Association between secondhand smoke exposure and ocular microbiome changes in children Authors: Qihang Sun, Xiangtian Ling, Yuzhou Zhang, Charlene C. Yim, Yu Peng, Yating Yang, Hei-Nga Chan, Xiujuan Zhang, Ka Wai Kam, Wai Kit Chu, Patrick Ip, Alvin L. Young, Stephen Kwok-Wing Tsui, Clement C. Tham, Chi Pui Pang, Li Jia Chen, Jason C. Yam Published: 2026 ABSTRACT Highlights•Secondhand smoke (SHS) exposure significantly alters the ocular surface microbiome (OSM) in children aged 3–18 years.•Microbial diversity is significantly reduced on ocular surface in SHS-exposed children.•Beneficial commensal taxa are depleted in SHS-exposed children.•Environmental stress-related metabolic pathways and immune-related pathways are significantly altered in SHS-exposed children. Introduction The ocular surface microbiome (OSM) is a complex and diverse community of microorganisms that reside on the ocular surface (JÁP et al., 2020). It plays an essential role in maintaining ocular health. A well-balanced OSM supports ocular surface homeostasis, preserves immune system functionality, and provides protection against pathogenic microorganisms (Tariq et al., 2025). Conversely, an imbalanced OSM can lead to the overgrowth of pathogenic microbes and the excessive production of pro-inflammatory or disease-promoting metabolites, thereby initiating or exacerbating the development of ocular diseases (Aragona et al., 2021). Alterations in the OSM have been implicated in various ocular disorders compared to healthy controls, including dry eye disease (Dong et al., 2019; Tong et al., 2022), allergic rhinoconjunctivitis (Yau et al., 2019), ocular infections such as trachoma (Pickering et al., 2019), bacterial (Shivaji et al., 2021), and fungal (Prashanthi et al., 2019) keratitis, diabetic retinopathy (Martins et al., 2004), and conjunctival mucosa-associated lymphoid tissue (Asao et al., 2019). The OSM is also associated with ocular health in children. Studies have shown that changes in the OSM are associated with dry eye disease in children and adolescents with diabetes (Chen et al., 2022). Alterations have been observed during the progression and treatment of trachoma in children (Pickering et al., 2019; Pickering et al., 2022). However, the relationship between the pediatric OSM and many ocular diseases remains poorly understood. Secondhand smoke (SHS) represents a significant environmental exposure in children’s home environments. Our previous systematic review demonstrated that SHS exposure is associated with a wide range of ocular disorders, including eye irritation, conjunctivitis, dry eye disease, uveitis, myopia, astigmatism, contact lens discomfort, age-related macular degeneration (AMD), glaucoma, thyroid eye disease, the ocular neurovascular structures, and corneal biomechanical parameters (Zhang et al., 2023). SHS exposure is also closely linked to children's ocular health. Our previous research found that exposure to SHS in children is related to choroidal thinning, retinal nerve fiber layer thinning, and alterations in the retinal vasculature (Cheung et al., 2023; Li et al., 2021; Yuan et al., 2019). Furthermore, SHS exposure is associated with an increased risk of myopia development (Zhang et al., 2023). Beyond its impact on ocular health, SHS exposure has been shown to influence the composition of the human microbiome, including the gut and nasal microbiomes (Jaspers, 2014; Pérez-Castro et al., 2023; Raber et al., 2023). Given the continuous exposure of the ocular surface mucosa to the external environment, environmental factors may influence ocular surface homeostasis and microbial composition more directly than in some other mucosal sites (Aragona et al., 2021; Petrillo et al., 2020). However, the impact of SHS on the OSM has not been explored. The goal of this study is to investigate the effects of SHS on OSM. By collecting samples from subjects of the Hong Kong Children Eye Study, we analyzed their OSM composition and compare the findings between the control group and the SHS exposure group. This investigation seeks to elucidate the influence of SHS on the OSM and provide insights into the potential mechanisms through which SHS contributes to the development of pediatric ocular diseases. Methods Subject recruitment Participants were recruited from the Hong Kong Children Eye Study. Eligible subjects included children and adolescents older than 3 years and younger than 18 years. Exclusion criteria were contact lens wear, recent antibiotic or hormonal medication use, a history of ocular trauma, congenital anomalies, or inability to undergo ophthalmic examinations. Recruitment and sample collection were conducted between May 2022 and December 2023. All participants received comprehensive ophthalmic and general physical examinations. For each participant, one eye was randomly selected for conjunctival sampling and microbiome analysis. Eye selection was not based on ocular laterality, clinical findings, or disease severity. A single-eye sampling strategy was adopted to avoid inter-eye correlation and maintain statistical independence between samples. The study