key: cord-333749-45v6b4tc
authors: Xie, Guogang; Ding, Fengming; Han, Lei; Yin, Dongning; Lu, Hongzhou; Zhang, Min
title: The role of peripheral blood eosinophil counts in COVID‐19 patients
date: 2020-06-20
journal: Allergy
DOI: 10.1111/all.14465
sha: 
doc_id: 333749
cord_uid: 45v6b4tc

BACKGROUND: Coronavirus disease 2019 (COVID‐19) emerged in Wuhan city and rapidly spread globally outside China. We aimed to investigate the role of peripheral blood eosinophil (EOS) as a marker in the course of the virus infection to improve the efficiency of diagnosis and evaluation of COVID‐19 patients. METHODS: 227 pneumonia patients who visited the fever clinics in Shanghai General Hospital and 97 hospitalized COVID‐19 patients admitted to Shanghai Public Health Clinical Center were involved in a retrospective research study. Clinical, laboratory, and radiologic data were collected. The trend of EOS level in COVID‐19 patients and comparison among patients with different severity were summarized. RESULTS: The majority of COVID‐19 patients (71.7%) had a decrease in circulating EOS counts, which was significantly more frequent than other types of pneumonia patients. EOS counts had good value for COVID‐19 prediction, even higher when combined with NLR. Patients with low EOS counts at admission were more likely to have fever, fatigue and shortness of breath, with more lesions in chest CT and radiographic aggravation, longer length of hospital stay and course of disease than those with normal EOS counts. Circulating EOS level gradually increased over the time, and was synchronous with the improvement of chest CT(12days vs 13days, P=0.07), later than that of body temperature(12days vs 10days, P=0.014), but was earlier than the negative conversion of nucleic acid assays(12days vs 17days, P=0.001). CONCLUSION: Peripheral blood EOS counts may be an effective and efficient indicator in diagnosis, evaluation and prognosis monitoring of COVID‐19 patients.

Since December 2019, many pneumonia patients with unknown cause have emerged in Wuhan, the capital city of Hubei province in China. Fever, fatigue, dry cough are common symptoms at the onset, and progressive dyspnea occurs in severe cases. The typical chest CT findings showed peripheral pulmonary plaques and interstitial lesions, which were very similar to viral pneumonia [1] . This pathogen has currently been named severe acute respiratory syndrome corona virus 2 (SARS-COV-2) , which has a phylogenetic similarity to SARS-CoV [2, 3] . By the time of writing, the number of patients infected by the virus has now reached 143000 and over 100 countries have reported confirmed cases.

To prevent transmission, how to figure out the potential suspected COVID-19 pneumonia patients and isolate them immediately is now the priority for physician in fever clinics in China. And also, for the confirmed COVID-19 patients, most patients have mild symptoms which may be indistinguishable clinically from common cold at the early stage of infection. However, in a median of 8 days from onset [4] , nearly fifteen to twenty percent of them will exacerbate with progressive dyspnea abruptly and rapidly develop into acute respiratory distress syndrome(ARDS) or end-organ failure. The characteristics of potential critical patients are still unclear. Therefore, how to use appropriately simple and effective method to screen out potentially serious patients is important for the prognosis of the disease.

In the process of diagnosing and treating COVID-19 patients, we found that peripheral blood eosinophils (EOS) significantly reduced among most patients regardless of the severity of the diseases at the early stage, which had not been reported in severe acute respiratory syndrome (SARS) and Middle East respiratory syndrome (MERS) [5, 6] , or the other types of pneumonia. Here

This article is protected by copyright. All rights reserved we aim to discuss this interesting phenomenon, and try to clarify its clinical significance in COVID-19. We hope our findings of the EOS in convenient routine blood test will be helpful for differential diagnosis and evaluating the prognosis of COVID-19 patients.

Study population (Figure 1 Secondly, we obtained data of hospitalized COVID-19 patients who admitted to

Public Health Clinical Center from January 20 to February 20, 2020. COVID-19 was confirmed according to WHO interim guidance [7] . This is a retrospective cohort study. The ethics committee of Shanghai General Hospital approved this study and granted a waiver of informed consent from study participants.

This article is protected by copyright. All rights reserved Data collection Characteristics of subjects from Shanghai General Hospital were collected as follows: age;

gender; duration of fever; accompanying symptoms of fever; COVID-19-related epidemiological history; body temperatures. Blood routine tests included following parameters: red blood cell (RBC) counts, hemoglobin, white blood cell (WBC) counts, percentage and absolute counts of neutrophils, lymphocytes, monocytes, EOS, and basophils, and C-reactive protein (CRP). Chest HRCT scans (slice thickness was 0.625mm, GE medical system) were performed.

