Characteristics and Outcomes of a Sample of Patients With COVID-19 Identified Through Social Media in Wuhan, China: Observational Study

Neurological complications and effects of COVID-19: Symptoms and conceivable mechanisms

A novel coronavirus, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was first identified in December 2019 in Wuhan, China. The new coronavirus disease (COVID-19) was declared a global pandemic by the World Health Organization (WHO) in March 2020. SARS-CoV-2 can invade the nervous system aside from infecting the respiratory system as its primary target. The most common nervous system symptoms of COVID-19 are stated as headache, myalgia, fatigue, nausea, vomiting, sudden and unexplained anosmia, and ageusia. More severe conditions such as encephalomyelitis, acute myelitis, thromboembolic events, ischemic stroke, intracerebral hemorrhage, Guillain-Barré-syndrome, Bell's palsy, rhabdomyolysis, and even coma have also been reported. Cohort studies revealed that neurological findings are associated with higher morbidity and mortality. The neurological symptoms and manifestations caused by SARS-CoV-2 and COVID-19 are examined and summarized in this article.

A new data integration framework for Covid-19 social media information

The Covid-19 pandemic presents a serious threat to people's health, resulting in over 250 million confirmed cases and over 5 million deaths globally. To reduce the burden on national health care systems and to mitigate the effects of the outbreak, accurate modelling and forecasting methods for short- and long-term health demand are needed to inform government interventions aiming at curbing the pandemic. Current research on Covid-19 is typically based on a single source of information, specifically on structured historical pandemic data. Other studies are exclusively focused on unstructured online retrieved insights, such as data available from social media. However, the combined use of structured and unstructured information is still uncharted. This paper aims at filling this gap, by leveraging historical and social media information with a novel data integration methodology. The proposed approach is based on vine copulas, which allow us to exploit the dependencies between different sources of information. We apply the methodology to combine structured datasets retrieved from official sources and a big unstructured dataset of information collected from social media. The results show that the combined use of official and online generated information contributes to yield a more accurate assessment of the evolution of the Covid-19 pandemic, compared to the sole use of official data.

Lung transplantation in a woman with paraquat poisoning that led to pulmonary fibrosis-Widely reported by the media: A case report

RATIONALE

Paraquat is an extremely toxic herbicide with a high mortality rate on poisoning. It can damage vital organs, such as the lungs, liver, heart, and kidneys. In this study, we report a case of pulmonary fibrosis after paraquat poisoning in a patient who underwent a lung transplant procedure after preoperative administration of corticosteroids and immunosuppressive agents and continuous noninvasive ventilation support therapy.

PATIENT CONCERNS

An 18-year-old student was hospitalized owing to diarrhea, chest pain, and gradually evolving dyspnea.

DIAGNOSES

Owing to the inability to estimate the intake concentration and dose, paraquat was only detected in the urine on the 13th day, resulting in rapid progression of the disease and severe pulmonary fibrosis.

INTERVENTIONS

Extensive media coverage has attracted the attention of all sectors of society. The patient received financial assistance; thus, she could receive a double-lung transplant with extracorporeal membrane oxygenation (ECMO) support on the 34th day after the poisoning.

OUTCOMES

Postoperatively, the girl was actively rehabilitated, adhered to anti-rejection medication, followed up regularly, and had a good prognosis.

LESSONS

Lung transplantation is currently the most effective treatment for pulmonary fibrosis, and mass media campaigns can provide economic support, influence potential organ donation, and provide such patients more chances to survive.

Heterogeneity and Risk of Bias in Studies Examining Risk Factors for Severe Illness and Death in COVID-19: A Systematic Review and Meta-Analysis

