第一部分 中低危急性胸痛患者临床特征及mHEART评分评估

背景：急性胸痛是急诊常见症状，其病因复杂，涵盖从高危心血管急症到低危非心源性疾病的广泛谱系，易导致高危患者漏诊或低危患者过度诊疗。流行病学数据显示，我国急诊胸痛患者中仅27.4%为急性冠脉综合征（Acute coronary syndrome，ACS），而63.5%为非心源性胸痛。传统评估工具如GRACE（Global Registry of Acute Coronary Events）、TIMI（Thrombolysis in Myocardial Infarction）评分多针对已确诊ACS患者，对未明确诊断的中低危患者适用性有限。HEART评分虽被推荐用于急诊胸痛分层，但其结合高敏肌钙蛋白（High-sensitivity cardiac troponin，hs-cTn）动态检测的改良版本——mHEART评分在中国人群中的预测价值尚未明确。本研究旨在探究中低危急性胸痛患者的临床特征、急诊转归及mHEART评分的分布情况。

方法：本研究采取前瞻性、多中心、随机单盲对照的试验设计，纳入北京市3家三甲医院2021年5月至2023年8月就诊的1910例中低危急性胸痛患者。入选标准：年龄≥18岁，胸痛发作24小时内，心电图无ST段抬高型心肌梗死（ST-segment–elevation myocardial infarction，STEMI）表现。排除标准：血流动力学不稳定、明确高危疾病（肺栓塞、主动脉夹层）、预期寿命<30天等。患者按区组随机法分为mHEART评分组（n=957）和常规诊疗组（n=953）。mHEART评分组基于mHEART评分（含病史、心电图、年龄、危险因素及0/3小时hs-cTn检测）指导诊疗，低危患者（0-3分）被建议尽早离院，相对高危患者（≥4分）根据临床情况做进一步的检查或治疗。常规诊疗组按照临床常规进行诊疗与护理，仅记录mHEART评分分值，用于区分高低危人群。mHEART研究的主要终点为急诊室留观时间与30天内心脏辅助检查率，次要终点为30天内主要不良心血管事件（Major adverse cardiovascular events，MACE），包含心血管死亡、心肌梗死（Myocardial infarction，MI）、血运重建及因心脏事件住院。本章节重点关注急性胸痛患者临床特征、症状特点、mHEART评分分布及急诊转归。

结果：1）基线特征：患者平均年龄60.0±12.9岁，男性56.2%，合并高血压（56.1%）、吸烟（37.7%）、糖尿病（26.8%）的比例较高，组间基线数据均衡。胸痛以胸骨后/心前区为主（74.8%），女性伴随症状（头晕、乏力）发生率显著高于男性（P=0.022）。2）mHEART评分分布：评分3分（23.8%）和4分（24.2%）占比最高，低危组（0-3分）占43.0%，相对高危组（≥4分）占57.0%。mHEART评分赋分来源中年龄与危险因素占比更高。3）急诊诊断与转归：81.6%的患者排除了ACS，15.9%患者收入院，83.1%患者出院返家，无急诊留观期间的死亡。

结论：高血压、吸烟、糖尿病为中低危胸痛患者最常见的伴随疾病，女性胸痛患者的非典型症状需要关注， 81.6%的中低危急性胸痛患者可排除ACS，总体转归良好。

第二部分 mHEART评分对中低危急性胸痛患者危险分层的应用价值

背景：胸痛是急诊科最常见的主诉之一，传统的医学检查、心电图和心脏生物标志物等方法无法简便地区分急性冠脉综合征（Acute coronary syndrome, ACS）患者与以胸痛为主诉的低危患者。多项研究表明，HEART评分能够有效识别低危胸痛患者，但仍然存在一定比例的潜在ACS漏诊率。在HEART评分的基础上引入0-3小时的连续高敏肌钙蛋白（High-sensitivity cardiac troponin，hs-cTn）检测流程，称之为改良的HEART（mHEART）评分。在中国人群中，关于HEART评分及其改良方法的研究相对较少。本研究旨在采用mHEART评分前瞻性、干预性地探讨mHEART评分对中低危急性胸痛患者危险分层的应用价值。

