Caspian Journal of

Introduction 
The acute respiratory disease stemming from the coronavirus disease 2019 (COVID-19) was declared a public health emergency worldwide by the World Health Organization (WHO) [1]. This virus was first identified in Wuhan, China, during the investigations on the disease outbreak [2]. In Iran, the first case was reported on February 17, 2020, reaching more than 205000 people until June 21, 2020, based on the official statistics [3]. The severity of the respiratory disease resulting from this virus can range from mild influenza to severe pneumonia, respiratory failure, and death [4]. Patients with mild symptoms usually improve within a week after a proper clinical intervention, whereas those with the severe form of the disease experience progressive respiratory distress due to the alveolar damage resulting from the virus, which may finally culminate in death [5]. Patients with the moderate type of the disease, particularly the elderly or individuals with underlying diseases, experience worse conditions and quickly move toward the severe type and crisis of the disease [6]. During the epidemic, because of the relatively unknown nature and high contagion speed of the disease, the speed in providing services became more important, and the need for providing medical health care heightened at once [7]. In such conditions, one of the wards influenced by the critical conditions due to COVID-19 was the intensive care unit (ICU), so the rate of hospitalization of patients with COVID-19 in the ICU was reported to be 30% [8]. The results of Kim’s study in America also indicated that about one-third of patients with COVID-19 needed to be hospitalized in the ICU [9]. The quick increase in the number of patients and their need for mechanical ventilation to prevent disease progression and respiratory failure highlighted the need for ICU wards around the world [10]. At present, also, after the declaration of the end of the COVID-19 pandemic by the WHO, due to the high costs and limited resources and ICU beds, it should be acted in such a way that these limited resources are provided to patients based on priority [11]. Hence, predictors for the patient’s requirement to be hospitalized in the ICU should be identified because the identification of these factors is a key component in providing a quicker and more precise diagnosis and the patient’s priority for hospitalization in these wards [12]. The importance of this issue is enhanced considering the virus’s rapid contagion and the long incubation period [13]. On the other hand, early and precise diagnosis of COVID-19 can lower the healthcare system burden [14].
Although various studies have dealt with assessing the predictors of the ICU hospitalization of patients with COVID-19, based on the researchers’ search, no study has considered all individual, clinical, laboratory, and respiratory factors to determine the likelihood of these patients’ ICU hospitalization. Furthermore, the results of related studies were also discrepant. The aim of the study is to determine individual, clinical, laboratory, and respiratory factors predicting the ICU hospitalization of patients with COVID-19.

Materials and Methods

Study type and population
This retrospective cross-sectional study, was performed on 386 patients with COVID-19 referring to Razi Hospital, affiliated with Guilan University of Medical Sciences, Rasht, Iran. Data were collected retrospectively by reviewing the medical records of patients meeting the inclusion criteria from December 20, 2020, to May 18, 2021. The sample size required to determine predictors of intensive care unit admission in patients with COVID-19 was determined to be 151 patients based on the predictive value of age on intensive care unit admission (odds ratio [OR]=3.35) in a pilot study with 95% confidence interval and 90% power. In addition to age, the effect of other individual, clinical, laboratory, and respiratory factors was also examined, therefore, considering 5 samples for each variable, the total number of samples to determine predictors of intensive care unit admission was determined to be 386 patients. The inclusion criteria were age 18 years, a definite diagnosis of COVID-19 based on clinical symptoms, a positive real-time polymerase chain reaction (RT-PCR) test from a nose or pharynx swab sample, and clinical findings in computed tomography (CT) scan and chest radiography, and having a complete medical record regarding the required information.

