Association of exposure to manganese and fine motor skills in welders - Results from the WELDOX II study

Abstract

The aim of this study was to evaluate the effect of exposure to manganese (Mn) on fine motor functions. A total of 48 welders and 30 unexposed workers as controls completed questionnaires, underwent blood examinations, and a motor test battery. The shift exposure of welders to respirable Mn was measured with personal samplers. For all subjects accumulations of Mn in the brain were assessed with T1-weighted magnetic resonance imaging. Welders showed normal motor functions on the Movement Disorder Society-Sponsored Revision of the Unified Parkinson Disease Rating Scale part III. Furthermore welders performed excellent on a steadiness test, showing better results than controls. However, welders were slightly slower than controls in motor tests. There was no association between fine motor test results and the relaxation rates R1 in globus pallidus and substantia nigra as MRI-based biomarkers to quantify Mn deposition in the brain.

1. Introduction

Exposure to manganese (Mn) has been associated with motor dysfunctions ranging from subclinical effects at low Mn exposure to clinical symptoms following high exposure (O’Neal and Zheng, 2015; Park, 2013). Patients suffering from manganism show extrapyramidal symptoms, which are similar but not identical to Parkinson’s disease (O’Neal and Zheng, 2015). Severe motor dysfunctions usually persist after cessation of Mn exposure (Bleich et al., 1999; Bouchard et al., 2007). Whether subclinical motor effects are reversible, is still subject of research (Bouchard et al., 2007).

Welders are exposed to welding fumes consisting of Mn and other compounds (Pesch et al., 2012). Due to the neurotoxic effects of Mn, the German occupational exposure limit for respirable Mn is set at 20μg/m³. Compliance with this threshold may be challenging in many welding processes (Pesch et al., 2012). Low personal Mn concentrations can be achieved by application of low-emission welding techniques, use of materials with low Mn content, improved exhaust ventilation, avoidance of work in confined space, and use of welding helmets with purified air supply. The effects of these measures were analyzed in the WELDOX I study (Pesch et al., 2012) and a subsequent intervention study (Lehnert et al., 2014).

Neuroimaging studies revealed that Mn deposits in the brain even at relatively low exposure concentrations and that T1-weighted magnetic resonance imaging (MRI) is suited to mirror these depositions (Chang et al., 2009; Dydak et al., 2011; Pesch et al., 2018). Mn has been shown to accumulate preferentially in the globus pallidus (GP) and the substantia nigra (SN) (Bowler et al., 2018; Chang et al., 2009; O’Neal and Zheng, 2015), which are both involved in the control of motor functions. Several studies analyzed the association between cumulative Mn exposure, concentrations of respirable or inhalable Mn or concentrations of blood Mn with motor functions (Meyer-Baron et al., 2009; Park, 2013; Pesch et al., 2017; Zoni et al., 2007), but very few studies have linked neuroimaging results with motor performances in welders yet (Bowler et al., 2018; Chang et al., 2009; Lewis et al., 2016). MRI-based biomarkers evaluating Mn deposits in the brain are more favorable than measuring Mn in blood, as MRI-based biomarkers indicate exposure of the previous 3–12 months (Chang et al., 2009; Ma et al., 2018), whereas Mn in blood is considered to be an indicator for short-term exposure (O’Neal and Zheng, 2015).

We studied the association between exposure to Mn and motor performances from the WELDOX II study encompassing 48 active welders and 30 controls. The accumulation of Mn in specific brain regions was assessed with the relaxation rates R1 (=1/T1) in GP and SN as MRI-based biomarkers to quantify Mn deposition at the target site. The association of airborne and systemic exposure to Mn with relaxation rates R1 in GP and SN in the WELDOX II study is presented in Pesch et al. (2018).

2. Material and methods

2.1. Study groups

In the neuroimaging study WELDOX II active and former welders, controls, patients with Parkinson disease and males with hemochromatosis were recruited between 2013 and 2016 in Germany as described before (Casjens et al., 2018; Pesch et al., 2018; van Thriel et al., 2017). For this analysis, we selected active welders and controls. The study exclusion criteria were younger than 45 years of age, and medical conditions that preclude performance of an MRI scan (claustrophobia, metal fragments in the eyes, carrying a cardiac pacemaker or cochlear implant, large tattoos). Controls had never worked in a profession with recognized exposure to Mn. Controls and welders were never diagnosed with a neurological disorder or any disease that may impair motor functions. For this analysis, we additionally excluded subjects older than 65 years, and patients in whom brain neoplasms or other pathologic changes were detected in the MRI. The final dataset for this analysis comprised 78 men (30 controls and 48 active welders).

