Eosinophils and basophils in severe fever with thrombocytopenia syndrome patients: Risk factors for predicting the prognosis on admission

Background Severe fever with thrombocytopenia syndrome (SFTS) virus (SFTSV) is an emerging tick-borne phlebovirus with a high fatality rate. Previous studies have demonstrated the poor prognostic role of eosinophils (EOS) and basophils (BAS) in predicting multiple viral infections. This study aimed to explore the role of EOS and BAS in predicting prognosis of patients with SFTS. Methodology A total of 194 patients with SFTS who were admitted to Yantai City Hospital from November 2019 to November 2021 were included. Patients’ demographic and clinical data were collected. According to the clinical prognosis, they were divided into survival and non-survival groups. Independent risk factors were determined by univariate and multivariate logistic regression analyses. Findings There were 171 (88.14%) patients in the survived group and 23 (11.86%) patients in the non-survived group. Patients’ mean age was 62.39 ± 11.85 years old, and the proportion of males was 52.1%. Older age, neurological manifestations, hemorrhage, chemosis, and increased levels of laboratory variables, such as EOS% and BAS% on admission, were found in the non-survival group compared with the survival group. EOS%, BAS%, aspartate aminotransferase (AST), direct bilirubin (DBIL), and older age on admission were noted as independent risk factors for poor prognosis of SFTS patients. The combination of the EOS% and BAS% had an area under the curve (AUC) of (0.82; 95% CI: 0.725, 0.932, P = 0.000), which showed an excellent performance in predicting prognosis of patients with SFTS compared with neutrophil-to-lymphocyte ratio (NLR), and both exhibited a satisfactory performance in predicting poor prognosis compared with De-Ritis ratio (AST/alanine aminotransferase (ALT) ratio). EOS% and BAS% were positively correlated with various biomarkers of tissue damage and the incidence of neurological complications in SFTS patients. Conclusion EOS% and BAS% are effective predictors of poor prognosis of patients with early-stage SFTS. The combination of EOS% and BAS% was found as the most effective approach.


Introduction
Severe fever with thrombocytopenia syndrome (SFTS) is an emerging tick-borne infectious disease caused by a novel phlebovirus (SFTS virus, SFTSV), which belongs to the family phe- [15]. Thus, it is urgent to concentrate on patients infected with SFTSV and to identify the associated risk factors to reduce the number of critically ill and fatal cases.
Previous studies have suggested that the primary role of eosinophils (EOS) and basophils (BAS) is associated with anti-parasitic and allergic reactions. Later, antiviral effects of EOS and BAS were confirmed, which were mainly reported in severe acute respiratory syndrome coronavirus 2 (SARS-COV-2), human immunodeficiency virus (HIV), influenza A viruses, and respiratory syncytial virus (RSV) [16][17][18][19][20][21][22]. Our retrospective study found significant differences in EOS% and BAS% of patients who were weakened compared with those who survived. We observed the elevated EOS% and BAS% in the non-survived group compared with the survived group, which was in contrast with other viral infections. They were also positively correlated with the frequency of neurological manifestations. The present study aimed to investigate EOS % and BAS% in the differential diagnosis and prognostic assessment of patients with SFTS using routine blood tests. In addition, it was attempted to explore the underlying mechanism and to elucidate its clinical significance of SFTS.

Ethics statement
This research was approved by the Ethics Committee of Beijing Ditan Hospital, Capital Medical University (Beijing, China; Approval No. DTEC-KY2022-022-01), and it was conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants.

Study design and patients' enrollment
This retrospective study included 226 SFTS patients from Yantai City Hospital (Yantai, China) between November 2019 and November 2021. The inclusion criteria were as follows: (1) Existence of epidemiological data; (2) Patients with fever (temperature >37.5˚C); (3) Occurrence of thrombocytopenia; (4) Patients with positive-serum nucleic acid test, immunoglobulin G (IgG) and/or IgM antibody for SFTSV, or SFTS isolated from specimens. However, 32 patients were excluded based on the following exclusion criteria: (1) Patients with other viral infections, such as coronavirus disease 2019  and hemorrhagic fever with renal syndrome (HFRS); (2) Patients with autoimmune diseases; (3) Patients with acute and chronic liver diseases; (4) Patients with blood disorders, such as leukemia and idiopathic thrombocytopenia; (5) Patients undergoing radiotherapy or chemotherapy for diverse types of cancer; (6) Patients receiving transfusion of blood products in two weeks; (7) Incomplete clinical data (Fig 1).
melena. Neurological sign was defined as appearance of at least one of the following changes: muscle tension, involuntary movements, and neural reflexes. The observational endpoint was defined as in-hospital death or discharge on improvement.

