Measuring air pollution from the 2021 Canary Islands volcanic eruption

The eruption of the Cumbre Vieja volcano on the island of La Palma (Canary Islands, Spain) began on September 19, 2021 and ended on December 13, 2021. It lasted continuously for 85 days with short periods of calm when lava did not exit the cone of the volcano. Vast amounts of volcanic material, including ash and gases, were emitted into the environment. This research focuses on these emissions. The main objective is to use available open-source data to examine the impact on regional and local air quality. Data from the following sources were used: 1) Copernicus Atmosphere Monitoring Service (CAMS) data was used to track the transfer of volcanic SO₂ in the troposphere in early October over long distances from the source of the eruption, including Western and Eastern Europe, across the Atlantic Ocean and the Caribbean; 2) Data from ground monitoring stations measured the concentrations of SO₂ and PM₁₀ near the source; 3) AErosol RObotic NETwork (AERONET) data from the La Palma station that showed high Aerosol Optical Depth (AOD) values (over 0.4) during the active phase of emissions on September 24 and 28, as well as on October 3; 4) Ångström Exponent (AE) values indicated the presence of particles of different sizes. On September 24, high AE values (>1.5), showed the presence of fine-mode fraction scattering aerosols such as sulfates; 5) Cloud Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) data additionally confirmed the presence of sulfate and dust aerosols in the atmosphere over the region. However, the influence of Saharan dust on the atmosphere of the entire region could not be excluded. This research helps forecast air pollution resulting from large-scale volcanic eruptions and associated health risks to humans.

Application of a clinical decision support system to assess the severity of the new coronavirus infection COVID-19

Purpose

To apply the clinical decision support system (CDSS) and evaluate its effectiveness in determining the prognosis of the new coronavirus infection COVID-19.

Methods

7118 outpatient and hospitalized cases with COVID-19 were analyzed, mean age 39.4±18.3 years, 52% men. The data was accumulated in the the Webiomed.DataSet service, which allows to accumulate a base of de-identified biomedical data from electronic health records. To test the severity of the COVID-19, the CDSS was connected to the 8 medical information systems in one region of the Russian Federation. For each risk factor (RF) of the unfavorable COVID-19 outcome the contribution to the risk was determined in points indicated in brackets: age over 60 (1), age over 80 (3), BMI of 30–34.9 kg/m2 (1), BMI of ≥35 kg/m2 (2), arterial hypertension (HTN) (1), diabetes mellitus (DM) (1), coronary artery disease (CAD) (2), cerebrovascular accident (CVA) (1), atrial fibrillation (AF) (1), pulmonary disease (1), cancer (1). Hospitalization and death were considered as unfavorable outcome. Each patient had risk level (high – two or more points, moderate – one, low – zero).

Results

64.2% was outpatient, age 34.8±17.3 years. 35.8% was hospitalized (mean age 47.8±15.1 y), 50 patients died (mean age 61.3±14.4 y, mortality 0.7%). Low risk had 74.9% outpatient treated patients, 26.4% – hospitalized, 26% – dead; average risk – 12.6%, 17.3%, 24%, high risk – 12.5%, 56.3%, 50% (respectively for subgroups). The RF incidence of poor prognosis in the groups: age over 60 years old – 9%, over 80 – 0.5%, HTN – 18.9%, DM – 5.2%, CAD – 3.9%, CVA – 1%, AF – 1.4%, COPD/asthma – 1.9%, cancer – 1.7%, obesity – 15.5%. In the hospitalized group: age over 60 years – 11.1%, over 80 – 1.6%, HTN – 13.3%, DM – 4.2%, CAD – 7.6%, CVA – 1.1%, AF – 1%, COPD/asthma – 1.3%, cancer – 1.3%, obesity – 13.2%. Among patient who died: age over 60 years – 54%, over 80 – 6%, AH – 50%, DM – 18%, CAD – 36%, CVA – 4%, AF – 12%, COPD/asthma – 6%, cancer – 4%, obesity – 30%. When comparing the incidence of RF in the high-risk group, a significant difference in hospitalized, dead, and patients treated outpatient was obtained for the following RF: age over 60 years (p<0.001), HTN (p<0.001), DM (0.004), CAD (p<0.001), AF (p<0.001), COPD and AD (p=0.043), obesity (p=0.031). In the moderate-risk group, the main RFs influencing the prognosis were age over 60 years (p<0.001), HTN (p=0.03) and obesity (p=0.004).

