If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Dept of Clinical Radiology, NHS Salisbury Foundation Trust, Wiltshire, United KingdomDept of Clinical Imaging, Al Wakra hospital, Hamad Medical Corporation, Qatar
Chest radiograph severity score (obtained by assigning 3 zones to each lung and the severity assessed based on presence or absence of an opacity) is simple to use and has excellent inter-reader reproducibility.
•
Baseline chest radiograph severity score assessed at the time of disease presentation correlates with clinical outcome in patients with COVID-19.
•
Chest radiograph severity score of 3 or more was independent predictor of hypoxemia and primary outcome.
Abstract
Objective
To measure the reliability and reproducibility of a chest radiograph severity score (CSS) in prognosticating patient's severity of disease and outcomes at the time of disease presentation in the emergency department (ED) with coronavirus disease 2019 (COVID-19).
Materials and methods
We retrospectively studied 1275 consecutive RT-PCR confirmed COVID-19 adult patients presenting to ED from March 2020 through June 2020. Chest radiograph severity score was assessed for each patient by two blinded radiologists. Clinical and laboratory parameters were collected. The rate of admission to intensive care unit, mechanical ventilation or death up to 60 days after the baseline chest radiograph were collected. Primary outcome was defined as occurrence of ICU admission or death. Multivariate logistic regression was performed to evaluate the relationship between clinical parameters, chest radiograph severity score, and primary outcome.
Results
CSS of 3 or more was associated with ICU admission (78 % sensitivity; 73.1 % specificity; area under curve 0.81). CSS and pre-existing diabetes were independent predictors of primary outcome (odds ratio, 7; 95 % CI: 3.87, 11.73; p < 0.001 & odds ratio, 2; 95 % CI: 1–3.4, p 0.02 respectively). No significant difference in primary outcome was observed for those with history of hypertension, asthma, chronic kidney disease or coronary artery disease.
Conclusion
Semi-quantitative assessment of CSS at the time of disease presentation in the ED predicted outcomes in adults of all age with COVID-19.
Coronavirus disease 2019 (COVID-19) is an infectious disease caused by a novel strain of severe acute respiratory syndrome corona virus-2 (SARS-COV-2).
Due to a higher viral reproduction number, infectious rate and mutant nature of the viral strains, the disease transmitted rapidly out of China. The World health organization (WHO) had to eventually declare this a global pandemic.
Accurate stratification of disease severity at the time of disease presentation by clinical assessment and with chest radiograph is paramount in accurate allocation of resources necessary for patient care. Chest radiograph is routinely performed in the emergency department (ED) for patients presenting with symptoms of upper or lower respiratory tract infection. Recent Cochrane meta-analysis concluded that chest radiograph is moderately sensitive and moderately specific for diagnosis of COVID-19.
Among the various methods that exists for assessing the severity of pneumonia in COVID-19, chest radiograph severity score is considered simple and reliable.
Although COVID-19 has a high incidence of morbidity and mortality in older patients, it was unclear if an older age group would affect the severity score in predicting hypoxemia, requirement for intensive care unit (ICU) admission or mechanical ventilation. In this scenario, it is important to assess the diagnostic and prognostic reproducibility of admission chest radiograph severity score in all adult patients presenting to the ED.
The purpose of this study was therefore to semi quantitatively assess the severity of COVID-19 pneumonia on chest radiograph at the time of disease presentation to the ED, across all adult age groups, and evaluate its prognostic role in planning patient management.
2. Materials and methods
This retrospective study was approved by the institution review board (MRC-01-20-1176) that waived the need for informed consent in all patients. The study was compliant with Standards for Reporting Diagnostic Accuracy initiative (STARD) recommendations.
