Journal Pre-proof

The Relationship between Plasma Al Levels and Multi-domain Cognitive

Performance among In-service Aluminum-exposed Workers at the SH
Aluminum Factory in China: A Cross-sectional Study

Shanshan Wang, Huaxing Meng, Nan Shang, Junhong Guo, Ting

Zhang, Shuhui Zhang, Yuqing Zhao, Huifang Zhang, Qinli Zhang,
Qiao Niu

PII: S0161-813X(19)30118-4
DOI: https://doi.org/10.1016/j.neuro.2019.10.011
Reference: NEUTOX 2547

To appear in: Neurotoxicology

Received Date: 8 May 2019

Revised Date: 24 October 2019
Accepted Date: 28 October 2019

Please cite this article as: Wang S, Meng H, Shang N, Guo J, Zhang T, Zhang S, Zhao Y,
Zhang H, Zhang Q, Niu Q, The Relationship between Plasma Al Levels and Multi-domain
Cognitive Performance among In-service Aluminum-exposed Workers at the SH Aluminum
Factory in China: A Cross-sectional Study, Neurotoxicology (2019),
doi: https://doi.org/10.1016/j.neuro.2019.10.011
This is a PDF file of an article that has undergone enhancements after acceptance, such as
the addition of a cover page and metadata, and formatting for readability, but it is not yet the
definitive version of record. This version will undergo additional copyediting, typesetting and
review before it is published in its final form, but we are providing this version to give early
visibility of the article. Please note that, during the production process, errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal
pertain.

© 2019 Published by Elsevier.

The Relationship between Plasma Al Levels and Multi-domain Cognitive
Performance among In-service Aluminum-exposed Workers at the SH Aluminum
Factory in China: A Cross-sectional Study

Shanshan Wanga,b, Huaxing Menga,b, Nan Shanga,b, Junhong Guob, Ting Zhanga,
Shuhui Zhanga, Yuqing Zhaoa, Huifang Zhanga, Qinli Zhanga , Qiao Niua*:
niuqiao55@163.com or Niuqiao55@sxmu.edu.cn

a
School of Public Health, Shanxi Medical University, Taiyuan, Shanxi 030001, China.

of
b
Department of Neurology, First Hospital, Shanxi Medical University, Taiyuan, Shanxi 030001,
China.

ro
*Corresponding Author: 56 Xin Jian Nan Lu, Taiyuan 030001, China
Tel: 86-351-4135029
Fax: 86-351-2027943
Email
-p
re
Highlights

 Al levels were measured in human plasma.

lP

 Multi-domain cognition was assessed in in-service Al-exposed workers.

 Negative correlations between the plasma Al levels and the total CDT scores and
executive/visuospatial cognitive performance, and positive correlations between the
plasma Al levels and CDT-position errors were observed.
na

 Dose-response relationships between plasma Al levels and total CDT scores,

executive/visuospatial cognitive performance and CDT-position errors were observed.
ur

ABSTRACT
Jo

Background: Aluminum (Al) exerts neurotoxic effects following overexposure. We previously

reported worse cognitive performance in workers exposed to Al than non-exposed individuals.
Cognition involves multiple domains. The effect of Al on multi-domain cognition has been studied
for decades, but still remains controversial.
Objective: To explore the relationship between plasma Al levels and multi-domain cognitive
performance among in-service aluminum-exposed workers at the SH Aluminum Factory in China
and identify possible types of early cognitive damage caused by exposure to aluminum.
Methods: Eight hundred thirty-one in-service aluminum-exposed workers at the SH Aluminum
Factory in China were investigated. The plasma Al concentrations were measured using inductively
coupled plasma-mass spectrometry (ICP-MS) and served as an internal exposure indicator. The
participants were divided into four subgroups based on the quartiles of plasma Al concentrations,
namely, the Q1, Q2, Q3, and Q4 subgroups. Cognitive function was assessed using the Mini-mental
State Examination (MMSE) and the clock-drawing test (CDT). Multi-domain cognition was
evaluated using sub-tests of the MMSE and the CDT.
Results: The average plasma Al concentration was 15.26 (8.28, 27.02) µg/L. The neurobehavioral
tests showed negative correlations between plasma Al levels and total CDT scores and
executive/visuospatial cognitive performance, and a positive correlation between plasma Al levels
and CDT-position errors (all P<0.05). Additionally, dose-response relationships between higher
plasma Al levels and lower total CDT scores, worse executive/visuospatial cognitive performance,
and more error rates in the CDT-position were observed (all Ptrend<0.05). However, no significant
correlations or trends were observed between plasma Al levels and other cognitive domains (all
P>0.05). The results from the multivariate logistic regression model and restricted cubic spline
models of dose-response relationships were consistent with the results obtained from the general

of
linear model. All potential confounders, such as age, marital status, education, income, type of work,
and smoking and drinking habits, were considered.

ro
Conclusion: Based on the results, aluminum exposure may exert a substantial effect on impairing
executive/visuospatial functions in multi-domain cognition at the early stage, particularly the
identification of spatial positions.
-p
Keywords: Aluminum-exposed workers, Plasma Al, Multi-domain cognition, Neurotoxic effects,
Executive/visuospatial impairment
re
1. Introduction

Aluminum (Al), a ubiquitous environmental and industrial contaminant, enters the human body
lP

through inhalation, ingestion and dermal uptake. Inhalation is the dominant route of occupational
exposure. The modern Al production industry mainly uses bauxite as a raw material, which is then
refined into Al hydroxide, further refined into alumina, and finally reduced to Al by electrolysis
(primary Al production). Large volumes of Al are recovered by smelting scrap Al (secondary Al
na

production). Workers on the production line mainly absorb aluminum by inhaling Al dust, which
has a bioavailability that is approximately 5-20 times higher than Al ingested in drinking water
(Priest, 2004; Krewski et al., 2007).
ur

The body burden of Al is predicted to increase over time (Priest, 2004) and Al exerts an adverse
effect on the nervous system, particularly during the progression of Alzheimer’s disease (AD) (Virk
et al., 2015; Kandimalla et al., 2016; Mirza et al., 2017; Xu et al., 2018) and mild cognitive
Jo

impairment (MCI), an early manifestation of AD (Petersen et al., 1999; Petersen, 2004). Thus, Al
exposure is a widespread concern. Many occupational epidemiological studies have reported worse
performance of workers exposed to Al on various neurobehavioral tests compared with non-exposed
individuals (Buchta et al., 2005; Zawilla et al., 2014). Our study group also reached the same
conclusion in a previous study (Lu et al., 2014). Additionally, a meta-analysis including nine studies
of occupational Al exposure also reported a negative correlation between urinary Al concentrations
and cognitive performance (Meyer-Baron et al., 2007).
Cognition involves multiple domains, including memory, language, attention, and executive/
visuospatial functions (Folstein et al., 1975). The effect of Al on multi-domain cognition has been
studied for decades, but still remains controversial, with some studies observing that memory loss
occurs first and others suggesting that the executive/visuospatial impairment may precede other
impairments (Mandal et al., 2012; Ahmed et al., 2016). Additionally, Al workers, a high-risk group,
may experience cognitive impairment much earlier than the general population. To date, however,
data on multi-domain cognitive impairments in Al workers in the early years of exposure remain
scarce. From the public health perspective, additional epidemiological evidence is needed for high-
risk groups, such as workers who are occupationally exposed to Al, in this research field.
Notably, the evaluation of Al concentrations in the environment is difficult, as the values may
vary in workplace air samples, and even a real-time individual monitoring of environmental
exposure may be less reliable than biological monitoring (Polizzi et al., 2002). Hence, an analysis
of urinary and plasma Al levels, which are presumed to be a good indicator of recent exposure, or
brain and cerebrospinal fluid Al levels, which are postulated to provide a better estimate of the Al
burden in the body and long-term exposure, is needed (Virk et al., 2015; Mohseni et al., 2016).
The aim of this study is to investigate the relationship between plasma Al levels and multi-

of
domain cognition among in-service Al workers at the SH Aluminum Factory, a large factory in
China, and to identify the functions displaying early cognitive damage in these individuals.

