Comparison of severe and general Mycoplasma pneumoniae pneumonia in children: a targeted next-generation sequencing based study

Abstract

Background

The incidence and severity of Mycoplasma pneumoniae pneumonia (MPP) in children have increased after the pandemic of COVID-19, raising public concern. However, factors that affect the severity of MPP have not been well described. This study aimed to investigate the influence of Mycoplasma pneumoniae characters, respiratory co-existing pathogens, and host response on the severity of MPP.

Methods

Clinical characteristics of 288 children hospitalized for MPP between November 2023 and July 2024 were analyzed retrospectively. Patients were divided into severe MPP (SMPP) and general MPP (GMPP) groups according to disease severity. Targeted next-generation sequencing (tNGS) was employed to analyze the respiratory pathogens of patients.

Results

Of 288 cases, there were 113 SMPP and 175 GMPP. Compared with GMPP group, children with SMPP had significantly higher levels of neutrophil percentage, Neutrophil to Lymphocyte Ratio (NLR), C-Reactive Protein (CRP), Alanine aminotransferase (ALT), Lactic dehydrogenase (LDH) and D-D dimer (all P < 0.05), which reflected a higher host immune response in SMPP group. Meanwhile, tNGS based pathogen analysis showed that Mycoplasma pneumoniae characters (A2063G mutation or not, pathogen concentration) were not associated with the severity of MPP (P > 0.05). Furthermore, co-existing pathogens analysis showed that the SMPP group had a lower number of co-detected pathogens, children with human respiratory syncytial virus (RSV) co-existing were more frequent in the SMPP group (all P < 0.05). Multivariate analysis showed that lower number of co-detected pathogens (decreased microbial diversity), RSV co-existing, elevated CRP, and LDH were independent risk factors for SMPP.

Conclusion

Collectively, our data showed that respiratory microbiome and host response may play important roles in the pathology of MPP, decreased microbial diversity, RSV co-existing, CRP, and LDH level can predict the severity of MPP.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-026-12914-6.

Background

Mycoplasma pneumoniae pneumonia (MPP) is a common community-acquired pneumonia in children [1]. Although most MPP children recover well, shown as general MPP (GMPP), some individuals may experience severe clinical manifestations, with local or systemic complications, which can be defined as severe MPP (SMPP) [2, 3]. After the COVID-19 pandemic, the incidence and severity of MPP increased in China and Europe, raising public concern about SMPP [4–6].

“Immunity debt” and immunological impairment after COVID-19 were thought to be reasons for the elevated incidence of SMPP [6, 7]. Most studies about SMPP also focused on the correlations of immune response and severity of MPP [8–12]. However, as an infectious disease, besides the host immune system, pathogen characteristics and respiratory microbiome may also influence the severity of the disease. Comprehensive analysis of pathogen-environment-host factors may give us a deeper understanding of the SMPP pathogenesis.

Mycoplasma pneumoniae (MP) attaches to human cells, releases community-acquired respiratory distress syndrome (CARDS) toxin, and causes cell damage directly or indirectly [13, 14]. Limited clinical studies have demonstrated that different strains of Mycoplasma pneumoniae had different CARDS expression or different macrolide resistance genes, which might be associated with the severity of MPP [15–17]. However, this conclusion was inconsistent in some other studies [18, 19], which calls for more research to know for sure.

The respiratory microbiome has important implications for pulmonary disease severity [20]. Different methods can be used for the study of respiratory microbiome in MPP. Traditional microorganism culture was proved to be laborious [21]. The 16 S rRNA gene sequencing method has been used in some studies to reveal the complex respiratory microbiome of MPP recently [22–24], however, its low resolution (only genus-level) restricts our understanding of the respiratory microbiome of MPP and some conclusions were also controversial in these studies. Metagenomic next-generation sequencing (mNGS) had a higher resolution but its high cost and high background noise also restricted its application. Because of its high efficiency and resolution, targeted next-generation sequencing (tNGS) was proved to be a promising technology in MPP respiratory microbiome analysis [25–27]. Here we used tNGS to analyze the pathogen characters and respiratory co-existing pathogens of MPP, combined with host immune response indicators, three core elements of MPP (pathogen-environment-host factors) were integrated and correlated with the severity of the disease.

