A risk scoring system to predict progression to severe pneumonia in patients with Covid-19

Platelet and large platelet ratios are useful in predicting severity of COVID-19

The role of platelets in coronavirus disease (COVID-19) severity requires further exploration. To determine whether the platelet index is useful in predicting COVID-19 severity, we compared the platelet index in patients with higher and lower oxygen requirements (≥ 4 L/min vs. < 4 L/min) and patients without COVID-19. We also analyzed the time course of the platelet index in each group. A total of 285 patients with COVID-19 and 36 without COVID-19 who were hospitalized at Fussa Hospital were analyzed. After matching for oxygen requirement at admission, multivariate analysis was performed. Platelets (≤ 16.6 × 104/μL) and platelet-large cell ratio (P-LCR) (≥ 27.8%) were significant factors influencing severity. Based on these factors, we created the Fussa platelet score, and the group with a Fussa platelet score ≥ 2 was significantly more likely to reach the 4 L/min oxygen requirement (event-free survival: Fussa platelet score ≥ 2 versus < 2, P < 0.00000001). Analysis of platelet index by time period showed a significant increase from 6-10 days after onset. The Fussa platelet score can be measured quickly, easily, and inexpensively in a clinic and may be useful in determining need for transfer to a critical care hospital.

Temporal evolution of chest radiographic appearances in COVID-19 with clinicoradiological associations: a multicentre United Kingdom resident-led study

AIM

To describe the (a) frequency of improving, static, and worsening chest radiograph (CXR) appearances; (b) differences in demographic, initial rudimentary haematological and CXR variables and these patterns; and (c) frequency of different trajectories of serial CXR evolution, in COVID-19 patients presenting consecutively.

MATERIALS AND METHODS

This multicentre retrospective study included all COVID-19 patients admitted from 1-30 April 2020, meeting the inclusion criteria across 24 (blinded) hospitals. Follow-up CXRs on admission, the subsequent (where available), and at 4-8 weeks were scored for the presence of parenchymal opacities across six zones. Three cohorts were defined: improved, static, and/or worsened. The chi-squared and Kruskal-Wallis tests were used to compare demographic, laboratory, and CXR variables. Trajectories of CXR evolution were assessed when all three CXRs were available (226 patients).

RESULTS

Of 452 included patients (median age 66 years, interquartile range 54.3-79, 262 men), 211 (46.7%) improved, 140 (31%) were static, and 101 (22.3%) worsened. Improving patients were more likely younger, with a classic COVID-19 radiograph and higher initial CXR zonal severity scores (both p<0.001), while worsening patients had lower initial lymphocyte counts (p=0.008). The most frequent trajectory was worsened then improved (n=63, 27.9%) followed by static then improved (n=46, 20.4%) and static (n=42, 18.6%).

CONCLUSION

Most patients with COVID-19 during the first wave of the pandemic demonstrated radiographic improvement; these patients were more likely younger with classic COVID-19 appearances and initially more extensive abnormality. Conversely, radiographic deterioration was associated with lower lymphocyte counts. The three most common trajectories were worsening then improvement, static then improvement, and static throughout.

New score to predict COVID-19 progression in vaccine and early treatment era: the COVID-19 Sardinian Progression Score (CSPS)

BACKGROUND

Several scores aimed at predicting COVID-19 progression have been proposed. As the variables vaccination and early SARS-CoV-2 treatment were systematically excluded from the prognostic scores, the present study's objective was to develop a new model adapted to the current epidemiological scenario.

METHODS

We included all patients evaluated by the Infectious Disease Unit in Sassari, with SARS-CoV-2 infection and without signs of respiratory failure at the first evaluation (P/F > 300). Disease progression was defined by the prescription of supplemental oxygen. In addition, variables related to demographics, vaccines, comorbidities, symptoms, CT scans, blood tests, and therapies were collected. Multivariate logistic regression modelling was performed to determine factors associated with progression; any variable with significant univariate test or clinical relevance was selected as a candidate for multivariate analysis. Hosmer-Lemeshow (HL) goodness of fit statistic was calculated. Odds ratio values were used to derive an integer score for developing an easy-to-use progression risk score. The discrimination performance of the risk index was determined using the AUC, and the best cut-off point, according to the Youden index, sensitivity, specificity, predictive value, and likelihood ratio, was chosen.

RESULTS

1145 patients [median (IQR) age 74 (62-83) years; 53.5% males] were enrolled; 336 (29.3%) had disease progression. Patients with a clinical progression were older and showed more comorbidities; furthermore, they were less vaccinated and exposed to preventive therapy. In the multivariate logistic regression analysis, age ≥ 60 years, COPD, dementia, haematological tumours, heart failure, exposure to no or one vaccine dose, fever, dyspnoea, GGO, consolidation, ferritin, De Ritis ≥ 1.2, LDH, and no exposure to early anti-SARS-CoV-2 treatment were associated with disease progression. The final risk score ranged from 0 to 45. The ROC curve analysis showed an AUC of 0.92 (95% CI 0.90-0.93) with a 93.7% specificity and 72.9% sensitivity. Low risk was defined when the cut-off value was less than 23. Three risk levels were identified: low (0-23 points), medium (24-35), and high (≥ 36).