This systematic review and meta-analysis synthesized the evidence on the impacts of demographics and comorbidities on the clinical outcomes of COVID-19, as well as the sources of the heterogeneity and publication bias of the relevant studies. Two authors independently searched the literature from PubMed, Embase, Cochrane library, and CINAHL on 18 May 2021; removed duplicates; screened the titles, abstracts, and full texts by using criteria; and extracted data from the eligible articles. The variations among the studies were examined by using Cochrane, Q.; I2, and meta-regression. Out of 11,975 articles that were obtained from the databases and screened, 559 studies were abstracted, and then, where appropriate, were analyzed by meta-analysis (n = 542). COVID-19-related severe illness, admission to the ICU, and death were significantly correlated with comorbidities, male sex, and an age older than 60 or 65 years, although high heterogeneity was present in the pooled estimates. The study design, the study country, the sample size, and the year of publication contributed to this. There was publication bias among the studies that compared the odds of COVID-19-related deaths, severe illness, and admission to the ICU on the basis of the comorbidity status. While an older age and chronic diseases were shown to increase the risk of developing severe illness, admission to the ICU, and death among the COVID-19 patients in our analysis, a marked heterogeneity was present when linking the specific risks with the outcomes.

COVID-19: A pluralistic and integrated approach for efficient management of the pandemic

The novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which triggered the ongoing pandemic, was first discovered in China in late 2019. SARS-CoV-2 is a respiratory virus responsible for coronavirus disease 2019 (COVID-19) that often manifests as a pneumonic syndrome. In the context of the pandemic, there are mixed views on the data provided by epidemiologists and the information collected by hospital clinicians about their patients. In addition, the literature reports a large proportion of patients free of pneumonia vs a small percentage of patients with severe pneumonia among confirmed COVID-19 cases. This raises the issue of the complexity of the work required to control or contain the pandemic. We believe that an integrative and pluralistic approach will help to put the analyses into perspective and reinforce collaboration and creativity in the fight against this major scourge. This paper proposes a comprehensive and integrative approach to COVID-19 research, prevention, control, and treatment to better address the pandemic. Thus, this literature review applies a pluralistic approach to fight the pandemic.

Psychological Symptoms in COVID-19 Patients: Insights into Pathophysiology and Risk Factors of Long COVID-19

There is growing evidence of studies associating COVID-19 survivors with increased mental health consequences. Mental health implications related to a COVID-19 infection include both acute and long-term consequences. Here we discuss COVID-19-associated psychiatric sequelae, particularly anxiety, depression, and post-traumatic stress disorder (PTSD), drawing parallels to past coronavirus outbreaks. A literature search was completed across three databases, using keywords to search for relevant articles. The cause may directly correlate to the infection through both direct and indirect mechanisms, but the underlying etiology appears more complex and multifactorial, involving environmental, psychological, and biological factors. Although most risk factors and prevalence rates vary across various studies, being of the female gender and having a history of psychiatric disorders seem consistent. Several studies will be presented, demonstrating COVID-19 survivors presenting higher rates of mental health consequences than the general population. The possible mechanisms by which the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) enters the brain, affecting the central nervous system (CNS) and causing these psychiatric sequelae, will be discussed, particularly concerning the SARS-CoV-2 entry via the angiotensin-converting enzyme 2 (ACE-2) receptors and the implications of the immune inflammatory signaling on neuropsychiatric disorders. Some possible therapeutic options will also be considered.

Clinical and inflammatory features based machine learning model for fatal risk prediction of hospitalized COVID-19 patients: results from a retrospective cohort study

OBJECTIVES

To appraise effective predictors for COVID-19 mortality in a retrospective cohort study.

METHODS

A total of 1270 COVID-19 patients, including 984 admitted in Sino French New City Branch (training and internal validation sets randomly split at 7:3 ratio) and 286 admitted in Optical Valley Branch (external validation set) of Wuhan Tongji hospital, were included in this study. Forty-eight clinical and laboratory features were screened with LASSO method. Further multi-tree extreme gradient boosting (XGBoost) machine learning-based model was used to rank importance of features selected from LASSO and subsequently constructed death risk prediction model with simple-tree XGBoost model. Performances of models were evaluated by AUC, prediction accuracy, precision, and F1 scores.

RESULTS

Six features, including disease severity, age, levels of high-sensitivity C-reactive protein (hs-CRP), lactate dehydrogenase (LDH), ferritin, and interleukin-10 (IL-10), were selected as predictors for COVID-19 mortality. Simple-tree XGBoost model conducted by these features can predict death risk accurately with >90% precision and >85% sensitivity, as well as F1 scores >0.90 in training and validation sets.

