Urinary styrene in the biological monitoring of styrene exposure.

The urinary excretion of styrene represents a promising indicator of exposure to this solvent. Nevertheless extensive research under field conditions is scant. In this investigation 214 styrene-exposed workers from 10 fiberglass-reinforced plastics factories were studied. Environmental monitoring was performed by personal passive sampling. Blood styrene and the urinary excretion of styrene and its main metabolites, mandelic acid (MA) and phenylglyoxylic acid (PGA), were measured. The correlation coefficient between the time-weighted average of environmental styrene and the mean urinary excretion of styrene was 0.88 (0.91 after logarithmic transformation), compared with the 0.82 and 0.78 of the end-of-shift MA and PGA values, respectively. A high correlation (0.86) was also found between styrene in the blood and urine. The results, obtained under field conditions with a large group of exposed workers, confirm the usefulness of the urinary excretion of styrene as an exposure index for the biological monitoring of styrene exposure.

Another problem in the biological monitoring of styrene exposure is the large interindividual variability caused for MA and PGA excretion by factors such as nonspecificity of the two metabolites, interference from other solvents, drugs and alcohol consumption, differences in individual metabolism, and the like (1, II). An index proposed for the biological monitoring of styrene exposure that possibly avoids some of the cited problems is the measurement of the urinary excretion of the solvent (12,13). Urinary styrene, like the urinary metabolites, mainly represents a weighted index of exposure, while the concentrations of the solvent in alveolar air represent instantaneous values (14).
Some preliminary data suggest that this index is reliable in predicting styrene exposure, the correlation coefficient between airborne styrene levels and urinary excretion varying between 0.88 for field conditions and 0.93 for volunteers (1,14). Moreover this test is specific and seems to be hardly affected by other solvents interfering in the biotransformation rate or by the presence of common metabolites (14) . After considering these preliminary data, we decided to study urinary styrene excretion in a group of workers from different factories producing products made of fiberglass-reinforced plastics. Our goal was to define, under field conditions, the usefulness of urinary styrene as an index of occupational exposure to styrene.

Subjects
Two hundred and fourteen workers (114 men and 100 women), mean age 28.1 (SD =10.6) years, from 10 fiberglass-reinforced plastics factories were examined. The workers were engaged in the hand production of products of various shapes and dimensions .
Information on the workers ' health status, smoking habits, and alcohol consumption was collected by means of a questionnaire during a medical examination . All of the workers were apparently healthy, and the daily average consumption of ethanol was always less than 50 g.

Environmental air sampling
For each examined worker, exposure was measured in their breathing zone with personal passive dosimeters (TK 200 Zambelli). Airborne styrene levels were determined on Thursday. This day was selected because the workers in several of the factories were engaged in the maintenance of plants on Friday (the last workday of the week, considered in various other studies), and therefore the exposure was not comparable with other days for aspects like quantity of styrene handled, work load, use of acetone, and the like.
A whole workshift (8 h) was evaluated. Morning and afternoon half shifts were monitored separately, each sampling period lasting 4 h. The only other solvent used in the workplaces was acetone. Nevertheless its environmental concentrations were always low, never exceeding 500 mg . m-3 •

Biological monitoring
On the same day of the environmental monitoring, at 0800, all of the workers were asked to empty their bladder. Then at 1200 (end of the first half shift) a sample of urine was collected. At 1300 the subjects emptied their bladder once more, and a specimen was again obtained at 1700 (end of shift) .
Only urine sample s of the whole half shifts (from 0800 to 1200 and from 1300 to 1700, respectively) were considered. If the workers had to urinate during a half shift, the sample was discarded . Five samples from the morning and 10 from the afternoon were excluded from the study for this reason.
Within 2 min of the voiding, in an atmosphere not polluted by styrene, 10 ml of the urine samples was transferred to 120-ml glass vials (Supelco Vials) with airtight plugs without silicone and kept refrigerated until the analysis .
For 100 of these workers , 5 ml of venous blood was collected from a brachial vein at the end of the morning shift, 10 to 15 min after the end of the styrene exposure. The samples were immediately injected into heparinized glass vials as described for urinary styrene . 176 For a smaller group of 65 workers MA, PGA, and styrene were measured in urine samples collected at the end of the shift and at 0800 on Friday (next morning sample). This subgroup was fully comparable to the total group in all relevant aspects (age, gender distribution, ethanol consumption, type and entity of exposure , work load , etc).