Active welders were recruited from 14 companies in North Rhine-Westphalia mainly from machinery and plant manufacturing. Controls were recruited from the general population of the greater Bochum area using newspaper advertisements. All participants were examined at the study center and for this study we analyzed the neuroimaging and motor function test results. A face-to-face interview was applied to assess socio-demographic characteristics, occupational history, current medications, and chronic diseases. Participants gave blood samples to determine carbohydrate-deficient transferrin (CDT) (Pesch et al., 2018). For controls, Mn in blood was measured in the sample taken at the study center, whereas for welders the sample was taken at their workplace.

2.2. Measurement of respirable welding fumes and manganese

Descriptions of the working conditions and the exposure data were gathered within the framework of the measurement system for exposure assessment of the German Social Accident Insurance (MGU) (Gabriel et al., 2010). In brief, respirable welding fumes and respirable Mn were sampled in the breathing zone of the welders over a period of four hours during a work shift (Pesch et al., 2018). Respirable Mn was determined by inductively coupled plasma mass spectrometry (ICP-MS) with a Perkin Elmer Elan DRC II and Perkin Elmer Nexion 350D (Waltham, Massachusetts).

2.3. Neuroimaging

At the study center MRI brain scans with a total scan time of one hour were performed on a 3 T Philips Achieva X-series whole-body clinical scanner (Philips Healthcare, Best, The Netherlands) using a 32-channel head coil. For anatomical reference high resolution T1-weighted 3D turbo field echo images (T1 TFE, TR/TE=8.4/3.9 ms, flip angle=8°, bandwidth=191 Hz/pixel, 220 slices, voxel size 1 mm³ isotropic, acquisition matrix: 240 × 240 mm, SENSE factor 2.5) were acquired. T1 maps were calculated by the variable flip angle method (Sabati and Maudsley, 2013) using two fast field echo images (T1 FFE, TR/TE=8.4/3.7ms, flip angles=3° and 17°, band width=191Hz/pixel, 160 slices, voxel size 1mm³ isotropic, acquisition matrix: 256 × 256mm, SENSE factor 2). Subsequently, the T1 maps were corrected for field inhomogeneity with the information from Dual-TR B1 maps. Using the corrected T1 maps, the R1 values of regions of interest (ROIs) were calculated as median of 1/T1 in selected areas of one single slice within the GP, SN or the white matter of the frontal lobe in both hemispheres. In the GP and the frontal lobe one ROI was placed in each hemisphere and in the SN two ROIs were placed in each hemisphere. To improve the quality of the data, two independent readers positioned the ROIs and for the statistical analysis the arithmetic means of R1 from both readers were calculated. Details on the measurements in ROIs can be found in Pesch et al. (2018).

2.4. Fine motor tests

The fine motor abilities of the study participants were assessed with the Motor Performance Series (MLS, Schuhfried, Mödling, Austria). The test battery with the MLS work panel comprised aiming, steadiness, line tracing, tapping and inserting pins. In the ‘aiming test’, the participants had to touch twenty holes with a diameter of 5mm spaced at a distance of 4mm one by one with a contact pen as fast as possible. The number and duration of mistakes and the overall duration of the task were recorded. This test measures the eye-hand coordination and the precision of arm-movements. In the ‘steadiness test’, the participants had to hold a contact pen for 32 seconds in a 5.8mm hole without touching the boundary and the number and duration of mistakes were documented. This test examines a person’s ability to maintain a precise arm-hand position. In the ‘line tracing test’, participants had to draw a stylus through a curvy course of a groove without touching the walls or bottom. The number and duration of mistakes and the overall duration of the task were recorded and the number of mistakes per second was calculated. The ‘line tracing test’ assesses the precision of arm-hand movements and of information processing. In the ‘inserting pins test’ the participants had to place 25 long pins positioned upright as 5 × 5 in a quadratic frame into 25 vertically aligned holes in the MLS work panel as quickly as possible and the time for completing this task was measured. The ‘line tracing test’ evaluates the speed of targeted movements, whereas the ‘tapping test’ examines the speed of untargeted movements. In the ‘tapping test’, participants had to tap a contact pen as often as possible on a 40mm² contact plate on the MLS work panel within 32 seconds and the number of hits was documented. Additionally to the MLS test battery, the spiralometry test (Kraus and Hoffmann, 2010) was applied to measure the kinetic tremor. After a training session, the participants had to draw twice a line over a preprinted spiral with a digital pen measuring the movements. Taken the data from the drawings and from the recordings of the digital pen, the tremor amplitude was estimated. All tests were performed with the dominant and non-dominant hand.

2.5. MDS-UPDRS3 motor scale

For all participants the Movement Disorder Society-Sponsored Revision of the Unified Parkinson Disease Rating Scale part III (MDS-UPDRS3) with a possible score range from zero to 132 was assessed and the examinations were videotaped. For validation of the MDS-UPDRS3 scores, the videos were rated by trained physicians (DW, MZ) as independent raters blinded for exposure history of the subjects.

