Toxicology Letters, 63 (1992) 333-343 0 1992 Elsevier Science Publishers B.V. All rights reserved 0378-4274/92/S 5.00
333
TOXLET 028 18
Urinalysis vs. blood analysis, as a tool for biological monitoring of solvent exposure*
Toshio Kawaia, Tomojiro Yasugia,b, Kazunori Mizunuma a,b, Shun’ichi Horiguchi” and Masayuki Ikedab “Osaka Occupational Health Service Center, Osaka (Japan)
and bDepartment of Public Health, Kyoto
University Faculty of Medicine, Kyoto (Japan)
(Received 22 June 1992) (Accepted 26 August 1992) Key words: Biological monitoring; Blood analysis; n-Hexane; Toluene; Occupational health; Urinalysis
SUMMARY Blood and urine samples were collected at the end of an 8-h workshift from 30 male workers exposed to a mixture of n-hexane, ethyl acetate and toluene (each being about 2 ppm as geometric means) and also from 20 nonexposed male workers. Blood samples were analyzed for n-hexane and toluene, and urine samples were analyzed for n-hexane, toluene, 2,5-hexanedione (both with and without hydrolysis) and hippuric acid. Based on the correlation between biological exposure indicators and solvent concentrations in air, sensitivity as an exposure indicator was compared between solvents in blood and solvents or metabolites in urine in terms of the lowest solvent concentration at which the exposed subjects can be statistically separated from the nonexposed. Both n-hexane and toluene in blood were sensitive enough to detect the exposure at 6.1 ppm and 1.4 ppm. respectively. n-Hexane exposure below 2 ppm was detectable also by urinalysis for 2,5-hexadione without hydrolysis. Urinary hippuric acid, however, failed to detect low toluene exposure under the conditions studied. Of additional interest is the fact that toluene in urine correlated significantly with toluene in air, which apparently deserves further study for confirmation.
INTRODUCTION
Biological monitoring of exposure is an important tool in occupational health. Among the various biological matrices for analysis, urine has been commonly ana-
Correspondence to: Professor M. Ikeda, Department of Public Health, Kyoto University Faculty of Medicine, Kyoto 606-01, Japan. * A part of this work was presented at the International Symposium on Biological Monitoring, held on 12-16 October 1992, in Kyoto, Japan.
334
Iyzed for its metabolite(s) in occupational health practice to solvent-exposed workers [I ,2], probably because urine samples are readily available under many circumstances. In recent years, however, the specificity of the urinalysis has been questioned especially when the intensity of exposure to solvent is reduced [3,4] due to the modernization of working environment, and attention toward blood analysis for the solvent per se has increased despite the inherent invasive nature of blood sampling, because blood sampling is also a common practice for the prevention of solvent poisoning. For example, the Ordonnance for Prevention of Organic Solvent Poisoning in Japan request hematology and serum biochemistry for health check-up of the workers exposed to selected organic solvents [5]. In the present study trials were made to compare the sensitivity of blood analysis and urinalysis, taking the two most popular solvents of n-hexane and toluene as an example. MATERIALS AND METHODS
Workers studied The survey was conducted in the second haif of a working week. The solventexposed workers (all men) were employed in an adhesive tape-producing plant for at least a year. Nonexposed controls (20 men) were recruited from offices with no known occupational exposure to organic solvents [6]. Determination of time-weighted average (TWA) intensy ofexposure to organic solvents Each of the exposed workers was equipped with a diffusive sampler (with carbon cloth as an adsorbent [7-91) on his chest for an entire 8-h shift. Analytical procedures for solvents adsorbed were as previously detailed [7-91. The additiveness formula [lo] was employed for integrated evaluation of combined solvent exposure. Analysis of blood and urine Workers were asked to pass urine 2-3 h before the end of the workshift. At the end of the shift, each of the workers was invited to a health examination room where no solvent vapor could be detected. Blood was drawn from the cubital vein within 10 min after leaving the workshop. Five ml of the blood was immediately taken into a 20-ml vial (for head-space gaschromatography (HS-CC); Hewlett Packard, Philadelphia, PA, USA together with 0.03 ml of anticoagulant solution (Anglot; Nippon Shoji, Osaka, Japan) and the vial was sealed with a Teflon-coated septum. The blood sampling was followed by the sampling of urine. Namely, immediately after sampling in a clinical urine cup, a 5-ml portion urine was taken to a HS-GC vial as was the case of blood analysis. The rest of the urine sample was kept at -30°C until analyzed for metabolites. HS-GC analysis of blood and urine for n-hexane and toluene was conducted by the method previously described fll]. In addition to the shift-end urine samples described above, 14 out of the 30 exposed workers also offered preshift urine
335
samples the next morning, some 16 h after the termination of the exposure. The preshift urine samples were subjected to 2,5-hexadione determination. After thawing, the stored urine samples were analyzed for hippu~c acid (a toluene metabolite) by the HPLC method of Kawai et al. [12], and 2,5-hexanedione both by the direct method (i.e., without acid hydrolysis [13]) and the hydrolysis method (i.e., with hydrolysis at pH 0.1 to 0.3 in the presence of sulfuric acid [13]), respectively. Creatinine concentration and specific gravity were measured by the conventional calorimetric and refractometric methods, respectively. The solvent and metabolite concentrations were expressed as observed (i.e., without any correction for urine density), or after correction for creatinine concen~ation [14] or a specific gravity of urine of 1.016 [ 151.The data on urine samples from 20 nonexposed men are cited from a previous publication [6].
