• Potential conflict of interest: Nothing to report.


The letter by Drs. Aslinia and Mindikoglu raises several points that needed to be clarified.

  • 1We included in our survey1 only patients with cirrhosis and hepatocellular carcinoma (HCC) who have been treated by radiofrequency ablation (RFA) for a first and previously untreated tumor. Therefore, the follow-up after treatment starts at the date of the first RFA session.
  • 2Articles reporting that 30% of patients with alcoholic liver diseases were infected by hepatitis C virus (HCV) date back to the period when HCV serology testing was introduced. This high percentage resulted from false positive serological results and under-recognition of the role of HCV infection in patients with high alcohol intake. In our study, all patients with positive HCV antibodies were also positive for HCV RNA, and we considered them as bearing HCV cirrhosis despite the fact that some had been heavy drinkers. The outcome of this subset of patients was similar to the whole HCV cirrhosis group, including antiviral treatment response, because most of them reduced or stopped their alcohol intake during follow-up. Conversely, the patients considered as having alcoholic cirrhosis were devoid of other causes of chronic liver disease, viral infection in particular. Some overlap might exist in this population between alcoholic steatohepatitis and nonalcoholic steatohepatitis, because the French definition of “a bon vivant” associates heavy drinking with excessive eating. Nevertheless, we considered only in this group patients with alcohol consumption over 80 g/day during more than 10 years (usually far more), a well-admitted but rather stringent threshold.
  • 3The authors are perfectly right to point out the potential and indeed very likely influence of the type of electrode used on the results. We ruled out the idea of including this parameter in a multivariate analysis for two reasons. The first one is the limited number of parameters we could take in account in a multivariate analysis according to the number of events in contrast with the numerous factors that could be involved in the prognosis. The second one is the discrepancy in tumoral size according to the type of probe used. The largest tumors were treated with the multipolar device (Prosurge, Celon Company),2 due to its increased capacity for destruction. In this setting, a comparison even by multivariate analysis would have been difficult.For underlining the role of technical points as to the type of probes and cooling systems, we previously performed a case-control study matching patients treated by internally cooled versus perfused electrodes according to different criteria, particularly tumoral size.3 This study showed that the use of perfused electrodes (Integra) was associated with a higher incidence of distant recurrences than was the use of internally cooled electrodes (Covidien). We had also shown in a pilot study that by using multipolar electrodes (Celon), complete ablation of tumors larger than 5 cm could be frequently achieved, a result out of reach of monopolar electrodes.2 Therefore, the use of a multipolar device for the treatment of smaller tumors such as those treated in our study might have reduced local recurrence rate. The heterogeneity of the techniques used is inherent to the retrospective nature of our study and the improvement of results with time following improvement in experience and technology is shared by other types of treatments such as surgery. This phenomenon needs to be stressed, because the overall results reported are likely to be less favorable than those expected today with RFA using up-to-date technique and technology.
  • 4From our point of view, matching patients subjected to a curative treatment like RFA with those having usually more advanced HCC and treated by a palliative one (e.g., chemoembolization) would make little sense. Comparing survival after RFA in patients who were eligible for resection with those for whom resection was contraindicated, we found a huge difference between the two groups in terms of 5-year overall survival rates (76% versus 27%, respectively). This result confirms that criteria usually used to select patients with HCC for resection (e.g., Barcelona Clinic Liver Cancer criteria) strongly affect survival of patient, whatever the type of treatment attempted (resection or RFA). In addition, the 76% for 5-year overall survival rate obtained with RFA in patients potentially eligible for resection is favorably comparable with those reported in surgical series. Therefore, we conclude that the indication of RFA as first-line treatment even for patients usually considered as good candidates for surgery was justified regarding its far lower complication rate and its repeatability in case of recurrence. Nevertheless, there is probably a subgroup of patients in whom RFA, although potentially curative, had little chance to achieve its goal: for example, patients with three tumors or who had an even limited vascular invasion. The reduced number of cases responding to this category did not allow us any case-control comparison with chemoembolization.

We thank the authors of the letter to have helped us clarify these points.

Gisele N'Kontchou*, Michel Beaugrand*, Olivier Seror†, * Hepatology Unit, Hôpital Universitaire Jean Verdier, Assistance Publique–Hôpitaux de Paris, Bondy, France, † Radiology Department, Hôpital Universitaire Jean Verdier, Assistance Publique–Hôpitaux de Paris, Bondy, France.