A 36-year-old male (Height: 170 cm, Body weight: 65 Kg, BMI: 22.5) who had cardiac arrest and was uncommunicative. Past medical history and family history were non-contributory. He had been smoking for 15 years and sometimes consumed alcohol a little. He was an office worker. Due to long working hours, he typically slept for only around three hours per night. One evening, after drinking alcohol with colleagues at a bar, he suddenly vomited and collapsed. When the ambulance arrived 6 mins later, he was unconscious (Glasgow Coma Scale was three) with VF. Despite 4 shocks and injections of adrenaline (3 mg total), he remained in VF state. He arrived at the hospital 32 mins after the emergency call.

Figure 1 shows the emergency room treatment. On admission to the emergency department, there was bloodstained fluid draining from the tracheal tube and oxygen saturation was undetectable. VF remained, but the pupillary light reflex was faintly present. To address the prolonged VF, we injected amiodarone (150 mg). During cannulation for VA-ECMO, we took chest X-ray which showed a white haze that may have been contusion or hemorrhage from chest compressions, or pulmonary edema from cardiac arrest. VA-ECMO was established at 54 mins after the emergency call, and blood tests revealed high lactate and severe acidosis. The likelihood of survival, let alone good neurological outcome, was considered very low. We then performed coronary angiography that demonstrated intact coronary arteries. After additional shocks, amiodarone and adrenaline,

spontaneous circulation returned, but the patient soon collapsed again, so we established intra-aortic balloon pumping (IABP). The volume out was 2000 ml of aspirated hemorrhage from the intubation tube and 300 ml of hemorrhage from the nasogastric tube. The volume of fluid infused was 11,000 ml plus, 1000 ml of red cell concentrate giving a total of >10,000 ml of fluid. Head computerized tomography (CT) showed clear gray-white matter differentiation, despite the prolonged cardiac arrest, and thoracic CT showed severe lung consolidation. The 12-lead electrocardiogram showed no problem, however there were many abnormal laboratory markers and the predictive values for neurological outcome were extremely poor.

Summary of emergency room treatment: The patient had cardiogenic shock, alveolar hemorrhage, severe respiratory disorder, hemorrhagic shock and severe coagulation disorder. Many factors predicted a poor neurological outcome, but we decided to treat aggressively with VA-ECMO and massive transfusion, based on the fact that the patient was young, newly married and had a faint pupillary reflex.

Figure 2 shows the progress after admission. We maintained a constant temperature between 36˚C and 37˚C as target temperature management (TTM) instead of mild therapeutic hypothermia for post cardiac arrest syndrome, because of bleeding tendency and prolonged cardiogenic shock. A massive infusion total of 28,000 ml in total was administered, including red cell concentrate and fresh frozen plasma to maintain circulation and coagulation. Adult respiratory distress syndrome occurred due to massive infusion. On the second day, the circulation had improved but respiratory function remained impaired, hence we changed the treatment VA- ECMO to VV-ECMO. We removed superabundant fluid using the extracorporeal ultrafiltration method (ECUM) of continuous hemodiafiltration, while monitoring lactate and Ht. The PaO2/FiO2 ratio (P/F ratio)

gradually improved with removal of fluid and the patient was completely weaned off the ventilator on the eighth day. Figure 3 shows the sequential chest X-rays. He commenced oral intake on the tenth day and hemodialysis ceased on the twenty-first day. He was ambulant and discharged on the twenty-sixth day.

He had been followed up for a year in the outpatient and found no complication. The patient returned to a social life equivalent to his preadmission level.

Summary of prognosis after admission: We maintained a constant temperature of between 36˚C and 37˚C as target temperature management (TTM) instead of mild therapeutic hypothermia, because of bleeding tendency and prolonged cardiogenic shock. We persevered through by switching V-A ECMO to V-V ECMO and improved oxygenation by aggressive fluid removal with ECUM at appropriate timing. We then performed renal replacement therapy for acute kidney injury (AKI).

Shinozaki, et al. reported that blood ammonia (NH₃) and lactate levels on hospital arrival predicted neurological outcome in patients with out-of-hospital cardiac arrest. The cut off value was 170 μg/dl for NH₃ and 12.0 mmol/l for Lactate [4] . Kasai, et al. also reported that the negative predictive value of NH₃ at levels of more than 192.5 μg/dl was 100% [5] . Although NH₃ was 251 μg/dl and Lactate was 15.8 mmol/l in our case, the neurological outcome was good. In our case, the presence of a pupillary light reflex may have predicted a favourable neurological outcome. When predicting the neurological outcome of a post cardiac arrest patient, we propose that the patient’s pupillary light reflex be taken into consideration in addition to NH₃ and lactate.

There have been few reports about switching VA-ECMO to VV-ECMO for respiratory disorder after ECPR [6] . In our case, there remained severe respiratory disorder from volume overload, lung contusion, alveolar hemorrhage and pulmonary atelectasis after recovering from severe shock, and we could improve respiratory function by switching from VA-ECMO to VV-ECMO. Since VA-ECMO increases after load by retrograde blood flow against the heart after recovering from severe shock, if severe respiratory disorder remains, it may be appropriate to switch to VV-ECMO.

There have been no reports about using ECUM to remove fluid from lung edema due to massive transfusion in acute phase, although there are some reports about ECUM for heart failure. We were able to remove fluid effectively by ECUM, monitoring the hematocrit and lactate levels to avoid excess. Removing fluid using diuretic drugs is often insufficient for volume overload due to massive transfusion in the acute phase. We were able to remove the intended amount of fluid by ECUM.