Long-term sero-epidemiological trends of Nine respiratory pathogens in a predominantly pediatric population in Southwest China: a multicenter retrospective study (2019–2023)

Abstract

Background

Acute respiratory tract infections (ARTIs) remain a major global health burden. Southwest China, with its complex geography and climate, represents a critical region for respiratory pathogen ecology. This study aimed to characterize the epidemiological dynamics of nine respiratory pathogens among ARTI patients in Southwest China across the pre-COVID-19 period (January 2019 to December 2019), the COVID-19 non-pharmaceutical interventions (NPIs) period (January 2020 to December 2022), and the post-NPIs period (January 2023 to December 2023).

Methods

A multicenter, retrospective study was conducted from January 2019 to December 2023, including 240,611 ARTI patients. Serum IgM antibodies against nine pathogens: Influenza A virus (FluA), Influenza B virus (FluB), Respiratory syncytial virus (RSV), Parainfluenza Virus (PIV), Adenovirus (ADV), Mycoplasma pneumoniae (MP), Chlamydia pneumoniae (CP), Legionella pneumophila (LP), and Coxiella burnetii (Q fever, QF) were detected by indirect immunofluorescence assay, and the findings were interpreted as serological positivity/recent exposure signals rather than definitive microbiological confirmation of acute infection. Epidemiological characteristics and temporal trends were analyzed using descriptive statistics and Join-Point regression.

Results

Monthly serological positivity rates showed a marked decline during the early phase of the NPIs period. However, the overall serological positivity rate surged to 30.07% in the post-NPIs period, significantly higher than the pre-COVID-19 (27.83%) and NPIs (27.87%) periods (p < 0.001). MP showed the highest serological positivity signal across all study periods. Pathogen responses diverged: FluB remained suppressed (declining to 0.58% in the post-NPIs period), whereas FluA and RSV resurged significantly. Children aged 28 days-4 years accounted for 49.56% of the cohort, indicating a predominantly pediatric testing population. The median age among serologically positive cases increased for several major pathogens, for example, FluA from 3 to 6 years and RSV from 0 to 2 years, despite higher pediatric susceptibility. Serological co-detection rates increased from 4.05% to 4.86% with a predominant shift from FluB-MP to PIV-MP. Join-Point analysis confirmed a sharp decline of FluB and a sustained increase of PIV in the NPIs period.

Conclusion

The NPIs were associated with substantial shifts in the serological positivity patterns of respiratory pathogens in Southwest China. While serology alone is insufficient to define the full complexity of acute respiratory pathogen ecology, these findings should be interpreted as serological exposure patterns rather than confirmed acute infections, and they provide a unique, large-scale sero-epidemiological map of recent exposure dynamics that serves as a valuable baseline for future molecular surveillance.

Clinical trial number: Not Applicable.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-026-13566-2.

Introduction

Acute respiratory tract infections (ARTIs) represent a leading cause of global infectious morbidity and mortality, imposing substantial healthcare and economic burdens worldwide, especially in children under five and older adults [1–3]. The etiology of ARTIs is diverse, involving viruses, bacteria, and atypical bacteria, whose circulation is influenced by complex interactions of host age, geography, and seasonal factors [4–6]. Since different respiratory pathogens often present with similar clinical symptoms, understanding their serological detection patterns and recent exposure profiles may provide useful epidemiological context, especially in large retrospective datasets where molecular confirmation is unavailable.

The implementation of non-pharmaceutical interventions (NPIs) during the SARS-CoV-2 pandemic disrupted the global circulation of common respiratory pathogens like Influenza and Respiratory syncytial virus [7–9]. Following the relaxation of restrictions, many regions reported a resurgence of respiratory infections, often with atypical timing, intensity, and pathogen interactions, raising concerns about an “immunity debt” due to reduced pathogen exposure during the NPIs period [10–12]. In China, stringent NPIs were maintained for nearly three years (January 2020 to December 2022), suppressing routine pathogen circulation and offering a unique context to evaluate the long-term impact of public health measures on respiratory pathogen ecology [13].