方法：mHEART研究是一项前瞻性、随机单盲对照的多中心研究。研究采取的mHEART评分将传统肌钙蛋白检测替换为hs-cTn检测，并额外增加了3h检测。筛选纳入的患者经过随机化分组后分别接受对应的干预措施。mHEART评分组根据评分指导诊疗流程，常规诊疗组遵循指南采取相应的诊疗策略，在入组后进行为期1个月的随访。研究的主要观察指标是本次就诊的急诊留观时间、30天随访期内的心脏辅助检查率。次要观察指标为30天内主要不良心血管事件（Major adverse cardiovascular events，MACE），包含心血管死亡、心肌梗死（Myocardial infarction，MI）、血运重建及因心脏事件住院。采用Kaplan-Meier生存曲线分别评估mHEART评分组和常规诊疗组的30天MACE发生率。采用多因素Cox分析评估干预措施与30天MACE之间的关系，计算风险比（Hazards ratios, HR）和95%置信区间（Confidence interval, CI）。计算特异度、敏感度、阴性预测值（Negative predictive value，NPV）、阳性预测值（Postive predictive value，PPV），评价mHEART评分预测中低危急性胸痛患者30天MACE的能力。

结果： 研究共纳入1947例满足入排标准的中低危急性胸痛患者。修正后的意向性治疗（Modified intent to treat，mITT）人群1819例患者，符合方案（Per Protocol，PP）集人群1533例患者。主要结果如下：1）急诊留观时间：mITT人群mHEART评分组整体急诊留观时间中位数为2.8小时，较常规诊疗组的2.6小时略长（P=0.042），其中，mHEART评分组中相对高危患者急诊留观时间较常规诊疗组延长，有显著差异（P=0.015），低危患者中两组差异不显著（P=0.250）。2）心脏辅助检查率：mITT人群mHEART评分组总检查率为38.5%，高于常规诊疗组的32.2%（P=0.005）。在总体及相对高危患者中，mHEART评分组的总检查率和影像学检查率均高于常规诊疗组（P＜0.05）；在PP人群低危患者中，mHEART评分组相比常规诊疗组表现出较低的总体检查率及影像学检查率（P＜0.05）。3）30天MACE发生率：整体30天MACE发生率较低，30天MACE发生率mITT人群为10.9%，PP人群为12.0%。在mITT人群中，mHEART评分组与常规诊疗组在总体、相对高危及低危人群中30天MACE发生率均无显著差异（P均>0.05）。mHEART评分对mHEART评分组30天MACE预测的敏感度为0.909，NPV为0.975。

结论：mHEART评分可以快速筛选出低危的胸痛患者，与临床常规诊疗效果相当。而对于中高危的胸痛患者，mHEART评分的应用延长了急诊留观时间、增加了心脏辅助检查率。相较于临床常规诊疗，使用mHEART评分指导诊疗未能实现在不影响30天MACE的前提下降低心脏辅助检查率、缩短急诊留观时间的目标，在诊疗效率和MACE发生率均未显示出优势。

第三部分 mHEART评分中的危险因素对胸痛患者预后的预测价值

背景：胸痛是急诊常见症状，其中约25%患者诊断为心血管疾病，但非特异性胸痛常被漏诊或误诊，导致不良预后风险。既往的研究中HEART评分主要被用于可疑急性冠脉综合征（Acute coronary syndrome, ACS）的中低危急性胸痛患者的进一步危险分层。HEART评分对可疑ACS的中低危胸痛患者30天内及6周的主要不良心血管事件（Major adverse cardiovascular events，MACE）发生均有较高的预测价值。然而，经过改良后的mHEART评分中哪些因素对中国中低危急性胸痛患者具有更好的预测价值仍有待于进一步评价，以期为临床提供简便有效的评价方法。本研究旨在评估mHEART评分中的相关危险因素对中低危可疑ACS的胸痛患者30天内MACE及冠状动脉狭窄程度的预测价值。

