Introduction
In December 2019, an unexpectedly high number of cases of atypical pneumonia with aggressive symptoms were observed in the Wuhan region of China. It was determined that the causative agent was a novel coronavirus not previously seen in humans and this virus was named Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). SARS-CoV-2, primarily transmitted from person to person, especially through the respiratory tract. The virus rapidly continued to spread, and on March 11, 2020 it was officially recognized as a 'pandemic' by the World Health Organization [1]. The clinical condition caused by the virus is named Coronavirus Disease 2019 (COVID-19) and typically consists of respiratory system symptoms that vary in severity. Patients may present with symptoms such as sore throat, runny nose, fever, cough, myalgia, arthralgia, headache, diarrhea and in severe cases, acute respiratory distress syndrome (ARDS), septic shock, multiple organ failure and even death [1, 2].
Studies have shown that approximately 80 % of patients diagnosed with COVID-19 develop a mild to moderate form of the disease while around 20 % of patients develop a severe form and about 4–6 % of patients require intensive care [3]. Since the beginning of the pandemic, the high number of admissions to emergency departments has made it necessary to promptly identify the critical patient group requiring critical care and refer them to intensive care.
In this study, patients were admitted to the emergency services’ intensive care unit with possible COVID-19 were evaluated with SARS-CoV-2 polymerase chain reaction (PCR) results, laboratory parameters, comorbidities, Acute Physiology and Chronic Health Evaluation-2 Scores (APACHE II), Sequential Organ Failure Assessment Scores (SOFA) and Pneumonia Severity Index Scores (PSI) to identify the similar characteristics of the critical patient group and determine the parameters that enable early detection.
Materials and methods
According to information obtained from the Hospital Information Management System, it was determined that 435 patients were admitted to the intensive care unit between January 1, 2021 and December 31, 2021. However, three of these patients were excluded due to under the age of 18 and 35 of them were excluded due to missing data. Total 397 patient data were retrospectively examined (fig. 1). Data of the included patients were accessed from both electronic records and intensive care patient files.Informed consent was obtained from all the patients, their families or legal representatives for their anonymized information to be published in this article.
Fig. 1. Flow Chart
The patients admitted to the intensive care unit were divided into two groups based on their PCR results, positive and negative. PCR negative group had at least 2 negative PCR results and negative COVID-19 antigen tests to rule out COVID-19 and they mostly diagnosed bacterial or influenza pneumonia. 28-day mortality difference between negative and positive groups was defined as the primary outcome. The impact of age, gender, comorbidities, laboratory values, SOFA, APACHE II and PSI scores on mortality was set as the secondary outcome.
Information regarding the patients’ medical history and comorbidities was obtained from the patient themselves, their relatives and by scanning the patient's history through the hospital information system. It was also obtained from the e-Nabız application of the Turkish Republic Ministry of Health. The PCR results of the patients were obtained from the Public Health Management System. Laboratory tests for the patients included complete blood count values, such as white blood cell count (WBC), neutrophils (Neu), lymphocytes (Lym), platelets (Plt), hemoglobin (Hgb), hematocrit (Hct), as well as kidney function tests [urea, blood urea nitrogen (BUN), creatinine], plasma electrolytes [sodium(Na), potassium(K)], glucose, liver enzymes [Aspartate Transaminase (AST), Alanine Transaminase (ALT)], total bilirubin, C-reactive Protein (CRP), D-dimer, and Ferritin. Blood gas analysis included pH, pO₂, pCO₂, HCO₃, and lactate measurements. Information about the need for mechanical ventilation, the fraction of inspired oxygen (FiO₂) being administered, patient vitals including systolic blood pressure, diastolic blood pressure, mean arterial pressure, oxygen saturation, heart rate, temperature and respiratory rate was obtained from the patients’ intensive care follow-up file or the hospital information system.
In the study, the initial test results obtained from patients who underwent multiple laboratory tests in the intensive care unit, the latest PCR results obtained and the initial vital signs were included. The mechanical ventilation status and FiO₂ value were determined within the first hour of the patient's admission to the intensive care unit.
