Development and Internal Validation of Machine Learning Models to Predict Mortality and Disability After Mechanical Thrombectomy for Acute Anterior Circulation Large Vessel Occlusion.

OBJECTIVE: Mechanical thrombectomy (MT) improves outcomes in patients with LVO but many still experience mortality or severe disability. We sought to develop machine learning (ML) models that predict 90-day outcomes after MT for LVO. METHODS: Consecutive patients who underwent MT for LVO between 2015-2021 at a Comprehensive Stroke Center were reviewed. Outcomes included 90-day favorable functional status (mRS 0-2), severe disability (mRS 4-6), and mortality. ML models were trained for each outcome using prethrombectomy data (pre) and with thrombectomy data (post). RESULTS: Three hundred and fifty seven patients met the inclusion criteria. After model screening and hyperparameter tuning the top performing ML model for each outcome and timepoint was random forest (RF). Using only prethrombectomy features, the AUCs for the RFpre models were 0.73 (95% CI 0.62-0.85) for favorable functional status, 0.77 (95% CI 0.65-0.86) for severe disability, and 0.78 (95% CI 0.64-0.88) for mortality. All of these were better than a standard statistical model except for favorable functional status. Each RF model outperformed Pre, SPAN-100, THRIVE, and HIAT scores (P < 0.0001 for all). The most predictive features were premorbid mRS, age, and NIHSS. Incorporating MT data, the AUCs for the RFpost models were 0.80 (95% CI 0.67-0.90) for favorable functional status, 0.82 (95% CI 0.69-0.91) for severe disability, and 0.71 (95% CI 0.55-0.84) for mortality. CONCLUSIONS: RF models accurately predicted 90-day outcomes after MT and performed better than standard statistical and clinical prediction models.

The influence of pre-reperfusion blood pressure on outcomes following mechanical thrombectomy for anterior circulation large vessel occlusion.

BACKGROUND: We evaluated how systolic blood pressure (SBP) and mean arterial pressure (MAP) parameters between presentation and reperfusion influence functional status and intracranial hemorrhage (ICH). METHODS: All patients who underwent MT for LVO at a single institution were reviewed. Independent variables included SBP and MAP measurements obtained on presentation, between presentation and reperfusion (pre-reperfusion), and between groin puncture and reperfusion (thrombectomy). Mean, minimum, maximum, and standard deviations (SD) for SBP and MAP were calculated. Outcomes included 90-day favorable functional status, radiographic ICH (rICH), and symptomatic ICH (sICH). RESULTS: 305 patients were included. Higher pre-reperfusion SBPmax was associated with rICH (OR 1.41, 95% CI 1.08-1.85) and sICH (OR 1.84, 95% CI 1.26-2.72). Higher SBPSD was also associated with rICH (OR 1.38, 95% CI 1.06-1.81) and sICH (OR 1.59, 95% CI 1.12-2.26). Greater SBPmax (OR 0.64, 95% CI 0.47-0.86), MAPmax (OR 0.72, 95% CI 0.52-0.97), SBPSD (OR 0.63, 95% CI 0.46-0.86), and MAPSD (0.63, 95% CI 0.45-0.84) during thrombectomy were associated with lower odds of 90-day favorable functional status. In a subgroup analysis, these associations were primarily limited to patients with intact collateral circulation. Optimal SBPmax cutoffs for predicting rICH were 171 (pre-reperfusion) and 179 mmHg (thrombectomy). Cutoffs for predicting sICH were 178 (pre-reperfusion) and 174 mmHg (thrombectomy). CONCLUSION: Greater maximum BP and variability in BP during the pre-reperfusion period are associated with unfavorable functional status and ICH after MT for anterior circulation LVO.

Mechanical Thrombectomy Global Access For Stroke (MT-GLASS): A Mission Thrombectomy (MT-2020 Plus) Study.

BACKGROUND: Despite the well-established potent benefit of mechanical thrombectomy (MT) for large vessel occlusion (LVO) stroke, access to MT has not been studied globally. We conducted a worldwide survey of countries on 6 continents to define MT access (MTA), the disparities in MTA, and its determinants on a global scale. METHODS: Our survey was conducted in 75 countries through the Mission Thrombectomy 2020+ global network between November 22, 2020, and February 28, 2021. The primary end points were the current annual MTA, MT operator availability, and MT center availability. MTA was defined as the estimated proportion of patients with LVO receiving MT in a given region annually. The availability metrics were defined as ([current MT operators×50/current annual number of estimated thrombectomy-eligible LVOs]×100 = MT operator availability) and ([current MT centers×150/current annual number of estimated thrombectomy-eligible LVOs]×100= MT center availability). The metrics used optimal MT volume per operator as 50 and an optimal MT volume per center as 150. Multivariable-adjusted generalized linear models were used to evaluate factors associated with MTA. RESULTS: We received 887 responses from 67 countries. The median global MTA was 2.79% (interquartile range, 0.70-11.74). MTA was <1.0% for 18 (27%) countries and 0 for 7 (10%) countries. There was a 460-fold disparity between the highest and lowest nonzero MTA regions and low-income countries had 88% lower MTA compared with high-income countries. The global MT operator availability was 16.5% of optimal and the MT center availability was 20.8% of optimal. On multivariable regression, country income level (low or lower-middle versus high: odds ratio, 0.08 [95% CI, 0.04-0.12]), MT operator availability (odds ratio, 3.35 [95% CI, 2.07-5.42]), MT center availability (odds ratio, 2.86 [95% CI, 1.84-4.48]), and presence of prehospital acute stroke bypass protocol (odds ratio, 4.00 [95% CI, 1.70-9.42]) were significantly associated with increased odds of MTA. CONCLUSIONS: Access to MT on a global level is extremely low, with enormous disparities between countries by income level. The significant determinants of MT access are the country's per capita gross national income, prehospital LVO triage policy, and MT operator and center availability.