CONCLUSION

We proposed the disease severity, age, serum levels of hs-CRP, LDH, ferritin, and IL-10 as significant predictors for death risk of COVID-19, which may help to identify the high-risk COVID-19 cases. KEY MESSAGES A machine learning method is used to build death risk model for COVID-19 patients. Disease severity, age, hs-CRP, LDH, ferritin, and IL-10 are death risk factors. These findings may help to identify the high-risk COVID-19 cases.

Diabetes, hypertension, body mass index, smoking and COVID-19-related mortality: a systematic review and meta-analysis of observational studies

OBJECTIVES

We conducted a systematic literature review and meta-analysis of observational studies to investigate the association between diabetes, hypertension, body mass index (BMI) or smoking with the risk of death in patients with COVID-19 and to estimate the proportion of deaths attributable to these conditions.

METHODS

Relevant observational studies were identified by searches in the PubMed, Cochrane library and Embase databases through 14 November 2020. Random-effects models were used to estimate summary relative risks (SRRs) and 95% CIs. Certainty of evidence was assessed using the Cochrane methods and the Grading of Recommendations, Assessment, Development and Evaluations framework.

RESULTS

A total of 186 studies representing 210 447 deaths among 1 304 587 patients with COVID-19 were included in this analysis. The SRR for death in patients with COVID-19 was 1.54 (95% CI 1.44 to 1.64, I2=92%, n=145, low certainty) for diabetes and 1.42 (95% CI 1.30 to 1.54, I2=90%, n=127, low certainty) for hypertension compared with patients without each of these comorbidities. Regarding obesity, the SSR was 1.45 (95% CI 1.31 to 1.61, I2=91%, n=54, high certainty) for patients with BMI ≥30 kg/m2 compared with those with BMI <30 kg/m2 and 1.12 (95% CI 1.07 to 1.17, I2=68%, n=25) per 5 kg/m2 increase in BMI. There was evidence of a J-shaped non-linear dose-response relationship between BMI and mortality from COVID-19, with the nadir of the curve at a BMI of around 22-24, and a 1.5-2-fold increase in COVID-19 mortality with extreme obesity (BMI of 40-45). The SRR was 1.28 (95% CI 1.17 to 1.40, I2=74%, n=28, low certainty) for ever, 1.29 (95% CI 1.03 to 1.62, I2=84%, n=19) for current and 1.25 (95% CI 1.11 to 1.42, I2=75%, n=14) for former smokers compared with never smokers. The absolute risk of COVID-19 death was increased by 14%, 11%, 12% and 7% for diabetes, hypertension, obesity and smoking, respectively. The proportion of deaths attributable to diabetes, hypertension, obesity and smoking was 8%, 7%, 11% and 2%, respectively.

CONCLUSION

Our findings suggest that diabetes, hypertension, obesity and smoking were associated with higher COVID-19 mortality, contributing to nearly 30% of COVID-19 deaths.

TRIAL REGISTRATION NUMBER

CRD42020218115.

COVID-19 Outcomes and Genomic Characterization of SARS-CoV-2 Isolated From Veterans in New England States: Retrospective Analysis

BACKGROUND

Clinical and virologic characteristics of COVID-19 infections in veterans in New England have not been described. The average US veteran is a male older than the general US population. SARS-CoV-2 infection is known to cause poorer outcomes among men and older adults, making the veteran population an especially vulnerable group for COVID-19.

OBJECTIVE

This study aims to evaluate clinical and virologic factors impacting COVID-19 outcomes.

METHODS

This retrospective chart review included 476 veterans in six New England states with confirmed SARS-CoV-2 infection between April and September 2020. Whole genome sequencing was performed on SARS-CoV-2 RNA isolated from these veterans, and the correlation of genomic data to clinical outcomes was evaluated. Clinical and demographic variables were collected by manual chart review and were correlated to the end points of peak disease severity (based on oxygenation requirements), hospitalization, and mortality using multivariate regression analyses.

