Levels and predictors of airborne and internal exposure to ... - Nature

Journal of Exposure Science and Environmental Epidemiology (2012) 22, 291 - 298 & 2012 Macmillan Publishers Limited All rights reserved 1559-0631/12

www.nature.com/jes

ORIGINAL ARTICLE

Levels and predictors of airborne and internal exposure to manganese and iron among welders Beate Pesch1, Tobias Weiss2, Benjamin Kendzia1, Jana Henry3, Martin Lehnert1, Anne Lotz1, Evelyn Heinze1, Heiko Udo Ka¨fferlein4, Rainer Van Gelder5, Markus Berges6, Jens-Uwe Hahn7, Markus Mattenklott8, Ewald Punkenburg9, Andrea Hartwig10, Thomas Bru¨ning11 and The Weldox Group We investigated airborne and internal exposure to manganese (Mn) and iron (Fe) among welders. Personal sampling of welding fumes was carried out in 241 welders during a shift. Metals were determined by inductively coupled plasma mass spectrometry. Mn in blood (MnB) was analyzed by graphite furnace atom absorption spectrometry. Determinants of exposure levels were estimated with multiple regression models. Respirable Mn was measured with a median of 62 (inter-quartile range (IQR) 8.4 -- 320) mg/m3 and correlated with Fe (r ¼ 0.92, 95% CI 0.90 -- 0.94). Inhalable Mn was measured with similar concentrations (IQR 10 -- 340 mg/m3). About 70% of the variance of Mn and Fe could be explained, mainly by the welding process. Ventilation decreased exposure to Fe and Mn significantly. Median concentrations of MnB and serum ferritin (SF) were 10.30 mg/l (IQR 8.33 -- 13.15 mg/l) and 131 mg/l (IQR 76 -- 240 mg/l), respectively. Few welders were presented with low iron stores, and MnB and SF were not correlated (r ¼ 0.07, 95% CI 0.05 to 0.20). Regression models revealed a significant association of the parent metal with MnB and SF, but a low fraction of variance was explained by exposure-related factors. Mn is mainly respirable in welding fumes. Airborne Mn and Fe influenced MnB and SF, respectively, in welders. This indicates an effect on the biological regulation of both metals. Mn and Fe were strongly correlated, whereas MnB and SF were not, likely due to higher iron stores among welders. Journal of Exposure Science and Environmental Epidemiology (2012) 22, 291--298; doi:10.1038/jes.2012.9; published online 29 February 2012 Keywords: biomonitoring; ferritin; iron; manganese; welding fume

INTRODUCTION Occupational exposure to manganeses (Mn) has been associated with neurotoxic effects along a continuum of dysfunctions,1,2 ranging from sub-functional neurotoxic effects associated with low Mn exposure to manganism as a neurodegenerative disease following high exposures.3 Because of its widespread use in steel production and other applications and also due to environmental exposures, Mn-induced neurotoxicity became one of the major fields of the Manganese Health Research Program.4 With regard to the neurotoxic effects of Mn, the American Conference of Governmental Industrial Hygienists (ACGIH) is proposing to lower the threshold limit value (TLV) from 200 mg/m3 to 20 mg/m3. Based on findings in epidemiological studies5 - 8 and a recent meta-analysis9 showing reduced performance of Mn-exposed workers predominantly in tests associated with motor functions, the German MAK commission has

recommended 200 mg/m3 for inhalable Mn and 20 mg/m3 for respirable Mn as MAK values. Compliance with occupational exposure limits (OELs) is challenging for welding. Exploration of welding databases revealed that airborne exposure is frequently above 200 mg/m3.10 Experimental investigations indicated that due to the low aerodynamic diameter of particles in the welding fume, most of the Mn is respirable.11 Only few small studies, however, measured Mn in the respirable fraction of the welding fume (for a review, see Hobson et al.12). According to our knowledge, no study measured respirable and inhalable Mn in parallel. Protective measures need to be evaluated with regard to compliance with OELs. Regression models for estimating the effects of predictors on the exposure levels imply large-sized studies with comprehensive data on airborne and internal exposure to Mn and Fe with detailed workplace descriptions. A recent study identified 27 welding

