Association between long-term PM2.5 exposure and risk of urinary tract infection in children: a nationwide longitudinal cohort study Authors: Jung Eun Choi, Hyemin Jang, Whanhee Lee, Jongmin Oh, Hae Soon Kim, Ji Hyen Lee, Jung Won Lee, Eunhee Ha Published: 2026 ABSTRACT AbstractBackgroundUrinary tract infection (UTI) is among the most common bacterial infections in children and can lead to both acute and chronic complications. Although fine particulate matter (PM2.5) is a known risk factor for cardiovascular, respiratory, and renal diseases, its role in pediatric urinary tract infections remains unclear; therefore, we examined the association between long-term PM2.5 exposure and UTI incidence in children using a nationwide longitudinal cohort.MethodsWe analyzed data from 395,032 children aged 0–19 years in the National Health Insurance Service–National Sample Cohort from 2002 to 2019. Annual average PM2.5 exposure was estimated at 1 km2 resolution using validated machine-learning ensemble models and assigned to residential districts. UTI and antibiotic-treated UTI were identified using ICD-10 codes and prescription records. Cox proportional-hazards models with time-varying exposures and covariates were used to estimate hazard ratios (HRs) with 95% confidence intervals (CIs). Subgroup and sensitivity analyses were conducted to assess effect modification and robustness.ResultsDuring follow-up, 28,129 UTI and 14,128 antibiotic-treated UTI cases occurred. A 5 µg/m3 increase in annual PM2.5 was associated with a higher risk of UTI (HR: 1.05, 95% CI: 1.02–1.08) and antibiotic-treated UTI (HR: 1.07, 95% CI: 1.03–1.11). The association between PM2.5 exposure and UTI risk was stronger among children living in rural areas. Our findings were consistent in sensitivity analyses, with 1-year lag exposure showing stronger associations.ConclusionLong-term exposure to PM2.5 was associated with an increased risk of both UTI and antibiotic-treated UTI in children. These findings suggest that air pollution may contribute to pediatric UTI risk through mechanisms involving systemic inflammation, oxidative stress, and immune dysregulation.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12889-026-28077-7.Keywords: Air pollution, Children, Cohort, Particulate matter, Urinary tract infection Introduction Urinary tract infection (UTI) is one of the most common causes of bacterial infections in the pediatric age group, affecting 3% to 7.5% of febrile children each year in both community and hospital settings. It has a burden on the child and parents that can lead to short-term complications, such as urosepsis and acute renal failure. UTI can lead to long-term complications, such as scarring of the kidney, hypertension, and even end-stage renal disease (ESRD) [1]. Air pollution is a well-known risk factor for adverse health outcomes and increased mortality, particularly due to its effects on the cardiovascular and respiratory systems [2]. Among various pollutants, fine particulate matter (PM2.5) has emerged as a significant environmental concern because of its ability to penetrate deep into the lungs and enter systemic circulation [2–5]. PM2.5 contains various harmful components, including dust, organic acids, heavy metals, and soil particles, some of which act as endocrine-disrupting chemicals and can compromise immune function [6]. These particles are known to induce systemic inflammation, oxidative stress, and adverse effects across multiple organ systems. Although most research on the harmful effects of air pollution has focused on cardiovascular and respiratory diseases, more recently, growing evidence has suggested a link between air pollution and renal dysfunction [7–12]. A recent systematic review also reported that long-term exposure to PM2.5 is associated with increased incidence and prevalence of chronic kidney disease (CKD) [13]. Several epidemiological studies have demonstrated that PM2.5 exposure is associated with decreased estimated glomerular filtration rate (eGFR), ESRD, and hospitalization due to acute renal injury [14–16]. These associations are believed to be mediated through biologically plausible mechanisms, including oxidative stress, immune dysregulation, and inflammation [17]. Previous studies suggest that PM2.5 exposure may increase susceptibility to both viral and bacterial infections by exerting