Development and validation of the ISARIC 4C Deterioration model for adults hospitalised with COVID-19: a prospective cohort study

Performance of the 4C and SEIMC scoring systems in predicting mortality from onset to current COVID-19 pandemic in emergency departments

The evolution of the COVID-19 pandemic has been associated with variations in clinical presentation and severity. Similarly, prediction scores may suffer changes in their diagnostic accuracy. The aim of this study was to test the 30-day mortality predictive validity of the 4C and SEIMC scores during the sixth wave of the pandemic and to compare them with those of validation studies. This was a longitudinal retrospective observational study. COVID-19 patients who were admitted to the Emergency Department of a Spanish hospital from December 15, 2021, to January 31, 2022, were selected. A side-by-side comparison with the pivotal validation studies was subsequently performed. The main measures were 30-day mortality and the 4C and SEIMC scores. A total of 27,614 patients were considered in the study, including 22,361 from the 4C, 4,627 from the SEIMC and 626 from our hospital. The 30-day mortality rate was significantly lower than that reported in the validation studies. The AUCs were 0.931 (95% CI: 0.90-0.95) for 4C and 0.903 (95% CI: 086-0.93) for SEIMC, which were significantly greater than those obtained in the first wave. Despite the changes that have occurred during the coronavirus disease 2019 (COVID-19) pandemic, with a reduction in lethality, scorecard systems are currently still useful tools for detecting patients with poor disease risk, with better prognostic capacity.

The SpO2/FiO2 Ratio Combined with Prognostic Scores for Pneumonia and COVID-19 Increases Their Accuracy in Predicting Mortality of COVID-19 Patients

Background: Identifying high-risk COVID-19 patients is critical for emergency department decision-making. Our study's primary objective was to identify new independent predictors of mortality and their predictive utility in combination with traditional pneumonia risk assessment scores and new risk scores for COVID-19 developed during the pandemic. Methods: A retrospective study was performed in two Italian University Hospitals. A multivariable logistic model was used to locate independent parameters associated with mortality. Results: Age, PaO2/FiO2, and SpO2/FiO2 ratios were found to be independent parameters associated with mortality. This study found that the Pneumonia Severity Index (PSI) was superior to many of the risk scores developed during the pandemic, for example, the International Severe Acute Respiratory Infection Consortium Coronavirus Clinical Characterisation Consortium (ISARIC 4C) (AUC 0.845 vs. 0.687, p < 0.001), and to many of the risk scores already in use, for example, the National Early Warning Score 2 (NEWS2) (AUC 0.845 vs. 0.589, p < 0.001). Furthermore, our study found that the Pneumonia Severity Index had a similar performance to other risk scores, such as CRB-65 (AUC 0.845 vs. 0.823, p = 0.294). Combining the PaO2/FiO2 or SpO2/FiO2 ratios with the risk scores analyzed improved the prognostic accuracy. Conclusions: Adding the SpO2/FiO2 ratio to the traditional, validated, and already internationally known pre-pandemic prognostic scores seems to be a valid and rapid alternative to the need for developing new prognostic scores. Future research should focus on integrating these markers into existing pneumonia scores to improve their prognostic accuracy.

Evaluation of the predictors and frequency of silent hypoxemia in COVID-19 patients and the gap between pulse oximeter and arterial blood gas levels: A cross-sectional study

Background

Silent hypoxemia is when patients do not experience breathing difficulty in the presence of alarmingly low O2 saturation. It could cause rapid deterioration and higher mortality rates among patients, so prompt detection and identifying predictive factors could result in significantly better outcomes. This study aims to document the evidence of silent hypoxemia in patients with COVID-19 and its clinical features.

Methods

A total of 78 hospitalized, nonintubated patients with confirmed COVID-19 infection were included in this study. Their O2 saturation was measured with a pulse oximeter (PO), and arterial blood gas (ABG) was taken. Demographic and clinical features were recorded. The Borg scale was used to evaluate dyspnea status, and patients with a score of less than two accompanied by O2 saturation of less than 94% were labeled as silent hypoxic. Univariate analysis was utilized to evaluate the correlation between variables and their odds ratio (OR) and 95% confidence interval (CI).

Results

Silent hypoxemia was observed in 20 (25.6%) of the participants. The average difference between the PO and ABG methods was 4.36 ± 3.43. Based on regression analysis, dyspnea and respiratory rate demonstrated a statistically significant correlation with the O2 saturation difference between PO and ABG (OR: 2.05; p = 0.026; 95% CI: 0.248-3.847 and OR: 0.144; p = 0.048, 95% CI: 0.001-0.286). Furthermore, the Borg scale (OR: 0.29; p = 0.009; 95% CI: 0.116-0.740) had a significant reverse correlation with silent hypoxia.

Conclusions

Silent hypoxemia can be a possible complication that affects some COVID-19 patients. Further care should be bestowed upon the younger population and those with underlying neurological or mental illnesses. Furthermore, the respiratory rate, pulse oximeter, and arterial blood gas O2 levels should be considered alongside each other.

Changes in hospital mortality in patients with cancer during the COVID-19 pandemic (ISARIC-CCP-UK): a prospective, multicentre cohort study

BACKGROUND

Patients with cancer are at greater risk of dying from COVID-19 than many other patient groups. However, how this risk evolved during the pandemic remains unclear. We aimed to determine, on the basis of the UK national pandemic protocol, how factors influencing hospital mortality from COVID-19 could differentially affect patients undergoing cancer treatment. We also examined changes in hospital mortality and escalation of care in patients on cancer treatment during the first 2 years of the COVID-19 pandemic in the UK.

METHODS

We conducted a prospective cohort study of patients aged older than 19 years and admitted to 306 health-care facilities in the UK with confirmed SARS-CoV-2 infection, who were enrolled in the International Severe Acute Respiratory and emerging Infections Consortium (ISARIC) WHO Clinical Characterisation Protocol (CCP) across the UK from April 23, 2020, to Feb 28, 2022; this analysis included all patients in the complete dataset when the study closed. The primary outcome was 30-day in-hospital mortality, comparing patients on cancer treatment and those without cancer. The study was approved by the South Central-Oxford C Research Ethics Committee in England (Ref: 13/SC/0149) and the Scotland A Research Ethics Committee (Ref 20/SS/0028), and is registered on the ISRCTN Registry (ISRCTN66726260).

FINDINGS

177 871 eligible adult patients either with no history of cancer (n=171 303) or on cancer treatment (n=6568) were enrolled; 93 205 (52·4%) were male, 84 418 (47·5%) were female, and in 248 (13·9%) sex or gender details were not specified or data were missing. Patients were followed up for a median of 13 (IQR 6-21) days. Of the 6568 patients receiving cancer treatment, 2080 (31·7%) died at 30 days, compared with 30 901 (18·0%) of 171 303 patients without cancer. Patients aged younger than 50 years on cancer treatment had the highest age-adjusted relative risk (hazard ratio [HR] 5·2 [95% CI 4·0-6·6], p<0·0001; vs 50-69 years 2·4 [2·2-2·6], p<0·0001; 70-79 years 1·8 [1·6-2·0], p<0·0001; and >80 years 1·5 [1·3-1·6], p<0·0001) but a lower absolute risk (51 [6·7%] of 763 patients <50 years died compared with 459 [30·2%] of 1522 patients aged >80 years). In-hospital mortality decreased for all patients during the pandemic but was higher for patients on cancer treatment than for those without cancer throughout the study period.

INTERPRETATION

People with cancer have a higher risk of mortality from COVID-19 than those without cancer. Patients younger than 50 years with cancer treatment have the highest relative risk of death. Continued action is needed to mitigate the poor outcomes in patients with cancer, such as through optimising vaccination, long-acting passive immunisation, and early access to therapeutics. These findings underscore the importance of the ISARIC-WHO pandemic preparedness initiative.

FUNDING

National Institute for Health Research and the Medical Research Council.

Development and Validation of Prediction Models for the 5-year Risk of Type 2 Diabetes in a Japanese Population: Japan Public Health Center-based Prospective (JPHC) Diabetes Study

BACKGROUND

This study aimed to develop models to predict the 5-year incidence of type 2 diabetes mellitus (T2DM) in a Japanese population and validate them externally in an independent Japanese population.

METHODS

Data from 10,986 participants (aged 46-75 years) in the development cohort of the Japan Public Health Center-based Prospective Diabetes Study and 11,345 participants (aged 46-75 years) in the validation cohort of the Japan Epidemiology Collaboration on Occupational Health Study were used to develop and validate the risk scores in logistic regression models.

RESULTS

We considered non-invasive (sex, body mass index, family history of diabetes mellitus, and diastolic blood pressure) and invasive (glycated hemoglobin [HbA1c] and fasting plasma glucose [FPG]) predictors to predict the 5-year probability of incident diabetes. The area under the receiver operating characteristic curve was 0.643 for the non-invasive risk model, 0.786 for the invasive risk model with HbA1c but not FPG, and 0.845 for the invasive risk model with HbA1c and FPG. The optimism for the performance of all models was small by internal validation. In the internal-external cross-validation, these models tended to show similar discriminative ability across different areas. The discriminative ability of each model was confirmed using external validation datasets. The invasive risk model with only HbA1c was well-calibrated in the validation cohort.

CONCLUSION

Our invasive risk models are expected to discriminate between high- and low-risk individuals with T2DM in a Japanese population.

