Potential dermal exposure to methyl methacrylate among dental technicians; variability and determinants in a field study

Abstract

Methyl methacrylate (MMA) is a commonly used chemical in dental work that can cause dermatitis. Nineteen dental technicians participated in a field study in which potential dermal exposure to MMA and exposure determinants, including glove use and MMA vapour in the breathing zone, were repeatedly monitored during three consecutive days. Using patches placed on various parts of their hands we observed that the fingers and palms of the dental technicians were exposed to MMA, and their forefingers were significantly more exposed than their ring fingers; this is based on pooled data for both left and right hands (p = 0.04). The exposure variability was greater between workers than within worker (i.e. day-to-day variability), but the between worker variability was to some extent explained by a model which included the tested determinants. Neither the amount of MMA vapours in the breathing zone nor glove use was consistently correlated with the dermal exposure. Thus, the effects of glove use and the distribution of exposure to MMA on the hands in working environments needs to be further investigated.

---

Aims of the investigation

Dental work involves exposure to toxic, irritating and sensitising chemicals, notably monomeric methyl methacrylate (MMA):^1,2 a volatile, flammable, colourless liquid that is mixed with prepolymerised MMA powder to form various acrylic products, including dental appliances such as dentures and orthodontic splints. MMA is considered a sensitizing agent,³ but it can also cause neurological symptoms.^1,4–6 MMA has been shown, in several studies,^7–11 to be capable of causing contact dermatitis on hands. Indeed, Aalto-Korte et al (2007)¹¹ have shown MMA and ethyleneglycol dimethacrylate to be the most common allergens affecting dental technicians. However, despite knowledge of the dermal effects of MMA we are only aware of one study¹² in which potential dermal exposure to MMA was assessed. Furthermore, in the cited study dental work was performed under simulated workplace conditions. It was shown that dermal exposure can occur via spillage of liquid MMA onto the hands, hand contact with the product, or by touching contaminated surfaces. However, exposure to MMA has not been previously studied, as far as we know, in the field among dental technicians at different dental laboratories.

The importance of taking valid, appropriate, dermal exposure measurements and assessing the effects of key determinants (factors, e.g. work tasks and practices, which increase or decrease the exposure) has been emphasized in recent papers.^13,14 Failure to do so will inevitably reduce the value of the information obtained, and the scope for thoroughly evaluating the dermal exposure hazards. In addition, knowledge of determinants could lead to improvements in working environments by identifying risk factors related to “actual” exposure and provide objective reasons for introducing appropriate preventive measures in workplaces. Examples of working environments in which dermal determinants have been evaluated are those of farmers using pesticides and insecticides,^15–17 asphalt roofing workers¹⁸ and workers in the rubber and wood pellets industries.^19,20 However, we are not aware of any previous study in which dental technicians' potential dermal exposure to MMA and its determinants were simultaneously evaluated. Knowledge of exposure and simultaneous determinants would improve understanding of dermal risks and how to improve preventive measures.

The ideal strategy for measuring dermal exposure to chemicals in a given environment and the interpretation of the acquired data strongly depend on the variability of the exposure between workers and within workers (i.e. day-to-day variability), and between body locations (i.e. between different parts of an exposed individual). According to a recent review by Vermeulen et al.,¹³ temporal and personal variability, i.e. the proportions of the total variation in dermal exposure accounted for by within and between worker variability components, are quite similar to the corresponding proportions for respiratory exposure. However, the between-body location variance component is dermal-specific, since spillage and contact with contaminanted surfaces can be significant for exposure of a defined body part. In the study earlier referred to by Liljelind et al.,¹² they selected measurement spots on the hands and lower arm, since they assumed that these skin areas would be the most contaminanted during the dental work tasks. The results showed that the hands were the main exposed skin area.

The aims of this study were to obtain measurements of potential dermal exposure to MMA on different parts of the hands of dental technicians working at several dental laboratories, to acquire estimates (from repeated measurements) of the between- and within-worker variance components of exposure and to assess the effects of possible exposure determinants.

