How to mend a broken heart: a major stab wound of the left ventricle

A 28-year-old male admitted with a stab wound under his left nipple, underwent emergency surgery because of confusion, a decreasing blood pressure and increasing tachycardia. A median sternotomy incision was made and after establishing cardiopulmonary bypass, a 7 cm wound in the left ventricle and a smaller wound in the left atrium were repaired. An injured segment of lung was resected and the left anterior descending and circumflex arteries were grafted after weaning from cardiopulmonary bypass was initially unsuccessful. Although the patient suffered a stroke, probably due to prehospital hypoperfusion, he eventually recovered without major sequelae. In addition to the case report we present a literature review of the last 15 years pertaining the management of penetrating cardiac injury.


Background
Onehundred and sixteen years ago Alex Cappelen repaired a penetrating injury of the left ventricle through a left anterior thoracotomy in Christiania (former name of Oslo), in one of the world`s least violent countries [1]. Cappelen`s operation is considered to be the first report of a cardiac surgical procedure. Today trauma centers all over the world perform complex cardiac repairs due to penetrating trauma but the mortality is still high [2][3][4][5].
We report the case of a young man who suffered a large stab wound (SW) in the left ventricle and left atrium in addition to a lung injury for approximately 2 h before undergoing reparative surgery. In addition we present a literature review of penetrating cardiac injuries from 1997 -2012 (Table 1). As data source we used all available English-language articles from peerreviewed journals in the Ovid MEDLINE and PubMed databases. The articles selected were relevant case reports, original articles and reviews focusing on the clinical presentation of penetrating cardiac injury, initial management, operative technique, complications and follow up.

Case presentation
A 28-year-old male was admitted to the emergency department (ED) with a 5 cm stab wound (SW) under his left nipple. Pre-hospital treatment included insertion of a left chest drain due to dyspnoea, but this was clamped during transport because of massive hemorrhage. On admission, he was self-ventilating, with palpable carotid pulses, but without a measurable blood pressure. He was agitated and pale with a Glasgow coma score of 12 since he could open his eyes, localize pain and speak. The blood pressure ranged from 80/60 to 100/60 mmHg after starting intravenous fluid therapy and he had a tachycardia of 100-120 beats per minute. When the clamp was removed from the chest drain, 650 ml of blood was rapidly drained. The chest x-ray showed persisting hemothorax and atelectasis and an additional drain was inserted. The arterial saturation varied from 86% to 98% and blood gas analysis showed a haemoglobin of 12.6 g/l, pH 7.17, base excess −9 and lactate 5.5 mmol/l. Focused Assessment with Sonography in Trauma (FAST) revealed no blood in the pericardium and upper abdomen. The neck veins were not distended and so the patient received transfusion of 1500 ml of crystalloid fluid and 250 ml of red cells. The blood pressure decreased as soon as the intravenous therapy was reduced, the tachycardia did not resolve and the patient was therefore transferred to the operating room.
After intubation, the ECG showed ST elevation and a median sternotomy incision was rapidly performed. The     pericardium was opened and although there was a clot ventral to the heart, there were no signs of cardiac tamponade. There was a 6 cm cut in the lateral pericardium corresponding to the stab wound in the chest and a 7 cm, almost transmural wound in the left ventricle, parallel to a major diagonal branch ( Figure 1). The wound was not bleeding. A 5 cm stab wound in the left lung ( Figure 2) was sutured and cardiopulmonary bypass (CPB) was established. The cardiac injury ended close to the origin of the left main stem and crossed the left atrium. The ventricular wound was repaired with single mattress sutures reinforced by strips of bovine pericardium (Figures 3, 4) without arresting the heart and without cross-clamping the aorta. The left atrium was sutured using 5/0 Prolene (Ethicon). Blood appeared in the tracheal tube and bronchoscopy revealed ongoing bleeding from the left lung which required resection of the lingula. Weaning from CPB was initially unsuccessful and we suspected that there had been injury to the left main stem either caused by the initial stab or by the hemostatic sutures. The left anterior descending artery was grafted using the internal mammary artery and a vein graft was anastomosed to the circumflex artery. The patient was thereafter successfully weaned from CPB. Post-operatively, the patient had signs of a stroke and a CT scan revealed a cerebral infarction. One week after surgery he was transferred to the neurological intensive care unit. After three weeks he was awake and self-ventilating. He was moved to his local hospital and was discharged after 6 weeks with only a minor deficit affecting the left upper extremity.