Regression lines were calculated by the least-square method. Analysis of variance (ANOVA) was employed for the comparison of means. RESULTS
Z~te~s~ty of exposure to organic solvents The measurements of TWA solvent concentrations in breathing zone air analysis showed that vapors of three solvents, i.e., Pt-hexane, ethyl acetate and toluene coexisted in the air. Thus, the vapor concentrations were expressed both individually and as the sum following an additiveness assumption [IO], taking 40,400 and 100 ppm as the occupational exposure limit @EL) for n-hexane, ethyl acetate and toluene [16]. The results are presented in Table I in terms of AM + ASD, GM (GSD) and the maximum value observed. It is clear from the table that the exposure of the workers
TABLE I INTENSITY OF EXPOSURE TO ORGANIC SOLVENTS Parameter
AM” ? ASDd GM” (GSD)’ Maximum
Solvent concentration n-hexane
EA”
toluene
sumb
5.8 t 7.7 2.2 (4.68) 35.1
5.2 t 7.7 2.1 (3.90) 30.0
3.1 i 2.8 2.1 (2.71) 12.0
0.19 rt 0.22 0.09 (3.83) 0.95
a Ethyl acetate. b Summation after an additiveness formula [lo), taking 40,400 and 100 ppm as occupational exposure limits for n-hexane, ethyl acetate and toluene, respectively 1161.’ Arithmetic mean (unit, ppm). d A~thmetic standard deviation (unit, ppm). a Geometric mean (unit, ppm). t Geometric standard deviation (~~ensionless).
336
was generally low, and below the corresponding OEL when evaluated individually, or even in combination (i.e., the sum in the worst case was 0.95). Correlation of solvent in blood and urine with the solvent in air The possible correlations were examined between the solvent concentration in the blood samples collected at the end of a workshift and the TWA concentration of the solvent in the breathing zone air during the shift. Similar examination was also carried out between the solvent concentration in the shift-end urine and the solvent in the air. The combined regression analyses with 30 exposed and 20 nonexposed subjects (Table II) showed that toluene in blood correlated significantly (P < 0.01) with toluene in air, and that such was also the case (r was approx. 0.7 with P < 0.01) for toluene in urine regardless of correction for urine density. Scatter diagrams between toluene in air and that in urine are depicted in Figure 1 for better understanding. n-Hexane in blood also correlated with n-hexane in air with a correlation coefficient of 0.639. n-Hexane in urine, however, failed to correlate with n-hexane in air, with a low correlation coefficient of well below 0.3, regardless of correction for urine density. Correlation of urinary metabolite with the solvent in air The correlation of two urinary metabolites, 2,5-hexanedione TABLE
II
CORRELATION
OF SOLVENT
Solvent in air (ppm)
n-Hexane
a a andp
from n-hexane and
IN BLOOD
OR URINE
Solvent
Correction
in blood @g/l)
for urine
or urine @g/l)
density
n-hexane
in blood
n-hexane n-hexane
in urine in urine
n-hexane
in urine
toluene toluene
in blood in urine
toluene
in urine
toluene
in urine
are the slope and the intercept
WITH
SOLVENT
IN AIR
Parameters
of regression*
a
B
r
2.99
15.15
1.15
16.45
0.639** 0.224
1.21 1.04
18.55 14.18
6.14 2.01
3.55 4.14
0.782**
none creatinine spec. gravity
3.45 2.44
3.28 3.05
0.696** 0.723**
none creatinine spec. gravity
on y-axis of a regression
(1.016)
(1.016)
0.197 0.229
0.650**
line. y = a.y + p. where x is time-weighted
average solvent concentration in breathing zone air and y is that in blood or urine collected at the end of the workshift. r is the correlation coefficient (**P < 0.01). The unit is ppm for the solvent in air, ,&l for solvent in blood or urine (as corrected for none or for a specific gravity of urine of 1.016) or PCs/g creatinine when corrected for urinary creatinine concentration. Data from 30 exposed and 20 nonexposed cases are subjected to regression analysis.
331
0
5
10
TOLUENE
IN AIR
15
20
(ppm)
Fig. 1. Correlation between toluene in air and toluene in urine. Time-weighted average toluene concentration in breathing zone air during an 8-h shift was measured by diffusive sampling. Urine samples were collected at the end of the shift, and analyzed for toluene. Urinary toluene concentrations are expressed as observed, or after correction for creatinine concentration or a specific gravity of urine of 1,016. The line in the middle is a regression line, two curves on both side of the line show the 95% confidence range of the group means, and the outer-most curves show the 95% confidence range for individual values. Each circle represents one case. Cases on the vertical axis are not shown to avoid congestion.
hippuric acid from toluene, with the levels of corresponding mother solvent in air was examined, and the results are summarized in Table III. Because it is known that the urinary level of 2,5-hexadione (HD) varies subject to the pretreatment with acid hydrolysis [13], the metabolite level was examined under two analytical conditions, i.e., one with the direct method without hydrolysis (HD/sHYD) and the other after hydrolysis at pH < 0.3 (HD/cHYD). Both HD/sHYD and HD/cHYD (i.e., HD regardless of hydrolysis pretreatment) correlated significantly (P < 0.01) with n-hexane in air (Table III) with a correlation coefficient of approx. 0.7 (< 0.6 for HD/cHYD when not corrected for urine density). In sharp contrast, no significant correlation
338
(P > 0.10) could be detected between toluene in air and hippuric in urine, independent of correction for urine density. Biological half-time of n-hexane in man
Urine samples were available from 14 workers both at the end of the shift of the day and immediately before the shift the next day, with a lapse of 16 h between sampling. Accordingly, trials were made to estimate the biological half-time of n-hexane in terms of disappearance of HD from the urine. In practice, such estimation, although crude, is considered possible only with HD/sHYD, because the background level was zero for HD/sHYD, whereas it was not negligible in the case of HD/cHYD [6]. The calculation gave a half-time value of 16.6 ? 12.9 h (AM -t ASD of 14 determinations). Further classification of the 14 subjects according to exposure intensity into >5 ppm, 5 to 10 ppm, and ~10 ppm groups showed that AM+ASD of the half-time was 11.9 + 8.5 h (n = 4), 18.3 & 15.3 h (n = 6) and 19.0 ? 11.3 h (n = 4), respectively. No significant (P > 0.05) difference among the means was detected by ANOVA. DISCUSSION
It is assumed in the present study that the metabolism of two solvents, n-hexane and toluene, are independent of each other; this assumption should be taken as quite
TABLE
III
CORRELATION
OF METABOLITE
Solvent in air (ppm)
Metabolite
IN URINE in urine
WITH
SOLVENT
Correction
for urine
density
IN AIR Parameters a
n-Hexane
n-Hexane
HD/sHYDb
none044
HD/sHYD HD/sHYD
creat.c &g/g treat.) spec. gravityd
@g/I)
I
B 6.9 1.9
1.5
0.696**
3.4
10.9
3.0
0.677** 0.646**
HD/cHYD’
noneWU
36.2
431.0
0.581**
HD/cHYD
treat. t&g/g treat.) spec. gravity @g/l)
39.8
316.8 385.4
0.734**
HBicHYD Toluene
of regression”
(unit)
hippuric
acid
hippuric
acid
none (mg/l) treat. (mg/g treat.)
hippuric
acid
spec. gravity
(mg/l)
50.1 -0.1 16.4
233.2 188.0
5.9
188.4
0.714** -0.018 0.274 0.093
a a, B, r and statistical significance of r are as in Table II. x is time-weighted average solvent concentration in respiratory zone air and y is metabolite in shift-end urine. Data from 30 exposed and 20 nonexposed of 2,5-hexanedione without hydrolysis. ’ Corcases are subjected to regression analysis. b Concentration rected for creatinine concentration. d Corrected for a specific gravity of urine of 1.016. e Concentration of 2,5-hexanedione after acid hydrolysis.
