Characteristics of Inflammatory Storm in BALF of Severe Pneumonia in Children: A Case Control Study


 Background: It is difficult to collect bronchoalveolar lavage fluid (BALF) from children. There are few studies on inflammatory storms in BALF of children with pneumonia. Therefore, we will retrospectively investigate characteristics of inflammatory storms in BALF of children with pneumonia in our hospital.Methods: This retrospective study included 161 children with pneumonia. Pediatric pneumonia was divided into two groups: mild and severe pneumonia, according to the clinic diagnosis. All clinic information and data came from our hospital’s database. Blood and BALF samples were collected and measured by Clinical laboratory and Clinical Molecular Medical Center. Flow cytometry was performed to detect the levels of cytokines, including IL-2, IL-4, IL-6, IL-10, IL-17A, TNF-α and IFN-γ. We performed receiver operator characteristic curve (ROC) and spearman correlation to evaluate the role of BALF inflammatory cytokines and cells in pneumonia.Results: 1) The levels of IL-6, IL-10, IL-17A, and TNF-α, and ratios of IL-6 to IL-10 and TNF-α to IL-10 in BALF of severe pneumonia group were higher than those of mild pneumonia group. 2) The number of inflammatory cells and the percentage of neutrophils in BALF of severe pneumonia group increased, and the percentage of macrophages decreased, compared with the mild pneumonia group. 3) ROC analysis showed that the levels of IL-6, IL-10 and TNF-α, and ratios of IL-6 to IL-10 and TNF-α to IL-10, the number of inflammatory cells and neutrophils, as well as the percentage of neutrophils exceeded the cut-off value (60.37pg/mL, 2.54pg/mL, 13.07pg/mL, 29.30, 1.57, 3140.00*109/L, 768.00*109/L, 43.00%, respectively) could be used to distinguish children with severe pneumonia from children with mild pneumonia. While the percentage of macrophages in BALF decreased to below the cut-off value (36.00%) also had the discrimination ability. 4) The BALF levels of IL-6, IL-10, TNF-α and IFN-γ were correlated, positively with the inflammatory cell count and the percentage of neutrophils, and negatively with the percentage of macrophages.Conclusion: In severe pediatric pneumonia, BALF inflammatory cytokines and inflammatory cell infiltration are more intense, especially above the cut-off value. This study demonstrates that these inflammatory signals are potential biomarkers for predicting the severity of pediatric pneumonia.Trial registration: This study approved by the Medical Research Ethics Committee of Children’s Hospital of Chongqing Medical University, registered in http://www.chictr.org.cn/, No. ChiCTR2000034048(registration date, June 22, 2020).

signals (mainly IL-6 and IL-10) help to identify the pathogen [8,9], evaluate the severity and prognosis of pneumonia [10][11][12]. Tocilizumab, an IL-6 receptor antagonist, is effective in treating patients with severe 2019 coronavirus disease (COVID-19) whose serum IL-6 is signi cantly elevated [13]. Previous researchers have found that in patients with pneumonia, certain immune function genes are up-regulated in lung aspirate cells and down-regulated in blood mononuclear cells [14], and the serum cytokines cannot fully and truly re ect the local immune responses [15]. The affected lung may be the source of cytokines [16].
Studies have pointed out that the levels of cytokines in bronchoalveolar lavage uid (BALF), such as IL-17A, are closely related to the severity of childhood mycoplasma pneumoniae (MP) pneumonia [17].
Therefore, we hypothesized that the cytokines and in ammatory cells in the lungs of children with severe pneumonia may have speci c changes, and local anti-in ammatory therapy and anti-cytokine receptor therapy may be suitable for children with severe pneumonia.
This study retrospectively analyzed the levels of cytokines (IL-2, IL-4, IL-6, IL-10, IL-17A, TNF-α, IFN-γ) and in ammatory cells in BALF in children with pneumonia. We evaluated the identi cation ability of cytokine levels and in ammatory cell levels in BALF for children with severe pneumonia. We also analyzed the relationship between cytokines and in ammatory cells in BALF in children with pneumonia. The results of the study will help us better understand the in ammatory storm in the lungs, and help us to diagnose and treat pediatric patients with severe pneumonia early.

