Neurotoxicology and Teratology 23 (2001) 473 480

Color vision and occupational toluene exposure$

ne dicte Stengelc, Donna Merglerd, Jean Claude Limassete, Dave Campagnaa,b, Be Franc ois Diebolde, Denis Michardf, Guy Huelb,*
Health Investigations Branch, Division of Health Studies, Agency for Toxic Substances and Disease Registry (ATSDR), Atlanta, GA, USA b pidemiologie et en Biostatistique, et de la Recherche Me dicale (INSERM-U472), Recherche en E Institut National de la Sante 16 av. P.V. Couturier, 94807 Villejuif Cedex, France c pide et de la Recherche Me dicale (INSERM-U170) Recherches E miologiques sur lEnvironnement et la Sante , Institut National de la Sante 94807 Villejuif Cedex, France d tude des Interactions Biologiques entre la Sante et lEnvironnement (CINBIOSE), Universite du Que bec a ` Montre al, Centre de Montreal, Canada H3C 3P8 e curite (INRS), 54501 Vandoeuvre les Nancy, France Institut National de Recherche et de Se f dical du Travail du Livre, 75006 Paris, France Service Me Received 20 November 2000; accepted 29 May 2001
a

Abstract We examined the relationship between acquired color vision loss and exposure to toluene and total hydrocarbons among 125 male workers. Seventy-two toluene-exposed printers were compared with 34 workers from the same photogravure plant with ambient background exposure, and with 19 workers from a bookbinding plant located in the same town (nonexposed). Environmental mean toluene exposure level at workstation was estimated from individual 8-h sampling. Historic exposure data from the last 30 years were used to construct two cumulative exposure indices, one for toluene and one for total hydrocarbons. Airborne toluene levels were overall lower than the current Threshold Limit Value (TLV) of 50 ppm. Color vision was assessed by the Lanthony D-15 desaturated panel. Color vision loss was quantitatively established by the Color Confusion Index (CCI) and classified by type of acquired dyschromatopsia according to Verriests classification. CCI was positively related to current airborne toluene levels, and cumulative exposure indices for toluene and total hydrocarbons (.18 r .35). Odds ratios of acquired dyschromatopsia were significant for current airborne toluene, toluene, and total hydrocarbon past exposure (1.27 [1.02 1.58], 1.21 [1.04 1.39], 1.15 [1.02 1.31], respectively). In conclusion, this study suggests that the Lanthony D-15 desaturated panel detects early neurotoxic effects among workers exposed to toluene. D 2001 Elsevier Science Inc. All rights reserved.
Keywords: Toluene; Hydrocarbons; Printing industry; Occupational exposure; Color perception; Human

1. Introduction Toluene (methylbenzene; Chemical Abstract Service No. 108-88-3) is a hydrocarbon and a solvent that is rapidly absorbed through the respiratory and gastrointestinal tracts and, to a lesser extent, through the skin. Toluene is excreted in the exhaled air in its unchanged form, and in the urine as the metabolite hippuric acid [3]. Following exposure to toluene, the primary human health concern is a dysfunction of the central nervous system. At
$ Use of trade names and commercial sources do not indicate endorsement by any of the participating agencies. * Corresponding author. Tel.: +33-1-45-59-50-74; fax: +33-1-45-5951-69. E-mail address: huel@vjf.inserm.fr (G. Huel).

low to moderate exposure in occupational settings, impaired cognitive, auditory, and neuromuscular functions have been observed [2,3]. In animal studies, toluene was responsible for subtle behavioral effects, ototoxicity, alterations in brain neurochemistry, and electrophysiological changes [2,3]. Exposure to toluene [10,12,31 33], as well as to other chemicals (see Refs. [11,13] for reviews), has also been associated with color vision loss. Impairments affected the macula or the lens in one of the eyes, unequally [15], and may depend on age [5,19,20,31] and alcohol consumption [5,21,28]. Visual function testing provides important information on early neurotoxic effects in general. We examined the relationship between occupational airborne toluene, cumulative indices for toluene and total hydrocarbons, and acquired color vision loss among three groups of male workers who were subjected to direct or

0892-0362/01/$ see front matter D 2001 Elsevier Science Inc. All rights reserved. PII: S 0 8 9 2 - 0 3 6 2 ( 0 1 ) 0 0 1 6 3 - 5

474

D. Campagna et al. / Neurotoxicology and Teratology 23 (2001) 473480

background exposure in a photogravure printing plant or were nonexposed and worked in a bookbinding plant.

