Decreased Eosinophil Counts and Elevated Lactate Dehydrogenase Predict Severe COVID-19 Patients with Underlying Chronic Airway Diseases

Lingling Yi Tongji Hospital of Tongji Medical College of Huazhong University of Science and Technology Dian Chen Tongji Hospital of Tongji Medical College of Huazhong University of Science and Technology Shuchen Zhang Tongji Hospital of Tongji Medical College of Huazhong University of Science and Technology Yuchen Feng Tongji Hospital of Tongji Medical College of Huazhong University of Science and Technology Wenliang Wu Tongji Hospital of Tongji Medical College of Huazhong University of Science and Technology Chenli Chang Tongji Hospital of Tongji Medical College of Huazhong University of Science and Technology Shengchong Chen Tongji Hospital of Tongji Medical College of Huazhong University of Science and Technology Guohua Zhen (  ghzhen@tjh.tjmu.edu.cn ) Tongji Hospital of Tongji Medical College of Huazhong University of Science and Technology

of COVID-19 con rmed cases and 603 691 deaths were reported in over 215 countries worldwide [1], demanding an urgent need for early identi cation for severe cases. The SARS-CoV-2 virus, which belongs to the betacoronavirus, is highly homologic (with 88% identity) to two bat-derived SARS-like coronaviruses, while more distant from SARS-CoV (around 79%) and Middle East respiratory syndrome coronavirus (MERS-CoV, around 50%) [2]. Clinical evidence has suggested that SARS-CoV can be transmitted from person to person via direct contact or through droplets from infected individuals [3,4].
SARS-CoV-2 is able to attack the respiratory system through binding the cell entry receptors angiotensinconverting enzyme 2 (ACE2) on airway epithelial cells and results in pneumonia and respiratory failure in critically ill patients.
Chronic bronchitis, chronic obstructive pulmonary disease (COPD), and asthma are common respiratory diseases with chronic airway in ammation [5][6][7][8][9]. Eosinophils, neutrophils, and macrophages in innate immune response signi cantly increase in the airway and lung during the initial phase of in ammation.
Subsequently, activated adaptive immunity leads to the recruitment of T and B lymphocytes. Th1, Th2, and Th17 cells play a crucial role in COPD, asthma, and chronic bronchitis, resulting in mucus overproduction and air ow obstruction [5,8]. Lymphocytopenia, however, has been reported in several studies in severe patients infected with SARS-CoV-2 [10][11][12]. Recently, circulating eosinophil counts were also reported to be decreased in COVID-19 patients, and associated with the severity of the disease [13,14]. Therefore, patients with underlying COPD, asthma, and chronic bronchitis may have different in ammatory states after SARS-CoV-2 infection compared to patients without chronic airway in ammation. [10][11][12][15][16][17]. In this study, we aimed to identify the potential predictors for the disease severity of COVID-19 patients with underlying chronic airway diseases including chronic bronchitis, COPD, and asthma.
In this retrospective cohort study, we reviewed medical records of 59 laboratory-con rmed COVID-19 patients with underlying chronic airway in ammation and compared the demographic, clinical, and radiological characteristics as well as laboratory results between severe and non-severe patients in this cohort. Potential predictors of disease severity were identi ed in the abnormal laboratory ndings using univariate and multivariate regression models.

Results
Demographics and clinical characteristics of non-severe and severe COVID-19 patients with chronic airway diseases A total of 1888 patients were admitted. Fifty-nine patients with underlying chronic airway in ammation, including COPD (0.95%), asthma (0.53%), and chronic bronchitis (1.64%) were con rmed to have SARS-CoV-2 infection. Thirty-three patients were classi ed as non-severe patients and twenty-six patients were classi ed as severe patients. Although COPD was more common in severe COVID-19 patients when compared with non-severe COVID-19 patients (42% vs. 21%), the difference was not statistically signi cant.
The median age for all patients was 71 years (interquartile range, 57-80) and more than half of them (54%) were over 70 years old. The majority (71%) of patients were male (Table 1). There was no signi cant difference in age and sex between non-severe and severe patients. Thirty-one (53%) patients had one or more comorbidities besides the three chronic airway diseases, with cardiovascular disease (46%) and endocrine system disease (15%) being the most common comorbidity. There were no signi cant differences in the presence of these comorbidities between the non-severe and severe COVID-19 patients. Half of the patients had smoking histories or current smokers. The most common symptoms were fever (83%), cough (73%), fatigue (47%) and dyspnea (42%). Dyspnea was more common in severe patients compared to non-severe patients (65% vs. 24%, p = 0.001) (  Table 2). Of note, signi cant differences in the expression of in ammation-related cytokines including interleukin (IL)-6, IL-8 and tumor necrosis factor (TNF)-α were observed between the two groups, which were dramatically increased in severe patients.

