Outcomes of COVID-19 patients admitted to Australian intensive care units during the early phase of the pandemic

Introduction

The COVID-19 pandemic has caused an unprecedented burden on intensive care units (ICUs) worldwide. In early case series, despite advanced ICU supports, including invasive mechanical ventilation and renal replacement therapy, ICU mortality rates were between 40-90%.^1-3 In a recently reported large scale randomised trial from the United Kingdom (UK), the mortality rate among invasively ventilated patients in the standard arm was 40%^4,5. ICU mortality rates for COVID-19 are substantially higher than reported in previous epidemics of viral pneumonitis, including the 2009 H1N1 influenza pandemic, where reported rates were between 10 and 30%.^6,7

Many reports of COVID-19 patient outcomes have come from healthcare systems whose capacity were exceeded with COVID-19 cases. In parts of China, Italy and New York, the rapid increase in COVID-19 cases permitted minimal time for preparation, with resultant shortages of resources, including beds⁸, equipment (including personal protective equipment and ventilators), and appropriately trained staff^9,10. The reported mortality may even have been underestimated, as many patients at the time of reporting were still undergoing treatment and whose final outcome was unknown^11,12.

In Australia, the first recorded case of COVID-19 was on 25 January 2020¹³, and by 5 July 2020, only 8566 confirmed cases had been reported¹⁴. As the pandemic took hold in Australia, the Short PeRiod IncideNce sTudy of Severe Acute Respiratory Infection (SPRINT-SARI Australia) was activated. This study aimed to collect comprehensive observational data on patients admitted to ICUs with COVID-19, to improve our understanding of the natural history of the disease, and to provide contemporary local data concerning ICU outcomes and resource utilisation.

Methods

Study design and setting

SPRINT-SARI Australia is a multi-centre, prospective, observational study of patients with COVID-19 admitted to participating ICUs in Australia. The study design, case report form (CRF) and protocol were developed in conjunction with the International Severe Acute Respiratory and Emerging Infection Consortium (ISARIC)¹⁵. A standardised CRF, which enabled a rapidly scalable data collection platform for acquiring clinical information and sharing, was developed in response to multiple outbreaks of severe acute respiratory infection over the past 10 years¹⁶. SPRINT-SARI Australia is supported by the Australian and New Zealand Intensive Care Society Clinical Trials Group (ANZICS CTG), and is coordinated by the Australian and New Zealand Intensive Care Research Centre (ANZIC-RC), Monash University. Participating sites across Australia were identified via the ANZICS CTG, or through previous affiliation with SPRINT-SARI Australia.

Australian ICUs are predominantly staffed by full time specialists. The decision to admit patients to ICU are largely governed by factors such as the likely response to treatment, the likely prognosis, and the long term outcome¹⁷. Standard nursing ratios of one nurse to one patient (1:1) for an ICU patient and one nurse to two (1:2) for high dependency patients exist in Australian ICUs¹⁸.

Participants

The study population included patients of all ages with a confirmed laboratory polymerase chain reaction (PCR) test for COVID-19 who had an index COVID-19 related admission to a participating ICU. Patients found to be PCR negative, or whose test was pending at the end of the study period, were excluded. Biological samples for PCR testing could be from the nasopharynx, trachea or lower airways via bronchoscopy, as per local policy¹⁹.

Data collection

Dedicated research staff at each ICU were responsible for screening all admissions for COVID-19 patients. Data from the electronic medical record or paper notes were entered into a database (REDCap, Vanderbilt University) without personal identifiers. The ANZIC-RC maintained the database and performed all analyses²⁰. Start-up meetings, detailed data dictionary, and quality checks were completed to ensure data quality and protocol standardisation, and to minimise bias.

Data collected included baseline demographic and clinical characteristics. The Acute Physiology and Chronic Health Evaluation II (APACHE-II) score and the Sequential Organ Failure Assessment (SOFA) score for the first 24 hours were calculated. Data on investigations, ICU treatments and interventions were collected daily until day 28. Outcomes were recorded as death or upon hospital discharge.

To quantify how comprehensive data collection was, we cross referenced SPRINT-SARI Australia admission data with that collected by the Commonwealth of Australia.¹⁴ This indicated 225 patients had been admitted to ICU with COVID-19 up to and including 5 July 2020. To determine maximum site occupancy, we divided the peak number of COVID-19 cases, by the total number of ICU beds at each study site (as reported by the Australian and New Zealand Intensive Care Society Centre for Outcomes and Resource Evaluation²¹), expressed as a proportion.