2. Methods 2.1. Participants A longitudinal study of exposure to toluene was conducted in a photogravure plant in France between January 1991 and May 1993 [26]. The study included 166 male workers among whom 92 were exposed to toluene and 74 were controls. The 92 exposed men worked with rotary printers, and were operators, handlers, winders, retouchers, or mechanics. The 74 control men were workers from two different subgroups. One group was selected from the same plant (n = 43) and included workers who were not directly exposed to toluene but were subjected to ambient background indirect exposure (engravers, forklift operators, compositors, and other unexposed personnel). The second group (n = 31) included participants who worked in a bookbinding plant in the same town and were not exposed to toluene. Both plants employed the same occupational physician. Three groups (exposed, with ambient exposure, and nonexposed) were thus considered for the purpose of this analysis. Color vision testing was performed in 1993, 2 weeks after the beginning of the second phase of the longitudinal study, and was thus not performed on the first 18 workers of the study due to logistical reasons. Additionally, 23 participants were excluded for conditions that could affect their color vision: six had diabetes, three took potentially ophthalmotoxic medication, six presented congenital colorblindness, and four had poor near visual acuity (visual resolution at 0.5 m superior to 1 min of arc or equivalent to 20/25 at near). One participant whose hobby exposed him to toluene was excluded. Three participants with unknown duration of employment were also excluded, leaving 125 participants for data analysis: 72 in the exposed group, 34 in the group with ambient exposure, and 19 in the nonexposed group. None of the participants had important ocular injuries. Among the participants included in this analysis, all worked for at least 1 year in the plant on 8-h rotating shifts, with successive week-day, -evening, and -night shifts. Current alcohol consumption was assessed by questionnaire in grams per day of beer, wine, and other alcoholic beverage and was classified as 0, 1 49, or ! 50 g/day. 2.2. Exposure assessment 2.2.1. Current exposure assessment Eight-hour personal air sampling in the breathing zone was performed during a workweek, once in March 1991 and once in June 1992. Air sampling was performed at all workstations, during all shifts, and for almost all participants who worked at the printing plant, whether they were directly

exposed or with an ambient exposure. Air samples were collected on charcoal tubes (SKC 226-01) with constantflow (50 ml/min) portable pumps over an 8-h shift. After desorption with carbon disulfide, the toluene content was analyzed by gas chromatography with flame ionization detection. Overall, 92% of the participants exposed had at least one air (8 h long) sample measurement, and 51% had two measurements. Among participants with an ambient exposure, 60% had at least one measurement, and 44% had two measurements. Because no change in work conditions occurred between March 1991 and June 1992, these measurements were considered to reflect the mean exposure levels during the study, and no significant difference was observed between 1991 and 1992 levels (data not shown). The current toluene exposure level was defined as the mean of the measurements for participants with two samples or the unique value for the others. Participants from the printing press without air sampling were attributed the mean value of all participants who worked at the same workstation. The exposure level of the external control participants from the bookbinding plant was assumed to be zero, and no exposure measurement was performed. Toluene exposure levels according to work assignment in this study have been published elsewhere [26]. 2.2.2. Assessment of past exposure to toluene and to total hydrocarbons The companys records allowed us to establish the history of their printing techniques. Letterpress printing ended in 1970, typesetting in 1979, and offset printing in 1980. Photogravure techniques, especially the composition of the ink-diluting solvents, also changed over time. Benzene disappeared from the formulation in 1952. Until 1968, the diluting solvent was a mixture of spirits composed of aromatic solvents (naphta petroleum distillates) containing toluene and xylene. Starting in 1968, the toluene content of ink solvents increased and reached 98% in 1982. The company had also used several generations of presses. The following two dates marked important modifications in the work conditions of the photogravure plant. In 1974, systems for solvent recovery and air exchange were installed on some machines. In 1982, giant printing presses went into operation; some were enclosed and located in a small building, others were not enclosed and were located in large workshops. Exposure to toluene decreased considerably from an average of 134 mg/m3 in the early sixties to 89 mg/m3 in the early nineties. Total hydrocarbon exposure decreased from 905 mg/m3 in the early sixties to 210 mg/m3 in the late seventies, before toluene became the main solvent used in 1983 [26]. Since 1960, regular area samples of both toluene and total hydrocarbon concentrations have been collected at fixed points near each press, twice a year. Sampling lasted on average 3 h. The samples were collected with charcoal tubes and analyzed with gas chromatography, a method very similar to that used for personal sampling in this study.