Material and methods

Subjects

Nineteen dental technicians participated in the study, ten men and nine women, working at four different dental laboratories in Sweden. All but one was right handed. Their work tasks included handling MMA fluid (approximately content of 99% MMA) which was mixed with prepolymerised MMA-powder. The amount MMA handled varies a lot from one appliance to another, depending on type and size, and if the dental technician was manufacturing new or repairing old appliances. Examples of dental appliances are orthodontic splints and dentures. When preparing dentures the dental technician kneads the dough with both hands before manually applying it into a mould. Each dental technician had an individual work bench and personally planned their workday. A typical work bench was applied with different compressed air driven tools for grinding and polishing cured MMA-materials and gypsum. The compressed air was also used to clean the dental appliances of dust. Tasks that involved larger amounts of MMA, like building new dentures, could be performed at a workbench equipped with a local exhaust.

Dermal exposure measurements

Potential dermal exposure was monitored on three consecutive days, as follows. Three 10 cm² (2.5 × 4 cm) patches made of active charcoal sandwiched between two layers of cotton fabric (Blücher GmbH, Germany) were attached with small safety pins or sewn onto both the left and right hands of a pair of cotton gloves. Patches were located at the forefinger tip, the ring finger tip and on the palm, as shown in Fig. 1. These locations were chosen since the forefinger and palm of the hand are known to be exposed areas during dental work.¹² The selection was done without considering any biological differences of skin among body parts. However, the ring finger was regarded as a “reference finger” that is likely to be less exposed than the forefinger. An individual wearing a pair of clean gloves (Vinyl Exam gloves from Evercare, Canada) attached the patches in a MMA-free environment, and removed them prior to extraction and analysis (see below). The dental technicians were asked to wear the cotton gloves during the whole working day. If this was not possible, they were asked to place them in clean glass bottles marked with their name and left or right hand, as appropriate, and closed with a tight screw lid for the period when they were not wearing them. Occasions when they did not wear cotton gloves included times when they were performing fine motoric grinding tasks and working with wet gypsum. If protective gloves were used, they were worn on top of the cotton gloves. In order to measure the airborne background exposure to MMA vapour, the workers wore a patch fastened with a safety pin at one side of the chest. Another patch placed inside a protective glove (of the same material as the gloves the workers used, i.e.vinyl, nitrile or latex) was worn on the opposite side of the chest to measure the background vapour that may diffuse through the protective gloves. The glove was unused and closed by a tight knot and fastened with a safety pin above the knot (thus, we did not perforate the closed portion of the glove). The instruction was to place the patch within the glove on the right side of the chest on all the workers. However, studies have demonstrated that the fastening of a sampler (e.g. a patch) at the right or left side in the breathing zone does not produce a statistical significantly difference in the estimation of the inhalation exposure. The substances considered are styrene and monoterpenes, which are as volatile as MMA.^21–23

Fig. 1 The positions of the patches on the hand (here attached to a cotton glove with small safety pins) referred in the text as sampling sites.

In order to assess the technicians' exposure determinants they were also asked to complete a questionnaire including questions about their use of protective gloves (always, occasionally or never), the brand of protective glove they wore (if used), MMA-work tasks and time expended on each task.

Field spikes were prepared daily by evenly spreading 2.8 mg or 5.6 mg MMA monomer (Aldrich, Germany, 99%) onto each of six patches. One field blank for each person and day was also prepared. After sampling, each patch was placed in a separate 20 ml glass bottle to which 10 ml of acetone (Burdick & Jackson, Fluka, Germany) had been added. The bottle was sealed with a screw lid equipped with a Teflon lining (CIAB, Sweden). In the analytic laboratory an internal standard (3 µl of 99.8% n-nonane, Fluka, Switzerland) was added with a calibrated pipette (Finnpipette, Thermo Labsystems, Finland) and the samples were automatically shaken for 30 min and stored at −18 °C with the patch left in the sample solution until analysis (within 10 days of the sampling). This procedure was followed at all but one of the dental laboratories, where the patches were stored in empty 20 ml glass bottle at + 8 °C for 1–4 days, and then at −18 °C at the analytical laboratory. The samples were extracted with 10 ml of acetone within ten days after sampling and treated as described above.

