Construction of a Predictive Model of Respiratory Endoscopic Intervention in Children with Lobar Pneumonia Caused by Mycoplasma Pneumoniae Infection


 Objectives: This study aimed to analyze the clinical features of children with lobar pneumonia caused by Mycoplasma pneumoniae (MP) infection, to provide a prognostic evaluation of interventional bronchoscopy, and to explore the independent risk factors for bronchoscopic intervention in children with lobar pneumonia caused by MP infection. We also aimed to construct an early-warning model of bronchoscopic intervention, so as to provide an objective evaluation tool for clinicians.Methods: We collected the clinical data from 533 children with lobar pneumonia caused by MP infection. The patients were divided into three groups according to the interventional indications for bronchoscopy and whether they were treated with bronchoscopic intervention, and the clinical features and prognosis of the three groups were compared. Binary logistic regression analysis was performed on the indicators with a significance value of P < 0.05, which we retrieved from the comparative analysis between the first two groups, in order to uncover the independent risk factors and regression equations concerning bronchoscopic intervention. The regression coefficient (β) of our regression model was then used to score related values in the model so as to construct a predictive scoring model of bronchoscopic intervention for the treatment of children with lobar pneumonia caused by MP infection.Results: Children with lobar pneumonia caused by MP infection who demonstrated absolute indications for bronchoscopy exhibited more severe clinical manifestations, and children without absolute indications for bronchoscopy achieved a better prognosis even without bronchoscopic intervention. To establish our early-warning model of bronchoscopic intervention for children with lobar pneumonia caused by MP infection, we utilized the following indices: a C-reactive protein (CRP) ≥ 20.94 mg/L (β1 = 2.253) received 3 points; a fever duration before bronchoscopy ≥ 6.5 d (β2 = 1.424), lactate dehydrogenase (LDH) ≥ 461.5 U/L (β3 = 1.246), or fever (β4 = 1.223), each received 2 points; and if complicated by pleural effusion (β5 = 0.841), 1 point, for a total score of 10 points.


Introduction
MP is the smallest prokaryotic pathogenic microorganism known. It is transmitted through droplets and direct contact and can cause disease in any season of the year (1). The incidence of Mycoplasma pneumoniae pneumonia (MPP) has gradually increased annually, and macrolide antibacterials have been used widely as treatment. However, the incidence of macrolide-resistant Mycoplasma pneumoniae pneumonia (MRMP) is increasing worldwide, severely limiting the treatment options for children with MPP (2, 3).
The chest imaging ndings in MPP vary, and lobar pneumonia caused by MP infection is considered to be one of the clinical features of refractory MPP (4). Respiratory endoscopy equipment and technology have seen progressive improvements in recent years, and endoscopy has gradually become an important treatment modality in treating various respiratory diseases in pediatrics (5).
However, children with lobar pneumonia caused by MP infection manifest clinical symptoms and pulmonary infections that are different from those of adults, and whether they should be actively treated with bronchoscopy requires further investigation. In order to better grasp the indications and contraindications of respiratory endoscopic intervention and to avoid over-or under-treatment, we must acknowledge an urgent need for corresponding early-warning models to standardize the indications of bronchoscopic intervention and to provide clinicians with objective assessment tools.

Patients and Data Collection
We herein selected children with lobar pneumonia due to MP infection as our research subjects, and they were all hospitalized at the Children's Hospital of Soochow University. Our study selection criteria were (1) meeting the diagnosis of MP infection, with serum MP-Immunoglobulin M (IgM) ≥ 1.1 COL and/or nasopharyngeal aspirates (NPAs) and/or BALF MP-DNA, determined by uorescence quantitative PCR, > 10 4 copies/ml; and (2) manifestation of respiratory symptoms with or without fever, consistent with the diagnosis of chest imaging of lobar pneumonia (6)-i.e., a chest radiograph or chest CT examination that revealed uniform consolidation of one or more segments/lobes of the lung. Exclusion criteria included individuals with contraindications to beroptic bronchoscopy; exhibiting bronchopulmonary dysplasia or malformations, repeated respiratory infections, aspiration pneumonia, or exogenous foreign bodies in the bronchus; and those with incomplete medical history data.
The clinical data collected included sex, age, date of hospitalization, pre-hospitalization course and preendoscopic heat course, number of hospitalization days, bronchoscopy time, clinical manifestations and signs, laboratory examinations, and lung imaging examinations.

