Impact of cognitive reserve on the relationship of lead exposure and neurobehavioral performance.

BACKGROUND: Cognitive reserve (CR)--a construct studied in many neurologic disorders--refers to the maintenance of cognitive performance in spite of ongoing underlying brain pathology. OBJECTIVE: We hypothesized that a dose-effect relationship would exist between chronic occupational lead exposure and cognitive effects in workers with low CR but not in workers with high CR and identical lead exposure, and that level of CR would not influence the relationship between lead exposure and motor performance. METHODS: We stratified currently employed lead smelter workers by Wide Range Achievement Test-R for reading (WRAT), a recognized measure of CR, into loCR and hiCR groups. From these two groups we matched 56 pairs on working lifetime weighted blood lead (TWA). We performed a factor analysis on 14 neuropsychological outcome variables. Within each CR group regression analyses after adjusting for age, alcohol use, and depression scale score tested for dose-effect relationships between TWA and outcome variables. RESULTS: Both CR groups had comparable age, years employed, alcohol use, and current blood lead levels. Factor analysis provided three factors and five tests used in the regression analyses. Significant dose-effect relationships between TWA and cognitive tests present only in the loCR group included Attention Factor and Digit Symbol. Both CR groups demonstrated significant dose-effect relationships on the Motor Factor. CONCLUSION: This study found that cognitive reserve protects against the effect of chronic lead exposure on select measures of cognitive performance but not on motor performance.

Differential effects of lead exposure on components of verbal memory.

AIMS: To determine if verbal learning and memory requiring acquisition and retention of information is differentially affected by lead exposure. METHODS: The Rey Auditory Verbal Learning Test (RAVLT), a test of verbal learning and memory, was administered to 256 English speaking lead smelter workers who had a mean (SD) age of 41 (9.4) years and employment duration of 17 (8.1) years. Lead exposure variables, based on up to 25 years of prior blood lead data, included a mean (SD) current blood lead (PbB) of 28 (8.8) microg/dl, working lifetime time weighted average blood lead (TWA) of 39 (12.3) microg/dl, and working lifetime integrated blood lead index (IBL) of 728 (434.4) microg-y/dl. Associations of these chronic and recent lead exposure variables with measures from the RAVLT were modelled through multiple linear regressions after controlling for age and educational achievement. RESULTS: PbB was not associated with any of the RAVLT variables. However, TWA and IBL contributed significantly to the explanation of variance of measures of encoding/storage and retrieval but not to immediate memory span, attention, and learning. Grouping study participants by RAVLT performance according to three recognised clinical memory paradigms showed significantly higher TWA and IBL in the group with "generalised memory impairment" after adjusting for age and educational achievement. We examined recall mechanisms in each group by serial position in the word list and found stronger primacy (recall of words from the beginning of the list) in the "no impairment" and "retrieval difficulties" groups while the "generalised memory impairment" group had better performance on recency (recall of words from the end of the list). CONCLUSIONS: Lead exposure over years and not PbB interfered with the organisation and recall of previously learned verbal material. Chronic lead exposure affects encoding/storage and retrieval of verbal information.

The factor structure of the Profile of Mood States (POMS) and its relationship to occupational lead exposure.

The Profile of Mood States (POMS) is an instrument composed of six mood scales that were developed using clinical populations. This study examines the replicability of the POMS factor structure in lead-smelter workers and its association with demographic variables and lead exposure. Four hundred sixty-seven current and retired smelter workers completed the POMS, and a factor analysis was performed on the individual items. Two multiple regressions in the currently employed workers (n = 340) with a mean age of 40 years (range, 20-63 years) were used to compare the relationship of the resulting POMS factors with demographic characteristics and two measures of occupational lead exposure: current blood lead level (mean, 28 micrograms/dL; range, 4-62 micrograms/dL) and working-lifetime integrated blood lead level (mean, 711 micrograms-yr/dL; range, 1-1537 micrograms-yr/dL). Factor analysis produced one factor, labeled "general distress," that was composed mainly of items from five of the six POMS subscales (anger, confusion, depression, fatigue, and tension) and a second factor labeled "psychological adjustment." Demographic variables, including education, age, and alcohol use, were significantly related to the "general distress" factor score. After the effects of the demographic variables were accounted for, the "general distress" factor was significantly related to integrated blood lead level but not current blood lead level. Integrated blood lead level, a measure of cumulative lead exposure that included prior high exposure, was significantly related to the POMS "general distress" factor, while a current blood lead level that was relatively low was unrelated. As expected, POMS psychological adjustment was not related to lead exposure. Factor analysis did not support the use of six separate POMS subscales in this occupational population.

