Review article: overview of medical treatments in unresectable hepatocellular carcinoma—an impossible meta-analysis?

Authors


Mathurin Service d’Hépatogastroentérologie, Groupe Hospitalier Pitié-Salpêtrière, 47-83 boulevard de l’Hôpital, 75651 Paris Cédex 13, France.

Abstract

Background:

Controversies surrounding medical treatment in patients with unresectable hepatocellular carcinoma continue to persist.

Aim:

To perform a meta-analysis of therapeutic modalities which had been evaluated in two or more randomized trials.

Methods:

Fifty-two randomized trials were studied; only 30 were included. This overview identified seven therapeutic modalities which had been evaluated in two or more trials: adriamycin, 5-fluorouracil, interferon, percutaneous ethanol injection, transarterial chemotherapy, the combination of lipiodol with transarterial chemotherapy, and tamoxifen.

Results:

Comparisons of survival between control groups showed substantial heterogeneity. There was no survival benefit at 1 year with adriamycin (mean difference 4%), 5-fluorouracil (mean difference −3%), percutaneous ethanol injection (mean difference 6%) or transarterial chemotherapy (mean difference −2%). For interferon, the survival benefit was significant with the Der Simonian & Laird method (mean difference 9%, 95% CI = 1–18%, P = 0.04) but not with the Peto et al. method (2.4 mean odds ratio, 95% CI = O.9–6.8). The meta-analysis of tamoxifen showed a borderline survival benefit (mean difference 25%, 95% CI = 0–49%, P = 0.05). However, in sensitivity analyses, the survival benefit of tamoxifen was no longer significant.

Conclusions:

No treatment has clearly proven efficacy in survival. 5-Fluorouracil, adriamycin and transarterial chemotherapy were not associated with survival benefit at 1 year. The number of randomized controlled trials was insufficient to enable a conclusion to be reached for interferon and percutaneous ethanol injection. Controversy persists concerning tamoxifen efficacy. Interferon and tamoxifen require new randomized controlled trials on a larger population of patients.

Ancillary