Tuberculosis (TB) is a contagious infection caused by Mycobacterium tuberculosis (1). Transmission of TB occurs when an infected person passes microbial aerosols into the alveoli of a new host through coughing. Drainage from any site of active TB can be a source of infection (2). TB poses a significant threat to public health worldwide (3). In 2019, there were 10 million TB case notifications and TB was one of the top ten causes of death, resulting in 1.4 million deaths worldwide. The fourth Sustainable Development Goal is to end the TB epidemic by 2030. However, TB incidence declined by only 9% from 2015 to 2019 (i.e., from 142 to 130 per 100,000 individuals) (4). Global TB targets are off track with the incidence increasing from 127 to 134 per 100,000 individuals between 2020 and 2021 (5).

TB is a disease of poverty (6). There is an apparent socio-economic gradient in TB burden across different locations, with the poorest being at the highest risk (7). Common risk factors for TB include human immunodeficiency virus (HIV) infection, smoking, diabetes, alcohol abuse, undernutrition, and indoor air pollution (8). Outdoor air pollution may also affect TB infection and/or disease progression (9). Particulate matter (PM) with aerodynamic diameters <2.5 μm (PM_2.5) and < 10 μm (PM₁₀) are the two main ambient air pollutants. PM₁₀ can be deposited in the bronchi and PM_2.5 can further reach the alveoli. The adverse impacts of the PM exposure on human health have been comprehensively discussed in epidemiological, experimental, and mechanism studies (10–13). As for the impact of outdoor exposure to PM on TB, a few studies have suggested that short-term exposure to ambient PM_2.5 is positively associated with TB incidence (14–16), while others have reported a statistically non-significant association (17, 18). Similarly, the short-term association between exposure to PM₁₀ and TB incidence has also been inconclusive (19–21). TB is a chronic infectious disease. Patients with relatively severe TB symptoms must be hospitalized. However, previous studies have seldom explored the association between PM exposure and hospital admission for TB (22).

China has a high TB burden, ranking third following India and Indonesia in 2019. In that year, 833,000 individuals (i.e., 58.3 per 100,000 individuals) developed TB in China, accounting for 8.4% of the global population (4). Of these, 775,764 had pulmonary tuberculosis (PTB), representing 93.1% of the total number of patients with TB in China (23). Hainan Province is a relatively less developed province in China. From 2016 to 2019, the PTB case notification rate in Hainan Province increased from 76.5 to 90.3 per 100,000 individuals, with an average annual increase of >4.0% (24). Further, Hainan Province had China’s fifth-highest notification rate of PTB cases in 2019. Probing the association between PM exposure and hospital admission for PTB in Hainan Province would enrich our knowledge of the impact of PM on population health in tropical areas, shed light on targeted intervention measures, and help reduce the PTB burden more effectively. Herein, we aimed to determine the associations of hospital admission for PTB with PM_2.5 and PM₁₀, and further identify the potentially vulnerable populations in Hainan Province, China from 2016 to 2019.

Table 1 shows the number and percentage of hospital admission for PTB in different subgroups in Hainan Province from 2016 to 2019. A total of 4,562 hospital admission for PTB were included in this study, all of which occurred on Hainan Island (Figure 1). Patients aged < 65, male, married, and those living in county-level divisions accounted for 76.2, 76.6, 79.7, and 67.7% of admissions, respectively. There were slightly more patients in the warm season (54.0%) than in the cold season. The average concentrations of PM_2.5 and PM₁₀ were 20.2 μg/m³ and 36.2 μg/m³, respectively during the study period (Table 2).

Figure 2 illustrates the ORs of hospital admission for PTB associated with each IQR increase in PM_2.5 and PM₁₀ across various lag days. We found that the effect of PM_2.5 at lag 0–8 days was the highest among the effects at different lags, and the same finding was observed for PM_10. Therefore, we mainly explored the effects of PM_2.5 and PM₁₀ at a lag of 0–8 days. Specifically, the risk of PTB hospitalization were, respectively, 1.155 (95% confidence interval [CI]: 1.041–1.282) and 1.142 (95% CI: 1.033–1.263) for an IQR increase in PM_2.5 and PM₁₀ at a lag of 0–8 days.

Figure 3 presents the effects of PM_2.5 on hospital admission for PTB in different subgroups. An IQR increase in PM_2.5 was associated with 1.323 (95% CI: 1.073–1.631), 1.149 (95% CI: 1.021–1.293), 1.160 (95% CI: 1.032–1.303), and 1.277 (95% CI: 1.089–1.498) hospital admission risks for PTB among patients aged ≥65 years, male patients, married individuals, and those living in prefecture-level cities, respectively. As for the disparity by season, the effect of PM_2.5 on hospital admission for PTB was statistically significant in the warm season (OR = 1.260, 95% CI: 1.066–1.488) but not in the cold season. The trends in the differences in the effects of PM₁₀ across subgroups were similar to those of PM_2.5, although the subgroup differences were not statistically significant (p > 0.05; Figures 3, 4). The effects of PM_2.5 on hospital admission for PTB seemed to be greater than those of PM₁₀ for most of the subgroups.

The sensitivity analysis results suggested that the main analysis findings were robust when changing the dfs for the natural cubic splines of temperature and relative humidity to 4–6 (Supplementary Tables S1, S2).

