Quiet reflection after effective action: hepatitis B viraemia and hepatocellular carcinoma recurrence


‘Follow every effective action with quiet reflection. From these quite reflections even greater actions emerge’

(Peter Drucker)

Hepatocellular carcinoma (HCC) represents a major health care challenge in the present era. Once diagnosed, the therapeutic options are limited in availability or wanting in efficacy. Although liver transplantation remains the best available option, it often cannot be availed because patients either present late in the course of their disease or outgrow transplantation criteria while waiting on the list. Lately, locally ablative therapies have gained immense popularity and have shown survival benefit in selected patients but are viable only in those with relatively small tumours (1). Similarly, transarterial chemoembolization (TACE) has shown a significant survival benefit but its usefulness is limited by modest efficacy rates and restriction to patients with early cirrhosis (2).

Although resection remains a useful treatment preference in non-cirrhotic patients with HCC (common in Asia), in the setting of cirrhosis it is used in <10% of patients. This is due to the presence of advanced liver cirrhosis and portal hypertension in the vast majority, which in turn exposes them to a high risk of decompensation after surgery (3). Moreover, after resection, recurrence of liver tumour can be expected in as many as 70% of patients within 5 years (4).

What alternatives do we have then that could make resection available for a wider range of patients? There are two manifest options – the first is to improve surgical techniques and develop better ways to select resection candidates, thereby condensing post-operative decompensation, and the second would be to reduce the recurrence rates after resection.

But do we really know who the best candidate for resection is? Although surgeons frequently use the Child–Pugh (CP) classification to assess liver function and perform resections in non-cirrhotic patients, CP class A and ‘good’ CP class B patients, it has become clear, both from clinical practice and from evolving evidence, that this system does not allow for accurate categorization of patients. Alternatively, two parameters have shown to be most predictive of decompensation post-liver resection including signs of clinically relevant portal hypertension and bilirubin >1 mg/dl. In the absence of these factors, most patients will not experience decompensation and survival may be as high as 70% at 5 years. On the other hand, survival drops to <50% at 5 years in the presence of clinically significant portal hypertension, and when both portal hypertension and elevated bilirubin are present, survival fares not better than 30% at 5 years (5). In addition, with better perioperative techniques, major hepatic resection has become a less morbid procedure and operative mortality in cirrhotic patients declines to below 5% (4).

After better patient selection and improved surgical care, our next goal is to prevent HCC from developing again in the remnant cirrhotic liver. Towards this, controlling the causative agent becomes the natural strategy. The landmark Risk Evaluation of Viral Load Elevation and Associated Liver Disease/Cancer-Hepatitis B Virus (REVEAL-HBV) study, which prospectively followed 3653 Taiwanese patients with HBV for over 11 years, showed that the incidence of HCC increases with HBV viral load in a dose–response relationship (6). While non-cirrhotic patients with normal alanine aminotransferase (ALT) were shown to be at significant risk, it was in those with persistently elevated HBV DNA levels during follow-up who exhibited the highest risk for HCC. Albert Einstein once said ‘Not everything that can be counted counts and not everything that counts can be counted’. Indeed, it is obvious now that counting serum HBV DNA certainly counts!

If the presence of cirrhosis and the level of HBV DNA are the most important epidemiological factors for the development of HCC, could sustained viral suppression reduce the incidence of HCC? A meta-analysis of seven non-randomized studies investigating the effect of interferon therapy in HBV patients showed a significant reduction in the incidence of HCC (7). To this effect, more impressive results were seen with lamivudine. In a trial involving advanced HBV-related cirrhotics, 651 patients were randomized to receive lamivudine or placebo for 5 years. The study required early termination because of a marked reduction in mortality and achievement of study end points (8). The incidence of HCC was 7.4% in the placebo arm vs 3.9% in the lamivudine arm.