Statistical analysis
Normally distributed data were expressed as mean ± standard deviation (x ± s), in which they were compared between groups using the independent-samples t-test, and one-way analysis of variance (ANOVA) was utilized for making comparison among multiple groups. Abnormally distributed data were expressed as median (M) with interquartile range (IQR), in which they were compared between groups using the Mann-Whitney U test, and Kruskal-Wallis test was utilized for making comparison among multiple groups. Categorical variables were expressed as percentage (n, %) and were analyzed by the χ2 test or the Fisher's exact test. Univariate and multivariate logistic regression analyses were performed to determine factors associated with the severity of SFTS. However, to identify independent prognostic factors for SFTS, variables with P-values less than 0.1 in the univariate logistic regression were imported into the multivariate logistic regression using the forward stepwise approach. Hosmer-Lemeshow test (H-L test) determined the model's good calibration (predictive accuracy). The predictive performance of the model for in-hospital mortality in early-stage was further evaluated by the receiver operating characteristic (ROC) curve analysis. The ROC curve analysis was used to calculate the optimal cut-off values for EOS% and BAS%. Finally, correlation matrixes were generated using the Spearman correlation coefficient, which did not make any assumption about the underlying distribution. The statistical analysis was conducted using SPSS 25.0 software (IBM, Armonk, NY, USA). A two-sided P < 0.05 was considered statistically significant.

SFTS patients' demographics and clinical characteristics
The study included 194 patients who were admitted to the Yantai City Hospital from November 2019 to November 2021. Patients were assigned into two groups depending on clinical outcomes, including 171 (88.14%) patients in the survival group and 23 (11.86%) patients in the non-survival group. For patients who were diagnosed with SFTS, their demographic and clinical characteristics are summarized in Tables 1-3. Patients' mean age was 62.39 ± 11.85 years old, in which patients in the non-survival group (71.22±11.76 years old) were older than those in the survival patient group (61.20±11.38 years old). Patients in the non-survival group had a shorter hospitalization than those in the survival group, in which 20 (87.0%) cases experienced shortened hospitalization. There were no significant differences between the two groups regarding gender, time from onset to admission, body temperature, history of tick bites, hypertensive disease, diabetes, coronary heart disease (CHD), and history of other diseases. Symptoms of digestive disorders (86.1%) accounted for the highest proportion, including poor appetite, nausea, vomiting, bloating, abdominal pain, and diarrhea, followed by fever (65.5%), and fatigue (63.9%). Compared with the survival group, higher incidence rates of chemosis, hemorrhage, and neurological manifestations were found in the non-survival group. A decrease in platelet (PLT) count was found in all the patients. In the non-survival group, higher levels of EOS%, BAS%, alanine aminotransferase (ALT), aspartate aminotransferase (AST), dehydrogenase (LDH), creatine kinase (CK), alkaline phosphatase (ALP), gamma-glutamyl transpeptidase (GGT), urea, creatinine (CREA), Creactive protein (CRP), procalcitonin (PCT), and lower lymphocyte (LYM)%, mean platelet volume (MPV), Ca 2+ , PLT, and albumin (ALB) were detected compared with those in the survival patients. No significant differences were detected between the two groups for the remaining indicators.

The predictive value of EOS%+BAS% for the prognosis on admission
In our study, the OR value did not significantly change, either after adjusting for age, gender, body temperature, arthralgia, hemorrhage, symptoms of digestive disorders, neurological symptoms and signs, hypertensive disease, and CHD, indicating that the combination of the EOS% and BAS% was a stable risk factor for prognosis of SFTS patients (S2 Table). According to the H-L test (P = 0.294), the combination of EOS% and BAS% had an excellent predictability for prognosis of patients with SFTS compared with NLR, and both had a satisfactory performance in predicting poor prognosis compared with De-Ritis ratio ( Fig 2B) (S3 Table).

