Nerve conduction, visual evoked responses and electroretinography in tunnel workers previously exposed to acrylamide and N-methylolacrylamide containing grouting agents

Available online at www.sciencedirect.com

Neurotoxicology and Teratology 30 (2008) 186 – 194 www.elsevier.com/locate/neutera

Nerve conduction, visual evoked responses and electroretinography in tunnel workers previously exposed to acrylamide and N-methylolacrylamide containing grouting agents Lars Ole Goffeng a,⁎, Mona Skard Heier b , Helge Kjuus a , Hans Sjöholm b , Kjell Aage Sørensen c , Vidar Skaug a a

Department of Occupational Medicine and Epidemiology, National Institute of Occupational Health, P.O. Box 8149 Dep, N-0033 Oslo, Norway b Department of Clinical Neurophysiology, Ullevål University Hospital, N-0407 Oslo, Norway c Veidekke Entreprenør AS, Occupational Health Service, P. O. Box 506 Skøyen, N-0214 Oslo, Norway Received 4 July 2007; received in revised form 11 December 2007; accepted 26 January 2008 Available online 7 February 2008

Abstract The study examines possible persisting effects on the peripheral nervous system and visual system in tunnel workers previously exposed to acrylamide and N-methylolacrylamide during grouting work. We compared neurophysiological function in 44 tunnel workers previously exposed during grouting operations (2–10 years post exposure), with 49 tunnel workers with no history of exposure to acrylamide. Nerve conduction velocities (NCV), distal delay, F-response and amplitude in median and ulnar nerves of the right arm, peroneal, sural and tibial nerves of the right leg, visual evoked response (VER) and electroretinography (ERG) were measured. VER and ERG were also performed in 24 subjects more recently exposed to acrylamide grout (16 months post exposure). Exposure to acrylamide containing grouts was assessed by questionnaires. A statistically significant reduction in the mean sensory NCV of the sural nerve (p = 0.005), as well as a non-significant reduction of sural amplitude was found in the previously exposed group compared to the control group. VER latencies to the onset of the occipital potential (N75) were prolonged in both exposed groups compared to the control group (p b 0.05). ERG 30 Hz flicker amplitude was reduced in the recently exposed group compared to the referents (p b 0.05). The results indicate slight subclinical, but persistent toxic effects in the sural nerve and the visual system in tunnel workers exposed to N-methylolacrylamide and acrylamide during grouting operations. © 2008 Elsevier Inc. All rights reserved. Keywords: Acrylamide; N-methylolacrylamide; Grouting agents; Visual system; Peripheral nervous system; Nerve conduction; Tunnel workers; Occupational exposure

1. Introduction Occupational exposure to acrylamide may present a hazard for workers. During the 1950s and 1960s, more than 150 cases of acrylamide-poisoning were reported primarily related to primary production of acrylamide from acrylonitrile [10,30,35], or to polymerisation of acrylamide to polyacrylamide [1,14]. The first known case of acrylamide-related health effects among construction workers applying acrylamide containing grouts to ⁎ Corresponding author. Tel.: +47 23 19 53 86; fax: +47 23 19 52 05. E-mail address: [email protected] (L.O. Goffeng). 0892-0362/$ - see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.ntt.2008.01.006

prevent water-leakage into tunnels was in France in 1970 [18]. Later, several reports have related symptoms among tunnel workers to use of such grouts [24,33]. Due to the toxic properties of acrylamide, grouts based on N-methylolacrylamide (NMA), which was regarded as less toxic, were developed. In tunnel construction, however, such grouts have lead to a leakage of both acrylamide and N-methylolacrylamide monomers, both because methylolacrylamide may be transformed to acrylamide at high pH of drainage water, and due to a low degree of polymerisation if the product is diluted in water [45]. Furthermore, the molecular structure of the two substances does not indicate different modes of toxic action.

L.O. Goffeng et al. / Neurotoxicology and Teratology 30 (2008) 186–194

We have recently reported slight neurotoxic subacute effects and possible genotoxic effects among Norwegian tunnel workers using N-methylolacrylamide based grouts [26,27]. The simultaneous incidence of acute neurotoxic effects among Swedish tunnel workers [19], led to a ban of N-methylolacrylamide based grouts in several countries. In both studies, the neurotoxic peripheral nervous system effects seemed mainly reversible. However, progressive impairment of sural sensory nerve conduction velocity (NCV) from examinations 6 months to 16 months post exposure was observed in the Norwegian study [26]. In previous reports, indications of persistent effects have been reported in severe cases [11,36,42]. Visual system effects following acrylamide exposure have been reported in experimental animal studies [34], and also in a human case report [33]. In the Swedish incident, one of seven observed cows drinking acrylamide-polluted water, showed abnormal pupillary light reflex, progressive retinal degeneration and changes in optic nerve discs, indicating degeneration of optic nerve neurons [15,16]. In Norway, NMA grouts were used in several large tunnel projects during the 1990s. This gave us an opportunity to study possible persistent nervous system effects in a group of workers exposed to such grouts remote in time. Thus, the aim of the present study was to examine possible persisting peripheral nervous system effects and to assess possible visual system effects, in a group of tunnel workers previously exposed to N-methylolacrylamide based grout during tunnel work. 2. Materials and methods 2.1. Subjects Forty-eight tunnel workers from four Norwegian companies, who had been working in tunnel projects where NMA grout had been used more than 2 years prior to the examination, were eligible for the study. They were identified from a registry of tunnel construction companies using acrylamide containing grouts, established in 1997 by the Norwegian Association for Building and Construction Industry for surveillance purposes. No measurements of acrylamide or NMA in the working