immunotoxic effects [18]. Children may be especially vulnerable to the health effects of air pollution due to their narrower airways, higher respiratory rates, and immature immune systems [19]. Both short- and long-term exposure to PM2.5 has been associated with increased risk and severity of infectious diseases, particularly upper respiratory tract infections [20]. In addition to respiratory illnesses, PM2.5 has been implicated in non-respiratory infections, such as UTIs and gastrointestinal disorders. However, the current evidence on these associations remains limited and inconsistent [21, 22]. Given growing concerns about the impact of air pollution on child health, this study aimed to investigate the association between long-term PM2.5 exposure and the risk of UTI in a nationwide longitudinal cohort of Korean children. Methods Study design and population The Korean National Health Insurance Service (NHIS) is a mandatory national healthcare system that covers 97% of the Korean population. It provides the NHIS–National Sample Cohort (NHIS–NSC) for research purposes, a population-representative sample comprising 2.2% of the total eligible South Korean Population from 2002 to 2019 [23]. Using the NHIS–NSC, we constructed a pediatric cohort comprising 428,135 individuals aged 0–19 years. Participants were followed from cohort entry until the first diagnosis of UTI (or antibiotic-treated UTI), death, the end of the calendar year in which they reached 19 years of age, or the end of the study period, whichever came first. Our cohort comprises four databases: qualification (e.g., age, sex, residence area), birth and deaths, medical treatments, and prescriptions [24]. In this study, we used these databases to identify demographic characteristics of the study participants and treatment information for UTI in children. Among 428,135 beneficiaries aged 0–19 at enrollment, we excluded 29,492 individuals enrolled after 2016 to ensure a minimum follow-up period of 4 years (2016–2019). We also excluded 3,611 individuals who lived in island areas such as Jeju and Ulleung, because our air pollution data covered only inland regions of South Korea. Therefore, a total of 395,032 individuals were included and followed up from 2002 to 2019 (Fig. 1). Flow diagram of included study subjects. Abbreviation: UTI, urinary tract infection; NHIS–NSC, National Health Insurance Service–National Sample Cohort The Institutional Review Board of the Ewha Womans University Seoul Hospital (IRB number: SEUMC 2022-05-034) approved the study. The requirement for informed patient consent was waived because the NHIS–NSC database is anonymized in accordance with strict confidentiality guidelines. Case definition In this study, we defined two UTI cases: UTI and antibiotic-treated UTI. First, we defined the onset of UTI using the corresponding primary diagnosis (International Classification of Diseases, 10th revision [ICD-10]) codes in the medical treatment database, namely N390 and N10, to capture hospital visits with any UTI onset. For the cases, we used the first hospital visit for each beneficiary with UTI diagnostic codes as the UTI onset. In addition, we defined antibiotic-treated UTI cases using prescription records and the medical treatment database. To address antibiotic-treated UTI cases that required treatment with medication, we defined the onset of the antibiotic-treated UTI for each participant as the first hospital visit, including both inpatient admissions and outpatient visits, with the primary diagnostic codes for UTI (N390 and N10) and antibiotic prescriptions (both oral and intravenous) recorded on the concurrent claim record. Lower UTI codes such as N30 (cystitis) were excluded to enhance diagnostic specificity in the claims-based outcome definition. Air pollutant data Because nationwide ground-based monitoring of PM2.5 in South Korea has been available since 2015, exposure data covering the entire study period were not available from regulatory monitoring networks [25]. Therefore, we relied on modeled air pollution data to assess long-term PM2.5 exposure. We used a nationwide air pollution data set that includes high-resolution PM2.5 estimates from machine-learning ensemble prediction models developed by the AiMS-CREATE network. Specifically, predicted monthly concentrations of ambient levels of PM2.5 at 1-km2 spatial resolution