2.6. Statistics

Median and interquartile range (IQR) were presented as measures of location and dispersion of continuous variables. Kruskal-Wallis tests were calculated to compare the study groups regarding Mn in blood, R1 in GP and R1 in SN. As Post-hoc analysis to the Kruskal-Wallis test the Dwass-Steel-Critchlow-Fligner multiple comparisons procedure was applied to help determine which pairs of study groups differ. Similar to Ringendahl (2002), a principal component analysis was calculated to extract factors from the test results of the MLS test battery and the spiralometry describing the fine motor skills of controls and welders. The Kaiser criterion (Eigenvalue >1) was chosen to determine the number of factors and a Varimax rotation was applied to identify each variable with a single factor. Each factor of fine motor skill was modeled with three multiple linear regression models to estimate the association with two different measures of the Mn exposure (model 1: study groups; model 2: R1 in GP and SN). For the first model three distinct Mn exposure groups were formed comprising controls as well as low and high exposed welders stratified by exposure to respirable Mn below or above 20μg/m³ which corresponds to the German occupational exposure limit. For the second model, Mn exposure was represented by the mean values of R1 in the GP of the two hemispheres and the mean values of R1 in the SN, respectively. All models were adjusted for age (continuous per 10 years) and education (intermediate school or university entrance level vs. lower secondary school or no qualification). Parameter estimates for these confounding variables and the intercept are presented for models of type one only. All calculations were performed with the statistical software SAS, version 9.4 (SAS Institute Inc., Cary, NC, USA).

2.7. Patients consent and study protocol approval

All participants provided written informed consent prior to participation. The study protocol was approved by the Ethics Committee of the Ruhr University Bochum, Germany (registration number 4762–13).

3. Results

3.1. Participants

Table 1 shows the characteristics of the 78 participants by study groups, comprising 30 controls, 23 welders with low Mn exposure (<20μg/m³), and 25 welders with high Mn exposure (≥20μg/m³). The subjects were mostly older employees with a median age of 51 years. Only three out of 25 highly exposed welders were never smokers. Every other high-exposed welder was a current smoker, in contrast to the reference group, where every third person was a current smoker. Controls had a higher educational level than welders. All welders had a blue-collar job as their longest-held occupation; whereas only 50% of controls had a blue-collar job. The proportion of self-reported alcohol consumption above 24g/day was higher in welders (low-exposed group 17.4%, high-exposed group 24%) than in controls (6.7%). The proportion of welders with carbohydrate-deficient transferrin (CDT) >2.6% was slightly higher than in controls.

Table 1:

Characteristics of 78 men, WELDOX II study, 2013–2015.

Characteristics	Controls (N=30)	Welders <20μg/m3 Mna (N=23)	Welders ≥20μg/m3 Mna (N=25)	Kruskal-Wallis test p-value
Age [years] (median, IQR b )	52 (49–56)	49 (47–56)	51 (47–55)
Smoking status (N, %)
Never	13 (43.3%)	7 (30.4%)	3 (12.0%)
Former	8 (26.7%)	6 (26.1%)	9 (36.0%)
Current	9 (30.0%)	10 (43.5%)	13 (52.0%)
School graduation (N, %)
Lower secondary school or no qualification	5 (16.7%)	16 (69.6%)	16 (64.0%)
Intermediate school or university entrance level	25 (83.3%)	7 (30.4%)	9 (36.0%)
Category of longest occupation (N, %)
Blue collar	15 (50.0%)	23 (100.0%)	25 (100.0%)
White collar	15 (50.0%)
Alcohol consumption >24g/day (N, %)
No	28 (93.3%)	19 (82.6%)	19 (76.0%)
Yes	2 (6.7%)	4 (17.4%)	6 (24.0%)
Carbohydrate-deficient transferrin [%] (N, %)
≤2.6	28 (93.3%)	21 (91.3%)	23 (92.0%)
>2.6	2 (6.7%)	2 (8.7%)	2 (8.0%)
Welding years [years] (median, IQR b )		28 (25–32)	26 (21–33)
Use of dust masks in the past
No		19 (82.6%)	21 (84.0%)
Yes		4 (17.4%)	4 (16.0%)
Respirable welding fume [mg/m3] (median, IQR b )		0.37 (0.36–0.41)	1.24 (0.78–2.20)
Respirable manganese [μg/m3] (median, IQR b )		5 (2–10)	86 (41–150)
Manganese in blood [μg/L] (median, IQR b )	6.5 (5.9–7.1)	7.4 (6.0–9.3)	8.4 (7.1–10.5)	0.0008
R1, substantia nigra [1/s] (median, IQR b ) c	0.76 (0.72–0.80)	0.74 (0.70–0.80)	0.76 (0.72–0.80)	0.73
R1, globus pallidus [1/s] (median, IQR b ) c	0.85 (0.82–0.91)	0.88 (0.84–0.93)	0.86 (0.83–0.97)	0.43

