key: cord-1032835-mkef5gm2 authors: Bastug, Aliye; Bodur, Hurrem; Erdogan, Serpil; Gokcinar, Derya; Kazancioglu, Sumeyye; Kosovali, Behiye Deniz; Ozbay, Bahadır Orkun; Gok, Gamze; Turan, Isil Ozkocak; Yilmaz, Gulsen; Gonen, Canan Cam; Yilmaz, Fatma Meric title: Clinical and laboratory features of COVID-19: Predictors of severe prognosis date: 2020-09-09 journal: Int Immunopharmacol DOI: 10.1016/j.intimp.2020.106950 sha: 685bc050a4fd995975c5f5a42a5f22a4ecc43d57 doc_id: 1032835 cord_uid: mkef5gm2 BACKGROUND: Coronavirus disease 2019 (COVID-19) emerged first in December 2019 in Wuhan, China and quickly spread throughout the world. Clinical and laboratory data are of importance to increase the success in the management of COVID-19 patients. METHODS: Data were obtained retrospectively from medical records of 191 hospitalized patients diagnosed with COVID-19 from a tertiary single-center hospital between March and April 2020. Prognostic effects of variables on admission among patients who received intensive care unit (ICU) support and those who didn’t require ICU care were compared. RESULTS: Patients required ICU care (n = 46) were older (median, 71 vs. 43 years), with more underlying comorbidities (76.1% vs. 33.1%). ICU patients had lower lymphocytes, percentage of large unstained cell (%LUC), hemoglobin, total protein, and albumin, but higher leucocytes, neutrophils, neutrophil–lymphocyte ratio (NLR), monocyte-lymphocyte ratio (MLR), platelet-lymphocytes ratio (PLR), urea, creatinine, aspartate amino transferase (AST), lactate dehydrogenase (LDH), and D-dimer when compared with non-critically ill patients (p < 0.001). A logistic regression model was created to include ferritin, %LUC, NLR, and D-dimer. %LUC decrease and D-dimer increase had the highest odds ratios (0.093 vs 5.597, respectively) to predict severe prognosis. D-dimer, CRP, and NLR had the highest AUC in the ROC analysis (0.896, 0.874, 0.861, respectively). CONCLUSIONS: The comprehensive analysis of clinical and admission laboratory parameters to identify patients with severe prognosis is important not only for the follow-up of the patients but also to identify the pathophysiology of the disease. %LUC decrease and D-dimer, NLR, and CRP increases seem to be the most powerful laboratory predictors of severe prognosis. A cluster of pneumoniae cases with unknown etiology emerged in Wuhan city of China in early December 2019. The causative agent was defined as a novel enveloped RNA beta-coronavirus, named as a severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) [1] . The illness was subsequently termed as the Coronavirus disease 2019 (COVID-19) by the World Health Organization (WHO). On 30 January 2020, WHO declared COVID-19 as a Public Health Emergency of International Concern [2] . Although most of the patients have a milder clinical course, some of them have severe pneumonia with high mortality rate requiring comprehensive care in ICU [3] . Therefore, determining the predictive indicators of a severe infection is of great importance. It might help to understand the clinical course of the pneumonia, but more importantly, we still need to clarify the pathophysiology of the disease. In the present study, we aimed to perform a comprehensive analysis of clinical, laboratory, and demographic characteristics of 191 patients with COVID-19, admitted to Ankara City Hospital, to determine the predictors of this serious illness. https://doi.org/10.1016/j.intimp.2020.106950 Received 6 July 2020; Received in revised form 19 August 2020; Accepted 27 August 2020 This retrospective case series was approved by the ethics board of Ankara City Hospital (No. E1-20-531). All consecutive COVID-19 patients admitted to Ankara City Hospital from March 13 to April 30 were enrolled. The diagnostic criteria of WHO interim guidance were used [3] . Demographic, clinical, laboratory data, and radiological findings were extracted from electronic medical records and case record forms. All laboratory results of the patients within the 24 h of admission were recorded. The baseline parameters were selected to predict the prognosis of the patients when they first presented to the hospital to help the clinicians identify patients who may need ICU care at some point of hospital stay and need a close follow-up. Data on the following laboratory parameters were collected: complete blood count parameters (CBC), urea, creatinine, estimated glomerular filtration rate (eGFR), total protein, albumin, AST, ALT, LDH, creatine kinase (CK), prothrombin time (PT), activated partial thromboplastin time (APTT), international normalized ratio (INR), fibrinogen, D-dimer, and ferritin. CBC was performed using the ADVIA 2120 Hematology System (Siemens Healthcare Diagnostics, Erlangen, Germany). Biochemical parameters defined above were measured using an Atellica Solution Immunoassay & Clinical Chemistry Analyzer (Siemens Healthcare Diagnostics, Erlangen, Germany). eGFR values were calculated using the