Selection of study areas and participants
A population-based cross-sectional survey was performed in three locations around Realgar Plant (Heshan Village, Wangyangqiao Community and Baiyangshan Village) between January to March 2019 in Shimen County, Hunan Province of China. Heshan Village, Wangyangqiao Community and Baiyangshan Village is 2 km, 5 km, 10km away from Realgar Plant, respectively (Fig.1). The livelihoods of the residents living in the three locations were approximately similar. Because the lack of data of external arsenic exposure level (arsenic concentrations of atmosphere, soil, water, etc.), the distance from the location to Realgar Plant can be used as representation for the previous level of arsenic exposure approximately. Participants were recruited from a medical examination program approved by the local government at Realgar Plant Hospital. Written informed consent forms were obtained prior to the present study from all participants. The individuals (≥18-year-old) who had lived in these three locations and owning the local household registration were recruited for this study. Individuals with hearing or vision difficulties, neurological conditions, severe mental disorders or end-stage physical diseases (e.g., Parkinson's disease, epilepsy, multiple sclerosis, stroke, schizophrenia, depressive disorder, heart failure, liver failure, respiratory failure, uremia, etc.) were excluded. The demographic characteristics of participants collected by questionnaire included: age, sex, height, weight, location, body mass index (BMI), education, personal medical history, and diastolic blood pressure (DBP), and systolic blood pressure (SBP).
Collection of blood sample and measurement of biochemical index
Fasting blood samples were collected between 8 to 9 a.m. Samples were stored in a cold-chain box and sent to Shimen People's Hospital to test fasting blood glucose and blood lipids. Indices of blood lipids examination included: total cholesterol (TC), high density lipoprotein cholesterol (HDL-C), low density lipoprotein cholesterol (LDL-C), and triglyceride (TG). Among the 1556 participants, 294 participants were unavailable or declined to provide blood samples. The total number of blood samples used for the biochemical index determination was thus 1262.
Cognitive function and sleep assessments
The general cognitive function was evaluated by using a Chinese version of MMSE which ranged from 0 to 30. The higher MMSE scores suggest the better cognitive function. As MMSE scores are susceptible to educational level, the cut-off score indicating cognitive impairment for different education levels are: ≤17 points for illiterate participants, ≤20 points for participants accepted primary education and ≤24 points for participants accepted secondary or higher education [13].
Sleep quality was assessed by Pittsburgh sleep quality index (PSQI). PSQI is a well-validated tool aimed to evaluate subjective sleep quality. It is a questionnaire with 19 self-related items, composed of 7 subsections: subjective sleep quality, sleep duration, sleep latency, sleep efficiency, sleep disturbances, usage of sleeping pills, and daytime dysfunction. Each subsection score ranges from 0 to 3, and the total score of PSQI is 21. A lower PSQI score indicated better sleep quality [14].
All assessments were conducted by trained researchers in the field of neurology according to the protocol. All of the researches were blind to the study areas and cognitive status of participants.
Arsenic exposure status assessment and diagnosis of endemic arseniasis
Previous study suggested that arsenic concentrations in hair samples can be used as an effective marker of arsenic intake [15]. Therefore, in this study, we collected hair samples (approximately 1cm in length) behind the ear and sealed them separately in envelopes. Samples were sent to the local Occupation Disease Prevention and Treatment Center for analyzing. The accurate arsenic concentrations in hair samples were measured by inductively coupled plasma-mass spectroscopy (ICP-MS) [15].
Endemic arsenicosis was diagnosed according on the diagnosis standards of medical and health professions formulated by National Health Commission of the People’s Republic of China (PRC) in 2015. Residents of endemic arsenicosis areas can be diagnosed as arsenicosis if they meet one of the following clinical characteristics: (i) existing unexplained papular, nodular or verrucous hyperkeratosis on the skin of palms and soles, or (ii) existing diffuse or scattered pigmentation spots and/or round depigmentation spot of varying sizes with blur edges on the skin of trunk or non-exposed parts. Meanwhile, arsenic concentrations in hair or urine samples were significantly higher than the normal reference values in non-endemic areas can be used as a reference indicator for the diagnosis of arsenicosis [16]. The arseniasis participants were unanimously diagnosed by two dermatologists in Xiangya Hospital and Centers for Disease Control (CDC) according to the protocol. All of the researchers were blind to the study areas and cognitive status of participants.
Statistical Analysis
Data from all questionnaires were processed by SPSS. Characteristics of participants were described using mean ± standard error for normal distributed continuous variables or median (25% quartile, 75% quartile) for skewed distributed continuous variables, and frequency (%) for categorical variables. When the sample distribution conformed to normality, differences between two continuous variables groups were assessed by using Student's t test. For the two categorical variables groups comparisons, Chi-square test was used to analyze the differences. Differences among means of multiple groups were assessed using one-way analysis of variance (ANOVA). For skewed distributed variables, a nonparametric test was used. Two-level Logistic regression analysis was used to determine whether expose to arsenic was a risk factor for cognitive impairment, results were presented as OR and adjusted OR with 95% confidence intervals (CI). We performed linear regression analysis to examine the association of the hair arsenic concentrations and MMSE scores. All analyses were performed using SPSS version 25.0 (SPSS Inc., Chicago, IL, USA). Comparisons between groups were considered to be statistically significant when p<0.05.