Extracellular vesicles are a late marker of inflammation, hypercoagulability and COVID-19 severity

Exacerbated inflammation and coagulation are a hallmark of COVID-19 severity. Extracellular vesicles (EVs) are intercellular transmitters involved in inflammatory conditions, which are capable of triggering prothrombotic mechanisms. Since the release of EVs is potentially associated with COVID-19-induced coagulopathy, the aim of this study was to evaluate changes in inflammation- and hypercoagulability-related EVs during the first month after symptom onset and to determine whether they are associated with disease severity. Blood samples of patients with mild or severe forms of the disease were collected on three occasions: in the second, third and fourth weeks after symptom onset for the quantification by flow cytometry of CD41A (platelet glycoprotein IIb/IIIa), CD162 (PSGL-1), CD31 (PECAM-1) and CD142 cells (tissue factor). Analysis of variance (ANOVA) with repeated measures, Kruskal-Wallis and correlation tests were used. Eighty-five patients were enrolled, 71% of whom had mild disease. Seventeen uninfected individuals served as controls. Compared to controls, both mild and severe COVID-19 were associated with higher EV-CD31+, EV-CD41+ and EV-CD142+ levels. All EV levels were higher in severe than in mild COVID-19 only after the third week from symptom onset, as opposed to C-reactive protein and D-dimer levels, which were higher in severe than in mild COVID-19 earlier during disease progression. EV levels were also associated with C-reactive protein and D-dimer levels only after the third week of symptoms. In conclusion, EVs expressing CD41A, CD31, TF, and CD162 appear as late markers of COVID-19 severity. This finding may contribute to the understanding of the pathogenesis of acute and possibly long COVID-19.

Comparison of eight prehospital early warning scores in life-threatening acute respiratory distress: a prospective, observational, multicentre, ambulance-based, external validation study

BACKGROUND

A myriad of early warning scores (EWSs) exist, yet there is a need to identify the most clinically valid score to be used in prehospital respiratory assessments to estimate short-term and midterm mortality, intensive-care unit admission, and airway management in life-threatening acute respiratory distress.

METHODS

This is a prospective, observational, multicentre, ambulance-based, external validation study performed in 44 ambulance services and four hospitals across three Spanish provinces (ie, Salamanca, Segovia, and Valladolid). We identified adults (ie, those aged 18 years and older) discharged to the emergency department with suspected acute respiratory distress. The primary outcome was 2-day all-cause in-hospital mortality, for all the patients or according to prehospital respiratory conditions, including dyspnoea, chronic obstructive pulmonary disease (COPD), COVID-19, other infections, and other conditions (asthma exacerbation, haemoptysis, and bronchoaspirations). 30-day mortality, intensive-care unit admission, and invasive and non-invasive mechanical ventilation were secondary outcomes. Eight EWSs, namely, the National Early Warning Score 2, the Modified Rapid Emergency Medicine Score, the Rapid Acute Physiology Score, the Quick Sequential Organ Failure Assessment Score, the CURB-65 Severity Score for Community-Acquired Pneumonia, the BAP-65 Score for Acute Exacerbation of COPD, the Quick COVID-19 Severity Index, and the Modified Sequential Organ Failure Assessment (mSOFA), were explored to determine their predictive validity through calibration, clinical net benefit as determined through decision curve analysis, and discrimination analysis (area under the curve of the receiver operating characteristic [AUROC], compared with Delong's test).

FINDINGS

Between Jan 1, 2020, and Nov 31, 2022, 902 patients were enrolled. The global 2-day mortality rate was 87 (10%); in proportion to various respiratory conditions, the rates were 35 (40%) for dyspnoea, nine (10%) for COPD, 13 (15%) for COVID-19, 28 (32%) for other infections, and two (2%) for others conditions. mSOFA showed the best calibration, a higher net benefit, and the best discrimination (AUROC 0·911, 95% CI 0·86-0·95) for predicting 2-day mortality, and its discrimination was statistically significantly more accurate (p<0·0001) compared with the other scores. The performance of mSOFA for predicting 2-day mortality was higher than the other scores when considering the prehospital respiratory conditions, and was also higher for the secondary outcomes, except for non-invasive mechanical ventilation.

INTERPRETATION

Our results showed that mSOFA outperformed other EWSs. The inclusion of mSOFA in prehospital decision making will entail a quick identification of patients in acute respiratory distress at high risk of deterioration, allowing prioritisation of resources and patient care.

FUNDING

Gerencia Regional de Salud, Public Health System of Castilla y León (GRS Spain).

TRANSLATION

For the Spanish translation of the abstract see Supplementary Materials section.

The association between vitamin D deficiency and the clinical outcomes of hospitalized COVID-19 patients

Background

Vitamin D deficiency is an emerging public health problem that affects more than one billion people worldwide. Vitamin D has been shown to be effective in preventing and reducing the severity of viral respiratory diseases, including influenza. However, the role of vitamin D in COVID-19 infection remains controversial. This study aimed to analyze the association of vitamin D deficiency on the clinical outcome of hospitalized COVID-19 patients.

Methods

A prospective cohort study was conducted among hospitalized COVID-19 patients at two COVID-19 referral hospitals in Indonesia from October 2021 until February 2022.

Results

The median serum 25(OH)D level in 191 hospitalized COVID-19 patients was 13.6 [IQR=10.98] ng/mL. The serum 25(OH)D levels were significantly lower among COVID-19 patients with vitamin D deficiency who had cardiovascular disease (p-value=0.04), the use of a ventilator (p-value=0.004), more severe COVID-19 cases (p-value=0.047), and mortality (p-value=0.002). Furthermore, serum 25(OH)D levels were significantly different between patients with mild and severe COVID-19 cases (p-value=0.019). Serum 25(OH)D levels in moderate and severe COVID-19 cases were significantly different (p-value=0.031). Lower serum 25(OH)D levels were significantly associated with an increased number of comorbidities (p-value=0.03), the severity of COVID-19 (p-value=0.002), and the use of mechanical ventilation (p-value=0.032). Mortality was found in 7.3% of patients with deficient vitamin D levels. However, patients with either sufficient or insufficient vitamin D levels did not develop mortality.

Conclusions

COVID-19 patients with vitamin D deficiency were significantly associated with having cardiovascular disease, mortality, more severe COVID-19 cases, and the used of mechanical ventilation. Lower serum 25(OH)D levels were associated with an increased number of comorbidities, COVID-19 severity, and the use of mechanical-ventilation. Thus, we suggest hospitalized COVID-19 patients to reach a sufficient vitamin D status to improve the clinical outcome of the disease.

Development and Validation of a Robust and Interpretable Early Triaging Support System for Patients Hospitalized With COVID-19: Predictive Algorithm Modeling and Interpretation Study

BACKGROUND

Robust and accurate prediction of severity for patients with COVID-19 is crucial for patient triaging decisions. Many proposed models were prone to either high bias risk or low-to-moderate discrimination. Some also suffered from a lack of clinical interpretability and were developed based on early pandemic period data. Hence, there has been a compelling need for advancements in prediction models for better clinical applicability.

OBJECTIVE

The primary objective of this study was to develop and validate a machine learning-based Robust and Interpretable Early Triaging Support (RIETS) system that predicts severity progression (involving any of the following events: intensive care unit admission, in-hospital death, mechanical ventilation required, or extracorporeal membrane oxygenation required) within 15 days upon hospitalization based on routinely available clinical and laboratory biomarkers.

METHODS

We included data from 5945 hospitalized patients with COVID-19 from 19 hospitals in South Korea collected between January 2020 and August 2022. For model development and external validation, the whole data set was partitioned into 2 independent cohorts by stratified random cluster sampling according to hospital type (general and tertiary care) and geographical location (metropolitan and nonmetropolitan). Machine learning models were trained and internally validated through a cross-validation technique on the development cohort. They were externally validated using a bootstrapped sampling technique on the external validation cohort. The best-performing model was selected primarily based on the area under the receiver operating characteristic curve (AUROC), and its robustness was evaluated using bias risk assessment. For model interpretability, we used Shapley and patient clustering methods.

RESULTS

Our final model, RIETS, was developed based on a deep neural network of 11 clinical and laboratory biomarkers that are readily available within the first day of hospitalization. The features predictive of severity included lactate dehydrogenase, age, absolute lymphocyte count, dyspnea, respiratory rate, diabetes mellitus, c-reactive protein, absolute neutrophil count, platelet count, white blood cell count, and saturation of peripheral oxygen. RIETS demonstrated excellent discrimination (AUROC=0.937; 95% CI 0.935-0.938) with high calibration (integrated calibration index=0.041), satisfied all the criteria of low bias risk in a risk assessment tool, and provided detailed interpretations of model parameters and patient clusters. In addition, RIETS showed potential for transportability across variant periods with its sustainable prediction on Omicron cases (AUROC=0.903, 95% CI 0.897-0.910).

CONCLUSIONS

RIETS was developed and validated to assist early triaging by promptly predicting the severity of hospitalized patients with COVID-19. Its high performance with low bias risk ensures considerably reliable prediction. The use of a nationwide multicenter cohort in the model development and validation implicates generalizability. The use of routinely collected features may enable wide adaptability. Interpretations of model parameters and patients can promote clinical applicability. Together, we anticipate that RIETS will facilitate the patient triaging workflow and efficient resource allocation when incorporated into a routine clinical practice.

Predicting invasive mechanical ventilation in COVID 19 patients: A validation study

INTRODUCTION

The decision to intubate and ventilate a patient is mainly clinical. Both delaying intubation (when needed) and unnecessarily invasively ventilating (when it can be avoided) are harmful. We recently developed an algorithm predicting respiratory failure and invasive mechanical ventilation in COVID-19 patients. This is an internal validation study of this model, which also suggests a categorized "time-weighted" model.

METHODS

We used a dataset of COVID-19 patients who were admitted to Rabin Medical Center after the algorithm was developed. We evaluated model performance in predicting ventilation, regarding the actual endpoint of each patient. We further categorized each patient into one of four categories, based on the strength of the prediction of ventilation over time. We evaluated this categorized model performance regarding the actual endpoint of each patient.

