Secondary Logo

Journal Logo

Case Reports

ECMO Flow as a Sign of Intraabdominal Hemorrhage After Prolonged CPR

Ranney, David*; Hatch, Sarah; Bonadonna, Desiree; Daneshmand, Mani*

Author Information
doi: 10.1097/MAT.0000000000000927
  • Free


Cardiopulmonary resuscitation (CPR) remains as one of the core components of advanced cardiac life support (ACLS), providing a survival advantage to those sustaining cardiac arrest.1,2 Despite the benefits of CPR, intrathoracic and intraabdominal injuries remain frequent complications of external cardiac massage (ECM) and are often unrecognized in the immediate postresuscitation period.3,4 The most common and expected injuries include fractures of the ribs and sternum, reported as high as 78.9% and 30.2% in one series, respectively.3 Less frequent injuries include liver lacerations resulting in hemoperitoneum, which has been observed in 0.6–2.2% of survivors and nonsurvivors undergoing CPR.5 These injuries often present more occultly given both the low index of clinical suspicion and the act of attributing hypotension to concomitant cardiogenic shock.

In the setting of prolonged CPR and acute refractory cardiogenic shock, emergent extracorporeal membrane oxygenation (ECMO) has become an increasingly utilized modality for cardiopulmonary support.6 While adequate circuit flow rates can often be established quickly, the presence of intraabdominal solid organ lacerations and hemoperitoneum may result in inadequate or fluctuating flow rates. As such, careful consideration should be given to the patient and circuit flow dynamics, which may raise suspicion for life-threatening–associated injuries. Here, we present two cases of liver lacerations attributed to ECM that were initially suspected based on ECMO circuit flows, and subsequently managed successfully.

Case Reports

Patient No. 1

A 65 year old male presented to the emergency department with a diagnosis of ST-segment elevation myocardial infarction (STEMI). His medical history was significant for multiple MIs with placement of several drug-eluting stents (DES), bilateral orthotopic lung transplantation for interstitial lung disease, steroid-induced diabetes mellitus, heparin-induced thrombocytopenia, and stage III chronic kidney disease. He was taken to the cardiac catheterization laboratory (i.e., Cath Lab) for emergent percutaneous coronary intervention (PCI). Coronary angiography revealed thrombosis of his prior coronary stents, after which he had one DES placed in the left anterior descending artery (LAD) and two DES placed in the right coronary artery (RCA). The patient then sustained a pulseless electrical activity (PEA) cardiac arrest and CPR was initiated. Return of spontaneous circulation (ROSC) occurred after a total of 45 minutes of ECM; however, persistent refractory cardiogenic shock prompted cannulation by the ECMO team, which was achieved just minutes after ROSC. Cannulation was performed via the right femoral artery and vein for venoarterial (VA) ECMO support, and circuit flow was initiated at 3.6 lpm. A bivalirudin infusion was initiated for circuit anticoagulation and the patient was taken to the intensive care unit (ICU) for ongoing management.

Approximately 60 minutes had elapsed from time of CPR initiation to ECMO initiation. Over the next several hours, the patient remained hypotensive with low ECMO flows despite having received over 60 units of blood products and administration of multiple vasopressors. Retroperitoneal hemorrhage related to ECMO cannulation was initially suspected; however, subsequent aortoiliac and central venous angiography in the operating room demonstrated no evidence of major vascular injury or extravasation. Meanwhile, ECMO flows remained low and the patient’s abdomen was noted to be tense and distended. Decompressive laparotomy was performed due to concern for abdominal compartment syndrome; however, 2 L of hemoperitoneum was evacuated and there was no evidence of retroperitoneal hematoma. Instead, a grade I/II laceration was found on the sixth segment of the liver as well as several abrasions to the anterior liver surface and a small laceration to the inferior pole of the spleen. Five hours had elapsed since the initiation of ECMO at this stage of the procedure. The abdomen was packed, vacuum dressed, and the patient was brought to the ICU for resuscitation. Because of ongoing bleeding from the vacuum dressing despite correction of his coagulopathy, the patient returned to the operating room, where a combination of packing, electrocautery, Surgicel (Ethicon Inc., Somerville, NJ), and Floseal (Baxter International, Deerfield, IL) was used to address these injuries. His coagulopathy improved overnight with further resuscitation and administration of blood products. Extracorporeal membrane oxygenation flows gradually increased to adequate and expected ranges. The abdominal fascia was closed on postoperative day 3, and he was successfully decannulated from ECMO on postoperative day 5. Postoperative chest radiographs were only notable for stable bilateral pleural effusions without any evidence of rib fracture (Figure 1).

Figure 1.
Figure 1.:
A: Chest radiograph of patient no. 1 on arrival to hospital. B: Chest radiograph of patient no. 1 after return of spontaneous circulation, intubation, and exploratory laparotomy; note presence of intraabdominal packing and bilateral diffuse pulmonary edema without evidence of rib fracture.