The present study evaluated the short-term effects of ambient PM_2.5 and PM₁₀ exposure on hospital admission for PTB in a tropical Chinese province. It was found that both PM_2.5 (OR = 1.155) and PM₁₀ (OR = 1.142) were positively associated with hospital admission for PTB, which was partially in line with the findings of previous studies that reported adverse effects of PM on PTB outcomes, such as incidence (19, 35) and mortality (36). However, Álvaro-Meca et al. found that the effect of PM₁₀ on hospital admission for PTB was not statistically significant in HIV-infected patients (22). The health impact of PM on PTB hospitalization may vary across different populations, warranting further investigation.

The mechanism linking PM exposure to hospital admission for PTB remains unclear, however, several plausible explanations exist. Fine PM can irritate the lung tissue, damage the tracheobronchial mucosa, and impair the body’s anti-mycobacterial immunity by inhibiting the synthesis and secretion of various key inflammatory mediators in the fight against M. tuberculosis (37). The susceptibility and infection of M. tuberculosis may be aggravated when exposed to high concentrations of PM because it induces alterations in protective host immune responses (38). Meanwhile, an experimental study suggested that exposure to PM_2.5 and PM₁₀ increased the intracellular growth of M. tuberculosis (39). Therefore, PM exposure may exacerbate M. tuberculosis infections in the lungs and increase risk of hospital admission for PTB.

We found that the effect of PM_2.5 on hospital admission for PTB was higher than that of PM₁₀, which was consistent with previous studies on respiratory disease hospitalizations (28, 40). An experimental study also reported that PM_2.5 induced higher oxidative stress and more malondialdehyde in rat lung epithelial cells and cells treated with PM_2.5 were more susceptible to apoptosis than PM₁₀ (41). The stronger effect of PM_2.5 on PTB outcomes may be attributed to its physicochemical characteristics such as a larger surface area and longer atmospheric suspension (42). At the same time, PM_2.5 can carry external impurities, such as metal ions, minerals, and polycyclic aromatic hydrocarbons into the fine bronchi and alveoli (43, 44), which may lead to various forms of inflammatory responses and impair cellular immune function in the lungs (45, 46).

In accordance with previous research (16), we found that patients aged ≥65 years were vulnerable to PM exposure. Older adults infected with M. tuberculosis, a potent stimulator of numerous inflammatory cytokines, tend to develop a range of inflammatory responses (47). With a decline in the innate immune response, PM inhalation may lead to overwhelming inflammation when superimposed on the inflammation associated with aging in the older adults patients with PTB (48).

Consistent with a previous study on TB incidence conducted in Shanghai (16), we found that PM significantly affected hospital admission for PTB in male patients but not in female patients. Polymorphisms of the Toll-like receptor 8 gene (gene identity: 51311) on the X chromosome are associated with susceptibility to PTB in males, and susceptibility to PTB also increases in male carriers of rs3764880 allele A (49). Additionally, males tend to consume alcohol and nicotine (50), which may have a significant inhibitory effect on tumor necrosis factor-α production and cell-mediated immunity, thereby exacerbating the PTB progression (51).

Our study found that the association between PM and hospital admission for PTB was statistically significant for married patients but not for others. Married people have less time to perform stress-relieving healthy physical activity for family-related tasks, including childcare (52). An Indian study found that single TB patients had higher quality of life scores (53). A cross-sectional study of treated PTB patients suggested that married PTB patients had lower mental-domain scores (54). Less physical activity, lower quality of life, and poor mental health may weaken the immune system, in which case PM inhalation may have more harmful health effects.

In our study, the effect of PM on hospital admission for PTB varied according to administrative division, with PM having a greater impact on patients living in prefecture-level cities. This finding is similar to that of a national study in China that explored the associations between major PM_2.5 components and TB incidence and mortality (55). Inconsistent socioeconomic characteristics across administrative regions may explain this result. The higher population density of prefecture-level cities (Supplementary Table S3) may facilitate the spread of respiratory viruses, such as influenza (56). Higher respiratory virus incidence and ambient PM concentration may exacerbate PTB (20, 38). Therefore, the effect of PM in prefecture-level cities tends to be more detrimental than that in county-level divisions.

We found that the effect of PM on hospital admission for PTB was statistically significant in the warm season but not in the cold season. The result can be interpreted as follows: First, M. tuberculosis’s rate of proliferation and growth increases significantly with increasing temperature (57). Second, on average, people spend more time outdoors (or with windows open) during the warm season, which can expose them to outdoor PM more often. Third, PM may reduce the phagocytosis of M. tuberculosis by peripheral blood mononuclear cells in warmer months but not in colder months (58).

Taken together, exposure to ambient PM_2.5 and PM₁₀ increases the risk of PTB hospital admission. PM with a smaller aerodynamic diameter is more hazardous. Our findings highlight the importance of reducing ambient PM concentrations to alleviate the PTB burden.

The data sets generated during and/or analyzed during the current study are available from the corresponding authors on reasonable request.

This study was approved by the Research Ethics Committee of Hainan General Hospital (No. 2021-314). Written informed consent was not required as the data was anonymized for the study. Consent was waived by the Research Ethics Committee.

P-PZ: formal analysis, methodology, software, visualization, and writing – original draft. YG: conceptualization, funding acquisition, investigation, project administration, resources, validation, and writing – original draft. G-ZZ: resources and writing – original draft. RL, X-BL, X-XF, JF, and FL: data curation. Y-PZ: conceptualization, supervision, and writing – review and editing. LL: conceptualization, funding acquisition, methodology, supervision, and writing – review and editing. All authors contributed to the article and approved the submitted version.

This work was supported by the National Natural Science Foundation of China [82003555] and Hainan Province Science and Technology Special Fund [ZDYF2021SHFZ079].

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.