Coagulation FX (factor X) is a serine protease that catalyzes the formation of fibrin clots. Although this maintains hemostasis, it can also result in pathological thrombi and emboli.^1,2 FXa (activated FX) inhibitors are efficacious for preventing deep venous thrombosis, pulmonary embolism,¹ and cardioembolic stroke related to nonvalvular atrial fibrillation.² However, the efficacy of FXa inhibitors for other forms of cardiovascular disease is not known.

Genetic variants related to circulating levels of a coagulation factor can be used as instrumental variables in Mendelian randomization analyses to study the effects of drugs that inhibit that coagulation factor.³ The aim of this work was to employ a genetic instrument for circulating FXa levels in an exploratory investigation into the effects of varying FXa levels on cardiovascular outcomes.

As the instrument for circulating FXa levels, we used the rs61753266 variant in the F10 gene that has been associated with plasma FXa levels at P=8×10^-15 in a study of 3301 European-ancestry individuals.⁴ Although the rs547138 variant in the F10 gene was associated with plasma FXa levels at P=6×10^-22 in the same study, it was also associated with Protein Z–dependent protease inhibitor at P=5×10^{-4 4}. Given the role of Protein Z–dependent protease inhibitor in inhibiting FXa, this likely represents a pleiotropic association that could bias Mendelian randomization analyses investigating the effect of FXa, and therefore this variant was not included. The rs61753266 variant we used had a weaker association with circulating Protein Z–dependent protease inhibitor (P=0.05).

We used the UK Biobank to perform analyses, considering the outcomes of coronary artery disease (CAD), peripheral artery disease, ischemic stroke, intracerebral hemorrhage, subarachnoid hemorrhage, deep vein thrombosis, and pulmonary embolism. Genetic association estimates for these outcomes were obtained from 367 570 unrelated European-ancestry participants. Cases were defined based on InternationalClassification of Diseases, Ninth and Tenth Revisions, Office of Population Censuses and Surveys Classification of Surgical Operations and Procedures (fourth revision), and participant self-reported data (Table). Only incident cases recorded until October 11, 2019 were considered. Relevant ethical approval for the UK Biobank study was obtained from the North West Multicentre Research Ethics Committee, and all participants provided informed consent. UK Biobank data are available on request (see the Acknowledgments section). The statistical code used is available from the corresponding author upon reasonable request.

Table. Associations of the Genetic Instrument for FXa Levels (rs61753266) With Cardiovascular Outcomes in the UK Biobank

FXa indicates activated factor X; HES, hospital episode statistics; ICD, International Classification of Diseases; and OPCS, Office of Population Censuses and Surveys.

* ICD-10 diagnoses were identified through both HES and death certificates (UK Biobank fields 40001 and 40002).

† Diagnoses made by a doctor (UK Biobank fields 6150 and 6152) and Medication for health condition (6177) were self-reported by participants by touchscreen entry; noncancer illness code (20002) and surgical operation code (20004) were self-reported by participants by interview with a nurse.

‡ Indicates that all subcodes are included.

Odds of each outcome per FXa decreasing allele of rs61753266 (frequency 0.3%) are detailed in the Table. Consistent with existing randomized-controlled trial data, the genetic instrument for lower FXa levels was associated with reduced deep vein thrombosis and pulmonary embolism risk.¹ This instrument is also associated with higher subarachnoid hemorrhage risk, in keeping with the increased bleeding risk associated with FXa inhibitors. These results, therefore, serve as a form of positive control, supporting the validity of the variant as an instrument for FXa inhibition. There was some suggestion that the genetic instrument for lower FXa levels may be associated with a reduced risk of ischemic stroke and peripheral artery disease, although the 95% CIs were broad. Inhibitors of FXa are known to be effective for reducing the risk of stroke related to nonvalvular atrial fibrillation.² The point estimates for the associations of the FXa lowering variant with CAD and intracerebral hemorrhage risk were close to the null.