De nitions 2.2.1 Classi cation Criteria
Our criteria re ected the 'Experts Consensus on Diagnosis and Treatment of Respiratory Endoscopy in Children with Refractory Pneumonia in China' Respiratory Endoscopy Intervention Indications (7) and are as follows. 1. Slow absorption of pneumonia: although the symptoms and signs of the patient improved after treatment with anti-infective drugs, chest X-ray and CT lesion absorption were less than 50% after two weeks. 2. Pneumonia that persisted for a lengthy time period: the course of the disease was prolonged (>2 weeks), and the condition was not alleviated after active treatment but deteriorated and persisted. 3.
Extensive erosion of airway mucosa, necrotic epithelial shedding, mucus plugs blocking the airways, and even bronchial shaping that could be observed under respiratory endoscopy. 4. Imaging showed unilateral emphysema, mediastinal emphysema, disappearance of bronchial in ation signs in unilateral or bilateral lung consolidation, tree bud signs, or other small airway diseases.

Establishment of Groups According to the Aforementioned Standards
For the Bronchoscopy Absolute Indication Group (Group A), children with lobar pneumonia caused by MP infection who met any of the four bronchoscopic intervention indications underwent bronchoscopy and lavage during hospitalization; and for the Bronchoscopy Relative Indication Group (Group B), children with lobar pneumonia caused by MP infection underwent bronchoscopy during their hospitalization on the basis of routine anti-infective and anti-in ammatory, cough and phlegm, and atomized inhalation treatments. Thin or occulent secretions were noted under the bronchoscope, but there were no signs of airway mucosal erosion, necrotic epithelial shedding, mucus plugs blocking the airway, or bronchial shaping. If none of the four indications for bronchoscopic intervention was met, children with lobar pneumonia caused by MP infection were treated with routine anti-infective and anti-in ammatory, cough and phlegm, and atomized inhalation treatments. Bronchoscopic intervention was not implemented during hospitalization (Group C). The three groups comprised 229, 166, and 138 children, respectively.

Prognostic Evaluation Criteria for the Three Groups of Children
Evaluation standards were as follows (8). (1) A normal body temperature: the body temperature was less than 37.4°C within 72 h. (2) The cough-degree classi cations were as follows: no cough recorded as 0 points; mild cough was an occasional or intermittent cough that did not affect the study or quality of the patient and was recorded as 1 point; moderate cough was between mild and severe and was assigned 2 points; severe cough was frequent or paroxysmal coughing day and night that affected sleep, study, or quality of life and was assigned 3 points. (3) The degree of rales in the lungs was assessed as follows. No rales in either lung was recorded as 0 points; a small amount of rales heard in both lungs was recorded as 1 point; rales that were audible and moderate in both lungs were recorded as 2 points; abundant rales were assigned 3 points. (4) The degree of absorption of lung lesions upon chest X-ray or CT was designated as follows: 0 points for complete absorption of lesions; 1 point for lesion absorption ≥ 1/2; 2 points for lesion absorption < 1/2; and 3 points for no absorption or aggravation of lesions.
Prognostic Evaluation Criteria (8): The children were re-examined using chest radiography or chest CT one week after admission to the hospital; and clinical symptoms, signs, and chest radiographic or chest CT changes were then used as the criteria. (1) For a child to be considered 'cured,' her/his body temperature needed to revert to normal, there was no cough or only a mild cough, lung signs disappeared, and chest Xray or CT showed that the lung lesions were basically absorbed. (2) 'Markedly effective' was assigned if the child had a normal body temperature, a moderate or lower degree of coughing, if lung signs disappeared or were only slightly present, and if chest X-ray or CT showed that the absorption of lung lesions was ≥1/2.
(3) An 'effective' rating was delivered if the child's body temperature was normal, the degree of coughing was moderate, the lung signs were alleviated, and the chest radiograph or chest CT indicated that the absorption of lung lesions was less than ½. (4) A rating of 'ineffective' indicated that the child's body temperature did not drop signi cantly, the child showed a degree of moderate or severe coughing, both lungs were full of rales or lung rales worse than before, and the chest X-ray or CT showed that the lung lesions were not absorbed or that they had worsened. (5) The total effective rate = (cured + markedly effective + effective cases)/total number of cases × 100%.