across the inland South Korea were obtained from a well-validated (cross-validated R2 for PM2.5: 0.87; for NO2: 0.75, which is used for sensitivity analysis) published model, and these predictions were estimated using ensemble models incorporating random forest, gradient boosting, and neural networks [26]. More information on the overall prediction model is provided in the appendix (p. 1). Monthly PM2.5 estimates were averaged to derive annual mean concentrations for each study year. We used these data to calculate the annual mean PM2.5 concentration for each administrative district (si/gun/gu), which represents the finest spatial resolution of residential information available in the NHIS-NSC, and to determine centroid points within each district’s boundaries for each year. Annual-average PM2.5 concentrations were then assigned to individuals based on their district of residence, with exposure updated longitudinally from baseline according to changes over follow-up. Covariates We considered confounding variables that could affect the association between PM2.5 and onsets of UTI in children. First, we used three individual-level variables from the NHIS-NSC files: age at cohort entry, sex, and household income status in three categories (low: decile 0–2, where decile 0 refers to recipients of medical aid; mid: decile 3–7; or high: decile 8–10). Household income status was treated as a time-varying variable according to calendar year. To address the potential confounding from regional differences, we also collected district-level indicators that include current smoking (%), number of beds in hospital per 1,000 people, untreated rate of medically necessary services, population living in developed areas (%), number of recipients of basic living per 1,000 people, and park areas (%). These regional variables were obtained from the Korea Community Health Survey, which is a cross-sectional health survey conducted by the Korean Center for Disease Control and Prevention since 2008 [27]. Additionally, to adjust for potential residual confounding from spatial and temporal trends, we included indicator variables for each region (Capital, Mid, and South) and each calendar year [22]. Statistical Analyses The proportional hazards assumption was assessed using Schoenfeld residual tests, and no substantial violation was identified. We then performed a Cox proportional-hazards model with time-varying exposures and covariates to estimate the association between time-varying annual exposure (i.e., current year) to PM2.5 and UTI onsets (for the first hospital visit for UTI or antibiotic-treated UTI, separately). All results are presented as hazard ratios (HRs) and the corresponding 95% confidence intervals (CIs) per 5 µg/m3 increase in PM2.5 [22, 28]. As mentioned above, all individual- and regional-level covariates were included in the model. There were no missing data for the variables included in the analysis. All statistical analyses were performed using SAS® version 9.4 (SAS Institute, Inc., Cary, NC, USA). In addition, to identify subpopulations with higher vulnerability, we did subgroup analyses. We evaluated potential effect modification by sex (male or female), income (low: decile 0–2, where decile 0 refers to recipients of medical aid; mid: decile 3–7; or high: decile 8–10) and urbanicity (urban: Gyeonggi-do and metropolitan cities—Seoul, Incheon, Daejeon, Sejong, Busan, Daegu, Gwangju, Ulsan; or rural: other areas). Finally, to assess any potential deviations from linearity in the concentration-response curve, we replaced the linear term of PM2.5 in the main model with a constrained thin plate regression spline function using the generalized additive model [29]. Sensitivity analysis We performed sensitivity analyses to examine whether our main results were robust to the selection of confounders, including the inclusion of older participants, the exposure time window, and potential outcome misclassification. First, analyses were performed using a broader definition of UTI outcome, which included both primary and secondary ICD-10 diagnoses, to capture a wider range of clinically relevant UTI cases. Second, a 1-year lag in PM2.5 exposure was applied to evaluate the association between antecedent exposure and subsequent UTI hospitalization, accounting for the temporal sequence between exposure and outcome. Third, we