key: cord-0715081-djpg6jx3
authors: Shiel, Emily; Miyakis, Spiros; Tennant, Elaine; Fernando, Shelanah; Kizny‐Gordon, Alice; Koh, Bryant; Findlay, Michael; Garnham, Katherine; Pilli, Shravya; Segboer, Hayden; Tallon, Jo; Kao, Joanna; Miller, Anne; Shore, Tim; Maher, Richard; Telford, Mark; Barclay, Kate; Harris, Ben; Newcombe, James; Hudson, Bernie; Figtree, Mel
title: Clinical Characteristics and Outcomes of COVID‐19 in a low‐prevalence, well‐resourced setting, Sydney, Australia
date: 2021-07-06
journal: Intern Med J
DOI: 10.1111/imj.15445
sha: 86f35b10082d78bcd5d264fb61f7f222307cc0f2
doc_id: 715081
cord_uid: djpg6jx3

BACKGROUND: The Northern Sydney Local Health District was one of the first health regions to be affected by COVID‐19 in Australia. AIM: To describe the clinical characteristics, risk factors and outcomes in our low‐prevalence Australian population. METHODS: A retrospective analysis of 517 laboratory‐confirmed COVID‐19 cases between January and June 2020. Patient information was collected as part of routine care within the COVID‐19 Virtual Hospital system. Outcomes examined were death, recovery at 30 days and intensive care unit (ICU) admission. RESULTS: The case fatality rate was 1.8%. Multivariate analysis showed factors independently associated with death, composite outcome of death/ICU admission or incomplete recovery at 30 days were age > 80 and presence of 2 or more comorbidities. Most cases acquired COVID‐19 through international (50.9%) or cruise ship travel (9.1%). Health care workers comprised 12.8% of the cohort and represented a disproportionately high percentage of the “unknown” source group (27.6%). The median incubation period was 5 days (IQR 3–8); one patient had an incubation period of 15 days. Hospitalisation was required in 11.8%, ICU admission in 2.1% and ventilation in 1.4%. A RALE score on CXR of >10 was independently associated with death. CONCLUSIONS: In this low prevalence, well‐resourced Australian setting, we report an overall low mortality. Factors associated with adverse patient outcomes on multivariate analysis were age greater than 80 and the presence of 2 or more comorbidities. This data can assist in early risk stratification of COVID‐19 patients, and in surge capacity planning for hospitals. This article is protected by copyright. All rights reserved.

Since the coronavirus disease 2019 pandemic, caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was officially declared on March 11 th , there have been 174 million confirmed infections and over 3.75 million deaths worldwide.(1) Australia's first COVID-19 case was identified on the 25 th January 2020, and there have since been over 30,000 confirmed cases and 910 deaths nationwide. (2) The Northern Sydney Local Health District (NSLHD), a health region in New South Wales (NSW), provides health services to a population of approximately 950,000 people. It experienced one of the earliest surges of COVID-19 cases in Australia, and most patients were managed through the NSLHD COVID-19 Virtual Hospital as outpatients.

Several studies globally report on clinical characteristics of COVID-19, with the most common symptoms of fever, cough, fatigue or myalgia. (3) (4) (5) (6) (7) (8) We are also seeing more varied symptoms, including gastrointestinal symptoms, (8) anosmia and dysgeusia (9) in high numbers of COVID-19 patients. Comorbidities such as obesity, hypertension, diabetes, cardiovascular disease, and older age are associated with more severe disease. (3, 6, 8, 10) We seek to characterise an Australian laboratory-confirmed COVID-19 population in a low prevalence setting, with particular focus on patient symptoms, comorbidities and outcomes. This information will help to identify high-risk COVID-19 patients, thereby assisting with appropriate triage and management of patients, and informing surge capacity planning.

Study approval was granted by the NSLHD Human Research and Ethics Committee and registered with the NSLHD Quality Unit and the Clinical Governance Unit, as a quality assurance project.

The Communicable Diseases Network Australia (CDNA) National Guidelines for Public Health Units (11) and emerging local epidemiological patterns were used to establish NSLHD COVID-19 clinics. Nasopharyngeal swabs were collected on symptomatic patients or those with epidemiological risk factors. Diagnosis of COVID-19 was made through detection of SARs-CoV-2 on PCR of respiratory tract specimens, using the AusDiagnostics (Mascot, Australia) SARS-CoV-2 multiplex-tandem polymerase chain reaction assay in the NSLHD public laboratory. Swabs processed in private laboratories were tested using a variety of other commercially available or in-house methods.

