APHA November 4 8 2006 Boston MA Neurological

APHA - November 4 -8, 2006 - Boston MA Neurological Risk Assessment: The Benchmark Dose Method and Deficits in Bridge Welders Robert M. Park, NIOSH*, Cincinnati OH Rosemarie Bowler, San Francisco State University Donald E. Eggerth, NIOSH* * Emily Diamond, The Wright Institute, Berkeley CA Katie J. Spencer, Miami University, Oxford OH Donald Smith and Robert Gwiazda UC, Santa Cruz CA Education and Information Division, Risk Evaluation Branch*/ Training Research and Evaluation Branch** , NIOSH, Cincinnati Disclaimer: The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the National Institute for Occupational Safety and Health Occupational risk assessment typically estimates excess life-time risk for onset of a chronic disease or fatal health outcome in relation to a fixed life-time workplace exposure. Excess lifetime risks are often considered acceptable if they are < 1/1000. For sub-clinical health effects such as respiratory or neurobehavioral changes measured as continuous variables, the Benchmark Dose (BMD) method examines exposure-induced shifts in outcome distributions (Fig. 1). But this poses some issues: 1) one must define a threshold for impairment (often the 5 th percentile of performance in an unexposed population) and then calculate at what exposure there is a 1, 5, or 10% excess impairment above that threshold. ; and 2) one must specify what increases in impairment are acceptable. Individuals falling below the 5 th percentile in a normal population may have other identifiable risk factors that themselves merit prevention; and allowing an increase of 5% impairment (above a background 5%) is equivalent to permitting a relative risk of 2. 0 and 50/1000 excess risk. Figure 1 16%ile (1 SD) in normals 10%ile in normals 5%ile in normals 1%ile in normals Performance in unexposed population ~N(100, 10) Performance in exposed population High manganese exposures in populations such as ferroalloy production have been associated with symptoms of parkinsonism and manganism, and signs and symptoms consistent with parkinsonism have been reported in some welder populations. Neurobehavioral assessments using a battery of tests were conducted for 48 California bridge welders exposed to manganese in confined spaces for up to two years. Exposures ranged 0. 1 -0. 4 mg/m 3. Results for four tests (Rey-Osterrieth Complex Figure Test, Working Memory Index, Stroop Color Word Test and Auditory Consonant Trigrams Test) were used in the benchmark dose procedure and reported previously (1, 2). That analysis, which used minimal work history, and environmental data covering only 17 months of the 2 yr period, observed excess impairment defined as performance below the 5 th percentile of normal, at the current Cal OSHA standard (0. 2 mg/m 3, 8 hr TWA), ranging 15 - 32%. These results suggested unacceptable risks of a possible parkinsonian syndrome at the current OSHA PEL for manganese (5. 0 mg/m 3, 15 min) and likely at the Cal OSHA PEL as well. RESULTS Linear regression models were used to predict managanese blood levels and test performance scores in relation to duration of Mn exposure and cumulative Mn exposure. Manganese blood levels increased with adjusted exposure duration [Mn(B)=6. 9800+0. 1507 adj. Dur, Rsq=0. 137, p=0. 0186] (Fig. 2). The association with cum. Mn was only very slightly stronger [Mn(B)= 7. 701+ 0. 7178 cum. Mn, Rsq=0. 139, p=0. 0177]. Performance on test ACT 3, a test of short-term auditory memory and divided attention (normal=0. 0), was strongly associated with adjusted duration [ACT 3=1. 050 -0. 1319 adj. Dur, Rsq=0. 26, p=0. 0002 one-tailed] (Fig. 3) and considerably less strongly with cumulative Mn [ACT 3=-. 04037 -0. 3197 cum. Mn, Rsq=0. 07, p=0. 04 one-tailed ]. Similarly, performance on the REYI, a test of shortterm memory and visuomotor coordination (normal=50), was also most associated with adj. exposure duration [REYI=50. 77 -0. 730 adj. Dur, Rsq=0. 11, p=0. 015 onetailed] (Fig. 4) and less strongly with an exposure metric. In the best fitting model cum Mn was calculated with a 30 day half-life [REYI=41. 35 -35. 42 cum. Mn(30), Rsq=0. 07, p=0. 045 one-tailed] (Fig. 5). In this model, the lowered estimate of baseline performance (intercept: 41. 4 vs 50. 8) could represent exposure misclassification or possibly residual performance deficit in workers whose only exposure was several months prior to neurobehavioral assessment. A similar pattern was observed for WMI, a test of memory, concentration and attention, with a somewhat stronger association with adj. exposure duration [WMI=102. 5 -0. 711 adj. Dur, Rsq=0. 12, p=0. 013 one-tailed] (Fig. 6) than with cumulative exposure (also with 30 day half-life) [WMI=93. 79 -41. 422 cum. Mn(30), Rsq=0. 10, p=0. 