In a stability study¹² the between-day variability obtained by repeatedly analysing three field samples was estimated to be 11–14% after 10 days and 3.5–8% after 20 days at −18 °C.

Analysis

Approximately 1.5 ml of the acetone extract of each shaken sample was transferred to an analytical vial, and a 2 µl portion was automatically injected by an HP7683 Series Injector into an HP6890 gas chromatograph (GC) equipped with a fused silica column (DB-5 J&W Scientific, USA, 30 m × 0.32 mm id, coated with 95% dimethyl-/5% phenylmethylsiloxane, film thickness 1.0 µm) and a flame ionisation detection (FID) detector. Their MMA contents were then analysed as previously described,¹² except that the GC oven temperature program started at 35 °C instead of 40 °C. The limits of detection (LOD) and quantification (LOQ) for a 10 cm² patch were 1.82 µg and 6.00 µg, respectively.¹² Samples with MMA contents below the LOQ but above LOD were assigned the measured value, while samples below LOD were assigned contents of LOD/2 µg.

To verify the FID identification of MMA, and calibrate the FID signals, the compound was also identified in 12 field samples with high levels of MMA (above LOQ) and 8 field samples with low levels of MMA (close or below to LOQ) using a mass spectrometric system consisting of an autoinjector (Agilent 7683B Series injector) and a GC (Agilent 6890N) equipped with the same column as above coupled to a mass spectrometer (Agilent Technologies 5975 inert XL MSD) operating in full-scan mode. MMA was identified by comparing the spectra obtained to library spectra (NIST '05). One of the most abundant fragment ions of MMA, m/z 100, was used to verify and calibrate the FID signals.

Statistical analysis

All of the following statistical procedures were performed using SAS software PC 9.1 (SAS Institute, Cary, NC, USA); the natural logarithms of the MMA concentrations were analysed due to obvious skewness and heteroscedasticity in the data. Mixed effects models (Proc Mixed) were used to investigate the effects of the fixed variables (see below) and to estimate variance components associated with the random effects of worker (between-worker variability) and the error term (within-worker variability). Mixed-effects models are convenient for evaluating the magnitudes of fixed effects while taking account of the correlated errors inherent in a repeated measures design, where multiple measurements are obtained from the same individual. The form of the models used here was:

ln(X_ij) = Y_ij = µ_y + α_i + β_j + χR_ij + ε_ij (1)

where X_ij represents the exposure level of the i-th worker on the j-th day and Y_ij (the natural logarithm of X_ij) is the sum of µ_y, α_i, β_j, χR_ij, ε_ij, representing the fixed underlying mean exposure (log space), the effect of the determinant “glove use” by the i-th worker, the random effect of the j-th worker, the effect of the determinant “background patch”, the amount of MMA on “the background patch” for the i-th worker on the j-th day, and the random effect of the j-th day of the i-th worker, respectively. β_j and ε_ij were assumed to be normally distributed and independent, with means of zero, and variances of σ_B² and σ_W² of the logged exposure, representing the between- and within-worker variance components, respectively. The main purpose of applying the mixed model was to determine whether the fixed effects of glove use and air levels of MMA significantly affected logged exposure levels. The goodness of fit of the final models was evaluated by investigating the residuals using q-q plots, scatter plots of residuals versus predicted values and Shapiro-Wilk's test. We found no strong discrepancies from a normal distribution in our data, when log MMA values were used. Differences between sample sites (Fig. 1) were evaluated by pairwise, univariate Student's T-tests; exposure means for each sample site (1–6, and left/right hand) and dental technician were examined. The Umeå University Ethics Committee approved the study (approval no. EPN dnr 07-060M).