By using picture archiving and communication system (PACS) search, radiology information system data was extracted from all chest radiograph examinations performed at our institution during the period from March 2020 through June 2020. There was a total of 8792 patients in whom a chest radiograph was performed in our ED during the study period. We identified 1275 consecutive adult patients with an ED encounter who met the inclusion criteria ((a) Reverse transcriptase-polymerase chain reaction (RT-PCR) pharyngeal swab confirmed for COVID-19 and (b) chest radiograph at the time of disease presentation in ED). We excluded patients, (1) with a known pre-existing lung disease like emphysema, tuberculosis, sarcoidosis, primary or secondary malignancies that would confound chest radiograph interpretation, (2) technical insufficient chest radiograph scan due to motion artifact, poor inspiration and rotated images that are uninterpretable, (3) patients in whom chest radiograph was not performed at the time of ED encounter, and (4) pediatric patient population (<18 years of age) (Fig. 1). All COVID-19 positive patients in our study sample had clinical and laboratory diagnosis, with at least one RT-PCR positive result for COVID-19.
Fig. 1Flow diagram of patient inclusion and exclusion in the present study. COVID-19 = coronavirus disease 2019, CXR-chest radiograph, MRN = medical record number, PACS=picture archiving and communication system, RBG-admission random blood glucose, RT-PCR – reverse transcription polymerase chain reaction.
2.2 Study variables, outcome and co-morbidity definitions
We recorded the information related to demographics (age, sex), co-morbidities (hypertension, diabetes mellitus, coronary artery disease, chronic kidney disease, asthma, human immunodeficiency virus and dyslipidemia), symptoms at the time of hospital presentation (fever, cough, malaise, headache, sore throat, running nose, dyspnea, diarrhea, vomiting, malaise), laboratory parameters (White blood cell count, C reactive protein, lactate dehydrogenase, HbA1C, random blood glucose), disease severity requirements (oxygen, ICU admission, non-invasive or invasive ventilation) and the outcome (discharged alive or death).
Pre-existing co-morbidity including diabetes was defined based on self-reporting by the patient, from the Hospital Information System (HIS) database or from ongoing medications. Primary outcome was defined as occurrence of admission to ICU (this includes all cases needing non-invasive or invasive mechanical ventilation) or death (scores of 5, 6, 7 on World Health Organization blueprint scale).
The mild disease group includes patients having mild or moderate symptoms and may or may not requiring supplemental nasal oxygen. The severe disease group includes patients that require ICU care due to deteriorating clinical status (with at least one or more of, (a) respiratory rate >30/min; (b) ratio of partial pressure of oxygen to fraction of inspired oxygen PaO2/FiO2 <300 mm Hg; (c) resting oxygen saturation <94 % on room air at sea level; (d) non-invasive or invasive mechanical ventilator requirements). The clinical disease group classification is summarized in Table 1.
Table 1Group classification based on clinical severity in COVID-19.
Clinically mild disease group
Clinically severe disease
Symptoms
Mild to moderate clinical symptoms
Severe to critical symptoms
Oxygen requirements
May or may not require oxygen
Requires continuous oxygen
Criteria
RR-Normal PaO2/FiO2 > 300 mm Hg SpO2 > 94 % on room air at sea level
RR > 30/min PaO2/FiO2 < 300 mm Hg SpO2 < 94 % on room air at sea level Non-invasive or invasive ventilator requirements
ICU admission
Not required
Required
ICU-Intensive care unit; PaO2/FiO2-ratio of arterial partial pressure of oxygen to fraction of inspired oxygen; RR-respiratory rate; SpO2-resting oxygen saturation.
We adopted the same scoring system in dividing each lung into three zones. The upper zone extends between lung apex and superior hilar markings, the middle zone extends between superior hila and inferior hila, the lower zone extends between inferior hila and the costophrenic angle. Each of the six zones in both lung fields was scored on the presence (score 1) or absence (score 0) of opacities like infiltrates, ground glassing or consolidation. The scores of each of the zones were then summed for a total score that ranges from 0 to 6 (Fig. 2).