CKD-EPI formula [4] . PT, APTT, INR, fibrinogen, and D-dimer were analyzed using the Sysmex CS-5100 System (Siemens Healthcare Diagnostics, Erlangen, Germany). Chest computed tomography (CT) and oro/nasopharyngeal swab sample for RT-PCR were obtained for all suspected patients in addition to routine blood tests. COVID-19 was diagnosed using consistent clinical manifestations, such as fever and respiratory symptoms, findings of pneumonia on CT, and/ or positive SARS-CoV-2 PCR results according to the WHO interim guidance [3] . Patients requiring treatment in an intensive care unit (ICU) on admission or at some point during hospital stay (ICU group) and those not needing ICU care (non-ICU group) were compared in terms of clinical and demographic data, radiological characteristics, and routine blood test results. Data are median (minimum value -maximum value) or n (%). P values comparing ICU patients and non-ICU patients. ICU = intensive care unit. NA = not applicable. NIV; non-invasive ventilation, IMV; invasive mechanical ventilation, *Single or multiple patchy ground glass opacities predominantly in the peripheral areas of the lungs, **Bilateral multi-lobar ground glass opacities and consolidation, *** Acute kidney disease was defined as 0.3 mg/dL increases of admission creatinine level compared to basal creatinine level. Statistical analyses were performed using the SPSS 21.0. The Kolmogorov-Smirnov test was performed to check the normality of the variables. Descriptive analysis was presented using mean ± SD for normally distributed variables and median (minimum-maximum value) for non-normally distributed variables. Demographic and laboratory data were compared between the groups using the Student's t test for parametric and the Mann-Whitney U test for non-parametric variables. Comparisons for categorical variables were executed using the chisquare test or the Fisher's exact test. Receiver operation characteristic (ROC) curve was performed to analyze the efficiency of the disease severity diagnosis. Binary logistic regression analysis was performed to define the independent predictors of the disease severity. To include the appropriate parameters in the logistic regression model, firstly we examined the bivariate relationships among parameters, and then clinically most relevant variables were included in the model. To eliminate the effects of potential confounders, namely age, gender, and chronic disease parameters included in the model, we used enter method and parameters with p values greater than 0.200 were excluded for the final model with the Forward LR test. The Hosmer-Lemeshow goodness of fit test was used. The odds ratio (OR) was calculated for significantly associated variables. Statistical significance was defined as p < 0.05. The study included 191 hospitalized patients with COVID-19 diagnosis according to WHO criteria [3] . In addition to consistent clinical symptoms and findings of COVID-19, all patients had PCR confirmation and/or radiologic findings. Of the patients, 148 (77.5%) had PCR confirmation and 171 (89.5%) had abnormalities on chest CT consistent with COVID-19. All SARS CoV-2 PCR negative patients (n = 43, 22.5%) had compatible signs, symptoms, and typical CT findings for COVID-19 without any alternative diagnosis, and 13 of whom were in the ICU group (Table 1) . They were compatible with the probable case criteria according to the WHO interim guidance [3] . Other possible respiratory pathogens were also excluded with the respiratory PCR panel, which yielded a negative result based on the Fast Track FTD Respiratory pathogens 21 [3] . We also wanted to evaluate the effect of each of hypertension and heart failure on severity. Diabetes mellitus, acute and chronic kidney disease, hypertension, congestive heart failure, and other cardiovascular diseases were all found significantly higher in the ICU group (p < 0.05 for each comorbidity). There were no significant differences for PCR results between groups (p < 0.284). However, advanced stage radiologic abnormalities (single/multiple patchy ground glass opacities, predominantly in the peripheral areas of the lungs) were more common in the ICU group [44 (95.6%) vs 36 (28.8%); p < 0.001]. The mean time frame from onset of symptoms to ICU admission was 7.09 ± 4.81 days. The length of hospital stay was significantly higher in the ICU group (p < 0.001). The mean ICU follow up time was 10.7 ± 8.16 days for critically ill patients. The mean Acute Physiology and Chronic Health Evaluation II (APACHE II) score was 19.22 ± 12.93 on the day of ICU admission (Table 1) . A total of 169 (88.5%) patients received antibiotic therapy [mostly azithromycin, 129 (67.5%)]. Doxycycline, ceftriaxone, and tigecycline were the other antibiotics used, respectively. Chloroquine was administered to 169 (88.5%) patients and favipiravir was given to 49 (25.7%) patients (Table 1) . Favipiravir was used mostly in severely/critically ill patients (78.3%). Oseltamivir was