Management of Hypothermic Cardiac Arrest with Hemoperitoneum from LUCAS Device: A Case Report

Introduction

Hypothermia is defined as a core temperature less than 35C, at which the body’s thermoregulatory abilities begin to fail [1]. Severe hypothermia with core temperature below 28C can lead to cardiac arrest through slowing of pacemaker cell depolarization and cardiomyocyte conduction [2,3]. Witnessed hypothermic cardiac arrest (HCA) is a rare phenomenon, with a 2019 literature review citing 214 published cases [4]. The survival rate at hospital discharge for witnessed HCA is estimated at 50–73% – compared with 27–35% in unwitnessed HCA – with around 90% experiencing full neurologic recovery [1,5,6].

Extracorporeal rewarming is central to successful treatment. We present a case of severe hypothermia with witnessed pre-hospital cardiac arrest successfully treated by cardiopulmonary bypass (CPB). The clinical course was complicated by hemoperitoneum caused by the device-assisted cardiopulmonary resuscitation (CPR), warranting exploratory laparotomy. Liver injury is a rare complication of CPR, occurring in an estimated 0.6% of cases [7]. We present the first published case of successful surgical management of iatrogenic hemoperitoneum in the context of HCA.

Case Report

A 30-year-old African-American man presented to our level-1 trauma center via emergency medical services (EMS) after being found by police on a roadside in −15°C weather. Past medical history was significant for human immunodeficiency virus (HIV), schizophrenia, and methamphetamine use. Upon EMS arrival, the patient suffered a witnessed pre-hospital cardiac arrest. The patient was endotracheally intubated by EMS, and CPR was initiated including chest compressions with a Lund University Cardiopulmonary Assist System (LUCAS) version 3 device. One defibrillation and 3 doses of epinephrine were administered during transport for ventricular fibrillation, but there was no return of spontaneous circulation (ROSC). Upon hospital arrival, the patient’s core temperature was 20°C, as measured by the esophageal temperature probe, and his Glasgow Coma Score (GCS) was 3T. His pupils were dilated, equal, and nonreactive bilaterally. Initial bloodwork showed potassium of 3.3 mEq/L, arterial pH of 7.2, base deficit of 6, and lactic acid of 7.0 mmol/L. Warmed trauma whole blood transfusion was initiated. Passive rewarming was started using warm blankets, a Bair Hugger™ (3M™), and an Arctic Sun™ 5000 (BD™). Due to the severe hypothermia, bilateral chest tubes were placed for active intrapleural rewarming. A Foley catheter was placed and continuous bladder irrigation started. Cardiothoracic surgery was consulted for extracorporeal rewarming, and the decision was made to proceed to the operating room with CPB.

In the operating room, CPR was briefly paused to perform wide sterile preparation with Betadine and draping from the neck to the knees. A standard median sternotomy was performed. The heart was exposed and a pericardial well was created. While preparing for cannulation, CPR was resumed with manual cardiac massage in an attempt to maintain perfusion. The patient was heparinized with 30 000 units given directly into the right atrium. The distal ascending aorta and right atrium were cannulated, and the patient was placed on CPB. Rewarming was initiated with a goal temperature of 37°C. ROSC was achieved at a body temperature of 31.7°C following 3 rounds of direct cardiac defibrillation. Time of ROSC was 215 minutes after initial witnessed arrest, 145 minutes after hospital arrival, and 20 minutes after initiation of extracorporeal life support (ECLS). Total CPR time was 195 minutes from witnessed arrest to initiation of ECLS. A timeline of major events in the course of care is provided in Table 1.

