糖尿病和高敏C反应蛋白水平影响消化系统恶性肿瘤发生风险的前瞻性队列研究

刘款; 李佳兴; 马超; 王万超; 田园; 董智刚; 韦稳强; 陈朔华; 吴寿岭; 刘四清

doi:10.3760/cma.j.cn115610-20241022-00461

糖尿病和高敏C反应蛋白水平影响消化系统恶性肿瘤发生风险的前瞻性队列研究

The influence of diabetes mellitus and high‑sensitivity C‑reactive protein on the risk of diges-tive system malignancy: a prospective cohort study

摘要

摘要:
目的探讨糖尿病（DM）和高敏C反应蛋白（Hs‑CRP）水平对消化系统恶性肿瘤发生风险的影响。
方法采用前瞻性队列研究方法。选取2006年华北理工大学附属开滦总医院等唐山市9家医院参加健康体检的93 928例受试者的临床资料。根据受试者是否合并DM和Hs‑CRP水平进行分组。无DM且Hs‑CRP≤3 mg/L受试者设为DM（-）CRP（-）组,无DM且Hs‑CRP>3 mg/L受试者设为DM（-）CRP（+）组,有DM且Hs‑CRP≤3 mg/L受试者设为DM（+）CRP（-）组,有DM且Hs‑CRP>3 mg/L受试者设为DM（+）CRP（+）组。由固定医师团队收集受试者资料。以2006年首次健康体检时间为随访起点,随访终点事件为发生消化系统恶性肿瘤或死亡,随访时间截至2021年12月31日。观察指标：（1）4组受试者临床资料比较。（2）受试者消化系统恶性肿瘤的发病情况及累积发病率。（3）DM和Hs‑CRP水平分别对消化系统恶性肿瘤发生风险的影响。（4）DM和Hs‑CRP水平共同对消化系统恶性肿瘤发生风险的影响。（5）敏感性分析。正态分布的计量资料多组间比较采用单因素方差分析；两两比较,方差齐采用最小显著性差异法检验,方差不齐采用Dunnett′s T3检验。偏态分布的计量资料多组间比较采用Kruskal⁃Wallis秩和检验,两两比较采用Dunn‑Bonferroni检验。计数资料多组间比较采用χ²检验,多重比较采用Bonferroni检验。采用Kaplan‑Meier法绘制累积发病曲线,Log‑rank检验进行累积发病率分析。采用Cox比例风险模型进行多因素分析。所有模型校正相关混杂因素。
结果（1）4组受试者临床资料比较。93 928例受试者中,DM（-）CRP（-）组70 743例,DM（-）CRP（+）组14 644例,DM（+）CRP（-）组6 425例,DM（+）CRP（+）组2 116例。4组受试者性别、年龄、空腹血糖、Hs‑CRP水平、三酰甘油、丙氨酸氨基转移酶、体质量指数、已婚、吸烟、饮酒、高中及以上学历、体育锻炼、高盐饮食、高脂饮食、乙型肝炎病毒表面抗原阳性、脂肪肝、肝硬化、胆囊结石、服用降糖药物、服用降脂药物比较,差异均有统计学意义（P<0.05）。（2）受试者消化系统恶性肿瘤的发病情况及累积发病率。93 928例受试者截至随访终点,新发消化系统恶性肿瘤2 008例,其中结直肠癌717例、肝癌456例、胃癌396例、食管癌195例、胰腺癌144例、胆囊或肝外胆管癌65例、小肠癌35例。DM（-）CRP（-）组、DM（-）CRP（+）组、DM（+）CRP（-）组、DM（+）CRP（+）组受试者消化系统恶性肿瘤的累积发病率分别为2.19%、2.42%、2.86%、3.59%,4组受试者累积发病率比较,差异有统计学意义（χ²=31.72,P<0.05）。（3）DM和Hs‑CRP水平分别对消化系统恶性肿瘤发生风险的影响。校正受试者混杂因素后,多因素分析结果显示：DM和Hs‑CRP>3 mg/L均是新发消化系统恶性肿瘤的独立影响因素（风险比=1.32、1.19,95%可信区间为1.13~1.56、1.06~1.33,P<0.05）。进一步分析发现,DM与Hs‑CRP>3 mg/L的交互作用有统计学意义（P<0.005）。（4）DM和Hs‑CRP水平共同对消化系统恶性肿瘤发生风险的影响。校正受试者混杂因素后,多因素分析结果显示：以DM（-）CRP（-）组为对照,DM（-）CRP（+）组、DM（+）CRP（-）组、DM（+）CRP（+）组新发消化系统恶性肿瘤风险均增加（风险比=1.14、1.23、1.79,95%可信区间为1.01~1.29、1.02~1.48、1.38~2.31,P<0.05）。在消化系统恶性肿瘤的特定部位分析中,以DM（-）CRP（-）组为对照,DM（-）CRP（+）组肝癌的发生风险增加（风险比=1.37,95%可信区间为1.07~1.75,P<0.05）；DM（+）CRP（-）组肝癌、胰腺癌的发生风险均增加（风险比=1.60、1.74,95%可信区间为1.16~2.21、1.00~3.02,P<0.05）；DM（+）CRP（+）组小肠癌、胰腺癌、结直肠癌的发生风险均增加（风险比=5.05、2.31、2.23,95%可信区间为1.57~16.21、1.00~5.31、1.54~3.24,P<0.05）。（5）敏感性分析。校正受试者混杂因素,排除3类受试者（随访1年内患消化系统恶性肿瘤103例、服用降糖药物2 370例、服用降脂药物915例）情况下,多因素分析结果显示：以DM（-）CRP（-）组为对照,DM（+）CRP（-）组、DM（+）CRP（+）组新发消化系统肿瘤的风险均增加（风险比_{排除随访1年内患消化系统恶性肿瘤}=1.26、1.66,95%可信区间为1.04~1.52、1.26~2.18,P<0.05；风险比_{排除服用降糖药物}=1.23、1.75,95%可信区间为1.02~1.49、1.31~2.33,P<0.05；风险比_{排除服用降脂药物}=1.24、1.80,95%可信区间为1.03~1.49、1.39~2.34,P<0.05）。
结论 DM和Hs⁃CRP>3 mg/L均是影响新发消化系统恶性肿瘤发生风险的独立因素；且DM与Hs⁃CRP存在交互协同作用促进消化系统恶性肿瘤发生。