Conclusions

The created CDSS allowed to stratify the risk of COVID-19 by the presence of cardiovascular risk factors and diseases, as well as by the presence of bronchopulmonary pathology and oncological diseases. The use of this CDSS allowed to route COVID-19 patients more effective. In addition to clinical criteria of the disease severity, the system allows to assess the prognosis quickly and hospitalize high-risk patients, or organize their careful monitoring in case of outpatient treatment.

Funding Acknowledgement

Type of funding sources: None.

Adapting neural network models to predict 10-year CVD development based on regional data calibration

Background

The used tools for prediction the individual risk of developing cardiovascular diseases and their complications using machine learning methods have proven better prognostic value in comparison with commonly used scales (e.g., Framingham, SCORE). To create such methods, the long-term accumulation of large amount of qualitative data are required. Moreover, to improve the accuracy of models, it is necessary to take into account regional characteristics that affect health: ethnic, nutritional characteristics, climatic conditions, living standards and medical care. These regional characteristics could significantly affect the development and outcomes of CVDs. However, the amount of regional data is not enough to build a qualitative model. Therefore, it is proposed to create models based on publicly available data and validate them on regional medical data sufficient for validation and calibration.

Methods

Two models were trained using data from the Framingham study. Model 1 was trained on 2 588 patient data and predicts a 10-year CVD probability according to the following risk factors: age, gender, cholesterol, HDL, smoking, SBP, and BP medications. Model 2 was trained on 4,363 patient data and predicts a 10-year death probability from CVD according to the following criteria: age, gender, cholesterol, smoking, SBP, BMI, heart rate. To retrain the obtained models, we used dataset created from data from patients in the northwestern part of Russia. The dataset consists of 438 patients, including the signs used in the trained models. This dataset includes CVD and death from it during a 10-year follow-up

Evaluation

We used randomized data splitting: divided the dataset into a training and a test set with an 80/20 proportion. The models was implement with keras convolution neural network (CNN) using 3 hidden layers. For data validation was used a 10 K-fold method.

Results

We compared the initial model metrics and those obtained after local data retraining. The accuracy of model 1 before retraining is 78%, after – 81.3%, the area under the ROC curve (AUC) before retraining: 0.77 (at 95% CI: 0.72–0.82C), after – 0.803. The accuracy of model 2 before retraining is 79%, after – 85.6%, the area under the ROC-curve (AUC) before retraining: 0.78 (at 95% CI: 0.72–0.82), after – 0.828.

Conclusion

Using this method of retraining predictive models, we can take into account local characteristics of the population and significantly increase the accuracy of predicting events. Expand the population to use the model according to local characteristics.

Funding Acknowledgement

Type of funding source: Private company. Main funding source(s): OOO K-SkAI

P1923Deep and machine learning models to improve risk prediction of cardiovascular disease using data extraction from electronic health records

Background

Advances in precision medicine will require an increasingly individualized prognostic evaluation of patients in order to provide the patient with appropriate therapy. The traditional statistical methods of predictive modeling, such as SCORE, PROCAM, and Framingham, according to the European guidelines for the prevention of cardiovascular disease, not adapted for all patients and require significant human involvement in the selection of predictive variables, transformation and imputation of variables. In ROC-analysis for prediction of significant cardiovascular disease (CVD), the areas under the curve for Framingham: 0.62–0.72, for SCORE: 0.66–0.73 and for PROCAM: 0.60–0.69. To improve it, we apply for approaches to predict a CVD event rely on conventional risk factors by machine learning and deep learning models to 10-year CVD event prediction by using longitudinal electronic health record (EHR).

