The present report showed the changing etiology of LC in Japan. To assess the recent real-world changes in the etiologies of LC in Japan, we conducted a nationwide survey in 2018 and the results were compared with those of the 2008 survey. Although nationwide surveys have been conducted in JSH annual meetings, these surveys have been sporadic and have had their own diagnostic criteria . This is the first study in which nationwide experts referred to the same criteria and determined the transition in the etiologies of LC in Japan. Our results suggested a recent change in the etiology of LC, with a decreased contribution of hepatitis virus-associated LC and an increased contribution of non-viral LC.
Regarding the etiologies of LC (Fig. 3), the total ratio of viral hepatitis-related LC was > 60%. When our results were compared to those in 2008, HBV-related LC and HCV-related LC decreased from 13.9 to 11.5% and 60.9 to 48.2%, respectively. The ratio of non-viral LC increased, the ratio of LC due to ALD increased from 13.6 to 19.9%, and that of NASH-related LC also increased from 2.1 to 6.3%. In addition, cholestasis-related LC increased slightly from 2.7% (PBC 2.4% and other cholestasis 0.3% in 2008) to 3.4%, and AIH also showed a small increase from 1.9 to 2.7%. These findings suggested a decreased ratio of viral hepatitis-related LC and an increased ratio of non-viral LC in Japan.
The data of the present study were generally consistent with the results of the 2008 survey. Regarding the LC patients diagnosed before 2008 in the current study, the ratios of HBV-related LC (13.7%), HCV-related LC (58.6%) and NASH-related LC (2.0%) were similar to those of the previous survey (HBV-related LC, 13.9%; HCV-related LC, 60.9%; NASH-related LC, 2.1%) ; however, the ratio of cryptogenic LC patients diagnosed before 2008 (5.6%) was higher relative to the 2008 survey (3.0%) . One possible explanation is that differences in the participating institutions influenced the results. Additionally, we have to note that each survey had its own classifications for LC patients, particularly for non-viral LC patients. The survey in 2008 classified non-viral LC into 10 categories, while the current study classified non-viral LC into nine etiologies. Thus, methodological differences between the two surveys might have somehow been involved in the discrepancy.
Regarding the geographic differences in the etiologies of LC patients, HCV-related LC remained the leading cause of LC in all areas (Fig. 4), and the characteristics of the etiologies in each area were generally in agreement with the previous survey. However, some differences were observed between the 2008 and 2018 surveys. In our results, the ratio of HBV-related LC seemed to be high in the ‘Kanto’ area. Additionally, the ratio of ALD-related LC was high in the ‘Kyushu’ area. We should again mention the methodological differences between the surveys. In the previous survey, the ‘Chubu’ area was divided into the ‘Hokuriku’ and ‘Chubu’ areas. In addition, ‘Tokyo’ had been categorized as one independent area separately from the ‘Kanto’ area, while ‘Okinawa’ prefecture was considered separately from the ‘Kyushu’ area. Similarly to the above-mentioned discrepancy regarding the etiologies of non-viral LC, some methodological differences may have caused the different results.
Overall, the comparison of the 2008 and 2018 surveys suggested a decreased ratio of viral hepatitis-related LC, particularly HCV-related LC and an increased ratio of non-viral LC, including NASH-related LC. Our results were consistent with those of previous reports that have suggested that non-viral liver diseases are increasingly contributing as etiologies of chronic liver diseases, such as the changing etiologies of hepatocellular carcinoma in Japan as well the United States of America [17,18,19]. To represent the recent real-world data, all data used in this study were obtained from geographical regions or areas throughout Japan. Even with the presence of some slightly different data due to methodological differences, we believe that the comparison of the results among multiple surveys could be beneficial for assessing the transition in the etiologies of LC. However, we also consider that caution is required when interpreting the results of comparisons among multiple surveys.
