Grade 1 Internal Carotid Artery Blunt Cerebrovascular Injury Persistence Risks Stroke With Current Management: An EAST Multicenter Study.

BACKGROUND: Higher blunt cerebrovascular injury (BCVI) grade and lack of medical therapy are associated with stroke. Knowledge of stroke risk factors specific to individual grades may help tailor BCVI therapy to specific injury characteristics. METHODS: A post-hoc analysis of a 16 center, prospective, observational trial (2018-2020) was performed including grade 1 internal carotid artery (ICA) BCVI. Repeat imaging was considered the second imaging occurrence only. RESULTS: From 145 grade 1 ICA BCVI included, 8 (5.5%) suffered a stroke. Grade 1 ICA BCVI with stroke were more commonly treated with mixed anticoagulation and antiplatelet therapy (75.0% vs 9.6%, P <.001) and less commonly antiplatelet therapy (25.0% vs 82.5%, P = .001) compared to injuries without stroke. Of the 8 grade 1 ICA BCVI with stroke, 4 (50.0%) had stroke after medical therapy was started. In comparing injuries with resolution at repeat imaging to those without, stroke occurred in 7 (15.9%) injuries without resolution and 0 (0%) injuries with resolution (P = .005). At repeat imaging in grade 1 ICA BCVI with stroke, grade of injury was grade 1 in 2 injuries, grade 2 in 3 injuries, grade 3 in 1 injury, and grade 5 in one injury. DISCUSSION: While the stroke rate for grade 1 ICA BCVI is low overall, injury persistence appears to heighten stroke risk. Some strokes occurred despite initiation of medical therapy. Repeat imaging is needed in grade 1 ICA BCVI to evaluate for injury progression or resolution.

Does treatment delay for blunt cerebrovascular injury affect stroke rate?: An EAST multicenter study.

BACKGROUND: The purpose of this study was to analyze injury characteristics and stroke rates between blunt cerebrovascular injury (BCVI) with delayed vs non-delayed medical therapy. We hypothesized there would be increased stroke formation with delayed medical therapy. METHODS: This is a sub-analysis of a 16 center, prospective, observational trial on BCVI. Delayed medial therapy was defined as initiation >24 hours after admission. BCVI which did not receive medical therapy were excluded. Subgroups for injury presence were created using Abbreviated Injury Scale (AIS) score >0 for AIS categories. RESULTS: 636 BCVI were included. Median time to first medical therapy was 62 hours in the delayed group and 11 hours in the non-delayed group (p < 0.001). The injury severity score (ISS) was greater in the delayed group (24.0 vs the non-delayed group 22.0, p <  0.001) as was the median AIS head score (2.0 vs 1.0, p <  0.001). The overall stroke rate was not different between the delayed vs non-delayed groups respectively (9.7% vs 9.5%, p = 1.00). Further evaluation of carotid vs vertebral artery injury showed no difference in stroke rate, 13.6% and 13.2%, p = 1.00 vs 7.3% and 6.5%, p = 0.84. Additionally, within all AIS categories there was no difference in stroke rate between delayed and non-delayed medical therapy (all N.S.), with AIS head >0 13.8% vs 9.2%, p = 0.20 and AIS spine >0 11.0% vs 9.3%, p = 0.63 respectively. CONCLUSIONS: Modern BCVI therapy is administered early. BCVI with delayed therapy were more severely injured. However, a higher stroke rate was not seen with delayed therapy, even for BCVI with head or spine injuries. This data suggests with competing injuries or other clinical concerns there is not an increased stroke rate with necessary delays of medical treatment for BCVI.

Endovascular Intervention in Internal Carotid Artery Blunt Cerebrovascular Injury: An EAST Multicenter Study.

BACKGROUND: Use of endovascular intervention (EI) for blunt cerebrovascular injury (BCVI) is without consensus guidelines. Rates of EI use and radiographic characteristics of BCVI undergoing EI nationally are unknown. METHODS: A post-hoc analysis of a prospective, observational study at 16 U.S. trauma centers from 2018 to 2020 was conducted. Internal carotid artery (ICA) BCVI was included. The primary outcome was EI use. Multivariable logistic regression was performed for predictors of EI use. RESULTS: From 332 ICA BCVI included, 21 (6.3%) underwent EI. 0/145 (0%) grade 1, 8/101 (7.9%) grade 2, 12/51 (23.5%) grade 3, and 1/20 (5.0%) grade 4 ICA BCVI underwent EI. Stroke occurred in 6/21 (28.6%) ICA BCVI undergoing EI and in 33/311 (10.6%) not undergoing EI (P = .03), with all strokes with EI use occurring prior to or at the same time as EI. Percentage of luminal stenosis (37.75 vs 20.29%, P = .01) and median pseudoaneurysm size (9.00 mm vs 3.00 mm, P = .01) were greater in ICA BCVI undergoing EI. On logistic regression, only pseudoaneurysm size was associated with EI (odds ratio 1.205, 95% CI 1.035-1.404, P = .02). Of the 8 grade 2 ICA BCVI undergoing EI, 3/8 were grade 2 and 5/8 were grade 3 prior to EI. Of the 12 grade 3 ICA BCVI undergoing EI, 11/12 were grade 3 and 1/12 was a grade 2 ICA BCVI prior to EI. DISCUSSION: Pseudoaneurysm size is associated with use of EI for ICA BCVI. Stroke is more common in ICA BCVI with EI but did not occur after EI use.