339
LSC-8
(B)
D /
t
0
n-HEXANE IN AIR (ppm)
Fig. 2. Graphical determination of LSC (lowest separation concentration) for n-hexane by analysis of blood and urine. (A) Shows the results of blood analysis; (B) the results of urinalysis. Conditions are as in Fig. 1, except for the outer-most curves which are not shown. Cases on the axis or out of the scales are not depicted. For definition of LX-B, LX-Ul, and LSCU2, see text.
probable especially when the exposure is low [6,17,18]. Co-presence of the third solvent, ethyl acetate, is virtually ignored because it is well known that this ester is readily hydrolyzed when absorbed [19,20] and decomposed to carbon dioxide and water. Based on this assumption, it is possible to compare the sensitivity of two approaches of biological monitoring, blood analysis and urinalysis, from the view point of the power to separate the exposed from the nonexposed subjects. For this purpose, a parameter of the lowest separation concentration (or LSC, in terms of solvent concentration in air, e.g., in ppm) was graphically calculated as the concentration at which the 95% lower limit of the distribution is equal to the 95% upper limit at 0 ppm. LSC was further specified as LSC-B and LSC-U when applied to blood analysis or urinalysis (B and U standing for blood or urine), respectively. In practice, LSC-U can be determined in two ways, one (LSC-Ul) just as in the case of LSC-B, and the other in
340
comparison with the upper 95% limit of the background level because certain metabolites, e.g., hippuric acid, are known to be present even in the urine from the nonexposed subjects (e.g., Ref. 6). Examples of graphical determination of LSC for n-hexane are shown in Figure 2 taking the observed, i.e., noncorrected values of 2,5-hexanedione after acid hydrolysis, as n-hexane metabolite concentrations, and that for toluene in Figure 3. Whereas the three LSCs can be determined in the former case, only LSC-B is available for toluene, because under the study conditions urinary hippuric acid correlates poorly with toluene in air with essentially no dose-dependent increase (Table III). This observation is in agreement with the findings in previous reports that hippuric acid only poorly correlated with toluene exposure among workers exposed to toluene at 26 ppm as a median [3], and that no correlation was observed in the daily variation in toluene exposure intensity and that of exposure-induced increments in hippuric acid concentrations among workers exposed to less than 10 ppm toluene [4]; the observed lack of correlation is apparently due to the rather high background level of hippuric acid (e.g., see Ref. 6).
TOLUENE IN AIR (Pam)
Fig. 3. Graphical determination of LSC and urine. (A) Shows the results of blood of Fig. 2. Note that both LSC-Ul and dependent increase
(lowest separation concentration) for toluene by analysis of blood analysis; (B) the results of urinalysis. Remarks are as in the legend LSC-U2 cannot be determined in B due to the lack of a dosein hippuric acid coupled with wide variation.
341
The results of LX dete~ination are s~marized in Table IV. It is evident from the values given in the table that LSC-B is essentially comparable to LX-UI in the case of n-hexane exposure, and smaller than LSC-U2 when urinary 2,5-hexanedione after acid hydrolysis (i.e., HD/cHYD in Table IV) is evaluated. In the case of toluene, LSC-B is as small as 1.4 ppm in contrast to the fact that no LSC-Ul or LSC-U2 (both based on urinary metabolite levels) is available due to the lack of correlation. Such superiority of blood analysis over urinalysis is a good confirmation, with a separate group of workers, of the previous observation [6], and in agreement with the opinion of Heinrich and Angerer [21] who studied methanol in blood and urine of methanolexposed workers and concluded that blood methanol is a better indicator of methanol exposure than urinary methanol. In practice, however, the invasive nature coupled with the longer biological half-time for the urinary metabolite (e.g., well over 10 h for urinary 2,5-hexanedione as observed in the present study) than for the mother solvent per se in blood is an disadvantage for blood analysis. Of particular interest is the fact that toluene in the shift-end urine correlates significantly (P < 0.01) with the TWA toluene exposure during the shift, in contrast to the fact that urinary n-hexane failed to do so (Table II). This might be a reflection of the TABLE IV LOWEST SEPARATION CONCENTRATION Analysis LSC”
Blood analysis LSC-B for solvent Urinalysis LX-U for solvent without correction corrected for creatinine corrected for spec. gravity (1.016) LSC-Ul for metabolite without correction corrected for creatinine corrected for spec. gravity (1.016) LSC-U2 for metabolite without correction corrected for creatinine corrected for spec. gravity (1.016)
(LSC) BY BLOOD ANALYSIS AND URINALYSIS Solvent n-hexane
toluene
6.1
1.4
NCb NC NC
1.9 3.0 1.8
HDlsHYD’, 5.1: HDlcHYD”, 7.4 HDlsHYD, 4.5: HDlcHYD, 4.2 HDlsHYD, 5.2: HDlcHYD, 4.5
NC NC NC
HD/sHYD, 1.1: HD/cHYD, 31.4 HD/sHYD, 1.9: HD/cHYD, 1I .7 HDlsHYD, 2.0: HDlcHYD, 10.0
NC NC NC
Values are for solvent concentration in ppm. a The lowest concentration at which the exposed are statistically separated from the nonexposed. LSC-B is by blood analysis, and LX-U is by urinalysis. LX-U1 is by comparison with the 95% upper limit at 0 ppm, and LSC-U2 is by comparison with the 95% upper limit of the background level. Also see Figs. 1 and 2. b Not calculable due to lack of correlation. ’ 2,5Hexanedione without hydrolysis. d 2,5-Hexanedione after acid hydrolysis.