1 Center of Epidemiology, Institute for Prevention and Occupational Medicine of the German Social Accident Insurance, Institute of the Ruhr-Universita¨t Bochum (IPA), Bu¨rkle-dela-Camp-Platz, Bochum, Germany; 2Department of Human Biomonitoring, Institute for Prevention and Occupational Medicine of the German Social Accident Insurance, Institute of the Ruhr-Universita¨t Bochum (IPA), Bu¨rkle-de-la-Camp-Platz, Bochum, Germany; 3Center of Medicine, Institute for Prevention and Occupational Medicine of the German Social Accident Insurance, Institute of the Ruhr-Universita¨t Bochum (IPA), Bu¨rkle-de-la-Camp-Platz, Bochum, Germany; 4Center of Toxicology, Institute for Prevention and Occupational Medicine of the German Social Accident Insurance, Institute of the Ruhr-Universita¨t Bochum (IPA), Bu¨rkle-de-la-Camp-Platz, Bochum, Germany; 5Division 1, Unit 1.3, Institute for Occupational Safety and Health of the German Social Accident Insurance (IFA), Alte Heerstrasse, Sankt Augustin, Germany; 6Division 3, Unit 3.1, Institute for Occupational Safety and Health of the German Social Accident Insurance (IFA), Alte Heerstrasse, Sankt Augustin, Germany; 7Division 2, Unit 2.1, Institute for Occupational Safety and Health of the German Social Accident Insurance (IFA), Alte Heerstrasse, Sankt Augustin, Germany; 8Division 2, Unit 2.3, Institute for Occupational Safety and Health of the German Social Accident Insurance (IFA), Alte Heerstrasse, Sankt Augustin, Germany; 9BG Holz und Metall, Seligmannallee, Hannover, Germany; 10Department of Food Chemistry and Toxicology, Karlsruhe Institute of Technology (KIT), Institute of Applied Biosciences, Kaiserstrasse, Karlsruhe, Germany; 11Institute for Prevention and Occupational Medicine of the German Social Accident Insurance, Institute of the Ruhr-Universita¨t Bochum (IPA), Bu¨rkle-de-la-Camp-Platz, Bochum, Germany. Correspondence to: Dr. Beate Pesch, Center of Epidemiology, Institute for Prevention and Occupational Medicine of the German Social Accident Insurance, Institute of the Ruhr-Universita¨t Bochum (IPA), Bu¨rkle-de-la-CampPlatz 1, 44789 Bochum, Germany. Tel.: þ 49 (0)234-3024536. Fax: þ 49 (0)234-3024505. E-mail: [email protected] Received 16 May 2011; accepted 3 October 2011; published online 29 February 2012

Exposure to manganese and iron among welders Pesch et al

292

studies of varying size for the estimation of predictors of airborne Mn exposure.12 Another study explored Mn and Fe in weldingfume databases.10 Similar statistical approaches are needed for estimating exposure-related determinants of systemic exposure to Mn in blood and iron metabolism. Mn is an essential trace element for a variety of physiological processes (for a review, see ATSDR13). Due to the structural similarity with iron (Fe), both transition metals share biological transport systems.14 Subjects with low iron stores show higher blood concentrations of Mn (MnB).15,16 Less is known about this association and iron homeostasis in welders who are exposed to both transition metals as constituents of the welding fume and may thus be prone to Mn and Fe overload.17 No guidance value has been set for MnB in German workers besides a biological reference value (BAR) of 15 mg/l. This threshold is the presumed 95th percentile of the distribution in the general population, but a representative survey on MnB in European or American men is yet lacking. WELDOX was conducted as a cross-sectional study among welders in various German industries with comprehensive exposure information to explore biological effects associated with metal exposure. Here, we analyzed the distributions of Mn in welding fume and blood, and explored potential predictors of the exposure levels as well as the associations with airborne Fe and serum ferritin (SF). MATERIALS AND METHODS Study Population Between 2007 and 2009, 241 welders were enrolled in a cross-sectional study in shipyards and other industries with data on airborne and internal metal exposure and detailed workplace information. Four welders in each shift were equipped with personal samplers before shift on a Tuesday, Wednesday, or Thursday. A trained team conducted the examination throughout the whole study period between 1400 and 1600 hours. The survey included a face-to-face interview, sampling of different biological specimens, and other investigations. All participants provided written informed consent. The study was approved by the Ethics Committee of the Ruhr University Bochum and was conducted in accordance with the Helsinki Declaration.

Characteristics of the Welding Materials and Worksites Exposure data were gathered within the framework of the exposuremonitoring and measurement system of the Institute for Occupational Safety and Health of the German Social Accident Insurance (IFA) and documented in the MEGA exposure database.18,19 Job tasks, welding process, use of protective measures (respiratory mask and powered airpurifying respirator (PAPR)), and other information were documented by in-person interviews and by the technicians performing the measurements of welding fume. Photos of the workplaces were used for an expert rating of the degree of confinement, the efficiency of the local exhaust ventilation (LEV), and the physical workload with binary scores. A larger distance of the nozzle from the breathing zone of the welders was considered as measure of inefficient ventilation. Information on welding material and electrodes was used for classification of the Mn content as ‘‘enriched,’’ if the content exceeded the median of 1.7%.

Air Sampling and Determination of Particulate Matter Welders were equipped with two personal samplers for collecting welding fume particles in the breathing zone inside of the helmets. Average duration of measurements was 3.5 h, ranging from 1.9 to 5 h. Cellulose nitrate filters (8 mm pore size, 37 mm diameter) were used to collect particles at a flow rate of 3.5 l/min. One sampler (PGP-EA) collected respirable welding fume, and the second sampler (GSP) collected inhalable particles.20 The filters were shipped to the central laboratory at IFA for gravimetric and metal analysis. Before weighing, the unloaded and particle-loaded collection media were conditioned for at least 1 day in the laboratory atmosphere. Environmental conditions such as humidity

were considered by calibration. The limits of detection (LODs) varied by duration of measurements, that is, with regard to total mass collected on the filters. In all, 90 gravimetric measurements of the respirable fraction and 47 measurements of the inhalable fraction of welding fume were below LOD.