Fig. 2(a) PA chest radiograph of a 27-year-old man with an unremarkable previous medical history, presented with four days history of fever and cough. Chest radiograph shows ill-defined opacity in left mid-zone with a total CSS of 1; (b) PA chest radiograph of a 37-year-old man with no significant past medical history, presented with three days history of fever, cough, and headache. Chest radiograph shows ill-defined opacities in bilateral mid and lower zones with a total CSS of 4; (c) PA chest radiograph of a 60-year-old man with known diabetes mellitus, presented with three days history of fever, cough, and shortness of breath followed by hospital admission after a positive COVID-19 RT-PCR test who subsequently required nasal oxygen during the hospital stay. Chest radiograph showed diffuse opacities in bilateral upper, mid, and lower zones with a total CSS of 6.
2.5 Chest radiograph severity score data collection
Chest radiograph image data were scored for CSS by 2 radiologists (1 & 2; H.S. and A.K. with 10 and 11 years of experience respectively) independent of each other. Two reviewers were blinded to disease severity and outcome. All patients in our study underwent either postero-anterior (PA) chest radiograph (1185 of 1275, 92.9 %) or antero-posterior (AP) chest radiograph (90 of 1275, 7.1 %). For statistical analysis, the overall CSS was taken randomly between the two readers.
2.6 Statistical analysis
All quantitative data presented as mean ± standard deviation (SD) or as median (inter-quartile range), while qualitative data presented in numbers and percentages (%). Mann-Whitney U test was applied to test the statistical significance of the difference in the median of continuous variables between two groups. The statistical significance of the difference in the proportion of categorical variables between groups was compared using the Chi-Square test. Cohens' K
coefficient was used to assess the agreement in chest radiograph severity score between the two chest radiologists. Multivariate logistic regression analysis was performed to explore the association of variables by estimating the unadjusted odds ratio for categorical outcome. The receiver operating curve analysis (ROC) was used to determine the area under curve, sensitivity, and specificity for the concordant score in relation to the outcome. Additionally, Youdens J statistic was performed to determine an appropriate cut-off score relevant to the outcome. Probability value (p-value) <0.05 was considered statistically significant and statistical analysis was done using the IBM SPSS V 20.0 software (IBM SPSS, USA).
3. Results
3.1 Patient overview and general characteristics
A total of 1275 adult patients positive for COVID-19 were included in the study (median age of 40 years, IQR 33 to 48, age range 18 to 96 years). The median duration of symptom onset to disease presentation at the ED was 4 days (IQR, 2–7). 272 patients (21.3 %) had one or more pre-existing co-morbidities. Hypertension was the most common co-morbidity in 182 cases (14.2 %) followed by diabetes mellitus type-2 in 175 cases (13.7 %). The most common clinical symptom at the time of disease presentation to the ED was fever 951 cases (74.5 %), followed by cough in 817 patients (64 %). A total of 89 (6.9 %) patients were admitted to ICU, 50 (3.9 %) required non-invasive or invasive mechanical ventilation, 5 (0.4 %) required extracorporeal membrane oxygenation (ECMO). 1259 (98.7 %) patients were discharged and 16 (1.3 %) died during their stay in the hospital. Patient demographics and basic summary based on clinical disease severity in Table 2.
Table 2Patient demographics and characteristics of SARS-COV-2 infection based on clinical disease severity. Data represented as number (percentage), mean (±SD) or median (interquartile range). In the column representing data related to all patients, the data that was available for statistical analysis is shown. CSS-chest radiograph severity score, ED-emergency department, LDG-lactate dehydrogenase, RBG-admission random blood glucose.
Based on clinical severity, the median age of mild disease group was 39 (IQR of 32–47) while in severe group the median age was 45 (IQR of 40–55), p < 0.001. The median chest radiograph severity score in clinically mild group was 1, and 4 in clinically severe group, p < 0.001.
Chest radiograph was scored by two radiologists with excellent inter-observer agreement (Cohens' K 0.95). The median CSS was 1 (IQR, 0 to 3). No patient had pneumothorax or pleural effusion.