This integration analysis may fill gap in our comprehensive understanding about factors associated with the severity of MPP and provide more efficient strategies for disease stratification and treatment.

Materials and methods

Study participants and data collection

This retrospective study was conducted at Shanghai Tongji Hospital, a tertiary general hospital in Shanghai, China from November 2023 to July 2024. Demographics, clinical data, and laboratory test results of patients aged less than 18 years old were retrieved from the inpatient electronic records and analyzed retrospectively. When more than one laboratory test was performed during a single hospitalization, only the results of the first sample collected at admission were included in the analysis. This study was conducted in accordance with the Declaration of Helsinki and approved by the ethics committee of Tongji Hospital, individual consent for this retrospective study was waived.

MPP diagnoses and disease severity

According to the 2023 edition of Guidelines for the Diagnosis and Treatment of Mycoplasma pneumoniae Pneumonia in Children, recommended by National Health Commission of the People’s Republic of China, patients were diagnosed with MPP if they met all the following criteria: (1) had symptom and signs of acute respiratory tract infections, with chest radiography confirmed pneumonia; (2) single serum MP antibody titer > 1:160 or ≥ 4-fold seroconversion of MP antibody titer; (3) positive detection of MP-DNA or MP-RNA.

Children were excluded if they: (1) had pre-existing congenital respiratory disease such as congenital bronchopulmonary abnormalities, or bronchiectasis; (2) had immune deficiencies, chronic diseases, heart diseases, or were using immunosuppressive drugs.

SMPP was defined as the presence of one or more of the following criteria [2, 3]: (1) poor general conditions; (2) high fever (> 39 °C) for more than 5 days or fever for more than 7 days without a declining trend in peak temperature; (3) increased breathing rate, wheezing, dyspnea, chest pain or hemoptysis, which might be signs of plastic bronchitis, necrotizing pneumonia, pleural effusion, pulmonary embolism, or acute asthma exacerbations; (4) extrapulmonary complications. (5) hypoxemia (SaO2 < 93% on room air); (6) imaging finding: multilobed, ≥ 2/3 infiltration of the lung or diffuse pan-bronchiolitis; (7) imaging finding: pneumonia lesions progressed for more than 50% with 1–2 days.

All the MPP children who didn’t meet the criteria for SMPP were considered to have GMPP.

Sample collection and targeted next-generation sequencing

Deep sputum from the respiratory tract of patients was collected in sterile containers with secure lids. After gargle, patients were instructed to expel sputum from the respiratory tract by forceful cough, making an effort to avoid contamination with saliva. For infants or young children unable to follow the instructions, negative pressure sputum suction was applied to collect sputum from the respiratory tract. Samples were collected and preserved at − 20 °C within 24 h for tNGS analysis.

DNA/RNA from the samples was extracted and purified using MagPure Pathogen DNA / RNA Kit (R6672-01B, Magen, Guangzhou, China) following the manufacturer’s instructions. Multiplex PCR library system (RP100TM, KingCreate, Guangzhou, China) and the next-generation sequencing platform KM MiniSeq Dx-CN (KingCreate, Guangzhou, Guangdong, China) were applied to detect 198 respiratory pathogens and more than 300 drug-resistant gene sites. The methods and workflows of tNGS have been described in detail previously [25].

The normalized read counts of detected microorganisms were compared with the read counts of known concentration internal reference in the same reaction system, and then the semiquantitative concentrations of detected microorganisms were calculated.

