第一部分 肥厚型心肌病的预后评估

 

(一)心率减速力和心率加速力在肥厚型心肌病风险预测中的价值

 

中文摘要

 

研究背景：

肥厚型心肌病(hypertrophic cardiomyopathy，HCM)患者常存在心脏自主神经功能异常，但其与不良预后的关系尚不明确。心率减速力(deceleration capacity, DC)和心率加速力(acceleration capacity, AC)是能够全面捕捉与心率减速和加速相关的振荡，从而对心脏整体自主神经张力进行综合评估的指标。

研究目的：

本研究旨在探讨DC和AC能否预测HCM患者的不良临床结局。

研究方法：

本研究是一项前瞻性观察性队列研究，于2013年至2022年间共纳入856例无血缘关系的HCM患者。研究的主要终点为全因死亡，次要终点为心血管死亡。DC和AC被分为三分位：低、中、高。所有患者均通过血液样本提取基因组DNA进行全基因组测序。入组患者中的697例接受了对比增强心脏磁共振成像。研究采用限制性立方样条(restricted cubic spline, RCS)结合Cox比例风险模型评估DC或AC与主要终点之间关系的线性假设。研究还采用Cox比例风险回归分析计算风险比(hazard ratios, HRs)和95%可信区间(confidence intervals, CIs)，以评估DC和AC与临床结局的关联。此外，研究通过受试者工作特征(receiver operator characteristic, ROC)曲线评估并比较DC、AC和传统心率变异性指标对主要终点的预测能力。

研究结果：

在中位随访3.0年(四分位数：2.3-3.9年)期间，本研究共记录了25例全因死亡事件、17例心血管死亡事件和6例心源性猝死事件。队列中有236例(27.6%)患者为基因型阳性。在接受对比增强心脏磁共振成像的患者中，582例(83.5%)存在晚期钆增强。DC与患者年龄、心率、左心房直径和纽约心脏协会(New York Heart Association, NYHA)心功能分级呈显著负相关，相关系数分别为-0.310、-0.133、-0.136和-0.122。相反，AC与这些指标呈显著正相关关系，相关系数分别为0.308、0.149、0.157和0.102。DC与AC之间存在强负相关关系，相关系数为-0.840。DC与主要终点风险呈非线性关系，而AC则呈线性关系。

随着DC水平的降低，患者的死亡风险变大，且风险在DC值低于4.6 ms时陡然增加。DC和AC均与全因死亡显著相关(DC：校正HR=0.733 [0.572-0.939], p = 0.014; AC：校正HR=1.442 [1.137-1.828], p = 0.003)。而仅有AC与心血管死亡显著相关(DC: 校正HR=0.795 [0.592-1.069], p = 0.130; AC: 校正HR=1.380 [1.034-1.841], p = 0.029)。将DC和AC作为三分位进行分析验证了上述结果。与高DC组相比，低DC组的全因死亡风险显著更高(校正HR=3.975 [1.215-13.006], p = 0.023)，虽然与心血管死亡无显著关联。高AC组的全因死亡风险则显著高于低AC组(校正HR=4.067 [1.247-13.269], p = 0.020)，也与心血管死亡无显著关联。

心率变异性指标与DC和AC均呈显著相关关系。ROC曲线分析显示，AC在预测全因死亡方面的曲线下面积显著大于相邻正常RR间期差异的均方根(the root mean square of the successive differences of all normal-to-normal differences, RMSSD)和差异超过50ms的正常RR间期占比(the proportion of normal-to-normal intervals more than 50ms different, pNN50)两项心率变异性指标(p = 0.032和0.026)。

研究结论：

低DC和高AC值与HCM患者全因死亡高风险独立相关。DC和AC可作为有效的电生理学指标，用于HCM患者的预后评估。

 

轻症梗阻性肥厚型心肌病患者早期室间隔切除术的远期预后

 

中文摘要

 

研究背景：

在肥厚型梗阻性心肌病(hypertrophic obstructive cardiomyopathy, HOCM)患者中，室间隔切除术可以完全缓解梗阻，被推荐用于治疗梗阻相关的药物难治性严重症状。然而，目前尚缺乏证据支持对轻症HOCM患者行室间隔切除术。