Determination of Manganese and Iron in Welding Fumes Mn and Fe were determined by inductively coupled plasma mass spectrometry (ICP-MS) with a Perkin Elmer Elan DRC II (Waltham, MA, USA). The filters were digested with 10 ml of a mixture of nitric acid and hydrochloric acid. This solution was heated for 2 h under reflux in a heating block at 130 1C. After cooling to room temperature, the solution was diluted with 10 ml of ultrapure water to dilute the viscous solution before ICP-MS analysis was carried out. ICP mass spectrometer was calibrated with different multielement standard solutions covering the whole range of analytes. The isotopes, 45Sc, 85Rb, and 165Ho, were used as internal standards. Overall, 5 measurements of respirable Mn, 3 measurements of inhalable Mn, 23 measurements of respirable Fe, and 9 measurements of inhalable Fe were below the limits of quantitation (LOQs).

Determination of Manganese in Whole Blood and Serum Ferritin Post-shift samples of serum and K-EDTA whole blood were shipped in temperature-controlled containers to the Institute for Prevention and Occupational Medicine at Ruhr University Bochum overnight. Aliquots were stored at 80 1C until processing. SF was measured by ADVIA Centaur (Siemens, Eschborn, Germany) at the Institute of Clinical Chemistry, Transfusion and Laboratory Medicine, BG University Hospital Bergmannsheil, Germany. SF determination was a chemiluminometric sandwich immunoassay, standardized to WHO standard 80/578. For the measurement of Mn in whole blood, 50 ml of blood was diluted with 950 ml 0.2% Triton-X-100 in 0.1 HNO3. The sample was analyzed using graphite furnace atomic absorption spectrometry (ZEEnit 700; Analytik Jena, Jena, Germany) at 2000 1C and 279.5 nm. The injection volume was 15 ml. Interference due to matrix effects was largely eliminated by means of ashing in the presence of oxygen (675 1C, 30 s), the Zeeman background compensation, and the standard addition procedure. Commercially available external material (RECIPE, Munich, Germany) was used for quality control. All MnB concentrations were above LOQ (1 mg/l). Within-series imprecision was 3.0%, between-series imprecision 8.4%. Accuracy of analytical results was ensured by successful participation in an international external quality assessment scheme for analyses in biological materials.

Statistical Analysis Median and inter-quartile range (IQR) were used to describe the distributions of the exposure variables. We refrained from presenting estimates of LOD or LOQ in the distributions of air measurements because of their strong dependence upon the duration of particle collection. The concentrations of the variables under study (Mn, Fe, MnB, and SF) were log-transformed for parametric tests due to the skewness of their distributions. Values oLOD or oLOQ were imputed according to the underlying lognormal data distribution using a maximum likelihood method and a bootstrap algorithm with 1000 runs. The imputed Pearson (r) and Spearman (rs) correlation coefficients were presented with their 95% confidence limits (CI). Multivariate regression models were applied to the exposure variables to estimate potential predictors of their concentrations. The regression coefficients were presented with 95% CI at the original scales as factors modifying the concentrations of the respective dependent variable. Adjusted R2 estimate the explained proportion of variance as measure of the model fit. All calculations were performed with the statistical software SAS, version 9.2 (SAS Institute, Cary, NC, USA).

RESULTS Exposure to Manganese and Iron in the Welding Fumes Characteristics of the study population are presented in Table 1. Median age of the welders was 41 years (range 19 -- 61 years).

Journal of Exposure Science and Environmental Epidemiology (2012), 291 - 298

& 2012 Macmillan Publishers Limited


293 Table 1.

Characteristics of 241 welders enrolled in the WELDOX study.

Variable Age (years) Median (range)

41 (19 - 61)

Smoking status Current Former Never

123 (51.0%) 62 (25.7%) 56 (23.2%)

Industries Shipyards Manufacture of containers and vessels Machine and tool building

56 (23.2%) 139 (57.7%) 46 (19.1%)

Welding process Gas metal arc welding with solid wire Flux-cored arc welding with shielding gas Tungsten inert gas welding Stick electrode or miscellaneous

95 47 66 33

(39.4%) (19.5%) (27.4%) (13.7%)

Material (electrode or base metal) Stainless steel Mild steel

154 (63.9%) 87 (36.1%)

Manganese content o1.7% Z1.7%

117 (48.5%) 124 (51.5%)

Respiratory protection Powered air-purifying respirator Respiratory mask

26 (10.8%) 49 (20.3%)

Efficient ventilation Expert rating

54 (22.4%)

Confined space Expert rating

23 (9.5%)

Physical workload Expert rating

Figure 1. Association between respirable manganese and iron in 241 welders.

197 (81.7%)

Every other welder was an active smoker (51.0%), 56 (23.2%) welders reported having never smoked , and 62 (25.7%) workers had quit smoking. The industries comprised shipyards (N ¼ 56 welders), manufacture of containers, vessels, and related products (N ¼ 139), and machine or tool building (N ¼ 46). Gas metal arc welding with solid wire (GMAW) (N ¼ 95), flux-cored arc welding with shielding gas (FCAW) (N ¼ 47), tungsten inert gas welding (TIG) (N ¼ 66), and metal arc welding with stick electrodes (N ¼ 20). Additional 13 welders performed more than one welding process during the shift. Figure 1 shows a close correlation between respirable Mn and Fe in the welding fumes (r ¼ 0.92, 95% CI 0.90 -- 0.94). The regression was linear at the log-transformed scales with 1.001 as coefficient for concentrations above LOQ. Table 2 depicts the distribution of the exposure variables in all welders and by major welding process. The median concentrations of welding fumes were 0.97 mg/m3 for respirable and 1.12 mg/m3 for inhalable aerosols, with higher concentrations for FCAW (6.87 mg/m3 and 6.24 mg/m3, respectively) and measurements frequently below LOD for TIG. The concentrations of Mn and Fe followed this pattern by welding process. Median respirable Mn was 62 mg/m3 (inhalable: 73 mg/m3) with higher concentrations for FCAW (600 mg/m3 and 585 mg/m3, respectively) and lower concentrations for TIG (8.45 mg/m3 and 10.3 mg/m3, respectively). Figure 2 shows similar concentrations for Mn in respirable and inhalable welding fumes with a linear regression at the log-transformed scales with & 2012 Macmillan Publishers Limited

