Identification of risk factors for mortality associated with COVID-19

Predicting COVID-19 severity: Challenges in reproducibility and deployment of machine learning methods

The increasing use of electronic health records (EHR) based computable phenotypes in clinical research is providing new opportunities for development of data-driven medical applications. Adopted widely in the United States and globally, EHRs facilitate systematic collection of patients' longitudinal information, which serves as one of the important foundations for artificial intelligence applications in medicine. Harmonization of input variables and outcome definitions is critically important for wider clinical applicability of artificial intelligence (AI) methodologies. In this review, we focused on Coronavirus Disease 2019 (COVID-19) severity machine learning prediction models and explored the pipeline for standardizing future disease severity model development using EHR information. We identified 2,967 studies published between 01/01/2020 and 02/15/2022 and selected 135 independent studies that had built machine learning prediction models to predict severity related outcomes of COVID-19 patients based on EHR data for the final review. These 135 studies spanning across 27 counties covered a broad range of severity related prediction outcomes. We observed substantial inconsistency in COVID-19 severity phenotype definitions among models in these studies. Moreover, there was a gap between the outcome of these models and clinician-recognized clinical concepts. Accordingly, we recommend that robust clinical input metrics, with outcome definitions which eliminate ambiguity in interpretation, to reduce algorithmic bias, mitigate model brittleness and improve generalizability of a universal model for COVID-19 severity. This framework can potentially be extended to broader clinical application.

A new parallel multi-objective Harris hawk algorithm for predicting the mortality of COVID-19 patients

Harris' Hawk Optimization (HHO) is a novel metaheuristic inspired by the collective hunting behaviors of hawks. This technique employs the flight patterns of hawks to produce (near)-optimal solutions, enhanced with feature selection, for challenging classification problems. In this study, we propose a new parallel multi-objective HHO algorithm for predicting the mortality risk of COVID-19 patients based on their symptoms. There are two objectives in this optimization problem: to reduce the number of features while increasing the accuracy of the predictions. We conduct comprehensive experiments on a recent real-world COVID-19 dataset from Kaggle. An augmented version of the COVID-19 dataset is also generated and experimentally shown to improve the quality of the solutions. Significant improvements are observed compared to existing state-of-the-art metaheuristic wrapper algorithms. We report better classification results with feature selection than when using the entire set of features. During experiments, a 98.15% prediction accuracy with a 45% reduction is achieved in the number of features. We successfully obtained new best solutions for this COVID-19 dataset.

SARS-CoV-2 and COVID-19: A Narrative Review

The World Health Organisation has reported that the viral disease known as COVID-19, caused by SARS-CoV-2, is the leading cause of death by a single infectious agent. This narrative review examines certain components of the pandemic: its origins, early clinical data, global and UK-focussed epidemiology, vaccination, variants, and long COVID.

A simple nomogram for predicting infectious diseases in adult kidney transplantation recipients

Objective

To investigate the risk factors of infectious diseases in adult kidney transplantation recipients and to establish a simple and novel nomogram to guide the prophylactic antimicrobial therapy.

Methods

Patients who received kidney transplantation between January 2018 and October 2021 were included in the study and were divided into a training and a testing set at a 1:1 ratio. Risk factors correlated to infectious diseases were selected using a Least Absolute Shrinkage and Selection Operator (LASSO) regression model. The prediction model was built by incorporating the variables selected by the LASSO model into a logistic regression equation. Calibration curves and receiver operating characteristic (ROC) curves were also applied to assess the model calibration and discrimination. A nomogram consisting of the selected factors was established to provide individualized risks of developing infections. Decision curve analysis (DCA) was adopted to estimate the net benefit and reduction in interventions for a range of clinically reasonable risk thresholds.

Results

In all, 863 adult kidney recipients were included in the study, and 407 (47.16%) of them developed infectious diseases during the 3-year follow-up period. A total of 8 variables were selected using LASSO regression and were retained for subsequent model construction and infection prediction. The area under the curve (AUC) was 0.83 and 0.81 in the training and testing sets, with high F scores of 0.76 and 0.77, sensitivity of 0.76 and 0.81, and specificity of 0.88 and 0.74, respectively. A novel nomogram was developed based on 8 selected predictors (requirement for albumin infusion, requirement for red blood cell infusion, triglyceride, uric acid, creatinine, globulin, neutrophil percentage, and white blood cells). The net benefit indicated that the nomogram would reduce unnecessary interventions at a wide range of threshold probabilities in both sets.

