This single centre, retrospective observational study was carried out from July 2020 to December 2020, COVID dedicated ICU of state-run tertiary care hospital in Firozabad, India. After approval from Institutional Ethics Committee (Ref. No-ASMC/IEC/2020-21/05) deidentified data of critically ill patients, who were laboratory confirmed for SARS-CoV-2 infection, and admitted to COVID dedicated ICU was reviewed.

Inclusion criteria: Data of all patients who were admitted to COVID dedicated ICU during the mentioned study period were included in this study.

Exclusion criteria: Data of those patients who were referred to other centres during study period were excluded.

According to the WHO guidance, [18] laboratory confirmation for SARS-CoV-2 was defined as a positive result of Real-Time Reverse Transcriptase-Polymerase Chain Reaction (rRT-PCR) assay of nasal and pharyngeal swabs. This guideline was implemented locally. Only those COVID-19 patients who had severe disease, were admitted to COVID dedicated ICU. Severe COVID-19 disease is defined by Ministry of Health and Family Welfare (MoHFW), India as clinical sign of pneumonia plus one of the following; respiratory rate >30 breaths/min, severe respiratory distress, Oxygen Saturation (Spo₂) <90% on room air [19]. Patients were discharged from ICU after clinical recovery and negative: RT-PCR report (after resolution of symptom).

In this study, clinical data were collected from indoor files of COVID ICU admissions. A total of 128 severe COVID-19 patients were admitted to COVID dedicated ICU from July 2020 to December 2020. Among these 128 severe COVID patients, eight patients were referred to other centres. All referrals were due to the patients or their family’s willingness to sought treatment at top corporate hospitals in metropolitan cities. So, sample size of this study was 120. Data of these admissions were collected and recorded- age, sex, co-morbidities and mode of oxygen delivery (invasive mechanical ventilation, non-invasive mechanical ventilation, high flow nasal canula). The number of patients who were discharged successfully and died during the treatment were recorded. Different age groups were defined as follows: upto 40 years; 41 to 60 years; 61 years and above.

Analyses was performed using “Statistical Analysis System” (SAS 9.4). Chi-square test was used to compare the outcomes of critically ill COVID patients with and without co-morbidity, in different age groups and in both sexes. The statistical significance is defined as p-value <0.05.

In this study, the data of 120 severely ill COVID-19 patients were reviewed who were admitted to COVID dedicated ICU between July 2020 to December 2020. By the 25^th of January 2021, all these 120 COVID ICU admissions have been disposed. Among these 120 admissions there were 90 male patients and 30 female patients with male to female ratio of 3:1. The age of patients varies from 17-88 years with a mean age of (58±15.29) years [Table/Fig-1].

In this study, 64 (53.3%) COVID ICU admissions have atleast one co-morbid condition. Hypertension was the most common co-morbid condition reported in 44 (36.6%) admissions, followed by type 2 diabetes mellitus which was found in 24 (20%) patients. Eighteen (15%) patients had Chronic Obstructive Pulmonary Disease (COPD) [Table/Fig-2].

All COVID-19 patients admitted to COVID dedicated ICU had moderate to severe ARDS (as per Berlin definition) [20]. The median baseline Pao₂/Fio₂ ratio was 124 (range 72 to 184). Chest radiograph was taken in all patients at the time of admission and daily thereafter. Bilateral infiltrates were found in radiographs of 86 (71.7%) patients and unilateral infiltrate in radiograph of 34 (28.3%) patients. Pleural effusion was not evident in any chest radiograph. Baseline laboratory investigations were done at the time of ICU admissions, which shows normal median total leucocyte count 7.6 (5.2-11.3)×10³/cmm, elevated median serum ferritin 614 (240-964) ng/mL and elevated median C-Reactive Protein (CRP) 108 (44-152) mg/L.