ro
2. Materials and Methods

2.1. Materials

2.1.1. Factory information

-p
re
This cross-sectional study was conducted in a large aluminum factory that sends its workers to
a designated hospital for an occupational health examination every year in SH city, China. The
lP

factory uses cryolite-alumina melting salt electrolysis in the modern electrolytic aluminum industry,
and it has several departments, including bauxite, alumina, Al electrolysis, thermoelectricity,
maintenance, transportation and marketing. Al levels in the drinking water were below the threshold
limits (200 µg/L) recommended by the World Health Organization (WHO). Although, we failed to
na

obtain permission from the factory to measure the concentration of aluminum dust in the air of the
workshop, it did not exceed the current national occupational health standard of China (4 mg/m3),
based on the report describing the evaluation of the occupational health evaluation unit responsible
ur

for the plant.

2.1.2. Participants
Jo

Initially, 905 participants were recruited from departments exposed to relatively high levels of
Al fumes and Al dusts from sources including alumina, Al electrolysis, and thermoelectricity. The
response rate of the questionnaire was 95.6%. Subsequently, we excluded 34 individuals based on
the inclusion and exclusion criteria. Finally, 831 individuals were included in our study. The
participants ranged in age from 20 to 59 years, with a length of Al service in the factory for ≥1 year.
In addition, all participants were male, of Han ethnicity and spoke Chinese. All participants wore
personal protective equipment during working hours, including a mouth cover, protective mask,
protective clothing and glasses.
 Exclusion criteria for the study were (1) any conditions that may cause cognitive impairment,
including hepatic or renal disorders, brain trauma, cerebrovascular diseases, epilepsy, Parkinson’s
disease and mental diseases; (2) any family history of dementia among first-degree relatives; (3)
any history of continuous medication with drugs containing aluminum, such as anti-acids or drugs
affecting the central nervous system; (4) workers with known poor vision and hearing.
All participants provided written informed consent. The study was approved by the Ethics and
Human Committees of Shanxi Medical University.

2.2. Methods

2.2.1. A self-designed questionnaire (The questionnaire is presented in Supplementary File 1)

A self-designed questionnaire was administrated through a face-to-face interview, and all the
interviewers were professional neurological physicians or highly trained medical graduate students.

of
Demographic information (age, sex, ethnicity, marital status, education and economic income) and
potential factors that may cause cognitive impairment (life-style habits, such as smoking and

ro
drinking, occupational history, family history, medication and medical history) were collected as
integrated and detailed data, as much as possible.
Men who had smoked ≥1 cigarette per day for ≥6 months and were still smoking were classified
-p
as smokers; Men who had smoked <1 cigarette per day or had stopped smoking for ≥6 months were
classified as ex-smokers, and individuals who had never smoked were classified as non-smokers.
Men who had drank alcohol at least once a week with an alcohol consumption per week ≥40 g
re
[alcohol consumption=volume (ml)*ethanol concentration (%)*0.8] for ≥6 months and were still
drinking were classified as drinkers. Men who had drank less than once a week with <40 g alcohol
consumption per week or had stopped drinking for ≥6 months were classified as ex-drinkers, and
lP

individuals who had never consumed alcohol were classified as non-drinkers.

2.2.2. Cognitive assessment

na

The participants were asked to undergo a face-to-face assessment of the Mini-mental State
Examination (MMSE) and the clock-drawing test (CDT) scales on the same day after completing
the self-designed questionnaire.
ur

MMSE (the scales are presented in Supplementary File 2) is the most commonly applied tool
for cognitive assessments and includes diverse domains covering orientation (orientation to time
and place), registration, recall, language, attention and calculation, and visuospatial construction. A
Jo

meta-analysis revealed a pooled sensitivity of 63.4% and specificity of 65.4% for identifying
individuals with a mild cognitive impairment among a healthy population (Mitchell, 2009). Scores
on the MMSE range from 0 (worst) to 30 (best) points.
The CDT is also quite frequently used as a screening tool for cognitive function due to its ease
of use and time-saving nature. It has multiple scoring methods, with an average sensitivity and
specificity of 85% (Shulman, 2000). Evidence from some studies has confirmed that a simpler
method is better among the multiple clock drawing scoring systems (Mainland et al., 2014).
Therefore, we adopted the ‘4-point scoring system’ with ‘0’ indicating the worst performance and
‘4’ indicating perfect performance in this study. The following instructions were provided to the
participants: ‘Please draw a clock face and place all the numbers on it. Now set the time to 15 past
9.’ Errors in the CDT were first categorized according to the criteria reported by Ricci and
colleagues (Ricci et al., 2016), who considered the factors of the circle (Ci), the numbers (N) and
the hands (H). We also considered the factor of centre of the circle (Ce). Then, we categorized all
18 types of errors observed in the CDT into seven clusters of errors using a new brief scoring method,
namely, distortion, position, omission, perseveration, reversal, code, and length (the details are
presented in Supplementary Files 3.1 and 3.2).
As the CDT mainly reflects executive/visuospatial abilities (Kim et al., 2009; Pinto et al., 2009;
Mitchell et al., 2010), we combined it with the visuospatial domain of the MMSE to assess
executive/visuospatial functions in multi-domain cognition (Schramm et al., 2002; Aprahamian et
al., 2010, Cacho et al., 2010). Domains of the cognitive functions of orientation (0-10 points),
registration (0-3 points), attention and calculation (0-5 points), recall (0-3 points), language (0-8
points), and executive/visuospatial abilities (0-5 points) were discussed.
The cut-off values for all scores were set to one standard deviation below the mean of the group

of
with the lowest plasma Al concentration, and values below the cut-off points were defined as
cognitive impairment, whereas values above the cut-off points were defined as normal cognition

ro
(Crook et al., 1986; Meyer et al., 2002). The MMSE scores were adjusted for education, as they are
impacted by the level of education.

2.2.3. Determination of plasma Al Concentrations -p

The Al exposure of each individual was measured by collecting a plasma sample in the morning
re
after an overnight fast, which was then stored at -80℃ until it was assayed on the same day as the
questionnaire, interview and cognitive evaluation were performed. The plasma Al concentrations
were analysed using ICP-MS (NexlON 300D, PerkinElmer, USA) with a unit of µg/L from plasma
lP

diluted 1:5 with 4% nitric acid after a 24 h incubation at room temperature. Each sample was
randomly measured twice in a blinded manner. The instrument was calibrated after every 10 samples
using an Al standard solution (Agilent, USA). The R2 value, linear range, the lowest limit of
detection (LOD), RSD, and the average recovery were 0.9999, 1-160 µg/L, 0.39 µg/L, 100.29%,
na

and 0.03-0.08%, respectively. The value of the solvent blank was deducted.