RESULTS

881 patients were included in the study; 96 of them were ventilated. AUC of the original algorithm is 0.87-0.94. The AUC of the categorized model is 0.95.

CONCLUSIONS

A minor degradation in the algorithm accuracy was noted in the internal validation, however, its accuracy remained high. The categorized model allows accurate prediction over time, with very high negative predictive value.

A randomized double-blind placebo-controlled trial of an inhibitor of plasminogen activator inhibitor-1 (TM5614) in mild to moderate COVID-19

An inhibitor of plasminogen activator inhibitor (PAI)-1, TM5614, inhibited thrombosis, inflammation, and fibrosis in several experimental mouse models. To evaluate the efficacy and safety of TM5614 in human COVID-19 pneumonia, phase IIa and IIb trials were conducted. In an open-label, single-arm trial, 26 Japanese COVID-19 patients with mild to moderate pneumonia were treated with 120-180 mg of TM5614 daily, and all were discharged without any notable side effects. Then, a randomized, double-blind, placebo-controlled trial was conducted in Japanese COVID-19 patients with mild to moderate pneumonia. The number of study participants was set to be 50 in each arm. Even after extension of the enrollment period, the number of study participants did not reach the initially intended sample size, and 75 patients were enrolled in the study. The total oxygenation scale from Day 1 to Day 14 as the primary endpoint was 1.5 in the TM5614 group vs 4.0 in the placebo group (p = 0.22), and the number of days of oxygen administration required as the secondary endpoint was 2.0 days in the TM5614 group vs 3.5 days in the placebo group (p = 0.34). Further studies will be necessary to verify the efficacy of PAI-1 inhibition for the treatment of COVID-19 pneumonia.Clinical trial registration: Two studies were conducted: a prospective, multicenter, open-label phase II study at https://jrct.niph.go.jp (jRCT2021200018) (First registration date 18/08/2020) and a prospective, multicenter, randomized, double-blind, placebo-controlled, phase II study at https://jrct.niph.go.jp (jRCT2021210006) (First registration date 28/05/2021).

Kinetics of C-reactive protein during extracorporeal membrane oxygenation

BACKGROUND

The exposure of blood to the artificial circuit during extracorporeal membrane oxygenation (ECMO) can induce an inflammatory response. C-reactive protein (CRP) is a commonly used biomarker of systemic inflammation.

METHODS

In this retrospective observational study, we analyzed results of daily plasma CRP measurements in 110 critically ill patients, treated with ECMO. We compared CRP levels during the first 5 days of ECMO operation, between different groups of patients according to ECMO configurations, Coronavirus disease 2019 (COVID-19) status, and mechanical ventilation parameters.

RESULTS

There was a statistically significant decrease in CRP levels during the first 5 days of veno-venous (VV) ECMO (173 ± 111 mg/L, 154 ± 107 mg/L, 127 ± 97 mg/L, 114 ± 100 mg/L and 118 ± 90 mg/L for days 1-5 respectively, p < 0.001). Simultaneously, there was a significant reduction in ventilatory parameters, as represented by the mechanical power (MP) calculation, from 24.02 ± 14.53 J/min to 6.18 ± 4.22 J/min within 3 h of VV ECMO initiation (p < 0.001). There was non-significant trend of increase in CRP level during the first 5 days of veno arterial (VA) ECMO (123 ± 80 mg/L, 179 ± 91 mg/L, 203 ± 90 mg/L, 179 ± 95 mg/L and 198 ± 93 for days 1-5 respectively, p = 0.126) and no significant change in calculated MP (from 14.28 ± 8.56 J/min to 10.81 ± 8.09 J/min within 3 h if ECMO initiation, p = 0.071).

CONCLUSIONS

We observed a significant decrease in CRP levels during the first 5 days of VV ECMO support, and suggest that the concomitant reduction in ventilatory MP may have mitigated the degree of alveolar stress and strain that could have contributed to a decrease in the systemic inflammatory process.

Safety and Outcome of High-Flow Nasal Oxygen Therapy Outside ICU Setting in Hypoxemic Patients With COVID-19

OBJECTIVE

High-flow nasal oxygen (HFNO) therapy is frequently applied outside ICU setting in hypoxemic patients with COVID-19. However, safety concerns limit more widespread use. We aimed to assess the safety and clinical outcomes of initiation of HFNO therapy in COVID-19 on non-ICU wards.

DESIGN

Prospective observational multicenter pragmatic study.

SETTING

Respiratory wards and ICUs of 10 hospitals in The Netherlands.

PATIENTS

Adult patients treated with HFNO for COVID-19-associated hypoxemia between December 2020 and July 2021 were included. Patients with treatment limitations were excluded from this analysis.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Outcomes included intubation and mortality rate, duration of hospital and ICU stay, severity of respiratory failure, and complications. Using propensity-matched analysis, we compared patients who initiated HFNO on the wards versus those in ICU. Six hundred eight patients were included, of whom 379 started HFNO on the ward and 229 in the ICU. The intubation rate in the matched cohort ( n = 214 patients) was 53% and 60% in ward and ICU starters, respectively ( p = 0.41). Mortality rates were comparable between groups (28-d [8% vs 13%], p = 0.28). ICU-free days were significantly higher in ward starters (21 vs 17 d, p < 0.001). No patient died before endotracheal intubation, and the severity of respiratory failure surrounding invasive ventilation and clinical outcomes did not differ between intubated ward and ICU starters (respiratory rate-oxygenation index 3.20 vs 3.38; Pa o2 :F io2 ratio 65 vs 64 mm Hg; prone positioning after intubation 81 vs 78%; mortality rate 17 vs 25% and ventilator-free days at 28 d 15 vs 13 d, all p values > 0.05).

CONCLUSIONS

In this large cohort of hypoxemic patients with COVID-19, initiation of HFNO outside the ICU was safe, and clinical outcomes were similar to initiation in the ICU. Furthermore, the initiation of HFNO on wards saved time in ICU without excess mortality or complicated course. Our results indicate that HFNO initiation outside ICU should be further explored in other hypoxemic diseases and clinical settings aiming to preserve ICU capacity and healthcare costs.

Clinical prediction models and the multiverse of madness

BACKGROUND

Each year, thousands of clinical prediction models are developed to make predictions (e.g. estimated risk) to inform individual diagnosis and prognosis in healthcare. However, most are not reliable for use in clinical practice.

MAIN BODY

We discuss how the creation of a prediction model (e.g. using regression or machine learning methods) is dependent on the sample and size of data used to develop it-were a different sample of the same size used from the same overarching population, the developed model could be very different even when the same model development methods are used. In other words, for each model created, there exists a multiverse of other potential models for that sample size and, crucially, an individual's predicted value (e.g. estimated risk) may vary greatly across this multiverse. The more an individual's prediction varies across the multiverse, the greater the instability. We show how small development datasets lead to more different models in the multiverse, often with vastly unstable individual predictions, and explain how this can be exposed by using bootstrapping and presenting instability plots. We recommend healthcare researchers seek to use large model development datasets to reduce instability concerns. This is especially important to ensure reliability across subgroups and improve model fairness in practice.

CONCLUSIONS

Instability is concerning as an individual's predicted value is used to guide their counselling, resource prioritisation, and clinical decision making. If different samples lead to different models with very different predictions for the same individual, then this should cast doubt into using a particular model for that individual. Therefore, visualising, quantifying and reporting the instability in individual-level predictions is essential when proposing a new model.

Prediction of oxygen supplementation by a deep-learning model integrating clinical parameters and chest CT images in COVID-19

PURPOSE

As of March 2023, the number of patients with COVID-19 worldwide is declining, but the early diagnosis of patients requiring inpatient treatment and the appropriate allocation of limited healthcare resources remain unresolved issues. In this study we constructed a deep-learning (DL) model to predict the need for oxygen supplementation using clinical information and chest CT images of patients with COVID-19.

MATERIALS AND METHODS

We retrospectively enrolled 738 patients with COVID-19 for whom clinical information (patient background, clinical symptoms, and blood test findings) was available and chest CT imaging was performed. The initial data set was divided into 591 training and 147 evaluation data. We developed a DL model that predicted oxygen supplementation by integrating clinical information and CT images. The model was validated at two other facilities (n = 191 and n = 230). In addition, the importance of clinical information for prediction was assessed.

RESULTS

The proposed DL model showed an area under the curve (AUC) of 89.9% for predicting oxygen supplementation. Validation from the two other facilities showed an AUC > 80%. With respect to interpretation of the model, the contribution of dyspnea and the lactate dehydrogenase level was higher in the model.

CONCLUSIONS

The DL model integrating clinical information and chest CT images had high predictive accuracy. DL-based prediction of disease severity might be helpful in the clinical management of patients with COVID-19.

Post-COVID-19 Condition in Children 6 and 12 Months After Infection

Importance

There is a need to understand the long-term outcomes among children infected with SARS-CoV-2.

Objective

To quantify the prevalence of post-COVID-19 condition (PCC) among children tested for SARS-CoV-2 infection in pediatric emergency departments (EDs).

Design, Setting, and Participants

Multicenter, prospective cohort study at 14 Canadian tertiary pediatric EDs that are members of the Pediatric Emergency Research Canada network with 90-day, 6-month, and 12-month follow-up. Participants were children younger than 18 years who were tested for SARS-CoV-2 infection between August 2020 and February 2022. Data were analyzed from May to November 2023.

Exposure

The presence of SARS-CoV-2 infection at or within 14 days of the index ED visit.

Main Outcomes and Measures

Presence of symptoms and QoL reductions that meet the PCC definition. This includes any symptom with onset within 3 months of infection that is ongoing at the time of follow-up and affects everyday functioning. The outcome was quantified at 6 and 12 months following the index ED visit.

