Abstract
BACKGROUND: Sleep duration has been associated with nonalcoholic fatty liver disease, but its association with liver cancer remains unknown. MATERIAL AND METHODS: In the prospective Women's Health Initiative Study, 139,368 postmenopausal women reported sleep habits at baseline (1993-1998). We ascertained 175 incident liver cancer cases during an average 13.8 years of follow-up through August 2014. We used multivariable Cox proportional hazard regression models to estimate a hazard ratio (HR) and its 95% confidence interval (95% CI) for risk of liver cancer in association with nocturnal sleep duration. RESULTS: Compared to women reporting 6-8 hours of sleep, the HR for liver cancer was 1.94 (95% CI 1.07-3.53) for women reporting ≥9 hours of sleep. Among the obese women, the HR associated with ≥9 hours of sleep was 3.18 (95% CI 1.84-8.60). The HR was 0.93 (95% CI 0.34-2.53) among nonobese women (p value for interaction = 0.18). Short sleep duration (≤5 hours) was not associated with liver cancer risk. CONCLUSION: Long sleep duration was associated with a moderate increase in liver cancer risk in obese postmenopausal women in the United States. Larger study is needed to confirm our observation on effect modification by adiposity status.