D. Campagna et al. / Neurotoxicology and Teratology 23 (2001) 473480

475

These measurements allowed us to calculate past hydrocarbon exposure level among the printers who worked on different generations of machines, and thus, to develop a job-exposure matrix. The detailed work history of each participant within the plant was recorded on a self-administered questionnaire. Combining data from the work histories and the matrix allowed us to assign to each printer the mean of the measurements for this machine for a given year. No past area measurements were available for jobs that were not associated with a specific machine. Thus, there were no measurements for mechanics and engravers, for example, whether or not they were exposed. To estimate past exposure for these jobs, we used the data from the two sets of personal measurements taken in 1991 and 1992, and calculated the relation between these exposure levels and that of the printers for each job category. Based on the assumption that this ratio remained stable over time, we assessed past exposure for each category, using the ratio with the mean of the measurements available for each year for the printers overall. Based on these estimated annual exposures, a cumulative exposure index (mg/m3 years) was calculated for all photogravure plant participants. 2.3. Visual function assessment The testing of visual function was performed within the first 3 h of a day work shift. Color vision and near visual acuity were examined monocularly in a darkened room under standardized illumination provided by a daylight 1150-lx fluorescent lamp positioned 0.3 m above the work plane. No time limit was imposed, and participants with a prescription for eyeglasses wore their own nontinted glasses or contact lenses. Workers were screened for adequate near visual acuity. Participants were required to have a minimal resolution of 1 min of arc at 0.5 m in each eye with the use of the National Optical Visual Chart placed at a distance of 0.4 m to be included in the analysis of this study. Color vision was assessed with the Lanthony D-15 hue desaturated panel (Luneau Ophtalmique, Paris, France). This test is based on the ability to recombine a set of 15 desaturated color caps according to a definite chromatic sequence. 2.3.1. Visual function quantitative evaluation A quantitative evaluation was performed by calculating the sum of the color differences of the caps adjacent to one another and the total color distance score (the sum of distances in color in color space between adjacent caps) using the formula developed by Bowman [6]. The Color Confusion Index (CCI) was calculated by dividing the participants total color distance score by a perfect score. The value 1 indicates a perfect score while higher values indicate color vision loss. Color vision data are presented as the mean CCI of the two eyes, the lowest CCI value being the best score (best eye), and the highest CCI the worst score (worst eye).

2.3.2. Visual function qualitative evaluation Color vision loss was classified within types of acquired dyschromatopsia based on Verriests classification: type III, a loss in the blue yellow range; type II, a combined loss in the blue yellow and red green ranges; and type I, a loss in the red green range [29] (see Ref. [11] for information concerning the classification of dyschromatopsia types). A person was classified into one of the categories if this person misplaced the caps by at least two caps. One cap inversion was classified as normal. Participants were classified as dyschromate (for qualitative analyses) if one of the three types of acquired loss was identified for one or both eyes. Participants were classified as nondyschromate if they had a normal color vision or a nonspecific type of dyschromatopsia. A loss was considered nonspecific when the observed error did not correspond to a specific type of acquired loss. 2.4. Statistical analysis Alcohol consumption, current toluene level at workstation, past toluene, and past total hydrocarbon indices were log-transformed due to skewed distributions. Paired Students t tests were used to examine the difference between the best and the worst eye. The relationship between exposure measurements, participant characteristics and CCI were assessed with Spearman correlation coefficients and were adjusted for age and alcohol consumption. Qualitative analysis of dyschromatopsia was performed using logistic multiple regression models with the dependent variable being dyschromatopsia. A comparison of mean values between worker groups was performed by t test and by the nonparametric Wilcoxon rank test. Statistical analyses were performed with the SAS statistical package (Version 6.12, SAS Institute, Cary, NC, USA). Two mathematical methods were used to assess the existence of threshold exposure levels. The first method was based on the absence of expected effect while the level of exposure increases (safety level), followed by an increasing impairment when the level of exposure continues to increase after the hypothetical horizontal line start to rise [7]. The second method was an interval regression threshold assessment model assuming that two different regression lines exist, one for participants with an exposure level lower than a fixed threshold value, and the other one for participants with an exposure level higher than the fixed threshold value [8].

3. Results 3.1. Population characteristics The average age for the whole population ranging from 21 to 55 years, and average consumption was 30 g/day, ranging from 0 Average duration of employment at the plant was 40 years, daily alcohol to 168 g/day. was 17 years,

476

D. Campagna et al. / Neurotoxicology and Teratology 23 (2001) 473480

Table 1 Participant characteristics, CCI, and dyschromatopsia by exposure groups Exposed (n = 72) Characteristics Age (years)y Alcohol consumption (g/day)z 0 1 49 ! 50 Duration of employment (years)y Color vision CCIz Mean of both eyes Worst eye Best eye Dyschromatopsia Normal Nonspecific type Specific type (Type I, II, or III) Type I Type II Type III 40 [22 55] 10 [0 83] (24) 34% (34) 48% (13) 18% 18 [1 36] Ambient exposure (n = 34) 43 [25 54]a 14 [0 117] (8) 24% (21) 62% (5) 14% 19 [2 37] Nonexposed (n = 19) 37 [24 55]a 17 [0 117] (4) 21% (9) 47% (6) 32% 8 [1 35]
0