All COVID ICU patients were monitored continuously by pulse oximetry. The highest mode of oxygen support in these patients was documented and it was determined by Spo₂, respiratory rate and breathing pattern of the patient and Arterial Blood Gas Analysis (ABG). The oxygen support was escalated and de-escalated according to the clinical condition of patients, and for this reason all deaths were reported in patients who were on invasive ventilation. The Extracorporeal Membrane Oxygenation (ECMO) facility is not available at the centre. Among these 120 patients, 40 (33.3%) patients were put on invasive ventilation whereas, 48 (40%) patients were managed with non-invasive ventilation. High-Flow Nasal Cannula (HFNC) was used initially in 48 patients (40%), but HFNC failure was reported in 13 cases and 3 patients complained of severe headache and lacrimation due to high flow of HFNC. All these 16 patients (13+3) were shifted to non-invasive ventilation, the use of high flow nasal canula was sufficient to manage 32 (26.7%) patients [Table/Fig-3].

Apart from oxygen support, severe COVID patients were treated with parenteral antibiotics guided by total leucocyte counts and in few cases by blood culture, parenteral methyl prednisolone, parenteral remdesivir, subcutaneous Low Molecular Weight Heparin (LMWH) and other supportive medications (vitamin C, zinc, paracetamol, Intravenous (IV) fluids, vasopressors etc.,). Treatment of co-morbid conditions was done accordingly.

Among these 120 COVID ICU admissions, one patient who was on invasive ventilation developed bilateral pneumothorax and was managed with bilateral intercostal drains but unfortunately died. One patient suddenly developed massive haematemesis and succumbed. Two patients developed urethral bleeding and one epistaxis, all were successfully managed by administrating IV tranexamic acid 500 mg 12 hourly till bleeding subsides and withholding LMWH, which was given 0.4 mg subcutaneously twice daily.

Among these 120 COVID ICU admissions, 103 patients were successfully treated and discharged from the ICU rest 17 patients died during the course of treatment. In this study, only one patient who had no co-morbid condition died due to development of bilateral pneumothorax and remaining 16 death were among patients who had atleast one co-morbid condition. In this study, it was observed that presence of co-morbid condition is a poor prognostic factor and it is found to be statistically significant (p=0.001) [Table/Fig-4].

Male sex has also appeared to be poor prognostic factor in critically ill COVID patients. Fifteen (16.7%) deaths were among 90 critically ill male COVID patients whereas, only 2 (6.7%) deaths were reported in 30 critically ill female COVID patients. However, it was not found to be statistically significant (p>0.05) [Table/Fig-5].

There was only one death under 40 years of age (mortality 6.7%). In age group of 41-60 years, 7 patients died (mortality 13.5%) and 45 were discharged successfully. In age groups 61 years and above 9 patients died (mortality 17%) and rest 44 patients were discharged successfully. However, difference in mortality of different age groups was not statistically significant (p>0.05) [Table/Fig-6]. The overall mortality in this COVID ICU study is found to be 14.2%.

In this study, critically ill COVID patients were predominantly male with a mean age of (58±15.29) years. An 87.5% of COVID ICU Patients were above 40 years of age. In comparison to this, study conducted in Italy by Grasselli G et al., shows 82% of critically ill patients were male and median age was 63 [21].

In the present study, 53.3% of critically ill patients had atleast 1 co-morbid condition, which is less as compared to 68% reported by Grasselli G et al., and Wang D et al., (72.2%) [21,22]. Authors reported higher percentage of chronic obstructive pulmonary disease (COPD) 15% in our patients as compared to 4% of study conducted by Grasselli G et al., [21]. The higher prevalence of COPD in current study is due to predominant habit of bidi smoking in local area and local population is mainly comprises of factory workers. In this study, bilateral infiltrates in chest radiograph were found in 71.7% patients and unilateral in 28.3% patients, this percentage was 87.5% and 12.5%, respectively in another study conducted in western India [23].

Baseline normal median leucocyte count 7.6 (5.2-11.3)×10³/cmm and elevated median serum ferritin 614 (240-964) ng/mL and C-Reactive Protein (CRP) 108(44-152) mg/L was also in accordance with study conducted in western India [23].

