Stab Wound of the Heart with Unusual Sequelae

Abstract

A 31-year-old woman was admitted to the emergency department with a stab wound to the heart. She was initially stable but rapidly developed hypotension. While the operating room and staff were in preparation, she underwent pericardiocentesis. She was then rushed to the operating room by the general surgical trauma team, who performed a bilateral anterior thoracotomy to control the bleeding. In the recovery room, the patient was still hypotensive, so cardiothoracic surgery was consulted.

An echocardiogram revealed severe hypokinesis of both ventricles. The cardiothoracic surgeons returned her to the operating room and discovered that the anterior pericardium had been completely removed by the trauma team. This had caused the posterior pericardium to form a “bowstring” that almost totally obstructed pulmonary venous return and restricted right ventricular outflow of blood, inducing right-sided heart failure. This pericardial string also strangulated the left atrium posteriorly, forming 2 compartments. We repositioned the patient's heart and implanted ventricular assist devices bilaterally to provide temporary circulatory support. The patient made a good recovery. We suggest that bilateral assist device placement can be beneficial in the recovery of a stunned but otherwise normal heart.

Penetrating injuries of the heart have multiple causes: most commonly, gunshot wounds, stab wounds, or blunt trauma. We propose a method to support the stunned and failing heart until recovery can take place, if the heart is otherwise normal. Cardiac herniation can be dangerous and fatal if not recognized as a sequela to emergency department thoracotomy.

Case Report

In November 2009, a 31-year-old woman was brought to the emergency department after sustaining two 0.5-cm stab wounds 1 cm to the left of the sternum in the 3rd intercostal space. She was awake and oriented, with a blood pressure of 137/70 mmHg and a heart rate of 120 beats/min. She had neck-vein distention, distal heart sounds, and deteriorating blood pressure (Beck's triad). While the operating room and staff were in preparation, pericardiocentesis was performed, whereby 60 mL of nonclotting blood was removed. The patient temporarily improved, then rapidly deteriorated and did not improve upon massive transfusion resuscitation in accordance with protocol. Before consulting the cardiac surgery team, the general surgical trauma team rushed her to the operating room and performed a left thoracotomy. Because they were unable to control the bleeding, they extended the thoracotomy across the sternum as a “clamshell” incision. Two small 0.5-cm stab wounds were noted proximal to the outflow tract on the anterior surface of the right ventricle (RV). Two pledgeted sutures easily controlled this bleeding. The patient was taken to the recovery room in fair condition but still required vasopressors. A subsequent transesophageal echocardiogram (TEE) showed a left ventricular ejection fraction (LVEF) of 0.30, with no evidence of cardiac tamponade (Fig. 1). However, the patient's condition continued to deteriorate, necessitating higher doses of vasopressors. Her urine output decreased, which was leading to anuria and severe lactic and metabolic acidosis.

Fig. 1 Transesophageal echocardiogram shows the mitral valve closed and the bowstring sign indicating constriction of the left atrium.

Several hours later, a cardiology consultation was obtained and a transthoracic echocardiogram (TTE) was performed. That study (Fig. 2) was negative for cardiac tamponade but revealed an LVEF of 0.10 to 0.15 and an akinetic RV. There was moderate mitral and tricuspid regurgitation and no evidence of traumatic septal defects or shunting. This was confirmed with TEE (Fig. 3), but no related diagnosis was made. An anomaly was found on both TTE and TEE (Figs. 2 and 3), but no related diagnosis was made. An intra-aortic balloon pump was placed but yielded little improvement. At this point, the patient's attending physicians called for a cardiac surgery consultation. We rapidly returned the patient to the operating room, because we knew that she would need further surgery. Yet we had no specific diagnosis and no explanation for her downward spiral. We did not have time to scrutinize the TEE (Fig. 3), but we would most likely have missed the diagnosis anyway.

LA = left atrium; LV = left ventricle

Fig. 2 Transthoracic echocardiogram shows the mitral valve open and the bowstring sign indicating constriction of the left atrium.

LA = left atrium; LV = left ventricle

Fig. 3 Transesophageal echocardiogram shows the mitral valve open and again the bowstring sign indicating constriction of the left atrium.

LA = left atrium; LV = left ventricle

A midline sternotomy was performed in anticipation of the need for cardiopulmonary bypass (CPB). The RV appeared to be markedly dilated and totally akinetic. The left ventricle (LV) was severely hypokinetic and barely contracting. The blood gas analysis revealed a pH of 7.01. A full dose of intravenous heparin (3 mg/kg) was administered, and CPB was promptly instituted via standard distal aortic and atrial venous cannulation with vacuum-assist drainage. Examination of the heart revealed no other pathologic condition. The coronary arteries were intact. We were treating the symptoms but still had no explanation of why the heart was failing.