Data collected from patients were used to calculate the neutrophil/lymphocyte ratio, platelet/lymphocyte ratio and neutrophil/platelet ratio in Microsoft Excel.
When evaluating comorbid conditions of the patients, those with a history of ischemic and hemorrhagic cerebrovascular events (CVE) were categorized as the CVE group, patients with known malignancies receiving radiotherapy, chemotherapy or those not undergoing treatment by their choice were categorized as the malignancy group. Patients with chronic neurological diseases such as Alzheimer's, Parkinson's, epilepsy or psychiatric histories such as anxiety disorders, major or minor depression, bipolar mood disorders, obsessive-compulsive behavior disorders were categorized under neuropsychiatric disorders group.
Using the data collected from the patients, APACHE II, SOFA and PSI scores were calculated from June 24 to June 26, 2024 through https://www.uptodate.com/contents/search?search=calculators.
The outcomes of the patients were determined using information obtained from the hospital information system of patient. For patients transferred to another intensive care units or wards, the outcomes were tracked through the hospital information system.
For statistical analysis, Jamovi version 2.3 was used. The normality of the data was tested using the Shapiro-Wilk test. For descriptive statistics, quantitative variables with a normal distribution were expressed as mean ± standard deviation. Quantitative variables with a non-normal distribution were expressed as median (interquartile range 25–75) and categorical variables were expressed using counts and percentages. Normal distributed variables were compared using the Student's t-test, non-normally distributed quantitative variables were compared using the Mann-Whitney U test and categorical variables were compared using the chi-square test or Fisher's exact test. Sensitivity of numerical data in predicting mortality was assessed using Receiver-operating characteristic ROC analysis. Subsequently, Area Under The Receiver Operating Characteristic (AUC) values were calculated. Youden’s index was used to determine optimal cutoff points. Positive likelihood ratio (PLR) and negative likelihood ratio (NLR) were calculated for each scoring system at this cutoff points. DeLong Test is performed to determine of difference between AUCs. Results with a p-value below 0.05 were considered statistically significant.
This study was deemed scientifically and ethically appropriate by the Kutahya Health Sciences University Non-Interventional Clinical Research Ethics Committee with decision number 37686 dated January 27, 2022 (Annex-1).
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Results
Among the 397 patients, 286 (72 %) had negative PCR results while 111 (28 %) had positive PCR results. In the negative group the median age was 72.50 (65–82) while in the positive group the median age was 73.00 (64–82) (p > 0,05). Among the patients in the negative group, 163 (57.0 %) were male and in the positive group 60 (54.1 %) patients were male (p > 0,05). The distribution of demographic data of the patients is presented in table 1.
Cerebrovascular event (CVE) and malignancy comorbidities were significantly higher in the PCR negative group. There is no statistically significant difference between the PCR groups for other comorbidities included hypertension (HT), diabetes mellitus (DM), coronary artery disease (CAD), congestive heart failure (CHF), atrial fibrillation (AF), acute or chronic renal failure (ARF/CRF), asthma or chronic obstructive pulmonary disease (COPD) and neuropsychiatric disease. The distribution of patients' comorbidities in the PCR groups is provided in table 1.
| Demographic data/Comorbidities | PCR– n = 286 (100 %) |
PCR+ n = 111 (100 %) |
p-value | |
|---|---|---|---|---|
| Age Median (IQR 25–75) |
72.50 (65-82) | 73 (64–82) | p = 0,99 | |
| Gender n, (%) |
Male | 163 (57 %) | 60 (54.1 %) | p = 0,59 |
| Female | 123 (43 %) | 51 (45.9 %) | ||
| HT | 140 (49 %) | 56 (50.5 %) | p = 0,78 | |
| CAD | 87 (30.4 %) | 27 (24.3 %) | p = 0,22 | |
| AKF/CKF | 108 (37.8 %) | 45 (40.5 %) | p = 0,61 | |
| DM | 106 (37.1 %) | 36 (32.4 %) | p = 0,38 | |
| CVE | 58 (20.3 %) | 9 (8,1 %) | p = 0,004 | |
| Malignancy | 57 (19.9 %) | 9 (8,1 %) | p = 0,005 | |
| CHF | 52 (18.2 %) | 13 (11,7 %) | p = 0,11 | |
| AF | 34 (11.9 %) | 12 (10,8 %) | p = 0,90 | |
| Asthma/COPD | 78 (27.3 %) | 29 (26,1 %) | p = 0,81 | |
| Neuropsychiatric disease | 34 (11.9 %) | 14 (12,6 %) | p = 0,84 | |
| Pearson’s Chi-Squared Test | ||||
Out of the 286 patients with a negative PCR result, 163 (57.0 %) deceased while 123 (43.0 %) survived. Among the 111 patients with a positive PCR result, 75 (67.6 %) deceased while 36 (32.4 %) survived. There was not significant difference in mortality between the PCR groups (p > 0.054). The distribution of patients' PCR results among outcome groups is presented in table 2.