方法：本研究基于急性胸痛患者早期危险分层及诊治流程的多中心随机对照研究做事后分析，纳入2021年5月至2023年8月就诊的可疑ACS的中低危胸痛患者903例，收集基线资料，包括人口学特征、心血管危险因素、症状特征、心电图结果和肌钙蛋白水平等，计算mHEART评分。评分0-3分为低风险，≥4分为高风险。通过电话或门诊随访30天内的MACE事件及心脏辅助检查结果。通过多因素Cox回归分析探究mHEART评分中各项因素对MACE事件的预测价值。在此基础上，删除年龄与危险因素两项，得到简化后的HET评分（包含症状评分、心电图评分及肌钙蛋白评分），比较HET评分与mHEART评分对30天MACE的预测价值。评价mHEART评分中不同的危险因素与30天MACE及重度冠脉狭窄（Severe coronary stenosis，SCS）的关联。使用多因素Cox回归模型分析mHEART评分与MACE的相关性，计算风险比（Hazards ratios, HR）和95%置信区间（Confidence interval, CI）。

通过受试者工作特征（Receiver operating characteristic, ROC）曲线计算曲线下面积（Area under the curve，AUC）评估预测价值。

结果：1）基线特征：根据有否MACE事件分为MACE组与无MACE组。MACE组较无MACE组男性比例更高（66.7% vs 55.3%），合并糖尿病（36.4% vs 25.1%）、高胆固醇血症（65.7% vs 42.7%）、吸烟（54.5% vs 36.2%）及饮酒史（33.3% vs 16.2%）更高（均P<0.05）。实验室指标中，MACE组空腹血糖、肌酸激酶同工酶（Creatine kinase isoenzymes，CK-MB）、高敏肌钙蛋白I（High-sensitivity cardiac troponin I，hs-cTnI）水平较无MACE组显著升高（P<0.001）。2）mHEART评分中不同的危险因素预测价值：症状评分（HR=4.567, 95%CI：2.310-9.032）和肌钙蛋白评分（HR=4.609, 95%CI:2.096-10.135）是MACE的独立预测因素，而年龄、心电图及危险因素评分未显示显著关联。mHEART评分对30天MACE的预测价值优于HET评分（AUC：0.782 vs 0.759），但两者均具有预测价值（P<0.001）。3）mHEART评分优化与简化：因症状和肌钙蛋白评分高分人群MACE发生率（46.3% vs 14.3%）、血运重建率（41.5% vs 12.4%）及影像学检查率（70.7% vs 38.6%）相较非肌钙蛋白及症状评分高分人群均显著更高（P<0.001）。4）mHEART评分与SCS的关联：mHEART评分≥5分的相对高危组中，冠脉狭窄>70%的比例显著高于低分组（70.7% vs 46.3%, P=0.002）。

结论：mHEART评分中症状与肌钙蛋白是预测30天MACE的主要因素。简化的HET评分对30天MACE事件也具有预测价值。mHEART评分≥5分与SCS具有相关性。

Part I. Clinical Characteristics and Assessment of Modified HEART Score in Patients with Low to Medium Risk Acute Chest Pain

Background: Acute chest pain is a common symptom in emergency medicine with a complex etiology covering a broad spectrum from high-risk cardiovascular emergencies to low-risk non-cardiogenic diseases, which can easily lead to underdiagnosis of high-risk patients or overdiagnosis of low-risk patients. Epidemiologic data show that only 27.4% of emergency chest pain patients in China have acute coronary syndrome (ACS), while 63.5% have non-cardiac chest pain. Traditional assessment tools such as GRACE (Global Registry of Acute Coronary Events) and TIMI (Thrombolysis in Myocardial Infarction) scores are mostly used for patients with confirmed ACS, and have limited applicability to low and medium-risk patients without a clear diagnosis. The HEART score is recommended for the classification of emergency chest pain, although it has been used for the classification of chest pain. Although the HEART score is recommended for emergency chest pain stratification, the predictive value of the mHEART score, a modified version of the HEART score combined with the dynamic detection of high-sensitivity cardiac troponin (hs-cTn), has not yet been clarified in the Chinese population. The aim of this study was to investigate the clinical characteristics, emergency department referral, and distribution of mHEART score in patients with low and intermediate-risk acute chest pain.