023 one-tailed]. Using the results of regression models, and deriving predicted distributions of performance as a function of exposure duration or cumulative exposure, the proportions of exposed populations falling below variously defined impairment thresholds were calculated. Benchmark doses – exposure levels sustained over a two year period – at which specified excess risk of impairment would be predicted are shown in Table 1. For a 5% excess risk of impairment above a threshold defined at the lower 5%ile in an unexposed population, the benchmark Mn concentrations ranged from 0. 030 mg/m 3 (REYI and cum. Mn) to 0. 092 mg/m 3 (WMI and Adj Duration). Defining impairment at a higher threshold of poor performance, that is including less severe deficits (e. g. , at 10 th %ile instead of 5 th %ile), produced lower benchmark doses. The predicted WMI distributions for various exposure levels are shown in Figure 7. Table 1 Excess Risk (%) ACT and adj. Exposure Duration METHODS 1 2. Park, R. M. , Bowler, R. M , Eggerth, D. E. , Diamond, E, Spence, r K. J. , Smith, D. , Gwiazda, R. “Issues in Neurological Risk Assessment for Occupational Exposures: The Bay Bridge Welders. ” Neuro. Toxicology 2006; 27: 373 -384. Bowler RM, Roels HA, Nakagawa S, Drezgic M, Diamond E, Park R, Koller W, Russell P Bowler RP, Mergler D, Bouchard M, Smith D, Gwiazda R, Doty RL. Dose-Effect Relations Between Manganese Exposure And Neurological, Neuropsychological And Pulmonary Function In Confined Space Bridge Welders. Occup Environ Med, accepted for publication Oct. 3, 2006. Blood Mn (ug/l) and adjusted welding duration (mos) Figure 3 ACT 3 Performance and adjusted welding duration (normal ~ N(0, 1) REYI and adj. Exposure Duration REYI and cum. Mn Exposure (30 d halflife) WMI and adj. Exposure Duration WMI and cum. Mn Exposure (30 d halflife) Figure 4 REYI Test and adjusted welding duration (normal ~ N(50, 10) ) Figure 6 WMI Test and adjusted welding duration (normal ~ N(100, 15) DISCUSSION The choice of what constitutes impairment requires clinical judgement for which there is ample precedent from inverstigations involving lead and other neurotoxicants. The issue of reversibility could not be addressed in the current study although follow-up of this population may provide some insights. The observation that duration of exposure was as good or better than cumualtive exposure as a predictor of neurobehavioral effects implies that considerable exposure misclassification may be present or that the exposure metrics, such as cumulative exposure, are not optimum. Figure 5 REYI Test and cumulative Mn (30 day half-life) CONCLUSION Neurobehavioral effects are being observed at far below the current OSHA standard for Mn (5. 0 mg/m 3) and probably below the Cal OSHA standard (0. 2 mg/m 3). Studies are needed to better characterize Mn exposure history and better decribe both the short- and long-term consequences of exposure to Mn fume. Figure 7 Benchmark Dose calculations (Mn, mg/m 3) ACT and cum. Mn Exposure For this analysis, a more detailed history of each worker’s activities over time was available (automatic vs. manual welding; location: in bridge piling or in associated chambers) as was environmental sampling data from the entire 2 yr period. The BMD procedure is reported for three outcomes: Auditory Consonant Trigrams Test (3 seconds), ACT 3, Rey-Osterrieth Complex Figure Test (immediate copy), REYI, and Working Memory Index, WMI. In every case, a lower test score represented poorer performance. Test performance was analyzed in relation to duration of employment on this bridge project, and in relation to duration actually exposed to welding with adjustment for average length of working day and days worked per week. The exposure metric used was cumulative exposure to total airborne manganese (Mn mg/m 3) as well as an alternate metric that was calculated assuming depletion of the manganese effect with various half-lives following exposure. For models based on durations, a global air concentration of 0. 20 mg/m 3 was assumed (the current Cal OSHA standard and a rough estimate of average exposures in this population). Figure 2 Distributions for 2 yrs at air concentrations: Mn=0. 0, 0. 02, 0. 05, 0. 10, 0. 15 mg/m 3 Impairment Threshold Assumed (%) 0. 15 1 5 10 1 0. 025 0. 012 0. 008 2 0. 042 0. 022 0. 015 5 0. 072 0. 042 0. 032 10 0. 102 0. 069 0. 055 1 0. 010 0. 007 0. 006 2 0. 020 0. 014 0. 012 5 0. 048 0. 033 0. 030 10 0. 086 0. 064 0. 058 1 0. 015 0. 008 0. 005 2 0. 032 0. 015 0. 012 Bay Bridge Project 5 0. 065 0. 039 0. 028 source: www. newbaybridge. org 10 0. 105 0. 069 0. 055 1 0. 009 0. 006 2 0. 018 0. 012 0. 011 5 0. 042 0. 030 0. 028 10 0. 078 0. 059 0. 055 1 0. 095 0. 028 0. 015 2 0. 132 0. 049 0. 028 5 0. 192 0. 059 10 0. 246 0. 139 0. 095 1 0. 039 0. 011 0. 007 2 0. 065 0. 022 0. 013 5 0. 118 0. 048 0. 032 10 >0. 250 0. 085 0. 060 0. 0 5 th %ile of normal San Francsico Normal population ~ N(100, 15)