His subsequent clinical course was complicated by acute kidney injury requiring dialysis, PEA arrest on postoperative day 30 in the setting of sepsis followed by ROSC, a related left pneumothorax and pleural effusion requiring pigtail catheter drainage, respiratory failure requiring temporary mechanical ventilation and tracheostomy, and persistent hypotension. He was neurologically intact following each of these events. He later succumbed to multiple hospital-acquired infections of the respiratory tract and bloodstream and was transitioned to comfort care. He later expired in-house after 4 months of hospitalization.

Patient No. 2

A 56 year old female presented to the emergency department by ambulance with sudden onset substernal chest pain radiating to her right shoulder. Her medical history was significant for systemic hypertension, type II diabetes, and hyperlipidemia. An electrocardiogram strip in the field revealed sinus rhythm with ST-segment elevation in the inferior leads with reciprocal changes in the anterior-septal leads, indicative of STEMI. Aspirin, heparin, and ticagrelor were administered on arrival and the patient was taken to the Cath Lab for emergent PCI. Coronary angiography revealed thrombus in both the left main (LM) coronary artery and LAD. Because of inadequate hemodynamics, an intraaortic balloon pump (IABP) was placed. Shortly thereafter, the patient sustained a cardiac arrest with ventricular fibrillation, and CPR was initiated. As the patient was not responding to ACLS measures, the ECMO team was activated. On arrival, the right femoral artery and vein were accessed and VA ECMO flow was initiated at 4.5 lpm, ending a total of 60 minutes of ECM. Cardiologists were able to complete PCI at this time with stenting of the LM and LAD.

After PCI, venous return to the circuit quickly diminished and hemoglobin was noted to be 5 g/dl, prompting immediate transfusion. Due to sporadic flows and suspicion for technical malfunction, the IABP was removed and replaced with an Impella via the left femoral artery in anticipation of rapidly transitioning from ECMO. Significant bleeding from the ECMO cannula site prompted consultation to vascular surgery and the patient was brought to the operating room for exploration. An injury to the profunda femoris was noted due to inadvertent profunda arterial cannulation. The cannula was relocated and the profunda was reimplanted into the superficial femoral artery. A subxiphoid pericardial window and chest tube was also placed for a large nonbloody pericardial effusion discovered on transesophageal echocardiography. Despite these interventions, hypovolemia persisted in the setting of low ECMO flows and new onset abdominal distention raised suspicion for an intraabdominal source of bleeding. The patient was taken to the operating room for exploratory laparotomy which revealed several liters of hemoperitoneum. Three and a half hours had now elapsed since time of ECMO initiation. A fracture of the lateral aspect of the left lobe of the liver was discovered and repaired with a combination of electrocautery, packing, clips, Surgical, and Tisseel (Baxter International). There was no evidence of retroperitoneal hematoma. The abdomen was closed and the patient was taken to the ICU for resuscitation and correction of coagulopathy. The patient was weaned and decannulated from ECMO with the Impella left in place. She was neurologically intact following each of these events and interventions.

Her subsequent postoperative course was complicated by respiratory failure requiring tracheostomy, which was later decannulated. She was discharged from the hospital 1 month after her admission. Of note, there was no evidence of rib fractures on any chest radiographs after her initial cardiac arrest and resuscitation.


In addition to skeletal fractures, intraabdominal visceral injury is known to occur as a complication of CPR.3 In particular, the close proximity of the liver to the ribs and sternum predispose this organ to blunt injury. Interestingly, neither of the patients in this series had evidence of rib or sternal fracture on subsequent chest radiographs, suggesting that liver injury had occurred solely from the blunt forces of ECM. This is likely further exacerbated by the prolonged duration of CPR and ECM in both patients. Incorrect technique of chest compressions may also contribute to intraabdominal organ injury. The American Heart Association (AHA) guidelines specifically designate the lower half of the sternum as the proper hand position for compressions,1 although provider fatigue and patient movement may result in migration to the epigastrium resulting in injury. As such, hand position should be as actively monitored by the code team leader as depth, rate, and quality of compressions.

The effects of coagulopathy and anticoagulation likely contributed to the massive hemoperitoneum in these patients as well. First, in preparation for PCI, both patients were loaded with antiplatelet agents before cardiac catheterization. Next, hemostasis is further impaired by the loading of anticoagulant before initiation of ECMO. These effects are then further exacerbated by the coagulopathy inherent to massive transfusion, hemorrhagic shock, and extracorporeal circulation.7,8 The results of these hemostatic insults are profound coagulopathy as evidenced by ongoing transfusion requirements and lack of intraoperative hemostasis despite multiple operative adjuncts and hemostatic biologic materials.

