Int Arch Occup Environ Hlth 35,193-200 (1975) © by Springer-Verlag 1975
An Automated Method for the Determination of Total Trichloro-Compounds in Human Urine ISAMU KANEKO', TSUYOSHI MIURA 2, TATSURO HIRAO2, and MASAYUKI IKEDA' 1 Department of Hygiene, Tohoku University School of Medicine, Sendai, and Nihon Technicon K K , Minato-ku, Tokyo Received February 13, 1975 / Accepted March 27, 1975 Summary
An automated method is developed for the determination of total
trichloro-compounds in human urine
Based on the oxidative conversion of total trichloro-compounds to trichloroacetic acid and colorimetric deter-
mination with Fujiwara reaction, the method is highly reproducible at the performance rate of 20 determinations per hr The results obtained by the automated method well agree with those by the manual method. Key words: Autoanalysis Solvents
Biological Monitoring
Chlorinated Hydrocarbon
Total Trichloro-Compounds.
INTRODUCTION
Significance of total trichloro-compounds determination in urine has been well established (Lehnert et al , 1974) as an index of exposure to several chlorinated solvents (Ikeda & Ohtsuji, 1972 a) including trichloroethylene, tetrachloroethylene (Ikeda et al , 1972 b; Ikeda & Imamura, 1973) as well as methylchloroform (Seki et al , 1974) The analytical procedures, originally time-consuming (Seto & Schultze, 1956 ; Tanaka & Ikeda, 1968), have been accordingly further improved to reduce the time necessary for the analysis (Imamura & Ikeda, 1973) Complexity of the manual procedures yet forms a bottleneck in practice and limits a wider application of biological monitoring to the prevention of possible adverse effects on the health of the workers It would be much more commodious if the entire analysis could be carried out in an automated system, dealing with more than 100 samples a day without requiring many human hands It is the purpose of the present study to develop a time and hand-saving automated system of urinalysis for total trichloro-compounds and trichloroacetic acid with the accuracy at least comparable to that of the manual analysis.
193
Introduction of the Auto Analyzer system is rather scarce in the field of industrial hygiene in relation to biological monitoring study except for the pioneer work of de Bary (1969) and Heistand & Todd (1972), who developed an automated method to determine phenol in urine and have been applying the method to the work-load control of workers.
MATERIALS AND METHODS Trichloroethanol was purchased from K & K Laboratories, Inc , Plainview, N Y , USA The purity was calculated as previously described (Tanaka & Ikeda, 1968) In the automated system, pyridine was diluted with water (pyridine: water = 2 : 1, v/v; prepared daily) and 1ON KOH was employed in the place of 7 8N solution The reasons will be given later Other reagents were the same as those of the manual system (Imamura & Ikeda, 1973). Urine samples were collected from factory workers who had been exposed to trichloroethylene The urine samples free of the metabolites (control urine) were obtained from those who had no known exposure to the chlorinated solvents (Ikeda & Ohtsuji, 1969). The equipment consists of the following Technicon Auto Analyzer components: a) a model IV sampler operating at 20 samples per hr with a 2:1 sample-wash ratio; b) a proportioning pump III provided with an air bar; c) 2 small heating baths with 1 min 50 sec incubation coil, one to be used at 960 for preheating and the other at 850 for colour development; d) a 80-feet heating bath for oxidation at 960 for 15 min; e) a colorimeter fitted with flow-cells of a 15 mm light path and the interference filters of a peak light transmission at a wavelength of 550 nm; and f) a standard chart recorder The principle employed in the automated analysis is based on what was developed in the manual analysis (Tanaka & Ikeda, 1968) Trichloroacetic acid (TCA) was colorimetrically measured by the Fujiwara reaction Total trichloro-compounds (TTC) was measured after chemical oxidation of trichloroethanol (TCE; either in the free form or as the glucuronide, i e. urochloralic acid) in the presence of chromium trioxide-nitric acid mixture The difference between the 2 results was attributed to TCE. The flow scheme of the automated method for TTC determiThe numbered fittings are standard nation is shown in Fig 1 Auto Analyzer parts The urine sample pumped is mixed with the oxidizing reagent and preheated at 960 The gas generated, together with air, is debubbled and the flow is subject to re-