Discussion
We report the case of a young male patient with a major cardiac stab wound combined with lung injury. Our patient was stabbed during a violent quarrel, thus being a typical stab victim, however, in Japan suicide attempts seem to be equally frequent [18,23]. In large series, gunshot wounds (GSW) are the predominant cause of cardiac penetrating trauma [2,4,6,29]. In Norway, this type of injury is obviously less common but still existing [37][38][39].    Knife is the most common weapon for stab injuries, followed by other sharp items such as screwdrivers [34], ice picks [19], chopsticks, pneumatic nailgun nails [14,20,40] but also curiosities as barb from a sting ray [28]. Fractured ribs or sternum are also reported to cause cardiac penetration [41]. Pneumatic nails might be shot without the patient noticing and cause surprise when detected by CT scan or eccocardiography imbedded in the heart [14,20]. The iatrogenic penetrations of the heart due to different medical devices (pacemaker leads, intracoronary stents, Amplatzer devices) are not discussed in this paper.
Penetrating cardiac wounds are mostly fatal either due to cardiac tamponade, exsanguination or coronary artery injury [1]. Clarke reports that of 1064 patients with stab wounds to the chest 104 were operated and 76 were found to have a cardiac injury [3]. The overall mortality was 10% giving an impression of low mortality in this particular group of cardiac injuries. However, when the data was put together with the mortuary report for the same time, the mortality for penetrating cardiac stab wounds was found to be 30%. Most of the studies are retrospective and the patient selection is determined by the survivors arriving at the hospital and ignorance of the mortuary data. Topal et al. report a mortality rate of 15% in 61 penetrating cardiac cases with predominantly stab wounds but state that "patients pronounced dead on arrival were not assessed in this study" [33]. The only known prospective study reports another reality with a mortality rate of 97% when multichamber penetrating injury is present [2]. Also Molina et al. reports high mortality (67%) in a cohort with mainly stab wounds throughout the last decennium [4].
Our patient maintained suboptimal circulation for approximately two hours before undergoing surgery. The time span taken into consideration, our patient was extremely lucky as the outcome is usually poor when the time from trauma to surgery increases [5,6]. An Israeli study of 14 patients reports 100% survival (9 SW, 2 GSW, 1 shrapnel injury and 1 multi trauma) with the mean time from injury to surgery of 37 min [7]. In addition to fast admission to surgery, this outstanding result may also be due to the fact that all patients had single chamber injuries and no coronary artery injury. According to Burack et al., patients with penetrating mediastinal trauma triage themselves between operative intervention or evaluation and observation as they present either stable or unstable on admission. In this retrospective study the authors present 207 patients of which 72 were unstable [10]. Of these 15% had cardiac injury with 18% survival when explored in the ED. The survival rate was 71% when patients with penetrating cardiac injury reached the operating room. All patients having cardiac injury in this study were unstable (authors criteria: traumatic cardiac arrest or near arrest and an emergency department thoracotomy (EDT); cardiac tamponade; ATLS grad III shock despite fluid resuscitation; chest tube output >1500 ml at insertion; chest tube output >500 ml in the initial hour; massive hemothorax after chest tube input). The study does not report the use of CPB.
In our patient, there was a large stab wound of the left ventricle running parallel to the diagonal artery as well as a stab wound in the left atrium. Regarding the location of penetrating cardiac injury, the right ventricle is the most common due to its ventral anatomical position,  followed by the left ventricle, right atrium and left atrium [2,3,11]. The patients with a single right ventricle injury are mostly salvagable whereas those with multichamber injuries have a very high mortality [2,4,21]. The concomitant injury of the lung in our patient is not a rarity [3]. Our patient did not suffer from cardiac tamponade as there was a large opening to the left pleural cavity through the wound in the pericardium. This probably saved his life, although profound hypovolemia can conceal signs of cardiac tamponade leading to delayed diagnosis [36]. However, cardiac tamponade in the reviewed studies is not a prognostic factor regarding survival [2,33].