| Outcome | p value | |||
|---|---|---|---|---|
| PCR Results | Deceased | Survived | Total | |
| Negative | 163 (57.0 %) | 123 (43.0 %) | 286 (100 %) | p = 0,054 |
| Positive | 75 (67.6 %) | 36 (32.4 %) | 111 (100 %) | |
| Total | 238 (59.9 %) | 159 (40.1 %) | 397 (100 %) | |
| Pearson’s Chi-Squared Test | ||||
In the PCR positive group, there was no significant difference in the median values of hematocrit, potassium, glucose and platelet/lymphocyte (Plt/lym) ratio between survivors and non-survivors (p > 0.05). Non-survivor group had higher median values of WBC, Neu, urea, BUN, creatinine, sodium, AST, ALT, CRP, D-Dimer, Neu/Lym ratio, and Neu/Plt ratio and lower median values of lymphocyte count, platelet count and hemoglobin level (p < 0.05). The examination of laboratory values and outcomes in the PCR-positive group is presented in table 3.
| Parameters | Outcome | n | Median | IQR (25–75 %) | p value |
|---|---|---|---|---|---|
| WBC (× 10³/uL) |
Deceased | 75 | 15.7 | (9.68–19.70) | < 0.001 |
| Survived | 36 | 7.46 | (6.09–10.25) | ||
| Neu (× 10⁹/L) |
Deceased | 75 | 13.25 | (7.87–17.70) | < 0.001 |
| Survived | 36 | 5.94 | (4.66–8.61) | ||
| Len (× 10⁹/L) |
Deceased | 75 | 0.64 | (0.36–1.06) | 0.020 |
| Survived | 36 | 0.8750 | (0.64–1.16) | ||
| Plt (× 10³/uL) |
Deceased | 75 | 164.00 | (117.0–244.0) | 0.003 |
| Survived | 36 | 208.50 | (188.75–303.50) | ||
| Hgb (g/dL) |
Deceased | 75 | 10.40 | (9.10–12.20) | 0.023 |
| Survived | 36 | 11.95 | (10.05–13.67) | ||
| Hct (%) |
Deceased | 75 | 34.00 | (29.10–38.20) | 0.104 |
| Survived | 36 | 37.00 | (32.07–42.50) | ||
| Urea (mg/dL) |
Deceased | 75 | 113.0 | (69.0–161.00) | < 0.001 |
| Survived | 36 | 49.50 | (34.50–76.75) | ||
| BUN (mg/dL) |
Deceased | 75 | 53.0 | (32.00–75.00) | < 0.001 |
| Survived | 36 | 23.0 | (16.25–36.00) | ||
| Creatinin (mg/dL) |
Deceased | 75 | 1.73 | (1.05–3.07) | < 0.001 |
| Survived | 36 | 1.045 | (0.77–1.26) | ||
| Sodium (mmol/L) |
Deceased | 75 | 143.00 | (138.0–140.75) | < 0.001 |
| Survived | 36 | 138.00 | (136.25–140.75) | ||
| Potassium (mmol/L) |
Deceased | 75 | 4.39 | (3.70–5.20) | 0.163 |
| Survived | 36 | 4.025 | (3.65–4.58) | ||
| Glucose (mg/dL) |
Deceased | 75 | 184.00 | (120–244) | 0.209 |
| Survived | 36 | 145.00 | (97.50–259.75) | ||
| AST (IU/L) |
Deceased | 75 | 43.00 | (26.00–88.00) | 0.020 |
| Survived | 36 | 29.50 | (22.00–51.50) | ||
| ALT (IU/L) |
Deceased | 75 | 33.00 | (19.00–61.00) | 0.023 |
| Survived | 36 | 21.00 | (15.25–34.50) | ||
| CRP (mg/L) |
Deceased | 75 | 188.40 | (84.00–241.20) | < 0.001 |
| Survived | 36 | 46.09 | (17.04–159.75) | ||
| D-dimer (mcg/L) |
Deceased | 75 | 4140.00 | (1848–4481) | < 0.001 |
| Survived | 36 | 1417.50 | (784.75–2898) | ||