Venous Injury in Patients with Blunt Traumatic Brain Injury: Retrospective Analysis of a National Cohort.

BACKGROUND: Cerebral venous injury (CVI) includes injury to a dural venous sinus or major vein and leads to poorer outcomes for patients with blunt traumatic brain injury (TBI). We sought to identify the incidence, associated factors, and outcomes associated with CVI in a large national cohort. METHODS: Adult patients with blunt TBI were identified from the National Trauma Databank (2013-2017). Outcomes included inpatient mortality, discharge disposition, stroke, length of stay (LOS), intensive care unit LOS, and duration of mechanical ventilation. Multivariate regression models were used to identify the association between exposure variables and CVI, as well as each outcome. RESULTS: There were 619,659 patients with blunt TBI who met the inclusion criteria. CVI occurred in 1792 (0.3%) patients. Mixed intracranial injury type had the strongest association with CVI (odds ratio [OR] 2.89, 95% confidence interval [CI] 2.38-3.50), followed by isolated TBI (OR 1.76, 95% CI 1.54-2.02) and skull fracture (OR 1.72, 95% CI 1.55-1.91). CVI was associated with increased odds of mortality (OR 1.38, 95% CI 1.19-1.60), nonroutine discharge (OR 1.26, 95% CI 1.12-1.40), and stroke (OR 1.95, 95% CI 1.33-2.86). It was also associated with longer LOS (ß 2.02, 95% CI 1.55-2.50) and intensive care unit LOS (ß 0.14, 95% CI 0.13-0.16). Among locations of venous injury, superior sagittal sinus injury had significant associations with mortality (OR 2.93, 95% CI 1.62-5.30) and nonroutine discharge disposition (OR 1.94, 95% CI 1.12-3.35), whereas the others did not. CONCLUSIONS: We identified a 0.3% incidence of CVI in all-comers with blunt TBI as well as several injury-related variables that may be used to guide investigation for dural venous sinus injury. CVI was associated with poorer outcomes, with superior sagittal sinus injury having the strongest association.

Outcomes associated with brain tissue oxygen monitoring in patients with severe traumatic brain injury undergoing intracranial pressure monitoring.

OBJECTIVE: Brain tissue oxygen monitoring combined with intracranial pressure (ICP) monitoring in patients with severe traumatic brain injury (sTBI) may confer better outcomes than ICP monitoring alone. The authors sought to investigate this using a national database. METHODS: The National Trauma Data Bank from 2013 to 2017 was queried to identify patients with sTBI who had an external ventricular drain or intraparenchymal ICP monitor placed. Patients were stratified according to the placement of an intraparenchymal brain tissue oxygen tension (PbtO2) monitor, and a 2:1 propensity score matching pair was used to compare outcomes in patients with and those without PbtO2 monitoring. Sensitivity analyses were performed using the entire cohort, and each model was adjusted for age, sex, Glasgow Coma Scale score, Injury Severity Score, presence of hypotension, insurance, race, and hospital teaching status. The primary outcome of interest was in-hospital mortality, and secondary outcomes included ICU length of stay (LOS) and overall LOS. RESULTS: A total of 3421 patients with sTBI who underwent ICP monitoring were identified. Of these, 155 (4.5%) patients had a PbtO2 monitor placed. Among the propensity score-matched patients, mortality occurred in 35.4% of patients without oxygen monitoring and 23.4% of patients with oxygen monitoring (OR 0.53, 95% CI 0.33-0.85; p = 0.007). The unfavorable discharge rates were 56.3% and 47.4%, respectively, in patients with and those without oxygen monitoring (OR 1.41, 95% CI 0.87-2.30; p = 0.168). There was no difference in overall LOS, but patients with PbtO2 monitoring had a significantly longer ICU LOS and duration of mechanical ventilation. In the sensitivity analysis, PbtO2 monitoring was associated with decreased odds of mortality (OR 0.56, 95% CI 0.37-0.84) but higher odds of unfavorable discharge (OR 1.59, 95% CI 1.06-2.40). CONCLUSIONS: When combined with ICP monitoring, PbtO2 monitoring was associated with lower inpatient mortality for patients with sTBI. This supports the findings of the recent Brain Oxygen Optimization in Severe Traumatic Brain Injury phase 2 (BOOST 2) trial and highlights the importance of the ongoing BOOST3 trial.