Results and discussion

Recovery from field spikes and analysis of field blanks

In a previous recovery study,¹² it was obtained a mean recovery from field spikes, prepared by adding MMA to patches and extracting them in acetone 20 minutes later, of 108%, with a standard deviation of 13%. All the samples were corrected for the recovery. However, in this study the patch samples from one dental laboratory could not be placed in a bottle with acetone after sampling. Therefore, a separate recovery study was performed and the preparation of the field spikes was done as following: we prepared 36 field spikes (12 each day, (6 “high” and 6 low “spikes”, as above)) and placed them in empty glass bottles. The mean recovery from these 36 field spikes was 79% (range 66–88%) with a standard deviation of 6.4%. (The mean recovery of low spikes was 81% and high spikes 77%.) All samples from this particular laboratory were corrected for 79% recovery.

The amounts in the spiked samples seems to be reasonably relevant since 38 collected samples (14%) had MMA amounts greater than 5.6 mg and 23 (8%) had amounts between 2.8 and 5.6 mg MMA.

The concentration of MMA was below the LOQ in 36 of the blank samples and very close to or slightly above the LOQ in the other ten. Hence, adjustments were made to the amounts found in the samples to account for the blank levels. A mean blank level for each measurement day was calculated and subtracted from the samples for each day, respectively, before the statistical analysis.

Dermal exposure to MMA

The key criterion for including data obtained from the technicians in the statistical analysis was that they had to have worked with MMA monomers during the study period; this excluded one of the female workers since she only worked with UV-curing material during the study days. In addition, the questionnaire showed that some of dental technicians had not worked with MMA materials during one or two of the three consecutive study days, and one person was on sick leave for two of the days, so data for these technicians on these days was either not collected or excluded from subsequent analysis.

Therefore, the results below are based on data obtained from 12 workers with complete measurement records from three days, four with records from two days and two from only one day, giving a total of 46 measurement days. The time estimates for different MMA work tasks were incomplete and therefore not included in subsequent statistical analysis.

The MMA contents of 276 dermal patch samples were analysed in total. Of these 18 (7%) showed MMA amounts below the LOQ but above LOD, and 20 (7%) was assigned LOD/2. These low amount samples are derived from individuals who had low exposure days and thus distributed evenly among the sample sites on the hands. Table 1 presents: the highest amount detected on any patch, the mean exposure and variance components for each sampling site (1–6); corresponding exposure levels for the left hand (summed values for sampling sites 1, 2 and 3), the right hand (summed values for sites 4, 5 and 6) and both hands; the arithmetic mean exposure at each of the dermal sites (which varied from 0.21 to 0.77 mg/cm²) and the geometric means (which were ca. 10-fold smaller than the arithmetic means); and the variance components for each sample site (1–6). The between- worker variability was two to six times higher than the within-worker variability. The variance components indicate that 66% to 85% of the total variability (σ²_B + σ²_W) was due to variations between workers (σ²_B) rather than between days for the same worker, and this distribution was consistent across all the sampling sites (data not shown). These findings are consistent with the results of studies by Kromhout and co-workers^13,24 in which a database of dermal exposure measurements (DERMDAT) was constructed from 20 surveys of workers from agricultural settings, the rubber industry, as well as coke-oven and paving workers. Further, in these cited studies^13,24 the variability components and factors influencing the exposure (determinants) were estimated and the results indicated that the tested determinants influenced the between-worker variance but not the within-worker variance. Statistical analysis of our data shows that the between-worker variability could be explained (but not significantly) by the amount of MMA found on the background patches and by glove usage, which reduced the between-worker variance component of approximately 50% and 30% for left and right hand, respectively (σ_B² in tables 1 and 2), while the within-worker variability was relatively unaffected by the determinants (σ_W² in tables 1 and 2). The small increase of the within-worker variability in our study is possibly because of the small samples available for our statistical analysis. Furthermore, the cited authors^13,24 attributed this to dermal exposure being “event-based” to a certain extent, occurring when workers touch contaminated surfaces and/or when occasional spills or splashes contaminate the workers' skin.