Results

Among the 5147 children at 6 months (1152 with SARS-CoV-2 positive tests and 3995 with negative tests) and 5563 children at 12 months (1192 with SARS-CoV-2 positive tests and 4371 with negative tests) who had sufficient data regarding the primary outcome to enable PCC classification, the median (IQR) age was 2.0 (0.9-5.0) years, and 2956 of 5563 (53.1%) were male. At 6-month follow-up, symptoms and QoL changes consistent with the PCC definition were present in 6 of 1152 children with positive SARS-CoV-2 tests (0.52%) and 4 of 3995 children with negative SARS-CoV-2 tests (0.10%; absolute risk difference, 0.42%; 95% CI, 0.02% to 0.94%). The PCC definition was met at 12 months by 8 of 1192 children with positive SARS-CoV-2 tests (0.67%) and 7 of 4371 children with negative SARS-CoV-2 tests (0.16%; absolute risk difference, 0.51%; 95% CI, 0.06 to 1.08%). At 12 months, the median (IQR) PedsQL Generic Core Scale scores were 98.4 (90.0-100) among children with positive SARS-CoV-2 tests and 98.8 (91.7-100) among children with negative SARS-CoV-2 tests (difference, -0.3; 95% CI, -1.5 to 0.8; P = .56). Among the 8 children with SARS-CoV-2 positive tests and PCC at 12-month follow-up, children reported respiratory (7 of 8 patients [88%]), systemic (3 of 8 patients [38%]), and neurologic (1 of 8 patients [13%]) symptoms.

Conclusions and Relevance

In this cohort study of children tested for SARS-CoV-2 infection in Canadian pediatric EDs, although children infected with SARS-CoV-2 reported increased chronic symptoms, few of these children developed PCC, and overall QoL did not differ from children with negative SARS-CoV-2 tests.

Prevalence and prognostic value of neurological affections in hospitalized patients with moderate to severe COVID-19 based on objective assessments

Neurological manifestations of coronavirus disease 2019 (COVID-19) have been frequently described. In this prospective study of hospitalized COVID-19 patients without a history of neurological conditions, we aimed to analyze their prevalence and prognostic value based on established, standardized and objective methods. Patients were investigated using a multimodal electrophysiological approach, accompanied by neuropsychological and neurological examinations. Prevalence rates of central (CNS) and peripheral (PNS) nervous system affections were calculated and the relationship between neurological affections and mortality was analyzed using Firth logistic regression models. 184 patients without a history of neurological diseases could be enrolled. High rates of PNS affections were observed (66% of 138 patients receiving electrophysiological PNS examination). CNS affections were less common but still highly prevalent (33% of 139 examined patients). 63% of patients who underwent neuropsychological testing (n = 155) presented cognitive impairment. Logistic regression models revealed pathology in somatosensory evoked potentials as an independent risk factor of mortality (Odds Ratio: 6.10 [1.01-65.13], p = 0.049). We conclude that hospitalized patients with moderate to severe COVID-19 display high rates of PNS and CNS affection, which can be objectively assessed by electrophysiological examination. Electrophysiological assessment may have a prognostic value and could thus be helpful to identify patients at risk for deterioration.

COVID-19: An evaluation of predictive scoring systems in South Africa

Background

| The Coronavirus Disease 2019 (COVID-19) pandemic, caused by SARS-CoV-2, has resulted in more than 700 million cases worldwide. Sepsis and pneumonia severity scores assist in risk assessment of critical outcomes in patients with COVID-19. This allows healthcare workers to triage patients, by using clinical parameters and limited special investigations, thus offering the most appropriate level of care.

Methods

| A retrospective cohort study of 605 adult patients hospitalised with moderate to severe COVID-19, at a tertiary state hospital in South Africa. Evaluating the utility of the CURB65, NEWS2 and ISARIC-4C Mortality Score, in predicting critical outcomes, using clinical characteristics on admission. Outcomes included in-hospital mortality, invasive mechanical ventilation, and intensive care unit admission (ICU). Performance of severity scores and risk factors was assessed by area under the receiver operator characteristics (AUROC) analysis and logistic regression.

Findings |

A total of 605 records were used, 129 (21 %) non-survivors, 101 (17 %) ICU admissions and 77 (13 %) requiring invasive ventilation. Greater odds of mortality was associated with moderate and severe risk groups of the CURB65, ISARIC-4C and NEWS2 score. Mortality AUROC curve analysis for the CURB65 score was 0·76 (95 % CI: 0·71-0·8), 0·77 (95 % CI: 0·73-0·81) for the ISARIC-4C and 0·77 (95 % CI: 0·73-0·82) for the NEWS2 score. The CURB65 score had a sensitivity of 86 % with 12·8 % mortality, ISARIC-4C score a sensitivity of 87·6 % with 8 % mortality and NEWS2 score a sensitivity of 92·2 % with 8·6 % mortality.

Interpretation |

In 605 hospitalised patients with moderate to severe COVID-19, predominantly infected by the ancestral strain, good performance of the NEWS2 and ISARIC-4C score in predicting in-hospital mortality was noted. The CURB65 score had a high mortality rate in its low-risk group suggesting unexplained risk factors, not accounted for in the score, thus limiting its utility in the South African setting.

Long-term outcomes of the global tuberculosis and COVID-19 co-infection cohort

BACKGROUND

Longitudinal cohort data of patients with tuberculosis (TB) and coronavirus disease 2019 (COVID-19) are lacking. In our global study, we describe long-term outcomes of patients affected by TB and COVID-19.

METHODS

We collected data from 174 centres in 31 countries on all patients affected by COVID-19 and TB between 1 March 2020 and 30 September 2022. Patients were followed-up until cure, death or end of cohort time. All patients had TB and COVID-19; for analysis purposes, deaths were attributed to TB, COVID-19 or both. Survival analysis was performed using Cox proportional risk-regression models, and the log-rank test was used to compare survival and mortality attributed to TB, COVID-19 or both.

RESULTS

Overall, 788 patients with COVID-19 and TB (active or sequelae) were recruited from 31 countries, and 10.8% (n=85) died during the observation period. Survival was significantly lower among patients whose death was attributed to TB and COVID-19 versus those dying because of either TB or COVID-19 alone (p<0.001). Significant adjusted risk factors for TB mortality were higher age (hazard ratio (HR) 1.05, 95% CI 1.03-1.07), HIV infection (HR 2.29, 95% CI 1.02-5.16) and invasive ventilation (HR 4.28, 95% CI 2.34-7.83). For COVID-19 mortality, the adjusted risks were higher age (HR 1.03, 95% CI 1.02-1.04), male sex (HR 2.21, 95% CI 1.24-3.91), oxygen requirement (HR 7.93, 95% CI 3.44-18.26) and invasive ventilation (HR 2.19, 95% CI 1.36-3.53).

CONCLUSIONS

In our global cohort, death was the outcome in >10% of patients with TB and COVID-19. A range of demographic and clinical predictors are associated with adverse outcomes.

Head-to-head comparison of six warning scores to predict mortality and clinical impairment in COVID-19 patients in emergency department

The aim was to evaluate the ability of six risk scores (4C, CURB65, SEIMC, mCHOSEN, QuickCSI, and NEWS2) to predict the outcome of patients with COVID-19 during the sixth pandemic wave in Spain. A retrospective observational study was performed to review the electronic medical records in patients ≥ 18 years of age who consulted consecutively in an emergency department with COVID-19 diagnosis throughout 2 months during the sixth pandemic wave. Clinical-epidemiological variables, comorbidities, and their respective outcomes, such as 30-day in-hospital mortality and clinical deterioration risk (a combined outcome considering: mechanical ventilation, intensive care unit admission, and/or 30-day in-hospital mortality), were calculated. The area under the curve for each risk score was calculated, and the resulting curves were compared by the Delong test, concluding with a decision curve analysis. A total of 626 patients (median age 79 years; 49.8% female) fulfilled the inclusion criteria. Two hundred and ninety-three patients (46.8%) had two or more comorbidities. Clinical deterioration risk criteria were present in 10.1% (63 cases), with a 30-day in-hospital mortality rate of 6.2% (39 cases). Comparison of the results showed that score 4C presented the best results for both outcome variables, with areas under the curve for mortality and clinical deterioration risk of 0.931 (95% CI 0.904-0.957) and 0.871 (95% CI 0.833-0.910) (both p < 0.001). The 4C Mortality Score proved to be the best score for predicting mortality or clinical deterioration risk among patients with COVID-19 attended in the emergency department in the following 30 days.

Differential expression of biomarkers in saliva related to SARS-CoV-2 infection in patients with mild, moderate and severe COVID-19

BACKGROUND

Severe COVID-19 is a disease characterized by profound dysregulation of the innate immune system. There is a need to identify highly reliable prognostic biomarkers that can be rapidly assessed in body fluids for early identification of patients at higher risk for hospitalization and/or death. This study aimed to assess whether differential gene expression of immune response molecules and cellular enzymes, detected in saliva samples of COVID-19 patients, occurs according to disease severity staging.

METHODS

In this cross-sectional study, subjects with a COVID-19 diagnosis were classified as having mild, moderate, or severe disease based on clinical features. Transcripts of genes encoding 6 biomarkers, IL-1β, IL-6, IL-10, C-reactive protein, IDO1 and ACE2, were measured by RT‒qPCR in saliva samples of patients and COVID-19-free individuals.

RESULTS

The gene expression levels of all 6 biomarkers in saliva were significantly increased in severe disease patients compared to mild/moderate disease patients and healthy controls. A significant strong inverse relationship between oxemia and the level of expression of the 6 biomarkers (Spearman's correlation coefficient between -0.692 and -0.757; p < 0.001) was found.