187

environment had been performed during injection work in companies represented in the registry. Information from the construction companies and occupational health department files were used to identify the participants for the study, based on their previous exposure history. Workers with the presumed highest exposure, based on number of projects, and amount of NMA grouts used in each project, were invited to participate. To limit the possibility of outcome selective recruitment into the study, the exposure assessment was based on described work tasks and length of grouting work, not on subjective exposure assessment by the workers themselves. Subjects with known persistent neurological disease, diabetes mellitus, or known alcohol or drug abuse were not eligible for the study. Four subjects with both previous and recent acrylamide exposure were excluded from the exposed group, due to potential incomplete peripheral nerve recovery following the more recent exposure [26]. Thus, 44 eligible subjects were included in the index group of workers with previous NMA exposure, all of them accepting to participate in the study. In addition, visual evoked response (VER) and electroretinography (ERG) was examined in another group of 24 more recently exposed tunnel workers (16 months post exposure). This group was selected from 73 workers who had taken part in railway tunnel construction during a grouting period terminating 16 months prior to a health examination. The 25 workers with highest exposure to acrylamide containing grout were identified and included. One of them was excluded due to diabetes. The complete procedure of inclusion, and NCV results from this group have been presented in detail elsewhere [26]. Fifty-three male tunnel workers not previously exposed to NMA grouts, randomly recruited from one of the four construction companies were invited to participate as control group. 50 of the 53 invited workers accepted to participate. One dropped out from parts of the study later on, leaving a control group of 49 subjects. The mean age of the 44 previously exposed workers were 47.9 years, compared to 43.7 years among the 24 recently exposed workers, and 44.6 years among the 49 controls (Table 1). Mean height in the previously exposed and control

Table 1 Distribution of age, life style factors and other exposure factors reported by the previously (n = 44) and recently (n = 24) acrylamide and NMA exposed groups and control group (n = 49) Group

Control (N = 49) Previously exposed (N = 44) Recently exposed (N = 24) a b c d

N = 48. N = 43. N = 23. N = 22.

Age (years)

Years in construction work

Present smokers

Alcohol N 5 l/year

Work with solvents

Previous lead exposure

Vibrating hand-tools

Whole-body vibration

Mean (SD)

Range

Mean (SD)

N (%)

Mean no. of cigarettes (SD)

N (%)

N (%)

N (%)

N (%)

N (%)

44.6 (9.7) 47.9 (9.5)

23–60 30–70

19.3 (10.7) 24.6 (12.5) b

24 (49.0) 18 (40.9)

15.3 (5.5) 13.1 (6.8)

12 (24.5) 2 (4.5)

30 (62.5) a 37 (84.1)

8 (16.3) 4 (9.1)

44 (91.7) a 43 (97.7)

25 (51.0) 26 (60.5) b

43.7 (8.7)

31–62

19.3 (7.8) c

11 (45.8)

12.8 (4.1)

8 (33.3)

11(47.8) c

2 (9.1) d

24 (100.0)

8 (36.4)

d

188

L.O. Goffeng et al. / Neurotoxicology and Teratology 30 (2008) 186–194

groups were 179.3 cm (SD: 6.6) and 179.4 cm (SD: 5.3), respectively. The previously exposed group had been engaged in tunnel construction or other construction work for 24.6 (SD = 12.5) years, the recently exposed group for 19.3 (SD = 7.8) years, and the controls for 19.3 (SD = 10.7) years. Eleven of the previously exposed workers were out of work at the time of examination due to causes unrelated to the outcome of the present study. The distribution of life style factors (smoking, alcohol consumption), and relevant occupational exposure factors are presented in Table 1. Previous concussion was reported by 18 exposed participants, and by 18 participants in the control group. Transient eye injury in one or both eyes was reported by 14 exposed participants and by 7 control group participants. Information on medication use among the study participants was collected. Medication reported to render a possible risk of PNS-affection was included as a variable in the initial analyses, but did not influence the results and was left out in the final analyses. The project was undertaken in close cooperation with the companies’ Occupational Health Services. The study was approved by the Norwegian Data Inspectorate and the Regional Committee for Medical Research Ethics. All the workers gave written informed consent to participate in the study. 2.2. Methods All participants were examined by questionnaires, neurophysiological measurements in the right arm and leg, and ERG and VER. The examinations took place during daytime. The exposed and control participants were mixed, to reduce the risk of systematic errors due to subtle differences in testing methods or equipment over time. The tester of neurophysiological function did not know whether the participants were members of the exposed group or the control group. 2.2.1. Questionnaires regarding symptoms and exposure Each participant received questionnaires regarding exposure, background factors, symptoms and potential confounders, like current alcohol consumption, smoking habits, previous experience with vibrating hand-tools, exposure to organic solvents and several other neurotoxins, and previous injuries. 2.2.2. Exposure assessment Altogether, 2600 t of an NMA-based grouting agent (RhocaGil, Siprogel) were used for grouting operations in tunnel construction in Norway in the period 1982–97, approximately half the amount used by the four companies recruiting subjects to the exposed group. Description of the grouting agent and working conditions have been presented elsewhere [26,45]. The included exposed tunnel workers also performed various other tasks in the tunnels, in addition to the grouting work. Detailed information on grout injection and tasks with potential acrylamide grout exposure was obtained by members of the project group (VS and HK) interviewing each exposed participant and assessing the individual exposure levels qualitatively, to supplement questionnaire information. They