Table 1 Estimated parameters of exposure for the technicians' left hands, right hands and both hands at each sampling site (Fig 1), based on equation 1, with no determinants (α_i and χR_ij) included in the model. The maximum MMA exposure levels are shown, but not minimum levels, since they were below quantifiable amounts at all sites

Site	Max exposure mg/cm^2 (raw data)	Mean, arithmetic mg/cm^2 (raw data)	Mean, geometric mg/cm^2 (anti-logged data)	Between-worker variance component (logged data, σB^2)	Within-worker variance component (logged data, σW^2)
1	4.36	0.426	0.0515	6.79	2.84
2	2.03	0.220	0.0257	7.29	2.96
3	2.27	0.212	0.0328	4.95	1.96
4	6.53	0.661	0.0266	7.99	2.96
5	6.46	0.768	0.0379	7.43	3.86
6	6.68	0.552	0.0251	6.87	1.17
Left hand	6.22	0.857	0.150	5.15	2.04
Right hand	19.0	1.98	0.121	6.80	2.16
Both hands	25.2	2.84	0.308	6.07	1.94

---

Table 2 Estimated parameters of exposure for the technicians' left hands, right hands and both hands at each sampling site, based on equation 1, with the determinants (α_i and χR_ij) included in the model. The asterisks indicate the significance of the estimate obtained for patches in air, and the fixed effect of glove use. For glove use 0 and 1 refer to “never” and “occasionally”, respectively. “Always” is the reference level and thus data for this class of glove use is not included in the table

Sample site	Between-worker variance component (logged data, σ^2B)	Within-worker variance component (logged data, σ^2W)	Glove use estimate for 0; 1 (logged data)	Background patch estimate (logged data)
Significance of the fixed effects (gloves and background patch): * p < 0.05; ** p < 0.01.
1	4.13	3.11	3.36; 3.32*	0.131
2	4.76	3.30	2.55; 2.51	0.414
3	2.91	1.84	1.83; 2.34	0.609**
4	6.02	3.12	3.89; 2.98	0.00368
5	5.75	4.12	3.36; 2.79	0.00860
6	5.18	1.20	2.73; 2.39	0.298
Left hand	2.62	2.26	2.33; 2.78*	0.413
Right hand	4.66	2.36	3.11; 2.88	0.196
Both hands	3.69	2.14	2.64; 2.87	0.303

---

The background patches provide indications of the amount of MMA in the air of the work environment, but the exposure to MMA is not related to direct spills on the hand. On eleven low dermal exposure measurement-days, the amount of MMA extracted from the background patches corresponded to more than 21% of the total amount of MMA extracted from the six patches on the respective workers' pairs of gloves, while the corresponding proportions for the other measurement-days were all less than 14% (mean 2.1%, median 0.5%; data not shown). The background patch data did not improve the models for the sample sites (table 2), except for one site; left palm of the hand. At this specific site, the amount MMA found seems correlated to the background levels, which might be explained by the fact that all but one of the dental technicians were right-handed. They used their right hand to carry contaminated bottles or utensils in their hands, while the left palm of the hand was less exposed by direct contact with contaminated objects. However, the work task “kneading dough” is an exception. Overall, the amount MMA found on the background patches is not correlated to the dermal exposure. Despite the fact that MMA is a volatile substance that is likely to be absorbed by patches sampling the air surrounding a dental technician who is handling it, we could not find that association. Thus, MMA levels in the surrounding air seem not to be a predictor of dermal exposure.