In Shanghai Public Health Clinical Center, we reviewed electronic records to collect clinical charts, nursing records, laboratory findings, and radiologic assessments for all patients. Radiologic assessments included chest radiography or computed tomography (CT). Laboratory and radiologic examinations were performed every 2-3 days. The end point was discharge from hospital or death. Epidemiological, demographic, clinical, laboratory, management, and outcome data were collected and recorded with standardized data collection forms.

Throat swab samples were obtained from all patients once a day and tested using real-time reverse transcriptase-polymerase chain reaction assays.

All data were recorded and checked separately by two qualified researchers.

This article is protected by copyright. All rights reserved Definitions Fever was defined as an axillary temperature above 37.5°C. Hypoxaemia was defined as arterial oxygen tension (PaO 2 ) below 60mmHg when breathing air, or PaO 2 over inspiratory oxygen fraction (FIO 2 ) of less than 300 mmHg. Severe and non-severe cases were defined according to WHO interim guidance [7] . Influenza pneumonia was diagnosed based on chest CT showing interstitial lesions, accompanied by flu-like symptoms (fever>38.5℃, accompanied with cough or

This article is protected by copyright. All rights reserved sore ts sore roat), and positive serum influenza A or B IgM. Decrease in circulating EOS counts was defined as the absolute value of peripheral blood EOS being below the lower limit of the normal range of the test (<0.02 ×10 9 /L), so was lymphopenia (<1.0 ×10 9 /L). Fatigue was defined as a feeling of extreme physical or mental tiredness. Shortness of breath, or dyspnea, was defined as a feeling of difficult or labored breathing that was out of proportion to the patient's level of physical activity.

Continuous variables were expressed as mean (SD) and compared with T test if they were normally, median (IQR) and compared with the Mann-Whitney U test if they were not; categorical variables were expressed as counts with percentages and compared by χ² test or Fisher's exact test.

Pearson correlation analysis was used to investigate the correlation of continuous variables. Receiver operating characteristics (ROC) curve analysis was performed to evaluate the diagnostic ability. A two-sided α of less than 0·05 was considered statistically significant. All the analyses were performed with the use of SPSS (version 20.0).

A total of 227 fever clinics outpatients of pneumonia were enrolled, 36 cases of suspected

This article is protected by copyright. All rights reserved patients isolated immediately, and finally 12 cases confirmed with COVID-19. The pathogen was not clearly identified in the remaining suspected COVID patients, but they were all received close observation and their conditions were significantly improved in the end. Meanwhile, 15 cases of influenza pneumonia, and 176 cases of other types of pneumonia were diagnosed. Conditions of the patients on admission were shown in Table 1 . There was no significant difference in age and gender among the patients in each group. Duration of fever was 1.43 ± 0.67 days for total subjects. No significant difference in the median interval from the onset of fever to hospital visit of each group of patients was showed. In addition to fever, the most common symptoms were cough and fatigue. Fatigue was more common in COVID-19 patients, suspected patients and influenza pneumonia patients than other types of pneumonia patients (66.7%, 58.3%, 66.7% vs 39.8%).

For blood parameters, lower counts of EOS were more frequently found in COVID-19 group compared with the other three groups. Three COVID-19 patients' circulating EOS vanished, while the patients in other groups rarely shared the same results. Lower percentages of white blood cells and neutrophils were found both in COVID-19 patients group and suspected patients group compared with the other types of pneumonia group, while no difference was seen in lymphocytes.

CRP was higher in other types of pneumonia subjects than in confirmed and suspected COVID-19 patients. (Table 1 ).

The data of EOS counts and neutrophils lymphocytes ratio (NLR) of all patients with COVID-19 (109 cases) and non COVID-19 pneumonia (215 cases) were extracted, whose predictive values were evaluated by means of ROC curves. The EOS counts had 

We collected 97 hospitalized patients with laboratory-confirmed COVID-19 including 85 non-severe patients and 12 severe patients in Shanghai Public Health Center for analysis. All non-