RESULTS

Of 476 veterans, 274 had complete and accessible charts. Of the 274 veterans, 92.7% (n=254) were men and 83.2% (n=228) were White, and the mean age was 63 years. In the multivariate regression, significant predictors of hospitalization (C statistic 0.75) were age (odds ratio [OR] 1.05, 95% CI 1.03-1.08) and non-White race (OR 2.39, 95% CI 1.13-5.01). Peak severity (C statistic 0.70) also varied by age (OR 1.07, 95% CI 1.03-1.11) and O2 requirement on admission (OR 45.7, 95% CI 18.79-111). Mortality (C statistic 0.87) was predicted by age (OR 1.06, 95% CI 1.01-1.11), dementia (OR 3.44, 95% CI 1.07-11.1), and O2 requirement on admission (OR 6.74, 95% CI 1.74-26.1). Most (291/299, 97.3%) of our samples were dominated by the spike protein D614G substitution and were from SARS-CoV-2 B.1 lineage or one of 37 different B.1 sublineages, with none representing more than 8.7% (26/299) of the cases.

CONCLUSIONS

In a cohort of veterans from the six New England states with a mean age of 63 years and a high comorbidity burden, age was the largest predictor of hospitalization, peak disease severity, and mortality. Non-White veterans were more likely to be hospitalized, and patients who required oxygen on admission were more likely to have severe disease and higher rates of mortality. Multiple SARS-CoV-2 lineages were distributed in patients in New England early in the COVID-19 era, mostly related to viruses from New York State with D614G mutation.

Chronic Diseases as a Predictor for Severity and Mortality of COVID-19: A Systematic Review With Cumulative Meta-Analysis

Introduction: Given the ongoing coronavirus disease 2019 (COVID-19) pandemic and the consequent global healthcare crisis, there is an urgent need to better understand risk factors for symptom deterioration and mortality among patients with COVID-19. This systematic review aimed to meet the need by determining the predictive value of chronic diseases for COVID-19 severity and mortality.

Methods: We searched PubMed, Embase, Web of Science, and Cumulative Index to Nursing and Allied Health Complete to identify studies published between December 1, 2019, and December 31, 2020. Two hundred and seventeen observational studies from 26 countries involving 624,986 patients were included. We assessed the risk of bias of the included studies and performed a cumulative meta-analysis.

Results: We found that among COVID-19 patients, hypertension was a very common condition and was associated with higher severity, intensive care unit (ICU) admission, acute respiratory distress syndrome, and mortality. Chronic obstructive pulmonary disease was the strongest predictor for COVID-19 severity, admission to ICU, and mortality, while asthma was associated with a reduced risk of COVID-19 mortality. Patients with obesity were at a higher risk of experiencing severe symptoms of COVID-19 rather than mortality. Patients with cerebrovascular disease, chronic liver disease, chronic renal disease, or cancer were more likely to become severe COVID-19 cases and had a greater probability of mortality.

Conclusions: COVID-19 patients with chronic diseases were more likely to experience severe symptoms and ICU admission and faced a higher risk of mortality. Aggressive strategies to combat the COVID-19 pandemic should target patients with chronic diseases as a priority.

Neurological Manifestations in COVID-19 Patients: A Meta-Analysis

Common symptoms such as dizziness, headache, olfactory dysfunction, nausea, vomiting, etc. in COVID-19 patients have indicated the involvement of the nervous system. However, the exact association of the nervous system with COVID-19 infection is still unclear. Thus, we have conducted a meta-analysis of clinical studies associated with neurological problems in COVID-19 patients. We have searched for electronic databases with MeSH terms, and the studies for analysis were selected based on inclusion and exclusion criteria and quality assessment. The Stats Direct (version 3) was used for the analysis. The pooled prevalence with 95% confidence interval of various neurological manifestations reported in the COVID-19 patients was found to be headache 14.6% (12.2-17.2), fatigue 33.6% (29.5-37.8), olfactory dysfunction 26.4% (21.8-31.3), gustatory dysfunction 27.2% (22.3-32.3), vomiting 6.7% (5.5-8.0), nausea 9.8% (8.1-11.7), dizziness 6.7% (4.7-9.1), myalgia 21.4% (18.8-24.1), seizure 4.05% (2.5-5.8), cerebrovascular diseases 9.9% (6.8-13.4), sleep disorders 14.9% (1.9-36.8), altered mental status 17.1% (12.3-22.5), neuralgia 2.4% (0.8-4.7), arthralgia 19.9% (15.3-25.0), encephalopathy 23.5% (14.3-34.1), encephalitis 0.6% (0.2-1.3), malaise 38.3% (24.7-52.9), confusion 14.2% (6.9-23.5), movement disorders 5.2% (1.7-10.4), and Guillain-Barre syndrome 6.9% (2.3-13.7). However, the heterogeneity among studies was found to be high. Various neurological manifestations related to the central nervous system (CNS) and peripheral nervous system (PNS) are associated with COVID-19 patients.