| Neu/lym | Deceased | 75 | 19.49 | (9.90–37.47) | < 0.001 |
| Survived | 36 | 7.37 | (3.70–12.88) | ||
| Plt/lym | Deceased | 75 | 284.44 | (130.00–541.37) | 0.701 |
| Survived | 36 | 277.77 | (180.02–367.75) | ||
| Neu/Plt | Deceased | 75 | 0.07 | (0.04–0.12) | < 0.001 |
| Survived | 36 | 0.02 | (0.02–0.04) | ||
| Mann-Whitney U Test | |||||
In the PCR positive group, the patients with ARF/CRF diagnosis had significantly higher mortality (p < 0.001). In patients with HT, CAD, CHF, AF, asthma/COPD, DM, CVE, malignancy and neuropsychiatric disease, there is no significant difference among outcome groups (p > 0.05). The relationship between comorbidities and outcomes in PCR-positive patients is presented in table 4.
| Comorbidities | Outcome | Total | p | |
|---|---|---|---|---|
| Deceased n = 75 (100 %) |
Survived n = 36 (100 %) |
|||
| HT | 37 (49.3 %) | 19 (52.8 %) | 56 (100.0 %) | 0.734 |
| CAD | 15 (20.0 %) | 12 (33.3 %) | 27 (100.0 %) | 0.125 |
| CHF | 9 (12.0 %) | 4 (11.1 %) | 13 (100.0 %) | 0.892 |
| AF | 9 (12.0 %) | 3 (8.3 %) | 12 (100.0 %) | 0.560 |
| Asthma/COPD | 23 (30.7 %) | 6 (16.7 %) | 29 (100.0 %) | 0.116 |
| ARF/CRF | 40 (53.3 %) | 5 (13.9 %) | 45 (100.0 %) | < 0.001 |
| DM | 22 (29.3 %) | 14 (38.9 %) | 36 (100.0 %) | 0.314 |
| CVE | 6 (8.0 %) | 3 (8.3 %) | 9 (100.0 %) | 0.952 |
| Malignancy | 7 (9.3 %) | 2 (5.6 %) | 9 (100.0 %) | 0.715 |
| Neuropsychiatric disease | 11 (14.7 %) | 3 (8.3 %) | 14 (100.0 %) | 0.543 |
| Pearson’s Chi-Squared Test and Fisher’s Exact Test | ||||
Among PCR-negative patients, the medians of APACHE II, SOFA and PSI scores in the deceased group were significantly higher (p < 0.001). Similarly, among PCR-positive patients, the median values of the APACHE II, SOFA and PSI scores in the deceased group were significantly higher (p < 0.001).
ROC analysis was performed to determine the relationship with APACHE II, SOFA and PSI scores of all patients and outcome groups. Then AUCs calculated. According to the Youden’s index, the optimal cutoff points were determined as 19 points for APACHE II (AUC = 0.96, p = 0.0001), 9 points for SOFA score (AUC = 0.96, p = 0.0001) and 81 points for PSI score (AUC = 0.91, p = 0.0001) (fig. 2). The cutoff points are significant based on sensitivity, specificity and the level of AUC. The optimal cutoff points for APACHE II, SOFA and PSI scores for all patients with sensitivity, specificity, positive likelihood ratio and negative likelihood ratio are presented in table 5. There were no significant difference between AUCs of APACHE II and SOFA score (DeLong Test, p = 0.850). The AUCs of PSI was significantly lower from AUC of APACHE II with AUC difference 0,050 (p < 0,001) and SOFA score with AUC difference 0,049 (p < 0,001).
