第一节 甲状腺功能异常住院心力衰竭患者的临床特征和预后分析

【目的】心血管系统是甲状腺激素的作用靶器官之一，甲状腺激素异常可影响心血管疾病的发生、发展，从而影响疾病预后。本文主要探讨甲状腺功能对住院心力衰竭（心衰）患者的影响作用。

【方法】本研究是一项单中心、回顾性研究，连续纳入 2009 年 3 月至 2018 年 6 月于中国医学科学院阜外医院心力衰竭重症监护病房住院的心衰患者 3733 例，所有患者都接受了包括人口学资料及病史采集、体格检查、常规 12 导联心电图、超声心动图、抽血化验等临床评估。患者出院后进行系统的门诊复查或电话随访，研究主要终点定义为全因死亡或心脏移植或植入左心室辅助装置的复合终点。采用Kaplan-Meier 生存曲线，分析评估不同甲状腺激素水平、不同甲状腺状态对预后的影响。限制性立方样条（RCS）回归分析用于评估甲状腺激素作为连续变量与复合终点之间的关系。采用单因素和多因素 Cox 比例风险回归模型计算甲状腺激素、不同甲状腺功能状态与心力衰竭患者复合终点的风险比。

【结果】根据甲状腺功能状态分为甲状腺功能正常组（n=1865，50.0%）、低三碘甲状腺原氨酸（T3）综合征组（n=610，16.3%）、甲状腺功能亢进（n=42，1.1%）、亚临床甲状腺功能亢进（n=223，6.0%）、甲状腺功能减低（n=164，4.4%）、亚临床甲状腺功能减低（n=237，6.3%）及其他（n=592，15.9%）7 组。中位随访时间为 2.75（1.00，4.94）年，最终随访时共记录了 1659 例终点事件。Kaplan-Meier 生存曲线提示低游离 T3（FT3）提示预后更差（P<0.001），促甲状腺激素（TSH）在正常范围者预后最好（P<0.001）。在 RCS 回归分析中，随着 FT3 的升高复合终点风险呈下降趋势，TSH 呈 U 型曲线分布。低 FT3[风险比（HR）=1.33；95%可信区间（CI）1.15~1.54；P<0.001，促甲状腺激素升高（HR=1.37；95% CI 1.15~1.64；P<0.001），低 T3 综合征（HR=1.39；95% CI 1.15~1.68；P<0.001），甲状腺功能亢进（HR=1.73；95% CI 1.00~2.98；P=0.048），亚临床甲状腺功能减低（HR=1.43；95% CI 1.13~1.82；P=0.003）和甲状腺功能减低（HR=1.76；95% CI 1.33~2.34；P<0.001）是全因死亡、心脏移植或植入左室辅助装置复合终点的独立危险因素。【结论】低 T3 综合征、甲状腺功能亢进、甲状腺功能减低和亚临床甲状腺功能减低与心衰的不良预后独立相关。

第二节 游离三碘甲状腺原氨酸/甲状腺素比值对住院心力衰竭患者预后的影响

【目的】探讨游离三碘甲状腺原氨酸/游离甲状腺素（FT3/FT4）比值对心力衰竭（心衰）患者预后的影响。【方法】本研究回顾性分析了 2009 年 3 月至 2018 年 6 月住院的 3527 例中国医学科学院阜外医院心力衰竭重症监护病房住院的心衰患者。根据 FT3/FT4 比值的中位数进行分为低 FT3/FT4 比值组（n=1764，FT3/FT4<0.21）和高 FT3/FT4 比值组(n=1763，FT3/FT4≥0.21)两组。收集患者临床资料，患者出院后进行系统的门诊复查或电话随访，研究主要终点定义为全因死亡或心脏移植或植入左心室辅助装置的复合终点。比较基线时不同 FT3/FT4 比值组患者临床资料差异，采用多因素 Cox 比例风险回归模型分析 FT3/FT4 比值与住院心衰患者预后的关系。

