Brand, Peter PhD; Bauer, Marcus MD; Gube, Monika MD; Lenz, Klaus Dipl Ing; Reisgen, Uwe Dr Ing; Spiegel-Ciobanu, Vilia Elena Dr Ing; Kraus, Thomas MD
From the Institute for Occupational and Social Medicine (Drs Brand, Bauer, Gube, and Kraus) and ISF—Welding and Joining Institute (Mr Lenz and Dr Reisgen), Aachen University of Technology, Aachen, Germany; and Institution for Statutory Accident Insurance and Prevention in the Woodworking and Metalworking Industry (Dr Spiegel-Ciobanu), Hannover, Germany.
Address correspondence to: Peter Brand, PhD, Institute for Occupational and Social Medicine, Aachen University of Technology, Pauwelsstr 30, D-52074 Aachen, Germany (firstname.lastname@example.org).
This project was funded by the Institution for Statutory Accident Insurance and Prevention in the Woodworking and Metalworking Industry (BGHM), Hannover, Germany, and by unrestricted grant (no. 360582) to the University Hospital Aachen, RWTH Aachen University.
Authors Brand, Bauer, Gube, Lenz, Ing, Reisgen, Ing, Spiegel-Ciobanu, Ing, and Kraus have no relationships/conditions/circumstances that present potential conflict of interest.
The JOEM editorial board and planners have no financial interest related to this research.
* Discuss previous findings linking emissions from metal inert gas brazing of zinc-coated materials to increased systemic inflammation.
* Outline the methods used in the new study to clarify the effects of occupational exposure to welding fumes containing different levels of zinc.
* Summarize the study findings, including their implications for safe exposure to zinc in welding fumes.
The German MAK Commission has recently proposed lower threshold limits for zinc-containing aerosols at workplaces.1 Whereas the workplace threshold limit for zinc oxide is at 5 mg m−3 in the Unites States, the MAK Commission proposes 0.1 mg m−3 as threshold limit for zinc (0.125 mg m−3 zinc oxide). It may be speculated that this threshold limit is difficult to observe in daily practice in some segments of metal working industries and crafts. More knowledge about effects of zinc-containing aerosols on human subjects would be helpful to evaluate the necessity of such low thresholds.
We have recently shown that inhalation of zinc oxide–containing welding fumes in a controlled experimental setting leads to asymptomatic systemic inflammation, which can be monitored the day after exposure by measuring the concentration of high-sensitivity C-reactive protein (hsCRP) in the blood.2 This biological effect was observed for metal inert gas (MIG) brazing of zinc-coated material at a zinc concentration of 1.5 mg m−3 in the welding fume. This experimental setting allows studying effects of well-defined welding fumes at different preselected fume concentrations in human subjects3 and therefore the evaluation of the “no observed effect level” (NOEL) for the welding fume under consideration. In the study presented here, the NOEL in respect to systemic inflammation as reflected by an increase of blood hsCRP was assessed in 12 healthy male subjects exposed to different concentrations of MIG brazing fumes of zinc-coated materials. Therefore, an adaptive study design with interims analyses after each exposure was performed. This study is part of a research project on health effects of welding fume particles that was designed in close cooperation with the Expert Committee “Metal and Surface Treatment” of the employers' liability insurance association, which supported this study.
Twelve healthy male subjects participated in this study. They were nonsmokers (9 never-smokers and 3 ex-smokers) and had no history of asthma or any other lung or cardiac disease. Table 1 shows the anthropometric and lung function data of the study population.4 The study protocol was approved by the ethics committee of the medical faculty of RWTH Aachen University. Informed written consent was obtained from each subject before inclusion.
