Risk factors affecting severe pneumococcal pneumonia in children: a retrospective study.

Save

Objective

To identify risk factors for severe Streptococcus pneumoniae pneumonia (SPP) in children, and to analyze the characteristics of co-infection, drug resistance, and vaccine serotype distribution.

Methods

A retrospective cohort study was conducted. The clinical data, laboratory parameters, etiological co-infections, antibiotic resistance, and vaccine serotype coverage rates were collected and compared between the severe SPP and non-severe SPP groups.

Results

A total of 203 children were included in the study. Univariate analysis revealed that the median age, fever duration, CRP, PCT, neutrophil-to-lymphocyte ratio (NLR), and platelet-to-lymphocyte ratio (PLR) were significantly higher in the severe SPP group than in the non-severe SPP group (P < 0.05). Multivariate analysis identified older age, elevated NLR, and longer fever duration as independent risk factors for severe SPP.

The area under the receiver operating characteristic (ROC) curve for the combined prediction of the three was 0.802, the highest. Severe SPP combined with mycoplasma pneumoniae (MP) infection rate was higher (P < 0.05), but the incidence of wheezing was instead lower in this subgroup (P < 0.05). Antibiotic resistance profiles did not differ significantly between groups, with the highest resistance rate to erythromycin (over 90%) and the lowest to penicillin (1.4%–1.8%).

The vaccination rate with the 13-valent pneumococcal conjugate vaccine (PCV13) was low in both groups (9.1% severe vs. 4.7% non-severe). There were no significant differences in PCV13 serotype coverage (58.2% severe vs.

55.4% non-severe).

Conclusion

Higher median age, elevated NLR, and longer duration of fever are independent risk factors for severe SPP in children, and their combination effectively predicts disease severity. Co-infection with MP may increase the risk of severe disease but alter the clinical phenotype. Clinical treatment should focus on the high macrolide resistance, and penicillins may be preferred.

Currently, the vaccination rate for PCV13 is low, and increasing this rate is the primary task to prevent this disease.

Clinical trial number

Not applicable.

Introduction

Streptococcus pneumoniae (SP) is one of the most common bacterial pathogens in children with community-acquired pneumonia (CAP) and a significant cause of mortality in children under 5 years of age worldwide, thereby placing a substantial disease burden on families and society [1]. SP is a common colonizer of the human nasopharynx, colonizing 27–65% of children [2]. SPP has a broad spectrum of clinical manifestations and can cause milder respiratory tract infections, such as otitis and sinusitis, as well as more severe and life-threatening invasive pneumococcal disease (IPD), which includes pneumonia (with or without sepsis) and meningitis [3].

Severe SPP easily leads to a variety of extrapulmonary complications and lacks specific manifestations. If timely and effective treatment is lacking, sequelae such as necrotizing pneumonia and obliterative bronchitis may occur, which is the leading cause of poor prognosis and death in children [1]. Therefore, early identification of high-risk children who may develop severe disease and timely and powerful clinical intervention are key to improving prognosis and reducing mortality.

Currently, numerous studies have focused on identifying risk factors for severe CAP in children; however, the evidence for specific risk factors associated with severe pneumonia caused by SP, a particular pathogen, remains relatively limited, and no consensus has been established. In addition, these risk factors may vary across different regions, medical levels, and PCV vaccination backgrounds. Therefore, analysis based on clinical data from local populations is urgently needed.

Inflammatory markers, such as NLR, a novel marker that comprehensively reflects the body’s inflammatory and immune status, have demonstrated good predictive value in various severe diseases in adults [4]. However, their application value in the field of SPP in children needs to be further verified.

In addition, mixed infections of SP and MP are pretty common in clinical practice. However, the interaction mechanism and clinical characteristics of this co-infection pattern, as well as whether it exacerbates the disease, differ. Consequently, the current study’s conclusions are inconsistent [5].

In terms of treatment, the global epidemic of SP antibiotic resistance, particularly the high macrolide resistance rate, poses a significant challenge to empirical clinical treatment [6]. At the same time, although PCV13 vaccination is effective in preventing invasive disease caused by vaccine serotypes, its vaccination coverage, serotype protection against pneumonia of varying severity, and the substitution of non-vaccine serotypes are all public health issues worthy of in-depth evaluation [7].