^aWelders were stratified in low and high exposed according to their shift concentration of respirable Mn below or above 20μg/m³

^bIQR: interquartile range

^cTotal N=71

3.2. Distribution of exposure

The distribution of exposure to welding fumes and Mn in the study groups is presented in Table 1. The median concentration to respirable welding fumes was 0.37mg/m³ in welders with low exposure to Mn and three times higher (1.24mg/m³) in welders with high exposure to Mn. The corresponding median concentrations of Mn in welding fumes were 5μg/m³ and 86μg/m³, respectively. Blood Mn concentrations differed significantly between the study groups (Kruskal-Wallis p value 0.0008). Welders with high exposure to Mn had a higher Mn blood concentration than controls (median 8.4 and 6.5μg/L; Dwass-Steel-Critchlow-Fligner p-value 0.0003). There were no significant differences in Mn blood concentrations between the remaining pairs of study groups (Dwass-Steel-Critchlow-Fligner p-values >0.05). Overall, the median R1 in the SN, calculated as mean of both hemispheres, was 0.76/s and the median R1 in the GP was 0.87/s. No significant differences of R1 in the SN or in the GP between the study groups could be observed (Kruskal-Wallis p values 0.73 and 0.43).

3.3. Motor test results

Table 2 shows the results of the motor tests in the three study groups. The median MDS-UPDRS3 scores of controls and welders were one and zero, respectively. 82% of all subjects had a score of zero or one.

Table 2:

Median and interquartile range of motor test results of 78 men.

Median (IQRa)	Controls (N=30)	Welders <20μg/m3 Mnb (N=23)	Welders ≥20μg/m3 Mnb (N=25)	Controls (N=30)	Welders <20μg/m3 Mnb (N=23)	Welders ≥20μg/m3 Mnb (N=25)
MDS-UPDRS3
▼ Motor examination [0–132]	1.0 (0.0–2.0)	0.0 (0.0–1.0)	0.0 (0.0–1.0)
	Dominant hand			Non-dominant hand
Aiming
▼ Mistakes	0 (0–1)	0 (0–1)	0 (0–1)	1 (0–3)	1 (0–2)	1 (0–4)
▼ Duration of mistakes [s]	0.00 (0.00–0.03)	0.00 (0.00–0.01)	0.00 (0.00–0.00)	0.05 (0.00–0.14)	0.02 (0.00–0.11)	0.01 (0.00–0.13)
▼ Duration overall [s]	8.2 (7.3–9.1)	8.8 (8.1–10.3)	9.1 (7.8–10.8)	8.5 (7.9–9.4)	10.3 (8.7–11.5)	10.5 (8.7–11.4)
Steadiness
▼ Mistakes	6 (1–14)	3 (1–6)	4 (3–8)	7 (3–19)	4 (2–10)	8 (4–15)
▼ Duration of mistakes [s]	0.2 (0.1–0.8)	0.1 (0.0–0.3)	0.1 (0.1–0.3)	0.5 (0.2–1.1)	0.2 (0.1–0.8)	0.5 (0.3–0.9)
Line tracing
▼ Mistakes	21 (16–25)	19 (13–23)	23 (20–27)	28 (22–34)	28 (22–35)	30 (25–35)
▼ Duration of mistakes [s]	1.6 (1.2–2.6)	1.7 (1.0–2.3)	1.9 (1.5–2.5)	2.5 (1.8–3.3)	3.1 (1.9–4.0)	3.3 (2.3–4.0)
▼ Duration overall [s]	19.9 (17.2–26.3)	22.2 (17.5–32.6)	25.7 (15.7–35.2)	21.4 (16.9–26.0)	26.5 (19.1–35.3)	24.1 (16.5–34.4)
▼ Mistakes per second	1.0 (0.8–1.6)	0.8 (0.4–1.4)	1.0 (0.6–1.8)	1.3 (0.8–1.9)	1.0 (0.6–1.7)	1.4 (0.8–2.0)
Inserting pins
▼ Duration overall [s]	43.1 (40.3–47.6)	44.0 (42.9–45.7)	47.9 (43.3–50.6)	43.4 (40.3–45.1)	46.9 (42.7–48.1)	46.7 (43.5–53.6)
Tapping
▲ Hits	212 (197–222)	207 (191–217)	204 (195–215)	192 (182–214)	177 (172–191)	184 (169–196)
Spiralometry
▼ Tremor amplitude [mm]	0.7 (0.6–0.8)	0.7 (0.7–0.8)	0.8 (0.6–0.9)	0.8 (0.7–1.0)	0.9 (0.7–1.1)	0.9 (0.8–1.1)

▼ Smaller values correspond with better test results; ▲ Higher values correspond with better test results

^aIQR: interquartile range

^bWelders were stratified in low and high exposed according to their shift concentration of respirable Mn below or above 20μg/m³

For all tests, the median test results of the dominant hand were better than the results of the non-dominant hand.