The question that follows then is whether viral suppression can reduce the recurrence of HCC post liver resection? In this issue of Liver International, Kim and colleagues report the effect of persistent HBV replication on tumour recurrence rates in 230 HCC patients undergoing curative resection. However, those who had received prior antiviral therapy, had fluctuating HBV DNA levels or had tumour recurrence within 12 month of resection were excluded. After a median follow-up of 35 months, the non-viraemic patients had a lower 5-year cumulative recurrence rate (54.7%) compared with those who were viraemic (72.9%; P=0.043). Moreover, in the multivariate analyses, sustained viraemia was independently associated with increased risk of recurrence. In terms of disease-free survival, the non-viraemic patients had significantly higher 1, 3 and 5-year disease-free survival rates compared with the viraemic group, although there was no significant difference in the overall survival. It is well established that tumour size, tumour number, vascular invasion, α-foetoprotein level, as well as other histological factors are important in determining the recurrence of HCC post-resection (9). However, very limited data are available on the importance of HBV viral load in tumour recurrence (10, 11). Most studies have only considered HBV DNA levels at one point in time (usually at the time of surgery) while the study by Kim et al. is the first to study the effect of sustained viraemia measured on several occasions, on tumour recurrence.

Nevertheless, a few issues in this important study oblige a proper perspective. Primarily, although the authors have excluded patients with recurrence in the first year after resection reasonably assuming that these are not true recurrences but occult subclinical intrahepatic metastases that were initially not detected, it is well known that such recurrences may appear as long as 3 years after resection (12). Therefore, by no means can one be certain that these were true de novo recurrences. Likewise, the authors excluded 22 patients with fluctuating HBV DNA levels with the clear intention of separating non-viraemic patients from persistently viraemic ones. However, by excluding such patients from the analysis the authors may have forfeited an opportunity to assess whether this subset of HCC patients behaved differently from those with persistently viraemic or non-viraemic ones.

Additionally, a major concern about this study is the fact that cirrhotic patients with active viraemia were not treated with antiviral therapy despite clear recommendations to the contrary (3). In defence, the authors state that at the time of study enrolment antiviral therapy in Korea was not available to patients with elevated HBV DNA levels if the ALT levels were not more than two times the upper limit of normal. It leaves us, therefore, with a festering doubt whether the treatment of such patients would have affected the overall survival analysis.

Furthermore, the cut-off HBV DNA level used for treatment initiation in this study was 105 copies/ml. The utilization of more sensitive polymerase chain reaction-based assays could have given the authors the contingency to differentiate between different levels of HBV DNA and their effect on recurrence and clinical outcomes. In addition, recent data suggest that a cut-off value of 104 copies/ml or even lower may be more accurate in predicting clinical outcomes (3). This would have helped us appreciate whether an additional antiviral agent is required to further suppress HBV DNA after reduction from levels >105 copies/ml. In turn, would such consummate suppression of the viral load translate into lesser HCC recurrences or perhaps demonstrate an outcome that is in between non-viraemic and persistently viraemic patients?

Finally, despite demonstrating the significance of HBV DNA on tumour recurrence and disease-free survival rates, the authors were unable to demonstrate a beneficial effect on overall survival. Certainly, a longer duration of follow-up could have better delineated the survival benefit in these patients. Interestingly, of the 54% of the viraemic group who experienced recurrence, the vast majority of these received chemotherapy, which perhaps exacerbated the underlying HBV disease, and thereby contributed to the mortality. Unfortunately, these patients received antiviral therapy only after HBV disease reactivation. Again, it cannot be forsaken that better results can be anticipated with pre-emptive treatment of HBV at much lower viraemia levels.

Other therapies are emerging that may be useful to reduce recurrence after resection. In this context, the effects of systemic chemotherapy and perioperative TACE have been studied but not shown benefit (13). Potentially effective adjuvant therapies such as internal radiation (14), adoptive immunotherapy (15) and acyclic retinoids (16) are being examined. Sorafenib, a tyrosine kinase inhibitor, has demonstrated significant survival advantage in advanced HCC (17) and has obtained approval from the Food and Drug Administration but it remains to be seen whether its combination with antiviral therapy would provide added benefit for those patients undergoing resection.

Meanwhile, the study in hand has shown us clearly that suppression of HBV DNA in HBV-infected patients seems to be significantly effective in reducing recurrence rates after tumour resection. Many questions remain unanswered and the impact of HBV treatment to prevent disease sequelae will only be clearly appreciated once further data evolve from clinical studies performed in the light of the recently updated HBV treatment guidelines. Until then it is the learners who will inherit the future.

This Editorial should have been published in the March 2008 issue alongside the following paper:

Kim BK, Park JY, Kim DY, et al. Persistent hepatitis B viral replication affects recurrence of hepatocellular carcinoma after curative resection. Liver Int 2008; 28: 393–401.

The Editor apologizes for this omission.