CONCLUSIONS

Biomarker gene expression determined in saliva samples still needs to be validated as a potentially valuable predictor of severe clinical outcomes early at the onset of COVID-19 symptoms.

Low CRB-65 Scores Effectively Rule out Adverse Clinical Outcomes in COVID-19 Irrespective of Chest Radiographic Abnormalities

Background: CRB-65 (Confusion; Respiratory rate ≥ 30/min; Blood pressure ≤ 90/60 mmHg; age ≥ 65 years) is a risk score for prognosticating patients with COVID-19 pneumonia. However, a significant proportion of COVID-19 patients have normal chest X-rays (CXRs). The influence of CXR abnormalities on the prognostic value of CRB-65 is unknown, limiting its wider applicability. Methods: We assessed the influence of CXR abnormalities on the prognostic value of CRB-65 in COVID-19. Results: In 589 study patients (71 years (IQR: 57-83); 57% males), 186 (32%) had normal CXRs. On ROC analysis, CRB-65 performed similarly in patients with normal vs. abnormal CXRs for predicting inpatient mortality (AUC 0.67 ± 0.05 vs. 0.69 ± 0.03). In patients with normal CXRs, a CRB-65 of 0 ruled out mortality, NIV requirement and critical illness (intubation and/or ICU admission) with negative predictive values (NPVs) of 94%, 98% and 99%, respectively. In patients with abnormal CXRs, a CRB-65 of 0 ruled out the same endpoints with NPVs of 91%, 83% and 86%, respectively. Patients with low CRB-65 scores had better inpatient survival than patients with high CRB-65 scores, irrespective of CXR abnormalities (all p < 0.05). Conclusions: CRB-65, CXR and CRP are independent predictors of mortality in COVID-19. Adding CXR findings (dichotomised to either normal or abnormal) to CRB-65 does not improve its prognostic accuracy. A low CRB-65 score of 0 may be a good rule-out test for adverse clinical outcomes in COVID-19 patients with normal or abnormal CXRs, which deserves prospective validation.

Hemodynamic, Oxygenation and Lymphocyte Parameters Predict COVID-19 Mortality

The mortality of COVID-19 patients has left the world devastated. Many scoring systems have been developed to predict the mortality of COVID-19 patients, but several scoring components cannot be carried out in limited health facilities. Herein, the authors attempted to create a new and easy scoring system involving mean arterial pressure (MAP), PF Ratio, or SF ratio-respiration rate (SF Ratio-R), and lymphocyte absolute, which were abbreviated as MPL or MSLR functioning, as a predictive scoring system for mortality within 30 days for COVID-19 patients. Of 132 patients with COVID-19 hospitalized between March and November 2021, we followed up on 96 patients. We present bivariate and multivariate analyses as well as the area under the curve (AUC) and Kaplan-Meier charts. From 96 patients, we obtained an MPL score of 3 points: MAP < 75 mmHg, PF Ratio < 200, and lymphocyte absolute < 1500/µL, whereas the MSLR score was 6 points: MAP < 75 mmHg, SF Ratio < 200, lymphocyte absolute < 1500/µL, and respiration rate 24/min. The MPL cut-off point is 2, while the MSLR is 4. MPL and MSLR have the same sensitivity (79.1%) and specificity (75.5%). The AUC value of MPL vs. MSLR was 0.802 vs. 0.807. The MPL ≥ 2 and MSLR ≥ 4 revealed similar predictions for survival within 30 days (p < 0.05). Conclusion: MPL and MSLR scores are potential predictors of mortality in COVID-19 patients within 30 days in a resource-limited country.

Evaluation of a remote monitoring service for patients with COVID-19 discharged from University College London Hospital

INTRODUCTION

In May 2020 a virtual ward for COVID-19 patients seen at University College London Hospital (UCLH) was established. The aim of this study was to see if specific factors can be used to predict the risk of deterioration and need for Emergency Department (ED) reattendance or admission.

METHODS

We performed a service evaluation of the COVID-19 virtual ward service at UCLH between 24/10/2020 and 12/2/2021. 649 patients were included with data collected on vital signs, basic measurements, and blood tests from their initial ED attendance, allowing calculation of ISARIC-4C mortality scores. Outcomes of interest were ED reattendance, facilitation of this by virtual ward physician, level of care if admitted, and death within 28 days of the first COVID-19 virtual ward appointment. Analysis was performed using Mann-Whitney U tests.

RESULTS

Reattendance rate to ED was 17.3% (112/649) of which 8% (51/649) were admitted. Half of ED reattendances were facilitated by the virtual ward service. Overall mortality was 0.92%. Patients who reattended ED, facilitated by the virtual ward service, had a higher mean CRP (53.63 vs 41.67 mg/L), presented to ED initially later in their COVID-19 illness (8 vs 6.5 days) and had a higher admission rate (61 vs 39%). The mean ISARIC-4C score was higher in the reattendance group compared to the non-reattendance group (3.87 vs 3.48, difference of 0.179, p = 0.003). The mean ISARIC-4C score was higher in the admission group than the non-reattendance group (5.56 vs 3.48, difference of 0.115, p = 0.003).

CONCLUSION

Identification of patient risk factors for reattendance following a diagnosis of COVID-19 in ED can be used to design a service to safely manage patients remotely. We found that the ISARIC -4C mortality score was associated with risk of hospital admission and could be used to identify those requiring more active remote follow up.

Multimodal deep learning for COVID-19 prognosis prediction in the emergency department: a bi-centric study

Predicting clinical deterioration in COVID-19 patients remains a challenging task in the Emergency Department (ED). To address this aim, we developed an artificial neural network using textual (e.g. patient history) and tabular (e.g. laboratory values) data from ED electronic medical reports. The predicted outcomes were 30-day mortality and ICU admission. We included consecutive patients from Humanitas Research Hospital and San Raffaele Hospital in the Milan area between February 20 and May 5, 2020. We included 1296 COVID-19 patients. Textual predictors consisted of patient history, physical exam, and radiological reports. Tabular predictors included age, creatinine, C-reactive protein, hemoglobin, and platelet count. TensorFlow tabular-textual model performance indices were compared to those of models implementing only tabular data. For 30-day mortality, the combined model yielded slightly better performances than the tabular fastai and XGBoost models, with AUC 0.87 ± 0.02, F1 score 0.62 ± 0.10 and an MCC 0.52 ± 0.04 (p < 0.32). As for ICU admission, the combined model MCC was superior (p < 0.024) to the tabular models. Our results suggest that a combined textual and tabular model can effectively predict COVID-19 prognosis which may assist ED physicians in their decision-making process.

Who to escalate during a pandemic? A retrospective observational study about decision-making during the COVID-19 pandemic in the UK

BACKGROUND

Optimal decision-making regarding who to admit to critical care in pandemic situations remains unclear. We compared age, Clinical Frailty Score (CFS), 4C Mortality Score and hospital mortality in two separate COVID-19 surges based on the escalation decision made by the treating physician.

METHODS

A retrospective analysis of all referrals to critical care during the first COVID-19 surge (cohort 1, March/April 2020) and a late surge (cohort 2, October/November 2021) was undertaken. Patients with confirmed or high clinical suspicion of COVID-19 infection were included. A senior critical care physician assessed all patients regarding their suitability for potential intensive care unit admission. Demographics, CFS, 4C Mortality Score and hospital mortality were compared depending on the escalation decision made by the attending physician.

RESULTS

203 patients were included in the study, 139 in cohort 1 and 64 in cohort 2. There were no significant differences in age, CFS and 4C scores between the two cohorts. Patients deemed suitable for escalation by clinicians were significantly younger with significantly lower CFS and 4C scores compared with patients who were not deemed to benefit from escalation. This pattern was observed in both cohorts. Mortality in patients not deemed suitable for escalation was 61.8% in cohort 1 and 47.4% in cohort 2 (p<0.001).

CONCLUSIONS

Decisions who to escalate to critical care in settings with limited resources pose moral distress on clinicians. 4C score, age and CFS did not change significantly between the two surges but differed significantly between patients deemed suitable for escalation and those deemed unsuitable by clinicians. Risk prediction tools may be useful in a pandemic to supplement clinical decision-making, even though escalation thresholds require adjustments to reflect changes in risk profile and outcomes between different pandemic surges.

Risk Factor for Superimposed Nosocomial Bloodstream Infections in Hospitalized Patients with COVID-19

Introduction

Corticosteroids are a component of the standard therapy for patients with coronavirus disease 2019 (COVID-19) because of the immunological dysregulation and hyperinflammation associated with the condition. This study aimed to evaluate the potential risk factors for nosocomial bloodstream infections in hospitalized patients with COVID-19, including the exploration of corticosteroid dosage and treatment duration.

Materials and Methods

A retrospective cohort study of hospitalized patients with COVID-19 was conducted in a tertiary care hospital. We performed univariate and multivariate analyses of various parameters to identify risk factors for nosocomial bloodstream infection.

Results

Of 252 patients, 19% had nosocomial bloodstream infections. The mortality rate of nosocomial bloodstream infections was 62.5%. Multivariate analysis revealed that male sex (odds ratio [OR] 3.43; 95% confidence interval [CI]: 1.60-7.33), receiving methylprednisolone (OR: 3.01; 95% CI: 1.24-7.31), receiving an equivalent dexamethasone dose of 6-12 mg/day (OR: 7.49; 95% CI: 2.08-26.94), and leukocytosis on admission (OR: 4.13; 95% CI: 1.89-9.01) were significant predictors of nosocomial bloodstream infections.

Conclusion

Unmodified risk variables for nosocomial bloodstream infections included male sex and leukocytosis at admission. Using methylprednisolone and obtaining a cumulative dosage of dexamethasone were adjusted risk variables associated with superimposed nosocomial bloodstream infections in hospitalized patients with COVID-19.