At the ‘aiming test’, controls were fastest with a median duration overall of 8.2s for the dominant hand and 8.5s for the non-dominant hand. The low-exposed welders were slower with 8.8s / 10.3s (dominant / non-dominant hand) followed by the high-exposed welders with 9.1s / 10.5s. At the ‘steadiness test’, low-exposed welders had the lowest median number of mistakes of all study groups along with a short median duration of mistakes 0.1s / 0.2s (dominant / non-dominant hand). In this test, the results of high-exposed welders rank second for the dominant hand. Controls were faster in line tracing and inserting pins than welders. In line tracing, high-exposed welders were worse than controls and low-exposed welders with making more mistakes and having a longer duration of mistakes. Controls performed slightly better than welders in tapping, having more hits, and in spiralometry, having smaller tremor amplitudes. With the dominant hand, high-exposed welders needed more time for line tracing and inserting pins, tapped less hits, and showed larger tremor amplitudes than controls and low-exposed welders. Using the non-dominant hand, low-exposed welders needed more time for line tracing and inserting pins and achieved less tapping hits than controls and high-exposed welders. All these results are based solely on descriptive measures.

3.4. Factor analysis of fine motor tests

Out of 22 variables describing the results of the motor tests of controls and welders, we extracted seven factors explaining 79% of the total variance (Table 3, Supplemental table S1). The factor ‘speed’ describes the speed of hand movements, mainly assessed by tapping hits and the overall duration of the ‘inserting pins test’ and of the ‘aiming test’, explaining 15.4% of the variance. The factor ‘steadiness’ explains 12.4% of the variance and this factor describes the hand steadiness which is mainly assessed by the ‘steadiness test’. The results of the ‘line tracing test’ divide mainly on the factors ‘line tracing precision’ and ‘line tracing speed’ explaining together 22% of the variance. The number and the duration of mistakes of the ‘aiming test’ from the dominant hand are the main influencing variables of the factor ‘aiming dominant hand’ and similarly the number and the duration of mistakes from the non-dominant hand of the factor ‘aiming non-dominant hand’. These two factors explain 19.6% of the variance. The factor ‘tremor’ describes the tremor of the hand which is mainly assessed by the results of the spiralometry test. This factor explains 9.6% of the variance.

Table 3:

Factor analysis of fine motor skills in 78 men (30 controls and 48 welders)

Factor	Factor description	Main variablesa	Variance explained
Factor 1	▲ Speed	▼ Aiming duration overall dom. hand (−47) and nd. hand (−55), ▼ Inserting pins duration overall dom. hand (−71) and nd. hand (−72), ▲ Tapping hits dom. hand (79) and nd. hand (78)	15.4%
Factor 2	▲ Steadiness	▼ Steadiness mistakes dom. hand (−89) and nd. hand (−76), ▼ Steadiness duration of mistakes dom. hand (−88) and nd. hand (−57)	12.4%
Factor 3	▲ Line tracing precision	▼ Line tracing mistakes dom. hand (−67) and nd. hand (−81), ▼ Line tracing duration of mistakes dom. hand (−52) and nd. hand (−77)	11.4%
Factor 4	▲ Line tracing speed	▼ Line tracing duration overall dom. hand (−89) and nd. hand (−94)	10.6%
Factor 5	▲ Aiming dom. hand	▼ Aiming mistakes dom. hand (−92), ▼ Aiming duration of mistakes dom. hand (−90)	9.9%
Factor 6	▲ Aiming nd. hand	▼ Aiming mistakes nd. hand (−90), ▼ Aiming duration of mistakes nd. hand (−92)	9.7%
Factor 7	▼ Tremor	▼ Tremor amplitude dom. hand (88) and nd. hand (82)	9.6%

dom. hand: dominant hand and nd. hand: non-dominant hand; ▼ Smaller values correspond with better test results; ▲ Higher values correspond with better test results

^aPresentation is restricted to variables with an absolute value of the factor loading above 45. A full length table is presented as supplemental table S1.

3.5. Association between exposure to manganese and factors of fine motor tests

The effect estimates of separate linear regression models for each of seven factors of fine motor skills are shown in Table 4.