also used empirically until the influenza PCR test was concluded. It was used significantly more in ICU patients, since they were critically ill. Steroid treatment was given only in 21.7% (10/46) of the ICU patients according to clinical judgement of the ICU specialist for critical illness related corticosteroid insufficiency. Anticoagulant treatment was given to all ICU patients routinely. However, it was not given to non-ICU patients since there was no advice for its implementation in COVID-19 patients at that time of the point. Immunomodulatory therapy (IL-1 or IL-6 inhibitor) was not given to the patients since there was a knowledge gap. There is not a precise treatment agent for COVID-19 to date. Supportive treatment is still the main intervention. Oxygen inhalation therapy with nasal cannula were administered in 49 (25.7%) patients. The need of nasal cannula oxygen therapy was significantly higher in ICU patients [(54.3% (n = 25) vs 16.6% (n = 24); p < 0.001)]. Among the ICU patients, high flow nasal oxygen therapy was required 2 patients (4.3%), noninvasive ventilation was needed in 9 patients (19.6%), and invasive mechanical ventilation was needed in 10 patients (21.7%). Death occurred in 20 (43.5%) critically ill patients in the ICU group (Table 1) . All other patients were clinically improved and discharged. Blood routine parameters were recorded on admission day for all patients and then compared between groups. There were several significant differences. The ICU group showed higher white-blood cell, neutrophil counts, neutrophil-lymphocyte ratio (NLR), monocyte-lymphocyte ratio (MLR), platelet-lymphocytes ratio (PLR), urea, creatinine, AST, LDH, PT, INR, fibrinogen, CRP, ferritin, D-dimer levels, as well as lower lymphocytes, %LUC, hemoglobin, hematocrit, total protein, and albumin (p < 0.001) ( Table 2) . A ROC curve analysis was used to determine the efficacy of various parameters predicting severe prognosis (Figs. 1, 2, and 3) . D-dimer, CRP, and NLR had the highest AUC in the ROC analysis (0.896; 0.874; 0.861, respectively). AUC, optimal cut-off, and sensitivity and specificity values of laboratory parameters are given in Table 3 . The binary logistic regression model included ferritin, %LUC, NLR, and D-dimer. The results of the logistic regression model are given in dimer, and NLR, and decreasing %LUC were independent predictors of the disease severity with the likelihood ratios shown in Table 4 . SARS-CoV-2 infection may lead to a wide clinical spectrum ranging from subclinical disease to severe pneumonia. Our initial findings showed that, laboratory parameters including %LUC, NLR, D-dimer, and CRP made the highest contribution to the prediction of the disease severity besides age, comorbidity presence, and symptoms on the admission. Of the patients enrolled in the study, 24.1% were critically ill. In line with Wang et al, the patients in the ICU group were older and showed more comorbidities compared to those in the non-ICU group. In addition, gender was not found as a determinative factor for critical illness [5] . On admission, prominent symptoms that occurred significantly more frequently in the ICU group were dry cough and dyspnea. The mean duration from the onset of symptoms to ICU requirement was 7.09 ± 4.81 days, which had been reported as 10 days median previously [5] . The evaluation of severe prognostic factors in that time course is important to prioritize patients, who may be in greater need of ICU. In our cohort, all ICU patients had pneumonia. As previously reported, the hallmark abnormalities on CT were bilateral multi-lobar ground glass opacities [5] , which were detected in the present study in 95.6% of the ICU patients. Patients in need of ICU care are the most vulnerable targets for death in COVID-19. Yang et al. reported 61.5% (32/52) mortality in critically ill patients [6] . In the present study, mortality rate was 43.5% (20/46) in the ICU group, whereas no death occurred in the non-ICU patients (0/145). Therefore, determining the prognostic severity criteria is fundamental to provide early intervention to patients who may require ICU support. In regard to the laboratory evaluation, SARS-CoV-2 causes alterations in some routine blood parameters. As for admission laboratory parameters, decreased hemoglobin, lymphocyte, %LUC, and albumin, and upregulated leucocytes, neutrophils, NLR, ferritin, D-dimer, creatinine, LDH, and CK levels are distinguishing features of the critically ill patients. Higher levels of leucocytes were prominent characteristics of the critically ill patients in the present study, in consistent with the meta-analysis, which reported higher white blood cell counts in patients with severe COVID-19 [7] . In contrast to this study, there was no significant differences between our groups regarding thrombocytes levels. Wang et al. reported no significant differences between ICU and non-ICU patients, regarding thrombocytes and monocytes, which is consistent