Study instruments and variables
The data collection tool included demographic information such as age, gender, height, weight, and blood type, clinical information including the patient’s symptoms at the time of admission (shortness of breath, cough, sputum, fever, headache, body pain, fatigue, gastrointestinal symptoms, chest pain, loss of sense of smell, a history of diabetes, blood pressure, cardiovascular disease, asthma, chronic lung disease, cancer, kidney disease, autoimmune disease, and smoking), a history of underlying diseases, taking medicines and their type, smoking, and laboratory tests such as complete blood count [CBC], biochemical tests, procalcitonin [PCT], platelet-to-lymphocyte ratio [PLR], lactate dehydrogenase [LDH], erythrocyte sedimentation [ESR], C-creative protein [CRP], neutrophil-to-lymphocyte ratio [NLR], sodium [Na], and potassium [K]), and respiratory indices (arterial blood gases [ABGS], oxygen saturation [SpO2], pressure of arterial oxygen to fraction of inspired oxygen [PaO2/FiO2], respiratory rate [RR], and SpO2 to FiO2 [SpO2/FiO2]). Body mass index (BMI) was calculated based on weight in kilogram divided by height in square of meter.

Statistical analysis
In order to determine the individual, clinical, laboratory, and respiratory factors of ICU hospitalization, the single-variable chi-square test and Fisher’s exact test, and for quantitative predictors in case of meeting normal distribution assumption, the independent t-test and otherwise, the Mann-Whitney U test were used. Stepwise logistic regression was used to determine adjusted OR with 95% confidence interval. The significance level of the tests was considered P<0.05. All statistical analysis were performed using SPSS software, version 16.

Results
Of total, 330 patients had a complete hospital record and included for the analysis. The overall hospitalization rate was 17% (56 out of 330). The majority of the patients were in the age group of 50-70 years (46.4%) and in terms of gender, the percentage of men and women was almost equal. The body mass index of nearly half of the patients was overweight. In terms of blood group, the majority of the COVID-19 patients, had blood group A (34.5%) and the majority were RH positive. According to Table 1, the percentage of hospitalization in the intensive care unit was significantly associated with sex (P=0.015) and BMI (P=0.017). The percentage of hospitalization in the male was significantly higher than that in the female (19% versus 10%). The rate of ICU hospitalization was highest among morbidly obese patients (42.9%) followed by overweighs (19%). Patients who were taking anti-asthma and pulmonary drugs, as well as those with a history of smoking had a higher probability of hospitalization than the others. The mean score of white blood cells, PLR, hemoglobin, LDH, and NLR were higher in ICU hospitalized patients. However, the mean and median scores of lymphocytes were lower in ICU hospitalized patients. The mean RR score was higher in ICU hospitalized patients than in patients not hospitalized in the ICU.



The levels of white blood cells (P=0.003), hemoglobin (P=0.05), PLR (P=0.002), lymphocytes (P<0.001), LDH (P<0.001), and NLR (P<0.001) were significantly associated with hospitalization in the intensive care unit admission. The mean and median white blood cells, PLR, hemoglobin, lactate dehydrogenase, and NLR were higher in the group of patients admitted to the intensive care unit than in COVID-19 patients who were not admitted to the intensive care unit. However, the mean and median lymphocytes in patients admitted to the intensive care unit were lower than in patients who were not admitted to the intensive care unit (Table 2).



Table 3 shows demographic and clinical predictors of hospitalization among COVID-19 patients. Among demographic variables, gender was significantly associated with hospitalization. Males were hospitalized in the ICU 2.1 times more than females (OR=2.06, P=0.016). Symptoms of shortness of breath (OR=9.5, P=0.002), taking anti-asthma and pulmonary drugs (OR=3.5, P=0.001), and a history of smoking (OR=3.4, P=0.003) were found as significant predictors related to ICU hospitalization in the final model so that the presence of shortness of breath elevated the odds of hospitalization in the ICU by 9.5 times, anti-asthma and pulmonary drugs by 3.5 times, and a history of smoking by 3.4 times. 



Respiratory and laboratory predictors of hospitalization are shown in Table 4. Among the investigated respiratory parameters, RR (OR=1.1, P=0.002), peripheral blood SpO2 (OR=0.919, P=0.013), and arterial oxygen relative pressure ratio to inspiratory oxygen fraction (OR=0.974, P<0.001) were considered predictors related to ICU hospitalization. In the final logistic regression model, neutrophils (OR=1.123, P<0.001) and LDH (OR=1.001, P<0.001) were considered laboratory predictors related to ICU hospitalization (Table 4).