Mental health outcomes of coronavirus infection survivors: A rapid meta-analysis

BACKGROUND

The current COVID pandemic is happening while the long-term effects of coronavirus infection remain poorly understood. The present article meta-analyzed mental health outcomes (anxiety, depression, etc.) from a previous coronavirus outbreak in China (2002).

METHOD

CNKI, Wanfang, PubMed/Medline, Scopus, Web of Science, Baidu Scholar, and Google Scholar were searched up to early June 2020 for articles in English or Chinese reporting mental illness symptoms of SARS patients. Main outcome measures include SCL-90, SAS, SDS, and IES-R scales. 29 papers met the inclusion criteria. The longest follow-up time included in the analysis was 46 months.

FINDINGS

The systematic meta-analysis indicated that mental health problems were most serious before or at hospital discharge and declined significantly during the first 12 months after hospital discharge. Nevertheless, average symptom levels remained above healthy norms even at 12 months and continued to improve, albeit slowly, thereafter.

INTERPRETATION

The adverse mental health impact of being hospitalized with coronavirus infection long outlasts the physical illness. Mental health issues were the most serious for coronavirus infected patients before (including) hospital discharge and improved continuously during the first 12 months after hospital discharge. If COVID-19 infected patients follow a similar course of mental health development, most patients should recover to normal after 12 months of hospital discharge.

From nicotine to the cholinergic anti-inflammatory reflex - Can nicotine alleviate the dysregulated inflammation in COVID-19?

The coronavirus SARS-CoV-2 of 2019 (COVID-19) causes a pandemic that has been diagnosed in more than 70 million people worldwide. Mild-to-moderate COVID-19 symptoms include coughing, fever, myalgia, shortness of breath, and acute inflammatory lung injury (ALI). In contrast, acute respiratory distress syndrome (ARDS) and respiratory failure occur in patients diagnosed with severe COVID-19. ARDS is mediated, at least in part, by a dysregulated inflammatory response due to excessive levels of circulating cytokines, a condition known as the "cytokine-storm syndrome." Currently, there are FDA-approved therapies that attenuate the dysregulated inflammation that occurs in COVID-19 patients, such as dexamethasone or other corticosteroids and IL-6 inhibitors, including sarilumab, tocilizumab, and siltuximab. However, the efficacy of these treatments have been shown to be inconsistent. Compounds that activate the vagus nerve-mediated cholinergic anti-inflammatory reflex, such as the α7 nicotinic acetylcholine receptor agonist, GTS-21, attenuate ARDS/inflammatory lung injury by decreasing the extracellular levels of high mobility group box-1 (HMGB1) in the airways and the circulation. It is possible that HMGB1 may be an important mediator of the "cytokine-storm syndrome." Notably, high plasma levels of HMGB1 have been reported in patients diagnosed with severe COVID-19, and there is a significant negative correlation between HMGB1 plasma levels and clinical outcomes. Nicotine can activate the cholinergic anti-inflammatory reflex, which attenuates the up-regulation and the excessive release of pro-inflammatory cytokines/chemokines. Therefore, we hypothesize that low molecular weight compounds that activate the cholinergic anti-inflammatory reflex, such as nicotine or GTS-21, may represent a potential therapeutic approach to attenuate the dysregulated inflammatory responses in patients with severe COVID-19.