Discussion
In this retrospective cohort study, we found that eosinophil counts less than 0.02 × 10 /L and LDH levels greater than 225 U/L on admission were associated with the severity of COVID patients with underlying chronic bronchitis, COPD and asthma. Moreover, eosinophil counts and LDH levels tend to return to a normal range in severe and non-severe patients after treatment. Eosinophil counts and LDH levels tend to return to normal range over time in non-severe patients We further analyzed the eosinophil counts and LDH levels in non-severe and severe COVID-19 patients with chronic bronchitis, COPD, and asthma, respectively. We found that there was a signi cant difference in eosinophil counts and LDH levels between severe and non-severe patients with chronic bronchitis and COPD whereas not in patients with asthma (Fig. 1). To observe the dynamic changes of eosinophil counts and LDH levels over time, we collected the eosinophil counts and LDH levels on the 5th, 10th, 15th, 20th, 25th, and 30th day after admission. We found that eosinophil counts increased over time both in severe and non-severe patients. Meanwhile, LDH decreased over time (Fig. 2). Severe patients showed a slower recovery rate than non-severe patients, especially eosinophil counts. Of note, both eosinophil counts and LDH levels recovered more slowly in severe patients with COPD than those in severe patients with chronic bronchitis and asthma. Our data suggest that, as the disease improved, eosinophil counts and LDH levels tend to return to normal range both in severe and non-severe patients, indicating a good therapeutic effect of patients with chronic airway diseases in COVID-19 treatment.
We further performed multivariate analysis for mortality in COVID-19 patients with chronic airway in ammation using the above four variables and found that eosinophil counts < 0.02 × 10 /L (odds ratio per 1 unit increase, 18 Fig. 1). This suggests that eosinopenia and elevated LDH are also potential predictors for the mortality of COVID-19 patients with underlying chronic airway diseases.
Circulating and tissue-resident eosinophils are associated with a variety of diseases, in which eosinophils participate in the pathological process and play a potent proin ammatory role, such as COPD, asthma, and chronic bronchitis. Previously, human eosinophil has been reported to play an important role in virus detection and defending through several Toll-like receptors (TLRs), including TLR1, TLR3, TLR4, TLR7, TLR9, and TLR10 [18][19][20][21]. Single-stranded RNA viruses, such as coronavirus, can be recognized by eosinophils in the airway tract through TLR7, whose subsequent stimulation triggers eosinophil cytokine expression and nitric oxide (NO) generation to promote viral clearance [19][20][21][22]. In view of elevated eosinophils in patients with chronic airway in ammation, COPD, asthma and chronic bronchitis have not yet been reported as major risk factors for the severity of SARS-CoV-2 infections. According to an ambispective cohort study of 548 COVID-19 patients, only 5 cases of asthma were identi ed, signi cantly lower than previously reported asthma prevalence in Wuhan (6.4%) [23][24][25][26]. patients [28]. A recent study has highlighted the signi cant role of CD101 − eosinophils in suppressing acute lung injury and respiratory failure [29]. Therefore, eosinophil could have helped patients with chronic airway in ammation escape from SARS-CoV-2 infections and has been identi ed as a probable potential indicator for prognosis in COVID-19. Jackson et al found a negative correlation between ACE2 expression in airway epithelium and peripheral blood eosinophil counts, which could explain the reason why severe patients were more vulnerable to SARS-CoV-2 infection [30]. Meanwhile, eosinopenia was more common in critically severe patients, suggesting that the resolution of eosinopenia could be a possible way to improve clinical status [31].
In our study, lower expression of eosinophil showed worse survival probability and eosinophil counts signi cantly decreased in severe COVID-19 patients with chronic bronchitis and COPD. No signi cant difference was observed in asthma patients, partly due to the limited sample size. We further explored dynamic changes of eosinophil counts in patients with chronic airway diseases in the course of COVID-19 and found that eosinophil counts gradually increased over time and returned to normal range in both severe and non-severe patients. It still remains unclear how eosinopenia takes place in COVID-19, but possible mechanisms of decreasing eosinophils could be inhibition of eosinopoiesis and egress of eosinophils from the bone marrow [32,33], the reduction of chemokine receptors or adhesion factors [34], and interferon (IFN) mediated eosinophil apoptosis during the virus infection [33].
LDH has long been reported to be associated with COPD, asthma, and chronic bronchitis and identi ed as a potential marker of chronic airway in ammation [35][36][37]. Meanwhile, a large number of studies reported elevated LDH levels in COVID-19, which could be a risk factor of mortality [10][11][12][38][39][40][41]. Zheng et al conducted a systematic literature review and meta-analysis including 4 studies, a total of 1286 cases, and found that LDH was statistically signi cantly higher in severe patients compared to nonsevere patients [38]. Elevated LDH in severe cases indicated diffuse lung injury and tissue damage [38,42], therefore, we hypothesized that LDH might be another predictor of chronic airway in ammation exacerbation in COVID-19. Kaplan-Meier survival analysis suggested the hazard of elevated LDH levels. Similar to eosinophil, LDH showed elevated levels in severe COVID-19 patients with chronic bronchitis and COPD, and gradually decreased over time in severe and non-severe COVID-19 patients.
Previous studies have identi ed older age as a risk factor of mortality in SARS, MERS, and COVID-19 [10][11][12][43][44][45]. However, in our study, age had no statistic difference between severe and non-severe patients, partly due to epidemiological characteristic in respiratory diseases with chronic airway in ammation, since such patients were commonly old regardless of disease severity. Lymphocytopenia was also associated with poor outcomes in our cohort (85%), which is consistent with other reports [40,46]. Impaired lymphogenesis or increased apoptosis could explain lymphocytopenia in severe cases of COVID-19 [47]. Of note, d-dimer levels greater than 1 µg/L were more common in severe patients compared to non-severe patients, which was reported as a risk factor for mortality of adult inpatients with COVID-19 [10].
Accumulating evidence reveals that cytokine storm plays a crucial role in the pathogenesis of COVID-19.
Our study had some limitations. Firstly, due to the retrospective study design, the accuracy of all laboratory results was dependent upon medical records. Observation bias might also exist in this study due to the limited sample size. Secondly, there could be a selection bias in the multivariate regression model when analyzing the risk factors.