This work represents an efficient way to explore the potential clinical applications of FXa inhibitors. The use of a genetic instrument for FXa levels helps overcome environmental confounding and reverse causation bias to allow causal inferences to be drawn. However, the approach also has limitations. Although the location of the variant we employ at the F10 locus and its association with plasma FXa levels both support its validity as an instrument for FXa levels,⁴ we cannot exclude the possibility that it affects risk of the considered outcomes through pathways independent of FXa levels, to bias the results of our analysis. Furthermore, as apparent from the confidence intervals of the results, our study had limited statistical power. Given the previously described association between our instrument for circulating FXa levels and CAD,⁵ and the role of FXa in inflammation, vascular remodeling, and fibrosis, our null finding for this outcome should be interpreted with caution. The discrepancy may, in part, be attributable to the criteria used to diagnose CAD—although Paraboschi et al⁵ considered angiographically confirmed cases and controls with no angiographic evidence of coronary atherosclerosis, our current study used International Classification of Diseases and Office of Population Censuses and Surveys codes, and self-report for case ascertainment (Table), with noncases considered as controls.

In conclusion, the findings of this Mendelian randomization study are consistent with the results of clinical trials in supporting an effect of FXa inhibition on reducing risk of venous thromboembolism and increasing risk of subarachnoid hemorrhage. The results did not support that FXa inhibition is associated with reduced CAD risk in a general population. Although further study is required to investigate the discrepancy with previous work,⁵ this information may be used to help prioritize future clinical trials.

Nonstandard Abbreviations and Acronyms

CAD	coronary artery disease
FX	factor X
FXa	activated Factor X
HES	hospital episode statistics
PAD	peripheral artery disease

coronary artery disease

factor X

activated Factor X

hospital episode statistics

peripheral artery disease

This research has been conducted using the UK Biobank resource (application 29202). The UK Biobank data is available on application (http://www.ukbiobank.ac.uk/register-apply).

Dr Gill is funded by the Wellcome Trust 4i Programme (203928/Z/16/Z) and British Heart Foundation Centre of Research Excellence (RE/18/4/34215) at Imperial College London. Dr Burgess is supported by Sir Henry Dale Fellowship jointly funded by the Wellcome Trust and the Royal Society (204623/Z/16/Z). The funding sources had no role in the design, acquisition of data, analysis, interpretation, or write up of this study.

Dr Gill is employed part-time by Novo Nordisk. The other author reports no conflicts.

For Sources of Funding and Disclosures, see page 552.

中文翻译：

---

使用与循环 FXa（活化因子 X）水平相关的遗传变异来代理 FXa 抑制对心血管结果的影响

Coagulation FX（因子 X）是一种丝氨酸蛋白酶，可催化纤维蛋白凝块的形成。虽然这可以维持止血，但也可能导致病理性血栓和栓子。^1,2 FXa（活化的 FX）抑制剂可有效预防与非瓣膜性心房颤动相关的深静脉血栓形成、肺栓塞、^{1和心源性卒中。2}然而，FXa 抑制剂对其他形式的心血管疾病的疗效尚不清楚。

与凝血因子循环水平相关的遗传变异可用作孟德尔随机分析中的工具变量，以研究抑制该凝血因子的药物的作用。³这项工作的目的是使用一种遗传工具来检测循环 FXa 水平，以探索不同 FXa 水平对心血管结果的影响。

作为循环 FXa 水平的工具，我们在对 3301 名欧洲血统个体的研究中使用了F10基因中的 rs61753266 变体，该变体与P = 8×10 ^-15的血浆 FXa 水平相关。^{4虽然在同一研究中，} F10基因中的 rs547138 变异与血浆 FXa 水平相关，P =6×10 ^-22，但它也与蛋白 Z 依赖性蛋白酶抑制剂相关，P =5×10 ^{-4 4}. 鉴于蛋白质 Z 依赖性蛋白酶抑制剂在抑制 FXa 中的作用，这可能代表了一种多效性关联，可能会偏向研究 FXa 影响的孟德尔随机分析，因此不包括该变体。我们使用的 rs61753266 变体与循环蛋白 Z 依赖性蛋白酶抑制剂的关联较弱（P = 0.05）。

我们使用英国生物银行进行分析，考虑到冠状动脉疾病 (CAD)、外周动脉疾病、缺血性中风、脑出血、蛛网膜下腔出血、深静脉血栓形成和肺栓塞的结局。这些结果的遗传关联估计来自 367 570 名无关的欧洲血统参与者。病例是根据国际疾病分类第九和第十修订版定义的，人口普查和调查办公室分类的外科手术和程序（第四次修订），以及参与者自我报告的数据（表）。仅考虑 2019 年 10 月 11 日之前记录的事件案例。英国生物银行研究的相关伦理批准来自西北多中心研究伦理委员会，所有参与者都提供了知情同意书。英国生物银行数据可应要求提供（见致谢部分）。所使用的统计代码可根据合理要求从通讯作者处获得。

