The patient was a 54-year-old man who was scheduled to undergo open subtotal gastrectomy for gastric cancer. He had no underlying diseases, except dyslipidemia, and no history of allergic responses to drugs. The patient was transferred to the operating room without any abnormal signs or symptoms. Midazolam (2 mg), lidocaine (20 mg), propofol (60 mg), and cisatracurium (8 mg) were administered intravenously for induction of general anesthesia. After endotracheal intubation but before skin incision, 1 g cefotetan was administered intravenously as a prophylactic antibiotic. A few minutes later, the patient’s heart rate suddenly dropped to less than 30 beats/min, according to electrocardiography monitoring, and he went into cardiac arrest. CPR was initiated immediately, and transesophageal echocardiography performed during CPR showed akinesia and low ejection fraction in the left ventricle. Because normal cardiac contractility was not recovered during CPR, veno-arterial ECMO was performed. The catheters for arterial inflow (catheter size: 15 Fr) and venous outflow (catheter size, 21 Fr) for ECMO were inserted through the right femoral artery and the left femoral vein, respectively. CPR including chest compressions was performed for 35 minutes. After return of spontaneous circulation, the patient’s blood pressure was 110/60 mmHg, his pulse rate was 70 beats/min, and the level of lactic acid reached 12.1 mmol/L. The patient was transferred to the intensive care unit (ICU) without performing the planned operation. His initial vital signs on admission to the ICU were 114/64 mmHg, 110 beats/min, and a lactic acid level of 9.7 mmol/L. Initial chest radiography in the ICU showed diffuse pulmonary edema and subsegmental atelectasis (Figure 1). Transthoracic echocardiography (TTE) performed on the day of ICU admission confirmed normal left and right ventricle contractility and 59% ejection fraction. Repositioning of the ECMO catheter was considered to correct the malposition; however, ECMO was stopped after 9 hours due to confirmation of normal cardiac function on TTE. Mechanical ventilation was maintained for 18 days. Owing to persistent loss of renal function, renal replacement therapy was continued after recovery. Twenty days after his cardiac arrest, abdominal computed tomography (CT) was performed for re-evaluation and therapeutic planning of the remaining malignant gastric mass. The examination revealed left fourth and fifth rib fractures as well as an aortic dissection extending from the proximal descending thoracic aorta to the abdominal aorta; these findings were not observed on previous evaluations and were detected incidentally. The DeBakey type III aortic dissection extended from the distal arch of the thoracic aorta to the proximal level of the renal artery involving the celiac trunk (Figure 2). Twenty-five days after CPR, the patient complained of atypical chest pain in the epigastric area, and angiographic CT and TTE were performed to evaluate the extent of the aortic dissection and associated complications compared with the previous abdominal CT scan. Because there were no interval changes, it was considered an uncomplicated type B aortic dissection with no sign of malperfusion of the major vessels. Medical treatment including antihypertensive drugs was continued. One month later, the patient was diagnosed with an anaphylactic reaction to cefotetan based on a positive response to a skin prick test.

CPR is a critical procedure for saving patients experiencing cardiac arrest by maintaining auxiliary cardiac circulation and ventilation until return of spontaneous cardiac circulation. In 1960, Kouwenhoven et al. [1] developed a method of external cardiac massage without thoracotomy for patients in cardiac arrest. Approximately 200,000 cases of in-hospital cardiac arrest occur each year in the United States[2]. The rate of survival to discharge is an estimated 15-20% [3,4]. The survival rate after in-hospital cardiac arrest has improved steadily owing to quality improvement efforts [4]. In addition, several studies have suggested that ECMO during CPR has improved the survival rates of in-hospital cardiac arrest [5-7]. Along with increased survival after CPR, there has been a corresponding increase in concern about CPR-related complications in survivors of cardiac arrest.

Resuscitation measures include invasive manipulations such as external cardiac compression, endotracheal intubation, and electric impulses for defibrillation. These manipulations support recovery of spontaneous circulation and termination of CPR, but patients may also experience traumatic injuries due to mechanical chest compression. In a study of approximately 1,000 autopsy cases, resuscitation-related injuries were reported in an estimated 21% to 65% of cases [8]. The most common complications after CPR are thoracic skeletal injuries, particularly fractures of the ribs (13% to 97%) and sternum (1% to 43%) [9]. Fractured ribs can cause mediastinal hematoma, hemothorax, pneumothorax, and pulmonary contusion [10]. Direct cardiac laceration and great vessel injuries are relatively rare but can lead to death [8].

Aortic dissection after CPR has rarely been reported in survivors [11-13]. Most blunt aortic injuries occur in the proximal descending aorta, where the relatively mobile aortic arch joins the fixed descending aorta via the ligamentum arteriosum. The shearing forces of sudden deceleration that occur in motor vehicle collisions can cause traumatic aortic dissection. Other mechanisms for this injury include sudden elevated intraluminal aortic pressure due to compression between the sternum and thoracic vertebrae [14]. The symptoms, clinical signs, and physical examination findings are nonspecific and difficult to differentiate from aortic dissection after CPR. A majority of CPR survivors cannot complain of symptoms because of decreased mentality due to hypoxic brain damage or a sedated state for post-CPR management, including mechanical ventilation. In these cases, the first clinical sign may be the onset of hemodynamic instability.

Chest radiography is typically used for screening, but it has a low sensitivity for diagnosing traumatic aortic injuries. Aortic injury after CPR may be suspected when chest radiography findings suggest mediastinal widening, loss of the aortopulmonary window, rightward tracheal/nasogastric tube deviation, left main stem bronchus depression, and left apical cap [15]. Imaging studies useful for diagnosis include CT and transesophageal echocardiography (TEE). Chest CT has high diagnostic sensitivity, exceeding 97% to 100% for aortic dissection [16]. Chest CT can also be used to evaluate other associated damage, such as thoracic skeletal injuries and lung parenchymal injuries. TEE can be performed at bedside in patients with hemodynamic instability to evaluate the myocardium for wall motion abnormalities. However, TEE has limitations such as dependency on operator skill and poor visualization of the distal ascending aorta and proximal arch [17].

Aortic dissection after CPR is rare but can be fatal. In the absence of previous evaluation, it can also be difficult to confirm a causal relationship between CPR-induced injury and the underlying disease that caused the sudden cardiac arrest. The 2010 American Heart Association (AHA) guidelines recommend harder (a depth of at least 5 cm) and faster compressions (a rate of at least 100 compressions per minute) than the 2005 guidelines [18]. Increasing the requirements of compression depth and rate may increase the risk of CPR-related complications, including life-threatening events. Still, to prevent complications, chest compressions should be performed correctly and in the accurate position rather than reducing compressive efforts [19]. With increasing rates of survival with CPR, potential traumatic injuries caused by chest compression should be considered. Accurate diagnosis and proper treatment of unavoidable CPR-related complications are important to decrease traumatic damages.