Current guidelines

Current guidelines for AIE stated that cardiac surgery should be considered without any delay in patients with heart failure, uncontrolled infection, annular abscess, or persistent high-risk of systemic embolism, except for patients with coma, brain hernia, or large major central lesions^[2-5][Table 1]. However, postponing the cardiac surgery for AIE with ICH for four weeks is still recommended. As for a relationship of neurological outcomes and the size of the brain lesions of stroke or ICH secondary to AIE, no study has ever been published.

Early studies

There have been many earlier studies regarding treating AIE patients with cerebral complications. In 1995, Eishi et al. analyzed 34 patients with ICH and reported that the risk of neurological deterioration is still significantly high 15 days and even four weeks after cerebral hemorrhage^[6]. In 1996, Gillinov et al. reported good outcomes in 7 patients with AIE and ICH, in whom the surgery was postponed for 2 to 3 weeks^[7]. In 2014, Wilbring et al. reported that a neurological deterioration was found in 33% of 6 patients with AIE and ICH who underwent surgery after 17 ± 24 days of diagnosis^[8]. In 2006, Ruttmann et al. reported a mortality of 66% after an early surgery of patients with ICH^[9]. In 2013, Garcia-Cabrera et al. reported that the mortality of patients with AIE and ICH who had surgery within 14 days was 75%, and that neurological exacerbation occurred in 50%^[10]. In 2016, A multi-center registry data analyzing the AIE patients complicated with stroke or ICH by Okita et al. demonstrated that patients with AIE and ICH who underwent open cardiac surgery had higher mortality when the surgery was performed within seven days after AIE and ICH onset than those who underwent surgery 8-21 days after the onset^[11]. These studies indicate that the mortality rate decreases when cardiac surgery for patients with ICH is delayed.

Recent studies

Conversely, recent studies by Shang et al. reported that cardiac surgery on 16 patients with AIE and ICH within four days after admission whose ICH size was smaller than 1-2 cm was safely performed^[12]. Yoshioka et al. analyzed 30 patients with AIE and preoperative ICH who underwent surgery^[13]. Irrespective of the timing of surgery, none of these 30 patients had neurological deterioration or died from neurologic complications. They also reported that postoperative neuroimaging did not demonstrate exacerbation of hemorrhagic lesions, even in 11 patients who underwent surgery within two weeks of ICH diagnosis. Kume et al. compared the postoperative outcomes of 25 patients with preoperative ICH with patients without ICH and they found that there was no significant difference in mortality or incidence of ICH worsening^[14]. They also found that there was no significant difference in ICH exacerbation between patients who were operated on within 14 days of ICH occurrence and those operated on after 14 days of ICH onset. A meta-analysis by Tam et al., including 27 observational studies, demonstrated no significant difference in mortality after cardiac surgery regarding the timing of surgery^[15]. These studies are summarized in Table 2.

Insights from neurology

ICH is classified into primary intraparenchymal hemorrhage [Figures 1 and 2], hemorrhagic transformation of ischemic infarcts [Figure 3], and rupture of intracranial infectious aneurysms [Figure 4]^[16]. Primary intraparenchymal hemorrhage is due to the septic erosion of the arterial wall with rupture. For non-IE primary intraparenchymal hemorrhage, several scoring systems were developed for predicting long-term outcomes. Many factors, such as lesion size, lesion location, neurologic function at presentation, age, and pre-morbid functional status may contribute to patients’ outcomes. An ICH with a volume of 30 mL, measured by computerized tomography (CT) scan, and intraventricular/infratentorial ICH tend to be associated with a worse prognosis^[17,18]. The hemorrhagic transformation of ischemic infarcts can be attributed to dysfunctions in the blood-brain barrier and the degradation of the extracellular matrix of the brain^[19]. While mycotic aneurysm is not included in the category of ICH, these lesions are common in AIE patients. They are considered to be a potential cause of subarachnoid hemorrhage [Figure 5]. In patients with cerebral aneurysms, neuro-surgical or endovascular intervention prior to cardiac surgery should be strongly considered^[20,21]. A recent study by Salaun et al. showed that patients with an undetermined etiology of ICH had significantly higher mortality compared to patients with ICH secondary to ruptured mycotic aneurysm or ischemic stroke^[22].

Figure 1. 76-year-old male with mitral prosthetic valve endocarditis. Parenchymal hemorrhage.

Figure 2. 68-year-old female with aortic valve infective endocarditis. Intraventricular hemorrhage.

Figure 3. 66-year-old male with aortic valve infective endocarditis. Hemorrhagic transformation of stroke after surgery.

Figure 4. 74-year-old male with aortic valve infective endocarditis. Cerebral mycotic aneurysm ruptured.

Figure 5. 37-year-old female with mitral infective endocarditis. Subarachnoid hemorrhage.