adhered to the principles of the Declaration of Helsinki and received approval from the Ethics Committee Board of the Chinese University of Hong Kong. Written informed consent was obtained from legal guardians, and oral assent was obtained from the children. Assessment of (SHS) exposure was conducted using a standardized parental questionnaire administered during participant recruitment. The questionnaire collected information regarding household smoking behavior after the child’s birth. Children were classified as SHS-exposed if at least one household member smoked in the home environment for ≥3 days per week from birth until the time of questionnaire completion. Participants without household smoking exposure during this period were classified as controls. DNA extraction, 16s amplification and sequencing For sample collection, one drop of sterile proxymetacaine hydrochloride (ProVain-POS Eye Drops 0.5% w/v) was applied in each eye. Two minutes later, a sterile, individually packaged Puritan flocked swab was gently rolled across the inferior conjunctival fornix from nasal to temporal side and back for at least 3 times. Samples were immediately placed into sterile tubes, kept on ice, and transferred to a −80 °C freezer within two hours. To minimize contamination, the swab was carefully handled to avoid contact with other ocular tissues. Environmental control swabs, with and without proxymetacaine hydrochloride, were also collected. DNA extraction was performed within two weeks using the DNeasy PowerSoil Pro Kit (Qiagen) following the manufacturer’s protocol. DNA concentrations were quantified with a Qubit 4 Fluorometer (Thermo Fisher). PCR amplification targeted the V3-V4 region of the bacterial 16S rRNA gene using primers 341F (CCTACGGGNGGCWGCAG) and 805R (GACTACHVGGGTATCTAATCC). High-throughput sequencing was performed on an Illumina MiSeq platform. Bioinformatics and statistical analysis Raw sequencing data were processed with the Qiime2 (2023.05) platform. (Hall and Beiko, 2018) After DADA2-based quality control and denoising, (Callahan et al., 2016) amplicon sequence variant (ASV) tables were generated for downstream analyses. Taxonomic classification was carried out using the q2-feature-classifier trained with study-specific primers and referenced against the SILVA 138 database. (Quast et al., 2013) Rare ASVs-those occurring only in a single sample or with fewer than ten total counts-were filtered out using R software. Contamination removal was conducted with the Decontam package. (Davis et al., 2018) Samples with fewer than 10,000 ASVs were excluded from analysis. Statistical analyses were carried out in R (version 4.2.0). Alpha diversity (Chao1, Shannon, Simpson indices) was calculated to assess microbial richness and evenness within samples using the phyloseq and vegan packages. Normality testing was performed prior to group comparisons. Comparisons between groups were conducted using Mann-Whitney U tests if alpha diversity indices did not follow a normal distribution. Beta diversity was evaluated using Bray-Curtis dissimilarity, Jensen-Shannon Divergence (JSD), and Jaccard indices, visualized with Principal Coordinate Analysis (PCoA) and Nonmetric Multidimensional Scaling (NMDS). Group differences were assessed via Permutational Multivariate Analysis of Variance (PERMANOVA). Taxonomic profiles were summarized using the phyloseq, DESeq2, and microbiomeMarker packages. Biomarker taxa associated with SHS exposure were identified using maaslin2 and randomForest. Differentially abundant taxa were further analyzed by Linear Discriminant Analysis (LDA) Effect Size (LEfSe), with an LDA score>2 considered significant. Predicted microbial functions were inferred using PICRUSt2 implemented in the ggpicrust package. Functional enrichment was compared with edgeR and DESeq2 and annotated with the MetaCyc and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases. Visualizations, including bar plots and heatmaps, were generated using dplyr and ggplot2. Multiple testing correction was performed using the Benjamini-Hochberg false discovery rate (FDR) method where applicable. Results Sample demographics, sequencing statistics, and data preprocessing A total of 432 children were enrolled in the study, including 111 children (25.7%) exposed to SHS and 321 children (74.3%) as healthy controls without SHS exposure. The demographic characteristics of the participants are summarized in Table 1. No statistically significant differences were observed in age (P = 0.65) or sex distribution (χ² =0.35, P = 0.55) between the two groups. Sequencing was successfully completed for all 432 samples. After quality filtering and denoising, a total of 15,779,081 high-quality reads were obtained. These reads were resolved into 10,930 ASVs. The average sequencing depth was 36,526 reads per sample (range: 10,860–84,174). The data of blank swabs, with and without proxymetacaine hydrochloride were presented in Supplementary Table S1.Table 1Demographic characteristics of study individuals (n = 432).Table 1 dummy alt textCharacteristicNo SHS Exposure (n = 321)SHS Exposure (n = 