did not know the neurophysiological test results at the time of the interview. Based on the combined interview and questionnaire information, we used a modified exposure-time index, developed in a previous study [26]. The exposure-time index was a sum score (range 0–14) based on frequency and level of exposure in different exposure-related work activities, multiplied with categories (0–4) of exposure duration. The index was dichotomised into high and low exposure subgroups with cutoff set at median value. 2.2.3. Nerve conduction measurements Nerve conduction measurements were performed under standardised conditions, with room temperature 23 ° C and skin temperatures N 30 ° C. Standard neurography from the right arm (median and ulnar nerves) and leg (tibial, sural, peroneal nerves), measuring motor and sensory NCV (m/s), motor (mV) and sensory (µV) amplitudes and F-responses (ms), was performed with surface electrodes both for stimulation and recording [25]. Motor conduction velocities were measured by orthodromic stimulation and sensory conduction velocities by antidromic stimulation. Amplitudes reduced to such a degree that potentials were not elicited during stimulation were scored as zero. The corresponding NCV and F-response could not be measured, and remained blank. Consequently, in the group comparison we also emphasized the number of non-elicited responses in addition to the mean amplitude. Visual evoked responses (VER) [3,13,20] were recorded with pattern reversal stimulation and stimulus frequency 2 Hz, using a VDU chessboard pattern with standardised luminance where the dark and light squares alternate, square size 30′. Each eye was stimulated separately, with the other eye covered, in a dark room. The averaged evoked cortical response from 200 stimuli was recorded from surface electrodes located at standard EEGplacements (O1, Oz and O2) over the occipital region, with a frontal reference (Fz). Latency to the maximum positive deviation from the baseline (P100) was measured (ms), besides occipital amplitude of the impulse (μV), as well as the latencies to the onset (N75), and duration, N75–N120 (ms), of the response potential. Analysis was based on the mean value for both eyes. Electroretinography (ERG) [3,13] was recorded with a surface electrode at the zygomatic arch just behind the retinal plane with a reference at the lower eye-lid in front of the retinal plane. Photoptic stimulation was given as flashes of white light with the frequencies 2 Hz and 30 Hz, with normal indoor light conditions. The amplitudes (μV) of the positive B-wave, which is generated by the photoreceptor cell of the retina and the adjoining bipolar cells, were measured at both frequencies. 2.3. Statistical analysis The SPSS Statistical Package (SPSS for Windows 14.0) was used for data analysis. In the total sample (N = 93), visual inspection of normal probability plots [17], besides Kolmogorov–Smirnov test [2,17], and calculating variable distribution skewness was used to assess the normality of variable distribution. A calculated skewness of 2 was set as cut-off for

L.O. Goffeng et al. / Neurotoxicology and Teratology 30 (2008) 186–194

choice of statistical methods for further data analysis, lower values indicating limited deviation from normal distribution. Sample distribution was significantly different from normal distribution (p b 0.05, Kolmogorov–Smirnov test) in total sample variables median and ulnar nerves amplitude and distal delay, and ERG 30 Hz amplitude. In median nerve amplitude and distal delay only, skewness was above cut-off. Group differences in these neurography variables and ERG 30 Hz amplitude were tested with both parametric (independent samples t-test) and non-parametric (Mann–Whitney U test) methods, rendering only minor differences. Consequently, for all continuous measures, arithmetic mean and standard deviation were calculated, and group differences were analysed with independent samples t-tests (Tables 2 and 4). The level of statistical significance was set two-tailed at p b 0.05. We corrected for the groups showing different distribution of registrations by applying the corrected values if Levene's test was significant (p b 0.05). Data were age-adjusted by applying ANCOVA-statistics including age as covariate in analysis of neurography group differences (Table 2), and by multiple linear regression,

189

backwards strategy, including exposure and age in the model, in the VER/ERG analysis (Tables 5 and 6). Multiple linear regression analysis, backwards strategy, was also chosen to further study the effect of independent variable exposure (dichotomous, 0–1) in combination with other factors, on dependent variables neurography (Table 3) and VER/ERG variables (table not shown). Age (continuous variable), and dichotomous (0–1) variables (acrylamide- and NMA exposure, exposure to organic solvents N 10 years, annual pure alcohol consumption N 5 l, and present smoking) were included in the model to control for possible confounders in all these analyses. In addition, height (cm, continuous variable) was included in the neurography analyses, and previous self-reported exposure to lead (dichotomous, 0–1) was included in VER/ERG analyses. Concussion without (1) or with (2) unconsciousness N30 min and perceived transient minor eye injury (dichotomous, 0–1), was included in the VER/ERG analyses of the previously acrylamide-exposed group only. The final model was based on an exclusion criterion of p N =0.1, and reporting levels of significance p b 0.05 and 0.01. To investigate dose– response relationship, regression analysis was applied with

Table 2 Motor and sensory NCV and amplitude, F-response and motor distal delay in the previously exposed group and the control group Previously exposed

Control

Crude

Age-adjusted

N

Mean (SD)

N

Mean (SD)

(95% CI of diff.)

p

(95% CI of diff.)

p

56.2 (5.3) 57.7 (7.7) 44.6 (5.3) 45.7 (6.2)

47 49 49 48

55.7 (5.9) 55.5 (6.4) 43.1 (4.7) 45.6 (4.3)

0.5 (− 1.9 to 2.2 (− 0.7 to 1.5 (− 0.5 to 0.0 (− 2.2 to

0.707 0.139 0.148 0.990

1.1 2.7 2.1 0.4

0.341 0.067 0.043 0.753

(2.8) (1.9) (2.4) (1.5)

49 49 49 49

6.7 (3.3) 6.5 (2.0) 4.3 (2.1) 3.5 (2.1)

0.3 (− 1.0 to 1.6) −0.5 (− 1.3 to 0.3) −0.4 (− 1.3 to 0.6) −0.7 (− 1.4 to 0.1)

0.621 0.211 0.422 0.085

0.5 (− 0.8 to 1.7) − 0.5 (−1.3 to 0.3) − 0.2 (−1.1 to 0.7) − 0.6 (−1.3 to 0.2)

0.464 0.247 0.702 0.149

50.9 (7.5) 58.0 (9.5) 44.2 (6.4)

49 45 48

53.7 (8.4) 58.9 (8.6) 48.4 (5.8)

−2.8 (− 6.1 to 0.6) −0.9 (− 4.8 to 3.1) −4.2 (− 6.8 to − 1.2)

0.109 0.667 0.002

− 1.9 (−5.1 to 1.4) − 0.5 (−4.5 to 3.5) − 3.7 (−6.3 to − 1.2)

0.256 0.801 0.005

Sensory amplitude (μV) – Median 43 – Ulnar 43 – Sural 43

30.5 (30.8) 12.0 (11.1) 6.1 (5.3)

49 49 49

31.4 (36.7) 13.1 (11.8) 8.2 (5.7)

−0.9 (− 15.0 to 13.3) −1.1 (− 5.8 to 3.7) −2.1 (− 4.4 to 0.2)

0.901 0.658 0.071

2.6 (− 11.1 to 16.3) − 0.1 (−4.7 to 4.6) − 1.6 (−3.9 to 0.6)

0.712 0.982 0.153

F-response (ms) – Median – Ulnar – Tibial – Peroneal

24.6 (2.1) 26.0 (2.1) 51.0 (7.0) 49.5 (4.9)