Statistical analysis

Normality test of quantitative data was applied before analysis. Quantitative normal or nearly normal data are expressed as the mean ± standard deviation. Quantitative skewness data are presented as the median (percentile: P25, P75), and categorical data are expressed as the frequency. T student test was used for comparison of quantitative normal data and Mann–Whitney U test was used for quantitative skewness data. The chi-square test or Fisher’s exact was used for unordered categorical data. For ordered categorical variables, the nonparametric test was used for comparison. Univariate analysis was performed to identify the differences between SMPP and GMPP. Logistic regression analysis was performed to identify risk factors associated with SMPP. Statistical significance was set as P < 0.05. Linearity of continuous independent variables were assessed using the BoxTidwell procedure. All data analyses were performed using SPSS (version 26.0; SPSS Inc., Chicago, IL, USA). The microbial classification chart was plotted by https://www.bioinformatics.com.cn (last accessed on 1 Nov 2024), an online platform for data analysis and visualization.

Results

Patient characteristics

Between November 2023 and July 2024, a total of 288 MPP patients were included in this study, including 128 males and 160 females. The median age was 8.0 years, ranging from 2 to 15 years old. The SMPP rate was 39.2% (113/288).

All patients were previously healthy. The difference of MP severity between preschool-aged children ( < = 6 years) and school-aged children (> 6 years) was not statistically significant (P = 0.43), shown in Table 1. Meanwhile, most of the co-detected pathogen frequencies between preschool-aged children and school-aged children were not statically different, such as Streptococcus pneumoniae, Haemophilus influenzae and Human respiratory syncytial virus (Supplement Table 1).

Table 1.

Clinical characteristics of patients

Characteristics	Total (N = 288)	SMPP (N = 113)	GMPP (N = 175)	P
Gender, n (%)
Male	128 (44.4)	48 (42.5)	80 (45.7)	0.59
Female	160 (55.6)	65 (57.5)	95 (54.3)
Age, years	8.0 (7.0–11.0)	8.0 (7.0–11.0)	8.0 (7.0-10.5)	0.22
<= 6years, n (%)	63 (21.9)	22 (19.5)	41 (23.4)	0.43
> 6 years, n (%)	225 (78.1)	91 (80.5)	134 (76.6)
Hemoglobin, g/L*	128.6 ± 9.9	127.8 ± 10.5	129.1 ± 9.5	0.28
Platelets, ×109/L*	250.6 ± 74.3	263.2 ± 70.1	249.0 ± 77.1	0.64
WBC, ×109/L	6.8 (5.6–8.5)	6.9 (5.7–8.5)	6.8 (5.6–8.6)	0.66
Neutrophil, %	64.0 (56.5–71.3)	66.0 (58.0–73.0)	63.0 (54.7–70.5)	0.01
Lymphocyte, %	26.0 (20.0–32.0)	25.0 (19.0–30.0)	27.0 (21.0-33.5)	0.02
NLR	2.5 (1.8–3.6)	2.7 (2.0–4.0)	2.3 (1.6–3.4)	0.02
CRP, mg/L	10.1 (5.9–19.8)	13.0 (7.3–26.0)	9.0 (4.7–17.0)	0.00
PCT, ng/mL	0.1 (0.0-0.1)	0.1 (0.0-0.1)	0.1 (0.0-0.1)	0.80
ALT, u/L	16.0 (14.0–20.0)	17.0 (14.0–22.0)	16.0 (13.0-19.5)	0.03
AST, u/L	36.0 (21.0-101.0)	36.0 (32.0–41.0)	36.0 (32.0–41.0)	0.88
LDH, u/L	280.0 (248.0-316.0)	287.0 (260.0-336.0)	272.0 (238.0-307.5)	0.00
CK-MB, u/L	1.1 (0.8–1.6)	1.0 (0.8–1.5)	1.1 (0.8–1.6)	0.98
Lactic acid, mmol/L	1.6 (1.2–1.8)	1.6 (1.3–1.7)	1.5 (1.2–1.8)	0.28
D-D dimer, mg/L	0.4 (0.3–0.6)	0.5 (0.3–0.8)	0.35 (0.3–0.5)	0.00
IL-6, pg/mL	14.0 (7.0–43.0)	14.0 (7.0–31.0)	15.0 (7.0-51.5)	0.22
IL2R, pg/mL	745.5 (576.0-933.0)	746.0 (614.0-978.0)	745.0 (548.5-895.5)	0.15
TNF, pg/mL	15.8 (12.9–19.2)	15.0 (12.5–19.0)	16.0 (13.1–20.0)	0.19

