Factors related to survival in Intensive Care Unit patients with Covid-19: a study from a single center in Brazil

Introduction: Analysis of the outcome of 268 ICU patients in a single-center, as well the impact of viral infection on patients with preexisting medical conditions and how these factors affected survival and hospital stay. Methodology: Patients admitted to the ICU from March-August, 2020 were retrospectively analyzed under the same protocol at Hospital Alemão Oswaldo Cruz, São Paulo, Brazil. Several factors were considered and the results were presented using 95% confidence intervals. For statistical significance, p <0.05 was adopted. Results: Patient median age was 72 years, 64,2 years for discharged patients and 79.9 years for those deceased (p<0.001). The most common comorbidities were: systemic arterial hypertension, diabetes, thyroid disease, cardiovascular and kidney disease. Predictors of survival through univariate analysis: myalgia (p=0.001), cerebrovascular disease (p=0.002), COPD (p=0.003), dementia (p=0.000), mechanical ventilation (p=0.000), dialysis (0.000), vasopressor use (0.000), SAPS3 (0.000), lymphopenia (p=0.004), elevated D-dimer (P=0.011), time in ICU before tracheostomy (p=0.002), and performing a tracheostomy (p=0.000). The independent predictors of mortality were: advanced age (p=0.003) and tracheostomy performed in ICU (p=0.002). Discussion: COVID-19 affects usually older adults, where there already is a higher fatality rate. Acute respiratory distress syndrome is the primary cause of death and <5% of patients were reported as experiencing co-infection at admission. Conclusion: age, vasopressor use in patients with tracheostomy, and systemic coronary disease, heart failure, neoplasia, and COPD, were found to be significantly associated with COVID-19 severity.


Introduction
In late December 2019, a few cases of a serious illness causing pneumonia and death were first reported in Wuhan, China. Soon after, the number of cases increased dramatically, spreading across China and then worldwide.
The primary cluster of patients was found to be connected with the Huanan South China Seafood Market in Wuhan. Coronavirus belongs to the family Coronaviridae (subfamily Coronavirinae), species capable of infecting a broad range of hosts, causing symptoms and diseases ranging from the common cold to severe and ultimately fatal illnesses, such as SARS and MERS 1,5 . Analysis of the viral genome has revealed that the new coronavirus is phylogenetically close to severe acute respiratory syndrome coronavirus (SARS-CoV), the causative agent of a viral outbreak in 2002 5 .
The World Health Organization (WHO) announced the official name of the disease as "coronavirus disease 2019 (COVID- 19)" and now publicly refers to the virus as "the COVID-19 virus" (formerly known as "2019-nCoV") 1 . The consequences for human health, the global economy, and the normal functioning of society have been unprecedented 2 .
The virus infects humans of all age groups, although the severe form of the disease is more common in older adults 2 . Patients with COVID-19 can be symptomatic or asymptomatic. According to most reports, mild symptoms occurred in about 81% of patients, including cough, sore throat, fever, myalgia, and moderate pneumonia 2,4 .
As of 8 August 2021, there was a total of 201.941.078 confirmed cases globally. Of 192 countries registered cases that reached an outcome, 4.282.732 resulted in mortality 3  The disease can lead to organ dysfunction-shock, severe acute respiratory syndrome (SARS), acute cardiac injury, and acute kidney injury (AKI)-and death 2 . While most COVID-19 patients will not require supportive care, 10-15% of patients develop acute respiratory distress that requires invasive ventilatory support 1 . Patients with the severe form of the disease had symptoms of severe pneumonia, dyspnea, and very low blood oxygen saturation (≤ 93%), which was observed in about 14% of cases 7,8 .
Critical symptoms occurred in about 5% of cases and included respiratory failure, multiorgan failure, and septic shock. Mild infections were observed to improve in a week, whereas severe cases experienced acute respiratory failure, even leading to sepsis and death. Most fatalities were reported in middle-aged/elderly populations with preexisting conditions, including diabetes, heart and kidney diseases, chronic obstructive pulmonary disease, cancer, and immune diseases 7 .
