CARDIAC SURGERY
Penetrating cardiac injuries: 28-year data analysis

Introduction

The sad truth is that much of our knowledge of penetrating chest trauma has arisen from the unfortunate predilection of Homo sapiens to poke holes in the chests of his fellows. The most ancient data about cardiac injuries were found as early as in the Edwin Smith papyrus written by the Egyptian Imhotep more than 5000 years ago [1]. Numerous attempts were made to treat patients with heart injuries in ancient times but none of them were successful [2-6].

In 1868, Fischer [7] published a monograph on the natural history of cardiac wounds, documenting a 10% survival rate. Successful experimental suture of the heart was demonstrated in 1882 and 1895 [8] but suggestions that cardiac suture could be performed in human beings was an unpopular view. The statement “surgeon who would attempt to suture the wound of the heart would lose the respect of his colleagues”, attributed to the well-known Viennese surgeon Theodor Billroth, reflected a widely held opinion [9]. The first attempt to repair cardiac injury was carried out by Axel H. Cappelen from Oslo (Norway) on September 4, 1895. However, on postoperative day 3 the patient died from sepsis [5, 10]. The second attempt to suture the heart was made in Rome (Italy) by G. Farina in March 1896. Unfortunately, it was unsuccessful as well [11].

A German surgeon from Frankfurt am Main, Ludwig Rehn, was the first to successfully perform cardiorrhaphy in September, 1896. Later he was able to collect 124 cases of heart suture with a survival rate of 40% [12].

In Lithuania the first successful cardiorrhaphy was performed in June 1926 [13].

The Department of General Thoracic Surgery in Vilnius University Hospital (as a separate unit) was established at the beginning of 1983. Since then general thoracic surgical service has been provided for patients from the city of Vilnius and its region.

Penetrating cardiac injury is a dramatic and often lethal form of trauma. The majority of patients are young persons. Numerous publications on the subject of injuries to the heart have been published; however, despite huge advances in medicine, only 6-30% of patients reach hospital alive [4, 14, 15]. Also, most often they reach the hospital in deep hemodynamic shock [4, 16, 17].

Nevertheless, prompt diagnosis and surgical repair can salvage patients who would otherwise be lost. Currently, most knowledge involving injuries of the heart and aorta comes from American studies or South Africa series, due to increased violence in these societies compared with Western Europe [14]. Since the transition to democracy in South Africa in 1994, there has been a gradual reduction in trauma [14]. Contrary to this, the transition to democracy in Lithuania in 1990 did not show such a tendency. Unfortunately, Lithuania and other parts of Europe are experiencing an increase in the incidence of penetrating traumas [18, 19].

Material and methods

A retrospective study of 186 patients with penetrating trauma to the heart was conducted over a period of 28 years (1983-2010). All the patients underwent surgery at the Department of General Thoracic Surgery of Vilnius University Hospital. Our facilities are located in the centre of Vilnius (the capital of Lithuania) with about 500 000 inhabitants. This hospital also provides medical services for the Vilnius region with a population of 100 000 inhabitants.

In order to evaluate factors influencing survival, data obtained for each patient included age, sex, time from the moment of trauma to admittance, and time “lost” in the hospital before emergency thoracotomy. Injury mechanisms, number of wounds, additional injuries, hemodynamic parameters and signs of life (SOL) on admission, as well as operative findings and presence of asystole were evaluated.

Signs of life were defined by the presence of any of the following: measurable blood pressure, palpable carotid pulse, pupillary response, spontaneous breathing efforts, and extremity movements (according to American College of Surgeons Committee on Trauma six estimated signs) [20]. However, we were unable to include the sixth sign, cardiac electrical activity, since this examination in some patients was not performed on admittance due to unknown reasons. Hemodynamic data were evaluated measuring systolic blood pressure and pulse rate.

Heart injury severity score was evaluated according to the American Association for the Surgery of Trauma (AAST) Heart injury scale, which includes six grades [15, 21, 22].

The statistical software SPSS 17.0 for Windows (SPSS Inc. Chicago, Illinois, USA) was employed. Due to the non-normal distribution of variables data are presented as median and interquartile range (IQR). Pearson Chi-square and Fisher exact tests were used to compare frequencies of categorical variables. To compare continuous variables Mann-Whitney test was used. Univariate and multivariate logistic regression analysis was used to calculate the impact of different factors on survival. Influence is shown by odds ratio (OR) and its 95% confidence interval (CI). Multivariate analysis was performed separately for the group of factors before surgery and for the group of factors identified during the operation. A value of p < 0.05 was considered to be significant.