Statistics

Data were extracted on 28 July 2020 and pertains to ICU admissions reported to SPRINT-SARI Australia between 27 February 2020 and 30 June 2020. The median ICU length of stay (LOS) was computed using Kaplan-Meier survival methods, censoring at the date of the last daily record, for patients without an ICU discharge date. Weibull survival regression analysis was used to assess risk factors for ICU mortality and LOS in survivors.²² Time to ICU mortality was defined as time from ICU admission to date of death, censoring at either ICU discharge or the date of the last daily record for patients alive and still in ICU. LOS in ICU among survivors was modelled as time from ICU admission to ICU discharge, censoring for both those who died and those still in ICU. Age, sex, APACHE-II and receipt of invasive ventilation were selected for inclusion in the multivariate models a priori, being highly clinically relevant. Remaining variables were then selected following a forward stepwise approach judged by the likelihood ratio test, with a 0.05 significance level used for variable removal and 0.01 significance level for variable inclusion.  A parametric survival model with a Weibull distribution was fit in order to incorporate ICU site random effects in the analysis²². Hazard ratios and p-values were reported. Our analysis included all available data. All proportions were adjusted if data were missing, and the total number of patients contributing data, for all analyses, are provided. We did not impute missing data. All analyses were performed using Stata version 16 (Stata Corp, College Station, Texas, United States of America (USA)) and R statistical software (R Core Team, 2019). 

Ethical approval

All the authors reviewed the manuscript and vouch for the accuracy and completeness of the data provided. Human Research Ethics Committee (HREC) approval for data collection, with a waiver of informed consent, was granted via the National Mutual Acceptance (NMA) scheme, through the Alfred (HREC/16/Alfred/59), or by separate applications to individual sites. Research Governance approval was granted by the Chief Health Officer (CHO) in South Australia and Victoria, and supported by the CHO in Queensland, under legislated public health powers. Individual site Research Governance approvals were granted at all sites where it was required.

Results

From 27 February to 30 June 2020, a total of 77 ICUs participated in SPRINT-SARI Australia, accounting for 1260/1503 (84%) of all public hospital ICU beds²¹. Forty-four sites contributed at least one confirmed COVID-19 patient, while 32 sites had no confirmed cases, and 1 site had not completed data entry. National ICU bed utilisation peaked on 5 April 2020 with 90 patients, with numbers falling to below 25 patients by the beginning of May (Figure 1 - available in PDF). Median peak ICU bed occupancy at each hospital was 14% (IQR 9-16, range 4-40) (Supplement Figure 1 - available in PDF). The ICU nurse to patient ratio was 1:1 (1766/2270 ICU days,77.80%) or 2:1 (171/2270 ICU days, 7.53%).

A total of 204 patients were included, representing 204/225 (90.7%) of the ICU cases in Australia (Supplement Figure 2 - available in PDF)¹⁴. 140 (68.6%) were males, 64 (31.4%) were female, and the median age was 63 years (IQR 53-72) (Table 1 - available in PDF). Most frequently reported comorbidities were obesity 80/204 (39.2%), diabetes 57/204 (27.9%), hypertension requiring angiotensin converting enzyme (ACE) inhibitor or angiotensin receptor blockade (ARB) 49/204 (24.0%) and chronic cardiac disease 40/204 (19.6%), while for 73/204(35.8%) of patients no comorbidities were recorded. The most prevalent symptoms at time of admission to hospital were fever, cough, shortness of breath, fatigue, myalgia and diarrhoea. Median duration from onset of first symptoms to hospital admission was 6 days (IQR 4-9), and from hospital admission to ICU admission was 1 day (IQR 0-3). Median APACHE-II and SOFA scores at 24 hours in ICU were 14 (IQR 10-18) and 6 (IQR 4-10) respectively. For 114/204 (55.9% ), infection was acquired through international travel, of whom almost half were cruise ship travellers (55/204, 27.0%). Close contact with a confirmed or probable case of infection was reported by 92/20445.1% (45.1%) of patients and 17/204 (8.3%) identified as a healthcare worker.

Invasive ventilation was provided for 119/204 (58.3%). Compared to non-ventilated patients, they were older (median 68 years; IQR 57-73 vs median 61 years; IQR 46-69), more likely to be obese (52/119, 43.7%  vs 28/85, 32.9%,), have diabetes (44/119, 37.0% vs 13/85, 15.3%) and chronic cardiac disease (27/119, 22.7%  vs 13/85, 15.3%), but were less likely to have chronic pulmonary disease (7/119, 5.9% vs 9/85, 10.6%) (Table 1 - available in PDF).