excluded community-level indicators from the main model. Lastly, we adjusted the annual average NO2 in the main model to investigate the potential confounding effect of a secondary pollutant. Results Descriptive statistics The total cohort dataset included 395,032 beneficiaries aged 0–19 years living in 226 districts across South Korea. Over the 18-year study period (2002–2019), we identified 28,129 first hospital visits for UTI and 14,128 cases of antibiotic-treated UTI, defined as UTI accompanied by antibiotic prescription. Baseline cohort characteristics are presented in Table 1. Among participants who developed UTI or antibiotic-treated UTI, the median time from baseline to disease onset was 4 years. On average, children were 6.7 years old at enrollment, with slightly more males (52.2%) than females (47.8%). Consistent with previous findings, UTI cases were more frequent in females than in males. Table 1Descriptive cohort characteristics for NHIS–NSC beneficiaries from 2002 to 2019CategoriesFull Cohorta (n = 395,032)UTI (n = 28,129)Antibiotic-treated UTI(n = 14,137)Air pollutant concentrationPM2.5 (µg/m3)29.7 ± 3.127.6 ± 3.127.5 ± 3.1Potential confoundersAge Years6.7 ± 6.55.6 ± 5.34.8 ± 5.0Sex Male206,254(52.2)12,854(45.7)6,276(44.4) Female188,778(47.8)15,275(54.3)7,861(55.6)Income status Low (decile 0–2)b31,261(8.2)2,725(9.7)1,346(9.5) Mid (decile 3–7)201,862(53.1)13,396(47.6)6,937(49.1) High (decile 8–10)146,819(38.6)12,008(42.7)5,854(41.4)Areasc Capital189,539(48.1)14,337(51.0)6,997(49.5) Mid53,470(13.6)3,940(14.0)1,923(13.6) South150,695(38.3)9,852(35.0)5,217(36.9)Community-level indicatorsCurrent smokers(%)25.6 ± 2.624.1 ± 2.824.0 ± 2.8Number of beds in hospital per 1,000 population10.4 ± 5.110.9 ± 5.811.0 ± 5.8Annual untreated rate of medically necessaryservices(%)13.0 ± 4.113.2 ± 4.113.2 ± 4.1% of population living in developed areas (each district)89.8 ± 18.591.1 ± 16.290.7 ± 16.5Number of recipients of basic living per 1,000 population33.0 ± 17.432.3 ± 15.632.5 ± 15.4Park area per person18.6 ± 22.417.4 ± 19.517.6 ± 19.4Abbreviation: UTI Urinary Tract Infection, NHIS–NSC National Health Insurance Service–National Sample Cohort, PM2.5 particulate matter < 2.5 μmValues are mean ± standard deviation, or n (%)a Descriptive characteristics were counted at enrollment for the entire cohort, and at onset for patients with UTIb Decile 0 of income refers to recipients of medical aidc Areas were divided by their approximate latitudes: Capital (Seoul, Incheon, Gyeonggi), Mid (Daejeon, Sejong, Chungcheong, Gangwon), and South (Busan, Daegu, Gwangju, Ulsan, Gyeongsang, Jeolla); missing counts are for Jeju Island, where air pollution is not projected Descriptive cohort characteristics for NHIS–NSC beneficiaries from 2002 to 2019 Annual untreated rate of medically necessary services(%) Abbreviation: UTI Urinary Tract Infection, NHIS–NSC National Health Insurance Service–National Sample Cohort, PM2.5 particulate matter < 2.5 μm Values are mean ± standard deviation, or n (%) a Descriptive characteristics were counted at enrollment for the entire cohort, and at onset for patients with UTI b Decile 0 of income refers to recipients of medical aid c Areas were divided by their approximate latitudes: Capital (Seoul, Incheon, Gyeonggi), Mid (Daejeon, Sejong, Chungcheong, Gangwon), and South (Busan, Daegu, Gwangju, Ulsan, Gyeongsang, Jeolla); missing counts are for Jeju Island, where air pollution is not projected Regarding socioeconomic distribution, 42.7% of UTI cases occurred among high-income families, whereas only 9.7% occurred among the lowest-income group. Geographically, the capital region (Seoul, Incheon, and Gyeonggi) accounted for nearly half of the UTI (51.0%) and antibiotic-treated UTI cases (49.5%), followed by the southern provinces. At the community level, UTI cases were observed in districts with slightly higher hospital bed density and a somewhat lower proportion of smokers, though these differences were modest. Air pollution exposure The geographical distributions of annual-average PM2.5 concentrations and UTI onsets are shown in Fig. 2. During the study period, the mean annual PM2.5 concentration was 29.7 µg/m3, substantially above the WHO recommended guideline of 5 µg/m3. In general, the capital (Northwest) and Southeast regions showed the highest number of UTI onsets and the highest PM2.5 