Epidemiologic study on Yusho, a Poisoning Caused by Ingestion of Rice Oil Contaminated with a Commercial Brand of Polychlorinated Biphenyls.

In October, 1968, an epidemic of a peculiar skin disease similar to chloracne was reported in Fukuoka-Ken (Fukuoka prefecture), Japan. The epidemic was later proved to have spread not only over Fukuoka-Ken but also over 20 other prefectures in the western part of Japan (Fig. 1). It produced 1,057 patients according to the latest tabulation (August, 1971) by the Ministry of Welfare. Soon after the epidemic was announced, a study group was organized by the staff of the faculties of medicine, pharmaceutical sciences, agriculture, and engineering of the Kyushu University and by the staff of the local health departments, to clarify the cause of the epidemic and to effectuate its control (1). We participated in the study group and conducted, as members of its epidemiologic study subgroup, an extensive epidemiologic investigation following the basic methodology of epidemiology (2). Fortunately, the cause of the epidemic was soon demonstrated to be the ingestion of a brand of rice oil contaminated with a commercial brand of polychlorinated biphenyls, and the disease was called "Yusho", namely oil disease. Although our observations and experiences are confined to the Yusho patients seen in FukuokaKen, we believe that reporting of them will help many people in the world who are deeply con-

Dark brown pigmentation of nails and skin, follicular accentuation, acne-form eruption, increased eye discharge, increased sweating at palms, and feeling of weakness were the most notable symptoms (2).

Descriptive Epidemiologic Studies
First, 325 patients seen in Fukuoka-Ken from October, 1968, to January, 1969, were analyzed in order to know their distribution characteristics. ODe of the most important characteristics readily noted was a distinct familiar aggregation. The 325 patients belonged to 112 families. As shown in Fig. 2 and Table 3, 99 percent of these patients were affected during 1968, while the remaining stated that they became ill in December, 1967. Fifty-five percent of the patients were concentrated in the 3 months from June to August, and no significant difference was noted between sexes in monthly distribution (Table 3). For geographical distribution of the patients, crude incidence rates of Yusho were calculated for 3 large cities and for the jurisdictional areas of the 22 local health departments of Fukuoka-Ken. Excepting 10 health departments where no cases of Yusho were reported, the rates varied considerably from 1.0 to 58.9 per 100,000. Examination of such geographical distribution failed to indicate any common socioeconomic or environmental factor which might be associated with the disease. The 325 patients consisted of 158 males and 167 females, indicating that both sexes were equally affected. More than 90 percent of them were younger than 50 years (Table 4). Age-and sexspecific incidence rates again indicated no significant sex difference but lower risks for both males and females in the age group over 60 years (Table 5).

Analytical Epidemiologic Studies
When our study started, a commercial brand of rice oil produced by K company (abbreviated as K rice oil) in Kitakyushu-Shi (Kitakyushu city), Fukuoka-Ken, had vaguely been suspected as a possible cause of the disease, because most patients with Yusho seemed to have used it. Therefore, a thorough investigation was undertaken to determine whether the patients had actually taken such special brand of oil before being affected. This was achieved by examining lot numbers appearing on the remaining containers of the oil used by some of the patients, shipping records of K company, and purchase and sale records at wholesale oil dealers' offices and at retail stores. It was soon disclosed that all the patients had used K rice oil, either canned (16.5 kg) or bottled (1.65 kg). Furthermore, an astonishing fact became evident: No matter where they lived, 166 of 170 patients who used only the canned K rice oil had used a very specific oil, produced or shipped by the company on February 5 and 6, 1968 (Table 6). For those who used bottled K rice oil, date of production or shipment could not be confirmed, because they had no old bottles at home. Examination of all available records concerning shipping, sale, and purchase at K company, wholesale dealers and retail stores, however, clearly indicated a possibility that 143 of 155 patients who used bottled K rice oil only had used a very specific rice oil produced or shipped from February 5 to 15,1968. This was because such oil had been shipped to and had reached the retail stores from which they had purchased their oil (Table 6). Thus, nearly all of the patients had had a very peculiar common experience of use or possible use of a rice oil produced or shipped by one company in a specific period of time. An attack rate as high as 63.9 percent was noted for those who consumed the specific canned oil.
In an additional survey, we examined whether those who regularly used K rice oil, but did not use the specific K rice oil produced or shipped  All these results suggested that Yusho was caused by use of the K rice oil that was produced or shipped from K company on February 5 and 6, 1968, or soon thereafter. Nevertheless, some other factors or agents might have been the primary or secondary cause of the epidemic. Therefore two case-control studies were done. In one, the per-  Table 7, only one of the 60 personal factors examined, namely habit of "eating fried foods or tempura nearly every day", was significantly more commonly seen among the patients than among the controls. In the latter case-control study, 69 households with Yusho patients were matched by place of residence with 207 control households without   (Table 8). Thus, the case-control studies clearly indicated that none of the factors tested except use of K rice oil could account for the disease.

