International journal of epidemiology最新文献

{"title":"Evaluating COVID-19 vaccine effectiveness across epidemiologic designs: test-negative case-control versus retrospective cohort analyses in a nationwide study","authors":"Layan Sukik,Pradipta Paul,Houssein H Ayoub,Peter Coyle,Patrick Tang,Mohammad R Hasan,Hadi M Yassine,Asmaa A Al Thani,Zaina Al-Kanaani,Einas Al-Kuwari,Andrew Jeremijenko,Anvar Hassan Kaleeckal,Ali Nizar Latif,Riyazuddin Mohammad Shaik,Hanan F Abdul-Rahim,Gheyath K Nasrallah,Mohamed Ghaith Al-Kuwari,Adeel A Butt,Hamad Eid Al-Romaihi,Mohamed H Al-Thani,Abdullatif Al-Khal,Hiam Chemaitelly,Laith J Abu-Raddad","doi":"10.1093/ije/dyag210","DOIUrl":"https://doi.org/10.1093/ije/dyag210","url":null,"abstract":"Abstract Background Observational studies are central to assessing vaccine effectiveness and informing public health policy. The two most common designs are test-negative case-control and cohort studies, yet direct comparisons of their performance using the same population, data sources, and analytic strategies remain rare. This study compares these two designs within a national population using harmonized analytic approaches. Methods Nationwide, matched test-negative case-control and retrospective cohort studies were conducted in Qatar between 1 January 2021 and 18 December 2021 to assess the effectiveness of the coronavirus disease (COVID-19) primary series of BNT162b2 and mRNA-1273 against any severe acute respiratory syndrome coronavirus 2 infection, symptomatic infection, and severe COVID-19. Results In the test-negative design, BNT162b2 effectiveness was 71.0% [95% confidence interval (CI): 70.1–71.9] against any infection, 75.6% (95% CI: 74.2–77.0) against symptomatic infection, and 96.3% (95% CI: 95.2–97.2) against severe COVID-19. In the cohort design, after adjustment for testing rate, corresponding effectiveness was 78.0% (95% CI: 77.1–78.9), 81.2% (95% CI: 79.8–82.5), and 98.0% (95% CI: 96.6–98.8), respectively. Both designs demonstrated similar, rapid, and largely linear waning of effectiveness in both magnitude and pattern. Test-negative design estimates lay between cohort estimates with and without adjustment for testing rate. Similar results were observed for mRNA-1273. Conclusions Both designs yielded consistent effectiveness estimates across vaccines and clinically relevant outcomes, with only modest differences related to handling of health-seeking behavior. These findings demonstrate that, when rigorously implemented using quality data and harmonized analytic strategies, both designs produce comparable effectiveness estimates.","PeriodicalId":14147,"journal":{"name":"International journal of epidemiology","volume":"36 1","pages":""},"PeriodicalIF":7.7,"publicationDate":"2026-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149020285","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Health behavior buffering the spillover effect of cancer on spousal depressive symptoms: causal evidence from a multicountry analysis","authors":"Yanshang Wang,Yiqi Xia,Chang Cai,Zhenyu Shi,Ping He","doi":"10.1093/ije/dyag209","DOIUrl":"https://doi.org/10.1093/ije/dyag209","url":null,"abstract":"Abstract Background Cancer, often described as a “family disease,” can negatively affect the mental health of spouses or partners through negative spillover effects. However, there is a lack of rigorous causal evidence regarding the spillover effects of cancer onset, and the dynamic pattern remains unclear. This study aims to assess the spillover effects of cancer onset on spousal depressive symptoms using a multicohort approach and examine the moderating role of a healthy lifestyle. Methods We conducted a pooled multicountry analysis utilizing 21 countries datasets. We employed a causal inference research design, combining matching with difference-in-differences (DID) as the primary identification strategy to analyze the causal effect and its heterogeneity. Moreover, we conducted the interaction term analysis to explore the role of health lifestyles. Using an event study methodology, we examined the dynamic pattern and validated the parallel trends assumption. Results Our findings reveal a robust causal link between cancer onset and spousal depressive symptoms. Further analysis indicates that healthy lifestyles serve as a buffer against this adverse spillover effect, particularly reflected in moderate alcohol use (β = −0.052, 95% confidence interval [CI] [−0.099, −0.005]), physical exercise (β = −0.116, 95% CI [−0.183, −0.049]), social activities (β = −0.126, 95% CI [−0.181, −0.071]), and tobacco use (β  =  0.100, 95% CI [0.029–0.171]). Additionally, the spillover effect exhibits a “bimodal” pattern, manifesting primarily in the short term (within 2 years) and later