Statistical Analyses
We used SPSS v. 24.0 statistical software for the data analysis. The measurement data that conformed to a normal distribution are expressed as means ± standard deviation. Two independent-sample t tests were used for comparisons between two groups, and one-way ANOVA was used to compare three groups. After the Bonferroni method corrected the signi cance level, a pairwise comparison was performed. The measurement data that did not conform to a normal distribution are presented as the median (25 th percentile, 75 th percentile), and we used the Mann-Whitney U test to compare two groups and the Kruskal-Wallis H test for three groups. The Bonferroni method was then used for comparisons between groups.
Counting data are expressed as percentages (%), with comparisons between groups executed with the χ 2 test and the Fisher exact-probability method depending upon which conditions were met. The differences between groups were statistically signi cant at P < 0.05. We used a binary logistic regression analysis (backward stepwise method) to establish a regression equation model to obtain the probability of bronchoscopic intervention for children with lobar pneumonia and drew a receiver operating characteristic (ROC) curve. Based on the β coe cient in the binary logistic regression model, the factors in the model were weighted to construct a predictive scoring system for children with lobar pneumonia caused by MP infection that required bronchoscopic intervention.

Clinical Characteristics
We enrolled a total of 533 children who met the inclusion criteria. In group A, there were 229 casesincluding 126 males (55.0%) and 103 females (45.0%)-and the male-to-female ratio was 1.22:1, with a median age of 5.92 years. In group B, there were 166 cases-including 82 males (49.4%) and 84 females (50.6%)-and the male-to-female ratio was 0.97:1, with a median age of 5.67 years. In group C, there were 138 cases-including 69 males (50.0%) and 69 females (50.0%)-and the male-to-female ratio was 1:1, with a median age of 5.83 years. There were no signi cant differences in either the average age or sex in the three groups of children (P > 0.05, Table 1). There were also no signi cant differences in the incidence rates of wheezing history, eczema history, allergic rhinitis history, or asthma history among the three groups of children ( Table 1). The ratio of disease duration prior to admission, fever duration before bronchoscopy, and fever peak in group A were all higher than in the latter two groups (P < 0.05, Table 1).
White blood cell count (WBC), neutrophil percentage (N%), C-reactive protein (CRP), lactate dehydrogenase (LDH), alanine aminotransferase (ALT), aspartate aminotransferase (AST), total value of Immunoglobulin A (IgA), complications with pulmonary atelectasis, and complications with pleural effusion in group A were all higher than in the latter two groups (P < 0.05, Table 2). MP-DNA copy numbers (medium-and high-load groups) in group A were higher than in the latter two groups (P < 0.05, Figure 1), while the percentages of CD3 + CD4 + and CD4 + /CD8 + cells in lymphocyte subpopulations were lower than in the latter two groups (P < 0.05, Table 2). The level of prealbumin (PA) in group A was also signi cantly lower than that in group B (after correction, P < 0.05, Table 2).
3.2 Evaluation of the Prognosis for the Three Groups of Children One Week After Admission to the Hospital The median hospitalization time for children in group A was 10 days, which was longer than the 8 days in group B and 7 days in group C (after correction, all P < 0.05), and that of group B was longer than that of group C (after correction, P < 0.05,). The fever subsidence time in group A was longer than in the latter two groups (after correction, all P < 0.05). On the 7th day of hospitalization, the scores for the degrees of coughing, moist pulmonary rales, and lesion absorption in group A were 1 (1, 2), 1 (0, 1), and 2 (2, 3), respectively; the scores in group B were 1 (1, 1), 0 (0, 1), and 1 (1, 2), respectively; and in group C, the scores were 1 (1, 1), 0 (0, 1), and 1 (1, 2), respectively-with the scores in group A being higher than those in groups B and C. The baseline levels for children in groups B and C were the same. We observed no differences in hospitalization time, fever subsidence time, the scores for degrees of coughing, moist pulmonary rales, or lesion absorption one week after hospitalization between the two groups (after correction, all P > 0.05).

