The association between diabetes mellitus and HCC: Noise or real? A meta-analysis of epidemiologic evidence.

Abstract

184 Background: Diabetes mellitus (DM) has been linked to an increase in the incidence of hepatocellular carcinoma (HCC). However, these studies have yielded controversial findings. This study evaluates the association between DM and risk of HCC in a meta-analysis. Methods: Two investigators independently conducted a search of studies on the association between DM and HCC from January 1970 through August 2010 on Medline. Identified articles were manually reviewed for additional references. Only full articles that provided risk estimates for the association were selected. The most adjusted risk estimates were abstracted and synthesized using the random-effects model. A stratified analysis by study design and study region has also been done. All studies were evaluated for publication bias before analysis. Results: A total of 35 studies, 20 case-control and 15cohort studies, were eligible for analysis. The overall summary risk ratio for HCC in DM was 1.60 (95% CI 1.49, 1.70; p<0.001). Among the case-control and cohort studies, the pooled risk estimate was 1.33 (95% CI 1.21, 1.46; p<0.001) and 2.24 (2.05, 2.44; p<0.001) respectively. By geographic location, the association remained significant among the European (15) and U.S. (10) based studies with risk ratios of 2.7 (95% CI 2.37, 3.05; p<0.001) and 1.58 (95% CI 1.45, 1.7) respectively. However, among the Japanese (10) based studies there was a nonsignificant 3% reduction in the risk of HCC (RR 0.97; 95% CI 0.71, 1.24; p>0.05). There was no evidence of publication bias among the studies included. Conclusions: The evidence in support of DM as an independent risk factor for HCC remains unclear. Though not significant, the finding among the Japanese based populations raises concerns over the potential role of environmental and/or genetic factors in the observed association. Further studies with great attention to potential confounders are warranted to definitively define the association between DM and HCC. No significant financial relationships to disclose.