From the standpoint of ECMO cannulation, the most anticipated complications resulting in hemorrhage include major vascular injuries such as inferior vena cava injury9 or retroperitoneal bleeding/hematoma because of arterial injury.10,11. In the absence of abdominal distention or other indications for laparotomy, these injuries are often suspected first, which may delay recognition of intraperitoneal injuries as was the case in these two patients. As such, it is important to have a high index of suspicion for intraperitoneal visceral injuries in patients who undergo CPR, particularly for long durations, in addition to the more commonly encountered cannulation-related injuries described above. Including these injuries earlier in the differential diagnosis is likely to minimize delays in diagnosis and treatment, particularly when imaging studies are unobtainable in the setting of hemodynamic compromise and massive transfusion. Furthermore, the inability to establish adequate ECMO circuit flow rates or the presence of sporadic or fluctuating flow rates are excellent indicators of hypovolemia, as was observed in these two cases. As such, a diagnosis of hemorrhage should be considered early, and before the administration of large volumes of crystalloid or colloid solutions, which are likely to worsen an already developing coagulopathy. In essence, patients successfully resuscitated following prolonged CPR and ECM should be evaluated and treated in a manner similar to that of patients sustaining blunt thoracoabdominal trauma. These injuries can be life-threatening and are often masked by concomitant cardiogenic shock as well as the presence of VA ECMO cannulation and its more frequently associated complications. Providers should thus maintain a high index of suspicion for intraabdominal solid organ injury, particularly in the presence of abnormal ECMO circuit flow characteristics. Contingent upon patient stability, cross-sectional or ultrasonographic imaging studies may be obtained after cannulation to assess for intraperitoneal hemorrhage. Neither patient in this series underwent imaging because of ongoing massive transfusion requirements and high clinical suspicion for major vascular injury. However, both patients were brought to our hybrid operating room where both abdominal exploration and arteriogram could be performed. The availability of the hybrid operating room can be extraordinarily useful in these instances, as demonstrated by these cases.


Although life-saving, prolonged CPR results in traumatic injury in a large proportion of patients surviving cardiac arrest. When ECMO is utilized in this setting, careful attention to circuit flow and serial abdominal examination may reveal early signs of intraabdominal hemorrhage, particularly from the liver. These injuries should be considered earlier in diagnostic algorithms, as equal counterparts to the more common complications of retroperitoneal hemorrhage and primary vascular injuries.


1. Neumar RW, Shuster M, Callaway CW, et al. Part 1: Executive Summary: 2015 American Heart Association Guidelines Update for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation 2015.132(18 suppl 2): S315–S367.
2. Sandroni C, Nolan J, Cavallaro F, Antonelli M. In-hospital cardiac arrest: Incidence, prognosis and possible measures to improve survival. Intensive Care Med 2007.33: 237–245.
3. Beom JH, You JS, Kim MJ, et al. Investigation of complications secondary to chest compressions before and after the 2010 cardiopulmonary resuscitation guideline changes by using multi-detector computed tomography: A retrospective study. Scand J Trauma Resusc Emerg Med 2017.25: 8.
4. Han KS, Lee SW, Park KH, Park JS, Jung JS, Yu CW, et al. Concealed resuscitation-related injuries as reversible cause of recurrent arrest following extracorporeal cardiopulmonary resuscitation. CJEM 2017.19: 404–409.
5. Zahn G, Hauck M, Pearson DA, Green JM, Heffner AC. Major hemorrhage from hepatic laceration after cardiopulmonary resuscitation. Am J Emerg Med 2015.33: 991.e3–991.e4.
6. Abrams D, Combes A, Brodie D. What’s new in extracorporeal membrane oxygenation for cardiac failure and cardiac arrest in adults? Intensive Care Med 2014.40: 609–612.
7. Simmons JW, Powell MF. Acute traumatic coagulopathy: Pathophysiology and resuscitation. Br J Anaesth 2016.117(suppl 3): iii31–iii43.
8. McManus ML, Kevy SV, Bower LK, Hickey PR. Coagulation factor deficiencies during initiation of extracorporeal membrane oxygenation. J Pediatr 1995.126: 900–904.
9. Bhaskar B, Mullany D, Parmar D, Ziengenfuss M, Shekar K. Successful conservative management of an iatrogenic ECMO cannula–related inferior vena cava injury. Anaesth Intensive Care 2015.43: 418–419.
10. Rupprecht L, Lunz D, Philipp A, Lubnow M, Schmid C. Pitfalls in percutaneous ECMO cannulation. Heart Lung Vessel 2015.7: 320–326.
11. Makdisi G, Wang IW. Extra corporeal membrane oxygenation (ECMO) review of a lifesaving technology. J Thorac Dis 2015.7: E166–E176.

ECMO; CPR; liver laceration; intraabdominal hemorrhage

Copyright © 2018 by the ASAIO