What social media told us in the time of COVID-19: a scoping review

With the onset of the COVID-19 pandemic, social media has rapidly become a crucial communication tool for information generation, dissemination, and consumption. In this scoping review, we selected and examined peer-reviewed empirical studies relating to COVID-19 and social media during the first outbreak from November, 2019, to November, 2020. From an analysis of 81 studies, we identified five overarching public health themes concerning the role of online social media platforms and COVID-19. These themes focused on: surveying public attitudes, identifying infodemics, assessing mental health, detecting or predicting COVID-19 cases, analysing government responses to the pandemic, and evaluating quality of health information in prevention education videos. Furthermore, our Review emphasises the paucity of studies on the application of machine learning on data from COVID-19-related social media and a scarcity of studies documenting real-time surveillance that was developed with data from social media on COVID-19. For COVID-19, social media can have a crucial role in disseminating health information and tackling infodemics and misinformation.

Association of smoking history with severe and critical outcomes in COVID-19 patients: A systemic review and meta-analysis

Introduction

The highly infectious coronavirus disease 2019 (COVID-19) has now rapidly spread around the world. This meta-analysis was strictly focused on the influence of smoking history on the severe and critical outcomes on people with COVID-19 pneumonia.

Methods

A systematic literature search was conducted in eight online databases before 1 February 2021. All studies meeting our selection criteria were included and evaluated. Stata 14.0 software was used to analyze the data.

Results

A total of 109 articles involving 517,020 patients were included in this meta-analysis. A statistically significant association was discovered between smoking history and COVID-19 severity, the pooled OR was 1.55 (95%CI: 1.41-1.71). Smoking was significantly associated with the risk of admission to intensive care unit (ICU) (OR=1.73, 95%CI: 1.36-2.19), increased mortality (OR=1.58, 95%CI: 1.38-1.81), and critical diseases composite endpoints (OR=1.61, 95%CI: 1.35-1.93), whereas there was no relationship with mechanical ventilation. The pooled prevalence of smoking using the random effects model (REM) was 15% (95%CI: 14%-16%). Meta-regression analysis showed that age (P=0.004), hypertension (P=0.007), diabetes (P=0.029), chronic obstructive pulmonary disease (COPD) (P=0.001) were covariates that affect the association.

Conclusions

Smoking was associated with severe or critical outcomes and increased the risk of admission to ICU and mortality in COVID-19 patients, but not associated with mechanical ventilation. This association was more significant for former smokers than in current smokers. Current smokers also had a higher risk of developing severe COVID-19 compared with non-smokers. More detailed data, which are representative of more countries, are needed to confirm these preliminary findings.

Evaluation and characterization of HSPA5 (GRP78) expression profiles in normal individuals and cancer patients with COVID-19

HSPA5 (BiP, GRP78) has been reported as a potential host-cell receptor for SARS-Cov-2, but its expression profiles on different tissues including tumors, its susceptibility to SARS-Cov-2 virus and severity of its adverse effects on malignant patients are unclear. In the current study, HSPA5 has been found to be expressed ubiquitously in normal tissues and significantly increased in 14 of 31 types of cancer tissues. In lung cancer, mRNA levels of HSPA5 were 253-fold increase than that of ACE2. Meanwhile, in both malignant tumors and matched normal samples across almost all cancer types, mRNA levels of HSPA5 were much higher than those of ACE2. Higher expression of HSPA5 significantly decreased patient overall survival (OS) in 7 types of cancers. Moreover, systematic analyses found that 7.15% of 5,068 COVID-19 cases have malignant cancer coincidental situations, and the rate of severe events of COVID-19 patients with cancers present a higher trend than that for all COVID-19 patients, showing a significant difference (33.33% vs 16.09%, p<0.01). Collectively, these data imply that the tissues with high HSPA5 expression, not low ACE2 expression, are susceptible to be invaded by SARS-CoV-2. Taken together, this study not only indicates the clinical significance of HSPA5 in COVID-19 disease and cancers, but also provides potential clues for further medical treatments and managements of COVID-19 patients.

