Trend in the prevalence of HBsAg positivity
Within the framework of this survey program of our society, HBsAg determination was first performed in 1971. In that investigation, HBsAg was called “Australia antigen (Au antigen).” This survey represented the first investigation of the epidemiological background of individual patients within the framework of this survey program [12]. According to the report on that survey, the total number of patients undergoing maintenance hemodialysis in Japan was 1994 as of the end of 1971. Of these patients, 1826 were undergoing hemodialysis using either the Kiil-type or Kolf-type dialyzer. The patients, both males and females, were predominantly from the age group of 26–30 years. Information on Au antigen (HBsAg) collected from 1055 of these patients revealed 103 positive cases and 952 negative cases, corresponding to a prevalence of HBsAg positivity of 9.76% [12] (prevalence of Au antigen positivity = number of Au antigen-positive patients ÷ [number of Au antigen-positive patients + number of Au antigen-negative patients]). After the investigation in 1971, the next investigation of HBsAg was conducted 28 years later, in 1999 [5]. The prevalence of HBsAg positivity in 1999 was 2.09%, which was much lower than the rate of 9.76% recorded in 1971.
From 1999 to 2007, the prevalence of HBsAg positivity remained almost unchanged at about 2% (Fig. 1). In the current survey conducted in 2018, which was 11 years after the last investigation, the prevalence of HBsAg positivity was 1.38%; this prevalence was markedly lower than the last recorded data in 2007. Exponential function regression was performed by plotting the prevalence of HBsAg positivity among all the dialysis patients at 5 time-points during the 1999–2007 period along the y-axis and the year along the x-axis, and the prevalence of HBsAg positivity (1.38%) in 2018 was considerably lower than the rate predicted by the regression formula (Fig. 2). This result indicates that the prevalence of HBsAg positivity among dialysis patients in Japan improved in a non-continuous manner during the 11-year period from 2007 to 2018. The reason for this change remains unexplained. Regarding the general population in Japan, as of the year 2000, the prevalence of HBsAg positivity was reportedly highest (1.37%) among subjects born between 1946 and 1950 and was lower in other age groups [13]. The cohort born between 1946 and 1950 was aged 50 to 54 years old in 2000. The age of this cohort as of 2018 was 68 to 72 years, which overlaps with the 70–74 age group (which had the highest proportion of positivity among the dialysis patients in the 2018 survey). The prevalence of HBsAg positivity among dialysis patients seems to be higher than the rate in the general population, even as of 2018, although we cannot arrive at a definitive conclusion about this tendency because of variances in the age distributions and years of investigation.
Dialysis vintage and the prevalence of HBsAg positivity
An analysis of the relationship between the dialysis vintage and the prevalence of HBsAg positivity revealed that the prevalence of HBsAg positivity was lowest in the group with a dialysis vintage of less than 2 years, with the rate increasing steadily with increases in the dialysis vintage thereafter (Fig. 3). Figure 12 graphically represents the relationship between the dialysis vintage and the prevalence of HBsAg positivity in the 2007 survey, enabling a comparison of the 2007 survey data with the data from the current survey [9] (Supplementary Table 9). Overall, the prevalence of positivity was higher in 2007 than in 2018. However, the results of the 2007 survey revealed that the prevalence of positivity increased gradually as the dialysis vintage increased, similar to the tendency observed in the 2018 survey.
The group of patients covered in the 2007 survey had started receiving maintenance dialysis in or before 2007. The group of patients with a dialysis vintage of 10 years or longer covered in the 2018 survey was initiated on maintenance dialysis beginning in or before 2008. These two groups of patients may thus be considered as constituting approximately the same cohort. A comparison of the prevalences of HBsAg positivity in the 2007 and 2018 surveys, paying close attention to the aforementioned patient cohort, revealed that the prevalence of positivity in the 2018 survey was lower than that in the 2007 survey. This trend was more marked among patients with a longer dialysis vintage. In other words, the prevalence of positivity in the group of patients with a dialysis vintage of 10–25 years in the 2018 survey accounted for 85% of the prevalence of positivity in the group of patients with a dialysis vintage of 0–15 years in the 2007 survey. These two groups of patients constitute approximately the same patient cohort. The prevalence of positivity in the group of patients with a dialysis vintage of 40 years or more in the 2018 survey accounted for about 63% of the prevalence of positivity in the group of patients with a dialysis vintage of 30 years or more in the 2007 survey. These two groups also constituted approximately the same patient cohort. These results may be interpreted as suggesting a tendency towards seroconversion (from HBsAg-positive to HBsAg-negative) or dropout because of death during the 10-year interval between the two surveys.