Patients
Children with pneumonia, who met the inclusion criteria and were hospitalized at the Children's Hospital of Chongqing Medical University from March 2019 to February 2020, were included. We collected the clinical and laboratory data of 161 children with pneumonia from our hospital's database. All children underwent bronchoalveolar lavage during hospitalization. Children were divided into severe and mild pneumonia groups.
This study approved by the Medical Research Ethics Committee of Children's Hospital of Chongqing Medical University, registered in http://www.chictr.org.cn/, No. ChiCTR2000034048(registration date, June 22, 2020).

De nition of pathogenic microorganisms
The de nition of pathogenic microorganisms was as follows[18-21]: 1) the pathogenic bacteria were cultured in pleural effusion or BALF (≥ +), 2) the nasopharyngeal aspirates (NPAs) obtained through arti cial airway cultured dominant bacteria (≥ ++), 3) the NPAs culture had the growth of dominant bacteria (≥ +++), 4) the bacterial nucleic acid test of NPAs or BALF were positive, 5) the detection of virus antigen or virus PCR of NPAs, BALF or pleural effusion was positive, or the serum virus IgM-speci c antibody was elevated (≥ 4 folds), 6) the serum MP IgM-speci c antibody was positive (≥ 1/160), the PCR of MP in NPAs, BALF or pleural effusion was positive, 7) the serum chlamydia pneumoniae (CP) IgMspeci c antibody was positive (≥ 1/40), the PCR of CP in NPAs, BALF or pleural effusion was positive.

Statistics
Statistical software SPSS version 25.0, MedCalc version 19.1.3 and GraphPad Prism version 8.0.1 were used for analyzing data. Data exceeded the LODL were represented as 1/2 of LODL for statistical processing. Non-normal data were expressed as median (interquartile range) and normal data were expressed as mean ± standard deviation. Chi-square test or Fisher's exact test was performed for classi ed variables, and Mann-Whitney U test was performed for non-parametric variables. The correlation between biomarkers was performed by Spearman correlation coe cient (r). The accuracy of biomarkers in identifying severe pneumonia was analyzed by receiver operator characteristic curve (ROC). The area under curve (AUC), represented the identi cation ability, and optimal cutoff values were calculated. And the higher the AUC, the higher the identi cation ability of biomarkers. The difference was statistically signi cant with double-tailed P-value < 0.05.

Patients Characteristics
This study retrospectively analyzed the data of 161 patients with pneumonia including 100 cases in the severe pneumonia group and 61 cases in the mild pneumonia group (Table 1). There were no signi cant differences in age, gender and weight between the two groups. The days of fever and hospitalization in the severe group were signi cantly longer than those in the mild group.
The red blood cell count (RBC) and hemoglobin level of the severe group decreased signi cantly, and the white blood cell count (WBC), platelet count and neutrophil percentage were not signi cantly different between the two groups. There was no signi cant difference in CRP levels between the two groups, while PCT levels of the severe group were signi cantly increased.
And in the severe group, the levels of alanine aminotransferase (ALT) and lactate dehydrogenase (LDH) increased signi cantly, while the levels of albumin were decreased signi cantly. The levels of brinogen and D-dimer in the severe group were also increased signi cantly.
Among children with severe pneumonia, 51 cases progressed to respiratory failure, 1 case developed acute respiratory distress syndrome (ARDS), 24 cases had pleural effusion, and 28 cases had pulmonary necrosis. 34 cases were treated with non-invasive continuous positive airway pressure and 1 case was treated with invasive mechanical ventilation (endotracheal intubation).
The microorganisms of children with pneumonia were shown in Table 2. Among the 161 children with pneumonia, the detection rate of MP was the highest; the top three detection rates were Streptococcus pneumoniae, Haemophilus in uenzae and Staphylococcus aureus; the virus detected rate adenovirus ranked rst, followed by respiratory syncytial virus. The pathogen detection rate of severe pneumonia group was higher than that of mild pneumonia group (100% vs. 85.24%). The pathogen detection of bacteria, virus, MP and CP showed no statistical difference between mild and severe pneumonia groups (P>0.05, Table 2). In the mild pneumonia group, at most two kinds of pathogenic microorganisms were detected simultaneously in mixed infection patients. In the severe pneumonia group, at most three kinds of pathogenic microorganisms were detected simultaneously in mixed infected patients.