Results

Over the 28-year period 193 patients were operated on for suspected cardiac injury. The detailed evaluation of data was applied to 186 persons who were operated on during the first 24 hours after trauma.

There were 165 males and 21 females. Median age was 33 years (IQR: 26-44). The vast majority of patients (139/186, 74.7%) reached the hospital by Vilnius city ambulance while 24 patients were transferred by Vilnius region ambulance service and the remaining 23 arrived by routine vehicles. The average time from the wounding moment until the call to the ambulance service was 28 (IQR: 9-74) minutes. Approximately 44 (IQR: 33.8-59.3) minutes passed from the emergency call until the patient’s arrival at the hospital. However, there was an extremely large difference between the city and regional ambulance characteristics observed: the average time from the wounding moment to arrival at the hospital in Vilnius city was 60 (IQR: 50-91) minutes, whereas in the Vilnius region it was 240 (IQR: 82-390) minutes. The transportation time of patients to the hospital in the city lasted approximately 41 (IQR: 31-53) minutes, whereas in the region it was 61 (IQR: 50.5-78) minutes.

One hundred and seventy-six patients sustained stab wounds and only ten had gunshot injuries. In 135 cases (72.6%) a single wound was found, while the remaining patients suffered from multiple wounds to the chest. In 34 cases (18.3%) heart injury was the result of a suicide attempt. Most patients (151/186, 81.2%) had wounds in the so-called “heart area” (precordium), which is shown in Figure 1. A similar dangerous area is shown by other authors [4, 22, 23].

All five SOL on admission were found in 145/186 (78%) patients. Forty (21.5%) patients had SOL from one to four, and only one patient had no signs of life on admission. The median blood pressure was 90 mm Hg (IQR: 60-111), and median pulse rate was 100 beats per minute (IQR: 86-118).

All 186 patients underwent operation for suspected heart injury. In 38 cases (20.4%) emergency thoracotomy was postponed. The main reasons for such delay were as follows: hemodynamically stable patients in relatively good general condition refused surgical intervention; in others laparotomy for intraabdominal bleeding was performed first. The most frequent operative approach was left anterolateral thoracotomy (157/186, 84.4%); right thoracotomy was used in 24/186 (12.9%) patients and longitudinal sternotomy in 5/186 (2.7%) cases. The mean time from admittance to emergency operation was 50 minutes (IQR: 30-80).

During the operation in 181 patients pericardiotomy was performed. Intrapericardial findings are presented in Table I. The heart injury, according to AAST, is graded from I to VI (Table II). No injury of the heart, pericardium or prepericardial fat at the operation was found in seven (3.8%) cases. They were all operated on for suspected heart injury but only injuries of other organs were found. In 17 cases (9.1%) during thoracotomy only prepericardial fat was injured, and in 24 cases (12.9%) only a wound of the pericardium was found (in three cases it was accompanied by cardiac tamponade). The right ventricle was the most often injured chamber – 75/186 (40.3%). Some patients had multiple wounds of the heart and 106 (57%) had associated injuries (Table III). In 22 patients laparotomy for intraabdominal bleeding was performed.

Cardiac arrest during the operation occurred in 31/186 (16.7%) persons. In 27 cases cardiac rhythm was restored by manual cardiac massage, whereas in four cases medications alone were enough. However, in 11 cases manual cardiac massage was ineffective and the patients died in the operating room.

Twenty-one out of 186 operated persons died: 11 in the operating theatre, 9 during the first 48 hours postoperatively and 1 patient seven days later. The overall survival rate was 88.7%. Survivors were compared to non-survivors with respect to the above survival predictors (Table IV). Univariate and multivariate logistic regression analysis were utilized to predict outcome for each clinical factor (Table V). Upon univariate analysis systolic blood pressure, all SOL on admission, and lower grade (I-III) of heart injury all predicted survival until hospital discharge. Mortality was predicted by cardiac tamponade, higher grade (IV-VI) of heart injury, injured right atrium and heart arrest during surgery. Multivariate analysis revealed that younger age and the presence of all SOL were independent predictors of survival until hospital discharge. Higher grade (IV-VI) of heart injury and heart arrest during operation independently predicted fatal outcome.

Discussion

In our study the overall survival rate of operated patients was 88.7%. Such a high percentage of survivors might mean that a significant number of injured persons did not reach the hospital in time. Similar observations are also presented by other authors [4, 16, 24]. In fact, the number of casualties with fatal outcome prior to admission is not known to us; however, some authors suggest that only 6-20% of victims reach hospital alive. The overall survival rate of penetrating heart trauma is estimated to be 10-92% [3, 16, 21, 24-27]. It depends not only on the severity of the injury but on other circumstances as well (population area, median time taken from injury to emergency department admission, associated injuries, etc.).

