key: cord-1003821-1h7pp77v
authors: Peyrony, Olivier; Marbeuf-Gueye, Carole; Truong, Vy; Giroud, Marion; Rivière, Clémentine; Khenissi, Khalil; Legay, Léa; Simonetta, Marie; Elezi, Arben; Principe, Alessandra; Taboulet, Pierre; Ogereau, Carl; Tourdjman, Mathieu; Ellouze, Sami; Fontaine, Jean-Paul
title: Accuracy of Emergency Department clinical findings for diagnostic of coronavirus disease-2019
date: 2020-05-21
journal: Ann Emerg Med
DOI: 10.1016/j.annemergmed.2020.05.022
sha: deedb117738d3a304da3d260a3910f476fc247b5
doc_id: 1003821
cord_uid: 1h7pp77v

Abstract: Objective We sought to describe the medical history and clinical findings of patients attending the ED with suspected COVID-19 and estimate the diagnostic accuracy of patients’ characteristics for predicting COVID-19. Methods We prospectively enrolled all patients tested for SARS-CoV-2 by RT-PCR in our ED from March 9, 2020 to April 4, 2020. We abstracted medical history, physical examination findings, and the clinical probability of COVID-19 (“low”, “moderate”, “high”) rated by emergency physicians depending on their clinical judgment. We assessed diagnostic accuracy of these characteristics for COVID-19 by calculating positive and negative likelihood ratios (LR+/LR-). Results We included 391 patients of whom 225 tested positive for SARS-CoV-2. RT-PCR was more likely to be negative when the emergency physician thought that clinical probability was low, and more likely to be positive when she or he thought that clinical probability was high. Patient-reported anosmia and the presence of bilateral B-lines on lung ultrasound had the highest LR+ for the diagnosis of COVID-19 (7.58; 95% CI 2.36–24.36 and 7.09; 95% CI 2.77–18.12 respectively). The absence of a high clinical probability determined by the emergency physician and the absence of bilateral B-lines on lung ultrasound had the lowest LR- for the diagnosis of COVID-19 (0.33; 95% CI 0.25–0.43 and 0.26; 95% CI 0.15–0.45 respectively). Conclusions Anosmia, emergency physician estimate of high clinical probability and bilateral B-lines on lung ultrasound increased the likelihood of identifying COVID-19 in patients presenting to the ED.

Covid-19, SARS-CoV-2, Emergency Department, clinical findings, clinical probability, lung ultrasound Introduction:

The novel Coronavirus disease of 2019 (COVID-19) outbreak has led to major reorganizations of emergency departments (EDs) in order to face the dramatic increase of patients suspected to be infected by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [1] . The clinical description of hospitalized patients has been reported in the literature [2] [3] [4] [5] [6] but no study has focused on clinical and diagnostic findings in the ED setting.

Among patients attending EDs, rapid triage of those suspected of COVID-19 is mandatory to appropriately isolate them and avoid secondary transmissions. Clinical diagnosis can be challenging since the disease may present with non-specific symptoms such as myalgia, cough or fever [3, 6] . Medical history and clinical presentations of COVID-19 patients attending EDs need to be precisely described to facilitate early recognition by emergency physicians and promptly trigger diagnostic procedures such as real-time reverse transcriptase polymerase chain reaction (RT-PCR).

The objectives of this study were to collect and describe the medical history and clinical findings of patients attending the ED with suspected COVID-19, and to assess utility of clinical parameters, physician gestalt (clinical judgment) and lung ultrasound (US) to accurately identify COVID-19 patients at the time of ED presentation.

This prospective observational study was conducted in the ED of Saint-Louis University Hospital, Paris, France. Starting on March 9, 2020, we prospectively enrolled a cohort of all COVID- 19 suspected adult patients that were tested for SARS-CoV-2. This study reports the results of patients enrolled until April 4, 2020. The study was approved by the Institutional Review Board of the French Speaking Society for Respiratory Medicine -Société de Pneumologie de Langue Française (number CEPRO 2020-014). Our study followed the Strengthening of Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines [7] .