The majority of patients were admitted to COVID ICU in need of respiratory support due to acute hypoxemic respiratory failure. In this study, 33.3% patients were intubated and put on invasive ventilation, this percentage of invasively ventilated patients is less as compared to other recently reported COVID ICU studies: 42% (Wuhan, China) [8], 47% (Wuhan, China) [22], 71% (Washington State, US) [7]. The need of invasive ventilation in our ICU was slightly higher than data reported by two Chinese studies from Wuhan with rates of 14.5% and 17%, respectively [3,24]. In this study, 40% of critically ill COVID patients were managed with non-invasive ventilation which is similar to the data collected from Wuhan, china (42%) [22]. In different studies use of noninvasive ventilation varied widely from 56% in another study of Wuhan China [8] to 19% in Washington State US [7].

In a retrospective review of COVID patients in two hospitals of Chongqing, China 27 patients experienced severe acute respiratory failure out of which 17 patients (63%) were treated with HFNC for first time. The HFNC failure was reported in 7 cases (41%) [25]. In this study, we initially put 48 patients (40%) on HFNC. A 32 COVID ICU patients (66.6% of initial HFNC trial) were successfully treated with HFNC, whereas 16 patients who were initially on HFNC shifted to non-invasive ventilation. Among these16 patients, 13 (27% of initial HFNC trial) were shifted to non-invasive ventilation because of HFNC failure and 3 patients (6.3% of initial HFNC trial) who were otherwise maintaining SpO₂ and PaO₂ complained severe headache and lacrimation due to high flow of HFNC, hence shifted to non-invasive ventilation.

In this study, one patient (0.83%) developed bilateral pneumothorax, which is in accordance with other retrospective COVID studies where reported incidence of pneumothorax is 1 to 2 % [8,13]. One patient was succumbed to massive haematemesis in this study reported incidence of upper GI bleed in COVID patients is 0.47% in a study conducted by Mauro A et al., [26]. In this study, two patients developed urethral bleeding and one epistaxis all were treated successfully by intravenous tranexamic acid and withholding LMWH. Bathula SSR et al., reported a case series of four COVID positive epistaxis patients, all were on LMWH [27].

Presence of co-morbid condition adversely affect outcome, as out of 17 COVID patient deaths, 16 patients had atleast one co-morbidity and only one non-co-morbid patient died due to development of bilateral pneumothorax. In this study, authors observed that presence of co-morbid condition is a poor prognostic factor and it is found to be statistically significant (p=0.001). In a study conducted by Sanyaolu A et al., it was reported that COVID-19 patients with co-morbidities were more likely to develop a more severe course and progression of the disease [28].

Male sex is also appeared to be poor prognostic factor in critically ill COVID patients. Only two death were reported in 30 critically ill female COVID patients (6.7%), whereas rest 15 death were among 90 critically ill male COVID patients (16.7%). However, it was not found to be statistically significant (p>0.05). Yanez ND et al., found that man had a higher risk of COVID 19 death than women [29].

There was only one death in under 40 years of age (mortality 6.7%), in 41 to 60 years of age mortality was 13.5% and in 61 years and above age group mortality was 17%. According to the study conducted by Yanez ND et al., the mortality rate of COVID-19 was strikingly higher in persons aged 65 years or older [29].

Compared to study conducted by Grasselli G et al., who reported 26% ICU mortality of critically ill COVID patients, mortality at the COVID dedicated ICU is 14.16% [21]. The lesser mortality in comparison to Italy is partially because we have younger population and partially because COVID outbreak was later in our locality so we had some knowledge regarding the nature and management of disease. At the centre mortality of critically ill COVID patients is decreasing with time.

Firstly, the patients comprised of small local population so data derived from this study is difficult to compare with other studies. Secondly, follow-up time is very less and the difficulties faced by the discharged patients some of which may have breathing problems are not known.

Among 17 COVID patient deaths, 16 patients had some co-morbidity and only one non-co-morbid patient died due to bilateral pneumothorax as a complication of invasive ventilation. Males were predominant among critically ill patients, mortality of male patients (16.7%) was higher than female patients (6.7%). Age specific mortality was highest (17%) in older patients, 61 years and above. Overall mortality of critically ill COVID patients at the ICU was 14.2%. From this study, it can be suggested that survival of critically ill COVID patients can further be improved by better management of their co-morbid conditions and avoiding complications of invasive ventilation. However, further multicentric studies with large sample size are needed to confirm these findings.