194
Flow rates (ml/min)
RECORDER COLORIMETER 15mm tubular flow cells 550 nm filters
PROPORTIONING PUMP m
Fig 1 Flow diagram for the automated determination of total trichloro-compounds. Numbered fittings are standard Technicon AutoAnalyzer parts coil figures indicate turns of coils
Numbers over
(P) stands for a pulse suppressor;
*=acidflex pump tube; **=silicone pump tube (inner diameter in inch); ***= Kel-F tube; ****=polyethylene tube; +=oxidant consists of 8g of chromium trioxide, 5 ml of water and 15 ml of concentrated (about 61%) nitric acid; *=pyridine-water mixture (2 : 1, v/v)
sampling before oxidation in a 80-feet heating bath at 960 for 15 min The digest is again debubbled, resampled and alkalified with 10 ON KOH The alkaline solution is mixed with pyridine-water (2 : 1, v/v) and heated at 850 for Fujiwara
reaction
The coloured pyridine layer is separated and clari-
fied by the addition of water, of which the extinction 550 nm is measured and recorded determined by comparing
The values
the observed extinction
to a cali-
bration curve prepared with TCA working standards Ikeda,
at
for samples are (Tanaka &
1968).
In order to measure TCA only, the tube from the sampler connected with the pump tube B in the proportioning pump, omitting the oxidation process (Fig 11). For the manual analysis, 15 min oxidation water bath coupled with 50 min employed
(Imamura & Ikeda,
is
in a boiling
colour development
at 650 was
1973).
195
RESULTS AND DISCUSSION
Development of the Automated Method Some urine sample is rich in salts, bubbles in the transfer tube when mixed with the strongly acid oxidant, and thus disturbs the constant alternating flow of air and solution. This annoyance could be eliminated by preheating after mixing with the oxidant The gas, probably carbon dioxide, is removed by a debubbler, and the sample flow is subject to resampling This preheating inevitably requires a thermo-resistant Kel-F tube connection Because of bubbling problems, control urine (i e , urine obtained from non-exposed subjects) is better than water as washing liquid in order to stabilize the system. The sample, 0 8 ml/min, is mixed with 0 58 ml/min of the oxidant, and the flow for oxidation through a heating bath at 960 takes about 15 min This oxidation condition was considered less intensive than that of the manual analysis in which the sample-oxidant mixture (1:1) is subject to heating in a boiling water bath for 15 min The experimental results to be described later, however, indicate that the authentic TCE added is quantitatively converted to TCA under the conditions employed Furthermore, no improvement in the oxidation rate was observed when 2 heating baths were connected in series to double the oxidation duration Heating condition for colour development, about 1 min and 50 sec at 850, is not critical in the AutoAnalyzer system. Preliminary experiments with the manual analysis revealed that the colour intensity in 5 ml of the pyridine layer after Fujiwara reaction is independent to the amount of water layer in the range of 3 5 to 7 5 ml as far as alkali concentration is constant, that the colour intensity is reduced when the alkali concentration is less than 7 8 N, and that precipitation after KOH addition (presumably potassium chromate and potassium nitrate) is prevented when the amount of water layer is 7.5 ml or more under the conditions studied Based on these observations, alkali and water of more amounts than those of the manual analysis are pumped into the system so that no precipitation will take place during the reaction. Pyridine is a potent solvent and a precaution is required to pump it in Only a silicone tube can tolerate the heavy duty of the pumping Further connection should be with glass tubes using short pieces of silicone tubes as connecting joints. Merging of pure pyridine with salt-rich aqueous phase sometimes caused irreversible precipitation inside the T-shaped
196