Methods: A prospective, multicenter, randomized single-blind controlled trial design was adopted in this study, and 1,910 patients with low-to-moderate-risk acute chest pain attending three tertiary hospitals in Beijing from May 2021 to August 2023 were included. Inclusion criteria: age ≥18 years, chest pain onset within 24 hours, no ST-segment-elevation myocardial infarction (STEMI) manifestations on electrocardiogram. Exclusion criteria: hemodynamic instability, definite high-risk diseases (pulmonary embolism, aortic coarctation), and life expectancy <30 days. Patients were divided into the mHEART score group (n=957) and the usual care group (n=953) according to the block randomization method. mHEART score group guided the care based on the mHEART score (including history, ECG, age, risk factors, and 0/3-hour hs-cTn test). low-risk patients (0-3 points) were advised to leave the hospital as early as possible, and relatively high-risk patients (≥4 points) according to the clinical situation Further examination or treatment was performed. The routine group was treated according to clinical practice, and only the mHEART score was recorded to differentiate between high- and low-risk groups. mHEART study's primary endpoints were the length of stay in the emergency room and the rate of cardiac assisted examination within 30 days, and the secondary endpoint was the rate of major adverse cardiovascular events (MACE) within 30 days, which included cardiovascular deaths and cardiovascular disease, and the rate of cardiac assisted examination within 30 days, and the rate of cardiac assisted examination within 30 days. MACE), encompassing cardiovascular death, myocardial infarction (MI), revascularization, and hospitalization for cardiac events. This section focuses on the clinical characteristics, symptomatic features, mHEART score distribution, and emergency referrals of patients with acute chest pain.

Results: 1) Baseline characteristics: the mean age of patients was 60.0±12.9 years, 56.2% were male, and there was a high percentage of comorbid hypertension (56.1%), smoking (37.7%), and diabetes mellitus (26.8%), with a balanced baseline data between groups. Chest pain was predominantly in the retrosternal/pre-cardiac region (74.8%), and the prevalence of concomitant symptoms (dizziness, malaise) was significantly higher in women than in men (P=0.022). 2) Distribution of mHEART scores: scores 3 (23.8%) and 4 (24.2%) were the highest in the group, with a low-risk group (0-3) accounting for 43.0%, and a relatively high-risk group (≥4) accounting for 57.0%. mHEART score assignment source had a higher proportion of age and risk factors.3) Emergency diagnosis and regression: 81.6% of patients were excluded from ACS, 15.9% were admitted to the hospital, 83.1% were discharged to home, and there were no deaths during emergency stay.

Conclusion: Hypertension, smoking, and diabetes mellitus were the most common concomitant diseases in patients with low and intermediate-risk chest pain, and atypical symptoms in female patients with chest pain required attention. ACS could be excluded in 81.6% of patients with low and intermediate-risk acute chest pain, and overall regression was favorable.

Part II. The Value of mHEART Score for Risk Stratification of Patients with Low to Medium Risk Acute Chest Pain

Background: Chest pain is one of the most common complaints in the emergency department, and traditional methods such as medical examination, electrocardiogram, and cardiac biomarkers are unable to easily differentiate patients with acute coronary syndrome (ACS) from low-risk patients whose main complaint is chest pain. Several studies have shown that the HEART score is effective in identifying patients with low-risk chest pain, but there is still a certain percentage of potential ACS underdiagnosis. The introduction of a 0-3 hour continuous High-sensitivity cardiac troponin (hs-cTn) testing process based on the HEART score is called the modified HEART (mHEART) score. There are relatively few studies on the HEART score and its modification in the Chinese population. The aim of this study was to prospectively and interventively investigate the application value of the mHEART score for risk stratification of patients with low and intermediate-risk acute chest pain using the mHEART score.

