Background: Observational studies show an association between telomere length and Hepatocellular carcinoma (HCC) risk, but the relationship is controversial. Particularly, it remains unclear whether the association is due to confounding or biases inherent in conventional epidemiological studies. Here, we applied Mendelian randomization approach to evaluate whether telomere length is causally associated with HCC risk. Methods: Individual-level data were from HBV-related HCC Genome-wide association studies (1,538 HBV positive HCC patients and 1,465 HBV positive controls). Genetic risk score, as proxy for actual measured telomere length, derived from nine telomere length-associated genetic variants was used to evaluate the effect of telomere length on HCC risk. Results: We observed a significant risk signal between genetically increased telomere length and HBV-related HCC risk (OR = 2.09, 95% CI 1.32-3.31, P = 0.002). Furthermore, a U-shaped curve was fitted by the restricted cubic spline curve, which indicated that either short or long telomere length would increase HCC risk (P = 0.0022 for non-linearity test). Subgroup analysis did not reveal significant heterogeneity between different age, gender, smoking status and drinking status groups. Conclusions: Our results indicated that a genetic background that favors longer or shorter telomere length may increase HBV-related HCC risk-a U-shaped association. (c) 2017 Elsevier Ltd. All rights reserved.