Hepatocellular Carcinoma and Its Risk Factors Their Annual Changes and Effects on the Age at Onset

A retrospective case-control study was carried out on patients with hepatocellular carcinoma (HCC) admitted between 1976 and 1985 to analyze the relationship between the cancer and risk factors. The following conclusions were reached: 1) the positive rate of Hepatitis B surface (HBs) antigen among the HCC patients decreased from 41% in the period 1976-1980 to 19% in the period 1981-1985. However, the risk of the patients positive for the antigen for HCC was higher in the second period (1:18.7) than in the first period (1:12.8), 2) drinking not only enhanced the risk for HCC 2.7 times, but also accelerated the onset of HCC by about 5 years, 3) blood transfusion enhanced the risk for HCC; HCC occurred in about 20 years on average after blood transfusion in the patients negative for the hepatitis B virus (HBV) and drinking history, and 4) smoking was suggested to be less related to the onset of HCC.


Introduction
In Japan, HCC is the third most frequent malignant tumor following gastric and pulmonary cancers and is increasing year by year.While the positive rate of HBV, a significant risk factor of HCC among HCC patients, is gradually decreasing (Japan Liver Cancer Seminar, 1979, 1986), the incidence of liver cirrhosis and HCC is increasing in proportion to an increase in alcohol consumption (Hirayama, 1983).To clarify the cause of the decrease in the positive rate of HBs antigen among HCC patients and to evaluate the effect of drinking on HCC, we evaluated these two risk factors and two other factors, blood transfusion and smoking, by a case-control study using hospital controls.The diagnosis of HCC was made histologically by biopsy or autopsy, or clinically by the assay of AFP, angiography, computed tomography and ultrasonography.

Subjects and Methods
To obtain 1:1 pairs, controls closest in age to the subjects were selected among the patients who met the following criteria; findings of the HBs antigen, confirmed histories of drinking, blood transfusion and smoking and the absence of chronic hepatic disorders.

I. Relationships between HCC and the risk factors
To examine the time-course of changes in the levels of the relationships between HCC and the risk factors, the patients and the controls were each divided into a group of 163 in the first period (5 years: January 1976-December 1980) and 303 each in the second period (5 years: January 1981-December 1985).The absence of deviations for sex difference and the age at diagnosis between the HCC and control groups and between the first and second period within each group was confirmed before analysis.
Since the positive rates for the risk factors might differ with the type of disease, the hospital controls were analyzed to rule out deviations for the type of disease between the two periods.

II. Relationship between HCC and drinking
To determine whether there was a dose-response relationship between HCC and drinking, the male patients negative for the HBs antigen were divided into the following 3 groups on the basis of alcohol consumption : none, heavy (more than 540 ml of sake/day approximately 80 ml of pure alcohol/day for more than 10 years) and moderate (an intermediate volume).
The male HCC patients were divided into the following 3 groups; positive for the HBs antigen without a history of blood transfusion and negative for the HBs antigen with or without a history of blood transfusion (the last, etiologically unknown group).The 3 groups were compared with respect to age distribution according to the presence or absence of drinking history.Frequency of risk factors i n cases and controls showed no significant difference (from 21% to 19%) in the control group.In the second period, the HCC group showed a significantly higher rate than that of the control group (p<0.001).The rate of history of blood transfusion decreased from 17% to 14% in the HCC group and from 7.4% to 7.6 % in the control group without any significant differences.

III . Relationship between blood transfusion and the onset of HCC
However, the rate in the HCC group was signif icantly higher than that in the control group in both periods (p<0.01).The rate of smoking history was 65% in both periods in the HCC group and decreased from 63% to 62% in the control group, with no significant difference in either group.There were no significant differences between the two groups.
4. Changes in the levels of the relationships between HCC and the risk factors from the first period to the second period (Table 4) The odds ratio representing the level of the relationship between HCC and the HBs antigen between the 163 HCC patients and 163 controls in the first period was 12.8(95% confidence interval: 5-36).The odds ratio in 303 pairs of HCC patients and controls in the second period was 18.7 (5.6-75).
The relationship between HCC and the HBs antigen was significant in both periods (p<0.001).The relationship between HCC and drinking history was calculated similarly.While the odds ratio in the first period was 1.3 (0.7-2.3) with no significant difference, that in the second period was 2.6(1.7-4.1) with a significant difference (p<0.001).The odds ratio for the history of blood transfusion was 2.6(1.2-5.8) in the first period and 2.1(1.2-3.7) in the second period, both values being significant (p<0.02).By contrast, the odds ratio for smoking history showed no significant differences, 1.1(0.6-2)and 1.2(0.8-1.7),respectively.

II. Relationship
between HCC and drinking history 1. Dose-response relationship between HCC and drinking (Table 5) The male HCC patients were analyzed to determine whether or not there was a dose-response relationship between HCC and drinking.
With the degree of risk in the non-drinking group as 1, the degree of relative risk was calculated.
The degree of relative risk was 1.3 in the moderate group and 2.7 in the heavy group, indicating a positive dose-response relationship, which was statistically significant (p<0.001).

