A Meta-Analysis of the Neuropsychological Effects of Occupational Exposure to Mercury Martin L. Rohling, Ph.D. University of South Alabama George J. Demakis, Ph.D. University of North Carolina – Charlotte

Abstract This paper reports a meta-analysis of 36 peer-reviewed published studies of the neuropsychological effects of occupational exposure to mercury, which yielded 43 independent samples. These studies included 2,512 exposed participants and 1,846 controls, for a total sample size of 4,358. Because the independent variables defining mercury exposure varied across studies, effect sizes were calculated for exposed versus non-exposed workers. Dose-response relations were considered for measures of mercury levels in urine (81% of studies reported), blood (42% of studies reported), and air samples (33% of studies reported). Level of exposure was also estimated by reported years of exposure (M = 11.3, SD = 5.6). Cohen’s d statistic yielded a statistically significant weighted study-mean effect size of -.23, p < .0001 for occupational mercury exposure. However, an effect this small is typically undetectable when evaluating individuals because it is smaller than the typical 95% confidence interval used for most neuropsychological measures.

None of the exposure variables analyzed reached statistical significance. The magnitude of self-reported symptoms (-.30) was slightly larger than that obtained from objective test scores (-.22), though the difference was not statistically significant. Also, the weighted mean effect size for psychomotor skills (-.34) was the largest in magnitude, whereas the weighted mean effect size for verbal comprehension measures had the smallest (-.06). However, an analysis of the differential effects of mercury exposure across cognitive domains found significant differences between verbal comprehension measures and all other domains. None of the other domains were significantly different from one another. The weighted study-mean effect size suggests that the prevalence of neuropsychological deficits due to occupational exposure to mercury is small and difficult to detect on an individual case-by-case basis.

Introduction Much of what we know regarding the level of mercury toxicity that is fatal comes from studying cases of acute community poisonings. For example, in 1956 in the Japanese community of Minamata, acute and fatal exposure to methylmercury was first diagnosed in near a polyvinyl chloride plastic plant that was discharging untreated effluent containing methyl mercury chloride into a nearby ocean bay. Once in the sediments of the sea, marine species absorbed the mercury, which contaminated of the ecosystem. Residents often consumed seafood from the contaminated waters. Some 39 years later, 2,252 individual had been officially diagnosed with Minamata disease, and over 1,000 had died from mercury toxicity.

Researchers have found that, for both humans and animals, the body initially retains about 80% of inhaled elemental mercury vapor. However, the gastrointestinal tract poorly absorbs liquid metallic mercury (less than 1%), whereas the respiratory tract’s rate of absorption is of elemental mercury is inversely proportional to particulate size. For orally ingested inorganic mercury, the gastrointestinal tract absorbs less than 10% on average, but considerable individual differences exist here too (WHO, 1991). The kidneys are the main depository for both elemental and inorganic mercury after inhalation (50-90% of the body burden). More mercury is transported to the brains of mice and monkeys after the inhalation of elemental mercury than after the intravenous injection of equivalent doses of the mercuric form (Kedziora & Duflou, 1995).

Purpose Our goal was to meta-analyze studies that included neuropsychological assessment, thereby providing data on the cognitive, psychomotor, sensory-perceptual, and psychological functioning of mercury exposed individuals. These domains were chosen because they have not been evaluated and compared in the previous research syntheses and because they are likely to be of particular interest to neuropsychologists. Studies of both currently exposed individuals and previously exposed individuals were included in this analysis to determine the degree to which mercury exposure results in long-term neuropsychological effects.

Coding Procedures We calculated effect sizes using Cohen’s d statistic. For analyses, each effect size was weighted by its standard error, a value that represents the precision of the effect size. All effect sizes were adjusted for sampling bias. Statistics and moderator analyses were performed on these weighted mean effects sizes. Finally, a 95% confidence interval (CI) was computed around each effect size mean to provide an estimate of the variability of d and to test whether d was statistically different from zero. The confidence interval was constructed by multiplying the standard error by the critical z value, which is a p value of .05 (1.96). Two coders (MLR and GJD) independently coded a non-redundant set of studies. These codes were then blindly double checked by the alternate coder. All disagreements were then resolved.