2.2.4. Quality control

ur

All test items were scored using the uniform guide language and scoring criteria by strictly
trained neurological physicians and medical graduate students. All experimental products were
Jo

plastic materials that had been soaked in 10% nitric acid for 3 days and then cleaned 7 times with
Milli Q water. The collection and processing of blood samples were conducted by professional
personnel. Plasma Al concentrations were measured using the same instrument and methods. The
data were subjected to double entry and validation.

2.2.5. Statistical analysis

All data were coded and entered using EpiData3.1 software (program design: Jens M. Lauritsen
and Michael Bruus, Odense, Denmark), and analysed with the SAS version 9.4 (SAS Institute, Inc.,
Cary, NC) and the SPSS 22.0 for windows (SPSS, Inc., Chicago, IL). Continuous variables are
presented as means ± standard deviations (SD) or medians and interquartile ranges (IQRs).
Categorical variables are presented as numbers and frequencies (%). Plasma Al concentrations
below the LOD are reported as the LOD divided by the square root of 2 (Hornung et al., 1990). A
multicollinearity diagnosis was applied to assess the presence of collinearity among the covariates,
with a variance inflation factor (VIF) ≤5 and tolerance (TOL) ≥0.2 indicating a lack of a strong
collinearity problem among covariates (Daoud, 2017) (Supplementary File 4). Comparisons of
differences in continuous variables among the four subgroups were performed using a general linear
model (analysis of covariance) and differences between categorical variables were analysed using
the Pearson chi-square test or Fisher’s exact test. Correlations of continuous variables were
determined by calculating partial correlation coefficients, and correlations of categorical variables
was calculated using Kendall’s tau-b correlation test. Natural log transformation was applied to
plasma Al concentrations in the analysis of partial correlations to correct for skewed distributions.
A multivariate logistic regression model and restricted cubic spline models were also used to

of
confirm the results of general linear model. The dose-response relationships between plasma-Al
levels and multi-domain cognition were explored using tests of the linear trends and spline analysis

ro
after adjusting for potential confounders. A P value <0.05 was considered statistically significant.

3. Results

3.1. The distribution of plasma Al concentrations

-p
re
The distribution of plasma Al concentrations in the 831 participants is shown in Table 1. The
average plasma Al concentration was 15.26 (8.28, 27.02) µg/L. The participants were divided into
four groups according to the quartiles of plasma Al concentrations: the Q1 (<8.28 µg/L), Q2 (8.28-
lP

15.26 µg/L), Q3 (15.26-27.02 µg/L), and Q4 (≥27.02 µg/L) subgroups.

3.2. The general characteristics of the participants

na

All participants in this study were male workers who resided in the locality perennially, with
an average age of 40.47±7.89 years, working duration of 16.10±9.65 years, and 11.03±2.10 years
of education. Of the 831 participants, 420 (50.5%) were smokers and 192 (23.1%) drinkers.
ur

The demographic characteristics are summarized in Table 2. No significant differences in age,

working duration, marriage status, education level, and income were observed among the four
groups (all P>0.05). Meanwhile, the type of work and smoking and drinking habits showed
Jo

significant differences (all P<0.05).

3.3. The relationship between the plasma Al concentration and multi-domain cognition

The adjusted mean total MMSE scores, total CDT scores, and the scores for multi-domain
cognition among the four groups are presented in Table 3. After adjusting for age, education, income,
marriage, type of work, smoking and drinking status, the general linear model revealed statistically
significant differences in the total CDT scores (P=0.012), attention and calculation (P=0.020), and
executive/visuospatial abilities (P=0.022) among the four groups. A linear test showed correlations
between higher plasma Al levels and lower total CDT scores (Ptrend=0.001) and worse
executive/visuospatial abilities (Ptrend=0.002). Additionally, the partial correlation analysis also
identified significant inverse correlations between the plasma Al concentrations and total CDT
scores (r=−0.083, P=0.019) and executive/visuospatial abilities (r=−0.072, P=0.042) after
adjusting for potential confounders. However, no significant trends or correlations were observed
between plasma Al concentrations and other cognitive domains (all P>0.05).
As shown in Table 4, we further performed a more detailed assessment of the
executive/visuospatial functions. Although we did not observe statistically significant differences
in all sub-tests of the CDT among the four groups, a dose-response relationship was observed for
the correlation between the plasma Al concentrations and CDT-position (Ptrend=0.014) and CDT-
omission (Ptrend=0.013) errors. As the plasma Al concentrations increased, the CDT-position and
CDT-omission error rates increased. Positive corrections were also observed between plasma Al
concentrations and CDT-position (r=0.078, P=0.014) and CDT-omission (r=0.078, P=0.013) errors.
However, no significant trends or correlations were observed between plasma Al concentrations and

of
the other sub-tests of the CDT (all P>0.05).
We subsequently used a multivariate logistic regression model to assess diverse domains of

ro
cognition and confirm the relationship between plasma Al concentrations and multi-domain
cognition (Table 5). The results of the multivariate logistic regression model were consistent with
the results of the general linear model. After adjusting for potential confounders, dose-response
-p
relationships between plasma Al concentrations and CDT performance (Ptrend<0.001),
executive/visuospatial abilities (Ptrend=0.019), and CDT-position errors (Ptrend=0.026) were
observed. As the plasma Al concentrations increased, the risk of the cognitive impairment increased.
re
However, no significant trends were observed for the correlation between plasma Al concentrations
and the CDT-omission errors after adjusting for potential confounders (Ptrend=0.178). Additionally,
no significant trends for the correlations between plasma Al concentrations and other cognitive
lP

functions were observed (all Ptrend>0.05, Supplementary File 5).

Next, we analysed plasma Al concentrations as a continuous variable using restricted cubic
spline models to further confirm the relationship between plasma Al concentrations and multi-
domain cognition. As shown in Fig. 1, as the plasma Al levels increased, ORs for CDT performance,
na

executive/visuospatial abilities, CDT-position errors, and language impairment displayed increasing

trends (P for the non-linear association<0.05). However, no significant trends in correlations
between plasma Al concentrations and other cognitive functions were observed (all P for the non-
ur

linear association>0.05, Supplementary File 6).

After further stratifying the participants by age, dose-response relationships between plasma
Al concentrations and CDT performance (Ptrend<0.001), executive/visuospatial abilities
Jo

(Ptrend=0.001), and CDT-position errors (Ptrend=0.010) were only observed in the group aged ≥40
years, after adjusting for potential confounders (Table 6). However, correlations between plasma Al
concentrations and other cognitive domains were not observed in any of the two age groups (all
Ptrend>0.05, Supplementary File 7).

Table 1

Distribution of plasma Al concentrations in the 831 participants.