【结果】总人群中年龄为（56.81±15.97）岁，男性为 2544 例（

72.1%）；中位随访时间为 2.79（1.00，5.03）年，最终随访时共记录了 1542 例终点事件。低 FT3/FT4组和高 FT3/FT4 组患者的年龄分别为 58.81±16.46，54.80±15.20（P<0.001），累积生存率分别为 38.4%，61.9%（P<0.001）。低 FT3/FT4 组患者左室射血分数较低，年龄、心率、N 末端 B 型脑钠肽原（NT-proBNP），肌钙蛋白 I、血肌酐、血钾等更高，而体重指数、白蛋白、血红蛋白、血钠、血脂等营养性指标更低。多因素Cox 回归模型分析显示，FT3[风险比（HR）=0.72，95%可信区间（CI） 0.63~0.84，P<0.001]、FT3/FT4（HR=0.76，95%CI 0.65~0.87，P<0.001）是心衰患者全因死亡、心脏移植或植入左心室辅助装置复合终点的相关因素。净重新分类指数（NRI）、整体鉴别指数（IDI）分析结果提示 FT3/FT4 比值联合对数转换的 NT-proBNP 均显著优于单独应用 FT3/FT4 比值（NRI=0.3，P<0.001； IDI=0.094，P<0.001）或 LogNT-proBNP（NRI=0.108，P<0.001；IDI=0.016，P<0.001）的预后价值。亚组分析提示 FT3/FT4 预测左室射血分数（LVEF）为<40%、40%~49%、≥50%患者复合终点风险的 HR（95%CI）分别为 0.91（0.77~1.08），0.83（0.50~1.39）， 0.65（0.50~0.85）（P 交互=0.045)。

【结论】低 FT3 和低 FT3/FT4 是住院心衰患者不良预后的重要危险因素，尤其在LVEF≥50%人群中 FT3/FT4 的预测作用更显著。

第三节 三碘甲状腺原氨酸与炎症指标的相关性及其对住院心力衰竭患者的影响

【目的】探索三碘甲状腺原氨酸（T3）与炎症指标之间的关系，及其对住院的心力衰竭（心衰）患者长期预后的潜在影响。

【方法】本研究为回顾性队列研究，连续纳入了 2009 年 3 月至 2018 年 6 月于中国医学科学院阜外医院心力衰竭重症监护病房住院的 2475 例心衰患者，根据甲状腺功能分为低 T3 综合征（LT3S）组（n=610，24.6%）和甲状腺功能（甲功）正常组（n=1865，75.4%）。出院后进行系统的门诊复查或电话随访，中位随访时间为 2.87（1.03，5.00）年，最终随访时共记录了 1048 例全因死亡。对比两组患者基线临床特征差异，采用 Cox 回归模型分析和 Kaplan-Meier 分析评估游离 T3（FT3）和高敏C 反应蛋白（hsCRP）对全因死亡风险的影响。

【结果】总人群中年龄（57±16）岁，其中男性 1823 例（73.7%）。LT3S 组患者白蛋白[（36.5±5.4）比（40.7±4.7）g/L]、血红蛋白[（129.4±25.1）比（140.6±20.6）g/L]、总胆固醇[3.6（3.0，4.4）比 4.2（3.5，4.9）mmol/L]低于甲功正常组（均 P<0.001），而年龄[（60.5±16.0）比（55.2±15.4）岁]、肌酐[105.0（83.6，137.0）比 87.8（75.6，106.3）mmol/L]、对数转化 N 末端 B 型利钠肽原[（8.2±1.3）比（7.2±1.4)）ng/L]高于甲功正常组（均 P<0.001）。Kaplan-Meier 生存分析提示低 FT3 组、高 hsCRP组患者累积生存率明显降低（P<0.001），低 FT3 联合高 hsCRP 亚组患者全因死亡风险最高（P 趋势<0.001）。限制性立方样条分析提示 FT3 与全因死亡呈负相关，而hsCRP 与全因死亡呈正相关。多因素 Cox 回归模型中，LT3S 是全因死亡的独立预测因子(HR=1.40，95% CI 1.16～1.69，P<0.001)。进一步对年龄、性别、体重指数、纽约心脏协会分级、冠心病、高血压、糖尿病，肌酐、hsCRP、血红蛋白、左心室射血分数及 N 末端 B 型利钠肽原等协变量进行了亚组分析，结果提示在 LT3S 与各亚组变量之间无明显相互作用（交互 P 值>0.05）。