Abstract:
Objective To investigate the influence of diabetes mellitus (DM) and high-sen-sitivity C‑reactive protein (Hs‑CRP) on the risk of digestive system malignancy.
Methods The pro-spective cohort study was conducted. The clinical data of 93 928 participants who participated health examination in 9 hospitals at Tangshan, including Kailuan General Hospital Affiliated to North China University of Science and Technology et al, in 2006 were selected. According to the presence or absence of DM and the level of Hs‑CRP, all participants were divided into 4 groups, including the DM(-)CRP(-) group defined as absence of DM and Hs‑CRP ≤3 mg/L, the DM(-)CRP(+) group defined as absence of DM and Hs‑CRP>3 mg/L, the DM(+)CRP(-) group defined as presence of DM and Hs‑CRP ≤3 mg/L, and the DM(+)CRP(+) group defined as presence of DM and Hs‑CRP >3 mg/L. The data of participants were collected by a fixed team of physicians. The first physical examination in 2006 was taken as the starting point for follow‑up. The end event of follow‑up was defined as the occurrence of digestive system malignancy or death, and the follow-up was up to December 31, 2021. Observation indicators: (1) comparison of clinical data among the 4 groups of participants; (2) the incidence and cumulative incidence rate of digestive system malignancy in participants; (3) influence of DM and Hs‑CRP level on the risk of digestive system malignancy; (4) the combined influence of DM and Hs‑CRP level on the risk of digestive system malignancy; (5) sensitivity analysis. Comparison of measurement data with normal distribution among multiple groups was conducted using the one‑way analysis of variance. For pairwise comparison, least significant difference test was used for homogeneity of variance, and Dunnett′s T3 test was used for heterogeneity of variance. Comparison of measurement data with skewed distribution among multiple groups was conducted using the Kruskal-Wallis rank sum test, and Dunn‑Bonferroni test was used for pairwise comparison. Comparison of count data among multiple groups was conducted using the chi‑square test, and Bonferroni test was used among multiple comparisons. The Kaplan‑Meier method was used to plot cumulative incidence curve, and Log⁃rank test was used for cumulative incidence rate analysis. The Cox proportional risk model was used for multivariate analysis. All models were adjusted for relevant confounders.
Results (1) Comparison of clinical data among the 4 groups of participants. Of the 93 928 participants, there were 70 743 cases in the DM(-)CRP(-) group, 14 644 cases in the DM(-)CRP(+) group, 6 425 cases in the DM(+)CRP(-) group, and 2 116 cases in the DM(+)CRP(+) group. There were significant differences in gender, age, fasting blood glucose, Hs‑CRP, triglyceride, alanine aminotransferase, body mass index, marrital status, smoking, drinking, high school degree or above, physical exercise, high salt diet, high fat diet, positive hepatitis B virus surface antigen, fatty liver, liver cirrhosis, gallstone, taking hypoglycemic drugs, taking lipid‑lowering drugs among the 4 groups of participants (P<0.05). (2) The incidence and cumulative incidence rate of digestive system malignancy in participants. At the end-up of follow‑up, 2 008 cases developed digestive system malignancy in the 93 928 participants, including 