stages (4–5 years) of cancer. Conclusion This study reveals a dynamic negative impact of cancer on spousal mental health, closely related to the cancer trajectory and spousal behaviors. It underscores the imperative for establishing a family-based integrated care model for cancer. Future research should explore the underlying mechanisms of these spillover effects as well as develop care models to conduct prospective impact assessments.","PeriodicalId":14147,"journal":{"name":"International journal of epidemiology","volume":"10 1","pages":""},"PeriodicalIF":7.7,"publicationDate":"2026-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149022354","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Age-specific vulnerability of kidney diseases to temperature: a nationwide case-crossover study.","authors":"Jinah Park, Jisung Choi, Ziyad Al-Aly, Kristie L Ebi, Jung Pyo Lee, Ho Kim, Whanhee Lee","doi":"10.1093/ije/dyag159","DOIUrl":"https://doi.org/10.1093/ije/dyag159","url":null,"abstract":"<p><strong>Background: </strong>Extreme temperatures are projected to increase; however, comprehensive mapping of associations between temperature extremes and kidney diseases across age groups and disease subtypes remains limited.</p><p><strong>Methods: </strong>Data were collected for all emergency department (ED) visits for kidney diseases across all ages from South Korea's national health insurance claims between 2011 and 2019. A time-stratified case-crossover design was used to assess the association between ambient temperature and ED visits for kidney diseases.</p><p><strong>Results: </strong>The study included around 1.9 million ED visits due to kidney diseases. Heat-related risks of kidney disease-related ED visits (99th vs. 75th percentile) were highest among adults aged 50-59 years [odds ratio (OR): 1.28, 95% confidence interval (CI): 1.22-1.34), while cold-related risks (1st vs. 25th percentile) were highest among those aged 30-39 years (OR: 1.13, 95% CI: 1.05-1.20). Heat-related hospitalization risk was primarily driven by urinary tract infection among individuals in their 30s (OR: 1.37; 95% CI: 1.09-1.72) and by acute kidney injury among those in their 40s (OR: 2.29; 95% CI: 1.52-3.44). From a cost-burden perspective, urinary tract infections among adults aged ≥70 years contributed the largest medical cost burden attributable to high and low temperatures.</p><p><strong>Conclusions: </strong>This nationwide study demonstrated age- and disease subtype-specific heterogeneity in temperature-related kidney disease risks. Our findings suggest that, in addition to older adults, younger and middle-aged populations should also be primarily considered in climate change adaptation strategies targeting kidney health.</p>","PeriodicalId":14147,"journal":{"name":"International journal of epidemiology","volume":"55 5","pages":""},"PeriodicalIF":5.7,"publicationDate":"2026-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148792779","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Spatial small-area estimation of under-five mortality in Ethiopia: mapping subnational inequalities to accelerate SDG achievement.","authors":"Solomon Keflie Assefa, Adem Tsegaw Zegeye, Belayneh Jejaw Abate, Dejen Kahsay Asgedom, Eliyas Addisu Taye, Endalew Minwuye Andargie, Eyob Akalewold Alemu, Halima Ayalew Kebede, Helen Birhan Alemaw, Simachew Getaneh Endalamew, Helina Mebratu Abegaz, Sofiya Ayalew Kebede","doi":"10.1093/ije/dyag158","DOIUrl":"https://doi.org/10.1093/ije/dyag158","url":null,"abstract":"<p><strong>Background: </strong>Ethiopia has achieved substantial reductions in under-five mortality, but national estimates mask important subnational variation. Reliable zonal estimates are needed to identify high-burden areas and guide local health planning. This study estimated under-five mortality across Ethiopian zones by using a Bayesian spatial small-area estimation (SAE) approach with geospatial covariates.</p><p><strong>Methods: </strong>An ecological analysis was conducted by using data from the 2019 Mini-Ethiopian Demographic and Health Survey. Under-five mortality risk was estimated at the zonal level through a Bayesian hierarchical spatial SAE model. Associations with selected area-level covariates were assessed. Smoothed relative risks (RRs) with corresponding 95% credible intervals (CrIs) were reported.</p><p><strong>Results: </strong>Marked geographic heterogeneity in under-five mortality was observed. Elevated risks were concentrated in peripheral lowland regions, particularly Somali, Afar, Benishangul-Gumuz, and Gambella. High-burden zones included Afar (Zones 1, 2, and 3), Metekel in Benishangul-Gumuz, Anguak and Nuer in Gambella, and Dollo, Faftan, and Liben in Somali. Population density was inversely associated with mortality (RR: 0.83, 95% CrI: 0.77-0.98), whereas longer travel time to the nearest city was associated with higher under-five mortality (RR: 1.24, 95% CrI: 1.02-1.44).