1.23 [1.00 1.81]c 1.28 [1.00 1.92]c 1.16 [1.00 1.70]c (26) 36% (9) 13% (37) 52%d (2) 3% (4) 6% (31) 43%

1.19 [1.00 1.72]b 1.23 [1.00 1.84]a 1.15 [1.00 1.72]b (13) 38% (2) 6% (19) 56%d (3) 9% (1) 3% (15) 44%

1.08 [1.00 1.36]b ,c 0 0 1.11 [1.00 1.41]a ,c b0,c0 1.05 [1.00 1.31] (13) 68% (2) 11% 0 (4) 21%d (0) 0% (0) 0% (4) 21%

0 0

Numbers in square brackets indicate the 5th and 95th percentiles, respectively. Numbers in parentheses indicate the number of measurements. Wilcoxon rank test: a greater than a0 ( P < .05); b greater than b0 ( P < .01); c greater than c0 ( P < .001). Chi-square: d greater than d0 ( P < .05). y Arithmetic mean. z Geometric mean.

ranging from 1 to 40 years. Participants with an ambient exposure were older than nonexposed participants, but the two groups were similar in age to the exposed one (Table 1). Daily alcohol consumption and duration of employment did not significantly differ between the three worker groups (Table 1). Worst and best eyes CCI values were randomly distributed among the left and the right eyes (data not shown). The distribution (mean, 5th and 95th percentiles) of worst, best, and mean CCI, and the frequencies of dyschromatopsia are shown by exposure groups in Table 1. Worst eye CCI was significantly higher than best eye CCI (data not shown; paired Students t test P < .0001). On average, toluene exposure level for rotary printers was a third of the French time-weighted average occupational exposure limit (i.e., 375 mg/m3 (100 ppm)), whereas ambient background exposure level was about one-tenth (Table 2). Current airborne toluene level was positively and significantly correlated with cumulative past toluene and past total hydrocarbon exposure indices ( r = .55, P < .001, n = 124 and r = .48, P < .001, n = 123, respectively). Past exposure indices were strongly correlated with each other (r = .96, P < .001, n = 123). 3.2. Potential confounding factors Workers age was positively and significantly correlated with past toluene and past total hydrocarbon exposure indices (r = .52, P = .0001, n = 124 and r = .57, P = .0001,

n = 123, respectively), and these correlations were stronger after exclusion of the nonexposed group (r = .59, P = .0001, n = 106 and r = .68, P = .0001, n = 105, respectively). No significant relationship was observed between age and airborne toluene. Workers age was positively correlated with each CCI measurement, but these relationships were significant for best eye CCI only (r =.19, P =.03) (Table 3). Quantitative daily alcohol consumption was not related to exposure levels or to any CCI measurement. CCI values did not differ significantly between participants with the highest daily alcohol consumption ( ! 50 g/day; n = 64) and the nondrinkers (n = 36) (data not shown). Significant positive correlations were observed between workers age and quantitative daily alcohol consumption (r = .28, P < .01,

Table 2 Exposure levels by exposure groups Type of exposure Current toluene exposure (mg/m3) Past cumulative toluene (mg/m3 years of exposure) Past cumulative total hydrocarbons (mg/m3 years of exposure) Exposed (n = 72) 136 [50 296] 1299 [200 6008] 1793 [200 21243] Ambient exposure (n = 34) 32 [16 74] 299 [34 1808] 534 [31 4586]

Geometric mean [5th and 95th percentiles]. Exposure levels were not measured and assumed to be equal to zero for the 19 nonexposed participants.

D. Campagna et al. / Neurotoxicology and Teratology 23 (2001) 473480 Table 3 Spearman correlation coefficients between exposure levels, participant characteristics, and CCI CCI n Characteristics Age Alcohol consumption Employment duration Exposure Current toluene exposure Past cumulative toluene Past cumulative total hydrocarbons crude crude crude 125 124 125 Mean two eyes r ( P) .16 (.07) .007 (.93) .13 (.16) Worst eye r ( P) .14 (.12) .005 (.96) .10 (.26)

477

Best eye r ( P) .19 (.03) .02 (.81) .17 (.06)

crude adjusteda crude adjusteda crude adjusteda

124 122 124 122 123 123

.25 .27 .35 .34 .30 .26

(.005) (.003) (.001) (.001) (.001) (.004)