Most respiratory pathogen surveillance studies in China have been limited to single-center studies or specific subpopulations, lacking a unified analysis across the pre-pandemic, NPIs, and post-NPIs periods [14–17]. This gap is particularly relevant in Southwest China, a region characterized by marked altitudinal variation, diverse climatic zones, and substantial geographic heterogeneity, all of which may influence respiratory pathogen circulation patterns [18, 19]. To address this, we conducted a multicenter, retrospective study in Southwest China from January 2019 to December 2023. This study aimed to characterize the sero-epidemiological trends of nine respiratory pathogens, including five viruses (Influenza A virus, Influenza B virus, Respiratory syncytial virus, Adenovirus, and Parainfluenza virus) and four atypical bacteria (Mycoplasma pneumoniae, Chlamydia pneumoniae, Legionella pneumophila, and Coxiella burnetii), across these different phases, focusing on serological positivity/exposure patterns, demographic distribution, seasonality, and serological co-detections. Our findings are intended to provide valuable insights for informing regional public health strategies, guiding clinical management, and optimizing vaccination policies in the post-pandemic era.

Methods

Ethical considerations

This multicenter, retrospective study was approved by the Biomedical Ethics Review Committee of West China Hospital, Sichuan University. Due to the retrospective nature of the study, the requirement for written informed consent was waived by the ethics committees of all participating centers. All procedures followed the ethical principles of the Declaration of Helsinki.

Study design and data sources

This retrospective multicenter study was based on routine clinical data from participating tertiary hospitals in Southwest China between January 2019 and December 2023. Eligible patients were care-seeking individuals presenting with symptoms of acute respiratory tract infection (ARTI), such as fever, cough, sore throat, sputum production, or other respiratory infection complaints, who underwent clinician-ordered IgM antibody testing for respiratory pathogens as part of etiological evaluation. The testing was performed in hospital-based clinical settings, including both outpatient and inpatient care, and the resulting cohort should therefore be interpreted as a care-seeking serologically tested population rather than a community-based surveillance population.

Demographic and laboratory data were extracted from the Hospital Information System (HIS) at each participating institution using a standardized protocol. All data were anonymized and de-identified prior to analysis to ensure confidentiality and compliance with privacy regulations.

Exclusion criteria:

1. Records with missing essential variables, including age, sex, testing date, study period classification, pathogen IgM result, or department information.

2. Patients younger than 28 days, as the dynamics of neonatal infections differ significantly from those of the general population [20].

3. For patients who underwent multiple tests during the same hospitalization or visit, only the earliest result was retained. If any were positive, the first positive result was selected for analysis.

Variables and statistical design

The study period was divided into three phases: (1) the pre-COVID-19 period (January 2019 to December 2019), (2) the non-pharmaceutical interventions (NPIs) period during COVID-19 public health restrictions (January 2020 to December 2022), and (3) the post-NPIs period (January 2023 to December 2023).

For seasonal analysis, standard meteorological seasons were used: winter (December-February), spring (March-May), summer (June-August), and autumn (September-November). The first two months of 2019 were classified as winter 2018, while only December was included for winter 2023. For the temporal trend analysis using Join-Point regression, the calendar quarters were defined as Q1 (January-March), Q2 (April-June), Q3 (July-September), and Q4 (October-December) to provide standardized, equidistant time intervals.

Patients were stratified into eight age groups: (1) 28 days-4 years, (2) 5–9 years, (3) 10–14 years, (4) 15–19 years, (5) 20–39 years, (6) 40–59 years, (7) 60–79 years, and (8) ≥80 years [21].

Hospital departments were categorized into eight groups based on study objectives and institutional structure: pediatrics (Peds), pulmonary and critical care medicine (PCCM), infectious diseases (ID), emergency department (ED), intensive care units (ICUs), internal medicine (IM), surgery (SUR), and other specialties (Others). The ICUs category included general ICUs, department-specific ICUs, and critical care medicine units.

Specimen collection and testing

Serum samples were collected by trained personnel. Indirect immunofluorescence assay was employed for the simultaneous detection of IgM antibodies against nine respiratory pathogens using the PNEUMOSLIDE IgM kit (Vircell S.L., Spain). The targeted pathogens were: Influenza A virus (FluA), Influenza B virus (FluB), Respiratory syncytial virus (RSV), Parainfluenza virus (PIV) types 1, 2, and 3, Adenovirus (ADV), Mycoplasma pneumoniae (MP), Chlamydia pneumoniae (CP), Legionella pneumophila (LP) serotype 1, and Coxiella burnetii (Q fever, QF). Given the retrospective nature of the cohort, a systematically collected PCR-confirmed subsample and a centralized external inter-laboratory quality-control dataset were not available for the present analysis. To promote inter-center reliability and minimize technical errors across the 5-year study period, all assay procedures were performed strictly according to the same manufacturer-provided standard operating procedures (SOPs) and instructions across all participating laboratories. Slides were interpreted under fluorescence microscopy by two independent senior laboratory technicians who were professionally trained and qualified under national clinical laboratory standards. Routine internal quality-control procedures were implemented at each center to monitor batch validity (detailed in Supplementary methods). Because this was a retrospective multicenter study based on routine clinical testing records, paired acute-convalescent sera, pathogen-specific IgG seroconversion data, PCR confirmation, and vaccination histories were not consistently available. Therefore, a positive IgM result in the present study was interpreted as serological positivity/recent exposure signal, rather than definitive proof of concurrent acute infection.