111)P valueSex (n, %)Male164 (51.1%)61 (55.0%)0.55Female157 (48.9%)50 (45.0%)Age, years (Mean ± SD)10.90 ± 2.4311.00 ± 2.080.65ASVs (Mean ± SD)36,242.6 ± 14,45837,344.2 ± 14,3520.46 Table 1Demographic characteristics of study individuals (n = 432).Table 1 dummy alt textCharacteristicNo SHS Exposure (n = 321)SHS Exposure (n = 111)P valueSex (n, %)Male164 (51.1%)61 (55.0%)0.55Female157 (48.9%)50 (45.0%)Age, years (Mean ± SD)10.90 ± 2.4311.00 ± 2.080.65ASVs (Mean ± SD)36,242.6 ± 14,45837,344.2 ± 14,3520.46 Demographic characteristics of study individuals (n = 432). Alpha diversity of ocular microbiome samples Significant differences were observed in alpha diversity indices including the Chao1 richness (Fig. 1A, P = 0.009), the Shannon index (Fig. 1B, P = 0.021) and the Simpson index (Fig. 1C, P = 0.015) of the OSM between the SHS exposure group and the control group.Fig. 1Alpha diversity of OSM samples of children from smoke-exposed households and from non-smoking households. The difference of diversity was analyzed using (A) Chao1 index, (B) Shannon index and (C) Simpson index. Group comparisons were performed using Mann-Whitney U tests.Fig 1 dummy alt text Alpha diversity of OSM samples of children from smoke-exposed households and from non-smoking households. The difference of diversity was analyzed using (A) Chao1 index, (B) Shannon index and (C) Simpson index. Group comparisons were performed using Mann-Whitney U tests. Beta diversity of ocular microbiome samples Beta diversity analysis was performed to evaluate differences in microbial community composition between the SHS exposure and control groups. PCoA revealed significant differences in community structure between the two groups (Fig. 2A, P = 0.01). NMDS results were consistent with the PCoA findings (Fig. 2B, P = 0.03). Consistent results were obtained using other beta diversity metrics (Table 2).Fig. 2Beta diversity of OSM samples of children from smoke-exposed households and from non-smoking households. The difference in community structures between the two groups were analyzed using (A) PcoA based on JSD and (B) NMDS based on Bray-Curtis dissimilarity.Fig 2 dummy alt textTable 2Analysis of beta diversity using different metrics.Table 2 dummy alt textMethodDistanceP valueFDR-adjusted P valuePCoABray-Curtis dissimilarity0.0330.033JSD0.0150.033Jaccard distance0.0260.033NMDSBray-Curtis dissimilarity0.0380.038JSD0.0190.038Jaccard distance0.0250.038 Beta diversity of OSM samples of children from smoke-exposed households and from non-smoking households. The difference in community structures between the two groups were analyzed using (A) PcoA based on JSD and (B) NMDS based on Bray-Curtis dissimilarity. Table 2Analysis of beta diversity using different metrics.Table 2 dummy alt textMethodDistanceP valueFDR-adjusted P valuePCoABray-Curtis dissimilarity0.0330.033JSD0.0150.033Jaccard distance0.0260.033NMDSBray-Curtis dissimilarity0.0380.038JSD0.0190.038Jaccard distance0.0250.038 Analysis of beta diversity using different metrics. Microbial community composition and structure in controls and SHS exposed children Microbial community composition was analyzed at both the phylum and genus levels. At the phylum level, 60 phyla were identified, while 1156 genera were annotated at the genus level. In the control group, the ocular microbiome was predominantly composed of Firmicutes (38.1%), Proteobacteria (31.3%) and Actinobacteriota (26.3%), collectively accounting for over 95% of total abundance (Fig. 3A). In the SHS exposure group, Firmicutes (34.0%) was also the most enriched phyla, whereas the abundance of Actinobacteriota (32.6%) was more than Proteobacteria (29.9%) (Fig. 3A). At the genus level, the five most abundant genera in the control group were Staphylococcus (24.4%), Corynebacterium (16.7%), unclassified Neisseriaceae-related taxa (11.7%), Streptococcus (6.1%), and Escherichia-Shigella (4.2%), accounting for over 60% of total abundance (Fig. 3B). In the SHS exposure group, the top five genera were Corynebacterium (23.5%), Staphylococcus (23.2%), unclassified Neisseriaceae-related taxa (11.7%), Streptococcus (6.0%), and Escherichia-Shigella (4.0%), also comprising over 60% of the total abundance (Fig. 3B).Fig. 3Taxonomic analysis of dominant species and differential abundance analysis in phylum level in OSM samples of children from smoke-exposed households and from non-smoking households. (A) Relative abundance of top 5 phyla in OSM samples. (B) Relative abundance of top 10 genera in OSM samples. (C) Differential abundance analysis using LEfSe identified phyla with significant differences (LDA score > 2) between the groups. (D) Differential abundance analysis using LEfSe identified genera with significant differences (LDA score > 2) between the groups.Fig 3 dummy alt text Taxonomic analysis of dominant species and differential abundance analysis in phylum level in OSM samples of children from smoke-exposed households and from non-smoking households. (A) Relative abundance of top 5 phyla in OSM samples. (B) Relative abundance of top 10 genera in OSM samples. (C) Differential abundance