Curvilinear relationship between blood lead level and reaction time. Differential association with blood lead fractions derived from exogenous and endogenous sources.

Prior studies demonstrate an inconsistent relationship between occupational inorganic lead exposure and simple reaction time (SRT) performance. In this study, we administered a computerized SRT test to 78 currently employed lead smelter workers and then investigated the relationship between different measures of blood lead and components of SRT performance. The measures of blood lead included current blood lead (PbB) and mathematically derived blood lead fractions from the environment (PbB-env) and from bone (PbB-bn). Measures of SRT performance, obtained from 44 trials with interstimulus intervals (ISIs) ranging from 1 to 10 seconds, included median SRT (SRT-md), mean SRT for ISIs between 1 and 5 seconds (SRT-1-5), and mean SRT for ISIs between 6 and 10 seconds (SRT-6-10). Multiple regression analysis after adjustment for age and education revealed a curvilinear relationship between PbB and SRT-md. As PbB increased from 0 to 30 micrograms dl-1, SRT-md decreased, and only with PbB levels above 30 micrograms dl-1 did SRT-md increase. PbB terms accounted for 13.7% of the variance in this SRT measure (P < 0.01). The longer ISI variable, SRT-6-10, was found to be more strongly related to PbB, to have lesser variability across ISIs, and to be unrelated to age. Additional multiple regression analysis to examine the relationship between components of SRT and the PbB fractions, PbB-env and PbB-bn, showed only PbB-env to account for significant variance in SRT-md, (14.4%, P < 0.01), SRT-1-5 (9.7%, P < 0.03), and SRT-1-6 (15%, P < 0.01). We conclude that the relationship between PbB and SRT is U-shaped, that the SRT measure SRT-6-10 has properties that make it the preferred measure of SRT performance in future studies, and finally that only PbB-env, and not PbB-bn, is related to components of SRT.

Differential contribution of current and cumulative indices of lead dose to neuropsychological performance by age.

We examined measures of lead dose that reflect intensity of exposure and cumulative exposure for differential association with neuropsychological functional domains and with neuropsychological performance by age. Eighty active lead smelter workers assessed for verbal memory and visuomotoric skills had a mean (range) age of 44 (24-64) years, duration of employment of 20 (4-26) years, education of 8 (0-13) grades, current blood lead (B-Pb) of 26 microgPb dl(-1) (13-43), working lifetime weighted average blood lead (TWA) of 42 microgPb dl(-1) (17-57), working lifetime integrated blood lead (IBL) of 903 microgPb-yr dl(-1) (81-1,436) and bone lead (bone-Pb) of 41 microgPb (g bone mineral)(-1) (-12-90). Multiple linear regression after adjusting for age and education showed a significant amount of variance in verbal memory (Rey Auditory Verbal Learning Test and Verbal Paired Associates) explained only by measures of lead dose intensity, B-Pb, and TWA. Visuomotoric ability (Grooved Pegboard [GP] and Digit Symbol) had significant variance accounted for by measures of both lead dose intensity, TWA, and of cumulative lead dose, IBL, and bone-Pb. The relationship between bone-Pb and GP was curvilinear in older workers, with increasingly slower performance above a bone-Pb threshold of 20 microgPb (g bone mineral)(-1). Examination of age-related change in GP with exposure using IBL, a surrogate for cumulative lead dose in the brain, showed enhanced age-related change in older workers with high IBL. These findings suggest that the older nervous system provides a substrate more susceptible to the effects of chronic lead exposure.

Relation of cumulative exposure to inorganic lead and neuropsychological test performance.

