Statistical Associations between Radiation Exposure and the Clinical Examination Data of Japanese Radiology Technicians

The associations between occupational irradiation, cigarette smoking , alcohol drinking and clinical examination data were investigated in Japanese male radiology technicians . The number of investigated examination items was 35, including 29 biochemical serum test, four hematological tests and systolic and diastolic blood pressure. The associations with each factor were evaluated using the multiple linear regression model. As single factors, radiation associated with urea nitrogen, alkaline phosphatase, monoamine oxidase and leukocyte count (four items), smoking associated with albumin-globulin index, zinc sulfate turbidity test, urea nitrogen, creatinine, neutral fat, amylase, serum iron, leukocyte count, hemoglobin and hematocrit (10 items), and drinking associated with creatinine, uric acid, glutamate oxaloacetate transaminase , leucine aminopeptidase, alkaline phosphatase and erythrocyte count (six items). As synergistic factors , the combination of radiation and smoking associated with nine items, radiation and drinking 10 items , smoking and drinking four items, and radiation, smoking and drinking two items. These results suggested that the number of items which radiation associated as single-factor were less than that of smoking and of drinking, however suggested that associations between radiation and examination data was synergistic when combined with smoking or drinking. J Epidemiol, 1995; 5 : 51-57.

Hisayoshi Kondol, Takashi Aoyama2, Tsutomu Sugahara3, Tetsuaki Hashimoto4, Yoichi Yamamoto5, and Yutaka Okumural The associations between occupational irradiation, cigarette smoking , alcohol drinking and clinical examination data were investigated in Japanese male radiology technicians .The number of investigated examination items was 35, including 29 biochemical serum test, four hematological tests and systolic and diastolic blood pressure.
The associations with each factor were evaluated using the multiple linear regression model.As single factors, radiation associated with urea nitrogen, alkaline phosphatase, monoamine oxidase and leukocyte count (four items), smoking associated with albumin-globulin index, zinc sulfate turbidity test, urea nitrogen, creatinine, neutral fat, amylase, serum iron, leukocyte count, hemoglobin and hematocrit (10 items), and drinking associated with creatinine, uric acid, glutamate oxaloacetate transaminase , leucine aminopeptidase, alkaline phosphatase and erythrocyte count (six items).As synergistic factors , the combination of radiation and smoking associated with nine items, radiation and drinking 10 items , smoking and drinking four items, and radiation, smoking and drinking two items.These results suggested that the number of items which radiation associated as single-factor were less than that of smoking and of drinking, however suggested that associations between radiation and examination data was synergistic when combined with smoking or drinking.J Epidemiol, 1995; 5 : 51-57.radiation, occupational exposure, radiology technicians , clinical examination data Effects of repeated exposure to low-dose ionizing radiation have not been well known and understood.For humans, irradiation increases the risk of cancer death, but it is unknown that the effects of radiation on the individuals while alive.If the radiation affects the health of individuals, it is considered that the values of some clinical examinations might change those levels with radiation dose.Radiology technicians have been subject to exposure to low-dose ionizing radiation repeatedly by their occupations.Among older Japanese radiology technicians, the cumulative dose is probably rather high due to the poor working conditions before 1955.The number of Japanese radiology technicians was approximately 22,000 in 1981.We considered it pertinent, therefore, to study the health of this professional population and to evaluate the risk from repeated exposure to low-dose ionizing radiation.From 1981 to 1985, we conducted a health study on about 1,000 members of the Japan Association of Radiology Technicians.This study was carried out in a research project in which the estimation of occupational radiation doses of Japanese radiology technicians and the mortality study were also included.The study involved clinical examinations as well as tests for visual acuity, electrocardiogram, chromosome aberration and physical strength.In this paper, we examined the relationships between the radiation exposure and the clinical examinations under considering cigarette smoking and alcohol drinking.