.27 .28 .34 .34 .29 .27

(.002) (.002) (.001) (.001) (.001) (.003)

.18 .22 .31 .28 .26 .20

(.04) (.02) (.001) (.002) (.003) (.03)

n: Number of measurements; r ( P): Spearman correlation coefficient ( P value). a Adjusted for age and alcohol consumption.

n = 124) and duration of employment ( r = .25, P < .01, n = 124). Employment duration was positively but not significantly correlated to the mean, worst, and best eye CCI (Table 3). A strong correlation was observed between workers age and employment duration (r = .86, P = .0001, n = 125). Exclusion of nonexposed participants only slightly modified these relationships. 3.3. Exposure levels and color vision Workers from the exposed group and from the group with an ambient exposure had significantly higher CCI values than the nonexposed participants (Table 1). The same differences were observed when using t tests (data not shown). Sixty of the 125 participants (48%) had profiles of acquired dyschromatopsia: 5 (4%) had type I (loss in the red green range), 5 (4%) had type II (red green and blue yellow loss in the same eye), and 50 (40%) had type III (loss in the blue yellow range) (Table 1). Thirteen (10%) other workers also presented color vision loss with no specific axis. The prevalence of having any dyschromatopsia (type I, II, or III) was significantly higher among exposed workers and participants with an ambient exposure (52% and 56%) than among nonexposed participants (21%) ( P =.02 and P =.01, respectively) (Table 1). Type I and II dyschromatopsia were not observed in the nonexposed group. Table 3 shows Spearman correlation coefficients between exposure levels and CCI measurements. Exposure levels (airborne toluene, past toluene, and past total hydrocarbons) were positively and significantly correlated with each CCI measurement (.18 r .35). The strongest and most significant correlations were observed for mean and worst eye CCI. Adjustment for age, daily alcohol consumption, and duration of employment did not modify these relationships. No threshold level of exposure to toluene or to cumulative

past toluene or total hydrocarbon indices below which homeostatic mechanisms provide protection from color vision change was observed (data not shown). When excluding nonexposed participants, positive correlations were observed between each exposure measurement (airborne toluene, past toluene, and past total hydrocarbons) and CCI indices (data not shown), but significant correlations were only observed with past toluene exposure index (mean, worst, and best eye CCI: r = .25, P < .01, n = 106; r = .23, P < .05, n = 106; r = .24, P < .01, n = 106, respectively). In logistic multiple regression models, the presence of acquired dyschromatopsia (type I, II, or III) was positively related to the different exposure measurements (Table 4). Age and alcohol consumption were not related to acquired dyschromatopsia, and adjusting for these two variables did not change these relationships.

4. Discussion We examined the effect of occupational airborne toluene exposure and past cumulative toluene and total hydrocarbon exposure on color vision in healthy male workers who were subjected to direct or ambient background

Table 4 Logistic regression models of acquired dyschromatopsia (type I, II, or III) according to exposure levels Dyschromatopsia Exposure levels Current toluene exposure Past cumulative toluene Past cumulative total hydrocarbons crude adjusteda crude adjusteda crude adjusteda OR 1.27 1.26 1.21 1.22 1.15 1.16 [95% CI] [1.02 1.58] [1.00 1.57] [1.04 1.39] [1.04 1.43] [1.02 1.31] [1.01 1.34]

OR: odd ratios. [95% CI]: 95% confidence interval. a Adjusted for age and alcohol consumption.

