A Machine Learning Approach for Predicting Real-time Risk of Intraoperative Hypotension in Traumatic Brain Injury.

BACKGROUND: Traumatic brain injury (TBI) is a major cause of death and disability. Episodes of hypotension are associated with worse TBI outcomes. Our aim was to model the real-time risk of intraoperative hypotension in TBI patients, compare machine learning and traditional modeling techniques, and identify key contributory features from the patient monitor and medical record for the prediction of intraoperative hypotension. METHODS: The data included neurosurgical procedures in 1005 TBI patients at an academic level 1 trauma center. The clinical event was intraoperative hypotension, defined as mean arterial pressure <65 mm Hg for 5 or more consecutive minutes. Two types of models were developed: one based on preoperative patient-level predictors and one based on intraoperative predictors measured per minute. For each of these models, we took 2 approaches to predict the occurrence of a hypotensive event: a logistic regression model and a gradient boosting tree model. RESULTS: The area under the receiver operating characteristic curve for the intraoperative logistic regression model was 0.80 (95% confidence interval [CI]: 0.78-0.83), and for the gradient boosting model was 0.83 (95% CI: 0.81-0.85). The area under the precision-recall curve for the intraoperative logistic regression model was 0.16 (95% CI: 0.12-0.20), and for the gradient boosting model was 0.19 (95% CI: 0.14-0.24). Model performance based on preoperative predictors was poor. Features derived from the recent trend of mean arterial pressure emerged as dominantly predictive in both intraoperative models. CONCLUSIONS: This study developed a model for real-time prediction of intraoperative hypotension in TBI patients, which can use computationally efficient machine learning techniques and a streamlined feature-set derived from patient monitor data.

Cytotoxic Effects of Common Irrigation Solutions on Chondrosarcoma and Giant Cell Tumors of Bone.

BACKGROUND: Irrigation is commonly used as an adjuvant treatment during the intralesional curettage of bone tumors. The goal of the present study was to analyze the in vitro cytotoxicity of commonly used irrigation solutions on chondrosarcoma and giant cell tumor (GCT) cells as there is no consensus on which solution leads to the greatest amount of cell death. METHODS: An in vitro evaluation was performed by exposing human GCT and human chondrosarcoma cell lines to 0.9% saline solution, sterile water, 70% ethanol, 3% hydrogen peroxide, 0.05% chlorhexidine gluconate (CHG), and 0.3% povidone iodine solutions independently for 2 and 5 minutes. A low-cytotoxicity control (LCC) and a high-cytotoxicity control (HCC) were established to determine the mean cytotoxicity of each solution and each solution's superiority to LCC and non-inferiority to HCC. RESULTS: The present study demonstrated that 0.05% CHG was non-inferior to the HCC when chondrosarcoma was exposed for 5 minutes and when GCT was exposed for 2 and 5 minutes (mean cytotoxicity, 99% to 102%) (p < 0.003 for all). Sterile water was superior to the LCC when chondrosarcoma was exposed for 5 minutes and when GCT was exposed for 2 minutes (mean, 28% to 37%) (p < 0.05). Sterile water (mean, 18% to 38%) (p < 0.012) and 3% hydrogen peroxide (mean, 7% to 16%) (p < 0.001) were both inferior to the HCC. The 3 other solutions were non-superior to the LCC (mean, -24% to -5%) (p < 0.023). CONCLUSIONS: In vitro irrigation in 0.05% CHG provided high cytotoxicity, comparable with the HCC. Therefore, the use of a 0.05% CHG solution clinically could serve as a potential chemical adjuvant during intralesional curettage of chondrosarcoma and GCT. CLINICAL RELEVANCE: In an effort to reduce the burden of residual tumor cells, irrigation solutions are often utilized as adjuvant local therapy. Use of a 0.05% CHG solution clinically could serve as a potential chemical adjuvant to intralesional curettage of chondrosarcoma and GCT. Further in vivo studies may be indicated to assess clinical outcomes and safety associated with the use of 0.05% CHG in the treatment of chondrosarcoma and GCT.

Implementing Primary Palliative Care in Post-acute nursing home care: Protocol for an embedded pilot pragmatic trial.

