Predicting respiratory decompensation in mechanically ventilated adult ICU patients.

Introduction: Mechanical ventilation is a life-saving treatment in the Intensive Care Unit (ICU), but often causes patients to be at risk of further respiratory complication. We created a statistical model utilizing electronic health record and physiologic vitals data to predict the Center for Disease Control and Prevention (CDC) defined Ventilator Associated Complications (VACs). Further, we evaluated the effect of data temporal resolution and feature generation method choice on the accuracy of such a constructed model. Methods: We constructed a random forest model to predict occurrence of VACs using health records and chart events from adult patients in the Medical Information Mart for Intensive Care III (MIMIC-III) database. We trained the machine learning models on two patient populations of 1921 and 464 based on low and high frequency data availability. Model features were generated using both basic statistical summaries and tsfresh, a python library that generates a large number of derived time-series features. Classification to determine whether a patient will experience VAC one hour after 35 h of ventilation was performed using a random forest classifier. Two different sample spaces conditioned on five varying feature extraction techniques were evaluated to identify the most optimal selection of features resulting in the best VAC discrimination. Each dataset was assessed using K-folds cross-validation (k = 10), giving average area under the receiver operating characteristic curves (AUROCs) and accuracies. Results: After feature selection, hyperparameter tuning, and feature extraction, the best performing model used automatically generated features on high frequency data and achieved an average AUROC of 0.83 ± 0.11 and an average accuracy of 0.69 ± 0.10. Discussion: Results show the potential viability of predicting VACs using machine learning, and indicate that higher-resolution data and the larger feature set generated by tsfresh yield better AUROCs compared to lower-resolution data and manual statistical features.

Alcohol and IL-6 Alter Expression of Synaptic Proteins in Cerebellum of Transgenic Mice with Increased Astrocyte Expression of IL-6.

Recent studies indicate that neuroimmune factors, including the cytokine interleukin-6 (IL-6), play a role in the CNS actions of alcohol. The cerebellum is a sensitive target of alcohol, but few studies have examined a potential role for neuroimmune factors in the actions of alcohol on this brain region. A number of studies have shown that synaptic transmission, and in particular inhibitory synaptic transmission, is an important cerebellar target of alcohol. IL-6 also alters synaptic transmission, although it is unknown if IL-6 targets are also targets of alcohol. This is an important issue because alcohol induces glial production of IL-6, which could then covertly influence the actions of alcohol. The persistent cerebellar effects of both IL-6 and alcohol typically involve chronic exposure and, presumably, altered gene and protein expression. Thus, in the current studies we tested the possibility that proteins involved in inhibitory and excitatory synaptic transmission in the cerebellum are common targets of alcohol and IL-6. We used transgenic mice that express elevated levels of astrocyte produced IL-6 to model persistently elevated expression of IL-6, as would occur in alcohol use disorders, and a chronic intermittent alcohol exposure/withdrawal paradigm (CIE/withdrawal) that is known to produce alcohol dependence. Multiple cerebellar synaptic proteins were assessed by Western blot. Results show that IL-6 and CIE/withdrawal have both unique and common actions that affect synaptic protein expression. These common targets could provide sites for IL-6/alcohol exposure/withdrawal interactions and play an important role in cerebellar symptoms of alcohol use such as ataxia.

Behavioral Effects of Neuronal, Parent-specific Commd1 Knockout in Mice.

In this study we focused on gene expression and behavioral differences in mice with brain-specific Commd1 knockout. Commd1 is an imprinted gene with preferential maternal expression, residing within a larger genomic region previously found to affect sensorimotor gating. In this study, individuals harboring a conditional Commd1 mutant allele were bred with Syn1-Cre animals, paying special attention to the parent of origin of the Commd1 mutation. Analysis of mRNA levels of Commd1 and phenotypic tests, including the open field, sensorimotor gating, and the forced swim test, were conducted on offspring with either maternally or paternally derived Commd1 knockout. We found that measurable Commd1 mRNA knockout occurred only in the maternally derived line and affected stereotypy and depressive-like behavior without differences in total locomotion compared to controls. Interestingly, we found that maternal knockout animals exhibited decreased time swimming and increased time immobile when compared to maternal and paternal wild type, and paternal knockout animals. However, there were no differences in climbing behavior between genotypes. This study demonstrates an in vivo behavioral role for Commd1 for the first time and demonstrates the need for careful interpretation of experimental results involving Cre-based knockout systems.

Acute respiratory symptoms in adults in general practice.

BACKGROUND: Community studies have shown that approximately 30% of patients with acute respiratory tract symptoms have no identifiable infective aetiology. This may not be applicable in general practice. OBJECTIVE: The purpose of this study was to determine the infective aetiology in patients who presented to primary care doctors with acute respiratory symptoms. METHODS: A prospective study was carried out in all nine primary care clinics belonging to the National Healthcare Group Polyclinics (NHGPs) in Singapore. The subjects comprised 594 consecutive patients (318 males, 276 females) aged > or = 21 years who presented with complaints of any one of cough, nasal or throat symptoms of <7 days duration. Data collection was through interview using structured questionnaire, physical examination, throat swabs for bacterial culture and nasal swabs for virus identification by immunofluorescence (IF) and polymerase chain reaction (PCR). Additional PCR was performed on a subsample of 100 patients. Patients were followed-up until resolution of symptoms. RESULTS: The aetiological diagnosis by infective agent is as follows: 150 patients (25.2%) had virus infections, of which 90.7% (136/150) were by rhinovirus. Fourteen patients (2.4%) had bacterial infections, of which 10 were due to group G streptococcus. Group A streptococcus was not detected. Nineteen patients with new pathogens were identified by further PCR. These included parainfluenza 4, human coronavirus OC43, adenovirus, enterovirus and Chlamydia pneumoniae. No pathogen could be identified in 49% of patients. There were no differences in clinical presentation and socio-demographic variables between patients who had viral infections and those in whom no pathogen could be identified. CONCLUSION: In about half of patients who presented at NHGPs, no pathogens could be identified even after PCR. A non-infective aetiology could be considered in these patients.