Number of Percentiles
Analytes
participants 5th 25th 50th 75th 95th
 Total plasma Al 831 < LOD 8.28 15.26 27.02 78.54

concentration

Stratification by age

<40 years 315 < LOD 7.96 15.87 33.62 88.54

≥40 years 514 < LOD 8.42 14.99 24.75 70.73

The LOD for the plasma Al concentrations was 0.39 µg/L.

Missing value—age: 2

Table 2

Demographic characteristics of the 831 participants in the studied groups.

Plasma Al levels (µg/L)

Characteristics P a or P b
Q1 (< 8.28) Q2 (8.28-15.26) Q3 (15.26-27.02) Q4 (≥ 27.02)

Number of participants 207 208 208 208

of
Age (years, mean ± SD) 40.25±8.32 41.52±7.87 40.46±7.96 39.66±7.33 0.110a

Age group (years), no. (%)

20-29 31 (15.0) 23 (11.1) 28 (13.5) 24 (11.5) 0.257b

ro
30-39 51 (24.6) 44 (21.2) 52 (25.0) 62 (29.8)

40-49 101 (48.8) 112 (53.8) 106 (51.0) 108 (51.9)

50-59

Working duration (years, mean ± SD)

Marital status, no. (%)

23 (11.1)

16.37±9.87
29 (13.9)

17.36±9.61
-p 22 (10.6)

15.84±10.31
13 (6.3)

14.85±8.63 0.063a
re
No 12 (5.8) 7 (3.4) 11 (5.3) 7 (3.4) 0.498b

Yes 195 (94.2) 201 (96.6) 197 (94.7) 201 (96.6)

Education (years, mean ± SD) 11.16±2.14 11.08±1.96 10.95±2.24 10.93±2.05 0.618a

lP

Education group (years), no. (%)

≤9 76 (36.7) 77 (37.0) 86 (41.3) 84 (40.4) 0.782b

9-12 107 (51.7) 111 (53.4) 98 (47.1) 105 (50.5)

na

> 12 24 (11.6) 19 (9.1) 24 (11.5) 18 (8.7)

Income, no. (%)

<1000 RMB of per capita monthly income 60 (29.0) 68 (32.7) 63 (30.3) 60 (28.8) 0.771b

≥1000 RMB of per capita monthly income 143 (69.1) 132 (63.5) 137 (65.9) 140 (67.3)
ur

Type of work, no. (%)

Electrolytic workers 58 (28.0) 61 (29.3) 83 (39.9) 89 (42.8) 0.002b

Jo

Non-electrolytic workers 149 (72.0) 147 (70.7) 125 (60.1) 119 (57.2)

Smoking status, no. (%)

Non-smokers 83 (40.1) 70 (33.7) 84 (40.4) 151 (72.6) <0.001b

Ex-smokers 5 (2.4) 2 (1.0) 6 (2.9) 10 (4.8)

Smokers 119 (57.5) 136 (65.4) 118 (56.7) 47 (22.6)

Drinking status, no. (%)

Non-drinkers 123 (59.4) 111 (53.4) 118 (56.7) 144 (69.2) 0.012b

Ex-drinkers 36 (17.4) 34 (16.3) 44 (21.2) 29 (13.9)

Drinkers 48 (23.2) 63 (30.3) 46 (22.1) 35 (16.8)

 P a: value determined using Student’s t-test.

P b: value determined using Pearson chi-square test.

Missing value—Q1 (age: 1, income: 4); Q2 (education: 1, income: 8); Q3 (income: 8); Q4 (age: 1, education: 1, income: 8).

Table 3

Comparisons of total MMSE scores, total CDT scores, and scores for multi-domain cognition among the four groups.

Mean plasma Al levels (µg/L) ± SD

Variables Q1 (< 8.28) Q2 (8.28-15.26) Q3 (15.26-27.02) Q4 (≥ 27.02) Pa Ptrend r Pb

Number of participants 207 208 208 208

Total MMSE scores 27.62±0.24 27.49±0.25 27.71±0.24 27.56±0.25 0.665 0.942 -0.004 0.912

Total CDT scores 2.83±0.12 2.71±0.12 2.64±0.12 2.52±0.12 0.012 0.001 -0.083 0.019

Multi-domain cognition

of
Orientation 9.65±0.10 9.63±0.10 9.63±0.10 9.77±0.10 0.233 0.145 0.023 0.523

Registration 2.99±0.02 2.99±0.02 3.00±0.02 3.00±0.02 0.696 0.239 0.017 0.633

ro
Recall 2.29±0.11 2.35±0.11 2.39±0.11 2.37±0.12 0.652 0.300 0.033 0.351

Attention and calculation 4.40±0.12 4.28±0.13 4.49±0.12 4.21±0.13 0.020 0.251 -0.041 0.245

Language 7.55±0.09 7.51±0.09 7.51±0.09 7.46±0.10 0.687 0.259 -0.025 0.489

Executive/visuospatial

abilities
3.57±0.14 3.44±0.14 3.33±0.14
-p 3.26±0.14 0.022 0.002 -0.072 0.042
re
P a: Analysis of covariance for differences among the four groups, after adjusting for age, education, income, marriage, type of work, and

smoking and drinking statuses.

Ptrend: Analysis of linear trends, after adjusting for age, education, income, marriage, type of work, and smoking and drinking statuses.
lP

P b: Analysis of partial correlations between plasma Al levels and total MMSE scores, total CDT scores, and scores for multi-domain

cognition, after adjusting for age, education, income, marriage, type of work, and smoking and drinking statuses.
na
ur
Jo
 Table 4

Comparisons of scores on sub-tests of the CDT among the four groups.

Plasma Al levels (µg/L), no. (%)

Variables P a or P b Ptrend r Pc
Q1 (< 8.28) Q2 (8.28-15.26) Q3 (15.26-27.02) Q4 (≥ 27.02)

Number of participants 207 208 208 208

Distortion errors 3 (1.4) 4 (1.9) 5 (2.4) 2 (1.0) 0.747b 0.812 -0.008 0.804

Position errors 52 (25.1) 58 (27.9) 69 (33.2) 73 (35.1) 0.099a 0.014 0.078 0.014

Perseveration errors 25 (12.1) 34 (16.3) 23 (11.1) 35 (16.8) 0.218a 0.407 0.026 0.408

Omission errors 33 (15.9) 48 (23.1) 49 (23.6) 55 (26.4) 0.067a 0.013 0.078 0.013

Same length errors 74 (35.7) 76 (36.5) 86 (41.3) 83 (39.9) 0.597a 0.253 0.036 0.253

Reversal errors 2 (1.0) 1 (0.5) 2 (1.0) 2 (1.0) 0.961b 1.000 0.005 0.868

Code errors 0 (0.0) 2 (1.0) 1 (0.5) 3 (1.4) 0.528b 0.203 0.046 0.144

of
P a: value determined using the Pearson chi-square test.

P b: value determined using Fisher’s exact test.

Ptrend: value determined using the linear-by-linear test.

ro
P c: value determined using Kendall’s tau-b correlation test.

-p
re
lP
na
ur
Jo
Table 5

The associations between plasma Al concentrations and the risks of cognitive impairments and multi-domain cognitive impairments among

the 831 occupationally exposed Al workers.