3.2 Association with clinical outcomes
In ROC curve analysis, CSS of 3 or more (78 % sensitivity and 73.1 % specificity; AUC 0.81; Fig. 3) was the better cut off associated with primary outcome. Patients who died had higher CSS, however there were not enough cases to demonstrate statistical significance (n = 16).
Fig. 3Receiver operating characteristics curve of CSS model performance with primary outcome. CSS-chest X-ray severity score.
Multivariate regression analysis was performed to assess factors associated with a primary outcome. CSS of 3 or more and pre-existing diabetes were found to be independent predictors of primary outcome (Odds ratio = 7; 95 % CI: 3.87–11.73, p < 0.001 & OR = 2; 95 % CI: 1–3.4, p 0.02 respectively, Table 3). On multivariate regression, after adjusting for age the CSS remained associated with primary outcome.
Table 3Multivariate regression analysis showing factors tested for association with clinical severity in COVID-19.
3.3 Secondary outcome data between CSS severity groups
Based on CSS cut off score of 3 by ROC curve analysis, the study sample was divided into mild CSS (<3) and severe CSS group (≥3) (Table 4). The median age was 38 (IQR 31–45) in mild and 45 (IQR of 38–52) in severe CSS, p < 0.001. Between the groups, statistical significance was observed on the need for oxygen, ICU, ventilator and ECMO, p < 0.05. In mild CSS group there was 1 death (0.1 %), while in severe CSS there were 15 (3.9 %) deaths, p < 0.001.
Table 4Secondary outcome data between comparison groups.
This study evaluated the prognostic value of initial chest radiograph in evaluating adult patients presenting to emergency department with COVID-19. Patients with worse COVID-19 pneumonia quantified by the chest radiograph severity score were more likely to have hypoxemia, and need ICU admission or mechanical ventilation. The chest radiograph severity score remained associated with worse outcomes in COVID-19 even after adjustment for age, indicating the usefulness of chest radiograph severity score across adults of all age.
The association between baseline semi quantitative radiographic scoring systems in COVID-19 and worse outcome has been consistently demonstrated by several studies in the past.
The association of chest radiographic findings and severity scoring with clinical outcomes in patients with COVID-19 presenting to the emergency department of a tertiary care hospital in Pakistan.
Our study is the largest study to date utilizing the CSS, which has been previously shown to be reliable for predicting the outcome in young and middle-aged patients <50 years of age.
Our study focused on using CSS, a simple and reliable scoring system when compared to other widely studied scoring systems like Radiographic assessment of lung edema (RALE/modified RALE) and Brixia.
The prognostic capacity of the radiographic assessment for lung edema score in patients with COVID-19 acute respiratory distress syndrome—an international multicenter observational study.
The latter scoring systems are more complex to apply in clinical practice as it involves assessing the nature and intensity of pneumonia and the inclusion of zonal scoring.
This could particularly be more challenging given the existing pressure on radiology workload and demanding clinical scenario leading to risks of burnouts.
Although baseline chest radiograph severity score was shown to be of predictive value in assessing worse outcome, the original study did not explore the performance of the scoring system in age groups of patients >50 years old.
In our study, despite adjusting for age, the performance of chest radiograph severity scores remained associated with primary outcome. Furthermore, many studies have assessed the extent and severity of lung involvement at initial disease presentation for predicting outcome. A study by Kaleemi et al used a chest X-ray severity score that was modified from RALE score, and found that increased initial chest X-ray severity score was associated with ICU admission and mortality,
The association of chest radiographic findings and severity scoring with clinical outcomes in patients with COVID-19 presenting to the emergency department of a tertiary care hospital in Pakistan.
In our study, patients with involvement of at least three or more lung zones were more likely to require mechanical ventilation or die and this finding was consistent with observations from prior reports.
The association of chest radiographic findings and severity scoring with clinical outcomes in patients with COVID-19 presenting to the emergency department of a tertiary care hospital in Pakistan.