connector, which hindered the constant flow Replacement with pyridine-water (2 : 1, v/v) clearly prevented the precipitation A mixture with lower water content (e g , pyridine : water = 3 : 1) was less effective. Extinction at 550 nm is about 83% of that at 530 nm, the wavelength with the maximum light absorption This disadvantage, however, is inevitable as interference filters are employed in the colorimeter, while a flow-cell with a longer light path, 15 mm, may serve to increase sensitivity The total flow-time from the sampler to the flow-cells is about 28 min. The inner capacity of the phase separator (B2 in Fig 1), employed to pick up pyridine phase selectively, appears to be rather large and may cause an undesirable mixing of colourdeveloped solution, resulting in a disturbance of a response. Improvement at this step will also help to achieve a higher sampling rate. Linearity of Colour Development in the Automated Method In order to examine if the intensity of the colour developed is linear to the TCA concentration, authentic TCA was dissolved up to the concentration of 96 mg/l in 4 control urine samples and the solutions were subject to the automated analysis The results are summarized in Fig 2 It is clear that the intensity of the colour developed is independent to the urine samples studied and proportional to the concentrations of TCA The same results were reproduced when TCE, up to 79 mg/l, was added to the control urine (Fig 2) Comparison between the 2 results with TCA and TCE on equi-molar basis revealed that colour intensity of TCE added is 101 7 ± 2 0 % (mean ±SD; calculated from the gradients of 4 regression lines) of the theoretical value, indicating that the oxidation condition is satisfactory for the practical use. Precision of the Automated Method Four determinations each were performed with the urine samples containing either 32 mg/l or 144 mg/l of authentic TCA to investigate the precision of the colour development process The coefficient of variation was 1 9% in the former case and it was 0 9% in the latter When 4 determinations each were carried out with the urine samples containing either 26 mg/l or 117 mg/l of authentic TCE in order to examine the precision of over-all procedure of the method, the results as shown in Fig 3 gave the coefficient of variations of 1 6% and 0 8%, respectively. 197
0.5
-0.4
E
0
Z
0.2
.E
I
0
20
40
60
80
100
Concentration (mg / I) of TCA and TCE
Fig 2
Development of colour lineary related to the amounts of TCA and TCE TCA and TCE were dissolved in the urine samples from 4 non-exposed individuals as shown by circles, squares, triangles and rhombs Open and solid symbols stand for TCA and TCE, respectively TCE are expressed as equimolar amounts of TCA Regression lines, Z = 0 00446 x (broken one) and z = O 00454 y (solid one), where x and y are concentrations of TCA and TCE, respectively, and Z is extinction measured, are calculated from the means of 4 determinations at each concentration
c C o O
o o
c
w
0
5
10 Time
15 ( min
20
25
30
)
Fig 3 Chart records of over-all determinations Four determinations each were carried out with the urine samples containing either (A) 26 mg/l or (B) 117 mg/l of TCE
198
Fig 4 Correlation between the results from the automated analysis and the manual analysis
Urine
samples, 54 in total, were collected from the factory workers exposed to trichloroethylene, and analyzed for TTC by both automated and manual methods. The line in the figure, y = 1.005 x + 0 480, is a calculated regression line where y is the value from the automated analysis and x is that from the manual analysis
The correlation
coefficient, r, is O 989 TTC (mg / I)by manual analysis
Correlation between the Results from the Manual and Automated Analysis Urine samples from the exposed subjects, 54 specimens in total, were analyzed for TTC by means of both manual and automated analyses, the results being summarized in a scatter diagram A close correlation exists between the 2 results as shown in Fig 4; the regression line is y = 1 005 x + 0 480, where
y : TTC mg/l as measured by the automated analysis, x
TTC mg/l as measured by the manual analysis,
and the correlation coefficient, r, is 0 989 (54 determinations) The results clearly indicate that an automated method at the sampling rate of 20 specimens per hr is established with the accuracy comparable to that of the time-consuming manual method.
Acknowledgements
Thanks are due to Dr Tokunaga of Kyoto Industrial Hygiene Association, Kyoto, Japan, Prof K Matsumoto of Nagasaki University Medical School, Nagasaki, Japan, and Dr N Ishihara of our depart-
ment for their collaboration in collecting urine samples.