Rhee et al. have described the following paradox: the more heavily injured persons reach the hospital alive, the higher is overall mortality [26]. Swedish authors present data from one of the busiest trauma centers in Sweden with the survival rate of 85.7%, and out of seven penetrating heart injuries only one was fatal [25]. On the other hand, if patients’ general condition remains stable for a couple of days, it is likely that the survival rate after surgery will reach nearly 100% [28].

Such a high survival rate in our country might be influenced by several reasons. Firstly, the call for an ambulance was delayed at the injury scene (the median time from injury to emergency call was 28 minutes!). Secondly, the median time taken for transferring the patient to the Emergency Department was 44 minutes (IQR: 33.8-59.3). The vast majority of patients who reached the hospital alive were transported by the city ambulance. This might mean that only a certain number of patients from the Vilnius region, predominantly in a stable condition, reached the hospital alive. According to Danish authors, the median time from injury to hospital admission was from 1 to 45 minutes [1]. A recent study in France by Pons et al. showed that the majority of injured persons reach the hospital during the first six hours and only a few of them during 24 hours [17]. In the data by Seamon from the USA, median time from injury to the emergency department was 19 minutes [27]. However, this time was limited to the area within a radius of 3.2 km. But according to the data from South Africa and India this time is usually about 2 hours [16].

In assessing the work of the ambulance service in our series, it is reasonable to point out that only one patient out of 186 was transported to the hospital already intubated and only in four cases was a central venous catheter inserted. The rest of the patients before entering the Emergency Department were treated routinely: infusion to peripheral vein, oxygen therapy, analgesics. In addition, some patients arrived at a hospital which had no surgical service and no staff needed in order to treat suspected cardiac injuries. In these cases much time was spent on finding the right hospital. Our data show that there is still much work to be done to improve the work of ambulance services.

In agreement with others, our results have shown that victims of penetrating cardiac injuries are predominantly young males [1, 3, 14, 16, 20, 24, 25, 30].

One hundred and seventy-six patients were victims of stab wounds (94.6%) and only ten persons (5.4%) sustained gunshot injuries. Patients with stab wounds predominate among persons admitted to hospitals in Europe, South Africa and Brazil [1, 3, 14-17, 24, 25, 29]. However, studies from the USA have shown more gunshot injuries [4, 27]. According to Campbell et al. [16] gunshot injuries of the heart were found in 33.6% of cases (407/1198). In this study the mortality rate before reaching the hospital for gunshot injuries was significantly higher than for stab wounds. Among persons who reached the hospital alive, gunshot injuries were found only in 5.7% of cases (4/70).

In our study most of the patients (72.6%) sustained a single wound. It was not a significant predictor of survival (OR 1.067, p = 0.9). According to the American researchers, there was more than one wound in the chest in about 60% of victims. These circumstances paradoxically led to better survival [27]. This paradox might be explained by a suggestion that there is a slim chance to survive at the injury scene in multiple injury cases, especially when first aid is delayed.

Wounding to the precordium (81.2% of patients) had no influence on survival (OR 1.406, p = 0.536). This is also emphasized by Seamon et al. [27]. According to Degiannis et al. data, 46% of patients with a single stab wound had it in the precordium. The mortality rate among patients with precordium stab wounds was 4%, whereas among persons with non-precordium stab wounds it was 25%, with a statistically significant difference (p = 0.03) [14]. Navsaria et al. [15] reported that a stab wound in the precordium resulted in heart injury in 60% of cases (in our study – 71.5%); on the other hand, 85% of the heart-injured persons had a wound in the precordium area (in our data – 81.2%).

Assessing hemodynamic data in patients at the Emergency Department, we found hypotension or immeasurable blood pressure in 84 cases (45.2%). Evaluating the possibility to survive, it was found that higher systolic blood pressure significantly predicts survival (OR 1.02, 95% CI: 1.009-1.031; p < 0.001). In the group of survivors, median systolic blood pressure on admission was 94 mm Hg, whereas in non-survivors it was estimated to be 70 mm Hg, p < 0.001. The same results were presented by Rodrigues et al. [3].

In our study pulse rate on admission had no prognostic value. In the American series it was found that normal pulse rate or tachycardia was estimated to be a true precursor of survival [27].