In the final model considering individual-social, laboratory, respiratory, and clinical variables fever (OR=4, P=0.005), a history of smoking (OR=6.5, P=0.002), RR (OR=1.2, P=0.004), and arterial oxygen relative pressure ratio to inspiratory oxygen fraction (OR=0.974, P<0.001) were the most important predictors related to ICU hospitalization in patients with COVID-19 so that in the final model, a history of smoking and RR were considered risk factors for ICU hospitalization and arterial oxygen relative pressure ratio to inspiratory oxygen fraction was considered a protective variable of ICU hospitalization, which the presence of fever increased the risk of ICU hospitalization by 4 times, a history of smoking by 6.5 times, and the increased RR by 1.2 times. However, by increasing the arterial oxygen relative pressure ratio to inspiratory oxygen fraction, the risk of ICU hospitalization decreased (Table 5).



Discussion
The results of this study indicated that gender and BMI had statistically significant relationships with ICU hospitalization, so the percentage of ICU hospitalized patients was higher in the group of males and patients with an excessively obese BMI. These findings were in line with Izquierdo et al.’s [15], Hatami et al.’s [3], and Ciceri et al.’s [16] studies. Herrera and Lesmes’s [17] study also revealed that obesity-related factors led to an increased risk of contracting COVID-19 and ICU hospitalization. Obesity involves a low-grade pro-inflammatory state, causing an impaired immune system and reducing its ability to respond to the respiratory infection of COVID-19 [18].
In the present study, the probability of ICU hospitalization was higher for patients with symptoms of shortness of breath. In this regard, based on Jain and Yuan’s study [19], cough (67.2%), fever (62.9%), and shortness of breath (61.2%) were the most common symptoms in the ICU hospitalized group. Also, in Hatami et al.’s [3] study, the most common clinical symptoms of ICU hospitalized patients included cough (62.6%), fever (55.9%), and shortness of breath (52.6%) in all patients.
The results of the current study demonstrated that the probability of ICU hospitalization was higher among patients taking anti-asthma and pulmonary drugs, as well as those with a history of smoking. Based on the results of Liu et al.’s study, smoking culminates in the progression of COVID-19 [20]. In their systematic review, Reddy et al. [21] reported that current smokers and patients with a history of smoking experienced an increased risk of contracting severe COVID-19, the need for mechanical ventilation and ICU hospitalization, and mortality. The increased intensification of infection in smoking patients can be attributed to the positive regulation of the angiotensin-converting enzyme 2 (ACE2) receptor, which is the main receptor for the COVID-19 virus to enter the host’s mucosa [22].
The mean and median scores of white blood cells, PLR, hemoglobin, LDH, and NLR were higher in ICU hospitalized patients. The results of our study were in line with Hatami et al.’s [3] and Carlino et al.’s [23] studies.
The mean RR score was higher in ICU hospitalized patients, but not the peripheral blood SpO2 and the arterial oxygen relative pressure ratio to inspiratory oxygen fraction, which are consistent with Hatami et al.’s [3], Ciceri et al.’s [16], and Carlino et al.’s [23] studies. 
Based on the current research results, among the individual-social, laboratory, respiratory, and clinical variables in the final model, fever, a history of smoking, RR, and arterial oxygen relative pressure ratio to inspiratory oxygen fraction were the most important predictors related to ICU hospitalization in patients with COVID-19. 
Among the respiratory parameters studied, RR, peripheral blood SpO2, and arterial oxygen partial pressure to inspiratory oxygen fraction ratio were considered as predictors associated with intensive care unit admission. Increasing RR increases the chance of admission, and increasing arterial oxygen partial pressure to inspiratory oxygen fraction ratio will reduce intensive care unit admission. In a study conducted by Ciceri et al. as primary predictors of COVID-19 clinical outcomes, mean SpO2 (SpO2=93% (range 60-99), with a PaO2/FiO2 ratio=267, were predictive factors [16]. Also, in a study conducted by Carlino et al., it was shown that patients in the ICU group had lower SpO2, PaO2, PaCO2, P/F, and higher RR [23], which is in line with the present study. In the present study, among the respiratory parameters studied, RR, peripheral blood SpO2, and the ratio of arterial oxygen partial pressure to inspiratory oxygen fraction were considered as predictors associated with intensive care unit admission. With increasing RR, the chance of admission increases, and increasing the ratio of arterial oxygen partial pressure to inspiratory oxygen fraction will reduce intensive care unit admission. In the study by Zhao et al., increasing RR and decreasing SpO2, smoking history, and increasing LDH and consumption Smoking was significantly associated with intensive care unit admission [24], which is consistent with the present study. In the present study, RR, the ratio of partial pressure of arterial oxygen to fraction of inspiratory oxygen are among the most important predictors of intensive care unit admission in patients with COVID-19. Among the laboratory parameters studied in the final logistic regression model, neutrophils and LDH were considered as laboratory predictors associated with intensive care unit admission. In a retrospective study by Zhao et al. in New York, the five main predictors of intensive care unit admission were identified as increased LDH and PCT, smoking history, decreased blood oxygen, and lymphocyte count [24]. The case of increased LDH in the laboratory examination section of the present study is consistent with it. In the study by Allenbach et al. in France, older age, poorer respiratory manifestations, higher C-reactive protein levels, and lower lymphocyte counts were associated with an increased risk of intensive care unit admission or death [25]. However, in the present study, neutrophils and LDH were considered laboratory predictors associated with intensive care unit admission. Among the sociodemographic variables and laboratory and respiratory parameters, as well as clinical variables in the final overall model, fever, smoking history, RR, and the ratio of arterial oxygen partial pressure to inspiratory oxygen fraction were among the most important predictors associated with intensive care unit admission in patients with COVID-19. In the study by Izquierdo et al., a combination of three clinical variables, namely age, temperature, and R, was the most important predictor of intensive care unit admission in patients with COVID-19 [15], which is in line with the present study. In a study by Chow et al. to evaluate a clinical tool for predicting critical illness and intensive care unit needs, the following factors were associated with the highest risk of critical illness and intensive care unit admission: Number of underlying diseases, body mass index, RR, white blood cell count, and percentage of lymphocytes, serum creatinine, and LDH [26]. This was consistent with the present study. According to current knowledge, those with higher RRs at the time of admission are at increased risk of severe outcomes [15]. Furthermore, according to previous studies, %SpO2<90 was the strongest predictor of ICU admission [27]. These findings support the notion that lung involvement in the form of pneumonia (interstitial inflammation and changes in alveolar ventilation) is the main pathophysiological mechanism of the disease and, consequently, hypoxemia in ICU patients [28].