Based on the above background, this study aimed to deeply analyze the risk factors affecting the severity of SPP in children through a retrospective cohort study, focus on evaluating the predictive efficacy of novel indicators such as NLR, and further reveal the differences in pathogen co-infection, antibiotic resistance spectrum in this region, and vaccine serotype distribution between critically ill and non-critically ill children, to provide a theoretical basis for early clinical identification of high-risk children and the development of precise prevention and control strategies.

Materials and methods

Patients

In this study, 273 children with SPP hospitalized in the Department of Respiratory Medicine, Changzhou Children’s Hospital Affiliated to Nantong University between February 2024 and February 2025 were retrospectively collected, and 203 children were finally included in the analysis after screening according to the inclusion and exclusion criteria (Fig. 1) for a retrospective cohort study. According to the diagnostic criteria for severe pneumonia in the “Guidelines for the Management of CAP in Children” formulated by the Respiratory Group of the Chinese Pediatric Society of the Chinese Medical Association [8], the groups were assessed within 24 h of admission: severe SPP group (55 cases) and the non-severe SPP group (148 cases).

Parents of all participating children provided written informed consent. This study was approved by the Ethics Committee of Changzhou Children’s Hospital (Changzhou Children’s Hospital Ethics Review 2024-003).

Diagnosis and inclusion criteria

Inclusion criteria were as follows: (1) age ≥ 28 days and ≤ 13 years; (2) meeting the diagnostic criteria of the “Guidelines for the Management of CAP in Children“ [8] (based on respiratory symptoms such as fever, cough, shortness of breath, pulmonary signs and imaging changes); (3) eligible respiratory specimens, including deep sputum bacterial culture to detect SP or bronchoalveolar lavage fluid metagenomic sequencing to clearly detect SP specific sequences. Note: In actual clinical situations, some critically ill children underwent more invasive alveolar lavage examinations to identify the pathogen, reflecting a more aggressive etiological exploration of severe cases in clinical practice.

Criteria for grouping of severe pneumonia

The grouping of this study was based on the clinical status of the child at the first comprehensive assessment (within 24 h) after admission. According to the diagnostic criteria for severe pneumonia in the Guidelines for the Management of CAP in Children [8], children who met any of the following criteria were divided into severe group, and the rest were included in the non-severe group: (1) poor general condition, refusal to eat, dehydration signs, or disturbance of consciousness; (2) respiratory distress and hypoxemia (SPO2 ≤ 92%); and (3) chest imaging showed imaging changes of polypneumonia/pleural effusion, or extrapulmonary complications.

Exclusion criteria were as follows

(1) diagnosis of nosocomial pneumonia; (2) severe clinical data loss; (3) Hospitalization < 24 h (initial severity assessment not possible); (4) combined with severe congenital diseases, immunodeficiency disease, hematological tumors, or other major chronic underlying diseases.

Outcome measures: for all laboratory parameters analyzed in this study, the first test value within 24 h after admission was used

Clinical data (gender, age, duration of fever, maximum body temperature, length of hospital stay, wheezing, history of PCV13 vaccination) and past medical history (presence or absence of underlying diseases, etc.) were collected; laboratory parameters were collected, mainly including immunoinflammatory parameters (CRP, PCT), coagulation-related parameters (D-dimer, FIB), blood biochemistry and myocardial enzyme spectrum parameters (albumin, ALT, AST, CK, CK-MB, LDH) and blood routine parameters (WBC, neutrophil percentage, lymphocyte percentage, PLT, absolute neutrophil count and absolute lymphocyte count) at admission, derived parameters NLR, PLR and PNR; microbiological data were recorded, including MP infection and other pathogen infection status to determine combined infection. Serotyping and antibiotic-susceptibility test results for SP isolates. Among them, PCV13 serotype coverage rate: 13 serotypes contained in PCV13 (1, 3, 4, 5, 6 A, 6B, 7 F, 9 V, 14, 18 C, 19 A, 19 F, 23 F) account for the proportion of all typed strains.