【结论】LT3S 是住院心衰患者不良预后的独立预测因子，FT3 联合 hsCRP 可提高对心衰全因死亡的预测价值。

Chapter 1 Clinical features and prognosis of hospitalized heart failure patients with thyroid dysfunction

Background: Cardiovascular system is one of the target organs of thyroid hormone. Abnormal thyroid hormone may lead to the occurrence and development of cardiovascular diseases, and affect the prognosis of diseases. Our study aimed to investigate the prognostic role of thyroid dysfunction in hospitalized HF patients.

Methods: We performed a single-center retrospective cohort study including hospitalized 3733 HF patients between March 2009 to June 2018. Clinical evaluations for all participants were performed, including the collection of demographic characteristics and medical history, vital signs, 12-lead electrocardiography, echocardiography, and blood tests. A systematic outpatient review or telephone follow-up was conducted after discharge. The primary endpoint was defined as a combination of all-cause death or heart transplantation or left ventricular assist device implantation. The effects of thyroid hormone on the risk of composite endpoint were evaluated by Cox regression analysis and Kaplan-Meier analysis. Restricted Cubic spline (RCS) regression analysis was used to evaluate the relationship between thyroid hormone as a continuous variable and the composite end point.

Results: According to the thyroid function status, they were divided into the normal thyroid function group (n=1865, 50.0%), low triiodothyronine (T3) syndrome group (n= 610, 16.3%), hyperthyroidism group (n=42, 1.1%), subclinical hyperthyroidism (n=223, 6.0%), hypothyroidism (n=164, 4.4%), subclinical hypothyroidism (n=237,6.3%) and others (n=592,15.9%). The median follow-up time was 2.75 (1.00, 4.94) years, a total of 1659 endpoint events were recorded. Kaplan-Meier survival curve suggested that low free T3 (FT3) was associated with worse prognosis (P<0.001), and normal thyroid stimulating hormone (TSH) indicated the best prognosis (P<0.001). In RCS regression analysis, FT3 had a negative relationship with the risk of composite end points, while TSH showed a U-shaped curve distribution. Low FT3 (HR=1.33; 95%CI 1.15~1.54; P<0.001), elevated TSH (HR=1.37; 95%CI 1.15~1.64; P<0.001) , LT3S (HR=1.39; 95%CI 1.15~1.68; P<0.001), overt hyperthyroidism (HR=1.73; 95%CI 1.00~2.98; P=0.048), subclinical hypothyroidism (HR=1.43; 95%CI 1.13~1.82; P=0.003) and overt hypothyroidism (HR=1.76; 95%CI 1.33~2.34; P<0.001) independently increased the risk of composite endpoint.

Conclusion: LT3S, overt hyperthyroidism, subclinical and overt hypothyroidism were independently associated with poor outcomes in HF.

Chapter 2 The prognostic value of free triiodothyronine/free thyroxine ratio in patients hospitalized with heart failure

Objective To investigate the effect of free triiodothyronine/free thyroxine (FT3/FT4) ratio on the prognosis of patients with heart failure (HF).

Methods A total of 3527 patients hospitalized in Heart Failure Center of Fuwai Hospital from March 2009 to June 2018 were analyzed in our study. Patients were divided into two groups according to median of FT3/FT4 ratio: low FT3/FT4 group (n=1764, FT3/FT4<0.21) and high FT3/FT4 group (n=1763, FT3/FT4≥0.21). The primary endpoint was defined as a composite end point of all-cause death or heart transplantation or implantation of a left ventricular assist device. The baseline characteristics of patients with different FT3/FT4 ratio groups were compared, and the relationship between FT3/FT4 ratio and prognosis of hospitalized patients with HF were analyzed by a multivariate Cox proportional hazard regression model.