342
difference in physical characteristics of the two solvents; toluene is described as being very slightly soluble, whereas n-hexane is insoluble in water, and n-hexane is more volatile (boiling point; 63°C) than toluene (llO.O’C) [22]. Little attention has been paid to toluene in urine as an exposure indicator [ 121. Nevertheless, Ghittori et al. 1233observed a close and significant correlation between TWA intensity of exposure to toluene and toluene concentration in shift-end urine with a correlation coefficient of 0.87 (P < 0.01) among toluene-exposed workers; urinary toluene at 100 ppm toluene exposure was estimated to be 225 pg.4, which is essentially the same as the present results (205 ,@I, see Table II). Although the present observation is not yet conclusive and should be taken as preliminary, the possible value of urinary toluene as a tool of toluene-exposure monitoring deserves further attention for conflation, when high sensitivity as an exposure indicator (LSC-U; about 2 ppm; Table IV), high practicality of FIG-GC, and noninvasive nature of urine collection are in combination taken into consideration. ACKNOWLEDGEMENTS
The authors are grateful to Professor M. Tati and Dr. S. Koshi of Japan Industrial Safety and Health Association, Tokyo, for their support in this work. REFERENCES 1 Lauwerys, R. (1983) Toluene. In: L. Ale&o, A. Berlin, R. Roi and M. Boni (Eds.), Human Biological Monitoring ofIndustrial ChemicaIs Series. Joint Research Centre Ispra Establishment, Ispra, Italy, pp. 179-17s. 2 Lauwerys, R. and Buchet, J.P. (1988) Bioiogical monitoring of exposure to benzene, toluene and xyiene. IARC Sci. Publ. 85,2(X-222. 3 Nise, G. (1992) Urinary excretion of a-cresol and hippuric acid after toluene exposure in rotogravure printing. Int. Arch. Occup. Environ. Health 63, 377.-381. 4 Johnai, H. and Sato, M. (1988) Exposure indices for painters exposed to toluene and xylene at low concentrations. Ind. Health 26, 197-202. 5 Ministry of Labor, the Government of Japan (1989) Ordonnance for the Prevention of Organic Solvent Poisoning, Ministry of Finance, Printing Bureau, Tokyo. 6 Kawai, T., Yasugi, T., Mizunuma, K., Horiguchi, S., Iguchi, H., Uchida, Y., Twani, 0. and Ikeda, M. (1992b) Comparative evaluation of urinalysis and blood analysis as a tool to detect exposure to organic solvents at low concentrations. Int. Arch. Occup. Environ. Health, in press. 7 Hirayama, T. and Ikeda, M. (1979) Applicability of carbon felt to the dosimetry of solvent vapor mixture. Am. Ind. Hyg. Assoc. J. 40, 1091-1096. 8 Ikeda, M., Kumai, M. and Aksoy, M. (1984) Application of carbon felt dosimetry to field studies distant from analyti~l laboratory. Ind. Her&h 2253-58. 9 Kasahara, M. and Ikeda, M. (1987) Spontaneous desorption of organic solvents from carbon cloth. Ind. Health 25, 73-81. 10 American Conference of Governmental Industrial Hygienists (1991) Threshold Limit Values and Biological Exposure Indices. ACGIH, Cincinnati, USA. 11 Kawai, T., Yasugi, T., Mizunuma, K., Horiguchi, S., Morioka, K., Miyashita, Y., Uchida, Y. and Ikeda, M. (1992a) Monitoring of workers exposed to a mixture of toluene, styrene and methanol
343
12
13
14 15 16 17 18 19 20
21 22 23
vapours by means of diffusive air sampling, blood analysis and urinalysis. Int. Arch. Occup. Environ. Health 63,429-435. Kawai, T., Mizunuma, K., Yasugi, T., Horiguchi, S., Uchida, Y., Iwami, O., Iguchi, H. and Ikeda, M. (1991) Urinary methylhippuric acid isomer levels after occupational exposure to xylene mixture. Int. Arch. Occup. Environ. Health 63, 69-75. Kawai, T., Misunuma, K., Yasugi, T., Uchida, Y. and Ikeda, M. (1990) The method of choice for the determination of 2,5-hexanedione as an indicator of occupational exposure to n-hexane. Int. Arch. Occup. Environ. Health 63,69-75. Jakubowski, M. and Kostrzewski, P. (1989) Excretion of methylbenzoic acid in urine as results of single and combined exposure to m-xylene. Polish J. Ind. Med. 2, 238-247. Rainsford, S.G. and Lloyd Davies, T.A. (1965) Urinary excretion of phenol by men exposed to vapour of benzene: A screening test. Br. J. Ind. Med. 22,21-26. Japan Association of Industrial Health (1991) Recommended occupational exposure limits. Jpn. J. Ind. Health 33,277-298. Wallen, M., Holm, S. and Nordqvist, M.B. (1978) Experimental human exposure to toluene. II. Int. Arch. Occup. Environ. Health 42, 105-l 17. Jackson, S. (1966) Creatinine in urine as an index of urinary excretion rate. Health Phys. 12, 843-850. Browning, E. (1965) Toxicity and Metabolism of Industrial Solvents. Elsevier, Amsterdam, pp. 522525. Mizunuma, K., Kawai, T., Yasugi, T., Horiguchi, S., Iwami, 0. and Ikeda, M. (1992) In vitro hydrolysis of methyl acetate, a limitation in application of head-space gaschromatography in biological monitoring of exposure. Toxicol. Lett., in press. Heinrich, R. and Angerer, J. (1982) Occupational chronic exposure to organic solvents. Int. Arch. Occup. Environ. Health 50, 341-349. Windholz, M., Budavari, S., Stroumtos, L.Y. and Fertig, M.N. (Eds.) (1976) The Merck Index, 9th Edn., Merck & Co., Rahway, NJ, USA. Ghittori, S., Imbiani, M., Pezzagno, G. and Capodaglio, E. (1987) The urinary concentration of solvents as a biological indicator of exposure: Proposal for biological equivalent exposure limit for nine solvents. Am. Ind. Hyg. Assoc. J. 48, 786790.