199
REFERENCES deBary, J L J : An automated method of determining phenol in urine as an index of benzene exposure Proceedings of Technicon England Symposium, pp.13 5 -1 3
7
(1969)
Heistand, R N , Todd, A S : Automated determination of total phenol in urine Amer industr Hyg Ass J 33, 378-381 (1972) Ikeda, M , Imamura, T : Biological half-life of trichloroethylene and 31, Int Arch Arbeitsmed tetrachloroethylene in human subjects 209-224 (1973) Ikeda, M , Ohtsuji, H : Hippuric acid, phenol and trichloroacetic acid levels in the urine of Japanese subjects with no known exposure to 26, 162-164 (1969) organic solvents Brit J industr Med Ikeda, M , Ohtsuji, H : A comparative study of the excretion of Fujiwara reaction-positive substances in urine of humans and rodents given Brit J. trichloro or tetrachloro-derivatives of ethane and ethylene industr Med
29,
99-104 (1972 a)
Ikeda, M , Ohtsuji, H , Imamura, T , Komoike, Y : Urinary excretion of total trichloro-compounds, trichloroethanol and trichloroacetic acid as a measure of exposure to trichloroethylene and tetrachloroethylene. Brit J industr Med 29, 328-333 (1972 b) Imamura, T , Ikeda, M : A time-saving procedure for the determination Int Arch Arbeitsof total trichloro-compounds in human urine samples 31, 333-338 (1973) med Lehnert, G , Morgan, A , Szadkowski, D , Zielhuis, R L : Halogenated hydrocarbon solvents: Long term effects and biological sampling in human beings
Int Arch Arbeitsmed
33, 251-255
(1974)
Seki, Y , Urashima, Y , Aikawa, H , Matsumura, H , Ichikawa, Y , Hiratsuka, F , Yoshioka, Y , Shimbo, S , Ikeda, M : Trichloro-compounds in the urine of humans exposed to methyl chloroform at sub-threshold levels
Int Arch Arbeitsmed
34, 39-49 (1974)
Seto, T A , Schultze, M O : Determination of trichloroethylene, tri28, chloroacetic acid, and trichloroethanol in urine Analyt Chem 1625-1629 (1956) Tanaka, S , Ikeda, M : A method for determination of trichloroethanol and trichloroacetic acid in urine Brit J industr Med 25, 214-219 (1968)
Prof
Masayuki Ikeda
Department of Hygiene Tohoku University School of Medicine Sendai 980, Japan
200
An Automated Method for the Determination of Total Trichloro-Compounds in Human Urine ISAMU KANEKO', TSUYOSHI MIURA 2, TATSURO HIRAO2, and MASAYUKI IKEDA' 1 Department of Hygiene, Tohoku University School of Medicine, Sendai, and Nihon Technicon K K , Minato-ku, Tokyo Received February 13, 1975 / Accepted March 27, 1975 Summary
An automated method is developed for the determination of total
trichloro-compounds in human urine
Based on the oxidative conversion of total trichloro-compounds to trichloroacetic acid and colorimetric deter-
mination with Fujiwara reaction, the method is highly reproducible at the performance rate of 20 determinations per hr The results obtained by the automated method well agree with those by the manual method. Key words: Autoanalysis Solvents
Biological Monitoring
Chlorinated Hydrocarbon
Total Trichloro-Compounds.
INTRODUCTION
Significance of total trichloro-compounds determination in urine has been well established (Lehnert et al , 1974) as an index of exposure to several chlorinated solvents (Ikeda & Ohtsuji, 1972 a) including trichloroethylene, tetrachloroethylene (Ikeda et al , 1972 b; Ikeda & Imamura, 1973) as well as methylchloroform (Seki et al , 1974) The analytical procedures, originally time-consuming (Seto & Schultze, 1956 ; Tanaka & Ikeda, 1968), have been accordingly further improved to reduce the time necessary for the analysis (Imamura & Ikeda, 1973) Complexity of the manual procedures yet forms a bottleneck in practice and limits a wider application of biological monitoring to the prevention of possible adverse effects on the health of the workers It would be much more commodious if the entire analysis could be carried out in an automated system, dealing with more than 100 samples a day without requiring many human hands It is the purpose of the present study to develop a time and hand-saving automated system of urinalysis for total trichloro-compounds and trichloroacetic acid with the accuracy at least comparable to that of the manual analysis.