OBJECTIVES: To determine if measures reflecting chronic occupational lead exposure are associated with performance on neuropsychological tests. METHOD: 467 Canadian male lead smelter workers (mean (SD) age 43.4 (11.00) years, education 9.8 (3.18) years, years of employment 17.7 (7.43), and current blood lead concentration (B-Pb) 27.5 (8.4) micrograms dl-1) were given a neuropsychological screening battery. Time weighted average (TWA) and time integrated blood levels (IBL) were developed from B-Pb records obtained through regular medical monitoring (mean (range) TWA 40.1 (4.0-66.4) micrograms dl-1, mean IBL 765.2 (0.6-1625.7) micrograms-y dl-1). 14 neuropsychological variables were included in three multivariate analyses of covariance, with each exposure variable as the grouping variable (high, medium, and low) and age, education, score on a measure of depressive symptoms, and self reported alcohol use as the covariates. Groups did not differ in history of neurological conditions. RESULTS: Neither the B-Pb, TWA, nor IBL was significant by multivariate analyses of covariance (MANCOVA). When years of employment, a suppressor variable, was included as a covariate, IBL exposure groups differed significantly on digit symbol, logical memory, Purdue dominant hand, and trails A and B. CONCLUSIONS: A dose-effect relation was found between cumulative exposure (IBL) and neuropsychological performance at a time when current B-Pb concentrations were low and showed no association with performance.

The impact of glutathione s-transferase M1 and cytochrome P450 1A1 genotypes on white-blood-cell polycyclic aromatic hydrocarbon-DNA adduct levels in humans.

Carcinogenic polycyclic aromatic hydrocarbons (PAHs) form DNA adducts via a complex metabolic activation pathway that includes cytochrome P450 (CYP) 1A1, whereas intermediate metabolites can be detoxified by conjugation through pathways including glutathione s-transferase M1 (GSTM1). PAH-DNA adducts can be measured in peripheral white blood cells (WBCs) and should reflect the net effect of competing activation and detoxification pathways and DNA repair as well as exposure. We have previously shown that WBC PAH-DNA adducts measured by an enzyme-linked immunosorbent assay (ELISA) were associated with recent, frequent consumption of charbroiled food among 47 nonsmoking wildland fire-fighters who provided two blood samples 8 wk apart. In the investigation reported here, which was performed in the same population, we measured the association between the GSTM1 null genotype, which results in loss of enzyme activity, and PAH-DNA adduct levels, hypothesizing that subjects with this genotype would have higher levels of DNA adducts because of their decreased ability to detoxify PAH metabolites. However, PAH-DNA adduct levels were nonsignificantly lower in subjects with the GSTM1 null genotype (n = 28) compared with other subjects (n = 19) (median 0.04 fmol/microgram DNA vs 0.07 fmol/microgram DNA, respectively, P = 0.45, Wilcoxon rank-sum test). Adduct levels were also lower in the nine subjects heterozygous or homozygous for the CYP1A1 exon 7 polymorphism (which codes for a valine rather than isoleucine and is thought to be associated with greater CYP1A1 activity) compared with the 38 wild-type subjects (P = 0.12). In the entire group, there was a positive association between consuming charbroiled food and PAH-DNA adduct formation (r = 0.24, P = 0.02, Spearman rank-order correlation). This association was weaker in the subgroup of subjects with the GSTM1 null genotype (r = 0.03, P = 0.84) and stronger among the remaining subjects (r = 0.57, P = 0.0002). These results suggest that the GSTM1 null genotype and CYP1A1 exon 7 polymorphism are not associated with increased susceptibility for PAH-DNA adduct formation in peripheral WBCs measured by ELISA in nonsmoking populations.

Relationship between bone lead and other indices of lead exposure in smelter workers.

The relationship between blood lead (B-Pb) indices and bone lead was determined in 81 smelter workers. Lead exposure had changed dramatically since the smelter opened 27 years ago; therefore, workers were stratified by years employed. Seventy workers, hired prior to 1979 (Group 1), had a mean (range) age of 46 years (35-64), mean years of employment 23 (16-27), mean years of education 8 (0-13), mean B-Pb 26 micrograms dl-1 (13-43), mean time-weighted average (TWA) B-Pb 44 micrograms dl-1 (23-57) and mean integrated blood lead index (IBL), 983 micrograms Pb year dl-1 (537-1437). Eleven workers, hired after 1978 (Group 2), had a mean age of 34 years (24-55), mean years of employment 11 (5-14), mean years of education 11 (8-12), mean B-Pb 26 micrograms dl-1 (8-13), mean TWA 33 micrograms dl-1 (17-44) and mean IBL 378 micrograms year dl-1 (81-555). Bone lead measured in the mid-tibia used K-X-ray fluorescence. Mean bone lead in Group 1 and Group 2 was 45 (-7-90) and 11 (-12-33) micrograms Pb (g bone mineral)-1, respectively. Linear regression analysis demonstrated a significant relationship between bone lead and B-Pb indices. B-Pb accounted for 10.8%, years employed 12.7%, TWA 31.4% and IBL 36% of the bone lead variance. Using a previously established relationship, the mean bone lead predicted a mean B-Pb content of approximately 8 micrograms dl-1 in Group 1 as compared to 2 micrograms dl-1 in Group 2. The mean B-Pb was 26 micrograms dl-1 in both groups despite differences in contribution from bone stores. Differential contribution of ambient air lead to B-Pb in the two groups of current workers with similar job titles may reflect use of personal protective equipment.