concentrations. Geographical distributions of annual average PM2.5 concentrations and incidence of urinary tract infection (UTI) and antibiotic-treated UTI in South Korea. A PM2.5 concentration (µg/m3) (B) Onset of UTI (C) Onset of antibiotic-treated UTI. Abbreviation: UTI, urinary tract infection; PM2.5, fine particulate matter Associations between PM2.5 exposure and UTI risk The exposure–response curves showed an approximately linear increase in risk below the mean PM2.5 concentration (29.7 µg/m3), whereas above this level the hazard ratio exhibited a slight decline followed by a sharp increase. Overall, the curves did not demonstrate a strictly linear relationship but rather a non-linear pattern, with a steeper increase at higher PM2.5 concentrations and wide confidence intervals, suggesting potential influence of sparse data or outliers in the upper exposure range (Fig. 3). Results from the Cox proportional hazards model are summarized in Table 2; Fig. 4. For every 5 µg/m3 increase in annual PM2.5, the hazard ratio (HR) was 1.05 (95% CI: 1.02–1.08) for UTI onset and 1.07 (95% CI: 1.03–1.11) for antibiotic-treated UTI. In general, a positive association between PM2.5 and UTI onset was observed across most subpopulations (Fig. 4). For UTI onset, the association was more prominent in females (HR: 1.06; 95% CI: 1.02–1.10), beneficiaries with low household income (HR: 1.09; 95% CI: 1.00–1.20), and those living in rural areas (HR: 1.16; 95% CI: 1.09–1.23), compared to the result of the total population. These patterns were similar across antibiotic-treated UTI cases; however, beneficiaries with high household income had the most significant HR for antibiotic-treated UTI (HR: 1.16; 95% CI: 1.09–1.24) (Table 2; Fig. 4). These subgroup differences should be interpreted with caution, as they may reflect differences in healthcare utilization and treatment recording rather than true biological heterogeneity. Urban–rural comparisons further highlighted disparities: children residing in rural areas experienced a higher risk of UTI onset (HR: 1.16; 95% CI: 1.09–1.23) than those in urban settings, for whom the association was null (HR: 0.99; 95% CI: 0.96–1.03). For antibiotic-treated UTI, the rural effect remained significant (HR 1.10; 95% CI: 1.02–1.20). These findings suggest that rural children may be particularly susceptible, potentially due to differences in baseline health, exposure patterns, or access to medical care. The exposure-response association between PM2.5 exposure and urinary tract infection. Note: All models were adjusted for age, sex, household income, urbanicity, region, smoking prevalence, hospital bed density, untreated service rate, basic living recipients, and park area. Abbreviation: PM2.5, fine particulate matter Table 2Association between long-term PM2.5 exposure per 5 µg/m3 increase and first hospital visit for UTI in specific subpopulationsCategoriesHR (95% CI)UTIAntibiotic-treated UTITotal1.05 (1.02–1.08)1.07 (1.03–1.11)Sex Male1.04 (1.00–1.09)1.08 (1.02–1.14) Female1.06 (1.02–1.10)1.06 (1.01–1.12)Income-level Low (decile 0–2)1.09 (1.00–1.20)1.04 (0.91–1.18) Mid (decile 3–7)1.02 (0.98–1.07)1.00 (0.94–1.06) High (decile 8–10)1.07 (1.02–1.12)1.16 (1.09–1.24)Urbanicity Urban0.99 (0.96–1.03)1.04 (0.99–1.10) Rural1.16 (1.09–1.23)1.10 (1.02–1.20)Abbreviations: UTI Urinary Tract Infection, PM particulate matter <2.5 μm, HR hazard ratio, CI confidence interval Association between long-term PM2.5 exposure per 5 µg/m3 increase and first hospital visit for UTI in specific subpopulations Abbreviations: UTI Urinary Tract Infection, PM particulate matter <2.5 μm, HR hazard ratio, CI confidence interval Association between long-term PM2.5 increases of 5 µg/m3 and first hospital visit for urinary tract infection in specific subpopulations. Note: The shaded gray band represents the lower and upper confidence interval bounds for the hazard ratio estimates in the overall group. Abbreviation: UTI, urinary tract infection; HR, hazard ratio; CI, confidence interval; PM2.5, fine particulate matter Sensitivity analyses Finally, sensitivity analyses confirmed the robustness of our results (Table S1). When UTI was defined more broadly to include secondary diagnostic codes, the associations persisted but were slightly attenuated (UTI, HR: 1.02; 95% CI: 0.99–1.04). Using a 1-year lag exposure model produced stronger