228. A Predictive Model of Perioperative Myocardial Infarction in Spine Surgery

BACKGROUND CONTEXT Emphasis has been placed upon reducing perioperative complications in spinal operations, of which cardiac complications remain among the most devastating for patients. Myocardial infarction (MI), along with its predictive factors, has been an understudied complication. PURPOSE To assess the incidence and risk factors for perioperative myocardial infarctions in spine surgery patients. STUDY DESIGN/SETTING Retrospective case control study PATIENT SAMPLE A total of 196,086 surgical spine patients OUTCOME MEASURES Predictors of MI following surgical spine intervention METHODS A patient was categorized as having sustained a postsurgical MI in this study if it occurred intraoperatively or within 30 days of surgery, manifested by documentation of ECG changes accordingly: ST elevation >1mm in two or more contiguous leads, new left bundle branch, new Q-wave in two or more contiguous leads, or new elevation in troponin greater than three times upper level of the reference range. The relationship between MI and non-MI spine patients was assessed using chi-squared and independent samples t-tests, as appropriate. Descriptive statistics, including frequency counts for categorical variables and means and standard deviations, were calculated to summarize demographics and clinical profiles such as spinal diagnoses and comorbidity. Univariate/multivariate analyses were run to assess predictive factors of MI in spine surgery patients. Logistic regression with stepwise model selection was employed to create a model to predict MI occurrence. RESULTS Of 196,523 patients (57.1±14.2 years, 48% female, 30.4±6.5 kg/m2) undergoing elective spine surgery, 436 patients had an acute MI intraoperatively or within 30 days postoperatively (Spine-MI) (69.07±10.4 years, 42% female, 30.39±6.22 kg/m2). Incidence of MI did not change significantly from 2010 to 2016 (0.2% to 0.3%, p=0.298). Spine-MI patients underwent significantly more fusions than elective spine surgical patients who did not have an MI (73.6% vs 58.4%, p<0.001), with an average of 1.03 levels fused. Spine-MI patients also had significantly more Smith-Peterson osteotomies (5.0% vs 1.8%, p<0.001) and three-column osteotomies (0.9% vs 0.2%, p<0.001), but had far fewer decompression-only procedures (26.4% vs 41.6%, p<0.001). Overall, patients who developed a perioperative MI underwent more revisions compared to all elective spine surgery patients (5.3% vs 2.9%, p=0.003). Spine-MI patients had significantly greater invasiveness scores (3.41 vs 2.73, p<0.001) and total operative time (211.6 vs 147.3 min, p<0.001). The average number of post-operative days until developing an MI was 5.27 days; 9.9% day of operation, 50.8% 1-3 days after, 20.8% 4-7 days 10.9% 8-15 days, 7.6% 16-30 days. Mortality rate for Spine-MI patients was 4.6% versus 0.05% in the entire elective spine surgical population (p<0.001). Multivariate modeling for Spine-MI predictors yielded an AUC of 83.7%, and included history of diabetes mellitus, cardiac arrest and peripheral vascular disease, past blood transfusion, dialysis-dependence, high preoperative platelet count, superficial surgical site infection and days from operation to discharge. CONCLUSIONS A model with good predictive capacity to predict MI after spine surgery now exists. Predictive modeling of myocardial infarction following spine operations can aid in risk-stratification of patients, consequently improving preoperative patient counseling and optimization in the peri-operative period. FDA DEVICE/DRUG STATUS This abstract does not discuss or include any applicable devices or drugs. Emphasis has been placed upon reducing perioperative complications in spinal operations, of which cardiac complications remain among the most devastating for patients. Myocardial infarction (MI), along with its predictive factors, has been an understudied complication. To assess the incidence and risk factors for perioperative myocardial infarctions in spine surgery patients. Retrospective case control study A total of 196,086 surgical spine patients Predictors of MI following surgical spine intervention A patient was categorized as having sustained a postsurgical MI in this study if it occurred intraoperatively or within 30 days of surgery, manifested by documentation of ECG changes accordingly: ST elevation >1mm in two or more contiguous leads, new left bundle branch, new Q-wave in two or more contiguous leads, or new elevation in troponin greater than three times upper level of the reference range. The relationship between MI and non-MI spine patients was assessed using chi-squared and independent samples t-tests, as appropriate. Descriptive statistics, including frequency counts for categorical variables and means and standard deviations, were calculated to summarize demographics and clinical profiles such as spinal diagnoses and comorbidity. Univariate/multivariate analyses were run to assess predictive factors of MI in spine surgery patients. Logistic regression with stepwise model selection was employed to create a model to predict MI occurrence. Of 196,523 patients (57.1±14.2 years, 48% female, 30.4±6.5 kg/m2) undergoing elective spine surgery, 436 patients had an acute MI intraoperatively or within 30 days postoperatively (Spine-MI) (69.07±10.4 years, 42% female, 30.39±6.22 kg/m2). Incidence of MI did not change significantly from 2010 to 2016 (0.2% to 0.3%, p=0.298). Spine-MI patients underwent significantly more fusions than elective spine surgical patients who did not have an MI (73.6% vs 58.4%, p<0.001), with an average of 1.03 levels fused. Spine-MI patients also had significantly more Smith-Peterson osteotomies (5.0% vs 1.8%, p<0.001) and three-column osteotomies (0.9% vs 0.2%, p<0.001), but had far fewer decompression-only procedures (26.4% vs 41.6%, p<0.001). Overall, patients who developed a perioperative MI underwent more revisions compared to all elective spine surgery patients (5.3% vs 2.9%, p=0.003). Spine-MI patients had significantly greater invasiveness scores (3.41 vs 2.73, p<0.001) and total operative time (211.6 vs 147.3 min, p<0.001). The average number of post-operative days until developing an MI was 5.27 days; 9.9% day of operation, 50.8% 1-3 days after, 20.8% 4-7 days 10.9% 8-15 days, 7.6% 16-30 days. Mortality rate for Spine-MI patients was 4.6% versus 0.05% in the entire elective spine surgical population (p<0.001). Multivariate modeling for Spine-MI predictors yielded an AUC of 83.7%, and included history of diabetes mellitus, cardiac arrest and peripheral vascular disease, past blood transfusion, dialysis-dependence, high preoperative platelet count, superficial surgical site infection and days from operation to discharge. A model with good predictive capacity to predict MI after spine surgery now exists. Predictive modeling of myocardial infarction following spine operations can aid in risk-stratification of patients, consequently improving preoperative patient counseling and optimization in the peri-operative period.