The aim of this study was to find the zinc concentration in welding fumes at which the onset of signs for systemic inflammation as previously found5 is located. This onset is supposed to be found between the value proposed by the MAK Commission (0.1 mg m−3 of zinc) and the concentration investigated in the previous study (1.5 mg m−3 of zinc).5 To find this onset of inflammation, an adaptive three-fold crossover study design was used. On three different study days, 12 subjects were exposed for 6 hours to different welding fume concentrations, and biological effects were measured before, after, and 24 hours after exposure. The time between each exposure was 1 week. Once an hour, each subject cycled inside the exposure laboratory for 10 minutes on a bicycle ergometer at 80 W to simulate physical work. At the first exposure day, subjects were exposed to a zinc concentration of 0.9 mg m−3 (mass of particles with diameters below 10 μm [PM10] = 1.43 mg m−3; Table 2). Then, an interims analysis was performed, and it was calculated whether the observed average increase of hsCRP from baseline to the value 24 hours after exposure was different from zero. If the difference was unequal to zero (effect), the zinc concentration in the welding fume was reduced to 0.43 mg m−3 (PM10 = 0.70 mg m−3), and all 12 subjects were exposed again. According to the result of the second interims analysis, the zinc concentration in the next exposure step was either reduced to 0.11 mg m−3 (PM10 = 0.18 mg m−3) or the study was aborted (Fig. 1). If the first interims analysis showed a difference equal to zero (no effect), fume concentration was increased in the next step to 1.2 mg m−3 (PM10 = 2.0 mg m−3) and then, in the next step, either increased to 1.5 mg m−3 (PM10 = 2.5 mg m−3) or the study was aborted. This study design allows locating the NOEL within the following zinc concentration intervals: 0 to 0.11, 0.11 to 0.43, 0.43 to 0.9, 0.9 to 1.2, and 1.2 to 1.5 mg m−3.
Welding fume exposure was performed in the Aachen Workplace Simulation Laboratory, which has been described in detail elsewhere.3 In brief, the Aachen Workplace Simulation Laboratory consists of two different units: the emission room, in which the welding fumes were generated, and the exposure laboratory, in which six test subjects at a time were exposed. Both units were connected by a ventilation system, which allowed regulating the concentration of welding fumes. Welding was performed intermittently for 45 seconds every 10 to 15 minutes. The technique to adjust the average particle mass concentration is illustrated in Fig. 2: The mean mass concentration (mean) is determined by the concentration offset (offset) and the concentration increase in each welding cycle (Δm). This increase is determined by the welding time and the parameters of the ventilation system. The offset is determined by the concentration value at which the next welding cycle is initiated. During the study, variation of the offset was used for the fine adjustment of the average mass concentration.
To assess the average particle mass concentration and to keep the exposure dose identical among the 12 subjects, the time integral (E) of exposure mass concentration C was calculated on-line for each individual subject:
Equation (Uncited)Image Tools
where t0 is the time of exposure start and t the current time. For each subject, exposure persisted until E reached the desired value (eg, 1.4 mg m−3 ⇒ 8.4 mgh m−3, 2.0 mg m−3 ⇒ 12 mgh m−3, 2.5 mg m−3 ⇒ 15 mgh m−3).
Welding Technique and Materials
Metal inert gas brazing was performed manually by an experienced welder below a funnel-shaped fume hood. The base material was a hot-dip zinc-coated steel sheet (EN 10346: DX51D+Z275) and the filler metal consisted of 96% copper, 1% manganese, and 3% silicium (CuSi3Mn) (Bercoweld S3, Bedra, Germany). As shielding gas, argon was used. Metal inert gas brazing was performed for 45 seconds every 10 to 15 minutes. The time interval between two welding cycles was modified to adjust the mass concentration offset.
Characterization of the Welding Fume
Particle mass concentration during welding fume exposure was continuously monitored using a tapered element oscillating microbalance (series 1400A, Thermo Scientific, Franklin, MA). Particle number–size distribution and total number were measured using a fast mobility particle sizer (model 3091, TSI, Shoreview, MN). For the continuous monitoring of welding-related gases, electrochemical sensors for NO, CO2, and CO (ADOS, Aachen, Germany) and a UV photometrical sensor for Ozon (series 1400A, Thermo Scientific, Franklin, MA, model 49i) were applied.