</p><p><strong>Conclusion: </strong>Under-five mortality remains unevenly distributed across Ethiopia, with persistent hotspots in remote and underserved regions. Improving geographic access to health services in these areas may help reduce inequalities in child survival and support progress toward the 2030 Sustainable Development Goals.</p>","PeriodicalId":14147,"journal":{"name":"International journal of epidemiology","volume":"55 5","pages":""},"PeriodicalIF":5.7,"publicationDate":"2026-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148812923","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Data Resource Profile: The Korean National Health Insurance Service Mother-Child Linked Database (NHIS-MCLD): a nationwide administrative database with a cohort design.","authors":"Jeong-Won Oh, Seung-Ah Choe, Eunseon Gwak, Seong Yong Park, Young June Choe, Sung-Il Cho","doi":"10.1093/ije/dyag206","DOIUrl":"https://doi.org/10.1093/ije/dyag206","url":null,"abstract":"","PeriodicalId":14147,"journal":{"name":"International journal of epidemiology","volume":"55 5","pages":""},"PeriodicalIF":5.7,"publicationDate":"2026-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149001774","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Maternal smoking and risk of childhood cancer: a report from the Childhood Cancer and Leukemia International Consortium","authors":"Allison Domingues, Logan G Spector, Giorgio Tettamanti, Maria Feychting, Atte Nikkilä, Anssi Auvinen, Wafa Alimam, Jad Abuhamed, Michael E Scheuer, Philip J Lupo, Julia E Heck, Johnni Hansen, Beth A Mueller, Heather Nelson, Richard MacLehose, Erin Marcotte","doi":"10.1093/ije/dyag225","DOIUrl":"https://doi.org/10.1093/ije/dyag225","url":null,"abstract":"Background Although tobacco is an established carcinogen and components of tobacco smoke may cross the placental membrane, past studies of maternal prenatal smoking and childhood cancer have resulted in mixed or null results. However, many studies have relied on case–control methods and may be susceptible to various biases. We aimed to elucidate the nature of the relationship between maternal prenatal smoking and childhood cancer through a multinational analysis of prospectively recorded record-linkage studies with probabilistic bias methods. Methods Data for this analysis came from existing population-based case–control record-linkage studies in the Childhood Cancer and Leukemia International Consortium, with studies from the USA, Sweden, Finland, and Denmark, including more than 20 000 cases and 300 000 controls. We assessed the association between maternal prenatal smoking and childhood cancer using logistic regression. Additionally, we conducted probabilistic bias analysis to investigate the effect of exposure misclassification using sensitivities and specificities from biomarker-based validation studies. Results Both before and after bias adjustment, maternal smoking was not found to be associated with most cancers investigated. Maternal prenatal smoking was inversely associated with paediatric melanoma [bias-adjusted odds ratio (OR) 0.68, 95% simulation interval (SI): 0.47–0.98] and skin carcinomas and other/unspecified carcinomas (bias-adjusted OR: 0.72, 95% SI: 0.52–0.96), and positively associated and unspecified malignant tumours (bias-adjusted OR: 1.50, 95% SI: 1.21–1.84). Conclusions This analysis supports previous findings that there is little evidence of an effect of maternal prenatal smoking on childhood cancer risk, with few potential exceptions. However, due to numerous other detrimental effects, public health efforts should continue to further reduce its prevalence.","PeriodicalId":14147,"journal":{"name":"International journal of epidemiology","volume":"22 1","pages":""},"PeriodicalIF":7.7,"publicationDate":"2026-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149150263","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The impact of temporal trends in stroke event rates and 30-day case fatality on trends in stroke mortality: an international comparative study.","authors":"Lee Nedkoff, F Lucy Wright, Nhung Nghiem, Solmaz Setayeshgar, Emmalin Buajitti, Aden C H Kwok, Dawit Zemedikun, Ximena Camacho, Samantha Stiles, Raphael Goldacre, Kate Smolina, Laura C Rosella, Tony Blakely, Philip Clarke","doi":"10.1093/ije/dyag157","DOIUrl":"10.1093/ije/dyag157","url":null,"abstract":"<p><strong>Background: </strong>Stroke mortality rates have declined in recent decades; however, data on contemporary drivers of this trend are limited. This study compared the contribution of trends in stroke event rates and 30-day case fatality to declines in stroke mortality from 2002 to 2018.