The effects of different EOS% levels on the clinical characteristics of SFTS patients
According to the results of the multivariate logistic regression model, EOS% was an independent risk factor for early death in patients with SFTS (OR, 3.215; 95% CI: 1.543-6.699). All patients were divided into EOS% low and EOS% high groups based on the cutoff value (0.35%). The EOS% high group had a higher fatality rate (28.8% vs. 5.6%, P = 0.000), a higher percentage of hospitalization �7 days (48.1% vs. 29.6% P = 0.016), and included more patients with neurological signs (21.2% vs. 9.9%, P = 0.038) compared with the EOS% low group. No significant     Continuous variable data are presented as median (interquartile ranges, IQR). Classified variable dates are presented as n/N (%), where N is the total number of patients with available data. P values comparing the group of survival and the group of non-survival. differences were found between the two groups in terms of age (P = 0.256), gender (P = 0.981), the highest body temperature (P = 0.853), and history of the tick bite (P = 0.136) ( Table 5).
Based on EOS count, patients were assigned into three groups: less than the lower limit of normal (group A), normal (group B), and greater than the upper limit of normal (group C). Group C was associated with a higher rate of death, older age, a shorter hospitalization, and a higher incidence of neurological symptoms and the presence of neurological signs than other  two groups. There were no significant differences among the three groups in terms of gender (P = 0.369), maximum body temperature (P = 0.943), or history of the tick bite (P = 0.072) (S4 Table). The absolute value of EOS was divided into >0 group and equal to 0 group. There were no significant differences in clinical outcomes (P = 0.076), age (P = 0.181), gender (P = 0.746), length of hospitalization (P = 0.133), the highest body temperature (P = 0.910), neurological symptoms (P = 0.076), and neurological signs (P = 0.066) between the two groups (S5 Table).

Correlation between circulating EOS% and the frequency of neurological manifestations in SFTS patients
Through Spearman correlation analysis, it was revealed that EOS% was positively correlated with the frequency of neurological symptoms in SFTS patients (r = 0.158, P = 0.028) and was positively correlated with the frequency of neurological signs (r = 0.180, P = 0.012) (S6 Table). Continuous variable data are presented as median (interquartile ranges, IQR).
Classified variable dates are presented as n/N (%), where N is the total number of patients with available data. P values comparing between the group of EOS% low and the group of EOS% high . https://doi.org/10.1371/journal.pntd.0010967.t005

The effects of different BAS% levels on the clinical characteristics of SFTS patients
BAS% was found as an independent risk factor for patients with early-stage SFTS (OR, 2.290; 95% CI: 1.156-4.535, P = 0.017). All patients were divided into BAS% low and BAS% hight groups based on the cutoff value (0.17%). The BAS% high group had a higher fatality rate (22.4% vs. 3.7%, P = 0.000), older age (65.60±11.09 vs. 59.89±11.87, P = 0.001), and a shorter hospitalization (9.0 days, IQR: 4.5-12.5 vs. 11.0 days, IQR: 6.0-13.0, P = 0.012) than the BAS% low group. No significant differences were found between the two groups in terms of gender (P = 0.051), the highest body temperature (P = 0.065), history of the tick bite (P = 0.207), neurological symptoms (P = 0.680), and neurological signs (P = 0.394) ( Table 6). According to basophilic count, patients were divided into two groups: normal range and greater than the upper limit of normal. The group of greater than the upper limit of normal was associated with a higher rate of death (50.0% vs. 10.1%, P = 0.007) and a higher incidence of neurological symptoms (50% vs. 10.2%, P = 0.008) and the presence of neurological signs (50% vs. 11.3%, P = 0.008) than the normal range group. No significant differences were found between the two groups in age (P = 0.145), gender (p = 0.890), the highest body temperature (P = 0.979), or history of the tick bite (P = 0.305) (S7 Table).