PPSV23 non-PCV13 serotype coverage: the 11 serotypes (2, 8, 9 N, 10 A, 11 A, 12 F, 15B, 17 F, 20, 22 F, 33 F) additionally covered by PPSV23 accounted for the proportion of all typed strains. Non-vaccine serotype coverage: proportion of serotypes not covered by PCV13 or PPSV23.

Criteria for diagnosing MP co-infection

In this study, the diagnosis of MP infection was primarily based on serological methods commonly used in clinical practice. A particle agglutination assay was used to detect serum MP-IgM antibody titers of ≥ 1:160 in the acute phase, or a fourfold or greater increase in double serum MP antibody titers during the course of the disease. All serologically positive cases were reviewed by senior clinicians in combination with the child’s medical history, clinical manifestations, and chest imaging characteristics to comprehensively determine MP as one of the causative pathogens of this infection.

Statistical analysis

Statistical analysis was performed using SPSS 25.0 software. Normality test (Shapiro-Wilk test) and homogeneity test were performed. Measurement data that met normal distribution and had equal variance were expressed as mean ± standard deviation.

An independent samples t-test was used to compare differences between groups. Measurement data not meeting normal distribution and having unequal variance were expressed as median (interquartile range) [M (P25, P75)]. An independent-samples nonparametric test (Mann-Whitney U rank-sum test) was used to compare differences.

Enumeration data were presented as rates (%), compared between groups using the chi-square test, and Fisher’s exact test was used when the theoretical frequency was < 5. Variables that were statistically significant and clinically meaningful in the univariate analysis were included in the multivariate logistic regression model for analysis. Regression analysis was performed using the forward stepwise method (Forward: LR) to identify independent risk factors for progression to severe disease in children with SPP, and their odds ratios (ORs) and 95% confidence intervals (95% CI) were calculated.

The goodness of fit of the final model was assessed using the Hosmer-Lemeshow test, with a P-value greater than 0.05 indicating a good model fit.

Furthermore, independent risk factors identified through logistic regression models were utilized to construct predictive models. ROC curves were used to analyze the predictive value of each independent risk factor and its combination for severe pneumonia. The Youden index was calculated to obtain the sensitivity and specificity, and the point with the maximum value was used as the optimal cutoff.

P < 0.05 was considered statistically significant.

Results

Patient screening and grouping process

As shown in Fig. 1 and 203 children with SPP, following screening, were included in the analysis and constituted the study cohort. According to the assessment, within 24 h after admission, the patients were divided into a severe SPP group (n = 55) and a non-severe SPP group (n = 148) based on the guideline criteria.

Fig. 1Study population

Etiological diagnostic features

All 203 children were microbiologically confirmed. Among them, the non-severe group (148 cases) was diagnosed by deep sputum culture. In the severe group, 48 cases (87.3%) were diagnosed by deep sputum culture.

Another 7 cases (12.7%) were diagnosed by metagenomic sequencing of bronchoalveolar lavage fluid due to severe disease or as part of the diagnostic and treatment process. There was a significant difference in the composition of diagnostic methods between the two groups (P < 0.001). Despite this difference, all children were definitively diagnosed with SPP by a specific process, fulfilling the core inclusion criteria for this study.

And this difference reflects a more aggressive etiological exploration practice for severe cases. Therefore, subsequent comparative analyses based on this grouping have a sound clinical and biological basis.

Baseline data

There was no significant difference in gender, WBC, neutrophil count, PLT, D-dimer, CK-MB, LDH, and other general data between the two groups (P > 0.05). The median age, duration of fever, maximum body temperature, length of hospital stay, CRP, PCT, neutrophil percentage, NLR, and PLR in the severe SPP were higher than those in the non-severe SPP group (P < 0.05). The incidence of wheezing, PNR, and albumin in the severe SPP was lower than that in the non-severe SPP group (P < 0.05).

The absolute difference in maximum body temperature between the two groups was slight, and the value of clinical differentiation was limited; therefore, they were excluded from the subsequent multivariate logistic regression analysis. See Table 1 and 2.

