 1.3-3.3). ROC analysis revealed C-statistics of 0.81 and 0.72 for 1-year and 5-year all-cause mortality, respectively (P < 0.001). CONCLUSIONS: CPPS is a simple and pragmatic clinical tool for quantifying risk of postoperative outcomes and mortality after common vascular surgery procedures. Further investigation is needed to validate the use of CPPS in enhancing existing predictors of patient outcomes and in serving as an adjunctive tool for determining resource allocation and discharge planning in patients who underwent vascular surgery.

Predictors and Consequences of Loss to Follow-up after Vascular Surgery.

BACKGROUND: Loss to follow-up (LTF) after surgery impacts quality of care and can adversely affect short- and long-term clinical outcomes. This study identifies modifiable factors contributing to LTF after vascular surgery and the factors' effect on short- and long-term clinical outcomes. METHODS: This is a retrospective single-center cohort study of 440 consecutive adult patients who underwent carotid endarterectomy, infrainguinal bypass, percutaneous lower extremity revascularization, or endovascular aortic aneurysm repair at Northwestern Memorial Hospital between November 2011 and November 2013. Twenty-six patients who died within 9 months after surgery were excluded because of competing risks with the study end points. Demographics, medical history and medications, hospitalization and procedure-related factors, and postoperative complications were collected from the medical record. The primary end point was LTF 1 month after surgery (LTF1M), defined as lack of an in-person outpatient visit with a vascular surgeon 1 month after the index procedure. Secondary outcomes were LTF 1 year after surgery (LTF1Y), defined as lack of an in-person outpatient visit with a vascular surgeon between 9 and 22 months after discharge, and overall 5-year survival. RESULTS: Overall LTF1M and LTF1Y rates were 27.3% and 46.8%, respectively. Kaplan-Meier analysis revealed no difference in survival based on the LTF1M status (P = 0.72), but patients who were LTF1Y had significantly worse survival at 5 years (P < 0.001). Seeing a nonvascular surgeon specialist at our institution (odds ratio (OR) 0.58, 95% confidence interval (CI): 0.35-0.94, P = 0.03) and having a reintervention (OR 0.17, 95% CI: 0.08-0.37, P < 0.001) were associated with decreased LTF1Y in a multivariable model. Overall mortality was more likely with LTF1Y (hazard ratio (HR) 3.27, 95% CI: 1.86-5.76, P < 0.001) and less likely with seeing another specialist at our institution (HR 0.38, 95% CI: 0.20-0.75, P = 0.005). CONCLUSIONS: LTF rates after vascular surgery are high and associated with poor long-term outcomes. Patients who did not see a nonvascular surgeon specialist at our institution had higher rates of LTF1Y and worse overall mortality, suggesting that improved integration of care can improve LTF and survival.