1.002 as estimate for the regression coefficient. With regard to the maximum allowable concentrations at workplaces in Germany (MAK values), 34% of the welders presented with inhalable Mn 4200 mg/m3, and 65% were exposed to respirable Mn 420 mg/m3. Welders applying TIG were frequently below these thresholds, whereas the majority of welders performing FCAW (89%) were observed with concentrations 420 mg/m3. Wearing PAPRs reduced median exposure to respirable Mn to 3.71 mg/m3 (Figure 3). Table 3 presents the estimates of potential other predictors of respirable Mn and Fe in the welding fumes. Exposure-related factors explained 72% of the variance of Mn and 73% of Fe, respectively. Efficient local exhaust ventilation reduced exposure significantly with factors of 0.38 for Mn (P ¼ 0.046) and 0.42 for Fe (Po0.0001). Marginally increased concentrations were found in confined spaces (Mn: exp(b) ¼ 1.87, P ¼ 0.06; Fe: exp(b) ¼ 1.64, P ¼ 0.09). The Mn content of the materials was an important exposure component. In comparison to GMAW with solid wire, TIG was associated with 0.08-fold lower Mn concentrations, and FCAW was related to 4.47-fold higher Mn concentrations. Welding of stainless steel in the electrodes or as base metal was found associated with lower airborne concentrations of both Mn and Fe in comparison to mild steel (Mn: exp(b) 0.59; Fe: exp(b) 0.31). Non-efficient ventilation resulted in higher exposure concentrations but of marginal significance. Welding in confined spaces was associated with a higher Mn concentration. Exposure to Manganese in Blood and Serum Ferritin Figure 4 shows no correlation between MnB and SF (r ¼ 0.07, 95% CI 0.05 to 0.20). The median blood concentrations were 10.30 mg/l for MnB and 131 mg/l for SF (Table 2). Using 20 mg/l as cutoff for low iron stores, eight welders were observed with postshift SF below that value. In contrast, 77 (32.0%) welders were found with SF 4200 mg/l. A total of 29 (12.0%) welders were 415 mg/l as presumed 95th percentile of the distribution of MnB in the general population. The differences in blood levels between the major welding processes reflected the pattern of respirable Mn and Fe but were less pronounced. Welders performing TIG had lower median concentrations (MnB: 8.71 mg/l, SF: 103 mg/l), and welders using FCAW were observed with higher concentrations (MnB: 13.09 mg/l, SF: 143 mg/l) (Table 2).



294 Table 2.

Exposure to welding fume, iron, and manganese in 241 welders during a working shift and blood concentrations of manganese and serum ferritin after shift presented with median, interquartile range in brackets, and the number of measurements (N) for respirable and inhalable particles (NR, NI) with measurements below the limit of detection (LOD; welding fume) or quantitation (LOQ; metals) in brackets.

N (NoLOD or LOQ) 3

Respirable welding fume (mg/m ) Inhalable welding fume (mg/m3) Respirable manganese (mg/m3) Inhalable manganese (mg/m3) Respirable iron (mg/m3) Inhalable iron (mg/m3) Manganese in blood (mg/l) Serum ferritin (mg/l)

241 (90) 203 (47) 241 (5) 230 (3) 241 (23) 230 (9) 241 241

Gas metal arc welding with solid wire NR ¼ 95; NI ¼ 93

Flux-cored arc welding with gas NR ¼ 47; NI ¼ 46

1.64 2.71 140 180 550 1110 11.48 165

6.87 6.24 600 585 1200 1652 13.09 143

Total 0.97 1.12 62 73 202 369 10.30 131

(n.d.; 3.42) (0.46; 3.73) (8.4; 320) (10; 340) (30; 910) (78; 1700) (8.33; 13.15) (76; 240)

Figure 2. Association between respirable and inhalable manganese in 241 welders.

Figure 5 depicts a non-linear association between respirable Mn and MnB (Spearman rs 0.44, 95% CI 0.33 -- 0.54). There was no obvious correlation between Mn in the welding fume and MnB up to 50 -- 100 mg/m3. Beyond that threshold we observed an increase of MnB with rising respirable Mn. Figure 6 shows a weak association between respirable Fe and SF (rs 0.23, 95% CI 0.11 -0.35). Table 4 presents the estimates of potential predictors of the internal exposure levels assessed as MnB and SF. Overall, exposure-related factors explained a small fraction of the variance of the biological exposure variables (MnB: R2 ¼ 0.16, SF: R2 ¼ 0.09). Due to the non-linear shape of the relation between respirable Mn and MnB, we transformed the Mn concentrations to their square-root values. Airborne Mn was a significant determinant of MnB (Po0.0001), and respirable Fe increased SF by a factor of 1.02/100 mg/m3 (P ¼ 0.001). An influence of respiratory masks on MnB and SF could not be detected. Physical workload increased MnB (P ¼ 0.04) but was not associated with SF. Age did not influence MnB within the observed age range. SF increased per 10 years of age 1.16-fold (95% CI 1.05 -- 1.29; P ¼ 0.005). DISCUSSION The majority of Mn is used in steel production for improving hardness. Steel contains up to 2.5% Mn. Special steel grades and