Fig. 2. Receiver operating characteristic curve for the predicted value of APACHE II, SOFA score, and PSI scores for the mortality
| Scoring System | Cutoff Points | Sensitivity 95% CI |
Specificity 95% CI |
PLR 95% CI |
NLR 95% CI |
AUC 95% CI |
*p |
|---|---|---|---|---|---|---|---|
| APACHE II | 19 | 0.89 (0.84–0.93) |
0.94 (0.89–0.97) |
14.16 (7.76–25.8) |
0,12 (0.08–0.17) |
0,96 (0.95–0.98) |
0.0001 |
| SOFA | 9 | 0.89 (0.84–0.93) |
0.96 (0.91–0.98) |
20.23 (9.79–41.8) |
0,11 (0.08–0.16) |
0,96 (0.94–0.98) |
0.0001 |
| PSI | 81 | 0.79 (0.74–0.83) |
0.89 (0.83–0.93) |
7.01 (4.52–10.9) |
0.23 (0.18–0.30) |
0.91 (0.88–0.94) |
0.0001 |
| ROC analysis performed. | |||||||
Discussion
According to the World Health Organization's data, as of June 2024, approximately 775 million people have been diagnosed with COVID-19 and around 7 million people have lost their lives [4]. New coronavirus variants continue to be identified as the days go by. While the mortality due to COVID-19 has significantly decreased since the introduction of vaccines, the risk of a new coronavirus variant emerging that is resistant to vaccines still exists. Despite advanced prophylaxis and treatment options, morbidity and mortality due to COVID-19 still persist and early diagnosis is particularly valuable for patients at high risk of developing critical illness. Initiating treatment early in this patient group is important for reducing mortality risk and preventing complications. Today, even if diagnosing COVID-19 in clinical practice is not as important as before, the virus is still continues to be a cause of mortality. Results of our study in a heterogeneous patient group including patients with positive and negative PCR results compatible with COVID-19 symptoms are valuable because they represent the patient population that clinicians directly encounter in clinical practice.
In PCR negative group 163 (57 %) of patients deceased while 123 (43 %) survived. In PCR positive group 75 (67.6 %) deceased and 36 (32.4 %) survived. There was no statistically significant difference in mortality among PCR groups (p = 0.054). Although the mortality rate is noticeably higher in the PCR-positive group, this difference is not statistically significant. There are some studies that report no significant difference in mortality between COVID-19 positive and negative patients who develop ARDS requiring mechanical ventilation [5]. Similarly, there were not significant difference between COVID-19 patients and other pneumonia etiologies [6]. However, some studies [7] showed that the mortality rate of COVID-19 patients is significantly higher from influenza group in the intensive care unit. There is no consensus in the literature. In our study, while the mortality rate is higher in COVID-19 diagnosed patients, this difference is not statistically significant. This may be explained by the relatively small number of patients.