The role of CPB has been debated in trauma surgery, espescially when it comes to penetrating cardiac wounds [6,21]. Some series present large cohorts of penetrating cardiac injury without use of CPB [3][4][5]. In case of complex cardiac injuries with multichamber lacerations the advantages of a bloodless and still operating field is obvious [6,20,21]. The required heparinisation for CPB might be deleterious in a trauma patient. However, if the bleeding source or sources can be controlled, the risk of further profound haemmorhage is low. On the other hand, full heparisation might cause severe morbidity, and CPB might initiate consumptive coagulopathy and profound systemic inflammatory reaction [28]. Off pump cardiopulmonary bypass is an alternative when it comes to coronary artery lesions [16,22,25]. Establishing CPB in arrested patients or patients in deep haemorrhagic shock is not favourable for the outcome [6]. It could be debated whether or not the aorta should have been cross-clamped in our patient during repair of the left ventricular wall and coronary bypass surgery, but the ECG changes and the suspicion of pre-existing ischemia due to sustained pre-operative hypoperfusion, persuaded us to leave the aorta unclamped in this particular case.
Peroperative fluorescent angiography is a reliable tool to identify suspect coronary artery involvement peroperatively either caused by the injury itself or the surgical procedure [15], unfortunately this technique was not available at our OR. Cardiac stabbings might lead to initially unidentified additional injuries which become apparent first several weeks to years later [8,18]. One study with a large series of patients report that these injuries seldom need surgical treatment [5]. There is consensus that echocardiographic assessment should be provided during the hospital stay [5,11].
On admission to the ED, our patient was given a high Glasgow coma score (GCS), yet post-operatively was found to have had a cerebral injury. Unfortunately, the patient was foreign, and despite speaking, nobody could assess his verbal response adequately. Furthermore, he received an intravenous injection of Ketalar a few minutes after admission, following which he needed assisted manual ventilation and was no longer able to communicate. The initial GCS was later reconsidered and probably the patient suffered from major hypoxia in the pre-hospital phase. Nevertheless the patients with lower GCS have poor outcome, Asensio still reports a high mortality rate (27%) for patients with Glasgow Coma Scale >8 [2]. However, in an emergency room thoracotomy material GCS was found to be a predictor of survival, despite none of the patients had a score >7 [29]. In our patient, it is possible that CPB might have caused cerebral injury by embolization or by giving an insufficient cerebral perfusion pressure. With pre-existing cerebral damage, the standard perfusion pressure during CPB in our patient (mean arterial pressure 50-60mmHg) might not have been high enough to meet the needs of the brain already damaged by hypoperfusion.
Patients with a simple penetrating cardiac injury might be successfully managed without a cardiac surgeon present [2,3]. However, repair of a severe wound of the left ventricle and the complications that can arise will require the surgical skills of a cardiac surgeon, as demonstrated in the present study and the likelihood of survival will be considerably increased by the immediate availability of a cardiac surgical service. The cases where initial tamponade was managed at a lower trauma care center with further transfer for definite surgery, witness of general surgeon`s competence of the initial management of these patients [13,28]. In our level I trauma center, a cardiothoracic surgeon in the trauma team has been practiced for decades and we believe provides optimal management of patients with penetrating cardiac trauma.

Conclusions
We present a complicated case of a young male patient with a chest stab wound who served the trauma team both diagnostic and treatment challenges. We provide the reader a review of literature of the last 15 years publications on penetrating cardiac injury, focusing on stab wounds. Our patient suffered a stroke which origin could be multigenetic, prehospital hypoperfusion, air emboli due to major lung injury and/or insufficient perfusion pressure or microemboli during the cardiopulmonary bypass. The patient in our study survived with minor sequelae due to coordinated work of the trauma team in charge. In conclusion, if the patient with a penetrating stab wound in the heart is not obviously dead on arrival, an attempt for cardiac repair should be done with or without CPB.