(n.d.; 3.34) (0.74; 5.02) (30; 320) (37; 360) (110; 1352) (249; 2000) (8.78; 13.52) (74; 283)

(3.36; 10.1) (1.05; 12.40) (270; 1100) (180; 1393) (480; 2600) (825; 3400) (10.32; 15.83) (92; 288)

Tungsten inert gas welding NR ¼ 66; NI ¼ 64 n.d. 0.56 8.45 10.3 19 77 8.71 103

(n.d.; 0.46) (n.d.; 0.93) (4; 22) (5.1; 26.5) (n.d.; 67) (32; 160) (7.15; 10.04) (73; 164)

Figure 3. Respirable manganese inside of 25 powered air-purifying respirators (PAPRs) and in the breathing zone of 137 GMAW welders not wearing PAPRs.

welding electrodes may have a higher Mn content, for example up to 10.5% as observed in our study. The Mn content of the materials, the emission rate of the welding process and many other factors contribute to a wide range of exposure to Mn. We confirmed the close correlations between Mn, Fe, and welding fumes as shown with data from large welding databases.10 The median concentrations of respirable Mn (62 mg/m3) and inhalable Mn (73 mg/m3) were comparable to a median of 60 mg/m3 from 114 Mn measurements of the database of The Welding Institute (accessible from http://www.twi.co.uk) and to a geometric mean of 97 mg/m3 in 96 Russian welders.21 Due to the skewed data distribution, the arithmetic mean of respirable Mn was 258 mg/m3 in our data set and corresponds to 270 mg/m3 reported as mean Mn from the welding database of the Occupational Safety and Health Administration (OSHA).10 Although large collections of welding-fume measurements have been assembled, we could not identify results on comprehensive side-by-side measurements of particle size-specific Mn concentrations among welders in the literature. We found similar Mn concentrations in the respirable (IQR 8 -- 320 mg/m3) and inhalable (IQR 10 -- 340 mg/m3) particle-size fractions of welding fume. Both particle-size-specific concentrations were strongly correlated. Respirable Mn comprised 7.9% of the mass of respirable welding




295 Table 3.

Potential predictors of the airborne concentrations of manganese and iron during a working shift among 215 welders.

Factor

Manganese (mg/m3)

Reference N

Intercept Flux-cored arc welding (FCAW) Tungsten inert gas welding (TIG) Stick electrode or miscellaneous Stainless steel Mn content Z1.7% Confined space Efficient ventilation Adjusted R2

42 66 29 128 110 36 48

N GMAW with solid wire

78

Mild steel Mn content o1.7% Non-confined Non-efficient or missing

87 105 179 167

Exp(b)

95% CI

P-value

124.48 4.47 0.08 0.63 0.59 2.32 1.87 0.38 0.72

84.62; 183.12 2.65; 7.54 0.05; 0.12 0.35; 1.14 0.40; 0.87 1.65; 3.26 1.05; 3.30 0.24; 0.60

o0.0001 o0.0001 0.12 0.008 o0.0001 0.06 0.046

Iron (mg/m3) Exp(b)

95% CI

P-value

1054.27 1.38 0.05 0.54 0.31

858; 1552 0.86; 2.21 0.04; 0.08 0.30; 0.98 0.21; 0.46

0.19 o0.0001 0.04 o0.0001

1.64 0.42 0.73

0.93; 2.90 0.27; 0.65

0.09 o0.0001

Figure 4. Association between manganese in blood and serum ferritin in 241 welders.

Figure 5. Association between respirable manganese and manganese in blood among 241 welders.

fume, but inhalable Mn 4.3% of the inhalable particles. These results support those derived in an experimental study of controlled Mn exposure where most of the Mn inside of helmets was respirable.11 The finding that most Mn is respirable is important for the compliance with gravimetric particle sizespecific OELs. For example, the maximum allowable concentration of 20 mg/m3 for respirable Mn is 10-fold lower than 200 mg/m3 for inhalable Mn as recommended by ACGIH and the German MAK commission. Hence, a higher proportion of welders (65%) were above the German MAK value for respirable Mn compared with 34% of welders with concentrations above the threshold for the inhalable fraction. All but one inhalable Mn concentrations measured in our study were below the US ceiling value of 5 mg/m3 for total fume as permissible exposure limit of OSHA (accessible from http://www.osha.gov). Hobson et al.12 applied multiple regression models to published means of Mn concentrations in welding fume for quantifying determinants of exposure levels. Our observations with individual measurements are overall in line with this regression analysis. We also revealed a good model fit with exposure-related factors for respirable Mn and Fe in welding fume and could demonstrate the major impact of the welding process. TIG resulted in much lower concentrations, where the median of respirable Mn was below the proposed threshold of 20 mg/m3. In contrast, FCAW with