333
TOXLET 028 18
Urinalysis vs. blood analysis, as a tool for biological monitoring of solvent exposure*
Toshio Kawaia, Tomojiro Yasugia,b, Kazunori Mizunuma a,b, Shun’ichi Horiguchi” and Masayuki Ikedab “Osaka Occupational Health Service Center, Osaka (Japan)
and bDepartment of Public Health, Kyoto
University Faculty of Medicine, Kyoto (Japan)
(Received 22 June 1992) (Accepted 26 August 1992) Key words: Biological monitoring; Blood analysis; n-Hexane; Toluene; Occupational health; Urinalysis
SUMMARY Blood and urine samples were collected at the end of an 8-h workshift from 30 male workers exposed to a mixture of n-hexane, ethyl acetate and toluene (each being about 2 ppm as geometric means) and also from 20 nonexposed male workers. Blood samples were analyzed for n-hexane and toluene, and urine samples were analyzed for n-hexane, toluene, 2,5-hexanedione (both with and without hydrolysis) and hippuric acid. Based on the correlation between biological exposure indicators and solvent concentrations in air, sensitivity as an exposure indicator was compared between solvents in blood and solvents or metabolites in urine in terms of the lowest solvent concentration at which the exposed subjects can be statistically separated from the nonexposed. Both n-hexane and toluene in blood were sensitive enough to detect the exposure at 6.1 ppm and 1.4 ppm. respectively. n-Hexane exposure below 2 ppm was detectable also by urinalysis for 2,5-hexadione without hydrolysis. Urinary hippuric acid, however, failed to detect low toluene exposure under the conditions studied. Of additional interest is the fact that toluene in urine correlated significantly with toluene in air, which apparently deserves further study for confirmation.
INTRODUCTION
Biological monitoring of exposure is an important tool in occupational health. Among the various biological matrices for analysis, urine has been commonly ana-
Correspondence to: Professor M. Ikeda, Department of Public Health, Kyoto University Faculty of Medicine, Kyoto 606-01, Japan. * A part of this work was presented at the International Symposium on Biological Monitoring, held on 12-16 October 1992, in Kyoto, Japan.
334
Iyzed for its metabolite(s) in occupational health practice to solvent-exposed workers [I ,2], probably because urine samples are readily available under many circumstances. In recent years, however, the specificity of the urinalysis has been questioned especially when the intensity of exposure to solvent is reduced [3,4] due to the modernization of working environment, and attention toward blood analysis for the solvent per se has increased despite the inherent invasive nature of blood sampling, because blood sampling is also a common practice for the prevention of solvent poisoning. For example, the Ordonnance for Prevention of Organic Solvent Poisoning in Japan request hematology and serum biochemistry for health check-up of the workers exposed to selected organic solvents [5]. In the present study trials were made to compare the sensitivity of blood analysis and urinalysis, taking the two most popular solvents of n-hexane and toluene as an example. MATERIALS AND METHODS
Workers studied The survey was conducted in the second haif of a working week. The solventexposed workers (all men) were employed in an adhesive tape-producing plant for at least a year. Nonexposed controls (20 men) were recruited from offices with no known occupational exposure to organic solvents [6]. Determination of time-weighted average (TWA) intensy ofexposure to organic solvents Each of the exposed workers was equipped with a diffusive sampler (with carbon cloth as an adsorbent [7-91) on his chest for an entire 8-h shift. Analytical procedures for solvents adsorbed were as previously detailed [7-91. The additiveness formula [lo] was employed for integrated evaluation of combined solvent exposure. Analysis of blood and urine Workers were asked to pass urine 2-3 h before the end of the workshift. At the end of the shift, each of the workers was invited to a health examination room where no solvent vapor could be detected. Blood was drawn from the cubital vein within 10 min after leaving the workshop. Five ml of the blood was immediately taken into a 20-ml vial (for head-space gaschromatography (HS-CC); Hewlett Packard, Philadelphia, PA, USA together with 0.03 ml of anticoagulant solution (Anglot; Nippon Shoji, Osaka, Japan) and the vial was sealed with a Teflon-coated septum. The blood sampling was followed by the sampling of urine. Namely, immediately after sampling in a clinical urine cup, a 5-ml portion urine was taken to a HS-GC vial as was the case of blood analysis. The rest of the urine sample was kept at -30°C until analyzed for metabolites. HS-GC analysis of blood and urine for n-hexane and toluene was conducted by the method previously described fll]. In addition to the shift-end urine samples described above, 14 out of the 30 exposed workers also offered preshift urine
335
samples the next morning, some 16 h after the termination of the exposure. The preshift urine samples were subjected to 2,5-hexadione determination. After thawing, the stored urine samples were analyzed for hippu~c acid (a toluene metabolite) by the HPLC method of Kawai et al. [12], and 2,5-hexanedione both by the direct method (i.e., without acid hydrolysis [13]) and the hydrolysis method (i.e., with hydrolysis at pH 0.1 to 0.3 in the presence of sulfuric acid [13]), respectively. Creatinine concentration and specific gravity were measured by the conventional calorimetric and refractometric methods, respectively. The solvent and metabolite concentrations were expressed as observed (i.e., without any correction for urine density), or after correction for creatinine concen~ation [14] or a specific gravity of urine of 1.016 [ 151.The data on urine samples from 20 nonexposed men are cited from a previous publication [6].
Regression lines were calculated by the least-square method. Analysis of variance (ANOVA) was employed for the comparison of means. RESULTS
Z~te~s~ty of exposure to organic solvents The measurements of TWA solvent concentrations in breathing zone air analysis showed that vapors of three solvents, i.e., Pt-hexane, ethyl acetate and toluene coexisted in the air. Thus, the vapor concentrations were expressed both individually and as the sum following an additiveness assumption [IO], taking 40,400 and 100 ppm as the occupational exposure limit @EL) for n-hexane, ethyl acetate and toluene [16]. The results are presented in Table I in terms of AM + ASD, GM (GSD) and the maximum value observed. It is clear from the table that the exposure of the workers
TABLE I INTENSITY OF EXPOSURE TO ORGANIC SOLVENTS Parameter
AM” ? ASDd GM” (GSD)’ Maximum
Solvent concentration n-hexane
EA”
toluene
sumb
5.8 t 7.7 2.2 (4.68) 35.1
5.2 t 7.7 2.1 (3.90) 30.0
3.1 i 2.8 2.1 (2.71) 12.0
0.19 rt 0.22 0.09 (3.83) 0.95
a Ethyl acetate. b Summation after an additiveness formula [lo), taking 40,400 and 100 ppm as occupational exposure limits for n-hexane, ethyl acetate and toluene, respectively 1161.’ Arithmetic mean (unit, ppm). d A~thmetic standard deviation (unit, ppm). a Geometric mean (unit, ppm). t Geometric standard deviation (~~ensionless).