Methods: The mHEART study was a prospective, randomized single-blind controlled multicenter study. The mHEART score adopted for the study replaced the traditional troponin assay with the hs-cTn assay and added an additional 3h assay. Patients were screened and randomized to receive the respective interventions. mHEART score group was guided by the scores, and the conventional treatment group followed the guidelines for the treatment strategy, with a follow-up period of 1 month after enrollment. The primary observations of the study were the duration of emergency room stay at the current visit, and the rate of ancillary cardiac examinations during the 30-day follow-up period. Secondary observations were major adverse cardiovascular events (MACE) within 30 days, which included cardiovascular death, myocardial infarction (MI), revascularization, and hospitalization for cardiac events. Kaplan-Meier survival curves were used to assess the 30-day MACE incidence rates in the mHEART score group and the usual care group, respectively. The association between the interventions and 30-day MACE was assessed using multifactorial Cox analysis, and risk ratios (Hazards ratios, HR) and 95% confidence intervals (CI) were calculated. Specificity, sensitivity, negative predictive value (NPV), and positive predictive value (PPV) were calculated to evaluate the ability of mHEART score to predict 30-day MACE in patients with low and intermediate-risk acute chest pain.

Results: A total of 1947 patients with low and intermediate-risk acute chest pain who met the entry criteria were included in the study. The Modified intent to treat (mITT) population was 1819 patients and the Per Protocol (PP) set population was 1533 patients. The main results were as follows: 1) emergency room stay: the overall median emergency room stay in the mHEART score group of the mITT population was 2.8 hours, which was slightly longer than that of 2.6 hours in the conventional treatment group (P=0.042), of which there was a significant difference in the prolongation of the emergency room stay in the relatively high-risk patients in the mHEART score group compared with that of the conventional treatment group (P=0.015), and in the low-risk patients, there was no significant difference between the two groups. The difference was not significant (P=0.250). 2) Cardiac auxiliary examination rate: the total examination rate in the mHEART score group of the mITT population was 38.5%, which was higher than that in the routine diagnosis and treatment group of 32.2% (P=0.005). In overall and relative high-risk patients, the mHEART score group had a higher overall examination rate and imaging rate than the conventional clinic group (P<0.05); in low-risk patients in the PP population, the mHEART score group demonstrated a lower overall examination rate and imaging rate compared with the conventional clinic group (P<0.05). 3) 30-day MACE incidence rate: the overall 30-day MACE incidence rate was relatively low, the 30-day MACE incidence rate was 10.9% in the mITT population and 12.0% in the PP population. In the mITT population, there was no significant difference in 30-day MACE incidence between the mHEART score group and the conventional treatment group in the overall, relative high-risk and low-risk populations (all P>0.05). mHEART score had a sensitivity of 0.909 and an NPV of 0.975 for the prediction of 30-day MACE in the mHEART score group.

Conclusion: The mHEART score can quickly screen out low-risk chest pain patients, which is comparable to the effect of routine clinical practice. For patients with medium- to high-risk chest pain, the application of mHEART score prolonged the emergency room stay and increased the rate of cardiac auxiliary examination. Compared with routine clinical care, the use of the mHEART score to guide care failed to achieve the goals of decreasing the rate of cardiac assisted examination and shortening the duration of emergency room stay without affecting the 30-day MACE, and did not show any advantage in terms of both care efficiency and MACE incidence.