In the survey conducted in 2018, the group of patients with a dialysis vintage of less than 10 years (who were initiated on maintenance dialysis in or after 2009) showed a tendency towards a gradual increase in the prevalence of HBsAg positivity as the dialysis vintage increased (Fig. 3). This suggests that, even at present, a certain number of patients newly develop hepatitis B infection after the initiation of hemodialysis.
Age and the prevalence of HBsAg positivity
In patients between the ages of 65 and 74 years, the prevalence of HBsAg positivity was highest. This result may be related to the tendency of the patients between the ages of 65 and 74 years to have a longest dialysis vintage (Supplementary Table 10).
Treatment method and the prevalence of HBsAg positivity
The prevalence of HBsAg positivity was highest among hemofiltration patients. However, the number of hemofiltration patients was very small (only 12), so this value should be interpreted with caution. The prevalence of HBsAg positivity among blood adsorption dialysis patients was second highest. This result may be related to these patients having very long dialysis vintages (Supplementary Table 11). Of note, the prevalence of HBsAg positivity among hemodiafiltration patients was lower than that among hemodialysis patients, even though the dialysis vintage of the hemodiafiltration patients tended to be longer than that of the hemodialysis patients. The dialysis vintage of peritoneal dialysis patients is considerably shorter than that of hemodialysis patients. However, the prevalence of HBsAg positivity among peritoneal dialysis patients was nearly equal to that among hemodialysis patients, even though peritoneal dialysis patients do not require extracorporeal circulation.
Trend in the prevalence of HCV-Ab positivity
When an exponential function regression analysis was conducted, with the prevalence of HCV-Ab positivity recorded at 7 time-points during the 1999–2007 period plotted along the y-axis and the year plotted along the x-axis, the prevalence of HCV-Ab positivity tended to decrease steadily over time, and the reported prevalence of HCV-Ab positivity in 2018 (4.7%) was approximately equal to the rate predicted by the regression curve (Fig. 5). This finding may be interpreted as suggesting that the prevalence of HCV-Ab positivity among all dialysis patients has continued to decrease at an approximately constant pace during the past 2 decades. However, we cannot rule out the possibility that this interpretation is erroneous for the reason discussed in the next section.
Dialysis vintage and the prevalence of HCV-Ab positivity
When the prevalence of HCV-Ab positivity was analyzed according to dialysis vintage, the prevalence of positivity remained almost unchanged at 4.0–4.7% when the dialysis vintage was less than 25 years, but the prevalence of positivity increased linearly as the dialysis vintage increased after 25 years (Fig. 6). The year 1993 was 25 years before 2018. The result mentioned above suggests that the prevalence of HCV-Ab positivity did not increase with time among the patients who began receiving maintenance dialysis around or after 1993. As reference data, Fig. 13 shows the relationship between the dialysis vintage and the prevalence of HCV-Ab positivity in the previous survey performed in 2007 [9] (Supplementary Table 12). In the survey conducted in 2007, the prevalence of positivity remained almost unchanged at 7.6–7.9% among patients with a dialysis vintage of less than 15 years, but it increased linearly as the dialysis vintage increased beyond 15 years. The year 1992 was 15 years before 2007. This result also suggests that the prevalence of HCV-Ab positivity did not increase among the patients who began receiving maintenance dialysis around or after 1992. Hepatitis C virus was discovered in 1989. After the discovery of this virus, the incidence of hepatitis C associated with blood transfusion decreased sharply [14, 15]. The treatment of renal anemia using an erythropoietin preparation in dialysis patients began to be covered by the national health insurance program in Japan in 1990 and transfusion therapy for anemia in dialysis patients decreased sharply thereafter. Consequently, the decrease in the prevalence of HCV-Ab positivity after the year 1992 or 1993 could be interpreted as reflecting these improvements in the treatment of anemia in dialysis patients.