Upon lifting the heart, we noted that the anterior pericardium had been completely removed with 1-cm margins, from phrenic nerve to phrenic nerve. Posterior to the heart, the pericardium was bunched up—thus acting as a “bowstring” that almost totally obstructed pulmonary venous return and therefore restricted the outflow of blood from the RV, causing right-sided heart failure. Such indentation can affect any chamber of the heart, but it is most likely to affect the left atrium. This “string” strangulated the left atrium posteriorly, forming 2 compartments. The entire heart fell into the left side of the chest.

After we placed the flattened pericardium beneath the heart in its proper location, cardiac function improved over the next several hours. However, the patient could not be weaned from CPB. The RV appeared to be more severely impaired than the LV. We placed a CentriMag® RV assist device (RVAD) (Thoratec Corporation; Pleasanton, Calif) from the right atrium to the pulmonary artery. Because the LV was unable to handle the amount of blood that was delivered by the RVAD, an LV assist device (LVAD) was then placed via the left atrium through the mitral valve into the LV, with the outflow into the ascending aorta. Cardiopulmonary bypass was then discontinued. To ensure that the heart would not herniate again, we sutured several large strips of Edwards bovine pericardium (Edwards Lifesciences Corporation; Irvine, Calif) from one edge of the native pericardium to the other edge. The sternum could not be approximated because of severe edema of the heart. The chest was therefore closed with a large 2-mm Gore-Tex® patch (W.L. Gore & Associates, Inc.; Flagstaff, Ariz). Multiple large-bore 32F Argyle drains were placed in the pericardium and anterior mediastinum. The patient was then returned to the recovery room in stable condition.

The next morning, the sternal patch appeared to be bulging. The patient was returned to the operating room to evacuate the underlying blood and irrigate the wound. The heart appeared to be contracting better. Over the next 4 days, the patient's condition continued to improve. She was partially sedated with a small dose of propofol, which did not compromise her hemodynamically. She was able to move all 4 extremities and follow simple commands. Her arterial blood gas levels were normal and she had excellent urine output. When the patient could be weaned down to 1 L of flow/min on the CentriMag with acceptable hemodynamic results, she was returned to the operating room for RVAD and LVAD explantation. On TEE, the heart function was almost normal and the cardiac chambers were much reduced in size. The sternum was amenable to approximation without difficulty, and the patient needed minimal vasopressors. The next day, the vasopressors were discontinued and she was extubated. On postoperative day 15, she was discharged from the hospital. Two weeks later, she had resumed her normal daily activities and was totally asymptomatic.

Discussion

Historical opinion concerning the invasive care of cardiac wounds has evolved from Boerhaave's judgment in the 18th century¹ that all penetrating injury to the heart is fatal.² Napoleon's surgeon, D.J. Larrey, decompressed a wound of the heart in 1829 by draining the pericardium.³ In 1883, Theodor Billroth reportedly stated that “the surgeon who should attempt to suture a wound of the heart would lose the respect of his colleagues.”⁴ Cappelen, in 1895, was the first person to suture the heart; the patient died less than 3 days later.⁵ In 1896, Paget expressed the opinion that “surgery of the heart has probably reached the limits set by nature to all surgery: no new methods and no new discovery can overcome the natural difficulties that attend a wound of the heart.”⁶ In 1897, Ludwig Rehn in Frankfurt, Germany, sutured an actively bleeding 1.5-cm wound of the RV; the patient recovered.⁷ In the United States, in 1902, Luther Hill of Montgomery, Alabama, first sutured the heart while repairing—on the kitchen table—a stab wound in a 13-year-old boy (that injury having been caused by a kerosene lamp).⁸

Management of cardiac injury is predicated upon the presenting condition of the patient on arrival at the hospital. Approximately 20% of all patients who reach the hospital have no obvious sign or symptom of cardiac injury.^9–12 When first seen, 80% to 90% of cardiac stab wounds develop cardiac tamponade, as opposed to 20% of gunshot wounds.¹³ Moreno and colleagues¹¹ observed that pericardial tamponade is a positive predictor of survival in penetrating cardiac injuries to both right and left ventricles. As a general rule, gunshot wounds are more lethal than stab wounds (77% vs 42%), a finding unchanged in a large retrospective study.¹⁴ A large prospective study by Asensio and colleagues¹⁵ reported an overall mortality rate of 63% (86% for gunshot and 32% for stab wounds); sinus rhythm and mechanism of injury at the time of pericardiotomy predicted a survival advantage.