All adult patients (age ≥ 18 years) that were tested for SARS-CoV-2 in our ED were included after giving oral consent. Cases were identified and enrolled 24 hours per day and 7 days per week during the study period by the attending emergency physician or resident who was in charge of the patient. Testing for SARS-CoV-2 in COVID-19 suspected patients was left to the clinician's discretion but most of the time, patients were tested when they were dyspneic or reported shortness of breath; had comorbidities that put them at risk of severe infection such as immunosuppression, chronic respiratory insufficiency, cardiovascular diseases, and obesity; if they were over 70 years old; or if they were too weak to be discharged home. Some patients without clinical suspicion of COVID-19 but needing hospitalization in non-COVID areas were also tested. Patients under the age of 70, with no comorbidities, and with no respiratory symptoms were not tested. Prior to the outbreak, our department received approximately 110 to 120 patients per day and this number dropped to approximately 50 to 60 from the beginning of the pandemic in March. During the study period, we tested approximately 10 to 20 patients per day. If patients attended the ED more than once, only the last visit was included. There were no other exclusion criteria.

When testing for SARS-CoV-2, the attending emergency medicine attending or resident physician was asked to report in a dedicated form the patient's medical history; Eastern Cooperative Oncology Group Performance Status (ECOG PS), which is a scale that describes patients' ability to care for themselves and perform daily activities, ranging from 0 ("fully active") to 4 ("completely disabled"); physical examination; and chest X-ray and lung US findings when those were performed. Lung US was performed with a pocket-size device (VSCAN, GE Healthcare). After medical history, physical examination, US, and chest Xrays, attending physicians were asked to rate the clinical probability of COVID-19 based on both their clinical judgment and a predefined three-level scale ("low", "moderate", "high").

As anosmia was reported in Europe around the end of March, this clinical sign was added to the form on March 24, 2020.

All study data and variables with their categories were defined prior to the beginning of the study. Four residents who were previously trained in data abstraction completed the forms using a dedicated spreadsheet. Age, gender, vital signs at ED arrival and any data that was missing in the forms were abstracted retrospectively from the patients' ED medical files with the exception of clinical probability that could only be determined prospectively. An emergency physician with expertise in research periodically monitored data abstraction.

When there was disagreement between abstractors or if data was ambiguous, this emergency physician, made the final decision. No assessment of inter-rater reliability was performed.

The gold standard for diagnosis was the result of SARS-CoV-2 RT-PCR via nasal swab (Cobas® SARS-CoV-2 Test, Roche, Meylan, France). The patients who initially had a negative RT-PCR in the ED but were tested positive in the next 48 hours were considered positive (initial false negative).

Continuous variables were reported as medians with their interquartile ranges (IQR). To assess the performances of each characteristic to accurately identify COVID-19, we calculated the sensitivity (Se), specificity (Sp), positive predictive value (PPV), negative predictive value (NPV), positive and negative likelihood ratios (LRs) and their 95% confidence intervals (CI). We also calculated PPVs and NPVs for each characteristic under a variety of prevalence assumptions (5%, 10%, 25%, 75%). We calculated post-test probabilities depending on both pre-test probabilities (5%, 10%, 25%, 50%, 75%) and the presence or absence of bilateral B-lines on lung US. Accuracy of the physician clinical probability in identifying patients with COVID-19 was assessed with a receiver operating characteristics (ROC) curve and by calculating the area under the curve (AUC) with its 95% CI. Data were analyzed with R 3.5.0 software (the R Foundation for Statistical Computing, Vienna, Austria).

During the study period, 400 patients were tested for SARS-CoV-2. After excluding nine patients who were tested two times during two ED visits, we included 391 patients. Among those patients, 225 (57.6%) were tested positive for SARS-CoV-2 (including five initial false negatives). Median age was 62 [48-71] years and 150 (38.4%) were females. General characteristics of these COVID-19 suspected patients are presented in Table 1 . Among patients with confirmed COVID-19, 67 (29.8%) were discharged home from the ED, 134 (59.5%) were hospitalized in wards, 22 (9.8%) were admitted to the ICU and 2 (0.9%) died in the ED.