OR (95% CI)

N (cognitive
Model 1 Model 2 Model 3
impairment/normal)

Total CDT scores

Q1 44/163 1.00 (Reference) 1.00 (Reference) 1.00 (Reference)

Q2 60/148 1.50 (0.96-2.35) 1.49 (0.95-2.34) 1.35 (0.85-2.16)

Q3 70/138 1.88 (1.21-2.92) 1.87 (1.20-2.90) 1.69 (1.06-2.67)

Q4 83/125 2.46 (1.60-3.79) 2.42 (1.57-3.73) 2.58 (1.60-4.17)

Ptrend <0.001 <0.001 <0.001

Executive/visuospatial

of
abilities

Q1 61/146 1.00 (Reference) 1.00 (Reference) 1.00 (Reference)

Q2 78/130 1.44 (0.95-2.16) 1.41 (0.94-2.13) 1.32 (0.86-2.03)

ro
Q3 87/121 1.72 (1.15-2.58) 1.71 (1.14-2.56) 1.50 (0.97-2.30)

Q4 89/119 1.79 (1.19-2.69) 1.77 (1.18-2.66) 1.82 (1.16-2.87)

Ptrend

CDT-position errors

Q1 52/155
0.022 0.025
-p 0.019

1.00 (Reference) 1.00 (Reference) 1.00 (Reference)

re
Q2 58/150 1.15 (0.75-1.78) 1.13 (0.73-1.76) 1.10 (0.70-1.73)

Q3 69/139 1.48 (0.97-2.27) 1.47 (0.96-2.25) 1.34 (0.86-2.09)

Q4 73/135 1.61 (1.06-2.46) 1.59 (1.04-2.43) 1.68 (1.05-2.67)

lP

Ptrend 0.033 0.038 0.026

CDT-omission errors

Q1 33/174 1.00 (Reference) 1.00 (Reference) 1.00 (Reference)

Q2 48/160 1.58 (0.97-2.59) 1.59 (0.97-2.61) 1.48 (0.89-2.47)

na

Q3 49/159 1.63 (1.00-2.66) 1.62 (0.99-2.65) 1.40 (0.84-2.33)

Q4 55/153 1.90 (1.17-3.07) 1.84 (1.13-2.99) 1.61 (0.95-2.72)

Ptrend 0.037 0.055 0.178

ur
Jo

Model 1: unadjusted.

Model 2: adjusted for age.

Model 3: adjusted for age, education, income, marriage, type of work, and smoking and drinking statuses.
 of
ro
-p
Fig.1. Restricted cubic spline models representing the associations of plasma Al concentrations with ORs for CDT performance,

executive/visuospatial abilities, CDT-position errors, and language impairment, after adjustment for age, education, income, marriage, type
re
of work, and smoking and drinking statuses. The reference group comprises individuals with normal cognition, dashed lines represent 95%

CIs, and the red knots represent plasma Al concentrations at the 25th, 50th, and 75th percentiles.
lP
na
ur
Jo
Table 6

The associations of plasma Al concentrations with the risk of cognitive impairment and multi-domain cognitive impairments among 831

occupational workers after stratification by age.

Age <40 years Age ≥40 years

N (cognitive N (cognitive

impairment/ AOR (95% CI) Ptrend impairment/ AOR (95% CI) Ptrend

normal) normal)

Total CDT

scores

Q1 20/62 1.00 (Reference) 24/100 1.00 (Reference)

Q2 27/40 1.53 (0.72-3.23) 33/108 1.28 (0.69-2.36)

Q3 26/54 1.03 (0.49-2.16) 0.281 44/84 2.26 (1.25-4.11) <0.001

Q4 27/59 1.63 (0.78-3.40) 55/66 3.49 (1.85-6.59)

of
Executive/visuo-

spatial abilities

ro
Q1 23/59 1.00 (Reference) 38/86 1.00 (Reference)

Q2 32/35 1.82 (0.88-3.73) 43/98 1.00 (0.58-1.71)

Q3 35/45 1.43 (0.71-2.89) 0.714 54/74 1.77 (1.03-3.04) 0.001

Q4

CDT-position

errors
28/58 1.40 (0.69-2.83) -p 59/62 2.44 (1.35-4.41)
re
Q1 23/59 1.00 (Reference) 29/95 1.00 (Reference)

Q2 22/45 1.07 (0.51-2.25) 36/105 1.15 (0.64-2.04)

Q3 24/56 0.79 (0.38-1.65) 0.728 45/83 1.75 (1.00-3.09) 0.01

lP

Q4 23/63 1.13 (0.55-2.31) 49/72 2.22 (1.21-4.10)

CDT-omission

errors

Q1 16/66 1.00 (Reference) 17/107 1.00 (Reference)

na

Q2 20/47 1.23 (0.55-2.76) 28/113 1.72 (0.87-3.40)

Q3 21/59 1.05 (0.48-2.32) 0.475 28/100 1.78 (0.90-3.55) 0.196

Q4 22/64 1.37 (0.63-2.97) 32/89 1.95 (0.94-4.02)

ur
Jo

AOR: adjusted for education, income, marriage, type of work, and smoking and drinking statuses.

Missing values—Q1 (age: 1); Q4 (age: 1).

4. Discussion

Aluminum is widespread in the environment, but poorly absorbed by the human body, and
excessive Al concentrations may cause toxicity. Based on this characteristic, the total burden of Al
in healthy individuals is 30-50 mg (ToxGuideTM for Aluminum, CAS# 7429-90-5), towards which
the brain makes an important contribution. However, brain tissue is impossible to obtain and
cerebrospinal fluid is difficult to obtain from normal populations. Plasma Al concentrations, an
indicator of the recent burden of Al in the body, was estimated as a proxy for the brain in the present
study. The median plasma Al concentrations following occupational exposure ranged from 9.6-33.5
µg/L in previous studies (Buchta et al., 2005; Kraus et al., 2006; Kiesswetter et al., 2007), and Klotz
et al. (Klotz et al., 2017) reported a concentration of approximately 13 µg/L in plasma when early
changes in neuropsychological tests were observed in occupationally exposed workers. In the
present study, the median plasma Al concentration (15.26 µg/L) of Al workers was at an
intermediate level compared to previous values; it was higher than in the report from Klotz (13 µg/L)
and in non-exposed individuals (<10 µg/L in plasma) in other studies (Kiesswetter et al., 2007).
Occupational epidemiologists have studied the relationship between Al exposure and cognitive
impairment, which is presumed be the initial symptom of dementia, for decades. In recent years,
multi-domain impairments have attracted increasing attention. Although various options for multi-
domain tests are available, the MMSE, the most widely used cognitive screening tool with 11 items

of
covering 7 domains, namely, orientation (orientation to time and place), registration, recall,
language, attention and calculation, and visuospatial construction skills, is recommended (Mitchell

ro
et al., 2010; Trzepacz et al., 2015). Nevertheless, it has some obvious limitations. The most
remarkable problem is its low sensitivity, which is attributed to a greater emphasis on verbal tests
but the absence of tests for executive/visuospatial functions (Schramm et al., 2002; Trzepacz et al.,
-p
2015). However, the impairments in the latter functions may also occur in patients with the early
stage of AD (Schramm et al., 2002; Ahmed et al., 2016). Hence, the CDT, which has been mainly
used to reflect executive/visuospatial functions (Kim et al., 2009; Mitchell et al., 2010), was
re
combined with the visuospatial assessment in MMSE to bridge this gap.
The cognitive status of individuals with or without impairments was determined based on the
cut-off points for the neurobehavioral scales. However, the cut-off points for the MMSE, CDT or
lP

multi-domain cognition are inconsistent (Mitchell, 2009). Therefore, one standard deviation below
the mean of the group with the lowest plasma Al concentration was set as the cut-off value in this
study (Meyer et al., 2002). Additionally, evidence from some cross-sectional studies has confirmed
that performance on cognitive tests is substantially affected by educational levels (Dick et al., 1984;
na