桌子。FXa 水平 (rs61753266) 遗传仪器与英国生物库心血管结果的关联

FXa 表示激活因子 X；HES，医院发作统计；ICD，国际疾病分类；和 OPCS，人口普查和调查办公室。

* ICD-10 诊断是通过 HES 和死亡证明（UK Biobank 字段 40001 和 40002）确定的。

† 医生的诊断（UK Biobank 字段 6150 和 6152）和健康状况药物治疗（6177）由参与者通过触摸屏输入进行自我报告；非癌症疾病代码（20002）和外科手术代码（20004）由参与者通过与护士面谈的方式自我报告。

‡ 表示包含所有子代码。

每个 FXa 减少等位基因 rs61753266（频率 0.3%）的每个结果的几率在表中详述。与现有的随机对照试验数据一致，用于降低 FXa 水平的遗传工具与降低深静脉血栓形成和肺栓塞风险相关。¹该仪器还与较高的蛛网膜下腔出血风险相关，与 FXa 抑制剂相关的出血风险增加一致。因此，这些结果可作为阳性对照的一种形式，支持该变体作为 FXa 抑制工具的有效性。有一些建议表明，降低 FXa 水平的遗传工具可能与降低缺血性中风和外周动脉疾病的风险有关，尽管 95% 的 CI 范围很广。已知 FXa 抑制剂可有效降低与非瓣膜性心房颤动相关的卒中风险。²降低 FXa 变异与 CAD 和脑出血风险相关性的点估计值接近于零。

这项工作代表了探索 FXa 抑制剂潜在临床应用的有效方法。使用 FXa 水平的遗传工具有助于克服环境混杂和反向因果偏差，从而可以得出因果推论。然而，该方法也有局限性。尽管我们在F10基因座上使用的变体的位置及其与血浆 FXa 水平的关联都支持其作为 FXa 水平工具的有效性，⁴我们不能排除它通过独立于 FXa 水平的途径影响所考虑结果的风险的可能性，以使我们的分析结果产生偏差。此外，从结果的置信区间可以明显看出，我们的研究具有有限的统计能力。鉴于之前描述的循环 FXa 水平仪器与 CAD ⁵之间的关联以及 FXa 在炎症、血管重塑和纤维化中的作用，我们对该结果的无效发现应谨慎解释。这种差异可能部分归因于用于诊断 CAD 的标准——尽管 Paraboschi 等人⁵考虑了经血管造影证实的病例和没有冠状动脉粥样硬化血管造影证据的对照，但我们目前的研究使用国际疾病分类和人口普查和调查办公室代码，以及用于病例确定的自我报告（表），非病例被视为对照。

总之，这项孟德尔随机化研究的结果与支持 FXa 抑制对降低静脉血栓栓塞风险和增加蛛网膜下腔出血风险的作用的临床试验结果一致。结果不支持 FXa 抑制与一般人群中 CAD 风险降低有关。尽管需要进一步研究来调查与先前工作的差异⁵ ，但这些信息可用于帮助确定未来临床试验的优先级。

非标准缩写和首字母缩略词

加元	冠状动脉疾病
外汇	因子 X
FXa	激活因子 X
他是	住院发作统计
软垫	外周动脉疾病

冠状动脉疾病

因子 X

激活因子 X

住院发作统计

外周动脉疾病

这项研究是使用 UK Biobank 资源（申请 29202）进行的。英国生物银行数据可在申请时获得 (http://www.ukbiobank.ac.uk/register-apply)。

Gill 博士由伦敦帝国理工学院的 Wellcome Trust 4i 计划 (203928/Z/16/Z) 和英国心脏基金会卓越研究中心 (RE/18/4/34215) 资助。Burgess 博士得到了由 Wellcome Trust 和皇家学会共同资助的 Henry Dale 爵士奖学金 (204623/Z/16/Z) 的支持。资金来源在本研究的设计、数据采集、分析、解释或撰写中没有任何作用。

Gill 博士在 Novo Nordisk 兼职。另一位作者报告没有冲突。

有关资金来源和披露，请参见第 552 页。