Conclusion
Our study reveals that eosinopenia and elevated LDH on admission are potential predictors of disease severity in adults with COVID-19 and underlying chronic airway diseases. These predictors may help clinicians identify the severe COVID-19 patients with chronic bronchitis, COPD, and asthma.

Study population and data collection
Subjects of this study were adults with COVID-19 and underlying chronic respiratory diseases (admission date from Jan 26 th to April 3 rd , 2020) in Sino-French New City Branch of Tongji hospital. COVID-19 patients were diagnosed according to World Health Organization (WHO) interim guideline [57]. All patients were con rmed by the positive ndings in reverse-transcriptase-polymerase-chain-reaction (RT-PCR) assay of SARS-CoV-2 RNA in throat swab specimens. The study was approved by the Institutional Review Board of Tongji Hospital, Huazhong University of Science and Technology.
The demographic information, clinical characteristics (included medical history, symptoms, comorbidities, smoking history, and allergic history) and radiological results of each patient were obtained from the electronic medical record system of Sino-French New City Branch of Tongji hospital and analyzed by three independent researchers. The severity of COVID-19 was staged according to the guidelines for diagnosis and treatment of COVID-19 published by Chinese National Health Committee (Version 5-7).

Statistical analysis
All data were analyzed with SPSS Statistics Software (version 26; IBM, New York, USA). The statistics for categorical variables were summarized as frequencies and percentages and were compared using Chisquare test or Fisher's exact test between different groups where appropriate. Continuous variables were described using median (IQR) and compared using Mann-Whitney U test. To explore the risk factors associated with disease severity, univariable and multivariable logistic regression models were used to estimate odds ratios and the 95% con dence intervals. A two-sided α of less than 0.05 was considered statistically signi cant.

Declarations
Authors' contributions LY and GZ conceptualized the study design. LY, DC, SZ, YF, WW, CC, SC collected demographic, clinical, and laboratory data. LY, DC, SZ, YF and GZ analyzed the data. LY and DC interpreted the results. LY, DC and GZ wrote the manuscript with all authors providing feedback for revision. All authors read and approved the nal report. U test was used. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Figure 2