研究目的：

本研究旨在明确早期通过室间隔切除术缓解左心室流出道梗阻是否能改善轻症HOCM患者的长期预后。

研究方法：

研究共纳入424例静息状态下流出道梗阻梯度(left ventricular outflow trend gradient, LVOTG)≥50 mmHg且症状轻微(纽约心脏协会[New York Heart Association， NYHA]心功能分级 I-II级，无晕厥)的HOCM患者。主要终点为全因死亡，次要终点为心血管死亡(包括心源性猝死、心力衰竭相关死亡和卒中相关死亡)。研究采用逆概率处理加权法(inverse probability of treatment weighting, IPTW)以控制组间基线的不平衡，随后通过最小绝对收缩和选择算子(least absolute shrinkage and selection operator, LASSO)惩罚回归模型筛选对终点最具预测性的变量，最后利用Cox比例风险模型调整可能影响判断室间隔切除术与临床结局之间关联性的混杂因素。最终使用风险比(hazard ratios, HRs)和95%可信区间(confidence intervals, CIs)评估室间隔切除术与临床结局之间的相关性。为验证主要分析结果的稳健性，我们利用1：1倾向性评分匹配(propensity score matching, PSM)进行敏感性分析，对PSM匹配队列亦采用LASSO惩罚的Cox比例风险回归模型调整混杂因素。

研究结果：

在中位随访5.1年(四分位数：3.4-8.5年)期间，37例患者(8.7%)发生了全因死亡事件，27例患者(6.4%)发生了心血管死亡事件。与未接受手术的患者相比，接受室间隔切除术的患者年龄更小(44.6±13.3岁 vs. 53.4±12.1岁，p < 0.001)，NYHA II级的比例更高(194例 [82.2%] vs. 117例 [62.2%]，p < 0.001)，存在胸闷症状的比例也更高(185例 [78.4%] vs.125例[66.5%]，p = 0.008)。

在IPTW有效平衡了关键协变量后，LASSO惩罚Cox比例风险回归模型也证实了手术的保护作用(全因死亡：调整后HR=0.238 [95%CI: 0.102-0.556], p = 0.001；心血管死亡：调整后HR=0.191 [95%CI: 0.065-0.567], p = 0.003)。除室间隔切除术外，左心室最大心室壁厚度、LVOTG、以及左心房直径也与终点事件显著相关。敏感性分析中，PSM匹配后的结果与IPTW加权后的结果一致(全因死亡：调整后HR=0.220 [95%CI: 0.058-0.840], p=0.027；心血管死亡：调整后HR=0.161 [95%CI: 0.031-0.828], p=0.029)。

研究结论：

本研究表明，对于静息LVOTG≥50 mmHg但尚未进展至严重症状的HOCM患者，早期室间隔切除术具有显著的保护作用。

 

第二部分 心肌病基因变异与表型数据库及基因诊断解读系统的建立

 

中文摘要

 

研究背景：

心肌病是心血管系统最常见的单基因遗传病，基因检测在心肌病患者及其家系的早期诊断、鉴别诊断、风险预测、临床诊疗指导及选择性生育等方面具有重要价值。然而，基因检测结果的判读已成为限制其临床应用的关键性瓶颈问题。

研究目的：

针对目前心肌病基因诊断面临的瓶颈，本研究通过建立适用于心肌病的基因变异与表型数据库，以及标准化的基因诊断解读系统，以促进基因诊断和风险预测在心肌病中的应用。

研究方法：

本研究通过系统全面地梳理肥厚型心肌病(hypertrophic cardiomyopathy, HCM)、扩张型心肌病(dilated cardiomyopathy, DCM)和致心律失常性心肌病(arrhythmogenic cardiomyopathy, ACM)的国内外文献，提取了公开的遗传及包括了预后的表型信息。本研究在美国医学遗传学与基因组学学会和分子病理学协会(American College of Medical Genetics and Genomics and the Association for Molecular Pathology, ACMG/AMP)指南框架下，针对心肌病和其致病基因的特殊特征进行优化，开发了适用于心肌病的基因变异与表型数据库和标准化的基因诊断解读系统。我们将该解读系统给出的致病性评级与ClinVar提供的致病性评级进行了对比，并利用数据库中的预后数据分析了代表性基因MYH7和MYBPC3不同位置变异携带患者的预后差异。心肌病基因变异与表型数据库及基因诊断解读系统已以网站的形式展现(https://cardiogene.bgi.com/#/home)。