The patient became hypoxic, with oxygen saturation in the high-70% range when trying to wean from bypass after ROSC was achieved. Fiberoptic bronchoscopy showed the tip of the endotracheal tube was in the right mainstem bronchus. The endotracheal tube was repositioned, but vitals remained tenuous. The bypass circuit was assessed, and it was noticed that returning blood volume was significantly lower than the volume being given to the patient. This prompted evaluation for other causes of instability. Bilateral chest cavities were explored showing minimal hemothorax. Physical examination showed the patient’s abdomen was tense and distended. Additional paralytic was given, but the abdomen remained tense, so intraoperative trauma surgery consultation was requested.

Focused assessment with sonography for trauma (FAST) examination was performed, and was positive for intraabdominal free fluid. The decision was made to perform an exploratory laparotomy. Dark venous hemoperitoneum was encountered upon entering the peritoneal cavity. The hemoperitoneum was evacuated with suction and 4-quadrant packing was performed. The patient was found to have a small, actively bleeding liver laceration under the falciform ligament, along with multiple palpable anterior rib fractures thought to be secondary to LUCAS CPR. Due to ongoing coagulopathy and cardiopulmonary instability, the laceration was packed with hemostatic gauze and temporary abdominal closure was performed with an AbThera (3MÔ). The patient was then successfully weaned off CPB but sternal closure was deferred due to severe cardiac and pulmonary edema as well as ongoing coagulopathic bleeding. Delayed sternal closure was performed with placement of a 24 Fr mediastinal Blake drain to gravity followed by packing and antimicrobial dressing placement. Arterial blood gas drawn immediately postoperatively showed pH of 7.29, pCO2 of 46 mmHg, pO2 of 289 mmHg, base deficit of 4, bicarbonate of 22 mmol/L, and oxygen saturation of 95%.

The patient was transferred to the intensive care unit (ICU) with 4 vasopressors required to maintain a mean arterial pressure (MAP) goal of >65 mmHg. Peak pressor requirements were norepinephrine at 0.10 mcg/kg/minute, epinephrine at 0.2 mcg/kg/minute, phenylephrine at 0.5 mcg/kg/minute, and vasopressin at 0.03 units/minute. Chest radiography demonstrated significant pulmonary edema, more pronounced on the right side (Figure 1). The intraoperative paralytic was allowed to wear off, and pupillometry showed reactive pupils bilaterally; paralytic therapy was then resumed due to the open chest. He continued to require high oxygenation settings on the ventilator to avoid hypoxia. Blood product resuscitation was continued in attempts to correct coagulopathy; however, there was persistent high-volume output from the temporary abdominal closure (1500 mL over approximately 4 hours). Due to concern for ongoing bleeding from the liver injury, the decision was made to reopen the abdominal incision at bedside. This was performed in the ICU with sterile technique and assistance of operating room personnel. The liver laceration showed ongoing oozing of blood. Hemostasis was achieved with horizontal mattress suturing with chromic suture and repacking of hemostatic agent. A new temporary abdominal closure dressing was placed. Following re-exploration, the patient had received a total of 3 units of packed red blood cells, 3 units of fresh frozen plasma, 2 units of platelets, and 5 units of cryoprecipitate. An additional 2 units of packed red blood cells were transfused at that time, and he did not require any further blood product transfusion throughout hospitalization.

The patient became oliguric overnight, with elevation in creatinine. Nephrology was consulted and continuous renal replacement therapy (CRRT) was started on hospital day 2 (or postoperative day 1). Vasopressor requirements gradually trended down and on hospital day 3, the patient was taken back to the operating room where the chest and abdomen underwent definitive closure without complication. Paralytic therapy was discontinued at that time. The patient remained intubated and care was continued in the ICU.

On hospital day 4, the patient began following commands after sedation was weaned. A transthoracic echocardiogram (TTE) demonstrated a left ventricular ejection fraction (LVEF) of 20–25%. EKG showed a QT interval of 538 ms. Peak troponin was 191 ng/mL at 23 hours after arrival (reference 0.000–0.033 ng/mL). A repeat TTE on hospital day 6 showed improvement in LVEF to 30%. The patient was diagnosed with hypothermia-induced heart failure with reduced ejection fraction. He was extubated on hospital day 5 after passing a spontaneous breathing trial; however, later that day he developed altered mental status and required reintubation due to hypercapnic respiratory failure. He was extubated a second time on hospital day 8. He demonstrated good respiratory and mental status thereafter and continued to progress toward recovery.