Once admitted to ICU, 79/204 (38.7%) of patients were commenced on invasive mechanical ventilation on day 1, while 54/204 (26.5%) were supported with high flow oxygen therapy (Figure 2 - available in PDF). The proportion of patients that were invasively ventilated increased to 94/113 (83.2%) by the end of the first week. Non-invasive ventilation (NIV) was used in 4/204 (1.9%) of patients on day 1 and 3/113 (2.7%) on day 7. The most common additional interventions were inotropes (111/204, 54.4%), neuromuscular blockade (87/204, 42.6%), prone positioning (56/204, 27.5%), and corticosteroids (58/204, 28.4%), while renal replacement therapy (23/204, 11.3%), and venovenous extracorporeal membrane oxygenation (ECMO) (2/204, 1.0%) were less common. Hydroxychloroquine was used in 32/204 (15.7%) of patients.

Hospital follow-up was complete for 194/204  (95%), with four patients having ongoing care in the ICU, and six in another area of the hospital. The median ICU LOS for invasively ventilated patients was 16 days (IQR 9-28) compared to 3 days (IQR 2-5) in those not requiring invasive mechanical ventilation (Figure 3 - available in PDF). Of mechanically ventilated patients, 27/119 (22.7%) stayed in ICU for 30 days or longer. Mixed-effects survival regression analysis of the number of days in ICU showed that the use of invasive ventilation (hazard ratio (HR) 0.07; 95% confidence interval (CI) 0.04-0.11; p<0.001) and renal replacement therapy (HR 0.44; 95% CI 0.24-0.80; p=0.007) were associated with a lower chance of early ICU discharge among survivors, after controlling for age, sex and APACHE-II (Supplement Tables 1 and 2 - available in PDF).

ICU mortality was 30/200 (15.0%; 95% CI 10.4-20.7). In those patients requiring invasive ventilation, mortality was 26/117 (22.2%; 95% CI 15.1-30.8), compared to 4/83 (4.8%; 95% CI 1.3-11.9) in those that were not. All but 2 of the deaths were in patients 60 years and over (Figure 4 - available in PDF). Table 2 (available in PDF) shows the univariate factors associated with ICU mortality; older age (over 64), chronic cardiac, pulmonary and kidney disease were associated with ICU mortality. When controlling for other factors (age, sex and invasive ventilation) initial severity of illness scored using APACHE-II (HR 1.15; 95% CI 1.09-1.21; p<0.001) and chronic cardiac disease (HR 3.38; 95% CI 1.46-7.83; p=0.004) were associated with ICU mortality (Table 3 - available in PDF).

Discussion

In this first report of Australian patients admitted to ICUs during the early phase of the COVID-19 pandemic, we found invasively ventilated patients had an ICU mortality of 22.2%, considerably lower than rates reported internationally, and despite a higher proportion of patients with complete outcomes. We also found that median ICU LOS was longer, with 22.7% of patients staying 30 days or longer.

Several factors may account for this difference. Reports from many European countries, the USA, and parts of China showed the numbers of COVID-19 patients increased rapidly allowing little preparation, quickly exceeding their health care capacity^2,11,12,23. The numbers of patients invasively ventilated on day 1 of their ICU admission in these countries were very high (70-90%). In our study, even at a peak of 90 patients, overall COVID-19 numbers remained low, and only 79/204 (38.7%)  of ICU cases were ventilated in the first 24 hours (Figure 2 - available in PDF). Moreover, patients were distributed across a large number of institutions, and the maximum number of COVID-19 patients at any one site (as a proportion of their total ICU beds), remained low (median 14% IQR 9-16) (Supplement Figure 1 - available in PDF).

 As the healthcare system in Australia never reached, let alone exceeded capacity during this period, it may be that in these circumstances, patients could access the ICU earlier in the course of their illness, thereby benefiting from interventions that have been associated with lower mortality rates^24,25. In a study from China, the rapid escalation of infections around the Wuhan epicentre was associated with increased mortality rates, when compared to other parts of China where the infection rate was slower²⁶.