Positive COVID-19 cases were either admitted via the Emergency Department (ED) or referred to the Infectious Diseases and Microbiology team at Royal North Shore Hospital, either directly by the laboratory or via the Public Health Unit (PHU). Patients who did not require admission were monitored by the Virtual Hospital. Contact identification and classification were performed according to the CDNA guidelines. (11) The NSLHD COVID-19 Virtual Hospital team, made up of doctors and nurses, contacted each patient within 24 hours of positive COVID-19 diagnosis. Initial assessment and risk stratification were based on World Health Organisation (WHO) and CDNA guidelines and known risk factors for severe disease including age, sex, comorbidities and symptoms. (3, 6, 8, (10) (11) (12) Patients had regular follow-up phone consultations and were referred to hospital if there were clinical concerns based on underlying risk factors, age, and clinical progress. The ED assessment then determined whether the patient required admission on clinical grounds. Patients who were unable to quarantine in their home were admitted to specialised quarantine hotels. When hospital was required, patients were managed across the four hospitals within the health district. The Virtual Hospital operated 7 days a week, from 0800-2200. After hours, patients were advised to contact their local ED.

Radiology and laboratory investigations were conducted according to each patient's clinical need.

All laboratory-confirmed COVID-19 patients within the NSLHD from January to June 2020 were retrospectively included in the study, if they had 30-day follow-up available. Patients were excluded if there was insufficient data available for analysis or if they were quarantined outside our district.

The electronic medical record for each patient, including Emergency Department assessments, COVID-19 clinic notes, inpatient admission and follow-up notes from the NSLHD COVID-19 Virtual Hospital team were reviewed retrospectively. Using a standardised form, data was collected on patient demographics, epidemiological risk, co-morbidities, clinical symptoms or signs, laboratory and radiological findings, need for hospitalisation or intensive care, and patient outcome (Appendix A). Body mass index (BMI) was assessed using standard classification.(13) Severity classification of COVID-19 was assessed using the World Health Organisation COVID-19 severity definition.(12)

The incubation period was defined as the interval between the earliest date of contact from a proven COVID-19 contact and the date of symptom onset. (11) This was only included if the date of exposure and source was clear. Most likely acquisition source was determined by the investigators through identification of patients' epidemiological risk factors. Close and casual contacts were defined as per CDNA guidelines. (11) Transmission was classified as "unknown" if a specific source could not be identified. Fever was defined as an axillary or forehead scan temperature of 37.5 o C or higher.(11) Comorbidities including immunodeficiency were defined according to the CDNA guidelines. (11) Study investigators contacted the patient at or after day 30 to determine recovery and to clarify information if required. Outcomes of the study were death, recovery at 30 days and the composite outcome of death or ICU admission. Patients were admitted to ICU based on clinical need. In addition, some patients with poor pre-morbid status were not deemed appropriate for ICU. Recovery was defined as the resolution of all symptoms and self-reported patient estimate of returning to 90% or above their premorbid function, when asked as binary questions.

Retrospective analysis of chest imaging was performed by two thoracic radiologists independently, involving a third radiologist when consensus was not achieved. Radiologists were blinded to the patients' outcome and to the original report. The Radiographic Assessment of Lung Edema (RALE) score was used to determine the extent of COVID-19 pulmonary involvement. (14) Statistics Data was analysed using SPSS for Windows, v. 26.0. (Armonk, NY: IBM Corp). For categorical variables, Pearson's Chi Square or Fisher's exact test was used; for continuous variables, Student's ttest or one-way analysis of variance with Bonferroni correction. 2-tailed tests of significance and p values <0.05 were considered. Univariate analysis for death, incomplete recovery at 30 days and composite end-point (death or ICU) identified significant factors which were then selected for inclusion in a logistic regression model, using a backward stepwise selection algorithm with entry and removal criteria of p≤0.05 and p≤0.1, respectively.

There were 520 patients in our original cohort, 3 were excluded due to insufficient data, leaving 517 patients for analysis. 49.9% of our cohort were female, with a median age of 48 years (range 3 months to 95 years) ( Table 1) . 265 patients (51.3%) acquired COVID-19 through international travel, while an additional 47 patients (9.1%) were infected following recent travel on a cruise ship. There were 108 cases (20.9%) of community acquired infection, of which, 83 (16.1%) were identified as close contacts, 22 (4.3%) as casual contacts and 3 (0.6%) that had no identified confirmed COVID-19 contact.