Methods

For machine learning, we applied logistic regression (LR) and recurrent neural networks with long short-term memory (LSTM) units as a deep learning algorithm. We extract from longitudinal EHR the following features: demographic, vital signs, diagnoses (ICD-10-cm: I21-I22.9: I61-I63.9) and medication. The problem in this step, that near 80 percent of clinical information in EHR is “unstructured” and contains errors and typos. Missing data are important for the correct training process using by deep learning & machine learning algorithm. The study cohort included patients between the ages of 21 to 75 with a dynamic observation window. In total, we got 31517 individuals in the dataset, but only 3652 individuals have all features or missing features values can be easy to impute. Among these 3652 individuals, 29.4% has a CVD, mean age 49.4 years, 68,2% female.

Evaluation

We randomly divided the dataset into a training and a test set with an 80/20 split. The LR was implemented with Python Scikit-Learn and the LSTM model was implemented with Keras using Tensorflow as the backend.

Results

We applied machine learning and deep learning models using the same features as traditional risk scale and longitudinal EHR features for CVD prediction, respectively. Machine learning model (LR) achieved an AUROC of 0.74–0.76 and deep learning (LSTM) 0.75–0.76. By using features from EHR logistic regression and deep learning models improved the AUROC to 0.78–0.79.

Conclusion

The machine learning models outperformed a traditional clinically-used predictive model for CVD risk prediction (i.e. SCORE, PROCAM, and Framingham equations). This approach was used to create a clinical decision support system (CDSS). It uses both traditional risk scales and models based on neural networks. Especially important is the fact that the system can calculate the risks of cardiovascular disease automatically and recalculate immediately after adding new information to the EHR. The results are delivered to the user's personal account.

Effect of rosiglitazone on capillary density and angiogenesis in adipose tissue of normoglycaemic humans in a randomised controlled trial

AIMS/HYPOTHESIS

Recent reports of decreased capillary density in the adipose tissue of obese individuals suggest that an imbalance of angiogenesis and adipogenesis may, in part, underlie insulin resistance. This study aimed to determine whether the insulin-sensitising peroxisome proliferator-activated receptor γ (PPARγ) activator rosiglitazone affects adipose tissue vascularisation in normal humans.

METHODS

A randomised, parallel-group, investigator-blinded placebo-controlled trial was conducted with normoglycaemic volunteers with BMI 27-43, recruited from the community at the University of Massachusetts Medical School, Worcester, MA, USA. Peri-umbilical adipose tissue biopsies were obtained before and after treatment for 6 weeks with rosiglitazone (8 mg once daily) or placebo, which were randomly allocated from a sequentially numbered list. The primary outcomes were adipocyte size and capillary density measured by immunohistochemistry, and angiogenic potential assessed by capillary sprout formation in Matrigel. Secondary outcomes were serum adiponectin, glycaemic, lipid and liver function variables.

RESULTS

A total of 35 individuals fulfilling the inclusion criteria were randomised, and complete before-vs-after analyses were achieved in 30 participants (13 and 17, placebo and rosiglitazone, respectively). Significant differences, assessed by paired two-tailed Student t tests, were seen in response to rosiglitazone for adipocyte size (3,458 ± 202 vs 2,693 ± 223 μm(2), p = 0.0049), capillary density (5.6 ± 0.5 vs 7.5 ± 0.5 lumens/field, p = 0.0098), serum adiponectin (14.3 ± 1.5 vs 28.6 ± 3.0 ng/ml, p < 0.0001) and alkaline phosphatase (1.04 ± 0.07 vs 0.87 ± 0.05 μkat/l, p = 0.001). A difference in angiogenic potential before and after treatment between the placebo and rosiglitazone groups was also seen (-23.88 ± 14 vs 13.42 ± 13, p = 0.029, two-tailed Mann-Whitney test).

