A Prospective Cohort Study of Antimony Exposure and Cognitive Impairment in Older Adults — China, 2017–2021

Antimony (Sb) has been identified as a novel neurotoxin that affects neurocognition in previous animal studies. However, there is limited research on the association between Sb exposure and cognitive impairment in humans. Given the increasing aging population, it is crucial to investigate the relationship between Sb and cognitive impairment in older adults. In this study, we utilized data from the Healthy Aging and Biomarkers Cohort Study (HABCS), a prospective cohort study. A total of 1,333 participants aged 65 years and older were recruited in 2017–2018 and followed up in 2020–2021. Blood Sb (B-Sb) and urine Sb (U-Sb) concentrations were measured using inductively coupled plasma mass spectrometry. Cognitive function was assessed using the validated Mini-Mental State Examination. During the follow-up period of 4,972.1 person-years, 241 cases of cognitive impairments were recorded. Cox regression models, adjusted for potential covariates, showed that the risk of cognitive impairment increased by 56.5% for each e-fold increase in U-Sb. Similar results were observed for B-Sb, with a 52.3% increase in the risk of cognitive impairment for each e-fold increase. Our findings suggest that reducing Sb exposure may help mitigate the burden of cognitive impairment, particularly in regions with high Sb pollution.

The participants were selected from the HABCS study (1), which was conducted in nine regions known for longevity between 2017 and 2018. Follow-up of participants took place between 2020 and 2021. Detailed information about the nine longevity regions can be found in the Supplementary Material. A total of 3,016 participants were initially recruited, but after excluding those under 65 years old, individuals without blood or urine Sb information, those with missing data on cognitive function, and those who were lost to follow-up or died, a final sample size of 1,333 participants was included (Supplementary Figure S1 and Supplementary Material). Cognitive function was assessed using the Chinese version of the Mini-Mental State Examination (MMSE). Participants completed a face-to-face interview involving 24 items on the MMSE questionnaire, with a maximum score of 30. Cognitive impairment was defined as follows: uneducated individuals with an MMSE score of less than 18, individuals with 1–6 years of education and an MMSE score of less than 20, and individuals with 6 or more years of education and an MMSE score of less than 24 (2). Furthermore, standardized questionnaires were used to collect sociodemographic characteristics, frequency of food consumption, and health-related information. Information on exposure to Sb and other metals was obtained through biological sample collection. Please refer to the supplementary material for details on Sb detection and covariate definitions (Supplementary Table S1 and Supplementary Material). Ethical approval for the study was obtained from the Ethics Review Committee of the National Institute of Environmental Health, Chinese Center for Disease Control and Prevention (No. 2017018).

The study participants were divided into two groups based on their cognitive function. Continuous variables with a normal distribution were presented as mean±standard deviation (SD), while those with a skewed distribution were described as median (P₂₅–P₇₅). The t-test or rank sum test was used to analyze continuous variables, while categorical variables were described as frequencies (percentages) and analyzed using the χ²-test. A correlation map was used to represent the relationship between the independent variable and the covariable. A strong correlation was defined as |r|>0.7, indicating the possibility of multicollinearity. However, no statistically significant correlations were observed between variables in the correlation map (Supplementary Figure S2). The associations between Sb exposure and the incidence of cognitive impairment were assessed using Cox proportional hazard models. The survival time for participants was calculated from the baseline survey to either the date of cognitive impairment or the end of the survey. Sb exposure levels were modeled as continuous variables using ln-transformed values or categorical variables stratified by high, medium, and low concentrations. Please refer to the Supplementary Material for specific methods used to divide concentration categories. To ensure the accuracy of the study, covariates were included in the Cox proportional hazard models. More details about the models can be found in the eMethods section in the Supplementary Material. The association between Sb exposure and cognitive function was flexibly modeled using restricted cubic splines (RCS). Subgroup analysis was conducted to explore the effects of the association in different populations, and sensitivity analysis was performed to assess the robustness of the observations. Additional information can be found in the supplementary materials. All statistical tests were two-tailed, and P values

A total of 1,333 participants were enrolled in the study from 2020 to 2021, with an average follow-up duration of 3.73±0.21 years. Among these participants, 1,092 (81.9%) had normal cognitive function, while 241 (18.1%) had cognitive impairment. The median (P₂₅–P₇₅) levels of B-Sb and U-Sb exposure were 2.98 (2.45–3.72) μg/L and 0.05 (0.05–0.19) μg/L, respectively (Supplementary Table S2).

After controlling for all covariates, we observed that the risk of cognitive impairment increased by 56.5% with each e-fold increase in U-Sb [hazard ratio (HR)=1.565, 95% confidence interval (CI): 1.230, 1.991]. Compared to the low U-Sb group, the HR (95% CI) for cognitive impairment in the high U-Sb group was 2.456 (1.536, 3.927) (Table 1, Model 4). The non-linear relationship between U-Sb and the risk of cognitive impairment in older adults was confirmed by restricted cubic spline analysis (P_nonlinearFigure 1A). Furthermore, the risk of cognitive impairment in older adults increased linearly with B-Sb exposure, with a 52.3% increased risk for each e-fold increase in B-Sb (HR=1.523, 95% CI: 1.100, 2.109) (Figure 1B, Table 1).

In the subgroup analysis, we observed a stronger impact of Sb exposure on cognitive impairment in individuals aged 65–79 years, men, and individuals who consume alcohol. Among smokers, the effect of U-Sb on cognitive impairment was higher compared to non-smokers. We found a significant interaction effect between B-Sb and sex, but no significant interactions were observed between age, cigarette smoking, alcohol drinking, and Sb exposure (Table 2). The results of four sensitivity analyses further supported the robustness of our findings (Supplementary Tables S3–S6).