66 patients (12.8%) were HCWs. The main acquisition source for HCWs was international or cruise ship travel (23 patients, 34.8%). 16 HCW patients (24.2%) had an unknown source, 12 (18.2%) acquired COVID-19 from hospital, 11 (16.7%) from community source and 4 (6.0%) from aged care facilities. Almost one third (27.6%) of 58 patients with unknown source in our cohort were HCWs. The median age of our HCW subset was 45 years (range 17 to 81 years), and 3 HCWs (4.5%) required hospitalisation. There were no deaths in our HCW subgroup.

Both date of exposure and symptom onset could be accurately identified in 100 cases. The median incubation period was 5 days (IQR 3-8), with one case having an incubation period of 15 days.

The main comorbidities in the cohort were hypertension (81 patients, 15.7%), asthma or other chronic lung condition (57 patients, 13.9%), obesity (38 patients, 12.9%) and diabetes (33 patients, 6.4%) ( Table 2 ). 21 patients (4.1%) were current smokers. The predominant symptoms of COVID-19 were cough (313 patients, 60.5%), fatigue (307 patients, 59.4%), subjective fever (254 patients, 49.1%), myalgia or arthralgia (191 patients, 49.1%), sore throat (213 patients, 41.2%) and anosmia (188 patients, 36.4%) or ageusia (177 patients, 34.2%).

The majority of our cohort did not require laboratory or radiological investigations. Blood tests were performed in 81 patients (15.6%), whilst 82 patients (15.9%) had imaging, with either chest x-ray (CXR) or chest computed tomography (CT).

A significant correlation between clinical and radiological severity was observed in the subset of 78 patients who had a CXR performed. The median RALE score for patients who were determined to have clinically severe COVID-19 was 16.4, compared with a median RALE score of 3.0 for those who were determined to be non-severe. Multivariate analysis of this subset demonstrated that age over 80 years and having a RALE score of over 10 were independently associated with death (p=0.011 for both age and RALE score) and with incomplete recovery at 30 days (p=0.047 for age and p=0.008 for RALE score).

The most common blood test abnormalities were elevated D-dimer, CRP and procalcitonin (Appendix B).

At the time of diagnosis, 11 patients (2.1%) were asymptomatic, 453 (87.6%) had mild disease, 43 (8.3%) had moderate disease and 10 (1.9%) had severe disease (Table 3) . During the course of their COVID-19 illness, 61 patients (11.8%) required hospitalisation, 11 patients (2.1%) required intensive care admission, and 7 patients (1.4%) required ventilation. The median hospital length of stay was 10 days. There were 9 deaths (case fatality rate 1.8%) with a median age of 86 years in this subset. Amongst the ICU patients, 3 (27.3%) died, all of whom required invasive ventilation, giving a mortality rate amongst those on ventilation 42.8% (3/7 patients).

Incomplete recovery was reported in 28 patients (5.5%) at 30 days. Persistent symptoms included pleuritic chest pain, cough, fatigue, headache, anosmia, upper respiratory tract symptoms and diarrhoea. A single case was recorded of each of the following complications: pulmonary embolism, transverse myelitis, brachial neuritis, and earlier need for permanent renal dialysis.

Age greater than 80 years (p <0.001), having 2 or more comorbidities (p <0.001) or being overweight or obese (p ≤0.022) were associated with death, composite outcome of death or ICU admission, and incomplete recovery at 30 days on univariate analysis (Table 4 ). Other characteristics and risk factors associated with the study end points on univariate analysis are also shown at Table 4 . Age greater than 80 years (RR 24.4, 95% CI 4.1-144.8) and having 2 or more comorbidities (RR 27.6, 95% CI 2.9-258.9) remained independently associated with death on multivariate analysis. Having 2 or more comorbidities (RR 20.3, 95% CI 5.2-79.8) was also independently associated with composite outcome of death or ICU admission. Aged greater than 80 (RR 5.6, 95% CI 1.9-16.7) and having 2 or more comorbidities (RR 4.3, 95% CI 1.7-11.1) were independently associated with incomplete recovery at 30 days. 97.1% (402 patients) of our cohort did not receive any COVID-19 specific treatment (Appendix C).