CONCLUSIONS/INTERPRETATION

Significant effects on adipose tissue vascular architecture occur after a short period of treatment with rosiglitazone in individuals with normal glucose tolerance. Improved adipose tissue vascularisation may, in part, mediate the therapeutic actions of this class of drugs.

TRIAL REGISTRATION

ClinicalTrials.gov NCT01150981

FUNDING

The study was funded by National Institutes of Health grant DK089101 to S. Corvera, and by pilot funding from the University of Massachusetts (UMASS) Center for Clinical Translational Sciences (M. Thompson, S. Malkani and S. Corvera). Morphology core services were supported by UMASS Diabetes Endocrine Research Center (DERC) grant DK32520.

A novel interface for on-line coupling of liquid capillary chromatography with matrix-assisted laser desorption/ionization detection

A novel interface has been developed which should allow the direct on-line coupling of liquid capillary chromatography with matrix-assisted laser desorption/ionization (MALDI) mass spectrometry detection. The interface employs continuous analyte/matrix co-crystallization onto a porous frit installed at a capillary end which is used as the target for MALDI. After separation, the analyte effluent is premixed with the MALDI matrix solution and introduced into the interface. The analyte/matrix mixture is co-crystallized onto the frit surface in the vacuum environment of the mass spectrometer. Continuous matrix/analyte crystallization and interface regeneration is accomplished by a combination of solvent flushing and laser ablation. The memory effect is negligible over a dynamic range of ca. 200. Several applications, including analysis of small peptides and combination with gel permeation chromatography (GPC), have indicated that the on-line MALDI interface does not sacrifice chromatographic or mass spectral resolution, and have demonstrated the possibility of a reliable LC-MALDI system. Copyright 1999 John Wiley & Sons, Ltd.

Characterization of polysiloxanes with different functional groups by time-of-flight secondary ion mass spectrometry

Polydimethylsiloxane (PDMS), polyhydromethylsiloxane (PHMS), and polymethylphenylsiloxane (PMPhS) have been studied by TOF-SIMS to investigate effects of functional group changes on polymer fragmentation mechanisms. Cyclic fragments are observed in the low mass range spectra of PDMS and PHMS, but not in the spectrum of PMPhS. Effects of functional group substitution on the fragmentation mechanisms of polysiloxanes are evident in the high mass range spectra (>1000 Da). Peaks of oligomers cationized by silver dominate the high mass range of the spectra of all low molecular weight polysiloxanes. However, fragmentation patterns of these samples are different. Neutral cyclic fragments cationized by silver are identified in the high mass range of the spectra of PDMS and PHMS, but not in the spectrum of PMPhS. The major fragments of PHMS and PMPhS are [oligomer-14+Ag](+). The PHMS spectrum also shows peaks [oligomer-28+Ag](+). These distinctive fragmentation patterns can be used to identify the polysiloxanes.

Use of the polymerase chain reaction in differentiating rinderpest field virus and vaccine virus in the same animals

In 1991, a disease with clinical signs indicative of rinderpest was reported in yaks in the former Soviet Union, near the border with Mongolia. At the peak of the epizootic, mortality among affected yaks was 32-42% in adults and 65% in animals less than one year old. Pathological samples were examined independently at two institutes in Russia. Both institutes confirmed the presence of rinderpest using complement fixation, agar gel diffusion and immunoassays. Since vaccination had been initiated to control an outbreak of a similar disease several months earlier, the later cases were possibly due to the vaccine and field rinderpest may not have been present. However, the disease had occurred in non-vaccinated animals and these were then vaccinated against the disease. Tissue samples obtained from these animals, which were examined at the Pirbright Laboratory using gel diffusion assays and specific nucleic acid probes, were found to be positive for rinderpest antigen and nucleic acid. Ribonucleic acid derived from the post-mortem tissue samples was amplified using the polymerase chain reaction and rinderpest-specific primers. Sequence analysis of the amplified deoxyribonculeic acid from the samples revealed the presence of two distinct virus strains, one identical to the Plowright rinderpest tissue culture vaccine and the other related to field strains of rinderpest virus circulating in Asia and the Middle East.