C reactive protein utilisation, a biomarker for early COVID-19 treatment, improves lenzilumab efficacy: results from the randomised phase 3 'LIVE-AIR' trial

OBJECTIVE

COVID-19 severity is correlated with granulocyte macrophage colony-stimulating factor (GM-CSF) and C reactive protein (CRP) levels. In the phase three LIVE-AIR trial, lenzilumab an anti-GM-CSF monoclonal antibody, improved the likelihood of survival without ventilation (SWOV) in COVID-19, with the greatest effect in participants having baseline CRP below a median of 79 mg/L. Herein, the utility of baseline CRP to guide lenzilumab treatment was assessed.

DESIGN

A subanalysis of the randomised, blinded, controlled, LIVE-AIR trial in which lenzilumab or placebo was administered on day 0 and participants were followed through Day 28.

PARTICIPANTS

Hospitalised COVID-19 participants (N=520) with SpO2 ≤94% on room air or requiring supplemental oxygen but not invasive mechanical ventilation.

INTERVENTIONS

Lenzilumab (1800 mg; three divided doses, q8h, within 24 hours) or placebo infusion alongside corticosteroid and remdesivir treatments.

MAIN OUTCOME MEASURES

The primary endpoint was the time-to-event analysis difference in SWOV through day 28 between lenzilumab and placebo treatments, stratified by baseline CRP.

RESULTS

SWOV was achieved in 152 (90%; 95% CI 85 to 94) lenzilumab and 144 (79%; 72 to 84) placebo-treated participants with baseline CRP <150 mg/L (HR: 2.54; 95% CI 1.46 to 4.41; p=0.0009) but not with CRP ≥150 mg/L (HR: 1.04; 95% CI 0.51 to 2.14; p=0.9058). A statistically significant interaction between CRP and lenzilumab treatment was observed (p=0.044). Grade ≥3 adverse events with lenzilumab were comparable to placebo in both CRP strata. No treatment-emergent serious adverse events were attributed to lenzilumab.

CONCLUSION

Hospitalised hypoxemic patients with COVID-19 with baseline CRP <150 mg/L derived the greatest clinical benefit from treatment with lenzilumab.

TRIAL REGISTRATION NUMBER

NCT04351152; ClinicalTrials.gov.

Stratifying risk outcomes among adult COVID-19 inpatients with high flow oxygen: The R4 score

BACKGROUND

High flow oxygen therapy (HFO) is a widely used intervention for pulmonary complications. Amid the coronavirus infectious disease 2019 (COVID-19) pandemic, HFO became a popular alternative to conventional oxygen supplementation therapies. Risk stratification tools have been repurposed -and new ones developed- to estimate outcome risks among COVID-19 patients. This study aims to provide a simple risk stratification system to predict invasive mechanical ventilation (IMV) or death among COVID-19 inpatients on HFO.

METHODS

Among 529 adult inpatients with COVID-19 pneumonia, we selected unadjusted clinical risk factors for developing the composite endpoint of IMV or death. The risk for the primary outcome by each category was estimated using a Cox proportional hazards model. Bootstrapping was used to validate the results.

RESULTS

Age above 62, eGFR under 60 ml/min, room air SpO2 ≤89 % upon admission, history of hypertension, history of diabetes, and any comorbidity (cancer, cardiovascular disease, COPD/ asthma, hypothyroidism, or autoimmune disease) were considered for the score. Each of the six criteria scored 1 point. The score was further simplified into 4 categories: 1) 0 criteria, 2) 1 criterion, 3) 2-3 criteria, and 4) ≥4 criteria. Taking the first category as the reference, risk estimates for the primary endpoint were HR; 2.94 [1.67 - 5.26], 4.08 [2.63 - 7.05], and 6.63 [3.74 - 11.77], respectively. In ROC analysis, the AUC for the model was 0.72.

CONCLUSIONS

Our score uses simple criteria to estimate the risk for IMV or death among COVID-19 inpatients with HFO. Higher category reflects consistent increases in risk for the endpoint.

Recovery from Covid-19 critical illness: A secondary analysis of the ISARIC4C CCP-UK cohort study and the RECOVER trial

Background

We aimed to compare the prevalence and severity of fatigue in survivors of Covid-19 versus non-Covid-19 critical illness, and to explore potential associations between baseline characteristics and worse recovery.

Methods

We conducted a secondary analysis of two prospectively collected datasets. The population included was 92 patients who received invasive mechanical ventilation (IMV) with Covid-19, and 240 patients who received IMV with non-Covid-19 illness before the pandemic. Follow-up data were collected post-hospital discharge using self-reported questionnaires. The main outcome measures were self-reported fatigue severity and the prevalence of severe fatigue (severity >7/10) 3 and 12-months post-hospital discharge.

Results

Covid-19 IMV-patients were significantly younger with less prior comorbidity, and more males, than pre-pandemic IMV-patients. At 3-months, the prevalence (38.9% [7/18] vs. 27.1% [51/188]) and severity (median 5.5/10 vs 5.0/10) of fatigue were similar between the Covid-19 and pre-pandemic populations, respectively. At 6-months, the prevalence (10.3% [3/29] vs. 32.5% [54/166]) and severity (median 2.0/10 vs. 5.7/10) of fatigue were less in the Covid-19 cohort. In the total sample of IMV-patients included (i.e. all Covid-19 and pre-pandemic patients), having Covid-19 was significantly associated with less severe fatigue (severity <7/10) after adjusting for age, sex and prior comorbidity (adjusted OR 0.35 (95%CI 0.15-0.76, p=0.01).

Conclusion

Fatigue may be less severe after Covid-19 than after other critical illness.

Validation of the qSOFA and CRB-65 in SARS-CoV-2-infected community-acquired pneumonia

Rationale

Prognostic accuracy of the quick sequential organ failure assessment (qSOFA) and CRB-65 (confusion, respiratory rate, blood pressure and age (≥65 years)) risk scores have not been widely evaluated in patients with SARS-CoV-2-positive compared to SARS-CoV-2-negative community-acquired pneumonia (CAP). The aim of the present study was to validate the qSOFA(-65) and CRB-65 scores in a large cohort of SARS-CoV-2-positive and SARS-CoV-2-negative CAP patients.

Methods

We included all cases with CAP hospitalised in 2020 from the German nationwide mandatory quality assurance programme and compared cases with SARS-CoV-2 infection to cases without. We excluded cases with unclear SARS-CoV-2 infection state, transferred to another hospital or on mechanical ventilation during admission. Predefined outcomes were hospital mortality and need for mechanical ventilation.

Results

Among 68 594 SARS-CoV-2-positive patients, hospital mortality (22.7%) and mechanical ventilation (14.9%) were significantly higher when compared to 167 880 SARS-CoV-2-negative patients (15.7% and 9.2%, respectively). All CRB-65 and qSOFA criteria were associated with both outcomes, and age dominated mortality prediction in SARS-CoV-2 (risk ratio >9). Scores including the age criterion had higher area under the curve (AUCs) for mortality in SARS-CoV-2-positive patients (e.g. CRB-65 AUC 0.76) compared to SARS-CoV-2 negative patients (AUC 0.68), and negative predictive value was highest for qSOFA-65=0 (98.2%). Sensitivity for mechanical ventilation prediction was poor with all scores (AUCs 0.59-0.62), and negative predictive values were insufficient (qSOFA-65=0 missed 1490 out of 10 198 patients (∼15%) with mechanical ventilation). Results were similar when excluding frail and palliative patients.

Conclusions

Hospital mortality and mechanical ventilation rates were higher in SARS-CoV-2-positive than SARS-CoV-2-negative CAP. For SARS-CoV-2-positive CAP, the CRB-65 and qSOFA-65 scores showed adequate prediction of mortality but not of mechanical ventilation.

An artificial intelligence approach for predicting death or organ failure after hospitalization for COVID-19: development of a novel risk prediction tool and comparisons with ISARIC-4C, CURB-65, qSOFA, and MEWS scoring systems

BACKGROUND

We applied machine learning (ML) algorithms to generate a risk prediction tool [Collaboration for Risk Evaluation in COVID-19 (CORE-COVID-19)] for predicting the composite of 30-day endotracheal intubation, intravenous administration of vasopressors, or death after COVID-19 hospitalization and compared it with the existing risk scores.

METHODS

This is a retrospective study of adults hospitalized with COVID-19 from March 2020 to February 2021. Patients, each with 92 variables, and one composite outcome underwent feature selection process to identify the most predictive variables. Selected variables were modeled to build four ML algorithms (artificial neural network, support vector machine, gradient boosting machine, and Logistic regression) and an ensemble model to generate a CORE-COVID-19 model to predict the composite outcome and compared with existing risk prediction scores. The net benefit for clinical use of each model was assessed by decision curve analysis.

RESULTS

Of 1796 patients, 278 (15%) patients reached primary outcome. Six most predictive features were identified. Four ML algorithms achieved comparable discrimination (P > 0.827) with c-statistics ranged 0.849-0.856, calibration slopes 0.911-1.173, and Hosmer-Lemeshow P > 0.141 in validation dataset. These 6-variable fitted CORE-COVID-19 model revealed a c-statistic of 0.880, which was significantly (P < 0.04) higher than ISARIC-4C (0.751), CURB-65 (0.735), qSOFA (0.676), and MEWS (0.674) for outcome prediction. The net benefit of the CORE-COVID-19 model was greater than that of the existing risk scores.

CONCLUSION

The CORE-COVID-19 model accurately assigned 88% of patients who potentially progressed to 30-day composite events and revealed improved performance over existing risk scores, indicating its potential utility in clinical practice.