Psychiatry was consulted on hospital day 10 and the patient endorsed the events leading to his admission as a suicide attempt. Also on hospital day10, he was transitioned from CRRT to intermittent hemodialysis. His kidney function did not recover, so a tunneled central venous catheter was placed on hospital day 16 to facilitate outpatient hemodialysis. He continued hemodialysis as an outpatient after discharge. Cardiology consultants recommended goal-directed medical therapy with outpatient referral to a heart failure clinic. Frostbite of the bilateral hands was noted and treated initially with topical ceramide cream (Figure 2). When demarcation began to occur, daily application of Betadine was started to provide antimicrobial coverage and facilitate autoamputation. The patient underwent rehabilitation with therapy teams and was discharged home on hospital day 22 with plans to continue hemodialysis 3 days per week. He denied suicidal ideations for several consecutive days prior to discharge and was deemed not a candidate for inpatient psychiatric admission. Chest radiography at discharge showed resolution of pulmonary edema (Figure 3). Outpatient follow-up demonstrated that he remained adherent with hemodialysis and had established care with a primary care provider. Psychiatric follow-up was arranged with an outside health system.

Discussion

Our case is interesting in that it was complicated by hemoperitoneum from a liver laceration secondary to CPR performed mostly with a LUCAS device. We primarily attribute the occurrence of the injury to the duration of CPR (195 minutes). In addition, the LUCAS device was paused and manual compressions were used during the placement of bilateral chest tubes. We suspect device placement was too low on the chest after it was repositioned, and that this misplacement predisposed the patient to solid organ injury. Liver lacerations from LUCAS have been reported in the literature before, including a severe case warranting left hepatectomy [8–10]. Providers caring for patients undergoing CPR should maintain a high clinical index of suspicion for hemothorax and hemoperitoneum secondary to rib fractures. Treatment with automated external chest compression devices is associated with 4-fold to 10-fold higher risk of liver injury compared with manual CPR [11,12]. This is significant because the known incidence of liver injury after manual CPR is rare and is estimated at 0.6% [7]. There is no available systematic review on the management or mortality associated with solid organ injury and hemoperitoneum secondary to CPR, and this literature gap presents a direction for future research. A single-center propensity-matched retrospective study compared manual compressions to the LUCAS device and found an equal rate of rib fractures but a significantly higher incidence of hemothorax with LUCAS (8.7% vs 0%); hemoperitoneum was not assessed by this study [13]. To date, no randomized controlled trials have demonstrated superiority of device-assisted compressions over manual compressions, although mechanical devices do offer the advantages of consistent high-quality compressions and preservation of human resources [13].

The decision to operate on this patient with severe HCA despite a prolonged CPR time of 195 minutes was influenced by a few key factors. First, the patient had a normal serum potassium of 3.3 mEq/L at the time of presentation. Hyperkalemia is significantly associated with mortality in severely hypothermic patients, and there are no reported cases of survival when serum potassium exceeds 12 mEq/L. Second, he presented in ventricular fibrillation, which is associated with favorable prognosis. A systematic review of 108 witnessed HCA patients showed that a presenting rhythm of ventricular tachycardia or ventricular fibrillation was associated with 84% survival as compared with asystole (70%) or pulseless electrical activity (43%) [4]. Finally, he experienced a witnessed pre-hospital arrest, which is associated with roughly double survival as compared with unwitnessed pre-hospital arrest [5,6].