Statistical analysis

The dependent variables included serological positivity (overall or pathogen-specific), and serological co-detection, while independent variables included age group, sex, season, patient type, clinical department, and study period. Categorical variables were summarized as frequencies and proportions. For descriptive comparisons, serological positivity and co-detection rates were compared across study period, age group, sex, season, patient type, and clinical department using Pearson’s chi-square test or Fisher’s exact test, as appropriate. Continuous variables (e.g., age) were assessed for distributional characteristics. Given the evident distribution of age in this large cohort, continuous variables were summarized as medians with interquartile ranges (IQR), and between-group comparisons were performed using non-parametric tests (Kruskal-Wallis test, followed by Dunn’s post hoc test where appropriate). The association between age and serological positivity rate was evaluated using Spearman’s analysis.

To further quantify the independent association between age groups and overall serological positivity while accounting for potential confounding, we performed a multivariable analysis. The prevalence ratio (PR) with a 95% confidence interval (CI) was estimated by the modified Poisson regression model with robust variance estimation due to the relatively high serological positivity rate (>20%) in this cohort [22]. Age group was the primary explanatory variable, with children aged 28 days-4 years as the reference group. Separate models were fitted for the pre-COVID-19, NPIs, and post-NPIs periods. Sex, season, patient type, clinical department, and hospital (as a fixed effect) were included as covariates.

Temporal trends were analyzed using the Join-Point Regression Program (version 5.4.0, Statistical Research and Applications Branch, National Cancer Institute, USA), with quarterly data as the unit. When statistically significant changes (joinpoints) were identified, linear regression models were fitted to estimate trend parameters for each segment. Linear trends were quantified as quarterly percentage change (QPC) using a natural log-linear model [ln(y) = xb]. Detailed model specifications and parameter settings are provided in the Supplementary Material.

All statistical analyses were performed using R software (version 4.4.3). Data visualization was implemented with the OriginPro 2021 (version 9.8.0.200) and UpSetR package. A two-sided P value < 0.05 was considered statistically significant.

Results

Epidemiological overview and overall detection trends

The study region in Southwest China encompasses a population of over 180 million people and covers a land area of about 1.12 million km² [23, 24]. A total of 240,611 patients presenting with ARTI symptoms, such as fever, cough, or sore throat, who underwent IgM antibody testing for nine respiratory pathogens between January 2019 and December 2023 were included in the analysis. Patient data were collected from three provinces in Southwest China, including Sichuan (65,781 cases), Yunnan (102,967 cases), and Guangxi (71,863 cases). The cohort had a median age of 5 years (IQR: 1–25). Children aged 28 days to 4 years constituted the largest age group (49.56%, 119,247/240,611), indicating a predominantly pediatric testing population (Table S1).

Overall, 68,231 individuals showed serological positivity for at least one pathogen (serological positivity rate: 28.36%), reflecting population-level recent exposure signals. MP showed the highest serological positivity signal (25.00%, 60,162/240,611) across all study periods. The overall serological positivity rate increased from the pre-COVID-19 (27.83%, 17,280/62,091) to the post-NPIs (30.07%, 16,354/54,390, p < 0.001) period (Table S3). Females consistently showed higher serological positivity rates than males across all periods (Table S2).