CONCLUSIONS

The proportion of patients with progression increases with high scores: the assessment of the risk could be helpful for clinicians to plan appropriate therapeutic strategies.

Prognostic models in COVID-19 infection that predict severity: a systematic review

Current evidence on COVID-19 prognostic models is inconsistent and clinical applicability remains controversial. We performed a systematic review to summarize and critically appraise the available studies that have developed, assessed and/or validated prognostic models of COVID-19 predicting health outcomes. We searched six bibliographic databases to identify published articles that investigated univariable and multivariable prognostic models predicting adverse outcomes in adult COVID-19 patients, including intensive care unit (ICU) admission, intubation, high-flow nasal therapy (HFNT), extracorporeal membrane oxygenation (ECMO) and mortality. We identified and assessed 314 eligible articles from more than 40 countries, with 152 of these studies presenting mortality, 66 progression to severe or critical illness, 35 mortality and ICU admission combined, 17 ICU admission only, while the remaining 44 studies reported prediction models for mechanical ventilation (MV) or a combination of multiple outcomes. The sample size of included studies varied from 11 to 7,704,171 participants, with a mean age ranging from 18 to 93 years. There were 353 prognostic models investigated, with area under the curve (AUC) ranging from 0.44 to 0.99. A great proportion of studies (61.5%, 193 out of 314) performed internal or external validation or replication. In 312 (99.4%) studies, prognostic models were reported to be at high risk of bias due to uncertainties and challenges surrounding methodological rigor, sampling, handling of missing data, failure to deal with overfitting and heterogeneous definitions of COVID-19 and severity outcomes. While several clinical prognostic models for COVID-19 have been described in the literature, they are limited in generalizability and/or applicability due to deficiencies in addressing fundamental statistical and methodological concerns. Future large, multi-centric and well-designed prognostic prospective studies are needed to clarify remaining uncertainties.

Rapid Assessment of COVID-19 Mortality Risk with GASS Classifiers

Risk prediction models are fundamental to effectively triage incoming COVID-19 patients. However, current triaging methods often have poor predictive performance, are based on variables that are expensive to measure, and often lead to hard-to-interpret decisions. We introduce two new classification methods that can predict COVID-19 mortality risk from the automatic analysis of routine clinical variables with high accuracy and interpretability. SVM22-GASS and Clinical-GASS classifiers leverage machine learning methods and clinical expertise, respectively. Both were developed using a derivation cohort of 499 patients from the first wave of the pandemic and were validated with an independent validation cohort of 250 patients from the second pandemic phase. The Clinical-GASS classifier is a threshold-based classifier that leverages the General Assessment of SARS-CoV-2 Severity (GASS) score, a COVID-19-specific clinical score that recently showed its effectiveness in predicting the COVID-19 mortality risk. The SVM22-GASS model is a binary classifier that non-linearly processes clinical data using a Support Vector Machine (SVM). In this study, we show that SMV22-GASS was able to predict the mortality risk of the validation cohort with an AUC of 0.87 and an accuracy of 0.88, better than most scores previously developed. Similarly, the Clinical-GASS classifier predicted the mortality risk of the validation cohort with an AUC of 0.77 and an accuracy of 0.78, on par with other established and emerging machine-learning-based methods. Our results demonstrate the feasibility of accurate COVID-19 mortality risk prediction using only routine clinical variables, readily collected in the early stages of hospital admission.

Rapid Assessment of COVID-19 Mortality Risk with GASS Classifiers

Risk prediction models are fundamental to effectively triage incoming COVID-19 patients. However, current triaging methods often have poor predictive performance, are based on variables that are expensive to measure, and often lead to hard-to-interpret decisions. We introduce two new classification methods that can predict COVID-19 mortality risk from the automatic analysis of routine clinical variables with high accuracy and interpretability. SVM22-GASS and Clinical-GASS classifiers leverage machine learning methods and clinical expertise, respectively. Both were developed using a derivation cohort of 499 patients from the first wave of the pandemic and were validated with an independent validation cohort of 250 patients from the second pandemic phase. The Clinical-GASS classifier is a threshold-based classifier that leverages the General Assessment of SARS-CoV-2 Severity (GASS) score, a COVID-19-specific clinical score that recently showed its effectiveness in predicting the COVID-19 mortality risk. The SVM22-GASS model is a binary classifier that non-linearly processes clinical data using a Support Vector Machine (SVM). In this study, we show that SMV22-GASS was able to predict the mortality risk of the validation cohort with an AUC of 0.87 and an accuracy of 0.88, better than most scores previously developed. Similarly, the Clinical-GASS classifier predicted the mortality risk of the validation cohort with an AUC of 0.77 and an accuracy of 0.78, on par with other established and emerging machine-learning-based methods. Our results demonstrate the feasibility of accurate COVID-19 mortality risk prediction using only routine clinical variables, readily collected in the early stages of hospital admission.