Table 4:

Separate linear regression models of factors of fine motor skills in 78 men

		Model 1					Model 2a
		Intercept	Age [10 years]	High vs. low educationc	Welders (<20μg m3 Mnb) vs. controls	Welders (≥20μg/m3 Mnb) vs. controls	R1, globus pallidusd	R1, substantia nigrad
Factor 1	β	2.74	−0.53	0.59	−0.40	−0.51	−1.31	0.68
▲ Speed	95% CI	(0.66; 4.81)	(−0.89; −0.16)	(0.13; 1.04)	(−0.95; 0.15)	(−1.04; 0.01)	(−3.59; 0.97)	(−2.31; 3.68)
	p value	0.011	0.0050	0.013	0.15	0.056	0.26	0.65
	R2			0.28				0.29
Factor 2	β	−1.71	0.20	0.51	0.83	0.43	1.73	−1.12
▲ Steadiness	95% CI	(−4.03; 0.62)	(−0.21; 0.61)	(0.000; 1.03)	(0.21; 1.44)	(−0.17; 1.02)	(−0.93; 4.39)	(−4.62; 2.37)
	p value	0.15	0.33	0.050	0.0093	0.16	0.20	0.52
	R2			0.10				0.03
Factor 3	β	1.53	−0.35	0.51	0.23	−0.14	1.46	−2.78
▲ Line tracing	95% CI	(−0.76; 3.82)	(−0.75; 0.05)	(0.000; 1.01)	(−0.38; 0.83)	(−0.72; 0.44)	(−0.97; 3.89)	(−5.96; 0.41)
precision	p value	0.19	0.088	0.049	0.46	0.63	0.23	0.086
	R2			0.13				0.11
Factor 4	β	1.19	−0.19	0.02	−0.38	−0.34	−0.57	1.22
▲ Line tracing	95% CI	(−1.21; 3.60)	(−0.61; 0.23)	(−0.51; 0.55)	(−1.02; 0.26)	(−0.95; 0.27)	(−3.19; 2.06)	(−2.23; 4.66)
speed	p value	0.32	0.38	0.94	0.24	0.27	0.67	0.48
	R2			0.04				0.01
Factor 5	β	−1.01	0.15	0.35	0.19	−0.05	0.02	−0.32
▲ Aiming	95% CI	(−3.41; 1.39)	(−0.27; 0.57)	(−0.19; 0.88)	(−0.44; 0.83)	(−0.66; 0.56)	(−2.66; 2.71)	(−3.84; 3.20)
dom. hand	p value	0.41	0.48	0.20	0.55	0.88	0.99	0.86
	R2			0.04				0.03
Factor 6	β	−0.20	−0.01	0.21	0.21	0.22	−0.50	−0.17
▲ Aiming	95% CI	(−2.63; 2.24)	(−0.44; 0.42)	(−0.33; 0.75)	(−0.43; 0.86)	(−0.40; 0.84)	(−3.17; 2.17)	(−3.67; 3.33)
nd. hand	p value	0.87	0.97	0.43	0.51	0.49	0.71	0.92
	R2			0.01				0.01
Factor 7	β	−0.79	0.13	0.04	0.08	0.29	−0.61	1.53
▼ Tremor	95% CI	(−3.22; 1.64)	(−0.30; 0.55)	(−0.50; 0.58)	(−0.57; 0.72)	(−0.33; 0.91)	(−3.28; 2.07)	(−1.98; 5.04)
	p value	0.52	0.56	0.88	0.81	0.35	0.65	0.39
	R2			0.02				0.01

▼ Smaller values correspond with better test results in main variables of the factor; ▲ Higher values correspond with better test results in main variables of the factor

CI: confidence interval; dom. hand: dominant hand; nd. hand: non-dominant hand

Model 1 analyzes the influence of exposure groups (high exposed welders, low exposed welders, and controls); model 2 the influence of brain accumulation of Mn on factors of fine motor skills.

^aAdjusted for age and education

^bWelders were stratified in low and high exposed according to their shift concentration of respirable Mn below or above 20μg/m³

^cIntermediate school or university entrance level vs. lower secondary school or no qualification

^dArithmetic mean of both hemispheres, total N=71

We did not observe a significant effect of manganese in terms of R1 in GP or SN on any factor of fine motor tests. Welders tended to be slower than controls, though differences were not statically significant (low-/ high-exposed welders vs controls: $\widehat{\beta }$ = −0.40 / −0.51, p = 0.15 / 0.06). And welders tended to perform better in ‘steadiness’ compared to controls. A lower R1 in the SN (but not the GP) was associated with a slightly better performance in ‘line tracing precision’ ($\widehat{\beta }$ = −2.78, p = 0.09).