Monthly testing volumes fluctuated, with notable declines early in the NPIs period (Fig. 1A). MP showed a consistently high serological positivity signal, accounting for over 60% of monthly seropositive cases. Most pathogens showed elevated activity during spring and winter. FluA, MP, LP, and QF exhibited peak serological positivity in spring, while FluB, RSV, PIV, and ADV peaked in winter. CP showed no distinct seasonality (Fig. 1B; Table S5). Pathogen-specific serological positivity rates varied significantly across periods (Fig. 1C; Table S3). MP showed higher IgM positivity in the NPIs (24.72%) and post-NPIs (27.71%) periods than in the pre-COVID-19 (23.20%) period (all p < 0.001). A more detailed annual breakdown of serological positivity rates is provided in Table S4, which allows further evaluation of temporal patterns in relation to previously reported PCR-based epidemiological trends [25]. Several pathogens experienced significant declines during the NPIs period compared to the pre-pandemic levels: FluA (0.66% to 0.35%), FluB (3.72% to 1.51%), RSV (1.95% to 0.88%), CP (0.20% to 0.07%), LP (1.14% to 0.41%), and QF (0.10% to 0.04%) (all p < 0.001). In the post-NPIs period, FluA and RSV rebounded in serological positivity signals to 0.82% and 1.10%, respectively (both p < 0.001), while FluB continued to decline to 0.58% (p < 0.001). Notably, CP, LP, and QF serological positivity rates in the post-NPIs period remained significantly lower than their pre-pandemic levels (all p < 0.001).

Fig. 1.

Temporal trends and epidemiological distribution of nine respiratory pathogens among patients with ARTI in Southwest China, 2019–2023. (A) monthly test volume and serological positivity rate. Stacked bars: monthly serologically positive cases (blue: male; pink: female). Grey area: monthly test volume. Red line: monthly serological positivity rate. Background shading marks the study periods: pre-COVID-19, NPIs, and post-NPIs, separated by dashed lines. (B) monthly pathogen composition and seasonal variation. Stacked area plot shows proportion of each pathogen. Black vertical lines separate periods; white dashed lines indicate seasonal boundaries. Heat bar: total seasonal serologically positive cases. (C) serological positivity rates by period for all pathogens combined and for each pathogen. Significance codes: ns, p > 0.05 (not significant); *, p < 0.05; **, p < 0.01; ***, p < 0.001

Age-specific distribution and shifts in serological positivity patterns

A strong negative correlation was observed between patient age and pathogen serological positivity rate across all periods (pre-COVID-19: ρ = −0.833; NPIs: ρ = −0.823; post-NPIs: ρ = −0.604; all p < 0.001), indicating a higher serological positivity burden among younger populations (Fig. 2A–C). Children aged 5–9 years consistently exhibited the highest serological positivity rates. Although the 15–19 years group showed relatively stable serological activity descriptively, its adjusted overall positivity did not differ significantly from the 28 days-4 years reference group across the three study phases (Fig. 3B; Table S2).

Fig. 2.

Age-specific distribution of respiratory pathogen serological positivity across study periods. (A-C) the number of tested cases, serologically positive cases, and serological positivity rates stratified by age group during the pre-COVID-19 (A), NPIs (B), and post-NPIs (C) periods. Wide bars: total number of tested cases; narrow bars: number of serologically positive cases; red line: serological positivity rate. The Spearman correlation coefficient (ρ) and P value shown for the association between age and serological positivity rate. (D-F) pathogen composition within age group during pre-COVID-19 (D), NPIs (E), and post-NPIs (F) periods

Fig. 3.

Age distribution and age-associated serological positivity patterns across study periods. (A) boxplots of age among serologically positive cases for each pathogen across study periods. Median age is labeled. Statistical significance between periods was assessed using the Kruskal-Wallis test followed by Dunn’s test; significance levels are shown as ns (p > 0.05, not significant), * (p < 0.05), ** (p < 0.01), and *** (p < 0.001). (B) forest plot of the association between age groups and overall serological positivity by study periods. Prevalence ratio (PR) with a 95% confidence interval (CI) was estimated by the modified Poisson regression model with robust variance and adjusted by sex, season, patient type, clinical department, and hospital (as a fixed effect). The reference group was children aged 28 days-4 years

MP dominated across all age groups, with the peak serological positivity rate in 5–9 years group (31.73%). During the NPIs period, PIV increased markedly, most notably among children aged 28 days to 4 years (from 2.95% to 17.96%, p < 0.001), surpassing FluB as the second most frequent pathogen. Conversely, FluB decreased substantially during NPIs, especially in children under 14 years. LP also declined consistently across age groups over time (Fig. 2D–F; Table S5).