Relationship
between the age at onset of HCC and drinking (Fig. 2

Discussion
The HBs antigen, drinking history, history of blood transfusion and smoking history were evaluated as four risk factors of HCC.Drinking and smoking histories differ significantly between male and female populations.
In addition, females show higher alcohol sensitivities (Nishimura et al. 1984).For these reasons, we examined the differences in the male/female ratio and age distribution between the first and second periods before analyzing the changes in the positive rates of the four factors in the past 10 years.Since there are diseases closely related to smoking and drinking, we investigated whether or not there were deviations for the disease categories in the control group between the first and second periods.
None of these analyses revealed significant differences, indicating that the effect of these compound factors was statistically negligible.The HBs antigen has been found to be closely related to HCC as a result of world-wide research in descriptive epidemiology (Yarrish et al. 1980;Hann et al. 1982) and analytical epidemiology (Bearsley et al. 1981;Sakuma et al. 1982) and more recently in genetic technology (Shfritz et al. 1981).In Japan, studies of HCC reported a gradual decrease in the positive rate of HBs antigen (Japan Liver Cancer Seminar, 1979Seminar, , 1986).This tendency is strongly supported by the findings in the present study.
The degree of relative risk for individuals positive for the HBs antigen to develop HCC (risk for HCC) differs according to the country and the research group, ranging from 1:223 to 1:3.8.There is no report on whether or not the decrease in the positive rate of HBs antigen among HCC patients in Japan is due to a decrease in the risk for HCC among them, or a relative decrease due to an increase in the incidence of HCC among individuals neg-ative for the HBs antigen.The present study revealed a significant decrease in the positive rate of HBs antigen among the HCC patients from 41% in the first period to 19% in the second period and a decrease from 4.9% to 1.7% in the control group.
Consequently, the risk for HCC among the patients positive for the HBs antigen showed an increase in the second period (12.8 to 18.7).These findings indicate that the decrease in the positive rate of HBs antigen among the HCC patients reflects a relative decrease due to an increase in the incidence of HCC among the patients negative for the HBs antigen, rather than a decrease in the risk for HCC among the patients positive for the HBs antigen.
Even if the risk for HCC remains unchanged, a decrease in HBV carriers in Japan seems to result in not only a continuing relative decrease, but also a continuing absolute decrease, in the incidence of HCC among individuals positive for the HBs antigen.
The risk for HCC in the first and second periods was 15 (7.1-33) , which is lower than the value reported by Sakuma et al. (1982), but is similar to that reported by Inaba et al. (1981).
Although the relationship between HCC and drinking was considered to be insignificant in the past, because the incidence of HCC was low (Mori and Shan, 1972) in Western countries where alcoholic liver cirrhosis is prevalent, HCC is now increasing (Ferenci et al. 1984;Rimkus and Dhom, 1986).In Japan, an increase in alcohol consumption has resulted in an increase in the incidence of alcoholic liver cirrhosis, drawing attention to its relation to HCC (Hirayama, 1983).Our study revealed a significant increase in the rate of drinking history among the HCC patients from 25% in the first period to 34% in the second period, but no significant difference (from 21% to 19%) in the control group.
The risk for HCC among the patients with drinking histories was 1.3 in the first period and 2.6 in the second period, with a significant difference in the latter.
It is speculated that an increase in the incidence of liver cirrhosis due to an increase in alcohol consumption and the longevity of cirrhotic patients with drinking histories as a result of progress in medical care have led to increases in the incidence of HCC and the risk for HCC.Investigation of the relationship between alcohol consumption and HCC in the male patients negative for the HBs antigen revealed an increase in the risk for HCC in proportion to alcohol consumption, with a significantly positive dose-response relationship : the risk for HCC among the habitual drinkers was 1.3 times as high as that among the non-drinking individuals, and that among the heavy drinkers, was 2.7 times as high.Analysis of the relationship between drinking and the age at onset of HCC according to the etiology showed that the patients with drinking histories were about 5 years younger in all three groups positive for the HBs antigen, with blood transfusion and with an unknown cause.Therefore, drinking not only enhances the risk of HCC, but also precipitates the onset of HCC.
Blood transfusion was significantly related to HCC in both the first and second periods, but smoking exhibited no such relationship.
Smoking has been investigated in relation to HCC in the epidemiological field.There are reports supporting (Trichopo los et al. 1980) and denying (Austin et al. 1986) the relationship among HCC patients negative for the HBs antigen.
The effect of smoking may have been rated too low because of analyzing only hospital controls on this study.However, smoking does not seem to be so strongly related to HCC as is HBV, blood transfusion or drinking.

Conclusion
1) The decrease in the positive rate of HBs antigen among the HCC patients reflects a relative decrease due to an increase in the HCC patients negative for the HBs antigen, rather than a decrease in the risk for HCC among the HBV carriers.
2) Drinking is not only a risk factor, but also a promoter of HCC, and has become a more significant risk factor.3) Smoking is not a significant risk factor of HCC.

Fig
Fig. 1.Frequency of risk factors in cases and controls.