MATERIALS AND METHODS
The health study on radiology technicians was performed for the members belonging to the Japan Association of Radiology Technicians.We offered free health examinations and announced them by distributing written notices and lecture classes to radiology technicians.The health examinations were performed in 24 prefectures which we could obtain the consent to participate in our study through Japan as shown in Figure 1.The subjects were taken from male technicians who came to be examined voluntarily.The total number of subjects was 1,014.The number and the percentage of participants in one prefecture ranged from 33 to 68 and from 2.9% to 27.9% (mean percentage 10.5%),respectively.A total of 37 diseases were observed among the subjects.Fifteen of these were considered to affect the data of examination items and thus were deleted from the analysis.As a result, 69 subjects who had these diseases were excluded and further analysis were performed with 945 subjects.The distribution of age in subjects was represented in Figure 2. Age ranged from 22 to 77 years, and the mean age was 47.3 years.The radiation dose for each radiology technician  was measured by film badge.However, the technicians had not worn film badges for personal radiation monitoring before 1960.Therefore, we estimated the radiation dose before 1960 in each technician based on the average dose of irradiation per exmination which was depended on their working conditions, the average number of daily examinations and the average monthly working days.Details of the method of dose estimation were discribed elsewhere1).The doses estimated were good correlation with those using the electron spin resonance (ESR) method with technician's teeth2).Cumulative radiation doses ranged from 0.01 Gy to 4.91 Gy, and the mean value was 0.31 Gy.The distribution of estimated cumulative exposure doses for individuals was represented in Figure 3. Cumulative dose was positively associated with age as shown in Figure 4.
For cigarette smoking, we defined as non-smokers who indicated that they had never smoked cigarettes.Exsmokers were those who were non-smokers at the time of investigation but indicated previous smoking.Smokers were those who had been smoked cigarettes continuously and were grouped with the smoking amount into eight categories.As a result, we classified smoking into 10 categories (non-smokers, ex-smokers and eight groups of smokers categorized to the smoking amount).
For alcohol drinking, we defined as non-drinkers who indicated that they had never or almost never drunk alcohol at ordinary times.Ex-drinkers were those who were non-drinkers at the time of investigation, but indicated previous alcohol drinking.Drinkers were those who regression was used to examine the associations with age , radiation dose, smoking, drinking and their interactions .The values of individual radiation dose were transformed logarithmically.The age was classified into 10 categories , and the values from 1 to 10 were assigned each category as shown in Table 1.It is impossible that the values of examination items in the population are deviated from the normal distribution when applying multiple linear regression model.The transformation of variables is often contrived so as to be normally distributed when they do not follow the normal distribution.Logarithmic transformation, square root transformation and reciprocal transformation were considered in this study.In order to test normality, we used the skewness with an expected value of zero for the normal distribution.If the normality was rejected by the skewness of crude values, the skewnesses of logarithmic transformed values, square root transformed had been drunk alcohol every day and were grouped with the drinking amount into four categories.As a result, we classified drinking into six categories (non-drinkers, exdrinkers and four groups of drinkers categorized to the drinking amount).
For each item of clinical examinations, multiple linear   2. The outliers leading to biased results were eliminated by use of the Grubbs-Smirnov test3-5).Statistical analyses were performed using the SAS statistical computer package.

RESULTS
After the transformation of variables to be normally distributed, the multiple linear regression model which used the stepwise method as the model-selection methods was fitted for each items.Partial regression coefficients which were significant (P<0.05) in the multiple linear regression were shown in Table 3.As single factors, radiation showed significant associations with urea nitrogen, alkaline phosphatase, monoamine oxidase and leukocyte count (four items), smoking showed with albuminglobulin index, zinc sulfate turbidity test, urea nitrogen, creatinine, neutral fat, amylase, serum iron, leukocyte count, hemoglobin and hematocrit (10 items), and alcohol drinking showed with creatinine, uric acid, glutamate oxaloacetate transaminase, leucine aminopeptidase, alkaline phosphatase and erythrocyte count (six items).The number of items which showed significant association with radiation was less than that of smoking and of alcohol drinking.The items showing synergism by radiation and smoking were total protein, albumin, gamma-glutamyl transpeptidase, alkaline phosphatase, creatine phosphokinase, chloride, total bilirubin, leukocyte count and hemoglobin (nine items).The items showing synergism by radiation and drinking were total protein, albumin, zinc sulfate turbidity test, beta-lipoprotein, cholinesterase, lactate dehydrogenase, gamma-glutamyl transpeptidase, amylase, sodium and leukocyte count (10 items).The items showing synergism by smoking and drinking were total cholesterol, inorganic phosphorus, calcium and hematocrit (four items).The items showing synergism by radiation, smoking and drinking were inorganic phosphorous and hemoglobin (two items).These results indicated that associations between radiation and each item might be synergistic when combined with smoking or drinking.