INTRODUCTION: Older adults with serious illness frequently receive post-acute rehabilitative care in nursing homes (NH) under the Part A Medicare Skilled Nursing Facility (SNF) Benefit. Treatment is commonly focused on disease-modifying therapies with minimal consideration for goals of care, symptom relief, and other elements of palliative care. INTERVENTION: The evidence-based Primary Palliative Care in Post-Acute Care (PPC-PAC) intervention for older adults is delivered by nurse practitioners (NP). PPC-PAC NPs assess and manage symptoms, conduct goals of care discussions and assist with decision making; they communicate findings with NH staff and providers. Implementation of PPC-PAC includes online and face-to-face training of NPs, ongoing facilitation, and a template embedded in the NH electronic health record to document PPC-PAC. OBJECTIVES: The objectives of this pilot pragmatic clinical trial are to assess the feasibility, acceptability, and preliminary effectiveness of the PPC-PAC intervention and its implementation for 80 seriously ill older adults newly admitted to a NH for post-acute care. METHODS: Design is a two-arm nonequivalent group multi-site pilot pragmatic clinical trial. The unit of assignment is at the NP and unit of analysis is NH patients. Recruitment occurs at NHs in Pennsylvania, New Jersey, Delaware, and Maryland. Effectiveness (patient quality of life) data are collected at two times points-baseline and 14-21 days. CONCLUSION: This will be the first study to evaluate the implementation of an evidence-based primary palliative care intervention specifically designed for older adults with serious illness who are receiving post-acute NH care.

Retrospective analysis of palliative metronomic chemotherapy in head and neck cancer.

BACKGROUND: Metronomic chemotherapy has shown promising results in selected patients of head and neck cancer in a small randomized study. This retrospective analysis was done to see whether the efficacy of metronomic chemotherapy in an unselected cohort of head and neck cancer patients is similar to that reported in the randomized study and the influence of the site and subsite of head and neck cancer on survival. MATERIALS AND METHODS: This was a retrospective analysis of head and neck cancer patients who received palliative metronomic chemotherapy between January 2013 and February 2015. The data of these patients were collected from our palliative chemotherapy database maintained in medical oncology outpatient department. The overall survival (OS) was calculated in days from the date of start of chemotherapy to the date of death. Patients who had not expired at last follow-up were censored during estimation of OS by Kaplan-Meier method. Factors affecting OS were identified by COX regression analysis. RESULTS: Over the stipulated time period, 340 patients received palliative metronomic chemotherapy. The median age of these patients was 48 years (22-90 years). The sites of tumor origin were oral cavity in 281 patients (82.6%), oropharynx in 33 patients (9.7%), larynx in seven patients (2.1%), hypopharynx in 12 patients (3.5%), and maxilla in seven patients (2.1%). Previous treatment was received by 286 patients (84.1%). The median time to failure was 3.5 months (interquartile range 2.0-6.0 months). The overall median survival was 155 days (95% confidence interval 140.2-169.8 days). Failure within 6 months of previous treatment was the most important factor influencing OS. There was a trend toward lower OS in patients with oral cancers (139 days vs. 210 days). Among the various oral cancer subsites, oral tongue primary had a lower OS. CONCLUSION: Oral metronomic chemotherapy has promising results when used in a selected cohort of patients but has dismal results in patients who failed within 6 months of previous treatment.

Carotid Plaque Lipid Content and Fibrous Cap Status Predict Systemic CV Outcomes: The MRI Substudy in AIM-HIGH.

OBJECTIVES: The aim of this study was to investigate whether and what carotid plaque characteristics predict systemic cardiovascular outcomes in patients with clinically established atherosclerotic disease. BACKGROUND: Advancements in atherosclerosis imaging have allowed assessment of various plaque characteristics, some of which are more directly linked to the pathogenesis of acute cardiovascular events compared to plaque burden. METHODS: As part of the event-driven clinical trial AIM-HIGH (Atherothrombosis Intervention in Metabolic Syndrome with Low HDL/High Triglycerides: Impact on Global Health Outcomes), subjects with clinically established atherosclerotic disease underwent multicontrast carotid magnetic resonance imaging (MRI) to detect plaque tissue composition and high-risk features. Prospective associations between MRI measurements and the AIM-HIGH primary endpoint (fatal and nonfatal myocardial infarction, ischemic stroke, hospitalization for acute coronary syndrome, and symptom-driven revascularization) were analyzed using Cox proportional hazards survival models. RESULTS: Of the 232 subjects recruited, 214 (92.2%) with diagnostic image quality constituted the study population (82% male, mean age 61 ± 9 years, 94% statin use). During median follow-up of 35.1 months, 18 subjects (8.4%) reached the AIM-HIGH endpoint. High lipid content (hazard ratio [HR] per 1 SD increase in percent lipid core volume: 1.57; p = 0.002) and thin/ruptured fibrous cap (HR: 4.31; p = 0.003) in carotid plaques were strongly associated with the AIM-HIGH endpoint. Intraplaque hemorrhage had a low prevalence (8%) and was marginally associated with the AIM-HIGH endpoint (HR: 3.00; p = 0.053). High calcification content (HR per 1 SD increase in percent calcification volume: 0.66; p = 0.20), plaque burden metrics, and clinical risk factors were not significantly associated with the AIM-HIGH endpoint. The associations between carotid plaque characteristics and the AIM-HIGH endpoint changed little after adjusting for clinical risk factors, plaque burden, or AIM-HIGH randomized treatment assignment. CONCLUSIONS: Among patients with clinically established atherosclerotic disease, carotid plaque lipid content and fibrous cap status were strongly associated with systemic cardiovascular outcomes. Markers of carotid plaque vulnerability may serve as novel surrogate markers for systemic atherothrombotic risk.