Dose-Response Relationship
To prove a causal relationship, a definite doseresponse relationship is needed. A rough estimate of the quantity of the specific K rice oil consumed by each patient and his family members was made disregarding their age, sex, amount of food intake, and possible loss of oil during and after cooking (4). Eighty of the 146 users of the specific canned K rice oil in question were believed to have consumed, individually, less than 720 ml. For these 80 light users, the attack rate of Yusho was 88 percent, while for those who were estimated to have used more than 720 ml, it was 100 percent (Table 9). It was also demonstrated that the proportion of severe cases of Yusho clearly increased with the amount of oil consumed. While the clinical severity of the disease was not found to differ significantly between the sexes, it did differ considerably according to age, as shown in Table 10 and Fig. 3. The proportion of severe cases among those aged 13 to 29 was significantly larger than that of other age-classes. Therefore, each of the three groups of users, with different levels of oil intake, was standardized for age using the age composition of the whole 146 users as standard. The figures, however, hardly changed from those shown in Table 9. Thus, a clear doseresponse relationship could be demonstrated, even though the estimates of the dose were inaccurate (4).

Toxic Agent
In view of all these epidemiologic observations, we concluded that the K rice oil of specific production or shipments had caused Yusho. Now, why was the oil toxic? The chemical study subgroup (chief: Prof. H. Tsukamoto, Faculty of Pharmaceutical Sciences, Kyushu University) demonstrated that the canned K rice oil produced or shipped on February 5, 1968, and used by some of the patients contained about 2,000 to  In this discovery, Prof. K. Inagami of the De-Gas chromatographic analysis revealed that a partment of Food Technology, Faculty of Agri-significant contamination of the bottled oil was culture, Kyushu University, and his associates limited only to those produced or shipped between made most admirable contributions as members February 7 to 10, 1968 (5). No analysis could be of the chemical study subgroup. The study made of the bottled K rice oil produced or shipped subgroup also demonstrated that the oil was not on February 5 and 6 because of the lack of samples. contaminated with toxic agents such as Cu, Ni, Similarly, 479 random samples of bottled K rice Zn, Co, As, Hg and Pentachlorophenol. Further-oil were analyzed for chlorine content by the more, it found that most of the components of X-ray fluorescence method with a count meter.
Kanechlor 400, particularly those corresponding Again, only the samples of February 7 to 10 to the peaks of higher retention times in the contained a large amount of chlorine (maximum gas chromatograms, were retained in the sebum, 462 ppm). None of the oils shipped in other subcutaneous fat, mesenterium, mesenteriolum, months were contaminated with more than a extraperitoneal adipose tissue, appendix ver-trace amount of chlorine. Thus, the results of miformis, heart, sternal marrow, small intestine, chemical studies completely coincided with those trachea, and other organs of patients, and can achieved by epidemiologic approaches. By the be transported into the fetus through the plaway, the Kanechlor had been used at the K centa (5,6,7). To examine whether only the K Company in the equipment for heating the procrice oil produced or shipped in the period in essed oil at a reduced pressure in order to remove question was contaminated with Kanechlor, the the odorous matters of the rice oil (Fig. 4). It is  believed that it must have leaked from the heating pipe and contaminated the oil, because small openings were discovered in the old pipe.
Amount of Kanechlor 400 Ingested by Patients One hundred forty-six patients who were proved to have used the contaminated canned K rice oil of February 5 and 6 by our investigation were used for the estimation of the amount of Kanechlor 400 ingested. As stated, the approximate amount of the oil consumed could be calculated for each of the patients. It was on average about 800 ml. Since the concentration of Kanechlor 400 in the oil was 2,000 to 3,000 ppm, the average amount of Kanechlor 400 ingested by a patient was estimated to be about 2 g (4). Similarly, the minimum dose of Kanechlor 400 consumed by a patient was estimated to be about 0.5g.