49 48 48 43

24.5 (2.9) 25.5 (2.5) 50.4 (6.7) 49.9 (5.3)

0.2 (− 0.9 to 1.2) 0.5 (− 0.4 to 1.5) 0.5 (− 2.4 to 3.5) −0.4 (− 2.7 to 2.0)

0.755 0.279 0.709 0.764

0.2 (− 0.9 to 1.2) 0.5 (− 0.5 to 1.5) 0.1 (− 2.8 to 3.0) − 0.6 (−2.9 to 1.6)

0.763 0.283 0.955 0.574

3.6 (0.5) 2.7 (0.5)

49 49

3.8 (0.7) 2.5 (0.4)

−0.2 (− 0.4 to 0.1) 0.2 (− 0.0 to 0.4)

0.119 0.088

− 0.2 (−0.5 to − 0.0) 0.1 (− 0.0 to 0.3)

0.047 0.142

Mean difference

Mean difference

a

Motor NCV (m/s) – Median 43 – Ulnar 43 – Tibial 43 – Peroneal 43 Motor amplitude (mV) – Median 43 – Ulnar 43 – Tibial 43 – Peroneal 43 Sensory NCV – Median – Ulnar – Sural

41 38 39

43 41 40 36

Motor distal delay (ms) – Median 43 – Ulnar 43 a

7.0 6.0 4.0 2.8

NCV = nerve conduction velocity.

2.8) 5.1) 3.6) 2.2)

(− 1.2 to 3.4) (− 0.2 to 5.6) (0.1 to 4.0) (− 1.9 to 2.6)

190

L.O. Goffeng et al. / Neurotoxicology and Teratology 30 (2008) 186–194

Table 3 Motor and sensory NCV and amplitude, F-response and motor distal delay in the previously exposed group and control group a Regression-coefficients b Constant

R2

Model p-value

Age

Height

Previous exposure

Present smoking

Alcohol consumption

Organic solvents

Motor NCV (m/s) – Median 64.9 ⁎⁎ – Tibial 87.5 ⁎⁎

− 0.19 ⁎⁎ − 0.20 ⁎⁎

– −0.20 ⁎

– 2.0 ⁎

– –

– –

– –

0.11 0.18

0.001 b0.001

Motor amplitude (mV) – Median − 18.4 n.s – Tibial 7.2⁎⁎

– − 0.07 ⁎⁎

0.14 ⁎⁎ –

– –

– –

– –

– –

0.08 0.08

0.008 0.006

Sensory NCV (m/s) – Median 65.8 ⁎⁎ – Sural 101.8 ⁎⁎

− 0.29 ⁎⁎ − 0.20 ⁎⁎

– −0.26 ⁎

– − 3.5 ⁎⁎

– 3.0 ⁎

– –

– –

0.12 0.30

0.001 b0.001

Sensory amplitude (μV) – Median 93.5 ⁎⁎ – Sural 69.0 ⁎⁎

− 1.21 ⁎⁎ − 0.20 ⁎⁎

– −0.30 ⁎⁎

– –

− 14.1 ⁎ –

– 2.9 n.s.

– 3.0 n.s.

0.14 0.24

0.001 b0.001

F-response (ms) – Median − 8.8 n.s. – Tibial − 49.3 ⁎

– 0.21 ⁎⁎

0.19 ⁎⁎ 0.51 ⁎⁎

– –

– − 2.6 ⁎

– –

– –

0.19 0.28

b0.001 b0.001

Motor distal delay (ms) – Median 3.1 ⁎⁎

0.02 ⁎

–

− 0.25

–

–

–

0.10

0.013

Backwards linear regression. Acrylamide/NMA exposure, age (years), height (cm), exposure to organic solvents, alcohol consumption N5 l, and smoking included in the model. b B-coefficient shown for variables included in the final model, exclusion criterion p N =0.100. ⁎ p b 0.05. ⁎⁎ p b 0.01. a

categories no (0), low (1) and high (2) as an expression of exposure to acrylamide containing grouts. 3. Results

3.2. Electroretinography (ERG) The 30 Hz flicker stimulation amplitude was reduced in both exposed groups, being statistically significant only among the recently exposed (7.4 μV, vs. 9.7 μV among the referents)

3.1. Neurographic measurements Most neurographic results for the previously exposed group did not differ significantly from those of the control group (Table 2). Mean sensory nerve conduction velocity (NCV) in the sural nerve of the leg, however, was significantly reduced in the previously exposed group, compared to the control group (44.2 vs. 48.4 m/s, p = 0.005). The mean sensory amplitude of the sural nerve was also reduced, including four participants with non-elicited responses, compared to only one in the control group. The reduction of sensory sural amplitude was not statistically significant, however, due to considerable standard deviation. In a linear regression model, including age and height, and exposure to acrylamide, organic solvents, alcohol consumption and cigarette smoking, age and height were, as expected, significant determinants for several of the outcome measures. Acrylamide exposure was a statistically significant predictor for reduced NCV in the sural nerve and increased motor NCV in the tibial nerve (Table 3). Dichotomising the acrylamide-exposed group into high and low exposure subgroups did not increase contribution of the exposure to the explained variance.

Table 4 Electroretinography (ERG) amplitudes, and visual evoked response (VER) duration, latency and amplitude in the recently and previously exposed groups, and control group Exposed groups Recently (N = 24) Electroretinography – 2 Hz amplitude 12.1 (2.9) a–b (μV) – 30 Hz amplitude 7.4 (2.5) b* a–b (μV) Visual evoked response – N75 latency (ms) 76.8 (11.2)* – P100 (ms) 105.2 (11.2) – Duration (ms) 57.6 (8.5)* – Amplitude (μV) 7.2 (2.7)* Mean value for both eyes. *Independent samples t-test, p b 0.05. a N = 48. b N = 22. c N = 42.