Biological Endpoint Parameters
Primary endpoint parameter was hsCRP as measured by immune nephelometry. Secondary endpoints were coagulation factor VIII and ristocetin cofactor measured in the blood, and cell counts (leukocytes, erythrocytes, platelets, neutrophil granulocytes, and eosinophil granulocytes).
Statistical Package for the Social Sciences (SPSS) software for Windows (version 20.0; Statistical Products and Service Solutions, Inc, Chicago, IL) was used to analyze the data. For each endpoint variable Yi, two differences were calculated: Yi,6-0—value directly after exposure minus value before exposure, and Yi,24-0—value 24 hours after exposure minus value before exposure.
Wilcoxon test was used to assess if Yi,24-0 for hsCRP is statistically different from zero. If the difference was significantly different from zero, an exposure-related biological effect was assumed.
Analysis of variance (parametric) and Kruskal-Wallis test (nonparametric) served to investigate whether the endpoint differences were dependent on the exposure scenario. To evaluate group differences, post hoc comparisons were performed using Tukey α adjustment. Results were considered as statistically significant if P < 0.05.
After the first interims analysis, it turned out that exposure with 0.9 mg m−3 zinc (PM10 = 1.4 mg m−3) did not lead to an hsCRP change different from zero (P = 0.85). Therefore, the next exposure was performed with a zinc concentration of 1.2 mg m−3 (PM10 = 2.0 mg m−3). Interims analysis showed that after this exposure, a significant increase of hsCRP was present (P < 0001). Therefore, the third exposure was performed with a zinc concentration of 1.5 mg m−3 (PM10 = 2.5 mg m−3), which, again lead to an increase of hsCRP (P = 0.026).
Characterization of Exposure
Figure 3 shows, for one of the groups, the time course of particle mass concentration on the three exposure days. Table 3 shows the actual average mass concentration on each exposure day. As can be seen, the given mass concentration can be adjusted with high precision. The deviation between the given concentration and the actual value was less than 1%. The average count median diameter of the particles during exposure was 113 ± 3 nm (mean of 1-hour averages ± SD). Details about the particle size distributions of MIG brazing welding processes can be found in Hartmann et al2 and Brand et al.6 Concentrations of welding-related gases were very low (ozone <50 μg/m3, NOx and CO not detectable).
Biological Endpoint Parameters
Both, analysis of variance (P = 0.024) and Kruskal-Wallis test (P = 0.001) showed a significant dependency of the difference hsCRP 24 hours after exposure versus hsCRP before exposure from the exposure dose (Fig. 3; Table 4). Post hoc analysis of group differences showed a significant difference between exposure concentration PM10 = 1.4 mg m−3 and PM10 = 2.0 mg m−3 (Fig. 4). In all but two measurements, C-reactive protein (CRP) was within the reference range (<5 mg/L). The change in coagulation factor VIII (24 hours after exposure versus before exposure) showed a slight dependency on exposure concentration in the analysis of variance (P = 0.045) but failed to reach statistical significance using the Kruskal-Wallis test (P = 0.083; Fig. 5). Post hoc analysis of group differences showed again a significant difference between exposure concentration PM10 = 1.4 mg m−3 and PM10 = 2.0 mg m−3. All other endpoint parameters showed no significant dependency on particle concentration.
It has been shown that for PM10 exposure concentration 2 and 2.5 mg m−3, MIG brazing fumes from zinc-coated materials induced an increase of CRP in the blood of healthy subjects, whereas a PM10 value of 1.4 mg m−3 induced no increase. Therefore, it may be concluded that in the setting investigated in this study, the NOEL for systemic inflammation is between 1.4 and 2 mg m−3. This means that for zinc oxide, the threshold level is between 1.1 and 1.5 mg m−3 and for zinc, between 0.9 and 1.2 mg m−3. The observed distinct increase of CRP in all but two cases did not result in CRP values greater than the reference range. Before exposure, the baseline CRP values were, in average, between 0.5 and 1 mg/L. Twenty-four hours after exposure, the average increased to 1.8 mg/L (PM10 = 2 mg m−3) and 2.0 mg/L (PM10 = 2.5 mg m−3).