Innovations in cardiopulmonary bypass (CPB) strategy

During a conventional CPB using a normal conventional dose of heparin, the coagulo-fibrinolytic system is greatly activated^[23]. Modifying the CPB strategy might be necessary to improve the surgical outcomes for AIE with ICH. Some authors have reported that using low-dose heparin (100 IU/kg) during cardiac surgery resulted in reduced blood loss and transfusion requirements^[24,25]. We reported the advantages of open cardiac surgery using nafamostat mesylate (NM; 6-amino-2 naphthalene-p-guanidinobenzoate dimethanesulfonate) as an alternative anticoagulant during CPB (NM-CPB) to reduce ICH exacerbation in AIE patients with stroke or ICH^[26,27]. NM has a strong inhibitory activity on various proteases in the coagulo-fibrinolytic system. Although its exact mechanism has not been elucidated, NM has a very short half-time and has potential inhibitory effects on the coagulation, fibrinolysis, and platelet aggregation system. Our anticoagulation protocol during NM-CPB has been that NM (1.0 mg/kg) was administered intravenously to assess for any allergic reaction after induction of anesthesia. An initial dose of heparin (50 IU/kg; 25% of normal dosage) was administrated to achieve the activated clotting time of over 400 s. After initiating the CPB, the NM was infused continuously through the venous line of the pump circuit at 1.8 mg/kg/h. The activated clotting time was maintained at 350-500 s by a fine titration of the NM dose, which was added at 0.8 mg/kg/h to the cardiotomy reservoir and at 1.0 mg /kg/h in the pump circuit. The activated clotting time was measured every 15 min. The strategy and methods of NM-CPB are shown in [Figure 6]. Sakamoto et al. reported that the timing of cardiac surgery in patients with AIE was 2.1 ± 1.2 days after the onset of AIE and the mortality was 33.3% with no worsening hemorrhages^[26]. Ota et al. reported a similar outcome when the timing of cardiac surgery was within 2.1 days from the onset of AIE^[27]. In both studies, the mortality was not related to the timing of cardiac surgery. Recently, in our study dealing with four patients with AIE and ICH^[28], the timing of cardiac surgery was 6.7 days after the onset of AIE and ICH, and fortunately the mortality was zero. NM can be a useful option as an anticoagulant in critical situations of cardiac surgery with bleeding tendencies involved by the CPB.

Figure 6. Methods of cardiopulmonary bypass using nafamostat mesylate (NM-CPB). Shown in the image is the pump circuit of NM-CPB. Nafamostat mesylate was infused continuously to the cardiotomy reservoir and venous reservoir at 0.8 mg/kg/h and 1.0 mg/kg/h, respectively. The rate of injection was changed according to the activated clotting time (ACT). Heparin: 50 IU/kg (25% of normal), ACT: 350-450 q 15 min, Mild hypothermia: (33 °C), Pump flow: 2.4-2.6 L/m²/min, Systemic pressure: 50 mmHg.

We showed the relationships between the size of infarction or hemorrhage and the timing of cardiac surgery for patients who underwent NM-CPB. Regarding acute stroke [Figure 7], in two patients who had normal CPB using the conventional dose of heparin, a hemorrhagic transformation of the brain lesions occurred postoperatively. However, another patient who had a stroke over 3000 mm² in size, had a successful early cardiac surgery without any exacerbation of stroke. As for ICH [Figure 8], there were four hospital death related to brain complications in two with NM-CPB and two with normal heparin CPB. The hospital mortality in patients with NM-CPB, and patients with normal heparin CPB were 15.3% and 28.5%, respectively. Among patients who underwent surgeries within seven days from the onset of the AIE, mortality in patients with NM-CPB and with normal heparin CPB were 16.6% (2/12), and 33.3% (1/3), respectively. This might indicate that an early surgery for ICH could be performed more safely by NM-CPB.

Figure 7. Relationships between the size of acute cerebral infarction and the interval from onset to operation (total n = 4 6, NM-CPB n = 32, normal CPB n = 14). There were two patients of exacerbation, where cardiac surgery was performed with cardiopulmonary bypass using a normal dose of heparin. Case 1: Multiple infarction and subdural hemorrhage with an area of 41 mm² turned to parenchymal hemorrhage from preoperative infarct lesion postoperatively. Case 2: Hemorrhagic infarction with an area of 633 mm² turned to ectopic subdural hematoma postoperatively. NM-CPB: Cardiopulmonary bypass using nafamostat mesylate; normal CPB: of cardiopulmonary bypass using heparin.

Figure 8. Relationships between the size of intracranial hemorrhage and the interval from onset to operation (total n = 20, NM-CPB n = 13, normal CPB n = 7). There were four hospital deaths related to brain complications, of which two underwent NM-CPB, while the other two underwent CPB using normal heparin. An early surgery within seven days from the onset was performed in 12 patients of NM-CPB and three patients of normal heparin CPB. The average size of intracranial hemorrhage was 723 ± 1069 mm. NM-CPB, cardiopulmonary bypass using nafamostat mesylate; normal CPB, of cardiopulmonary bypass using heparin.