193
Introduction of the Auto Analyzer system is rather scarce in the field of industrial hygiene in relation to biological monitoring study except for the pioneer work of de Bary (1969) and Heistand & Todd (1972), who developed an automated method to determine phenol in urine and have been applying the method to the work-load control of workers.
MATERIALS AND METHODS Trichloroethanol was purchased from K & K Laboratories, Inc , Plainview, N Y , USA The purity was calculated as previously described (Tanaka & Ikeda, 1968) In the automated system, pyridine was diluted with water (pyridine: water = 2 : 1, v/v; prepared daily) and 1ON KOH was employed in the place of 7 8N solution The reasons will be given later Other reagents were the same as those of the manual system (Imamura & Ikeda, 1973). Urine samples were collected from factory workers who had been exposed to trichloroethylene The urine samples free of the metabolites (control urine) were obtained from those who had no known exposure to the chlorinated solvents (Ikeda & Ohtsuji, 1969). The equipment consists of the following Technicon Auto Analyzer components: a) a model IV sampler operating at 20 samples per hr with a 2:1 sample-wash ratio; b) a proportioning pump III provided with an air bar; c) 2 small heating baths with 1 min 50 sec incubation coil, one to be used at 960 for preheating and the other at 850 for colour development; d) a 80-feet heating bath for oxidation at 960 for 15 min; e) a colorimeter fitted with flow-cells of a 15 mm light path and the interference filters of a peak light transmission at a wavelength of 550 nm; and f) a standard chart recorder The principle employed in the automated analysis is based on what was developed in the manual analysis (Tanaka & Ikeda, 1968) Trichloroacetic acid (TCA) was colorimetrically measured by the Fujiwara reaction Total trichloro-compounds (TTC) was measured after chemical oxidation of trichloroethanol (TCE; either in the free form or as the glucuronide, i e. urochloralic acid) in the presence of chromium trioxide-nitric acid mixture The difference between the 2 results was attributed to TCE. The flow scheme of the automated method for TTC determiThe numbered fittings are standard nation is shown in Fig 1 Auto Analyzer parts The urine sample pumped is mixed with the oxidizing reagent and preheated at 960 The gas generated, together with air, is debubbled and the flow is subject to re-
194
Flow rates (ml/min)
RECORDER COLORIMETER 15mm tubular flow cells 550 nm filters
PROPORTIONING PUMP m
Fig 1 Flow diagram for the automated determination of total trichloro-compounds. Numbered fittings are standard Technicon AutoAnalyzer parts coil figures indicate turns of coils
Numbers over
(P) stands for a pulse suppressor;
*=acidflex pump tube; **=silicone pump tube (inner diameter in inch); ***= Kel-F tube; ****=polyethylene tube; +=oxidant consists of 8g of chromium trioxide, 5 ml of water and 15 ml of concentrated (about 61%) nitric acid; *=pyridine-water mixture (2 : 1, v/v)
sampling before oxidation in a 80-feet heating bath at 960 for 15 min The digest is again debubbled, resampled and alkalified with 10 ON KOH The alkaline solution is mixed with pyridine-water (2 : 1, v/v) and heated at 850 for Fujiwara
reaction
The coloured pyridine layer is separated and clari-
fied by the addition of water, of which the extinction 550 nm is measured and recorded determined by comparing
The values
the observed extinction
to a cali-
bration curve prepared with TCA working standards Ikeda,
at
for samples are (Tanaka &
1968).
In order to measure TCA only, the tube from the sampler connected with the pump tube B in the proportioning pump, omitting the oxidation process (Fig 11). For the manual analysis, 15 min oxidation water bath coupled with 50 min employed
(Imamura & Ikeda,
is
in a boiling
colour development
at 650 was
1973).