This article is protected by copyright. All rights reserved severe patients and 9 severe patients had been discharged, while three severe patients had received mechanical ventilation and one of them had died finally. The median age diabetes (5.2%), which were comparable in two groups. The most common symptoms were cough (68.0%), followed by fever (60.8%), fatigue (38.1%), sputum production (37.1%) and shortness of breath (21.6%). Symptoms presented by subjects in severe patients group including fever (91.7%), cough (66.7%) and shortness of breath (83.3%), similar to the results of other studies [8] , while only 12.9% of the non-severe patients developed shortness of breath. On admission, all patients with chest CT scan had abnormal results, the majority (83.5%) had multilobular lesions, which could be seen in most COVID-19 severe patients (91.7%). The representative chest CT images showed bilateral multiple lobular and subsegmental areas of groundglass opacities or consolidation. The median interval from the onset of symptoms to hospital admission for all patients was 7 days (IQR, [5] [6] [7] [8] with no significant difference between severe and non-severe patients(p=0.314), followed by median time of hospitalize as 12 days (IQR, [9] [10] [11] [12] [13] [14] , and the total disease course was 19 days(IQR, [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] . The length of hospitalization and the total course of illnesses in severe patients were significantly longer than non-severe patients(p=0.045), as expected ( Table 2) .

EOS counts decreased below 0.02 ×10 9 /L in 71.7% of COVID-19 patients, including 45.4% of patients dropped below the lower limit of detection, and the EOS counts Accepted Article decreased below the lower This article is protected by copyright. All rights reserved limit of detection in all 12 severe patients. In contrast, the reduction of lymphocytes in these two patients' groups was not showed significant difference, with 35.3% in nonsevere patients and 41.4% in severe patients, respectively. However, there was a positive correlation between EOS and lymphocyte levels in the two groups of patients (r=0.414, p<0.05). We also found significant differences in neutrophil-to-lymphocyte ratio (NLR) (p=0.026) and plasma D-dimer level (p=0.014) between severe and non-severe patients ( Table 3 ).

We divided COVID-19 patients into two groups based on the circulating EOS counts when admitted to the hospital, low EOS group(< 0. 

This article is protected by copyright. All rights reserved We started to calculate the course of the disease from the onset. Available data of peripheral

This article is protected by copyright. All rights reserved blood EOS counts from all discharged patients were superimposed according to the distribution. The average value of the EOS counts in the same course was regarded as the value of the day, and the incidence of EOS reduction(The ratio of the number of EOS counts below 0.02×10 9 /L to the total number of EOS counts) on every day since onset was also calculated.

It showed that EOS in peripheral blood decreased significantly from the onset. In the first 4

days of the course, EOS level of the patients was lower than normal significantly with over 80%

of EOS counts below 0.02×10 9 /L on the fourth day. EOS level gradually increased over the time, and fully recovered and reached its peak on the 16th day. At the same time, the incidence of EOS reduction progressively decreased and dropped below the lower limit of detection on the 18th day, suggesting that all patients' EOS counts return to normal level ( Figure 3A ). However, lymphocytes did not show such a significant trend ( Figure   3B ).

We also compared the recovery trend of EOS in severe and non-severe patients. The EOS counts of severe patients began to recover slowly after ten days since onset, while of non-severe patients recover much faster from the trough point on the 4th day ( Figure   3C ). Thus, there were obvious differences in recover speed according to disease severity.

Assessment on the improvement of disease was according to normal body temperature, improvement of chest CT evidence and negative conversion of nucleic acid assays. We compared the recovery time of EOS counts in peripheral blood with the above three indicators, and found that the recovery time of EOS counts was slightly shorter than that of chest CT (12days, Figure 3F、I ).

We also superimposed the incidence of the recovery (The ratio of the number of normal cases to the total cases) of the above indicators including body temperature, improvement of chest CT and EOS counts, observing their changing trends according to the course of the disease. We can clearly observe that fully recovered of the body temperature on 15th day, which in the first place, the second place was EOS on 18th day, and then followed chest CT image on 19th days in all discharged patients ( Figure 3J ).

In previous reports, patients hospitalized with SARS-COV-2 infection showed leucopenia and lymphopenia [4, 9] , which is similar to SARS and MERS infections and indistinguishable from other viral respiratory infections such as influenza [10] . Our study discovered that fewer patients

This article is protected by copyright. All rights reserved with SARS-COV-2 appeared leukopenia, less than half had lymphopenia, while nearly three-quarters had a reduction in circulating EOS, even disappeared at the onset of disease, and regardless the severity of the disease. It was a unique characteristic compared with other types of pneumonia, and may have a role in diagnosis in COVID-19 patients, which was not mentioned before.