Babies Born to Patients and Non-affected
Wives of Patients Thirteen women consisting of 11 with Yusho and 2 unaffected wives of patients were shown to have delivered 10 live born and two stillbom babies from February 15 to December 31, 1968 (8,10,11). As shown in Table 11, nine of them had unusually grayish, dark-brown stained skin and similar pigmentation of the gingiva and nails was noted in five of them. Increased eye discharge was also notable in most of them. Histological examination of a stillborn fetus showed a marked hyperkeratosis and atrophy of epidermis and cystic dilatation of hair follicle, especially at the head. A marked increase of melanin pigments in the basal cells of epidermis was also noted (9). Since no such symptoms at birth or stillbirth has been experienced in Japan during the period in which their incidence among the patients was extremely high, and also since none of the mothers had had any unusual experiences, for instance, in use of drugs or in health status, these unusual phenomena could be considered to have been caused by ingestion of the contaminated oil. However, a clear dose-response relationship between the oil consumption and such phenomena could not be shown because of the limited number of cases. The amount of the contaminated oil consumed during pregnancy ranged from 0.    (Table 11) (8). Twelve of 13 fetuses were smaller than the national standards, and 4 of them were small-for-dates babies (8,10). As they grew older, the stain in their skin gradually faded (11). No evidence has so far been obtained with regard to any physical and mental retardation of the babies, but we suggest that a particularly careful and prolonged follow-up observation should be maintained for their future (8).
Growth of Affected Children To examine whether Yusho disturbs children's growth, the affected school children, 23 boys and 19 girls, were compared in 1967, 1968, and 1969 with their 719 healthy classmates matched by sex. The gains of the affected boys in both height and weight decreased significantly after the poisoning, while the affected girls showed no definite change in this respect (12).

Current State of Patients
More than 3 years have passed since the epidemic was reported. It is a heart-breaking fact that many patients with Yusho are still tortured by the sickness because no curative treatments have been discovered yet. In summer of 1970, a mass clinical examination of the Yusho patients was carried out in Fukuoka-Ken, and the clinical state of the patients examined were compared with their previous findings observed in summer of 1969. As shown in Table 12, about a half of the 159 patients, for whom the comparison was feasible, were shown to have clinically improved signs, while the remaining half showed no such favorable signs, and more than 10 percent of the patients were even worsening (13). It should be noted, furthermore, that even many of those who appeared to be clinically improving had several serious complaints, such as persistent headache, general fatigue and feeling of weakness, numbness in limbs, weight loss, and others (14). All these facts clearly indicate that recovery from Yusho is extremely difficult. To cure them as quickly as possible and to prevent another epidemic of Yusho, all possible world-wide cooperative efforts are highly desired.
From the middle of February to the end of March, 1968, an epizootic of a strange disease closely resembling chick edema disease occurred involving more than two million chickens in the western part of Japan (15,18). More than 400,000 chickens were reported to have died. The disease was characterized by such clinical signs as labored breathing, droopiness, ruffled feathers, high mortality and decreased egg production (15,18). Autopsy revealed a marked subcutaneous edema, hydropericardium, ascites, and pulmonary edema, muscular ecchymosis in the thorax or inside of the thigh, and yellowish mottled appearance of the liver (15,17,18). An epizootiological investigation demonstrated that the disease was caused by feeding chickens with specific lots of commercial brands of formula chicken feeds manufactured by two companies. In view of the fact that the K rice oil used by Yusho patients had been contaminated with Kanechlor 400, the toxic feeds were analyzed for chlorinated hydrocarbons and proved to contain Kanechlor 400 (16,19,20). The experimental reproduction of the disease by administering this chemical mixture was successful (16,20,21). The reason the chicken feeds were contaminated with Kanechlor 400 could be reasonably understood by the following facts: both of the chicken feed manufacturers had been using "Dark oil" produced by the K company (Fig. 3) as an ingredient of their formula feeds, and they had actually used "Dark oil" purchased from the company from February 6 to 27, 1968. The remaining sample of "Dark oil," used for production of the toxic formula feeds, was also shown to contain about 1,300 ppm of Kanechlor 400 (16,20). Thus, it became clear that both the epidemic of Yusho and the epizootic of the disease among chickens had been very closely associated, being caused by intake of the oils contaminated with Kanechlor 400. It is very unfortunate, however, that the epidemic of Yusho could not have been prevented even though a large scale epizootic of such an unusual chicken disease had preceded the incidence of Yusho by more than 6 months.