Results The age of the total population was (56.81±15.97) years, and 2544 cases (72.1%) were males. The median follow-up time was 2.79 (1.00, 5.03) years and a total of 1542 endpoint events were recorded at the final follow-up. The mean ages of patients in low FT3/FT4 group and high FT3/FT4 group were 58.81±16.46, 54.80±15.20 (P<0.001), and the cumulative survival rates were 38.4% and 61.9%, respectively (P<0.001). Low FT3/FT4 group had lower left ventricular ejection fraction, higher age, heart rate, N-terminal B-type brain natriuretic peptide (NT-proBNP), higher troponin I, serum creatinine and blood potassium, and lower nutritional indexes including lower body mass index, albumin, hemoglobin, sodium, blood lipid. Multivariate Cox regression model analysis showed that FT3 (HR=0.72, 95%CI: 0.63~0.84, P<0.001), FT3/FT4 (HR=0.76, 95%CI: 0.65~0.87, P<0.001) was associated with all-cause death, heart transplantation, or LVAD implantation in patients with HF. The results of net reclassification index (NRI) and integrated discrimination improvement (IDI) analysis indicated that the predictive value of the log (NT-proBNP) combined with FT3/FT4 ratio was significantly better than that of FT3/FT4 ratio alone (NRI=0.3, P<0.001; IDI=0.094, P<0.001) or Log (NT-proBNP) alone (NRI=0.108, P<0.001; IDI=0.016, P<0.001). Subgroup analysis suggested that the FT3/FT4 ratio was more significant in predicting the risk of composite endpoints in patients with LVEF≥50% (HR=0.65, 95%CI: 0.50~0.85, P for interaction=0.045).

Conclusion Low FT3 and low FT3/FT4 are important risk factors for poor prognosis in hospitalized HF patients, especially in patients with LVEF≥50%.

Chapter 3 Correlation between triiodothyronine and inflammatory factors and effect on hospitalized heart failure patients

Objective To investigate the association between triiodothyronine (T3) and inflammatory factors, and its potential effect on long-term outcomes in hospitalized patients with heart failure (HF). 

Methods A total of 2 475 patients with HF admitted to the Heart Failure Care Unit were consecutively enrolled in this retrospective cohort study from December 2006 to June 2018. Patients were divided into the low T3 syndrome group (n=610, 24.6%) andthe normal thyroid function group (n=1 865, 75.4%). The median follow-up time was 2.87 (1.03, 5.00) years. A total of 1 048 all-cause deaths were recorded at the final follow-up. The effects of free T3 (FT3) and high-sensitivity C-reactive protein (hsCRP) on the risk of all-cause death were evaluated by Cox regression analysis and Kaplan-Meier analysis.

Results The age of the total population was (57±16) years, 1 823 cases (73.7%) were male. Compared to those with normal thyroid function, albumin [(36.5±5.4) vs. (40.7±4.7) g/L], hemoglobin [(129.4±25.1) vs. (140.6±20.6) g/L], total cholesterol [3.6 (3.0, 4.4) vs. 4.2 (3.5, 4.9) mmol/L] (all P<0.001) were lower, Whereas age [(60.5±16.0) vs. (55.2±15.4) years], creatinine [105.0 (83.6, 137.0) vs. 87.8 (75.6, 106.3) mmol/L], log N-terminal B-type natriuretic peptide (NT-proBNP) [(8.2±1.3) vs. (7.2±1.4) ng/L] were higher in LT3S patients (all P<0.001). In Kaplan-Meier survival analysis, patients with lower FT3 and higher hsCRP had significantly lower cumulative survival (P<0.001), and lower FT3 combined with higher hsCRP subgroup had the highest risk of all-cause death (P_trend <0.001). Restricted cubic spline analysis showed that FT3 had negative relationship with all-cause death, while hsCRP had positive relationship with all-cause death. In multivariate Cox regression analysis, LT3S was an independent predictor of all-cause mortality (HR=1.40, 95%CI 1.16-1.69, P<0.001). Subgroup analysis of covariates such as age, sex, body mass index, New York Heart Association grade, coronary heart disease, hypertension, diabetes, creatinine, hsCRP, hemoglobin, left ventricular ejection fraction, and NT-proBNP showed no significant relationship between LT3S and subgroup variables (interaction P value <0.05).

Conclusion LT3S is an independent predictor of poor prognosis in patients with heart failure. FT3 combined with hsCRP improve the predictive value of all-cause death in hospitalized patients with heart failure.