The workers classified their glove use as never (three workers), occasionally (nine workers) or always (six workers) in their responses to the questionnaire. Glove use did not significantly explain the amount MMA collected on the sample sites, except for sample site 1(left forefinger) and left hand when added to the model (table 2). In the model including both of these determinants the effect of glove use is separated from the effect of background MMA, and it provides clear indications that in these two cases workers who occasionally used gloves were more heavily dermally exposed to MMA than the workers who always used gloves (p < 0.05, respectively). This shows that at one out of the six sample sites the gloves had a protective effect for the workers who always used gloves in comparison to those who used them “occasionally”. Remarkably, the estimated dermal exposure of individuals who never used gloves was only slightly higher than those of individuals who occasionally used gloves. For the left hand and both hands the estimates were even lower for the “never” group (table 2). One explanation could be the years of work tenure, i.e. work experience, which might have an impact on the exposure, but we have no information about years of experience for each worker. Further, it may also be at least partly because the former group only included three workers, a sample too small to facilitate reliable statistical analysis.

Overall, there is no convincing effect of glove use, which is not surprising since the gloves do not protect the workers from dermal MMA contamination. Several published papers have shown that MMA readily and rapidly penetrates the type of gloves the dental technicians use i.e.vinyl, nitrile and latex gloves.^25–28 The breakthrough time for neat MMA is less than two minutes for all three glove types used by the technicians in our study.²⁷ According to the responses to the questionnaire, the shortest work tasks involving handling MMA were estimated by the workers to last four minutes (data not shown). The workers also often kept the contaminated gloves on after they had finished their MMA work tasks, which extended the potential dermal exposure time.

MMA levels in one of the patches attached to the individuals' chests, representing background MMA exposure, and 36 (78%) attached to their chests within gloves, were below the LOQ. The means for these sets of patches were 0.018 mg/cm² (max. 0.33 mg/cm²) and 0.0028 mg/cm² (max. 0.037 mg/cm²), respectively, and we extracted quantifiable amounts of MMA from patches placed in gloves on the chest on 10 measurement-days (data not shown). Further, a poor association was found between the amounts in the background patches and the patches within gloves (r² = 0.41). The patches in the glove may have been exposed through occasional splashes or diffusion of MMA through the glove from contaminated hands and utensils that touched the outside of the glove. Another possible explanation is that evaporated MMA may have diffused through the gloves, but this possibility has to be further investigated.

The mean exposure measurements obtained for each sample site (1–6) from the 16 dental technicians for whom measurements had been obtained on at least two days were compared using a paired t-test. No significant differences were found in the measurements between the right and left forefingers (sites 1 and 5), nor between the right and left ring fingers (sites 2 and 4). However, a significant difference was found between the summed exposure of both forefingers and the summed exposure of both ring fingers (p = 0.043). Further, there was a significant difference between the exposure measurements for the left forefinger and left ring finger (p = 0.019), but not between the corresponding right hand measurements. The ring finger can be regarded as a reference finger since it may not be used as frequently as the forefinger in the dental technicians' work tasks, although the validity of this hypothesis has to be further examined. In a previous Swedish study, dermatitis was found more frequently on the left hands than the right hands of dental personnel.²⁹ However, we did not find a significant difference between the exposure on the technicians' left and right hands (p = 0.078) (according to measurements from all three sample sites on the two hands), and the mean exposure for all the workers was twice as high (arithmetic mean) on the right hand than the left hand (table 1). Thus, the distribution of exposure on technicians' hands, and its associations with clinical dermal symptoms warrants further investigation, especially in field studies.

Conclusion

This study highlights the importance of applying a sampling strategy that includes the use of repeated measurements to assess the variability of exposure, carefully selecting measurement sites, and considering potential determinants.

Acknowledgements

The Swedish Council for Working Life and Social Research supported this study. We are grateful to the students Frida Nilsson and Ellen Askne (Dental Technology Programme, Faculty of Odontology, University Hospital of Northern Sweden, S-901 85 Umeå, Sweden). We are also grateful to Associate Professor Kåre Eriksson, Occupational and Environmental Medicine, University Hospital of Northern Sweden, S-901 85 Umeå, for reviewing the paper.