Schwamm et al., 1987; Tombaugh et al., 1992). Thus, the cut-off points for the MMSE used to
determine cognitive impairment in the present study were adjusted for educational levels. Regarding
the selection of covariates, we chose age, education, income, marriage, type of work, and smoking
ur

and drinking statuses as covariates in the present study, based on previous research from our study
group and some other related publications, which suggest that these factors may be correlated with
cognition(Lu et al., 2013; Petersen et al., 2014; Hu et al., 2017; Escher et al., 2019).
Jo

Visuospatial function is associated with temporal-parietal regions, and executive function may
play a key role in the successful performance of visuospatial tests (Ahmed et al., 2016). Lee et al.
(Lee et al., 2008) reported a close relationship between poor CDT performance in patients with AD
and changes in the right parietal cortex using FDG-PET. According to Eknoyan et al. (Eknoyan et
al., 2012), frontoparietal circuits, which may play a role in coordinating the processing of
visuospatial information and the executive functions of a clock, and frontostriatal circuits are needed
to accurately plan and draw a clock face. Thus, the CDT would be expected to reflect the main
domains of executive/visuospatial abilities, which are mediated by the parietal cortex. In the present
study, higher plasma Al levels were associated with worse performance on the CDT. Similar results
were obtained from multivariate logistic regression analyses and restricted cubic spline models
analyses, where we observed an increasing risk of cognitive impairment as the plasma Al levels
increased. This result is consistent with the findings from a previous study by Polizzi et al. (Polizzi
et al., 2002). After a further analysis of multi-domain cognition, we also identified a dose-response
relationship between plasma Al levels and executive/visuospatial cognitive impairment, but not in
other domains in the early stage. Zawilla et al. (Zawilla et al., 2014) reported worse performance on
tests of memory, language naming, verbal fluency (reflect executive functions & verbal memory)
and visuospatial functions in 54 Al-exposed workers compared with their 51 reference controls,
while we only observed a negative correlation between Al exposure and executive/visuospatial
functions in this study. The potential explanation for this difference is the differences in types of Al
particles, geographical areas, and habits. Moreover, we analysed a larger sample size of 831
participants, and the previous study reported a higher age level of 46 years. Some other conflicting
reports also exist (Letzel et al., 2000; Buchta et al., 2003; Meyer-Baron et al., 2007; Yang et al.,
2015). Hence, the evidence from this study that the executive/visuospatial functions may be the

of
main domain affected by Al exposure in the early stage requires further study with older workers
and more accurate imaging data.

ro
A qualitative analysis of CDT data is considered a deeper reflection of patients’ cognitive status
than a quantitative analysis (Cahn et al., 1996; Parsey et al., 2011). Patients with cognitive
impairments produce more errors on sub-tests of the CDT than normal individuals, such as improper
-p
numbers and hands and an incorrect circle (Lee et al., 2008). In the present study, the main types of
errors on the sub-tests of the CDT were same-length errors, position errors, and omission errors.
However, we only observed a dose-response relationship between plasma Al levels and position
re
errors, suggesting that exposure to Al may exert a major early adverse effect on fine motor control,
the correct recognition of spatial positions and visuospatial planning. Regrettably, relevant reports
for a comparison of these effects are currently unavailable.
lP

Since cognition is strongly correlated with age (Petersen et al., 2014), we stratified participants
into two subgroups by age to further explore the relationship between plasma Al levels and
executive/visuospatial cognitive impairment. Interestingly, although we did not observe a
relationship between age and the executive/visuospatial cognitive impairment (data not shown), an
na

older age may increase the risk effect of aluminium exposure on executive/visuospatial impairment.
For that after stratifying participants by age, the group aged ≥40 years, but not the group aged <40
years, showed a positive correlation between plasma Al concentrations and the risk of
ur

executive/visuospatial cognitive impairment.

The greatest innovation of this study is the combined use of the CDT and visuospatial tests of
the MMSE, which may increase the sensitivity of the assessment of multi-domain impairments
Jo

(Schramm et al., 2002). The more detailed analysis of CDT is also a highlight. Moreover, the present
study provides large-scale occupational epidemiological data. Of course, some limitations must be
noted. First, further evidence, such as brain magnetic resonance imaging or electroencephalogram,
was not available to identify specific functional areas of the brain associated with the cognitive
impairment. We will perform these analyses in our next stage of research. Second, we lack
additional evidence describing the health of individuals who were not exposed. Third, plasma Al
concentrations may not accurately reflect long-term exposure and the Al burden in the body, but
only serve as a good indicator of recent aluminum exposure, since the plasma is only a system for
transporting Al. Further research is needed to address these limitations.
5. Conclusions

A negative correlation is observed between plasma Al levels and cognitive performance. In the
early stage, exposure to Al may exert a major effect on impairing executive/visuospatial functions
in multi-domain cognition, particularly the identification of spatial positions.

Financial conflicts of interest

The authors have no actual or potential conflicts of interest to declare.

Declarations of interest

None.

of
Acknowledgments

ro
We would like to thank the National Natural Science Foundation of China (81430078 and
81703202) for providing financial support for this study. We are grateful to the participants who
gave their time to help us perform the study. Furthermore, we would like to thank all the medical
-p
professionals for their assistance with the collection of blood samples.
re
lP
na
ur
Jo
References

Ahmed, S., Brennan, L., Eppig, J., Price, C.C., Lamar, M., Delano-Wood, L., et al., 2016. Visuoconstructional
Impairment in Subtypes of Mild Cognitive Impairment. Appl Neuropsychol Adult 23, 43-52. https://doi.org/
10.1080/23279095.2014.1003067.
Aprahamian, I., Martinelli, J.E., Neri, A.L., Yassuda, M.S., 2010. The accuracy of the Clock Drawing Test compared
to that of standard screening tests for Alzheimer's disease: results from a study of Brazilian elderly with
heterogeneous educational backgrounds. Int Psychogeriatr 22, 64-71. https://doi.org/10.1017/
S1041610209991141. Epub 2009 Oct 9.
Buchta, A.M., Kiesswetter, B.E., Schaper, B.M., Zschiesche, C.W., Schaller, D.K., Kuhlmann, A.A., Letzel, A.S.,
2005. Neurotoxicity of exposures to aluminium welding fumes in the truck trailer construction industry.
Environ Toxicol Pharmacol 19, 677-685. https://doi.org/10.1016/j.etap.2004.12.036.
Buchta, M., Kiesswetter, E., Otto, A., Schaller, K.H., Seeber, A., Hilla, W., et al., 2003. Longitudinal study
examining the neurotoxicity of occupational exposure to aluminium-containing welding fumes. Int Arch Occup

of
Environ Health 76, 539-548. https://doi.org/10.1007/s00420-003-0450-9.
Cacho, J., Benito-Leon, J., Garcia-Garcia, R., Fernandez-Calvo, B., Vicente-Villardon, J.L., Mitchell, A.J., 2010.