analysis using LEfSe identified phyla with significant differences (LDA score > 2) between the groups. (D) Differential abundance analysis using LEfSe identified genera with significant differences (LDA score > 2) between the groups. Differential abundance analysis identified phyla and genera with significant differences between the groups (Fig. 3C). At the phylum level, the control group had higher abundances of Firmicutes, Bacteroidota, Fusobacteriota and Campilobacterota (Fig. 3C). At the genus level, the control group had higher abundances of Lactobacillus, Anaerococcus, Leptotrichia and >10 genera (Fig. 3C). No phyla or genera were enriched in the SHS group. Predicted functional profiling of microbial communities To further explore the potential functional implications of the observed shifts in bacterial composition, predicted functional profiling of the microbial communities was performed. The MetaCyc analysis revealed that the metabolic pathways in the SHS exposure group were primarily enriched in l-methionine salvage cycle, biphenyl degradation, toluene degradation and heparin degradation (Fig. 4A). The KEGG pathway bar plot highlighted significant differences in the abundance of functional pathways between the SHS exposure and control groups. The significant pathways enriched in the SHS exposure group were related to immune functions such as complement, T cell receptor, B cell receptor, MAPK signaling pathway and several cancers (Fig. 4B).Fig. 4Predicted functional profiling of differential microbial communities between children from smoke-exposed households and from non-smoking households. (A) Bar plot showing the top 20 functional pathways enriched in the SHS exposure group compared to the control group, based on MetaCyc annotations. (B) Bar plot of KEGG pathways displaying the top 20 differential functions between the SHS exposure and control groups.Fig 4 dummy alt text Predicted functional profiling of differential microbial communities between children from smoke-exposed households and from non-smoking households. (A) Bar plot showing the top 20 functional pathways enriched in the SHS exposure group compared to the control group, based on MetaCyc annotations. (B) Bar plot of KEGG pathways displaying the top 20 differential functions between the SHS exposure and control groups. Discussion This study provides the first comprehensive characterization of the effect of SHS exposure on the OSM in children, revealing alterations in specific microbial taxa and predicted functional pathways. Significant differences in both alpha diversity and beta diversity metrics were observed between SHS-exposed and control groups, along with differential taxonomic and functional profiles, suggesting that SHS exposure may influence the overall structure and diversity of the ocular surface microbiome in children. Notably, significant differences were observed in multiple alpha diversity indices, including Chao1 richness, Shannon diversity, and Simpson diversity, indicating that SHS exposure is associated with altered microbial richness and evenness of the OSM. Consistently, beta diversity analyses using multiple distance metrics also demonstrated significant separation between SHS-exposed children and controls, suggesting a distinct microbial community structure associated with SHS exposure. These findings are consistent with reports showing reduced or altered microbial diversity in the oral, nasal, and respiratory microbiomes of smoke-exposed individuals (O'Dwyer et al., 2016; Yu et al., 2017). The ocular surface is a low-biomass, immunologically active environment, and chronic exposure to SHS-derived pollutants may disrupt microbial homeostasis by inducing local oxidative stress, epithelial barrier dysfunction, or immune modulation, thereby reshaping both microbial diversity and composition. At the phylum level, the OSM in both groups was dominated by Firmicutes, Proteobacteria and Actinobacteriota, consistent with previous reports in healthy children (Cavuoto et al., 2019). However, distinct compositional shifts were observed in SHS-exposed children, characterized by a relative reduction in Firmicutes and an increased of Actinobacteriota. LEfSe analysis further revealed that several phyla, including Firmicutes, Bacteroidota, and Fusobacteriota, were significantly enriched in the control group, suggesting a relative depletion of these taxa in SHS-exposed children. These phyla include many commensal organisms that contribute to mucosal immune regulation and epithelial homeostasis (Cao et al., 2022; Jordan et al., 2023; Li et al., 2025; Yang et al., 2025; Zafar and Saier, 2021), and their reduction may reflect a disruption of the normal ocular microbial ecosystem. At the genus level, the overall dominant taxa such as Staphylococcus, Corynebacterium, Streptococcus, unclassified Neisseriaceae-related taxa, and Escherichia-Shigella were broadly similar between groups, suggesting relative stability of commonly detected ocular microbial taxa. However, differential abundance analysis demonstrated that numerous genera, including Lactobacillus and Rubellimicrobium, were