The median age among serologically positive cases increased significantly from the pre-COVID-19 to NPIs period for most viruses: FluA increased from 3 years (IQR: 1–9) to 6 years (IQR: 2–43; p < 0.001), FluB from 3 years (IQR: 1–7) to 5 years (IQR: 1–50; p < 0.001), RSV from 0 years (IQR: 0–3) to 2 years (IQR: 0–7; p < 0.001), PIV from 3 years (IQR: 1–18) to 3 years (IQR: 1–7; p = 0.005), ADV from 2 years (IQR: 1–5) to 3 years (IQR: 1–13; p < 0.001), and MP from 4 years (IQR: 2–10) to 4 years (IQR: 2–10; p = 0.028). In contrast, LP decreased from 8 years (IQR: 3–41) to 6 years (IQR: 2–13; p < 0.001). The shift toward older age groups persisted into the post-NPIs period for FluB, PIV, ADV, and MP. Notably, QF, which had the fewest seropositive cases, showed a dramatic decrease from 42 years (IQR: 3–60) in 2019 to 4 years (IQR: 3–4) in 2023 (p = 0.001) (Fig. 3A). To further assess whether the observed age gradient remained after adjustment for other epidemiological factors, we performed multivariable modified Poisson regression analyses stratified by study period (Fig. 3B). The adjusted results were broadly consistent with the descriptive trends, but also revealed several additional features. In particular, the 15–19 years group did not differ significantly from the 28 days-4 years reference group after multivariable adjustment. In contrast, all adult groups showed lower positivity than the youngest children. Notably, this age-related difference became less marked in the post-NPIs period, especially among the oldest adults (≥80 years), who demonstrated the lowest PR (95% CI) of 0.36 (0.32–0.42) in pre-COVID-19, and 0.56 (0.52–0.61) in the post-NPIs period. Detailed pathogen-specific age-pattern curves are shown in Supplementary Material (Figure S1).

Department-specific positivity and pathogen dynamics

Analysis of case distribution across hospital departments revealed a consistent concentration of seropositive cases in Pediatrics (Peds) department throughout the study period, accounting for an average of 51.61% of serologically positive cases (Fig. 4). Serological positivity rates increased significantly from 2019 to 2023 in both Peds (36.56% to 40.59%, p < 0.001) and Pulmonary and Critical Care Medicine (PCCM) departments (19.74% to 27.24%, p < 0.001). More moderate but significant increases were further observed in Infectious Diseases (26.05% to 30.44%, p < 0.001) and Internal Medicine (22.14% to 23.48%, p < 0.001) (Table S2). This departmental pattern was primarily supported by a limited number of dominant pathogen trends. MP remained the principal contributor across departments and showed sustained increases in several major clinical settings, including Peds, PCCM, Infectious Diseases, and Internal Medicine. In addition, PIV increased markedly during the NPIs period and contributed substantially to the pediatric seropositive burden. Detailed pathogen-specific departmental trends are shown in Supplementary Material (Figure S2).

Fig. 4.

Department-specific distribution of respiratory pathogens. Sankey diagram: the distribution of serologically positive cases across study periods, clinical departments, and pathogens. Width proportional to composition ratio

Join-Point regression analysis of temporal trends

Join-Point regression analysis of quarterly data from 2019 to 2023 identified significant trend transition points for most pathogens, revealing the specific timing of changes associated with public health interventions (Fig. 5). FluB decreased sharply from Q3 2020 to Q2 2021 (quarterly percentage change [QPC] = −58.13%; p < 0.001) without recovery. In contrast, PIV increased dramatically from Q1 2019 to Q1 2021 (QPC = 38.24%, p < 0.001). ADV declined until Q2 2021 (QPC = −8.15%; p = 0.026), surged until Q1 2022 (QPC = 53.93%, p = 0.012), then decreased steadily (QPC = −21.68%, p = 0.007). Bacterial pathogens CP and LP showed early-pandemic declines from Q1 2019 to Q4 2021 (QPC = −16.88%, p = 0.002) and to Q2 2021 (QPC = −14.96%, p < 0.001), respectively. MP showed a modest overall upward trend across the study period (average quarterly percentage change [AQPC] = 1.34%, p = 0.007) with no significant joinpoints. No significant joinpoints were identified for FluA, likely due to substantial variability in its quarterly data. QF was excluded from analysis due to insufficient case numbers (Fig. 5A–I; Table S6).

Fig. 5.