Smoking and risk of negative outcomes among COVID-19 patients: A systematic review and meta-analysis

INTRODUCTION

COVID-19 has major effects on the clinical, humanistic and economic outcomes among patients, producing severe symptoms and death. Smoking has been reported as one of the factors that increases severity and mortality rate among COVID-19 patients. However, the effect of smoking on such medical outcomes is still controversial. This study conducted a comprehensive systematic review and meta-analysis (SR/MA) on the association between smoking and negative outcomes among COVID-19 patients.

METHODS

Electronic databases, including PubMed, EMBASE, Cochrane Library, Science Direct, Google Scholar, were systematically searched from the initiation of the database until 12 December 2020. All relevant studies about smoking and COVID-19 were screened using a set of inclusion and exclusion criteria. The Newcastle-Ottawa Scale was used to assess the methodological quality of eligible articles. Random meta-analyses were conducted to estimate odds ratios (ORs) with 95% confidence interval (CIs). Publication bias was assessed using the funnel plot, Begg's test and Egger's test.

RESULTS

A total of 1248 studies were retrieved and reviewed. A total of 40 studies were finally included for meta-analysis. Both current smoking and former smoking significantly increase the risk of disease severity (OR=1.58; 95% CI: 1.16-2.15, p=0.004; and OR=2.48; 95% CI: 1.64-3.77, p<0.001; respectively) with moderate appearance of heterogeneity. Similarly, current smoking and former smoking also significantly increase the risk of death (OR=1.35; 95% CI: 1.12-1.62, p=0.002; and OR=2.58; 95% CI: 2.15-3.09, p<0.001; respectively) with moderate appearance of heterogeneity. There was no evidence of publication bias, which was tested by the funnel plot, Begg's test and Egger's test.

CONCLUSIONS

Smoking, even current smoking or former smoking, significantly increases the risk of COVID-19 severity and death. Further causational studies on this association and ascertianing the underlying mechanisms of this relation is warranted.

Prognostic Assessment of COVID-19 in the Intensive Care Unit by Machine Learning Methods: Model Development and Validation

BACKGROUND

Patients with COVID-19 in the intensive care unit (ICU) have a high mortality rate, and methods to assess patients' prognosis early and administer precise treatment are of great significance.

OBJECTIVE

The aim of this study was to use machine learning to construct a model for the analysis of risk factors and prediction of mortality among ICU patients with COVID-19.

METHODS

In this study, 123 patients with COVID-19 in the ICU of Vulcan Hill Hospital were retrospectively selected from the database, and the data were randomly divided into a training data set (n=98) and test data set (n=25) with a 4:1 ratio. Significance tests, correlation analysis, and factor analysis were used to screen 100 potential risk factors individually. Conventional logistic regression methods and four machine learning algorithms were used to construct the risk prediction model for the prognosis of patients with COVID-19 in the ICU. The performance of these machine learning models was measured by the area under the receiver operating characteristic curve (AUC). Interpretation and evaluation of the risk prediction model were performed using calibration curves, SHapley Additive exPlanations (SHAP), Local Interpretable Model-Agnostic Explanations (LIME), etc, to ensure its stability and reliability. The outcome was based on the ICU deaths recorded from the database.

RESULTS

Layer-by-layer screening of 100 potential risk factors finally revealed 8 important risk factors that were included in the risk prediction model: lymphocyte percentage, prothrombin time, lactate dehydrogenase, total bilirubin, eosinophil percentage, creatinine, neutrophil percentage, and albumin level. Finally, an eXtreme Gradient Boosting (XGBoost) model established with the 8 important risk factors showed the best recognition ability in the training set of 5-fold cross validation (AUC=0.86) and the verification queue (AUC=0.92). The calibration curve showed that the risk predicted by the model was in good agreement with the actual risk. In addition, using the SHAP and LIME algorithms, feature interpretation and sample prediction interpretation algorithms of the XGBoost black box model were implemented. Additionally, the model was translated into a web-based risk calculator that is freely available for public usage.

CONCLUSIONS

The 8-factor XGBoost model predicts risk of death in ICU patients with COVID-19 well; it initially demonstrates stability and can be used effectively to predict COVID-19 prognosis in ICU patients.