717 cases of colorectal cancer, 456 cases of liver cancer, 396 cases of gastric cancer, 195 cases of esophageal cancer, 144 cases of pancreatic cancer, 65 cases of gallbladder cancer or extrahepatic cholangiocarcinoma, 35 cases of small bowel cancer. The cumulative incidence rates of digestive system malignancy were 2.19%, 2.42%, 2.86%, 3.59% in participants of the DM(-)CRP(-) group, DM(-)CRP(+) group, DM(+)CRP(-) group, DM(+)CRP(+) group, respectively, showing a significant difference among the 4 groups (χ²=31.72, P<0.05). (3) Influence of DM and Hs‑CRP level on the risk of digestive system malignancy. After adjusting for the confounding factors of the participants, results of multivariate analysis showed that DM and Hs‑CRP >3 mg/L were independent influencing factors for the incidence of digestive system malignancy (hazard ratio=1.32, 1.19, 95% confidence interval as 1.13-1.56, 1.06-1.33, P<0.05). Futher analysis showed that there was a significant difference in interaction between DM and Hs‑CRP >3 mg/L (P<0.05). (4) The combined influence of DM and Hs‑CRP level on the risk of digestive system malign-ancy. After adjusting for confounding factors, results of multivariate analysis showed that using the DM(-)CRP(-) group as the control group, the risk of incidence of digestive system malignancy increased in the DM(-)CRP(+) group, DM(+)CRP(-) group, and DM(+)CRP(+) group, respectively (hazard ratio=1.14, 1.23, 1.79, 95% confidence interval as 1.01-1.29, 1.02-1.48, 1.38-2.31, P<0.05). In the site‑specific analysis of digestive system malignancy, using the DM(-)CRP(-) group as the control group, the risk of incidence of liver cancer increased in the DM(-)CRP(+) group (hazard ratio=1.37, 95% confidence interval as 1.07-1.75, P<0.05), the risk of incidence of liver cancer and pancrea-tic cancer increased in the DM(+)CRP(-) group (hazard ratio=1.60, 1.74, 95% confidence interval as 1.16-2.21, 1.00-3.02, P<0.05), the risk of incidence of small bowel cancer, pancreatic cancer and colorectal cancer increased in the DM(+)CRP(+) group (hazard ratio=5.05, 2.31, 2.23, 95% confidence interval as 1.57-16.21, 1.00-5.31, 1.54-3.24, P<0.05). (5) Sensitivity analysis. After adjusting for confounding factors of excluding 3 types of participants (103 cases of digestive system malignancy within 1 year of follow‑up, 2 370 cases of taking glucose‑lowering drugs, and 915 cases of taking lipid‑lowering drugs), results of multivariate analysis showed that using the DM(-)CRP(-) group as the control group, the risk of incidence of digestive system malignancy increased in the DM(+)CRP(-) group, and DM(+)CRP(+) group, respectively (hazard ratio_{excluding cases of digestive system malignancy within 1 year of follow-up}=1.26, 1.66, 95% confidence interval as 1.04-1.52, 1.26-2.18, P<0.05; hazard ratio_{excluding cases taking glucose-lowering drugs}=1.23, 1.75, 95% confidence interval as 1.02-1.49, 1.31-2.33, P<0.05; hazard ratio_{excluding cases taking lipid-lowering drugs}=1.24, 1.80, 95% confidence interval as 1.03-1.49, 1.39-2.34, P<0.05).
Conclusions DM and Hs‑CRP >3 mg/L are independent influencing factors for the incidence of digestive system malignancy. There is an interation and synergistic effect between DM and Hs-CRP to promote the incidence of digestive system malignancy.

HTML全文

参考文献(40)

施引文献

资源附件(0)