Join-Point regression analysis of quarterly serological positivity rates for nine respiratory pathogens, 2019–2023. (A-I) quarterly serological positivity rates (red points) and fitted Join-Point regression trends (colored segments) for FluA (A), FluB (B), RSV (C), PIV (D), ADV (E), MP (F), CP (G), LP (H), and QF (I). QPC: quarterly percentage change for each segment; statistical significance: *, p < 0.05. Grey bars: the absolute number of quarterly serologically positive cases. No Join-Point analysis for QF due to insufficient sample size

Changing patterns of pathogen serological Co-detections

Although single-pathogen positivity predominated, serological co-detections (defined as the detection of IgM antibodies against two or more respiratory pathogens in the same serum sample) represented a significant and evolving component. The overall serological co-detection rate increased from 4.05% (2,512/62,091) in the pre-COVID-19 period to 4.86% (6,033/124,130) during the NPIs period (p < 0.001), remaining elevated at 4.82% (2,622/54,390) in the post-NPIs period. Double co-detections constituted the most frequent co-detection type (Table S3).

Viral pathogens (FluA, FluB, PIV, and ADV) were more frequently involved in co-detections (Table S5). The proportion of higher-order serological co-detections (involving three or more pathogen-specific IgM antibodies) increased during the NPIs period for FluA, FluB, RSV, ADV, MP, and LP. For RSV and ADV, these higher-order co-detections accounted for a larger proportion of co-detection events than double co-detections during the NPIs period. In contrast, PIV and QF showed decreasing trends of higher-order co-detections (Fig. 6A).

Fig. 6.

Dynamics of pathogen co-detection patterns across study periods. (A) distribution of co-detection types (double, triple, quadruple, and quintuple+ co-detections) for each pathogen in different periods. (B-D) UpSet plots illustrating the frequent co-detection combinations (with a frequency ≥ 10) during the pre-COVID-19 (B), NPIs (C), and post-NPIs (D) periods. Connected dots indicate pathogen co-detection combinations; vertical bars: frequency of each combination; horizontal bars: number of co-detection events for each pathogen. Colors indicate different co-detection types

The predominant co-detection combinations shifted markedly. Before the pandemic, FluB-MP was the most common pair, accounting for 38.69% (972/2,512). During the NPIs period, PIV-MP became dominant (52.88%, 3,190/6,033; p < 0.001), while FluB-MP declined markedly to 8.40% (507/6,033, p < 0.001). This period also saw the emergence of higher-order combinations among the top patterns, such as RSV-PIV-MP and PIV-ADV-MP. In the post-NPIs period, the dominance of PIV-MP was further reinforced (59.12%, 1,550/2,622), while FluA-MP increased markedly (9.12%, 239/2,622), replacing FluB-MP as a major co-detection pattern (Fig. 6B–D; Table S7).

Discussion

This multicenter study delineates the profound impact of NPIs on the epidemiology of nine respiratory pathogens in Southwest China from 2019 to 2023. Our findings reveal significant shifts in pathogen serological positivity/exposure patterns, age distribution, departmental burden, and co-detection patterns, underscoring the lasting ecological consequences of public health measures in this distinct region.

The relatively high serological positivity of Mycoplasma pneumoniae (MP) IgM observed in this study should be interpreted with caution in the context of known limitations of serological assays. First, MP-specific IgM may persist for weeks to months after acute infection, and in some cases even longer, particularly in pediatric populations [26]. This “antibody tailing” effect suggests that IgM positivity does not necessarily reflect an ongoing acute infection and may lead to overestimation of the true acute infection prevalence. Second, part of this sustained MP-IgM signal may have been influenced by non-specific serological reactivity or polyclonal IgM responses induced by other common pediatric infections such as Epstein-Barr virus and Cytomegalovirus, which could not be systematically excluded in this retrospective dataset. Nevertheless, our annual analysis (Table S4) further revealed a transient suppression of MP in 2020 followed by a rapid rebound in 2021, which aligns with the regional endemicity reported in Southwest China (in Chengdu and Chongqing) [25, 27]. Therefore, the persistently high MP positivity likely reflects a combination of prolonged antibody persistence, background seropositivity, and genuine regional circulation, rather than definitive evidence of uninterrupted acute MP predominance.