Conclusions

Adult kidney transplantation recipients are high-risk hosts for infectious diseases. The novel nomogram consisting of 8 factors reveals good predictive performance and may promote the reasonable antimicrobial prescription. More external validations are required to confirm its effectiveness for further clinical application.

Retrospective Study of Aging and Sex-Specific Risk Factors of COVID-19 with Hypertension in China

Background

Coronavirus disease 2019 (COVID-19) has been a global threat that pushes healthcare to its limits. Hypertension is one of the most common risk factors for cardiovascular complications in COVID-19 and is strongly associated with disease severity and mortality. To date, clinical mechanisms by which hypertension leads to increased risk in COVID-19 are still unclear. Furthermore, additional factors might increase these risks, such as the consideration of age and sex, which are of interest when in search of personalized treatments for hypertensive COVID-19 patients.

Methods

We conducted a retrospective cohort study of 543 COVID-19 patients in seven provinces of China to examine the epidemiological and clinical characteristics of COVID-19 in this population and to determine risk factors of hypertensive COVID-19 patients. We also used univariable and multivariable logistic regression methods to explore the risk factors associated with hypertensive COVID-19 patients in different age and sex subgroups.

Results

Among the enrolled COVID-19 patients, the median age was 47 years (interquartile range (IQR) 34.0–57.0), and 99 patients (18.23%) were over 60 years old. With regard to comorbidities, 91 patients (16.75%) were diagnosed with hypertension, followed by diabetes, coronary disease, and cerebrovascular disease. Of the hypertensive COVID-19 patients, 51 (56.04%) were male. Multivariable analysis showed that old age, comorbid diabetes or coronary heart disease on admission, increased D-dimer, increased glucose, and decreased lymphocyte count were independent risk factors associated with hypertensive COVID-19 patients. Elevated total bilirubin (odds ratio [OR]: 1.014, 95% confidence interval [CI]: 0.23–1.05; p = 0.043) and triglycerides (OR: 1.173, 95% CI: 0.049–1.617; p = 0.007) were found to be associated with elderly hypertensive COVID-19 patients. In addition, we found that decreased lymphocytes, basophil, high-density lipoprotein, and increased fibrinogen and creatinine were related to a higher risk of disease severity in male patients. The most common abnormal clinical findings pertaining to female hypertensive COVID-19 patients were hemoglobin, total bile acid, total protein, and low-density lipoprotein.

Conclusions

Factors associated with increased risk of hypertensive COVID-19 patients were identified. Results to the different age and sex subgroups in our study will allow for better possible personalized care and also provide new insights into specific risk stratification, disease management, and treatment strategies for COVID-19 patients with hypertension in the future.

Associations of pre-existing cardiovascular morbidity with severity and the fatality rate in COVID-19 patients: a systematic review and meta-analysis

Objectives

The aim of this study was to evaluate the association of pre-existing cardiovascular comorbidities, including hypertension and coronary heart disease, with coronavirus disease 2019 (COVID-19) severity and mortality.

Methods

PubMed, ScienceDirect, and Scopus were searched between January 1, 2020, and July 18, 2020, to identify eligible studies. Random-effect models were used to estimate the pooled event rates of pre-existing cardiovascular disease comorbidities and odds ratio (OR) with 95% confidence intervals (95% CIs) of disease severity and mortality associated with the exposures of interest.

Results

A total of 34 studies involving 19,156 patients with COVID-19 infection met the inclusion criteria. The prevalence of pre-existing cardiovascular disease in the included studies was 14.0%. Pre-existing cardiovascular disease in COVID-19 patients was associated with severe outcomes (OR, 4.1; 95% CI, 2.9 to 5.7) and mortality (OR, 6.1; 95% CI, 2.9 to 12.7). Hypertension and coronary heart disease increased the risk of severe outcomes by 3 times (OR, 3.2; 95% CI, 2.0 to 3.6) and 2.5 times (OR, 2.5; 95% CI, 1.7 to 3.8), respectively. No significant publication bias was indicated.