336
was generally low, and below the corresponding OEL when evaluated individually, or even in combination (i.e., the sum in the worst case was 0.95). Correlation of solvent in blood and urine with the solvent in air The possible correlations were examined between the solvent concentration in the blood samples collected at the end of a workshift and the TWA concentration of the solvent in the breathing zone air during the shift. Similar examination was also carried out between the solvent concentration in the shift-end urine and the solvent in the air. The combined regression analyses with 30 exposed and 20 nonexposed subjects (Table II) showed that toluene in blood correlated significantly (P < 0.01) with toluene in air, and that such was also the case (r was approx. 0.7 with P < 0.01) for toluene in urine regardless of correction for urine density. Scatter diagrams between toluene in air and that in urine are depicted in Figure 1 for better understanding. n-Hexane in blood also correlated with n-hexane in air with a correlation coefficient of 0.639. n-Hexane in urine, however, failed to correlate with n-hexane in air, with a low correlation coefficient of well below 0.3, regardless of correction for urine density. Correlation of urinary metabolite with the solvent in air The correlation of two urinary metabolites, 2,5-hexanedione TABLE
II
CORRELATION
OF SOLVENT
Solvent in air (ppm)
n-Hexane
a a andp
from n-hexane and
IN BLOOD
OR URINE
Solvent
Correction
in blood @g/l)
for urine
or urine @g/l)
density
n-hexane
in blood
n-hexane n-hexane
in urine in urine
n-hexane
in urine
toluene toluene
in blood in urine
toluene
in urine
toluene
in urine
are the slope and the intercept
WITH
SOLVENT
IN AIR
Parameters
of regression*
a
B
r
2.99
15.15
1.15
16.45
0.639** 0.224
1.21 1.04
18.55 14.18
6.14 2.01
3.55 4.14
0.782**
none creatinine spec. gravity
3.45 2.44
3.28 3.05
0.696** 0.723**
none creatinine spec. gravity
on y-axis of a regression
(1.016)
(1.016)
0.197 0.229
0.650**
line. y = a.y + p. where x is time-weighted
average solvent concentration in breathing zone air and y is that in blood or urine collected at the end of the workshift. r is the correlation coefficient (**P < 0.01). The unit is ppm for the solvent in air, ,&l for solvent in blood or urine (as corrected for none or for a specific gravity of urine of 1.016) or PCs/g creatinine when corrected for urinary creatinine concentration. Data from 30 exposed and 20 nonexposed cases are subjected to regression analysis.
331
0
5
10
TOLUENE
IN AIR
15
20
(ppm)
Fig. 1. Correlation between toluene in air and toluene in urine. Time-weighted average toluene concentration in breathing zone air during an 8-h shift was measured by diffusive sampling. Urine samples were collected at the end of the shift, and analyzed for toluene. Urinary toluene concentrations are expressed as observed, or after correction for creatinine concentration or a specific gravity of urine of 1,016. The line in the middle is a regression line, two curves on both side of the line show the 95% confidence range of the group means, and the outer-most curves show the 95% confidence range for individual values. Each circle represents one case. Cases on the vertical axis are not shown to avoid congestion.
hippuric acid from toluene, with the levels of corresponding mother solvent in air was examined, and the results are summarized in Table III. Because it is known that the urinary level of 2,5-hexadione (HD) varies subject to the pretreatment with acid hydrolysis [13], the metabolite level was examined under two analytical conditions, i.e., one with the direct method without hydrolysis (HD/sHYD) and the other after hydrolysis at pH < 0.3 (HD/cHYD). Both HD/sHYD and HD/cHYD (i.e., HD regardless of hydrolysis pretreatment) correlated significantly (P < 0.01) with n-hexane in air (Table III) with a correlation coefficient of approx. 0.7 (< 0.6 for HD/cHYD when not corrected for urine density). In sharp contrast, no significant correlation
338
(P > 0.10) could be detected between toluene in air and hippuric in urine, independent of correction for urine density. Biological half-time of n-hexane in man
Urine samples were available from 14 workers both at the end of the shift of the day and immediately before the shift the next day, with a lapse of 16 h between sampling. Accordingly, trials were made to estimate the biological half-time of n-hexane in terms of disappearance of HD from the urine. In practice, such estimation, although crude, is considered possible only with HD/sHYD, because the background level was zero for HD/sHYD, whereas it was not negligible in the case of HD/cHYD [6]. The calculation gave a half-time value of 16.6 ? 12.9 h (AM -t ASD of 14 determinations). Further classification of the 14 subjects according to exposure intensity into >5 ppm, 5 to 10 ppm, and ~10 ppm groups showed that AM+ASD of the half-time was 11.9 + 8.5 h (n = 4), 18.3 & 15.3 h (n = 6) and 19.0 ? 11.3 h (n = 4), respectively. No significant (P > 0.05) difference among the means was detected by ANOVA. DISCUSSION
It is assumed in the present study that the metabolism of two solvents, n-hexane and toluene, are independent of each other; this assumption should be taken as quite
TABLE
III
CORRELATION
OF METABOLITE
Solvent in air (ppm)
Metabolite
IN URINE in urine
WITH
SOLVENT
Correction
for urine
density
IN AIR Parameters a
n-Hexane
n-Hexane
HD/sHYDb
none044
HD/sHYD HD/sHYD
creat.c &g/g treat.) spec. gravityd
@g/I)
I
B 6.9 1.9
1.5
0.696**
3.4
10.9
3.0
0.677** 0.646**
HD/cHYD’
noneWU
36.2
431.0
0.581**
HD/cHYD
treat. t&g/g treat.) spec. gravity @g/l)
39.8
316.8 385.4
0.734**
HBicHYD Toluene
of regression”
(unit)
hippuric
acid
hippuric
acid
none (mg/l) treat. (mg/g treat.)
hippuric
acid
spec. gravity
(mg/l)
50.1 -0.1 16.4
233.2 188.0
5.9
188.4
0.714** -0.018 0.274 0.093
a a, B, r and statistical significance of r are as in Table II. x is time-weighted average solvent concentration in respiratory zone air and y is metabolite in shift-end urine. Data from 30 exposed and 20 nonexposed of 2,5-hexanedione without hydrolysis. ’ Corcases are subjected to regression analysis. b Concentration rected for creatinine concentration. d Corrected for a specific gravity of urine of 1.016. e Concentration of 2,5-hexanedione after acid hydrolysis.