Control group Previously (N = 43)

(N = 49)

13.5 (5.0)

13.2 (4.9) a

8.6 (4.2) c

9.7 (4.8) a

75.8 (6.4) c* 102.4 (7.5) c 58.3 (10.4) c* 5.1 (2.7) c

72.0 (7.3) 102.1 (6.5) 63.8 (10.8) 5.8 (2.4)

L.O. Goffeng et al. / Neurotoxicology and Teratology 30 (2008) 186–194

(Table 4). In this group inclusion of acrylamide- and NMA exposure and age in a linear regression model, confirmed acrylamide exposure as a significant contributing factor to reduced 30 Hz flicker stimulation amplitude (Table 6). 2 Hz flicker stimulation amplitude showed no significant differences between the two exposed groups and the control group. 3.3. Visual evoked response (VER) The VER latencies to the onset of the occipital potential (N75) were significantly prolonged in both the exposed groups compared to the controls (Table 4). Response duration was not prolonged. Acrylamide exposure contributed significantly in explaining prolonged latencies to the onset of the occipital potential (N75) in both the exposed groups (Tables 5 and 6) when age was also included in a linear regression model. However, the explained variance from the model was low. By including additional potential confounders in the model (eye injuries, concussion, alcohol consumption, cigarette smoking, exposure to organic solvents or lead) this pattern was not changed (R2 = .11, contribution of exposure to N75 latency in the previously exposed group: p = 0.006). Splitting the exposed groups into different exposure categories, based on reported exposure duration, or by using the described exposure-time index, revealed no correlation to degree of exposure. 4. Discussion The results indicate slight subclinical, but persisting toxic effects in the sural nerve and the visual system in tunnel workers exposed to N-methylolacrylamide and acrylamide during grouting operations. Persisting PNS effects following acrylamide exposure have been reported in severe cases [11,36,42]. Visual system effects have been reported in animal studies Table 5 Electroretinography (ERG) amplitudes, and visual evoked response (VER) in the previously exposed group and control group (N = 93) a Regression-coefficients b Constant ⁎⁎ ERG 2 Hz amplitude a–b (μV) 30 Hz amplitude a–b (μV) VER N75 latency (ms) P100 (ms) Duration (ms) Amplitude (μV)

R2

Model p-value

Age

Previous exposure

13.3

–

–

0.00

–

13.4

− 0.09

–

0.04

0.062

72.0 102.2 74.0 5.5

– – − 0.23 –

3.8 ⁎ – − 4.7 ⁎ –

0.07 0.00 0.10 0.00

0.011 – 0.008 –

a Mean value for both eyes. Backwards linear regression. Acrylamide/NMA exposure (0–1) and age included in the model. b B-coefficient shown for variables included in the final model, exclusion criterion p N =0.100. ⁎ p b 0.05. ⁎⁎ p b 0.01.

191

Table 6 Electroretinography (ERG) amplitudes, and visual evoked response (VER) in the recently exposed group and control group (N = 73) a Regression-coefficients b

ERG 2 Hz amplitude a–b (μV) 30 Hz amplitude a–b (μV) VER N75 latency (ms) P100 (ms) Duration (ms) Amplitude (μV)

R2

Model p-value

Constant ⁎⁎

Age

Recent exposure

12.9

–

–

0.00

–

9.7

–

−2.4 ⁎

0.07

0.031

72.0 103.1 63.8 8.6

– – – − 0.06

4.8 ⁎ – − 6.2 ⁎ 1.3 ⁎

0.06 0.00 0.08 0.11

0.032 – 0.017 0.016

a Mean value for both eyes. Backwards linear regression. Acrylamide/NMA exposure (0–1) and age included in model. b B-coefficient shown for variables included in the final model, exclusion criterion p N =0.100. ⁎ p b 0.05. ⁎⁎ p b 0.01.

[16,34] and human case studies, but to our knowledge not in epidemiologically designed studies. Although the results seem biologically plausible, methodological aspects of the study need to be discussed. In particular, the lack of exposure response relationships and the potential heterogeneity between the groups will be addressed. 4.1. Study population Recruitment of participants to the exposed groups was based purely on described work tasks and length of grouting work, not on subjective exposure assessment by the workers themselves. By using these more “objective” exposure criteria as a basis for recruitment, the possibility of outcome selective recruitment into the study should be reduced. For comparability reasons, regarding potential confounders, such as alcohol intake, previous and current health and work-related exposures, we chose another group of tunnel workers as control group. The high participant rate of eligible subjects into all three groups makes further selection bias less likely. The older age groups were over represented in the previously exposed group, however, and many participants had been involved in tunnel projects finished several years prior to the onset of the present study. This may explain why more people in this group were out of work at the time of examination. The observed exposurerelated outcomes, however, were not substantially influenced by employment status. As most neurophysiological outcome parameters are agerelated [28,37], age was included and adjusted for in all analyses. In most instances increasing age was significantly related to reduced peripheral nerve NCV and amplitudes. Heavy smoking [5,21] and alcohol consumption [44] may influence visual function. The proportion of present smokers was comparable in the two groups, while more participants reported

192

L.O. Goffeng et al. / Neurotoxicology and Teratology 30 (2008) 186–194

annual alcohol consumption N 5 l in the control group. More control workers than exposed workers reported previous exposure to lead. This could contribute to a blurring of a potential health effect from acrylamide exposure, but lead exposure primarily affects retinal rod function [8,9]. Visual function in addition to PNS effects may be influenced by previous exposure to organic solvents, reported more often in the previously exposed group [22]. Due to this, these covariates were included into the analyses. Nerve conduction velocities, and particularly F-latencies, are correlated to height. As mean height was the same in both groups, this parameter only changed exposure-related outcome estimates marginally when included in the analyses, indicating limited confounding from this factor in the present study. 4.2. Exposure characterization Questionnaire based surrogate variables of exposure duration and cumulative exposure did not reveal any exposure response relationships, nor did dichotomising the exposed group in a high and low exposure sub-group, based on reported exposure level and duration. This could weaken the probability of the observed group differences to be caused by acrylamide exposure. However, the previously exposed workers had been acrylamide exposed for many periods, some of them relatively short, and often several years back in time. Most workers took part in many different working tasks, with different exposure levels. Consequently, assessment of duration and level of their exposure was difficult, as was also stated by the participants. The exposure estimates based on this subjective reporting are therefore considered less accurate. The crude exposed/unexposed group allocation may consequently be the most relevant exposure measure in the present study. 4.3. Neurophysiological findings Reduced potential amplitudes are usually signs of axonal lesions or neuronal damage, whereas reduced conduction velocities or prolonged latencies are signs of lesions to the myelin. In our study we found small, but significant changes, some of them suggesting axonal lesions, others suggestive of demyelinisation, in peripheral nerves and the visual system. Among the non-significant differences in neurophysiological function, several changes, for instance NCV in the median, ulnar and tibial nerves, might indicate trends in favour of the exposed group. As there is no physiological rationale for a positive effect of acrylamide on nerve function, these findings are considered as truly insignificant. On the other hand, nonsignificant findings indicating a deletory effect of acrylamide in the exposed group do not exclude a slight, but possible real trend. 4.3.1. Peripheral nerves The significantly reduced sural sensory NCV in the previously exposed group suggests a toxic effect on myelin in this nerve. The non-significantly reduced sensory amplitude (6.1 vs. 8.2 µV) in the same nerve, including four non-elicited