478

D. Campagna et al. / Neurotoxicology and Teratology 23 (2001) 473480

exposure in a photogravure plant or were not exposed and worked in a bookbinding plant. Toluene levels measured among exposed workers (mean 157 mg/m3, 5th and 95th percentiles [50 296] mg/m3) and participants with ambient exposure (35 mg/m3, [16 74] mg/m3) were overall lower than the Threshold Limit Value Time Weighted Average (TLV TWA) proposed by the American Conference of Governmental Industrial Hygienists (ACGIH) of 50 ppm (188 mg/m3) [1]. Our study shows a positive relationship between current airborne toluene measured at the workstations and color vision loss. Positive relationships with color vision loss were also observed with past cumulative toluene and total cumulative hydrocarbon exposure. These relationships were observed with or without adjustment for age and alcohol consumption. In a similar study, a subclinical reduction in color vision was observed in 33 toluene exposed rubber workers, compared with 16 participants not exposed to ophtalmotoxic substances [10]. A positive relationship between color perception and a cumulative exposure index, but not urinary toluene levels, was also found [10]. Others [31,33] found significantly higher CCI values among 46 shoe factory workers and 37 photogravure printing press workers exposed to toluene [average exposure level of 495 mg/m3 (132 ppm)] compared with 90 participants with background exposure [average exposure level of 120 mg/ m3 (32 ppm)]. In another study, color vision loss was more frequent in workers predominantly exposed to toluene in an ink-manufacturing plant compared to controls [16]. Additionally, visual evoked potentials have been shown to be altered in printing press workers exposed to long-term low concentrations of toluene [30]. Optic neuropathies with dyschromatopsia, blindness, or retinal impairment have also been reported in participants who chronically sniffed toluene based glue [27]. Not all studies, however, report similar results. No effect on color vision was found in 59 male photogravure workers exposed to unspecified levels of airborne toluene (blood toluene ranging from < 0.22 to 7.37 mg/l) for periods of 1 month to 36 years (mean of 10 years) in the workplace [23]. In another recent study, an acute exposure to toluene (mean blood toluene concentration of 4.92 mg/l) was not associated with an impairment of color vision among eight male photogravure workers [22]. Color vision was also not impaired in a group of 82 print shop workers exposed to low levels of acetone, isopropyl alcohol, toluene, xylene, and 2-ethoxyethanol [4]. No color vision loss was expected in that study as the exposure was low and the TLV for the mixture of chemicals for the most exposed workers was determined to be at 6% of the ACGIH TLV (obtained by applying the ACGIH recommended formula for mixtures) [18]. For comparison, the current exposure in our ambient exposure group was about 19% of the TLV, and lower than in the past (data not shown). Finally, no effect on color vision was noted in a large investigation among 261 workers exposed to toluene [geometric mean of 173 mg/m 3

(46 ppm)] and to tetrachloroethylene and 120 controls [24]. These negative results may indicate that exposure was insufficient to produce an effect, or that the instrument used in some of these studies (new Lanthony Panel [14]) may not be as sensitive to the type of color vision loss resulting from toluene exposure as the one we used (desaturated color vision test Lanthony D-15 [15]) [17]. This second hypothesis may be more relevant to the last reported negative study [24] as toluene exposure levels in this survey were higher than those encountered in our population. The desaturated color vision test Lanthony D-15 [15] used in our study requires placing caps in order of chromatic similarity (from blue, to green, to yellow and red). In the Nakatsuka et al. study, participants did only part of the new Lanthony Test [14], separating colored from achromatic (grey scaled) caps, then ordering the achromatic caps [11]. The two tests (Desaturated Lanthony D-15 and new Lanthony Test) may indeed investigate different physiological pathways of color vision [17]. In contrast to the large negative study [24] that examined the effect of toluene on the type of acquired dyschromatopsia (blue yellow loss and red green loss) [27], our findings are based on a quantitative color vision loss estimated by a chromatic confusion index proposed by Bowman [6]. No threshold level of exposure to toluene or to cumulative past toluene or total hydrocarbon indices was identified in our population. However, one of the methods [7] used in our study previously identified threshold levels of exposure to styrene below which homeostatic mechanisms provide protection from color vision change [7,13]. In our study, a higher proportion of participants was affected with acquired dyschromatopsia (type I, II, or III) in the exposed worker group (52% among exposed participants and 56% among the group with an ambient exposure) than in the nonexposed group (21%). The type of color vision loss provides an indication of neural damage area. Type III loss, the most commonly observed dyschromatopsia in exposed workers in our study (43%), reflects external retinal changes, while type II is indicative of a more advanced damage involving internal retinal layers and/or the optic nerve (see Ref. [11] review article for information concerning retinal change and pathology associated with Type II and III dyschromatopsia). Other studies also reported higher prevalence of type II and type III dyschromatopsia among workers exposed to toluene compared to nonexposed workers [32,33]. While examining the relationship between CCI scores and exposure levels, we took into account several covariates that have been reported to influence acquired color vision loss. The expected relationship between age and CCI [5,19,20,31] was only observed for the best eye, not for the worst eye or the mean of both eyes, suggesting that the age effect was small compared to exposure effect. Agerelated color vision loss may result from changes in the lens or in the macula [25]. The expected increase in chromatic confusion with alcohol consumption [5,21,28] was, how-