195
RESULTS AND DISCUSSION
Development of the Automated Method Some urine sample is rich in salts, bubbles in the transfer tube when mixed with the strongly acid oxidant, and thus disturbs the constant alternating flow of air and solution. This annoyance could be eliminated by preheating after mixing with the oxidant The gas, probably carbon dioxide, is removed by a debubbler, and the sample flow is subject to resampling This preheating inevitably requires a thermo-resistant Kel-F tube connection Because of bubbling problems, control urine (i e , urine obtained from non-exposed subjects) is better than water as washing liquid in order to stabilize the system. The sample, 0 8 ml/min, is mixed with 0 58 ml/min of the oxidant, and the flow for oxidation through a heating bath at 960 takes about 15 min This oxidation condition was considered less intensive than that of the manual analysis in which the sample-oxidant mixture (1:1) is subject to heating in a boiling water bath for 15 min The experimental results to be described later, however, indicate that the authentic TCE added is quantitatively converted to TCA under the conditions employed Furthermore, no improvement in the oxidation rate was observed when 2 heating baths were connected in series to double the oxidation duration Heating condition for colour development, about 1 min and 50 sec at 850, is not critical in the AutoAnalyzer system. Preliminary experiments with the manual analysis revealed that the colour intensity in 5 ml of the pyridine layer after Fujiwara reaction is independent to the amount of water layer in the range of 3 5 to 7 5 ml as far as alkali concentration is constant, that the colour intensity is reduced when the alkali concentration is less than 7 8 N, and that precipitation after KOH addition (presumably potassium chromate and potassium nitrate) is prevented when the amount of water layer is 7.5 ml or more under the conditions studied Based on these observations, alkali and water of more amounts than those of the manual analysis are pumped into the system so that no precipitation will take place during the reaction. Pyridine is a potent solvent and a precaution is required to pump it in Only a silicone tube can tolerate the heavy duty of the pumping Further connection should be with glass tubes using short pieces of silicone tubes as connecting joints. Merging of pure pyridine with salt-rich aqueous phase sometimes caused irreversible precipitation inside the T-shaped
196
connector, which hindered the constant flow Replacement with pyridine-water (2 : 1, v/v) clearly prevented the precipitation A mixture with lower water content (e g , pyridine : water = 3 : 1) was less effective. Extinction at 550 nm is about 83% of that at 530 nm, the wavelength with the maximum light absorption This disadvantage, however, is inevitable as interference filters are employed in the colorimeter, while a flow-cell with a longer light path, 15 mm, may serve to increase sensitivity The total flow-time from the sampler to the flow-cells is about 28 min. The inner capacity of the phase separator (B2 in Fig 1), employed to pick up pyridine phase selectively, appears to be rather large and may cause an undesirable mixing of colourdeveloped solution, resulting in a disturbance of a response. Improvement at this step will also help to achieve a higher sampling rate. Linearity of Colour Development in the Automated Method In order to examine if the intensity of the colour developed is linear to the TCA concentration, authentic TCA was dissolved up to the concentration of 96 mg/l in 4 control urine samples and the solutions were subject to the automated analysis The results are summarized in Fig 2 It is clear that the intensity of the colour developed is independent to the urine samples studied and proportional to the concentrations of TCA The same results were reproduced when TCE, up to 79 mg/l, was added to the control urine (Fig 2) Comparison between the 2 results with TCA and TCE on equi-molar basis revealed that colour intensity of TCE added is 101 7 ± 2 0 % (mean ±SD; calculated from the gradients of 4 regression lines) of the theoretical value, indicating that the oxidation condition is satisfactory for the practical use. Precision of the Automated Method Four determinations each were performed with the urine samples containing either 32 mg/l or 144 mg/l of authentic TCA to investigate the precision of the colour development process The coefficient of variation was 1 9% in the former case and it was 0 9% in the latter When 4 determinations each were carried out with the urine samples containing either 26 mg/l or 117 mg/l of authentic TCE in order to examine the precision of over-all procedure of the method, the results as shown in Fig 3 gave the coefficient of variations of 1 6% and 0 8%, respectively. 197