ro
Does the combination of the MMSE and clock drawing test (mini-clock) improve the detection of mild
Alzheimer's disease and mild cognitive impairment? J Alzheimers Dis 22, 889-896.
https://doi.org/10.3233/JAD-2010-101182.
-p
Cahn, D.A., Salmon, D.P., Monsch, A.U., Butters, N., Wiederholt, W.C., Corey-Bloom, J., Barrett-Connor, E., 1996.
Screening for dementia of the alzheimer type in the community: the utility of the Clock Drawing Test. Arch
Clin Neuropsychol 11, 529-539. https://doi.org/10.1016/0887-6177(95)00041-0.
re
Crook, T., Bartus, R.T., Ferris, S.H., Whitehouse, P., Cohen, G.D., Gershon, S., 1986. Age‐ associated memory
impairment: Proposed diagnostic criteria and measures of clinical change — report of a national institute of
mental health work group. Dev Neuropsychol 2, 261-276. https://doi.org/10.1080/87565648609540348.
lP

Daoud, J.I., 2017. Multicollinearity and Regression Analysis. Journal of Physics: Conference Series 949, 012009.
https://doi.org/10.1088/1742-6596/949/1/012009.
Dick, J.P., Guiloff, R.J., Stewart, A., Blackstock, J., Bielawska, C., Paul, E.A., Marsden, C.D., 1984. Mini-mental
state examination in neurological patients. J Neurol Neurosurg Psychiatry 47, 496-499. https://doi.org/
na

10.1136/jnnp.47.5.496.
Eknoyan, D., Hurley, R.A., Taber, K.H., 2012. The clock drawing task: common errors and functional neuroanatomy.
J Neuropsychiatry Clin Neurosci 24, 260-265. https://doi.org/10.1176/appi.neuropsych.12070180.
ur

Escher, C., Jessen, F., 2019. [Prevention of cognitive decline and dementia by treatment of risk factors]. Nervenarzt
90, 921-925. https://doi.org/10.1007/s00115-019-0759-6.
Folstein, M.F., Folstein, S.E., McHugh, P.R., 1975. "Mini-mental state". A practical method for grading the cognitive
Jo

state of patients for the clinician. J Psychiatr Res 12, 189-198. https://doi.org/10.1016/0022-3956(75)90026-6.
Hornung, R.W., Reed, L.D., 1990. Estimation of Average Concentration in the Presence of Nondetectable Values.
App Occup Env Hyg 5, 46-51. https://doi.org/10.1080/1047322X.1990.10389587.
Hu, C., Yu, D., Sun, X., Zhang, M., Wang, L., Qin, H., 2017. The prevalence and progression of mild cognitive
impairment among clinic and community populations: a systematic review and meta-analysis. Int Psychogeriatr
29, 1595-1608. https://doi.org/10.1017/S1041610217000473.
Kandimalla, R., Vallamkondu, J., Corgiat, E.B., Gill, K.D., 2016. Understanding Aspects of Aluminum Exposure in
Alzheimer's Disease Development. Brain Pathol 26, 139-154. https://doi.org/10.1111/bpa.12333.
Kiesswetter, E., Schaper, M., Buchta, M., Schaller, K.H., Rossbach, B., Scherhag, H., et al., 2007. Longitudinal
study on potential neurotoxic effects of aluminium: I. Assessment of exposure and neurobehavioral
performance of Al welders in the train and truck construction industry over 4 years. Int Arch Occup Environ
Health 81, 41-67. https://doi.org/10.1007/s00420-007-0191-2.
Kim, Y.S., Lee, K.M., Choi, B.H., Sohn, E.H., Lee, A.Y., 2009. Relation between the clock drawing test (CDT) and
structural changes of brain in dementia. Arch Gerontol Geriatr 48, 218-221. https://doi.org/10.1016/
j.archger.2008.01.010.
Klotz, K., Weistenhã¶Fer, W., Neff, F., Hartwig, A., Van, T.C., Drexler, H., 2017. The Health Effects of Aluminum
Exposure. Dtsch Arztebl Int. 114, 653-659. https://doi.org/10.3238/arztebl.2017.0653.
Kraus, T., Schaller, K.H., Angerer, J., Hilgers, R.D., Letzel, S., 2006. Aluminosis--detection of an almost forgotten
disease with HRCT. J Occup Med Toxicol 1, 4. https://doi.org/10.1186/1745-6673-1-4.
Krewski, D., Yokel, R.A., Nieboer, E., Borchelt, D., Cohen, J., Harry, J., et al., 2007. Human health risk assessment
for aluminium, aluminium oxide, and aluminium hydroxide. J Toxicol Environ Health B Crit Rev 10 Suppl 1,
1-269. https://doi.org/10.1080/10937400701597766.

of
Lee, D.Y., Seo, E.H., Choo, I.H., Kim, S.G., Lee, J.S., Lee, D.S., et al., 2008. Neural correlates of the Clock Drawing
Test performance in Alzheimer's disease: a FDG-PET study. Dement Geriatr Cogn Disord 26, 306-313.

ro
https://doi.org/10.1159/000161055.
Lee, K.S., Kim, E.A., Hong, C.H., Lee, D.W., Oh, B.H., Cheong, H.K., 2008. Clock drawing test in mild cognitive
impairment: quantitative analysis of four scoring methods and qualitative analysis. Dement Geriatr Cogn Disord
26, 483-489. https://doi.org/10.1159/000167879. -p
Letzel, S., Lang, C.J., Schaller, K.H., Angerer, J., Fuchs, S., Neundorfer, B., Lehnert, G., 2000. Longitudinal study
of neurotoxicity with occupational exposure to aluminum dust. Neurology 54, 997-1000. https://doi.org/
re
10.1212/WNL.54.4.997.
Lu, X.T., Liang, R.F., Jia, Z.J., Wang, H., Song, W.F., Li, Q, Y., Niu, Q., 2013. [Effect of aluminum exposure on
cognitive function in electrolytic workers and its influential factors]. Zhonghua Lao Dong Wei Sheng Zhi Ye
lP

Bing Za Zhi 31, 113-116. https://doi.org/10.3760/cma.j.issn.1001-9391.2013.02.009.

Lu, X., Liang, R., Jia, Z., Wang, H., Pan, B., Zhang, Q., Niu, Q., 2014. Cognitive disorders and tau-protein expression
among retired aluminum smelting workers. J Occup Environ Med 56, 155-160. https://doi.org/
10.1097/JOM.0000000000000100.
na

Mainland, B.J., Amodeo, S., Shulman, K.I., 2014. Multiple clock drawing scoring systems: simpler is better. Int J
Geriatr Psychiatry 29, 127-136. https://doi.org/10.1002/gps.3992.
Mandal, P.K., Joshi, J., Saharan, S., 2012. Visuospatial perception: an emerging biomarker for Alzheimer's disease.
ur

J Alzheimers Dis 31 Suppl 3, S117-135. https://doi.org/10.3233/JAD-2012-120901.