Analytical methods
Styrene was measured by a gas chromatograph (HP 5880 A) connected to a mass selective detector (HP 5970 A). Details on the analytical methods have been reported elsewhere (14).
The activated charcoal of passive dosimeters was desorbed with 5 ml of carbon disulfide and kept at room temperature (20°C) for I h, during which it was periodically shaken. Desorption liquid (0.5 ml) was injected into the gas chromatograph-mass selective detection unit. Vials containing urine (or blood) were stored in a refrigerator (4°C) until the analysis, which was always performed within a few days. For 2 h before the analysis the vials were kept at 37°C and periodically shaken to speed the separation ad equilibrium of the styrene vapor between the urine or blood and the headspace air.
Two milliliters of the headspace of the vials was then injected (2 ml gastight syringe Hamilton) into a O.5-mlloop of the gas chromatograph.
Urine samples for the MA and PGA measurements were kept frozen until the analysis by a gas chromatographic method (15).
The detection limit of the technique was 0.0048 umol . I-I. The mean recovery at three different styrene concentrations ranged from 94 to 103%. The interassay coefficient of variation for styrene was 4.2% (for 10 determinations), the mean value being 1022/lmol·I-I. Some volatile chemicals, such as styrene, appear to be eliminated in the kidney by a diffusion proce ss determined by the equilibration of partial pressures in urine and plasma (16). As a result, the urine:blood ratio equals the urine-blood distribution coefficient, and the concentration of the determinant in urine is independent of urinary output. Adjusting for creatinine does not seem to be justified if the excretion mechanism of the determinant differs from the excretion mechani sm of creatinine (17); therefore our urinary styrene data are presented without correction. Also for the urinary styrene metabolites (MA and PGA), the results were not corrected as, according to some authors (18,19), an adju stment for variat ions in urine flow using densit y or creat inine seems to offer no definite advantages.

Statistical analysis
All of the statistical analyses were performed with the SPSS-PC+TM software of the statistical package for the social sciences. A simple linear regression was performed with environmental styrene as the independent variable and urinary styrene or its urinary metabolites as the dependent variables. The confidence limits were calculated for the expected values of the dependent variable.

Results
The results of the environmental and biological monitoring are summarized in table 1. For the whole group, the mean time-weighted (TWA) exposure (843.8 umol . m? or 87.9 mg . m:') was about onethird of the current threshold limit value (TLV), as the TWA , proposed by ACGm (213 mg· m-3 or 2044.931 Ilmol · m-'), but the range of the values varied from 23 up to 7397 umol . m-3 • Correspondingly, a wide variation was observed also for the styrene concentration of the urine and blood. Exposure during the morning and afternoon shifts was similar.
The correlation coefficients (r) between airborne styrene and urinary styrene are presented in table 2; the corresponding equations are reported in figures 1-4. We have considered both the relationship between each half-shift exposure and the corresponding urinary styrene values (morning exposure versus urinary styrene at 1200 and afternoon exposure versus urinary styrene at 1700, respecti vely) (figures 1 Scand J Work Environ Health 1993, vol 19, no 3 and 2), and 8-h TWA exposure versus mean urinary styrene levels (urinary styrene at 1200 + urinary styrene at 170012) (figure 3). In every case the correlation coefficients ranged from 0.82 to 0.88.
Since the distribution of both the environmental and biological values was fairly skewed, a logarithmic transformation of the variables was performed. A slight increase in the correlation coefficients was observed , as shown in table 2. The correlation coefficient between the environmental and blood styrene concentrations was also high (r = 0.86; 0.89 after logarithmic transformation) (table 2). Remarkably, the urinary excretion of styrene (1200 samples) was strictly related (r = 0.86) to the blood styrene values obtained at the same time ( figure 4). The results of the environmental and biological monitoring performed in the subgroup of 65 workers are reported in table 3. Tables 4 and 5 show the correlation coefficients and the regression equations between airborne styrene and the exposure indices. The correlation coefficients for urinary styrene were close to those observed for the whole group. MA and PGA in the urine samples taken at the end of the shift correlated better with the styrene exposure (r =0.82 and 0.78, respectively) than the values obtained the next  Environmental styrene JLmol • m-3 Environmental styrene (TWA) urnol -rn? morning (r =0.46-0.53). The correlation coefficients between the metabolite concentrations in the sample s taken the next morning and the TWA airborne styrene levels were similar to those obtained between the same metabolites and the morning and afterno on environmental concentr ations of the solvent (table 4).

