Identification of parameters in routine blood and coagulation tests related to the severity of COVID-19

Objective: This study aimed to identify the predictive value of simple markers in routine blood and coagulation tests for the severity of coronavirus disease 2019 (COVID-19). Methods: A total of 311 consecutive COVID-19 patients, including 281 patients with mild/moderate COVID-19 and 30 patients with severe/life-threatening COVID-19, were retrospectively enrolled. Logistic modeling and ROC curve analyses were used to assess the indexes for identifying disease severity. Results: Lymphocyte and eosinophil counts of COVID-19 patients in the severe/life-threatening group were significantly lower than those of patients in the mild/moderate group (P < 0.001). Coagulation parameters, high-sensitivity C-reactive protein (hsCRP) levels and procalcitonin levels were higher in the severe/life-threatening group compared with the mild/moderate group (all P < 0.05). Univariate and multivariate logistic models revealed that hsCRP and fibrinogen degradation products (FDPs) were predictors of severe COVID-19 (OR = 1.072, P = 0.036; and OR = 1.831, P = 0.036, respectively). The AUROCs of hsCRP and FDP for predicting severe/life-threatening COVID-19 were 0.850 and 0.766, respectively. The optimal cutoffs of hsCRP and FDP for the severe/life-threatening type of COVID-19 were 22.41 mg/L and 0.95 µg/ml, respectively. Conclusion: Serum CRP and FDP levels are positively related to the severity of COVID-19. This finding indicates that CRP and FDP levels may potentially be used as early predictors for severe illness and help physicians triage numerous patients in a short time.


Introduction
An outbreak of coronavirus disease 2019  caused by person-to-person transmission of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has spread rapidly worldwide with a spectrum of disease ranging from asymptomatic to critical illness [1]. Globally, greater than 40 million cases of COVID-19 have now been reported to the WHO with greater than one million deaths [2]. Given that COVID-19 patients with mild illness or asymptomatic carriers are prone to transmit the virus, the number of new and severe cases is increasing rapidly every day [3]. Due to the genomic homologies of coronavirus pathogens, COVID-19 exhibits a similar clinical course and pathological features as severe acute respiratory syndrome (SARS) and Middle East respiratory syndrome (MERS) [4][5][6]. Although most patients with COVID-19 are mild and can recover gradually after two weeks, approximately 15-20% of these patients develop severe illness [7,8]. Moreover, patients with severe disease may quickly develop acute respiratory distress syndrome (ARDS), acute heart injury, secondary infection, and even Ivyspring International Publisher multiple organ failure [1,6]. Therapeutic measures, such as antiviral treatment, can contribute to alleviating severe cases of COVID-19 and improve the prognosis [9]. Therefore, it is critically important to quickly and accurately recognize the severity of COVID-19 during the early stages of disease according to simple clinical parameters, so treatment can be initiated for severe cases in a timely manner.
Recently, various studies reported that laboratory changes, including hematologic parameters and coagulation indicators, correlated with poor prognosis in patients with severe pneumonia caused by SARS-CoV-2 [10,[11][12][13]. Hematologic parameters, including leucocytes, neutrophil-to-lymphocyte ratio (NLR) and monocyte-to-lymphocyte ratio (MLR), have been reported as indicators in diagnosing and stratifying the risk of infectious diseases [14]. A recent study indicated that the NLR may represent a reliable marker for evaluating the severity of COVID-19 [15]. Moreover, coagulation dysfunction in COVID-19 patients is associated with an increased risk of death [16]. SARS-CoV-2 may activate the innate immune system to clear the virus. However, an excessive immune response can cause an inflammatory storm, destroy microcirculation, activate the coagulation system and lead to disseminated intravascular coagulation (DIC) [17]. Currently, some coagulation parameters, such as D-dimer and fibrin degradation product (FDP), have been described as prognostic factors of severe disease [10,12]. However, studies on the predictive performances of hematologic parameters and coagulation markers in COVID-19 patients are on-going, and the clinical value of these markers in predicting severe illness needs to be further confirmed.
Therefore, in this retrospective study, we aimed to examine the correlation between hematologic parameters and coagulation markers and the severity of COVID-19. We also sought further identify any biomarkers that could serve as potential predictors of severe illness.

Ethics statement
The study protocol and informed consent documents were reviewed and approved by the Ethics Committee of Shanghai Public Health Clinical Center, Fudan University (YJ-2020-S111-02). All COVID-19 patients provided written informed consent during admission.

Patients
A total of 311 patients with COVID-9 confirmed by the SARS-CoV-2 nucleic acid test (RT-PCR) who were admitted to Shanghai Public Health Clinical Center, Fudan University from January 20 to March 15 were enrolled in our retrospective study. Clinical and laboratory data of the patients were collected from medical records. The laboratory data, including routine blood test parameters, high-sensitivity C-reactive protein (hsCRP) measurement, coagulation tests, and erythrocyte sedimentation rate (ESR), were investigated.