研究结果：

本研究最终纳入HCM 856篇HCM相关文献， 435篇DCM相关文献， 288篇ACM相关文献。研究共纳入了7255个患者个体表型，89个患者群体表型，以及62074个变异位点。根据三种心肌病的疾病特征，本研究选择了对疾病预后影响较大且文献报道频繁的表型以及硬终点事件进行收录。在ACMG/AMP指南规则的基础上，本研究细化或更新了PVS1、PS3、PS4、PM1、PM2、PM4、PM5、PP2、以及PP3证据项规则。在证据项汇总时，本研究使用基于贝叶斯框架的证据项赋分规则赋予变异致病性级别。数据库共收录了23401个HCM报道病例，其中17925个病例携带致病/疑似致病位点；收录了7450个DCM报道病例，其中4771个病例携带致病/疑似致病位点；收录了5689个ACM报道病例，其中4057个病例携带致病/疑似致病位点。

基于62074个收录的变异位点，将本数据库致病性分级与ClinVar提供的致病性分级进行了比较。在本数据库中被分类为致病/疑似致病的3573个变异中，有2477(69.3%)在ClinVar中也被分级为致病/疑似致病，而有709 (19.8%)为ClinVar未收录的位点。在ClinVar中认为是意义未明或证据冲突的32658个变异中，383个(1.2%)变异在本系统中被纠正为致病/疑似致病，1002个(3.1%)变异在本系统中被纠正为良性/疑似良性。在我们收录的1387个ClinVar未收录的变异位点中，共有709个(51.1%)变异评级为致病/疑似致病，671个(48.4%)变异评级为意义未明，7个(0.5%)变异评级为良性/疑似良性。

 根据编码序列位置对数据库纳入的MYH7和MYBPC3基因变异位点进行了定位分析。MYH7基因的全部变异位点以及致病/疑似致病位点广泛分布在基因编码区域内，但主要集中在22号外显子及之前。MYBPC3 基因的全部变异位点及致病/疑似致病位点也广泛分布在该基因的全部基因编码区，未呈现明显的区域聚集现象。MYH7在外显子15和23附近的变异携带患者的死亡率稍高，而MYBPC3在外显子17和30附近达到了携带患者的死亡率曲线顶点，提示了这些位置的突变可能表现为更差的临床预后。

研究结论：

本研究开发了全球首个针对心肌病的基因变异与表型数据库及基因诊断解读系统，有望显著提高心肌病基因诊断的标准化水平。作为目前最大的综合了中英文文献来源数据的心肌病遗传-表型数据库及基因诊断解读系统，可为心肌病的风险预测提供新的参考。

 

Part I: Prognostic Assessment of Hypertrophic Cardiomyopathy

 

(1) The Value of Deceleration Capacity and Acceleration Capacity in Risk Prediction for Hypertrophic Cardiomyopathy

 

Abstract

 

Background: Abnormal cardiac autonomic function is commonly observed in hypertrophic cardiomyopathy (HCM) patients; however, its association with prognosis remains unclear. Deceleration capacity (DC) and acceleration capacity (AC) of heart rate are novel indices that can capture all deceleration and acceleration-related oscillations of heart rate, offering a comprehensive assessment of overall tonic autonomic activity.

Objective: This study aims to determine whether DC and AC can predict adverse outcomes in HCM patients.

Methods: In this prospective observational cohort study, 856 unrelated HCM patients were enrolled between 2013 and 2022. The primary outcome was all-cause death, while the secondary outcome was cardiovascular death. DC and AC were categorized into tertiles: low, middle, and high. All patients involved in the study underwent whole genome sequencing. Contrast cardiac magnetic resonance imaging was performed in 697 patients. Restricted cubic spline (RCS) analysis with Cox proportional hazard models was performed to evaluate the linearity regarding the relationship between DC or AC and the primary outcome. Cox proportional hazards regression analysis was employed to calculate the hazard ratios (HRs) and 95% confidence intervals (CIs) for evaluating the associations of DC and AC with clinical outcomes. Receiver-operator characteristic (ROC) curves were generated to assess and compare predictive abilities of DC, AC, and heart-rate variability.