Correlation between circulating BAS% and clinical parameters of SFTS patients
Through Spearman correlation analysis, it was revealed that BAS% was

Discussion
In the present study, EOS% and BAS% were for the first time used as variables to predict clinical outcomes of early-stage SFTS patients. The combination of EOS% and BAS% exhibited a satisfactory predictive performance compared with previously reported measures related to clinical outcomes. We also found that EOS% and BAS% were associated with neurological symptoms and signs. Patients mainly presented with thrombocytopenia, liver dysfunction, elevated biomarkers of tissue damage, and a higher frequency of neurological-related manifestations, particularly in non-survived patients, which is in parallel with previous studies [3], [23,24]. EOS are bone marrow-derived leukocytes. As research has progressed, a comprehensive understanding of the critical role of EOS in immunity and host defense has emerged [20,25]. EOS has been used as an indicator for disease progression and outcomes. In our cohort, EOS% was noted as an independent risk factor for death on admission and was positively associated with neurological signs and/or symptoms. Eosinophilic Cationic Protein (ECP) is one of the main components of EOS, and the level of ECP was reported to be positively correlated with EOS% and was associated with neurological damage [26,27]. ECP can alter the permeability of cell membranes, subsequently causing calcium influx, which can ultimately lead to cell apoptosis. [28] In addition, Peng et al. showed that the increased intracellular cation levels lead to the sequential activation of the caspase-9, pro-caspase-3 and 8, inducing apoptosis of neuronal cells [29]. In our study, EOS% was found to be positively correlated with the incidence of neurological symptoms and signs, with no significant correlation with delirium, stupor, somnolence, and coma, which could be related to the inadequate number of cases with associated symptoms.
Basophils are an essential component of innate immunity and are also a promoter of type 2 immune responses, which play a role in parasitic infections, allergic reactions, and viral infections. In our study, BAS% was identified as a predictor of poor prognosis of early-stage SFTS patients, which was consistent with studies on the COVID-19 [30,31]. However, indifferent to COVID-19 [32], the relative elevation of basophils in the non-survived group compared with that in the survived group may be related to tick bite transmission. [33] With multiple pattern recognition receptors on basophils, such as Dendritic Cell-Specific Intercellular adhesion molecule 3-Grabbing Nonintegrin (DC-SIGN) and C-type lectin, basophils may provide a stable cellular basis for HIV capture and transmission [21,34]. Importantly, several studies found that SFTS enters host cells via these two receptors, thereby involving basophils as one of the target cells for SFTS [35,36]. Studies have shown that SFTSV infection drove macrophage differentiation skewed to M2 phenotype, which facilitated virus shedding, and resulted in viral spread [37]. Interleukin-4 (IL-4) production from basophils can contribute to the differentiation of macrophages towards the M2 phenotype. [38] In addition to promoting M2 macrophage differentiation, IL-4 can cause microvascular infiltration and a procoagulant state through remodeling and upregulation of the expression levels of vascular cell adhesion molecule-1 (VCAM-1) and monocyte chemoattractant protein-1 (MCP-1), which may result in damage to the endothelium [39]. IL-4 induces T cell differentiation towards the TH2 phenotype, and a significant correlation of Th1/Th2 with disease severity in SFTS patients was reported [40,41]. In addition, the activation of BAS may cause the release of large amounts of cytokines, such as IL-6 and IL-8, which are essential components of the cytokine storm and are associated with the poor prognosis of SFTS patients [42,43]. Hence, we hypothesized that organ failure in SFTS patients could be attributed to immune dysfunction associated with the involvement of BAS.
Prediction of the clinical outcomes by innate immune cells and immune checkpoints has been frequently reported. Studies have shown that immune checkpoints are associated with viral escape from host immunity [44,45]. In the study of COVID-19, programmed deathligand 1 (PD-L1), one of the immune checkpoints, was highly expressed in EOS and BAS in severe patients, and it was positively correlated with sequential organ failure assessment (SOFA) scores, providing a new idea for subsequent studies on SFTS [46].
There are still some limitations in this study. Firstly, the small sample size should be noted, as well as the lack of viral load data, and there was no validation cohort. Secondly, the cerebrospinal fluid of patients was not examined to assess the cause of neurological symptoms because of thrombocytopenia. Finally, the role of EOS and BAS in systemic tissue damage in SFTS patients was not investigated. Hence, it is essential to eliminate the abovementioned limitations.
In conclusion, both EOS% and BAS% were found as independent risk factors for poor prognosis of patients with early-stage SFTS, and combination of EOS% and BAS% was the most effective approach. EOS% and BAS% are rapid, simple, effective, and inexpensive prognostic markers, and they may be efficacious for diagnosing and treating a variety of diseases.
Supporting information S1 Fig. EOS% and BAS% were elevated in the non-survival group compared to the survival group and were positively correlated with the incidence of neurological complications, and their combination was highly predictive of the prognosis of SFTS patients. (Created with BioRender.com). (TIF) S1