Comparison of neurobehavioral function in workers exposed to a mixture of organic and inorganic lead and in workers exposed to solvents.

The neurobehavioral effects of lead (organic and inorganic) and organic solvents were compared in 386 U.S. workers (52 reference, 190 lead, and 144 solvent workers). The association between neurobehavioral test performance and duration of exposure to lead or solvents was also examined and compared. The neurobehavioral test battery consisted of examiner and computer-administered neurobehavioral tests, a test of olfactory function, and questionnaires that assessed neuropsychiatric symptoms. Adjusted mean differences on the neurobehavioral test scores were estimated by comparing the exposed group to the referent group using linear regression and adjusting for premorbid intellectual ability, age, and race. Both lead and solvents were associated with diminished neurobehavioral performance in all neurobehavioral areas tested. Specifically, while lead and solvent exposure had the same magnitude of adverse effects on tests of manual dexterity, lead exposure was associated with greater adverse effects on memory and learning tests but with less adverse effects on executive/motor tests and on a test of olfaction than solvent exposure. An elevated number of neuropsychiatric symptoms was reported by 7% of the referent group, 43% of the lead group, and 15% of the solvent group. For exposure duration of < or = 10 years, more neurobehavioral decrements were found in the solvent group relative to the lead group. However, for exposure duration of > or = 18 years, the lead group showed more decrements than the solvent group. Overall, these data suggest differences in neurobehavioral functioning between the lead (organic and inorganic) and solvent exposed workers examined in this study.

Contribution of occupation and diet to white blood cell polycyclic aromatic hydrocarbon-DNA adducts in wildland firefighters.

Wildland (forest) firefighters are exposed to a wide range of carcinogenic polycyclic aromatic hydrocarbons (PAH) in forest fire smoke. PAH undergo metabolic activation and can subsequently bind to DNA. In this study, we investigated the association between occupational and dietary PAH exposures and the formation of WBC PAH-DNA adducts in a population of wildland firefighters. An enzyme-linked immunosorbent assay using an antiserum elicited against benzo(a)pyrene-modified DNA was used to measure PAH-DNA adducts in WBC obtained from 47 California firefighters at two time points, early and late in the 1988 forest fire season. PAH-DNA adduct levels were not associated with cumulative hours of recent firefighting activity. However, firefighters who consumed charbroiled food within the previous week had elevated PAH-DNA adduct levels, which were related to frequency of charbroiled food intake. These findings suggest that dietary sources of PAH contribute to PAH-DNA adduct levels in peripheral WBC and should be evaluated when using this assay to assess occupational and environmental PAH exposure.

Decrements in neurobehavioral performance associated with mixed exposure to organic and inorganic lead.

In 1990, 222 current employees of a chemical facility in the eastern United States that manufactured tetraethyl lead were administered a neurobehavioral test battery, tests of olfactory function and peripheral vibration threshold, and questionnaires that assessed neuropsychiatric symptoms. A cumulative variable of exposure to inorganic and organic lead was derived from 12 years of personal industrial hygiene sampling data and an occupational history interview that assessed work in each of 29 exposure zones in the lead area. The range of assigned exposure intensities in these 29 zones was 4-119 micrograms/m3 for organic lead and 1-56 micrograms/m3 for inorganic lead. Cumulative lead exposure and exposure duration were defined as categorical variables (four groups) in multiple linear regression models. The adjusted mean differences in neurobehavioral test scores were estimated by comparing the average scores of the moderate, high, and highest exposure groups with the low exposure (reference) group. After adjustment for premorbid intellectual ability, age, race, and alcohol consumption, neurobehavioral test scores were poorer as measures of both cumulative lead exposure and exposure duration increased; many of the associations evidenced dose-response relations. Associations were observed in most cognitive and functional domains tested, but were most common in two domains: manual dexterity and verbal memory/learning. On the affected neuropsychologic measures, the groups with the highest exposure averaged scores 5-22% lower than those of the reference groups. Overall, the data revealed generally consistent and coherent associations between two measures of lead exposure and poorer neurobehavioral test performance.