Potential and limitations of machine meta-learning (ensemble) methods for predicting COVID-19 mortality in a large inhospital Brazilian dataset

The majority of early prediction scores and methods to predict COVID-19 mortality are bound by methodological flaws and technological limitations (e.g., the use of a single prediction model). Our aim is to provide a thorough comparative study that tackles those methodological issues, considering multiple techniques to build mortality prediction models, including modern machine learning (neural) algorithms and traditional statistical techniques, as well as meta-learning (ensemble) approaches. This study used a dataset from a multicenter cohort of 10,897 adult Brazilian COVID-19 patients, admitted from March/2020 to November/2021, including patients [median age 60 (interquartile range 48-71), 46% women]. We also proposed new original population-based meta-features that have not been devised in the literature. Stacking has shown to achieve the best results reported in the literature for the death prediction task, improving over previous state-of-the-art by more than 46% in Recall for predicting death, with AUROC 0.826 and MacroF1 of 65.4%. The newly proposed meta-features were highly discriminative of death, but fell short in producing large improvements in final prediction performance, demonstrating that we are possibly on the limits of the prediction capabilities that can be achieved with the current set of ML techniques and (meta-)features. Finally, we investigated how the trained models perform on different hospitals, showing that there are indeed large differences in classifier performance between different hospitals, further making the case that errors are produced by factors that cannot be modeled with the current predictors.

Smart healthcare: A prospective future medical approach for COVID-19

COVID-19 has greatly affected human life for over 3 years. In this review, we focus on smart healthcare solutions that address major requirements for coping with the COVID-19 pandemic, including (1) the continuous monitoring of severe acute respiratory syndrome coronavirus 2, (2) patient stratification with distinct short-term outcomes (eg, mild or severe diseases) and long-term outcomes (eg, long COVID), and (3) adherence to medication and treatments for patients with COVID-19. Smart healthcare often utilizes medical artificial intelligence (AI) and cloud computing and integrates cutting-edge biological and optoelectronic techniques. These are valuable technologies for addressing the unmet needs in the management of COVID. By leveraging deep learning/machine learning capabilities and big data, medical AI can perform precise prognosis predictions and provide reliable suggestions for physicians' decision-making. Through the assistance of the Internet of Medical Things, which encompasses wearable devices, smartphone apps, internet-based drug delivery systems, and telemedicine technologies, the status of mild cases can be continuously monitored and medications provided at home without the need for hospital care. In cases that develop into severe cases, emergency feedback can be provided through the hospital for rapid treatment. Smart healthcare can possibly prevent the development of severe COVID-19 cases and therefore lower the burden on intensive care units.

Inflammatory Biomarkers Associated with In-Hospital Mortality in Critical COVID-19 Patients

The COVID-19 pandemic poses global healthcare challenges due to its unpredictable clinical course. The aim of this study is to identify inflammatory biomarkers and other routine laboratory parameters associated with in-hospital mortality in critical COVID-19 patients. We performed a retrospective observational study on 117 critical COVID-19 patients. Following descriptive statistical analysis of the survivor and non-survivor groups, optimal cut-off levels for the statistically significant parameters were determined using the ROC method, and the corresponding Kaplan-Meier survival curves were calculated. The inflammatory parameters that present statistically significant differences between survivors and non-survivors are IL-6 (p = 0.0004, cut-off = 27.68 pg/mL), CRP (p = 0.027, cut-off = 68.15 mg/L) and IL-6/Ly ratio (p = 0.0003, cut-off = 50.39). Additionally, other statistically significant markers are creatinine (p = 0.031, cut-off = 0.83 mg/dL), urea (p = 0.0002, cut-off = 55.85 mg/dL), AST (p = 0.0209, cut-off = 44.15 U/L), INR (p = 0.0055, cut-off = 1.075), WBC (p = 0.0223, cut-off = 11.68 × 10⁹/L) and pH (p = 0.0055, cut-off = 7.455). A survival analysis demonstrated significantly higher in-hospital mortality rates of patients with values of IL-6, IL-6/Ly, AST, INR, and pH exceeding previously mentioned thresholds. In our study, IL-6 and IL-6/Ly have a predictive value for the mortality of critically-ill patients diagnosed with COVID-19. The integration of these parameters with AST, INR and pH could contribute to a prognostic score for the risk stratification of critical patients, reducing healthcare costs and facilitating clinical decision-making.