</p><p><strong>Methods: </strong>Linked hospitalization and mortality data were obtained from Western Australia (WA), England, New Zealand, British Columbia (BC), and Ontario. Age-standardized and age-specific stroke mortality and event rates and 30-day case fatality were calculated and age-adjusted trends estimated from Poisson and binomial regression models. The relative contribution of trends in stroke events and 30-day case fatality to trends in stroke mortality were calculated.</p><p><strong>Results: </strong>There were 2 248 242 stroke events identified. Age-adjusted reductions in event rates ranged from -0.96%/year [95% confidence interval (CI) - 1.22, 0.70%] in men in WA to -2.51%/year (95% CI -2.61, -2.40%) in women in Ontario. Age-standardized 30-day case fatality was highest in England in 2002 (men 43.7%, women 45.5%), but declined at a greater rate than in other jurisdictions. Declines in stroke mortality ranged from -3.3%/year to -5.5%/year. Reductions in 30-day case fatality were the major contributor to declines in stroke mortality [range: men, 63.5% (Ontario)-80.9% (England); women, 55.3% (BC)-73.1% (England)]. In people aged <65 years, stroke mortality declines were almost completely driven by reductions in 30-day case fatality with increasing or attenuated event rates among those aged 35-64 years.</p><p><strong>Conclusion: </strong>Improvements in 30-day case fatality had a greater impact on stroke mortality than trends in event rates, suggesting that further gains could be achieved with targeted prevention. Adverse trends in event rates in people aged <65 years could signal future increases in stroke mortality.</p>","PeriodicalId":14147,"journal":{"name":"International journal of epidemiology","volume":"55 5","pages":""},"PeriodicalIF":5.7,"publicationDate":"2026-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13653606/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148989771","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"All covariates are not created equal in regression calibration: a review of methods on using biomarkers to calibrate dietary intake.","authors":"Wenze Tang, Zihan Qian, Xiao Gu, Walter C Willett, Molin Wang","doi":"10.1093/ije/dyag169","DOIUrl":"10.1093/ije/dyag169","url":null,"abstract":"<p><p>Regression calibration is a widely used method for correcting bias in regression coefficient estimates caused by measurement error in continuous exposures. In nutritional epidemiology, recent studies often use nutrient-derived biomarkers to calibrate true dietary intakes, which are typically assessed through small feeding studies or weighed dietary records. However, biomarkers may themselves lie on the causal pathway between intake and outcome, violating the surrogacy assumption that underpins the validity of regression calibration. We evaluate regression calibration-based methods that have been applied or proposed for calibrating dietary intake using biomarkers, focusing on their validity and relative efficiency. Our assessment combines analytic bias quantification, simulation studies, and a real-data application. Two approaches are generally valid when using mediators to calibrate the true exposure: (1) the expanded calibration method, which recovers the total effect via the product method from mediation analysis, and (2) the two-stage calibration method. In simulations reflecting realistic scenarios, the expanded calibration method demonstrated superior efficiency for a continuous outcome. Finally, we show that estimating the effect of dietary intake using a calibrated biomarker is valid when the mean biomarker calibration model contains a linear term in true intake with coefficient one, possibly along with additional covariates.</p>","PeriodicalId":14147,"journal":{"name":"International journal of epidemiology","volume":"55 5","pages":""},"PeriodicalIF":5.7,"publicationDate":"2026-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13499690/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148792811","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Association between ozone and emergency department visits in people with disabilities in South Korea: a nationwide case-crossover study.","authors":"Jinah Park, Ayoung Kim, Michelle L Bell, Ziyad Al-Aly, Kristi L Ebi, Sooyoung Kim, Seoyeong Ahn, Cinoo Kang, Ho Kim, Whanhee Lee","doi":"10.1093/ije/dyag189","DOIUrl":"https://doi.org/10.1093/ije/dyag189","url":null,"abstract":"<p><strong>Background: </strong>Ozone is known to have harmful health impacts, and its concentration is projected to rise with climate change; however, little is known about how ozone impacts the health of people with disabilities and its interaction with temperature. We aimed to estimate the association between ozone exposure and emergency department (ED) visits, stratified by disability status and type.