A 78-year-old woman with no personal or family history of abnormal bleeding noticed a bruise on her thigh 2 weeks after her second COVID-19 (SARS-CoV-2) mRNA vaccination (Pfizer-BioNTech; day 1; not shown). Otherwise, she had no serious adverse events following vaccination. Her skin bruise persisted. On day 38, a new skin bruise appeared on her left hand. On day 50, she visited a hospital because of pain and swelling in the left hand (Figure 1A). She had no evidence of any malignancies or autoimmune disorders and no medication intake. Her SARS-CoV-2 antigen was negative. The hematoma was removed to treat her compartment syndrome of the left hand. However, she had persistent postoperative bleeding and red blood cell units were transfused (2 units/day for 3 days). On day 55, she was transferred to our hematology department.

Her physical examination revealed ecchymosis on her hind upper arms and left knee (not shown). Computed tomography (CT) scans showed intra-articular bleeding in her left shoulder (Figure 1B) and asymptomatic central nervous system (CNS) bleeding (Figure 1C). Accordingly, the patient received 2 units of fresh frozen plasma for 4 days with no apparent effect.

Blood tests revealed normocytic anemia (red cell count 2.40 × 10⁶/μL, reference range 3.86–4.92 × 10⁶/μL; hemoglobin 7.8 g/dL, 11.6–14.8 g/dL; hematocrit 22.8%, 35.1%–44.4%) with normal platelet count (159 × 10³/μL, 158–348 × 10³/μL), prothrombin time (10.9 s, 9.5–13.5 s), and activated partial thromboplastin time (25.9 s, 25.0–38.0 s). Fibrinogen level was 288 mg/dL (155–415 mg/dL). Factor VIII/8 (F8) activity was >201%, (62%–145%), von Willebrand factor (VWF) activity 241% (50%–150%), and its antigen level was >201% (50%–150%). F8 inhibitor was negative. Hepatitis B surface antigen, Hepatitis C antibody, and HIV-1 and -2 antibodies were negative or nonreactive. No monoclonal protein was detected by initial screening.

While factor XIII/13 (F13) antigen level was slightly reduced (59%, reference range >70%), its activity was below the detection limit (<3%, 70%–140%; measured by an ammonia release assay using a Berichrom FXIII Kit (Sysmex Corporation, Kobe, Japan) at a commercial laboratory service (SRL Ltd., Hachioji, Japan). Administration of F13 concentrates (1200 units/day for 5 days) significantly improved her symptoms, and she was discharged home on day 74. In an experimental work-up conducted by the Japanese collaborative research group for autoimmune coagulation factor deficiencies (AiCFD), a 5-step dilution mixing test with a healthy control plasma showed an F13 inhibitor pattern (Figure 1E) and anti-F13-A subunit autoantibodies were detected by immunochromatography and enzyme-liked immunosorbent assay (Figure 1G,H, respectively). Therefore, we made a definite diagnosis of autoimmune F13 deficiency (AiF13D), according to the ISTH/SSC criterion 2015.¹ Immediately, prednisolone (0.5 mg/kg daily) was administered orally on day 91, but she noticed a new bruise on her right hand on day 100. Since F13 activity was still low (14%), she was re-hospitalized on day 103. Screening CT did not detect any internal bleeding. Prednisolone was increased to 1 mg/kg, and intravenous (IV) immunoglobulin (IVIg, 400 mg/kg for 5 days) was administered. However, she was found lying on the floor in the early morning on day 109. She had a mild disturbance of consciousness but did not complain of headache. She had no elevated blood pressure, vomiting, paralysis, or other local neurological deficits. She was given 1200 units of F13 concentrates and her F13 activity increased to 33%; however, she died of CNS bleeding (i.e., cerebral hemorrhage and subarachnoid hemorrhage) in the afternoon on the same day after approximately 10 h (Figure 1D). Since the patient's family refused to autopsy, the exact cause and/or trigger of CNS bleedings remained unknown.