References

1. P. A. Leggat and U. Kedjarune, Int Dent J, 2003, 53, 126–131.
2. N. Torbica and S. Krstev, Occup Environ Med, 2006, 63, 145–148.
3. Swedish National Board of Occupational Safety & Health. Occupational Exposure Limits, 2005;17. Stockholm, Sweden (in Swedish).
4. M. Donaghy, G. Rushworth and J. M. Jacobs, Neurology, 1991, 41, 1112–1116.
5. D. R. Sadoh, M. K. Sharief and R. S. Howard, Br Dent J, 1999, 186(8), 380–381.
6. A. M. Seppäläinen and R. Rajaniemi, Am J Ind Med, 1984, 5, 471–477.
7. T. Rustemeyer and P. J. Frosch, Contact Dermatitis, 1996, 34, 125–133.
8. L. Kanerva, K. Alanko, T. Estlander, R. Jolanki, A. Lahtinen and A. Savela, Contact Dermatitis, 2000, 42, 175–176.
9. K. Wrangsjö, C. Swartling and B. Meding, Contact Dermatitis, 2001, 45, 158–163.
10. A. T. Goon, M. Isaksson, E. Zimerson, C. L. Goh and M. Bruze, Contact Dermatitis, 2006, 55(4), 219–226.
11. K. Aalto-Korte, K. Alanko, O. Kuuliala and R. Jolanki, Contact Dermatitis, 2007, 57(5), 324–330.
12. I. E. Liljelind, K. A. Eriksson, L. O. Nilsson, I. B. Jonsson and Y. I. Burstrom, J. Environ. Monit., 2005, 7(5), 519–523.
13. R. Vermeulen, P. Stewart and H. Kromhout, Scand J Work Environ Health, 2002, 28(6), 371–385.
14. H. Kromhout, W. Fransman, R. Vermeulen, M. Roff and J. J. van Hemmen, Ann Occup Hyg, 2004, 48(3), 187–196.
15. J. de Cock, D. Heederik, H. Kromhout, J. S. M. Boleij, F. Hoek, H. Wegh and E. Tjoe Ny, Am Ind Hyg Assoc J, 1998, 59, 158–165.
16. P. A. Stewart, T. Fears, B. Kross, L. Ogilvie and A. Blair, Scand J Work Environ Health, 1999, 25(1), 33–38.
17. L. E. Blanco, A. Aragón, I. Lundberg, C. Lidén, C. Wesseling and G. Nise, Ann Occup Hyg, 2005, 49(1), 17–24.
18. M. D. McClean, R. D. Rinehart, A. Sapkota, J. M. Cavallari and R. F. Herrick, J Occup Environ Hyg, 2007, 4(Suppl 1), 118–126.
19. R. Vermeulen, R. P. Bos and H. Kromhout, Mutat Res, 2001, 490(1), 27–34.
20. K. Hagström, C. Lundholm, K. Eriksson and I. Liljelind, Ann Occup Hyg., 2008, 52, 685–694.
21. R. F. Malek, J. M. Daisy and B. S. Cohen, Appl Occup Environ Hyg, 1999, 14, 777–784.
22. K. Eriksson and L. Wiklund, J. Environ. Monit., 2004, 6, 563–568.
23. K. Eriksson, I. Liljelind, J. Fahlén and E. Lampa, Ann Occup Hyg, 2005, 49(6), 529–533.
24. H. Kromhout and R. Vermeulen, Ann Occup Hyg, 2001, 45(4), 257–273.
25. E.-C. Lönnroth and I. E. Ruyter, JOSE, 2003, 9(3), 289–99.
26. E.-C. Lönnroth, H. Wellendorf and I. E. Ruyter, Eur J Oral Sci, 2003, 111, 440–446.
27. H. Andreasson, A. Boman, S. Johnsson, S. Karlsson and L. Barregård, Eur J Oral Sci, 2003, 111, 529–535.
28. M. Nakamura, H. Oshima and Y. Hashimoto, J Prosthet Dent, 2003, 90, 81–85.
29. E.-C. Lönnroth and H. Shahnavaz, Swed Dent J, 1998, 22, 23–32.

---