It is well known that exposure to zinc-containing aerosol particles can induce toxic and inflammatory reactions in cells, animals, and humans in which exposures to high concentrations of zinc aerosols may induce zinc fume fever.7–11 Kuschner et al12 exposed 15 healthy subjects for 3 hours to zinc oxide fumes with median mass concentrations of 33 mg m−3 (range, 20 to 42 mg m−3). They found increased levels of tumor necrosis factor α in the bronchoalveolar lavage fluid, an indicator for inflammation. Fine et al13 exposed 13 subjects for 2 hours to two different mass concentrations of zinc oxide (2.5 and 5 mg m−3). They found, for the higher concentration, increased levels of plasma interleukin 6 and body temperature, but they observed no increase in tumor necrosis factor α in the blood. Beckett et al14 exposed 12 human subjects for 2 hours to 0.5 mg m−3 zinc oxide administered either as fine or ultrafine particles. They found no effects of the exposure on hematological parameters, inflammation markers in the blood, and cardiac electrophysiology. Our recent study,5 finally, has shown that MIG brazing fume with a zinc oxide concentration of 1.9 mg m−3 induced an increase of hsCRP in 12 healthy subjects. From these studies, it may be concluded that the NOEL for zinc oxide is, in respect to systemic inflammation, less than 1.9 mg m−3. The German MAK Commission1 took the study of Beckett et al14 as a basis for the determination of the workplace threshold concentration. They extrapolated the data for a work shift duration of 8 hours and concluded that 0.1 mg m−3 zinc should be safe. The data presented in this study indicate that even higher concentrations of zinc should be safe—at least no increase of hsCRP as a sensitive indicator for systemic inflammation could be detected for zinc oxide concentrations less than 1.5 mg m−3 (zinc concentration, 1.2 mg m−3).
In a previous study, it was found that the zinc content of the welding fume emitted from a MIG brazing process with a CuSi3Mn welding wire on zinc-coated material is 60% of the total mass. Nevertheless, long-term stability of this finding was not measured, because it is known from previous studies15 that for given materials and welding parameters, welding fume emission rates and elemental composition are quite stable.
Emissions from MIG brazing do not contain zinc alone. Analysis of the emitted aerosol particles has shown that particles contain 60% zinc and 17% of copper.5 Like many metals, copper has adverse effects on cells and animals16,17 and may lead to impaired host defense and inflammation. Therefore, it cannot be excluded that the observed reactions in healthy subjects may be modulated by the copper within the fume. Riley et al18 studied the effects of various metals individually and in combinations in epithelial cell. They found that zinc is able to diminish the negative effect of copper but has an additive effect with nickel. This means that threshold limits given for single components of workplace emissions are in principle problematical. The advantage of this study design is that we investigated emissions from a real welding process with a certain industrial significance (eg, automotive industry and its components suppliers). For this process, MIG brazing with a CuSi3Mn welding wire on zinc-coated steel, the NOEL for systemic inflammation (as detected by hsCRP) was determined. For other processes—even with the same zinc content—effects may be different.
In this article, it has been shown that for the fumes emitted from a MIG brazing process with a CuSi3Mn welding wire on zinc-coated steel, the NOEL for systemic inflammation as detected by an increase of hsCRP is found for welding fume concentrations (PM10) between 1.4 and 2 mg m−3 (containing 0.9 to 1.2 mg m−3 zinc).
