Abstract: Background　The correlation between metabolic diseases and thyroid nodules has been mainly explored in community populations，but rarely in university faculty. Moreover，since most of these studies are based on cross-sectional data，in which the association of metabolic diseases and thyroid nodules may not be precisely measured due to potential individual confounders such as gender，age，lifestyle，disease history，and family history. Objective　To further examine the association of metabolic diseases and thyroid nodules using the panel data of health checkups in university faculty. Methods　In this study，we collected the panel data of four-year（from 2015 to 2018） health checkups of a university's faculty in Beijing，including gender，age，height，weight，body mass index，blood pressure，fasting plasma glucose，total cholesterol （TC），triacylglycerol （TG），low-density lipoprotein cholesterol （LDL-C），high-density lipoprotein cholesterol （HDL-C），serum uric acid，thyroid B-mode ultrasound，liver B-mode ultrasound，etc. We calculated the detection rates of thyroid nodules，hypertension，abnormal glucose metabolism，hypercholesterolemia，hypertriglyceridemia，hyper-LDL-cholesterolemia，hypo-HDL-cholesterolemia，hyperuricemia and fatty liver. The two-way fixed effect model was used to measure the strength of association between metabolic diseases and thyroid nodules after controlling the potential confounders such as individual effect and temporal effect. Results　In total，4 725 cases were included for analysis. Among them，2 413 cases of thyroid nodules were detected. In 2018，the detection rate of thyroid nodules was 56.95% （684/1 201）. The top three metabolic diseases detected were dyslipidemia〔42.96% （516/1 201）〕，overweight and obesity〔42.96% （516/1 201）〕，and fatty liver 〔33.64% （404/1 201）〕. Men had higher detection rates of hypertension，hyper-LDL-cholesterolemia，hypertriglyceridemia，hypo-HDL-cholesterolemia，abnormal glucose metabolism，overweight and obesity，hyperuricemia and fatty liver than women （P<0.01）. The middle-aged（41-50 years old） and older groups（≥51 years old） had higher detection rates of thyroid nodules，hypertension，and abnormal glucose metabolism（P<0.05）. Moreover，those aged ≥51 years had higher detection rates of thyroid nodules，hypertension，hypercholesterolemia，hyper-LDL-cholesterolemia，hypertriglyceridemia，abnormal glucose metabolism，overweight and obesity，hyperuricemia and fatty liver than those aged 41-50 or ≤ 40 years （P<0.05）. After controlling for gender and age，it was found that the detection rate of hypertension，hypertriglyceridemia，or overweight and obesity was positively correlated with that of thyroid nodules with varying degrees of strength in the two-way fixed effect model for all participants （P<0.10）. Hypertension caused a 3.3% increase in the detection rate of thyroid nodules and passed the test at the 10% significance level. Hypertriglyceridemia caused a 2.6% increase in the detection rate of thyroid nodules and passed the test at the 10% significance level. Overweight and obesity caused a 1.6% increase in the detection rate of thyroid nodules and passed the test at the 1% significance level. Conclusion　The prevalence of metabolic diseases and thyroid nodules were high. The prevalence of thyroid nodules was increased with age in the university faculty. Those with hypertension，hypertriglyceridemia，overweight and obesity may be more susceptible to thyroid nodules，hence，early screening may be a measure to prevent thyroid nodules and thyroid cancer for them.

Key words: Thyroid nodules, Metabolic diseases, University faculty, Health checkup, Panel data, Fixed effect model

摘要： 背景　针对代谢性疾病与甲状腺结节的相关性，既往研究多以社区人群为研究对象开展初步探索，但鲜有聚焦高校教师群体的深度分析，且多数研究基于横断面数据进行，由于代谢性疾病与甲状腺结节发生率可能受到患者性别、年龄、生活方式、疾病史、家族史等个体因素的干扰，对两者相关性尚无法进行精准推断。目的　聚焦高校教师群体，利用面板数据分析代谢性疾病对甲状腺结节发生的影响效应，进一步提高研究结论的可靠性。方法　收集2015—2018年北京市某高校教师健康体检的面板数据，体检资料包括性别、年龄、身高、体质量、体质指数（BMI）、血压、空腹血糖（FPG）、总胆固醇（TC）、三酰甘油（TG）、低密度脂蛋白胆固醇（LDL-C）、高密度脂蛋白胆固醇（HDL-C）、血尿酸（SUA）、甲状腺B超、肝脏B超等，计算其甲状腺结节、高血压、糖代谢异常、高总胆固醇血症、高三酰甘油血症、高低密度脂蛋白胆固醇血症、低高密度脂蛋白胆固醇血症、高尿酸血症、脂肪肝检出率。采用双向固定效应模型，控制高校教师个体差异和时间趋势的潜在干扰，分析代谢性疾病对甲状腺结节发生的影响效应。结果　4年累计获得有效研究样本4 725例次，其中甲状腺结节检出2 413例次。2018年，甲状腺结节检出率为56.95%（684/1 201），代谢性疾病检出率较高的为血脂异常50.79%（610/1 201）、超重肥胖42.96%（516/1 201）、脂肪肝33.64%（404/1 201）等。男性高血压、高低密度脂蛋白胆固醇血症、高三酰甘油血症、低高密度脂蛋白胆固醇血症、糖代谢异常、超重肥胖、高尿酸血症、脂肪肝检出率均高于女性（P<0.01）。中生代教师（41~50岁）和资深教师（≥51岁）甲状腺结节、高血压、糖代谢异常检出率高于青年教师（≤40岁）（P<0.05）。资深教师（≥51岁）甲状腺结节、高血压、高总胆固醇血症、高低密度脂蛋白胆固醇血症、高三酰甘油血症、糖代谢异常、超重肥胖、高尿酸血症、脂肪肝检出率高于青年教师（≤40岁）和中生代教师（41~50岁）（P<0.05）。基于全样本双向固定效应模型分析发现，在控制高校教师性别、年龄影响的基础上，高血压、高三酰甘油血症、超重肥胖检出率对甲状腺结节检出率有不同程度的正向影响（P<0.10）。患高血压可导致甲状腺结节检出率升高3.3%，并通过10%显著性检验；患高三酰甘油血症可引起甲状腺结节检出率上升2.6%，并通过10%显著性检验；患超重肥胖可导致甲状腺结节检出率增高1.6%，并通过1%显著性检验。结论　北京市某高校教师代谢性疾病、甲状腺结节检出率较高，甲状腺结节检出率随高效教师年龄增长而增加，高血压、高三酰甘油血症、超重肥胖会导致高校教师甲状腺结节发生率增加，因此应尽早识别高危人群达到早期预防甲状腺结节和甲状腺癌的目的。

关键词: 甲状腺结节, 代谢性疾病, 高校教师, 健康体检, 面板数据, 固定效应模型