Table 1Baseline characteristics of the study populationsevere SPP(n = 55)non-severe SPP(n = 148)t/Z/χ2-valuep-valueGender (M/F)34/2183/650.5410.462Age, years3.83(2.75,4.55)1.92(0.81,3.75)-4.853<0.001Fever duration (d)4.00(2.50,5.00)1.00(0.00,3.00)-5.080<0.001Max Temp (℃)39.42 ± 0.6939.04 ± 0.77-29230.004Length of hospital stay, d8.00(6.00,9.00)6.00(5.00,8.00)-4.146<0.001Incidence of wheezing(n, %)9(16.4%)53(35.8%)7.1490.008Data are presented as mean ± standard deviation, n (%), or median (interquartile range)

Table 2Laboratory test results in study populationsevere SPP(n = 55)non-severe SPP(n = 148)t/Z-valuep-valueCRP (mg/L)13.38(6.58,29.84)7.80(3.90,21.62)-2.5180.012WBC (×109/L)10.11(6.15,13.28)9.73(7.30,13.26)-0.5900.555PCT (ng/mL)0.08(0.03,0.21)0.04(0.02,0.09)-2.9880.003N (%)60.09 ± 16.4048.71 ± 17.90-4.115<0.001PLT (×109/L)302.55 ± 100.54318.10 ± 105.760.9430.347NLR2.24(1.12,3.46)1.19(0.63,2.15)-3.862<0.001PLR103.62(73.05,161.08)87.70(55.99,111.58)-3.3050.001PNR54.67(31.19,85.40)70.67(44.46,107.29)-2.3580.018D-dimer (mg/L)0.39(0.21,0.69)0.33(0.21,0.52)-1.1750.240FIB (g/L)3.82 ± 1.233.34 ± 1.14-2.5910.010Albumin (g/L)41.16 ± 3.9843.51 ± 2.934.000<0.001CK-MB (U/L)31.00(22.50,53.00)39.00(29.00,55.00)-1.6430.100LDH (U/L)367.00(316.50,457.50)356.00(312.50,444.00)-0.4110.681Data are presented as mean ± standard deviation or median (interquartile range); WBC: White blood cell count(4.3–11.3 × 109/L); CRP: C-reactive protein(0–10 mg/L); PCT: Procalcitonin (< 0.05 ng/mL); PLT: Platelets(167–453 × 109/L); NLR: Neutrophil/lymphocyte ratio; PNR: Platelet/neutrophil ratio; PLR: Platelet/lymphocyte ratio; N: Neutrophil percentage(31–70%); D-dimer (< 1 mg/L); FIB: Fibrinogen(2–4 g/L); Albumin (39–54 g/L); CK-MB: creatine kinase isoenzyme(0–25 U/L); LDH: Lactic dehydrogenase(109–245 U/L)

Logistic multivariate analysis affecting the occurrence of severe pneumonia in children infected with SP

Univariate analysis showed that median age, duration of fever, maximum body temperature, CRP, PCT, NLR, PLR, PNR, neutrophil percentage, and incidence of wheezing were significantly associated with the occurrence of severe SPP in children (P < 0.05). Multivariate logistic regression analysis revealed that higher median age, higher NLR, and longer fever duration were risk factors for severe SPP in children (P < 0.05). The results of the Hosmer-Lemeshow goodness-of-fit test for this multivariate logistic regression model showed that χ2 = 7.388, P = 0.495, suggesting good model calibration.

See Table 3.