The clinical course of COVID-19 demonstrated the complexity of the COVID-19

profile with different clinical presentations. Clinical manifestations ranged from asymptomatic cases to patients with mild and severe symptoms, with or without pneumonia [11] .It was difficult to distinguish the radiologic manifestations between patients infected SARS-COV-2 and other respiratory pathogens, and almost half COVID-19 patients' temperature was normal at the beginning. Our analysis of fever clinic patients with pneumonia found that EOS counts of peripheral blood in patients with COVID-19 were significantly reduced, which was further confirmed by the data in hospitalized COVID-19 patients. Data analysis showed that decreased EOS counts were more common in COVID-19 patients than other types of pneumonia, and no significant difference was identified between severe and non-severe patients, which was also mentioned in patients from Wuhan [12, 13] or outside Wuhan [14] . Nearly half patients, especially for the severe patients could not be detected circulating EOS at all, demonstrating that the decreased EOS counts may be an important diagnostic clue for SARS-CoV-2 infection in suspected patients with atypical symptoms and radiographic infiltration with or without lymphopenia. The good value of EOS counts for COVID -19 prediction was showed in our study, when combined with NLR, the predictive value was even higher, indicating the advantage of the E (EOS counts < 0.017 ×10 9 /L) NL (NLR<2.425) model over EOS alone in COVID-19 prediction. Currently, COVID-19

This article is protected by copyright. All rights reserved patients were confirmed or excluded by nucleic acid assay, while false negative was unavoidable [15] . In our study, according to peripheral blood EOS counts or ENL model, though the nucleic acid was negative, the possibility of false negative should also be considered.

So far, chest CT infiltration range might be helpful to figure out the potential critical patients at the early stage, but non-severe patients also presented with diffuse abnormal damage without hypoxemia, while some patients with a small part of lung involved at the beginning quickly developed to severe cases. Some researchers reported that patients with age ≥ 50 and NLR ≥ 3.13 facilitated severe illness [16] . In our study, the majority of mild COVID-19 patients had a significant decrease in EOS level from the onset, and then increased gradually. The maximum EOS reduction was on 4th day from onset and began to restore in two-thirds of non-severe patients within the following 3 days, while all severe patients remained undetected. The restore of EOS counts was almost synchronous with the improvement of chest CT, and it was after that of body temperature, but was earlier than the negative conversion of nucleic acid assays. Patients with low EOS counts at admission were more likely to have fever, fatigue and shortness of breath with more deterioration of lung by CT scan than those with normal EOS counts, suggesting that low EOS counts may be related to severe conditions. The insight can also be convinced by the monitor of EOS in severe patients. That means we should pay more attention to monitoring the circulating EOS, and if it failed to be increased timely, and remained at low level, severe stage may

This article is protected by copyright. All rights reserved develop soon. Meanwhile, increase of circulating EOS may indicates the favorable prognosis in COVID-19 patients.

Why circulating EOS disappeared at the onset of COVID-19 patients regardless of severity is still mysterious. EOS develop in the bone marrow microenvironment from multipotent hematopoietic stem cells, which give rise to a population of unique eosinophil-committed progenitors that are capable of terminally differentiating into mature EOS [17] . The intestinal and urogenital tracts [18] . EOS increase highly in parasitic infection and allergic diseases, while significantly decrease in patients with acute infectious diseases such as typhoid, with major surgery and burns, and sepsis. A recent study showed that numbers of mature eosinophils in the blood and bone marrow markedly declined compared with baseline after endotoxin administration for four hours, whereas numbers of all eosinophil progenitors did not change [19] . In COVID-19, the reduction of peripheral blood EOS began early regardless of severity may be caused by decrease of bone marrow release and increase organ recruitment. Most viral infections caused decrease in circulating EOS in the blood except human immunodeficiency virus infection [20] . In a variety of viral infections, including viral myocarditis and respiratory syncytial virus (RSV) pneumonia [21] , tissue EOS in the absence of blood EOS has been described. Whether EOS play a role in antiviral defense, are responsible for tissue

This article is protected by copyright. All rights reserved destruction, or are simply recruited to sites of tissue damage is unknown. Data from experimental murine infection with RSV and in influenza A support the hypothesis that EOS play a predominantly protective role [22, 23] . When examining the role of EOS in antiviral host defense, researchers found that EOS, along with neutrophils, were recruited to the lung tissue early in the course of infection and following infection and preceded the developments of respiratory symptoms [24] . Percopo [22] patients so far [25, 26] . It had been observed for a long time that viral infection was closely related to bone marrow suppression [27] ,which was mainly caused by the direct or immune damage of the virus to bone marrow stem cells or stromal cells, resulting in microenvironment destroy. Whether SARS-COV-2 could also affect bone marrow function through the above mechanism and cause the decrease of peripheral blood eosinophils was still unknown. On the other hand, Huang [4] the stress and then EOS release in bone marrow was suppressed [28] . Corticosteroids

This article is protected by copyright. All rights reserved promote cell death and clearance of EOS, and impair EOS survival and differentiation [29] , but intra-nasal corticosteroid (including spray) seemed not to affect the course of COVID-19 [30] .