1. MAK Commission. Zink und seine anorganischen Verbindungen, Toxikologisch-arbeitsmedizinische Begründungen für MAK-Werte [in German]. Weinheim, Germany: Wiley-VCH; 2010.
2. Hartmann L, Bauer M, Bertram J, et al. Assessment of the biological effects of welding fumes emitted from metal inert gas welding processes of aluminium and zinc-plated materials in humans. Int J Hyg Environ Health. 2013 May 29 [epub ahead of print]. doi:10.1016/j.ijheh.2013.04.008.
3. Brand P, Havlicek P, Steiners M, et al. Exposure of healthy subjects with emissions from a gas metal arc welding process: part 1—exposure technique and external exposure. Int Arch Occup Environ Health. 2013;86:25–30.
4. Quanjer PH, Tammeling GJ, Cotes JE, Pedersen OF, Peslin R, Yernault JC. Lung volumes and forced ventilatory flows. Eur Respir J. 1993;6:5–40.
5. Hartmann L, Bauer M, Bertram J, et al. Assessment of the biological effects of welding fumes emitted from metal inert gas welding processes of aluminium and zinc-plated materials in humans. Int J Hyg Environ Health. In press.
6. Brand P, Lenz K, Reisgen U, Kraus T. Number size distribution of fine and ultrafine fume particles from various welding processes. Ann Occup Hyg. 2013;57:305–313.
7. Mueller EJ, Seger DL. Metal fume fever—a review. J Emerg Med. 1985;2:271–274.
8. Cosma G, Fulton H, DeFeo T, Gordon T. Rat lung metallothionein and heme oxygenase gene expression following ozone and zinc oxide exposure. Toxicol Appl Pharmacol. 1992;117:75–80.
9. Xia T, Kovochich M, Liong M, et al. Comparison of the mechanism of toxicity of zinc oxide and cerium oxide nanoparticles based on dissolution and oxidative stress properties. ACS Nano. 2008;2:2121–2134.
10. Zhang J, Song W, Guo J, Sun Z, Ding F, Gao M. Toxic effect of different ZnO particles on mouse alveolar macrophages. J Hazard Mater. 2012;219/220:148–155.
11. Ross D. Welders' metal fume fever. J Soc Occup Med. 1974;24:125–129.
12. Kuschner WG, D'Alessandro A, Wong H, Blanc PD. Early pulmonary cytokine responses to zinc oxide fume inhalation. Environ Res. 1997;75:7–11.
13. Fine JM, Gordon T, Chen LC, Kinney P, Falcone G, Beckett WS. Metal fume fever: characterization of clinical and plasma IL-6 responses in controlled human exposures to zinc oxide fume at and below the threshold limit value. J Occup Environ Med. 1997;39:722–726.
14. Beckett WS, Chalupa DF, Pauly-Brown A, et al. Comparing inhaled ultrafine versus fine zinc oxide particles in healthy adults: a human inhalation study. Am J Respir Crit Care Med. 2005;171:1129–1135.
15. Reisgen U, Olschok S, Lenz K, Spiegel-Ciobanu EV. Ermittlung von Schweißrauchdaten und Partikelkenngrößen bei verzinkten Werkstoffen. Schweißen und Schneiden. 2012;64:788–796.
16. Cho WS, Duffin R, Poland CA, et al. Metal oxide nanoparticles induce unique inflammatory footprints in the lung: important implications for nanoparticle testing. Environ Health Perspect. 2010;118:1699–1706.
17. Kim JS, Adamcakova-Dodd A, O'Shaughnessy PT, Grassian VH, Thorne PS. Effects of copper nanoparticle exposure on host defense in a murine pulmonary infection model. Part Fibre Toxicol. 2011;8:29.
18. Riley MR, Boesewetter DE, Kim AM, Sirvent FP. Effects of metals Cu, Fe, Ni, V, and Zn on rat lung epithelial cells. Toxicology. 2003;190:171–184.
Copyright © 2014 by the American College of Occupational and Environmental Medicine