</p><p><strong>Methods: </strong>ED visit data spanning 2015-19 and 2022-23 were collected from the Korean National Health Insurance Service. We included the 3-day moving average (the same day, 1 day prior, and 2 days prior; lag 0-2) ozone concentration at the 8-hour maximum and adjusted for daily mean temperature and national holidays. Using a time-stratified case-crossover design and conditional logistic regression, we estimated the association between ozone concentration and natural-cause ED visits. Furthermore, an interaction term between temperature strata and ozone was included to assess potential effect modification across temperature value.</p><p><strong>Results: </strong>Overall, 3 484 877 ED visits were recorded. An interquartile range (∼22 ppb) increase in daily 8-hour maximum ozone (lag 0-2) was associated with a 1.94% increase [95% confidence interval (CI): 1.64-2.25] and a 1.93% increase (95% CI: 1.62-2.25) in ED visits during the warm season (April-September) and the entire year (January-December), respectively, among persons with disabilities. These estimates were higher than those observed among persons without disabilities [1.36% (95% CI: 0.92-1.80) and 1.42% (95% CI: 1.03-1.82) during the warm season and the entire year, respectively]. Greater risks were observed in brain lesion disorders, developmental and mental disabilities, ED visits due to respiratory diseases during June, and on days with high temperatures.</p><p><strong>Conclusions: </strong>These findings underscore the importance of tailoring strategies to specific disability types and early-season public health interventions-beginning in April or May-incorporating both ozone levels and daily temperature to effectively mitigate the heightened risks identified in this population.</p>","PeriodicalId":14147,"journal":{"name":"International journal of epidemiology","volume":"55 5","pages":""},"PeriodicalIF":5.7,"publicationDate":"2026-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148880394","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Prenatal exposure to PM2.5 components and fetal growth: identifying critical gestational weeks and key toxic components in a Chinese birth cohort.","authors":"Qiannan Wei, Li Wen, Jianling Shen, Qingqing Wang, Wenzhuo Li, Yan Jin, Xueting Liu, Peiyao Chen, Wenjie He, Qiansheng Hu, Zhenyan Han, Weiwei Lin","doi":"10.1093/ije/dyag185","DOIUrl":"10.1093/ije/dyag185","url":null,"abstract":"<p><strong>Background: </strong>Current evidence remains inconclusive regarding the developmental toxicity of specific fine particulate matter (PM2.5) components during critical fetal growth periods, particularly concerning key toxic components and exposure windows.</p><p><strong>Methods: </strong>This prospective birth cohort included 2680 mother-newborn pairs with 9339 growth ultrasound measurements [abdominal circumference, head circumference (HC), femur length (FL), estimated fetal weight (EFW)] in Guangzhou, China, from 2017 to 2020. Weekly exposures to PM2.5 components-black carbon (BC), organic matter (OM), nitrate (NO3-), sulfate (SO42-), and ammonium (NH4+)-were quantified by using the Tracking Air Pollution in China database (TAP) real-time air-pollutant-exposure model. Distributed lag models with generalized estimating equations were employed to identify critical exposure windows and machine-learning approaches were used to assess constituent mixture effects and identify dominant toxic components.</p><p><strong>Results: </strong>FL, HC, and EFW demonstrated component-specific vulnerability. FL reduction at gestational Weeks 12-15 was most strongly associated with BC exposure, followed by OM and SO42-; HC deficits in Weeks 5-12 and 22-24 were predominantly linked to NO3- exposure; EFW reductions during Weeks 6-8 were most sensitive to NO3- exposure. BC and NO3- were identified as the most impactful constituents.</p><p><strong>Conclusion: </strong>Our findings identify gestational Weeks 12-15 for BC exposure and Weeks 5-12 and 22-24 for NO3- exposure as critical vulnerable windows for adverse fetal development associated with PM2.5 components. These results support the need for further research into component-specific air-quality guidelines targeting sensitive gestational periods.</p>","PeriodicalId":14147,"journal":{"name":"International journal of epidemiology","volume":"55 5","pages":""},"PeriodicalIF":5.7,"publicationDate":"2026-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148891602","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}