with our findings [5] . SARS-CoV-2 causes increases in inflammatory factors, which results in anemia due to the destruction of red blood cell (RBC) and decreased erythrogenesis [8] . Red blood cell counts were significantly lower in critically ill patients in our cohort. The decreased hemoglobin levels were reported in severe COVID-19 patients in compatible with our results [8] . Lymphopenia and higher NLR have been reported as predictors for severe prognosis [6, [9] [10] [11] . Although lymphocytosis is an expected finding for viral infections, one of the possible reasons for depletion of lymphocytes in SARS-CoV-2 infection is that lymphocytes are the target for virus since the angiotensin converting enzyme 2 (ACE 2) receptor of the virus is expressed on lymphocytes [7, 12] . The other possible reasons for lymphopenia are as follows: the migration of lymphocytes from peripheral bloods to the lung, defective hematopoiesis, and the apoptosis of lymphocytes as a response to hyperinflammation [8, 13] . Lymphopenia was a more distinctive feature of the critically ill patients in our cohort as reported previously [10] . The presence of prominent lymphopenia may predict disease severity. The ROC analysis revealed the optimal cut-off value for lymphocytes as 0.980x10 9 per L (AUC;0.718), with 74.5% sensitivity and 60% specificity. Fan et al. reported that the lymphopenic patients had a few reactive lymphocytes as lymphoplasmacytoid [10] . Large unstained cells are reported to include activated lymphocytes and peroxidase-negative cells. Lower %LUC levels were a prominent feature of the ICU patients in our cohort. The optimal cut-off value predicting the severe illness was determined as 1.450% (AUC;0.681) with 66.2% sensitivity and 62.2% specificity by the ROC analysis. Logistic regression analysis revealed that %LUC was negatively correlated with the independent predictor of disease severity with an odds ratio of 0.063. According to the statistical model including NLR, ferritin, %LUC, and D-dimer, %LUC had the highest independent contribution to the model to predict severe diagnosis. The results of our study might be an indicator of the lack of the immune response to increase the activated lymphocytes, which we observed in non-ICU patients. The lack of the defined immune response might be an important factor for the patients to experience a more severe disease. Neutrophilia may result from the cytokine storm triggered by SARS-CoV-2. Although, the exact role of neutrophils is not known in viral infections, pro-inflammatory mediators produced by activated neutrophils have a detrimental role for the host [14] . In the meta-analysis of Zeng et al., higher neutrophil levels were correlated with the severity of COVID-19 7 . Consistently, our findings also suggest that higher level of neutrophils may be used to predict severe cases. SARS-CoV-2-triggered hyper inflammation seems to increase NLR, which promotes severe prognosis. A possible reason for this may be the reactive oxygen radicals released from neutrophils causing the cell's DNA damage [11] . Several studies have highlighted the predictive value of NLR for illness severity [7, 11, 15] . Our analysis also revealed that, NLR, MLR, and PLR might be used as a predictive diagnostic tool for determining patients needing ICU support. Previous studies have mentioned that the validation of PLR value is needed [7] . In our cohort, the optimal cut-off value for PLR was determined as 175 with the highest AUC, sensitivity, and specificity (0.715, 75.6% and 57.2%, respectively). The AUC of NLR reached the highest value (0.861) at the optimal cut-off value of 3.2 to predict the severe prognosis, independently. Our findings are consistent with previous reports suggesting higher NLR as a predictive nomogram for severe COVID-19 infection [11] . Hyper coagulation and disseminated intravascular coagulopathy (DIC) are other prominent features of viral infections, which may cause various complications [16] . Hypercoagulation may be due to the hyper inflammatory response. A markedly increased levels of fibrinogen and D-dimer were reported as predictive factors for severe prognosis, in consistent with our findings [5, 17] . Monitoring the fibrinogen activity may be helpful for clinicians to identify the patients with high risk for severe COVID-19, since it is not usually altered in other viral infections on admission. The optimum cut-off value for D-dimer for predicting a severe disease was determined as 0.565 mg/L using the ROC curve ( Fig. 3 ) (85.7% sensitivity; 80.6% specificity). Binary logistic regression analysis showed that the higher level of D-dimer compared to the cutoff value (0.565 mg/L) was also significant (p = 0.012, OR:7.139; 95% CI: 1.542-33.047) after the adjustment of ferritin, %LUC, and NLR. Ferritin was also determined as an independent predictive factor for severe infection (p = 0.006). D-dimer and PT were reported to be positively correlated with mortality [18] . Anticoagulant therapy with low molecular weight heparin to decrease the risk of venous thromboembolism and DIC was reported to be associated with better prognosis in patients with COVID-19 [18] . Although lungs are the main target organ for SARS-CoV-2, end organ damage is not limited to lungs. Lymphocytes are known to inhibit hyper immune responses occurring due to viral infection. For this reason, the lack of efficient lymphocyte levels occurring via SARS-CoV-2 infection results in increased cytokines and exacerbated inflammatory responses, which leads to damages in liver and kidney in addition to lungs [19, 20] . Lymphopenia and CRP were reported as independent predictors for hepatic injury in patients with COVID-19 [21] . In the present study, elevated CRP, AST, and LDH, and decreased total protein and albumin were found as significant laboratory parameters in the ICU group on hospital admission. Kidney is another target organ for the virus since the ACE2 receptors of SARS-CoV-2 is highly expressed on kidney tubule cells [22] . The increased creatinine and decreased eGFR levels were significantly different in the ICU group. Compared with non-critically ill patients, the ICU patients had several laboratory abnormalities as discussed above with the literature findings. However, to the best of our knowledge, there exists no study defining the importance of decreasing %LUC levels in predicting poor prognosis. In this study, our aim was to identify a useful marker which may be helpful in making an accurate prediction of severe prognosis, and we observed that %LUC value is an independent predictor that negatively correlated with poor prognosis with an odds ratio of 0.063. Therefore, it may be an addition to the list of severe prognosis predictors that can easily be obtained since it is a parameter reported in routine complete blood cell (CBC) tests. All findings of our study may be the result of cellular immune deficiency, hyper coagulation, kidney, and hepatic injury. In conclusion, our findings indicate that the %LUC decrease and the D-dimer, NLR, and CRP increases appear to be the most powerful laboratory predictors of severe prognosis for COVID-19. Monitoring the predictors of severity may assist clinicians to identify and follow-up patients with higher risk for progression. AB, HB, SK, BOO, DG, BDK, IOT, and CCG are clinicians and provided clinical details. SE and GY conducted statistical analysis and composed the figure. AB, FMY, and HB wrote and edited the manuscript. All authors were involved in the gathering of the manuscript and approved the final version. Genomic characterisation and epidemiology of 2019 novel coronavirus: implications for virus origins and receptor binding World Health Organization declares global emergency: A review of the 2019 novel coronavirus (COVID-19) Clinical management of severe acute respiratory infection (SARI) when COVID-19 disease is suspected: interim guidance A new equation to estimate glomerular filtration rate Clinical course and outcome of 107 patients infected with the novel coronavirus, SARS-CoV-2, discharged from two hospitals in Wuhan, China Longitudinal association between markers of liver injury and mortality in COVID-19 in China Can we predict the severity of COVID-19 with a routine blood test? Polish archives of internal medicine Abnormalities of peripheral blood system in patients with COVID-19 in Hematological findings and complications of COVID-19 Hematologic parameters in patients with COVID-19 infection The diagnostic and predictive role of NLR, d-NLR and PLR in COVID-19 patients High expression of ACE2 receptor of 2019-nCoV on the epithelial cells of oral mucosa The laboratory tests and host immunity of COVID-19 patients with different severity of illness The role of neutrophils in the upper and lower respiratory tract during influenza virus infection of mice Neutrophil-to-lymphocyte ratio and lymphocyte-to-C-reactive protein ratio in patients with severe coronavirus disease 2019 (COVID-19): A metaanalysis Coagulation disorders in coronavirus infected patients: COVID-19, SARS-CoV-1, MERS-CoV and lessons from the past Clotting factors in COVID-19: epidemiological association and prognostic values in different clinical presentations in an Italian cohort Anticoagulant treatment is associated with decreased mortality in severe coronavirus disease 2019 patients with coagulopathy Liver impairment associated with disease progression in COVID-19 patients Longitudinal characteristics of lymphocyte responses and cytokine profiles in the peripheral blood of SARS-CoV-2 Risk factors related to hepatic injury in patients with corona virus disease 2019, medRxiv Single-cell RNA sequencing data suggest a role for angiotensin-converting enzyme 2 in kidney impairment in patients infected with 2019-novel coronavirus This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Supplementary data to this article can be found online at https:// doi.org/10.1016/j.intimp.2020.106950.