International trends in primary liver cancer incidence from 1973 to 2007

International trends in PLC incidence rates during the period 1973–2007 showed that the PLC incidence rates increased in most European, American and Oceanian populations, although these age-standardized PLC incidence rates in 2003–2007 were much lower than these in Asia. Meanwhile, PLC incidence rates decreased in Asian populations, although their age-standardized PLC incidence rates in 2003–2007 were the highest in the world.

PLC is a common cancer, particularly in Asia countries such as China, Japan and Singapore (Chinese). Among these countries, PLC is closely associated with hepatitis virus infection (HBV infection in China and Singapore, HCV infection in Japan) and exposure to aflatoxin (in China). In our study, the decreasing trends in China and Singapore may be attributed to some public health measures [24-27]. HBV vaccination was incorporated into the national childhood immunization program by China and Singapore from the middle 1980s to the early 1990s. The immunization coverage with three doses of HBV vaccine was 70.7%-95.0% in 1999 [28,29]. Several studies also reported the decreases in PLC incidence rates in China, particularly in Shanghai and in younger age groups [24,30]. Another study in Taiwan showed that the age- and sex- adjusted rate ratios for individuals aged 5 to 29 years decreased by more than 80% for HCC incidence from 1977–1980 to 2001–2004 [25]. In Singapore, Chia et al. [26] suggested that a general declining trend in liver cancer incidence was especially notable in local-born Chinese. Although the measure had no an effect on general population, we expect it will play an important role in the reduction of PLC incidence rates in the coming decades. Moreover, dietary aflatoxin exposure declined in the high-incidence areas of PLC seemed to have contributed to the decrease in PLC incidence in China [31]. A study in Qidong, China [31], where aflatoxin exposures were common, had reported that the decreasing liver cancer incidence in population over 25 years could mainly be attributable to the reduction of exposure to aflatoxin from 1980 to 2008.

In Japan, there were different trends between 1973–1992 and 1988–2007. The increasing trends started in 1973–1977 and reached peak in 1988–1992. This was thought to be in part due to the spread of HCV infection, which began in the 1920s and increase after World War II with an explosion in parenteral amphetamine use and paid blood donation [32,33]. Although APC did not significantly decrease during 1988–2007, the decline in PLC incidence had been continuously seen from 1988–1992 in our analysis. Stiffening of legal penalties against amphetamine use starting in 1954 and conversion from paid to voluntary blood donation in late 1960s might have reduced HCV transmission [34]. After the discovery of HCV RNA, HCV screening tests for first- and second-generation HCV antibodies started in 1989 and 1992, respectively [35]. These tests were adopted by the Red Cross Blood Center for screening blood, which further decreased the risk of post-transfusion hepatitis. The Japanese government has taken urgent comprehensive countermeasures against hepatitis (HBV and HCV) and HCC since 2002 [33]. Therefore, these measures would provide a significant contribution to decrease the number of patients suffering from HCV-related liver diseases including PLC.

In contrast, PLC was not a very common cancer in European, American and Oceanian countries where there were no epidemic regions of HBV infection. However, an increasing trend of PLC incidence rates was seen in most of these populations which were partly due to the widespread HCV infection associated with drug use, exposure to contaminated blood transfusion and/or needles used for medical purposes [36]. The natural history of PLC indicated that the time between exposure to HCV and development of HCC is about 30 years [37]. HCV infections were found in 30-50% of HCC patients in the United States and 44-66% of HCC patients in Italy [38]. Both of these countries had the highest PLC incidence rates in their own continent. The different peak years of HCV infection prevalence for each country were likely responsible for the respective peaks in PLC incidence rates. The increasing trends in PLC incidence rates in the United States could be attributed to increased HCV exposures by contact with contaminated blood and injection drug use during the 1960s and 1970s [39]. In Italy, the upsurge of liver cancer incidence after 1970s was in part attributable to HCV infection caused by the re-use of disposable syringes among intravenous drug users without proper disinfection [40]. Meanwhile, in several studies conducted in Western countries, 30 to 40% of patients with HCC did not have chronic infection with HBV or HCV, suggesting alternative causes [5]. Other factors including alcohol [41,42], obesity [43,44], and non-alcoholic fatty liver disease (NAFLD) [5] might be contributed to the increasing trend. In population-based cohort studies in the United States and Scandinavia [44-46], HCC was 1.5 to 2.0 times as likely to develop in obese persons as in those who were not obese. NAFLD, which is present in up to 90% of all obese persons and up to 70% of persons with type 2 diabetes, has been proposed as a possible risk factor for HCC [47]. Although there were still some difficult problems in the latency period from exposure to these factors and PLC development, more emphasis should be recommended to control these factors.

The advent of precise diagnostic tests may increase recognition of the disease, which accounts for a rising incidence, rather than a true increase in its occurrence [48]. Ultrasonography, measurement of serum alpha-fetoprotein, and computed tomography scanning have been routinely used since the early 1980s, which should lead to an increase in the number of hepatic biopsies conducted. However, the percentage of histologically confirmed PLC has not increased significantly during the study period in these countries which had an increasing trend in PLC incidence rates. In addition, females in Poland (from 1973–1977 to 2003–2007) and males in Spain (from 1982–1987 to 2003–2007) also exhibited a decreasing trend of PLC incidence rates. The reasons for this decreasing trend remain unclear.

This study has several strengths. The data were abstracted from large, well-established registries throughout the world. For the first time, data covering 35 years were analyzed to describe the variation of international trends in PLC incidence rates, which may stimulate further etiologic research. In addition, incidence rates for particular histologic subtypes of PLC in different populations were examined separately. One limitation of this study was that trends by histologic subtypes were not examined. The variation of ICD coding might influence the interpretation of our results. In our study, ICD coding contained ICD-8 (Malignant neoplasm of liver and intrahepatic bile ducts, specified as primary), ICD-9 (Malignant neoplasm of liver, specified as primary) and ICD-10 (Malignant neoplasm of liver and intrahepatic bile ducts, specified as primary). CC was not included in ICD-9 (period from 1978–1982 to 1993–1997). Therefore, the changes in PLC rates mainly reflect changes of HCC. Our study was also limited by the lack of nationwide cancer registries in some countries, thus registration data might not accurately reflect the true patterns in the respective countries.