Does Early Surgical Decompression in Cauda Equina Syndrome Improve Bladder Outcome?: Comments on a Recent Study by Srikandarajah et al.

STUDY DESIGN AND OBJECTIVES: This study by Srikandarajah et al is a retrospective cohort study of 200 CES patients. Although parts of the study appear to be well done we identified serious problems that impacted their conclusions. SUMMARY OF BACKGROUND DATA: There is strong consensus that patients with incomplete cauda equina syndrome (CESI) should be operated upon expeditiously to prevent progression to CES with urinary retention (CESR). There is controversy concerning optimal timing of surgery in patients who have entered CESR. Some studies conclude urgent surgery for CESR does nothing to improve chance of recovery; others conclude surgery should be done within a 48 or 24-hours window after onset of bladder paralysis. Srikandarajah et al concluded that in CESI patients, decompressive surgery within 24 hours of onset of autonomic symptoms reduces bladder dysfunction at follow-up. In CESR patients, "no statistically significant difference in outcome was observed" regarding timing of surgery. METHODS: We analyzed the methods used by Srikandarajah et al to collect and analyze their data. RESULTS: The primary problem with their study is that it does not consider deterioration to CESR that occurs in CESI patients over time. We also found serious problems with the methods, implementation (including numeric errors), and interpretation of the statistical analysis. CONCLUSION: The authors' conclusion that in CESR patients, "no statistically significant difference in outcome was observed" regarding timing of surgery is not justified as a representation of their findings, because absence of statistical significance does not mean acceptance of the null hypothesis of "no effect". Their numeric results do show elevated risk of delay. We do not want to detract from the basic importance of this study, because it emphasizes the importance of operating expeditiously on patients with CESI. However, the authors' conclusions are compromised by the methodological problems. LEVEL OF EVIDENCE: N/A.

A review on host-pathogen interactions: classification and prediction.

The research on host-pathogen interactions is an ever-emerging and evolving field. Every other day a new pathogen gets discovered, along with comes the challenge of its prevention and cure. As the intelligent human always vies for prevention, which is better than cure, understanding the mechanisms of host-pathogen interactions gets prior importance. There are many mechanisms involved from the pathogen as well as the host sides while an interaction happens. It is a vis-a-vis fight of the counter genes and proteins from both sides. Who wins depends on whether a host gets an infection or not. Moreover, a higher level of complexity arises when the pathogens evolve and become resistant to a host's defense mechanisms. Such pathogens pose serious challenges for treatment. The entire human population is in danger of such long-lasting persistent infections. Some of these infections even increase the rate of mortality. Hence there is an immediate emergency to understand how the pathogens interact with their host for successful invasion. It may lead to discovery of appropriate preventive measures, and the development of rational therapeutic measures and medication against such infections and diseases. This review, a state-of-the-art updated scenario of host-pathogen interaction research, has been done by keeping in mind this urgency. It covers the biological and computational aspects of host-pathogen interactions, classification of the methods by which the pathogens interact with their hosts, different machine learning techniques for prediction of host-pathogen interactions, and future scopes of this research field.

Summary of clinical and laboratory data of study subjects with and without DCE-MRI plaque measurements in the AIM-HIGH clinical trial.

This brief data article summarizes the clinical risk factors and laboratory data of a group of subjects recruited for the AIM-HIGH trial (Atherothrombosis Intervention in Metabolic Syndrome with Low HDL/High Triglycerides and Impact on Global Health Outcomes) and an associated magnetic resonance imaging (MRI) substudy. The sample is restricted to those on statin therapy at the time of enrollment and data are presented stratified by whether dynamic contrast enhanced MRI (DCE-MRI) markers of carotid plaque vascularity and inflammation were available or not. The data provided herein are directly related to the article "Longer Duration of Statin Therapy is Associated with Decreased Carotid Plaque Vascularity by Magnetic Resonance Imaging" [2].