Table 3Logistic regression analysis of independent risk factors for severe pneumonia in children infected with SPVall’s Chi-squarep-valueOR95%CIMedian Age5.6810.0171.272(1.044,1.549)Fever duration17.043<0.0011.449(1.215,1.728)NLR7.2080.0071.205(1.052,1.381)NLR: Neutrophil/lymphocyte ratio

ROC curves were used to evaluate the predictive value of median age, duration of fever, and NLR level for the development of severe pneumonia in children infected with SP

Median age, duration of fever, and NLR level had a significant predictive value for the development of severe SPP in children (P < 0.001). The area under the curve (AUC) was 0.722, 0.729, and 0.676, respectively. The optimal cutoff value of median age was 1.29 years, with a predictive sensitivity of 96.4% and a specificity of 41.9%; the optimal cutoff value of fever time was 2.5 days, with a predictive sensitivity of 74.5% and a specificity of 62.2%; the optimal cutoff value of NLR was 1.395, with a predictive sensitivity of 70.9% and a specificity of 57.4%; and the predictive value of the combination of the three indicators increased, with an AUC of 0.802, a predictive sensitivity of 80.0%, and a specificity of 68.2% (P < 0.001).

See Table 4; Fig. 2.

Table 4Predictive value of risk factors for severe pneumonia in children infected with SPAUCCritical valueSensitivity(%)Specificity(%)p-value95%CIMedian Age0.7221.29096.40041.900<0.0010.650–0.794Fever duration0.7292.50074.50062.200<0.0010.645–0.812NLR0.6761.39570.90057.400<0.0010.597–0.756Three indicators combined0.8020.21580.00068.200<0.0010.737–0.867AUC: Area under the curve; NLR: Neutrophil/lymphocyte ratio

Fig. 2ROC curves for predicting severe pneumonia in children infected with SP

Infections with other pathogens in children with SPP

There were 21 children (38.2%) with severe SPP combined with MP infection and 35 children (23.6%) with non-severe SPP combined with MP infection, and the difference between the two groups had statistical significance (P < 0.05); there was no significant difference in the ratio of children with respiratory syncytial virus (RSV), influenza virus (IV), human parainfluenza viruses (HPIVs), rhinovirus (RV), adenovirus (ADV) and MP-viral infection between the two groups (P > 0.05). The rate of wheezing in children with severe SPP combined with MP infection was 14.3%, which was significantly lower than that in non-severe SPP combined with MP infection (40.0%), and the difference had statistical significance (P < 0.05); See Table 5.

Table 5Infections with other pathogens in children with SPPsevere SPP(n = 55)non-severe SPP(n = 148)χ2-valuep-valueCombined MP(n, %)21(38.2%)35(23.6%)4.2400.039Combined RSV(n, %)6(10.9%)10(6.8%)0.9520.329Combined IV(n, %)1(1.8%)7(4.7%)0.8980.343Combined HPIVs(n, %)0(0%)4(2.7%)1.5160.218Combined RV(n, %)2(3.6%)4(2.7%)0.1220.727Combined ADV(n, %)2(3.6%)4(2.7%)0.1220.727Combined MP-virus(n, %)4(7.3%)8(5.4%)0.2510.616Incidence of wheezing associated with MP(n, %)3(14.3%)14(40.0%)4.1050.043MP: Mycoplasma pneumoniae; RSV: Respiratory syncytial virus; IV: Influenza virus; HPIVs: human parainfluenza viruses; RV: Rhinovirus; ADV: adenovirus

Resistance of SP

The resistance rate of erythromycin was the highest in severe SPP group and non-severe SPP group, up to 100.0% and 98.0%, respectively; followed by clindamycin, doxycycline, cotrimoxazole and cefuroxime, the resistance rate was 98.2%, 78.2%, 63.6% and 49.1% in severe SPP group and 90.5%, 74.3%, 62.8% and 52.7% in non-severe SPP group, respectively, and the differences were not statistically significant (P > 0.05); the resistance rate of cefepime, meropenem, ceftriaxone, amoxicillin/clavulanic acid and penicillin was 5.5%, 5.5%, 1.8% and 1.8% in severe SPP group, and 8.1%, 4.1%, 3.4% and 1.4% in non-severe SPP group, respectively, and there was no significant difference between the two groups (P > 0.05); the resistance rate of SP strains to rifampicin, linezolid, moxifloxacin and vancomycin was 0% in the two groups. One SP strain was levofloxacin-resistant in the non-severe SPP group but not in the severe SPP group. See Table 6.