Various autoimmune diseases have been reported to develop after SARS-CoV-2 mRNA vaccination. To date, seven cases of autoimmune F8 deficiency (AiF8D; or acquired hemophilia A), which is the most frequent AiCFD, have been reported.^2-6 On the contrary, AiF13D is an extremely rare AiCFD, and this is the world's first case of AiF13D that is likely to have been triggered by SARS-CoV-2 mRNA vaccination. The exact mechanism(s) responsible for these immuno-hematological complications is unknown, but the onset time of our case is similar to other immune-related adverse events such as nephritis, vasculitis, etc. following vaccination. Although the temporal relationship of the development of AiF8D following SARS-CoV-2 mRNA vaccination may suggest a possible etiologic contribution,⁶ it is difficult² and challenging⁵ to establish the causal relationship between the vaccine and AiF8D, and the emergence of F8 inhibitors post vaccination is most likely to be a coincidence,^{2, 3, 5} as the authors of previous reports have acknowledged. Nevertheless, it should be noted that the world's first case with AiF8D recurrence after SARS-CoV-2 infection has also been reported, suggesting a close link between SARS-CoV-2 virus and the hemostatic mechanism.⁷

The age of our AiF13D patient (78 years) and the onset time from vaccination was similar to those of the seven AiF8D cases (67–86 years; 14 days vs. 4–21 days after the first or second dose, respectively). However, her bleeding symptoms, that is, CNS hemorrhages, were much more severe than all the AiF8D cases except for one patient with hemothorax caused by an accidental fall with chest and shoulder contusion.³ There were no clear reasons or triggers for neither the first CNS bleeding episode on day 55 nor the second episode on day 109 in our patient. It is well known that cases of congenital F13 deficiency frequently present with CNS bleeding, which is the leading cause of death. Therefore, the attending physician of any AiF13D patient should pay close attention to the development of CNS bleeding.

Like our AiF13D patient, an AiF8D patient also died of active arterial bleeding due to acute gall bladder rupture.³ This may be because he had denied any further intervention following arterial coiling, which only transiently stopped the bleeding. Although two AiF8D patients had cessation of bleeding without using any hemostatic agents, most patients achieved hemostasis by administration of the so-called bypassing medicines such as recombinant activated factor VII/7, activated prothrombin complex concentrates, and F8 inhibitor bypassing activity. On the contrary, since there is no bypassing drug in F13 replacement therapy, our AiF13D patient continued to receive plasma-derived F13 concentrates. Although the recombinant F13-A subunit preparation has been on the market for a long time, it cannot be used in Japan (or even worldwide) because it is not covered by public medical insurance, which is an important problem to be solved in the future.

Six out of seven AiF8D cases recovered with corticosteroids alone or in combination with rituximab or cyclophosphamide. One AiF8D patient was given IVIg, which was considered effective for autoimmune VWF deficiency (or autoimmune VW disease), probably due to its simultaneous relapse.⁴ In this case, a large hemorrhage plaque occurred as early as the fourth day of the second inoculation, and her hemostasis was obtained by F8/VWF replacement therapy. Our AiF13D patient was less responsive to IVIg infusions and eventually developed fatal bleeding, that is, CNS hemorrhage. As mentioned above,⁴ recrudescence of autoimmune hematologic disorders is not uncommon, and the attending physician should closely follow-up on such patients.

About 10 years ago, we formed our Japanese collaborative research group on AiCFD and have been conducting a nationwide survey to make a definitive diagnosis of patients with unexplained bleeding disorders in Japan. There are many patients with thrombosis related to SARS-CoV-2 infection, which has been a worldwide pandemic for the past 2 years; however, very few patients have bleeding. In particular, AiCFD after SARS-CoV-2 vaccination is an extremely rare disease, and it is difficult to diagnose and treat; therefore, clinicians may need to raise awareness of this disease.

In conclusion, AiF13D should be differentiated from bleeding due to other causes after SARS-CoV-2 mRNA vaccination. This is because the disease requires special laboratory tests, such as the ammonia release assay and amine incorporation assay, and specific hemostatic therapy by employing F13 concentrates along with immunosuppressants.

This study was approved by the institutional review boards of Yamagata University of Medicine and Oji General Hospital. All procedures were conducted in accordance with the Declaration of Helsinki.