The second potential reason is that our study cohort differs from that of other countries in several key ways. First, the source of infection was most commonly from international travel (114/204, 55.9%), compared to locally acquired infection in many other countries^1,11. This ‘travelling’ population may be ‘healthier’ compared to cases due to community spread, where patients, such as those in long term care facilities, have been infected¹. The median age of our cohort (63 years; IQR 53-72) was similar to ICU cohorts reported from the USA (64 years)¹² and Italy (63 years)², but younger than the UK (73 years)¹¹ population. In addition, our cohort had fewer comorbidities, the number of which has been shown to increase the risk of ICU admission, the need for mechanical ventilation or death²⁷.

We also found the ICU LOS of invasively ventilated patients was prolonged (median 16 days; IQR 9-28) compared to studies from the UK (9.7 days)¹¹ and USA (12 days)¹². While the use of other ICU supports, including rates of proning (28-40%), neuromuscular blockade (39%), renal replacement therapy (20%) and ECMO (0-3%) was similar to our findings, it is plausible that the prolonged duration of support in Australian ICUs was only possible because the system had capacity to provide it. This very long, resource-intense period of treatment has important implications for ongoing service provision and planning. The natural history of COVID-19 in Australian ICUs during this early phase appeared to be a resource heavy, protracted admission with a low mortality rate. Given the recent increase in COVID-19 cases, it remains to be seen whether this will persist, particularly with changes in epidemiology due to wider community transmission. Of note, our data implies a substantial resource burden in caring for these patients that, if not met, may also impact outcomes.²⁷

Strengths and limitations

This study had comprehensive coverage of ICUs in Australia, providing unique, nationally representative data in a health care system operating within its capacity. The data were collected using a standardised CRF, with experienced data collectors, and included daily data fields for the first 28 days. The long duration of follow-up was complemented with near complete outcome data. The limitations include the observational nature of the study, with inevitable confounding among therapeutic factors associated with mortality and other characteristics.

We did not collect data on COVID-19 patients who were not admitted to the ICU, and as such, our study does not explore the complex decision making process around ICU admission. While this may limit the generalisability of our findings, we chose a pragmatic approach, so as to rapidly collect, analyse, and report data on the interventions and outcomes of confirmed COVID-19 cases admitted to ICU in Australia. In addition, as this study focussed primarily on ICU interventions, treatments provided prior to ICU admission were not included in analysis, and may have impacted the primary outcome. Finally, the COVID-19 pandemic is ongoing in Australia, and it is uncertain whether these data will be representative of future cohorts.

Conclusion

During the early phase of the pandemic in Australia, patients admitted to ICU with COVID-19 had lower mortality and longer length of stay than reported from other regions. These findings reinforce the importance of ensuring adequate local ICU capacity, particularly given the recent increase in COVID-19 cases.