Table 6Antibiotic sensitivity patterns of SP isolatesresistancesevere SPP(n = 55)non-severe SPP(n = 148)χ2-valuep-valueErythromycin(n, %)55(100.0%)145(98.0%)1.1320.287Penicillin(n, %)1(1.8%)2(1.4%)0.0600.806Ceftriaxone(n, %)3(5.5%)5(3.4%)0.4570.499Cefuroxime(n, %)27(49.1%)78(52.7%)0.2090.647Doxycycline(n, %)43(78.2%)110(74.3%)0.3210.571Meropenem(n, %)3(5.5%)6(4.1%)0.1860.667Clindamycin(n, %)54(98.2%)134(90.5%)3.4210.064Amoxicillin/Clavulanic Acid(n, %)1(1.8%)6(4.1%)0.6020.438Cefepime(n, %)3(5.5%)12(8.1%)0.4130.521Levofloxacin(n, %)0(0%)1(0.7%)0.3730.541Cotrimoxazole(n, %)35(63.6%)93(62.8%)0.0110.917

SP vaccine composition

The serotype distribution in the two groups showed similar composition, with the top five serotypes being 19 F, 6B, 23 A, 19 A, and 23 F. The serotype coverage rate of PCV13 vaccine was 58.2% and 55.4% in severe SPP group and non-severe SPP group; the serotype coverage rate of PPSV23 non-PCV13 vaccine was 5.5% and 2.7% in severe group and non-severe group; the proportion of non-vaccine serotypes was 29.1% and 35.8% in severe group and non-severe group; PCV13 vaccination rates were low in both groups, 9.1% in severe group and 4.7% in non-severe group; the above differences were not statistically significant (P > 0.05). See Table 7.

Table 7SP vaccine compositionsevere SPP(n = 55)non-severe SPP(n = 148)χ2-valuep-valuePCV13 covered serotypes(n, %)32(58.2%)82(55.4%)0.1260.723PPSV23 non-PCV13 covered serotypes(n, %)3(5.5%)4(2.7%)0.9120.340Non-vaccine serotypes(n, %)16(29.1%)53(35.8%)0.8070.369PCV13 Inoculation Rate5(9.1%)7(4.7%)1.3710.242PCV13: 13-valent pneumococcal conjugate vaccine; PPSV23: 23-valent pneumococcal polysaccharide vaccine

Discussion

In this study, we systematically investigated risk factors, etiological characteristics, and vaccine protective effects that contribute to the progression of SPP to severe disease in children through a retrospective cohort analysis, providing an important basis for early clinical identification of high-risk children and for the development of prevention and treatment strategies.

First, our univariate analysis revealed that children with severe SPP had significantly higher levels of inflammatory markers (e.g., CRP, PCT, neutrophil percentage, NLR, PLR) than those in the non-severe group. In comparison, nutritional and immune status parameters (e.g., albumin, PNR) were significantly lower, consistent with previous studies and suggesting that excessive inflammatory responses and a relatively weak immune status are important pathophysiological bases for disease severity [9, 10]. However, it is noteworthy that absolute levels of CRP and PCT were relatively low at admission, even in the severe group.

On the one hand, this may be related to the fact that the early presentation of the child and the use of antibiotics before referral partially inhibit the peak inflammation; on the other hand, it may also reflect the differences in the intensity of inducing a systemic inflammatory response among different SP serotypes [11]. Despite these factors that may decrease absolute levels, intergroup comparisons showed that inflammatory markers remained significantly higher in critically ill children. This finding is consistent with the conclusion of a recent extensive retrospective study of 28,043 children with CAP, which confirmed that a CRP level greater than 8 mg/L was an independent risk factor for progression to severe CAP in children [12].

In addition, it has been noted that CRP and PCT, as key indicators of systemic inflammation, are positively correlated with the severity of pneumonia; however, there is insufficient evidence to suggest that they are specific in distinguishing pneumonia caused by different pathogens [10].