Conclusion
The results of the current study revealed that among the individual-social, clinical, laboratory, and respiratory variables, fever, a history of smoking, high RR, and arterial oxygen relative pressure ratio to inspiratory oxygen fraction were the most crucial predictors related to ICU hospitalization in patients with COVID-19. Since the determination of predictive factors is a critical factor in diagnosing the disease prognosis and taking timely and proper treatment actions, empowering nurses is, therefore, one of the most important predictors of ICU hospitalization of patients with COVID-19.

Ethical Considerations

Compliance with ethical guidelines
The research was approved by the Research Ethics Committee of Guilan University of Medical Sciences, Rasht, Iran (Code: IR.GUMS.REC.1396.251).

Funding
The paper was extracted from the master's thesis of Yaser Zolfi Parsheh, approved by the Department of Nursing, School of Guilan University of Medical Sciences, Rasht, Iran. This research was supported by the research project, funded by the Guilan University of Medical Sciences, Rasht, Iran (Grand No.: 2872). 

Authors' contributions
Conceptualization and supervision: Yaser Zolfi Parsheh, Masoomeh Adib Rahim Abadi, and Mohammad Haghighi; Methodology: Tahereh Khaleghdoost Mohamadi; Data collection: Yaser Zolfi Parsheh;Data analysis: Ehsan Kazemnezhad Leyli; Funding acquisition and resources: Masoomeh Adib Rahim Abadi; Investigation, and writing: All authors. 

Conflict of interest
The authors declared no conflict of interest.

Acknowledgements
The researchers sincerely appreciate the cooperation and support of the Department of Research and Technology of Guilan University of Medical Sciences, Rasht, Iran. The authors would like to thank the patients who had fun with them.



References