Factors associated with stroke formation in blunt cerebrovascular injury: An EAST multicenter study.

BACKGROUND: Stroke risk factors after blunt cerebrovascular injury (BCVI) are ill-defined. We hypothesized that factors associated with stroke for BCVI would include medical therapy (i.e., Aspirin), radiographic features, and protocolization of care. METHODS: An Eastern Association for the Surgery of Trauma-sponsored, 16-center, prospective, observational trial was undertaken. Stroke risk factors were analyzed individually for vertebral artery (VA) and internal carotid artery (ICA) BCVI. Blunt cerebrovascular injuries were graded on the standard 1 to 5 scale. Data were from the initial hospitalization only. RESULTS: Seven hundred seventy-seven BCVIs were included. Stroke rate was 8.9% for all BCVIs, with an 11.7% rate of stroke for ICA BCVI and a 6.7% rate for VA BCVI. Use of a management protocol (p = 0.01), management by the trauma service (p = 0.04), antiplatelet therapy over the hospital stay (p < 0.001), and Aspirin therapy specifically over the hospital stay (p < 0.001) were more common in ICA BCVI without stroke compared with those with stroke. Antiplatelet therapy over the hospital stay (p < 0.001) and Aspirin therapy over the hospital stay (p < 0.001) were more common in VA BCVI without stroke than with stroke. Percentage luminal stenosis was higher in both ICA BCVI (p = 0.002) and VA BCVI (p < 0.001) with stroke. Decrease in percentage luminal stenosis (p < 0.001), resolution of intraluminal thrombus (p = 0.003), and new intraluminal thrombus (p = 0.001) were more common in ICA BCVI with stroke than without, while resolution of intraluminal thrombus (p = 0.03) and new intraluminal thrombus (p = 0.01) were more common in VA BCVI with stroke than without. CONCLUSION: Protocol-driven management by the trauma service, antiplatelet therapy (specifically Aspirin), and lower percentage luminal stenosis were associated with lower stroke rates, while resolution and development of intraluminal thrombus were associated with higher stroke rates. Further research will be needed to incorporate these risk factors into lesion specific BCVI management. LEVEL OF EVIDENCE: Prognostic and Epidemiologic, Level IV.

Readmission for venous thromboembolism after emergency general surgery is underreported and influenced by insurance status.

BACKGROUND: Prior studies of venous thromboembolism (VTE) after emergency general surgery (EGS) are not nationally representative nor do they fully capture readmissions to different hospitals. We hypothesized that different-hospital readmission accounted for a significant number of readmissions with VTE after EGS and that predictive factors would be different for same- and different-hospital readmissions. METHODS: The 2014 Nationwide Readmissions Database was queried for nonelective EGS hospitalizations. The outcomes were readmission to the index or different hospitals within 180 days with VTE. Multivariate logistic regressions identified risk factors for readmission to index and different hospitals with VTE, reported as odds ratios with their 95% confidence intervals. Patients were excluded if during the index admission they expired, developed a VTE, had a vena cava filter placed, or did not have at least 180 days of follow-up. RESULTS: Of 1,584,605 patients meeting inclusion criteria, 1.3% (n = 20,963) of patients were readmitted within 180 days with a VTE. Of these, 28% (n = 5,866) were readmitted to a different hospital. Predictors overall for readmission with VTE were malignancy, prolonged hospitalization, age, and being publicly insured. However, predictors for readmission to a different hospital are based on hospital characteristics, including for-profit status, or procedure type. CONCLUSIONS: Nearly one in three readmissions with VTE after EGS occurs at a different hospital and may be missed by current quality metrics that only capture same-hospital readmission. Such metrics may underestimate for-profit hospital postoperative VTE rates relative to public and nonprofit hospitals, potentially affecting benchmarking and reimbursement. Postdischarge VTE rate is associated with insurance status. These findings have implications for policy and prevention programming design. LEVEL OF EVIDENCE: Epidemiological study, level III.

Comparing Machine Learning Algorithms for Predicting Acute Kidney Injury.

Prior studies have used vital signs and laboratory measurements with conventional modeling techniques to predict acute kidney injury (AKI). The purpose of this study was to use the trend in vital signs and laboratory measurements with machine learning algorithms for predicting AKI in ICU patients. The eICU Collaborative Research Database was queried for five consecutive days of laboratory measurements per patient. Patients with AKI were identified and trends in vital signs and laboratory values were determined by calculating the slope of the least-squares-fit linear equation using three days for each value. Different machine learning classifiers (gradient boosted trees [GBT], logistic regression, and deep learning) were trained to predict AKI using the laboratory values, vital signs, and slopes. There were 151,098 ICU stays identified and the rate of AKI was 5.6 per cent. The best performing algorithm was GBT with an AUC of 0.834 ± 0.006 and an F-measure of 42.96 per cent ± 1.26 per cent. Logistic regression performed with an AUC of 0.827 ± 0.004 and an F-measure of 28.29 per cent ± 1.01 per cent. Deep learning performed with an AUC of 0.817 ± 0.005 and an F-measure of 42.89 per cent ± 0.91 per cent. The most important variable for GBT was the slope of the minimum creatinine (30.32%). This study identifies the best performing machine learning algorithms for predicting AKI using trends in laboratory values in ICU patients. Early identification of these patients using readily available data indicates that incorporating machine learning predictive models into electronic medical record systems is an inevitable requisite for improving patient outcomes.