Further multivariate logistic regression analysis, which controlled for confounding factors, identified higher median age, longer duration of fever, and elevated NLR as independent risk factors for severe disease in children with SPP. In this study, we found that older children (but still within the range of infants or preschool children) had a higher risk of severe disease, which seems contrary to some classical perceptions (i.e., younger age is a high-risk factor for severe disease). This may be related to an increased risk of community exposure, exposure to pathogen loads, and possibly greater virulence in older children [13].

Notably, recent systematic reviews on IPD clearly state that the outcome of disease is influenced by both host characteristics (e.g., age) and bacterial virulence factors (including serotypes) [14]. This suggests that age is not a simple linear relationship with risk of severe disease. The duration of fever directly reflects the body’s ability to control the infection and the persistence of the inflammatory response.

As an easily obtainable composite inflammatory index, the NLR encompasses both neutrophils and lymphocytes, which can sensitively reflect the degree of innate immunity and inflammatory response in the body, and has been confirmed to have good application value in evaluating the severity and prognosis of infectious diseases, such as sepsis [4]. However, the current high-level evidence is still insufficient for its specific predictive value in SPP. The results of this study may shed new light on this.

Our ROC curve analysis revealed that combining median age (> 1.29 years), duration of fever (> 2.5 days), and NLR (> 1.395) improved the predictive efficacy for severe disease, surpassing that of any single index. This combination provided a simple and effective risk-assessment model for the early and rapid identification of high-risk children in clinical settings. For children with these characteristics, clinicians should be vigilant and take more aggressive interventions to prevent their condition from progressing to severe disease.

Second, we observed that nearly half of the children in both groups were co-infected with MP or RSV, likely because viral respiratory infections were the primary contributing factor to SP [15]. This phenomenon is consistent with previous findings [1]. At the same time, we also observed that the proportion of severe SPP combined with MP infection was significantly higher than that of the non-severe group.

A notable finding is that, among children with MP infection, wheezing was significantly lower in the severe SPP group than in the non-severe SPP group. It has been demonstrated that age is a significant factor in the development of wheezing. Younger children are more likely to experience wheezing symptoms after suffering from pneumonia due to airway narrowing and an immature immune system [16].

The mean age of children in the non-severe group was younger in this study. At the same time, that study also highlighted that children with simple bacterial pneumonia (such as SPP) were more likely to experience wheezing symptoms than those with MP infection [16]. Another study reported that MP infection was 12% more likely to be associated with wheezing freedom than infection with other pathogens, and wheezing-free status was a statistically significant diagnostic indicator [17].

Thus, this may suggest that in the context of SP co-infection with MP, the pathophysiological mechanism of severe pneumonia focuses more on a strong systemic inflammatory response and tissue damage [18] rather than wheezing due to airway hyperreactivity. This conclusion is also consistent with the finding that some studies have identified a history of wheezing as a negative predictor of severe pneumonia [19].

Resistance rates to SP vary geographically among classes of antibiotics and change over time, with macrolide resistance rates being particularly high in Asia (> 90%) and penicillin non-susceptibility rates being high in Asia, Europe, and South American countries (45–50%) [20]. Therefore, antibiotic therapy for SP infection should be guided by local antibiotic susceptibility profiles. Regarding drug resistance, there was no significant difference in the antibiotic resistance spectrum between the two groups.

The resistance rate to penicillin was very low (1.4%–1.8%), and there were no cases of resistance to rifampicin, linezolid, moxifloxacin, vancomycin, or other key drugs. Notably, high susceptibility (resistance 3%–5%) to third-generation cephalosporins (ceftriaxone) remains. Among them, the penicillin resistance rate of SP strains in this study was lower than that in Shanghai (4.9%) in East China [1].

These data strongly support the use of penicillins or third-generation cephalosporins as credible options for empirical therapy until susceptibility results are returned. At the same time, vancomycin or linezolid should be reserved for severe or suspected drug-resistant bacterial infections. However, the high resistance rates to macrolides, such as erythromycin and clindamycin, are consistent with worldwide and domestic surveillance data [1], strongly suggesting that macrolides alone should be avoided in empirical treatment, especially in severe cases.

The only case of emergence of levofloxacin-resistant strains, although individual, reminds us of the need for careful monitoring of the pediatric use of fluoroquinolones. These data provide direct evidence for updating antimicrobial steering protocols for SP in children in this region.