If the patient is hemodynamically stable and there is a suspicion of cardiac involvement, whether by location of the penetrating injury or by symptoms of cardiac injury, the principles of care should include the placement of a large-bore intravenous line for access and orders for an electrocardiogram (ECG), chest radiograph, and echocardiogram.^16,17 A radiograph might show an enlarged cardiac silhouette, with or without a hemothorax or pneumothorax. On the ECG, ST-T wave changes might be present. If the patient presents with the classic Beck's triad (distended neck veins, low blood pressure, and distant heart sounds) yet remains stable enough, TTE would be the best way to diagnose suspected cardiac injuries. On the other hand, if the patient is unstable, movement to the operating room should be expedited for the creation of a pericardial window or for definitive exploration. Pericardiocentesis is perhaps appropriate as a temporary measure, until the operating room becomes available.^18,19

Historical experimental work showed that death can result from an acute pericardial pressure of 12 to 22 cm H₂O or a volume of 150 cc of acutely accumulated pericardial fluid.^20,21 In a study of animals, removal of only a small portion of the pericardial fluid led to a much lower mortality rate; and rapid infusion of fluid to raise the central venous pressure (CVP) enabled these animals to withstand intrapericardial pressures that were ≤146% of baseline.²² About 80 to 100 mL of fluid can be acutely tolerated in physiologic negative intrapericardial pressure (IPP) without hemodynamic compromise. A high CVP/IPP gradient is therefore needed to prevent systolic blood pressure compromise.²³

Failure to obtain blood from the pericardium is of no diagnostic significance, but the aspiration of nonclotting blood most probably indicates cardiac injury. Griswold and Maguire stated in 1942 that “if tamponade is serious enough to require aspiration, open operation with suture of the heart and drainage of the pericardium into the pleura is the treatment of choice.”²⁴

Herniation of the heart is an uncommon occurrence,^25,26 and a congenital opening in the pericardium is rare. Postpneumonectomy cardiac herniation is more common, especially with intrapericardial ligation of the pulmonary artery and veins.^25,26 In blunt trauma, cardiac herniation results from rupture of the pericardium due to intense pressure from compression of the pericardium between the sternum and the spine. The more common use of minimally invasive coronary artery bypass (MIDCAB) surgery through the left side of the chest has led us to expect more reports of herniation of the heart, consequent to inadequate closure of the pericardium. We have actually seen this after an uneventful MIDCAB, and it is characterized by dramatic postoperative hemodynamic deterioration.

As in other parts of the body, herniation might cause no symptoms or it might give rise to obstruction or strangulation. When the latter develops, signs of low cardiac output and elevation of CVP or ECG signs of rotation of the heart can be seen. Chest radiographs might show distortion of the cardiac shadow or a change of heart position. With this, our first reported experience of posterior pericardial herniation, we recommend a TEE study to look for a constrictive bowstring sign, a band across one of the chambers of the heart, most commonly the left atrium. Cardiac herniation is most often missed, until postmortem examination.

Summary

As very nearly illustrated by this case, the development of cardiac herniation after injury repair can be fatal. Prompt diagnosis and surgical management can save lives. If time permits, the most reliable test is an echocardiogram and the most reliable procedure is pericardiocentesis, both for diagnosis and temporary treatment. If the patient's instability does not permit these interim procedures, his or her rapid transport to the operating room for pericardial window or definitive treatment is most appropriate. We had an unfortunate sequela (cardiac herniation) after complete removal of the anterior pericardium, which had been emergently performed through a bilateral anterior thoracotomy clamshell incision. Although an LV apical herniation can be anticipated after an emergent left thoracotomy for cardiac injury repair or resuscitation, constriction of the left atrium and pulmonary venous inflow should also be kept in mind if the anterior pericardium is opened widely. Aggressively resuscitated cardiac ventricles can be severely dilated and therefore more susceptible to herniation, a factor that can contribute to hemodynamic failure. Posterior heart strangulation (identified by the bowstring sign) should be suspected when a large pericardial incision has been made and the patient's rapid decline occurs in the absence of another reasonable explanation. Echocardiography or surgical exploration can prove to be life-saving. Although CPB support has been useful in early and late management of complex cardiac wounds,²⁷ RVAD and LVAD placement can enable the stunned myocardium to recover in select circumstances.