Results: During a median follow-up of 3.0 years (IQR: 2.3 - 3.9 years), 25 cases of all-cause death, 17 cases of cardiovascular death, and 6 cases of sudden cardiac death were recorded. Among the cohort, 236 (27.6%) patients were genotype positive. Of the 697 patients who underwent concomitant contrast cardiac magnetic resonance imaging, 582 (83.5%) exhibited late gadolinium enhancement. DC demonstrated significant negative correlations with age, heart rate, left atrial diameter, and NYHA class, yielding correlation coefficients of -0.310, -0.133, -0.136, and -0.122, respectively. Conversely, AC showed positive correlations with these indices, with correlation coefficients of 0.308, 0.149, 0.157, and 0.102, respectively. A strong negative correlation was observed between DC and AC, with a correlation coefficient of -0.84. A non-linear association between DC and the risk of all-cause mortality was found, while AC demonstrated a linear relationship. The risk of mortality increased with lower levels of DC, plateauing at values below 4.6 ms. DC and AC were strongly associated with all-cause death (DC: adjusted HR=0.733 [0.572-0.939], p = 0.014; AC: adjusted HR=1.442 [1.137-1.828], p = 0.003). However, only AC was significanlty associated with cardiovascular death (DC: adjusted HR=0.795 [0.592-1.069, p = 0.130; AC: adjusted HR=1.380 [1.034-1.841], p = 0.029). These findings were further confirmed by categorizing DC and AC into tertiles. Individuals in the low DC group exhibited a significantly higher risk of all-cause death (adjusted HR=3.975 [1.215-13.006], p = 0.023) but not of cardiovascular death, compared to the high DC group after fully adjustment. While those in the high AC group demonstrated an increased risk for all-cause death (adjusted HR=4.067 [1.247-13.269], p = 0.020) than those with low AC levels, and also no significant association was found for cardiovascular mortality. Indices of heart rate variability showed significant correlations with both DC and AC. ROC curve analysis revealed that the area under the curve for the root mean square of the successive differences of all normal-to-normal differences (RMSSD) and the proportion of normal-to-normal intervals more than 50 ms different (pNN50) in predicting all-cause mortality was significantly smaller than that for AC (p = 0.032 and 0.026, respectively), although the comparison with DC did not reach statistical significance.

Conclusion: Low DC and high AC values are independently associated with an elevated risks of all-cause mortality in HCM patients. These findings suggest that DC and AC may serve as effective electrophysiological markers for risk prediction in this population.

 

(2) Prognostic Analysis of Septal Myectomy for Mildly Symptomatic Hypertrophic Obstructive Cardiomyopathy

 

Abstract

 

Background: In patients with severe hypertrophic obstructive cardiomyopathy (HOCM), septal myectomy is strongly indicated for intractable severe symptoms. However, there is a lack of evidence supporting the survival benefits of myectomy for HOCM individuals with mild symptoms.

Objective: The aim of this study is to clarify whether early surgical relief of left ventricular outflow tract obstruction by myectomy provides clinical benefit for the long-term prognosis of HOCM patients with mild symptoms.

Methods: A total of 424 HCM patients with resting left ventricular outflow trend gradient (LVOTG) ≥50 mmHg and mild symptoms (New York Heart Association [NYHA] I-II without syncope) were enrolled. The primary outcome was all-cause death, while the secondary outcome was cardiovascular death (including sudden cardiac death, heart failure-related death, and stroke-related death). Inverse probability of treatment weighting (IPTW) was employed to balance baseline characteristics between groups, followed by least absolute shrinkage and selection operator (LASSO)-penalized Cox proportional hazards regression to identify the most predictive variables for endpoints and adjust for potential confounding factors that might affect the assessment of the relationship between myectomy and clinical outcomes. Hazard ratios (HRs) and 95% confidence intervals (CIs) are for assessing the associations between septal myectomy and clinical outcomes. To validate the robustness of the primary analysis, we performed sensitivity analysis using 1:1 propensity score matching (PSM), with LASSO-penalized Cox regression also applied to the PSM-matched cohort.

Results: Over a median follow-up of 5.1 years [IQR: 3.4-8.5 years]), 37 patients (8.7%) experienced all-cause death and 27 patients (6.4%) experienced cardiovascular death. Compared to patients who did not undergo surgery, those who underwent septal myectomy were younger (44.6 ± 13.3 years vs. 53.4 ± 12.1 years, p < 0.001), had a higher proportion of NYHA class II (194 [82.2%] vs. 117 [62.2%], p < 0.001), and exhibited a higher prevalence of chest tightness (185 [78.4%] vs. 125 [66.5%], p = 0.008). Before adjustment, septal myectomy was associated with lower risks of both all-cause death (HR=0.346 [95% CI: 0.164-0.730], p = 0.005) and cardiovascular death (HR=0.301 [95% CI: 0.126-0.724], p = 0.007). After IPTW effectively balancing key covariates, LASSO-penalized Cox proportional hazards regression confirmed the protective effect of surgery (all-cause death: adjusted HR=0.238 [95% CI: 0.102-0.556], p = 0.001; cardiovascular death: adjusted HR=0.191 [95% CI: 0.065-0.567], p = 0.003). Besides myectomy, maximal left ventricular wall thickness, LVOTG and left atrial diameter were also significantly associated with two endpoint events. In sensitivity analysis, PSM yielded consistent results with IPTW (all-cause death: adjusted HR=0.220 [95% CI: 0.058-0.840], p = 0.027; cardiovascular death: adjusted HR=0.161 [95% CI: 0.031-0.828], p = 0.029).