Conclusion

COVID-19 patients with pre-existing cardiovascular comorbidities have a higher risk of severe outcomes and mortality. Awareness of pre-existing cardiovascular comorbidity is important for the early management of COVID-19.

Serum hydroxybutyrate dehydrogenase and COVID-19 severity and mortality: a systematic review and meta-analysis with meta-regression

Alterations in cardiac and renal biomarkers have been reported in coronavirus disease 19 (COVID-19). We conducted a systematic review and meta-analysis to investigate serum concentrations of hydroxybutyrate dehydrogenase (HBDH), a combined marker of myocardial and renal injury, in hospitalized COVID-19 patients with different disease severity and survival status. We searched PubMed, Web of Science and Scopus, between December 2019 and April 2021, for studies reporting HBDH in COVID-19. Risk of bias was assessed using the Newcastle–Ottawa scale, publication bias was assessed with the Begg’s and Egger’s tests, and certainty of evidence was assessed using GRADE. In 22 studies in 15,019 COVID-19 patients, serum HBDH concentrations on admission were significantly higher in patients with high disease severity or non-survivor status when compared to patients with low severity or survivor status (standardized mean difference, SMD = 0.90, 95% CI 0.74 to 1.07, p < 0.001; moderate certainty of evidence). Extreme between-study heterogeneity was observed (I² = 93.5%, p < 0.001). Sensitivity analysis, performed by sequentially removing each study and re-assessing the pooled estimates, showed that the magnitude and the direction of the effect size were not substantially modified. A significant publication bias was observed. In meta-regression, the SMD of HBDH concentrations was significantly associated with markers of inflammation, sepsis, liver damage, non-specific tissue damage, myocardial injury, and renal function. Higher HBDH concentrations were significantly associated with higher COVID-19 severity and mortality. This biomarker of cardiac and renal injury might be useful for risk stratification in COVID-19. (PROSPERO registration number: CRD42021258123).

Risk profiles for negative and positive COVID-19 hospitalized patients

COVID-19 is a viral infection that affects people differently, where the majority of cases develop mild symptoms, some people require hospitalization, and unfortunately, a small number of patients perish. Hence, identifying risk factors is critical for physicians to make treatment decisions. The purpose of this article is to determine whether unsupervised analysis of risk factors in positive and negative COVID-19 subjects can aid in the identification of a set of reliable and clinically relevant risk profiles. Positive and negative patients hospitalized were randomly selected from the Mexican Open Registry between March and May 2020. Thirteen risk factors, three distinct outcomes, and COVID-19 test results were used to categorize registry patients. As a result, the dataset was reported via 6144 different risk profiles for each age group. The unsupervised learning method is proposed in this study to discover the most prevalent risk profiles. The data was partitioned into discovery (70%) and validation (30%) sets. The discovery set was analyzed using the partition around medoids (PAM) method, and the stable set of risk profiles was estimated using robust consensus clustering. The PAM models' reliability was validated by predicting the risk profile of subjects from the validation set and patients admitted in November 2020. In the validation set, the clinical relevance of the risk profiles was evaluated by determining the prevalence of three patient outcomes: pneumonia diagnosis, ICU admission, or death. Six positive and five negative COVID-19 risk profiles were identified, with significant statistical differences between them. As a result, PAM clustering with consensus mapping is a viable method for discovering unsupervised risk profiles in subjects with severe respiratory health problems.

Estimated pulse wave velocity (ePWV) as a potential gatekeeper for MRI-assessed PWV: a linear and deep neural network based approach in 2254 participants of the Netherlands Epidemiology of Obesity study

