Pancreatic cancer in chronic hepatitis B

In this issue of the journal Iloeje et al. (1) report on the unexpected finding of an increased risk of pancreatic cancer in patients with chronic hepatitis B (CHB). These data come from a large scale prospectively designed longitudinal cohort study in Taiwan looking at various population health parameters. This is the same population that constituted the REVEAL study population (2). The REVEAL study was an analysis of the incidence of hepatocellular carcinoma (HCC), cirrhosis and liver disease outcomes in patients with CHB. (As an aside, it is worth noting that public health authorities in Taiwan have sponsored or undertaken numerous studies looking at the health of their population. They have extensive and properly managed databases from which they can draw data such as that presented here. Few other countries can match this kind of epidemiological data.) The population health study recruited nearly 24 000 subjects, and followed them prospectively for more than 10 years. Data on the relationship between hepatitis B viral replication and the development of cirrhosis and HCC has already been published (2–4). The cohort that made up this study on pancreatic cancer was slightly different than the REVEAL cohort. The overall study population for the REVEAL study was about 23 800, whereas for the pancreatic cancer study nearly 2000 subjects were excluded because of confounding factors. These exclusions reduced the number of patients with hepatitis B to 3930, still a large enough cohort to provide meaningful data. Patients with pancreatic cancer were identified by cross linkage with national death records, again testament to the excellence of health records in Taiwan. There were 48 cases of pancreatic cancer, with an incidence of 12.7/100 000/year among those without hepatitis B infection, and 20.4/100 000/year among those with hepatitis B. This gave an odds ratio of 1.95, i.e. those with hepatitis B were about twice as likely to develop pancreatic cancer as those who were not infected. However, the absolute risk increase was only 0.08%. This is a very small increase in absolute risk. The increased risk of pancreatic cancer was confined to those who were under age 50 at recruitment. Does this mean that older hepatitis B carriers get pancreatic cancer less often than younger subjects? This is unlikely, because for all cancers related to infection the longer the duration of infection the higher the incidence of cancer. A more likely explanation is that competing causes of death in the over 50s, primarily deaths related to liver disease in those infected with hepatitis B, were responsible for the apparent decrease in risk of pancreatic cancer. Patients at risk for pancreatic cancer died of their liver disease before the pancreatic cancer could strike. This study was conducted in Taiwan, a country in which hepatitis B is endemic. The authors carefully evaluated the possibility that the epidemiology of pancreatic cancer was different in Taiwan than elsewhere, and that these results might have been influenced by other factors that are not present, e.g. in North America or Europe. The authors addressed this point by comparing pancreatic cancer incidence in their study with that in North America. The overall incidence in their study in the hepatitis B uninfected was similar to the incidence in the USA, lending credence to the results in hepatitis B. The association between hepatitis B infection and pancreatic cancer has been assessed previously (5, 6). The first study was part of a large cohort study in Korea, in which the risk of pancreatic cancer was similar in those with and without hepatitis B infection. This study was subject to criticism, in that patients with liver disease were excluded, so that the contribution that liver disease rather than hepatitis B infection might confer could not be tested. Secondly, the hepatitis B status was known for only about two-thirds of the subjects, and thirdly, this study included all pancreatic neoplasms, including neuroendocrine tumours. In a second study an association was identified for patients with a prior infection with hepatitis B, as determined by the presence of anti-HBc. None of the patients with cancer of the pancreas in this study were positive for HBsAg. This was a retrospective case–control analysis, in which anti-HBc was compared in patients with pancreatic cancer and matched controls. The adjusted odds ratio was 2.5, indicating that there was a statistically significant association between anti-HBc and the pancreatic cancer. Thus, does the current study end the debate? Probably not, nor is a similar study likely to be conducted in the future. However, since the absolute increase in risk of pancreatic cancer is so small, and the findings are unlikely to ever change clinical practice, the lack of certainty about this finding is not something to lose sleep over. This study does raise some questions. Are the findings specific to pancreatic cancer, or is there a general increase in non-liver cancer in patients with CHB. This has not so far been documented for any other cancers other than pancreas. The second question that this study raises is whether there is an increased risk of non-liver malignancies only in hepatitis B, or whether the same phenomenon can be seen in other liver diseases. El-Serag et al. (7) evaluated risk of cancer of the pancreas and intrahepatic cholangiocarcinoma in patients with chronic hepatitis C. They found that after adjusting for confounders there was no association between hepatitis C infection and pancreatic cancer, but there was an increased risk of intrahepatic cholangiocarcinoma of about a 2.5-fold. Others have also shown that intrahepatic cholangiocarcinoma is more prevalent in patients with cirrhosis than in the normal population (8). How might hepatitis B virus (HBV) infection increase the risk of developing pancreatic cancer? There are several possible scenarios. There may be direct infection of pancreatic tissue, or the state of chronic liver disease might somehow alter intracellular metabolism in the pancreatic cells, making malignant transformation more likely. Hepadnaviral DNA can be found in pancreatic tissue. This was shown with both duck hepatitis B and woodchuck hepatitis virus in their respective hosts (9–13). The presence of viral RNA or single stranded viral DNA suggested that HBV was replicating in these tissues. There is also data in humans indicating that HBV antigens and viral DNA can be detected in pancreatic tissue. Replicative forms of HBV RNA were identified in the pancreas in a patient with fulminant hepatitis B (14), and integrated HBV DNA has been demonstrated in the pancreas of a hepatitis B carrier in whom the pancreas was histologically normal (15). Partially double-stranded DNA found in this case might have been from contaminating blood, rather than from the pancreas itself, and therefore does not conclusively indicate the presence of viral replication in the pancreas. There are several possible mechanisms whereby HBV could cause pancreatic cancer. If the mechanism of malignant transformation was by direct infection, one would expect some evidence of cellular injury. All other cancers associated with infectious organisms (HCC, cervical cancer, gastric cancer) are associated with injury to the target tissue. There does not seem to be sufficient evidence to incriminate this mechanism. Another possibility is that HBV DNA integrated into the cellular genome in pancreatic cells might be directly oncogenic, either by disrupting the function of one or more anti-oncogenes, or stimulating the activity of one or more oncogenes. Findings in man have suggested the presence of integrated viral DNA in the pancreas (14). However, nothing more is known about this. These findings of this study raise the question whether patients with chronic liver disease should be screened for malignancies other than HCC. As far as pancreatic cancer is concerned, at least in some patients the pancreas can be visualized on ultrasound, although ultrasound is not the best method of imaging the pancreas. This study reported in this issue of the journal found a relative risk of about 2. The absolute increase in risk was only 0.08%. Thus, 12 500 subjects would have to be screened each year to find a single pancreatic cancer. Clearly, this does not justify any additional screening over what is already being performed as screening for liver cancer. These results are provocative, but the magnitude of increased risk is too small to warrant any change to the way patients are evaluated or managed.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.