Discussion
Thus far little attention has been paid to gender-related differences in the correlations between environmental styrene levels and exposure indices. For this reason we examined male and female workers separately in a preliminary analysis, but no differences were observed between the two groups. As an example, the regression lines between the TWA environmental levels of styrene and the mean urinary styrene were y =0.29 x + 129.7 and y =0.27 x + 154.2 for the men and women, respectively. The difference was not significant. The same was true for all of the other exposure indices. For this reason, all of the data were pooled in the rest of the analyses . The observed wide variability in the TWA levels of airborne styrene (20-7400 umol . rrr" ) is the rule in the reinforced plastics industry, as previously documented (20)(21)(22). In the examined group, composed of workers who were exposed to a wide range of styrene levels and who perform ed different job s in 10 different factories, we confirmed our preliminary results ( 12,23). The correlation between environmental styrene and urinary excretion of the solvent was very good, as shown in table 2 and figures 1-3. The correlation with exposure was of the same order, or possibly better, as those observed for metabolites in other field studies reported in the literature (8,10,(24)(25)(26).
Furtherm ore, the styrene concentrations of the samples collected during each half shift were closely related to the environmental exposure during the same work period. The styrene in the urine sample taken at 1200 represented the morning exposure well, while the styrene in the urine sample s taken at 1700 Sum of urinary phenylglyoxylic acid and mandelic acid concentrations (mol· 1-1), next morning 58 4.88 6.14 0.14-29.15 • Geometric mean. b Geometric standard deviation.
e End of afternoon half shift = end of shift.
d Mean of the morning half -shift and end-of-shift samples combined .   represented the afternoon exposure well. For an evaluation of the exposure during the whole shift, the mean of the urinary styrene excretion at 1200 and 1700 must be considered (table 2, figures 1-3).
As observed by others for metabolites (26--28), also for urinary styrene, the logarithm ic transformation of data further improved the correlation coefficient. As an example, the correlation was increased from 0.88 to 0.9 I for the relation between the TWA 180 exposure and the urinary styrene concentration (table 2).
The regres sion line between the exposure and the urinary styrene levels did not start from the origi n of the Cartesian axis ( figure 3). This phenom enon suggests the existence of a body burden of styrene on Thursday. Such an accumulation of styrene during the workweek was not unexpected accordin g to pharmacokinetic data (29,30) and has also been re-ported earlier (7,14,31). We also observed that the intercept of the regression line of the afternoon was nearly 40% higher than that of the morning regression line. This result suggests an increase in body burden during the day (figures I and 2).
Confounding by acetone exposure was unlikely since values equivalent to the exposure levels measured for this compound have not earlier been found to interfere with styrene metabolism or elimination (32).
It was not possible for us to evaluate the effect of alcohol consumption on styrene metabolism. Nevertheless, the current alcohol intake of the workers was fairly moderate «50 g per day). According to our data, the biological equivalent exposure limit (BEEL) for urinary styrene, corresponding to the current TLV-TWA of the ACGIH (50 ppm), is 727.7 nrnol-I-I; this limit is close to the BEEL of 768 nmol . I-I previously proposed (12,23).
A high correlation (r = 0.89 after logarithmic transformation) was observed between the urinary styrene and blood styrene levels of the samples collected a few minutes (10 to 15 min) after the end of exposure ( figure 4) . This finding supports the hypothesis that urinary styrene may be representative of the internal dose of styrene.
Recent data derived from a physiological model also suggest that the concentration of the solvent in the brain, one of the main target organs, is predicted well by urinary levels (33). In consequence, urinary styrene values are likely to represent an index of the active dose of the solvent, rather than a mere exposure index . This possibility is also supported by our preliminary observations on the effects of the solvent on color vision (34). Data obtained in the subgroup of 64 workers were similar to the results of the whole group (table 4). The styrene concentrations of the urine samples collected at the end of the half shifts correlated well with the exposure during the corresponding period of time (r = 0.88 and 0.86, respectively, for the morning and afternoon shifts), and the TWA exposure was related to the mean urinary styrene levels (r = 0.88). The regression lines were also similar (table 5).
The correlations between airborne styrene and its metabolite concentrations were generally a little lower than for urinary styrene, but of the same order currently reported by others (8,10,25,26) . In our end-shift samples we observed no definite improvement in the correlation coefficients when the sum of MA and PGA was used instead of MA only, while PGA alone was less related to the exposure levels. These results are substantially in agreement with those of other studies (I , 26,35). Finally, unlike what has been reported by others, the next-morning metabolite values were definitely less related to the external exposure than the end-of-shift values were. This result, together with the similarity of the correlations between the next-morning metabolite levels and the TWA of the morning and afternoon Scand J Work Environ Health 1993. vol 19, no 3 airborne levels of styrene, suggests that there was a relatively low variability in the environmental styrene concentration (or in the individual exposure) during the workday (I).
Our results suggest that urinary styrene represents a good exposure index for styrene. The collection of the samples needs more care than for metabolites, but this care is largely compensated by the specificity and the easily achievable relation to the dose also under field conditions. This index is valid even when only a single half shift can be evaluated.
Nevertheless, it must be stressed that the biological limit that we calculated (728 nmol . I-I) was based on an environmental limit. Factors such as work load, previous exposure, drugs and alcohol consumption, and individual variability have been proved to interfere with the absorption and metabolism of the solvent. Furthermore it is clear that even body burden may affect the BEEL value. For these reasons, to define effective biological exposure limits further, research on effect indices is required. From this point of view , the studies on color vision loss (34) seem to represent a promising beginning.