The ‘speed’ decreased with growing age (age [10 years]: $\widehat{\beta }$ = −0.53, p = 0.005). Similarly, younger subjects were superior in ‘line tracing precision’ compared to older subjects though not significant (age [10 years]: $\widehat{\beta }$ = −0.35, p = 0.09). Education showed an effect on ‘speed’, ‘steadiness’, and ‘line tracing precision’ with highly educated subjects showing a better performance (high vs low education: $\widehat{\beta }$ ranging from 0.51 to 0.59, p value from 0.01 to 0.05).

The applied statistical models on ‘line tracing speed’, ‘aiming dominant hand’, ‘aiming non-dominant hand’, and ‘tremor’ had very low coefficients of determination (R² below 0.05) and none of the investigated exposure variables showed an effect on these factors.

A sensitivity analysis excluding six subjects with carbohydrate-deficient transferrin above the clinical reference value for heavy alcohol consumption of 2.6% revealed similar results (data not shown). Another sensitivity analysis including smoking and alcohol consumption per g/day in all regression models revealed similar results (Supplemental table S2).

4. Discussion

In the WELDOX II study, no association between cerebral Mn exposure and motor functions could be observed. Though the majority of welders were exposed above the current threshold for respirable Mn of 20μg/m³, no differences in MRI-based biomarkers for Mn (R1 in GP, R1 in SN) were found between welders and controls. In MDS-UPDRS3, welders had very low scores and showed normal motor functions. The dexterity of welders and controls were further evaluated by six motor tests resulting in 22 variables, which could be reduced to seven factors. Welders performed better in the factor ‘steadiness’ than controls, but no effect of the MRI-based biomarkers for Mn on this factor was observed. Controls showed a slightly better performance in the factor ‘speed’ than welders, though this factor was not associated with the MRI-based exposure indices. There was no association between Mn exposure, assessed through MRI or through exposure groups, and any other factor, besides ‘steadiness’ and ‘speed’, measuring the fine motor functions.

4.1. Strengths and limitations

Strengths of our study are the assessment of Mn deposition in the brain and the investigation of its association with motor functions. So far, neuroimaging studies characterizing the exposure of welders are associated with high efforts and costs and are therefore rare and mostly small. An extensive motor test battery was used to investigate possible deficits and neurotoxic effects of Mn exposure. Limitations include the cross-sectional design of the study, the rather small sample size, a possible healthy worker effect, a possible selection bias through recruitment of controls by newspaper advertisements, and differences in education between welders and controls, which may confound on the observed associations. No differences between any of the study groups in terms of the MRI measures could be observed which points to a low brain exposure. The investigation of associations between Mn exposure and motor functions seems here to be limited to a low exposure setting.

4.2. Mn exposure and brain deposition of Mn

The concentrations of respirable welding fume and Mn exposure in the WELDOX II study are in line with the previous WELDOX I study on 241 German welders (Pesch et al., 2012). As expected, blood Mn concentrations were higher in welders than in controls. We stratified the welders in low and high exposed according to the shift concentration of respirable Mn below or above 20μg/m³. Several studies demonstrated that the R1 signals in GP and SN are suitable MRI indicators of Mn deposition in the brain and might depict the Mn exposure of the last few months in a non-linear fashion (Bowler et al., 2018; Lee et al., 2015; Long et al., 2014; Ma et al., 2018). Although many welders were exposed above the current threshold for respirable Mn of 20μg/m³ in the WELDOX II study, we observed no differences between any of the study groups in terms of R1 in the GP and R1 in the SN. A detailed discussion about the effect of respirable Mn on the R1 signals in GP and SN in the WELDOX II study is published by Pesch et al. (2018).

4.3. Welders performed excellent on MDS-UPDRS3

A recent prospective study in 886 U.S. welding-exposed workers found a dose-dependent increase of the Unified Parkinson Disease Rating Scale motor subsection part 3 (UPDRS3) with cumulative Mn exposure (Racette et al., 2017). Similarly, a study in environmentally Mn-exposed subjects observed a higher risk of scoring abnormally (score >0) on the UPDRS3 in the exposed group (N=100) compared to a reference group (N=90) (Kim et al., 2011). In contrast, in our much smaller study in 78 men the German welders of the WELDOX II study performed excellently on the MDS-UPDRS3 and the variation of the score in welders and controls was low. Our results are in line with a prospective study in welding trainees (N=56), though, which did not detect an association between cumulative exposure and UPDRS3 score after adjusting for possible learning effects (Baker et al., 2015). In the WELDOX II study, the scores of welders and controls are low on MDS-UPDRS3. A population-based study by (Keezer et al., 2016) reports a higher mean MDS-UPDRS3 score but when their result is adjusted for age (≤65 years) and sex (men only), the estimated mean MDS-UPDRS3 score is also comparatively low.