Analysis of OSHA inspection data with exposure monitoring and medical surveillance violations.

Occupational Safety and Health Administration (OSHA) inspection data from the Integrated Management Information System (IMIS) enforcement data base are presented for lead, ethylene oxide, and formaldehyde for fiscal years 1985, 1987, and 1989, and are discussed with emphasis on exposure monitoring or medical surveillance section violations. These data suggest that the exposure monitoring section of these standards is more commonly used to cite workplaces below these standards than is the medical surveillance section. Medical surveillance violations more commonly resulted in fines, but there were no differences in the magnitude of the fines for exposure monitoring or for medical surveillance violations. Implications of these findings are discussed.

Comparison of computerized and examiner-administered neurobehavioral testing techniques.

To better understand the relationship of newly developed computerized neurobehavioral tests to their examiner-administered counterparts, performance on four pairs of tests was compared among 185 solvent-exposed workers in the paint manufacturing industry. Two approaches to the comparison of the test methods (computer or examiner) were employed: (1) the multitrait-multimethod correlation matrix and (2) linear regression modeling of the difference scores between pairs of tests (same functional domain, different test formats). The data revealed that the validity criteria of the multitrait-multimethod technique were met for the tests of simple visual reaction time and attention-concentration (digit symbol substitution). Linear regression analyses demonstrated that the digit symbol substitution and visual memory test pairs were differentially affected by age, vocabulary, solvent exposure, or manual dexterity. Overall, this latter approach holds promise as a means of understanding test format similarities and differences.

Pulmonary function and respiratory symptoms in wildland firefighters.

We studied cross-seasonal changes in pulmonary function and respiratory symptoms in 52 wildland firefighters in Northern California. The mean cross-seasonal change in forced expiratory volume in 1 second (FEV1) was -1.2% (95% confidence interval [CI] -0.5%, -2.0%) with a corresponding mean change in forced expiratory volume (FVC) of -0.3% (95% CI 0.4%, -1.0%). Decreases in FEV1 and FVC were most strongly associated with hours of recent fire-fighting activity (P = .002 and .01, respectively). When the study group was divided into three categories based on recent fire-fighting activity, firefighters in the high activity category (mean +/- SE, 73 +/- 7 hours of fire-fighting in previous week) had a -2.9% (130 mL) change in FEV1 and a -1.9% (102 mL) change in forced vital capacity (FVC). There was a significant cross-seasonal increase in most respiratory symptoms evaluated. Several symptoms (eye irritation, nose irritation, and wheezing) were associated with recent fire-fighting. These findings suggest that wildland firefighters experience a small cross-seasonal decline in pulmonary function and an increase in several respiratory symptoms. Research is under way to identify the fire conditions and specific components of exposure that produce pulmonary irritants, and to examine the potential reversibility of acute pulmonary change.

Anti-tick saliva antibody: a biologic marker of tick exposure that is a risk factor for Lyme disease seropositivity.

Anti-tick saliva antibody (ATSA) has potential as a biologic marker of exposure to tick bites. In 1989, we conducted a cross-sectional study of 304 outdoor workers in Monmouth County, New Jersey, to evaluate associations between self-reported tick exposure, ATSA status, and Lyme disease antibody status. ATSA levels 1) were correlated with an index of tick exposure on the basis of three self-reported measures of tick exposure and outdoor hours worked per week (p = 0.01); 2) were consistently higher in pet owners compared with persons without pets (p = 0.03); and 3) when examined by duration since last tick bite, peaked at 3-5 weeks after tick bite and then declined (p = 0.06). ATSA levels dichotomized at the 75th percentile (approximately two standard deviations above the mean ATSA optical density of 25 subjects who denied recent tick exposure) were associated with self-reported tick exposure: adjusted odds ratios for high ATSA levels were 4.2 (95% confidence interval (CI) 0.9-18.9) for moderate (versus none) tick exposure and 5.8 (95% CI 1.2-27.2) for high (versus none) tick exposure. Finally, high ATSA levels were associated with Lyme disease seropositivity, with an adjusted odds ratio of 3.2 (95% CI 1.3-7.6). The data suggest that ATSA is a biologic marker of tick exposure that is a risk factor for Lyme disease seropositivity.

A quantitative approach to the characterization of cumulative and average solvent exposure in paint manufacturing plants.