3.3
The ratios of pro-in ammatory to anti-in ammatory cytokine in BALF of severe group are higher than that of mild group It is important to maintain a balance between pro-in ammatory and anti-in ammatory cytokines, so we calculated the ratio of proin ammatory cytokines to anti-in ammatory cytokines in BALF of children with pneumonia. In our study, the IL-6/IL-10 ratio (34.48, 13.82~66.67) and TNF-α/IL-10 ratio (0.91, 1.16~2.64) in BALF of severe group were signi cantly higher than those of mild group (Fig.1b). There was no signi cant difference in the IL-17A/IL-10 ratio between the two groups (P>0.05, Fig.1b).

Discussion
At present, pneumonia is still one of the reasons that threaten the health of children in the world. The clinical manifestations of children with pneumonia are very complicated due to the infection by a variety of pathogenic microorganisms. As we all known, in ammation is not only related to the invasion of pathogenic microorganisms, but also related to the body's in ammatory immune response [22]. We should look for a detection or treatment method for the in ammatory immune response of pneumonia, so that we can quickly identify the critically ill children and take more effective treatments. The intense in ammatory storm in patient with severe pneumonia has aroused widespread concern. And our study further demonstrates that the cytokines and in ammatory cells in BALF are important indicators for re ecting the severity of pneumonia.
Our study shown that compared with children with mild pneumonia, children with severe pneumonia had longer days of fever and hospitalization. The signi cant decrease of RBC and hemoglobin level in children with severe pneumonia indicated a decrease of oxygen carrying capacity, which was related to more severe hypoxia symptoms. The in ammatory markers (PCT, LDH) in children with severe pneumonia also increased signi cantly. And the abnormal liver function and blood coagulation function in children with severe pneumonia suggested that severe patients had damage external lung organ.
Children with severe pneumonia had abnormal in ammatory immune responses. In our study, the levels of IL-6, IL-10, IL-17A, TNF-α and the number of in ammatory cells (including neutrophils, macrophages and lymphocytes) in BALF of children with severe pneumonia were signi cantly higher than those of children with mild pneumonia. These results were similar to previous study [23,24]. Interestingly, compared with children with mild pneumonia, the percentage of macrophages in BALF of children with severe pneumonia was decreased relatively, while the percentage of neutrophils increased signi cantly. It suggests that undue in ltration of neutrophils is an important mechanism of lung in ammation in children with severe pneumonia.
The levels of IL-6 are associated with the severity of infection and in ammation [25,26],as well as the severity of adult pneumonia [27]. In this study, compared with children with mild pneumonia, children with severe pneumonia had more severe in ammatory cell in ltration in lungs. Immune cells, such as activated alveolar macrophages and lymphocytes, and damaged endothelial cells, could secrete IL-6 and chemokines that elicit neutrophils migration into the lungs [28,29]. Early recruitment of neutrophils was the key to initiating effective host defenses [30]. IL-6 level in BALF of children with pneumonia was positively correlated with the percentage of neutrophils. On the one hand, IL-6 could promote the survival of neutrophils [31], enhance the bactericidal activity of neutrophils, and expand the in ammatory effect to eliminate pathogens [32]. However, undue release of IL-6 caused excessive in ammatory damage, which in turn leads to lung damage and aggravation of the disease [33].
IL-10 is a cytokine with anti-in ammatory effect [34]. IL-10 could inhibit the effect of activated alveolar macrophages on T helper cells, reduce the release of pro-in ammatory cytokines, and relieve the body's in ammatory response [34]. There was a signi cant positive correlation between IL-6 level and IL-10 level, which may be due to that IL-6 itself can induce IL-10 expression. In this study, IL-10 level in BALF of children with severe pneumonia were signi cantly increased, suggesting that IL-10 is also one of the indicators for evaluating the severity of pediatric pneumonia. Studies had shown that IL-10 could inhibit the overproduction of neutrophils [35]. However, in our study, IL-10 level in BALF of children with pneumonia was positively correlated with the percentage of neutrophils. It might be because IL-6 promotes the expression of IL-10 and also promotes the aggregation of neutrophils. So, the level of IL-10, the number and the percentage of neutrophils in BALF can re ect the intensity of the body's in ammation. Our results also showed that IL-10 level in BALF was negatively correlated with the percentage of macrophages, which might be related to the inhibition of macrophages by IL-10 [34].
The level of TNF-α was also positively correlated with the level of IL-6 and IL-10, and the percentage of neutrophils. TNF-α seems to have the same effect as IL-6 in amplifying the in ammatory response. TNFα, secreted by activated alveolar macrophages, is also the major factor affecting the migration of neutrophils to in ammatory sites. Under the synergistic effect of IL-17A and TNF-α, neutrophils migrated rapidly to the in ammation site and persist [36]. Blocking TNF-α could signi cantly down-regulate in ammatory [37]. Like IL-6 and IL-10, TNF-α is also an indicator of the severity of in ammation in the body. TNF-α was a promoter of lung pathological damage [38]. TNF-α could induce macrophage and neutrophils to release chemokines [39]. And TNF-α could stimulate the release of platelet activating factors to cause vasodilation, and it could promote the adhesion and migration of leukocytes to in ammatory sites [32]. It could lead to more severe in ammatory damage in lung tissue. Therefore, the level of TNF-α may also indicate the severity of pediatric pneumonia.