WBC-White Blood Cells, NLR-Neutrophil to Lymphocyte Ratio, CRP-C Response Protein, ESR-Erythrocyte Sedimentation Rate, PCT-Procalcitonin, ALT-Alanine aminotransferase, AST-Aspartate transaminase, LDH-Lactic dehydrogenase, CKMB-Creatine kinase isoenzyme MB, IL6-Interleukin-6,IL2R-Interleukin-2 receptor, TNF-Tumor Necrosis Factor, M. pneumoniae-Mycoplasma pneumoniae

* only Hemoglobin and Platelets meet the normality test and presented with mean ± SD, other Quantitative skewness data are presented as the median (percentile: P25, P75)

Comparison of host response between SMPP and GMPP

As shown in Table 1, compared with the GMPP group, the SMPP group had a higher percentage of neutrophils and a lower percentage of lymphocytes in the peripheral whole blood cell test, hence leading to a higher Neutrophil to Lymphocyte Ratio (NLR) in the SMPP group (P = 0.02).

The serum levels of C-Reactive Protein (CRP) were higher in the SMPP group than in the GMPP group (P = 0.02), but no significant difference was found in Procalcitonin (PCT) levels between the two groups (P = 0.80).

In addition, biochemical examination indicates the SMPP group had a higher level of Alanine aminotransferase (ALT), Lactic dehydrogenase (LDH), and D-D dimer than the GMPP group (P = 0.03, 0.00, 0.00).

However, no significant difference was found in serum cytokine Interleukin-6 (IL-6), Interleukin-2 receptor (IL-2R), and Tumor Necrosis Factor (TNF) (all P > 0.05) between the two groups.

Comparison of MP characters between SMPP and GMPP

For macrolide-resistant genes, only A2063G was found and no other mutations were found in all samples. The mutation ratios were high in both groups, but the macrolide-resistant gene mutation was not associated with disease severity (P = 0.54), Fig. 1A.

Fig. 1.

Comparison of MP characters between SMPP and GMPP group. (A) The percentage comparison of macrolide-resistant gene (A2063G mutation) in SMPP and GMPP group. (B) The comparison of MP normalized read counts between SMPP and GMPP groups. The semiquantitative concentrations of MP were compared in the underlying chart

For MP DNA concentration, no significant difference was found in normalized read counts between SMPP and GMPP groups (P = 0.30). Meanwhile, the difference in MPP semiquantitative concentrations between the two groups was also insignificant (P = 0.80), Fig. 1B.

Comparison of respiratory co-existing pathogens between SMPP and GMPP

Co-existing pathogens analysis showed that single MP infection was more frequent in the SMPP group (P = 0.03), and the SMPP group had a lower number of co-detected pathogens than the GMPP group (P = 0.03), Fig. 2A.

Fig. 2.

Comparison of respiratory co-existing pathogens between SMPP and GMPP group. (A) Cases of different numbers of co-existing pathogens between the SMPP and GMPP groups. (B) Comparison of co-detected pathogen percentage at kingdom level. (C) Comparison of co-detected pathogen percentage at the genus level. (D) Comparison of co-detected pathogen percentage at the species level

Further analysis showed that at the kingdom level, the SMPP group had a slightly higher percentage of fungus co-detection and a lower percentage of bacteria co-detection, but the differences between the two groups were insignificant (P = 0.63), Fig. 2B and Supplement Table 2.

Analysis at the genus level showed that the SMPP group had a slightly higher percentage of Haemophilus, Fusobacterium, Orthopneumovirus, and Candida co-detection, lower percentage of Streptococcus, Staphylococcus and Enterovirus co-detection, but only the difference in percentage of Orthopneumovirus was significant between two groups (P = 0.02), Fig. 2C and Supplement Table 3.