In contrast, most viral pathogens were suppressed during the NPIs period, consistent with global observations [7, 9]. However, their post-NPIs trajectories diverged. The rebound of FluA and RSV in 2023 aligns with the classic “immunity debt” concept [10, 28], whereas the sustained suppression of FluB may reflect the global diminution of B/Yamagata lineage [29, 30], viral interference, or a more profound loss of population immunity [31, 32]. Bacterial pathogens CP, LP, and QF also remained suppressed, suggesting prolonged NPIs effects on their transmission cycles [33, 34]. The low IgM detection rates observed for some respiratory viruses (e.g., IAV and RSV), particularly in young children, do not necessarily indicate low epidemiological importance. For these viral pathogens, molecular assays are generally more suitable for acute diagnosis, whereas single-serum IgM may miss early infections or perform suboptimally in reinfection settings [35]. Because repeated exposure and reinfection are common for RSV and influenza, the subsequent IgM response may be blunted or absent. Furthermore, because PCR and serology capture different diagnostic windows, single-serum IgM may still provide supportive retrospective information on recent exposure in real-world clinical settings where patient presentation timing varies. Accordingly, the low IgM positivity of these viral pathogens should be interpreted as a limitation of serology-based detection rather than as evidence against their circulation.

A distinctive finding from Southwest China was the anomalous surge of PIV during the NPIs period, potentially explained by “virus-virus interference” theory [36, 37] where suppression of dominant viruses (e.g., RSV and FluA) may have created an ecological vacuum for PIV. Local environmental factors, such as specific meteorological conditions (e.g., temperature and relative humidity) conducive to PIV transmission, may have also modulated this unusual activity [38, 39].

Our analysis further revealed not only age-related shifts over time, but also a strong pediatric-driven structure of the dataset itself. Because children younger than 4 years accounted for nearly half of the tested population, the overall serological patterns observed in this study were strongly influenced by pediatric testing dynamics. Against this background, the strong negative correlation between age and pathogen serological positivity underscores the concentration of serological positivity in younger populations. The upward shift in the median age among serologically positive cases for several major viral pathogens during the NPIs period may reflect delayed primary exposure and the accumulation of susceptible individuals in older children and adults [11, 40]. In addition, the RSV increase observed in individuals aged ≥ 80 years remains epidemiologically relevant and may have implications for elderly vaccination strategies [41].

Departmental analysis showed a concentrating burden in Pediatrics (Peds) and Pulmonary and Critical Care Medicine (PCCM) departments in post-NPIs, reflecting both high pediatric susceptibility and an increase in severe cases requiring specialized care, potentially straining resources.

The evolution of serological co-detection patterns further suggests that the respiratory pathogen profile changed across the three study phases. The increase in the overall serological co-detection rate during the NPIs period, together with the replacement of FluB-MP by PIV-MP as the most frequently observed co-detection pattern, indicates that the overlap of pathogen-specific serological signals shifted after the implementation of public health interventions. However, these findings must be interpreted with strict microbiological caution. Because the present analysis was based on single-serum IgM detection, serological co-detection does not indicate simultaneous acute multi-pathogen infection. Rather, these overlapping signals likely reflect a combination of independent antibody kinetics, recent sequential exposures, the prolonged persistence of MP-IgM, and non-specific polyclonal cross-reactivity. In particular, the emergence of complex higher-order co-detections should not be overinterpreted as evidence of direct biological synergy or concurrent viral-bacterial interaction. Nevertheless, the redistribution of these dominant co-detection patterns remains epidemiologically informative. It reflects broader macro-level changes in age-related exposure structures, seasonal circulation disruptions, and the relative shifting contributions of viral and atypical bacterial serological activity across the different pandemic phases.

Join-Point regression provided precise temporal insights into the impact of NPIs: the sharp FluB decline without recovery contrasted with the dramatic PIV increase, illustrating the heterogeneous effects of the NPIs period [8, 9]. For MP, the absence of significant quarterly joinpoints suggests that the serological series did not show abrupt inflection points at the quarterly level. However, this finding should be interpreted together with the annual descriptive analysis (Table S4) and the known limitations of MP-IgM serology.

Our study has several important limitations that must be acknowledged. First and foremost, our reliance on single-serum IgM detected by indirect immunofluorescence, rather than molecular confirmation (PCR) or paired acute-convalescent sera, means our data represent recent serological exposures rather than definitive acute etiologies. This methodology is subject to bidirectional biases: it may underestimate early acute infections due to delayed seroconversion and blunted responses in viral re-infections, while simultaneously overestimating the apparent acute burden due to long-term antibody persistence and non-specific cross-reactivity from infections like Epstein-Barr virus and Cytomegalovirus [42, 43]. Furthermore, serology inherently lacks pathogen subtyping capability [44]. Second, because our data were derived from care-seeking populations at tertiary hospitals, this introduces a selection bias toward symptomatic or more severe cases, which may not perfectly reflect asymptomatic or mild community circulation. Third, the lack of comprehensive vaccination histories and pathogen genetic data limits our ability to explore the mechanistic drivers behind certain epidemiological shifts. Fourth, the tested pathogen panel was restricted to the nine pathogens routinely included in the clinical IgM assay and did not include several important respiratory viruses, especially SARS-CoV-2 and Enterovirus/Rhinovirus. Therefore, the present study does not provide a complete picture of the full respiratory pathogen landscape in Southwest China.