shielding gas exposes the welders, on average, to concentrations of about 600 mg/m3, whereas neither LEV nor other technological improvements are currently capable of achieving compliance with 20 mg/m3. This high-emission process has been employed in the welding of large steel parts such as in ship building, where enclosed processes, such as in car production, are less feasible. New technologies with improved automation may reduce Mn exposure. Flynn and Susi10 explored welding databases, with particular focus on enclosure and ventilation. Our study confirmed that confined space and inefficient LEV increase exposure to welding fume. Whereas welding databases contain information whether ventilation was installed, we assessed additionally the efficiency of LEV by expert rating of photos taken at the workplaces. The photographic documentation of the workplaces revealed that the nozzle of the local exhaust ventilation was frequently not efficiently positioned with regard to the welding plume. LEV can be improved, for example by integration into the torch pipe. Overall, both confined space and inefficient LEV can increase the concentrations of respirable Mn and Fe at least about two-fold. Also other factors such as the individual working habits can contribute to the wide range of Mn concentrations.22 However, special preventive measures are needed to allow compliance of common welding




296

operations such as GMAW and FCAW with recommended OELs for respirable Mn. In a subgroup of welders wearing PAPRs, we could demonstrate a considerable reduction of respirable Mn to a median of about 4 mg/m3 inside the helmets. However, few welders with PAPRs presented with respirable Mn above 20 mg/m3. For comparison, ambient air contains 0.3 mg/m3 or less (ATSDR13 2008). This protective equipment can hardly be used in certain high-exposure settings where a confined space may hinder the movement of the welder, for example, in the double bottom of ships. It is important to note that the majority of welders worked without respirators or PAPRs. Mn-containing aerosols in welding fume are yet poorly characterized.22 Metal oxides in welding aerosols are usually small in size, with particle diameters frequently below 50 nm.23 Mn occurs in welding fume as oxides that can form spinels or spheric particles together with other metals, including iron. It is unknown how this complex chemistry influences the bioavailability of Mn and Fe.24 MnB and SF are common measures of the systemic homeostasis of Mn and Fe. Mn or iron overload during welding may disturb this homeostasis. We found significant associations of the post-shift concentrations of MnB and SF with the airborne concentrations of the parent metal in the respirable particle fraction of the welding fumes measured during the shift. Increased concentrations of MnB have been frequently observed in exposed vs non-exposed workers, for example in welders21 and in smelter

Figure 6. Association between respirable iron and serum ferritin in 241 welders. Table 4.

workers.25,26 We determined about 10 mg/l as median MnB in German welders that was similar to the average concentration reported for US welders in bridge building27 but lower than in Korean welders (15.5 mg/l).28 A lower average concentration of 8.6 mg/l in Russian welders might be explained by a longer time lag between shift exposure and blood sampling.21 Higher means were reported in alloy-production workers (12 mg/l).26,29,30 A receiver operating characteristic analysis determined 10 mg/l as 95th% percentile in a group of unexposed South African workers in comparison to smelters.26 Every other welder in our study was above that value. A level of 15 mg/l has been suggested as an estimate of the 95th percentile in the German population. About 20% of the welders were above that cut point. However, MnB has not yet been investigated in the German population with a representative survey. To the best of our knowledge, only the Korean health survey (KNHANES) included the determination of MnB in men and reported an average of 12 mg/l (Kim and Lee, 2011). Differences at group level may also indicate influences of diet, iron status, and other factors, which might impair to derive a common threshold for an increased MnB by occupational exposure at individual level. The cross-sectional design has been commonly employed in the investigation of biological markers in occupational settings or in population-based studies.31 Choosing a study design is usually a compromise, for example with the feasibility in the field. Repeated air measurements and examinations are challenging in occupational settings, in particular in the long term. The cross-sectional design has limited value in the exploration of temporal effects.32 To proof an accumulation during the week, we analyzed whether the week day influenced the concentrations of MnB and SF. We found no indication for such a short-term trend (data not shown), but indications for a chronic effect were found because MnB and SF were lower in TIG welders. Welders performed predominantly either TIG as a low-emission process or high-emission procedures during the previous years. Mn and Fe are ubiquitous metals in nature and essential for crucial functions in the maintenance of biological systems.14 Hence, homeostasis of the body burden with both metals employs complex biological regulation mechanisms. Besides other functions, Mn is a cofactor of important enzymes, such as manganese superoxide dismutase, and Fe is, for example, incorporated into heme. The strong biological regulation of Mn and Fe may be the reason why exposure-related factors explained only a small fraction of the variance of MnB and SF. However, we still observed a significant influence of the parent metals on both blood parameters. SF increased slightly with respirable Fe, without signs of a threshold. The association between Mn and MnB appeared with a non-linear shape with rising respirable Mn only above 50 -100 mg/m3. Although we observed a significant association of post-shift SF with Fe measurements during the working shift and relatively lower SF concentrations in TIG welders, the IQR (76 -- 240 mg/l) was similar to the distribution of SF in US males (IQR 87 -- 222 mg/l) from

Potential predictors of the concentrations of manganese in whole blood and serum ferritin after a working shift among 241 welders.

