Abstract: Background The correlation between sleep duration and heart failure has been established，yet the majority of studies are observational in nature，rendering causal inference difficult due to confounding variables. Objective To explore the causal relationship between both inadequate and excessive sleep durations and heart failure using two-sample Mendelian randomization. Methods The research data was derived from the genome-wide association study（GWAS）summary database，wherein inadequate sleep duration consisted of 106，192 samples and excessive sleep duration consisted of 34，184 samples. Appropriate single nucleotide polymorphisms were selected as instrumental variables，and two-sample Mendelian randomization analyses were conducted using inverse variance weighted，weighted median，Weighted mode and MR-Egger regression to evaluate the causal relationship between sleep duration and heart failure，as measured by the odds ratio. Sensitivity analyses were performed using heterogeneity testing，horizontal pleiotropic test，and leave-one-out analysis. Results 23 SNPs were identified as being associated with inadequate sleep duration，while 5 SNPs were associated with exessive sleep duration. The IVW analysis revealed that inadequate sleep duration is a risk factor for heart failure ［OR（95%CI）=2.347（1.209-4.555）］，while excessive sleep duration may be a protective factor ［OR（95%CI）=0.774（0.141-4.249）］，though the latter finding did not reach statistical significance. Sensitivity analysis，after removal of SNPs with multiple testing correction，demonstrated the robustness of MR results. Conclusion There is a positive causal relationship between short sleep duration and heart failure，that is，short sleep duration increases the risk of heart failure.

Key words: Sleep, Sleep duration, Heart failure, Mendelian randomization, Causal inference

摘要： 背景 睡眠时间与心力衰竭相关，但大多数研究为观察性研究，由于混杂因素过多难以进行因果关联的推断。目的 采用两样本孟德尔随机化方法评估睡眠时间过短和睡眠时间过长与心力衰竭的因果关联。方法 研究数据来自全基因组关联分析（GWAS）汇总数据库，其中睡眠时间过短包含106192例样本，睡眠时间过长34184例样本，分别选取合适的SNPs作为工具变量，采用逆方差加权法、加权中位数法和MR-Egger回归法等进行两样本孟德尔随机化分析，以OR值评估睡眠时间与心力衰竭之间的因果关联。采用异质性检验、水平多效性检验、leave-one-out检验等进行敏感性分析。结果 筛选到与睡眠时间过短相关的SNPs有23个，与睡眠时间过长相关的SNPs有5个，IVW结果显示睡眠时间过短是心力衰竭发生的危险因素［OR（95%CI）=2.347（1.209~4.555）］，而睡眠时间过长可能是心力衰竭发生的保护因素［OR（95%CI）=0.774（0.141~4.249）］，但结果无统计学意义，在剔除存在水平多效性的SNP后进行敏感性分析，表明MR结果具有稳健性。结论 睡眠时间过短与心力衰竭存在正向因果关联，即睡眠时间过短会增加心力衰竭的患病风险。

关键词: 睡眠, 睡眠时间, 心力衰竭, 孟德尔随机化, 因果推断