DISCUSSION
The associations with radiation have been investigated on human beings in various aspects.Seltser and Sartwe116) and Matanoski et al.7) indicated that the longevity of American radiologists was not shorter than that of other physicians.Court-Brown and Doll8) also reported that the longevity of British radiologists was not shorter than that of controls.For radiology technicians, Miller and Jablon9) reported that a longevity shortening was not observed in X-ray technologists in the United States.
With regard to radiation-induced cancer, Wang et al.10) reported that diagnostic X-ray workers in China revealed a 21% greater incidence of cancer than physicians who did not routinely use X-rays.
With regard to radiation-induced aging, Beisky et al.11) and Conard et al.12) investigated the A-bomb survivors and Marshall Islands people respectively, and reported that radiation did not accelerate aging.
With regard to mortality, Aoyama et al. 13,14) reported that the observed number of death from all causes in Japanese radiology technicians was less than the expected number.However, they reported that the observed number of death from cancer was more than that of the expected.In the A-bomb survivors, Shimizu et al.15) reported that the excess death from the cardiovascular diseases and cerebrovascular diseases was observed in a high dose exposed group, and this indicated that A-bomb radiation may have caused diseases other than cancer also.response and the estimated relative risk at 1 Gy was 1.14 (95% CI : 1.04-1.27).Furthermore, Shimizu et al.15) reported that the mortality from cirrhosis among the A-bomb survivors increased in the higher-radiation-group whose estimated dose ranged from 2 to 3 Gy and the estimated relative risk was 1.16.From these reports, the associations between alkaline phosphatase, monoamine oxidase and radiation observed in this study may be due to slight liver disorder by repeated radiation exposure.It is considered that the increasing of the level of uria nitrogen is either renal or extrarenal.According to calculations for uria nitrogen-creatinine index, the prevalence of person whose value was less than 10 was 13.2% in all subjects, and that was 0.9% in those whose level of urea nitrogen was greater than 20.From these results, it seems that the increasing of uria nitrogen observed in the present study is not caused by radiation-induced kidney disorder.It is known that the leukocyte count decreases by high dose radiation which impair bone marrow, and our result is opposed to this fact.Technicians' radiation dose may be too low to impair bone marrow.It is also known that the leukocyte count increases by physical and mental stress, and the technicians whose radiation doses are relatively high may be under more stress than those whose radiation doses are low.
With regard to the associations with cigarette smoking, the induction of lung cancer, bronchitis and pulmonary emphysema and ischemic heart disease were reported21-23) The associations with cigarette smoking observed in this study were increases of the levels of albumin-globulin index, neutral fat, serum iron, leukocyte count, hemoglobin and hematocrit and decreases of the levels of zinc sulfate tubidity test, urea nitrogen, creatinine and amylase.Howell24), Corre et al.25) and Friedman et al .26)reported that leukocyte count was significantly higher in smokers than in non-smokers.Nordenberg et al.27) reported that cigarette smoking caused an upward shift of hemoglobin level.The results of this study for leukocyte count and hemoglobin confirmed the above reports.It seems likely that habitual smoking causes chronic hypoxia and results in increases in the level of hemoglobin and hematocrit.Doll et al28).reported that cigarette smoking increased the risk to hepatic cirrhosis, and a decrease of the level of creatinine and increases of the levels of neutral fat and serum iron were all observed in liver disorders.From these results, smokers may have mild liver disorders and thereby show changed the levels of these items.
With regard to the associations with alcohol drinking, inductions of liver cancer, hypertension and stroke were reported 29-31).The associations with drinking observed in this study were increases of the levels of uric acid, glutamate oxaloacetate transaminase, leucine aminopeptidase and decreases of the levels of creatinine, alkaline phosphatase and erythrocyte count.Persons with liver disorders tend to have a high level of uric acid, glutamate oxaloacetate transaminase and leucine aminopeptidase and a low level of creatinine.
Drinkers may have mild liver disorders due to chronic alcoholic damage and thereby show changed levels.As the disorders of hematopoietic organ involving alcohol, it is known anemia which results from poor or insufficient nutrition.The result of this study for erythrocyte count is consistent with this fact.

Figure 4 .
Figure 4. Correlation diagram between radiation dose and age .

Table 1 .
Classified age groups and values.
Radiation DoseFigure3.Distribution of cumulative radiation dose.

Table 2 .
Transformation with the smallest skewness.A/G index : albumin-globlin index, ZTT : zinc sulfate turbidity test, TTT : thymol turbidity test, LDH : lactate dehydrogenase, GOT : glutamate oxaloacetate transaminase, GPT : glutamate pyruvate transaminase, LAP: leucine aminopeptidase, gamma-GTP : gamma-glutamyl transpeptidase, A1P : alkaline phosphatase, CPK : creatine phosphokinase, MAO : monoamine oxidase values and reciprocal transformed values were calculated, and non-transformation or transformation of variables with the smallest skewness was adopted.The number of items with minimum skewness was seven for nontransformation, 12 for logarithmic transformation, eight for square root transformation and eight for reciprocal transformation as shown in Table

Table 3 .
The partial regression coefficients of each factor.