Patient-reported symptoms, physical examination, lung ultrasound and chest X-ray findings in patients with or without COVID-19 are summarized in Table 2 Emergency physicians rated the clinical probability for 273 patients. Table 3 shows the proportion of patients with or without COVID-19 depending on the emergency physician's clinical probability. RT-PCR was more likely to be negative for SARS-CoV-2 when the emergency physician thought that clinical probability was low, and more likely to be positive when she or he thought that clinical probability was high. Figure 1 prevalence) and the results of lung US.

Our study has several limitations. First, not every patient was tested for SARS-CoV-2 and testing was left to the clinician's discretion. However, despite the absence of clear predefined inclusion criteria, testing was performed in the vast majority of cases in patients who had severe symptoms such as dyspnea, reported shortness of breath, presented with comorbidities (e.g., immunosuppression or chronic pulmonary or cardiovascular diseases), or were over 70 years old. Thus, our results may not be valid in other populations such as young people without comorbidities and those with very few symptoms. Also, as our results are issued from a single center in France, they may not be generalizable to other centers. Second, it is possible that patients did not systematically report their symptoms, and this might have lowered the estimates of their prevalence. Third, we may have underestimated anosmia because we only added this to the form on March 24, 2020. Fourth, lung ultrasound was not systematic and occurred in only 22.3% of the patients, which contributed to the large 95% CI observed. It is possible that emergency physicians performed lung US in the most severe cases and that the accuracy of this exam in predicting COVID-19 is likely to be lower in patients with fewer symptoms. Lastly, our gold-standard to diagnose COVID-19 was based on the RT-PCR which may have false-negatives. To address this, we also evaluated patients who were initially tested negative for SARS-CoV-2 but were then hospitalized and secondarily tested positive and considered them as positives.

To the best of our knowledge, this is the first prospective study that both described patientreported symptoms and physical examination findings in a large cohort of COVID-19 suspected patients attending the ED and estimated the accuracy of clinical findings for the diagnosis of COVID-19.

At the beginning of the epidemic in France, knowledge of the COVID-19 clinical picture was mainly extrapolated from cases in Wuhan, China. Emergency physicians prepared to face an unknown disease and attend to patients with non-specific influenza-like symptoms with clinical signs of lung infection, or acute respiratory failure for the most severe cases [3, 6] .

Since then, studies have been published about the epidemiology [1] , the risk factors for severe disease [4] and the description of critically ill patients [8] [9] [10] [11] infected with SARS-CoV-2, but there is limited data regarding the accuracy of clinical findings in patients with suspected COVID-19. Moreover, most studies have focused on patients who were already hospitalized for the disease, which may differ from the ED population.

Prompt identification of possible cases is mandatory to avoid the spread of the virus by patients with mild or non-specific symptoms [12] . Therefore, emergency physicians need to be cautious when they evaluate such patients and be aware of some pitfalls. Whereas fever (>37.3 °C) was reported in more than 90% of the patients hospitalized with COVID-19 [4, 6] , we found that, even if 78.2% of the patients reported fever, only 34.4% of the patients had ≥ 38 °C and 56.7% had ≥ 37.5 °C at ED triage. It is possible that the temperature was initially lowered by antipyretics and subsequently rose during the ED stay. Nevertheless, temperature should not be used in isolation to exclude COVID-19 [13] . Consistent with other studies, we found that the most frequent reported symptoms were fever, cough, dyspnea and myalgia [2] [3] [4] [5] [6] . Gastrointestinal symptoms were present in 23.6% of our patients, whereas one small study of 18 patients found a rate of 17% [14] , while larger cohorts have reported these symptoms in less than 10% [2] [3] [4] [5] [6] . Anosmia was reported by 13.8% of the patients in our cohort and was the most specific sign of COVID-19. It is likely that we underestimated this sign, which was not initially described in the Chinese literature and was reported in Europe around the end of March [15] [16] [17] .