Data collection
Routine blood test results; hsCRP; coagulation parameters, including APTT, FDP, fibrinogen (FIB), PT, international normalized ratio (INR), prothrombin time activity (PTA), and D-dimer; and ESR were determined on the first day after admission. All laboratory tests were conducted in the Department of Medical Laboratory, Shanghai Public Health Clinical Center, Fudan University.

Definition
Two inflammatory predictors, including the monocyte-to-lymphocyte ratio (MLR) and neutrophilto-lymphocyte ratio (NLR), were included in this analysis.

Outcomes
Based on the 7 th trial version of the "Diagnosis and treatment program of novel coronavirus-infected pneumonia" released by the National Health Commission of the People's Republic of China [18], COVID-19 patients were divided into mild/moderate and severe/life-threatening groups. Patients meeting any of the following criteria were diagnosed with severe/life-threatening type: 1) polypnea with a respiratory rate ≥ 30 breaths per minute; 2) finger oxygen saturation ≤ 93% in a resting state; 3) PaO2/FiO2 ≤ 300 mmHg; 4) respiratory failure requiring mechanical ventilation; 5) emerging shock; and 6) combined with other organ failure requiring ICU care. Severe/life-threatening signs persisting for up to one week were considered severe disease.

Statistical analysis
Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS version 22.0). The continuous variables are presented as the mean ± SD or median (range), which were compared using the independent t test or the Mann-Whitney U test, respectively. Categorical variables are presented as proportions and were compared using the chi-square test. Logistic regression models were used to assess the correlation between the laboratory parameters and disease severity. The area under the receiver operating characteristic curve (AUROC) and its 95% confidence interval (CI) were calculated to evaluate the prediction efficiency of the potential predictors. Decision curve analysis was conducted using the rmda package in the R program [19]. A two-sided P<0.05 was considered statistically significant.

Clinical characteristics
A total of 311 patients (162 males and 149 females) with COVID-19 were enrolled in the study. Among them, 281 (90.4%) patients were in the mild/moderate group, and 30 (9.6%) were in the severe/life-threatening group.

Routine blood tests and coagulation data
Inflammation and coagulation parameters on admission were compared between the mild/ moderate and severe/life-threatening groups ( Table  2). No significant differences in WBC counts, RBC counts, hemoglobin, monocyte counts, INR, PT or ESR were observed between the two groups (P > 0.05 for all). The blood classified counts of the severe/life-threatening patients exhibited an increase in neutrophils (median 3.54×10 9 /L vs. 2.88×10 9 /L, P = 0.005) and a decrease in lymphocytes (median 0.68 ×10 9 /L vs 1.15×10 9 /L, P < 0.001) and eosinophils (median 0 vs. 0.01×10 9 /L, P < 0.001). A similar difference was observed for platelet counts (median 146.0×10 9 /L vs 179.

Predictive performances of serum inflammation and coagulation parameters for COVID-19 severity
The performances of hsCRP and FDP in predicting the severity of COVID-19 were calculated using ROC curves. The ROC curves of hsCRP and FDP are presented in Figure 1. The AUROCs of the indexes for predicting COVID-19 severity are shown in Table 4. The AUROCs of hsCRP and FDP for predicting severe/life-threatening COVID-19 were 0.850 (95% CI = 0.798-0.893) and 0.766 (95% CI = 0.714-0.812), respectively. Based on the estimated AUROCs to predict the severity of COVID-19, the performance of hsCRP versus FDP was comparable (P = 0.054). In addition, we conducted decision curve analysis to further investigate the clinical application values of the two markers. The decision curve analyses of hsCRP and FDP for predicting the severity of COVID-19 are presented in Figure 2A. The results showed that hsCRP exhibited the highest net benefit at any given threshold. Based on the above findings, we further plotted clinical impact curves to evaluate the clinical impact. The results demonstrated good predictive power of the two markers (Figure 2B and C).

Predictive thresholds and accuracies of serum indexes for severe/life-threatening COVID-19
The predictive thresholds and accuracies of serum indexes for severe/life-threatening COVID-19 are presented in Table 5. Maximizing the sum of sensitivity and specificity, the optimal cutoffs of hsCRP and FDP were 22.41 mg/L and 0.95 µg/ml, respectively, for predicting the severe/lifethreatening type of COVID-19. To obtain a sensitivity and specificity of at least 90%, the cutoffs of hsCRP were 9.85 mg/L and 46.42 mg/L, respectively, and the cutoffs of FDP were 0.78 µg/ml and 2.57 µg/ml, respectively.