In our study, the median values of hematocrit, potassium, glucose and Platelet-to-Lymphocyte (Plt/Lym) ratio were not significantly different between survivors and non-survivors in the COVID-19 group (p > 0.05). However, in the non-survivor group, the median values of WBC, Neu, urea, BUN, creatinine, sodium, AST, ALT, CRP, D-Dimer, Neu/Lym ratio and Neu/Plt ratio were significantly higher, while the median values of lymphocytes, platelets, and hemoglobin levels were lower. Similar to our study it was reported [8] that among COVID-19 positive patients, WBC, Neu, CRP values, D-Dimer, AST, ALT, and creatinine levels were significantly higher while lymphocyte and platelet levels were significantly lower in the deceased group. The relevant study mentioned that disseminated intravascular coagulation frequently developed in the deceased patient group and this condition was consistent with elevated D-Dimer levels and decreased platelet counts. Additionally, it was noted that Macrophage Activation Syndrome occurring in the critical patient group led to acute kidney injury and liver damage, causing elevated liver enzymes such as AST and ALT, as well as increased plasma levels of creatinine. CRP was associated with hyperinflammation state and patients with high CRP levels had high risk of respiratory failure [9]. In another study [10] it was reported that decreased lymphocyte and platelet count, elevated CRP, D-Dimer, AST and ALT and creatinine levels were associated with a poor prognosis. The relevant study highlights that Angiotensin-converting Enzyme-2 receptors on lymphocytes target lymphocytes against the virus, and this process results in lymphopenia. Lymphocytes are the primary cells responsible for generating an immune response against viral pathogens and lymphopenia is indicated to result in an inadequate immune response to the virus, hence being associated with a poor prognosis. Elevated D-Dimer levels are indicative of venous thromboembolism and lead to ventilation-perfusion mismatch, which in turn causes end-organ damage, poor prognosis and an increased risk of mortality [11]. Consistent with the results in our study, there are numerous studies [12, 13] that support the association of increased WBC, Neu, CRP, AST and ALT, creatinine and decreased platelet and lymphocyte counts with high risk of mortality. In some studies [14] Neu/Lym and Neu/Plt ratios were found to be significantly higher in the deceased group, similar to our study. However, in the same study, the Plt/Lym ratio was found to be higher in the non-survivor group compared to survivors, unlike our study. There are numerous studies in the literature examining the relationship between laboratory values and mortality in COVID-19 patients. The results identified in our study are consistent with the literature [15–18].
In COVID-19 group mortality was found significantly higher in patients who had ARF/CRF, while no significant effect on mortality was observed for other comorbidities such as HT, DM, CAD, CVE, AF, CHF, malignancy, asthma/COPD and neuropsychiatric disease. Among the COVID-19 positive patients in the intensive care unit with hypertension, the duration of stay in the intensive care unit is longer and the risk of mortality is higher [19]. Additionally, DM, CAD and CVE are associated with increased mortality. Rehatta and colleagues [20] demonstrated that HT, DM, renal failure, CHF, CVE and chronic lung disease significantly increased the risk of mortality in COVID-19 patients. However, the same study did not find a significant difference on mortality for malignancy, CAD and neuropsychiatric disease. Renal failure, history of cerebrovascular events and malignancy were found as the risk factors for mortality in COVID-19 patients [21]. While mortality risk does not increase in patients with epilepsy and Alzheimer's Disease among neurologic disorders, there is an increase risk with Parkinson's disease [22]. Depression and other various psychiatric and neurological diseases was found to be associated with mortality [23]. There is no consensus in studies examining the relationship between comorbidities and COVID-19 patient mortality. In our study, patients with a history of AF, COPD or asthma, malignancy and neuropsychiatric diseases showed higher mortality rates; however, this difference was not statistically significant. This may be attributed to the relatively low number of patients in these subgroups. Patients with CAD had a lower mortality rate but the difference was not statistically significant. This result may be attributed to the impact of other comorbidities on mortality. Our findings are consistent with the literature.
Among patients with a negative PCR result, the deceased group had statistically higher APACHE II, SOFA and PSI scores (p < 0.001). There are numerous studies in the literature that examine intensive care scoring systems and patient outcomes. One study reported that APACHE II scores were significantly higher in the intensive care patients in the deceased group but unlike our study, SOFA scores did not show a significant difference between the deceased and survived patients [24]. In contrast some studies [25] showed that APACHE II and SOFA scores both were higher in the deceased group, similarly to our study. Alavi-Moghaddam and colleagues [26] reported that in community-acquired pneumonia cases, as the PSI and Confusion, Urea, Respiratory Rate, Blood Pressure and Age Above or Below 65 Years (also known as CURB-65) score increased, the risk of mortality also increased. The results obtained in our study are in line with the literature.