Previous reports have attributed a range of neurobehavioral effects to low-level, occupational solvent exposure. These studies have generally been limited in their exposure assessments and have specifically lacked good estimates of exposure intensity. In the present study, the authors describe the development of two exposure variables that quantitatively integrate industrial hygiene sampling data with estimates of exposure duration--a cumulative exposure (CE) estimate and a lifetime weighted average exposure (LWAE) estimate. Detailed occupational histories were obtained from 187 workers at two paint manufacturing plants. Historic industrial hygiene sampling data for total hydrocarbons (a composite variable of the major neurotoxic solvents present) were grouped according to 20 uniform, temporally stable exposure zones, which had been defined during plant walk-through surveys. Sampling at the time of the study was used to characterize the few zones for which historic data were limited or unavailable. For each participant, the geometric mean total hydrocarbon level for each exposure zone worked in was multiplied by the duration of employment in that zone; the resulting products were summed over the working lifetime to create the CE variable. The CE variable was divided by the total duration of employment in solvent-exposed jobs to create the LWAE variable. The explanatory value of each participant's LWAE estimate in the regression of simple visual reaction time (a neurobehavioral test previously shown to be affected by chronic solvent exposure) on exposure was compared with that of several other exposure variables, including exposure duration and an exposure variable based on an ordinal ranking of the exposure zones.(ABSTRACT TRUNCATED AT 250 WORDS)

Solvent-associated olfactory dysfunction: not a predictor of deficits in learning and memory.

OBJECTIVE: The aim of the study was to evaluate associations between olfactory dysfunction and aberrations in learning and memory after chronic occupational exposure to mixed hydrocarbon solvents. METHOD: This was a cross-sectional, epidemiologic study of 187 paint manufacturing workers. The authors administered quantitative tests of olfactory function (the University of Pennsylvania Smell Identification Test) and neurobehavioral function (eight computer- and examiner-administered tests of learning and memory) to workers for whom detailed information was available on lifetime occupational exposure to solvents. RESULTS: Olfactory function test scores were positively correlated with performance on seven of eight of the tests of learning and memory in bivariate analyses. After adjustment for important confounding variables (i.e., age, vocabulary score, and cumulative exposure to hydrocarbon solvents) with multiple linear regression, olfactory function scores predicted performance only on the Wechsler Memory Scale Delayed Logical Memory test. CONCLUSIONS: Overall, the data did not reveal that olfactory dysfunction was correlated with decrements in various memory functions.

Dose-related subclinical neurobehavioral effects of chronic exposure to low levels of organic solvents.

One hundred eight-seven workers (mean age +/- SD; 42 +/- 8.8 years) in two paint manufacturing plants were examined to determine if sustained low-level exposure to mixed organic solvents resulted in the painters' syndrome (a psycho-organic syndrome). The test battery consisted of a medical and occupational questionnaire, the Present State Examination, the Zung Depression Scale, the Scandinavian Questionnaire 16, a neuropsychological battery, and vibration thresholds. Solvent exposure, expressed as total hydrocarbon of combined selected solvents, was quantitated using 13-15 years of personal breathing zone samplings. Linear regression analysis controlling for several confounding variables demonstrated significant correlations between increasing exposure to mixed organic solvents and neurobehavioral performance for vibration threshold and several neuropsychological tests. Dose-related effects of chronic solvent exposure on neurobehavioral outcomes (all subclinical) were shown, but "typical" symptoms characteristic of the painter's syndrome were not found.

Solvent-associated decrements in olfactory function in paint manufacturing workers.

To assess the effects of low-level organic solvent exposure on olfactory function, a cross-sectional study in paint manufacturing workers was undertaken. Workers in two paint manufacturing facilities (N = 187) were tested using the University of Pennsylvania Smell Identification Test (UPSIT), a standardized, quantitative test of olfactory function. Industrial hygiene air samples over the past 13-15 years revealed that average solvent exposures in these plants were 2-40% of the existing threshold limit values for the three chemicals measured. Stratification by smoking status revealed evidence of dose-related decrements in olfactory function (p = .01) only in non-smokers. Furthermore, those nonsmoking workers in the highest exposure category had UPSIT scores below the fifth percentile for their age. These results suggest that solvents may cause nervous system dysfunction at lower levels than previously suspected, and that the olfactory system may be a critical target organ for the neurotoxic effects of solvents and other chemicals.