The mortality rates are approximately 2%-3%. Older age, presence of comorbidities such as hypertension, diabetes, cardiovascular disease, chronic lung disease, cancer, higher d-dimer and C-reactive protein, and lower lymphocyte levels are associated with higher mortality 7,8 . Therefore, there is an urgent need for effective and specific antiviral treatment. Currently, supportive care measures such as ventilation oxygenation and fluid management remain the standard of care. Several clinical trials are currently trying to identify the most potent drug or combination against the disease and it is strongly recommended to enroll patients in ongoing trials 3,6 .
Mechanical ventilation (MV) for patients infected with the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is associated with prolonged airway intubation and high worldwide mortality of at least 50-67% 3 .
The development cycle of a vaccine production against SARS-CoV-2 moved remarkably fast given the major pandemic issue that has emerged and major international vaccine funding agencies are supporting the multitude of innovative ongoing efforts. Nowadays, vaccination is being applied around the world with optimistic results, reducing the number of infected people, hospitalizations, and deaths 6 .
In this light, the present original research describes the outcome in 268 ICU patients in a single-center, as well analyzing the effects of viral infection on preexisting medical conditions such as hypertension, diabetes, cardiovascular diseases, obesity, chronic obstructive pulmonary disease, kidney disease, cancer, and how these factors affected survival and hospital stay.

Methodology
We retrospectively analyzed patients included in this study (n=268), who were admitted to ICU between March 18th, 2020 to August 30th, 2020. All patients were analyzed under the same protocol at Hospital Alemão Oswaldo Cruz, São Paulo, Brazil.
The patients had COVID-19 infection documented by nasal pharyngeal swab for reverse transcriptase-polymerase chain reaction (rt-PCR) assay and developed severe respiratory failure requiring admission to ICU. Data were collected following a medical record review of each patient chart. The institutional review board approved this study (#4.849.542).
Patients with the following characteristics were considered eligible for analysis: 1) Covid-19 confirmed by test, 2) admission to ICU.
Several COVID-19 related factors were considered, including age, gender, symptoms before hospitalization, comorbidities, vasopressors use, radiological findings, use of high flow nasal catheter, prone position, ECMO, tracheostomy, and blood tests.
In addition, we collected variables such as type and duration of symptoms when entering ICU, FIO2 and PaO2/ FiO2 relation, and the total time of hospitalization.

Statistical Analyses
For database preparation and descriptive analysis, the Statistical Package for the Social Sciences software (SPSS Inc., Chicago, IL, USA), version 17.0 for Windows, was used.
The results were generated through tables and graphs. Categorical variables are expressed in frequencies and percentages -n (%). Continuous variables with normal distribution were expressed as means and standard deviations; and those with nonnormal distribution, in median and interquartile ranges. The normality of the numerical variables was verified through descriptive statistics, graphical analysis, and the Kolmogoronov-Sminov test.
In the comparison between groups (hospital discharge vs. death) for the numerical variables, the independent t-test was used, when the variables had a normal distribution and the Mann-Whitney test for those with asymmetric distribution. In the comparison between groups and categorical variables, the chi-square (X2) test was used, when the distribution had an N in each category with less than five individuals, Fischer's exact test was used. The COX regression model was used to evaluate the predictive variables for mortality in patients with COVID-19. After univariate analysis, independent variables were inserted into the COX model, remaining in the model if they remained significant (p <0.05). The manual procedure for inserting and removing variables was adopted. The results were presented using the hazard ratio and its respective 95% confidence intervals. The Kaplan-Mayer was used for patients' survival curves. For statistical inferences, p <0.05 was adopted for all analyses.

Results
Two hundred patients (74.6%) were discharged from the hospital (139 male and 34 female), and 68 patients (44 male and 24 female) were not discharged or died (25.4%).