References

1. Bhatraju PK, Ghassemieh BJ, Nichols M, et al. Covid-19 in Critically Ill Patients in the Seattle Region - Case Series. N Engl J Med 2020;382:2012-22.
2. Grasselli G, Zangrillo A, Zanella A, et al. Baseline Characteristics and Outcomes of 1591 Patients Infected With SARS-CoV-2 Admitted to ICUs of the Lombardy Region, Italy. JAMA 2020;doi: 10.1001/jama.2020.5394.
3. Yang X, Yu Y, Xu J, et al. Clinical course and outcomes of critically ill patients with SARS-CoV-2 pneumonia in Wuhan, China: a single-centered, retrospective, observational study. Lancet Respir Med 2020;8:475-81.
4. Coronavirus Disease ( COVID-19) Communicable Diseases Network Australia (CDNA) National Guidelines for Public Health Units. 12 Aug 2020. https://www1.health.gov.au/internet/main/publishing.nsf/Content/7A8654A8CB144F5FCA2584F8001F91E2/$File/COVID-19-SoNG-v3.7.pdf (accessed 19 Aug 2020).
5. Recovery Collaborative Group, Horby P, Lim WS, et al. Dexamethasone in Hospitalized Patients with Covid-19 - Preliminary Report. N Engl J Med 2020;doi: 10.1056/NEJMoa2021436.
6. ANZIC Influenza Investigators, Webb SA, Pettila V, et al. Critical care services and 2009 H1N1 influenza in Australia and New Zealand. N Engl J Med 2009;361:1925-34.
7. Nguyen-Van-Tam JS, Openshaw PJ, Hashim A, et al. Risk factors for hospitalisation and poor outcome with pandemic A/H1N1 influenza: United Kingdom first wave (May-September 2009). Thorax 2010;65:645-51.
8. Grasselli G, Pesenti A, Cecconi M. Critical Care Utilization for the COVID-19 Outbreak in Lombardy, Italy: Early Experience and Forecast During an Emergency Response. JAMA 2020;doi: 10.1001/jama.2020.4031.
9. Murthy S, Gomersall CD, Fowler RA. Care for Critically Ill Patients With COVID-19. JAMA 2020;doi: 10.1001/jama.2020.3633.
10. Xie J, Tong Z, Guan X, Du B, Qiu H, Slutsky AS. Critical care crisis and some recommendations during the COVID-19 epidemic in China. Intensive Care Med 2020;46:837-40.
11. Docherty AB, Harrison EM, Green CA, et al. Features of 20 133 UK patients in hospital with covid-19 using the ISARIC WHO Clinical Characterisation Protocol: prospective observational cohort study. BMJ 2020;369:m1985 (doi: 10.1136/bmj.m1985).
12. Richardson S, Hirsch JS, Narasimhan M, et al. Presenting Characteristics, Comorbidities, and Outcomes Among 5700 Patients Hospitalized With COVID-19 in the New York City Area. JAMA 2020;doi: 10.1001/jama.2020.6775.
13. Caly L, Druce J, Roberts J, et al. Isolation and rapid sharing of the 2019 novel coronavirus (SARS-CoV-2) from the first patient diagnosed with COVID-19 in Australia. Med J Aust 2020;212:459-62.
14. Covid-National Incident Room Surveillance Team. COVID-19, Australia: Epidemiology Report 16 (Reporting week to 23:59 AEST 17 May 2020). Commun Dis Intell (2018) 2020;44 (doi: 10.33321/cdi.2020.44.45).
15. International Severe Acute Respiratory and Emerging Infection Consortium (ISARIC), Short PeRiod IncideNce sTudy of Severe Acute Respiratory Infection (SPRINT-SARI) webpage. https://isaric.tghn.org/sprint-sari/ (accessed 19 Aug 2020).
16. International Severe Acute Respiratory and Emerging Infections Consortium (ISARIC) COVID-19 Report 19 May 2020. https://media.tghn.org/medialibrary/2020/05/ISARIC_Data_Platform_COVID-19_Report_19MAY20.pdf (accessed 19 Aug 2020).
17. Warrillow S, Austin D, Cheung W, et al. ANZICS guiding principles for complex decision making during the COVID-19 pandemic. Crit Care Resusc 2020;22:98-102.
18. Australian and New Zealand Intensive Care Society (ANZICS) COVID-19 Guidelines, Version 2. https://www.anzics.com.au/wp-content/uploads/2020/04/ANZI_3367_Guidelines_V2.pdf (accessed 19 Aug 2020).
19. Flint AWJ, Bailey M, Reid CM, et al. Preoperative identification of cardiac surgery patients at risk of receiving a platelet transfusion: The Australian Cardiac Surgery Platelet Transfusion (ACSePT) risk prediction tool. Transfusion 2020.
20. Australian and New Zealand Intensive Care Research Centre, Monash University, webpage. https://www.monash.edu/medicine/sphpm/anzicrc/home (accessed 19 Aug 2020).
21. Australian and New Zealand Intensive Care Society (ANZICS) Centre for Outcome and Resource Evaluation Adult Patient Database (APD) Activity Report 2017‐2018. https://www.anzics.com.au/wp-content/uploads/2019/10/2018-ANZICS-CORE-Report.pdf (accessed 19 Aug 2020).
22. Therneau TM, Grambsch PM. Modeling Survival Data: Extending the Cox Model (ISBN 978-1-4757-3294-8). 1st ed. New York, NY: Springer; 2000:350 pages.
23. Guan WJ, Ni ZY, Hu Y, et al. Clinical Characteristics of Coronavirus Disease 2019 in China. N Engl J Med 2020;382:1708-20.
24. Peng Y, Xu B, Sun B, Han G, Zhou YH. Importance of timely management of patients in reducing fatality rate of coronavirus disease 2019. J Infect Public Health 2020;13:890-2.
25. Litton E, Bucci T, Chavan S, et al. Surge capacity of intensive care units in case of acute increase in demand caused by COVID-19 in Australia. Med J Aust 2020;212:463-7.
26. Ji Y, Ma Z, Peppelenbosch MP, Pan Q. Potential association between COVID-19 mortality and health-care resource availability. Lancet Glob Health 2020;8:e480.
27. Mitra AR, Fergusson NA, Lloyd-Smith E, et al. Baseline characteristics and outcomes of patients with COVID-19 admitted to intensive care units in Vancouver, Canada: a case series. CMAJ 2020;192:E694-E701