Development of a Risk Score for AKI onset in COVID-19 Patients: COV-AKI Score

PURPOSE

Acute Kidney Injury (AKI) in COVID-19 patients is associated with increased morbidity and mortality. In the present study, we aimed to develop a prognostic score to predict AKI development in these patients.

MATERIALS AND METHODS

This was a retrospective observational study of 2334 COVID 19 patients admitted to 23 different hospitals in Brazil, between January 10th and August 30rd, 2020. The primary outcome of AKI was defined as any increase in serum creatinine (SCr) by 0.3 mg/dL within 48 h or a change in SCr by ≥ 1.5 times of baseline within 1 week, based on Kidney Disease Improving Global Outcomes (KDIGO) guidelines. All patients aged ≥ 18 y/o admitted with confirmed SARS-COV-2 infection were included. Discrimination of variables was calculated by the Receiver Operator Characteristic Curve (ROC curve) utilizing area under curve. Some continuous variables were categorized through ROC curve. The cutoff points were calculated using the value with the best sensitivity and specificity.

RESULTS

A total of 1131 patients with COVID-19 admitted to the ICU were included. Patients mean age was 52 ± 15,8 y/o., with a prevalence of males 60% (n = 678). The risk of AKI was 33% (n = 376), 78% (n = 293) of which did not require dialysis. Overall mortality was 11% (n = 127), while for AKI patients, mortality rate was 21% (n = 80). Variables selected for the logistic regression model and inclusion in the final prognostic score were the following: age, diabetes, ACEis, ARBs, chronic kidney disease and hypertension.

CONCLUSION

AKI development in COVID 19 patients is accurately predicted by common clinical variables, allowing early interventions to attenuate the impact of AKI in these patients.

Real-world effectiveness of sotrovimab and remdesivir for early treatment of high-risk hospitalized COVID-19 patients: A propensity score adjusted retrospective cohort study

Early treatment of high-risk COVID-19 patients may prevent disease progression. However, there are limited data to support treatment of hospitalized or fully vaccinated patients with mild-to-moderate disease. In this retrospective cohort study, we studied the effect of early use of sotrovimab and remdesivir in high-risk hospitalized COVID-19 patients. We included PCR-confirmed COVID-19 patients admitted to the National Centre for Infectious Diseases who presented within the first 5 days of illness, and who were not requiring oxygen or ICU care at presentation. Sotrovimab- and remdesivir-treated groups were compared with control (no early treatment). A multiple propensity-score adjusted multivariable regression analysis was conducted with a composite primary endpoint of in-hospital deterioration (oxygen requirement, ICU admission, or mortality). Of 1118 patients, 841 were in the control group, 106 in the sotrovimab group and 169 in the remdesivir group. The median age was 63 years (IQR 46-74 years) and 505 (45.2%) were female. In unvaccinated patients, both remdesivir and sotrovimab treatment were protective (adjusted odds ratio [aOR] 0.19, 95% CI 0.064-0.60 and 0.18 [95% CI 0.066-0.47]), respectively. Contrarily, among the vaccinated patients there was no significant treatment effect with early remdesivir treatment (aOR 2.51, 95% CI 0.83-7.57, p = 0.10). Remdesivir and sotrovimab treatment, given early in the disease course to unvaccinated high-risk patients, was effective in reducing the risk of in-hospital deterioration and severe disease. This effect was not seen in fully vaccinated patients, which may be due to a small sample size or residual confounding.

Non-Typical Clinical Presentation of COVID-19 Patients in Association with Disease Severity and Length of Hospital Stay

BACKGROUND

This study aimed to investigate the incidence of non-typical symptoms in ambulatory patients with mild-to-moderate COVID-19 infection and their potential association with disease progression.

MATERIALS AND METHODS

Data on the symptomatology of COVID-19 patients presenting to the fast-track emergency department were collected between March 2020 and March 2021. Fever, cough, shortness of breath, and fatigue-weakness were defined as "typical" symptoms, whereas all other symptoms such as nasal congestion, rhinorrhea, gastrointestinal symptoms, etc., were defined as "non-typical".

RESULTS

A total of 570 COVID-19 patients with a mean age of 42.25 years were included, the majority of whom were male (61.3%; N = 349), and were divided according to their symptoms into two groups. The mean length of hospital stay was found to be 9.5 days. A higher proportion of patients without non-typical symptoms were admitted to the hospital (p = 0.001) and the ICU (p = 0.048) as well. No significant differences were observed between non-typical symptoms and outcome (p = 0.685). Patients who did not demonstrate at least one non-typical symptom had an extended length of stay (p = 0.041). No statistically significant differences in length of hospital stay were associated with individual symptoms.

CONCLUSION

With the possible exception of gastrointestinal symptoms, non-typical symptoms of COVID-19 at baseline appear to predispose to a milder disease.

Early biochemical analysis of COVID-19 patients helps severity prediction

COVID-19 pandemic has put the protocols and the capacity of our Hospitals to the test. The management of severe patients admitted to the Intensive Care Units has been a challenge for all health systems. To assist in this challenge, various models have been proposed to predict mortality and severity, however, there is no clear consensus for their use. In this work, we took advantage of data obtained from routine blood tests performed on all individuals on the first day of hospitalization. These data has been obtained by standardized cost-effective technique available in all the hospitals. We have analyzed the results of 1082 patients with COVID19 and using artificial intelligence we have generated a predictive model based on data from the first days of admission that predicts the risk of developing severe disease with an AUC = 0.78 and an F1-score = 0.69. Our results show the importance of immature granulocytes and their ratio with Lymphocytes in the disease and present an algorithm based on 5 parameters to identify a severe course. This work highlights the importance of studying routine analytical variables in the early stages of hospital admission and the benefits of applying AI to identify patients who may develop severe disease.

Construction and validation of a deterioration model for elderly COVID-19 Sub-variant BA.2 patients

Rationale

COVID-19 pandemic has imposed tremendous stress and burden on the economy and society worldwide. There is an urgent demand to find a new model to estimate the deterioration of patients inflicted by Omicron variants.

Objective

This study aims to develop a model to predict the deterioration of elderly patients inflicted by Omicron Sub-variant BA.2.

Methods

COVID-19 patients were randomly divided into the training and the validation cohorts. Both Lasso and Logistic regression analyses were performed to identify prediction factors, which were then selected to build a deterioration model in the training cohort. This model was validated in the validation cohort.

Measurements and main results

The deterioration model of COVID-19 was constructed with five indices, including C-reactive protein, neutrophil count/lymphocyte count (NLR), albumin/globulin ratio (A/G), international normalized ratio (INR), and blood urea nitrogen (BUN). The area under the ROC curve (AUC) showed that this model displayed a high accuracy in predicting deterioration, which was 0.85 in the training cohort and 0.85 in the validation cohort. The nomogram provided an easy way to calculate the possibility of deterioration, and the decision curve analysis (DCA) and clinical impact curve analysis (CICA)showed good clinical net profit using this model.

Conclusion

The model we constructed can identify and predict the risk of deterioration (requirement for ventilatory support or death) in elderly patients and it is clinically practical, which will facilitate medical decision making and allocating medical resources to those with critical conditions.

Machine learning predicts the short-term requirement for invasive ventilation among Australian critically ill COVID-19 patients

OBJECTIVE(S)

To use machine learning (ML) to predict short-term requirements for invasive ventilation in patients with COVID-19 admitted to Australian intensive care units (ICUs).

DESIGN

A machine learning study within a national ICU COVID-19 registry in Australia.

PARTICIPANTS

Adult patients who were spontaneously breathing and admitted to participating ICUs with laboratory-confirmed COVID-19 from 20 February 2020 to 7 March 2021. Patients intubated on day one of their ICU admission were excluded.

MAIN OUTCOME MEASURES

Six machine learning models predicted the requirement for invasive ventilation by day three of ICU admission from variables recorded on the first calendar day of ICU admission; (1) random forest classifier (RF), (2) decision tree classifier (DT), (3) logistic regression (LR), (4) K neighbours classifier (KNN), (5) support vector machine (SVM), and (6) gradient boosted machine (GBM). Cross-validation was used to assess the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity of machine learning models.

RESULTS

300 ICU admissions collected from 53 ICUs across Australia were included. The median [IQR] age of patients was 59 [50-69] years, 109 (36%) were female and 60 (20%) required invasive ventilation on day two or three. Random forest and Gradient boosted machine were the best performing algorithms, achieving mean (SD) AUCs of 0.69 (0.06) and 0.68 (0.07), and mean sensitivities of 77 (19%) and 81 (17%), respectively.

CONCLUSION

Machine learning can be used to predict subsequent ventilation in patients with COVID-19 who were spontaneously breathing and admitted to Australian ICUs.

A prediction model for COVID-19 liver dysfunction in patients with normal hepatic biochemical parameters

Coronavirus disease 2019 (COVID-19) patients with liver dysfunction (LD) have a higher chance of developing severe and critical disease. The routine hepatic biochemical parameters ALT, AST, GGT, and TBIL have limitations in reflecting COVID-19-related LD. In this study, we performed proteomic analysis on 397 serum samples from 98 COVID-19 patients to identify new biomarkers for LD. We then established 19 simple machine learning models using proteomic measurements and clinical variables to predict LD in a development cohort of 74 COVID-19 patients with normal hepatic biochemical parameters. The model based on the biomarker ANGL3 and sex (AS) exhibited the best discrimination (time-dependent AUCs: 0.60-0.80), calibration, and net benefit in the development cohort, and the accuracy of this model was 69.0-73.8% in an independent cohort. The AS model exhibits great potential in supporting optimization of therapeutic strategies for COVID-19 patients with a high risk of LD. This model is publicly available at https://xixihospital-liufang.shinyapps.io/DynNomapp/.