Independent Risk Factors for Bronchoscopic Intervention in Children with Lobar Pneumonia Caused by MP Infection
The indicators with a P < 0.05 between groups A and B were analyzed by backward stepwise binary logistic regression analysis, and our results showed that fever duration before bronchoscopy (odds ratio [ Table 3). The abovementioned ve variables were included in the best regression equation, and the regression equation for predictive probability was logit (P) = −7.532 + 0.166X 1 + 0.047X 2 + 0.004X 3 + 1.528X 4 + 1.375X 5 . (Note: X n , the independent variable; X 1 , fever duration before bronchoscopy; X 2 , CRP; X 3 , LDH; X 4 , fever; X 5 , pleural effusion.)

The Predictive Value of Independent Risk Factors in Children with Lobar Pneumonia Caused by MP Infection Who Require Bronchoscopy
The critical values for the independent factors were used to assign the values for each factor. The fever duration before bronchoscopy ≥ 6.5 d, CRP ≥ 20.94 mg/L, LDH ≥ 461.5 U/L, fever, and complication with pleural effusion were the risk factors for bronchoscopic intervention in children with MP-infected lobar pneumonia; and the logistic regression analysis after assigning values to each risk factor was statistically signi cant (P < 0.05, Table 4). The ROC curve was plotted with the predicted probability value of the regression model, and the AUC was 0.860 (95% CI, 0.824-0.897; P=0.000), indicating that the predicted probability model had an upper-middle diagnostic accuracy (Figure 1).
According to the score, 395 children with lobar pneumonia caused by MP infection who also underwent bronchoscopic intervention (groups A and B) were divided into high-risk (7-10 points), medium-risk (4-6 points), and low-risk groups (0-3 points): there were 129 cases in the high-risk group, 146 cases in the medium-risk group, and 120 cases in the low-risk group; and 117 cases (90.71%) in the high-risk group, 92 cases (63.01%) in the medium-risk group, and 20 cases (16.67%) in the low-risk group with absolute indications for bronchoscopy. When the scores for children with lobar pneumonia caused by MP infection reached 6, 7, 8, 9, or 10 points, the possibilities for bronchoscopic intervention were 83.78%, 87.69%, 88.33%, 97.22%, and 100%, respectively ( Figure 2).