一名没有异常出血个人或家族史的 78 岁女性在第二次 COVID-19 (SARS-CoV-2) mRNA 疫苗接种后 2 周注意到她的大腿上有瘀伤（辉瑞-BioNTech；第 1 天；未显示） . 否则，她在接种疫苗后没有出现严重的不良事件。她的皮肤瘀伤持续存在。第 38 天，她的左手出现了新的皮肤瘀伤。第 50 天，她因左手疼痛和肿胀到医院就诊（图 1A）。她没有任何恶性肿瘤或自身免疫性疾病的证据，也没有服用任何药物。她的 SARS-CoV-2 抗原呈阴性。去除血肿以治疗她的左手筋膜室综合征。然而，她有持续的术后出血，并且输注了红细胞单位（2 单位/天，共 3 天）。第 55 天，她被转到我们的血液科。

她的体格检查显示她的后上臂和左膝有瘀斑（未显示）。计算机断层扫描 (CT) 扫描显示她的左肩关节内出血（图 1B）和无症状的中枢神经系统 (CNS) 出血（图 1C）。因此，该患者连续 4 天接受 2 个单位的新鲜冰冻血浆，但没有明显效果。

血液检查显示正细胞性贫血（红细胞计数 2.40 × 10 ⁶ /μL，参考范围 3.86–4.92 × 10 ⁶ /μL；血红蛋白 7.8 g/dL，11.6–14.8 g/dL；血细胞比容 22.8%，35.1%–44.4%）血小板计数正常（159 × 10 ³ /μL，158-348 × 10 ³ /μL），凝血酶原时间（10.9 s，9.5-13.5 s）和活化部分促凝血酶原激酶时间（25.9 s，25.0-38.0 s）。纤维蛋白原水平为 288 mg/dL (155–415 mg/dL)。VIII/8因子（F8）活性>201%，（62%–145%），血管性血友病因子（VWF）活性241%（50%–150%），其抗原水平>201%（50%– 150%）。F8 抑制剂为阴性。乙型肝炎表面抗原、丙型肝炎抗体和 HIV-1 和 -2 抗体均为阴性或无反应性。初步筛选未检测到单克隆蛋白。

虽然因子 XIII/13 (F13) 抗原水平略有降低（59%，参考范围 >70%），但其活性低于检测限（<3%，70%–140%；通过氨释放测定使用Berichrom FXIII 试剂盒（Sysmex Corporation，Kobe，Japan）在商业实验室服务（SRL Ltd.，Hachioji，Japan）。给予 F13 浓缩物（1200 单位/天，持续 5 天）显着改善了她的症状，她于第 74 天。在日本合作研究小组针对自身免疫性凝血因子缺乏症 (AiCFD) 进行的一项实验性工作中，使用健康对照血浆进行的 5 步稀释混合试验显示 F13 抑制剂模式（图 1E）和抗 F13 -通过免疫层析和酶样免疫吸附试验检测到亚单位自身抗体（图1G，H，分别）。因此，根据 2015 年 ISTH/SSC 标准，我们明确诊断为自身免疫性 F13 缺乏症 (AiF13D)。¹在第 91 天立即口服泼尼松龙（每天 0.5 mg/kg），但在第 100 天她注意到右手上有新的瘀伤。由于 F13 活性仍然很低（14%），她在第 103 天再次住院. 筛查 CT 未检测到任何内出血。泼尼松龙增加至 1 mg/kg，并给予静脉内 (IV) 免疫球蛋白（IVIg，400 mg/kg，持续 5 天）。然而，在第 109 天的清晨，人们发现她躺在地板上。她有轻微的意识障碍，但没有抱怨头痛。她没有血压升高、呕吐、瘫痪或其他局部神经功能缺损。给予她 1200 单位的 F13 浓缩物，她的 F13 活性增加到 33%；然而，她死于中枢神经系统出血（即，脑出血和蛛网膜下腔出血）在大约 10 小时后的同一天下午（图 1D）。由于患者家属拒绝尸检，中枢神经系统出血的确切原因和/或触发因素仍然未知。