potentials, may indicate a possible axonal damage. This corresponds to what was found in the group of recently acrylamide-exposed construction workers, where reduced sural sensory NCV and amplitude was observed 16 months compared to 4 months post exposure [26]. These independent but corresponding NCV results among two groups of tunnel workers with remote and recent exposure to acrylamide may indicate exposure-related mixed demyelinating and axonal changes in the sural nerve, and is in accordance with a recent review on acrylamide toxicity [41]. Our investigation further confirms earlier studies showing that the sural nerve appears to be more vulnerable than other peripheral nerves to toxins [25] and metabolic disturbances like diabetes [31]. The observed effect on sensory nerve conduction may have been underestimated since conduction velocity was not included in the statistics for individuals where sensory responses were too low for accurate measurements. Reduced sensory amplitude of the nerve action potential of long nerves has often been considered as the most prominent symptom of acrylamide exposure [11], reflecting axonal neuropathy. However, normal sural nerve action potential amplitude varies by a factor of 3–4 due to variation in number of large myelinated fibres, making the amplitude less well suited for screening in individuals exposed to toxic substances [28]. However, certain sensory nerve terminals have shown to be more vulnerable than adjacent motor nerve terminals [38]. Neither larger fibre diameter alone [12,38,47] nor fibre length [38] can fully explain this, and there has been a shift in attention from the axonal transport mechanisms to a focus on the properties of the nerve endings themselves, or impulse transmission, as the main locus of effect from acrylamide exposure [32]. Methodologically it is thus worth considering that activity in the most distal segments of the sensory nerves or the sensory receptors are not registered during sensory NCV, even though abnormalities may begin in, or be limited to, those regions [46]. Peripheral nerves consist of myelinated and unmyelinated nerve fibres with a wide spectrum of diameters and hence large differences in conduction velocities. Conventional NCV measures the very fast conducting myelinated large diameter fibres of 9–14μm [28]. In the sural nerve these axons make up less than 30% of all myelinated fibres [39]. The roughly 10% NCV-reduction in the present sample is most probably due to demyelinisation, but may also be caused by selective axonal affection with conduction block in the largest diameter fibres [12], or by initial axonal degeneration of distal parts of fibres, with secondary demyelination (mixed neuropathy) [11,28]. Signs of peripheral nervous system damage from acrylamide intoxication appear slowly, first in the feet and hands, later in the legs and arms [43]. The most distal parts of long and large diameter nerve fibres are especially vulnerable and first undergo slow, progressive retrograde degeneration, termed “dying back” [38]. This sequential progression, however, was not observed in our previous study [26]. In that study, most PNS effects were observed 4 months post exposure with recovery of the shorter nerves of the arms during the period from 4 to 16 months post exposure, while NCV and amplitude of the sural sensory nerve

L.O. Goffeng et al. / Neurotoxicology and Teratology 30 (2008) 186–194

showed a delayed effect, with reduction in amplitude and NCV from 4 to 16 months post exposure. In the present study a similar sural sensory nerve effect persists 2–10 years post exposure. In experimental studies animals have generally been exposed orally, by inhalation, or by injections [43]. The construction workers in the present study have been exposed through skin contact, probably primarily through wet hands and arms. Thus a local effect in the nerves of the arms may have preceded a systemic effect accounting for the delayed sural deterioration [26]. 4.3.2. Visual system The main findings from the visual system are reduced amplitude to cone specific 30 Hz stimulation in the recently exposed group, without similar reduction in the previously exposed group, and prolonged optic nerve VER N75 latency in both the previously and recently exposed groups. Electroretinography (ERG) response represents total retinal activity elicited by a light stimulus [13]. Although both rods and cones respond to 2 Hz flicker stimulation, this frequency primarily measures rod function. Normal 2 Hz ERG responses in both exposed groups and the controls indicate spared rod function. At 30 Hz flicker stimulation rod function is eliminated and only cone function is measured. The cones are photoreceptor cells found in high densities in the central part of the retina. They are especially important in visual acuity and colour vision. Reduced amplitudes at ERG 30 Hz flicker stimulation were found in both exposed groups, but were statistically significant only in the recently exposed group, indicating a slight cone dysfunction. As the recently exposed and the previously exposed groups represent different participants and not a follow up of the same group, the difference in ERG-results may simply express a different degree of retinal affection in the two groups. However, the more marked degree of cone dysfunction in the recently exposed group compared to those previously exposed (Table 4) may also be interpreted as a possible sign of partial recovery of cone function. This supports a suggestion that the neural retina may be more plastic than previously believed [23]. There was a slight statistically significant increase in VER P75-latency of the optic nerves in both the previously and recently exposed groups, suggesting a slight, persisting toxic effect on the myelin of the fastest conducting optic nerve fibres. Animal studies, however, have shown both demyelinisation and axonal damage in the optic nerve after acrylamide exposure: In monkeys sacrificed immediately after orally presented acrylamide exposure, a pattern showing axonal swellings most prominent in distal optic tract fibres within lateral geniculate nucleus, besides disproportionately thin myelin sheaths, with no involvement of retinal ganglion cells or axonal changes in retinal nerve fibre layer was seen [6]. On the other hand, monkeys sacrificed after 6–8 months, showed loss of retinal ganglion cells most prominent in the fovea, and of retinogeniculate optic nerve axons [7], with no reported signs of demyelinisation, indicating a ”dying back”-pattern of axonal changes, with a delayed retinal persisting toxic effect of exposure to acrylamide and N-methylolacrylamide.