associations, particularly for antibiotic-treated UTI (HR: 1.05; 95% CI: 1.01–1.10), suggesting that cumulative or delayed effects of PM2.5 may be important in disease development. Models that excluded regional indicators or additionally adjusted for NO2 exposure yielded results consistent with the primary analyses, indicating that residual spatial confounding and co-pollutant effects were unlikely to explain the observed associations fully. Discussion This study investigated the association between long-term exposure to PM2.5 and the incidence of UTI in children, using a nationwide longitudinal cohort of 395,032 individuals aged 0 to 19 years, followed from 2002 to 2019. Exposure to higher levels of PM2.5 was associated with an increased risk of UTI in children. This association was observed for both overall UTI incidence and for more severe cases requiring antibiotic treatment, suggesting that long-term exposure to PM2.5 may increase vulnerability to bacterial infections beyond the respiratory system. Previous studies have attempted to establish an association between PM2.5 and kidney diseases. However, to date, no study has specifically examined its relationship with pediatric UTIs. Existing research has predominantly focused on CKD, acute kidney injury (AKI), and alterations in kidney function markers, such as eGFR, in relation to PM2.5 exposure. For instance, several cohort and cross-sectional studies have reported that long-term exposure to ambient PM2.5 is associated with increased CKD incidence and progression, and with a higher risk of hospital admissions for AKI in older adults. For example, a cohort study in the United States found a significant association between long-term exposure to air pollution and hospital admissions for AKI, even at lower levels of exposure [22]. Similarly, a time-series analysis conducted in China demonstrated that short-term exposure to PM2.5 and PM10 was significantly associated with increased urinary disease morbidity, with the most substantial effect observed over a cumulative exposure period of 11 days [17]. In Korea, a nationwide cross-sectional study among adults reported that long-term exposure to PM10 and NO2 was significantly associated with decreased eGFR, indicating impaired renal function, although the association with CKD was not significant after adjustment for covariates [30]. A cross-sectional study among Chinese women found that household air pollution was significantly associated with an increased risk of urinary incontinence, suggesting that indoor air pollution exposure is linked to impaired bladder function [31]. A recent study based on the UK Biobank cohort of approximately 375,000 adults reported a strong association between long-term PM2.5 exposure and UTI incidence, with a hazard ratio of 1.43 per 5 µg/m3 increase and a non-linear exposure–response relationship [9]. Our findings are generally consistent with previous findings supporting an association between long-term PM2.5 exposure and UTI risk. The relatively smaller effect estimates observed in our pediatric population may reflect differences in age distribution, biological susceptibility, exposure distribution, and population characteristics between children and adults, as well as differences in outcome definitions and exposure assessment methods. Several potential biological mechanisms may explain how PM2.5 exposure increases susceptibility to UTI. First, systemic pathways have been proposed. Inhaled PM2.5 can be absorbed into the bloodstream, leading to systemic inflammation, oxidative stress, and immune dysregulation. Studies have shown that PM can induce oxidative stress and systemic inflammation upon entering the respiratory tract [32–34]. Specifically, PM2.5 exposure has been reported to impair macrophage phagocytic activity, reduce mucosal barrier integrity, and dysregulate cytokine production, thereby increasing susceptibility to infections [32]. These immune alterations may compromise the host’s ability to clear uropathogens efficiently, increasing vulnerability to bacterial infections, including UTIs [3]. Second, renal-specific pathways have also been suggested. PM2.5 may activate the renin–angiotensin–aldosterone system and the bradykinin cascade, promote oxidative stress and DNA damage in renal tissues, and ultimately contribute to autophagy dysfunction and fibrosis. Such