D. Campagna et al. / Neurotoxicology and Teratology 23 (2001) 473480

479

ever, not observed in this population of workers. This may suggest that toluene ophthalmotoxicity hides the effect of alcohol consumption. Similar exposure effect relations were observed for mean and worst CCI, and weaker correlations were observed with best eye CCI. To our knowledge, there is no physiological reason to believe that best eye or worst eye CCI is more representative of early alterations than mean chromatic discrimination. However, a stronger association with worst eye CCI was reported among workers exposed to styrene [9]. Particular attention was paid to the comparability of the three exposure groups in this study, and color vision was tested using standardized operating procedure under identical conditions. All participants were male, and the groups had similar alcohol consumption. Controls and exposed participants worked rotating shifts and were all tested at the beginning of a day shift. The exposure of the ambient participants contrasted markedly with that of the exposed workers and was only one-fifth as high. The nonexposed participants were not exposed to neurotoxicants. Finally, biases related to socioeconomic status are unlikely, because the controls were recruited among workers having similar occupational activity. On the other hand, the exposed group in the present study was composed of printing industry workers whose color discrimination was probably better than that of the general population because of their work, which makes constant demands on color vision [12]. For this reason, selection of a control group from workers who frequently need to rely on color vision may be more appropriate to study color vision among printing workers [12]. In conclusion, this study showed positive relationships between color vision loss and (1) airborne toluene, (2) past cumulative toluene, and (3) past total cumulative hydrocarbon exposure, among male workers. These findings suggest that the Lanthony D-15 Hue Desaturated color vision panel detects early neurotoxic effects among workers exposed to toluene. As suggested by Geller and Hudnell [11], further studies on color vision assessment should use the Lanthony D-15 saturated and the Lanthony D-15 desaturated tests to evaluate the effect of neurotoxicants and the respective sensitivity of these two tests [11]. Our results also show that an impairment of color vision can appear at toluene exposure levels below the actual ACGIH TLV TWA of 50 ppm (188 mg/m3), raising the question of whether this limit is sufficiently protective. Finally, the observation of correlations between impaired color perception and cumulative solvent exposure is supportive of recent findings [10] and suggests that there is a progressive loss of color vision as exposure is continued.

would also like to thank Mrs. Forestier, Neels, and Prevost, occupational nurses, for their help in scheduling ne e who was in charge of the appointment and Mrs. Ce exposure data management. We also thank Dr. FagotCampagna for her valuable comments in the preparation of this manuscript.

References
[1] ACGIH, 2001 TLVs1 and BEIs1, American Conference of Governmental Industrial Hygienists Inc., Cincinnati, 2001, p. 192. [2] P. Arlien-Sborg, Solvent neurotoxicity, CRC Press, Boca Raton, FL, 1992. [3] Agency for Toxic Substances and Disease Registry (ATSDR), Toxicological profile for Toluene, Atlanta, U.S. Department of Health and Human Services, 2000, p. 312. [4] B. Baird, J. Camp, W. Daniell, J. Antonelli, Solvents and color discrimination ability, nonreplication of previous findings, JOM 36 (1994) 747 751. [5] L. Blain, D. Mergler, La dyschromatopsie chez des personnes expo es professionnellement aux solvants organiques, J. Fr. Ophthalmol. se 2 (1986) 127 133. [6] K.J. Bowman, A method for quantitative scoring of the Farnsworth Panel D-15, Acta Ophthalmol. 60 (1982) 907 916. [7] D. Campagna, F. Gobba, D. Mergler, T. Moreau, C. Galassi, A. Cavalleri, G. Huel, Color vision loss among styrene-exposed workers; neurotoxicological threshold assessment, Neurotoxicology 17 (1996) 367 374. [8] D. Campagna, G. Huel, F. Girard, J. Sahuquillo, P. Blot, Environmental lead exposure and activity of d-aminolevulinic acid dehydratase (ALA-D) in maternal and cord blood, Toxicology 134 (1999) 143 152. langer, G. Truchon, C. Ostiguy, [9] D. Campagna, D. Mergler, G. Huel, S. Be D. Drolet, Visual dysfunction among styrene exposed workers, Scand. J. Work Environ. Health 21 (1995) 382 390. [10] A. Cavalleri, F. Gobba, E. Nicali, V. Fiocchi, Dose-related color vision impairment in toluene-exposed workers, Arch. Environ. Health 55 (2000) 399 404. [11] A.M. Geller, K. Hudnell, Critical issues in the use and analysis of the Lanthony desaturate color vision test, Neurotoxicol. Teratol. 19 (1997) 455 465. [12] M. Gonzalez, M. Velten, A. Cantineau, Increased acquired dyschromatopsia among solvent-exposed workers: An epidemiology study on 249 employees of an aluminium-foil printing factory, Int. Arch. Occup. Environ. Health 71 (1998) 317 324. [13] R. Kishi, S. Tozaki, Y.-Y. Gong, Impairment of neurobehavioral function and color vision loss among workers exposed to low concentration of styrene a review of literatures, Industrial Health 38 (2000) 120 126. [14] P. Lanthony, Manual of new color test, Lanthony Selon Munsell, Luneau, Paris, 1975. [15] P. Lanthony, A. Dubois-Poulsen, Le Farnsworth-15 Desature, Bull. Soc. Ophthalmol. Fr. 73 (1973) 861 866. langer, M. Bachand, S. Be dard, C. Brabant, S. De [16] S. Legault-Be Grosbois, D. Mergler, Perte de discrimination chromatique chez ` une exposition complexe et variable aux des travailleurs soumis a solvants organiques, Arch. Mal. Prof. 49 (1988) 475 482. [17] D. Mergler, Behavioral neurophysiology: quantitative measures of sensory toxicity: Chap. 47, in: L. Chang, W. Slikker (Eds.), Neurotoxicology: Approaches and Methods, San Diego, Academic Press, 1995, pp. 727 736. [18] D. Mergler, Color vision: replication, rather than nonreplication of findings, J. Occup. Environ. Med. 37 (1995) 789 790. langer, S. De Grobois, N. Vachon, Chromal focus of [19] D. Mergler, S. Be