This the largest Australian cohort study of COVID-19 outcomes and characteristics to date. The majority of cases in our cohort were acquired overseas, with most managed as outpatients through the NSLHD Virtual Hospital. This Virtual Hospital was established to manage COVID-19 in our local health district and will now be utilised for non-COVID-19 diagnoses. We report a cohort with relatively mild disease and low case fatality rate, in a well-resourced setting. Hospital-in-the-home services have been found to provide safe and resource efficient care to patients in other Australian settings. (15) (16) (17) The median incubation period of 5 days seen in this cohort is consistent with international studies and guidelines.(5, 18) However, one patient did have an incubation period of 15 days, falling outside of the widely recommended 14-day quarantine period. There have been other cases published with incubation periods beyond 14 days. (19) The incubation period could only be calculated in less than one fifth of our cohort, predominantly due to the main acquisitions being international, where source could not be identified.

COVID-19 remains a challenging clinical diagnosis. The most commonly reported manifestations in our cohort were non-specific symptoms of cough, fatigue and fever, consistent with Australian and international studies. (3-8, 17, 20) Most patients were afebrile on admission and characterised as mild severity or asymptomatic. The variability of initial symptoms can result in difficulty identifying those at risk of deterioration. Our study identified the following risks for adverse outcome: age over 80 years, presence of 2 or more comorbidities, and being overweight or obese. These findings are similar to large international studies and recent Australian studies where age, male sex, and comorbidities, including obesity, were associated with worse overall outcomes. (3, 8, 10, 21) We note a relatively high percentage of HCWs in our study. Although there may be a testing bias, this shows that HCWs are a key vulnerable group where early identification is critical and public health measures including vaccination should be targeted. The majority of HCWs in our cohort acquired COVID-19 in the community, rather than a nosocomial setting. This predominance of community acquisition amongst HCWS is a consistent finding in different health care settings. (22, 23) The significant difference found in our cohort between the RALE score in those with severe vs nonsevere disease, is in line with previous reports (14) and, together with abnormal pathology investigations, can be utilised to identify a subset of patients with severe disease who require closer follow-up.

Hospitalisation was required in 11.8% of our cases, with ICU requirement appropriate and necessary in 2.1%. These figures can assist health systems to prepare for future COVID-19 surges. Very few patients in our cohort received COVID-19 specific treatment, reflecting the limited available evidence during the early stages of the pandemic. This cohort predated any established effective therapy for COVID-19 including the results of the Randomised Evaluation of Covid-19 Therapy (RECOVERY) trial showing efficacy of dexamethasone.(24) Treatment used in several patients in our cohort, including hydroxychloroquine alone or in combination with azithromycin, has been later shown to have no effect and can even increase COVID-19 associated mortality.(25) Despite small numbers, our ICU mortality figures are higher than reported in other Australian ICU studies. (20, 26) Our case fatality rate of 1.8% compares favourably to the Australian wide rate of 3.0% (2) . We hypothesise that the comparatively lower mortality rate seen in this study is reflective of several factors. Firstly, the rates of testing in this health district were relatively high, thus enabling the early detection of asymptomatic cases that increases the denominator in our overall calculations. Furthermore, the relatively low prevalence in the NSLHD allowed a swift and sustained health response from the PHU, microbiology laboratories, hospital systems and ICUs, thus facilitating containment, rapid diagnosis and effective medical management. The general population's compliance with public health measures such as physical distancing, enforced periods of lockdown, along with tight border control, have contributed to the low prevalence and sustained response. Finally, the relatively low rates of comorbidities and good pre-morbid function (27) seen in the cohort are likely to have had an impact on lowering the case fatality rate.

It is noted, however, that the case fatality rate in this study is slightly higher than the NSW overall rate of 1.0%. (2) We suggest that this may be reflective of a subset of patients from a single aged care facility in which a COVID-19 outbreak was experienced during the study period. Five of the nine patients who died from COVID-19 were from this aged care facility. Our analysis of the subgroup of aged care was underpowered, however, we know that aged care facilities account for a large proportion of COVID-19 deaths globally and in Australia. (28) This association is likely to be largely driven by the association between age and COVID-19 mortality, as we noted in our study.