Part III. Predictive Value of Risk Factors in the mHEART Score for Prognosis in Patients with Chest Pain

Background: Chest pain is a common symptom in emergency medicine, of which approximately 25% of patients are diagnosed with cardiovascular disease, but nonspecific chest pain is often missed or misdiagnosed, leading to the risk of poor prognosis. The HEART score has been used in previous studies mainly for further risk stratification of low and intermediate-risk patients with acute chest pain in suspected acute coronary syndrome (ACS). The HEART score has a high predictive value for the occurrence of Major adverse cardiovascular events (MACE) both within 30 days and at 6 weeks in intermediate- and low-risk chest pain patients with suspected ACS. However, which factors of the modified mHEART score have better predictive value for Chinese patients with low and intermediate-risk acute chest pain still need to be further evaluated, with the aim of providing a simple and effective evaluation method for clinical use. The aim of this study was to evaluate the predictive value of the relevant risk factors in the mHEART score for MACE and the degree of coronary artery stenosis within 30 days in chest pain patients with low to intermediate risk of suspected ACS.

Methods: In this study, based on the post-hoc analysis of a multicenter randomized controlled study of early risk stratification and diagnostic and treatment processes in patients with acute chest pain, 903 patients with low and intermediate risk of chest pain suspected of ACS who were seen from May 2021 to August 2023 were included, and baseline data including demographic characteristics, cardiovascular risk factors, symptomatic characteristics, electrocardiographic findings, and troponin levels were collected to calculate the mHEART score. Scores 0-3 were categorized as low risk and ≥4 as high risk. MACE events and ancillary cardiac findings within 30 days were followed up by telephone or outpatient clinic. The predictive value of each factor in the mHEART score for MACE events was explored by multifactorial Cox regression analysis. On this basis, the two items of age and risk factors were deleted to obtain the simplified HET score (including symptom score, electrocardiogram score, and troponin score), and the predictive values of HET score and mHEART score for 30-day MACE were compared. To evaluate the association of different risk factors in mHEART score with 30-day MACE and severe coronary stenosis (SCS). The association between mHEART score and MACE was analyzed using multifactorial Cox regression models, and risk ratios (Hazards ratios, HR) and 95% confidence intervals (CI) were calculated.

The predictive value was assessed by calculating the Area under the curve (AUC) from the Receiver operating characteristic (ROC) curve.

Results: 1) Baseline characteristics: subjects were categorized into MACE and no MACE groups based on the presence or absence of MACE events. the MACE group had a higher proportion of males than the no MACE group (66.7% vs. 55.3%), and a history of comorbidities of diabetes mellitus (36.4% vs. 25.1%), hypercholesterolemia (65.7% vs. 42.7%), tobacco use (54.5% vs. 36.2%), and alcohol use (33.3% vs. 16.2%). 33.3% vs 16.2%) were higher (all P<0.05). Among the laboratory parameters, fasting blood glucose, creatine kinase isoenzymes (CK-MB), and high-sensitivity cardiac troponin I (hs-cTnI) levels were significantly higher in the MACE group compared with the no-MACE group (P<0.001). 2) The predictive value of different risk factors in the mHEART score: symptom score (HR=4.567, 95% CI: 2.310-9.032) and troponin score (HR=4.609, 95% CI: 2.096-10.135) were independent predictors of MACE, whereas age, electrocardiogram, and risk factor scores did not show significant associations. The predictive value of the mHEART score for 30-day MACE was better than that of the HET score (AUC: 0.782 vs 0.759), but both had predictive value (P<0.001). 3) Optimization and simplification of the mHEART score: due to symptomatic and troponin scores in the high-scoring population the incidence of MACE (46.3% vs 14.3%), haemotransfusion (41.5% vs 12.4%) and imaging (70.7% vs 38.6%) were significantly higher in the non-troponin and symptom-scored high-score group (P<0.001). 4) Association of mHEART score with SCS: The proportion of coronary stenosis >70% was significantly higher in the relatively high-risk group with mHEART score ≥5 than in the low group (70.7% vs 46.3%, P=0.002).

Conclusion: Symptoms and troponin in the mHEART score are major predictors of 30-day MACE. The simplified HET score also had predictive value for 30-day MACE events. mHEART score ≥5 correlated with SCS.