Although absolute IgM positivity rates may be influenced by methodological limitations, the use of a consistent assay platform and standardized procedures across all centers and time periods suggests that such biases are likely systematic and stable. Therefore, the observed temporal trends are less likely to reflect diagnostic artifacts and more likely to represent relative changes in population-level exposure dynamics, supporting the internal validity of the longitudinal comparisons in this study. Accordingly, while absolute IgM positivity rates may be overestimated, the temporal variations in serological positivity are more likely to reflect relative changes in population-level exposure patterns rather than fluctuations in assay performance or diagnostic bias. Unlike PCR, which detects nucleic acids in a narrow window, IgM reflects the host-specific immune response and may provide supportive retrospective information on recent exposure in real-world clinical settings where patient presentation timing varies. Taken together, these findings provide a unique, large-scale regional “sero-epidemiological map” that serves as a necessary baseline for future prospective validation using molecular tools. Importantly, this study does not aim to replace molecular diagnostics, but rather to provide a complementary sero-epidemiological perspective on long-term population-level exposure patterns.

Conclusions

In conclusion, this study provides a large-scale regional sero-epidemiological profile of recent respiratory pathogen exposure across the three pandemic phases in Southwest China. Our findings suggest substantial temporal shifts, strong pediatric predominance, and evolving serological co-detection patterns following COVID-19 NPIs. However, because serology alone cannot fully resolve the complexity of acute respiratory pathogen ecology, these observations should be interpreted cautiously within the methodological boundaries of single-serum IgM-based surveillance, and should be viewed as an important regional baseline that supports future surveillance that integrates molecular multiplex detection and vaccination data, thereby informing public health policy in the post-COVID-19 era.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviation

ARTIs

Acute Respiratory Tract Infections

NPIs

Non-Pharmaceutical Interventions

FluA

Influenza A virus

FluB

Influenza B virus

RSV

Respiratory Syncytial Virus

PIV

Parainfluenza Virus

ADV

Adenovirus

MP

Mycoplasma pneumoniae

CP

Chlamydia pneumoniae

LP

Legionella pneumophila

QF

Q Fever

HIS

Hospital Information System

Peds

Pediatrics

PCCM

Pulmonary and Critical Care Medicine

ID

Infectious Diseases

ED

Emergency Department

ICUs

Intensive Care Units

IM

Internal Medicine

SUR

Surgery

IQR

Interquartile Range

QPC

Quarterly Percentage Change

AQPC

Average Quarterly Percentage Change

PR

Prevalence Ratio

Authors’ contributions

Y.X. and L.Z. designed and coordinated the study. L.Z., L.L., M.D., X.X., J.H., C.T., Y.L., Y.H., C.W., Y.D, W.H., D.H., L.J., B.N., Y.L., and X.L. collected the data. L.Z. and Y.H. analyzed the results and wrote the manuscript.

Funding

This work was supported by the 1. 3. 5 project for disciplines of excellence from West China Hospital of Sichuan University [ZYGD23036] and a horizontal research project fund: Epidemiological Investigation of Respiratory Pathogens in Southwest China and a Study on the Benefits of Different Combination Strategies[ NO.311231682]. The funding agencies had no involvement in the study design, data collection, analysis, interpretation of results, manuscript drafting, or the decision to submit the article for publication.

Data availability

The aggregate data and statistical results that support the findings of this study are available within the article and its supplementary information. The underlying patient-level data are not publicly shared to protect patient confidentiality but can be made available from the corresponding author upon reasonable request and with permission from the participating institutions.

Declarations

Ethics approval and consent to participate

Ethical approval for this study was obtained from the ethics committees of all participating centers: West China Hospital, Sichuan University: 2023–47. Affiliated Hospital of Youjiang Medical College for Nationalities: YYFY-LL-2025–303. Liuzhou Workers’ Hospital: KY2022309. Kunming Children’s Hospital: 2023–03-112-K01. The Second People’s Hospital of Yibin: 2023–231-01. Suining Central Hospital: KYLLKS20230140.

Consent for publication

Not Applicable.

Competing interests

The authors report there are no competing interests to declare.