On a global scale, our case fatality rate compares favourably to countries such as the United Kingdom, China and Italy, with case fatality rates of 2.9%, 4.7% and 3.0% respectively; and similarly to the United States with a case fatality rate of 1.8%. (29) Limitations of our study include the retrospective collection of patient self-reported recovery, which makes it liable to recall bias. Additionally, we do not have the complete data detailing dates of patient follow-up comparative to their diagnosis, making it difficult to determine the extent of the recall bias. Ascertainment bias may have been introduced where residents of NSLHD were managed by alternate health districts whilst in hotel quarantine. The data sets for patient BMI and recorded fever were incomplete, limiting the power of multivariate analysis. Analysis of laboratory and radiological investigations were limited, with less than 16 per cent requiring further investigations. This reflects the overall mild disease in our cohort. Our study was conducted in an area of high socioeconomic status with lower rates of obesity and smoking than the average Australian community, therefore limiting generalisability to the Australian population. Similarly, this data is from last year, limiting generalisability in such a fast-moving field, with unprecedented rate of new data emerging.

We report an Australian cohort analysis of 517 laboratory-confirmed COVID-19 patients with predominantly mild disease and a low mortality in a low prevalence, well-resourced setting. Factors associated with poor patient outcomes on multivariate analysis were age over 80 years and the presence of two or more comorbidities. Our data can inform risk stratification in the COVID-19 Virtual Hospital to determine resource allocation and early hospital admission. Rates of hospitalisation, ICU admission and ventilation requirement in this cohort, have assisted NSLHD in surge capacity planning for subsequent COVID-19 waves. 

World Health Organization. WHO Coronavirus Disease (COVID-19) Dashboard 2020

Australian Government Department of Health. Coronavirus (COVID-19) current situation and case numbers 2021

Clinical characteristics of 3,062 COVID-19 patients: a meta-analysis

Review of the Clinical Characteristics of Coronavirus Disease 2019 (COVID-19)

Clinical Characteristics of Coronavirus Disease 2019 in China

Clinical Characteristics of Covid-19 in New York City

Presenting Characteristics, Comorbidities, and Outcomes Among 5700 Patients Hospitalized With COVID-19 in the

UK patients with COVID-19 using the ISARIC WHO Clinical Characterisation Protocol

Features of anosmia in COVID-19

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

World Health Organization. Clinical management of COVID-19: Interim guidance. 2020. 13

Chest X-ray in new Coronavirus Disease 2019 (COVID-19) infection: findings and correlation with clinical outcome

Virtual health care for community management of patients with COVID-19

Hospital-in-the-Home experience of first 23 COVID-19 patients at a regional NSW hospital

Respiratory failure, clinical course and community management of COVID-19 patients in a large Australian cohort

Incubation period of COVID-19: a rapid systematic review and meta-analysis of observational research

Is a 14-day quarantine period optimal for effectively controlling coronavirus disease 2019 (COVID-19)?

Outcomes for patients with COVID-19 admitted to Australian intensive care units during the first four months of the pandemic

Outcomes for the first wave of hospitalised patients with COVID-19 in the South Australian context: a retrospective audit

Characteristics of healthcare workers infected with COVID-19: A cross-sectional observational study

Acquisition of COVID-19 by health care workers: the importance of nonpatient workplace sources

Dexamethasone in Hospitalized Patients with Covid-19

Effect of hydroxychloroquine with or without azithromycin on the mortality of coronavirus disease 2019 (COVID-19) patients: a systematic review and meta-analysis. Clinical microbiology and infection : the official publication of the European Society of Clinical Microbiology and Infectious Diseases

Impact of COVID-19 on an Australian intensive care unit: lessons learned from South Australia

Mortality associated with COVID-19 in care homes: international evidence. (2020) Article in LTCcovid.org, International Long-Term Care Policy Network

Coronavirus Resource Centre. Mortality Analyses 2021

Appendix C 97.1% of our cohort (402 patients) did not receive any COVID-19 specific treatment. Of the 15 patients who did receive specific treatment, 5 received hydroxychloroquine alone, 7 received hydroxychloroquine and azithromycin in combination, 2 received lopinavir/ritonavir and hydroxychloroquine in combination, and one received lopinavir/ritonavir alone. Only 8 of these 15 patients completed their treatment as planned, with prolonged QTc interval (42.9%) and gastrointestinal upset (28.6%) being the main side effects. One patient had their treatment stopped due to prescriber concern regarding potential drug interaction causing prolonged QTc, while another patient's treatment was stopped due to concern regarding the lack of evidence for the COVID-19 specific treatment.