Meyer-Baron, M., Schaper, M., Knapp, G., van Thriel, C., 2007. Occupational aluminum exposure: evidence in
support of its neurobehavioral impact. Neurotoxicology 28, 1068-1078. https://doi.org/10.1016/
Jo

j.neuro.2007.07.001.
Meyer, J., Xu, G., Thornby, J., Chowdhury, M., Quach, M., 2002. Longitudinal analysis of abnormal domains
comprising mild cognitive impairment (MCI) during aging. J Neurol Sci 201, 19-25. https://doi.org/ 10.1016/
s0022-510x(02)00159-4.
Mirza, A., King, A., Troakes, C., Exley, C., 2017. Aluminium in brain tissue in familial Alzheimer's disease. J Trace
Elem Med Biol 40, 30-36. https://doi.org/10.1016/j.jtemb.2016.12.001.
Mitchell, A.J., 2009. A meta-analysis of the accuracy of the mini-mental state examination in the detection of
dementia and mild cognitive impairment. J Psychiatr Res 43, 411-431. https://doi.org/10.1016/
j.jpsychires.2008.04.014.
Mitchell, A.J., Malladi, S., 2010. Screening and case-finding tools for the detection of dementia. Part II: evidence-
based meta-analysis of single-domain tests. Am J Geriatr Psychiatry 18, 783-800. https://doi.org/10.1097/
JGP.0b013e3181cdecd6.
Mitchell, A.J., Malladi, S., 2010. Screening and case finding tools for the detection of dementia. Part I: evidence-
based meta-analysis of multidomain tests. Am J Geriatr Psychiatry 18, 759-782. https://doi.org/10.1097/
JGP.0b013e3181cdecb8.
Mohseni, H.K., Chettle, D.R., 2016. A History of In Vivo Neutron Activation Analysis in Measurement of
Aluminum in Human Subjects. J Alzheimers Dis 50, 913-926. https://doi.org/10.3233/JAD-150595.
Parsey, C.M., Schmitter-Edgecombe, M., 2011. Quantitative and qualitative analyses of the clock drawing test in
mild cognitive impairment and Alzheimer disease: evaluation of a modified scoring system. J Geriatr Psychiatry
Neurol 24, 108-118. https://doi.org/10.1177/0891988711402349.
Petersen, R.C., 2004. Mild cognitive impairment as a diagnostic entity. J Intern Med 256, 183-194. https://doi.org/
10.1111/j.1365-2796.2004.01388.x.
Petersen, R.C., Caracciolo, B., Brayne, C., Gauthier, S., Jelic, V., Fratiglioni, L., 2014. Mild cognitive impairment:

of
a concept in evolution. J Intern Med 275, 214-228. https://doi.org/10.1111/joim.12190.
Petersen, R.C., Smith, G.E., Waring, S.C., Ivnik, R.J., Tangalos, E.G., Kokmen, E., 1999. Mild cognitive impairment:

ro
clinical characterization and outcome. Arch Neurol 56, 303-308. https://doi.org/10.1089/109493102321018169.
Pinto, E., Peters, R., 2009. Literature review of the Clock Drawing Test as a tool for cognitive screening. Dement
Geriatr Cogn Disord 27, 201-213. https://doi.org/10.1159/000203344.
-p
Polizzi, S., Pira, E., Ferrara, M., Bugiani, M., Papaleo, A., Albera, R., Palmi, S., 2002. Neurotoxic effects of
aluminium among foundry workers and Alzheimer's disease. Neurotoxicology 23, 761-774. https://doi.org/
10.1016/S0161-813X(02)00097-9.
re
Priest, N.D., 2004. The biological behaviour and bioavailability of aluminium in man, with special reference to
studies employing aluminium-26 as a tracer: review and study update. J Environ Monit 6, 375-403.
https://doi.org/ 10.1039/b314329p.
lP

Ricci, M., Pigliautile, M., D'Ambrosio, V., Ercolani, S., Bianchini, C., Ruggiero, C., et al., 2016. The clock drawing
test as a screening tool in mild cognitive impairment and very mild dementia: a new brief method of scoring
and normative data in the elderly. Neurol Sci 37, 867-873. https://doi.org/10.1007/s10072-016-2480-6.
Schramm, U., Berger, G., Muller, R., Kratzsch, T., Peters, J., Frolich, L., 2002. Psychometric properties of Clock
na

Drawing Test and MMSE or Short Performance Test (SKT) in dementia screening in a memory clinic
population. Int J Geriatr Psychiatry 17, 254-260. https://doi.org/10.1002/gps.585.
Schwamm, L.H., Van Dyke, C., Kiernan, R.J., Merrin, E.L., Mueller, J., 1987. The Neurobehavioral Cognitive Status
ur

Examination: comparison with the Cognitive Capacity Screening Examination and the Mini-Mental State
Examination in a neurosurgical population. Ann Intern Med 107, 486-491. https://doi.org/10.7326/0003-4819-
107-4-486.
Jo

Shulman, K.I., 2000. Clock-drawing: is it the ideal cognitive screening test? Int J Geriatr Psychiatry 15, 548-561.
https://doi.org/10.1002/1099-1166(200006)15:6<548::aid-gps242>3.0.co;2-u.
Tombaugh, T.N. ,McIntyre, N.J., 1992. The mini-mental state examination: a comprehensive review. J Am Geriatr
Soc 40, 922-935. https://doi.org/10.1111/j.1532-5415.1992.tb01992.x.
Trzepacz, P.T., Hochstetler, H., Wang, S., Walker, B., Saykin, A.J., 2015. Relationship between the Montreal
Cognitive Assessment and Mini-mental State Examination for assessment of mild cognitive impairment in older
adults. BMC Geriatr 15, 107. https://doi.org/10.1186/s12877-015-0103-3.
Virk, S.A., Eslick, G.D., 2015. Aluminum Levels in Brain, Serum, and Cerebrospinal Fluid are Higher in
Alzheimer's Disease Cases than in Controls: A Series of Meta-Analyses. J Alzheimers Dis 47, 629-638.
https://doi.org/10.3233/JAD-150193.
Xu, L., Zhang, W., Liu, X., Zhang, C., Wang, P., Zhao, X., 2018. Circulatory Levels of Toxic Metals (Aluminum,
Cadmium, Mercury, Lead) in Patients with Alzheimer's Disease: A Quantitative Meta-Analysis and Systematic
Review. J Alzheimers Dis 62, 361-372. https://doi.org/10.3233/JAD-170811.
Yang, X., Yuan, Y., Lu, X., Yang, J., Wang, L., Song, J., et al., 2015. The Relationship Between Cognitive
Impairment and Global DNA Methylation Decrease Among Aluminum Potroom Workers. J Occup Environ
Med 57, 713-717. https://doi.org/10.1097/JOM.0000000000000474.
Zawilla, N.H., Taha, F.M., Kishk, N.A., Farahat, S.A., Farghaly, M., Hussein, M., 2014. Occupational exposure to
aluminum and its amyloidogenic link with cognitive functions. J Inorg Biochem 139, 57-64. https://doi.org/
10.1016/j.jinorgbio.2014.06.003.

of
ro
-p
re
lP
na
ur
Jo