Shoe and Field Surface Risk Factors for Acute Lower Extremity Injuries Among Female Youth Soccer Players.

OBJECTIVE: To describe acute lower extremity injuries and evaluate extrinsic risk factors in female youth soccer. DESIGN: Nested case-control study. SETTING: Youth soccer clubs in Seattle, WA. PARTICIPANTS: Female soccer players (n = 351) ages 11 to 15 years randomly selected from 4 soccer clubs from which 83% of their players were enrolled with complete follow-up for 92% of players. INTERVENTIONS: Injured players were interviewed regarding injury, field surface, shoe type, and position. Uninjured controls, matched on game or practice session, were also interviewed. MAIN OUTCOME MEASURES: The association between risk factors and acute lower extremity injury using logistic regression to estimate odds ratios (OR) and 95% confidence intervals (CI). RESULTS: One hundred seventy-three acute lower extremity injuries occurred involving primarily the ankle (39.3%), knee (24.9%), and thigh (11.0%). Over half (52.9%) recovered within 1 week, whereas 30.2% lasted beyond 2 weeks. During practices, those injured were approximately 3-fold (OR, 2.83; 95% CI, 1.49-5.31) more likely to play on grass than artificial turf and 2.4-fold (95% CI, 1.03-5.96) more likely to wear cleats on grass than other shoe and surface combinations. During games, injured players were 89% (95% CI, 1.03-4.17) more likely to play defender compared with forward. CONCLUSIONS: Half of the acute lower extremity injuries affected the ankle or knee. Grass surface and wearing cleats on grass increased training injuries. CLINICAL RELEVANCE: The majority, 64%, of female youth soccer players' acute injuries involve the ankle and knee and injury prevention strategies in this age group should target these areas. When considering playing surfaces for training, communities and soccer organizations should consider the third-generation artificial turf a safe alternative to grass.

Longer duration of statin therapy is associated with decreased carotid plaque vascularity by magnetic resonance imaging.

OBJECTIVE: Plaque neovasculature is a major route for lipoprotein and leukocyte ingress into plaques, and has been identified as a risk factor for carotid plaque disruption. Vp, a variable derived from pharmacokinetic modeling of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI), correlates with plaque neovasculature density. Because lipid-lowering therapy has been associated with regression of neovasculature in animal models, we sought to determine clinical correlates of carotid plaque neovasculature (as assessed by Vp) in participants on statin therapy for established cardiovascular disease. METHODS: 98 participants from an AIM-HIGH sub-study underwent DCE-MRI of their carotid arteries. Expert readers who were blinded to all clinical variables analyzed the MR images to measure carotid plaque Vp in all participants. Associations between Vp and duration of statin therapy and other clinical risk factors were analyzed. RESULTS: Prior duration of statin treatment at enrollment ranged from <1 year (21%) 1-5 years (40%) and >5 years (39%). In univariate analyses, shorter duration of statin therapy (P = 0.01), the presence of metabolic syndrome (P = 0.02), and higher body mass index (P = 0.01) and lipoprotein(a) (P = 0.01) were all significantly associated with higher baseline Vp values. In multivariate analyses, significant associations remained between shorter duration of statin therapy (P = 0.004) and lipoprotein(a) (P = 0.04). CONCLUSIONS: These are the first human, in vivo findings suggesting a relationship between duration of statin therapy and regression of carotid plaque neovasculature. Future longitudinal studies are warranted both to confirm this finding and to address whether changes in neovasculature may translate into change in risk for plaque disruption. CLINICALTRIALS. GOV IDENTIFIERS: NCT00880178, NCT01178320 and NCT00120289.

A randomized, double-blind phase I/IIa study of intranasal glutathione in Parkinson's disease.

BACKGROUND: Depletion of reduced glutathione is associated with PD and glutathione augmentation has been proposed as a disease-modifying strategy. The aim of this study was to determine the safety and tolerability of intranasal reduced glutathione in individuals with PD. METHODS: Thirty individuals with PD were randomized to either placebo (saline), 300 mg/day, or 600 mg/day of intranasal glutathione in three divided daily doses. Follow-up visits included side effect screening of PD symptoms and cognition, blood chemistry, sinus irritation, and hyposmia. Tolerability was measured by frequency and severity of reported adverse events, compliance, and withdrawals from the study. RESULTS: After 3 months, there were no substantial differences between groups in the number of adverse events reported or observed among all safety measures assessed. All groups met tolerability criteria. CONCLUSIONS: These data support the safety and tolerability of intranasal glutathione in this population. Pharmacokinetic and dose-finding studies are warranted.