In our study, among patients with a positive PCR result, the deceased group had significantly higher APACHE II, SOFA and PSI scores. Beigmohammadi and colleagues [27] stated that an increase in APACHE II and SOFA scores increased the risk of mortality. Another study emphasized that APACHE II score more successful in predicting mortality compared to the SOFA score [28]. We found there was no significant difference between APACHE II and SOFA score AUCs (DeLong Test, p > 0,05) so both scoring systems successful in predicting mortality. The primary factor determining mortality in COVID-19 patients is respiratory system involvement, but mortality prediction can be successfully achieved using the SOFA score. An increase in the PSI score of patients with COVID-19 pneumonia proportionally increased the patient's risk of death [29]. Same study reports that mortality rate was 17.35 % in PSI class 4 patients and 74.78 % in PSI class 5. In our study, the median PSI score for the deceased group was 167 with the majority of them falling into PSI class 5. The median PSI score for the surviving group was 95 with most of them falling into class 3. PSI score is successful in predicting mortality in community-acquired pneumonias, as well as in COVID-19-related pneumonias, with a high predictive value [29]. Our results are in line with the literature.
ROC analysis is performed and Youden’s Index is used to determine optimal cutoff points for APACHE II, SOFA and PSI scores. 19 points for APACHE II (AUC = 0.96, p = 0.0001), 9 points for SOFA score (AUC = 0.96, p = 0.0001) and 81 points for PSI score (AUC = 0.91, p = 0.0001) are determined as optimal cutoff points. An APACHE II score≥ 19 points is 89 % sensitive and 94 % specific for mortality with PLR of 14,03 (%95 CI; 7,69–25,61) and NLR of 0,13 (0,09–0,18). A SOFA score ≥ 9 points is 89 % sensitive and 96 % specific for mortality with PLR of 20,23 (9,79–41,80) and NLR of 0,11 (0,08–0,16). A PSI score≥ 81 points is 79 % sensitive and 89 % specific for mortality and has a PLR of 7,01 (4,52–10,9) and a NLR of 0,23 (0,18–0,30). In a study [30] it is found that a SOFA score of 2 or higher is 84 % sensitive and 63 % specific for 30-day mortality (AUC = 0.80, p < 0.05). Niaz and colleagues [31], in their published study, set the cutoff point for the SOFA score at 7 and found it to be 75 % sensitive and 95 % specific for mortality (AUC = 0.89). In a study examining the relationship between the PSI score and mortality [32], a 30-day mortality prediction was generated with 75 % sensitivity and 47 % specificity in patients with a PSI group 4 or higher (AUC = 0.69). Optimal cutoff point for the APACHE II score is determined as 15 (AUC = 0.88, p < 0.001) and used it to predict mortality with 85 % sensitivity and 77 % specificity [33]. Zou and colleagues (24) found that an APACHE II score equal or more than 17 had a sensitivity of 96 % and specificity of 86 % for mortality prediction (AUC = 0.96, p < 0.05). The optimal cutoff point for the SOFA score was determined as 3 (AUC = 0.86, p < 0.05). When compared to the literature, the cutoff points in our study allow for more successful mortality prediction with higher sensitivity, specificity and AUC values.
Limitations
Some limitations of our study include its retrospective nature and being conducted at a single center. The absence of a dialysis unit has resulted in the inability to admit patients with urgent dialysis needs to our intensive care unit. Additionally, due to the COVID-19 pandemic, patients with a known diagnosis of COVID-19 were directly referred from the emergency department to pandemic intensive care units for isolation, and therefore, patients with a confirmed COVID-19 diagnosis were not admitted to our intensive care unit.
Conclusion
In our study, there was no statistically significant difference in mortality between patients with COVID-19 and those with respiratory distress due to different etiologies in the intensive care unit. Only renal failure significantly increased mortality, other comorbidities did not. Considering all patients, APACHE II, SOFA and PSI scores were significantly higher in the deceased patients and the cutoff points were determined as 19 points for APACHE II, 9 points for SOFA score and 81 points for PSI score. Multicenter studies with larger patient populations are needed in this regard.
Disclosure. The authors declare no competing interests.
Author contribution. All authors according to the ICMJE criteria participated in the development of the concept of the article, obtaining and analyzing factual data, writing and editing the text of the article, checking and approving the text of the article.
Ethics approval. This study was approved by the Kutahya Health Sciences University Non-Interventional Clinical Research Ethics Committee with decision number 37686 dated January 27, 2022 (Annex-1).
Funding source. This study was not supported by any external sources of funding.
Data availability statement. The data sets analysed are available from the corresponding author.
Informed consent. Informed consent was obtained from all the patients for their anonymized information to be published in this article.