4.4. The fine motor test battery measures factors of motor dexterity

We also chose the MLS test battery (Neuwirth and Benesch, 2012) and the spiralometry test (Kraus and Hoffmann, 2010) to measure the motor dexterity of welders and controls. We combined the motor test results using a factor analysis approach and found similar factors as described by Neuwirth and Benesch (2012) and by Ringendahl (2002) for the MLS test battery, although the testing procedures and studied subjects were not identical.

4.5. Welders performed better on steadiness test than controls

We observed a group difference in the factor ‘steadiness’ comprising all steadiness test measures with welders showing better test results than the reference group. This finding stands in contrast to two studies where occupational Mn-exposed subjects performed worse in hand steadiness tests than controls (Bast-Pettersen et al., 2004; Roels et al., 1999). However, these workers were employed in a dry-alkaline battery plant (Roels et al., 1999) or in Mn-alloy-production plants (Bast-Pettersen et al., 2004), whereas several studies on welders also observed a good to excellent performance on steadiness tests in comparison to a reference group (Bast-Pettersen et al., 2000; Ellingsen et al., 2015; Pesch et al., 2017; Wastensson et al., 2012). Because we did not observe an effect of the MRI-based markers for Mn, the observed group differences are probably not caused by Mn exposure. Instead, the group difference might be explained by a selection of workers into or out of the welding profession or a job-related training effect, as accurate welding lines need a steady hand. Additionally, the exposures of workers at dry-alkaline battery plants and at Mn-alloy-production plants are probably different to the exposures of welders regarding particle sizes and solubility of Mn species. The hypothesized selection and training effects might similarly effect the factor ‘tremor’, where we did not observe a group difference between welders and controls.

4.6. Welders are slightly slower than controls in motor tests but not associated with Mn biomarkers

Whether low exposed welders show deficits in the function of motor speed is still unknown. A meta-analysis on this subject found significantly lower performance scores in finger tapping tests for Mn-exposed workers (Meyer-Baron et al., 2009). However, it has been demonstrated later that one of the included study results was biased by the study subjects’ alcohol consumption (Ellingsen et al., 2008; Ellingsen et al., 2015). Similarly, a large study on elderly men found an increased risk of an impaired performance in the tapping test for men with a higher cumulative exposure to Mn (Pesch et al., 2017). However, in the WELDOX II study welders showed slightly worse results in the factor ‘speed’ compared to a reference group but an association with cerebral Mn exposure could not be determined. Lower education was also associated with a worse performance on the factor ‘speed’. Although we adjusted the model for education, there might still be problems with confounding as welders were lower educated as controls in the WELDOX II study,

4.7. Age and education are confounders for motor function tests

Overall, age and education are important confounders in the analysis of motor functions. In our study motor speed decreased with increasing age, which is in line with findings from other studies (Kraus et al., 2000; Pesch et al., 2017). Steadiness on the other hand was not affected by age in our study corresponding to observations from other studies analyzing subjects in the working age (Buchta et al., 2005; Kiesswetter et al., 2009). Contrasting results were found in the elderly where performances in the steadiness test decreased in age groups beyond working age (Kraus et al., 2000; Pesch et al., 2017). Education was another strong confounder in our study where better scores on the factors ‘speed’, ‘steadiness’, and ‘line tracing precision’ were associated with higher secondary school qualifications. Pesch et al. (2017) reported a similar effect of occupational education on fine motor tests in elderly men, though Kiesswetter et al. (2007; 2009) did not observe differences in motor performances by higher school qualification. However, the education index of their subjects was in general low with most subjects having only graduated from primary school. Our study controls were on average better educated than welders. As higher education is associated with a better motor performance, welders might show worse results than controls on tests on motor functions due to confounding with education. Though we adjusted all models for education, we cannot rule out possible residual confounding.

4.8. Conclusion

In the WELDOX II study, where the majority of studied welders were exposed above the current threshold for respirable Mn of 20μg/m³, no difference in cerebral Mn exposure measured with T1-weighted MRI images between welders and controls could be observed which points to a low brain exposure in welders. In this setting, no association was found between Mn exposure measured with T1-weighted MRI images and fine motor functions. The good performance of welders in the steadiness test may be due to selection and training effects. Fine motor skills may have been influenced by the subjects’ age and education. Because the welders in WELDOXII were comparatively young, any conclusion about the long-term neurotoxicity of Mn is not possible as we cannot rule out that detrimental effects arise later in life. Larger and prospective studies are therefore needed to better evaluate the long-term neurotoxicity of Mn.

Highlights.

• Accumulations of manganese in the brain assessed with T1-weighted MRI

• No clear association between manganese exposure and fine motor functions

• Welders showed normal motor functions on the MDS-UPDRS part III scores

• Welders performed better on a steadiness test than controls

• Limitations: cross-sectional, newspaper recruitment, disparities in education