We speculate that the lungs of children with severe pneumonia are relatively insu cient in antiin ammatory cytokines, which is well demonstrated in our study by the levels of anti-in ammatory cytokines. The ratios of IL-6 to IL-10 and TNF-α to IL-10 in BALF of children with severe pneumonia were signi cantly elevated, indicating the excessive release of pro-in ammatory cytokines and the relative insu ciency of anti-in ammatory cytokines in the lungs. The imbalance of pro-in ammatory and antiin ammatory effects exacerbated lung damage. Other studies supported this nding [11,40].
In addition, in our study, an increase in the number of in ammatory cells in BALF was also found in children with severe pneumonia. Pro-in ammatory cytokines could help extend the in ammatory response by increasing the total number of in ammatory cells, especially neutrophils [41]. Compared with children with mild pneumonia, the percentage of macrophages in BALF in children with severe pneumonia decreased signi cantly. This decline was relative, because the number of macrophages in children with severe pneumonia was also increased signi cantly. While, a decrease in the activity of macrophages was also found in bacteria-infected mice, and this change was closely related to abnormal cytokine signals [42]. Microbial infection could change the function of macrophages and impair their phagocytic and killing ability [43]. In our study, the undue cytokines might inhibit the proliferation and activation of macrophages to avoid more harmful in ammation. Because the levels of cytokines in BALF were negatively correlated with the percentage of macrophages. Both undue accumulation of neutrophils and relative insu ciency of macrophages were associated to lung injury. Unfortunately, due to the lack of data of macrophage activity in BALF in this study, it is not clear whether the decrease in the percentage of macrophages is accompanied by a decrease in macrophage activity.
ROCs of cytokines and in ammatory cells in BALF were used to distinguish severe pneumonia from mild pneumonia. We found that in BALF the levels of IL-6, IL-10 and TNF-α, the ratio of IL-6 to IL-10, the ratio of TNF-α to IL-10, the number of in ammatory cells and neutrophils, the percentage of neutrophils, increased above the threshold could distinguish severe pediatric pneumonia. And the percentage of macrophage decreased exceeding the threshold could also distinguish pediatric severe pneumonia. As a result, the signals of in ammatory cytokines and cells in BALF are important biomarkers for re ecting the severity of pediatric pneumonia, helping to identify critically ill children more accurately and choose individualized treatment regimens.
And, suppressing of in ammatory storm and regulating of cytokine imbalance should be one of the new therapeutic targets. If the condition of children with pneumonia permits, bronchoalveolar lavage is a promising treatment option to remove and reduce undue cytokines and in ammatory cells for alleviating lung damage. For severe pediatric patients, appropriate anti-in ammatory treatments (including glucocorticoids and immunoglobulins) should be used as an option. Meduri,G.U and colleagues' study [44] showed that long-term use of glucocorticoids could reduce systemic in ammatory response in adults patients with ARDS. Local anti-in ammatory treatments are also worth considering. And inhaled hormones may be one of the important anti-in ammatory treatments, because aerosol inhalation of budesonide could effectively prevent the occurrence of pneumonia after thoracotomy in adults [45]. Of course, the use of glucocorticoids should be more personalized and discreet to achieve greater bene ts for children with pneumonia. Other studies reported that intravenous IL-6 receptor antagonists are bene cial for COVID -19 patients[46, 47], so local anti-cytokine receptor therapy may be applicable for critical pediatric pneumonia.
Our data demonstrated that the cytokines (IL-2, IL-4, IL-6, IL-10, IL-17A, TNF-α, IFN-γ) in BALF of children with pneumonia were positively pairwise correlated. It was also supported by previous studies[48 -50], showing a cascade of cytokine responses. Moreover, the levels of cytokines (IL-6, IL-10, TNF-α, IFN-γ) in BALF in children with pneumonia were positively correlated with the number of in ammatory cells and the percentage of neutrophils. This indicated that the pro-in ammatory effect of cytokines was associated with the increase of the total number of in ammatory cells in the lungs, especially the accumulation of neutrophils. Previous studies indicated that excessive accumulation and continuous activation of neutrophils could lead to greater cytokine storms and harmful in ammatory responses [22,41].
There are some limitations in this study. Previous studies [8,9] found that cytokine pro les also played a role in the identifying of pathogens. However, due to individual differences and few children with pneumonia with pathogen mono-infection in our study, no speci c pro les of pathogen-related cytokines were found. In addition, because of the ethical restrictions on collecting BALF from healthy children, and it is di cult to determine that relatively healthy children who underwent bronchoscopy and collected BALF had no lung in ammation at all, this study did not set up a healthy child control group.