Analysis at species level showed that the SMPP group had a slightly higher percentage of Haemophilus influenzae, Fusobacterium nucleatum, Human respiratory syncytial virus (RSV) and Candida tropicalis co-detection, lower percentage of Staphylococcus aureus, Streptococcus pneumoniae, Rhinovirus and Candida parapsilosis co-detection, but only the difference in percentage of Human RSV was significant between two groups (P = 0.02), Fig. 2D and Supplement Table 4.

Multivariate logistic regression analysis of factors associated with SMPP

Integrated the environment-host factors identified in the univariate analysis above, after excluding possible co-linearity and testing the linearity of continuous independent variables, NLR, CRP, ALT, LDH, Numbers of co-detected pathogens, RSV co-infection were included for further analysis. Multivariate logistic regression analysis showed that lower number of co-detected pathogens, RSV co-existing, elevated CRP and LDH were independent risk factors for SMPP, shown in Table 2.

Table 2.

Multivariate logistic regression analysis for factors associated with SMPP

Factors	Partial regression coefficient	SE	Wald χ2	OR	95% CI		P
					Lower	Upper
Numbers of co-detected pathogens	-0.201	0.082	5.971	0.82	0.70	0.96	0.02
RSV co-infection	2.449	0.892	7.536	11.57	2.02	66.49	0.01
CRP	0.015	0.007	4.635	1.02	1.00	1.03	0.03
LDH	0.006	0.002	6.442	1.01	1.00	1.01	0.01

Discussion

Although MPP was often thought to be a mild and self-limited disease, an increased number of severe cases were reported after the pandemic of COVID-19, raising public concern [4–7]. Factors that influence the severity of MPP were still not fully described, hindering disease stratification and treatment. Here, by using tNGS, a new respiratory pathogen analysis approach, we analyzed the pathogen-environment-host factors associated with disease severity comprehensively, hoping to characterize risk factors for SMPP.

Extrapulmonary complications and response to corticosteroid treatment implied that MPP might be an immune-mediated disease [28–30]. Neutrophils, Lymphocytes, CRP, and PCT are key members of our immune system and play crucial roles in the pathology of MPP [31–33]. Consistent with previous studies [8, 10, 34], our study demonstrated that SMPP had a higher percentage of Neutrophils, NLR, and CRP, which reflected more serious host immune responses in SMPP. However, the predicted value of PCT for SMPP was still controversial [11, 12], in our study, PCT values were normal and had no significant difference between groups, so it could not be used as a biomarker of SMPP. Meanwhile, some cytokines might also be involved in the immune response of MPP and were reported as predictors of SMPP in previous studies [9, 35], but for cytokines detected in our study, no significant difference was found between SMPP and GMPP groups.

Besides immune cells and cytokines, serum ALT, LDH, CKMB, or D-D dimer, which reflect organ injury or vascular endothelium damage, were also considered to be biomarkers of over-activated immune response and correlated with MPP severity [11, 12]. In our study, the ALT, LDH, and D-D dimer were higher in the SMPP group, reflecting a higher host immune response in the SMPP group indirectly.

MP strains harboring macrolide resistance gene mutation may have poor response to macrolide treatment, hence leading to SMPP [16], but there were still controversial results [18, 19]. In our study, most MP strains had an A2063G mutation, and no relationship was found between the A2063G mutation and the severity of the disease.

Besides macrolide resistance gene mutation, MP load was also found to be associated with disease severity previously [36, 37], but in our study, no relationship was found between MP concentration and disease severity, the sample depth, time, high prevalence of macrolide resistance and the MP concentration calculation method of tNGS may mask the MP load difference in our study, unified sampling and standard MP concentration calculation method may help us better understand the relationship between MP load and MP severity.

Changes in the lung microbiome, including decreased microbial diversity and dysbiosis may promote lung infection of MP [20]. Consistent with previous studies [22, 23], our work demonstrated that lower number of co-detected pathogens (decreased microbial diversity) might correlated with the severity of MPP, which supported reduced broad-spectrum antibiotic use in MP patients.