The presence of underlying comorbidities can be reasonably suggested as drivers for worse outcome. The study from Center for Disease Control and Prevention's analysis of adults <49 years of age demonstrated that underlying co-morbidities like hypertension, diabetes, obesity, asthma and immunosuppressive disease are precursors for worse outcomes in COVID-19.
Hospitalization rates and characteristics of patients hospitalized with laboratory-confirmed coronavirus disease 2019 - COVID-NET, 14 states, march 1–30, 2020.
In our study, although hypertension was the most common comorbidity, there was no significant difference on the primary outcome, and our finding was somewhat incongruent with other reports that noted hypertension is associated with worse prognosis in COVID-19.
Hospitalization rates and characteristics of patients hospitalized with laboratory-confirmed coronavirus disease 2019 - COVID-NET, 14 states, march 1–30, 2020.
However, pre-existing diabetes was an independent factor associated with outcomes of our interest, and this observation was consistent with other reports.
Hospitalization rates and characteristics of patients hospitalized with laboratory-confirmed coronavirus disease 2019 - COVID-NET, 14 states, march 1–30, 2020.
Our study has several limitations. First, this study was retrospective and may introduce observer bias in assessing the disease outcome. Second, there has been a male gender bias in our study, as our hospital was designated for male in-patient admission, and there was a separate in-patient admission facility for females elsewhere. However, recent studies have shown there was no significant difference between the gender leading to morbidity or mortality in COVID-19.
the collected data do not have information on BMI. Fourth, the collected data do not have information on the vital signs at the time of admission, number of days in ICU and number of days requiring ventilation. Fifth, on multivariate regression analysis the preexisting diabetes and random blood glucose were associated with primary outcome, however due to possible collinearity between RBG and pre-existing diabetes, and the variability of RBG related to steroid or diabetic medications, our focus was mainly on pre-existing diabetes in furthering the discussion. Sixth, while the overall cohort of patients was higher in number than previous studies the number of deaths was not large enough to assess if CSS or any of the comorbidities was statistically significant in predicting increased risk of death.
In the present state, it is difficult to infer the intrinsic severity of virus variants due to difference in population immunity and vaccination status.
Despite the waves, variants and breakthrough infections, chest radiograph is likely to remain a simple valuable tool to prognosticate outcome in patients with COVID-19. Overall, our results may enable clinicians to predict worse outcomes by assessing the number of lung zones involved through the use of CSS along with clinical parameters and to appropriately triage adult patients presenting to the ED with COVID-19.
Funding
Authors have not received any funding for this study.
Credit authorship contribution statement
Guarantor of integrity of study A.N, D.K. Study concept, study design, manuscript drafting A.N. Revision for important intellectual content; editing and approval of final version of submitted manuscript; agrees to ensure any questions related to the work are appropriately resolved, all authors. Literature research, O·K, H.S, A.K, A.T, M.H, D.K., M.M, A.N. Chest radiograph severity scoring, H.S., A.K. Data manager and data collection: M.Z, B.J. Methodology and statistical analysis, V.A., D·S, A.N.
Declaration of competing interest
Authors have no disclosure or no relevant relationships.
References
Hosseiny M.
Kooraki S.
Gholamrezanezhad A.
et al.
Radiology perspective of coronavirus disease 2019 (COVID-19): lessons from severe acute respiratory syndrome and Middle East respiratory syndrome.
The association of chest radiographic findings and severity scoring with clinical outcomes in patients with COVID-19 presenting to the emergency department of a tertiary care hospital in Pakistan.
The prognostic capacity of the radiographic assessment for lung edema score in patients with COVID-19 acute respiratory distress syndrome—an international multicenter observational study.
Hospitalization rates and characteristics of patients hospitalized with laboratory-confirmed coronavirus disease 2019 - COVID-NET, 14 states, march 1–30, 2020.
00317-5&id=gr1.jpg){kind=link}
00317-5&id=gr2.jpg){kind=link}
00317-5&id=gr3.jpg){kind=link}