摘要目的:探讨开滦研究人群恶性肿瘤确诊前后总胆固醇差值对患者全因死亡的影响。方法:采用前瞻性队列研究方法，选择2006年6月至2015年12月健康体检期间发现恶性肿瘤，且确诊恶性肿瘤前后总胆固醇资料完整的开滦（集团）有限责任公司在职和退休职工作为观察对象。通过医疗保险系统、开滦总医院病案管理信息系统检索受试者恶性肿瘤首次确诊时间，采用SAS 9.4软件提取受试者肿瘤确诊前后2年内体检资料。根据恶性肿瘤确诊前后总胆固醇差值四分位数分组：第1分位组为总胆固醇差值<-0.93 mmol/L（823例）；第2分位组为总胆固醇差值≥-0.93 mmol/L且<-0.21 mmol/L（811例）；第3分位组为总胆固醇差值≥-0.21 mmol/L且<0.49 mmol/L（832例）；第4分位组为总胆固醇差值≥0.49 mmol/L（833例）。采用发病密度计算第1、2、3、4分位组的死亡率，采用Cox比例风险模型分析各组全因死亡的影响因素。结果:中位随访时间2.46年，共发生1 564例全因死亡事件。第1、2、3、4分位组死亡例数分别为481、440、333、310例，死亡率分别为230.10/千人年、163.90/千人年、115.34/千人年和83.44/千人年（ χ2=604.62， P<0.001）。校正混杂因素后，Cox比例风险模型分析显示，与第1分位组相比，第2、3、4分位组发生全因死亡的 HR（95% CI）分别为0.86（0.76~0.98）、0.62（0.54~0.72）、0.58（0.50~0.67）（均 P趋势<0.001）。不同部位肿瘤患者中，校正混杂因素后，与第1分位组比较，第4分位组肿瘤部位除甲状腺、乳腺、前列腺和睾丸、泌尿系统外，均 P趋势<0.05。 结论:开滦研究人群中，恶性肿瘤确诊后总胆固醇水平较确诊前下降幅度小及升高是全因死亡的保护因素。

abstractsObjective:To explore the effects of pre- and post-diagnosis of malignant neoplasms total cholesterol difference on all-cause death in the Kailuan study population.Methods:A prospective cohort study method was used to observe on the job and retired workers who were diagnosed with malignant neoplasms during the healthy physical examination and had the complete data of total cholesterol pre- and post-diagnosis of malignant neoplasms at Kailuan (Group) Limited Liability Corporation from June 2006 to December 2015. Medical insurance system and medical records case management system of Kailuan General Hospital were used to retrieve the confirmed diagnosis time of participants with malignant neoplasms. SAS 9.4 software was used to extract the physical examination data of participants diagnosed as malignant neoplasms before and after 2 years. The whole participants were divided into 4 groups according to the quartile of the total cholesterol difference pre- and post-diagnosis of cancer. The first quartile group: the total cholesterol difference <-0.93 mmol/L (823 cases), the second quartile group: the total cholesterol difference≥-0.93 mmol/L and <-0.21 mmol/L (811 cases), the third quartile group: the total cholesterol difference≥-0.21 mmol/L and <0.49 mmol/L (832 cases), the forth quartile group: the total cholesterol difference≥0.49 mmol/L (833 cases). The incidence density was used to calculate the mortality of different total cholesterol difference quartile groups, and the Cox proportional hazards model was used to analyze the influencing factors of all-cause death of all quartile groups.Results:Totally, 1 564 cases had all-cause death during median 2.46 years of follow-up time. There were 481, 440, 333, 310 death cases respectively of 4 quartile groups. The death rates of all quartile groups were 230.10/1 000 person-years, 163.90/1 000 person-years, 115.34/1 000 person-years, and 83.44/1 000 person-years respectively ( χ2 = 604.62, P < 0.001). After adjusting for confounding factors, the Cox proportional hazards model analysis showed that compared with the first quartile group, the hazard ratio with the 95% confidence interval of all-cause death in the second, third, and fourth quartile groups was 0.86 (0.76-0.98), 0.62(0.54-0.72) and 0.58 (0.50-0.67) respectively (all Ptrend < 0.001). After adjusting for confounding factors of patients with cancer in different location, the forth quartile group except for cancer in thyroid, breast, prostate, testis and urinary system, all Ptrend was less than 0.05 compared with that of the first quartile group. Conclusion:Among Kailuan study population, the smaller decline or even rise in total cholesterol level of post-diagnosis compared with that of pre-diagnosis is a protective factor of all-cause death.