The accuracy of handheld pre-arrest breath test instruments as a predictor of the evidential breath alcohol test results.

Drunk driving is a serious threat to public safety. All available and appropriate tools for curbing this threat should be employed to their full extent. The handheld pre-arrest breath test instrument (PBT) is one tool for identifying the alcohol-impaired driver and enforcing drunk driving legislation. A set of data was evaluated (n = 1779) where the PBT instrument was employed in drunk driving arrests to develop a multivariate predictive model. When maintained and operated by trained personnel, the PBT provides a reasonable estimate of the evidential test result within the relevant forensic range (95% prediction interval:  ± 0.003 g/210 L). ROC analysis shows that a multivariate model for PBT prediction of the evidentiary alcohol concentration above versus below the legal limit of 0.08 g/210 L has excellent performance with an AUC of 0.96. These results would be of value in evidential hearings seeking to admit the PBT results in drunk driving trials.

Evaluation of the Zachery Lystedt Law among female youth soccer players.

BACKGROUND: Despite recent increased awareness about sports concussions, few studies have evaluated the effect of concussion laws on concussion outcomes among young athletes. The purpose of our study was to evaluate the effect of the Washington State Zachery Lystedt Concussion Law on playing with concussion symptoms and being evaluated by a health care provider. METHODS: We performed a prospective cohort study of 351 elite female soccer players, aged 12 to 15 years, from 33 randomly selected youth soccer teams in the Puget Sound region of Washington State from 2008 to 2012. The Washington State Zachery Lystedt Concussion Law went into effect on July 1, 2009. Among concussed players (N = 59), we assessed the risk of playing with symptoms, the evaluation by a health care professional, and receiving a concussion diagnosis before and after the law was passed using logistic regression to estimate odds ratios and 95% CIs. RESULTS: The majority of concussed players (59.3%) continued to play with symptoms, and we found no statistically significant difference in the proportion of players who played with symptoms before and after the law was passed. Only 44.1% of concussed players were evaluated by a health care provider, with no difference before and after the law was passed. Among those evaluated by a health care professional, players were 2.1-fold (95% CI, 1.0-10.1) more likely to receive a concussion diagnosis after the law was passed. CONCLUSION: The majority of concussed female youth soccer players report playing with symptoms. Legislation mandating concussion education and evaluation prior to returning to play was not associated with an increase in concussion evaluations by health care providers.

Reproducibility of a quantitative cutaneous cytological technique.

BACKGROUND: Cutaneous cytology is a valuable tool for diagnosis of canine superficial pyoderma. Current published reproducible techniques are semiquantitative. HYPOTHESIS/OBJECTIVES: The aim of this study was to evaluate the reproducibility of a quantitative method for skin surface cytology in dogs with superficial pyoderma. ANIMALS: Impression smears were collected from five normal dogs and 20 dogs with clinical and cytological evidence of superficial pyoderma. METHODS: Four investigators evaluated 10 oil immersion fields (OIF) on 25 slides, selecting fields with inflammatory cells, nuclear streaming and or keratinocytes under ×10 magnification. Investigators repeated blinded evaluations of all slides at least twice. For each OIF, polymorphonuclear leukocytes (PMNs), intracellular (IC) cocci, extracellular (EC) cocci, IC rods, EC rods and yeast were quantified. Nuclear streaming was scored as present or absent. For each parameter, within-reader and between-reader agreements were expressed by the intraclass correlation (ICC) value (≤0.20 poor, 0.21-0.40 fair, 0.41-0.60 moderate, 0.61-0.80 good and 0.81-1.00 excellent) or kappa statistic (κ). RESULTS: Reproducible parameters included: PMNs (ICC = 0.58), nuclear streaming (ICC = 0.68), EC cocci (ICC = 0.64) and IC cocci (ICC = 0.32). When qualified as present or absent, within-reader κ for IC cocci was 0.71. The method demonstrated 93% sensitivity in identifying dogs with superficial pyoderma and 51% specificity in identifying normal dogs according to established criteria. However, if criteria for normal dogs were limited to the absence of PMNs and IC bacteria, sensitivity of 64% and specificity of 98% were demonstrated. CONCLUSIONS AND CLINICAL IMPORTANCE: For several parameters, including PMNs, nuclear streaming, EC cocci and IC cocci, a reproducible, quantitative cytological technique was identified.