Risk Scores for Myocardial Infarction and Major Adverse Cardiac Event following Major Amputation for Limb Ischemia with Internal VQI Validation.

Abstract

Objective: Numerous cardiac risk evaluation models exist for patients undergoing major vascular surgical interventions. These models, however, do not consider patients undergoing major amputation for limb ischemia. The purpose of this study is to create risk scores for myocardial infarction and composite adverse cardiac events following major amputation performed for limb ischemia.

Methods: The first step was univariable analysis for the outcomes of perioperative MI and MACE for patients undergoing non-emergent major amputation for limb ischemia (primary risk score study group N=10,260, validation group N=5130). Univariable analysis was conducted with Chi-squared testing for categorical variables and student t-test for comparison of means of ordinal variables. Next, binary logistic regression analysis was performed for the outcomes of perioperative MI and MACE utilizing variables which had achieved a univariable P value ≤ 0.1. Using this regression, it was determined which variables have a multivariable association for the outcomes as defined by a multivariable regression P value of ≤ .05. Weighted cumulative event scores were then created for the outcomes of perioperative MI and MACE. Variables with a multivariable P-value ≤ .05 on the regression analysis were included in the scores and weighted based on their respective regression beta-coefficient in a point scale. AUC analysis and Hosmer-Lemeshow goodness of fit was also conducted for the validation cohort using the same risk score variables and scoring system as the primary study group. Supplementary machine learning analysis was performed to reinforce variable importance.

Results: Variables with a significant (P<.05) multivariable association and meeting inclusion or the MACE risk score included : advancing age (aOR 1.02/unit time; P<.001); female sex (aOR 1.28, P=.01); asymptomatic CAD (aOR 1.36, P=.009); symptomatic CAD (aOR 1.28, P=.049); CABG greater than 5 years ago (aOR 1.37, P=.014); class II CHF (aOR 1.30, P=.05); class III CHF (aOR 1.75, P=.002); Class IV CHF (aOR 5.26, P<.001); COPD (aOR 1.29, P=.012); ESRD (aOR 2.20, P<.001); and renal insufficiency (aOR 1.76, P<.001). Regarding the risk score for MACE following major amputation for limb ischemia, patients with risk scores ≤ 1 experienced MACE in just 2.6% of cases. The MACE rate rose in an exponential fashion with rising risk score with rates of 26.9% for patients with scores 16 and higher indicating a tenfold increased risk. In MI analysis, the following variables achieved a multivariable p-value ≤ and were thus ultimately included in the risk score for MI : advancing age (aOR 1.02/unit time, P=.007); asymptomatic CAD (aOR 1.52, P=.022); symptomatic CAD (aOR 1.56, P=.038); history of CABG greater than 5 years ago (aOR 1.51, P=.029); Class III CHF (aOR 1.84, P=.019); Class IV CHF (aOR 3.07, P=.002); ESRD on dialysis (aOR 1.95, P<.001); renal insufficiency (aOR 1.95, P<.001); and not being on an antiplatelet preoperative (protective aOR 0.65, P=.007). Regarding the risk score for MI, patients with risk scores of ≤ 0 had MI rates of just 0.5% with steep escalation noted with advancing risk score as patients with risk scores of 10 and higher had MI rates of 6.0% indicating a 12-fold higher risk of MI. Risk score AUC values were 0.7 and 0.71 respectively. Patients with MACE and MI perioperatively had survival rates as low as 42% by the two-year mark versus 67%-69% in those without (P<.001). Machine learning confirmed the importance of the key variables and achieved AUC values ranging from 0.77 to 0.94.

Conclusions: Risk scores for perioperative MI and MACE during hospitalization for major amputation due to limb ischemia have been created with accurate internal validation. The data has the potential to impact preoperative and perioperative patient management to reduce adverse event rates. The most impactful variables increasing risk of MI and MACE are advancing age, history of class III or IV CHF, history of CAD, CABG over 5 years ago, renal insufficiency and ESRD.