Co-existing pathogens may influence the susceptibility of MP to host, affecting the severity of MPP. So, there might be two types of co-detected pathogens for MPP: “probiotic” or “harmful microorganism” [20]. Our comparison at the kingdom, genus, and species levels found that SMPP and GMPP had different microbial spectrum (SMPP had higher percentage of Haemophilus influenzae, Fusobacterium nucleatum, RSV co-detection and Candida tropicalis and lower percentage of Staphylococcus aureus, Streptococcus pneumoniae, Rhinovirus, and Candida parapsilosis co-detection), this indicated that they may play different roles in the pathology of MPP. Previous studies found that the number of reads for RSV in MP-positive patients was higher than that in MP-negative patients, SMPP patients were more likely to have RV co-infection than GMPP [7, 38], combined with our data, we infer that RSV may interact with MP and it might be a indicator for SMPP, more studies are needed to elucidate the RSV and MP interaction in these patients.

Besides, immune statuses and respiratory tract colonization may be different in children of different ages, especially between preschool-aged and school-aged children. Younger children are more susceptible to RSV, leading to more severe disease, age might be a co-factor for disease severity, this might influence our understanding about the relationship between RSV co-infection and MP severity. However, in our study, no difference of disease severity and RSV infection frequencies between preschool-aged and school-aged children were observed, this might be due to most of the MP patients included in our study were school-aged children, the influence of age as a co-variate was masked because of the special age distribution in our study.

Finally, by integrating the environment-host factors identified in the univariate analysis above, we proposed that lower number of co-detected pathogens (decreased microbial diversity), RSV co-existing, elevated CRP and LDH might be independent risk factors for SMPP.

Our work has some limitations, Firstly, our retrospective study came from a single pandemic, seasonality may influence the pathogen species of respiratory infections, the MP strains were also relatively identical, which might influence the analysis of correlations between MP strain characters and MPP severity. So, studies including more different MP strains from different regions and times are needed in the future; secondly, the pathogens detected by tNGS in our study were relatively few, they may not be able to present all the respiratory microbiome and distinguish infection from colonization, although tNGS was reported to be efficiency and have comparable sensitivity and specificity with mNGS [27], potential bias may still exist in our study. Thirdly, the data on antibiotic use were incomplete in our study, which may hinder our understanding about the influence of the respiratory microbiome on the severity of MPP, meanwhile, due to the retrospective nature of our study, the causal relationship between the microbiome and SMPP remained unclear, more studies, especially high quality prospective studies are needed in the future.

Conclusion

Our work demonstrated co-existing pathogens (lower number of co-detected pathogens, RSV co-existing) and host response (higher CRP or LDH level) might be independent risk factors for SMPP, which might be helpful for better clinical strategy.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

MP

Mycoplasma pneumoniae

MPP

Mycoplasma pneumoniae pneumonia

SMPP

Severe Mycoplasma pneumoniae pneumonia

GMPP

General Mycoplasma pneumoniae pneumonia

mNGS

Metagenomic next-generation sequencing

tNGS

Targeted next-generation sequencing

CARDS

Community-acquired respiratory distress syndrome

WBC

White Blood Cells

NLR

Neutrophil to Lymphocyte Ratio

CRP

C-Reactive Protein

PCT

Procalcitonin

ALT

Alanine aminotransferase

AST

Aspartate transaminase

LDH

Lactic dehydrogenase

CKMB

Creatine kinase isoenzyme MB

IL6

Interleukin-6

IL2R

Interleukin-2 receptor

TNF

Tumor Necrosis Factor

Author contributions

Xiaohong Qiao conceived and designed the study, Xiaoting Song and Xiaoxiao Song collected the data, Shuanglong Lu and Ning Zhou analyzed the data and wrote the manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

This study was supported by Pediatric Service Brand Construction Project of Shanghai Putuo District, China.

Data availability

The datasets generated and/or analyzed during the current study are available in the Figshare repository (10.6084/m9.figshare.27650574).

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the Declaration of Helsinki and approved by the ethics committee of Tongji Hospital, individual consent for this retrospective study was waived because it was retrospective, there was no risk of harm to subjects, and all patients were anonymous.

Competing interests

The authors declare no competing interests.