Our data showed that 145 (78%) patients had all five SOL on admission. It was an independent predictor of survival (OR 6.829, 95% CI: 2.557-18.242; p < 0.001). It was found in 82.4% of survivors and 42.9% of non-survivors, p < 0.001. Some authors also conclude that the more SOL found on admission the greater opportunity to survive [20, 22, 27].

Emergency surgical repair remains the single procedure which enables victims’ lives to be saved [4, 24]. In our series 148 patients (79.6%) were operated on immediately after admission for suspected cardiac injury. For the rest the operation was delayed due to numerous reasons. In some clinics and hospitals patients with suspected heart injury were being operated on at the emergency department operating room [4, 14, 20, 22, 27]. All the patients in our series were operated on in the operating theatre of the Department of General Thoracic Surgery. It takes only a few minutes to reach it from the Emergency Department (some operations were started within 10 minutes after the patient’s arrival). There is a 24-hour general thoracic surgeon on duty at our hospital, so the delay is minimized. In our opinion, there is no strong necessity to spend a huge amount of money in order to arrange an emergency department operating room, having well-equipped operating rooms nearby. According to Asensio et al., the survival rate is higher in patients operated on in a routine way in comparison with emergency department surgery [22]. This might be explained by the fact that most patients operated on at the emergency department are in a critical condition, and emergency room thoracotomy is often considered a controversial “last chance” method of resuscitation.

What incision should be chosen when operating on a patient with suspected cardiac injury? The data in the literature are controversial. Some authors prefer longitudinal sternotomy [14, 15, 21, 24], while others advocate for thoracotomy, especially for patients in a terminal condition [3, 22, 23, 25, 27, 29]. According to Kang et al. [4], the easiest way to reach the heart is longitudinal sternotomy. However, we performed sternotomy only in a few cases. In all other cases preference was given to anterolateral thoracotomy.

The most frequent finding during the operation was hemopericardium or cardiac tamponade. However, in nine cases hydropericardium was found. According to Pons et al., two cases of hemopericardium and three cases of hydropericardium were found out of 13 patients with precordium stab wounds who arrived in a stable general condition, and thoracotomy was performed for suspected heart injury [17]. Danish authors found hemopericardium in all the reported cases [1]. Campbell et al. [16] reported that in 9.8% of cases (117/1198) hemopericardium was found without visible injury of the heart or coronary vessels [in our series it was found in 9.1% (17/186) of cases].

Cardiac tamponade led to significantly worse prognosis in our series. It was found in 29.1% of survivors and in 52.4% of non-survivors (p = 0.031). The influence of this symptom on survival rate according to the published data is ambiguous [16, 22, 25]. Some authors presume that in case of cardiac tamponade acute exsanguination does not occur. This situation provides some stability in the patient’s general condition. However, there is no information concerning the duration of the above condition available. But some authors [4, 14, 16] did find a better survival rate in patients with verified cardiac tamponade. However, when this “bright period” ends, the cardiac function deteriorates immediately [22]. Evaluating our data, it became evident that most patients with cardiac tamponade did not reach the hospital in the shortest time (in the so-called “golden minutes”). It was on rare occasions that severely injured persons with cardiac tamponade reached the hospital in due time, which explains the lowest survival rate in this group. The right ventricle was the most frequently injured cardiac chamber (40.3%). The same data are provided by other authors [4, 16, 21, 24].

Conclusions

One must suspect heart injury when there is a wound in the precordium. Delay in calling the ambulance and long transportation time worsen the prognosis for survival. Our data have shown that all SOL on admission and younger age are independent predictors of survival, whereas higher grade of heart injury and heart arrest during surgery independently predict fatal outcome.

References

1. Rashid MA, Lund JT. Trauma to the heart and thoracic aorta: Copenhagen experience. Interactive Cardiovascular and Thoracic Surgery 2003; 2: 53-57.

2. Hurt R. The history of Cardiothoracic Surgery. From early times. The Parthenon publishing group. 1996: 399-403.

3. Rodrigues AJ, Furlanetti LL, Faidiga GB, Scarpelini S, Barbosa Evora PR, de Andrade Vincente WV. Penetrating cardiac injuries: a 13-year retrospective evaluation from a Brazilian trauma center. Interact Cardiovasc Thorac Surg 2005; 4: 212-215.

4. Kang N, Hsee L, Rizoli S, Alison P. Penetrating cardiac injury: Overcoming the limits set by nature. Injury 2009; 40: 919-927.

5. Soreide K, Soreide JA. Axel H. Cappelen, MD (1858-1919): First Suture of a Myocardial Laceration from a Cardiac Stab Wound. J Trauma 2006; 60: 653-654.