The Hypothermia Outcome Prediction after ECLS (HOPE) score is a validated survival estimation tool for patients presenting with HCA [14,15]. It can be quickly calculated using 6 variables including age, sex, asphyxiation, initial serum potassium, temperature upon arrival, and duration of CPR until start of ECLS. The HOPE survival probability for our patient was 78%. A drastically unfavorable HOPE score may influence the decision to proceed to the operating room in similar cases.

We encountered disproportionately low venous return when attempting to wean from CPB, which led to examination of the abdomen and discovery of hemoperitoneum. In the setting of inadequate venous return to the bypass circuit, a systematic approach to problem-solving is invaluable. Lines and cannulas should be assessed, then vasoplegia or ongoing hemorrhage should be considered. The approach to hypoxia when weaning from CPB should be similarly systematic. Key considerations include adequacy of lung expansion, ventilator settings, airway secretions, pleural space pathology, anemia, and clinically significant valvular regurgitation.

The initial choice of CPB rather than extracorporeal membrane oxygenation (ECMO) was driven by the patient’s unstable condition on arrival. The patient had profound vital instability and physiologic derangements which were more readily correctable on a CPB circuit than on an ECMO circuit. ECMO is less readily available than CPB but is associated with greater mortality benefit [15]. Bjertnaes et al performed a meta-analysis of 23 observational case series comprising 464 patients with pre-hospital HCA and found that ECMO had a higher survival rate at discharge than CPB (44%, n=219 vs 31%, n=235) and concluded that survival was 41% more likely with ECMO than with CPB [16]. However, in their study, patients initially resuscitated with CPB and subsequently placed on ECMO for failure to wean were sorted into the ECMO group, suggesting the proportion of CPB-treated patients was higher than reported. Altogether, the evidence indicates that CPB is the mainstay initial therapeutic modality and that ECMO becomes a preferred strategy for patients with failure to wean. It must be noted that continuous ECMO requires systemic anticoagulation, which is undesirable for coagulopathic hypothermic patients with ongoing bleeding, as in our case.

Severe hypothermia is defined as core body temperature below 28°C. Austin et al performed a systematic review including 47 cases of accidental hypothermia treated with extracorporeal rewarming and found a median body temperature of 24.6°C upon presentation [interquartile range 22.2°C to 26°C] [17]. Walpoth et al assessed 73 patients who suffered HCA using data from the International Hypothermia Registry and found a median body temperature of 25.0°C upon hospital arrival [interquartile range 22.5°C to 27.4°C] [18]. These data suggest that our patient, who presented with a core temperature of 20°C, was within the lowest quartile of body temperature for survivable HCA. The lowest recorded survivable body temperature following HCA is 13.7°C in the adult population and 11.8°C in the pediatric population [19–21].

Although we report a rare complication of treatment, this report highlights the need for conscious and intentional placement of resuscitative devices, as too low of a placement can lead to severe intra-abdominal consequences. The case is generalizable in that our reported complication could occur during any CPR with mispositioned chest compressions, regardless of the etiology or circumstances leading to cardiac arrest. The specific context of this patient’s presentation and treatment does not greatly alter the generalizability of the complication with regard to the technical aspect of device placement. However, our report is limited in that it is a single patient experience and is therefore not powered to allow for conclusions about the incidence of hemoperitoneum with device-assisted chest compressions or the incidence of hemoperitoneum that requires surgical intervention. More large-scale studies are needed to clarify the epidemiology of this uncommon resuscitative complication.

Conclusions

Pre-hospital HCA is a complex condition with severe sequelae including arrhythmic instability leading to cardiopulmonary arrest, pulmonary edema, and coagulopathy. While ECMO has demonstrated superior survival benefits over CPB, its limited availability necessitates alternative management strategies in resource-limited settings. Clinicians should maintain a high index of suspicion for occult injuries related to CPR, particularly those associated with automated chest compression devices such as LUCAS, which carry a significantly higher risk of liver injury. Additionally, delayed sternal closure can play a critical role in managing cardiopulmonary edema and coagulopathy following extracorporeal rewarming.