renal-specific mechanisms provide a biologically plausible link between air pollution exposure and increased UTI risk. However, recent findings on these mechanisms have been inconsistent, highlighting the need for further investigation [13, 35]. Third, experimental evidence from animal studies suggests that fine particulate matter can translocate from the respiratory system into systemic circulation and accumulate in renal tissue [36]. Sub-chronic PM2.5 exposure has been associated with renal inflammation and tissue injury, and nanoparticle tracing studies have demonstrated direct renal deposition of inhaled particles [37]. Such renal involvement may plausibly increase susceptibility to urinary tract infections by impairing local immune responses or renal barrier function. We observed differences in the association between long-term PM2.5 exposure and UTI by urbanicity, with higher hazard ratios observed in rural areas than in urban areas (Table 2; Fig. 4). However, these subgroup findings should be interpreted cautiously. The observed differences may reflect regional variation in healthcare access, diagnostic practices, antibiotic prescribing behavior, socioeconomic conditions, and other unmeasured contextual factors rather than true biologic susceptibility. One potential explanation is that the dominant sources and chemical composition of PM2.5 vary between urban and rural environments. Urban PM2.5 is primarily influenced by traffic- and industrial-related combustion sources, whereas rural PM2.5 more often reflects contributions from soil and agricultural dust, biomass burning, and agricultural activities. Differences in PM2.5 composition may be associated with heterogeneity in health effects across regions. However, because information on PM2.5 chemical constituents was not available in this study, we were unable to evaluate source- or composition-specific effects. Future studies incorporating PM2.5 composition data and contextual regional factors may help to assess further regional differences in the association between air pollution exposure and UTI. In addition, regional variation in healthcare access, socioeconomic conditions, and other unmeasured contextual factors may have contributed to the observed differences. The strengths of this study include its large cohort of 395,032 participants and its extended follow-up period based on retrospective nationwide data, which may support the generalizability of the findings to the Korean pediatric population. The availability of detailed health insurance data enabled accurate identification of both UTI events and antibiotic-treated UTI cases. However, the present study has some limitations. First, PM2.5 exposure levels were assigned based on administrative districts; individual-level differences in actual exposure, such as time spent indoors, commuting to daycare, kindergarten, or school, and the use of air filtration, were not fully accounted for. This may have led to misclassification of exposure and measurement error. In addition, because annual PM2.5 exposure estimates were assigned according to calendar year, some participants with events occurring early in the year may have been assigned exposure values that partially reflected post-event exposure periods, potentially resulting in temporal exposure misclassification. Second, although UTI is an acute condition, we applied a minimum follow-up period to facilitate the evaluation of long-term PM2.5 exposure. This design choice may be perceived as arbitrary and could introduce selection bias by excluding individuals with shorter follow-up. Therefore, the results should be interpreted with caution. Third, although extensive covariate adjustment was performed, residual confounding by unmeasured individual-level factors remains possible. Important pediatric UTI-related factors, including congenital urinary tract abnormalities, hygiene practices, daycare attendance, breastfeeding history, obesity, passive smoking exposure, and parental education, were not available in this claims-based database. Some of these factors may be associated with both residential environmental conditions and susceptibility to infection, potentially contributing to residual confounding. Future studies using large prospective birth cohorts or linked datasets incorporating childcare, educational, and environmental information are needed to