6. Westaby S, Bosher C. Landmarks in Cardiac Surgery. ISIS Medical Media, Oxford 2000: 19-20.

7. Fischer G. Die Wunden des Herzens und des Herzbeutels. Arch Klin Chir 1868; 9: 571-7.

8. Beck CS. Wounds of the heart. Arch Surg 1926; 13: 205-27.

9. Absolon KB. Theodor Billroth and cardiac surgery. J Thorac Cardiovasc Surg 1983; 86: 451-452.

10. Asensio JA, Petrone P, Pereira B, Pena D, Prichayudh S, Tsunoyama T, Ruiz F, Marttos A, Capin A, De Marchena E. Penetrating cardiac injuries: a historic perspective and fascinating trip through time. J Am Coll Surg 2009; 3: 462-472.

11. Farina G. Discussion. Centralbl Chir 1896; 23: 1224. As quoted by Beck CS. Wounds of the heart. Arch Surg 1926; 13: 205-227.

12. Rehn L. Zur chirurgie des Herzens und des Herzbeutels. Arch Klin Chir 1907; 83: 723-730.

13. Dumcius A. Review of history of Lithuanian heart surgery. Medicine 2002; 38 (2 suppl.): 5-10.

14. Degiannis E, Loogna P, Doll D, Bonanno F, Bowley DM, Smith MD. Penetrating cardiac injuries: Recent experience in South Africa. World J Surg 2006; 30: 1258-1264.

15. Navsaria PH, Nicol AJ. Haemopericardium in stable patients after penetrating injury: Is subxiphoid pericardial window and drainage enough? A prospective study. Injury 2005; 36: 745-750.

16. Campbell NC, Thomson SR, Muckart DJJ, Meumann CM, Van Middelkoop I, Botha JBC. Review of 1198 cases of penetrating cardiac trauma. Br J Surg 1997; 84: 1737-1740.

17. Pons F, Lang-Lazdunski L, de Kerangal X, Chapuis O, Bonnet PM, Jancovici R. The role of videothoracoscopy in management of precordial thoracic penetrating injuries. Eur J Cardiothoracic Surg 2002; 22: 7-12.

18. Crawford R, Kasem H, Bleetmen A. Traumatic pericardial tamponade: relearning old lessons. J Accid Emerg Med 1997; 14: 252-257.

19. Krug EG, Mercy JA, Gahlberg LL. The world report on violence and health. Lancet 2002; 360: 1083-1088.

20. Cawich SO, Mitchell DIG, Williams EW, Mcfarlane ME, Martin A, Plummer JM,

Blake G, Newnham MS, Brown H. Emergency department thoracotomy in Jamaica: A case controlled study. Int J Surg 2007; 5: 311-315.

21. Navid F, Gleason TG. Great vessel and cardiac trauma: diagnostic and management strategies. Semin Thorac Cardiovasc Surg 2008; 20: 31-38.

22. Asensio JA, Soto SN, Forno W, Roldan G, Petrone P, Salim A, Rowe V, Demetriades D. Penetrating cardiac injuries: a complex challenge. Injury 2001; 32: 533-543.

23. Salapa M, Hutan M, Zelenak J. Penetrating cardiac injury. Bratisl Lek Listy 2001; 102: 467-469.

24. Harris DG, Papagiannopoulos KA, Pretorius J, Van Rooyen T, Rossouw GJ. Current evaluation of cardiac stab wounds. Ann Thorac Surg 1999; 68: 2119-2122.

25. Rashid MA, Wikstrom T, Ortenwall P. Cardiac injuries: a ten-year experience. Eur J Surg 2000; 166: 18-21.

26. Rhee PM, Foy H, Kaufmann C, Areola C, Boyle E, Maier RV, Jurkovich G. Penetrating cardiac injuries: a population-based study. J Trauma 1998; 45: 366-370.

27. Seamon MJ, Fisher CA, Gaughan JP, Kulp H, Dempsey DT, Goldberg AJ. Emergency department thoracotomy: Survival of the least expected. World J Surg 2008; 32: 604-612.

28. Harris DG, Janson JT, Van Wyk J, Pretorius J, Rossouw GJ. Delayed pericardial effusion following stab wounds to the chest. Eur J Cardiothorac Surg 2003; 23: 473-476.

29. Madiba TE, Thomson SR, Mdlalose N. Penetrating chest injuries in the firearm era. Injury 2001; 32: 13-16.

30. Ibirogba S, Nicol AJ, Navsaria PH. Screening helical computed tomographic scanning in haemodynamic stable patients with transmediastinal gunshot wounds. Injury 2007; 38: 48-52.