193

The animal studies show structural changes, mainly in the early phases after exposure. Our study, however, demonstrates functional changes at a later stage after exposure since even our recently exposed group of tunnel workers is examined 16 months after exposure cessation, and indicates demyelinisation of the optic nerve or optic tract. The initial process leading to this is not shown and can only be subject to speculation. Demyelinisation may occur due to primary changes in the oligodendrocyte or myelin sheath itself, but may also be secondary to axonal changes like axonal swelling [6]. A significant reduction of amplitude would then be expected in the acute phase. This was not found in our study, but may have been present at an earlier stage. Our results may thus be in accordance with experimental animal studies [7] indicating effects from acrylamide primarily centrally in the retina, and in optic nerve axons [4,29,40]. 5. Conclusion The results of the present study indicate slight, persistent subclinical nervous system affection of the sural nerve and signs of retinal and optic nerve impairment. To our experience, this is the first human group study to show persisting visual system effects following acrylamide exposure. Uncertainty due to study sample heterogeneity, and failure to establish an exposure response relationship, may weaken the probability of the effects to be caused by exposure to acrylamide and N-methylolacrylamide containing grouting agents. However, the results are strengthened by compatible findings in two different samples examined at different time intervals after exposure cessation, and by corresponding results from experimental animal studies. Acknowledgements The study was supported by the Confederation of Norwegian Business and Industry Work Environment Fund, Scandinavian Rock Group ANS (SRG), Gardermobanen AS, and Rhodia PPMC. The funding sources had no role in study design, data collection, analysis or interpretation of data, or in the writing of the paper or in the decision to submit the paper for publication. References [1] R.B. Auld, S.F. Bedwell, Peripheral neuropathy with sympathetic overactivity from industrial contact with acrylamide, Can. Med. Assoc. J. 96 (1967) 652–654. [2] A. Cavalleri, F. Gobba, E. Nicali, V. Fiocchi, Dose-related color vision impairment in toluene-exposed workers, Arch. Environ. Health 55 (2000) 399–404. [3] K.H. Chiappa (Ed.), Evoked Potentials in Clinical Medicine, third ed., Lippincott-Raven Publishers, Philadelphia, 1997. [4] F. Di Russo, S. Pitzalis, G. Spinoti, T. Aprile, F. Patria, D. Spinelli, S.A. Hillyard, Identification of the neural sources of the pattern-reversal VEP, NeuroImage 24 (2005) 874–886. [5] C. Erb, T. Nicaeus, M. Adler, J. Isensee, E. Zrenner, H.J. Thiel, Colour vision disturbances in chronic smokers, Graefe. Arch. Clin. Exp. Ophthalmol. 237 (1999) 377–380. [6] T.A. Eskin, L.W. Lapham, J.P.J. Maurissen, W.H. Merigan, Acrylamide effects on the macaque visual system. II. Retinogeniculate morphology, Invest. Ophthalmol. Vis. Sci. 26 (1985) 317–329.

194

L.O. Goffeng et al. / Neurotoxicology and Teratology 30 (2008) 186–194

[7] T.A. Eskin, W.H. Merigan, Selective acrylamide-induced degeneration of color opponent ganglion cells in macaques, Brain Res. 378 (1986) 379–384. [8] D.A. Fox, L.W.-F. Chu, Rods are selectively altered by lead: II. Ultrastructure and quantitative histology, Exp. Eye Res. 46 (1988) 613–625. [9] D.A. Fox, D.B. Farber, Rods are selectively altered by lead: I. Electrophysiology and biochemistry, Exp. Eye Res. 46 (1988) 597–611. [10] A. Fujita, J. Shibata, H. Kato, Y. Amani, D. Itomy, E. Suzuki, T. Nakazawa, T. Takahashi, Clinical observations of three cases of acrylamide poisoning, Nippon Ijo Shimpo 1869 (1960) 37–40 (in Japanese). [11] P.M. Fullerton, Electrophysiological and histological observations on peripheral nerves in acrylamide poisoning in man, J. Neurol. Neurosurg. Psychiatry 32 (1969) 186–192. [12] P.M. Fullerton, J.M. Barnes, Peripheral neuropathy in rats produced by acrylamide, Br. J. Ind. Med. 23 (1966) 210–221. [13] T. Ganes, Retina, synsbaner og synscortex. Neurofysiologiske undersøkelser [Retina, visual pathways and the visual cortex. Neurophysiological examinations], Tidsskr. Nor. Lægeforen. 107 (1987) 1774–1776. [14] T.O. Garland, M.W.H. Patterson, Six cases of acrylamide poisoning, Br. Med. J. 4 (1967) 134–138. [15] A.C. Godin, B. Bengtsson, R. Niskanen, E. Tareke, M. Törnqvist, K. Forslund, Acrylamide and N-methylolacrylamide poisoning in a herd of Charolais crossbreed cattle, Vet. Rec. 151 (2002) 724–728. [16] A.C. Godin, R.R. Dubielzig, E. Giuliano, B. Ekesten, Retinal and optic nerve degeneration in cattle after accidental acrylamide intoxication, Vet. Ophtalmol. 3 (2000) 235–239. [17] M. Gonzalez, M. Velten, A. Cantineau, Increased acquired dyschromatopsia among solvent-exposed workers: an epidemiology study on 249 employees of an aluminum-foil printing factory, Int. Arch. Occup. Environ. Health 71 (1998) 317–324. [18] J. Graveleau, P. Loirat, V. Nusinovici, Polynévrite par l’acrylamide, Rev. Neurol. 123 (1970) 62–65. [19] L. Hagmar, M. Törnqvist, C. Nordander, I. Rosén, M. Bruze, A. Kutiainen, et al., Health effects of occupational exposure to acrylamide using hemoglobin adducts as biomarkers of internal dose, Scand. J. Work, Environ. Health 27 (2001) 219–226. [20] A.M. Halliday, Evoked Potentials in Clinical Testing, Churchill Livingstone, Edinburgh, 1993. [21] I.F. Hepsen, C. Evereklioglu, Defective visual field in chronic heavy smokers, Acta Ophthalmol. Scand. 79 (2001) 53–56. [22] A. Iregren, M. Andersson, P. Nylén, Color vision and occupational chemical exposures. I. An overview of tests and effects, Neurotoxicology 23 (2002) 719–733. [23] B.W. Jones, C.B. Watt, J.M. Frederick, W. Baehr, C.-K. Chen, E.M. Levine, A.H. Milam, M.M. Lavail, R.E. Marc, Retinal remodelling triggered by photoreceptor degenerations, J. Comp. Neurol. 464 (2003) 1–16. [24] C.M. Kesson, A.W. Baird, D.H. Lawson, Acrylamide poisoning, Postgrad. Med. J. 53 (1977) 16–17. [25] J. Kimura (Ed.), Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd edition, Oxford University Press, New York, 2001. [26] H. Kjuus, L.O. Goffeng, M.S. Heier, H. Sjöholm, S. Øvrebø, V. Skaug, B. Paulsson, M. Törnqvist, S. Brudal, Effects on the peripheral nervous system of tunnel workers exposed to acrylamide and N-methylolacrylamide, Scand. J. Work, Environ. Health 30 (2004) 21–29. [27] H. Kjuus, I.L. Hansteen, D. Ryberg, L.O. Goffeng, S. Øvrebø, V. Skaug, Chromosome aberrations in tunnel workers exposed to acrylamide and