Pulse wave velocity (PWV) assessed by magnetic resonance imaging (MRI) is a prognostic marker for cardiovascular events. Prediction modelling could enable indirect PWV assessment based on clinical and anthropometric data. The aim was to calculate estimated-PWV (ePWV) based on clinical and anthropometric measures using linear ridge regression as well as a Deep Neural Network (DNN) and to determine the cut-off which provides optimal discriminative performance between lower and higher PWV values. In total 2254 participants from the Netherlands Epidemiology of Obesity study were included (age 45–65 years, 51% male). Both a basic and expanded prediction model were developed. PWV was estimated using linear ridge regression and DNN. External validation was performed in 114 participants (age 30–70 years, 54% female). Performance was compared between models and estimation accuracy was evaluated by ROC-curves. A cut-off for optimal discriminative performance was determined using Youden’s index. The basic ridge regression model provided an adjusted R² of 0.33 and bias of < 0.001, the expanded model did not add predictive performance. Basic and expanded DNN models showed similar model performance. Optimal discriminative performance was found for PWV < 6.7 m/s. In external validation expanded ridge regression provided the best performance of the four models (adjusted R²: 0.29). All models showed good discriminative performance for PWV < 6.7 m/s (AUC range 0.81–0.89). ePWV showed good discriminative performance with regard to differentiating individuals with lower PWV values (< 6.7 m/s) from those with higher values, and could function as gatekeeper in selecting patients who benefit from further MRI-based PWV assessment.

How often and to what extent do admitted COVID-19 patients have signs of cardiac injury?

Background

COVID-19 can cause myocardial injury in a significant proportion of patients admitted to the hospital and seems to be associated with worse prognosis. The aim of this review was to study how often and to what extent COVID-19 causes myocardial injury and whether this is an important contributor to outcome with implications for management.

Methods

A literature search was performed in Medline and Embase. Myocardial injury was defined as elevated cardiac troponin (cTn) levels with at least one value > 99th percentile of the upper reference limit. The primary outcome measure was mortality, whereas secondary outcome measures were intensive care unit (ICU) admission and length of hospital stay.

Results

Four studies and one review were included. The presence of myocardial injury varied between 9.6 and 46.3%. Myocardial injury was associated with a higher mortality rate (risk ratio (RR) 5.54, 95% confidence interval (CI) 3.48–8.80) and more ICU admissions (RR 3.78, 95% CI 2.07–6.89). The results regarding length of hospital stay were inconclusive.

Conclusion

Patients with myocardial injury might be classified as high-risk patients, with probably a higher mortality rate and a larger need for ICU admission. cTn levels can be used in risk stratification models and can indicate which patients potentially benefit from early medication administration. We recommend measuring cTn levels in all COVID-19 patients admitted to the hospital or who deteriorate during admission.

Supplementary Information

The online version of this article (10.1007/s12471-021-01571-w) contains supplementary material, which is available to authorized users.

Impact of liver enzymes on SARS-CoV-2 infection and the severity of clinical course of COVID-19

BACKGROUND AND AIM

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the virus responsible for the current pandemic, can have multi-organ impact. Recent studies show that liver injury could be a manifestation of the disease, and that liver disease could also be related to a worse prognosis. Our aim was to compare the characteristics of patients with severe coronavirus disease 2019 (COVID-19) due to SARS-CoV-2 who required intubation versus stable hospitalized patients to identify the early biochemical predictive factors of a severe course of COVID-19 and subsequent requirement for intubation, specifically in Mexican.

METHODS

This was an observational case-control study nested in a cohort study. Complete medical records of patients admitted for confirmed COVID-19 at a tertiary level center in Mexico City were reviewed. Clinical and biochemical data were collected, and the characteristics of patients who required invasive mechanical ventilation (IMV) (cases) were compared with stable hospitalized patients without ventilation (controls).

RESULTS

We evaluated 166 patients with COVID-19 due to SARS-CoV-2 infection; 114 (68.7%) were men, the mean age was 50.6 ± 13.3 years, and 27 (16.3%) required IMV. The comparative analysis between cases and controls showed (respectively) significantly lower blood oxygen saturation (SpO₂) (73.5 ± 12.0% vs. 83.0 ± 6.8%, P < 0.0001) and elevated alanine aminotransferase (ALT) (128 (14-1123) IU/L vs. 33 (8-453) IU/L, P = 0.003), aspartate aminotransferase (AST) (214 (17-1247) vs. 44 (12-498) IU/L, P = 0.001), lactic dehydrogenase (LDH) (764.6 ± 401.9 IU/L vs. 461.0 ± 185.6 IU/L, P = 0.001), and D-dimer (3463 (524-34,227) ng/mL vs. 829 (152-41,923) ng/mL, P = 0.003) concentrations. Patients in the cases group were older (58.6 ± 12.7 years vs. 49.1 ± 12.8 years, P=0.001). Multivariate analysis showed that important factors at admission predicting the requirement for IMV during hospitalization for COVID-19 were AST ≥250 IU/L (odds ratio (OR) = 64.8, 95% confidence interval (CI) 7.5-560.3, P < 0.0001) and D-dimer ≥ 3500 ng/mL (OR = 4.1, 95% CI 1.2-13.7, P=0.02).