Conclusions: This study demonstrates that myectomy treatment is independently associated with a lower risk of all-cause death and cardiovascular death in HOCM patients with an LVOTG over 50 mmHg and mild symptoms.

 

Part II: The Establishment of a Cardiomyopathy Genetic Variant and Phenotype Database along with an Interpretation System for Genetic Diagnosis

 

Abstract

 

Background: Cardiomyopathy is the most common monogenic inherited disease of the cardiovascular system. Genetic testing holds significant value in the early diagnosis, risk stratification, clinical management, and selective reproduction of patients with cardiomyopathy and their families. However, the interpretation of genetic testing results has become a critical obstacle limiting its clinical application.

Objective: To address the current challenges in the genetic diagnosis of cardiomyopathy, this study aims to establish a database of variants and phenotypes specific in cardiomyopathy, and a standardized genetic diagnostic interpretation system, thereby promoting the application of genetic diagnosis in cardiomyopathy.

Methods: We systematically reviewed the literature on hypertrophic cardiomyopathy (HCM), dilated cardiomyopathy (DCM), and arrhythmogenic cardiomyopathy (ACM), extracting available genetic and phenotypic information. Under the framework of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP) guidelines, we developed the first database of variants and phenotypes for cardiomyopathy, and established a standardized genetic diagnostic interpretation system. We compared the pathogenicity classifications provided by this interpretation system with those from ClinVar and analyzed prognostic differences among patients carrying variants at different loci of the representative genes MYH7 and MYBPC3 using the prognostic data in the database. The cardiomyopathy genetic variant and phenotype database, along with the genetic diagnostic interpretation system, have been presented as a website (https://cardiogene.bgi.com/#/home).

Results: The database included 856 HCM, 435 DCM, and 288 ACM literatures in total. The database encompasses 7,255 individuals, 89 cohorts, and 62,074 variants. Phenotypes and endpoints frequently reported in the literature and prognostically significant for the diseases were included. Based on the ACMG/AMP guidelines, we refined or updated the rules including PVS1, PS3, PS4, PM1, PM2, PM4, PM5, PP2, and PP3. A Bayesian framework-based scoring system was used to determine the classification of variants. The database encompassed a total of 23,401 reported HCM cases, among which 17,925 cases carried pathogenic/likely pathogenic variants; 7,450 reported DCM cases were included, with 4,771 cases carrying pathogenic/likely pathogenic variants; and 5,689 reported ACM cases were documented, among which 4,057 cases carried pathogenic/likely pathogenic variants.

Based on 62,074 included variants, we compared the pathogenicity classification in our database with that provided by ClinVar. Among the 3,573 variants classified as pathogenic/likely pathogenic in our database, 2,477 (69.3%) were also categorized as pathogenic/likely pathogenic in ClinVar, while 709 (19.8%) were variants not included in ClinVar. Among the 32,658 variants classified as variants of uncertain significance or with conflicting interpretations in ClinVar, 383 (1.2%) variants were reclassified as pathogenic/likely pathogenic in our system, and 1,002 (3.1%) variants were reclassified as benign/likely benign in our system. Among the 1,387 ClinVar-unreported variants included in our database, a total of 709 (51.1%) variants were rated as pathogenic/likely pathogenic, 671(48.4%) variants were rated as variants of uncertain significance, and 7(0.5%) variants were rated as benign/likely benign.

A localization analysis was performed on the variants of MYH7 and MYBPC3 based on coding sequence positions. All variants and pathogenic/likely pathogenic variants in MYH7 were widely distributed across the coding region but were primarily concentrated before exon 22. All variants and pathogenic/likely pathogenic variants in MYBPC3 were broadly distributed across its entire coding region without significant regional clustering. Variants near exons 15 and 23 of MYH7 were associated with slightly higher patient mortality, while variants near exons 17 and 30 of MYBPC3 peaked in patient mortality curves, suggesting that variants at these locations may correlate with worse clinical outcomes.

Conclusion: This study developed the first variant and phenotype database and genetic diagnostic interpretation system specifically for cardiomyopathy, which is expected to significantly improve the standardization of genetic diagnosis in cardiomyopathy. As the largest integrated genetic-phenotype database for cardiomyopathy combining Chinese and English literature sources, it provides a new reference for the risk prediction in cardiomyopathy.