[28] [29]

[30] [31]

[32] [33]

[34]

[35] [36] [37]

[38]

[39] [40]

[41]

[42]

[43] [44] [45]

[46] [47]

N-methylolacrylamide, Scand. J. Work, Environ. Health 31 (2005) 300–306. C. Krarup, F. Buchthal, Conduction studies in peripheral nerve, Neurobehav. Toxicol. Teratol. 7 (1985) 319–323. Z. Kubová, J. Szanyi, J. Langrová, J. Kremlácek, M. Kuba, K. Honegr, Motion-onset and pattern-reversal visual evoked potentials in diagnostics of neuroborreliosis, J. Clin. Neurophysiol. 23 (2006) 416–420. A.S. Kuperman, The Pharmacology of Acrylamides, Cornell University Graduate School, Itcaca, New York, 1957, pp. 1–86. J.R. Larsen, H. Sjoholm, K.F. Hanssen, L. Sandvik, T.J. Berg, K. DahlJorgensen, Optimal blood glucose control during 18 years preserves peripheral nerve function in patients with 30 years' duration of type 1 diabetes, Diabetes Care 26 (2003) 2400–2404. R.M. LoPachin, C.D. Balaban, J.F. Ross, Acrylamide axonopathy revisited, Toxicol. Appl. Pharmacol. 188 (2003) 135–153. C. Mapp, M. Mazzotta, G.B. Bartolucci, L. Fabbri, La neuropatia da acrilamide: prime osservazioni in Italia [Neuropathy due to acrylamide: first observations in Italy], Med. Lav. 68 (1977) 1–12 (In Italian). W.H. Merigan, E. Barkdoll, J.P.J. Maurissen, T.A. Eskin, L.W. Lapham, Acrylamide effects on the macaque visual system. I. Psychophysics and electrophysiology, Invest. Ophthalmol. Vis. Sci. 26 (1985) 309–316. P. Morviller, Propos sur un toxique industriel peu connu en France: L’acrylamide, Arch. Mal. Prof. Med. Trav. Secur. Soc. 30 (1969) 527–530. J.E. Myers, I. Macun, Acrylamide neuropathy in a South African factory: an epidemiologic investigation, Am. J. Ind. Med. 19 (1991) 487–493. M. Onofrj, A. Thomas, D. Iacono, G. D’Andreamatteo, C. Paci, Agerelated changes of evoked potentials, Neurophysiol. Clin. 31 (2001) 83–103. H.H. Schaumburg, H.M. Wisniewski, P.S. Spencer, Ultrastructural studies of the dying-back process. I. Peripheral nerve terminal and axon degeneration in systemic acrylamide intoxication, J. Neuropath. Exp. Neurol. 33 (1974) 260–284. J.M. Shefner, F. Buchthal, C. Krarup, Slowly conducting myelinated fibers in peripheral neuropathy, Muscle Nerve 14 (1991) 534–542. H. Shigeto, S. Tobimatsu, T. Yamamoto, T. Kobayashi, M. Kato, Visual evoked cortical magnetic responses to checkerboard pattern reversal stimulation: a study on the neural generators of N75, P100 and N145, J. Neurol. Sci. 156 (1998) 186–194. A. Shipp, G. Lawrence, R. Gentry, T. McDonald, H. Bartow, J. Bounds, N. Macdonald, H. Clewell, B. Allen, C. Van Landingham, Acrylamide: review of toxicity data and dose–response analyses for cancer and noncancer effects, Crit. Rev. Toxicol. 36 (2006) 481–608. E.A. Smith, F.W. Oehme, Acrylamide and polyacrylamide: a review of production, use, environmental fate and neurotoxicity, Rev. Environ. Health 9 (1991) 215–228. P.S. Spencer, H.H. Schaumburg, Nervous system degeneration produced by acrylamide monomer, Environ. Health Perspect 11 (1975) 129–133. G. Verriest, P. Francq, P. Perart, Results of colour vision tests in alcoholic and in mentally disordered subjects, Ophthalmologica 180 (1980) 247–256. M. Weideborg, T. Kallqvist, K.E. Odegard, L.E. Sverdrup, E.A. Vik, Environmental risk assessment of acrylamide and methylolacrylamide from a grouting agent used in the tunnel construction of Romeriksporten, Norway, Water Res. 35 (2001) 2645–2652. A.J. Wilbourn, Nerve conduction studies. Types, components, abnormalities, and value in localization, Neurol. Clin. N. Am. 20 (2002) 305–338. A.J. Wilbourn, Sensory nerve conduction studies, J. Clin. Neurophysiol. 11 (1994) 584–601.