significantly enriched in the control group. Lactobacillus has been reported to suppresses inflammatory stress-induced premature cellular senescence of human retinal pigment epithelium cells (Yamazaki et al., 2020), protect against blue light-induced retinal degeneration (Tsao et al., 2025) and improve dry eye symptoms (Lee et al., 2023). Rubellimicrobium has been detected lower in abundance in the conjunctiva of MALT lymphoma patients compared to healthy controls (Asao et al., 2019). Overall, these genera have been reported as commensal or potentially beneficial members at the ocular surface. Their reduction in SHS-exposed children suggests a loss of protective microbial functions, which may predispose the ocular surface to dysbiosis or inflammatory responses. Importantly, no bacterial taxa were found to be significantly enriched in the SHS exposure group, indicating that SHS may exert a broadly suppressive or disruptive effect on commensal microbial communities rather than promoting overgrowth of specific dominant pathogens. Beyond taxonomic shifts, predicted functional profiling suggested SHS-associated alterations in biological pathways. MetaCyc analysis showed enrichment of metabolic pathways involved in l-methionine salvage cycle, biphenyl degradation, heparin degradation and toluene degradation, which are linked to xenobiotic metabolism (Haiser and Turnbaugh, 2013) and stress adaptation. (Wang and Pan, 2016) These predicted functional shifts may reflect microbial responses to chemical constituents of SHS, such as aromatic hydrocarbons and aldehydes, which can select for bacteria with detoxification capabilities. (Leung et al., 2023; Martin et al., 2019) KEGG pathway analysis further revealed enrichment of immune-related pathways in SHS-exposed children, including complement, T cell receptor, B cell receptor, TGF-beta signaling pathway and MAPK signaling pathway, as well as pathways associated with cancer development. These findings may be clinically relevant because dysregulation of T cell receptor signaling has been implicated in ocular inflammatory diseases such as uveitis. (Deng et al., 2022) However, the present results are based on predicted functional inference from 16S rRNA sequencing data and should therefore be interpreted cautiously. Similarly, the TGF-β and MAPK pathways have been reported to participate in ocular surface inflammation and fibrosis, contributing to the pathogenesis of dry eye disease (Lin et al., 2023; Wang et al., 2025) and age-related macular degeneration. (Muraleva and Kolosova, 2023; Wang et al., 2019) The predicted functional enrichment observed here may suggest potential microbiome-associated adaptations to SHS exposure. However, these inference-based findings do not provide direct evidence of microbial metabolic activity or host immune modulation, and further validation using metagenomics, metatranscriptomics, or metabolomics is required. Several limitations of this study should be acknowledged. First, although significant associations were identified, the cross-sectional design precludes causal inference between SHS exposure and OSM alterations. Second, functional predictions based on 16S rRNA sequencing provide only indirect insights into microbial metabolic and immune-related activities. Future studies employing shotgun metagenomics, metatranscriptomics, or metabolomics would enable more precise functional and strain-level characterization. Third, this study focused exclusively on bacterial communities, while viruses, fungi, and other microbial eukaryotes may also play important roles in ocular surface health and warrant further investigation. In addition, residual confounding factors such as indoor air quality, environmental pollutants, socioeconomic status, and hygiene practices may influence the ocular surface microbiome and were not comprehensively assessed in the present study. These factors should be carefully controlled in future investigations. Finally, the topical anesthetic formulation used for conjunctival sampling contained benzalkonium chloride (BAK), a preservative with known antimicrobial activity. Although conjunctival swabs were collected shortly after anesthetic administration using a standardized protocol across all participants, and 16S rRNA sequencing detects bacterial DNA rather than bacterial viability, the potential influence of BAK on microbial composition in low-biomass ocular surface samples cannot be completely excluded. In conclusion, this study demonstrates that SHS exposure in children is associated with significant alterations in OSM diversity, community structure, taxonomic composition, and predicted functional pathways. Rather than promoting enrichment of specific pathogenic taxa, SHS exposure appears to be associated with disruption of commensal microbial communities and predicted functional shifts related to xenobiotic metabolism and immune signaling. These findings underscore the sensitivity of the pediatric OSM to SHS and highlight the importance of reducing SHS exposure to protect children’s ocular health. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.