As described in the section on HBsAg, the group of patients with a dialysis vintage of 10–24 years covered by the 2018 survey can be considered being approximately equivalent to the group of patients with a dialysis vintage of 0–14 years covered by the 2007 survey. The prevalence of HCV-Ab positivity in this group was 4.1–4.7% in the 2018 survey, which was about 2/3 of the rate of 7.6–7.8% estimated in the corresponding group in the 2007 survey. In general, patients are unlikely to test negative for HCV-Ab once they have tested positive. Here, we may assume that the low prevalence of HCV-Ab positivity in the current survey can be attributed to a reduction in HCV-Ab-positive patients because of death. If this assumption is valid, the prevalence of positivity could be considered as showing a tendency towards a gradual reduction with increasing dialysis vintage. However, the survey results did not indicate such a trend. Furthermore, the prevalence of positivity in the group of patients with a dialysis vintage of less than 10 years was also 4.0–4.5%, which was approximately equal to the rate in the patient group with a dialysis vintage of 10–24 years. We may therefore judge that the dropout of HCV-Ab-positive patients because of death was not responsible for the lower prevalence of HCV-Ab positivity in the current survey in the group of patients that was initiated on dialysis in or after 1992–1993, compared with the prevalence of positivity documented in the same group in the previous survey conducted in 2007. The reason for the lower prevalence of HCV-Ab positivity in the current survey, compared with that in the 2007 survey, remains uncertain. One potential explanation is that the methods used for HCV-Ab testing have changed during the last decade. HCV antibody testing methods can be classified into three generations (first, second, and third) depending on the region of the viral antigen used for antibody detection. Later generations of testing methods detect a wider range of antigens. The latest 3rd generation inspection methods were developed in the late 1990s, and these 3rd generation inspection methods became widespread in the 2000s. In general, the difference in detection sensitivity between the 2nd generation test method and the 3rd generation test method for HCV antibody is thought to be minimal [16]. However, the difference between the prevalence of HCV-Ab positivity in the 2007 survey and that in the 2018 survey may partly reflect the difference in detection sensitivity between 2nd generation and 3rd generation HCV antibody test methods.
As mentioned in the preceding section, the prevalence of HCV-Ab positivity among all dialysis patients included in the 2018 survey was approximately equal to the rate predicted from the changes over time for the prevalence of HCV-Ab positivity during the 1999–2007 period. If we assume that the prevalence of HCV-Ab positivity at present is, for some reason or other, equivalent to about 2/3 of the prevalence of HCV-Ab positivity recorded with the testing method available in 2007, the prevalence of positivity in 2018 as predicted based on the changes over time during the 1999–2007 period should be about 2/3 of the predicted value, i.e., equal to about 3%. In practice, however, the prevalence of positivity recorded in 2018 was 4.7%. This could be interpreted as suggesting that the magnitude of the reduction in the prevalence of HCV-Ab positivity among dialysis patients overall in the 2018 survey was not as large as the reduction predicted by the changes in the prevalence of HCV-Ab positivity during the 1999–2007 period.
Age and the prevalence of HCV-Ab positivity
Similar to the prevalence of HBsAg positivity, in patients between the ages of 65 and 74 years, the prevalence of HCV-Ab positivity was highest. This result may be related to the tendency of the patients between the ages of 65 and 74 years to have a longest dialysis vintage (Supplementary Table 10).
Treatment method and the prevalence of HCV-Ab positivity
The prevalence of HCV-Ab positivity was extremely high among patients undergoing blood adsorption dialysis. This observation is probably related to the very long dialysis vintages of these patients, since the prevalence of HCV-Ab positivity was significantly higher among patients with a dialysis vintage of over 25 years (Fig. 6, Supplementary Table 4). Of note, unlike the prevalence of HBsAg positivity, the prevalence of HCV-Ab positivity in peritoneal dialysis patients was much lower than the prevalence of HCV-Ab positivity among hemodialysis patients. The dialysis vintage of peritoneal dialysis patients is considerably shorter than that of hemodialysis patients. And as mentioned before, the prevalence of HCV-Ab positivity increased much more than that of HBsAg positivity along with the dialysis vintage. These situations may have affected the difference in the prevalences of HCV-Ab and HBsAg positivity among the patients.