This study again highlights the shortcomings of vaccination. PCV13 vaccination rates were very low in both groups (9.1% in the severe group and 4.7% in the non-severe group), indicating that an effective immune barrier had not yet been formed. PCV13 coverage of serotypes from both groups of isolates exceeded 55%, highlighting that increasing its coverage rate is currently the most cost-effective preventive measure.

This coverage is generally consistent with recent global surveillance data (PCV13 serotypes cause approximately 52.1% of IPD cases in children under 5 years of age) [11], but is lower than the average reported in other regions of China of 68.4% [21]. In terms of serotype distribution, the dominant serotype composition of the two groups in this study was similar, with the top five serotypes being 19 F (20% severe vs. 16.8% non-severe), 6B, 23 A, 19 A, and 23 F.

This finding is consistent with the results of a recent systematic review covering IPD versus non-IPD in mainland China, which reported the highest proportion of serotype 19 F at 22.0% [22]. Notably, the proportion of PPSV23 non-PCV13 serotypes was somewhat higher in the severe group, although the difference did not reach statistical significance. Non-PCV13 serotypes, including PPSV23 non-PCV13 serotypes and non-vaccine serotypes, together accounted for more than one-third of both groups, although this difference was not statistically significant.

This finding may suggest that serotype replacement is occurring in response to the pressure of PCV13 vaccination. This is in line with global trends, where up to 50% of IPD cases are caused by non-PCV13 serotypes in regions where PCV13 is widely used [11]. Therefore, while efforts are being made to increase PCV13 vaccination rates, vigilance and continuous monitoring of serotype shifts are essential.

Limitations of this study

First, as a single-center retrospective study, selection bias and information bias are unavoidable, and extrapolating conclusions requires further verification in multicenter studies. In this study, differences were observed in etiological diagnosis methods between the two groups. Specifically, some children in the severe group were diagnosed using metagenomic sequencing of bronchoalveolar lavage fluid, whereas all children in the non-severe group relied on sputum culture.

This reflects the reality of more aggressive and precise etiological exploration of severe cases in clinical practice, but may theoretically introduce bias with inconsistent diagnostic sensitivity. Nevertheless, given the greater sensitivity and specificity of metagenomic sequencing, and that all tests clearly point to SP, the main conclusion of this study, regarding risk factors for severe disease, remains based on a clear SP-infected population and has important clinical reference value. Future prospective studies may use uniform, highly sensitive diagnostic criteria for further validation.

Second, in this study, MP co-infection was primarily determined by serological detection of MP-IgM antibodies, combined with clinical diagnosis. However, this method still cannot completely rule out the possibility of false-positive MP-IgM. This limitation may lead to some overestimation of the rate of MP co-infection.

However, it should be emphasized that both groups in this study used the same diagnostic criteria; therefore, the conclusions of the group comparisons were relatively less affected. Future studies should employ more specific diagnostic methods, such as PCR based on respiratory specimens, for further validation.

Third, this study did not conduct a comprehensive etiological test in all children, so the impact of co-infection with other potential pathogens on disease severity cannot be ruled out completely.

Conclusion

Higher median age, longer duration of fever, and elevated NLR level are independent risk factors for the development of severe SPP in children. The combination of the three has a higher predictive value for early identification of high-risk children.Co-infection with SP and MP may increase the risk of severe disease by exacerbating pulmonary parenchymal inflammation rather than inducing wheezing, and clinical vigilance is needed for this atypical but more severe co-infection pattern.In this study, the isolates maintained high sensitivity to penicillin but severe resistance to macrolides, and it is recommended to preferentially use penicillins according to the susceptibility results, avoid macrolides alone, and retain vancomycin and linezolid as the cornerstone of treatment of severe infections.The PCV13 vaccination rate of the study population is very low, and increasing the vaccination rate is the fundamental strategy to prevent SPP and its severe disease. At the same time, it is necessary to remain vigilant for the risk of infection from non-vaccine serotypes, including PPSV23 non-PCV13 serotypes.