Current stroke risk assessment tools presume the impact of risk factors is linear and cumulative. However, both novel risk factors and their interplay influencing stroke incidence are difficult to ...reveal using traditional additive models. The goal of this study was to improve upon the established Revised Framingham Stroke Risk Score and design an interactive Non-Linear Stroke Risk Score. Leveraging machine learning algorithms, our work aimed at increasing the accuracy of event prediction and uncovering new relationships in an interpretable fashion. A two-phase approach was used to create our stroke risk prediction score. First, clinical examinations of the Framingham offspring cohort were utilized as the training dataset for the predictive model. Optimal Classification Trees were used to develop a tree-based model to predict 10-year risk of stroke. Unlike classical methods, this algorithm adaptively changes the splits on the independent variables, introducing non-linear interactions among them. Second, the model was validated with a multi-ethnicity cohort from the Boston Medical Center. Our stroke risk score suggests a key dichotomy between patients with history of cardiovascular disease and the rest of the population. While it agrees with known findings, it also identified 23 unique stroke risk profiles and highlighted new non-linear relationships; such as the role of T-wave abnormality on electrocardiography and hematocrit levels in a patient's risk profile. Our results suggested that the non-linear approach significantly improves upon the baseline in the c-statistic (training 87.43% (CI 0.85-0.90) vs. 73.74% (CI 0.70-0.76); validation 75.29% (CI 0.74-0.76) vs 65.93% (CI 0.64-0.67), even in multi-ethnicity populations. The clinical implications of the new risk score include prioritization of risk factor modification and personalized care at the patient level with improved targeting of interventions for stroke prevention.
Limited real-world data exist comparing each non-vitamin K antagonist oral anticoagulant (NOAC) to warfarin in patients with nonvalvular atrial fibrillation who have had a previous ischemic stroke or ...transient ischemic attack.
Using MarketScan claims from January 2012 to June 2015, we identified adults newly initiated on oral anticoagulation, with ≥2 diagnosis codes for nonvalvular atrial fibrillation, a history of previous ischemic stroke/transient ischemic attack, and ≥180 days of continuous medical and prescription benefits before anticoagulation initiation. Three analyses were performed comparing 1:1 propensity score-matched cohorts of apixaban versus warfarin (n=2514), dabigatran versus warfarin (n=1962), and rivaroxaban versus warfarin (n=5208). Patients were followed until occurrence of a combined end point of ischemic stroke and intracranial hemorrhage (ICH) or major bleed, switch/discontinuation of index oral anticoagulation, insurance disenrollment, or end of follow-up. Mean follow-up was 0.5 to 0.6 years for all matched cohorts.
Using Cox regression, neither apixaban nor dabigatran reduced the combined primary end point of ischemic stroke or ICH (hazard ratio HR, 0.70; 95% confidence interval CI, 0.33-1.48 and HR, 0.53; 95% CI, 0.26-1.07) and had nonsignificant effect on hazards of major bleeding (HR, 0.79; 95% CI, 0.38-1.64 and HR, 0.58; 95% CI, 0.26-1.27) versus warfarin. Rivaroxaban reduced the combined end point of ischemic stroke or ICH (HR, 0.45; 95% CI, 0.29-0.72) without an effect on major bleeding (HR, 1.07; 95% CI, 0.71-1.61). ICH occurred at rates of 0.16 to 0.61 events per 100 person-years in the 3 NOAC analyses, with no significant difference for any NOAC versus warfarin.
Results from our study of the 3 NOACs versus warfarin in nonvalvular atrial fibrillation patients with a previous history of stroke/transient ischemic attack are relatively consistent with their respective phase III trials and previous stroke/transient ischemic attack subgroup analyses. All NOACs seemed no worse than warfarin in respect to ischemic stroke, ICH, or major bleeding risk.