Our study has some deficiencies. Firstly, laboratory and radiologic examinations could not be performed every day, so these data were not continuous monitoring results.

Secondly, the sample size of severe patients enrolled in our study was relatively small.

In conclusion, peripheral blood EOS counts may be a more convenient and effective indicator in addition to other blood parameters and CT scan in diagnosis and evaluation of COVID-19 patients. Circulating EOS level lower than 0.02×10 9 /L may not only play a role in identifying suspect patients, which would be better combined with NLR in confirming the diagnosis, but also be important in predicting severe status and indicating the favorable prognosis.

Moreover, the test of peripheral blood EOS counts is fast, simply and inexpensively, which can be chosen as an effective monitor parameter during the diagnosis and treatment of COVID-19. White blood cell (4.0-10.0×10 9 /L) < 4 4-10 > 10 Neutrophils (2.0-6.0×10 9 /L) < 2 2-6 > 6 Lymphocytes (1.0-3.5×10 9 /L) < 1 ≥ 1 Monocytes (0.1-0.6×10 9 /L) < 0.6 ≥ 0.6 Eosinophils (0.02-0. 

A novel coronavirus from patients with pneumonia in China

Genomic characterisation and epidemiology of 2019 novel coronavirus: implications for virus origins and receptor binding

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

SARS: clinical features and diagnosis

Commentary: Middle East respiratory syndrome coronavirus (MERS-CoV): announcement of the Coronavirus Study Group

Clinical management of severe acute respiratory infection when novel coronavirus (2019-nCoV) infection is suspected: interim guidance

Clinical Characteristics of Coronavirus Disease 2019 in China

Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study

Accepted Article This article is protected by copyright. All rights reserved

The continuing 2019-nCoV epidemic threat of novel coronaviruses to global health-The latest 2019 novel coronavirus outbreak in Wuhan

Eleven faces of coronavirus disease 2019

Clinical characteristics of 140 patients infected by SARS-CoV-2 in Wuhan

Dysregulation of immune response in patients with COVID-19 in Wuhan, China

COVID-19 in a designated infectious diseases hospital outside Hubei Province, China

Laboratory diagnosis and monitoring the viral shedding of 2019-nCoV infections

Neutrophil-to-Lymphocyte Ratio Predicts Severe Illness Patients with 2019 Novel Coronavirus in the Early Stage

Contributions of Eosinophils to Human Health and Disease

Eosinophils and eosinophil-associated diseases: an update

Differentiation and activation of eosinophils in the human bone marrow during experimental human endotoxemia

Eosinophilia in patients infected with human immunodeficiency virus

Accepted Article This article is protected by copyright. All rights reserved

Eosinophils capture viruses, a capacity that is defective in asthma

Activated mouse eosinophils protect against lethal respiratory virus infection

Eosinophils Promote Antiviral Immunity in Mice Infected with Influenza A Virus

Eosinophils, eosinophil ribonucleases, and their role in host defense against respiratory virus pathogens

COVID-19 Autopsies

Role of Bone Marrow in Pathogenesis of Viral Infections

Efficacy and safety of glucocorticoids in the treatment of severe community-acquired pneumonia: A meta-analysis

Targeting eosinophils in allergy, inflammation and beyond

Intranasal corticosteroids in allergic rhinitis in COVID-19 infected patients: An ARIA-EAACI statement

where N is the total number of patients with available data. p values comparing the group of nonsevere and severe patients are from χ² test or Mann-Whitney U test. The primary composite end point was discharged from hospital

We would like to acknowledge all health-care workers involved in the diagnosis and treatment of patients in Shanghai. We thank the Shanghai General Hospital for coordinating data collection for patients with pneumonia; we thank Shanghai Public Health Clinical Center for sharing data collection.None of the authors have any conflict of interest to declare.(Dr. Xie has nothing to disclose. Dr. Ding has nothing to disclose. Dr. Han has nothing to disclose.Dr. Yin has nothing to disclose. Dr. Lu has nothing to disclose. Dr. Zhang has nothing to disclose.)

This article is protected by copyright. All rights reserved