Effects of Saturated Fat, Polyunsaturated Fat, Monounsaturated Fat, and Carbohydrate on Glucose-Insulin Homeostasis: A Systematic Review and Meta-analysis of Randomised Controlled Feeding Trials.

BACKGROUND: Effects of major dietary macronutrients on glucose-insulin homeostasis remain controversial and may vary by the clinical measures examined. We aimed to assess how saturated fat (SFA), monounsaturated fat (MUFA), polyunsaturated fat (PUFA), and carbohydrate affect key metrics of glucose-insulin homeostasis. METHODS AND FINDINGS: We systematically searched multiple databases (PubMed, EMBASE, OVID, BIOSIS, Web-of-Knowledge, CAB, CINAHL, Cochrane Library, SIGLE, Faculty1000) for randomised controlled feeding trials published by 26 Nov 2015 that tested effects of macronutrient intake on blood glucose, insulin, HbA1c, insulin sensitivity, and insulin secretion in adults aged ≥18 years. We excluded trials with non-isocaloric comparisons and trials providing dietary advice or supplements rather than meals. Studies were reviewed and data extracted independently in duplicate. Among 6,124 abstracts, 102 trials, including 239 diet arms and 4,220 adults, met eligibility requirements. Using multiple-treatment meta-regression, we estimated dose-response effects of isocaloric replacements between SFA, MUFA, PUFA, and carbohydrate, adjusted for protein, trans fat, and dietary fibre. Replacing 5% energy from carbohydrate with SFA had no significant effect on fasting glucose (+0.02 mmol/L, 95% CI = -0.01, +0.04; n trials = 99), but lowered fasting insulin (-1.1 pmol/L; -1.7, -0.5; n = 90). Replacing carbohydrate with MUFA lowered HbA1c (-0.09%; -0.12, -0.05; n = 23), 2 h post-challenge insulin (-20.3 pmol/L; -32.2, -8.4; n = 11), and homeostasis model assessment for insulin resistance (HOMA-IR) (-2.4%; -4.6, -0.3; n = 30). Replacing carbohydrate with PUFA significantly lowered HbA1c (-0.11%; -0.17, -0.05) and fasting insulin (-1.6 pmol/L; -2.8, -0.4). Replacing SFA with PUFA significantly lowered glucose, HbA1c, C-peptide, and HOMA. Based on gold-standard acute insulin response in ten trials, PUFA significantly improved insulin secretion capacity (+0.5 pmol/L/min; 0.2, 0.8) whether replacing carbohydrate, SFA, or even MUFA. No significant effects of any macronutrient replacements were observed for 2 h post-challenge glucose or insulin sensitivity (minimal-model index). Limitations included a small number of trials for some outcomes and potential issues of blinding, compliance, generalisability, heterogeneity due to unmeasured factors, and publication bias. CONCLUSIONS: This meta-analysis of randomised controlled feeding trials provides evidence that dietary macronutrients have diverse effects on glucose-insulin homeostasis. In comparison to carbohydrate, SFA, or MUFA, most consistent favourable effects were seen with PUFA, which was linked to improved glycaemia, insulin resistance, and insulin secretion capacity.

Trends in trans fatty acids reformulations of US supermarket and brand-name foods from 2007 through 2011.

INTRODUCTION: Although some US food manufacturers have reduced trans fatty acids (TFA) in their products, it is unknown how much TFA is being reduced, whether pace of reformulation has changed over time, or whether reformulations vary by food type or manufacturer. METHODS: In 2007, we identified 360 brand-name products in major US supermarkets that contained 0.5 g TFA or more per serving. In 2008, 2010, and 2011, product labels were re-examined to determine TFA content; ingredients lists were also examined in 2011 for partially hydrogenated vegetable oils (PHVO). We assessed changes in TFA content among the 270 products sold in all years between 2007 and 2011 and conducted sensitivity analyses on the 90 products discontinued after 2007. RESULTS: By 2011, 178 (66%) of the 270 products had reduced TFA content. Most reformulated products (146 of 178, 82%) reduced TFA to less than 0.5 g per serving, although half of these 146 still contained PHVO. Among all 270 products, mean TFA content decreased 49% between 2007 and 2011, from 1.9 to 0.9 g per serving. Yet, mean TFA reduction slowed over time, from 30.3% (2007-2008) to 12.1% (2008-2010) to 3.4% (2010-2011) (P value for trend < .001). This slowing pace was due to both fewer reformulations among TFA-containing products at start of each period and smaller TFA reductions among reformulated products. Reformulations also varied substantially by both food category and manufacturer, with some eliminating or nearly eliminating TFA and others showing no significant changes. Sensitivity analyses were similar to main findings. CONCLUSIONS: Some US products and food manufacturers have made progress in reducing TFA, but substantial variation exists by food type and by parent company, and overall progress has significantly slowed over time. Because TFA consumption is harmful even at low levels, our results emphasize the need for continued efforts toward reformulating or discontinuing foods to eliminate PHVO.