Conclusion
In summary, our data shows that the signals of cytokines and in ammatory cells in BALF have a unique role in evaluating the severity of in ammatory and pneumonia in children. The undue release of cytokines and in ammatory cells in BALF of children with severe pneumonia are more serious than that of children with mild pneumonia. And both undue accumulation of neutrophils and insu ciency of macrophage are associated with lung injury. ROC analysis showed that the levels of IL-6, IL-10 and TNF-α, the ratios of IL-6 to IL-10 or TNF-α to IL-10, the number of in ammatory cells and neutrophils, the percentage of neutrophils increased over the threshold could be used to distinguish children with severe pneumonia from children with mild pneumonia. Moreover, the decrease in the percentage of macrophages in BALF below the threshold also has identi cation ability on severe pediatric pneumonia. Consequently, the signals of cytokines and in ammatory cells in BALF are potentially important biomarkers for re ecting the severity of pneumonia. And suppressing of in ammatory storms and regulating of cytokine imbalance (such as, anti-in ammatory therapy, anti-cytokine receptor therapy) may be new treatment chooses for children with pneumonia.

Declarations Con ict of interest
The authors declare no con ict of interest.

Authors' contributions
Chang Shu participated in the study design and article revision. Qu Bei Li, Yang Yang, Zhi Hua Zhao, Jun Jie Tan and Guo Fu collected and analyzed the data, Fang Deng wrote this manuscript. Xiao Hua Liang and Hui Ling Cao provided technical support. All authors read and approved the nal manuscript.