better address these factors. Fourth, although susceptibility to UTI and responses to air pollution exposure may vary across developmental stages, we were unable to perform detailed age-stratified analyses. Because of the dynamic cohort structure, individuals contributed unequal amounts of follow-up time across age groups, potentially leading to biased or unstable estimates in fine age strata. Accordingly, the findings should be interpreted as average associations across childhood and adolescence rather than age-specific effects. Lastly, because UTI outcomes were identified using claims-based diagnostic records, milder UTI cases, asymptomatic bacteriuria, or culture-negative diagnoses may not have been consistently captured, particularly in outpatient pediatric settings. Therefore, outcome misclassification remains possible, and the findings may not fully represent the entire clinical spectrum of pediatric UTIs. Therefore, the observed associations should be interpreted with caution, and causal inference cannot be established. Future studies should incorporate individual-level exposure assessments, clinical validation of outcomes, and biomarker-based investigations to understand the underlying mechanisms and reduce residual confounding. Conclusions We observed an association between long-term exposure to PM2.5 and an increased risk of UTIs in a large cohort of Korean Children. These findings suggest that air pollution may contribute to the development of pediatric UTIs, potentially through mechanisms involving systemic inflammation, oxidative stress, and immune dysregulation. Our results suggest an association between PM2.5 exposure and pediatric UTI risk, while highlighting the need for cautious interpretation of subgroup differences given potential differences in healthcare utilization and outcome ascertainment. Public health efforts to reduce air pollution, along with further research to investigate causal relationships and underlying mechanisms, are warranted. Although annual PM2.5 concentrations in Korea have gradually declined over the past decade, the pace of improvement has recently slowed, and current levels remain substantially higher than the WHO air quality guideline of 5 µg/m3. Therefore, continued efforts to further reduce ambient PM2.5 concentrations are needed to protect vulnerable pediatric populations from preventable health risks, including UTI. Supplementary Information Supplementary Material 1 Acknowledgements This study was supported by the project “Institute of Ewha-Seoul Clinical Laboratories (SCL) for Environmental Health (IESEH)” and Research of Environmental Examination Model for Children and Women (No. 1-2022-0205-001-2). Authors’ contributions JEC, HMJ, WHL, JWL, HSK, and EHH conceived and designed the study. HMJ performed the formal analysis. HMJ, WHL, JMO, and JEC contributed to data curation and methodology. JEC was responsible for project administration. HMJ, WHL, and JEC contributed to data visualization. JEC, HMJ, JWL, and EHH drafted the manuscript. All authors contributed to reviewing and editing the manuscript. EHH and JWL supervised the study. All authors read and approved the final manuscript. Funding None. Data availability Data are available from the National Health Insurance Service’s Big Data Office (No. NHIS-2024-2-137). The authors had no special access privileges and have no right to share or provide the data. To request the database, visit https://nhiss.nhis.or.kr. Declarations Ethics approval and consent to participate This study was approved by the Institutional Review Board of Ewha Womans University Seoul Hospital (IRB No: SEUMC 2022-05-034). The requirement for informed consent was waived because this study used anonymized secondary data from the National Health Insurance Service database. All procedures were conducted in accordance with the principles of the Declaration of Helsinki and relevant guidelines and regulations. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Jung Won Lee, Email: jwped@ewha.ac.kr. Eunhee Ha, Email: eunheeha@ewha.ac.kr. References Associated Data Data Availability Statement Data are available from the National Health Insurance Service’s Big Data Office (No. NHIS-2024-2-137). The authors had no special access privileges and have no right to share or provide the data. To request the database, visit https://nhiss.nhis.or.kr.