• Potential conflict of interest: Nothing to report.


We thank Dr. Duberg and Dr. Hultcrantz for their interest in our work. We agree with them that the apparent fall in mortality rates from hepatocellular carcinoma (HCC) in Sweden during the mid-1990s may reflect changes in coding practices more than real variations in mortality, as we already stated in our article.1

They suggested to include liver cancer unspecified together with HCC in the analysis of mortality trends. Unspecified liver cancers, however, include a heterogeneous group of (mainly secondary) neoplasms, which is broadly variable across countries and time periods.2, 3 It is therefore difficult to make any consistent inference on such a heterogeneous group.

More important, most other countries considered did not show sharp changes in HCC mortality over the 25-year period analyzed. It is therefore unlikely that the problem noted for Sweden by ourselves and Dr. Duberg and Dr. Hultcrantz applies to a similar extent to most other countries as well.

Nevertheless, we can only stress again the importance of interpreting with due caution mortality and incidence data on (primary) liver cancer, given the problems in reliability and validity of certification data for this neoplasm, mainly due to the difficulty in distinguishing primary and secondary liver neoplasms.

Cristina Bosetti*, Fabio Levi†, Paolo Boffetta‡, Franca Lucchini†, Eva Negri*, Carlo La Vecchia* §, * Istituto di Ricerche Farmacologiche “Mario Negri”, Milan, Italy, † Unité d'Épidémiologie du Cancer et Registres Vaudois et Neuchâtelois des Tumeurs, Institut de Médecine Sociale et Préventive (IUMSP), Universitè de Lausanne, Lausanne, Switzerland, ‡ International Agency for Research on Cancer, Lyon, France, § Istituto di Biometria e Statistica Medica, G.A. Maccacaro,, Università degli Studi di Milano, Milan, Italy.