HCV-Ab and HCV-RNA
An analysis of the distribution of HCV-Ab and HCV-RNA among dialysis patients revealed that 92.6% of the patients tested negative for both HCV-Ab and HCV-RNA, while 2.7% tested positive for both (Fig. 7). As reference data, Fig. 14 shows the distribution of HCV-Ab and HCV-RNA in the survey conducted in 2007 (analyzing 63,098 patients; Supplementary Table 13). In 2007, 85.6% of the patients tested negative for both HCV-Ab and HCV-RNA, while 9.0% tested positive for both. Thus, the negativity rate for both antibody and RNA was 7.0 percentage points higher in 2018 than in 2007, while the prevalence of positivity for both was 6.3 points lower in 2018 than in 2007. Thus, one can say that the prevalence of positivity for both antibody and RNA had decreased in 2018 to 1/3 or lower, compared with the rate recorded in 2007. For a long time, no treatment was available for dialysis patients with hepatitis C. In 2015, however, treatment with direct-acting antivirals (DAAs) began to be covered by the national health insurance program in Japan. At present, the treatment of hepatitis C using DAAs in dialysis patients is also covered by the national health insurance program. The clinical use of DAAs for the treatment of hepatitis C may have contributed to the sharp drop in the percentage of patients who tested positive for both HCV-Ab and HCV-RNA during the 11-year period from 2007 to 2018. To test the validity of this hypothesis, we conducted the following analysis.
In the current survey, HCV-RNA data was available for 50.2% of the HCV-Ab-positive patients, while similar data was available for only 32.7% of the HCV-Ab-negative patients. These results suggest that HCV-RNA measurements are less likely to be conducted for HCV-Ab-negative patients. Thus, the difference in the percentage of patients who have undergone HCV-RNA testing among HCV-Ab-negative patients might have caused a bias in the prevalence of HCV-RNA positivity. In other words, if the percentage of patients undergoing HCV-RNA determination among the HCV-Ab-negative patients increased, the number of HCV-RNA-negative patients would also increase, causing a bias (reduction) in the prevalence of HCV-RNA positivity among all the patients. To avoid such a bias, we calculated the prevalence of HCV-RNA positivity among the HCV-Ab-positive patients alone in each survey year and analyzed the changes over time during the 8-year period in which the HCV-Ab and HCV-RNA statuses were determined (Fig. 8) [5,6,7,8,9,10,11]. The rate increased each year until 2003 but then began to decrease from 2006 onwards. Figure 15 shows a scatter diagram, with the prevalence of HCV-RNA positivity plotted along the y-axis and the year plotted along the x-axis. Because a tendency towards a reduction in the prevalence of HCV-RNA positivity began to be noted in 2006, it was difficult to judge based on this illustration alone whether the prevalence of positivity in 2018 was lower than the rate anticipated from the trend noted before 2007, i.e., whether the start of the clinical application of DAAs in 2015 caused a decrease in the number of HCV-RNA-positive patients.
ALT (GPT)
An analysis of the serum ALT level in HBsAg-positive and HBsAg-negative patients revealed that the mean serum ALT level was slightly lower in the HBsAg-negative patients (14.57 IU/L) than in the HBsAg-positive patients (15.49 IU/L); however, large differences in the distribution of patients with low and high serum ALT levels were not seen between the two groups (Fig. 9). In a similar analysis of the serum ALT level conducted among HCV-Ab-positive and HCV-Ab-negative patients, the mean serum ALT level was slightly lower in the HCV-Ab-negative patients (14.51 IU/L) than in the HCV-Ab-positive patients (16.36 IU/L), although large differences in the distribution of patients with high and low serum ALT levels were not seen between the two groups (Fig. 10). These results indicate that neither the HBsAg status nor the HCV-Ab status exerted a significant impact on the serum ALT level in dialysis patients.
When the serum ALT levels in HCV-RNA-positive and HCV-RNA-negative patients were analyzed only among HCV-Ab-positive patients, the mean level was higher in the HCV-RNA-positive patients (19.09 IU/L) than in the HCV-RNA-negative patients (16.60 IU/L), and an analysis of the patient distribution revealed a lower percentage of patients with a serum ALT level of less than 10 IU/L and a higher percentage of patients with a serum ALT level in the range of 15–59 IU/L in the HCV-RNA-positive group, compared with the HCV-RNA-negative group (Fig. 11). These findings indicate that seropositivity for HCV-RNA among the HCV-Ab-positive patients tended to be associated with hepatic impairment.