One characteristic point of the current survey was that nationwide experts referred to the same criteria and the transition in the etiology of LC patients were evaluated according to the defined standards. In addition, we assessed the change not only in the distribution of etiologies (Fig. 5) but also in the real numbers of new LC patients (Figs. 6 and 7). Regarding the data on the proportions of etiologies (Fig. 5), the proportions of non-viral LC, including ALD-related and NASH-related LC, seemed to have increased in the last decade. However, the results did not demonstrate an increase in the number of non-viral LC patients in Japan. According to the survey on the real numbers of patients (Fig. 6), the number of patients who were newly diagnosed with non-viral LC increased, while the number of patients newly diagnosed with viral hepatitis-related LC, especially HCV-related LC, decreased over time, suggesting that both the decrease in the viral hepatitis-related LC and the increase in non-viral LC contributed to changing the distribution of the LC etiologies in Japan (Fig. 5). Although some papers that evaluated the real numbers of LC patients reported survey results in specific regions of Japan or China [20, 21], the current cohort is the largest yet and describes the real-world nationwide data in Japan.
As described in “Introduction”, Japan has a high prevalence of viral hepatitis, and even before the establishment of the GHSS on viral hepatitis, various measures for the prevention and treatment of viral hepatitis were intensively implemented under national programs . We consider that various national strategies implemented since 2008 as well as advances in antiviral treatments have helped to reduce the rate of hepatitis virus-related disease, since HCV elimination due to antiviral treatments, such as interferon or direct-acting antivirals (DAAs), could prevent progression to LC in HCV-infected patients. HCV infection was widespread around 60 years ago in Japan; thus, Japanese patients have been suffering from this disease for longer in comparison to patients in other countries, suggesting that the transition of Japanese patients with HCV-related liver diseases could help to identify future global trends in HCV-related diseases . In addition, HCV-infected patients tend to be elderly in Japan, and the number of patients is expected to decrease due to the improved antiviral treatment in combination with a natural reduction over time . Our real-world data demonstrating a reduced contribution of HCV infection to the etiology of LC in Japan may help predict the future trends in other countries.
The present study was associated with several limitations. First, this survey was conducted according to two JSH-edited books. Although HBV and/or HCV infection could be serologically diagnosed, the diagnostic criteria for other non-viral liver diseases, such as NASH, were not completely unified. In addition, the guideline was published for a wide range of physicians. Thus, the classification was determined according to the clinical diagnosis, and histological assessment was not mandatory in the current survey. Although our results were generated under the standards defined by hepatologists, care should be taken when applying our results to other countries, especially for data regarding the etiology of non-viral LC. Second, the current survey was focused on the changing trends in the etiology of LC and we did not request detailed clinical data. This simplicity was advantageous for accumulating a large number of LC patients; however, we could not scrupulously report the clinical features of LC patients. Regarding the age of LC patients, since the data available in the current survey were limited to the mean age of LC patients from each institute, we could only show the increasing trend in the mean age (Supplementary Fig. 2), even though analyzing the background characteristics, such as sex and mean age of the LC patients, according to the etiology would have been more useful. Third, the retrospective data were obtained directly from the participating institutions, and older data were difficult to retrieve in a number of institutions. Our clinical data with the real number of patients (Figs. 6 and 7) were obtained from 36 hospitals without any missing or incomplete data. Although these hospitals were located in all areas of Japan, it should be noted that the results were obtained from a limited number of enrolled patients (N = 18,358; Fig. 2). In addition, the present study employed a retrospective design. The transition of patient numbers according to the etiology should be prospectively examined among the same institutions. We analyzed the data of the 36 hospitals that were able to provide the number of LC patients newly diagnosed each year from 2008 to 2016 without missing any years because these hospitals were considered to have continuously accumulated their own data independently of the current survey, suggesting that their data were simulated the prospective cohort to some extent. However, our results were not equivalent to prospective data.
In summary, we showed the survey results regarding the transition in the etiologies of LC in Japan. Our results suggested a decreased ratio of hepatitis virus-associated LC, particularly HCV-related LC and an increase in the ratio of non-viral LC, including NASH-related LC in the real world.