Discussion
The present study mainly demonstrated that hsCRP and FDP provided the greatest contribution to the prediction of critical illness by examining the results of hematologic parameters and coagulation function of COVID-19 patients.
Hematologic indexes, including lymphocyte, neutrophil, eosinophil counts, and NLR, helped to stratify COVID-19 patients. Our results corresponded to previous studies [20][21][22]. Qin et al. [21] reported that compared to mild cases, severe cases tended to have lower lymphocyte counts, higher WBCs and NLRs, and lower percentages of monocytes, eosinophils and basophils in a cohort of 450 COVID-19 patients. A meta-analysis including 3377 COVID-19 patients showed that patients with severe and lethal illness had significantly higher WBC and lower lymphocyte and platelet counts [22]. The NLR is typically used to identify bacterial infection severity and the prognosis of cardiovascular disease [23,24]. For COVID-19 patients, elevated NLR is an independent predictor for poor clinical outcome [15,25]. The combination of NLR and MLR potentially helped to predict severe COVID-19 with an AUROC of 0.93 [26]. In addition, SARS-CoV-2 causes an increase in inflammation, which leads to anemia owing to the destruction of RBCs and decreased erythrogenesis [26,27]. In contrast to these studies, no significant differences were noted between our groups with respect to RBC levels. Previous studies reported that COVID-19 patients with platelet counts of less than 100×10 9 /L only accounted for 5%; however, platelet counts of less than 150×10 9 /L were identified in 70-95% of patients with severe COVID-19 [6,7]. Our findings were also consistent with those of these studies. Thus, thrombocytopenia in patients with COVID-19 did not seem to be a key predictor of COVID-19 progression [6,28].
Elevated serum CRP levels in infected patients may be potential biomarkers for the diagnosis of infectious diseases [29]. Currently, studies have reported that higher CRP levels are associated with adverse aspects of COVID-19, such as the development of ARDS, elevated troponin T levels, myocardial injury, and death [11,30,31]. In our study, CRP levels in patients with severe/life-threatening COVID-19 were significantly increased compared with mild/moderate cases, which is consistent with earlier similar studies [32,33]. Furthermore, our study demonstrated very good performance of hsCRP in predicting the critical stage of COVID-19 with an AUROC of 0.85 and a cutoff level of 22.41 mg/L. In a large study that included 220 patients, the authors indicated that the AUROC of hsCRP for the early diagnosis of pneumonia complicated by sepsis was 0.82 with a cutoff of 55 mg/L [34]. A study by Herold et al. [35] indicated that IL-6 and CRP highly predicted the need for invasive ventilation with maximal values > 80 pg/ml and > 97 mg/l, respectively. These results indicated that CRP could be used as a single tool to stratify COVID-19 patients. Clinically, elevated CRP levels may be an early indicator of nosocomial infection in patients with slow recovery of COVID-19 and may inform doctors to provide empirical antibiotic treatment as soon as possible to prevent deterioration of prognosis [7,36].  Coagulation is a very organized process involving the interaction of endothelial cells, platelets and coagulation factors. Abnormal coagulation appears to be an important issue in COVID-19 patients. Recent data support that COVID-19 patients are at high risk of developing disseminated intravascular coagulation (DIC) or even death [7,10]. Higher levels of D-dimer and FDP levels are associated with poor prognosis in patients with COVID-19 [10]. A meta-analysis reported that elevated fibrinogen and FDP levels on admission were associated with an increased risk of poor outcome in COVID-19 patients [37]. The present study found that compared to patients with mild/moderate COVID-19, the PTA value in patients with severe/life-threatening illness was lower, whereas APTT, D-dimer, FDP, and FIB levels were higher. These results confirmed findings from earlier similar studies [10,12]. Additionally, our study indicated that FDP was a risk factor associated with the development of critical COVID-19; more importantly, we identified that the AUROC of the FDP level was 0.71 with a cutoff of 0.95 µg/ml. These findings strengthen the suggestion that routine monitoring of coagulation parameters may represent a potentially useful tool to improve the early diagnosis of critical COVID-19 and establish an accurate therapeutic strategy [38,39]. It has been widely demonstrated that coagulation is activated and accelerated in response to infections because this mechanism may increase the physiological response [40,41]. The mechanisms of coagulation dysfunction in COVID-19 are complex and potentially include direct injury of endothelial cells, imbalance of inflammatory response, over activation of immune system, ischemia-reperfusion injury, and drug factors [42].
This study has several limitations. First, because this study was a single-center and retrospective study, the characteristics of the enrolled patients may not be representative, and the findings should be validated in clinical studies with more power. Second, early treatment could help promote the recovery of patients with severe COVID-19. Future studies could further investigate the dynamic changes in in hsCRP and FDP in the treatment of patients with critical COVID-19 and the impacts of these markers on disease prognosis. Third, we did not examine the imaging index to assess the value of patients' imaging parameters and the dynamic changes in predicting severe/life-threatening cases. Finally, our study did not enroll pediatric patients. Previous studies have demonstrated that children of all ages can develop COVID-19. Compared with adult patients, pediatric patients tend to have a mild COVID-19 course with a good prognosis [43,44].
In conclusion, CRP and FDP serum levels are positively related to the severity of COVID-19. This finding indicates that CRP and FDP may potentially be used as early predictors for severe illness, and this information may help physicians triage numerous patients in a short time.