Factor

Manganese (mg/l) N

Intercept Respirable manganese Respirable iron Respiratory mask Physical workload Age Adjusted R2

Sqrt (mg/m3) Per 100 mg/m3 49 197 Per 10 years

Ferritin (mg/l)

Exp(b)

95% CI

P-value

8.57 1.01

6.97; 10.53 1.01; 1.02

o 0.0001

0.91 1.14 0.98

0.80; 1.03 1.01; 1.30 0.94; 1.03 0.16


0.14 0.04 0.48

Exp(b)

95% CI

58.53

35.83; 95.64

1.02 1.09 1.04 1.16

1.01; 1.03 0.82; 1.45 0.78; 1.40 1.05; 1.29 0.09

P-value

0.001 0.53 0.78 0.005



NHANES III.33 For German men, corresponding data on the distribution of SF in the general population could not be identified. Differences may be found with regard to low and high iron stores. Whereas population studies express concern on nutritional deficiencies in essential metals for some part of the population, welders are likely prone to an inhalative iron overload. Hence, only few welders were observed with low SF. Every third welder was presented with SF 4200 mg/l. It is current consensus that low iron stores result in higher MnB. However, less is known on Mn uptake in welders with regard to iron overload. In contrast to findings from population surveys in Norwegian women and Korean men, we observed no correlation between MnB and SF in welders.15,16 Organ distribution of Mn is dependent on the interaction with the iron metabolism, because divalent metals share transporter proteins.34 The divalent metal transporter 1 (DMT1) was found abundantly expressed also in the lung.35 A detailed analysis of the iron status in welders will be the topic of another publication.

CONCLUSION Concerns about neurotoxic effects of Mn guided governmental agencies to propose the lowering of OELs, in particular 20 mg/m3 for respirable Mn. These OELs are challenging for welding, where most of Mn occurs as small-sized and thus respirable aerosols. Hence, a large fraction of welders are exposed to Mn above the proposed threshold. Although MnB was associated with respirable Mn, biological regulation of this essential metal seems to control systemic exposure up to a certain exposure level. Few welders presented with low iron stores. This may explain that we could not confirm a negative association between MnB and SF, which has been observed in the general population. More research is needed to clarify the interdependency of Mn and Fe with regard to organ distribution and neurodegeneration.

CONFLICT OF INTEREST The authors declare no conflict of interest.

ACKNOWLEDGEMENTS The WELDOX study was financially supported by the German Social Accident Insurance (DGUV). We thank the technicians of DGUV, who measured the welding fume, all staff working for the DGUV measurement system, and all welders and companies having participated. We gratefully acknowledge the field team from the Institute for Prevention and Occupational Medicine of DGUV (IPA), especially Sandra Schoeneweis, Hans Berresheim, and Hannelore Ramcke-Kruell. We thank Christoph van Thriel from the Leibniz Research Centre for Working Environment and Human Factors, Dortmund (Germany) for his helpful comments on the manuscript.

REFERENCES 1 Martin C.J. Manganese neurotoxicity: connecting the dots along the continuum of dysfunction. Neurotoxicology 2006: 27: 347 - 349. 2 Mergler D., Baldwin M., Belanger S., Larribe F., Beuter A., and Bowler R., et al. Manganese neurotoxicity, a continuum of dysfunction: results from a community based study. Neurotoxicology 1999: 20: 327 - 342. 3 Couper J. On the effects of black oxide of manganese when inhaled into the lungs. Brit Ann Med Pharm 1837: 1: 41 - 42. 4 Aschner M., Lukey B., and Tremblay A. The Manganese Health Research Program (MHRP): status report and future research needs and directions. Neurotoxicology 2005: 27: 733 - 736. 5 Bast-Pettersen R., Ellingsen D.G., Hetland S.M., and Thomassen Y. Neuropsychological function in manganese alloy plant workers. Int Arch Occup Environ Health 2004: 77: 277 - 287. 6 Lucchini R., Apostoli P., Perrone C., Placidi D., Albini E., and Migliorati P., et al. Long-term exposure to ‘‘low levels’’ of manganese oxides and neurofunctional changes in ferroalloy workers. Neurotoxicology 1999: 20: 287 - 297.