据报道，在接种 SARS-CoV-2 mRNA 后会出现各种自身免疫性疾病。迄今为止，已报告了 7 例自身免疫性 F8 缺乏症（AiF8D；或获得性血友病 A），这是最常见的 AiCFD。^2-6相反，AiF13D 是一种极为罕见的 AiCFD，这是世界上第一例可能由 SARS-CoV-2 mRNA 疫苗接种引发的 AiF13D 病例。导致这些免疫血液学并发症的确切机制尚不清楚，但我们病例的发病时间与疫苗接种后的其他免疫相关不良事件如肾炎、血管炎等相似。尽管 SARS-CoV-2 mRNA 疫苗接种后 AiF8D 发展的时间关系可能表明可能的病因学贡献，⁶很难²和具有挑战性⁵确定疫苗和 AiF8D 之间的因果关系，而疫苗接种后 F8 抑制剂的出现很可能是巧合，2、3、5^正如之前报告的作者所承认的那样。尽管如此，应该指出的是，世界上首例感染 SARS-CoV-2 后 AiF8D 复发的病例也有报道，这表明 SARS-CoV-2 病毒与止血机制之间存在密切联系。⁷

我们的 AiF13D 患者的年龄（78 岁）和接种疫苗的发病时间与 7 例 AiF8D 病例的相似（67-86 岁；分别在第一剂或第二剂后 14 天和 4-21 天）。然而，她的出血症状，即中枢神经系统出血，比所有 AiF8D 病例都严重得多，除了一名患者因胸部和肩部挫伤意外跌倒导致血胸。³在我们的患者中，第 55 天的第一次 CNS 出血发作和第 109 天的第二次发作都没有明确的原因或触发因素。众所周知，先天性 F13 缺乏的病例经常出现中枢神经系统出血，这是导致死亡的主要原因。因此，任何 AiF13D 患者的主治医师都应密切关注中枢神经系统出血的发展。

与我们的 AiF13D 患者一样，一名 AiF8D 患者也因急性胆囊破裂而死于活动性动脉出血。³这可能是因为他拒绝在动脉盘绕后进行任何进一步的干预，这只是暂时止血。尽管两名 AiF8D 患者在没有使用任何止血剂的情况下停止了出血，但大多数患者通过施用所谓的旁路药物（如重组活化因子 VII/7、活化凝血酶原复合物浓缩物和 F8 抑制剂旁路活性）实现了止血。相反，由于在 F13 替代疗法中没有绕过药物，我们的 AiF13D 患者继续接受血浆衍生的 F13 浓缩物。虽然重组F13-A亚基制剂上市已久，但由于不在公共医疗保险范围内，无法在日本（甚至世界范围内）使用，这是未来亟待解决的重要问题.

7 个 AiF8D 病例中有 6 个单独使用皮质类固醇或与利妥昔单抗或环磷酰胺联合治疗。一名 AiF8D 患者接受了 IVIg，这被认为对自身免疫性 VWF 缺乏症（或自身免疫性 VW 疾病）有效，可能是由于其同时复发。⁴本例中，早在第二次接种的第四天就出现大面积出血斑块，通过F8/VWF替代疗法止血。我们的 AiF13D 患者对 IVIg 输注反应较差，最终出现致命性出血，即 CNS 出血。如前所述，⁴自身免疫性血液病复发并不少见，主治医师应密切随访此类患者。

大约 10 年前，我们成立了日本 AiCFD 合作研究小组，并一直在进行一项全国性调查，以对日本不明原因出血性疾病患者做出明确诊断。SARS-CoV-2感染相关血栓患者较多，近2年全球大流行；然而，极少数患者有出血。尤其是SARS-CoV-2疫苗接种后的AiCFD是一种极为罕见的疾病，诊断和治疗难度较大；因此，临床医生可能需要提高对这种疾病的认识。

总之，AiF13D 应与 SARS-CoV-2 mRNA 疫苗接种后其他原因引起的出血区分开来。这是因为该疾病需要特殊的实验室测试，例如氨释放测定和胺掺入测定，以及通过使用 F13 浓缩物和免疫抑制剂进行的特定止血治疗。

这项研究得到了山形医科大学和王子综合医院的机构审查委员会的批准。所有程序均按照赫尔辛基宣言进行。