CONCLUSIONS

Our study confirms the importance of monitoring liver enzymes in hospitalized patients with COVID-19; seriously ill patients have significantly elevated AST and D-dimer concentrations, which have prognostic implications in the SARS-CoV-2 disease course.

Prediction models for diagnosis and prognosis of covid-19: systematic review and critical appraisal

OBJECTIVE

To review and appraise the validity and usefulness of published and preprint reports of prediction models for diagnosing coronavirus disease 2019 (covid-19) in patients with suspected infection, for prognosis of patients with covid-19, and for detecting people in the general population at increased risk of covid-19 infection or being admitted to hospital with the disease.

DESIGN

Living systematic review and critical appraisal by the COVID-PRECISE (Precise Risk Estimation to optimise covid-19 Care for Infected or Suspected patients in diverse sEttings) group.

DATA SOURCES

PubMed and Embase through Ovid, up to 1 July 2020, supplemented with arXiv, medRxiv, and bioRxiv up to 5 May 2020.

STUDY SELECTION

Studies that developed or validated a multivariable covid-19 related prediction model.

DATA EXTRACTION

At least two authors independently extracted data using the CHARMS (critical appraisal and data extraction for systematic reviews of prediction modelling studies) checklist; risk of bias was assessed using PROBAST (prediction model risk of bias assessment tool).

RESULTS

37 421 titles were screened, and 169 studies describing 232 prediction models were included. The review identified seven models for identifying people at risk in the general population; 118 diagnostic models for detecting covid-19 (75 were based on medical imaging, 10 to diagnose disease severity); and 107 prognostic models for predicting mortality risk, progression to severe disease, intensive care unit admission, ventilation, intubation, or length of hospital stay. The most frequent types of predictors included in the covid-19 prediction models are vital signs, age, comorbidities, and image features. Flu-like symptoms are frequently predictive in diagnostic models, while sex, C reactive protein, and lymphocyte counts are frequent prognostic factors. Reported C index estimates from the strongest form of validation available per model ranged from 0.71 to 0.99 in prediction models for the general population, from 0.65 to more than 0.99 in diagnostic models, and from 0.54 to 0.99 in prognostic models. All models were rated at high or unclear risk of bias, mostly because of non-representative selection of control patients, exclusion of patients who had not experienced the event of interest by the end of the study, high risk of model overfitting, and unclear reporting. Many models did not include a description of the target population (n=27, 12%) or care setting (n=75, 32%), and only 11 (5%) were externally validated by a calibration plot. The Jehi diagnostic model and the 4C mortality score were identified as promising models.

CONCLUSION

Prediction models for covid-19 are quickly entering the academic literature to support medical decision making at a time when they are urgently needed. This review indicates that almost all pubished prediction models are poorly reported, and at high risk of bias such that their reported predictive performance is probably optimistic. However, we have identified two (one diagnostic and one prognostic) promising models that should soon be validated in multiple cohorts, preferably through collaborative efforts and data sharing to also allow an investigation of the stability and heterogeneity in their performance across populations and settings. Details on all reviewed models are publicly available at https://www.covprecise.org/. Methodological guidance as provided in this paper should be followed because unreliable predictions could cause more harm than benefit in guiding clinical decisions. Finally, prediction model authors should adhere to the TRIPOD (transparent reporting of a multivariable prediction model for individual prognosis or diagnosis) reporting guideline.

SYSTEMATIC REVIEW REGISTRATION

Protocol https://osf.io/ehc47/, registration https://osf.io/wy245.

READERS' NOTE

This article is a living systematic review that will be updated to reflect emerging evidence. Updates may occur for up to two years from the date of original publication. This version is update 3 of the original article published on 7 April 2020 (BMJ 2020;369:m1328). Previous updates can be found as data supplements (https://www.bmj.com/content/369/bmj.m1328/related#datasupp). When citing this paper please consider adding the update number and date of access for clarity.