There is a higher incidence of males in both groups. The longer the hospital stay, the greater the chance of death (figure 1).
The mean time duration of symptoms before hospital admission was 5.5 days, seven days (4-9) for discharged patients, and four days (2-6.75) for the patients not discharged from the hospital (p<0,001). The last group mentioned had a longer hospital stay, showing a faster unfavorable outcome for these patients. We noticed that shorter symptom duration usually progresses fastest to severe disease presentation.
The median age was 72 years, 64,2 years (53-74) for patients who were discharged, and 79.9 years (71.4-88.4) for those not discharged (p<0.001). The group of patients who was not discharged from the hospital was composed of the oldest ones (table 1).
Regarding lung affection before ICU admission, 21 patients had no specific CT SciELO Preprints -This document is a preprint and its current status is available at: https://doi.org/10.1590/SciELOPreprints.4453 The median days on mechanical ventilation was 12 (7.5-16), hospitalization total time of 17 (10-31) and 18.5 (10-37), and ICU time before tracheostomy 11.5 (7-18) and 19 (14.5-25.5), for those discharged and not discharged groups. M=Median; IIQ= Interval inter-quartil; m= mean; DP=standard deviation; * = Mannwhitney Test; ** =qui-quadrado Test; *** =T independente Test; α = exact fischer Test; From Table 1, collected data points towards: -Older age in the group that was not discharged from the hospital had worst prognosis; -A shorter duration of symptoms pre-hospitalization in the deceased group; -A greater incidence of myalgia in the group that was discharged from the hospital; -Concerning comorbidities, as comorbidities such as cardiovascular disease, neoplasia, COPD, and dementia, were highly prevalent among the group with more deceased patients; -During ICU stay, a higher prevalence of mechanical ventilation, vasopressor use, dialysis, and tracheostomy in the group that presented more deaths; -The SAPS3 was higher in patients who died at ICU admission; -Lymphopenia, elevated D-dimer, and the duration in the ICU pre-TQT were higher in the group with more deaths.
The univariate analysis (table 2)    factor (p=0.001); tracheostomy performed in ICU was a mortality predictor (p=0.002), as shown in (table 2 and 3).  26 . The most common clinical findings are fever, cough, and fatigue with some laboratory findings such as increased serum ferritin, D-dimers, and C reactive protein (CRP) 27 . Some studies reported that risk factors associated with the development of acute respiratory distress syndrome and death included older age, neutrophilia, organ dysfunction, coagulopathy, and elevated D-dimer levels 45 .
Previous ICU studies found mortality rates of 62% (China) and 67% (USA), but these figures had not accounted for many who were still in the ICU 41,42. . However, these numbers may reflect the pandemics beginning, a time when there was not much knowledge about the virus and disease outcome. Analyzed data found a lower mortality rate (25.4%), likely associated with improved healthcare due to literature on disease management and treatment.
Older adults are more affected and there is also a higher fatality rate in this subset of patients. Acute respiratory distress syndrome (ARDS) is the primary cause of death in COVID-19 28 and a recent scope review found that for COVID-19, < 5% of patients were reported as experiencing bacterial/fungal co-infection at admission, but the development of secondary infections during ICU admission is common 29,30 .
Age was correlated with increased mortality on univariate (p<0.001) and multivariate (p=0.003) analysis as well. Mortality rate was of 34% (n=68), although some authors reported mortality as high as 49% in patients with critical illness 35  Early reports have suggested an incubation period of two to 14 days, with clinical presentations ranging from mild infection to severe disease to fatal illness [9][10][11] . The most commonly reported symptoms are cough, fever, and dyspnea [12][13][14][15] . Myalgia and gastrointestinal symptoms, including diarrhea and nausea or vomiting, are also common 8 . Myalgia was found in 25% of patients and was a survival predictor on univariate analysis (p<0.001).