Abstract Background Prediction models for cardiovascular events and cardiovascular death in patients with established cardiovascular disease are not generally available. Methods Participants from the ...prospective REduction of Atherothrombosis for Continued Health (REACH) Registry provided a global outpatient population with known cardiovascular disease at entry. Cardiovascular prediction models were estimated from the 2-year follow-up data of 49,689 participants from around the world. Results A developmental prediction model was estimated from 33,419 randomly selected participants (2394 cardiovascular events with 1029 cardiovascular deaths) from the pool of 49,689. The number of vascular beds with clinical disease, diabetes, smoking, low body mass index, history of atrial fibrillation, cardiac failure, and history of cardiovascular event(s) <1 year before baseline examination increased risk of a subsequent cardiovascular event. Statin (hazard ratio 0.75; 95% confidence interval, 0.69-0.82) and acetylsalicylic acid therapy (hazard ratio 0.90; 95% confidence interval, 0.83-0.99) also were significantly associated with reduced risk of cardiovascular events. The prediction model was validated in the remaining 16,270 REACH subjects (1172 cardiovascular events, 494 cardiovascular deaths). Risk of cardiovascular death was similarly estimated with the same set of risk factors. Simple algorithms were developed for prediction of overall cardiovascular events and for cardiovascular death. Conclusions This study establishes and validates a risk model to predict secondary cardiovascular events and cardiovascular death in outpatients with established atherothrombotic disease. Traditional risk factors, burden of disease, lack of treatment, and geographic location all are related to an increased risk of subsequent cardiovascular morbidity and cardiovascular mortality.
Background There is a paucity of contemporary data estimating the incidence of major adverse cardiovascular events (MACE) in patients with established atherosclerotic disease or multiple risk factors ...managed in routine practice. We estimated 1- and 4-year incidences of MACE and the association between MACE and vascular beds affected in these patients. Methods and Results Using US IBM MarketScan data from January 1, 2013 to December 31, 2017, we identified patients ≥45 years old with established coronary artery disease, cerebrovascular disease, peripheral artery disease, or the presence of ≥3 risk factors for atherosclerosis during 2013 with a minimum of 4 years of follow-up. We calculated 1- and 4-year incidences of MACE (cardiovascular death or hospitalization for myocardial infarction or ischemic stroke). A Cox proportional hazards regression model adjusted for age and sex was used to evaluate the association between vascular bed number/location(s) affected and MACE. We identified 1 302 856 patients with established atherosclerotic disease or risk factors for atherosclerosis. Coronary artery disease was present in 16.9% of patients, cerebrovascular disease in 7.6%, peripheral artery disease in 13.6%, and risk factors for atherosclerosis only in 66.0%. The 1- and 4-year incidences of MACE were 1.4% and 6.9%, respectively. At 4 years, MACE was more frequent in patients with atherosclerotic disease in a single (hazard ratio=1.51, 95% CI=1.48-1.55), 2-(hazard ratio=2.35, 95% CI=2.27-2.44), or all 3 vascular beds (hazard ratio=3.30, 95% CI=2.97-3.68) compared with having risk factors for atherosclerosis. Conclusions Patients with established atherosclerotic disease or who have multiple risk factors and are treated in contemporary, routine practice carry a substantial risk for MACE at 1- and 4- years of follow-up. MACE risk was shown to vary based on the number and location of vascular beds involved.
Implementation of prehospital stroke triage is a public policy intervention that can have an immediate impact on acute stroke care in a region. OBJECTIVE To evaluate the impact that a citywide policy ...recommending prehospital triage of patients with suspected stroke to the nearest primary stroke center had on intravenous tissue plasminogen activator (tPA) use in Chicago, Illinois.
Retrospective multicenter cohort study from September 1, 2010, to August 31, 2011 (6 months before and after intervention that began March 1, 2011).
Ten primary stroke center hospitals in Chicago.
All admitted patients with stroke and transient ischemic attack. INTERVENTION Prehospital triage policy of patients with stroke to primary stroke centers.
Intravenous tPA use (measured as a fraction of patients with ischemic strokes arriving through the emergency department). RESULTS There were 1075 stroke and transient ischemic attack admissions in the pretriage period and 1172 in the posttriage period. Patient demographic characteristics including age, sex, and risk factors were similar between the 2 periods (mean age, 65 years; 53% female). Compared with the pretriage period, use of emergency medical services increased from 30.2% to 38.1% (P < .001) and emergency medical services prenotification increased from 65.5% to 76.5% (P = .001) after implementation. Rates of intravenous tPA use were 3.8% and 10.1% (P < .001) and onset-to-treatment times decreased from 171.7 to 145.7 minutes (P = .03) in the pretriage and posttriage periods, respectively. Stroke unit admission, symptomatic intracranial hemorrhage rates, and in-hospital mortality were not significantly different between periods. Adjusting for mode of arrival, prehospital notification, and onset-to-arrival time, the posttriage period was independently associated with increased tPA use for patients with ischemic stroke presenting through the emergency department (adjusted odds ratio = 2.21; 95% CI, 1.34-3.64).
Implementation of a prehospital stroke triage policy in Chicago resulted in significant improvements in emergency medical services use and prenotification and more than doubled intravenous tPA use at primary stroke centers.