Using ISARIC 4C mortality score to predict dynamic changes in mortality risk in COVID-19 patients during hospital admission

As SARS-CoV-2 infections continue to cause hospital admissions around the world, there is a continued need to accurately assess those at highest risk of death to guide resource use and clinical management. The ISARIC 4C mortality score provides mortality risk prediction at admission to hospital based on demographic and physiological parameters. Here we evaluate dynamic use of the 4C score at different points following admission. Score components were extracted for 6,373 patients admitted to Barts Health NHS Trust hospitals between 1st August 2020 and 19th July 2021 and total score calculated every 48 hours for 28 days. Area under the receiver operating characteristic (AUC) statistics were used to evaluate discrimination of the score at admission and subsequent inpatient days. Patients who were still in hospital at day 6 were more likely to die if they had a higher score at day 6 than others also still in hospital who had the same score at admission. Discrimination of dynamic scoring in those still in hospital was superior with the area under the curve 0.71 (95% CI 0.69-0.74) at admission and 0.82 (0.80-0.85) by day 8. Clinically useful changes in the dynamic parts of the score are unlikely to be associated with subject-level measurements. Dynamic use of the ISARIC 4C score is likely to provide accurate and timely information on mortality risk during a patient's hospital admission.

Predicting major complications in patients undergoing laparoscopic and open hysterectomy for benign indications

BACKGROUND

Hysterectomy, the most common gynecological operation, requires surgeons to counsel women about their operative risks. We aimed to develop and validate multivariable logistic regression models to predict major complications of laparoscopic or abdominal hysterectomy for benign conditions.

METHODS

We obtained routinely collected health administrative data from the English National Health Service (NHS) from 2011 to 2018. We defined major complications based on core outcomes for postoperative complications including ureteric, gastrointestinal and vascular injury, and wound complications. We specified 11 predictors a priori. We used internal-external cross-validation to evaluate discrimination and calibration across 7 NHS regions in the development cohort. We validated the final models using data from an additional NHS region.

RESULTS

We found that major complications occurred in 4.4% (3037/68 599) of laparoscopic and 4.9% (6201/125 971) of abdominal hysterectomies. Our models showed consistent discrimination in the development cohort (laparoscopic, C-statistic 0.61, 95% confidence interval [CI] 0.60 to 0.62; abdominal, C-statistic 0.67, 95% CI 0.64 to 0.70) and similar or better discrimination in the validation cohort (laparoscopic, C-statistic 0.67, 95% CI 0.65 to 0.69; abdominal, C-statistic 0.67, 95% CI 0.65 to 0.69). Adhesions were most predictive of complications in both models (laparoscopic, odds ratio [OR] 1.92, 95% CI 1.73 to 2.13; abdominal, OR 2.46, 95% CI 2.27 to 2.66). Other factors predictive of complications included adenomyosis in the laparoscopic model, and Asian ethnicity and diabetes in the abdominal model. Protective factors included age and diagnoses of menstrual disorders or benign adnexal mass in both models and diagnosis of fibroids in the abdominal model.

INTERPRETATION

Personalized risk estimates from these models, which showed moderate discrimination, can inform clinical decision-making for people with benign conditions who may require hysterectomy.

External validation of the COVID-19 4C mortality score in an urban United States cohort

BACKGROUND

Identifying patients at risk for mortality from COVID-19 is crucial to triage, clinical decision-making, and the allocation of scarce hospital resources. The 4C Mortality Score effectively predicts COVID-19 mortality, but it has not been validated in a United States (U.S.) population. The purpose of this study is to determine whether the 4C Mortality Score accurately predicts COVID-19 mortality in an urban U.S. adult inpatient population.

METHODS

This retrospective cohort study included adult patients admitted to a single-center, tertiary care hospital (Philadelphia, PA) with a positive SARS-CoV-2 PCR from 3/01/2020 to 6/06/2020. Variables were extracted through a combination of automated export and manual chart review. The outcome of interest was mortality during hospital admission or within 30 days of discharge.

RESULTS

This study included 426 patients; mean age was 64.4 years, 43.4% were female, and 54.5% self-identified as Black or African American. All-cause mortality was observed in 71 patients (16.7%). The area under the receiver operator characteristic curve of the 4C Mortality Score was 0.85 (95% confidence interval, 0.79-0.89).

CONCLUSIONS

Clinicians may use the 4C Mortality Score in an urban, majority Black, U.S. inpatient population. The derivation and validation cohorts were treated in the pre-vaccine era so the 4C Score may over-predict mortality in current patient populations. With stubbornly high inpatient mortality rates, however, the 4C Score remains one of the best tools available to date to inform thoughtful triage and treatment allocation.

External Validation of Mortality Scores among High-Risk COVID-19 Patients: A Romanian Retrospective Study in the First Pandemic Year

Background: We aimed to externally validate three prognostic scores for COVID-19: the 4C Mortality Score (4CM Score), the COVID-GRAM Critical Illness Risk Score (COVID-GRAM), and COVIDAnalytics. Methods: We evaluated the scores in a retrospective study on adult patients hospitalized with severe/critical COVID-19 (1 March 2020–1 March 2021), in the Teaching Hospital of Infectious Diseases, Cluj-Napoca, Romania. We assessed all the deceased patients matched with two survivors by age, gender, and at least two comorbidities. The areas under the receiver-operating characteristic curves (AUROCs) were computed for in-hospital mortality. Results: Among 780 severe/critical COVID-19 patients, 178 (22.8%) died. We included 474 patients according to the case definition (158 deceased/316 survivors). The median age was 75 years; diabetes mellitus, malignancies, chronic pulmonary diseases, and chronic kidney and moderate/severe liver diseases were associated with higher risks of death. According to the predefined 4CM Score, the mortality rates were 0% (low), 13% (intermediate), 27% (high), and 61% (very high). The AUROC for the 4CM Score was 0.72 (95% CI: 0.67–0.77) for in-hospital mortality, close to COVID-GRAM, with slightly greater discriminatory ability for COVIDAnalytics: 0.76 (95% CI: 0.71–0.80). Conclusion: All the prognostic scores showed close values compared to their validation cohorts, were fairly accurate in predicting mortality, and can be used to prioritize care and resources.

Development and validation of an early warning model for hospitalized COVID-19 patients: a multi-center retrospective cohort study

Background

Timely identification of deteriorating COVID-19 patients is needed to guide changes in clinical management and admission to intensive care units (ICUs). There is significant concern that widely used Early warning scores (EWSs) underestimate illness severity in COVID-19 patients and therefore, we developed an early warning model specifically for COVID-19 patients.

Methods

We retrospectively collected electronic medical record data to extract predictors and used these to fit a random forest model. To simulate the situation in which the model would have been developed after the first and implemented during the second COVID-19 ‘wave’ in the Netherlands, we performed a temporal validation by splitting all included patients into groups admitted before and after August 1, 2020. Furthermore, we propose a method for dynamic model updating to retain model performance over time. We evaluated model discrimination and calibration, performed a decision curve analysis, and quantified the importance of predictors using SHapley Additive exPlanations values.

Results

We included 3514 COVID-19 patient admissions from six Dutch hospitals between February 2020 and May 2021, and included a total of 18 predictors for model fitting. The model showed a higher discriminative performance in terms of partial area under the receiver operating characteristic curve (0.82 [0.80–0.84]) compared to the National early warning score (0.72 [0.69–0.74]) and the Modified early warning score (0.67 [0.65–0.69]), a greater net benefit over a range of clinically relevant model thresholds, and relatively good calibration (intercept = 0.03 [− 0.09 to 0.14], slope = 0.79 [0.73–0.86]).

Conclusions

This study shows the potential benefit of moving from early warning models for the general inpatient population to models for specific patient groups. Further (independent) validation of the model is needed.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40635-022-00465-4.

Development and validation of a dynamic 48-hour in-hospital mortality risk stratification for COVID-19 in a UK teaching hospital: a retrospective cohort study

OBJECTIVES

To develop a disease stratification model for COVID-19 that updates according to changes in a patient's condition while in hospital to facilitate patient management and resource allocation.

DESIGN

In this retrospective cohort study, we adopted a landmarking approach to dynamic prediction of all-cause in-hospital mortality over the next 48 hours. We accounted for informative predictor missingness and selected predictors using penalised regression.

SETTING

All data used in this study were obtained from a single UK teaching hospital.

PARTICIPANTS

We developed the model using 473 consecutive patients with COVID-19 presenting to a UK hospital between 1 March 2020 and 12 September 2020; and temporally validated using data on 1119 patients presenting between 13 September 2020 and 17 March 2021.

PRIMARY AND SECONDARY OUTCOME MEASURES

The primary outcome is all-cause in-hospital mortality within 48 hours of the prediction time. We accounted for the competing risks of discharge from hospital alive and transfer to a tertiary intensive care unit for extracorporeal membrane oxygenation.

RESULTS

Our final model includes age, Clinical Frailty Scale score, heart rate, respiratory rate, oxygen saturation/fractional inspired oxygen ratio, white cell count, presence of acidosis (pH <7.35) and interleukin-6. Internal validation achieved an area under the receiver operating characteristic (AUROC) of 0.90 (95% CI 0.87 to 0.93) and temporal validation gave an AUROC of 0.86 (95% CI 0.83 to 0.88).

CONCLUSIONS

Our model incorporates both static risk factors (eg, age) and evolving clinical and laboratory data, to provide a dynamic risk prediction model that adapts to both sudden and gradual changes in an individual patient's clinical condition. On successful external validation, the model has the potential to be a powerful clinical risk assessment tool.

TRIAL REGISTRATION

The study is registered as 'researchregistry5464' on the Research Registry (www.researchregistry.com).