The need for mechanical ventilation (p=0.000) and longer ICU stay (p=0.002), were also correlated to a worse prognosis on univariate analysis. According to other authors, 97% of patients on invasive mechanical ventilation died in a multi-center study conducted early in the Wuhan outbreak, mortality is affected by local practices, and larger studies are awaited 43 . The same study reported that 53% of deaths were related to respiratory failure 43 .
Recent reports 12,15,16,17 suggest that approximately 14% to 29% of hospitalized patients with COVID-19 pneumonia require intensive care, primarily for respiratory support in the setting of hypoxic respiratory failure, with acute respiratory distress syndrome (ARDS) developing in 33% of hospitalized patients at a median time from symptom onset of eight days 9 . In these reports 12,13 , critically ill patients were older, more likely to be male and to have underlying comorbidities. The mortality rate ranged from 8.7% to 21% among those patients admitted with pneumonia 12,14-16 . These findings support the observations of earlier studies, which found a high percentage of hospitalized patients of advanced age with preexisting conditions, hypertension being the most common [18][19][20] .
COVID-19 rapidly spread throughout the state of São Paulo and has disproportionately affected the population, who have high rates of co-morbid conditions and a mean BMI of 30. The obese patients had a high incidence of unfavorable outcomes, as reported previously by other authors. In the present study, obesity was present in 27.9% of the deceased group. Although it is a considerable number of patients, it was not important in univariate analysis (p=0.748).
In the initial reports from Wuhan, China, during the early stages of the pandemic, shortness of breath was reported in 54% of patients and was associated with a composite endpoint of admission to an ICU, use of mechanical ventilation, and death 19 .
A similar prevalence of dyspnea was reported in 21 critically ill patients in Washington State and the COVID-19-Associated Hospitalization Surveillance Network database 21 .
In our series, dyspnea at presentation was associated with hospitalization and the need for ICU management -it was the most prevalent symptom, especially for those who were not discharged (n=55/80.9%). The need for vasopressors drugs might be due to one or the combination of such factors: muscle blockers, septic shock, myocarditis, or other myocardial dysfunctions. The scenario with such findings denotes severely ill patients. The need for vasopressors had an impact on survival on multivariate analysis (p=0.000).
Tracheostomy is a common procedure in critically ill patients who require an extended period of MV. The use of tracheostomy can facilitate weaning from MV and potentially increase the availability of intensive care unit (ICU) beds. Analysis shows that 11% of the patients who were discharged from the hospital had tracheostomy performed in the ICU. On the other hand, 30.9% (21/68) of the deceased group underwent tracheostomy, showing a higher incidence of such procedure in the worse patients group. According to collected data and previous publications, performing tracheostomy did not impact the natural history of these patients, although those who underwent the procedure had a worse prognosis (p=0.000). On the multivariate analysis, the patients who had tracheostomy had an impact on survival as well (p=0.002). This could be explained by tracheostomy performed in patients in sub-optimal conditions for the procedure, advanced age, those who had fewer conditions for MV weaning, or with more comorbidities. There are no guidelines for COVID-19 patients in MV who should undergo or not to tracheostomy in current literature.

Conclusions
In conclusion, age, the need for vasopressor medications in patients who underwent tracheostomy, and underlying comorbidities, such as systemic coronary disease, heart failure, neoplasia, and COPD, were found to be significantly associated with COVID-19 severity. These pre-existing conditions could increase the susceptibility of such individuals to COVID-19. Recognizing these risk factors could help clinicians reduce mortality by identifying patients with poor prognoses at an early stage. Lymphopenia and elevated D-dimer were also related to bad prognosis and worse disease course. A long time in ICU in this cohort was a bad prognosis marker as well.
The spectrum of described diseases in this study are similar to those from the Centers for Disease Control and Prevention's COVID-19-Associated Hospitalization Surveillance Network, the New York City area, and China. Moreover, data analysis confirmed that shorter symptom duration is associated to worst disease prognosis.