Table 1 - Demographic & exposure info on participants Exposed Participants Non-exposed Participants t or X2 M SD n M SD n p Mn N 58.4 82.5 43 51.3 51.8 43 .4341 Mn Age 41.2 8.7 37 41.5 8.4 31 .7924 Mean Ed 13.1 4.5 19 12.8 4.3 18 .1687 % Male 71.6 38.9 32 74.1 34.3 26 .9183 Mn Exp (yr) 11.3 5.6 30 --- --- --- --- Last Exp (yr) 2.6 7.0 34 --- --- --- --- Air (mg/M3) 52.9 28.0 14 .00 .00 3 < .0001 Blood (mg/L) 11.3 12.4 18 3.0 1.8 11 .0192 Urine (mg/L) 43.2 52.2 35 3.4 2.7 22 < .0001

Table 3 - Wt ESs (M) for domains of functioning, including NP domains DV-Type d SE n p Z p Q Psychomtr -.34 .08 38 .0001 .9048 Psych -.33 .10 21 .0004 .4516 Cognitive -.25 .09 28 .0026 .9997 Sens/Perc -.04 .14 10 .3859 .9998 Self-Rep -.30 .04 22 .0001 .0001 Obj Test -.22 .03 42 .0001 .0151 Executive -.66 .20 2 .0005 .0120 Visuosp -.30 .06 12 .0001 .0136 Learn/Mem -.24 .05 21 .0001 .0001 Attention -.24 .04 27 .0001 .0456 Process Spd -.22 .04 19 .0001 .0001 Verb Comp -.06 .06 11 .1587 .0006

Results The wt study-mn ES for Hg exposure was -.23 (SE = .03. Weighting by the inverse of the estimated variance caused the overall study-mn ES to decrease from -.33 to -.23. This suggests that the larger and more reliable studies generated ES that tended to be smaller than the less reliable studies. Despite the shrinkage, the overall study-mn ES for NP outcome was still found to be significantly different from zero, z = 7.22, p < .0001. When the ES was transformed into a point-biserial correlation coefficient (Rosenthal’s r) between exposed and control participants the resulting correlation was -.11. This indicates that 1.2% of the variance in ES can be accounted for by mercury exposure.

Hg in Air. There were no significant correlation between outcome and measures of mercury in air (Rho = .53, p = .07). Hg in Blood. There was no significant association between these two variables (Rho = -.29, p = .3040). Hg in Urine. There was no significant association between these two variables (Rho = -.21, p = .2594). Type of Hg. The wt study-mn ESs did not significantly differ from one another (-.28 vs. -.18, respectively). ESs by Moderator. None of the categories were significantly different from one another. Neuropsych Domains. There were significant wt study-mn ESs for the domains of executive functioning (-.66), visuospatial processing (-.30), learning and memory (-.24), attention (-.24), and processing speed (-.22). Objective Test vs. Self-Reported. The wt study-mn ES for objective testing was -.22 (SE = .03), which was not significantly different from the self-reported symptoms mean of -.30 (SE = .04).

Discussion When examining the influence of mercury on NP functioning defined dichotomously (exposed vs. unexposed), we found an ES indicative of mild impairment (d = -.23). This wt ES was sign. different from 0 and homogeneous The total amount of variance that mercury exposure appears to account for in the effect size distribution is relatively small (1.2%). Dose-Response Analyses None of the time measures revealed a pattern of impairment related to exposure in the expected direction. When examining air, blood, and urine we were able to replicate the ratios that other agencies have reported (WHO, 1991), as well as those generated by other researchers (Meyer-Baron et al., 2002). However, relationships between these objective measures of mercury exposure and outcome with nonparametric regression analyses were not significant.