339
LSC-8
(B)
D /
t
0
n-HEXANE IN AIR (ppm)
Fig. 2. Graphical determination of LSC (lowest separation concentration) for n-hexane by analysis of blood and urine. (A) Shows the results of blood analysis; (B) the results of urinalysis. Conditions are as in Fig. 1, except for the outer-most curves which are not shown. Cases on the axis or out of the scales are not depicted. For definition of LX-B, LX-Ul, and LSCU2, see text.
probable especially when the exposure is low [6,17,18]. Co-presence of the third solvent, ethyl acetate, is virtually ignored because it is well known that this ester is readily hydrolyzed when absorbed [19,20] and decomposed to carbon dioxide and water. Based on this assumption, it is possible to compare the sensitivity of two approaches of biological monitoring, blood analysis and urinalysis, from the view point of the power to separate the exposed from the nonexposed subjects. For this purpose, a parameter of the lowest separation concentration (or LSC, in terms of solvent concentration in air, e.g., in ppm) was graphically calculated as the concentration at which the 95% lower limit of the distribution is equal to the 95% upper limit at 0 ppm. LSC was further specified as LSC-B and LSC-U when applied to blood analysis or urinalysis (B and U standing for blood or urine), respectively. In practice, LSC-U can be determined in two ways, one (LSC-Ul) just as in the case of LSC-B, and the other in
340
comparison with the upper 95% limit of the background level because certain metabolites, e.g., hippuric acid, are known to be present even in the urine from the nonexposed subjects (e.g., Ref. 6). Examples of graphical determination of LSC for n-hexane are shown in Figure 2 taking the observed, i.e., noncorrected values of 2,5-hexanedione after acid hydrolysis, as n-hexane metabolite concentrations, and that for toluene in Figure 3. Whereas the three LSCs can be determined in the former case, only LSC-B is available for toluene, because under the study conditions urinary hippuric acid correlates poorly with toluene in air with essentially no dose-dependent increase (Table III). This observation is in agreement with the findings in previous reports that hippuric acid only poorly correlated with toluene exposure among workers exposed to toluene at 26 ppm as a median [3], and that no correlation was observed in the daily variation in toluene exposure intensity and that of exposure-induced increments in hippuric acid concentrations among workers exposed to less than 10 ppm toluene [4]; the observed lack of correlation is apparently due to the rather high background level of hippuric acid (e.g., see Ref. 6).
TOLUENE IN AIR (Pam)
Fig. 3. Graphical determination of LSC and urine. (A) Shows the results of blood of Fig. 2. Note that both LSC-Ul and dependent increase
(lowest separation concentration) for toluene by analysis of blood analysis; (B) the results of urinalysis. Remarks are as in the legend LSC-U2 cannot be determined in B due to the lack of a dosein hippuric acid coupled with wide variation.
341
The results of LX dete~ination are s~marized in Table IV. It is evident from the values given in the table that LSC-B is essentially comparable to LX-UI in the case of n-hexane exposure, and smaller than LSC-U2 when urinary 2,5-hexanedione after acid hydrolysis (i.e., HD/cHYD in Table IV) is evaluated. In the case of toluene, LSC-B is as small as 1.4 ppm in contrast to the fact that no LSC-Ul or LSC-U2 (both based on urinary metabolite levels) is available due to the lack of correlation. Such superiority of blood analysis over urinalysis is a good confirmation, with a separate group of workers, of the previous observation [6], and in agreement with the opinion of Heinrich and Angerer [21] who studied methanol in blood and urine of methanolexposed workers and concluded that blood methanol is a better indicator of methanol exposure than urinary methanol. In practice, however, the invasive nature coupled with the longer biological half-time for the urinary metabolite (e.g., well over 10 h for urinary 2,5-hexanedione as observed in the present study) than for the mother solvent per se in blood is an disadvantage for blood analysis. Of particular interest is the fact that toluene in the shift-end urine correlates significantly (P < 0.01) with the TWA toluene exposure during the shift, in contrast to the fact that urinary n-hexane failed to do so (Table II). This might be a reflection of the TABLE IV LOWEST SEPARATION CONCENTRATION Analysis LSC”
Blood analysis LSC-B for solvent Urinalysis LX-U for solvent without correction corrected for creatinine corrected for spec. gravity (1.016) LSC-Ul for metabolite without correction corrected for creatinine corrected for spec. gravity (1.016) LSC-U2 for metabolite without correction corrected for creatinine corrected for spec. gravity (1.016)
(LSC) BY BLOOD ANALYSIS AND URINALYSIS Solvent n-hexane
toluene
6.1
1.4
NCb NC NC
1.9 3.0 1.8
HDlsHYD’, 5.1: HDlcHYD”, 7.4 HDlsHYD, 4.5: HDlcHYD, 4.2 HDlsHYD, 5.2: HDlcHYD, 4.5
NC NC NC
HD/sHYD, 1.1: HD/cHYD, 31.4 HD/sHYD, 1.9: HD/cHYD, 1I .7 HDlsHYD, 2.0: HDlcHYD, 10.0
NC NC NC
Values are for solvent concentration in ppm. a The lowest concentration at which the exposed are statistically separated from the nonexposed. LSC-B is by blood analysis, and LX-U is by urinalysis. LX-U1 is by comparison with the 95% upper limit at 0 ppm, and LSC-U2 is by comparison with the 95% upper limit of the background level. Also see Figs. 1 and 2. b Not calculable due to lack of correlation. ’ 2,5Hexanedione without hydrolysis. d 2,5-Hexanedione after acid hydrolysis.