Other findings like bilateral crackles on lung auscultation or the rapid need for high levels of oxygen delivery at ED arrival were highly suggestive of COVID-19, especially among middle-aged or older patients with comorbidities such as diabetes or cardiovascular diseases [18] . As this disease induces endotheliitis, leading to vascular derangements [19] , it is very likely that new symptoms involving multiple organs such as neurological [20] or skin disorders [21] will emerge in the future.

Besides clinical signs, lung imaging and particularly CT-scans have shown to have a high sensitivity for the diagnosis of COVID-19, particularly in severe cases, and may be valuable in patients with high clinical probability but negative RT-PCR [22] [23] [24] . Another option might be to perform lung ultrasound, allowing rapid diagnosis and severity assessment at patient bedside for suspected COVID-19 [25, 26] . In our study, the presence or absence of bilateral B-lines with a pocket-size ultrasound device had the higher LR+ and a low LR-, respectively.

In a study with 20 patients with COVID-19, Xing et al found that all patients showed pleural line abnormalities, and bilateral B-lines on lung US, regardless of the severity or stage of the disease [27] . In another retrospective study that included 30 patients with COVID-19, interstitial pulmonary edema was present on lung US in 90.0% of the cases [28] . Nonetheless, literature on this topic is scarce and more data is needed [29] .

In the present study, we found that emergency physician clinical judgment was quite accurate and that only 19. 

Epidemiology Based on Current Evidence

Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

Risk Factors Associated With Acute Respiratory Distress Syndrome and Death in Patients With Coronavirus Disease

Clinical findings in a group of patients infected with the 2019 novel coronavirus (SARS-Cov-2) outside of Wuhan, China: retrospective case series

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration

Baseline Characteristics and Outcomes of 1591 Patients Infected With SARS-CoV-2 Admitted to ICUs of the Lombardy Region

Clinical course and outcomes of critically ill patients with SARS-CoV-2 pneumonia in Wuhan, China: a single-centered, retrospective, observational study

Covid-19 in Critically Ill Patients in the Seattle Region -Case Series

Covid-19 in South Korea -Challenges of Subclinical Manifestations

Updating the diagnostic criteria of COVID-19 "suspected case" and "confirmed case" is necessary

Epidemiologic features and clinical course of patients infected with SARS-CoV-2 in Singapore

Isolated sudden onset anosmia in COVID-19 infection. A novel syndrome?

Anosmia and ageusia: common findings in COVID-19 patients

Olfactory and gustatory dysfunctions as a clinical presentation of mild-to-moderate forms of the coronavirus disease (COVID-19): a multicenter European study

Endothelial cell infection and endotheliitis in COVID-19

Neurologic Manifestations of Hospitalized Patients With Coronavirus Disease

Diagnostic performance between CT and initial real-time RT-PCR for clinically

Correlation of Chest CT and RT-PCR Testing in Coronavirus Disease 2019 (COVID-19) in China: A Report of 1014 Cases

Our Italian Experience Using Lung Ultrasound for Identification, Grading and Serial Follow-up of Severity of Lung Involvement for Management of Patients with COVID-19

Proposal for international standardization of the use of lung ultrasound for COVID-19 patients; a simple, quantitative, reproducible method

Lung ultrasound findings in patients with COVID-19 pneumonia

A Clinical Study of Noninvasive Assessment of Lung Lesions in Patients with Coronavirus Disease-19 (COVID-19) by Bedside Ultrasound

2020; in press 29 -MJ Fiala. A Brief Review of Lung Ultrasound in COVID-19: Is it Useful?

The authors wanted to thank Pierre Bourrier for his precious learnings about echostethoscopy and the medical emergency team of Saint-Louis hospital for their active participation in this study.

The study was approved by the Institutional Review Board of the French Speaking society for respiratory medicine -Société de Pneumologie de Langue Française (number CEPRO 2020-014).

The authors wish to thank Pierre Bourrier for his valuable teachings about ultrasound and the medical emergency team of Saint-Louis hospital for their active participation in this study.