Acknowledgments We are grateful to the management and to all the workers of the plants who participated in this study. We

480

D. Campagna et al. / Neurotoxicology and Teratology 23 (2001) 473480 acquired chromatic discrimination loss and solvent exposure among printshop workers, Toxicology 49 (1988) 341 348. D. Mergler, L. Blain, Assessing color vision loss among solventexposed workers, Am. J. Ind. Med. 12 (1987) 195 203. D. Mergler, L. Blain, J. Lemair, F. Lalande, Colour vision impairment and alcohol consumption, Neurotoxicol. Teratol. 10 (1988) 255 260. A. Muttray, V. Wolters, D. Jung, J. Konietzko, Effects of high dose of toluene on color vision, Neurotoxicol. Teratol. 21 (1999) 41 45. A. Muttray, V. Wolters, O. Mayer-Popken, K.H. Schicketanz, J. Konietzko, Effect of subacute occupatioanl exposure to toluene on color vision, Int. J. Occup. Med. Environ. Health 8 (1995) 339 345. H. Nakatsuka, T. Watanabe, Y. Takeuchi, N. Hisanaga, E. Shibata, H. Suzuki, Y.H. Huang, Z. Chen, Q.S. Qu, M. Ikeda, Absence of blue yellow color vision loss among workers exposed to toluene or tetrachloroethylene, mostly at levels below occupational exposure limits, Int. Arch. Environ. Health 64 (1992) 113 117. J. Pokorny, V.C. Smith, A.J. Verriest, Congenital and acquired colour vision defects, Grune and Stratton, New York, NY, 1979. B. Stengel, S. Cenee, J.-C. Limasset, F. Diebold, D. Michard, P. Druet, D. Hemon, Immunologic and renal markers among photogravure printers exposed to toluene, Scand. J. Work Environ. Health 24 (1998) 276 284. [27] N. Toyonaga, E. Adachi-Usami, H. Yamazaki, Clinical and electrophysiological findings in three patients with toluene dependency, Doc. Ophthalmol. 73 (1989) 201 207. [28] E. Valic, T. Waldhor, C. Konnaris, A. Michitsch, C. Wolf, Acquired dyschromatopsia in combined exposure to solvents and alcohol, Int. Arch. Occup. Environ. Health 70 (1997) 403 406. [29] F. Verriest, On acquired deficiency of colour vision with special reference to its detection and classification by means of the test of Farnsworth, Vision Res. 1 (1961) 201 219. [30] A. Vrca, D. Bozicevic, V. Karacic, R. Fuchs, D. Prpic-Majic, M. Malinar, Visual evoked potentials in individuals exposed to long-term low concentrations of toluene, Arch. Toxicol. 69 (1995) 337 340. [31] M. Zavalic, Z. Mandic, R. Turk, A. Bogadi-Sare, D. Plavec, Quantitative assessment of color vision impairment in workers exposed to toluene, Am. J. Ind. Med. 33 (1998) 297 304. [32] M. Zavalic, Z. Mandic, R. Turk, A. Bogadi-Sare, D. Plavec, M. Gomzi, L.J. Skender, Assessment of colour vision impairment in male workers exposed to toluene generally above occupational exposure limits, Occup. Med. 48 (1998) 175 180. [33] M. Zavalic, Z. Mandic, R. Turk, A. Bogadi-Sare, D. Plavec, L.J. Skender, Qualitative color vision impairment in toluene-exposed workers, Int. Arch. Occup. Environ. Health 71 (1998) 194 200.

[20] [21] [22] [23]

[24]

[25] [26]