342
difference in physical characteristics of the two solvents; toluene is described as being very slightly soluble, whereas n-hexane is insoluble in water, and n-hexane is more volatile (boiling point; 63°C) than toluene (llO.O’C) [22]. Little attention has been paid to toluene in urine as an exposure indicator [ 121. Nevertheless, Ghittori et al. 1233observed a close and significant correlation between TWA intensity of exposure to toluene and toluene concentration in shift-end urine with a correlation coefficient of 0.87 (P < 0.01) among toluene-exposed workers; urinary toluene at 100 ppm toluene exposure was estimated to be 225 pg.4, which is essentially the same as the present results (205 ,@I, see Table II). Although the present observation is not yet conclusive and should be taken as preliminary, the possible value of urinary toluene as a tool of toluene-exposure monitoring deserves further attention for conflation, when high sensitivity as an exposure indicator (LSC-U; about 2 ppm; Table IV), high practicality of FIG-GC, and noninvasive nature of urine collection are in combination taken into consideration. ACKNOWLEDGEMENTS
The authors are grateful to Professor M. Tati and Dr. S. Koshi of Japan Industrial Safety and Health Association, Tokyo, for their support in this work. REFERENCES 1 Lauwerys, R. (1983) Toluene. In: L. Ale&o, A. Berlin, R. Roi and M. Boni (Eds.), Human Biological Monitoring ofIndustrial ChemicaIs Series. Joint Research Centre Ispra Establishment, Ispra, Italy, pp. 179-17s. 2 Lauwerys, R. and Buchet, J.P. (1988) Bioiogical monitoring of exposure to benzene, toluene and xyiene. IARC Sci. Publ. 85,2(X-222. 3 Nise, G. (1992) Urinary excretion of a-cresol and hippuric acid after toluene exposure in rotogravure printing. Int. Arch. Occup. Environ. Health 63, 377.-381. 4 Johnai, H. and Sato, M. (1988) Exposure indices for painters exposed to toluene and xylene at low concentrations. Ind. Health 26, 197-202. 5 Ministry of Labor, the Government of Japan (1989) Ordonnance for the Prevention of Organic Solvent Poisoning, Ministry of Finance, Printing Bureau, Tokyo. 6 Kawai, T., Yasugi, T., Mizunuma, K., Horiguchi, S., Iguchi, H., Uchida, Y., Twani, 0. and Ikeda, M. (1992b) Comparative evaluation of urinalysis and blood analysis as a tool to detect exposure to organic solvents at low concentrations. Int. Arch. Occup. Environ. Health, in press. 7 Hirayama, T. and Ikeda, M. (1979) Applicability of carbon felt to the dosimetry of solvent vapor mixture. Am. Ind. Hyg. Assoc. J. 40, 1091-1096. 8 Ikeda, M., Kumai, M. and Aksoy, M. (1984) Application of carbon felt dosimetry to field studies distant from analyti~l laboratory. Ind. Her&h 2253-58. 9 Kasahara, M. and Ikeda, M. (1987) Spontaneous desorption of organic solvents from carbon cloth. Ind. Health 25, 73-81. 10 American Conference of Governmental Industrial Hygienists (1991) Threshold Limit Values and Biological Exposure Indices. ACGIH, Cincinnati, USA. 11 Kawai, T., Yasugi, T., Mizunuma, K., Horiguchi, S., Morioka, K., Miyashita, Y., Uchida, Y. and Ikeda, M. (1992a) Monitoring of workers exposed to a mixture of toluene, styrene and methanol
343
12
13
14 15 16 17 18 19 20
21 22 23
vapours by means of diffusive air sampling, blood analysis and urinalysis. Int. Arch. Occup. Environ. Health 63,429-435. Kawai, T., Mizunuma, K., Yasugi, T., Horiguchi, S., Uchida, Y., Iwami, O., Iguchi, H. and Ikeda, M. (1991) Urinary methylhippuric acid isomer levels after occupational exposure to xylene mixture. Int. Arch. Occup. Environ. Health 63, 69-75. Kawai, T., Misunuma, K., Yasugi, T., Uchida, Y. and Ikeda, M. (1990) The method of choice for the determination of 2,5-hexanedione as an indicator of occupational exposure to n-hexane. Int. Arch. Occup. Environ. Health 63,69-75. Jakubowski, M. and Kostrzewski, P. (1989) Excretion of methylbenzoic acid in urine as results of single and combined exposure to m-xylene. Polish J. Ind. Med. 2, 238-247. Rainsford, S.G. and Lloyd Davies, T.A. (1965) Urinary excretion of phenol by men exposed to vapour of benzene: A screening test. Br. J. Ind. Med. 22,21-26. Japan Association of Industrial Health (1991) Recommended occupational exposure limits. Jpn. J. Ind. Health 33,277-298. Wallen, M., Holm, S. and Nordqvist, M.B. (1978) Experimental human exposure to toluene. II. Int. Arch. Occup. Environ. Health 42, 105-l 17. Jackson, S. (1966) Creatinine in urine as an index of urinary excretion rate. Health Phys. 12, 843-850. Browning, E. (1965) Toxicity and Metabolism of Industrial Solvents. Elsevier, Amsterdam, pp. 522525. Mizunuma, K., Kawai, T., Yasugi, T., Horiguchi, S., Iwami, 0. and Ikeda, M. (1992) In vitro hydrolysis of methyl acetate, a limitation in application of head-space gaschromatography in biological monitoring of exposure. Toxicol. Lett., in press. Heinrich, R. and Angerer, J. (1982) Occupational chronic exposure to organic solvents. Int. Arch. Occup. Environ. Health 50, 341-349. Windholz, M., Budavari, S., Stroumtos, L.Y. and Fertig, M.N. (Eds.) (1976) The Merck Index, 9th Edn., Merck & Co., Rahway, NJ, USA. Ghittori, S., Imbiani, M., Pezzagno, G. and Capodaglio, E. (1987) The urinary concentration of solvents as a biological indicator of exposure: Proposal for biological equivalent exposure limit for nine solvents. Am. Ind. Hyg. Assoc. J. 48, 786790.