Discussion
MP has received increasing attention in recent years as the principal pathogen causing lobar pneumonia in children. When the effect of conventional anti-infective treatment is not favorable, bronchoscopic intervention therapy for lobar pneumonia caused by MP infection attains increased clinical application. Through this study, we found that children with lobar pneumonia caused by MP infection who demonstrated an absolute indication for bronchoscopy exhibited more severe clinical manifestations and that children without absolute indications for bronchoscopy achieved a better prognosis even without bronchoscopic intervention. Therefore, not all children with lobar pneumonia caused by MP infection should receive bronchoscopic intervention as treatment.
In order to avoid over-or under-treatment in clinical practice, the indications for bronchoscopic treatment of lobar pneumonia caused by MP infection need to be rigorously understood. Therefore, we used binary logistic regression analysis to screen out the independent risk factors for lobar pneumonia caused by MP infection that required respiratory endoscopic intervention and then established an early-warning model.
Our results showed that fever duration before bronchoscopy ≥ 6.5 d, CRP ≥ 20.94 mg/L, LDH ≥ 461.5 U/L, fever, and complication with pleural effusion were the risk factors for bronchoscopic intervention in children with MP-infected lobar pneumonia. After MP infects the body, it stimulates a series of heat sources and causes the body to continue to generate fever/high fever. Severe MP infection can cause damage to the mucosal-ciliated columnar epithelium of the airways and can even produce shedding. High secretion of airway mucus then leads to the destruction of mucosal-ciliary system function and impairs airway clearance. Airway secretions provide an excellent culture medium for bacteria and viruses, which can easily lead to mixed infections (9). Mixed infections can then aggravate the systemic in ammatory response, heat-peak height, fever time, and hospital stay in children with lobar pneumonia caused by MP infection (10)(11)(12)(13). CRP is an acute phase, non-speci c reactive protein that is signi cantly augmented during tissue injury or infection, and studies have shown that CRP can be signi cantly elevated when MP infection is combined with lung consolidation and pleural effusion (14). The levels of CRP and LDH are also positively correlated with the severity of MP infection (15)(16)(17), and MP infection is the primary cause of pleural effusion in children (18)(19)(20). After MP infection, speci c antibodies induce an autoantibody reaction, pleural disease, and nally brinous exudates, which often represent a stronger immune response.
Studies have shown that age, duration of fever, CRP, and LDH are independent risk factors for intrabronchial mucus in children with MPP (21); when children with MPP have CRP ≥ 12.27 mg/L and LDH ≥ 462.65 U/L, there is a possibility of the formation of mucus plugs in the bronchus (22). Other studies have shown that CRP > 44 mg/L and LDH > 480 U/L are independent risk factors for children with refractory MPP undergoing multiple bronchoscopic interventions. In one scoring system for refractory MPP in children undergoing multiple bronchoscopic interventions, CRP > 44 mg/L and LDH > 480 U/L were assigned 1 point each (23). In this study, we established an early-warning model of bronchoscopic intervention treatment for children with lobar pneumonia caused by MP infection in which CRP ≥ 20.94 mg/L elicited 3 points, fever duration before bronchoscopy ≥ 6.5 d received 2 points, LDH ≥ 461.5 U/ L received 2 points, fever received 2 points, and complication with pleural effusion received 1 point, for a total score of 10 points. When the score was ≥6, the possibility of bronchoscopic intervention was >80%, and the higher the score, the greater the probability of bronchoscopic intervention. This study possessed some limitations. First, this was a retrospective study, and the selected research subjects may have shown selection bias. Factors such as the long research timespan, the limitations of the test levels, and the geographical limitations also exerted certain in uences on the establishment of the predictive model. Second, the early-warning model established by our research is still only in the initial stage, and it needs to be further tested and improved in the clinic to provide clinicians with a simple, convenient, and valuable scoring system-which should then be bene cial in clinical treatment.
In conclusion, we herein presented an early-warning model of respiratory endoscopic intervention in children with lobar pneumonia caused by Mycoplasma pneumoniae infection and showed that the evaluative effect of the early-warning model was acceptable (with an AUC for the predictive score of 0.859) -with suitable accuracy and clinical practicability.

Declarations
Ethics approval and consent to participate This study protocol was approved by the Ethical Review Committee of Children's Hospital of Soochow University with judgment's reference number 2020CS078. Authors reporting data collection from humans, interviews con rm that all experiments were performed in accordance with relevant guidelines and regulations. Written consent was obtained from all their legal guardian(s) before data collection.A statement to con rm that all methods were carried out in accordance with relevant guidelines and regulations.

Availability of data and materials
All data generated or analysed during this study are included in this published article.

Competing interests
The authors declare that they have no competing interests.  Tables   Table 1 Comparison of clinical characteristics among the three groups of children.  Group C, bronchoscopic intervention was not provided during hospitalization. △, χ 2 value obtained from Pearson test; , H value obtained from Kruskal-Wallis H test; ▲, F value obtained from one-way ANOVA, with post-hoc pairwise comparisons made after the Bonferroni method was used to correct the signi cance level. a, P < 0.05 after correction compared with groups A and B; b, P < 0.05 after correction compared between groups A and C; c, P < 0.05 after correction compared between groups B and C. The copy number was divided into three groups according to the MP-DNA of the child's sputum specimen: low-copy group, MP-DNA copy number ≤ 10 4 /mL; medium-copy group, MP-DNA copy number 10 4 -10 6 /mL; high-copy group, MP-DNA copy number > 10 6 /mL. CRP,C-reactive protein; LDH, Lactate dehydrogenase. Table 4 Logistic regression analysis after the assignment of each independent risk factor.  Table 5 Assignment  Regression model and receiver operating characteristic (ROC) curve of predictive score.

Figure 2
Percentage of children with lobar pneumonia caused by Mycoplasma pneumoniae (MP) infection in each scoring group who required bronchoscopic intervention.