297 7 Myers J.E., teWaterNaude J., Fourie M., Zogoe H.B., Naik I., and Theodorou P., et al. Nervous system effects of occupational manganese exposure on South African manganese mineworkers. Neurotoxicology 2003a: 24: 649 - 656. 8 Myers J.E., Thompson M.L., Ramushu S., Young T., Jeebhay M.F., and London L., et al. The nervous system effects of occupational exposure on workers in a South African manganese smelter. Neurotoxicology 2003c: 24: 885 - 894. 9 Meyer-Baron M., Knapp G., Schaper M., and van T.C. Performance alterations associated with occupational exposure to manganese--a meta-analysis. Neurotoxicology 2009: 30: 487 - 496. 10 Flynn M.R., and Susi P. Manganese, iron, and total particulate exposures to welders. J Occup Environ Hyg 2010: 7: 115 - 126. 11 Harris M.K., Ewing W.M., Longo W., DePasquale C., Mount M.D., and Hatfield R., et al. (Manganese exposures during shielded metal arc welding (SMAW) in an enclosed space. J Occup Environ Hyg 2005: 2: 375 - 382. 12 Hobson A., Seixas N., Sterling D., and Racette B.A. Estimation of Particulate Mass and Manganese Exposure Levels among Welders. Ann Occup Hyg 2011: 55: 113 - 125. 13 ATSDR (Agency for Toxic Substances and Disease Registry). Draft Toxicological Profile for Manganese. US Department of Health and Human Services, Public Health Service, ATSDR, Atlanta, GA, September 2008. www.atsdr.cdc.gov/ toxprofiles/tp.asp?id=102&tid=23. 14 Fitsanakis V.A., Zhang N., Garcia S., and Aschner M. Manganese (Mn) and iron (Fe): interdependency of transport and regulation. Neurotox Res 2010: 18: 124 - 131. 15 Kim Y., and Lee B.K. Iron deficiency increases blood manganese level in the Korean general population according to KNHANES 2008. Neurotoxicology 2011: 32(2): 247 - 254. 16 Meltzer H.M., Brantsaeter A.L., Borch-Iohnsen B., Ellingsen D.G., Alexander J., and Thomassen Y., et al. Low iron stores are related to higher blood concentrations of manganese, cobalt and cadmium in non-smoking, Norwegian women in the HUNT 2 study. Environ Res 2010: 110: 497 - 504. 17 Lu L., Zhang L.L., Li G.J., Guo W.R., Liang W.N., and Wei Z. [Altered systemic iron metabolism in welders exposed to manganese]. Zhonghua Lao Dong Wei Sheng Zhi Ye Bing Za Zhi 2006: 24: 31 - 34. 18 Gabriel S., Sun Y., and Chen W.H. [On-the-job Insurance League of Germany monitoring system for hazardous substances in working places]. GefahrstoffeReinhaltung der Luft - Air Quality Control 2010: 70: 43 - 49. 19 Stamm R. MEGA-database: one million data since 1972. Appl Occup Environ Hyg 2001: 16: 159 - 163. 20 Breuer D., Hahn J.U., Hober D., Emmel C., Musanke U., and Ruhl R., et al. Air sampling and determination of vapours and aerosols of bitumen and polycyclic aromatic hydrocarbons in the Human Bitumen Study. Arch Toxicol 2011: 85(Suppl 1): S11 - S20. 21 Ellingsen D.G., Dubeikovskaya L., Dahl K., Chashchin M., Chashchin V., and Zibarev E., et al. Air exposure assessment and biological monitoring of manganese and other major welding fume components in welders. J Environ Monit 2006: 8: 1078 - 1086. 22 Smargiassi A., Baldwin M., Savard S., Kennedy G., Mergler D., and Zayed J. Assessment of exposure to manganese in welding operations during the assembly of heavy excavation machinery accessories. Appl Occup Environ Hyg 2000: 15: 746 - 750. 23 Berlinger B., Benker N., Weinbruch S., L’Vov B., Ebert M., and Koch W., et al. Physicochemical characterisation of different welding aerosols. Anal Bioanal Chem 2011: 399: 1773 - 1780. 24 McMillan G. Is electric arc welding linked to manganism or Parkinson’s disease? Toxicol Rev 2005: 24: 237 - 257. 25 Cowan D.M., Fan Q., Zou Y., Shi X., Chen J., and Aschner M., et al. Manganese exposure among smelting workers: blood manganese-iron ratio as a novel tool for manganese exposure assessment. Biomarkers 2009: 14: 3 -- 16. 26 Myers J.E., Thompson M.L., Naik I., Theodorou P., Esswein E., and Tassell H., et al. The utility of biological monitoring for manganese in ferroalloy smelter workers in South Africa. Neurotoxicology 2003b: 24: 875 - 883. 27 Bowler R.M., Roels H.A., Nakagawa S., Drezgic M., Diamond E., and Park R., et al. Dose-effect relationships between manganese exposure and neurological, neuropsychological and pulmonary function in confined space bridge welders. Occup Environ Med 2007: 64: 167 - 177. 28 Chang Y., Song H.J., Lee J.J., Seo J.H., Kim J.H., and Lee H.J., et al. Neuroplastic changes within the brains of manganese-exposed welders: recruiting additional neural resources for successful motor performance. Occup Environ Med 2010: 67: 809 - 815. 29 Bouchard M., Mergler D., Baldwin M., Panisset M., and Roels H.A. Neuropsychiatric symptoms and past manganese exposure in a ferro-alloy plant. Neurotoxicology 2007: 28: 290 - 297. 30 Bouchard M., Mergler D., Baldwin M.E., and Panisset M. Manganese cumulative exposure and symptoms: a follow-up study of alloy workers. Neurotoxicology 2008: 29: 577 - 583.



298 31 Ceppi M., Gallo F., and Bonassi S. Study design and statistical analysis of data in human population studies with the micronucleus assay. Mutagenesis 2011: 26: 247 - 252. 32 Fowke J.H. Issues in the design of molecular and genetic epidemiologic studies. J Prev Med Public Health 2009: 42: 343 - 348. 33 Menke A., Muntner P., Fernandez-Real J.M., and Guallar E. The association of biomarkers of iron status with mortality in US adults. Nutr Metab Cardiovasc Dis 2011; (e-pub ahead of print).

34 Gunshin H., Mackenzie B., Berger U.V., Gunshin Y., Romero M.F., and Boron W.F., et al. Cloning and characterization of a mammalian proton-coupled metal-ion transporter. Nature 1997: 388: 482 - 488. 35 Brain J.D., Heilig E., Donaghey T.C., Knutson M.D., Wessling-Resnick M., and Molina R.M. Effects of iron status on transpulmonary transport and tissue distribution of Mn and Fe. Am J Respir Cell Mol Biol 2006: 34: 330 -- 337.

