Liu Tong, Wang Wanchao, Cui Haozhe, Sun Miaomiao, Wang Yiming, Liu Xining, Cao Liying, Liu Hai, Liu Siqing
Department of Hepatobiliary Surgery, Kailuan General Hospital Affiliated to North China University of Science and Technology.
Department of Graduate School, North China University of Science and Technology.
Medicine (Baltimore). 2019 Jul;98(30):e16369. doi: 10.1097/MD.0000000000016369.
Previous studies have demonstrated a positive relationship between liver cancer and diabetes mellitus. However, elevated fasting blood glucose (FBG) itself may be a risk factor for the development of hepatocellular carcinoma (HCC) rather than diabetes, and during the follow-up period, death is an event that may occur before the occurrence of HCC, which should be dealt with competing risk models. Our study aims to investigate the relationship between FBG and new-onset HCC by using competing risk regression models.We prospectively studied the relationship between FBG concentrations and risk of HCC in a cohort of 93,447 participants who were free of prior HCC, and whose demographic characteristics and biochemical parameters were recorded. Cox proportional hazards regression models and competing risk regression models were used to evaluate the association between FBG concentrations and risk of incident HCC.A total of 302 participants were diagnosed with HCC among 93,447 subjects during 810,499 person-years of follow-up. The multivariable hazard ratios (HRs) [95% confidence interval (95% CI)] for the association of FBG and log(FBG) with HCC were 1.07 (1.01∼1.12), 1.84 (1.23∼2.74) in an analysis adjusted for other potential variables. In the multivariable adjusted analysis, participants who were in 4.82 mmol/L≤FBG≤5.49 mmol/L group and FBG >5.49 mmol/L group would have increased the risk of HCC by 47% and 69%, respectively. In a cause-specific hazard model (CS model), the multivariable HRs (95% CI) for the association of FBG with HCC were 1.46 (1.09∼1.98), 1.69 (1.27∼2.27) in the multivariable adjusted analysis. Similar results were also observed in sub-distribution hazard function model (SD model) with corresponding multivariate HRs (95% CI) of 1.46 (1.09∼2.00), 1.69 (1.25∼2.27) in 4.82 mmol/L≤FBG≤5.49 mmol/L group and FBG >5.49 mmol/L group, respectively.Higher FBG concentrations itself were positively associated with new-onset HCC in the Cox proportional hazards regression models and competing risk models. FBG concentrations can be used as a scientific and important way to identify individuals with a higher risk of HCC and control of FBG concentrations might serve as a possible way to decrease the risk of HCC among Chinese population.Trial registration: ChiCTR-TNRC-11001489. Registered August 24, 2011 (retrospectively registered).
以往的研究表明肝癌与糖尿病之间存在正相关关系。然而，空腹血糖（FBG）升高本身可能是肝细胞癌（HCC）发生的一个危险因素，而非糖尿病，并且在随访期间，死亡是可能在HCC发生之前出现的事件，对此应采用竞争风险模型进行处理。我们的研究旨在通过使用竞争风险回归模型来探究FBG与新发HCC之间的关系。我们前瞻性地研究了93447名无既往HCC且记录了人口统计学特征和生化参数的参与者队列中FBG浓度与HCC风险之间的关系。采用Cox比例风险回归模型和竞争风险回归模型来评估FBG浓度与新发HCC风险之间的关联。在810499人年的随访期间，93447名受试者中共有302人被诊断为HCC。在对其他潜在变量进行调整的分析中，FBG和log(FBG)与HCC关联的多变量风险比（HRs）[95%置信区间（95%CI）]分别为1.07（1.01～1.12）、1.84（1.23～2.74）。在多变量调整分析中，FBG在4.82 mmol/L≤FBG≤5.49 mmol/L组和FBG＞5.49 mmol/L组的参与者发生HCC的风险分别增加47%和69%。在病因特异性风险模型（CS模型）中，多变量调整分析中FBG与HCC关联的多变量HRs（95%CI）分别为1.46（1.09～1.98）、1.69（1.27～2.27）。在亚分布风险函数模型（SD模型）中也观察到了类似结果，在4.82 mmol/L≤FBG≤5.49 mmol/L组和FBG＞5.49 mmol/L组中相应的多变量HRs（95%CI）分别为1.46（1.09～2.00）、1.69（1.25～2.27）。在Cox比例风险回归模型和竞争风险模型中，较高的FBG浓度本身与新发HCC呈正相关。FBG浓度可作为识别HCC高风险个体的一种科学且重要的方法，控制FBG浓度可能是降低中国人群HCC风险的一种可行方法。试验注册：ChiCTR-TNRC-11001489。于2011年8月24日注册（追溯注册）。