0.5
-0.4
E
0
Z
0.2
.E
I
0
20
40
60
80
100
Concentration (mg / I) of TCA and TCE
Fig 2
Development of colour lineary related to the amounts of TCA and TCE TCA and TCE were dissolved in the urine samples from 4 non-exposed individuals as shown by circles, squares, triangles and rhombs Open and solid symbols stand for TCA and TCE, respectively TCE are expressed as equimolar amounts of TCA Regression lines, Z = 0 00446 x (broken one) and z = O 00454 y (solid one), where x and y are concentrations of TCA and TCE, respectively, and Z is extinction measured, are calculated from the means of 4 determinations at each concentration
c C o O
o o
c
w
0
5
10 Time
15 ( min
20
25
30
)
Fig 3 Chart records of over-all determinations Four determinations each were carried out with the urine samples containing either (A) 26 mg/l or (B) 117 mg/l of TCE
198
Fig 4 Correlation between the results from the automated analysis and the manual analysis
Urine
samples, 54 in total, were collected from the factory workers exposed to trichloroethylene, and analyzed for TTC by both automated and manual methods. The line in the figure, y = 1.005 x + 0 480, is a calculated regression line where y is the value from the automated analysis and x is that from the manual analysis
The correlation
coefficient, r, is O 989 TTC (mg / I)by manual analysis
Correlation between the Results from the Manual and Automated Analysis Urine samples from the exposed subjects, 54 specimens in total, were analyzed for TTC by means of both manual and automated analyses, the results being summarized in a scatter diagram A close correlation exists between the 2 results as shown in Fig 4; the regression line is y = 1 005 x + 0 480, where
y : TTC mg/l as measured by the automated analysis, x
TTC mg/l as measured by the manual analysis,
and the correlation coefficient, r, is 0 989 (54 determinations) The results clearly indicate that an automated method at the sampling rate of 20 specimens per hr is established with the accuracy comparable to that of the time-consuming manual method.
Acknowledgements
Thanks are due to Dr Tokunaga of Kyoto Industrial Hygiene Association, Kyoto, Japan, Prof K Matsumoto of Nagasaki University Medical School, Nagasaki, Japan, and Dr N Ishihara of our depart-
ment for their collaboration in collecting urine samples.
199
REFERENCES deBary, J L J : An automated method of determining phenol in urine as an index of benzene exposure Proceedings of Technicon England Symposium, pp.13 5 -1 3
7
(1969)
Heistand, R N , Todd, A S : Automated determination of total phenol in urine Amer industr Hyg Ass J 33, 378-381 (1972) Ikeda, M , Imamura, T : Biological half-life of trichloroethylene and 31, Int Arch Arbeitsmed tetrachloroethylene in human subjects 209-224 (1973) Ikeda, M , Ohtsuji, H : Hippuric acid, phenol and trichloroacetic acid levels in the urine of Japanese subjects with no known exposure to 26, 162-164 (1969) organic solvents Brit J industr Med Ikeda, M , Ohtsuji, H : A comparative study of the excretion of Fujiwara reaction-positive substances in urine of humans and rodents given Brit J. trichloro or tetrachloro-derivatives of ethane and ethylene industr Med
29,
99-104 (1972 a)
Ikeda, M , Ohtsuji, H , Imamura, T , Komoike, Y : Urinary excretion of total trichloro-compounds, trichloroethanol and trichloroacetic acid as a measure of exposure to trichloroethylene and tetrachloroethylene. Brit J industr Med 29, 328-333 (1972 b) Imamura, T , Ikeda, M : A time-saving procedure for the determination Int Arch Arbeitsof total trichloro-compounds in human urine samples 31, 333-338 (1973) med Lehnert, G , Morgan, A , Szadkowski, D , Zielhuis, R L : Halogenated hydrocarbon solvents: Long term effects and biological sampling in human beings
Int Arch Arbeitsmed
33, 251-255
(1974)
Seki, Y , Urashima, Y , Aikawa, H , Matsumura, H , Ichikawa, Y , Hiratsuka, F , Yoshioka, Y , Shimbo, S , Ikeda, M : Trichloro-compounds in the urine of humans exposed to methyl chloroform at sub-threshold levels
Int Arch Arbeitsmed
34, 39-49 (1974)
Seto, T A , Schultze, M O : Determination of trichloroethylene, tri28, chloroacetic acid, and trichloroethanol in urine Analyt Chem 1625-1629 (1956) Tanaka, S , Ikeda, M : A method for determination of trichloroethanol and trichloroacetic acid in urine Brit J industr Med 25, 214-219 (1968)
Prof
Masayuki Ikeda
Department of Hygiene Tohoku University School of Medicine Sendai 980, Japan
200
